Timothy Pratt, Charles W. Bostian, Jeremy E. Allnutt - Satellite Communications-Wiley Indian (2003)

This Wiley Student Edition is part of a continuing
program of paperbound textbooks especially designed

for students in developing countries at a reduced price.
Second Edition
THIS BOOK IS FOR SALE ONLY IN THE COUNTRY TO
Pratt
WHICH IT IS FIRST CONSIGNED BY WILEY INDIA
PVT. LTD. AND SHOULD NOT BE
Bostian
RE-EXPORTED. Allnutt
SATELLITE
FOR SALE ONLY IN :
SATELLITE COMMUNICATIONS
INDIA, BANGLADESH, NEPAL, PAKISTAN, SRI
LANKA AND BHUTAN
Wiley Acing the GATE Series

A Complete Solution for all GATE Preparation Needs
AS PER LATEST GATE SYLLABUS
COMMUNICATIONS
Timothy Pratt
Charles Bostian
Jeremy Allnutt
www.wileyindia.com
SPECIAL INDIA EDITION

The content of this book may have been
modiﬁed to suit Indian context.
ISBN978-93-88991-29-2
eISBN: 938899129-X
SECOND
EDITION
Wiley India Pvt. Ltd.
Customer Care +91 120 6291100
csupport@wiley.com
www.wileyindia.com
9 789388 991292
www.wiley.com
fm.qxd 29/08/02 19:04 Page i
Satellite
Communications
fm.qxd 29/08/02 19:04 Page ii
fm.qxd 29/08/02 19:04 Page iii
Satellite
Communications
Second Edition
Timothy Pratt
Charles W. Bostian
Department of Electrical & Computer Engineering
Virginia Polytechnic Institute and State University
Jeremy E. Allnutt
Department of Electrical & Computer Engineering
George Mason University
John Wiley & Sons

Satellite Communications
Second Edition
Authorized reprint by Wiley India Pvt. Ltd., 4436/7, Ansari Road, Daryaganj,
New Delhi – 110002.
Copyright © 2003 by John Wiley & Sons, Inc. All rights reserved.
All rights reserved. AUTHORIZED REPRINT OF THE EDITION PUBLISHED
BY JOHN WILEY & SONS, INC. No part of this book may be reproduced in any
form without the written permission of the publisher.
Limits of Liability/ Disclaimer of Warranty: The publisher and the author make
no representations or warranties with respect to the accuracy or completeness of
the contents of this work and specifically disclaim all warranties, including without
limitation warranties of fitness for a particular purpose. No warranty may be
created or extended by sales or promotional materials. The advice and strategies
contained herein may not be suitable for every situation. This work is sold with the
understanding that the publisher is not engaged in rendering legal, accounting,
or other professional services. If professional assistance is required, the services of
a competent professional person should be sought. Neither the publisher nor the
author shall be liable for damages arising herefrom. The fact that an organization
or website is referred to in this work as a citation and/or a potential source of
further information does not mean that the author or the publisher endorses the
information the organization or website may provide or recommendations it may
make. Further, readers should be aware that Internet websites listed in this work
may have changed or disappeared between when this work was written and when it
is read.
Wiley also publishes its books in a variety of electronic formats. Some content that
appears in print may not be available in electronic books. For more information
about Wiley products, visit our website at: www.wiley.com.
Authorized India Edition
ISBN: 978-81-265-0833-4
ISBN: 978-93-88991-29-2 (ebk)

fm.qxd 29/08/02 19:04 Page v
Dedication
To our wives: Maggie, Frieda, and Norma
fm.qxd 29/08/02 19:04 Page vi
fm.qxd 29/08/02 19:04 Page vii
ABOUT THE AUTHORS
Timothy Pratt is a professor in the Bradley Department of Electrical and Computer En-
gineering at Virginia Tech, where he has been a faculty member since 1981. He received
his B.Sc. and Ph.D. degrees in electrical engineering from the University of Birmingham,
United Kingdom, and has taught communications subjects in the United Kingdom and the
United States. His research interests are in satellite communications, position location,
and avionics. Dr. Pratt is a senior member of the IEEE and a member of the IEE (London).
Charles W. Bostian is Clayton Ayre Professor of Electrical and Computer Engineering at

Virginia Tech where he has been a faculty member since 1969. His primary research in-
terests are in the areas of wireless communications and radiowave propagation. He is co-
author of the Wiley Text, Solid State Radio Engineering, published in 1980. Professor
Bostian received his degrees in electrical engineering from North Carolina State Univer-
sity and is a fellow of the IEEE.
Jeremy E. Allnutt is a professor in the Electrical and Computer Engineering Department

of George Mason University and Director of the MS in Telecommunications Program. His
primary interest is radiowave propagation effects on satellite links, which he pursued at
research establishments in England and Canada, before working at INTELSAT in the
United States from 1979 to 1994. Prior to joining George Mason University in 2000, he
was a professor at the University of York, England, and at Virginia Tech. Dr. Allnutt ob-
tained his B.Sc. and Ph.D. in Electrical Engineering from Salford University, England,
and is a Fellow of IEE and Senior member IEEE.
vii
fm.qxd 29/08/02 19:04 Page viii
fm.qxd 29/08/02 19:04 Page ix
PREFACE
There have been many new developments in satellite communications technology since
the first edition of this text was published in 1986. However, the underlying principles of
the transmission of radio signals via satellites remain the same. Thus the basic material
in the textbook relating to satellite orbits, look angles, transponders on communications
satellites, link budget calculations, SN and BER for analog and digital links, multiple
access techniques, error control, and the propagation of radio waves through the earth’s
atmosphere remains as important now as in 1986. What has changed is that new applica-
tions have been developed for satellite communication systems, and new satellites and
terminals have been built to implement the new systems.
The second edition of Satellite Communications makes no attempt to describe all
of the satellite systems now in operation. That would require an encyclopedia. The text
concentrates on the principles of satellite communication systems with the aim of pro-
viding the reader with a sound understanding of how a satellite communication system
successfully transfers information from one earth station to another. The first edition of
this text was written by Charles Bostian and Tim Pratt to support the courses we taught
on satellite communication. The book found wide popularity, both as a text for students
in senior year or beginning graduate courses at universities, and as a basic reference for
practicing engineers. In the second edition, we are honored to be joined by our friend
and colleague Jeremy Allnutt, with whom we have worked on satellite systems for over
25 years. He contributed the chapters on orbital mechanics, propagation, nongeostationary
satellite systems and VSAT networks.
Much material that was included in the first edition has been omitted in the second
to make way for chapters covering VSAT systems, LEO and NGSO systems, direct broad-
cast television, and satellite navigation. The advent of personal communications via low
earth orbit (LEO) satellites was not anticipated when the text was written in 1984, nor the
development of direct broadcasting from satellites using digital transmission. The growth
of very small aperture terminal (VSAT) systems has also occurred since 1986, and has
led to application of many of the techniques discussed in the first edition. The Global Po-
sitioning System (GPS) has become the dominant radio navigation aid, using a constel-
lation of 24 satellites to provide accurate position location everywhere on earth. Perhaps
the greatest change in technology over the past fifteen years has been the transition from
analog to digital transmission techniques. The transition is almost complete in the United
States, with only the distribution of video signals to cable TV head ends remaining as a
last bastion of analog transmission. The section in Chapter 5 of the first edition that cov-
ered FDM/FM/FDMA systems has been retained as an appendix because such systems
continue in operation in some parts of the world.
The emphasis throughout the text is on digital transmission techniques; Chapter 5
reviews the basic theory of digital radio transmission, which is fundamental to all digital
satellite systems. In parallel with the transition to digital satellite transmission, great
changes have occurred in terrestrial communication systems. Optical fibers were just start-
ing to come into use in 1986, and the Internet was still in its infancy. Cellular telephones
were barely in use. Many of the developments in terrestrial communication systems have
carried over to satellite systems, and much of the technology that was new in 1986 has
now matured and has been well described elsewhere.
ix
fm.qxd 29/08/02 19:04 Page x
x PREFACE
Within the United States, satellites are not used for telephone (voice) links. The de-
velopment of terrestrial optical fiber links has made satellite transmission of telephone
traffic uneconomic, and the delay associated with GEO satellite link is a nuisance.
Domestic satellites serving the United States now carry video signals for distribution to
cable TV companies or direct to homes and serve networks of VSAT stations linked to
central hubs in major cities. The development of direct to home satellite broadcast tele-
vision (DBS-TV) has had a major impact on the marketplace. In the United States, digi-
tal DBS-TV transmissions are now received in 15 million homes (2001 figure), and in
Europe a similar number of homes receive satellite television programming. Video distri-
bution, to cable companies and direct to home, accounts for more than half of all the
worldwide earnings from satellite communication systems.
The authors would like to thank their colleagues and students who, over the years,
have made many valuable suggestions to improve this text. Their advice has been heeded,
and the second edition of Satellite Communication is the better for it. Many more worked
examples have been added to the second edition to illustrate how calculations are carried
out for each topic.
fm.qxd 29/08/02 19:04 Page xi
CONTENTS
Preface ix
1. Introduction 1
1.1 Background 1
1.2 A Brief History of Satellite Communications 3
1.3 Satellite Communications in 2000 6
1.4 Overview of Satellite Communications 15
1.5 Summary 16
References 16
2. Orbital Mechanics and Launchers 17

2.1 Orbital Mechanics 17
Developing the Equations of the Orbit 17
Kepler’s Three Laws of Planetary Motion 22
Describing the Orbit of a Satellite 23
Locating the Satellite in the Orbit 25
Locating the Satellite with Respect to the Earth 27
Orbital Elements 29
Example 2.1.1 Geostationary Satellite Orbit Radius 29
Example 2.1.2 Low Earth Orbit 29
Example 2.1.3 Elliptical Orbit 30
2.2 Look Angle Determination 30
The Subsatellite Point 31
Elevation Angle Calculation 32
Azimuth Angle Calculation 34
Specialization to Geostationary Satellites 35
Visibility Test 36
Example 2.2.1 Geostationary Satellite Look Angles 36
2.3 Orbital Perturbations 38
Longitudinal Changes: Effects of the Earth’s Oblateness 39
Inclination Changes: Effects of the Sun and the Moon 40
Example 2.3.1 Drift with a Geostationary Satellite 42
2.4 Orbit Determination 42
2.5 Launches and Launch Vehicles 43
Expendable Launch Vehicles (ELVs) 44
Placing Satellites into Geostationary Orbit 48
2.6 Orbital Effects in Communications Systems Performance 49
Doppler Shift 49
Example 2.6.1 Doppler Shift for a LEO Satellite 50
Range Variations 51
Solar Eclipse 51
Sun Transit Outage 53
xi
fm.qxd 29/08/02 19:04 Page xii
xii CONTENTS
2.7 Summary 54
References 54
Problems 55
3. Satellites 57
3.1 Satellite Subsystems 57
Attitude and Orbit Control System (AOCS) 57
Telemetry, Tracking, Command and Monitoring (TTC&M) 59
Power System 59
Communications Subsystems 59
Satellite Antennas 59
3.2 Attitude and Orbit Control System (AOCS) 60
Attitude Control System 60
Orbit Control System 66
3.3 Telemetry, Tracking, Command, and Monitoring 68
Telemetry and Monitoring System 68
Tracking 68
Command 70
3.4 Power Systems 71
3.5 Communications Subsystems 72
Description of the Communications System 72
Transponders 75
3.6 Satellite Antennas 80
Basic Antenna Types and Relationships 80
Example 3.6.1 Global Beam Antenna 82
Example 3.6.2 Regional Coverage Antenna 83
Satellite Antennas in Practice 83
3.7 Equipment Reliability and Space Qualification 87
Space Qualification 87
Reliability 88
Redundancy 90
3.8 Summary 92
References 93
Problems 93
4. Satellite Link Design 96

4.1 Introduction 96
4.2 Basic Transmission Theory 100
Example 4.2.1 104
Example 4.2.2 104
4.3 System Noise Temperature and G/T Ratio 105
Noise Temperature 105
Calculation of System Noise Temperature 107
Example 4.3.1 110
Example 4.3.2 110
Noise Figure and Noise Temperature 111
Example 4.3.3 112
G/T Ratio for Earth Stations 112
Example 4.3.4 112
fm.qxd 29/08/02 19:04 Page xiii
CONTENTS xiii
4.4 Design of Downlinks 112

Link Budgets 113
Link Budget Example: C-Band Downlink for Earth Coverage Beam 115
4.5 Satellite Systems Using Small Earth Stations 117
Direct Broadcast TV 118
Example 4.5.1 123
4.6 Uplink Design 124
Example 4.6.1 127
4.7 Design for Specified CN: Combining CN and CI Values in Satellite Links 127
Example 4.7.1 129
Overall (CN)0 with Uplink and Downlink Attenuation 129
Uplink and Downlink Attenuation in Rain 130
Uplink Attenuation and (CN)up 130
Downlink Attenuation and (CN)dn 131
System Design for Specific Performance 131
Satellite Communication Link Design Procedure 131
4.8 System Design Examples 132
System Design Example 4.8.1 133
Ku Band Uplink Design 133
Ku Band Downlink Design 134
Rain Effects at Ku Band 135
Summary of Ku Band Link Performance 137
System Design Example 4.8.2 Personal Communication System Using
Low Earth Orbit Satellites 137
Inbound Link: Mobile Terminal to Gateway Station 141
Mobile Terminal to Satellite Link 142
Satellite to Gateway Station Link 143
Outbound Link 144
Downlink CN Budget 145
Optimizing System Performance 146
Link Margins with FEC 147
Rain Attenuation at Ku Band 147
Path Blockage at L-Band 149
Summary of L-band Mobile PCS System Performance 149
4.9 Summary 150
References 150
Problems 151
5. Modulation and Multiplexing Techniques for Satellite Links 156

5.1 Frequency Modulation 157
Waveform Equation for FM 158
Bandwidth of FM Signals: Carson’s Rule 159
Baseband SN Ratio for FM Signals 159
Pre-emphasis and de-emphasis 161
Pre-emphasis 162
5.2 Analog FM Transmission by Satellite 164
Television Signals 165
SN Ratios for FM Video Transmission 167
Example 5.2.1 168
fm.qxd 29/08/02 19:04 Page xiv
xiv CONTENTS
FM Threshold 168
SCPC FM Links 169
Example 5.2.2 170
Data Transmission Using Analog FM Channels 170
Example 5.2.3 171
5.3 Digital Transmission 172
Baseband Digital Signals 172
Baseband Transmission of Digital Data 172
Band-pass Transmission of Digital Data 179
Example 5.3.1 181
Example 5.3.2 181
Transmission of QPSK Signals through a Bandlimited Channel 182
Example 5.3.3 185
Example 5.3.3 185
5.4 Digital Modulation and Demodulation 187
Terminology 187
Modulation and Coding 187
Bit and Symbol Error Rates 188
Binary Phase Shift Keying (BPSK) 189
Probability of a Symbol Error 191
BPSK Bit Error Rate 194
QPSK Bit Error Rate 194
Example 5.4.1 195
Example 5.4.2 197
Generation of Quadrature Phase Shift Keying (QPSK) Signals 198
QPSK Variants 199
5.5 Digital Transmission of Analog Signals 201
Sampling and Quantizing 201
Nonuniform Quantization: Compression and Expansion 204
Signal-to-Noise Ratio in Digital Voice Systems 206
Digital Television 208
5.6 Time Division Multiplexing 209
TDM Terminology: The U.S. T1 24-Channel System 209
Other TDM Systems 211
Channel Synchronization in TDM 212
5.7 Summary 212
References 213
Problems 214
6. Multiple Access 221
6.1 Introduction 221
6.2 Frequency Division Multiple Access (FDMA) 223
Intermodulation 226
Intermodulation Example 228
Calculation of CN with Intermodulation 230
Example 6.2.1 Power Sharing in FDMA 231
Example 6.2.2 Channel Capacity with Demand Access FDMA 232
6.3 Time Division Multiple Access (TDMA) 233
Bits, Symbols, and Channels 234
TDMA Frame Structure 235
fm.qxd 29/08/02 19:04 Page xv
CONTENTS xv
Example 6.3.1 TDMA in a Fixed Station Network 237

Reference Burst and Preamble 238
Unique Word 239
Guard Times 241
Synchronization in TDMA Networks 242
Transmitter Power in TDMA Networks 243
Example 6.3.2 TDMA in a VSAT Network 244
Example 6.3.2 TDMA in a Fixed Earth Station Network 244
Satellite Switched TDMA 246
6.4 Onboard Processing 246
Baseband Processing Transponders 247
Satellite Switched TDMA with Onboard Processing 248
6.5 Demand Access Multiple Access (DAMA) 249
Example 6.5.1 FDMA-SCPC-DA 252
6.6 Random Access 254
6.7 Packet Radio Systems and Protocols 254
6.8 Code Division Multiple Access (CDMA) 257
Spread Spectrum Transmission and Reception 258
DS-SS CDMA Capacity 262
Example 6.8.1 CDMA in a Fixed Earth Station Network 263
Example 6.8.2 CDMA in an LEO Satellite Network 263
Example 6.8.3 GPS 264
6.9 Summary 266
References 267
Problems 267
7. Error Control for Digital Satellite Links 273
7.1 Error Detection and Correction 273
7.2 Channel Capacity 275
7.3 Error Control Coding 277
Example 7.3.1 278
Linear and Cyclic Block Codes 279
Golay Codes 280
7.4 Performance of Block Error Correction Codes 281
7.5 Convolutional Codes 282
7.6 Implementation of Error Detection on Satellite Links 284
Example 7.6.1 287
7.7 Concatenated Coding and Interleaving 288
7.8 Turbo Codes 290
7.9 Summary 292
References 292
Problems 293
8. Propagation Effects and their Impact on Satellite–Earth Links 295
8.2 Quantifying Attenuation and Depolarization 298
Example 8.2.1 301
8.3 Propagation Effects that Are Not Associated with Hydrometeors 306
Atmospheric Absorption 307
Cloud Attenuation 308
fm.qxd 29/08/02 19:04 Page xvi
xvi CONTENTS
Tropospheric Scintillation and Low Angle Fading 308

Faraday Rotation in the Atmosphere 310
Ionospheric Scintillations 312
8.4 Rain and Ice Effects 312
Characterizing Rain 312
Rain Climate Maps 314
Rainfall Rate Exceedance Contour Maps 315
Raindrop Distributions 315
8.5 Prediction of Rain Attenuation 317
Example 8.5.1 319
Example 8.5.2 323
Calculation of Long-Term Statistics for NGSO Systems 324
Scaling Attenuation with Elevation Angle and Frequency 325
Cosecant Law 325
Example 8.5.3 325
Squared Frequency Scaling Law 326
Example 8.5.4 326
ITU-R Long-Term Frequency Scaling of Rain Attenuation 326
8.6 Prediction of XPD 326
Canting Angle 328
Tilt Angle 328
Example 8.6.1 330
Example 8.6.2 331
Ice Crystal Depolarization 332
Rain Effects on Antenna Noise 332
Example 8.6.3 333
8.7 Propagation Impairment Countermeasures 333
Attenuation 333
Power Control 334
Signal Processing 335
Diversity 335
Depolarization 337
8.8 Summary 338
References 339
Problems 340
9. VSAT SYSTEMS 343

9.2 Overview of VSAT Systems 345
9.3 Network Architectures 347
One-Way Implementation 347
Split-Two-Way (Split IP) Implementation 347
Two-Way Implementation 348
9.4 Access Control Protocols 349
Delay Considerations 351
9.5 Basic Techniques 354
Multiple Access Selection 355
Signal Formats 362
Modulation, Coding, and Interference Issues 362
fm.qxd 29/08/02 19:04 Page xvii
CONTENTS xvii
9.6 VSAT Earth Station Engineering 366

Antennas 366
Transmitters and Receivers 367
9.7 Calculation of Link Margins for a VSAT Star Network 370
9.8 System Design Procedure: Example 9.1 372
Description of System 373
System Parameters 373
Preliminary Calculations 374
Link CN Ratios 375
Inbound Links 376
Inbound Links with 270 Channels 378
Outbound Links 379
System Analysis 380
9.9 Some New Developments 383
9.10 Summary 384
References 385
Problems 385
10. Low Earth Orbit and Non-Geostationary Satellite Systems 388

10.2 Orbit Considerations 391
Equatorial Orbits 391
Inclined Orbits 392
Elliptical Orbits 394
Molniya Orbit 396
Radiation Effects 398
Sun Synchronous Orbit 403
10.3 Coverage and Frequency Considerations 406
General Aspects 406
Frequency band 406
Elevation Angle Considerations 408
Number of Beams per Coverage 411
Off-Axis Scanning 412
Determination of Optimum Orbital Altitude 418
Radiation Safety and Satellite Telephones 420
Projected NGSO System Customer Service Base 420
10.4 Delay and Throughput Considerations 421
10.5 System Considerations 423
Incremental Growth 424
Interim Operations 424
Replenishment Options 424
End-to-End System Implementation 425
10.6 Operational NGSO Constellation Designs 425
Ellipso 425
Globalstar 426
New ICO 428
Iridium 428
Orbcomm 429
Skybridge 429
fm.qxd 29/08/02 19:04 Page xviii
xviii CONTENTS
Teledesic 430
Example 10.6.1 System Design 432
10.7 Summary 434
References 434
Problems 435
11. Direct Broadcast Satellite Television and Radio 439

11.1 C-Band and Ku-Band Home Satellite TV 440
11.2 Digital DBS TV 441
11.3 DBS-TV System Design 447
11.4 DBS-TV Link Budget 449
11.5 Error Control in Digital DBS-TV 450
11.6 Master Control Station and Uplink 452
11.7 Installation of DBS-TV Antennas 454
11.8 Satellite Radio Broadcasting 455
11.9 Summary 456
References 457
12. Satellite Navigation and the Global Positioning System 458

12.2 Radio and Satellite Navigation 461
12.3 GPS Position Location Principles 463
Position Location in GPS 464
GPS Time 466
12.4 GPS Receivers and Codes 467
The CA Code 468
12.5 Satellite Signal Acquisition 470
12.6 GPS Navigation Message 472
12.7 GPS Signal Levels 473
12.8 Timing Accuracy 475
12.9 GPS Receiver Operation 476
12.10 GPS CA Code Accuracy 480
Dilution of Precision: HDOP, VDOP, and GDOP 481
12.11 Differential GPS 482
12.12 Summary 484
References 485
Problems 485
Appendix A Decibels in Communications Engineering 487
Appendix B FDM/FM/FDMA Analog Telephone Transmission 491

Baseband Voice Signal 491
Voice Signal Multiplexing 493
Frequency Modulation with Multiplexed Telephone Signals 496
Bandwidth Calculation for FDM/FM Telephone Signals 497
Telephone Performance Specifications 498
Practical Examples 499
Example B.1 499
fm.qxd 29/08/02 19:04 Page xix
CONTENTS xix
Example B.2 503

References 503
Appendix C Complementary Error Function erfc(x) and Q Function Q(z) 504

Equivalence Formulas and Tables of Values 504
References 504
Appendix D The Simple Attenuation Model 507

Example D.1 511
References 512
Glossary 513
Index 522
c01.qxd 13/07/02 14:02 Page 1
CHAPTER 1
INTRODUCTION
1.1 BACKGROUND
Two developments in the twentieth century changed the way people lived: the automobile
and telecommunications. Prior to the widespread availability of personal automobiles, in-
dividuals had to travel on foot, by bicycle, or on horseback. Trains provided faster travel
between cities, but the lives of most people were centered on their hometowns and im-
mediate surroundings. A journey of 100 miles was a major expedition for most people,
and the easy mobility that we all take for granted in the twenty-first century was unknown.
Before the telegraph and telephone came into widespread use, all communication was
face-to-face, or in writing. If you wanted to talk to someone, you had to travel to meet
with that person, and travel was slow and arduous. If you wanted to send information, it
had to be written down and the papers hand carried to their destination.
Telecommunication systems have now made it possible to communicate with virtu-
ally anyone at any time. Early telegraph and telephone systems used copper wire to carry
signals over the earth’s surface and across oceans, and high frequency (HF) radio, also
commonly called short wave radio, made possible intercontinental telephone links. Arti-
ficial earth satellites have been used in communications systems for more than 35 years
and have become an essential part of the world’s telecommunications infrastructure. Satel-
lites allow people to talk by telephone and exchange electronic mail from anywhere in
the world and to receive hundreds of TV channels in their homes.
The origins of satellite communications can be traced to an article written by
Arthur C. Clarke in the British radio magazine Wireless World in 19451. At the time,
Clarke was serving in the British Royal Air Force, and was interested in long-distance
radio communication. He later became famous as the author of 2001: A Space Odyssey,
and other science fiction books2. In 1945, HF radio was the only available method for
radio communication over transcontinental distances, and it was not at all reliable.
Sunspots and ionospheric disturbances could disrupt HF radio links for days at a time.
Telegraph cables had been laid across the oceans as early as the mid-1800s, but cables
capable of carrying voice signals across the Atlantic did not begin service until 1953.
Clarke suggested that a radio relay satellite in an equatorial orbit with a period of 24 h
would remain stationary with respect to the earth’s surface and make possible long-
distance radio links. At the time Clarke wrote, there were no satellites in orbit nor rockets
powerful enough to launch them. But his ideas for what we now know as a geostationary
satellite system were not science fiction, as the launch of the Russian satellite Sputnik
in 1957 was to prove. In 1965 the first geostationary satellite, Early Bird, began
to provide telephone service across the Atlantic Ocean, fulfilling Clarke’s vision of
20 years earlier.
Satellite communication systems were originally developed to provide long-distance
telephone service. In the late 1960s, launch vehicles had been developed that could place
a 500 kg satellite in geostationary earth orbit (GEO), with a capacity of 5000 telephone
circuits, marking the start of an era of expansion for telecommunication satellites.
Geostationary satellites were soon carrying transoceanic and transcontinental telephone calls.
1
c01.qxd 13/07/02 14:02 Page 2
2 CHAPTER 1 INTRODUCTION
For the first time, live television links could be established across the Atlantic and Pacific
oceans to carry news and sporting events.
The geostationary orbit is preferred for all high capacity communication satellite
systems because a satellite in GEO appears to be stationary over a fixed point on the
ground. It can establish links to one-third of the earth’s surface using fixed antennas at
the earth stations. This is particularly valuable for broadcasting, as a single satellite can
serve an entire continent. Direct broadcast satellite television (DBS-TV) and the distri-
bution of video signals for cable television networks are the largest single revenue source
for geostationary satellites, accounting for $17 B in revenues in 1998. By year 2001, nearly
200 GEO communication satellites were in orbit, serving every part of the globe. Although
television accounts for much of the traffic carried by these satellites, international and
regional telephony, data transmission, and Internet access are also important. In the pop-
ulated parts of the world, the geostationary orbit is filled with satellites every 2° or 3°
operating in almost every available frequency band.
GEO satellites have grown steadily in weight, size, lifetime, and cost over the years.
Some of the largest satellites launched to date are the KH and Lacrosse surveillance satel-
lites of the U.S. National Reconnaissance Office weighing an estimated 13,600 kg (30,000
lb)3. By 2000, commercial telecommunications satellites weighing 6000 kg with lifetimes
of 15 years were being launched into geostationary orbit at a typical cost around $125 M
for the satellite and launch. The revenue earning capacity of these satellites must exceed
$20 M per year for the venture to be profitable, and they must compete with optical fibers
in carrying voice, data, and video signals. A single optical fiber can carry 4.5 Gbps, a ca-
pacity similar to that of the largest GEO satellites, and optical fibers are never laid singly
but always in bundles. But GEO satellites can compete effectively on flexibility of delivery
point. Any place within the satellite coverage can be served by simply installing an earth
terminal. To do the same with a fiber-optic link requires fiber to be laid. Fiber-optic trans-
mission systems compete effectively with satellites where there is a requirement for high
capacity or, equivalently, when the user density exceeds the required economic threshold.
GEO satellites have been supplemented by low and medium earth orbit satellites for
special applications. Low earth orbit (LEO) satellites can provide satellite telephone and
data services over continents or over the entire world, and by 2000 three systems were in
orbit or nearing completion, with a total of 138 LEO satellites. LEO satellites are also
used for earth imaging and surveillance. Although not strictly a satellite communications
system, the Global Positioning System (GPS), which uses 24 medium earth orbit (MEO)
SIDEBAR
The high capacity of both optical fibers and satellites, To put the reduction in the cost of an international tele-
and the steady move of telecommunications traffic phone call in perspective, we must remember that in-
from analog signals to digital has lowered the cost of comes have risen significantly over this time period.
long-distance telephone calls and increased enor- In the 1950s, a typical blue-collar wage was $1.50 per
mously the number of circuits available. In 1960, prior hour, so a blue-collar worker had to work for 40 min
to the advent of satellite communications, the United to pay for a call to Europe, ignoring any tax deduc-
States had 550 overseas telephone circuits. Calls to tions. In 2000, the average worker in the United States
Europe cost more than $1 per minute at 1960 prices, earned $11.00 per hour, and had to work less than 1
and had to be placed through an operator, with delays min to pay for the international call. The United States
of many hours being common. In 2000, virtually all now has hundreds of thousands of overseas telephone
international calls could be dialed by the end user, and circuits, and video links daily carry live news reports
rates to Europe had dropped to below $0.10 per minute. from all over the globe.
c01.qxd 13/07/02 14:02 Page 3
1.2 A BRIEF HISTORY OF SATELLITE COMMUNICATIONS 3
satellites, has revolutionized navigation. GPS receivers have become a consumer product.
Eventually every car and cellular telephone will have a GPS receiver built into it so that
drivers will not get lost and emergency calls from cellular phones will automatically carry
information about the phone’s location.
1.2 A BRIEF HISTORY OF SATELLITE

COMMUNICATIONS
Satellite communications began in October 1957 with the launch by the USSR of a small
satellite called Sputnik I. This was the first artificial earth satellite, and it sparked the space
race between the United States and the USSR. Sputnik I carried only a beacon transmit-
ter and did not have communications capability, but demonstrated that satellites could be
placed in orbit by powerful rockets. The first satellite successfully launched by the United
States was Explorer I, lofted from Cape Canaveral on January 31, 1958 on a Juno I rocket.
The first voice heard from space was that of President Eisenhower, who recorded a brief
Christmas message that was transmitted back to earth from the Project Score satellite in
December 1958. The Score satellite was essentially the core of the Atlas ICBM (inter-
continental ballistic missile) booster with a small payload in the nose. A tape recorder on
Score had a storage capacity that allowed a 4 min message received from an earth station
to be retransmitted. The batteries on Score failed after 35 days in orbit.
After some early attempts to use large balloons (Echo I and II) as passive reflec-
tors for communication signals, and some small experimental satellite launches, the first
true communications satellites, Telstar I and II, were launched in July 1962 and May 1963.
The Telstar satellites were built by Bell Telephone Laboratories and used C-band transpon-
ders adapted from terrestrial microwave link equipment. The uplink was at 6389 MHz and
the downlink was at 4169 MHz, with 50-MHz bandwidth. The satellites carried solar cells
and batteries that allowed continuous use of the single transponder, and demonstrations
of live television links and multiplexed telephone circuits were made across the Atlantic
Ocean, emphatically demonstrating the feasibility of satellite communications.
The Telstar satellites were launched into what is now called a medium earth orbit,
with periods of 158 and 225 min. This allowed transatlantic links to operate for about 20 min
while the satellite was mutually visible. The orbits chosen for the Telstar satellites took
them through several bands of high energy radiation which caused early failure of the elec-
tronics on board. However, the value of communication satellites had been demonstrated
and work was begun to develop launch vehicles that could deliver a payload to geosta-
tionary orbit, and to develop satellites that could provide useful communication capacity.
On July 24, 1961, U.S. President John F. Kennedy defined the general guidelines
of U.S. policy in regard to satellite communications and made the first unambiguous ref-
erences to a single worldwide system. On December 20, 1961, the U.S. Congress rec-
ommended that the International Telecommunications Union (ITU) should examine the
aspects of space communications for which international cooperation would be necessary.
The most critical step was in August 1962, when the U.S. Congress passed the Commun-
ications Satellite Act. This set the stage for commercial investment in an international
satellite organization and, on July 19, 1964, representatives of the first 12 countries to in-
vest in what became Intelsat (the International Telecommunications Satellite Organiza-
tion) signed an initial agreement. The company that represented the United States at this
initial signing ceremony was Comsat, an entity specifically created to act for the United
States within Intelsat. It should be remembered that, at this point, the Bell System had a
complete monopoly of all long-distance telephone communications within the United
c01.qxd 13/07/02 14:02 Page 4
States. When Congress passed the Communications Satellite Act, the Bell System was
specifically barred from directly participating in satellite communications, although it was
permitted to invest in Comsat.
Comsat essentially managed Intelsat in the formative years and should be credited
with the remarkable success of the international venture. The first five Intelsat series of
satellites (INTELSAT I through V) were selected, and their procurement managed, by
teams put in place under Comsat leadership. Over this same phase, though, large portions
of the Comsat engineering and operations groups transferred over to Intelsat so that, when
the Permanent Management Arrangements came into force in 1979, many former Comsat
groups were now part of Intelsat.
In mid-1963, 99% of all satellites had been launched into LEO. LEO, and the slightly
higher medium earth orbit (MEO), were much easier to reach than GEO with the small
launchers available at that time. The intense debate was eventually settled on launcher re-
liability issues rather than on payload capabilities. The first 6 years of the so-called space
age was a period of both payload and launcher development. The new frontier was very
risky, with about one launch in four being fully successful. The system architecture of the
first proposed commercial communications satellite system employed 12 satellites in an
equatorial MEO constellation. Thus, with the launch failure rate at the time, 48 launches
were envisioned to guarantee 12 operational satellites in orbit. Without 12 satellites in or-
bit, continuous 24-h coverage could not be offered. Twenty-four hours a day, seven days
a week—referred to as 247 operation—is a requirement for any successful communica-
tions service. A GEO systems architecture requires only one satellite to provide 247 op-
eration over essentially one-third of the inhabited world. On this basis, four launches would
be required to achieve coverage of one third of the earth; 12 for the entire inhabited world.
Despite its unproven technological approach, the geostationary orbit was selected by the
entities that became Intelsat.
The first Intelsat satellite, INTELSAT I (formerly Early Bird) was launched on April
16, 1965. The satellite weighed a mere 36 kg (80 lb) and incorporated two 64 GHz
transponders, each with 25-MHz bandwidth. Commercial operations commenced between
Europe and the United States on June 28, 1965. Thus, about 2 decades after Clarke’s land-
mark article in Wireless World, GEO satellite communications began. Intelsat was highly
successful and grew rapidly as many countries saw the value of improved telecommuni-
cations, not just internationally but for national systems that provided high quality satellite
communications within the borders of large countries.
Canada was the first country to build a national telecommunication system using
GEO satellites. Anik 1A was launched in May 1974, just 2 months before the first U.S.
domestic satellite, WESTAR 1. The honor of the first regional satellite system, however,
goes to the USSR Molniya system of highly elliptic orbit (HEO) satellites, the first of
which was launched in April 1965 (the same month as INTELSAT I). Countries that are
geographically spread like the USSR, which covers 11 time zones, have used regional
satellite systems very effectively. Another country that benefited greatly from a GEO re-
gional system was Indonesia, which consists of more than 3000 islands spread out over
more than a thousand miles. A terrestrially based telecommunication system was not eco-
nomically feasible for these countries, while a single GEO satellite allowed instant com-
munications region wide. Such ease of communications via GEO satellites proved to be
very profitable. Within less than 10 years, Intelsat was self-supporting and, since it was
not allowed to make a profit, it began returning substantial revenues to what were known
as its Signatories. Within 25 years, Intelsat had more than 100 Signatories4 and, in early
2000, there were 143 member countries and Signatories that formed part of the interna-
tional Intelsat community.
c01.qxd 13/07/02 14:02 Page 5
1.2 A BRIEF HISTORY OF SATELLITE COMMUNICATIONS 5
The astonishing commercial success of Intelsat led many nations to invest in their
satellite systems. This was particularly true in the United States. By the end of 1983, tele-
phone traffic carried by the U.S. domestic satellite systems earned more revenue than the
Intelsat system. Many of the original Intelsat Signatories had been privatized by the early
1990s and were, in effect, competing not only with each other in space communications,
but with Intelsat. It was clear that some mechanism had to be found whereby Intelsat could
be turned into a for-profit, private entity, which could then compete with other commer-
cial organizations while still safeguarding the interests of the smaller nations that had come
to depend on the remarkably low communications cost that Intelsat offered. The first step
in the move to privatizing Intelsat was the establishment of a commercial company called
New Skies and the transfer of a number of Intelsat satellites to New Skies.
In the 1970s and 1980s there was rapid development of GEO satellite systems for
international, regional, and domestic telephone traffic and video distribution. In the United
States, the expansion of fiber-optic links with very high capacity and low delay caused vir-
tually all telephone traffic to move to terrestrial circuits by 1985. However, the demand
for satellite systems grew steadily through this period, and the available spectrum in
C band was quickly occupied, leading to expansion into Ku band. In the United States,
most of the expansion after 1985 was in the areas of video distribution and VSAT (very
small aperture terminal) networks. By 1995 it was clear that the GEO orbit capacity at
Ku band would soon be filled, and Ka-band satellite systems would be needed to handle
the expansion of digital traffic, especially wide band delivery of high-speed Internet data.
SES, based in Luxemburg, began two-way multimedia and Internet access service in west-
ern and central Europe at Ka band using the Astra 1H satellite in 20016. Several Ka-band
satellite systems are expected to be operational in the United States by 20037,8.
The ability of satellite systems to provide communication with mobile users had
long been recognized, and the International Maritime Satellite Organization (Inmarsat)
has provided service to ships and aircraft for several decades, although at a high price.
LEO satellites were seen as one way to create a satellite telephone system with worldwide
coverage; numerous proposals were floated in the 1990s, with three LEO systems eventu-
ally reaching completion by 2000 (Iridium, Globalstar, and Orbcomm). The implementation
of a LEO and MEO satellite system for mobile communication has proved much more
costly than anticipated, and the capacity of the systems is relatively small compared to
SIDEBAR
The first step in the move to privatizing Intelsat was Intelsat is currently (2000) in the process of re-
the establishment of a commercial company called newing its major assets through the purchase of up to
New Skies. New Skies is based in the Netherlands seven INTELSAT IX satellites from SS-Loral to re-
and, on 30 November 1998, six satellites were trans- place the current fleet of INTELSAT VI, and some of
ferred from Intelsat ownership to New Skies. There the INTELSAT VII, satellites. Each of these satellites
was one INTELSAT V series satellite (IS-513 at carries the equivalent of 96 units of 36 MHz band-
183° E), one INTELSAT VII series satellite (IS-703 width. The satellites will be located at 62° E, 60° E,
at 57° E), two INTELSAT VIII series satellites 335.5° E, 325.5° E, 332.5° E, 342° E, and 328.5° E.
(IS-803 at 338.5° E and IS-806 at 319.5° E), the More details on the Intelsat fleet of satellites can be
INTELSAT-K satellite (in inclined orbit at 338.5° E), found at http://www.intelsat.int. Intelsat is moving
and a new satellite designed for direct broadcast serv- forward with plans to privatize the remainder of the
ices (K-TV at 95° E). New Skies has as their prime organization in the 2002/2003 time frame. Any reor-
businesses plan the provisioning of TV services, both ganization will contain strong safeguards for smaller
distribution and direct to home. users to the system.
c01.qxd 13/07/02 14:02 Page 6
GEO satellite systems, leading to a higher cost per transmitted bit. Satellite telephone sys-
tems were unable to compete with cellular telephone systems because of the high cost
and relatively low capacity of the space segment. The Iridium system, for example, cost
over $5 B to implement, but provided a total capacity for the United States of less than
10,000 telephone circuits. Iridium Inc. declared bankruptcy in early 2000, having failed
to establish a sufficiently large customer base to make the venture viable. The entire Irid-
ium system was sold to Iridium Satellite LLC for a reported $25 M, approximately 0.5%
of the system’s construction cost. The future of the other LEO and MEO satellite telephone
systems also seemed uncertain at the time this book was written.
Satellite navigation systems, notably the Global Positioning System, have revolu-
tionized navigation and surveying. The Global Positioning System took almost 20 years
to design and fully implement, at a cost of $12 B. By 2000, GPS receivers could be built
in Original Equipment Manufacturer (OEM) form for less than $25, and the worldwide
GPS industry was earning billions of dollars from equipment sales and services. In the
United States, aircraft navigation will depend almost entirely on GPS by 2010, and blind
landing systems using GPS will also be available. Accurate navigation of ships, especially
in coastal waters and bad weather, is also heavily reliant on GPS. Europe is building a
comparable satellite navigation system called Gallileo.
1.3 SATELLITE COMMUNICATIONS IN 2000
Tables 1.1, 1.2, and 1.3 list the majority of the GEO, MEO, and LEO communication
satellites in orbit in 2000. The list is not exhaustive, and excludes satellites used solely
for military communications and surveillance, and those used primarily for weather fore-
casting and earth imaging. Not all the communications satellites are included, and exper-
imental and scientific satellites are omitted. In all, Tables 1.1 and 1.2 list a total of 172
geostationary communication satellites. When other satellites in geostationary orbit are
considered, there were close to 200 GEO satellites in operation in 2000 (Table 1.4).
GEO satellites have always been the backbone of the commercial satellite commu-
nications industry. Large GEO satellites can serve one-third of the earth’s surface, and can
carry up to 4 Gbps of data, or transmit up to 16 high power direct broadcast satellite tele-
vision (DBS-TV) signals, each of which can deliver several video channels. The weight
and power of GEO satellites have also increased. In 2000 a large GEO satellite could
weigh 10,000 kg (10 tons), might generate 12 kW of power, and carry 60 transponders,
with a trend toward even higher powers but lower weight. For example, in 2001 Space
System/Loral contracted with APT Satellite Company Ltd. in Hong Kong to build the
Apstar-V satellite, a GEO satellite serving Asia with a mass of 4845 kg when injected
into geostationary orbit and an expected lifetime of 13 years. Apstar-V will generate an
initial power of 10.6 kW, and carry 38 C-band transponders with 60-W output power and
16 Ku-band transponders at 141 W each5. Satellites generating 25 kW and carrying
antennas with hundreds of beams are planned for the time frame 2005–2010.
Television program distribution and DBS-TV have become the major source of rev-
enue for commercial satellite system operators, earning more than half of the industry’s
$30 B revenues for 1998. By the end of 2000 there were over 14 million DBS-TV cus-
tomers in the United States. The high capacity of GEO satellites results from the use of
high-power terrestrial transmitters and relatively high gain earth station antennas. Earth
station antenna gain translates directly into communication capacity, and therefore into
revenue. Increased capacity lowers the delivery cost per bit for a customer. Systems with
fixed directional antennas can deliver bits at a significantly lower cost than systems using
TABLE 1.1 GEO Satellite Systems: U.S. Operators (after 3, 5)
Organization Satellites Type Transponders Orbit location
American Mobile AMSC-1 Mobile land 16 L band 101° W

Satellite Corp., communications
Reston, VA
c01.qxd 13/07/02 14:02 Page 7
http://www.AmMobile.com
Columbia Columbia 515 Telecommunications 12 C band 37.7° W
Communications Corp., (formerly Intelsat 515) 12 Ka band
Bethesda, MD
http://www.tdrss.com TDRSS-5, 6 Data relay 4, 12 C band 174.3° E, 47° W
Comsat Corp., Inmarsat 2F, 1–4 Mobile 9 GEO satellites Many locations
Bethesda, MD ship and aircraft L band
http://www.comsat.com Inmarsat 3F, 1–5 communications See Inmarsat entry
Comsat is the U.S.
provider of Inmarsat
and Intelsat Telecommunications 19 GEO satellites Many locations
Intelsat services C band and Ku band
See Intelsat entry
Directv Inc., DBS-1, 2, 3, Direct to home Ku band BSS band 110° W
El Segundo, CA Directv-3R digital television DBS-1, 2, 3, 3R:
http://www.directv.com broadcasting 16 HP transponders 101° W
Tempo 2 Tempo 2: 11
transponders 119° W
Echostar Echostar 1–5 Direct to home Ku band BSS band Echostar 1, 2: 119° W
Communications Corp., digital television 16 transponders per Echostar 3: 61.5° W
Littleton, CO broadcasting satellite Echostar 4: 148° W
http://www.dishnetwork.com Echostar 5: 110° W
GE Americom, GE-1, 1A, 2–8 Telecommunications, 24 C band 79° W through 139° W
Princeton, NJ video distribution, Up to 28 Ku band
http://www.geamericom.com GE-1E broadcasting, 16 Ku band 5° E
VSAT networks
GE Gstar 4 16 Ku band 105° W
(continued )
7
8
TABLE 1.1 (continued )
Organization Satellitesa Type Transponders Orbit location
GE-1A and GE-5 are

Satcom C-1, 3, 4, 5 designated for 24 C band 131° W through 139° W
c01.qxd 13/07/02 14:02 Page 8
broadcasting only
Satcom K-2 16 Ku band 85° W
Spacenet 3, 4 18 C band 85° W, 83° W
6 Ku band
Loral Skynet, Telstar 4, 5, 6, 7 Broadcasting, 24 C band 89° W, 97° W, 93° W,
Bedminster, NJ video distribution, 16 to 28 Ku band 129° W
http://www.loralskynet.com Telstar 11, 12 telecommunications 34, 38 Ku band 37.5° W, 34° W
PanAmSat Corp., Galaxy 1RR, 5, 6, 9 Telecommunications 24 C band 133° W, 125° W,
Greenwich, CT 74° W, 123° W
http://www.panamsat.com Galaxy 3R Telecommunications 24 C band, 8 Ku band 95° W
Galaxy 7 Telecommunications 8 C band, 32 Ku band 91° W
Galaxy 8L DBS-TV broadcasting 91° W
to Latin America Up to 24 C band
Up to 36 Ku band
PAS 1, 3R, 4, 5, 6B, 7 Telecommunications 16 C band, 16 Ku band 43° W through 68.5° W
24 C band, 24 Ku band
PAS 2, 8 Telecommunications 36 Ku band 169° W, 166° W
PAS 6 DBS-TV broadcasting 43° W
to South America 14, 14, 19 Ku band
SBS-4, 5, 6 Telecommunications 77° W, 123° W,
74° W
WorldSpace Corp., AfriStar Audio broadcasting 6 HP Ku band 21° E
Washington, DC
http://www.worldspace.com AmeriStar Broadcasting 95° W
AsiaStar Broadcasting 105° E
a
Telecommunications means any form of signal that can be sent through a satellite transponder, including analog and digital voice, data, and video.
For more complete information about these satellite systems consult reference 6.
There are 71 GEO satellites listed in the above table.
TABLE 1.2 GEO Satellite Systems: Non-U.S. and International Operators (after 3, 5)
ACeS Asia Cellular Garuda 1 Mobile communications 140 with spot beams 123° E
Satellite, Indonesia
www.acesinternational.com
c01.qxd 13/07/02 14:02 Page 9
Arab Satellite Arabsat 2A Broadcasting 22 C band, 12 Ku band 26° E

Communications
Organization, Arabsat 2B Broadcasting, 22 C band, 12 Ku band 30.5° E
Riyadh, Saudi Arabia telecommunications
www.arabsat.com Arabsat 3A Broadcasting 20 Ku band 26° E
Asia Broadcasting and L-Star 1 Broadcasting 32 Ku band 126° E
Communications
Network, Ltd., L-Star 2 Telecommunications 32 Ku band 126° E
Bangkok, Thailand
Asia Satellite Asiasat 1 Broadcasting 24 C band 122° E
Telecommunications Co. Asiasat 2 24 C band, 9 Ku band 100.5° E
Ltd., Asiasat 3S Telecommunications 28 C band, 16 Ku band 105.5° E
Hong Kong, PRC
www.asiasat.com
Broadcasting Satellite BSat-1A, BSat-1B DBS-TV 4 Ku band 110° E
System Corp.,
Tokyo, Japan BS-3N Broadcasting 3 Ku band 109.85° E
Deutsche Telekom DFS Kopernikus (1, 2) Broadcasting, 10 Ku band, 1 Ka band 23.5 ° E, 28.5° E
Geschaftsbereich telecommunications
Rundfunk,
Bon-Bad Godesburg,
Germany
www.dtag.de
Embratel, Brasilsat A2 DBS-TV 24 C band 92° W
Rio De Janiero, Brazil broadcasting
Brasilsat B1 28 C band, 1 X band 70° W
Brasilsat B2 28 C band, 1 X band 65° W
Brasilsat B3 28 C band 84° W
(continued )
9
10
Eutelsat, Eutelsat 1 F-4, F-5 Broadcasting, 10 Ku band (2 spare) 25.5° E, 21.5° E
c01.qxd 13/07/02 14:02 Page 10
Paris, France telecommunications

www.eutelsat.com
Eutelsat 2 F-1, F-2, F-3, Broadcasting, 16 Ku band (8 spare) 48° E, 12.5° E, 36° E,
F-4 telecommunications 10° E
16° E, 7° E
Eutelsat W2, W3 Broadcasting, 24 Ku band
telecommunications
Hot Bird DBS-TV 16 Ku band 13° E
Hot Bird 2, 3, 4 DBS-TV 20 Ku band All at 13° E
Inmarsat Ltd., Inmarsat 2F-1, 2F-2, Mobile L band, 179° E, 98° W, 65° E
London, UK 2F-3, 2F-4 telecommunications demand assigned 109° E
www.inmarsat.org
Inmarsat 3F-1, 3F-2, 64° E, 15.5° E, 178° E
3F-3, 3F-4, 3F-5 54° W, 25° E
Intelsat, Intelsat 505 Broadcasting, 21 C band, 6 Ku band 72° E
Washington, DC, USA telecommunications
www.intelsat.int Intelsat 510, 511 26 C band, 6 Ku band, 33° E, 330.5° E
1 L band
Intelsat 601, 602, 603, 64 C band, 24 Ku band 325.5° E, 62° E, 335.5° E
604, 605 60° E, 332.5° E
Intelsat 701, 702, 704, 42 C band, 20 Ku band 180° E, 177° E, 66° E
705, 709 342° E, 310° E
Intelsat 706, 707 Broadcasting, 42 C band, 28 Ku band 307° E, 359° E
telecommunications
Intelsat 801, 802, 804 64 C band, 12 Ku band 328.5° E, 174° E, 64° E
Intelsat 805 36 C band, 6 Ku band 304.5° E
Japan Satellite Systems JCSat-1B, JCSat-2 Telecommunications 32 Ku band 150° E, 154° E
Inc.,
Tokyo, Japan JCSat-3 Broadcasting 12 C band, 28 Ku band 128° E
www.jcsat.co.jp
JCSat-4A Telecommunications 32 Ku band 124° E
Korea Telecom, Koreasat 1, 2 Broadcasting 15 Ku band 116° E, 113° E
Korea
www.kt.co.kr Koreasat 3 Telecommunications 30 Ku band, 3 Ka band 116° E
Mabuhay Philippines Agila 2 DBS-TV, 30 C band, 24 Ku band 146° E
Satellite Corp., broadcasting,
Makati City, Philippines telecommunications
NahuelSat, SA, Nahuel 1 Broadcasting 18 Ku band 71.8° W
Buenos Aires, Argentina
c01.qxd 13/07/02 14:02 Page 11
www.nahuelsat.com.ar Nahuel 2 Telecommunications 46 C band, 36 Ku band 81° W

New Skies Satellites, NS 513 Telecommunications 42 C band, 12 Ku band 183° E
N.V.,
The Hague, Netherlands NS5 703 Broadcasting, 46 C band, 20 Ku band 57° E
www.newskiessat.com telecommunications
NS5 803 Broadcasting, 64 C band, 12 Ku band 21.5° W
telecommunications
NSS 806 Broadcasting, 36 C band, 6 Ku band 40.5° W
telecommunications
NSS K Broadcasting, 16 Ku band 21.5° W
telecommunications
PT Pasifik Satelit Garuda 1 Mobile communications 140 spot beams 123° E
Nusantara,
Bekasi, Indonesia Broadcasting, 4 Ku band 113° E
www.psn.co.id Palapa C1 telecommunications
Sino Satellite SinoSat 1 Broadcasting, 24 C band, 14 Ku band 123° E
Communications Co., telecommunications
Ltd., Beijing, PRC
www.sinosat.com
Societe Europenne des Astra 1A, 1B DBS-TV, multimedia 16 Ku band 19.2° E
Satellites, SA (SES), Astra 1C, 1D DBS-TV, multimedia 20 Ku band 19.2° E
Betzdorf, Luxembourg Astra 1E, 1F DBS-TV, multimedia 20, 22 Ku band 19.2° E
www.astra.lu Astra 1G DBS-TV, multimedia 30 Ku band 19.2° E
Astra 2A, 2B DBS-TV, multimedia 32, 30 Ku band 28.2° E
Astra 2G DBS-TV, multimedia 16 Ku band 28.2° E
(continued )
11
12
Space Communications Superbird A Broadcasting, 23 Ku band, 2 Ka band 158° E

Corp., telecommunications
Tokyo, Japan Superbird B 23 Ku band, 2 Ka band 162° E
c01.qxd 13/07/02 14:02 Page 12
www.superbird.co.jp
Superbird C 23 Ku band, 2 Ka band 144° E
Spacecom Satellite Amos 1 DBS-TV, 7 Ku band 4° W
Communication telecommunications
Services,
Ramat-Gan, Israel
www.spacecom.co.il
Swedish Space Corp., Sirius 1 DBS-TV 5 Ku band 5° E
Solna, Sweden
www.ssc.se Sirius 2 DBS-TV, VSAT 32 Ku band 5° E
networks
Sirius 3 DBS-TV 15 Ku band
Telenor Satellite Thor 1 Broadcasting, 5 Ku band 1° E
Services AS, telecommunications
Oslo, Norway Thor 2 15 Ku band 1° E
www.telenor.com
Thor 3 14 Ku band 1° E
Telesat Canada, Anik E1 Broadcasting, 14 C band, 12 Ku band 111.1° W
Gloucester, ON, Canada telecommunications
www.telesat.ca Anik E2 14 C band, 12 Ku band 107.3° W
TMI Communications, MSat 1 Mobile communications 16 L band, 1 Ku band 106.5° W
Ottawa, Canada
www.tmisolutions.com
Turk Telekom, Turksat 1B, 1C Broadcasting, 31.3° E, 42° E
Ankara, Turkey telecommunications
There are 101 satellites listed in Table1.2.

TABLE 1.3 LEO and MEO Satellite Systems (after 3, 5)
Organization Satellites Type Transponders Constellation orbit

c01.qxd 13/07/02 14:02 Page 13
Globalstar, 48 LEO satellites with Mobile communications, 16 spot beams within Six orbital planes
San Jose, CA, USA four spares in orbit satellite telephones, footprint. Each beam inclined at 52°, eight
www.globalstar.com all digital has multiple 1.25-MHz satellites per plane
channels with 1 to 13 1413 km altitude
channels per beam. (763 nm)
Multiple access through
CDMA. L- and S-band
links to mobiles
Iridium LLC, 66 LEO satellites with Mobile communications, 48 spot beams with Six orbital planes
Washington, DC, USA seven spares in orbit satellite telephones, seven RF channels in inclined at 84.6°, eleven
www.iridium.com all digital 8 MHz. L-band links satellites per plane
to mobiles. Ka-band 898 km altitude
links to Gateways. (485 nm)
22 GHz satellite cross
links. Multiple access
through FDMA/TDMA
Orbcomm Global L.P., 28 LEO satellites Data transmission to Bent pipe transponder 24 satellites in 45°
Dulles, VA, USA handheld and mobile with earth coverage inclined orbits.
www.orbcomm.com terminals beams. Data rate up to Two in 70° inclined
2400 bps in 0.1-s bursts orbits, two inclined 108°
vhf links to mobiles
(uplink 148 MHz,
downlink 137 MHz)
13
c01.qxd 13/07/02 14:02 Page 14
TABLE 1.4 Other Satellite Systems
System Satellites Type and lifetime Application Orbits
Global positioning Navstar GPS Design lifetime 7.5 years Navigation, Six orbital
System (GPS), 13 through 21 early warning planes with
operated by U.S. 22 through 40 Design lifetime 10 years four satellites
Air Force per plane at
Useful web sites: 43, 44, 45 Design lifetime 10 years 20,200 km
www.navcen.uscg.mil altitude.
www.laafb.af.mil/SMC/ All satellites broadcast Inclination
CZ/homepage/ CDMA signals on two of orbital
http://gps.faa.gov/ L-band frequencies plane is 55°
http://www.spacecom.
af.mil
low gain antennas, such as those designed for use by mobile users. Consequently, GEO
satellites look set to be the largest revenue earners in space for the foreseeable future.
Figure 1.1 shows the estimated growth in revenue from all satellite communication
services, projected to 2010.
All radio systems require frequency spectrum, and the delivery of high-speed data
requires a wide bandwidth. Satellite communication systems started in C band, with an
allocation of 500 MHz, shared with terrestrial microwave links. As the GEO orbit filled
up with satellites operating at C band, satellites were built for the next available frequency
band, Ku band. There is a continuing demand for ever more spectrum to allow satellites
to provide new services, with high speed access to the Internet forcing a move to Ka-band
and even higher frequencies. Access to the Internet from small transmitting Ka-band earth
stations located at the home offers a way to bypass the terrestrial telephone network and
achieve much higher bit rates. SES began two-way Ka-band Internet access in Europe in
1998 with the Astra-K satellite, and the next generation of Ka-band satellites in the United
States will offer similar services.
200
Worldwide revenue in billions of $U.S.
100
0
19802000 2010
1990
Year
FIGURE 1.1 Growth of worldwide revenues from satellite communications
1980 through 2010. Beyond 2000, the curve is a projection.
c01.qxd 13/07/02 14:02 Page 15
1.4 OVERVIEW OF SATELLITE COMMUNICATIONS 15
Successive World Radio Conferences have allocated new frequency bands for com-
mercial satellite services that now include L, S, C, Ku, K, Ka, V, and Q bands. Mobile satel-
lite systems use vhf, uhf, L, and S bands with carrier frequencies from 137 to 2500 MHz,
and GEO satellites use frequency bands extending from 3.2 to 50 GHz. Despite the growth
of fiber-optic links with very high capacity, the demand for satellite systems continues to in-
crease. Satellites have also become integrated into complex communications architectures that
use each element of the network to its best advantage. Examples are VSAT/WLL (very small
aperture terminals/wireless local loop) in countries where the communications infrastructure
is not yet mature and GEO/LMDS (local multipoint distribution systems) for the urban fringes
of developed nations where the build-out of fiber has yet to be an economic proposition.
1.4 OVERVIEW OF SATELLITE

COMMUNICATIONS
Satellite communication systems exist because the earth is a sphere. Radio waves travel
in straight lines at the microwave frequencies used for wideband communications, so a
repeater is needed to convey signals over long distances. Satellites, because they can link
places on the earth that are thousands of miles apart, are a good place to locate a repeater,
and a GEO satellite is the best place of all. A repeater is simply a receiver linked to a
transmitter, always using different radio frequencies, that can receive a signal from one
earth station, amplify it, and retransmit it to another earth station. The repeater derives
its name from nineteenth century telegraph links, which had a maximum length of about
50 miles. Telegraph repeater stations were required every 50 miles in a long-distance link
so that the Morse code signals could be re-sent before they became too weak to read.
The majority of communication satellites are in geostationary earth orbit, at an
altitude of 35,786 km. Typical path length from an earth station to a GEO satellite is
38,500 km. Radio signals get weaker in proportion to the square of the distance trav-
eled, so signals reaching a satellite are always very weak. Similarly, signals received on
earth from a satellite 38,500 km away are also very weak, because of limits on the
weight of GEO satellites and the electrical power they can generate using solar cells. It
costs roughly $25,000 per kilogram to get a geostationary satellite in orbit. This obvi-
ously places severe restrictions on the size and weight of GEO satellites, since the high
cost of building and launching a satellite must be recovered over a 10 to 15 year lifetime
by selling communications capacity.
Satellite communication systems are dominated by the need to receive very weak
signals. In the early days, very large receiving antennas, with diameters up to 30 m, were
needed to collect sufficient signal power to drive video signals or multiplexed telephone
channels. As satellites have become larger, heavier, and more powerful, smaller earth
station antennas have become feasible, and Direct Broadcast Satellite TV (DBS-TV) re-
ceiving systems can use dish antennas as small as 0.5 m in diameter.
Satellite systems operate in the microwave and millimeter wave frequency bands,
using frequencies between 1 and 50 GHz. Above 10 GHz, rain causes significant attenu-
ation of the signal and the probability that rain will occur in the path between the satellite
and an earth station must be factored into the system design. Above 20 GHz, attenuation
in heavy rain (usually associated with thunderstorms) can cause sufficient attenuation that
the link will fail.
For the first 20 years of satellite communications, analog signals were widely used,
with most links employing frequency modulation (FM). Wideband FM can operate at
low carrier-to-noise ratios (C/N), in the 5 to 15 dB range, but adds a signal-to-noise
c01.qxd 13/07/02 14:02 Page 16
improvement so that video and telephone signals can be delivered with signal-to-noise
ratios (S/N) of 50 dB. The penalty for the improvement is that the radio frequency (RF)
signal occupies a much larger bandwidth than the baseband signal. In satellite links, that
penalty results because signals are always weak and the improvement in signal-to-noise
ratio is essential.
The move toward digital communications in terrestrial telephone and data trans-
mission has been mirrored by a similar move toward digital transmission over satellite
links. In the United States, only TV distribution at C band remains as the major analog
satellite transmission system. Even this last bastion of analog signaling seems destined to
disappear as cable TV stations switch over to digital receivers that allow six TV signals
to be sent though a single Ku-band transponder. More importantly, dual standards per-
mitting the transmission of not only digital TV but also high definition TV (HDTV), will
eventually remove analog TV from consideration.
Almost all other signals are digital—telephony, data, DBS-TV, radio broadcasting,
and navigation with GPS all use digital signaling techniques. All of the LEO and MEO
mobile communication systems are digital, taking advantage of voice compression tech-
niques that allow a digital voice signal to be compressed into a bit stream at 4.8 kbps.
Similarly, MPEG 2 (Moving Picture Coding Expert Group) and other video compression
techniques allow video signals to be transmitted in full fidelity at rates less then 6.2 Mbps.
1.5 SUMMARY
Satellite communication systems have become an services, because the use of high gain fixed anten-
essential part of the world’s telecommunications in- nas at earth stations maximizes the capacity of the
frastructure, serving billions of people with tele- satellite. Over the years, there has been a trend away
phone, data, and video services. Despite the growth from trunk communications using very large earth
of fiber-optic links, which have much greater ca- station antennas toward delivery from more power-
pacity than satellite systems and a lower cost per ful satellites to individual users using much smaller
bit, satellite systems continue to thrive and invest- antennas. LEO and MEO satellites are used for mo-
ment in new systems continues. Satellite services bile communications and navigation systems and,
have shifted away from telephony toward video and as the need for Geographic Information Systems
data delivery, with television broadcasting directly grows with a variety of applications, LEO earth im-
to the home emerging as one of the most powerful aging satellites have the potential to provide strong
applications. GEO satellites carry the majority of revenue streams.
REFERENCES
1. A. C. CLARKE, “Extra-terrestrial Relays,” Wireless World, 5. Aviation Week and Space Technology, Aerospace Source
pp. 305–308, 1945. Book, McGraw-Hill, New York, Vol. 154, No. 3,
2. A. C. CLARKE, 2001: A Space Odyssey, New American pp. 161–179 and pp. 249–266, Jan. 15, 2001.
Library, New York. 6. http://www.astra.lu
3. Aviation Week and Space Technology, Aerospace Source 7. www.astrolink.com
Book, McGraw-Hill, New York, Vol. 153, No. 3, January 8. www.hns.com.spaceway
17, 2000.
4. D. W. E. REES, “Satellite Communications: The First
Quarter Century of Service,” John Wiley & Sons, New
York, 1989.
c02.qxd 13/07/02 14:10 Page 17
CHAPTER 2
ORBITAL MECHANICS
AND LAUNCHERS
2.1 ORBITAL MECHANICS
Developing the Equations of the Orbit

This chapter is about how earth orbit is achieved, the laws that describe the motion of an
object orbiting another body, how satellites maneuver in space, and the determination of
the look angle to a satellite from the earth using ephemeris data that describe the orbital
trajectory of the satellite.
To achieve a stable orbit around the earth, a spacecraft must first be beyond the bulk
of the earth’s atmosphere, i.e., in what is popularly called space. There are many defini-
tions of space. U.S. astronauts are awarded their “space wings” if they fly at an altitude
that exceeds 50 miles (80 km); some international treaties hold that the space frontier
above a given country begins at a height of 100 miles (160 km). Below 100 miles, per-
mission must be sought to over-fly any portion of the country in question. On reentry, at-
mospheric drag starts to be felt at a height of about 400,000 ft (76 miles 122 km).
Most satellites, for any mission of more than a few months, are placed into orbits of at
least 250 miles (400 km) above the earth. Even at this height, atmospheric drag is sig-
nificant. As an example, the initial payload elements of the International Space Station
(ISS) were injected into orbit at an altitude of 397 km when the shuttle mission left those
modules on 9 June 1999. By the end of 1999, the orbital height had decayed to about
360 km, necessitating a maneuver to raise the orbit. Without onboard thrusters and suffi-
cient orbital maneuvering fuel, the ISS would not last more than a few years at most in
such a low orbit. To appreciate the basic laws that govern celestial mechanics, we will be-
gin first with the fundamental Newtonian equations that describe the motion of a body.
We will then give some coordinate axes within which the orbit of the satellite can be set
and determine the various forces on the earth satellite.
Newton’s laws of motion can be encapsulated into four equations:
s ut 1 12 2at 2 (2.1a)
v2 u2 2at (2.1b)
v u at (2.1c)
P ma (2.1d)
where s is the distance traveled from time t 0; u is the initial velocity of the object at
time t 0 and v the final velocity of the object at time t; a is the acceleration of the ob-
ject; P is the force acting on the object; and m is the mass of the object. Note that the ac-
celeration can be positive or negative, depending on the direction it is acting with respect
to the velocity vector. Of these four equations, it is the last one that helps us understand
the motion of a satellite in a stable orbit (neglecting any drag or other perturbing forces).
Put into words, Eq. (2.1d) states that the force acting on a body is equal to the mass of
17
c02.qxd 13/07/02 14:10 Page 18
18 CHAPTER 2 ORBITAL MECHANICS AND LAUNCHERS
the body multiplied by the resulting acceleration of the body. Alternatively, the resulting
acceleration is the ratio of the force acting on the body to the mass of the body. Thus, for
a given force, the lighter the mass of the body, the higher the acceleration will be. When
in a stable orbit, there are two main forces acting on a satellite: a centrifugal force due to
the kinetic energy of the satellite, which attempts to fling the satellite into a higher orbit,
and a centripetal force due to the gravitational attraction of the planet about which the
satellite is orbiting, which attempts to pull the satellite down toward the planet. If these
two forces are equal, the satellite will remain in a stable orbit. It will continually fall to-
ward the planet’s surface as it moves forward in its orbit but, by virtue of its orbital ve-
locity, it will have moved forward just far enough to compensate for the “fall” toward the
planet and so it will remain at the same orbital height. This is why an object in a stable
orbit is sometimes described as being in “free fall.” Figure 2.1 shows the two opposing
forces on a satellite in a stable orbit1.
Force mass acceleration and the unit of force is a Newton, with the notation
N. A Newton is the force required to accelerate a mass of 1 kg with an acceleration of
1 m/s2. The underlying units of a Newton are therefore (kg) m/s2. In Imperial Units,
one Newton 0.2248 ft lb. The standard acceleration due to gravity at the earth’s sur-
face is 9.80665 103 km/s2, which is often quoted as 981 cm/s2. This value decreases
The satellite has a mass, m,

and is traveling with velocity,
v, in the plane of the orbit
2
FOUT = mv
r
FIN = GM2Em
r
ME
FIGURE 2.1 Forces acting on a satellite in a stable orbit around the earth (from Fig. 3.4 of
reference 1). Gravitational force is inversely proportional to the square of the distance between
the centers of gravity of the satellite and the planet the satellite is orbiting, in this case the
earth. The gravitational force inward (FIN, the centripetal force) is directed toward the center of
gravity of the earth. The kinetic energy of the satellite (FOUT, the centrifugal force) is directed
diametrically opposite to the gravitational force. Kinetic energy is proportional to the square of
the velocity of the satellite. When these inward and outward forces are balanced, the satellite
moves around the earth in a “free fall” trajectory: the satellite’s orbit. For a description of the
units, please see the text.
c02.qxd 13/07/02 14:10 Page 19
2.1 ORBITAL MECHANICS 19
with height above the earth’s surface. The acceleration, a, due to gravity at a distance r
from the center of the earth is1
a mr 2 km/s2 (2.1)
where the constant is the product of the universal gravitational constant G and the mass
of the earth ME.
The product GME is called Kepler’s constant and has the value 3.986004418
105 km3/s2. The universal gravitational constant is G 6.672 1011 Nm2/kg2 or 6.672
1020 km3/kg s2 in the older units. Since force mass acceleration, the centripetal
force acting on the satellite, FIN, is given by
FIN m 1mr 2 2 (2.2a)
m 1GME r 2 2 (2.2b)
In a similar fashion, the centrifugal acceleration is given by1
a v2 r (2.3)
which will give the centrifugal force, FOUT, as
FOUT m 1v2r2 (2.4)
If the forces on the satellite are balanced, FIN FOUT and, using Eqs. (2.2a) and (2.4),
m mr 2 m v2 r
hence the velocity v of a satellite in a circular orbit is given by
v 1mr2 12 (2.5)
If the orbit is circular, the distance traveled by a satellite in one orbit around a planet is
2r, where r is the radius of the orbit from the satellite to the center of the planet. Since
distance divided by velocity equals time to travel that distance, the period of the satellite’s
orbit, T, will be
T 12pr2 v 12pr2 3 1mr2 12 4
Giving
T 12pr 32 2 1m12 2 (2.6)
Table 2.1 gives the velocity, v, and orbital period, T, for four satellite systems that
occupy typical LEO, MEO, and GEO orbits around the earth. In each case, the orbits are
TABLE 2.1 Orbital Velocity, Height, and Period

of Four Satellite Systems
Orbital height Orbital velocity Orbital period

Satellite system (km) (km/s) (h min s)
Intelsat (GEO) 35,786.03 3.0747 23 56 4.1

New-ICO (MEO) 10,255 4.8954 5 55 48.4
Skybridge (LEO) 1,469 7.1272 1 55 17.8
Iridium (LEO) 780 7.4624 1 40 27.0
Mean earth radius is 6378.137 km and GEO radius from the center of the
earth is 42,164.17 km.
c02.qxd 13/07/02 14:10 Page 20
z
FIGURE 2.2 The initial coordinate
system that could be used to de-
scribe the relationship between the
Earth rotation earth and a satellite. A Cartesian
Satellite coordinate system with the geo-
graphical axes of the earth as the
r principal axes is the simplest coor-
dinate system to set up. The rota-
tional axis of the earth is about the
c Equatorial plane
axis cz, where c is the center of the
earth and cz passes through the
geographic north pole. Axes cx, cy,
and cz are mutually orthogonal
axes, with cx and cy passing
y
through the earth’s geographic
equator. The vector r locates the
moving satellite with respect to the
x center of the earth.
circular and the average radius of the earth is taken as 6378.137 km1. A number of coor-
dinate systems and reference planes can be used to describe the orbit of a satellite around
a planet. Figure 2.2 illustrates one of these using a Cartesian coordinate system with the
earth at the center and the reference planes coinciding with the equator and the polar axis.
This is referred to as a geocentric coordinate system.
With the coordinate system set up as in Figure 2.2, and with the satellite mass m
located at a vector distance r from the center of the earth, the gravitational force F on the
satellite is given by
GME m r
F (2.7)
r3
Where ME is the mass of the earth and G 6.672 1011 Nm2/kg2. But force mass
acceleration and Eq. (2.7) can be written as
d2r
Fm (2.8)
dt 2
From Eqs. (2.7) and (2.8) we have
r d2r
3m (2.9)
r dt 2
Which yields
d2r r
2 m0 (2.10)
dt r3
This is a second-order linear differential equation and its solution will involve six
undetermined constants called the orbital elements. The orbit described by these orbital
elements can be shown to lie in a plane and to have a constant angular momentum. The
solution to Eq. (2.10) is difficult since the second derivative of r involves the second de-
rivative of the unit vector r. To remove this dependence, a different set of coordinates can
c02.qxd 13/07/02 14:10 Page 21
z
z0
FIGURE 2.3 The orbital plane coor-

dinate system. In this coordinate sys-
tem, the orbital plane of the satellite
is used as the reference plane. The
orthogonal axes x0 and y0 lie in the
c y0 orbital plane. The third axis, z0, is
perpendicular to the orbital plane.
The geographical z-axis of the earth
(which passes through the true North
Pole and the center of the earth, c)
does not lie in the same direction as
the z0 axis except for satellite orbits
that are exactly in the plane of the
x0 geographical equator.
be chosen to describe the location of the satellite such that the unit vectors in the three
axes are constant. This coordinate system uses the plane of the satellite’s orbit as the ref-
erence plane. This is shown in Figure 2.3.
Expressing Eq. (2.10) in terms of the new coordinate axes x0, y0, and z0 gives
d 2x0 d 2y0 m1x0 x̂0 y0 ŷ0 2
x̂0 a b a 2 b 0
1x20 y20 2 32
2 ŷ0 (2.11)
dt dt
Equation (2.11) is easier to solve if it is expressed in a polar coordinate system rather than
a Cartesian coordinate system. The polar coordinate system is shown in Figure 2.4.
With the polar coordinate system shown in Figure 2.4 and using the transformations
x0 r0 cos f0 (2.12a)
y0 r0 sin f0 (2.12b)
x̂0 r̂0 cos f0 f̂0 sin f0 (2.12c)
ŷ0 f̂0 cos f0 r̂0 sin f0 (2.12d)
and equating the vector components of r0 and 0 in turn in Eq. (2.11) yields
d 2r0
2 r0 a b 2
df0 m
(2.13)
dt dt r0
y0
FIGURE 2.4 Polar coordinate system in the plane

of the satellite’s orbit. The plane of the orbit coin-
cides with the plane of the paper. The axis z0 is
r0 straight out of the paper from the center of the
earth, and is normal to the plane of the satellite’s
φ0 orbit. The satellite’s position is described in terms
of the radius from the center of the earth r0 and the
x0
z0 angle this radius makes with the x0 axis, 0.
c02.qxd 13/07/02 14:10 Page 22
and
d 2f0
r0 a 2 b 2a ba b0
dr0 df0
(2.14)
dt dt dt
Using standard mathematical procedures, we can develop an equation for the radius
of the satellite’s orbit, r0, namely
p
r0
1 e cos 1f0 u0 2
(2.15)
Where 0 is a constant and e is the eccentricity of an ellipse whose semilatus rectum p is

given by
p 1h2 2 m (2.16)
and h is magnitude of the orbital angular momentum of the satellite. That the equation of
the orbit is an ellipse is Kepler’s first law of planetary motion.
Kepler’s Three Laws of Planetary Motion

Johannes Kepler (1571–1630) was a German astronomer and scientist who developed his
three laws of planetary motion by careful observations of the behavior of the planets in
the solar system over many years, with help from some detailed planetary observations
by the Hungarian astronomer Tycho Brahe. Kepler’s three laws are
1. The orbit of any smaller body about a larger body is always an ellipse, with the cen-
ter of mass of the larger body as one of the two foci.
2. The orbit of the smaller body sweeps out equal areas in equal time (see Figure 2.5).
t2
A 12
t3
A 34
E
t1
t4
FIGURE 2.5 Illustration of Kepler’s second law of planetary motion. A satellite is in orbit
about the planet earth, E. The orbit is an ellipse with a relatively high eccentricity, that is,
it is far from being circular. The figure shows two shaded portions of the elliptical plane in
which the orbit moves, one is close to the earth and encloses the perigee while the other
is far from the earth and encloses the apogee. The perigee is the point of closest ap-
proach to the earth while the apogee is the point in the orbit that is furthest from the
earth. While close to perigee, the satellite moves in the orbit between times t1 and t2 and
sweeps out an area denoted by A12. While close to apogee, the satellite moves in the orbit
between times t3 and t4 and sweeps out an area denoted by A34. If t1 t2 t3 t4 then
A12 A34.
c02.qxd 13/07/02 14:10 Page 23
SIDEBAR
Kepler’s laws were subsequently confirmed, about 50 the concept of gravitational attraction. The work was
years later, by Isaac Newton, who developed a math- published in the Philosophiae Naturalis Principia
ematical model for the motion of the planets. New- Mathematica in 1687. At that time, Latin was the in-
ton was one of the first people to make use of differ- ternational language of formally educated people,
ential calculus, and with his understanding of gravity, much in the way English has become the international
was able to describe the motion of planets from a language of e-mail and business today, so Newton’s
mathematical model based on his laws of motion and Principia was written in Latin.
3. The square of the period of revolution of the smaller body about the larger body
equals a constant multiplied by the third power of the semimajor axis of the orbital
ellipse. That is, T 2 (4 2a3) where T is the orbital period, a is the semimajor
axis of the orbital ellipse, and is Kepler’s constant. If the orbit is circular, then a
becomes distance r, defined as before, and we have Eq. (2.6).
Describing the orbit of a satellite enables us to develop Kepler’s second two laws.
Describing the Orbit of a Satellite

The quantity 0 in Eq. (2.15) serves to orient the ellipse with respect to the orbital plane
axes x0 and y0. Now that we know that the orbit is an ellipse, we can always choose x0
and y0 so that 0 is zero. We will assume that this has been done for the rest of this
discussion. This now gives the equation of the orbit as
p
r0 (2.17)
1 e cos f0
The path of the satellite in the orbital plane is shown in Figure 2.6. The lengths a and b
of the semimajor and semiminor axes are given by
a p 11 e2 2 (2.18)
b a11 e2 2 12 (2.19)
The point in the orbit where the satellite is closest to the earth is called the perigee
and the point where the satellite is farthest from the earth is called the apogee. The perigee
and apogee are always exactly opposite each other. To make 0 equal to zero, we have
chosen the x0 axis so that both the apogee and the perigee lie along it and the x0 axis is
therefore the major axis of the ellipse.
The differential area swept out by the vector r0 from the origin to the satellite in
time dt is given by
dA 0.5r 02 a b dt 0.5hdt
df0
(2.20)
dt
Remembering that h is the magnitude of the orbital angular momentum of the satellite,
the radius vector of the satellite can be seen to sweep out equal areas in equal times. This
is Kepler’s second law of planetary motion. By equating the area of the ellipse (ab) to
the area swept out in one orbital revolution, we can derive an expression for the orbital
period T as
T 2 14p2a3 2 m (2.21)
c02.qxd 13/07/02 14:10 Page 24
yo
ro
φo Perigee
Apogee a C O xo
ae
a (1 + e) a (1 − e)
FIGURE 2.6 The orbit as it appears in the orbital plane. The point O is the center of
the earth and the point C is the center of the ellipse. The two centers do not coincide
unless the eccentricity, e, of the ellipse is zero (i.e., the ellipse becomes a circle and
a b). The dimensions a and b are the semimajor and semiminor axes of the orbital
ellipse, respectively.
This equation is the mathematical expression of Kepler’s third law of planetary mo-
tion: the square of the period of revolution is proportional to the cube of the semimajor
axis. (Note that this is the square of Eq. (2.6) and that in Eq. (2.6) the orbit was assumed
to be circular such that semimajor axis a semiminor axis b circular orbit radius
from the center of the earth r.) Kepler’s third law extends the result from Eq. (2.6), which
was derived for a circular orbit, to the more general case of an elliptical orbit. Equa-
tion (2.21) is extremely important in satellite communications systems. This equation
determines the period of the orbit of any satellite, and it is used in every GPS receiver
in the calculation of the positions of GPS satellites. Equation (2.21) is also used to find
the orbital radius of a GEO satellite, for which the period T must be made exactly equal
to the period of one revolution of the earth for the satellite to remain stationary over a
point on the equator.
An important point to remember is that the period of revolution, T, is referenced to
inertial space, namely, to the galactic background. The orbital period is the time the or-
biting body takes to return to the same reference point in space with respect to the galac-
tic background. Nearly always, the primary body will also be rotating and so the period
of revolution of the satellite may be different from that perceived by an observer who is
standing still on the surface of the primary body. This is most obvious with a geostation-
ary earth orbit (GEO) satellite (see Table 2.1). The orbital period of a GEO satellite is ex-
actly equal to the period of rotation of the earth, 23 h 56 min 4.1 s, but, to an observer
on the ground, the satellite appears to have an infinite orbital period: it always stays in
the same place in the sky.
c02.qxd 13/07/02 14:10 Page 25
To be perfectly geostationary, the orbit of a satellite needs to have three features:

(a) it must be exactly circular (i.e., have an eccentricity of zero); (b) it must be at the
correct altitude (i.e., have the correct period); and (c) it must be in the plane of the equa-
tor (i.e., have a zero inclination with respect to the equator). If the inclination of the
satellite is not zero and/or if the eccentricity is not zero, but the orbital period is cor-
rect, then the satellite will be in a geosynchronous orbit. The position of a geosynchro-
nous satellite will appear to oscillate about a mean look angle in the sky with respect
to a stationary observer on the earth’s surface. The orbital period of a GEO satellite,
23 h 56 min 4.1 s, is one sidereal day. A sidereal day is the time between consecutive
crossings of any particular longitude on the earth by any star, other than the sun1. The
mean solar day of 24 h is the time between any consecutive crossings of any particular
longitude by the sun, and is the time between successive sunrises (or sunsets) observed
at one location on earth, averaged over an entire year. Because the earth moves round
the sun once per 365 1⁄4 days, the solar day is 1440365.25 3.94 min longer than a
sidereal day.
Locating the Satellite in the Orbit

Consider now the problem of locating the satellite in its orbit. The equation of the orbit
may be rewritten by combining Eqs. (2.15) and (2.18) to obtain
a11 e2 2
r0 (2.22)
1 e cos f0
The angle 0 (see Figure 2.6) is measured from the x0 axis and is called the true anom-
aly. [Anomaly was a measure used by astronomers to mean a planet’s angular distance
from its perihelion (closest approach to the sun), measured as if viewed from the sun. The
term was adopted in celestial mechanics for all orbiting bodies.] Since we defined the pos-
itive x0 axis so that it passes through the perigee, 0 measures the angle from the perigee
to the instantaneous position of the satellite. The rectangular coordinates of the satellite
are given by
x0 r0 cos f0 (2.23)
y0 r0 sin f0 (2.24)
As noted earlier, the orbital period T is the time for the satellite to complete a rev-
olution in inertial space, traveling a total of 2 radians. The average angular velocity
is thus
h 12p2 T 1m12 2 1a32 2 (2.25)
If the orbit is an ellipse, the instantaneous angular velocity will vary with the position of
the satellite around the orbit. If we enclose the elliptical orbit with a circumscribed cir-
cle of radius a (see Figure 2.7), then an object going around the circumscribed circle with
a constant angular velocity would complete one revolution in exactly the same period
T as the satellite requires to complete one (elliptical) orbital revolution.
Consider the geometry of the circumscribed circle as shown in Figure 2.7. Locate
the point (indicated as A) where a vertical line drawn through the position of the satellite
intersects the circumscribed circle. A line from the center of the ellipse (C) to this point
(A) makes an angle E with the x0 axis; E is called the eccentric anomaly of the satellite.
c02.qxd 13/07/02 14:10 Page 26
yo axis
A
a
yo
E
xo axis
C O xo
Orbit
Circumscribed Circle
FIGURE 2.7 The circumscribed circle and the eccentric anomaly E. Point O is the center of
the earth and point C is both the center of the orbital ellipse and the center of the circum-
scribed circle. The satellite location in the orbital plane coordinate system is specified by (x0,
y0). A vertical line through the satellite intersects the circumscribed circle at point A. The
eccentric anomaly E is the angle from the x0 axis to the line joining C and A.
It is related to the radius r0 by

r0 a11 e cos E2 (2.26)
Thus
a r0 ae cos E (2.27)
We can also develop an expression that relates eccentric anomaly E to the average
angular velocity , which yields
h dt 11 e cos E2 dE (2.28)
Let tp be the time of perigee. This is simultaneously the time of closest approach to the
earth; the time when the satellite is crossing the x0 axis; and the time when E is zero. If
we integrate both sides of Eq. (2.28), we obtain
h1t tp 2 E e sin E (2.29)
The left side of Eq. (2.29) is called the mean anomaly, M. Thus
M h1t tp 2 E e sin E (2.30)
The mean anomaly M is the arc length (in radians) that the satellite would have traversed
since the perigee passage if it were moving on the circumscribed circle at the mean an-
gular velocity .
If we know the time of perigee, tp, the eccentricity, e, and the length of the semi-
major axis, a, we now have the necessary equations to determine the coordinates (r0, 0)
c02.qxd 13/07/02 14:10 Page 27
and (x0, y0) of the satellite in the orbital plane. The process is as follows
1. Calculate using Eq. (2.25).
2. Calculate M using Eq. (2.30).
3. Solve Eq. (2.30) for E.
4. Find r0 from E using Eq. (2.27).
5. Solve Eq. (2.22) for 0.
6. Use Eqs. (2.23) and (2.24) to calculate x0 and y0.
Now we must locate the orbital plane with respect to the earth.
Locating the Satellite with Respect

to the Earth
At the end of the last section, we summarized the process for locating the satellite at
the point (x0, y0, z0) in the rectangular coordinate system of the orbital plane. The lo-
cation was with respect to the center of the earth. In most cases, we need to know
where the satellite is from an observation point that is not at the center of the earth.
We will therefore develop the transformations that permit the satellite to be located
from a point on the rotating surface of the earth. We will begin with a geocentric equa-
torial coordinate system as shown in Figure 2.8. The rotational axis of the earth is the
zi axis, which is through the geographic North Pole. The xi axis is from the center of
the earth toward a fixed location in space called the first point of Aries (see Figure 2.8).
This coordinate system moves through space; it translates as the earth moves in its or-
bit around the sun, but it does not rotate as the earth rotates. The xi direction is always
the same, whatever the earth’s position around the sun and is in the direction of the
first point of Aries. The (xi, yi) plane contains the earth’s equator and is called the
equatorial plane.
Angular distance measured eastward in the equatorial plane from the xi axis is
called right ascension and given the symbol RA. The two points at which the orbit
zi
FIGURE 2.8 The geocentric

equatorial system. This geocentric
system differs from that shown in
Figure 2.1 only in that the xi axis
points to the first point of Aries.
The first point of Aries is the di-
rection of a line from the center
δ
of the earth through the center of
the sun at the vernal equinox
(about March 21 in the Northern
Hemisphere), the instant when
RA the subsolar point crosses the
yi
equator from south to north. In
the above system, an object may
be located by its right ascension
xi RA and its declination .
c02.qxd 13/07/02 14:10 Page 28
penetrates the equatorial plane are called nodes; the satellite moves upward through
the equatorial plane at the ascending node and downward through the equatorial
plane at the descending node, given the conventional picture of the earth, with north
at the top, which is in the direction of the positive z axis for the earth centered coor-
dinate set. Remember that in space there is no up or down; that is a concept we are
familiar with because of gravity at the earth’s surface. For a weightless body in space,
such as an orbiting spacecraft, up and down have no meaning unless they are defined
with respect to a reference point. The right ascension of the ascending node is called
. The angle that the orbital plane makes with the equatorial plane (the planes inter-
sect at the line joining the nodes) is called the inclination, i. Figure 2.9 illustrates these
quantities.
The variables and i together locate the orbital plane with respect to the equato-
rial plane. To locate the orbital coordinate system with respect to the equatorial coordi-
nate system we need , the argument of perigee west. This is the angle measured along
the orbit from the ascending node to the perigee.
Standard time for space operations and most other scientific and engineering pur-
poses is universal time (UT), also known as zulu time (z). This is essentially the mean
solar time at the Greenwich Observatory near London, England. Universal time is meas-
ured in hours, minutes, and seconds or in fractions of a day. It is 5 h later than Eastern
Standard Time, so that 07:00 EST is 12:00:00 h UT. The civil or calendar day begins
at 00:00:00 hours UT, frequently written as 0 h. This is, of course, midnight (24:00:00)
on the previous day. Astronomers employ a second dating system involving Julian days
and Julian dates. Julian days start at noon UT in a counting system whereby noon on
December 31, 1899, was the beginning of Julian day 2415020, usually written 241 5020.
These are extensively tabulated in reference 2 and additional information is in reference
14. As an example, noon on December 31, 2000, the eve of the twenty-first century, is
the start of Julian day 245 1909. Julian dates can be used to indicate time by append-
ing a decimal fraction; 00:00:00 h UT on January 1, 2001—zero hour, minute, and
zi
Satellite
Perigee
yi
ω
Ω Ascending node
xi
FIGURE 2.9 Locating the orbit in the geocentric equatorial system. The satellite penetrates
the equatorial plane (while moving in the positive z direction) at the ascending node. The
right ascension of the ascending node is and the inclination i is the angle between the
equatorial plane and the orbital plane. Angle , measured in the orbital plane, locates the
perigee with respect to the equatorial plane.
c02.qxd 13/07/02 14:10 Page 29
second for the third millenium A.D.—is given by Julian date 245 1909.5. To find the
exact position of an orbiting satellite at a given instant in time requires knowledge of
the orbital elements.
Orbital Elements
To specify the absolute (i.e., the inertial) coordinates of a satellite at time t, we need to
know six quantities. (This was evident earlier when we determined that a satellite’s equa-
tion of motion was a second order vector linear differential equation.) These quantities
are called the orbital elements. More than six quantities can be used to describe a unique
orbital path and there is some arbitrariness in exactly which six quantities are used. We
have chosen to adopt a set that is commonly used in satellite communications: eccentric-
ity (e), semimajor axis (a), time of perigee (tp), right ascension of ascending node (),
inclination (i), and argument of perigee (). Frequently, the mean anomaly (M) at a given
time is substituted for tp.
EXAMPLE 2.1.1 Geostationary Satellite Orbit Radius

The earth rotates once per sidereal day of 23 h 56 min 4.09 s. Use Eq. (2.21) to show that the radius
of the GEO is 42,164.17 km as given in Table 2.1.
Answer Equation (2.21) gives the square of the orbital period in seconds
T 2 14p2a3 2 m
Rearranging the equation, the orbital radius a is given by
a3 T 2m 14p 2 2
For one sidereal day, T 86,164.09 s. Hence
a3 186,164.12 2 3.986004418 105 14p2 2 7.496020251 1013 km3
a 42,164.17 km
This is the orbital radius for a geostationary satellite, as given in Table 2.1.
EXAMPLE 2.1.2 Low Earth Orbit

The Space Shuttle is an example of a low earth orbit satellite. Sometimes, it orbits at an altitude of
250 km above the earth’s surface, where there is still a finite number of molecules from the at-
mosphere. The mean earth’s radius is approximately 6378.14 km. Using these figures, calculate the
period of the shuttle orbit when the altitude is 250 km and the orbit is circular. Find also the linear
velocity of the shuttle along its orbit.
Answer The radius of the 250-km altitude Space Shuttle orbit is (re h) 6378.14 250.0
6628.14 km
From Eq. 2.21, the period of the orbit is T where
T 2 14p2a3 2 m 4p2 16628.142 33.986004418 105 s2

2.88401145 107 s2
Hence the period of the orbit is
T 5370.30 s 89 min 30.3 s.

c02.qxd 13/07/02 14:10 Page 30
This orbit period is about as small as possible. At a lower altitude, friction with the earth’s atmos-
phere will quickly slow the Shuttle down and it will return to earth. Thus, all spacecraft in stable
earth orbit have orbital periods exceeding 89 min 30 s.
The circumference of the orbit is 2a 41,645.83 km.
Hence the velocity of the Shuttle in orbit is
2paT 41,645.835370.13 7.755 km/s
Alternatively, you could use Eq. (2.5): v (r)12. The term 3.986004418 105 km3/s2 and
the term r (6378.14 250.0) km, yielding v 7.755 km/s.
Note: If and r had been quoted in units of m3/s2 and m, respectively, the answer would have been
in meters/second. Be sure to keep the units the same during a calculation procedure.
A velocity of about 7.8 km/s is a typical velocity for a low earth orbit satellite. As the alti-
tude of a satellite increases, its velocity becomes smaller.
EXAMPLE 2.1.3 Elliptical orbit

A satellite is in an elliptical orbit with a perigee of 1000 km and an apogee of 4000 km. Using a
mean earth radius of 6378.14 km, find the period of the orbit in hours, minutes, and seconds, and
the eccentricity of the orbit.
Answer The major axis of the elliptical orbit is a straight line between the apogee and perigee,
as seen in Figure 2.7. Hence, for a semimajor axis length a, earth radius re, perigee height hp, and
apogee height ha,
2a 2re hp ha 2 6378.14 1000.0 4000.0 17,756.28 km
Thus the semimajor axis of the orbit has a length a 8878.14 km. Using this value of a in Eq. (2.21)
gives an orbital period T seconds where
T 2 14p 2a 3 2 m 4p2 18878.072 33.986004418 105 s2
6.930872802 107 s2
T 8325.1864 s 138 min 45.19 s 2 h 18 min 45.19 s
The eccentricity of the orbit is given by e, which can be found from Eq. (2.27) by consider-
ing the instant at which the satellite is at perigee. Referring to Figure 2.7, when the satellite is at
perigee, the eccentric anomaly E 0 and r0 re hp. From Eq. (2.27), at perigee
r0 a11 e cos E 2 and cos E 1
Hence
re hp a11 e2
e 1 1re hp 2 a 1 7,378.148878.14 0.169
2.2 LOOK ANGLE DETERMINATION
Navigation around the earth’s oceans became more precise when the surface of the globe
was divided up into a gridlike structure of orthogonal lines: latitude and longitude. Lat-
itude is the angular distance, measured in degrees, north or south of the equator and
longitude is the angular distance, measured in degrees, from a given reference
longitudinal line. At the time that this grid reference became popular, there were two
major seafaring nations vying for dominance: England and France. England drew its ref-
erence zero longitude through Greenwich, a town close to London, England, and France,
c02.qxd 13/07/02 14:10 Page 31
2.2 LOOK ANGLE DETERMINATION 31
SIDEBAR
Frequencies and orbital slots for new satellites are for a particular service is deemed to have protection
registered with the International Frequency Regis- from newcomers. Any other organization filing to
tration Board (IFRB), part of the ITU located in carry the same service at, or close to, that orbital lo-
Geneva. The initial application by an organization cation (within 2°) must coordinate their use of the
or company that wants to orbit a new satellite is frequency bands with the first organization. The first
made to the national body that controls the alloca- user may cause interference into subsequent filer’s
tion and use of radio frequencies—the FCC in the satellite systems, since they were the first to be
United States, for example—which must first ap- awarded the orbital slot and frequencies, but the later
prove the application and then forward it to the filers’ satellites must not cause interference with the
IFRB. The first organization to file with the IFRB first user’s system.
not surprisingly, drew its reference longitude through Paris, France. Since the British
Admiralty chose to give away their maps and the French decided to charge a fee for
theirs, it was not surprising that the use of Greenwich as the zero reference longitude
became dominant within a few years. [It was the start of .com market dominance through
giveaways three centuries before E-commerce!] Geometry was a much older science
than navigation and so 90° per quadrant on the map was an obvious selection to make.
Thus, there are 360° of longitude (measured from 0° at the Greenwich Meridian, the
line drawn from the North Pole to the South Pole through Greenwich, England) and
90° of latitude, plus being measured north of the equator and minus south of the equa-
tor. Latitude 90° N (or 90°) is the North Pole and latitude 90° S (or 90°) is the South
Pole. When GEO satellite systems are registered in Geneva, their (subsatellite) location
over the equator is given in degrees east to avoid confusion. Thus, the INTELSAT pri-
mary location in the Indian Ocean is registered at 60° E and the primary location in the
Atlantic Ocean is at 335.5° E (not 24.5° W). Earth stations that communicate with satel-
lites are described in terms of their geographic latitude and longitude when developing
the pointing coordinates that the earth station must use to track the apparent motion of
the satellite.
The coordinates to which an earth station antenna must be pointed to communi-
cate with a satellite are called the look angles. These are most commonly expressed
as azimuth (Az) and elevation (El), although other pairs exist. For example, right
ascension and declination are standard for radio astronomy antennas. Azimuth is
measured eastward (clockwise) from geographic north to the projection of the satellite
path on a (locally) horizontal plane at the earth station. Elevation is the angle meas-
ured upward from the local horizontal plane at the earth station to the satellite path.
Figure 2.10 illustrates these look angles. In all look angle determinations, the precise
location of the satellite is critical. A key location in many instances is the subsatellite
point.
The Subsatellite Point

The subsatellite point is the location on the surface of the earth that lies directly between
the satellite and the center of the earth. It is the nadir pointing direction from the satel-
lite and, for a satellite in an equatorial orbit, it will always be located on the equator. Since
geostationary satellites are in equatorial orbits and are designed to stay “stationary” over
c02.qxd 13/07/02 14:10 Page 32
Local vertical
Path to
satellite
El
North
Az
Projection of
path onto local
horizontal plane
East
FIGURE 2.10 The definition of elevation (El) and azimuth (Az). The elevation angle is
measured upward from the local horizontal at the earth station and the azimuth angle is
measured from true north in an eastward direction to the projection of the satellite path
onto the local horizontal plane.
the earth, it is usual to give their orbital location in terms of their subsatellite point. As
noted in the example given earlier, the Intelsat primary satellite in the Atlantic Ocean
Region (AOR) is at 335.5° E longitude. Operators of international geostationary satellite
systems that have satellites in all three ocean regions (Atlantic, Indian, and Pacific) tend
to use longitude east to describe the subsatellite points to avoid confusion between using
both east and west longitude descriptors. For U.S. geostationary satellite operators, all of
the satellites are located west of the Greenwich meridian and so it has become accepted
practice for regional systems over the United States to describe their geostationary satellite
locations in terms of degrees W.
To an observer of a satellite standing at the subsatellite point, the satellite will ap-
pear to be directly overhead, in the zenith direction from the observing location. The
zenith and nadir paths are therefore in opposite directions along the same path (see
Figure 2.11). Designers of satellite antennas reference the pointing direction of the satel-
lite’s antenna beams to the nadir direction. The communications coverage region on the
earth from a satellite is defined by angles measured from nadir at the satellite to the
edges of the coverage. Earth station antenna designers, however, do not reference their
pointing direction to zenith. As noted earlier, they use the local horizontal plane at the
earth station to define elevation angle and geographical compass points to define az-
imuth angle, thus giving the two look angles for the earth station antenna toward the
satellite (Az, El).
Elevation Angle Calculation

Figure 2.12 shows the geometry of the elevation angle calculation. In Figure 2.12, rs
is the vector from the center of the earth to the satellite; re is the vector from the cen-
ter of the earth to the earth station; and d is the vector from the earth station to the
satellite. These three vectors lie in the same plane and form a triangle. The central
angle measured between re and rs is the angle between the earth station and the
c02.qxd 13/07/02 14:10 Page 33
Nadir direction
Sub Zenith direction
FIGURE 2.11 Zenith and nadir pointing directions. The line joining the satellite and the
center of the earth, C, passes through the surface of the earth at point Sub, the subsatellite
point. The satellite is directly overhead at this point and so an observer at the subsatellite
point would see the satellite at zenith (i.e., at an elevation angle of 90°). The pointing direc-
tion from the satellite to the subsatellite point is the nadir direction from the satellite. If the
beam from the satellite antenna is to be pointed at a location on the earth that is not at the
subsatellite point, the pointing direction is defined by the angle away from nadir. In general,
two off-nadir angles are given: the number of degrees north (or south) from nadir; and the
number of degrees east (or west) from nadir. East, west, north, and south directions are
those defined by the geography of the earth.
Satellite
Local horizontal
d rs
El
FIGURE 2.12 The geometry
of elevation angle calculation.
ψ
re γ The plane of the paper is the
Earth station plane defined by the center of
Center the earth, the satellite, and the
of earth
earth station. The central angle
is . The elevation angle El is
measured upward from the
local horizontal at the earth
station.
c02.qxd 13/07/02 14:10 Page 34
satellite, and
is the angle (within the triangle) measured from re to d. Defined so that
it is nonnegative, is related to the earth station north latitude Le (i.e., Le is the num-
ber of degrees in latitude that the earth station is north from the equator) and west lon-
gitude le (i.e., le is the number of degrees in longitude that the earth station is west
from the Greenwich meridian) and the subsatellite point at north latitude Ls and west
longitude ls by
cos 1g2 cos 1Le 2 cos 1Ls 2 cos 1ls le 2 sin 1Le 2 sin 1Ls 2 (2.31)
The law of cosines allows us to relate the magnitudes of the vectors joining the cen-
ter of the earth, the satellite, and the earth station. Thus
1 2
d rs c 1 a b 2 a b cos 1g2 d
re 2 re
(2.32)
rs rs
Since the local horizontal plane at the earth station is perpendicular to re, the elevation
angle El is related to the central angle
by
El c 90° (2.33)
By the law of sines we have

rs d

sin 1c2 sin 1g2
(2.34)
Combining the last three equations yields
rs sin 1g2
cos 1El2
d
sin 1g2
1 2
(2.35)
c 1 a b 2 a b cos 1g2 d
re 2 re
rs rs
Equations (2.35) and (2.31) permit the elevation angle El to be calculated from knowl-
edge of the subsatellite point and the earth station coordinates, the orbital radius rs, and
the earth’s radius re. An accurate value for the average earth radius is 6378.137 km1 but
a common value used in approximate determinations is 6370 km.
Azimuth Angle Calculation

Because the earth station, the center of the earth, the satellite, and the subsatellite point
all lie in the same plane, the azimuth angle Az from the earth station to the satellite is the
same as the azimuth from the earth station to the subsatellite point. This is more difficult
to compute than the elevation angle because the exact geometry involved depends on
whether the subsatellite point is east or west of the earth station, and in which of the hemi-
spheres the earth station and the subsatellite point are located. The problem simplifies
somewhat for geosynchronous satellites, which will be treated in the next section. For the
general case, in particular for constellations of LEO satellites, the tedium of calculating
the individual look angles on a second-by-second basis has been considerably eased by a
range of commercial software packages that exist for predicting a variety of orbital
c02.qxd 13/07/02 14:10 Page 35
SIDEBAR
A popular suite of software employed by many the launch vehicle. Hughes used two lunar flybys to
launch service contractors is that developed by An- provide the necessary additional velocity to circu-
alytical Graphics: the Satellite Tool Kit3. The core larize the orbit at geostationary altitude. A number
program in early 2001, STK 4.0, and the subsequent of organizations offer web sites that provide orbital
subseries, was used by Hughes to rescue AsiaSat3 plots in a three-dimensional graphical format with
when that satellite was stranded in a highly ellipti- rapid updates for a variety of satellites (e.g., the
cal orbit following the failure of an upper stage in NASA site4).
dynamics and intercept solutions (see reference 13 for a brief review of 10 software
packages available in early 2001).
Specialization to Geostationary Satellites

For most geostationary satellites, the subsatellite point is on the equator at longitude ls,
and the latitude Ls is 0. The geosynchronous radius rs is 42,164.17 km1. Since Ls is zero,
Eq. (2.31) simplifies to
cos 1g2 cos 1Le 2 cos 1ls le 2 (2.36)
Substituting rs 42,164.17 km and re 6,378.137 km in Eqs. (2.32) and (2.35) gives
the following expressions for the distance d from the earth station to the satellite and the
elevation angle El at the earth station
d 42,164.1731.02288235 0.30253825 cos 1g2 4 1 2 km (2.37)
sin 1g2
cos 1El2
31.02288235 0.30253825 cos 1g2 4 12
(2.38)
For a geostationary satellite with an orbital radius of 42,164.17 km and a mean earth ra-
dius of 6378.137 km, the ratio rsre 6.6107345 giving
El tan1 3 16.6107345 cos g2 sin g4 g (2.39)
To find the azimuth angle, an intermediate angle must first be found. The inter-
mediate angle permits the correct 90° quadrant to be found for the azimuth since the
azimuthal angle can lie anywhere between 0° (true north) and clockwise through 360°
(back to true north again). The intermediate angle is found from
tan 1ls le 2
a tan1 c d
sin 1Le 2
(2.40)
Having found the intermediate angle , the azimuth look angle Az can be found from:
Case 1: Earth station in the Northern Hemisphere with
(a) Satellite to the SE of the earth station: Az 180° (2.41a)
(b) Satellite to the SW of the earth station: Az 180° (2.41b)
Case 2: Earth station in the Southern Hemisphere with
(c) Satellite to the NE of the earth station: Az (2.41c)
(d) Satellite to the NW of the earth station: Az 360° (2.41d)
c02.qxd 13/07/02 14:10 Page 36
rs
El
re
Subsatellite cos γ
point
Earth
station
re
γ
FIGURE 2.13 The geometry of the

Center visibility calculation. The satellite is said to
of earth
be visible from the earth station if the
elevation angle El is positive. This requires
that the orbital radius rs be greater than the
ratio re /cos( ) where re is the radius of the
earth and is the central angle.
Visibility Test
For a satellite to be visible from an earth station, its elevation angle El must be above
some minimum value, which is at least 0°. A positive or zero elevation angle requires that
(see Figure 2.13)
re
rs
cos 1g2
(2.42)
This means that the maximum central angular separation between the earth station and
the subsatellite point is limited by
g cos1 a b
re
(2.43)
rs
For a nominal geostationary orbit, the last equation reduces to 81.3° for the satellite
to be visible.
EXAMPLE 2.2.1 Geostationary Satellite Look Angles

An earth station situated in the Docklands of London, England, needs to calculate the look angle
to a geostationary satellite in the Indian Ocean operated by Intelsat. The details of the earth station
site and the satellite are as follows:
Earth station latitude and longitude are 52.0° N and 0°.
Satellite longitude (subsatellite point) is 66.0° E.
c02.qxd 13/07/02 14:10 Page 37
Step 1: Find the central angle

cos 1g2 cos 1Le 2 cos 1ls le 2
cos 152.02 cos 166.02 0.2504
yielding 75.4981°
The central angle is less than 81.3° so the satellite is visible from the earth station.
Step 2: Find the elevation angle El
El tan1 3 16.6107345 cos g2 sin g4 g
tan1 3 16.6107345 0.25042 sin 175.49812 4 75.4981
5.847°
Step 3: Find the intermediate angle
tan 1ls le 2
a tan1 c d
sin 1Le 2
tan1 3 1tan 166.0 022 sin 152.02 4
70.667°
Step 4: Find the azimuth angle
The earth station is in the Northern Hemisphere and the satellite is to the southeast of the
earth station. From Eq. (2.41a), this gives
Az 180° a 180 70.667 109.333°1clockwise from true north2
Note that, in the example above, the elevation angle is relatively low (5.85°).
Refractive effects in the atmosphere will cause the mean ray path to the satellite to bend
in the elevation plane (making the satellite appear to be higher in the sky than it actually
is) and to cause the amplitude of the signal to fluctuate with time. These aspects are dis-
cussed more fully in the propagation effects chapter. While it is unusual to operate to a
satellite below established elevation angle minima (typically 5° at C band, 10° at Ku band,
and in most cases, 20° at Ka band and above), many times it is not possible to do this.
Such cases exist for high latitude regions and for satellites attempting to reach extreme
east and west coverages from their given geostationary equatorial location. To establish
whether a particular satellite location can provide service into a given region, a simple
visibility test can be carried out, as shown earlier in Eqs. (2.42) and (2.43).
A number of geosynchronous orbit satellites have inclinations that are much larger
than the nominal 0.05° inclination maximum for current geosynchronous satellites. (In
general, a geosynchronous satellite with an inclination of 0.1° may be considered to
be geostationary.) In extreme cases, the inclination can be several degrees, particularly
if the orbit maneuvering fuel of the satellite is almost exhausted and the satellite’s
position in the nominal location is only controlled in longitude and not in inclination.
This happens with most geostationary communications satellites toward the end of their
operational lifetime since the reliability of the payload, or a large part of the payload,
generally exceeds that of the lifetime of the maneuvering fuel. Those satellites that can
no longer be maintained in a fully geostationary orbit, but are still used for communi-
cations services, are referred to as inclined orbit satellites. While they now need to have
tracking antennas at the earth terminals once the inclination becomes too large to allow
the satellite to remain within the 1-dB beamwidth of the earth station antennas, sub-
stantial additional revenue can be earned beyond the normal lifetime of the satellite.
Those satellites that eventually reach significantly inclined orbits can also be used to
communicate to parts of the high latitude regions that were once beyond reach, but only
c02.qxd 13/07/02 14:10 Page 38
for a limited part of the day. The exceptional reliability of electronic components in
space, once they have survived the launch and deployment sequences, has led space-
craft designers to manufacture satellites with two end-of-life criteria. These are: end of
design life (EODL), which refers to the lifetime expectancy of the payload components
and end of maneuvering life (EOML), which refers to the spacecraft bus capabilities,
in particular the anticipated lifetime of the spacecraft with full maneuver capabilities in
longitude and inclination.
Current spacecraft are designed with fuel tanks that have a capacity that usually
significantly exceeds the requirement for EODL. Once the final mass of the spacecraft
(without fuel) is known, a decision can be made as to how much additional fuel to load
so that the economics of the launch and the anticipated additional return on investment
can be balanced. Having additional fuel on board the spacecraft can be advantageous
for many reasons, in addition to adding on-orbit lifetime. In many cases, satellites are
moved to new locations during their operational lifetime. Examples for this are open-
ing up service at a new location with an older satellite or replacing a satellite that has
had catastrophic failure with a satellite from a location that has fewer customers. Each
maneuver, however, consumes fuel. A rule of thumb is that any change in orbital loca-
tion for a geostationary satellite reduces the maneuvering lifetime by about 1 month.
Moving the satellite’s location by 1° in longitude takes as much additional fuel as mov-
ing the location by 180°: both changes require an acceleration burn, a drift phase, and
a deceleration burn. The 180° location change will clearly take longer, since the drift
rates are the same in both cases. Another use for additional fuel is to allow for orbital
perturbations at any location.
2.3 ORBITAL PERTURBATIONS
The orbital equations developed in Section 2.1 modeled the earth and the satellite as point
masses influenced only by gravitational attraction. Under these ideal conditions, a “Kep-
lerian” orbit results, which is an ellipse whose properties are constant with time. In prac-
tice, the satellite and the earth respond to many other influences including asymmetry of
the earth’s gravitational field, the gravitational fields of the sun and the moon, and solar
radiation pressure. For low earth orbit satellites, atmospheric drag can also be important.
All of these interfering forces cause the true orbit to be different from a simple Kepler-
ian ellipse; if unchecked, they would cause the subsatellite point of a nominally geosyn-
chronous satellite to move with time.
Historically, much attention has been given to techniques for incorporating addi-
tional perturbing forces into orbit descriptions. The approach normally adopted for com-
munications satellites is first to derive an osculating orbit for some instant in time (the
Keplerian orbit the spacecraft would follow if all perturbing forces were removed at that
time) with orbital elements (a, e, tp, , i, ). The perturbations are assumed to cause the
orbital elements to vary with time and the orbit and satellite location at any instant are
taken from the osculating orbit calculated with orbital elements corresponding to that time.
To visualize the process, assume that the osculating orbital elements at time t0 are (a0, e0,
tp, 0, i0, 0). Then assume that the orbital elements vary linearly with time at constant
rates given by (dadt, dedt, etc.). The satellite’s position at any time t1 is then calculated
from a Keplerian orbit with elements
1t t0 2, e0 1t1 t0 2, etc.
da de
a0
dt 1 dt
c02.qxd 13/07/02 14:10 Page 39
2.3 ORBITAL PERTURBATIONS 39
This approach is particularly useful in practice because it permits the use of either
theoretically calculated derivatives or empirical values based on satellite observations.
As the perturbed orbit is not an ellipse, some care must be taken in defining the or-
bital period. Since the satellite does not return to the same point in space once per revo-
lution, the quantity most frequently specified is the so-called anomalistic period: the
elapsed time between successive perigee passages. In addition to the orbit not being a per-
fect Keplerian ellipse, there will be other influences that will cause the apparent position
of a geostationary satellite to change with time. These can be viewed as those causing
mainly longitudinal changes and those that principally affect the orbital inclination.
Longitudinal Changes:
Effects of the Earth’s Oblateness
The earth is neither a perfect sphere nor a perfect ellipse; it can be better described as a
triaxial ellipsoid1. The earth is flattened at the poles; the equatorial diameter is about 20 km
more than the average polar diameter. The equatorial radius is not constant, although the
noncircularity is small: the radius does not vary by more than about 100 m around the equa-
tor1. In addition to these nonregular features of the earth, there are regions where the
average density of the earth appears to be higher. These are referred to as regions of mass
concentration or Mascons. The nonsphericity of the earth, the noncircularity of the equa-
torial radius, and the Mascons lead to a nonuniform gravitational field around the earth.
The force on an orbiting satellite will therefore vary with position.
For a low earth orbit satellite, the rapid change in position of the satellite with re-
spect to the earth’s surface will lead to an averaging out of the perturbing forces in line
with the orbital velocity vector. The same is not true for a geostationary (or geosynchro-
nous) satellite. A geostationary satellite is weightless when in orbit. The smallest force on
the satellite will cause it to accelerate and then drift away from its nominal location. The
satellite is required to maintain a constant longitudinal position over the equator, but there
will generally be an additional force toward the nearest equatorial bulge in either an east-
ward or a westward direction along the orbit plane. Since this will rarely be in line with
the main gravitational force toward the earth’s center, there will be a resultant component
of force acting in the same direction as the satellite’s velocity vector or against it, depending
on the precise position of the satellite in the GEO orbit. This will lead to a resultant ac-
celeration or deceleration component that varies with longitudinal location of the satellite.
Due to the position of the Mascons and equatorial bulges, there are four equilib-
rium points in the geostationary orbit: two of them stable and two unstable. The stable
points are analogous to the bottom of a valley, and the unstable points to the top of a hill.
If a ball is perched on top of a hill, a small push will cause it to roll down the slope into
a valley, where it will roll backwards and forwards until it gradually comes to a final stop
at the lowest point. The satellite at an unstable orbital location is at the top of a gravity
hill. Given a small force, it will drift down the gravity slope into the gravity well (valley)
and finally stay there, at the stable position. The stable points are at about 75° E and
252° E and the unstable points are at around 162° E and 348° E1. If a satellite is perturbed
slightly from one of the stable points, it will tend to drift back to the stable point without
any thruster firings required. A satellite that is perturbed slightly from one of the unstable
points will immediately begin to accelerate its drift toward the nearer stable point and,
once it reaches this point, it will oscillate in longitudinal position about this point until
(centuries later) it stabilizes at that point. These stable points are sometimes called the
graveyard geosynchronous orbit locations (not to be confused with the graveyard orbit for
a geosynchronous satellite, which is the orbit to which the satellite is raised once the
c02.qxd 13/07/02 14:10 Page 40
satellite ceases to be useful). Note that, due to the nonsphericity of the earth, etc., the sta-
ble points are neither exactly 180° apart, nor are the stable and unstable points precisely
90° apart.
Inclination Changes:
Effects of the Sun and the Moon
The plane of the earth’s orbit around the sun—the ecliptic—is at an inclination of 7.3° to
the equatorial plane of the sun (Figure 2.14). The earth is titled about 23° away from the
normal to the ecliptic. The moon circles the earth with an inclination of around 5° to the
equatorial plane of the earth. Due to the fact that the various planes—the sun’s equator,
the ecliptic, the earth’s equator (a plane normal to the earth’s rotational axis), and the
moon’s orbital plane around the earth—are all different, a satellite in orbit around the earth
will be subjected to a variety of out-of-plane forces. That is, there will generally be a net
acceleration force that is not in the plane of the satellite’s orbit, and this will tend to try
to change the inclination of the satellite’s orbit from its initial inclination. Under these
conditions, the orbit will precess and its inclination will change.
The mass of the sun is significantly larger than that of the moon but the moon is con-
siderably closer to the earth than the sun (see Table 2.2). For this reason, the acceleration
force induced by the moon on a geostationary satellite is about twice as large as that of the
sun. The net effect of the acceleration forces induced by the moon and the sun on a
Moon
Earth’s
5° Eq. plane
Earth
Satellite
23°
7.3°
Sun’s
Eq. plane
Sun
FIGURE 2.14 Relationship between the orbital planes of the sun, moon, and earth. The
plane of the earth’s orbit around the sun is the ecliptic. The geostationary orbit plane (the
earth’s equatorial plane) is about 23° out of the ecliptic, and leads to maximum out-of-
geostationary-orbit-plane forces at the solstice periods (approximately June 21 and
December 21). The orbit of the moon is inclined about 5° to the earth’s equatorial plane. The
moon revolves around the earth in 27.3 days, the earth (and the geostationary satellite)
rotates once about 24 h, and the earth revolves around the sun every 365.25 days. In
addition, the sun—which has a greater girth at the equator than at the poles—has its equator
inclined about 7.3° to the ecliptic. All of these various angular differences and orbital periods
lead to conditions where all of the out-of-plane gravitational forces are in one direction with
respect to the equatorial (geostationary orbital) plane at a given time as well as to conditions
where the various gravitational out-of-plane forces partially cancel each other out. The
precessional forces that cause the inclination of the geostationary satellite’s orbit to move
away from the equatorial plane therefore vary with time.
c02.qxd 13/07/02 14:10 Page 41
2.3 ORBITAL PERTURBATIONS 41
TABLE 2.2 Comparative Data for the Sun, Moon, and Earth
Mean radius Mass Mean orbit radius Spin period
Sun 696,000 km 333,432 units 30,000 light years 25.04 earth days
Moon 3,476 km 0.012 units 384,500 km 27.3 earth days
Earth 6,378.14 km 1.0 units 149,597,870 km 1 earth day
The orbit radius refers to the center of the home galaxy (Milky Way) for the sun, center
of earth for the moon, and center of the sun for the earth, respectively.
geostationary satellite is to change the plane of the orbit at an initial average rate of change
of 0.85°/year from the equatorial plane1.
When both the sun and moon are acting on the same side of the satellite’s orbit, the
rate of change of the plane of the geostationary satellite’s orbit will be higher than aver-
age. When they are on opposite sides of the orbit, the rate of change of the plane of the
satellite’s orbit will be less than average. Examples of maximum years are 1988 and 2006
(0.94°/year) and examples of minimum years are 1997 and 2015 (0.75°/year)1. These rates
of change are neither constant with time nor with inclination. They are at a maximum
when the inclination is zero and they are zero when the inclination is 14.67°. From an ini-
tial zero inclination, the plane of the geostationary orbit will change to a maximum in-
clination of 14.67° over 26.6 years. The acceleration forces will then change direction at
this maximum inclination and the orbit inclination will move back to zero in another 26.6
years and out to 14.67° over a further 26.6 years, and so on.
In some cases, to increase the orbital maneuver lifetime of a satellite for a given
fuel load, mission planners deliberately place a satellite planned for geostationary orbit
into an initial orbit with an inclination that is substantially larger than the nominal 0.05°
for a geostationary satellite. The launch is specifically timed, however, so as to set up the
necessary precessional forces that will automatically reduce the inclination “error” to close
to zero over the required period without the use of any thruster firings on the spacecraft.
This will increase the maneuvering lifetime of the satellite at the expense of requiring
greater tracking by the larger earth terminals accessing the satellite for the first year or so
of the satellite’s operational life.
Under normal operations, ground controllers command spacecraft maneuvers to
correct for both the in-plane changes (longitudinal drifts) and out-of-plane changes (in-
clination changes) of a satellite so that it remains in the correct orbit. For a geostationary
satellite, this means that the inclination, ellipticity, and longitudinal position are controlled
so that the satellite appears to stay within a “box” in the sky that is bounded by 0.05°
in latitude and longitude over the subsatellite point. Some maneuvers are designed to cor-
rect for both inclination and longitude drifts simultaneously in the one burn of the ma-
neuvering rockets on the satellite. In others, the two maneuvers are kept separate: one
burn will correct for ellipticity and longitude drift; another will correct for inclination
changes. The latter situation of separated maneuvers is becoming more common for two
reasons. The first is due to the much larger velocity increment needed to change the plane
of an orbit (the so-called north–south maneuver) as compared with the longitude/ellipticity
of an orbit (the so-called east–west maneuver). The difference in energy requirement is
about 10 :1. By alternately correcting for inclination changes and in-plane changes, the
attitude of the satellite can be held constant and different sets of thrusters exercised for
the required maneuver.
The second reason is the increasing use of two completely different types of thrusters
to control N–S maneuvers on the one hand and E–W maneuvers on the other. In the
c02.qxd 13/07/02 14:10 Page 42
mid-1990s, one of the heaviest items that was carried into orbit on a large satellite was
the fuel to raise and control the orbit. About 90% of this fuel load, once on orbit, was to
control the inclination of the satellite. Newer rocket motors, particularly arc jets and ion
thrusters, offer increased efficiency with lighter mass. In general, these low thrust, high
efficiency rocket motors are used for N–S maneuvers leaving the liquid propellant thrusters,
with their inherently higher thrust (but lower efficiency) for orbit raising and in-plane
changes. In order to be able to calculate the required orbit maneuver for a given satellite,
the controllers must have an accurate knowledge of the satellite’s orbit. Orbit determina-
tion is a major aspect of satellite control.
EXAMPLE 2.3.1 Drift with a Geostationary Satellite
A quasi-GEO satellite is in a circular equatorial orbit close to geosynchronous altitude. The quasi-
GEO satellite, however, does not have a period of one sidereal day: its orbital period is exactly
24 h—one solar day. Calculate
(i) the radius of the orbit
(ii) the rate of drift around the equator of the subsatellite point in degrees per (solar) day.
An observer on the earth sees that the satellite is drifting across the sky.
(iii) Is the satellite moving toward the east or toward the west?
Answer Part (i) The orbital radius is found from Eq. (2.21), as in worked Example 2.2.1. Equa-
tion (2.21) gives the square of the orbital period in seconds (remembering that T here is one solar day)
T 2 14p 2a3 2 m
Rearranging the equation, the orbital radius a is given by
a3 T 2m 14p2 2 186,4002 2 3.986004418 1054p2
7.5371216 1013 km3
a 42,241.095 km
Part (ii) The orbital period of the satellite (one solar day) is longer than a sidereal day by 3 min
55.9 s 235.9 s. This will cause the subsatellite point to drift at a rate of 360° 235.986400 per
day or 0.983° per day.
Part (iii) The earth moves toward the east at a faster rate than the satellite, so the drift will appear
to an observer on the earth to be toward the west.
2.4 ORBIT DETERMINATION
Orbit determination requires that sufficient measurements be made to determine uniquely

the six orbital elements needed to calculate the future orbit of the satellite, and hence cal-
culate the required changes that need to be made to the orbit to keep it within the nomi-
nal orbital location. Three angular position measurements are needed because there are
six unknowns and each measurement will provide two equations. Conceptually, these can
be thought of as one equation giving the azimuth and the other the elevation as a function
of the six (as yet unknown) orbital elements.
The control earth stations used to measure the angular position of the satellites also
carry out range measurements using unique time stamps in the telemetry stream or com-
munications carrier. These earth stations are generally referred to as the TTC&M (teleme-
try tracking command and monitoring) stations of the satellite network. Major satellite
networks maintain their own TTC&M stations around the world. Smaller satellite systems
generally contract for such TTC&M functions from the spacecraft manufacturer or from
c02.qxd 13/07/02 14:10 Page 43
2.5 LAUNCHES AND LAUNCH VEHICLES 43
the larger satellite system operators, as it is generally uneconomic to build advanced

TTC&M stations with fewer than three satellites to control. Chapter 3 discusses TTC&M
systems.
2.5 LAUNCHES AND LAUNCH VEHICLES
A satellite cannot be placed into a stable orbit unless two parameters that are uniquely
coupled together—the velocity vector and the orbital height—are simultaneously correct.
There is little point in obtaining the correct height and not having the appropriate veloc-
ity component in the correct direction to achieve the desired orbit. A geostationary satel-
lite, for example, must be in an orbit at a height of 35,786.03 km above the surface of the
earth (42,164.17-km radius from the center of the earth) with an inclination of zero de-
grees, an ellipticity of zero, and a velocity of 3074.7 m/s tangential to the earth in the
plane of the orbit, which is the earth’s equatorial plane. The further out from the earth the
orbit is, the greater the energy required from the launch vehicle to reach that orbit. In any
earth satellite launch, the largest fraction of the energy expended by the rocket is used to
accelerate the vehicle from rest until it is about 20 miles (32 km) above the earth. To make
the most efficient use of the fuel, it is common to shed excess mass from the launcher as
it moves upward on launch: this is called staging. Figure 2.15 gives a schematic of a Proton
launch from the Russian Baikonur complex at Kazakhstan, near Tyuratam.
Most launch vehicles have multiple stages and, as each stage is completed, that por-
tion of the launcher is expended until the final stage places the satellite into the desired tra-
jectory. Hence the term: expendable launch vehicle (ELV). The Space Shuttle, called the
Space Transportation System (STS) by NASA, is partially reusable. The solid rocket boosters
are recovered and refurbished for future missions and the shuttle vehicle itself is flown
back to earth for refurbishment and reuse. Hence the term: reusable launch vehicle (RLV)
for such launchers. More advanced launch vehicles are being developed that would pro-
vide both single stage to orbit (SSTO) and RLV capabilities. The NASA series of X-33 and
X-34 test vehicles form the public portion of this quest (see the NASA home page4).
25:00 4th
10:00 3rd stage stage
separation roll/allign
05:41 2nd stage

separation
06:10 3rd stage
05:34 Payload ignition
fairing jettison
Major Events from GTO to final User Handoff

00:21 Roll
1:27:00 Reach GTO
02:07 1st stage
3:59:10 Completion of programed turns
Separation/
6:58:00 Completion of the compensation turn
2nd stage ignition 7:09:20 Second 4th stage ignition (2 sec)
7:09:50 Spacecraft separates from 4th stage, GEO
7:10+ Handoff to User
Lift-off
FIGURE 2.15 Schematic of a Proton launch (after reference 5).

c02.qxd 13/07/02 14:10 Page 44
There are also a number of private ventures that aim to achieve RLV capabilities in
the first decade of the twenty-first century. Two excellent web sites to keep abreast of
these, and related space issues, are those maintained by Spaceviews6 and Orbreport7. Of
equal importance to the orbital height the satellite is intended for is the inclination of the
orbit that the spacecraft needs to be launched into.
The earth spins toward the east. At the equator, the rotational velocity of a sea level
site in the plane of the equator is (2 radius of the earth)(one sidereal day) 0.4651
km/s. This velocity increment is approximately 1000 mph (1610 km/h). An easterly
launch from the equator has a velocity increment of 0.465 km/s imparted by the rotation
of the earth. A satellite in a circular, equatorial orbit at an altitude of 900 km requires an
orbital velocity of about 7.4 km/s tangential to the surface of the earth. A rocket launched
from the equator needs to impart an additional velocity of (7.4 0.47) km/s 6.93 km/s:
in other words, the equatorial launch has reduced the energy required by about 6%. This
equatorial launch “bonus” led to the concept of a sea launch by Hughes and Boeing8. If
the launch is not to be into an equatorial orbit, the payload capabilities of any given rocket
will reduce as the inclination increases.
A satellite launched into a prograde orbit from a latitude of degrees will enter an
orbit with an inclination of degrees to the equator. If the satellite is intended for geo-
stationary orbit, the satellite must be given a significant velocity increment to reorient the
orbit into the earth’s equatorial plane. For example, a satellite launched from Cape
Canaveral at 28.5° N latitude requires a velocity increment of 366 m/s to attain an equa-
torial orbit from a geosynchronous orbit plane of 28.5°. Ariane is launched from the Guiana
Space Center in French Guiana, located at latitude 5° S in South America, and SeaLaunch
can launch from the equator. The lower latitude of these launch sites results in significant
savings in the fuel used by the apogee kick motor (AKM).
Expendable Launch Vehicles (ELVs)

1998 was an important year for ELVs: it was the year when the number of commercial
launches in the United States surpassed the number of government launches for the first
time9. The gap between commercial and government launches will continue to grow. The
Teal Group estimated in mid-1999 that 1447 satellites would be launched worldwide be-
tween 2000 and 2009 on 850 to 900 launch vehicles10. At an average cost of $100 M per
launch, this represents a business worth about $ 90 B over 10 years. Of these 1447 satel-
lites, 893 were considered commercial ventures with the remainder split between military
and civilian government spacecraft. There is therefore a healthy market for ELVs and a
number of companies, consortia, and national entities are seeking to enter this expanding
field. Reference 15 contains a good survey of the ELVs being developed for the twenty-
first century. Figure 2.16 presents a rough comparison between the main launch vehicles
SIDEBAR
The STS can launch approximately 65,000 lb. (29,478 shuttle is rarely used to launch civilian payloads, its
kg) into a standard 28.5° orbital inclination at an or- mission being confined to military payloads [e.g.,
bital height of about 200 km from the Kennedy Space TDRSS (Tracking and Data Relay Satellite System)
Flight Center in Cape Canaveral. If the Vandenburg satellites], joint ventures with other agencies [e.g.,
Air Force Base launch site in California still had the ESA (European Space Agency) Spacelab facility],
capability of launching the Shuttle, the payload ca- “big science” missions (e.g., the X-ray telescope
pability of the Shuttle for a polar launch (inclination Chandra), and International Space Station flights. The
90°) would be reduced to 32,000 lb (14,512 kg). vast majority of the satellite launches are therefore
Since the Challenger accident in January 1996, the conducted by expendable launch vehicles.
1996 Market $50 $60 $80 $90 $75 $95 $105 $85 $95 $110 – $120
Prices, $M
16 16
c02.qxd 13/07/02 14:10 Page 45
LAUNCH 14 14
VEHICLES
12 12
PERFORMANCE
× 1000 Lb
to GTO (i = 7°) 10 10
8 8
6 6
4 4
2 2
DELTA II-7925 ARIANE 40 ATLAS II ATLAS IIA ARIANE 42P ARIANE 44P ATLAS IIAS ARIANE 42L ARIANE 44LP ARIANE 44L TITAN III ARIANE 5
First Flight 1990 1990 1991 1992 1991 1991 1993 1993 1988 1989 1989 1996
Performance Kg (lb)
GTO, i = 7 deg 1,820 (1) 2,050 2,810 (2) 3,050 (2) 2,840 3,320 3,700 (2) 3,380 4,060 4,520 5,000 (1) 5,000-6,700
200 × 35,786 Km
(108 × 19,323 NM)
(4,010) (4,520) (6,200) (6,710) (6,260) (7,320) (8,150) (7,450) (8,950) (9,965) (11,000) (11-13,900)
Payload Max 2.54 3.65 3.65 3.65 3.65 3.65 3.65 3.65 3.65 3.65 3.65 4.57
Diameter m (ft) (8.3) (12) (12) (12) (12) (12) (12) (12) (12) (12) (12) (15)
Launch Site CCAS KOUROU CCAS CCAS KOUROU KOUROU CCAS KOUROU KOUROU KOUROU CCAS KOUROU
(Projected) (VAFB) (VAFB) (VAFB) (VAFB)
(1) DELTA & TITAN @ i = 28 deg

(2) ATLAS GTO REFERENCE ORBIT IS 167 × 35,788 KM (90 × 19,324 NM), i = 27 deg
FIGURE 2.16 Representative ELVs (after reference 5). CCAS, Cape Canaveral Air Station; VAFB, Vandenburg Air Force Base.
45
c02.qxd 13/07/02 14:10 Page 46
$M
200
H-2
180
160
140
Atlas IIAR
120 Atlas IIAS 13.9
Atlas IIA Ariane 5
100 Atlas II Ariane 44L
80 Delta III
Zenit 3
60
Long March 3B
Proton D-1-e
40 Delta II
20
0
4 6 8 10 12
Pounds into Geostationary Transfer Orbit
FIGURE 2.17 Launch vehicle market price vs performance, 1996 prices (after reference 5).
The launch vehicles have been normalized to a launch into geostationary transfer orbit at an
inclination of 28°. The trend line for launchers is shown as $12,000 per pound. Note that
Long March, Zenit, and Proton are well below this trend line, mainly due to aggressive pric-
ing objectives to break into a market long dominated by U.S. and European launchers.
used for Geostationary Transfer Orbit (GTO) injection during the 1990s, plus the Ariane 5
launcher. The 1996 pricing of these vehicles is shown in Figure 2.17. Not included in these
data are the advanced Chinese launch vehicles being developed for both unmanned and
manned missions in the twenty-first century. The largest of these Chinese launch vehicles
rivals the Ariane 5 vehicle with a geostationary transfer orbit capability of 26,000 lb.
TABLE 2.3 Some Next Generation Launchers Compared with Ariane 44

and Atlas IIAS Baseline Vehicles (1999 Prices)
Weight to Total cost Lead time Max. payload Launch

Launcher orbit (kg) ($M) (months) diameter (m) latitude (°)
Ariane 44 4000 130 36 3.65 5.2

Ariane 5 6800 120 36 4.57 5.2
Atlas IIAS 3700 100 36 3.45 28.5
Atlas IIIA 4120 125 36 4.19 28.5
Atlas IIIB 4500 135 48 4.19 28.5
Atlas V 6500 150* 48 5.40 28.5
Delta III 3800 130 36 4.00 28.7
Delta IV(small) 2177 60* 36* 3.00 28.7
Delta IV(med.) 4173 120* 36* 4.00 28.7
Delta IV(heavy) 13200 400* 48* 5.00 28.7
Titan III 4500 260 36 3.65 28.6
Titan IV 5700 435 48 4.57 28.6
Proton M 4800 80 24 3.68 51.6
* These data are estimated values.

The Atlas V and Proton M vehicles are planned for operational flights beginning in 2002 or
2003. The Delta IV family of launch vehicles will become operational from 2002 to 2004.
c02.qxd 13/07/02 14:10 Page 47
2.5 LAUNCHES AND LAUNCH VEHICLES 47
TABLE 2.4 Some Launch Vehicle Selection Factors
Price/cost
Reliability
Recent launch success/failure history
Dependable launch schedule
Urgency of your launch requirement
Performance
Spacecraft fit to launcher (size, acoustic, and vibration environment)
Flight proven (see recent launch history)
Safety issues
Launch site location
Availability
What is the launcher backlog of orders?
What is the launch site backlog of launches?
Market issues
What will the market bear at this particular time?
It can be seen from Figure 2.17 that there was a well-established trend line of about
$25,000 per kg into GTO prior to the introduction of the Chinese Long March and the
Russian Zenit and Proton vehicles. The pricing of the Chinese and Russian launchers re-
flected an aggressive marketing strategy to break into the launch services field. Ariane 5
was the first of the next-generation launchers aimed at both large, single payloads into
GTO and multiple payload injection into LEO and MEO. Some more next-generation
launchers are shown in Table 2.3 on the previous page. It is anticipated that the bulk of
the large satellite launches will be conducted with Atlas V, Delta IV, and Ariane vehicles
and their Russian and Chinese equivalents over the first 2 decades of the twenty-first cen-
tury. The decision on which particular rocket to use in a given situation will depend on a
variety of factors. Some of these are set out in Table 2.4.
The decision-making routine using the above criteria is shown in Figure 2.18.
Launch Vehicle Selection Factors
• Cost to manufacturer
• “Performance”, or throw-weight to orbit
• Reliability
• Price/cost • Schedule dependability
• Reliability • Market forces
– Recent failures • Insurance
• Dependable launch schedule
– Urgency of the customer
• Performance
• Spacecraft fit
• Flight proven
• Safety
• Launch site location
• Availability—Launch site; vehicle; schedule;
• Market conditions—What the market will bear
FIGURE 2.18 Schematic of the decision making process to select a rocket for a
given satellite requirement (after reference 5).
c02.qxd 13/07/02 14:10 Page 48
FIGURE 2.19 Illustration of the

Apogee: AKM fires at this point
GTO/AKM approach to geosta-
tionary orbit (not to scale). The
combined spacecraft and final
rocket stage are placed into low
earth orbit (LEO) around the
earth. After careful orbit determi-
GEO nation measurements, the final
LEO GTO
stage is ignited in LEO and the
spacecraft inserted into a transfer
orbit that lies between the LEO
and the geostationary orbit alti-
tude: the so-called geostationary
transfer orbit or GTO. Again, af-
ter more careful orbit determina-
Perigee: tion, the apogee kick motor
GTO insertion starts here (AKM) is fired on the satellite
and the orbit is both circularized
at geostationary altitude and the
inclination reduced to close to
zero. The satellite is then in GEO.
Some of the launch vehicles deliver the spacecraft directly to geostationary orbit
(called a direct-insertion launch) while others inject the spacecraft into a geostationary
transfer orbit (GTO). Spacecraft launched into GTO must carry additional rocket motors
and/or propellant to enable the vehicle to reach the geostationary orbit. There are three
basic ways to achieve geostationary orbit.
Placing Satellites into Geostationary Orbit

Geostationary Transfer Orbit and AKM The initial approach to launching geo-
stationary satellites was to place the spacecraft, with the final rocket stage still attached,
into low earth orbit. After a couple of orbits, during which the orbital elements are measured,
the final stage is reignited and the spacecraft is launched into a geostationary transfer or-
bit. The GTO has a perigee that is the original LEO orbit altitude and an apogee that is the
GEO altitude. Figure 2.19 illustrates the process. The position of the apogee point is close
to the orbital longitude that would be the in-orbit test location of the satellite prior to it be-
ing moved to its operational position. Again, after a few orbits in the GTO while the or-
bital elements are measured, a rocket motor (usually contained within the satellite itself)
is ignited at apogee and the GTO is raised until it is a circular, geostationary orbit. Since
the rocket motor fires at apogee, it is commonly referred to as the apogee kick motor
SIDEBAR
The first successful GEO satellite was Syncom, lite’s deployable elements (e.g., solar panels, an-
launched in 1963. Hughes Corporation built the tennas) were stowed and locked in place to avoid
satellite and the spacecraft was spin-stabilized damage while the AKM accelerated the satellite to
while it was in geostationary transfer orbit. In this GEO. Hughes patented the technique of spin stabi-
way, the satellite was correctly aligned for the lizing the spacecraft in GTO. To avoid infringing
apogee motor firing. The apogee motor was fairly this patent, other satellite manufacturers developed
powerful and the apogee burn was only for a few a new way to achieve GEO, known as a slow orbit
minutes. During this apogee burn, all of the satel- raising technique.
c02.qxd 13/07/02 14:10 Page 49
2.6 ORBITAL EFFECTS IN COMMUNICATIONS SYSTEMS PERFORMANCE 49
FIGURE 2.20 Illustration of slow orbit

raising to geostationary orbit (not to scale).
The combined spacecraft and final rocket
stage are placed into low earth orbit (LEO)
GTO around the earth. As before (see Figure
2.19), the spacecraft is injected into GTO but,
LEO in this case, once the satellite is ejected from
the final rocket stage, it deploys many of the
elements that it will later use in GEO (solar
panels, etc.) and stabilizes its attitude using
thrusters and momentum wheels, rather
Successive
than being spin-stabilized. The higher power
orbit raisings
from GTO thrusters are then used around the apogee
to raise the perigee of the orbit until the
orbit is circular at the GEO altitude. At the
same time as the orbit is being raised, the
thruster firings will be designed gradually to
GEO
reduce the inclination to close to zero.
(AKM). The AKM is used both to circularize the orbit at GEO and to remove any incli-
nation error so that the final orbit of the satellite is very close to geostationary.
Geostationary Transfer Orbit with Slow Orbit Raising In this procedure,

rather than employ an apogee kick motor that imparts a vigorous acceleration over a few min-
utes, the spacecraft thrusters are used to raise the orbit from GTO to GEO over a number of
burns. Since the spacecraft cannot be spin-stabilized during the GTO (so as not to infringe
the Hughes patent), many of the satellite elements are deployed while in GTO, including the
solar panels. The satellite has two power levels of thrusters: one for more powerful orbit rais-
ing maneuvers and one for on-orbit (low thrust) maneuvers. Since the thrusters take many
hours of operation to achieve the geostationary orbit, the perigee of the orbit is gradually
raised over successive thruster firings. The thruster firings occur symmetrically about the
apogee although they could occur at the perigee as well. The burns are typically 60 to 80 min
long on successive orbits and up to six orbits can be used. Figure 2.20 illustrates the process.
In the first two cases, AKM and slow orbit raising, the GTO may be a modified or-
bit with the apogee well above the required altitude for GEO. The excess energy of the
orbit due to the higher-than-necessary altitude at apogee can be traded for energy required
to raise the perigee. The net energy to circularize the orbit at GEO is therefore less and
the satellite can retain more fuel for on-orbit operations.
Direct Insertion to GEO This is similar to the GTO technique but, in this case,
the launch service provider contracts to place the satellite into GEO. The final stages of
the rocket are used to place the satellite directly into GEO rather than the satellite using
its own propulsion system to go from GTO to GEO.
2.6 ORBITAL EFFECTS IN COMMUNICATIONS

SYSTEMS PERFORMANCE
Doppler Shift
To a stationary observer, the frequency of a moving radio transmitter varies with the trans-
mitter’s velocity relative to the observer. If the true transmitter frequency (i.e., the
c02.qxd 13/07/02 14:10 Page 50
frequency that the transmitter would send when at rest) is fT, the received frequency fR is
higher than fT when the transmitter is moving toward the receiver and lower than fT when
the transmitter is moving away from the receiver. Mathematically, the relationship [Eq.
(2.44a)] between the transmitted and received frequencies is
fR fT ¢f VT
(2.44a)
fT fT vp
or
¢f VT fTc VTl (2.44b)
where VT is the component of the transmitter velocity directed toward the receiver, vp c
the phase velocity of light (2.9979 108 3 108 m/s in free space), and is the wave-
length of the transmitted signal. If the transmitter is moving away from the receiver, then
VT is negative. This change in frequency is called the Doppler shift, the Doppler effect, or
more commonly just “Doppler” after the German physicist who first studied the phenome-
non in sound waves. For LEO satellites, Doppler shift can be quite pronounced, requiring
the use of frequency-tracking receivers. For geostationary satellites, the effect is negligible.
EXAMPLE 2.6.1 Doppler Shift for a LEO Satellite

A low earth orbit satellite is in a circular polar orbit with an altitude, h, of 1000 km. A transmitter
on the satellite has a frequency of 2.65 GHz. Find
(i) The velocity of the satellite in orbit
(ii) The component of velocity toward an observer at an earth station as the satellite appears over
the horizon, for an observer who is in the plane of the satellite orbit.
(iii) Hence, find the Doppler shift of the received signal at the earth station. Use a mean earth ra-
dius value, re, of 6378 km.
The satellite also carries a Ka-band transmitter at 20.0 GHz.
(iv) Find the Doppler shift for this signal when it is received by the same observer.
Answer Part (i) The period of the satellite is found from Eq. (2.21):
T 2 14p2a3 2 m
T 2 4p 2 16378 10002 33.986004418 10 5
3.977754 107 s2
T 6306.94 s
The circumference of the orbit is 2a 46,357.3 km so the velocity of the satellite in orbit is vs
where
vs 46,357.36306.94 7.350 km/s
Part (ii) The component of velocity toward an observer in the plane of the orbit as the satellite ap-
pears over the horizon is given by vr vs cos , where is the angle between the satellite velocity
vector and the direction of the observer at the satellite. The angle can be found from simple geom-
etry to be
cos u re 1re h2 63787378 0.8645
Hence the component of satellite velocity toward the observer is
vr vs cos u 6.354 km/s 6354 m/s
Part (iii) The Doppler shift of the received signal is given by Eq. (2.44b). Hence, for this satel-
lite and observer, with a transmitter frequency of 2.65 GHz, 0.1132 m, and the Doppler shift
c02.qxd 13/07/02 14:10 Page 51
in the received signal is

¢f VTl 63540.1132 56,130 Hz 56.130 kHz
Part (iv) A Ka-band transmitter with frequency 20.0 GHz has a wavelength of 0.015 m. The cor-
responding Doppler shift at the receiver is
¢f VTl 63540.015 423.60 kHz
Doppler shift at Ka band with a LEO satellite can be very large and requires a fast frequency-
tracking receiver. Ka-band LEO satellites are better suited to wideband signals than narrowband
voice communications.
Range Variations
Even with the best station-keeping systems available for geostationary satellites, the position
of a satellite with respect to the earth exhibits a cyclic daily variation. The variation in po-
sition will lead to a variation in range between the satellite and user terminals. If time di-
vision multiple access (TDMA) is being used, careful attention must be paid to the timing
of the frames within the TDMA bursts (see Chapter 6) so that the individual user frames
arrive at the satellite in the correct sequence and at the correct time. Range variations on
LEO satellites can be significant, as can path loss variations. While guard times between
bursts can be increased to help in any range and/or timing inaccuracies, this reduces the
capacity of the transponder. The on-board capabilities of some satellites permit both tim-
ing control of the burst sequence and power level control of individual user streams.
Solar Eclipse
A satellite is said to be in eclipse when the earth prevents sunlight from reaching it, that
is, when the satellite is in the shadow of the earth. For geostationary satellites, eclipses
occur during two periods that begin 23 days before the equinoxes (about March 21 and
about September 23) and end 23 days after the equinox periods. Figure 2.21 from refer-
ence 11 and Figure 2.22 from reference 12 illustrate the geometry and duration of the
eclipses. Eclipses occur close to the equinoxes, as these are the times when the sun, the
earth, and the satellite are all nearly in the same plane.
During full eclipse, a satellite receives no power from its solar array and it must op-
erate entirely from its batteries. Batteries are designed to operate with a maximum depth
of discharge; the better the battery, the lower the percentage depth of discharge can be. If
the battery is discharged below its maximum depth of discharge, the battery may not re-
cover to full operational capacity once recharged. The depth of discharge therefore sets the
power drain limit during eclipse operations. Nickel–Hydrogen batteries, long the mainstay
of communications satellites, can operate at about a 70% depth of discharge and recover
fully once recharged. Ground controllers perform battery-conditioning routines prior to
eclipse operations to ensure the best battery performance during the eclipse. The routines
consist of deliberately discharging the batteries until they are close to their maximum depth
of discharge, and then fully recharging the batteries just before eclipse season begins.
The eclipse season is a design challenge for spacecraft builders. Not only is the main
power source withdrawn (the sun) but also the rapidity with which the satellite enters and
exits the shadow can cause extreme changes in both power and heating effects over rela-
tively short periods. Just like a common light bulb is more likely to fail when the current
is switched on as opposed to when it is under steady state conditions, satellites can suf-
fer many of their component failures under sudden stress situations. Eclipse periods are
c02.qxd 13/07/02 14:10 Page 52
Satellite in sun transit outage

Satellite in eclipse
Sun Earth
shadow
Earth R
C
S
A
Sun
-ray
D at e direct
quin ion
ox
Geostationary orbit
FIGURE 2.21 Eclipse geometry (Source: J. J. Spilker, Jr., Digital Communications by
Satellite, Prentice Hall, p. 144, copyright © 1977, Pearson Education, Upper Saddle River,
NJ, reprinted with permission). During the equinox periods around the March 21 and
September 23, the geostationary plane is in the shadow of the earth on the far side of the
earth from the sun. As the satellite moves around the geostationary orbit, it will pass
through the shadow and undergo an eclipse period. The length of the eclipse period will
vary from a few minutes to over an hour (see Figure 2.22), depending on how close the
plane of the geostationary orbit is with respect to the center of the shadow thrown by
the earth.
Day of the year Day of the year

60 70 80 90 100 110 240 250 260 270 280 290
80 Full 80
shadow
70 Half 70
shadow
60 60
Eclipse time (min.)
Eclipse time (min.)
50 50
40 40
30 30
20 20
10 10
0 0
1 11 21 31 10 20 17 27 7 17 28 7
March April September October August
Date Date
FIGURE 2.22 Dates and duration of eclipses. (Source: James Martin, Communications
Satellite Systems, Prentice Hall, p. 37, copyright © 1978 Pearson Education, Upper Saddle
River, NJ. Reprinted with permission.)
c02.qxd 13/07/02 14:10 Page 53
therefore monitored carefully by ground controllers, as this is when most of the equip-
ment failures are likely to occur.
Sun Transit Outage

During the equinox periods, not only does the satellite pass through the earth’s shadow
on the “dark” side of the earth, but the orbit of the satellite will also pass directly in front
of the sun on the sunlit side of the earth (Figure 2.23). The sun is a “hot” microwave
source with an equivalent temperature of about 6000 to 10,000 K, depending on the time
within the 11-year sunspot cycle, at the frequencies used by communications satellites
(4 to 50 GHz). The earth station antenna will therefore receive not only the signal from
the satellite but also the noise temperature transmitted by the sun. The added noise
temperature will cause the fade margin of the receiver to be exceeded and an outage will
occur. These outages may be precisely predicted. For satellite system operators with more
The sun
Geostationary
orbit
Communications
signal
Thermal noise from

the sun
Earth station
FIGURE 2.23 Schematic of sun outage conditions. During the equinox periods, not only
does the earth’s shadow cause eclipse periods to occur for geostationary satellites, during
the sunlit portion of the orbit, there will be periods when the sun appears to be directly
behind the satellite. At the frequencies used by communications satellites (4 to 50 GHz), the
sun appears as a hot noise source. The effective temperature of the sun at these frequen-
cies is on the order of 10,000 K. The precise temperature observed by the earth station an-
tenna will depend on whether the beamwidth partially, or completely, encloses the sun.
c02.qxd 13/07/02 14:10 Page 54
than one satellite at their disposal, traffic can be off-loaded to satellites that are just out
of, or are yet to enter, a sun outage. The outage in this situation can therefore be limited
as far as an individual user is concerned. However, the outages can be detrimental to op-
erators committed to operations during daylight hours.
2.7 SUMMARY
Newton’s laws of motion explain the forces on a satel- (UTC), otherwise referred to as GMT, is given. The
lite in orbit. The balance between the force pulling a use of Julian days, which begin at noon, was intro-
satellite inward to the earth—i.e. gravity—and that duced by astronomers to allow them to make obser-
trying to fling a satellite away from the earth—kinetic vations overnight without having the day change on
energy—is a fine one. To achieve stable orbit, a satel- them (as normal UTC days do at midnight).
lite must have the correct velocity, be traveling in the Locating the satellite in its orbit is a complex
right direction, and be at the right height for its process, with a number of possible frames of refer-
velocity. As the orbital height increases, the gravita- ence. Different approaches are discussed. Procedures
tional acceleration decreases, the orbital velocity for calculating the look angles from the earth to a geo-
decreases, and the period of the satellite increases. Cal- stationary satellite are given. The natural forces that
culation procedures for obtaining the period of a satel- act on a satellite to cause orbital perturbations are set
lite and its velocity are set out. It is seen that Kepler’s out and the need for orbital maneuvers explained. The
constant, the product of the universal gravitational important difference between orbital maneuver life
constant, G, and the mass of the earth ME, is funda- and orbital design life is explained. Details on launch
mental to many of the equations that give the forces procedures and launch vehicles are provided, with
on the satellite and the velocity of the satellite in its typical launch campaign information set out. The two
orbit. Kepler’s three laws describing the motion of one basic methods of launching geostationary satellites
body orbiting another are given and the terminology are described, one using an apogee kick motor and
employed in satellite ephemeris data is explained. The the other a slow orbit raising technique. Finally,
relationship between the astronomers’ use of Julian Doppler shift, range variations, solar eclipse, and sun
dates and Julian days and the Universal Time Constant transit outage are reviewed.
REFERENCES
1. GARY D. GORDON and WALTER L. MORGAN, Principles 9. R. DEKOK, “Spacelift in and beyond the next millen-
of Communications Satellites, John Wiley & Sons, nium,” Launchspace, p. 6, May/June 1999.
ISBN 0-471-55796-X, 1993. 10. As reported in jeff@spaceviews.com in July 1999,
2. The American Ephemeris and Nautical Almanac, U.S. the biweekly update from the web site given in ref-
Government Printing Office, Washington, DC (published erence 6.
annually). 11. J. J. SPILKER, Jr., Digital Communications by Satellite,
3. http://www.stk.com Prentice-Hall, Englewood Cliffs, NJ, 1977.
4. The NASA liftoff home page is http://liftoff.msfc.nasa. 12. JAMES MARTIN, Communications Satellite Systems,
gov/realtime/JTrack/Spacecraft.html The home page al- Prentice-Hall, Englewood Cliffs, NJ, 1978.
lows you to see the International Space Station, weather 13. D. M. RUSSELL, “Browsing orbital analysis tools,”
and research satellites, and the Shuttle track if it is in Launchspace, Vol. 3, No. 6, p. 24, December 1, 1998.
orbit. The page specializing in three-dimensional graph- 14. JAMES R. WERTZ and WILEY J. LARSON, eds., Space Mis-
ical views of satellites is http://liftoff.msfc.nasa. gov/re- sion Analysis and Design, 3rd Ed., Kluwer Academic
altime/jtrack/3d/Jtrack3d.html Publishers, 2001. ISBN 0-7923-5901-1.
5. Private communication, EE 4644 Spring 1997, DAVID 15. Aviation Week and Space Technology, Vol. 151, No. 24,
WALSH and CLIF GROVES. December 13, 1999. Special issue on 21st century
6. http://www.spaceviews.com launch vehicles.
7. http://www.orbreport.com is a site dedicated to the 16. “Chinese Rockets and R&D Advances.” Aviation Week
space transportation industry and is an element of ISIR, and Space Technology, Vol. 155, No. 20, pp. 54 – 55,
International Space Industry Report. November 12, 2001.
8. K. ROUNDTREE, “Launching Payloads by Sea,” Launch-
space, pp. 38–39, May/June 1999; see also in the Hughes
web site at http://www.hcisat.com
c02.qxd 13/07/02 14:10 Page 55
PROBLEMS 55
PROBLEMS
1. Explain what the terms centrifugal and centripetal 5. An observation satellite is to be placed into a
mean with regard to a satellite in orbit around the circular equatorial orbit so that it moves in the same
earth. direction as the earth’s rotation. Using a synthetic aper-
A satellite is in a circular orbit around the earth. ture radar system, the satellite will store data on sur-
The altitude of the satellite’s orbit above the surface face barometric pressure, and other weather related pa-
of the earth is 1400 km. (i) What are the centripetal rameters, as it flies overhead. These data will later be
and centrifugal accelerations acting on the satellite in played back to a controlling earth station after each trip
its orbit? Give your answer in m/s. (ii) What is the around the world. The orbit is to be designed so that
velocity of the satellite in this orbit? Give your an- the satellite is directly above the controlling earth sta-
swer in km/s. (iii) What is the orbital period of the tion, which is located on the equator, once every 4 h.
satellite in this orbit? Give your answer in hours, min- The controlling earth station’s antenna is unable to op-
utes, and seconds. Note: Assume the average radius erate below an elevation angle of 10° to the horizontal
of the earth is 6378.137 km and Kepler’s constant has in any direction. Taking the earth’s rotational period to
the value 3.986004418 105 km3/s2. be exactly 24 h, find the following quantities:
2. A satellite is in a 322-km high circular orbit. De- a. The satellite’s angular velocity in radians per second.
termine: b. The orbital period in hours.
a. The orbital angular velocity in radians per second; c. The orbital radius in kilometers.
b. The orbital period in minutes; and d. The orbital height in kilometers.
c. The orbital velocity in meters per second. e. The satellite’s linear velocity in meters per second.
Note: Assume the average radius of the earth is f. The time interval in minutes for which the con-
6378.137 km and Kepler’s constant has the value trolling earth station can communicate with the satel-
3.986004418 105 km3/s2. lite on each pass.
3. The same satellite in Problem 2 above (322-km 6. What is the difference, or are the differences, be-
circular orbit) carries a 300-MHz transmitter. tween a geosynchronous satellite and a geostationary
a. Determine the maximum frequency range over satellite orbit? What is the period of a geostationary
which the received signal would shift due to Doppler satellite? What is the name given to this orbital pe-
effects if received by a stationary observer suitably riod? What is the velocity of a geostationary satellite
located in space. Note: The frequency can be shifted in its orbit? Give your answer in km/s.
both up and down, depending on whether the satel- A particular shuttle mission released a TDRSS satel-
lite is moving toward or away from the observer. You lite into a circular low orbit, with an orbital height of
need to determine the maximum possible change in 270 km. The shuttle orbit was inclined to the earth’s
frequency due to Doppler (i.e., 2 f). equator by approximately 28°. The TDRSS satellite
b. If an earth station on the surface of the earth at needed to be placed into a geostationary transfer orbit
mean sea level, 6370 km from the center of the earth, (GTO) once released from the shuttle cargo bay, with
can receive the 300-MHz transmissions down to the apogee of the GTO at geostationary altitude and the
an elevation angle of 0°, calculate the maximum perigee at the height of the shuttle’s orbit. (i) What was
Doppler shift that this station will observe. Note: In- the eccentricity of the GTO? (ii) What was the period
clude the earth’s rotation and be sure you consider of the GTO? (iii) What was the difference in velocity
the maximum possible Doppler shift for a 322-km of the satellite in GTO between when it was at apogee
circular orbit. and when it was at perigee? Note: Assume the average
4. What are Kepler’s three laws of planetary motion? radius of the earth is 6378.137 km and Kepler’s con-
Give the mathematical formulation of Kepler’s third stant has the value 3.986004418 105 km3/s2.
law of planetary motion. What do the terms perigee 7. For a variety of reasons, typical minimum eleva-
and apogee mean when used to describe the orbit of tion angles used by earth stations operating in the
a satellite orbiting the earth? commercial fixed services using satellites (FSS) com-
A satellite in an elliptical orbit around the earth munications bands are as follows: C band 5°; Ku band
has an apogee of 39,152 km and a perigee of 500 km. 10°; and Ka band 20°.
What is the orbital period of this satellite? Give your (i) Determine the maximum and minimum range
answer in hours. Note: Assume the average radius of in kilometers from an earth station to a geostationary
the earth is 6378.137 km and Kepler’s constant has satellite in the three bands. (ii) To what round-trip
the value 3.986004418 10 5 km3/s2. signal propagation times do these ranges correspond?
c02.qxd 13/07/02 14:10 Page 56
You may assume the signal propagates with the ve- between two earth stations on the surface of the earth,
locity of light in a vacuum even when in the earth’s which are not simultaneously visible to any single
lower atmosphere. satellite in the system, by using the ISL equipment to
8. Most commercial geostationary communications link up the satellites. In this question, the effects of
satellites must maintain their orbital positions to ray bending in the atmosphere may be ignored, pro-
within 0.05° of arc. If a geostationary satellite cessing delays on the satellites may initially be as-
meets this condition (i.e., it has an apparent motion sumed to be zero, the earth may be assumed to be
0.05° of arc N–S and 0.05° of arc E–W, as meas- perfectly circular with a flat (i.e., not hilly) surface,
ured from the center of the earth), calculate the maxi- and the velocity of the signals in free space (whether
mum range variation to this satellite from an earth in the earth’s lower atmosphere or essentially in a vac-
station with a mean elevation angle to the center of uum) may be assumed to be the velocity of light in a
the satellite’s apparent motion of 5°. You may assume vacuum.
that the equatorial and polar diameters of the earth (i) What is the furthest apart two geostationary
are the same. satellites may be so that they can still communicate
with each other without the path between the two
9. An interactive experiment is being set up between
satellites being interrupted by the surface of the earth?
the University of York, England (approximately
Give your answer in degrees longitude between the
359.5° E, 53.5° N) and the Technical University of
subsatellite points. (ii) If the longest, one-way delay
Graz, Austria (approximately 15° E, 47.5° N) that will
permitted by the ITU between two earth stations
make use of a geostationary satellite. The earth sta-
communicating via a space system is 400 ms, what
tions at both universities are constrained to work only
is the furthest apart two geostationary satellites may
above elevation angles of 20° due to buildings, etc.,
be before the transmission delay of the signal from
near their locations. The groups at the two universi-
one earth station to the other, when connected through
ties need to find a geostationary satellite that will be
the ISL system of the two satellites, equals 400 ms?
visible to both universities simultaneously, with both
The slant path distance between each earth station and
earth stations operating at, or above, an elevation an-
the geostationary satellite it is communicating with
gle of 20°. What is the range of sub-satellite points
may be assumed to be 40,000 km. (iii) If the satel-
between which the selected geostationary satellite
lites in part (ii) employ onboard processing, which
must lie?
adds an additional delay of 35 ms in each satellite,
10. The state of Virginia may be represented roughly what is the maximum distance between the ISL-
as a rectangle bounded by 39.5° N latitude on the linked geostationary satellites now? (iv) If both of the
north, 36.5° N latitude on the south, 76.0° W longi- two earth stations used in parts (ii) and (iii) must ad-
tude on the east, and 86.3° W longitude on the west. ditionally now send the signals over a 2500-km opti-
If a geostationary satellite must be visible throughout cal fiber line to the end user on the ground, with an
Virginia at an elevation angle no lower than 20°, what associated transmission delay in the fiber at each end
is the range of longitudes within which the subsatel- of the link, what is the maximum distance between
lite point of the satellite must lie? the ISL-linked geostationary satellites now? You may
11. A geostationary satellite system is being built assume a refractive index of 1.5 for the optical fiber
which incorporates intersatellite links (ISLs) between and zero processing delay in the earth station equip-
the satellites. This permits the transfer of information ment and end user equipment.
c03.qxd 28/08/02 19:34 Page 57
CHAPTER 3
SATELLITES
Maintaining a microwave communication system in orbit in space is not a simple prob-

lem, so communications satellites are very complex, extremely expensive to purchase, and
also expensive to launch. A typical large geostationary satellite, for example, is estimated
to cost around $125 M, on station (see Chapter 2). The cost of the satellite and launch are
increased by the need to dedicate an earth station to the monitoring and control of the
satellite, at a cost of several million dollars per year. The revenue to pay these costs is ob-
tained by selling the communication capacity of the satellite to users, either by way of
leasing circuits or transponders, or by charging for circuit use, as in the international
telephone and data transmission service.
Communications satellites are usually designed to have a typical operating life-
time of 10 to 15 years. The operator of the system hopes to recover the initial and op-
erating costs well within the expected lifetime of the satellite, and the designer must
provide a satellite that can survive the hostile environment of outer space for that long.
In order to support the communications system, the satellite must provide a stable plat-
form on which to mount the antennas, be capable of station keeping, provide the required
electrical power for the communication system, and also provide a controlled tempera-
ture environment for the communications electronics. In this chapter we discuss the sub-
systems needed on a satellite to support its primary mission of communications. We
also discuss the communications subsystem itself in some detail, and other problems
such as reliability.
The emphasis throughout this chapter is on satellites in geostationary orbit. Com-
munications satellites for low earth orbit are in most cases quite similar to small GEO
satellites and have similar requirements. The discussion of satellites in this chapter is
necessarily brief. For more details of the many subsystems used on satellites and their
construction and operation the reader should refer to reference 1. Much information
about individual satellites can be found on the web sites of satellite manufacturers and
operators. See Table 1.1–1.4 in Chapter 1 for an extensive listing, which includes Web
addresses.
3.1 SATELLITE SUBSYSTEMS
The major subsystems required on the satellite are given below. Figure 3.1 shows an ex-
ploded view of a typical geostationary (GEO) satellite with several of the subsystems
indicated.
Attitude and Orbit Control System (AOCS)

This subsystem consists of rocket motors that are used to move the satellite back to the
correct orbit when external forces cause it to drift off station and gas jets or inertial devices
that control the attitude of the satellite.
57
c03.qxd 28/08/02 19:35 Page 58
Diameter:
238 cm (93 in.)
Overall Height:
701 cm (275 in.)
Weight:
785 kg in orbit
(1732 lb)
Telemetry & command bicones
Receiver reflector
Nutation damper
Receive feed
horns
Transmit odd channel Transmit even
reflector channel reflector
Odd channel
feed horns Even channel
transmit feed horns
Global transmit transmit
horn Spot beam output multiplexer
Global
receive horn Global output Input filters
Telemetry horns Looking
multiplexer
forward
Beacon
TWTA transmitter
Bearing and power EPCs Despun shelf

transfer assembly
(BAPTA) Communication receivers Encoder/decoder
Earth sensors
Inboard Conical sunshield Position and orientation
sunshield propellant tanks (4) Command
Radial jet receivers
Sun sensor
Battery controller
Axial Radial jet
jet
Axial jet
Earth sensor
Booster adapter
Spinup jet
Spun electronics stack
FIGURE 3.1 Exploded view of a spinner satellite based on the Boeing (Hughes) HS 376 design.
INTELSAT IVA (courtesy of Intelsat)
58
c03.qxd 28/08/02 19:35 Page 59
3.1 SATELLITE SUBSYSTEMS 59
Telemetry, Tracking, Command, and

Monitoring (TTC&M)
These systems are partly on the satellite and partly at the controlling earth station1. The
telemetry system sends data derived from many sensors on the satellite, which monitor the
satellite’s health, via a telemetry link to the controlling earth station. The tracking system
is located at this earth station and provides information on the range and the elevation and
azimuth angles of the satellite. Repeated measurement of these three parameters permits
computation of orbital elements, from which changes in the orbit of the satellite can be de-
tected. Based on telemetry data received from the satellite and orbital data obtained from
the tracking system, the control system is used to correct the position and attitude of the
satellite. It is also used to control the antenna pointing and communication system config-
uration to suit current traffic requirements, and to operate switches on the satellite.
Power System
All communications satellites derive their electrical power from solar cells. The power is
used by the communications system, mainly in its transmitters, and also by all other elec-
trical systems on the satellite. The latter use is termed housekeeping, since these subsystems
serve to support the communications system.
Communications Subsystems
The communications subsystem is the major component of a communications satellite, and
the remainder of the satellite is there solely to support it. Frequently, the communications
equipment is only a small part of the weight and volume of the whole satellite. It is usu-
ally composed of one or more antennas, which receive and transmit over wide bandwidths
at microwave frequencies, and a set of receivers and transmitters that amplify and retrans-
mit the incoming signals. The receiver–transmitter units are known as transponders. There
are two types of transponder in use on satellites: the linear or bent pipe transponder that
amplifies the received signal and retransmits it at a different, usually lower, frequency, and
the baseband processing transponder which is used only with digital signals, that converts
the received signal to baseband, processes it, and then retransmits a digital signal.
Satellite Antennas
Although these form part of the complete communication system, they can be considered
separately from the transponders. On large GEO satellites the antenna systems are very
complex and produce beams with shapes carefully tailored to match the areas on the earth’s
surface served by the satellite. Most satellite antennas are designed to operate in a single
frequency band, for example, C band or Ku band. A satellite which uses multiple frequency
bands usually has four or more antennas.
The subsystems listed above are discussed in more detail in this chapter. There are
other subsystems that are not discussed here, but which are essential to the operation of
the satellite—the thermal control system that regulates the temperature inside a satellite,
for example. The reader who is interested in spacecraft design should refer to the literature
of that field, particularly the IEEE Transactions on Aerospace and Electronic Systems2
and the American Institute of Aeronautics and Astronautics Transactions and annual
Conference Proceedings 3,4. Only a brief review of the subsystems that support the com-
munication mission is included here.
c03.qxd 28/08/02 19:35 Page 60
60 CHAPTER 3 SATELLITES
3.2 ATTITUDE AND ORBIT CONTROL

SYSTEM (AOCS)
The attitude and orbit of a satellite must be controlled so that the satellite’s antennas point
toward the earth and so that the user knows where in the sky to look for the satellite. This
is particularly important for GEO satellites since the earth station antennas that are used
with GEO satellites are normally fixed and movement of the satellite away from its ap-
pointed position in the sky will cause a loss of signal. There are several forces acting on
an orbiting satellite that tend to change its attitude and orbit, as discussed in Chapter 2.
The most important are the gravitational fields of the sun and the moon, irregularities in
the earth’s gravitational field, solar pressure from the sun, and variations in the earth’s
magnetic field.
Solar pressure acting on a satellite’s solar sails and antennas, and the earth’s mag-
netic field generating eddy currents in the satellite’s metallic structure as it travels through
the magnetic field, tend to cause rotation of the satellite body. Careful design of the struc-
ture can minimize these effects, but the orbital period of the satellite makes many of the
effects cyclic, which can cause nutation (a wobble) of the satellite. The attitude control
system must damp out nutation and counter any rotational torque or movement.
The presence of gravitational fields from the sun and the moon cause the orbit of a
GEO satellite to change with time. At GEO orbit altitude, the moon’s gravitational force
is about twice as strong as the sun’s. The moon’s orbit is inclined to the equatorial plane
by approximately 5°, which creates a force on the satellite with a component that is nor-
mal to the satellite’s orbit. The plane of the earth’s rotation around the sun is inclined by
23° to the earth’s equatorial plane. As discussed in Chapter 2, there is a net gravitational
pull on the satellite that tends to change the inclination of the satellite’s orbit, pulling it
away from the earth’s equatorial plane at an initial rate of approximately 0.86° per year.
The orbital control system of the satellite must be able to move the satellite back into the
equatorial plane before the orbital inclination becomes excessive. LEO satellites are less
affected by gravitational fields of the sun and moon. Since they are much closer to the
earth than GEO satellites, the earth’s gravity is much stronger, and the pull from the sun
and moon are proportionately weaker.
The earth is not quite a perfect sphere. At the equator, there are bulges of about 65 m
at longitudes 162° E and 348° E, with the result that a satellite is accelerated toward one
of two stable points in the GEO orbit at longitude 75° E and 252° E, as shown in Figure 3.2.
To maintain accurate station keeping, the satellite must be periodically accelerated in the
opposite direction to the forces acting on it. This is done as a sequence of station-keeping
maneuvers, using small rocket motors (sometimes called gas jets or thrusters) that can be
controlled from the earth via the TTC&M system.
Attitude Control System

There are two ways to make a satellite stable in orbit, when it is weightless. The body of
the satellite can be rotated, typically at a rate between 30 and 100 rpm, to create a gyro-
scopic force that provides stability of the spin axis and keeps it pointing in the same di-
rection. Such satellites are known as spinners. The popular Hughes 376 (now Boeing 376)
satellite is an example of a spinner design. Alternatively, the satellite can be stabilized by
one or more momentum wheels. This is called a three-axis stabilized satellite, of which
the Hughes (Boeing) 701 series is an example. The momentum wheel is usually a solid
metal disk driven by an electric motor. Either there must be one momentum wheel for
each of the three axes of the satellite, or a single momentum wheel can be mounted on
c03.qxd 28/08/02 19:35 Page 61
3.2 ATTITUDE AND ORBIT CONTROL SYSTEM (AOCS) 61
Drift
Drift
Satellite stable point 75° E

E
N
DIA
MERI
W ICH 0°
Satellite semistable EEN
GR
point 165° E
Satellite semistable
point 15° W W
Equator
65-m bulge
Synchronous
orbit
Satellite stablepoint 105° W
Drift
Drift
FIGURE 3.2 Forces on a synchronous satellite.
gimbals and rotated to provide a rotational force about any of the three axes5,6. Increas-
ing the speed of the momentum wheel causes the satellite to precess in the opposite di-
rection, according the principle of conservation of angular momentum. Figure 3.3 shows
examples of both the spinner and three-axis design of satellite.
The spinner design of satellite is typified by many satellites built by the Hughes
Aircraft Corporation for domestic satellite communication systems. As shown in Fig-
ures 3.1 and 3.3a, the satellite consists of a cylindrical drum covered in solar cells that
contains the power systems and the rocket motors. The communications system is mounted
at the top of the drum and is driven by an electric motor in the opposite direction to the
rotation of the satellite body to keep the antennas pointing toward the earth. Such satel-
lites are called despun.
SIDEBAR
In the early days of satellite communication despun an- no RF rotating joints are needed in the communication
tennas were not used, so antennas with a circular sym- system. Electrical power and some control signals must
metric pattern were employed. These antennas have be brought through the despin bearing assembly with
low gain and are now used only for basic TTC&M sys- slip rings. The design of the bearing and slip rings to
tems that must operate regardless of the satellite’s ori- guarantee friction-free operation for 10 to 15 years in
entation. By despinning the antennas and transponders, the total vacuum of outer space is a challenge.
c03.qxd 28/08/02 19:35 Page 62
(a )
FIGURE 3.3 (a) A spin-

ner satellite, INTELSAT IV
A (b) A three-axis stabi-
(b) lized satellite, INTELSAT
V (courtesy of Intelsat).
The satellite is spun up by operating small radial gas jets mounted on the periphery
of the drum, at an appropriate point in the launch phase. The despin system is then brought
into operation so that the main TTC&M antennas point toward the earth. The main TTC&M
system operates at 64 GHz on the Intelsat satellite, with a 2-GHz backup system for use
during the launch phase. A variety of liquid propulsion mixes have been used for the gas
jets, the most common being a variant of hydrazine (N2H4), which is easily liquefied under
c03.qxd 28/08/02 19:35 Page 63
pressure, but readily decomposes when passed over a catalyst7. Increased power can be ob-
tained from the hydrazine gas jets by electrically heating the catalyst and the gas. Satellites
that use liquid fuel thrusters have standardized on bipropellant fuels, that is fuels that mix
together to form the thruster fuel. The most common bipropellents used for thruster oper-
ations are mono-methyl hydrazine and nitrogen tetroxide, although standard hydrazine is
still used in place of mono-methyl hydrazine by some satellite manufacturers. The bipro-
pellants are hypogolic: that is they ignite spontaneously on contact, and so do not need
either a catalyst or a heater. By adjusting the flow of the bipropellants, pulses of thrust can
be generated at the correct time and in the correct direction.
There are two types of rocket motors used on satellites. The traditional bipropellant
thruster described above, and arc jets or ion thrusters. The fuel that is stored on a GEO
satellite is used for two purposes: to fire the apogee kick motor that injects the satellite
into its final orbit, and to maintain the satellite in that orbit over its lifetime. If the launch
is highly accurate, a minimum amount of fuel is used to attain the final orbit. If the launch
is less accurate, more fuel must be used up in maneuvering the satellite into position, and
that reduces the amount left for station keeping. A new development in thrusters uses a
high voltage source to accelerate ions to a very high velocity, thus producing thrust. The
ion engine thrust is not large, but because the engine can be driven by power from the
solar cells it saves on expendable fuel. Ion engines can also be used to slowly raise a GEO
satellite from a transfer orbit to GEO orbit as described in Chapter 2, although the process
takes months rather than hours as with a conventional rocket engine.
Arc jets or ion thrusters are mainly used for north–south station keeping, which is where
the greatest use of fuel is required for station-keeping maneuvers, and became operational on
the Hughes (Boeing) 600 series of satellite buses. Arc jets or ion thrusters lack the total thrust
required to move satellites quickly (e.g., for major longitudinal changes in position) but a
small, continuous thrust is adequate to maintain N–S and E–W position keeping.
In a three-axis stabilized satellite, one pair of gas jets is needed for each axis to pro-
vide for rotation in both directions of pitch, roll, and yaw. An additional set of controls,
allowing only one jet on a given axis to be operated, provides for velocity increments in
the X, Y, and Z directions. When motion is required along a given axis, the appropriate
gas jet is operated for a specified period of time to achieve the desired velocity. The op-
posing gas jet must be operated for the same length of time to stop the motion when the
satellite reaches its new position. Fuel is saved if the velocity of the satellite is kept small,
but progress toward the destination is slow.
Let us define a set of reference Cartesian axes (XR, YR, ZR) with the satellite at the
origin, as shown in Figure 3.4. The ZR axis is directed toward the center of the earth and
is in the plane of the satellite orbit. It is aligned along the local vertical at the satellite’s
subsatellite point. The XR axis is tangent to the orbital plane and lies in the orbital plane.
The YR axis is perpendicular to the orbital plane. For a satellite serving the Northern
Hemisphere, the directions of the XR and YR axes are nominally east and south.
Rotation about the XR, YR, and ZR axes is defined as roll about the XR axis, pitch
about the YR axis, and yaw about the ZR axis, in exactly the same way as for an aircraft
or ship traveling in the X direction. The satellite must be stabilized with respect to the ref-
erence axes to maintain accurate pointing of its antenna beams. The axes XR, YR, and ZR
are defined with respect to the location of the satellite; a second set of Cartesian axes, X,
Y, Z, as shown in Figure 3.4, define the orientation of the satellite. Changes in a satellite’s
attitude cause the angles , , and in Figure 3.4 to vary as the X, Y, and Z axes move
relative to the fixed reference axes XR, YR, and ZR. The Z axis is usually directed toward
a reference point on the earth, called the Z-axis intercept. The location of the Z-axis in-
tercept defines the pointing of the satellite antennas; the Z-axis intercept point may be
c03.qxd 28/08/02 19:35 Page 64
Earth
Equator Z
S
N E
ZR
XR
Yaw Roll
X
W Spacecraft
Pitch
Y YR
S
(a )
φ
ZR Orbital
θ
path
XR
θ
ψ X
Y YR
(b )
FIGURE 3.4 (a) Forces on a satellite. (b) Relationship between axes of a satellite.
moved to repoint all the antenna beams by changing the attitude of the satellite with the
attitude control system.
In a spinner-type satellite, the axis of rotation is usually the Y axis, which is main-
tained close to the YR axis, perpendicular to the orbital plane. Pitch correction is required
only on the despun antenna system and can be obtained by varying the speed of the de-
spin motor. Yaw and roll are controlled by pulsing radially mounted jets at the appropri-
ate instant as the body of the satellite rotates.
Attitude control of a three-axis stabilized satellite requires an increase or a decrease
in the speed of the inertia wheel. If a constant torque exists about one axis of the satel-
lite, a continual increase or decrease in momentum wheel speed is necessary to maintain
the correct attitude. When the upper or lower speed limit of the wheel is reached, it must
be unloaded by operating a pair of gas jets and simultaneously reducing or increasing
the wheel speed. Closed-loop control of attitude is employed on the satellite to maintain
c03.qxd 28/08/02 19:35 Page 65
North
Earth Fields of view

Upper sensor
Lower sensor
Voltage
South Upper sensor
output
Lower sensor
output
Time Error to N No error Error to S
FIGURE 3.5 Principle of N–S control of a spinner satellite using infrared Earth sensors.
the correct attitude. When large, narrow beam antennas are used, the whole satellite may
have to be stabilized within 0.1° on each axis. The references for the attitude control
system may be the outer edge of the earth’s disk, as observed with infrared sensors, the
sun, or one or more stars.
Figure 3.5 illustrates how an infrared sensor on the spinning body of a satellite can
be used to control pointing toward the earth. Figure 3.6 shows a typical control system
loop using the technique illustrated in Figure 3.5. The control system will be more com-
plex for a three-axis stabilized satellite and may employ an onboard computer to process
the sensor data and command the gas jets and momentum wheels.
Despin control system (E–W)

Upper
sensor Earth E–W Antenna Antenna
Lower pulse control drive position
input unit servo sensor
sensor
Attitude control system (N–S)
Earth Attitude Gas jet N–S

pulse control control jet
comparator system unit controls
6 GHz 4 GHz
horn horn
Command Control Telemetry Telemetry
receiver system transmitter formatter
Command and telemetry system

FIGURE 3.6 Typical onboard control system for a spinner satellite.
c03.qxd 28/08/02 19:35 Page 66
Correct position for

spacecraft in equatorial plane
Earth
W
Equatorial
plane
E
Inclined orbital
plane
S
Spacecraft position
in inclined orbital FIGURE 3.7 Satellite in
plane inclined orbit.
Orbit Control System

As discussed in Chapter 2, a geostationary satellite is subjected to several forces that tend
to accelerate it away from its required orbit. The most important, for the geostationary
satellite, are the gravitation forces of the moon and the sun, which cause inclination of
the orbital plane, and the nonspherical shape of the earth around the equator, which causes
drift of the subsatellite point. There are many other smaller forces that act on the satellite
causing the orbit to change. Accurate prediction of the satellite position a week or 2 weeks
ahead requires a computer program with up to 20 force parameters; we shall restrict our
discussion here to the two major effects.
Figure 3.7 shows a diagram of an inclined orbital plane close to the geostationary
orbit. For the orbit to be truly geostationary, it must lie in the equatorial plane, be circu-
lar, and have the correct altitude. The various forces acting on the satellite will steadily
pull it out of the correct orbit; it is the function of the orbit control system to return it to
the correct orbit. This cannot be done with momentum wheels since linear accelerations
are required. Gas jets that can impart velocity changes along the three references axes of
the satellite are required.
If the orbit is not circular, a velocity increase or decrease will have to be made along
the orbit, in the X-axis direction in Figure 3.4. On a spinning satellite, this is achieved by
pulsing the radial jets when they point along the X axis. On a three-axis stabilized satel-
lite, there will usually be two pairs of X-axis jets acting in opposite directions, one pair
of which will be operated for a predetermined length of time to provide the required ve-
locity change. The orbit of a geostationary satellite remains approximately circular for
long periods of time and does not need frequent velocity corrections to maintain circularity.
Altitude corrections are made by operating the Z-axis gas jets.
The inclination of the orbit of a satellite that starts out in a geostationary orbit
increases at an average rate of about 0.85° per year, with an initial rate of change of
c03.qxd 28/08/02 19:35 Page 67
SIDEBAR
The precessional forces that cause inclination changes at the start of operations. Some communications sys-
can be used to increase satellite station-keeping life- tems do not need very tight station-keeping tolerance,
time by deliberately launching the satellite so that the either because the earth segment can track the satel-
inclination is not zero, but the precessional forces lite accurately or because no tracking is required (as
will act to reduce the inclination to close to zero. Dur- would be the case for omnidirectional antennas). GEO
ing this period, the E–W station-keeping tolerance is satellites in such systems may relax their inclination
closely maintained. Once the inclination is close to (N–S) tolerance but may never relax the E–W station
zero, normal station keeping is started to maintain a keeping tolerance as this would lead to unacceptable
tight orbit control in both axes. In this way, approxi- interference into other systems. GEO satellites in such
mately a year of additional maneuvering lifetime may relaxed orbits are sometimes called inclined-orbit
be obtained for each degree the inclination is nonzero satellites.
inclination for a satellite in an equatorial orbit between 0.75° to 0.94° per year (see Chap-
ter 2). Most GEO satellites are specified to remain within a box of 0.05° and so, in prac-
tice, corrections, called a north–south station-keeping maneuver are made every 2 to 4
weeks to keep the error small. It has become normal to split the E–W and N–S maneu-
vers so that at intervals of 2 weeks the E–W corrections are made first and then after 2
more weeks, the N–S corrections are made. If arc jets or ion thrusters are used for N–S
station-keeping maneuvers, these tend to operate almost continuously since their thrust
levels are low when compared with traditional liquid fueled engines.
Correcting the inclination of a satellite orbit requires more fuel to be expended than
for any other orbital correction. This places a weight penalty on those satellites that must
maintain very accurate station keeping, and reduces the communications payload they can
carry. As much as half the total satellite weight at launch may be station keeping fuel
when the satellite’s expected lifetime on orbit is 15 years.
East–west station keeping is effected by use of the X-axis jets of the satellite. For a
satellite located away from the stable points at 75° E and 252° E, a slow drift toward these
points will occur. Typically, the X-axis jets are pulsed every 2 or 3 weeks to counter the
drift and add a small velocity increment in the opposite direction. The satellite then drifts
through its nominal position, stops at a point a fraction of a degree beyond it, and then
drifts back again. East–west station keeping requires only a modest amount of fuel and is
necessary on all geostationary communications satellites to maintain the spacing between
adjacent satellites. With orbital locations separated by 2° or 3°, east–west drifts in excess
of a fraction of a degree cannot be tolerated, and most GEO satellites are held within
0.05° of their allotted longitude.
Some communications satellites such as the Russian Molniya series are not in geo-
stationary orbit. (Molniya means lightning in Russian.) Early Molniya satellites were
launched into a highly elliptical 12-h orbit with a large (65°) inclination angle to provide
communication to northerly latitudes like Siberia. The Russian satellite gave its name to
any satellite in a highly elliptical inclined orbit.
Low earth orbit (LEO) and medium earth orbit (MEO) satellites also need AOC
systems to maintain the correct orbit and attitude for continuous communication. Be-
cause of the much stronger gravitational force of the earth in LEO orbit, attitude stabi-
lization is often accomplished with a rigid gravity gradient boom. This is a long pole
that points toward the center of the earth, providing damping of oscillations about the
satellite’s z axis by virtue of the difference in gravitational field at the top of the pole
and at the bottom.
c03.qxd 28/08/02 19:35 Page 68
3.3 TELEMETRY, TRACKING, COMMAND,

AND MONITORING
The TTC&M system is essential to the successful operation of a communications satellite.

It is part of the satellite management task, which also involves an earth station, usually
dedicated to that task, and a group of personnel. The main functions of satellite manage-
ment are to control the orbit and attitude of the satellite, monitor the status of all sensors
and subsystems on the satellite, and switch on or off sections of the communication sys-
tem. The TTC&M earth station may be owned and operated by the satellite owner, or it
may be owned by a third party and provide TTC&M services under contract.
On large geostationary satellites, some repointing of individual antennas may be
possible, under the command of the TTC&M system. Tracking is performed primarily by
the earth station. Figure 3.8 illustrates the functions of a controlling earth station.
Telemetry and Monitoring System

The monitoring system collects data from many sensors within the satellite and sends
these data to the controlling earth station. There may be several hundred sensors located
on the satellite to monitor pressure in the fuel tanks, voltage and current in the power con-
ditioning unit, current drawn by each subsystem, and critical voltages and currents in the
communications electronics. The temperature of many of the subsystems is important and
must be kept within predetermined limits, so many temperature sensors are fitted. The
sensor data, the status of each subsystem, and the positions of switches in the communi-
cation system are reported back to the earth by the telemetry system. The sighting devices
used to maintain attitude are also monitored via the telemetry link: this is essential in case
one should fail and cause the satellite to point in the wrong direction. The faulty unit must
then be disconnected and a spare brought in, via the command system, or some other
means of controlling attitude devised.
Telemetry data are usually digitized and transmitted as phase shift keying (PSK) of
a low-power telemetry carrier using time division techniques. A low data rate is normally
used to allow the receiver at the earth station to have a narrow bandwidth and thus main-
tain a high carrier to noise ratio. The entire TDM frame may contain thousands of bits of
data and take several seconds to transmit. At the controlling earth station a computer can
be used to monitor, store, and decode the telemetry data so that the status of any system
or sensor on the satellite can be determined immediately by the controller on the earth.
Alarms can also be sounded if any vital parameter goes outside allowable limits.
Tracking
A number of techniques can be used to determine the current orbit of a satellite. Velocity
and acceleration sensors on the satellite can be used to establish the change in orbit from
the last known position, by integration of the data. The earth station controlling the satel-
lite can observe the Doppler shift of the telemetry carrier or beacon transmitter carrier
to determine the rate at which range is changing. Together with accurate angular meas-
urements from the earth station antenna, range is used to determine the orbital elements.
Active determination of range can be achieved by transmitting a pulse, or sequence of
pulses, to the satellite and observing the time delay before the pulse is received again.
The propagation delay in the satellite transponder must be accurately known, and more
than one earth station may make range measurements. If a sufficient number of earth sta-
tions with an adequate separation are observing the satellite, its position can be established
c03.qxd 28/08/02 19:35 Page 69
3.3 TELEMETRY, TRACKING, COMMAND, AND MONITORING 69
Satellite
Satellite
TTC&M antenna
Receive Transmit
antenna antenna
Tele-
Telemetry
command
receiver
transmitter
Tracking
system
Data
processor
Computer
for attitude Controller FIGURE 3.8 Typical
and orbital control tracking, telemetry,
Ephemeris command and
data monitoring system.
by triangulation from the earth station by simultaneous range measurements. With pre-
cision equipment at the earth stations, the position of the satellite can be determined
within 10 m.
Ranging tones are also used for range measurement. A carrier generated on board
the satellite is modulated with a series of sine waves at increasing frequency, usually
harmonically related. The phase of the sine wave modulation components is compared
at an earth station, and the number of wavelengths of each frequency is calculated. Am-
biguities in the numbers are resolved by reference to lower frequencies, and prior knowl-
edge of the approximate range of the satellite. If sufficiently high frequencies are used,
perhaps even the carrier frequency, range can be measured to millimeter accuracy. The
technique is similar to that used in the terrestrial telurometer and in aircraft radar
altimeters.
c03.qxd 28/08/02 19:35 Page 70
Command
A secure and effective command structure is vital to the successful launch and operation
of any communications satellite. The command system is used to make changes in atti-
tude and corrections to the orbit and to control the communication system. During launch,
it is used to control the firing of the apogee kick motor and to spin up a spinner or extend
the solar sails and antennas of a three-axis stabilized satellite.
The command structure must possess safeguards against unauthorized attempts to
make changes to the satellite’s operation, and also against inadvertent operation of a control
due to error in a received command. Encryption of commands and responses is used to
provide security in the command system. A typical system of the type shown in Figure 3.8
will originate commands at the control terminal of the computer. The control code is con-
verted into a command word, which is sent in a TDM frame to the satellite. After check-
ing for validity in the satellite, the word is sent back to the control station via the teleme-
try link where it is checked again in the computer. If it is found to have been received
correctly, an execute instruction will be sent to the satellite so that the command is exe-
cuted. The entire process may take 5 or 10 s, but minimizes the risk of erroneous commands
causing a satellite malfunction.
The command and telemetry links are usually separate from the communication sys-
tem, although they may operate in the same frequency band (6 and 4 GHz). Two levels
of command system are used in the Intelsat satellite: the main system operates in the 6-
GHz band, in a gap between the communication channel frequencies; the main telemetry
system uses a similar gap in the 4-GHz band. The TTC&M antennas for the 64 GHz
system can be seen in Figure 3.1 on the satellite. These are earth-coverage horns, so the
main system can be used only after correct attitude of the satellite is achieved.
During the launch phase and injection into geostationary orbit, the main TTC&M
system may be inoperable because the satellite does not have the correct attitude or has
not extended its solar sails. A backup system is used at this time, which controls only the
most important sections of the satellite. A great deal of redundancy is built into this sys-
tem, since its failure will jeopardize the entire mission. Near omnidirectional antennas are
used at either UHF or S band (2–4 GHz), and sufficient margin is allowed in the signal-
to-noise ratio (SN) at the satellite receiver to guarantee control under the most adverse
conditions. The backup system provides control of the apogee kick motor, the attitude
control system and orbit control thrusters, the solar sail deployment mechanism (if fitted),
and the power conditioning unit. With these controls, the satellite can be injected into
SIDEBAR
Controlling a satellite in orbit is a complex process The satellite drifted round the earth over a period of
which requires considerable care. In one case, an 21⁄2 months, and was eventually recovered by a team
incorrect sequence of command instructions caused of experts using large antennas in Australia and the
loss of control of the Olympus satellite, a European United States to send telemetry commands to the
large GEO satellite used for experiments in the 3020 satellite. The solar cells provided short bursts of
GHz band. The E–W thrusters fired for a lengthy pe- power as they rotated past the sun’s direction, al-
riod, causing the satellite to drift toward the east at lowing commands to be sent for a few seconds every
5° per day, and the satellite also began rotating with 90 s. The rotation of the satellite was eventually
a period of 90 s. All communication with the satel- stopped and Olympus returned to its correct location,
lite was lost, and the batteries discharged fully be- but with shortened life expectancy due to the loss of
cause the solar sails no longer pointed at the sun. station-keeping fuel.
c03.qxd 28/08/02 19:35 Page 71
3.4 POWER SYSTEMS 71
geostationary orbit, turned to face the earth, and switched to full electrical power so that
handover to the main TTC&M system is possible. In the event of failure of the main
TTC&M system, the backup system can be used to keep the satellite on station. It is also
used to eject the satellite from geostationary orbit and to switch off all transmitters when
the satellite eventually reaches the end of its useful life.
3.4 POWER SYSTEMS
All communications satellites obtain their electrical power from solar cells, which con-
vert incident sunlight into electrical energy. Some deep space planetary research satellites
have used thermonuclear generators to supply electrical power, but because of the danger
to people on the earth if the launch should fail and the nuclear fuel be spread over an
inhabited area, communications satellites have not used nuclear generators.
The sun is a powerful source of energy. In the total vacuum of outer space, at
geostationary altitude, the radiation falling on a satellite has an intensity of 1.39 kW/m2.
Solar cells do not convert all this incident energy into electrical power; their efficiency is
typically 20 to 25% at beginning of life (BOL) but falls with time because of aging of the
cells and etching of the surface by micrometeor impacts. Since sufficient power must be
available at the end of life (EOL) of the satellite to supply all the systems on board, about
15% extra area of solar cells is usually provided as an allowance for aging.
A spin-stabilized satellite usually has a cylindrical body covered in solar cells.
Because the solar cells are on a cylindrical surface, half of the cells are not illuminated
at all, and at the edges of the illuminated half, the low angle of incidence results in little
electrical power being generated. The output from the solar cells is slightly higher than
would be obtained with normal incidence on a flat panel equal in area to the projected
area of the cylinder, that is, its width times its height. The cells that are not illuminated
by sunlight face cold space, which causes them to cool down. The solar cells on a spinner
satellite have a lower temperate than those on solar sails, which increases their efficiency
somewhat. Early satellites were of small dimensions and had relatively small areas of solar
cells. More recently, large communications satellites for direct broadcast operation gen-
erate up to 6 kW from solar power.
A three-axis stabilized satellite can make better use of its solar cell area, since the
cells can be arranged on flat panels that can be rotated to maintain normal incidence of
the sunlight. Only one-third of the total area of solar cells is needed relative to a spinner,
with some saving in weight. A primary advantage, however, is that by unfurling a folded
solar array when the satellite reaches geostationary orbit, power in excess of 10 kW can
be generated with large arrays. To obtain 10 kW from a spinner requires a very large body
on which to place the solar cells, which may then exceed the maximum payload dimensions
of the launch vehicle.
Solar sails must be rotated by an electric motor once per 24 h to keep the cells in
full sunlight. This causes the cells to heat up, typically to 50° to 80°C, which causes a
drop in output voltage. In the spinner design, the cells cool down when in shadow and
run at 20° to 30°C, with somewhat higher efficiency. The bombardment of the sails by
protons and electrons is also more severe, and a thicker layer of glass may be needed to
slow down deterioration of the cells, with a consequent weight penalty. A rotary joint must
be used with each solar sail to transfer current from the rotating sail to the body of the
satellite.
The satellite must carry batteries to power the subsystems during launch and dur-
ing eclipses. Eclipses occur twice per year, around the spring and fall equinoxes, when
c03.qxd 28/08/02 19:35 Page 72
the earth’s shadow passes across the satellite, as illustrated in Figures 2.21 and 2.22. The
longest duration of eclipse is 70 min, occurring around March 21 and September 21 each
year. To avoid the need for large, heavy batteries, part or all of the communications system
load may be shut down during eclipse, but this technique is rarely used when telephony
or data traffic is carried. TV broadcast satellites may not carry sufficient battery capac-
ity to supply their high-power transmitters during eclipse, and may shut down. By lo-
cating the satellite 20° W of the longitude of the service area, the eclipse will occur after
1 A.M. local time for the service area, when shutdown is more acceptable. Batteries are
usually of the nickel–hydrogen type which do not gas when charging and have good re-
liability and long life, and can be safely discharged to 70% of their capacity. A power-
conditioning unit controls the charging current and dumps excess current from the solar
cells into heaters or load resistors on the cold side of the satellite. Sensors on the bat-
teries, power regulator, and solar cells monitor temperature, voltage, and current and sup-
ply these data to both the onboard control system and the controlling earth station via
the telemetry downlink. Typical battery voltages are 20 to 50 V with capacities of 20 to
100 ampere-hours.
3.5 COMMUNICATIONS SUBSYSTEMS
Description of the Communications System

A communications satellite exists to provide a platform in geostationary orbit for the
relaying of voice, video, and data communications. All other subsystems on the satellite
exist solely to support the communications system, although this may represent only a
small part of the volume, weight, and cost of the satellite in orbit. Since it is the com-
munications system that earns the revenue for the system operator, communications satel-
lites are designed to provide the largest traffic capacity possible. The growth in capacity
is well illustrated in Figure 3.9 for the Intelsat system. Successive satellites have become
larger, heavier, and more costly, but the rate at which traffic capacity has increased has
been much greater, resulting in a lower cost per telephone circuit or transmitted bit with
each succeeding generation of satellite. The satellite transponders have limited output
power and the earth stations are at least 36,000 km away from a GEO satellite, so the
received power level, even with large aperture earth station antennas, is very small and
rarely exceeds 1010 W. For the system to perform satisfactorily, the signal power must
exceed the power of the noise generated in the receiver by between 5 and 25 dB, depending
on the bandwidth of the transmitted signal and the modulation scheme used. With low
power transmitters, narrow receiver bandwidths have to be used to maintain the required
signal-to-noise ratios.
Early communications satellites were fitted with transponders of 250 or 500 MHz
bandwidth, but had low gain antennas and transmitters of 1 or 2 W output power. The
earth station receiver could not achieve an adequate signal-to-noise ratio when the full
bandwidth was used with the result that the system was power limited. Later genera-
tions of communications satellites have transponders with greatly increased output
power—up to 200 W for DBS-TV satellites—and have steadily improved in bandwidth
utilization efficiency, as seen in Figure 3.9. The total channel capacity of a satellite
that uses a 500-MHz band at 64 GHz can be increased only if the bandwidth can be
increased or reused. The trend in high-capacity satellites has been to reuse the avail-
able bands by employing several directional beams at the same frequency (spatial
frequency reuse) and orthogonal polarizations at the same frequency (polarization
c03.qxd 28/08/02 19:35 Page 73
Spacecraft INTELSAT I INTELSAT II INTELSAT III INTELSAT IV INTELSAT IV-A INTELSAT V INTELSAT VI
Year of first 1965 1967 1968 1971 1975 1980 1986 (planned)
launch
Dimensions 0.71 m dia × 1.42 m dia × 1.42 m dia × 2.38 m dia × 2.38 m dia × 15.27 m across 3.6 m dia ×
0.59 m high 0.67 m high 1.98 m high 7.01 m high 7.01 m high solar sails × 11.7 m high
6.71 m high
On orbit weight 34 kg 76 kg 152 kg 595 kg 786 kg 1020 kg 1800 kg

End of life 46 W 85 W 125 W 569 W 708 W 1220 W 2100 W
primary power
Total bandwidth 50 MHz 130 MHz 360 MHz 450 MHz 720 MHz 2250 MHz 3360 MHz
Notional capacity 240 240 1500 5000 11,000 plus 24,000 plus 33,000 plus
two-way telephone 2 TV channels 2 TV channels 2 TV channels
circuits
Design lifetime 1.5 years 3 years 5 years 7 years 7 years 10 years 10 years
Spacecraft cost $3.6 M $3.5 M $4.5 M $14 M $18 M $25 M $140 M (first
five satellites)
Launch cost $4.6 M $4.6 M $6 M $20 M $20 M $23 M ?
Cost per telephone $23,000 $11,000 $1,600 $810 $494 $200 ?
circuit year
Contractor Hughes Hughes TRW Hughes Hughes Ford Aerospace Hughes
73
FIGURE 3.9 Illustration of the growth in size and weight of Intelsat satellites over 3 decades.
c03.qxd 28/08/02 19:35 Page 74
frequency reuse). Large GEO satellites also use both the 64 GHz and 1411 GHz
bands to obtain more bandwidth; for example, some GEO satellites have achieved an
effective bandwidth of 2250 MHz within a 500-MHz band at 64 GHz and a 250-MHz
band at 1411 GHz by a combination of spatial and polarization frequency reuse, and
later generations of Intelsat satellites have achieved up to sevenfold reuse of their fre-
quency bands.
The designer of a satellite communication system is not free to select any frequency
and bandwidth he or she chooses. International agreements restrict the frequencies that
may be used for particular services, and the regulations are administered by the appro-
priate agency in each country—the Federal Communication Commission (FCC) in the
United States, for example. Frequencies allocated to satellite services are listed in Tables
4.1 and 4.2 in Chapter 4. The bands currently used for the majority of services are 64
GHz and 1411 GHz, with 3020 GHz coming into service. The 64 GHz band was ex-
panded from 500-MHz bandwidth in each direction to 1000-MHz by a World Adminis-
trative Radio Conference in 1979, but other services share the new part of the band and
may cause interference to the satellite communication link. Similar bandwidth is avail-
able at 1411 GHz, and Ka-band satellites will be launched in the first decade of 2000 to
exploit the wider bandwidths available in the 3020 GHz bands. A different frequency is
required for the transmit path (normally the higher frequency) at the satellite because the
high-power transmit signal would overload the receiver if they both operated at the same
frequency.
The 500-MHz bands originally allocated for 64 and 1411 GHz satellite com-
munications have become very congested and are now completely filled for some seg-
ments of the geostationary orbit, such as that serving North America. Extension of the
bands to 1000 MHz will eventually provide greater capacity as the new frequencies
come into use. Many systems now use 1411 GHz for TV broadcast and distribution,
and 3020 GHz systems are introducing Internet-like services from GEO. The standard
spacing between GEO satellites was originally set at 3°, but under regulations covering
North America and much of the rest of the world, the spacing has been reduced to 2°.
The move to 2° spacing opened up extra slots for new satellites in the 64 and 1411
GHz bands.
Satellite systems designed for Ku band (1411 GHz) and Ka band (3020 GHz),
have narrower antenna beams, and better control of coverage patterns than satellites us-
ing C band (64 GHz). As the available orbital slots for GEO satellites have filled up
with satellites using the 64 and 1411 GHz bands, attention has focused on the use of
SIDEBAR
Ka-band satellite links must accept larger outage times vidual users very well. The user wants high-speed, high
due to rain than 1411 GHz links, and are therefore bet- volume data delivery from the Internet to send video
ter suited to data transmission than voice. The trans- clips and high resolution photographs (known generi-
mission of data can be delayed when a rain fade affects cally as multimedia), but does not send equivalent vol-
the link, but a telephone conversation will be ended if umes of data into the Internet. A lower capacity link
the link fails for more than a few seconds. In the time from the user is adequate, and this suits Ka-band links
frame 2001–2010 several Ka-band satellite systems will with large hub antennas and small user terminals. The
come into service. The major market appears to be data link will suffer occasional outages, perhaps for a total
relay, particularly Internet access. Satellites support of 25 h in a year, when heavy rain is falling through the
asymmetric data links such as Internet service to indi- path between the satellite and the user’s small terminal.
c03.qxd 28/08/02 19:35 Page 75
3.5 COMMUNICATIONS SUBSYSTEMS 75
the 3020 GHz band. Originally, this band had 3 GHz bandwidth allocated to satellite
services, but part of the band was reallocated to the Land Multipoint Distribution Ser-
vice (LMDS). Approximately 2 GHz of bandwidth is still available for satellite systems
at Ka band on an exclusive or shared basis, which is equal to the combined allocations
of C and Ku bands. However, propagation in rain becomes a major factor at frequencies
above 10 GHz. Attenuation (in dB) in rain increases at roughly the square of the fre-
quency, so at 20 GHz rain attenuation is four times larger, in dB, than at 10 GHz. It is
almost impossible to provide rain attenuation margins larger than 20 dB in any system,
and many Ka-band systems will have margins of only 4 or 5 dB when operating to small
earth station antennas.
Transponders
Signals (known as carriers) transmitted by an earth station are received at the satellite by
either a zone beam or a spot beam antenna. Zone beams can receive from transmitters
anywhere within the coverage zone, whereas spot beams have limited coverage. The re-
ceived signal is often taken to two low noise amplifiers and is recombined at their output
to provide redundancy. If either amplifier fails, the other one can still carry all the traf-
fic. Since all carriers from one antenna must pass through a low noise amplifier, a failure
at that point is catastrophic. Redundancy is provided wherever failure of one component
will cause the loss of a significant part of the satellite’s communication capacity.
Figure 3.10 shows a simplified block diagram of a satellite communication subsys-
tem for the 64 GHz band. The 500-MHz bandwidth is divided up into channels, often
36 MHz wide, which are each handled by a separate transponder. A transponder consists
of a band-pass filter to select the particular channel’s band of frequencies, a downconverter
to change the frequency from 6 GHz at the input to 4 GHz at the output, and an output
amplifier. The communication system has many transponders, some of which may be
spares; typically 12 to 44 active transponders are carried by a high-capacity satellite. The
transponders are supplied with signals from one or more receive antennas and send their
outputs to a switch matrix that directs each transponder band of frequencies to the
appropriate antenna or antenna beam. In a large satellite there may be four or five beams
to which any transponder can be connected. The switch setting can be controlled from the
earth to allow reallocation of the transponders between the downlink beams as traffic pat-
terns change.
In the early satellites such as INTELSAT I and II, one or two 250-MHz bandwidth
transponders were employed. This proved unsatisfactory because of the nonlinearity of
the traveling wave tube transmitter used at the output of the transponder, and later GEO
satellites have used up to 44 transponders each with 36, 54, or 72 MHz bandwidth. The
reason for using narrower bandwidth transponders is to avoid excessive intermodulation
problems when transmitting several carriers simultaneously with a nonlinear transmitter,
as discussed in Chapter 6. Intermodulation distortion is likely to occur whenever a high
power amplifier is driven close to saturation. Since we generally want to have more than
one earth station transmitter sending signals via a satellite, one solution would be to pro-
vide one transponder for each earth station’s signal. In the case of the Intelsat global sys-
tem, this could result in a requirement for as many as 100 transponders per satellite. As
a compromise, 36 MHz has been widely used for transponder bandwidth, with 54 and 72
MHz adopted for some satellites.
Many domestic satellites operating in the 64 GHz band carry 24 active transpon-
ders. The center frequencies of the transponders are spaced 40 MHz apart, to allow
c03.qxd 28/08/02 19:35 Page 76
Transmit
3720 3760 3880 3880 3920 3960 4000 4040 4080 4120 4160
1 3 5 9 11 13 15 17 19 21 23
4199
T/M
3740 3820 3860 3900 3940 3980 4020 4060 4100 4140 4180
3701 2 6 8 10 12 14 16 18 20 22 24
T/M
3700 Frequency MHz 4200

Receive
5945 5985 6025 6065 6105 6145 6185 6225 6265 6305 6345 6385
1 3 5 7 9 11 13 15 17 19 21 23
6424
CMD
5965 6005 6045 6085 6125 6165 6205 6245 6285 6325 6365 6405
2 4 6 8 10 12 14 16 18 20 22 24
5925 Frequency MHz 6425

Frequency plan. Note: Number below channel center frequency refers to transponder
identity.
TWTA 1
V
Input 5 Output
multi- 9
multi- Output
13 D
plexer plexer filter V
17
Input Receiver/driver TWTA 21
filter (redundant) TWTA 3
Input 7 Output
11 Output V
multi- 15
multi-
plexer plexer filter
19
TWTA 23
H
TWTA 2 D
Input 6 Output
multi- 10
multi- Output
14 D H
plexer plexer filter
18
Input Receiver/driver TWTA 22
filter (redundant) TWTA 4
Input 8 Output
multi- 12
multi- Output H
16 filter
plexer 20 plexer
TWTA 24
FIGURE 3.10 Transponder arrangement of RCA’s SATCOM satellites and frequency plan.
The translation frequency is 2225 MHz. [Reproduced with permission from W. H. Braun
and J. E. Keigler, “RCA Satellite Networks: High Technology and Low User Cost,” Proc-
eedings of the IEEE, 72, 1483–1505 (November 1984). Copyright © 1984 IEEE.]
guard bands for the 36 MHz filter skirts. With a total of 500 MHz available, a single
polarization satellite can accommodate 12 transponders across the band. When freq-
uency reuse by orthogonal polarizations is adopted, 24 transponders can be accommo-
dated in the same 500 MHz bandwidth. Traditional linear transponder-type satellites
now have sixfold reuse (INTELSAT VI and IX) or even sevenfold reuse (INTELSAT
VIII) at C band. The reuse is achieved through microwave switch interconnections be-
tween subbeams. Internet-like satellites need a plethora of beam interconnections—more
than 50 in most cases. The only way to achieve this level of beam /path interconnections
is via on board processing (OBP).
c03.qxd 28/08/02 19:35 Page 77
14-GHz
11-GHz
Dual polarized
Dual polarized
Ch. 1-2, 5-6, 7-12 Upconverters
Uplink R×1 Downlink
R×2 I/P TWTA’s
MUX and
R×3 O/P MUX
Spot Spot
beam R×4 beam
antennas antennas
6-GHz
Dual polarized 4-GHz
O/P Dual polarized
TWTA’s
MUX
Receivers I/P
6-GHz MUX
Ch. 1-2, 3-4, 5-6

7, 8, 9
Switch O/P
TWTA’s
matrix MUX
Receivers I/P
6-GHz MUX Downlink
Uplink zone
zone beam TWTA’s O/P beam
Ch. 1-2, 3-4, 5-6, 7-8
antenna MUX antenna
6-GHz
Receivers I/P 4-GHz
6-GHz MUX O/P
TWTA’s
Global MUX
beam
Ch. 9, 10, 11, 12, 7-8 Global
horn
beam horn
FIGURE 3.11 Simplified block diagram of an INTELSAT V communication system. Note
that the switch matrix allows many possible interconnections between uplink beams and
downlink transmitters. (Courtesy C. F. Hoeber, Ford Aerospace and Communications Corp.)
Figure 3.11 shows a simplified diagram of the communication system carried by

INTELSAT V satellites. The later series of Intelsat satellites use a similar arrangement.
The bulk of the traffic is carried by the 64 GHz section, with a total bandwidth of 2000
MHz available by frequency reuse. The switch matrix allows a very large number of vari-
ations in connecting the 6-GHz receivers to the 4-GHz transmitters, and also intercon-
nects the 64 and 1411-GHz sections. This provides Intelsat with a great deal of flexibility
in setting up links through the satellite.
When more than one signal shares a transponder (using frequency division multiple
access, FDMA) the power amplifier must be run below its maximum output power to
maintain linearity and reduce intermodulation products. The degree to which the trans-
mitter output power is reduced below its peak output is known as output backoff: in FDMA
systems, 2 to 7 dB of output backoff is typically used, depending on the number of ac-
cesses to the transponder and the extent to which the characteristics of the HPA have been
linearized. Backoff results in a lower downlink carrier-to-noise ratio at the earth station
with FDMA when multiple accesses to each transponder are required. Time division mul-
tiple access (TDMA) can theoretically be used to increase the output power of transpon-
ders by limiting the transponder to a single access. However, most TDMA systems are
hybrid FDMA–TDMA schemes known as multifrequency TDMA (MF-TDMA), in which
several TDMA signals share the transponder bandwidth using FDMA. Linearity of the
HPA remain an issue for MF-TDMA systems.
c03.qxd 28/08/02 19:35 Page 78
6 GHz 4 GHz 4 GHz 4 GHz

LNA BPF Downconverter BPF LPA HPA
6 GHz uplink 4 GHz downlink

Local
antenna antenna
oscillator
2225 MHz
FIGURE 3.12 Simplified single conversion transponder (bent pipe) for 64 GHz band.
Figure 3.12 shows a typical single conversion bent pipe transponder of the type used
on many satellites for the 64 GHz band. The output power amplifier is usually a solid
state power amplifier (SSPA) unless a very high output power (50 W) is required, when
a traveling wave tube amplifier (TWTA) would be used9. The local oscillator is at 2225
MHz to provide the appropriate shift in frequency from the 6-GHz uplink frequency to the
4-GHz downlink frequency, and the band-pass filter after the mixer removes unwanted fre-
quencies resulting from the down-conversion operation. The attenuator can be controlled
via the uplink command system to set the gain of the transponder. Redundancy is provided
for the high-power amplifiers (HPA) in each transponder by including a spare TWT or
solid-state amplifier (SSPA) that can be switched into circuit if the primary power ampli-
fier fails. The lifetime of HPAs is limited, and they represent the least reliable component
in most transponders. Providing a spare HPA in each transponder greatly increases the
probability that the satellite will reach the end of its working life with all its transponders
still operational. Transponders can also be arranged so that there are spare transponders
available in the event of a total failure. The arrangement is known as M for N redundancy.
For example, it is common to have 16 for 10 redundancy or even 14 for 10. That is, 16 (or
14) output amplifiers are connected in a ring such that any of the 10 signals can pass
through them. Thus, 6 (or 4) amplifiers are acting as back-up amplifiers while 10 are on
line. Most HPAs have bandwidths much larger than the allocated frequency band and so
it matters little which signals are passing through them. At Ku band, ring redundancy is
still used, but it is much more like 2 for 1, that is, one spare for every active unit.
Transponders for use in the 1411-GHz bands normally employ a double frequency
conversion scheme as illustrated in Figure 3.13. It is easier to make filters, amplifiers, and
equalizers at an intermediate frequency (IF) such as 1100-MHz than at 14 or 11 GHz, so
the incoming 14-GHz carrier is translated to an IF of around 1 GHz. The amplification
and filtering are performed at 1 GHz and a relatively high-level carrier is translated back
to 11 GHz for amplification by the HPA.
Stringent requirements are placed on the filters used in transponders, since they must
provide good rejection of unwanted frequencies, such as intermodulation products, and also
have very low amplitude and phase ripple in their pass bands. Frequently a filter will be
followed by an equalizer that smoothes out amplitude and phase variations in the pass band.
Phase variation across the pass band produces group delay distortion, which is particularly
troublesome with wideband FM signals and high-speed phase shift keyed data transmissions.
A considerable increase in the communications capacity of a satellite can be achieved
by combining onboard processing with switched-beam technology. A switched-beam satel-
lite generates a narrow transmit beam for each earth station with which it communicates,
and then transmits sequentially to each one using time division multiplexing of the signals.
The narrow beam has to cover only one earth station, allowing the satellite transmit antenna
to have a very high gain compared to a zone-coverage antenna. A narrow scanning beam
can also be used, or a combination of fixed and scanning beams. Unless the satellite has a
c03.qxd 28/08/02 19:35 Page 79
14 GHz 1 GHz 1 GHz

LNA BPF Downconverter BPF IF amplifier
13 GHz
14 GHz
Local
uplink
oscillator
antenna
10 GHz
Local 1 GHz 11 GHz 11 GHz
oscillator BPF HPA downlink
antenna
Upconverter
FIGURE 3.13 Simplified double conversion transponder (bent pipe) for
1411 GHz band.
zone-coverage receiver antenna, data storage is required at the satellite since it communi-
cates with only one earth station at a time. The high gain antennas used in switched-beam
systems raise the EIRP (effective isotropically radiated power) of the satellite transmitter
and thus increase the capacity of the downlink. Switched beam systems on GEO satellites
work best at Ka band where the wavelength is short enough that the limited dimensions of
antennas on the satellite still allow beams of less than 0.4° beamwidth to be generated.
Multiple beam antennas with baseband processing transponders are used on GEO and LEO
satellites providing service to mobile terminals and handheld telephones10. The low gain
of the near omnidirectional antenna of a mobile earth station must be compensated by a
high gain antenna on the satellite, necessitating the use of multiple beam antennas.
It is possible to conserve uplink bandwidth by using different modulation techniques
on the uplink and downlink and by providing a baseband processor on the satellite. A high
level modulation such as 16-QAM with four bits per symbol can be used on the link be-
tween the satellite and a large earth station to improve bandwidth efficiency. This approach
has been adopted in the Astrolink and Spaceway 3020 GHz satellites16,17.
Onboard processing may also be used to advantage to switch between the uplink
access technique (e.g., MF-TDMA) and the downlink access technique (e.g., TDM) so
that small earth stations may access each other directly via the satellite. The processor can
provide the data storage needed for a switched-beam system and also can perform error
correction independently on the uplink and downlink. A typical arrangement of the com-
munication system for a satellite employing onboard processing is shown in Figure 3.14.
Multiple beam Multiple beam

uplink antenna downlink antenna
FEC Baseband FEC
Receiver Demodulator decoder Processor encoder Modulator Transmitter
Rx Demod FEC OBP FEC Mod Tx
On-board controller
FIGURE 3.14 Onboard processing transponder.
c03.qxd 28/08/02 19:35 Page 80
3.6 SATELLITE ANTENNAS
Basic Antenna Types and Relationships

Four main types of antennas are used on satellites. These are
1. Wire antennas: monopoles and dipoles.
2. Horn antennas.
3. Reflector antennas.
4. Array antennas.
Wire antennas are used primarily at VHF and UHF to provide communications for
the TTC&M systems. They are positioned with great care on the body of the satellite in
an attempt to provide omnidirectional coverage. Most satellites measure only a few wave-
lengths at VHF frequencies, which makes it difficult to get the required antenna patterns,
and there tend to be some orientations of the satellite in which the sensitivity of the
TTC&M system is reduced by nulls in the antenna pattern.
An antenna pattern is a plot of the field strength in the far field of the antenna
when the antenna is driven by a transmitter. It is usually measured in decibels (dB) be-
low the maximum field strength. The gain of an antenna is a measure of the antenna’s
capability to direct energy in one direction, rather than all around. Antenna gain is de-
fined in Chapter 4, Section 4.2. At this point, it will be used with the simple definition
given above. A useful principle in antenna theory is reciprocity. Reciprocity means that
an antenna has the same gain and pattern at any given frequency whether it transmits or
receives. An antenna pattern measured when receiving is identical to the pattern when
transmitting.
Figure 3.15 shows typical satellite antenna coverage zones. The pattern is frequently
specified by its 3-dB beamwidth, the angle between the directions in which the radiated
17°
Global beam Spot beams
Phased array Vertical

antenna polarization
Horizontal
polarization
Multiple spot beams

Orthogonally
and scanning beams
polarized beams
FIGURE 3.15 Typical satellite antenna patterns and coverage zones. The antenna for the
global beam is usually a waveguide horn. Scanning beams and shaped beams require
phased array antennas or reflector antennas with phased array feeds.
c03.qxd 28/08/02 19:35 Page 81
3.6 SATELLITE ANTENNAS 81
70
60
Latitude (degrees)
50
−7−6 −5 −4 −3 −2 −1
40
10 0 10 20 30
Longitude (degrees)
FIGURE 3.16 Typical coverage patterns for Intelsat satellites over the Atlantic Ocean.
(or received) field falls to half the power in the direction of maximum field strength. How-
ever, a satellite antenna is used to provide coverage of a certain area, or zone on the earth’s
surface, and it is more useful to have contours of antenna gain as shown in Figure 3.16.
When computing the signal power received by an earth station from the satellite, it
is important to know where the station lies relative to the satellite transmit antenna con-
tour pattern, so that the exact EIRP can be calculated. If the pattern is not known, it may
be possible to estimate the antenna gain in a given direction if the antenna boresight or
beam axis direction and its beamwidth are known.
Horn antennas are used at microwave frequencies when relatively wide beams are
required, as for global coverage. A horn is a flared section of waveguide that provides an
aperture several wavelengths wide and a good match between the waveguide impedance
and free space. Horns are also used as feeds for reflectors, either singly or in clusters.
Horns and reflectors are examples of aperture antennas that launch a wave into free space
from a waveguide. It is difficult to obtain gains much greater than 23 dB or beamwidths
narrower than about 10° with horn antennas. For higher gains or narrow beamwidths a
reflector antenna or array must be used.
Reflector antennas are usually illuminated by one or more horns and provide a larger
aperture than can be achieved with a horn alone. For maximum gain, it is necessary to
generate a plane wave in the aperture of the reflector. This is achieved by choosing a
reflector profile that has equal path lengths from the feed to the aperture, so that all the
energy radiated by the feed and reflected by the reflector reaches the aperture with the
same phase angle and creates a uniform phase front. One reflector shape that achieves this
with a point source of radiation is the paraboloid, with a feed placed at its focus. The
c03.qxd 28/08/02 19:35 Page 82
paraboloid is the basic shape for most reflector antennas, and is commonly used for earth
station antennas. Satellite antennas often use modified paraboloidal reflector profiles to
tailor the beam pattern to a particular coverage zone. Phased array antennas are also used
on satellites to create multiple beams from a single aperture, and have been used by Irid-
ium and Globalstar to generate up to 16 beams from a single aperture for their LEO mobile
telephone systems11.
Some basic relationships in aperture antennas can be used to determine the ap-
proximate size of a satellite antenna for a particular application, as well as the antenna
gain. More accurate calculations are needed to determine the exact gain, efficiency, and
pattern of a satellite antenna, and the interested reader should refer to one of the many
excellent texts in this field for details.12–14 The following approximate relationships will
be used here to guide the selection of antennas for a communications satellite.
An aperture antenna has a gain G given by
G hA 4pAl2 (3.1)
where A is the area of the antenna aperture in meters, is the operating wavelength in
meters, and A is the aperture efficiency of the antenna. The aperture efficiency A is not
easily determined, but is typically in the range 55 to 68% for reflector antennas with single
feeds, lower for antennas with shaped beams. Horn antennas tend to have higher effi-
ciencies than reflector antennas, typically in the range 65 to 80%. If the aperture is circular,
as is often the case, Eq. (3.1) can be written as
G hA 1pDl2 2 (3.2)
where D is the diameter of the circular aperture in meters.
The beamwidth of an antenna is related to the aperture dimension in the plane in
which the pattern is measured. A useful rule of thumb is that the 3 dB beamwidth in a
given plane for an antenna with dimension D in that plane is
u3 dB 75lD degrees (3.3)
where 3 dB is the beamwidth between half power points of the antenna pattern and D is
the aperture dimension in the same units as the wavelength . The beamwidth of a horn
antenna may depart from Eq. (3.2) quite radically. For example, a small rectangular horn
will produce a narrower beam than suggested by Eq. (3.2) in its E plane and a wider
beamwidth in the H plane.
Since both Eqs. (3.2) and (3.3) contain antenna dimension parameters, the gain and
beamwidth of an aperture antenna are related. For antennas with A 60%, the gain is
approximately
G 33,000 1u3 dB 2 2 (3.4)
where 3 dB is in degrees and G is not in decibels. If the beam has different beamwidths in
orthogonal planes, 3 dB should be replaced by the product of the two 3 dB beamwidths. Val-
ues of the constant in Eq. (3.3) vary between different sources, with a range 28,000 to 35,000.
The value 33,000 is typical for reflector antennas used in satellite communication systems.
EXAMPLE 3.6.1 Global Beam Antenna

The earth subtends an angle of 17° when viewed from geostationary orbit. What are the dimensions
and gain of a horn antenna that will provide global coverage at 4 GHz?
If we design our horn to give a circularly symmetric beam with a 3-dB beamwidth of 17°
using Eq. (3.2)
Dl 75 1u3 dB 2 4.4
c03.qxd 28/08/02 19:35 Page 83
At 4 GHz, 0.075 m, so D 0.33 m (just over 1 ft). If we use a circular horn excited
in the TE11 mode, the beamwidths in the E and H planes will not be equal and we may be forced
to make the aperture slightly smaller to guarantee coverage in the E plane. A corrugated horn
designed to support the HE hybrid mode has a circularly symmetric beam and could be used in this
application. Waveguide horns are generally used for global beam coverage. Reflector antennas are
not efficient when the aperture diameter is less than 8.
Using Eq. (3.4), the gain of the horn is approximately 100, or 20 dB, at the center of the
beam. However, in designing our communication system we will have to use the edge of beam gain
figure of 17 dB, since those earth stations close to the earth’s horizon, as viewed from the satellite,
are close to the 3 dB contour of the transmitted beam.
EXAMPLE 3.6.2 Regional Coverage Antenna

The continental United States (48 contiguous states) subtends an angle of approximately 6° 3°
when viewed from geostationary orbit. What dimension must a reflector antenna have to illuminate
half this area with a circular beam 3° in diameter at 11 GHz?
Can a reflector be used to produce a 6° 3° beam? What gain would the antenna have?
Using Eq. (3.2), we have for a 3° circular beam
Dl 753 25
and with 0.0272 m, D 0.68 m (just over 2 ft). The gain of this antenna, from Eq. (3.3) is
approximately 35 dB.
To generate a beam with different beamwidths in orthogonal planes we need an aperture with
different dimensions in the two planes. In this case, a rectangular aperture 25 12.5 would gen-
erate a beam 6° 3°, and would have a gain of 32 dB, approximately. In order to illuminate such
a reflector, a horn with unequal beamwidths is required, since the reflector must intercept most of
the radiation from the feed if it is to have an acceptable efficiency. Rectangular, or more commonly
elliptical, outline reflectors are used to generate unequal beamwidths. When orthogonal polariza-
tions are to be transmitted or received, it is better to use a circular reflector with a distorted profile
to broaden the beam in one plane, or a feed cluster to provide the appropriate amplitude and phase
distribution across the reflector.
Satellite Antennas in Practice

The antennas of a communications satellite are often a limiting element in the complete
system. In an ideal satellite, there would be one antenna beam for each earth station, com-
pletely isolated from all other beams, for transmit and receive. However, if two earth sta-
tions are 300 km apart on the earth’s surface and the satellite is in geostationary orbit,
their angular separation at the satellite is 0.5°. For 3 dB to be 0.5°, D must be 150,
which requires an aperture diameter of 11.3 m at 4 GHz. Antennas this large have been
flown on satellites (ATS-6 deployed a 2.5 GHz, 10-m diameter antenna, for example), and
large unfurled antennas are used to create multiple spot beams from GEO satellites serv-
ing mobile users. However, at 20 GHz, an antenna with D 150 is only 1.5 m wide,
and such an antenna can readily be flown on a 3020 GHz satellite. A phased array feed
is used to create many 0.5° beams which can be clustered to serve the coverage zone of
the satellite.
To provide a separate beam for each earth station would also require one antenna
feed per earth station if a multiple-feed antenna with a single reflector were used. A
compromise between one beam per station and one beam for all stations has been used
in many satellites by using zone-coverage beams and orthogonal polarizations within the
same beam to provide more channels per satellite. Figure 3.3b shows a GEO satellite that
c03.qxd 28/08/02 19:35 Page 84
West
East
spot beam
spot beam
West East
zone beam zone beam
West
East
hemi beam
hemi beam
Subsatellite point
FIGURE 3.17 Contour plot of the spot beam of ESA’s OTS satellite projected onto the
earth. The contours are in 1 dB steps, normalized to 0 dB at the center of the beam.
(Courtesy of ESA.)
has four reflector antennas. Each reflector is illuminated by a complex feed that provides
the required beam shape to permit communication between earth stations within a given
coverage zone. Figure 3.17 shows the coverage zones provided by a typical Intelsat satel-
lite. The largest reflector on the satellite transmits at 4 GHz and produces the “peanut”
shaped patterns for the zone beams, which are designed to concentrate the transmitted
energy onto densely populated areas such as North America and western Europe where much
telecommunications traffic is generated. The smaller antennas are used to provide hemi-
sphere transmit and receive beams, and the 1411 GHz spot beams. In addition, there are
horn antennas providing global beam coverage.
Countries such as the United States create an enormous demand for communica-
tion services, and a number of domestic satellite communication systems have been
established to meet that demand. In 2000 the geostationary orbit had domestic satellites
spaced every 2°, operating at 64 GHz and 1411 GHz from longitude 60° W to 140°
W. This encompasses all orbital locations that can be simultaneously viewed by earth
stations in the United States and Canada, and each operator has been given a limited
number of orbital slots in which to place a satellite. As a result, there is a great deal of
pressure on the operating companies to obtain the maximum number of channels per
satellite in order to give the operator the greatest possible revenue-earning capacity. This
has encouraged the development of frequency reuse antennas by means of orthogonal
polarizations and multiple beams, the combination of 64 and 1411 GHz communication
systems on one satellite, and the use of multilevel digital modulation and TDMA to in-
crease capacity.
SIDEBAR
Some of the largest commercial GEO satellites pro- 200 spot beams for mobile services and Internet ac-
posed to date are the Inmarsat I-4 series19. These satel- cess. Three Inmarsat I-4 satellites are to be built by
lites are based on the Matra Marconi Space Eurostar the European consortium Astrium at a total cost of
3000 platform, a 5 metric ton three-axis stabilized $700 M, with two scheduled for launch in 2003–2004.
satellite with 9 kW onboard power. The satellites have The third satellite is a spare.
c03.qxd 28/08/02 19:35 Page 85
FIGURE 3.18 Intelsat VI satellite on station.
The requirements of narrow antenna beams with high gain over a small cover-
age zone leads to large antenna structures on the satellite. Frequently, the antennas in
their operating configuration are too large to fit within the shroud dimensions of the
launch vehicle, and must be folded down during the launch phase. Once in orbit, the
antennas then can be deployed. In many larger satellites, the antennas use offset pa-
raboloidal reflectors with clusters of feeds to provide carefully controlled beam shapes.
The feeds mount on the body of the satellite, close to the communications subsystem,
and the reflector is mounted on a hinged arm. Figure 3.18 shows an example of this
design of antenna for the INTELSAT VI satellite. For launch, the solid reflectors fold
down to provide a compact structure; in orbit, the hinged arms are swung out and
locked in place to hold the reflectors in the correct position. When the satellite is in
geostationary orbit it is weightless, so very little energy is required to move the large
reflector.
SIDEBAR
One interesting idea is the inflatable antenna, sev- can be sprayed with a metallic coating made up of
eral examples of which have been flown experi- very small particles of aluminum to create a re-
mentally18. The antenna can be squeezed into a small flecting surface for electromagnetic (EM) waves.
space for launch and inflated from a pressurized gas Inflatable antennas can be made very large without
bottle when the satellite is in orbit. Once inflated, a a significant weight penalty, and are therefore at-
foam material emitted along with the inflation gas tractive for any satellite requiring multiple narrow
hardens to make a rigid structure. Plastic materials beams.
c03.qxd 28/08/02 19:35 Page 86
ATS-6
in orbit
1. After separation 2. Solar array booms extended 3. Solar array panels extended
4. 30-ft reflector deploys 5. Fully deployed configuration

FIGURE 3.19 Deployment sequence of ATS-6 10-m antenna. (Courtesy of NASA.)
Figure 3.19 shows the deployment sequence used for the 30-ft antenna carried by
ATS-6: the antenna was built as a series of petals that folded over each other to make a
compact unit during launch, which then unfurled in orbit. The solar sails folded down
over the antenna, and were deployed first. Springs or pyrotechnic devices can be used to
provide the energy for deployment of antennas or solar sails, with a locking device to en-
sure correct positioning after deployment. Similar unfurlable antennas are used on GEO
satellites that provide satellite telephone service at L band using multiple narrow beams.
c03.qxd 28/08/02 19:35 Page 87
3.7 EQUIPMENT RELIABILITY AND SPACE QUALIFICATION 87
3.7 EQUIPMENT RELIABILITY AND

SPACE QUALIFICATION
Communications satellites built in the 1980s and 1990s have provided operational lifetimes
of up to 15 years. Once a satellite is in geostationary orbit, there is little possibility of re-
pairing components that fail or adding more fuel for station keeping. The components that
make up the satellite must therefore have very high reliability in the hostile environment
of outer space, and a strategy must be devised that allows some components to fail without
causing the entire communication capacity of the satellite to be lost. Two separate ap-
proaches are used: space qualification of every part of the satellite to ensure that it has a
long life expectancy in orbit and redundancy of the most critical components to provide
continued operation when one component fails.
Space Qualification
Outer space, at geostationary orbit distances, is a harsh environment. There is a total vac-
uum and the sun irradiates the satellite with 1.4 kW of heat and light on each square me-
ter of exposed surface. Where surfaces are in shadow, heat is lost to the infinite sink of
space, and surface temperature will fall toward absolute zero. Electronic equipment can-
not operate at such extremes of temperature and must be housed within the satellite and
heated or cooled so that its temperature stays within the range 0° to 75°C. This requires
a thermal control system that manages heat flow throughout a GEO satellite as the sun
moves around once every 24 h. Thermal problems are equally severe for a LEO satellite
that moves from sunlight to shadow every 100 min.
The first stage in ensuring high reliability in a satellite is by selection and screen-
ing of every component used. Past operational and test experience of components indi-
cates which components can be expected to have good reliability. Only components that
have been shown to have high reliability under outer space conditions will be selected.
Each component is then tested individually (or as a subsystem) to ensure that it meets its
specification. This process is known as quality control or quality assurance and is vital
in building any equipment that is to be reliable. Once individual components and sub-
systems have been space qualified, the complete satellite must be tested as a system to
ensure that its many systems are reliable.
When a satellite is designed, three prototype models are often built and tested. The
mechanical model contains all the structural and mechanical parts that will be included
in the satellite and is tested to ensure that all moving parts operate correctly in a vacuum,
over a wide temperature range. It is also subjected to vibration and shock testing to sim-
ulate vibration levels and G forces likely to be encountered on launch. The thermal model
contains all the electronics packages and other components that must be maintained at the
correct temperature. Often, the thermal, vacuum, and vibration tests of the entire satellite
will be combined in a thermal vacuum chamber for what is known in the industry as a
shake and bake test. The antennas are usually included on the thermal model to check for
distortion of reflectors and displacement or bending of support structures. In orbit, an an-
tenna may cycle in temperature from above 100°C to below 100°C as the sun moves
around the satellite. The electrical model contains all the electronic parts of the satellite
and is tested for correct electrical performance under total vacuum and a wide range of
temperatures. The antennas of the electrical model must provide the correct beamwidth,
gain, and polarization properties.
Testing carried out on the prototype models is designed to overstress the system and
induce failure in any weak components: temperature cycling will be carried out to 10%
beyond expected extremes; structural loads and G forces 50% above those expected in
c03.qxd 28/08/02 19:35 Page 88
flight may be applied. Electrical equipment will be subjected to excess voltage and cur-
rent drain to test for good electronic and thermal reliability. The prototype models used
in these tests will not usually be flown. A separate flight model (or several models) will
be built and subjected to the same tests as the prototype, but without the extremes of tem-
perature, stress, or voltage. Preflight testing of flight models, while exhaustive, is designed
more to cause failure of parts, rather than to check that they will operate under worst-case
conditions.
Space qualification is an expensive process, and one of the factors that makes large
GEO satellites expensive. Some low earth orbit satellites have been built successfully us-
ing less expensive techniques and relying on lower performance in orbit. LEO satellite
systems require large numbers of satellites that are generally less expensive than large
GEO satellites. The Iridium system, for example, was designed with 66 operational satel-
lites in its constellation to provide continuous worldwide coverage, with at least eight
spare satellites in orbit at any time. If one operational satellite fails, a spare is moved in
to take its place. This allowed Iridium satellites to be built with a higher probability of
failure than a GEO satellite. Experimental satellites have also been built using low cost
techniques. The University of Surrey, U.K., for example, built a series of digital store and
forward satellites that were used by radio amateurs and others which each cost less than
$1 M15. Most of the components on the Surrey satellites were not space qualified, but
were selected carefully to ensure best possible lifetimes at reasonable cost and then the
entire satellite was subjected to shake and bake tests.
Many of the electronic and mechanical components that are used in satellites are
known to have limited lifetimes, or a finite probability of failure. If failure of one of these
components will jeopardize the mission or reduce the communication capacity of the satel-
lite, a backup, or redundant, unit will be provided. The design of the system must be such
that when one unit fails, the backup can automatically take over or be switched into op-
eration by command from the ground. For example, redundancy is always provided for
traveling wave tube amplifiers used in the transponders of a communications satellite, as
these are known to have a limited lifetime.
The success of the testing and space qualification procedures used by NASA has
been well illustrated by the lifetime achieved by many of its scientific satellites. Satellites
designed for a specific mission lasting 1 or 2 years have frequently operated successfully
for up to 25 years. Sufficient reliability was designed into the satellite to guarantee the
mission lifetime such that the actual lifetime has been much greater. In the next section
we will look at how reliability can be quantified.
Reliability
We need to be able to calculate the reliability of a satellite subsystem for two reasons:
we want to know what the probability is that the subsystem will still be working after
a given time period, and we need to provide redundant components or subsystems where
the probability of a failure is too great to be accepted. The owner of a satellite used for
communications expects to be able to use a predetermined percentage of its communi-
cations capacity for a given length of time. Amortization of purchase and launch costs
will be calculated on the basis of an expected lifetime. The manufacturers of satellites
must provide their customers with predictions (or guarantees) of the reliability of the
satellite and subsystems: to do this requires the use of reliability theory. Reliability the-
ory is a mathematical attempt to predict the future and is therefore less certain than
other mathematical techniques that operate in absolute terms. The application of relia-
bility theory has enabled satellite engineers to build satellites that perform as expected,
c03.qxd 28/08/02 19:35 Page 89
End of
life
Burn-in
Probability
of
failure
0
Time
FIGURE 3.20 Bathtub curve for probability of failure.
at acceptable construction costs. It should be noted, however, that the cost of a satellite
is very high compared to other equipment with a comparable number of components: a
large GEO satellite costs around $100 M to build, close to the cost of a Boeing 747 jet
airliner. The cost is acceptable because of the high revenue-earning capability of the
satellite.
The reliability of a component can be expressed in terms of the probability of fail-
ure after time t, PF(t). For most electronic equipment, probability of failure is higher at
the beginning of life—the burn-in period—than at some later time. As the component
ages, failure becomes more likely, leading to the bathtub curve shown in Figure 3.20.
Components for satellites are selected only after extensive testing. The aim of
the testing is to determine reliability, causes of failure, and expected lifetime. The re-
sult is a plot similar to Figure 3.20. Testing is carried out under rigorous conditions,
representing the worst operating conditions likely to be encountered in space, and may
be designed to accelerate failure in order to shorten the testing duration needed to
determine reliability. Units that are exposed to the vacuum of space are tested in a vac-
uum chamber, and components subjected to sunlight are tested under equivalent radi-
ant heat conditions. The initial period of reduced reliability can be eliminated by a
burn-in period before a component is installed in the satellite. Semiconductors and
integrated circuits that are required to have high reliability are subjected to burn-in
periods from 100 to 1000 h, often at a high temperature and excess voltage to induce
failures in any suspect devices and to get beyond the initial low reliability part of the
bathtub curve.
SIDEBAR
The bathtub curve is familiar to owners of auto- out to replace parts that are known to wear most
mobiles. A new car may have defects when it is de- quickly. For example, spark plugs and drive belts
livered, and errors in manufacturing may lead to may be replaced every 40,000 miles. The skill in
components failing soon after purchase. This is one owning an automobile is to judge the time at which
of the reasons that manufacturers offer warranties for the vehicle is starting up the end-of-life portion of
the initial life of their products. Once these defects the bathtub curve. That is the optimum time to dis-
have been overcome, by repair or replacement, reli- pose of the vehicle, and the worst possible time to
ability improves for a number of years until me- buy one. Preventive maintenance is not possible with
chanical parts start to wear and failures occur. In an a geostationary satellite, at present, so other strate-
automobile, preventive maintenance can be carried gies must be adopted.
c03.qxd 28/08/02 19:35 Page 90
The reliability of a device or subsystem is defined as

Ns 1t2 Number of surviving components at time t
R1t2 (3.5)
N0 Number of components at start of test period
The numbers of components that failed in time t is Nf (t) where
Nf 1t2 N0 Ns 1t2 (3.6)
From the engineering viewpoint, what we need to know is the probability of any
one of the N0 components failing: this is related to the mean time before failure (MTBF).
Suppose we continue testing devices until all of them fail. The ith device fails after time
ti where
1 N0
MTBF m a ti (3.7)
N0 i1
The average failure rate , is the reciprocal of the MTBF, m. If we assume that is a
constant, then
Number of failures in a given time
l
Number of surviving components
1 ¢Nf 1 dNf
l 1 MTBF (3.8)
Ns ¢t Ns dt
Failure rate is often given as the average failure rate per 109 h. The rate of failure, dNfdt,
is the negative of the rate of survival dNsdt, so we can redefine as
1 dNs
l (3.9)
Ns dt
By definition from Eq. (3.5), the reliability R is NsN0, so
1 d 1 dR
l 1N0 R2 (3.10)
N0 R dt R dt
A solution of Eq. (3.10) is
R elt (3.11)
Thus the reliability of a device decreases exponentially with time, with zero relia-
bility after infinite time, that is, certain failure. However, end of useful life is usually taken
to be the time tl, at which R falls to 0.37(1e), which is when
tl 1 l m (3.12)
The probability of a device failing, therefore, has an exponential relationship to the
MTBF and is represented by the right-hand end of the bathtub curve.
Redundancy
The equations in the preceding section allow us to calculate the reliability of a given de-
vice when we know its MTBF. In a satellite, many devices are used, each with a different
MTBF, and failure of one device may cause catastrophic failure of a complete subsystem.
If we incorporate redundant devices, the subsystem can continue to function correctly. We
can define three different situations for which we want to compute subsystem reliability:
c03.qxd 28/08/02 19:35 Page 91
R1 R2 R3 RN
(a)
R1
R2
(b )
R1 R2
R3 R4
(c)
R1
S1 R2 S2
FIGURE 3.21 Redundancy connections.

R3 (a) Series connection. (b) Parallel connec-
tion. (c) Series/parallel connection.
(d ) (d) Switched connection.
series connection, used in solar cells arrays, parallel connection, used to provide redun-
dancy of the high power amplifiers in satellite transponders, and a switched connection,
often used to provide parallel paths with multiple transponders. These are illustrated in
Figure 3.21; also shown is a hybrid arrangement, a series/parallel connection, widely used
in electronic equipment.
The switched connection arrangement shown in Figure 3.21d is also referred to as
ring redundancy since any component can be switched in for any other. Switches S1 and
S2 are a little more complicated than as shown, affording the choice of multiple paths in
an M for N ring redundancy configuration. The important point to note is that the active
devices (R1, R2, p , Rn) have sufficient bandwidth, power output range, etc., to be able
to handle any of the channels that might be switched through to them. Most TWTAs and
SSPAs are such wideband, large power range devices.
An example of parallel redundancy for the HPA of a 64 GHz bent pipe transpon-
der is shown in Figure 3.22. The transponder translates incoming signals in the 6-GHz
uplink band by 2225 MHz and retransmits them in the 4-GHz band. The high-power out-
put stage of the transponder has two parallel TWT amplifiers. One TWTA will be switched
off, but must present a matched load when both off and on. If one TWTA fails, the other
is switched on either automatically, or by command from earth. The TWT is a thermionic
device with a heated cathode and a high voltage power supply. In common with other
c03.qxd 28/08/02 19:35 Page 92
TWTA
#1
6 GHz 4 GHz 4 GHz Combiner

LNA BPF Downconverter BPF LPA
S C
6 GHz Local 4 GHz

Splitter downlink
uplink oscillator
antenna 2225 MHz antenna
Redundant TWTA
4 GHz HPAs #2
FIGURE 3.22 Redundant TWTA configuration in HPA of a 64 GHz bent

pipe transponder.
thermionic devices such as cathode ray tubes and magnetrons, they have a relatively short
MTBF. Although the MTBF may be 50,000 h, this is the period after which 50% of such
devices will have failed, on average.
The parallel connection of two TWTs, as shown in Figure 3.22 raises the reliabil-
ity of the amplifier stage to 0.60 at the MTBF period, assuming zero probability of a short
circuit. A lifetime of 50,000 h is approximately 6 years of continuous operation, which is
close to the typical design lifetime of a satellite. To further improve the reliability of the
transponders, a second redundant transponder may be provided with switching between
the two systems. Note that a combination of parallel and switched redundancy is used to
combat failures that are catastrophic to one transponder channel and to the complete com-
munication system.
3.8 SUMMARY
Satellites that carry communications relays must Satellites often employ frequency reuse, either by
provide a stable platform in orbit. Large GEO satel- using the same frequencies again in spatially sepa-
lites have payload design lives that exceed 10 years rated beams or by using the same frequencies in
and sufficient fuel to provide a maneuvering life- orthogonal polarizations within the same beam.
time that typically exceeds 15 years. The satellite Sometimes, both reuse techniques are used simulta-
must carry a number of subsystems to support its neously. Frequency reuse allows the same RF spec-
communications mission. The attitude and orbital trum to be used more than once to increase the
control system keeps the satellite in the correct orbit satellite’s capacity.
and on station, and pointing in the correct direction. Antennas are a limiting factor in all radio commu-
The telemetry, tracking, command, and monitoring nication systems. Very complex antennas have been
system allows an earth station to control the sub- developed for satellites to provide multiple beams and
systems in the satellite and to monitor their health. orthogonal polarizations from a single antenna.
The power system provides the electrical energy Reflector antennas with clustered feeds and phased
needed to run the satellite (housekeeping) and the array antennas are used to generate shaped and mul-
communications system. Solar cells generate the tiple beams. Reliability is an important issue in satel-
electrical power, a power conditioning unit controls lites. Redundancy can be used to provide additional
its distribution, and batteries provide power during receivers and high-power amplifiers that can take over
launch and eclipses. when a unit fails.
c03.qxd 28/08/02 19:35 Page 93
PROBLEMS 93
REFERENCES
1. WALTER L. MORGAN, GARY D. GORDON, Communication 9. J. WOLKENSTEIN and J. N. LAPADRE, “Solid State Power
Satellites Handbook, Wiley Interscience, New York, 1989. Amplifiers Replacing TWTs in C-Band Satellites,” RCA
2. IEEE Transactions on Aerospace and Electronic Systems, Engineer, 27, 7, October/November 1982.
AS-1 through AS-37, 1963–2000, Institute of Electronic 10. ROY MAUGER and CATHERINE ROSENBERG, “QoS Guar-
and Electrical Engineers, New York. antees for Multimedia Services on a TDMA-based Satel-
3. “Astronautics and Aeronautics,” Journal of the American lite Network,” IEEE Communications Society Magazine
Institute of Aeronautics and Astronautics, Vol. 1 through Vol. 35, No. 7, July 1997.
37, 1963–2000, AIAA, 1633 Broadway, New York. 11. www.globalstar.com
4. J. ALPER and J. N. PELTON, eds., “The Intelsat Global 12. W. L. STUTZMAN and G. A. THIELE, Antenna Theory and
Satellite System,” Progress in Astronautics and Aero- Design, John Wiley & Sons, New York, 1981.
nautics, Vol. 93, The American Institute of Aeronautics 13. A. W. RUDGE, K. MILNE, A. D. OLVER, and P. KNIGHT,
and Astronautics, New York, 1984. Handbook of Antenna Design, Volume 1, IEE Electro-
5. KEIGLER, W. LINDORTER, and L. MUHLFELDER, “Stable magnetic Wave Series No. 15, Peter Perigrinus Ltd.,
Attitude Control for Synchronous Communications Satel- Stevenage, Herts, UK, 1983.
lites,” in Progress in Astronautics and Aeronautics, 33, 14. S. SILVER, ed., Microwave Antenna Theory and Design,
MIT Press, Cambridge, MA, 1974. originally published as Vol. 12 of the MIT Radiation
6. G. E. SCHMIDT, Jr., “Magnetic Attitude Control for Geo- Laboratory Series, 1949. (Republished by Peter Peri-
synchronous Satellite,” Proceedings of the AIAA Com- grinus Ltd., Stevenage, Herts, UK, 1984.)
munications Satellite Systems Conference, San Diego, 15. www.ee.surrey.ac.uk/SSC/SSHP/sshp.html
CA, April 1978, pp. 110–112. 16. www.astrolink.com
7. E. W. SCHMIDT, Hydrazine and Its Derivatives, Wiley, 17. www.hns.com.spaceway
New York, 1984. 18. “Inflatable Structures Taking to Flight,” Aviation Week,
8. W. H. BRAUN and J. E. KEIGLER, “RCA Satellite Networks: pp. 60–62, January 25, 1999.
High Technology and Low User Cost,” Proceedings of the 19. “Astrium Signs Inmarsat Deal,” Aviation Week, p. 39,
IEEE, 72, 1483–1505, November 1984. May 22, 2000.
PROBLEMS
1. The telemetry system of a geostationary commu- and that the effective solar radiation absorbing area of
nications satellite samples 100 sensors on the space- the solar cells is equal to the cross-sectional area of
craft in sequence. Each sample is transmitted to earth the drum.
as an 8-bit word in a TDM frame. An additional 200 b. If the solar cells degrade by 15% over the lifetime
bits are added to the frame for synchronization and of the satellite, so that the end-of-life output power
status information. The data are then transmitted at a is 85% of the beginning-of-life output power, what
rate of 1 kilobit per second using BPSK modulation is the output of the solar cells immediately after
of a low-power carrier. launch?
a. How long does it take to send a complete set of c. If the drum covered in solar cells of the spinner
samples to earth from the satellite? design had been replaced by solar sails that rotated
b. Including the propagation delay, what is the longest to face the sun at all times, what area of solar sails
time the earth station operator must wait between would have been needed? Assume that cells on solar
a change in a parameter occurring at the spacecraft sails generate only 90% of the power of cells on a
and the new value of that parameter being received spinner due to their higher operating temperature.
via the telemetry link? (Assume a path length of 3. A Direct Broadcast Television (DBS-TV) satellite
40,000 km.) is in geostationary orbit. The electrical power required
2. A spinner satellite has solar cells wrapped round a to operate the satellite and its transmitters is 4 kW.
cylindrical drum 3.00 m in diameter, with a height of Two designs of satellite can be used: three axis
5.0 m on station. The drum is rotated at 60 rpm to stabilized with solar cells and a spinner.
spin-stabilize the satellite. At the end of life, the solar a. A three-axis stabilized satellite has two solar sails
cells are required to deliver 4.0 kW of electrical power. of equal area that rotate to face the sun at all times.
a. Calculate the efficiency of the solar cells at end The efficiency of the solar cells at end of life is pre-
of life. Assume an incident solar power of 1.39 kW/m2, dicted to be 15%.
c03.qxd 28/08/02 19:35 Page 94
Calculate the area of cells required by the GEO satel- 6. A geostationary satellite provides communications
lite, and the length of each sail if its width is 2.0 m. within the United States at Ku band. The antennas on
b. A spinner design of DBS-TV satellite is made up the satellite have beamwidths of 6° in the E–W
from a drum coated in solar cells. The drum has a di- direction and 3° in the N–S direction. A separate an-
ameter of 3.5 m. The efficiency of the solar cells is tenna is used for transmitting in the 11 GHz band and
predicted to be 18% at end of life. Since half the so- receiving in the 14 GHz band.
lar cells are in darkness, and some are weakly illu- a. Find the dimensions and estimate the gain of the
minated by the sun, the effective area of solar cells transmitting antenna in the N–S and E–W directions.
on a spinner is equal to the diameter of the satellite b. Find the dimensions and estimate the gain of the
multiplied by the height of the solar cells on the drum. receiving antenna in the N–S and E–W directions.
Calculate the height of the drum to provide the
6. The state of Virginia can be represented approxi-
required 4 kW of electrical power.
mately on a map as an area bounded by 39.5° N
4. Batteries make up a significant part of the in-orbit latitude, 36.5° N latitude, 76.0° W longitude, and
weight of a communications satellite but are needed 83.0° W longitude. A geostationary satellite located
to keep the communications system operating during at 79.5° W longitude has an antenna with a spot beam
eclipses. A direct broadcast TV satellite requires 500 W that covers all of Virginia at a downlink center fre-
of electrical power to operate the housekeeping quency of 11.155 MHz. In this problem you will
functions of the satellite and 5 kW to operate its 16 estimate the antenna dimensions subject to two dif-
high-power transponders. The longest duration of an ferent assumptions. In both cases use an aperture
eclipse is 70 min, during which time the batteries must efficiency of 65%.
provide power to keep the satellite operating, but the a. The antenna is a circular parabolic reflector gen-
batteries must not discharge below 70% of their erating a circular beam with a 3-dB beamwidth equal
capacity. The satellite bus operates at 48 V. to the diagonal of the area bounding the state of
a. What is the current that must be supplied by the Virginia. Estimate the length of the diagonal by meas-
power conditioning unit to keep the satellite operating uring the distance on a map of the United States, and
normally? calculate the beamwidth of the antenna from simple
b. Battery capacity is rated in ampere hours, the geometry. Hence determine the diameter of the an-
product of the current (in amps) that the battery can tenna on the satellite in meters and its approximate
supply multiplied by the length of time that this gain in decibels.
current can be supplied before the battery is fully b. The antenna is an elliptical parabolic reflector with
discharged. The satellite batteries must not discharge 3-dB beamwidths in the N–S and E–W directions that
beyond 70% of their rated capacity during eclipse. are equal to the height and the width of the area
Find the battery capacity required for this DBS-TV bounding the state of Virginia. Find the N–S and E–W
satellite. dimensions from a map of the United States, and use
c. If batteries weigh 1.25 kg per ampere-hour of ca- geometry to calculate the required 3-dB beamwidths
pacity, how much weight on this satellite is devoted of the satellite antenna. Calculate the approximate
to batteries? gain of the antenna.
d. If half of the transponders are shut down during 7. The state of Pennsylvania is approximately 1° wide
eclipse, what saving in battery weight is achieved? (E–W) by one-half degree high (N–S) when viewed
5. A geostationary satellite provides service to a re- from geostationary orbit at a longitude of 75° W. Cal-
gion which can be covered by the beam of an antenna culate:
on the satellite with a beamwidth of 1.8°. The satel- a. The dimensions of a downlink Ku-band antenna on
lite carries transponders for Ku band and Ka band, a geostationary satellite with 3-dB beamwidths equal
with separate antennas for transmit and receive. For to the width and height of Pennsylvania. Use a fre-
center frequencies of 14.011.5 GHz and 30.020.0 quency of 11.0 GHz. Identify the dimensions as E–W
GHz, determine the diameters of the four antennas on and N–S.
the satellite. b. The dimensions of an uplink Ka-band antenna on a
a. Find the diameters of the two transmitting anten- geostationary satellite with 3-dB beamwidths equal to
nas. Specify the diameter and calculate the gain at the width and height of Pennsylvania. Use a frequency
each frequency. of 30.0 GHz. Identify the dimensions as E–W and N–S.
b. Find the diameters of the two receiving antennas. c. Suppose that the maximum dimension of the satel-
Specify the diameter and calculate the gain at each lite at launch is 3 m wide, determined by the shroud
frequency. of the ELV. Describe in a paragraph how you would
c03.qxd 28/08/02 19:35 Page 95
PROBLEMS 95
launch the satellites in (a) and (b) above carrying: (a) c. Assuming that all 16 beams from the satellite anten-
the Ku-band antenna, and (b) the Ka-band antenna. nas have equal beamwidths, determine the beamwidth
8. A constellation of low earth orbit satellites has an of one beam. (Hint: Draw a circle representing the cov-
altitude of 1000 km. Each satellite has two multiple erage area and fit 16 circles representing the 3-dB
beam antennas that generate 16 beams. One antenna beamwidths of the beams inside the first circle.)
is used to transmit at 2.4 GHz and the other antenna d. Find the gain and the dimensions of each antenna
receives at 1.6 GHz. on the satellite.
a. Using simple geometry, find the coverage angle of 9. Calculate the total power radiated by the sun in
the satellite antenna when the lowest elevation angle watts and in dBW.
for an earth station is 10°. (Hint: Draw a diagram of Hint: The sun is 93 million miles (about 150 mil-
the earth and the satellite and use the law of sines to lion kilometers) from the earth. At that distance, the
solve the angles in a triangle.) sun produces a flux density of 1.39 kW/m2. This
b. Estimate the coverage area over the surface of the power density is present over all of a sphere with a
earth, in km. radius of 150 million km.
c04.qxd 28/08/02 19:37 Page 96
CHAPTER 4
SATELLITE LINK DESIGN
The design of a satellite communication system is a complex process requiring compro-

mises between many factors to achieve the best performance at an acceptable cost. We
will first consider geostationary satellite systems, since GEO satellites carry the vast
majority of the world’s satellite traffic.
4.1 INTRODUCTION
Figure 2.17 of Chapter 2 shows that the cost to build and launch a GEO satellite is about
$25,000 per kg. Weight is the most critical factor in the design of any satellite, since the
heavier the satellite the higher the cost, and the capital cost of the satellite must be re-
covered over its lifetime by selling communication services. The overall dimensions of
the satellite are critical because the spacecraft must fit within the confines of the launch
vehicle. When stowed for launch, the diameter of the spacecraft typically must be less
than 3.5 m. Most large GEO satellites use deployable solar panels and antennas, but the
antenna reflectors require accurate surfaces and are not folded for launch. This limits the
maximum aperture dimension to about 3.5 m. As in most radio systems, antennas are a
limiting factor in the capacity and performance of the communication system.
The weight of a satellite is driven by two factors: the number and output power of
the transponders on the satellite and the weight of station-keeping fuel. As much as half
the total weight of satellites intended to remain in service for 15 years may be fuel. High
power transponders require lots of electrical power, which can only be generated by solar
cells. Increasing the total output power of the transponders raises the demand for electrical
power and the dimensions of the solar cells, adding more weight to the satellite.
Three other factors influence system design: the choice of frequency band, atmos-
pheric propagation effects, and multiple access technique. These factors are all related, with
the frequency band often being determined by what is available. Tables 4.1 and 4.2 tabulate
SIDEBAR
The information carrying capacity of any radio com- operation has been toward larger satellites with high
munication link is determined by the RF power at the output powers and large antennas and smaller earth
receiver input. Large antennas are needed to receive stations, exemplified by VSAT networks and DBS-
weak signals, and the signals from satellites in geo- TV. Direct to home satellite television broadcasting
stationary orbit are invariably weak. Early satellites requires millions of receiving terminals, so these are
were small and light, carrying small antennas and made small and low cost, with the result that the DBS-
transponders with low output powers, which resulted TV satellites are large and expensive. The designer of
in very weak signals at the earth’s surface. The earth a satellite communication system must work to min-
stations required for communication with these satel- imize the capital cost of the entire system, and must
lites were large and expensive, with 25-m diameter also ensure that sufficient revenue can be earned from
antennas and high-power transmitters. The trend the system to recover the large capital cost of build-
over the 35 years that GEO satellites have been in ing and launching satellites.
96
c04.qxd 28/08/02 19:37 Page 97
4.1 INTRODUCTION 97
TABLE 4.1 Major Frequency Allocations for Fixed Satellite Service and
Broadcasting Satellites
Frequency Fixed satellite service Broadcasting satellites
2320–2345 MHz Radio broadcasting

2500–2535 Down region III
2500–2655 Down region II Down region II
2655–2690 Down region II Down region II
Up region II, III
3400–3700 Down
3700–4200 Down
4500–4800 Down, up
5725–5850 Up region I
5850–5925 Up region I
5925–7075 Up
7250–7450 Down, government
8215–8400 Up, government
10.7–11.7 GHz Down Up region I
11.7–12.2 Down region II Down regions I and III
12.2–12.7 Down regions I and II,
U.S. Direct Broadcast Satellite TV
12.50–12.75 Up region I and II, down region I
12.75–13.25 Up
14.00–14.25 Up
14.25–14.50 Up
14.5–14.8 Up
17.3–17.7 Up
17.7–18.6 Down
18.1–18.6 Down
18.6–18.8 Down regions I and III
18.8–19.7 Down
27.0–27.5 Up regions II and III
27.5–29.5 Up
30.0–31.0 Up
37.5–39.5 Down
39.5–40.5 Down
40.5–42.5 Down
42.5–43.5 Up
47.2–50.2 Up
50.4–51.4 Up
71.0–75.5 Up
81–84 Down
84–86 Down
92–95 Up
102–105 Down
149–164 Down
202–217 Up
231–241 Down
265–275 Up
Regions I, II, and III are regions of the earth’s surface defined in the International Telecommunication
Union’s Radio Regulations. Region I covers Europe, Africa, and northern Asia. Region II covers North
and South America, and Region III covers the remainder of Asia.
c04.qxd 28/08/02 19:37 Page 98
98 CHAPTER 4 SATELLITE LINK DESIGN
TABLE 4.2 Major Frequency Allocations for Mobile Satellite Services
Aeronautical Maritime Land mobile

Frequency mobile mobile and other services
137–138 MHz Down, shared

148–149.9 Up, shared
149.9–150.05 Up, shared
399.9–400.05 Up
400.15–401 Down, shared
406–406.1 Emergency beacons
890–896 Region II (limited use)
Shared with cellular radio
1559–1610 Navigation satellite, down
1530–1535 Down Down region I only
1535–1544 Down
1544–1545 Down
1545–1555 Down
1555–1559 Down
1559–1610 Navigation satellite, down
1610–1625.5 Navigation satellite, up
1625.5–1631.5 Up
1631.5–1634.5 Up
1634.5–1645.5 Up Shared
1645.5–1646.5 Up Up Up
1646.5–1656.5 Up
1656.5–1660 Up
1660.0–1660.5 Up
2483.5–2500 Down Down Down
5.00–5.25 GHz Up Up Up
7.30–7.75 Up, government Up, government Up, government
15.4–15.7 Down Down Down
29.5–31.0 Up Up Up
43.5–45.5 Up, government Up, government Up, government
45.5–47.0 Up Up Up
71.0–74.0 Up Up Down
95.0–100
134–142
190–200
252–265
Regions I, II, and III are regions of the earth’s surface defined by the International
Telecommunications Union. (See Table 4.1 for an explanation of their geographic locations.)
the most important frequencies allocated for satellite communications. The major bands are
the 64 GHz, 1411 GHz, and 3020 GHz bands. (The uplink frequency is quoted first, by
convention.) However, over much of the geostationary orbit there is already a satellite using
both 64 GHz and 1411 GHz every 2°. This is the minimum spacing used for satellites in
c04.qxd 28/08/02 19:37 Page 99
4.1 INTRODUCTION 99
GEO to avoid interference from uplink earth stations. Additional satellites can only be ac-
commodated if they use another frequency band, such as 3020 GHz. Rain in the atmosphere
attenuates radio signals. The effect is more severe as the frequency increases, with little at-
tenuation at 4 and 6 GHz, but significant attenuation above 10 GHz. Attenuation through
rain (in decibels) increases roughly as the square of frequency, so a satellite uplink operat-
ing at 30 GHz suffers four times as much attenuation as an uplink at 14 GHz.
Low earth orbit (LEO) and medium earth orbit (MEO) satellite systems have simi-
lar constraints to GEO satellite systems, but require more satellites which each serve a
smaller area of the earth’s surface. Although the satellites are much closer to the earth
than GEO satellites and therefore produce stronger signals, this advantage is usually lost
since the earth terminals need low gain omnidirectional antennas because the position of
the satellite is continually changing. LEO and MEO satellites use multiple beam anten-
nas to increase the gain of the satellite antenna beams, and also to provide frequency reuse.
Mobile satellite terminals must operate with low gain antennas at the mobile unit,
and at as low a RF frequency as can be obtained. The link between the satellite and the
major earth station (often called a hub station) is usually in a different frequency band as
it is a fixed link. Figure 4.1 shows an illustration of a maritime satellite communication
Maritime
satellite
6-GHz uplink
1.6-GHz uplinks
from ships
4-GHz downlink
Interconnection
to terrestrial Shore
network station
90-ft dish
1.5-GHz downlinks
to ships
2-m dish
FIGURE 4.1 A maritime satellite communication system.

c04.qxd 28/08/02 19:37 Page 100
system using a GEO satellite and L-band links to mobiles, with C-band links to a fixed
hub station.
All communication links are designed to meet certain performance objectives, usu-
ally a bit error rate (BER) in a digital link or a signal-to-noise ratio (SN) in an analog
link, measured in the baseband channel. The baseband channel is where an information
carrying signal is generated or received; for example, a TV camera generates a baseband
video signal, and a TV receiver delivers a baseband video signal to the picture tube to
form the images that the viewer watches. Digital data are generated by computers at
baseband, and BER is measured at baseband.
The baseband channel BER or SN ratio is determined by the carrier-to-noise ratio
(CN) at the input to the demodulator in the receiver. In most satellite communications
applications, the CN ratio at the demodulator input must be greater than 6 dB for the
BER or SN objective to be achieved. Digital links operating at CN ratios below 10 dB
must use error correction techniques to improve the BER delivered to the user. Analog
links using frequency modulation (FM) require wideband FM to achieve a large
improvement in SN ratio relative to CN ratio.
The CN ratio is calculated at the input of the receiver, at the output terminals (or
port) of the receiving antenna. RF noise received along with the signal and noise generated
by the receiver are combined into an equivalent noise power at the input to the receiver,
and a noiseless receiver model is used. In a noiseless receiver, the CN ratio is constant
at all points in the RF and IF chain, so the CN ratio at the demodulator is equal to the
CN ratio at the receiver input. In a satellite link there are two signal paths: an uplink
from the earth station to the satellite, and a downlink from the satellite to the earth sta-
tion. The overall CN at the earth station receiver depends on both links, and both, there-
fore must achieve the required performance for a specified percentage of time. Path
attenuation in the earth’s atmosphere may become excessive in heavy rain, causing the
CN ratio to fall below the minimum permitted value, especially when the 3020 GHz
band is used, leading to a link outage.
Designing a satellite system therefore requires knowledge of the required per-
formance of the uplink and downlink, the propagation characteristics and rain attenuation
for the frequency band being used at the earth station locations, and the parameters of
the satellite and the earth stations. Additional constraints may be imposed by the need
to conserve RF bandwidth and to avoid interference with other users. Sometimes, all of
this information is not available and the designer must estimate values and produce ta-
bles of system performance based on assumed scenarios. It is usually impossible to de-
sign a complete satellite communication system at the first attempt. A trial design must
first be tried, and then refined until a workable compromise is achieved. This chapter
sets out the basic procedures for the design of satellite communication links, and in-
cludes design examples for a digital TV link using a GEO satellite and quadrature
phase shift keying (QPSK) modulation, and a LEO satellite system for personal
communication.
4.2 BASIC TRANSMISSION THEORY
The calculation of the power received by an earth station from a satellite trans-
mitter is fundamental to the understanding of satellite communications. In this section,
we discuss two approaches to this calculation: the use of flux density and the link
equation.
c04.qxd 28/08/02 19:37 Page 101
4.2 BASIC TRANSMISSION THEORY 101
Area A m2
Isotropic source
EIRP = P t watts
Distance R m
Flux density F watts/m2

FIGURE 4.2 Flux density produced by an isotropic source.
Consider a transmitting source, in free space, radiating a total power Pt watts uni-
formly in all directions as shown in Figure 4.2. Such a source is called isotropic; it is an
idealization that cannot be realized physically because it could not create transverse elec-
tromagnetic waves. At a distance R meters from the hypothetical isotropic source trans-
mitting RF power Pt watts, the flux density crossing the surface of a sphere with radius
R is given by
Pt
F W/m2 (4.1)
4pR 2
All real antennas are directional and radiate more power in some directions
than in others. Any real antenna has a gain G(), defined as the ratio of power per
unit solid angle radiated in a direction to the average power radiated per unit solid
angle
P1u2
G1u2 (4.2)
P0 4p
where
P() is the power radiated per unit solid angle by the antenna
P0 is the total power radiated by the antenna
G() is the gain of the antenna at an angle 1
The reference for the angle is usually taken to be the direction in which maxi-
mum power is radiated, often called the boresight direction of the antenna. The gain of
the antenna is then the value of G() at angle 0°, and is a measure of the increase
in flux density radiated by the antenna over that from an ideal isotropic antenna radiating
the same total power. For a transmitter with output Pt watts driving a lossless antenna
with gain Gt, the flux density in the direction of the antenna boresight at distance R
meters is
PtGt
F W/m2 (4.3)
4pR 2
The product PtGt is often called the effective isotropically radiated power or EIRP,
and it describes the combination of transmitter power and antenna gain in terms
of an equivalent isotropic source with power PtGt watts, radiating uniformly in all
directions.
c04.qxd 28/08/02 19:37 Page 102
Isotropic source Incident flux density

EIRP = P t watts F watts/m2
Receiver
Pr
Receiving antenna with

Area A m2, gain G r
FIGURE 4.3 Power received by an ideal antenna with area
A m2. Incident flux density is F Pt /4R 2 W/m2. Received power
is Pr F A Pt A/4R 2 W.
If we had an ideal receiving antenna with an aperture area of A m2, as shown in

Figure 4.3, we would collect power Pr watts given by
Pr F A watts (4.4)
A practical antenna with a physical aperture area of Ar m2 will not deliver the power
given in Eq. (4.4). Some of the energy incident on the aperture is reflected away from the
antenna, and some is absorbed by lossy components. This reduction in efficiency is
described by using an effective aperture Ae where
Ae hA Ar (4.5)
and A is the aperture efficiency of the antenna2. The aperture efficiency A accounts for
all the losses between the incident wavefront and the antenna output port: these include
illumination efficiency or aperture taper efficiency of the antenna, which is related to the
energy distribution produced by the feed across the aperture, and also other losses due to
spillover, blockage, phase errors, diffraction effects, polarization, and mismatch losses.
For parabolodial reflector antennas, A is typically in the range 50 to 75%, lower for small
antennas and higher for large Cassegrain antennas. Horn antennas can have efficiencies
approaching 90%. Thus the power received by a real antenna with a physical receiving
area Ar and effective aperture area Ae m2 is
PtGt Ae
Pr watts (4.6)
4pR 2
Note that this equation is essentially independent of frequency if Gt and Ae are con-
stant within a given band; the power received at an earth station depends only on the EIRP
of the satellite, the effective area of the earth station antenna, and the distance R.
A fundamental relationship in antenna theory2 is that the gain and area of an antenna
are related by
G 4pA e l2 (4.7)
where is the wavelength (in meters) at the frequency of operation.
Substituting for Ae in Eq. (4.6) gives
PtGtGr
Pr
14pRl2 2
watts (4.8)
c04.qxd 28/08/02 19:37 Page 103
4.2 BASIC TRANSMISSION THEORY 103
This expression is known as the link equation, and it is essential in the calcula-
tion of power received in any radio link. The frequency (as wavelength, ) appears in
this equation for received power because we have used the receiving antenna gain, in-
stead of effective area. The term [4R]2 is known as the path loss, Lp. It is not a loss
in the sense of power being absorbed; it accounts for the way energy spreads out as an
electromagnetic wave travels away from a transmitting source in three-dimensional
space.
Collecting the various factors, we can write
EIRP Receiving antenna gain

Power received watts (4.9)
Path loss
In communication systems, decibel quantities are commonly used to simplify equa-

tions like Eq. (4.9). In decibel terms, we have
Pr EIRP Gr Lp dBW (4.10)

where
EIRP 10 log10 1Pt Gt 2 dBW
Gr 10 log10 14pAe l2 2 dB
Path loss Lp 10 log10 3 14pRl2 2 4 20 log10 14pRl2 dB
Equation (4.10) represents an idealized case, in which there are no additional losses
in the link. It describes transmission between two ideal antennas in otherwise empty space.
In practice, we will need to take account of a more complex situation in which we have
losses in the atmosphere due to attenuation by oxygen, water vapor, and rain, losses in
the antennas at each end of the link, and possible reduction in antenna gain due to mis-
pointing. All of these factors are taken into account by the system margin but need to be
calculated to ensure that the margin allowed is adequate. More generally, Eq. (4.10) can
be written
Pr EIRP Gr Lp La L ta L ra dBW (4.11)
where
La attenuation in atmosphere
Lta losses associated with transmitting antenna
Lra losses associated with receiving antenna
The conditions in Eq. (4.11) are illustrated in Figure 4.4. The expression dBW
means decibels greater or less than 1 W (0 dBW). The units dBW and dBm (dB greater
or less than 1 W and 1 mW) are widely used in communications engineering. EIRP,
being the product of transmitter power and antenna gain is often quoted in dBW.
Note that once a value has been calculated in decibels, it can readily be scaled if
one parameter is changed. For example, if we calculated Gr for an antenna to be 48 dB,
at a frequency of 4 GHz, and wanted to know the gain at 6 GHz, we can multiply Gr by
(64)2. Using decibels, we simply add 20 log(64) or 20 log(3) 20 log(2) 9.5 6
3.5 dB. Thus the gain of our antenna at 6 GHz is 51.3 dB.
Appendix A gives more information on the use of decibels in communications
engineering.
c04.qxd 28/08/02 19:37 Page 104
Satellite
EIRP = Pt G t
Loss L ta
Path loss L p
Atmospheric
Earth station
loss L a
antenna gain G r
Loss L ra
LNA
Received
power Pr
FIGURE 4.4 A satellite link. LNA, low noise amplifier.
EXAMPLE 4.2.1
A satellite at a distance of 40,000 km from a point on the earth’s surface radiates a power of 10 W
from an antenna with a gain of 17 dB in the direction of the observer. Find the flux density
at the receiving point, and the power received by an antenna at this point with an effective area
of 10 m2.
Using Eq. (4.3)
F PtGt 14pR 2 2 10 50 14p 14 107 2 2 2 2.49 1014 W/m2
The power received with an effective collecting area of 10 m2 is therefore
Pr 2.49 1013 W
The calculation is more easily handled using decibels. Noting that 10 log10 4 11.0 dB
F in dB units 10 log10 1PtGt 2 20 log10 1R2 11.0
27.0 152.0 11.0
136.0 dB1W/m2 2
Then Pr 136.0 10.0 126 dBW.
Here we have put the antenna effective area into decibels greater than 1 m2 (10 m2 10 dB
greater than 1 m2).
EXAMPLE 4.2.2
The satellite in Example 4.2.1 operates at a frequency of 11 GHz. The receiving antenna has a gain
of 52.3 dB. Find the received power.
c04.qxd 28/08/02 19:37 Page 105
4.3 SYSTEM NOISE TEMPERATURE AND G/T RATIO 105
Using Eq. (4.10) and working in decibels

Pr EIRP Gr path loss 1dBW2
EIRP 27.0 dBW
Gr 52.3 dB
Path loss 14pRl2 2 20 log10 14pRl2 dB
20 log10 3 14p 4 107 2 12.727 102 2 4 dB 205.3 dB
Pr 27.0 52.3 205.3 126.0 dBW
We have the same answer as in Example 4.2.1 because the figure of 52.3 dB is the gain of
a 10 m2 aperture at a frequency of 11 GHz.
Equation (4.10), with other parameters for antenna and propagation losses, is com-
monly used for calculation of received power in a microwave link and is set out as a link
power budget in tabular form using decibels. This allows the system designer to adjust
parameters such as transmitter power or antenna gain and quickly recalculate the received
power.
The received power, Pr, calculated by Eqs. (4.6) and (4.8) is commonly referred
to as carrier power, C. This is because most satellite links use either frequency mod-
ulation for analog transmission or phase modulation for digital transmission. In both
of these modulation systems, the amplitude of the carrier is not changed when the data
are modulated onto the carrier, so received carrier power C is always equal to received
power Pr.
4.3 SYSTEM NOISE TEMPERATURE

AND G/T RATIO
Noise Temperature
Noise temperature is a useful concept in communications receivers, since it provides a
way of determining how much thermal noise is generated by active and passive devices
in the receiving system. At microwave frequencies, a black body with a physical temper-
ature, Tp degrees kelvin, generates electrical noise over a wide bandwidth.
The noise power is given by3
Pn kTpBn (4.12)
where
k Boltzmann’s constant 1.39 1023 J/K 228.6 dBW/K/Hz
Tp physical temperature of source in kelvin degrees
Bn noise bandwidth in which the noise power is measured, in hertz
Pn is the available noise power (in watts) and will be delivered only to a load that
is impedance matched to the noise source. The term kTp is a noise power spectral density,
in watts per hertz. The density is constant for all radio frequencies up to 300 GHz.
We need a way to describe the noise produced by the components of a low noise
receiver. This can conveniently be done by equating the component to a black body radi-
ator with an equivalent noise temperature, Tn kelvins. A device with a noise temperature
of Tn kelvins (symbol K, not K) produces at its output the same noise power as a black
body at a temperature Tn degrees kelvin followed by a noiseless amplifier with the same
c04.qxd 28/08/02 19:37 Page 106
gain as the actual device. The description of a low noise component by an equivalent noise
source at the input of a noiseless amplifier is very useful because we can add noise tem-
peratures to determine the total noise power in a receiver, as shown in the following analy-
sis. Note that the unit of noise temperature is kelvins, not degrees kelvin. This distinction
is often lost on the suppliers of consumer satellite broadcast receiving equipment.
In satellite communication systems we are always working with weak signals (be-
cause of the large distances involved) and must make the noise level as low as possible
to meet the CN ratio requirements. This is done by making the bandwidth in the receiver,
usually set by the IF amplifier stages, to be just large enough to allow the signal (carrier
and sidebands) to pass unrestricted, while keeping the noise power to the lowest value
possible. The bandwidth used in Eq. (4.12) should be the equivalent noise bandwidth. Fre-
quently we do not know the equivalent noise bandwidth and use the 3-dB bandwidth of
our receiving system instead. The error introduced by using the 3-dB bandwidth is small
when the filter characteristic of the receiver has steep sides.
Noise temperatures from 30 K to 200 K can be achieved without physical cooling
if GaAsFET (gallium arsenide field effect transistor) amplifiers are employed. GaAsFET
amplifiers can be built to operate at room temperature with noise temperatures of 30 K at
4 GHz and 100 K at 11 GHz. Typically, noise temperatures increase with frequency, and
an LNA for a 20 GHz receiver might have a noise temperature of 150 K. One might ask
how an amplifier can have a noise temperature that is lower than its physical temperature.
Noise temperature simply relates the noise produced by an amplifier to the thermal noise
from a matched load at the same physical temperature placed at the input to the ampli-
fier. If the amplifier produced no noise at all, its noise temperature would be 0 K. If the
amplifier produces less noise than a matched load at the same physical temperature, its
noise temperature will be lower than its physical temperature.
To determine the performance of a receiving system we need to be able to find the total
thermal noise power against which the signal must be demodulated. We do this by deter-
mining the system noise temperature, Ts. Ts is the noise temperature of a noise source, located
at the input of a noiseless receiver, which gives the same noise power as the original receiver,
measured at the output of the receiver and usually includes noise from the antenna.
If the overall end-to-end gain of the receiver is Grx (Grx is a ratio, not in decibels)
and its narrowest bandwidth is Bn Hz, the noise power at the demodulator input is
Pno kTs BnGrx watts (4.13a)
where Grx is the gain of the receiver from RF input to demodulator input.
The noise power referred to the input of the receiver is Pn where
Pn kTs Bn watts (4.13b)
Let the antenna deliver a signal power Pr watts to the receiver RF input. The signal
power at the demodulator input is PrGrx watts, representing the power contained in the
carrier and sidebands after amplification and frequency conversion within the receiver.
Hence, the carrier-to-noise ratio at the demodulator is given by
C Pr Grx Pr
(4.14)
N kTs BnGrx kTs Bn
The gain of the receiver cancels out in Eq. (4.14), so we can calculate CN ratios
for our receiving terminals at the antenna output port. This is convenient, because a link
budget will find Pr at this point. Using a single parameter to encompass all of the sources
of noise in a receiving terminal is very useful because it replaces several sources of noise
in the receiver by a single system noise temperature, Ts.
c04.qxd 28/08/02 19:37 Page 107
Signal from
satellite
Receiving
antenna
LNA BPF Mixer BPF IF amplifier
Pr IF output
G rf Gm G IF
Local oscillator
FIGURE 4.5 Simplified earth station receiver. BPF, bandpass filter.
Calculation of System Noise Temperature

Figure 4.5 shows a simplified communications receiver with an RF amplifier and single fre-
quency conversion, from its RF input to the IF output. This is the form used for all radio re-
ceivers, with few exceptions, known as the superhet (short for superheterodyne). The superhet
receiver has three main subsystems: a front end (RF amplifier, mixer and local oscillator) an
IF amplifier (IF amplifiers and filters), and a demodulator and baseband section.
The RF amplifier in a satellite communications receiver must generate as little noise
as possible, so it is called a low noise amplifier or LNA. The mixer and local oscillator
form a frequency conversion stage that downconverts the RF signal to a fixed intermediate
frequency (IF), where the signal can be amplified and filtered accurately.
SIDEBAR
In general, it is difficult to make good narrowband could also receive another RF signal at a frequency
filters with a ratio of bandwidth to center frequency fsignal flo which would produce an output from the
less than 1%. If we want a 10-MHz bandwidth filter mixer at fif. This is called an image frequency. It is
for a signal received from a satellite at 11 GHz, it blocked by a band-pass filter in the RF amplifier that
is very difficult to implement the filter at the RF fre- is wide enough to pass the wanted range of signal
quency of 11 GHz, where the filter bandwidth is less frequencies but has high attenuation for the image
than 0.1% of the center frequency. We would instead frequencies.
down convert the 11-GHz signal to a frequency One further advantage of the superhet receiver
around 1 GHz, where the 10-MHz bandwidth is 1% design is that tuning of the receiver can be done
of the IF frequency. This is the advantage of the su- with the local oscillator. The IF is at a fixed fre-
perhet receiver design: very accurate filters can be quency, and the local oscillator frequency is varied
used by converting the signal to a convenient to select the wanted signal. The front end is fol-
frequency. lowed by an IF amplifier stage which contains band-
The downconversion in the front end is achieved pass filters that exactly match the spectrum of the
by multiplying the received signal and the local os- received signal. In many small earth stations, known
cillator frequency in a nonlinear device—a mixer. as very small aperture terminals (VSATs), the LNA,
Multiplication of two RF signals creates products at LO, and first IF amplifier and filters are all included
their sum and difference frequencies; the frequency in a single package called a low noise block con-
of the local oscillator (LO) is usually set to fsignal verter (LNB) located immediately behind the an-
fif. This is called low side injection. The receiver tenna feed.
c04.qxd 28/08/02 19:37 Page 108
Signal from satellite
Receiving
antenna
First
LNA BPF Mixer BPF IF amplifier
Pr
G rf
First local
oscillator
Second IF
BPF Mixer BPF amplifier Demodulator
Tunable Baseband
Second local output
channel
oscillator
select filter
(tunable)
FIGURE 4.6 Double conversion earth station receiver. The first downconversion shifts
signals in a 500-MHz band to the first IF range 900–1400 MHz. The second downconverter
has a tunable local oscillator and channel selection filter to select the wanted transponder
signal in the second IF centered at 70 MHz.
Many earth station receivers use the double superhet configuration shown in Fig-
ure 4.6 which has two stages of frequency conversion. The front end of the receiver is
mounted behind the antenna feed and converts the incoming RF signals to a first IF in the
range 900 to 1400 MHz. This allows the receiver to accept all the signals transmitted from
a satellite in a 500-MHz bandwidth at C band or Ku band, for example. The RF ampli-
fier has a high gain and the mixer is followed by a stage of IF amplification. This section
of the receiver is called a low noise block converter (LNB). The 900–1400 MHz signal is
sent over a coaxial cable to a set-top receiver that contains another down-converter and a
tunable local oscillator. The local oscillator is tuned to convert the incoming signal from
a selected transponder to a second IF frequency. The second IF amplifier has a bandwidth
matched to the spectrum of the transponder signal. Direct broadcast satellite TV receivers
at Ku band use this approach, with a second IF filter bandwidth of 20 MHz.
The equivalent circuits in Figure 4.7a can be used to represent a receiver for the pur-
pose of noise analysis. The noisy devices in the receiver are replaced by equivalent noise-
less blocks with the same gain and noise generators at the input to each block such that
the block produces the same noise at its output as the device it replaces. The entire receiver
is then reduced to a single equivalent noiseless block with the same end-to-end gain as the
actual receiver and a single noise source at its input with temperature Tn. The total noise
power at the output of the IF amplifier of the receiver in Figure 4.7a is given by
Pn GIF kTIF Bn GIFGmkTmBn GIFGmGRF kBn 1TRF Tin 2 (4.15)
where GRF, Gm, and GIF are the gains of the RF amplifier, mixer, and IF amplifier, and
TRF, Tm, and TIF are their equivalent noise temperatures. Tin is the noise temperature of the
antenna, measured at its output port.
c04.qxd 28/08/02 19:37 Page 109
Noiseless RF Noiseless Noiseless IF

amplifier mixer amplifier
Tin Pn
+ Gain + Gain + Gain
G RF Gm GIF
TRF Tm TIF
Equivalent noise sources

(a )
Noiseless
Noiseless
lossy device
receiver
Tin Pn Tin Pn
Gain
Gain +
+ Gl
G RF .Gm .G IF
Noise
Noise
source
source
Ts Tno
(b ) (c )
FIGURE 4.7 (a) Noise model of receiver. The noisy amplifiers and downconverter have
been replaced by noiseless units, with equivalent noise generators at their inputs. (b) Noise
model of receiver. All noisy units have been replaced by one noiseless amplifier, with a sin-
gle noise source Ts as its input. (c) Noise model for a lossy device. The lossy device has
been replaced by a lossless device, with a single noise source Tno at its output.
Equation (4.15) can be rewritten as

Pn GIFGmGRF 3 1kTIFBn 2 1GRFGm 2 1kTmBn 2 GRF 1TRF Tin 2 4
GIFGmGRF kBn 3TRF Tin TmGRF TIF 1GRFGm 2 4 (4.16)
The single source of noise shown in Figure 4.7b with noise temperature Ts gener-
ates the same noise power Pn at its output if
Pn GIFGmGRF kTsBn (4.17)
The noise power at the output of the noise model in Figure 4.7b will be the same as the
noise power at the output of the noise model in Figure 4.7a if
kTs Bn kBn 3 1Tin TRF TmGIF TIFGmGRF 2 4
Hence the equivalent noise source in Figure 4.7b has a system noise temperature Ts where
Ts 3Tin TRF Tm GRF TIF 1GmGRF 2 4 (4.18)
Succeeding stages of the receiver contribute less and less noise to the total system
noise temperature. Frequently, when the RF amplifier in the receiver front end has a high
gain, the noise contributed by the IF amplifier and later stages can be ignored and the
system noise temperature is simply the sum of the antenna noise temperature and the LNA
c04.qxd 28/08/02 19:37 Page 110
noise temperature, so Ts Tantenna TLNA. Note that the values for component gains in
Eq. (4.18) must be linear ratios, not in decibels.
The noise model shown in Figure 4.7b replaces all the individual sources of noise
in the receiver by a single noise source at the receiver input. This assumes that all the noise
comes in from the antenna or is internally generated in the receiver. In some circum-
stances, we need to use a different model to deal with noise that reaches the receiver after
passing through a lossy medium. Waveguide and rain losses are two examples. When rain-
drops cause attenuation, they radiate additional noise whose level depends on the
attenuation. We can model the noise emission as a noise source placed at the “output” of
the atmosphere, which is the antenna aperture. The noise model for an equivalent output
noise source is shown in Figure 4.7c, and produces a noise temperature Tno given by
Tno Tp 11 Gl 2 (4.19)
where Gl is the linear gain (less than unity, not in decibels) of the attenuating device or
medium, and Tp is the physical temperature in degrees kelvin of the device or medium.
For an attenuation of A dB, the value of Gl is given by
Gl 10 A10 (4.20)
EXAMPLE 4.3.1
Suppose we have a 4-GHz receiver with the following gains and noise temperatures:
Tin 25 K GRF 23 dB
TRF 50 K GIF 30 dB
TIF 1000 K
Tm 500 K
Calculate the system noise temperature assuming that the mixer has a gain Gm 0 dB.
Recalculate the system noise temperature when the mixer has a 10-dB loss. How can the noise
temperature of the receiver be minimized when the mixer has a loss of 10 dB?
The system noise temperature is given by Eq. (4.18)
Ts 3 25 50 15002002 110002002 4 87.5 K
If the mixer had a loss, as is usually the case, the effect of the IF amplifier would be greater. For
Gm 10 dB, the linear value is Gm 0.1 as a ratio. Then
Ts 3 25 50 15002002 11000202 4 137.5 K
The lowest system noise temperatures are obtained by using a high gain LNA. Suppose we
increase the LNA gain in this example to GRF 50 dB, giving ratio GRF 105.
Ts 3 25 50 1500105 2 11000104 2 4 75.1 K
The high gain of the RF LNA amplifier has made the system noise temperature almost as low as it
can go: Ts Tin TRF 75 K in this example. LNAs for use in satellite receivers usually have
gains in the range 40–55 dB.
EXAMPLE 4.3.2
The system illustrated in Example 4.3.1 has an LNA with a gain of 50 dB. A section of lossy wave-
guide with an attenuation of 2 dB is inserted between the antenna and the RF amplifier. Find the
new system noise temperature for a waveguide temperature of 300°K.
The waveguide loss of 2 dB (ratio 1.58) can be treated as a gain, Gl, that is less than unity:
(Gl 11.58 0.631). The lossy waveguide attenuates the incoming noise and adds noise
c04.qxd 28/08/02 19:37 Page 111
generated by its own ohmic loss. The equivalent noise generator placed at the output of the section
of waveguide that represents the noise generated by the waveguide has a noise temperature Twg,
where
Twg Tp 11 Gl 2 30011 0.6312 110.7 K
The waveguide attenuates the noise from the antenna, so Tin 0.63 25 15.8 K
The new system noise temperature, referred to the input of the LNA, is
Ts 315.8 110.7 50 1500105 2 11000104 2 4 176.6 K
We can refer the system noise temperature to the antenna output port by dividing the above
result by Gl. This transfers the noise source from the LNA input to the waveguide input.
Ts 176.60.631 279.9 K
The new system noise temperature is 5.7 dB higher than the system noise temperature with-
out the lossy waveguide.
Note that when the system noise temperature is low, each 0.1 dB of attenuation
ahead of the RF amplifier will add approximately 6.6 K to the system noise temper-
ature. (Using the formula in Example 4.3.2 with Tp 290°K, Gl 0.1 dB 0.977
and Tno 290 0.023 6.6 K.) This is the reason for placing the front end of the
receiver at the output of the antenna feed. Waveguide losses ahead of the LNA can
have a disastrous effect on the system noise temperature of low noise receiving
systems.
Noise Figure and Noise Temperature

Noise figure is frequently used to specify the noise generated within a device. The oper-
ational noise figure (NF) is defined by the following formula4:
1SN2 in
NF
1S N2 out
(4.21)
Because noise temperature is more useful in satellite communication systems, it is best

to convert noise figure to noise temperature, Td. The relationship is
Td T0 1NF 12 (4.22)
where the noise figure is a linear ratio, not in decibels, and where T0 is the reference
temperature used to calculate the standard noise figure—usually 290 K. NF is fre-
quently given in decibels and must be converted to a ratio before being used in
Eq. (4.18).
Table 4.3 gives a comparison between noise figure and noise temperature over the
range encountered in typical systems.
TABLE 4.3 Comparison of Noise Temperature and Noise Figure
Noise temperature (K) 0 20 40 60 80 100 120 150 200 290

Noise figure (dB) 0 0.29 0.56 0.82 1.06 1.29 1.50 1.81 2.28 3.0
Noise temperature (K) 400 600 800 1,000 1,500 2,000 3,000 5,000 10,000
Noise figure (dB) 3.8 4.9 5.8 6.5 7.9 9.0 10.5 12.6 15.5
c04.qxd 28/08/02 19:37 Page 112
EXAMPLE 4.3.3
An amplifier has a quoted noise figure of 2.5 dB. What is its equivalent noise temperature? Using
Eq. (4.22)
Td 29011.78 12 226 K
This value of noise temperature could then be used in Eq. (4.17), with other appropriate data, to
calculate system noise temperature.
G/T Ratio for Earth Stations

The link equation can be rewritten in terms of (CN) at the earth station
c dc d c dc d c d
C PtGtGr l 2 PtGr l 2 Gr
(4.23)
N kTsBn 4pR kBn 4pR Ts
Thus CN r Gr Ts, and the terms in the square brackets are all constants for a given satel-
lite system. The ratio Gr Ts, which is usually quoted as simply GT in decibels, with units
dB/K, can be used to specify the quality of a receiving earth station or a satellite receiv-
ing system, since increasing Gr Ts increases the received CN ratio.
Satellite terminals may be quoted as having a negative GT which is below 0 dB/K.
This simply means that the numerical value of Gr is smaller than the numerical value of
Ts.
EXAMPLE 4.3.4
An earth station antenna has a diameter of 30 m, has an overall efficiency of 68%, and is used to
receive a signal at 4150 MHz. At this frequency, the system noise temperature is 79 K when the
antenna points at the satellite at an elevation angle of 28°. What is the earth station GT ratio un-
der these conditions? If heavy rain causes the sky temperature to increase so that the system noise
temperature rises to 88 K, what is the new GT value?
First calculate the antenna gain. For a circular aperture:
Gr hA4pAl2 hA 1pDl2 2
At 4150 MHz, 0.0723 m. Then
G 0.68 1p 300.07232 2 1.16 106 or 60.6 dB
Converting Ts into dBK
Ts 10 log 79 19.0 dBK
GT 60.6 19.0 41.6 dB/K
If Ts 88 K in heavy rain, GT 60.6 19.4 41.2 dB/K
4.4 DESIGN OF DOWNLINKS
The design of any satellite communication is based on two objectives: meeting a mini-
mum CN ratio for a specified percentage of time, and carrying the maximum revenue
earning traffic at minimum cost. There is an old saying that “an engineer is a person who
can do for a dollar what any fool can do for one hundred dollars.” This applies to satel-
lite communication systems. Any satellite link can be designed with very large antennas
to achieve high CN ratios under all conditions, but the cost will be high. The art of good
c04.qxd 28/08/02 19:37 Page 113
4.4 DESIGN OF DOWNLINKS 113
system design is to reach the best compromise of system parameters that meets the spec-
ification at the lowest cost. For example, if a satellite link is designed with sufficient mar-
gin to overcome a 20-dB rain fade rather than a 3-dB fade, earth station antennas with
seven times the diameter are required.
All satellite communications links are affected by rain attenuation. In the 64
GHz band the effect of rain on the link is small. In the 1411 GHz (Ku) band, and
even more so in the 3020 GHz (Ka) band, rain attenuation becomes all important.
Satellite links are designed to achieve reliabilities of 99.5 to 99.99%, averaged over a
long period of time, typically a year. That means the CN ratio in the receiver will fall
below the minimum permissible value for proper operation of the link for between 0.5
and 0.01% of the specified time; the link is then said to suffer an outage. The time pe-
riod over which the percentage of time is measured can be a month, sometimes the
“worst month” in attenuation terms, or a year. Rain attenuation is a very variable phe-
nomenon, both with time and place. Chapter 8 discusses the prediction of path atten-
uation and provides ways to estimate the likely occurrence of outages on a given link.
In this chapter we will simply assume certain rain attenuation statistics to use in
examples of link design.
C-band links can be designed to achieve 99.99% reliability because the rain at-
tenuation rarely exceeds 1 or 2 dB. The time corresponding to 0.01% of a year is 52
min; at this level of probability the rain attenuation statistics are usually not stable and
wide fluctuations occur from year to year. Outages occur in heavy rain, usually in thun-
derstorms, and thunderstorm occurrence varies widely. A link designed to have out-
ages totaling 52 min each year may well have outages of several hours one year and
none the next. Most Ka-band links cannot be designed to achieve 99.99% reliability
because rain attenuation generally exceeds 10 dB, and often 20 dB, for 0.01% of the
time. Outage times of 0.1 to 0.5% of a year (8 to 40 h) are usually tolerated in Ka-
band links. The allowable outage time for a link depends in part on the traffic carried.
Telephone traffic needs real-time channels that are maintained for the duration of a
call, so C band or Ku band is used for voice channels with sufficient link margin that
outage times are small. Internet transmissions are less affected by short outages and
generally do not require a real-time channel, making Ka band better suited for Internet
access.
Link Budgets
CN ratio calculation is simplified by the use of link budgets. A link budget is a tabular
method for evaluating the received power and noise power in a radio link. Link budgets
invariably use decibel units for all quantities so that signal and noise powers can be cal-
culated by addition and subtraction. Since it is usually impossible to design a satellite link
at the first attempt, link budgets make the task much easier because, once a link budget
has been established, it is easy to change any of the parameters and recalculate the result.
Tables 4.4a and 4.4b show a typical link budget for a C-band downlink using a global
beam on a GEO satellite and a 9-m earth station antenna.
The link budget must be calculated for an individual transponder, and must be re-
peated for each of the individual links. In a two-way satellite communication link there
will be four separate links, each requiring a calculation of CN ratio. When a bent pipe
transponder is used the uplink and downlink CN ratios must be combined to give an
overall CN. In this section we will calculate the CN ratio for a single link. Later ex-
amples in this chapter demonstrate the evaluation of a complete satellite communication
system.
c04.qxd 28/08/02 19:37 Page 114
TABLE 4.4a C-Band GEO Satellite Link Budget in Clear Air
C-band satellite parameters

Transponder saturated output power 20 W
Antenna gain, on axis 20 dB
Transponder bandwidth 36 MHz
Downlink frequency band 3.7–4.2 GHz
Signal FM-TV analog signal
FM-TV signal bandwidth 30 MHz
Minimum permitted overall C/N in receiver 9.5 dB
Receiving C-band earth station
Downlink frequency 4.00 GHz
Antenna gain, on axis, 4 GHz 49.7 dB
Receiver IF bandwidth 27 MHz
Receiving system noise temperature 75 K
Downlink power budget
Pt Satellite transponder output power, 20 W 13.0 dBW
Bo Transponder output backoff 2.0 dB
Gt Satellite antenna gain, on axis 20.0 dB
Gr Earth station antenna gain 49.7 dB
Lp Free space path loss at 4 GHz 196.5 dB
Lant Edge of beam loss for satellite antenna 3.0 dB
La Clear air atmospheric loss 0.2 dB
Lm Other losses 0.5 dB
Pr Received power at earth station 119.5 dBW
Downlink noise power budget in clear air
k Boltzmann’s constant 228.6 dBW/K/Hz
Ts System noise temperature, 75 K 18.8 dBK
Bn Noise bandwidth, 27 MHz 74.3 dBHz
N Receiver noise power 135.5 dBW
C/N ratio in receiver in clear air
C/N Pr N 119.5 dBW (135.5 dBW) 16.0 dB
Link budgets are usually calculated for a worst case, the one in which the link will
have the lowest CN ratio. Factors which contribute to a worst case scenario include: an
earth station located at the edge of the satellite coverage zone where the received signal
is typically 3 dB lower than in the center of the zone because of the satellite antenna pat-
tern, maximum path length from the satellite to the earth station, a low elevation angle at
TABLE 4.4b C-Band Downlink Budget in Rain
Prca Received power at earth station in clear air 119.5 dBW

A Rain attenuation 1.0 dB
Prain Received power at earth station in rain 120.5 dBW
Nca Receiver noise power in clear air 135.5 dBW
Nrain Increase in noise temperature due to rain 2.3 dB
Nrain Receiver noise power in rain 133.2 dBW
C/N ratio in receiver in rain
C/N Prain Nrain 120.5 dBW (133.2 dBW) 12.7 dB
c04.qxd 28/08/02 19:37 Page 115
4.4 DESIGN OF DOWNLINKS 115
the earth station giving the highest atmospheric path attenuation in clear air, and maximum
rain attenuation on the link causing loss of received signal power and an increase in re-
ceiving system noise temperature. The edge of the coverage pattern of the satellite an-
tenna and the longest path usually go together. However, when a satellite has a multiple
beam antenna, this may not always be the case. Earth station antennas are assumed to be
pointed directly at the satellite, and therefore operate at their on-axis gain. If the antenna
is mispointed, a loss factor is included in the link budget to account for the reduction in
antenna gain.
The calculation of carrier to noise ratio in a satellite link is based on the two equa-
tions for received signal power and receiver noise power that were presented in Sections
4.1 and 4.2. Equation 4.11 gives the received carrier power in dB watts as
Pr EIRP Gr Lp La Lr Lt dBW (4.24)
A receiving terminal with a system noise temperature TsK and a noise bandwidth
Bn Hz has a noise power Pn referred to the output terminals of the antenna where
Pn kTsBn watts (4.25)
The receiving system noise power is usually written in decibel units as
N k Ts Bn dBW (4.26)
where k is Boltzmann’s constant (228.6 dBW/K/Hz), Ts is the system noise temperature
in dBK, and Bn is the noise bandwidth of the receiver in dBHz. Note that because we are
working in units of power, all decibel conversions are made as 10 log10(Ts) or 10 log10(Bn).
The 20 log10 factor used in the calculation of path loss results from the (4R)2 term in
the path loss equation.
Link Budget Example: C-Band Downlink

for Earth Coverage Beam
The satellite used in this example (see Tables 4.4a and 4.4b) is in geostationary earth or-
bit and carries 24 C-band transponders, each with a bandwidth of 36 MHz. The downlink
band is 3.7–4.2 GHz and the satellite uses orthogonal circular polarizations to provide an
effective RF bandwidth of 864 MHz. The satellite provides coverage of the visible earth,
which subtends an angle of approximately 17 from a satellite in a geostationary orbit,
by using a global beam antenna. Since antenna beamwidth and gain are linked together
[3 dB beamwidth 1133,000 G2 where G is a ratio, not in decibels], the on-axis gain
of the global beam antenna is approximately 20 dB. However, we must make the link
budget calculation for an earth station at the edge of the coverage zone of the satellite
where the effective gain of the antenna is 3 dB lower, at 17 dB. The CN ratio for the
downlink is calculated in clear air conditions and also in heavy rain.
An antenna with a gain of 20 dB has an effective aperture diameter of 5.6 wave-
lengths [G (D)2], which gives D 0.42 m at a frequency of 4 GHz. The calculation
of CN ratio is made at a mid-band frequency of 4 GHz.
The saturated output power of the transponder is 20 W 13 dBW. We will as-
sume an output back-off of 2 dB, so that the power transmitted by the transponder is 11
dBW. Hence the on-axis EIRP of the transponder and antenna is PtGt 11 20 31
dBW. The transmitted signal is a single 30-MHz bandwidth analog FM-TV channel in
this example. Following common practice for analog TV transmission, the receiver noise
bandwidth is set to 27 MHz, slightly less than the 30-MHz bandwidth of the FM-TV
signal.
c04.qxd 28/08/02 19:37 Page 116
The receiving earth station has an antenna with an aperture diameter of 9 m and a
gain of 49.7 dB at 4 GHz, and a receiving system noise temperature of 75 K in clear air
conditions. The GT ratio for this earth station is GT 49.7 10 log10 75 30.9 dBK1.
The maximum path length for a GEO satellite link is 40,000 km, which gives a path loss
of 196.5 dB at 4 GHz ( 0.075 m). We must make an allowance in the link budget for
some losses that will inevitably occur on the link. At C band, propagation losses are small,
but the slant path through the atmosphere will suffer a typical attenuation of 0.2 dB in
clear air. We will allow an additional 0.5-dB margin in the link design to account for mis-
cellaneous losses, such as antenna mispointing, polarization mismatch, and antenna degra-
dation, to ensure that the link budget is realistic.
The earth station receiver CN ratio is first calculated for clear air conditions, with
no rain in the slant path. The CN ratio is then recalculated taking account of the effects
of rain. The minimum permitted overall CN ratio for this link is 9.5 dB, corresponding
to the FM threshold of an analog satellite TV receiver. Table 4.4a shows that we have a
downlink CN of 16.0 dB in clear air, giving a link margin of 6.5 dB. This link margin
is available in clear air conditions, but will be reduced when there is rain in the slant path.
Heavy rain in the slant path can cause up to 1 dB of attenuation at 4 GHz, which
reduces the received power by 1 dB and increases the noise temperature of the receiving
system. Using the output noise model discussed in the previous section with a medium
temperature of 273 K, and a total path loss for clear air plus rain of 1.2 dB (ratio of 1.32),
the sky noise temperature in rain is
Tsky 273 11 1 1.322 66 K
In clear air the sky noise temperature is about 13 K, the result of 0.2 dB of clear air
attenuation. The noise temperature of the receiving system has therefore increased by
(66 13) K 53 K to 75 53 K 128 K with 1 dB rain attenuation in the slant path,
from a clear air value of 75 K. This is an increase in system noise temperature of 2.3 dB.
We can now adjust the link budget very easily to account for heavy rain in the slant
path without having to recalculate the CN ratio from the beginning. The received carrier
power is reduced by 1 dB because of the rain attenuation and the system noise tempera-
ture is increased by 2.3 dB. Table 4.4b shows the new downlink budget in rain.
The CN ratio in rain has a margin of 3.2 dB over the minimum permissible CN
ratio of 9.5 dB for an analog FM-TV transmission. The CN margin will translate into a
higher than needed SN ratio in the TV baseband signal, and can be traded off against
earth station antenna gain to allow the use of smaller (and therefore lower cost) antenna.
We should always leave a small margin for unexpected losses if we want to guarantee a
particular level of reliability in the link. In this case, we will use a 2-dB margin and ex-
amine how the remaining 1.2 dB of link margin can be traded against other parameters
in the system.
A reduction in earth station antenna gain of 1.2 dB is a reduction in the gain
value, as a ratio, of 1.32. Antenna gain is proportional to diameter squared, so the
diameter of the earth station antenna can be reduced by a factor of 11.32 1.15, from
9 m to 7.8 m.
We could transmit a QPSK signal from the satellite instead of an analog FM sig-
nal. Using the 27-MHz noise bandwidth receiver, we could transmit a digital signal at 54
Mbps using QPSK, but would require a minimum CN ratio in the receiver of 14.6 dB,
allowing a 1-dB implementation margin and a minimum BER of 106. The link margin
would be 1.9 dB under heavy rain conditions, so we would need to increase the earth
station antenna diameter by a factor of 1.55 to 13.9 m to provide a CN ratio of 14.6 dB
under heavy rain conditions. A 54-Mbps digital signal could carry seven digital TV signals
c04.qxd 28/08/02 19:37 Page 117
4.5 SATELLITE SYSTEMS USING SMALL EARTH STATIONS 117
using MPEG-2 compression, a much more attractive proposition than carrying a single
analog FM-TV signal, although at the cost of a larger earth station antenna.
Global beam antennas are not widely used, although most Intelsat satellites carry
them. Regional TV signal distribution is much more common, so the C-band link in
Tables 4.4a and 4.4b is more likely to use a regional antenna, serving the United States,
for example, with a 6 by 3 beam. The gain of a typical satellite antenna providing cov-
erage of the 48 contiguous states is 32.0 dB on axis (G 33,0001 2), which is
12.0 dB higher than the on-axis gain of a global beam. Using the link budget in Tables
4.4a and 4.4b, we can trade the extra 12-dB gain of a regional coverage satellite antenna
for a reduction in earth station antenna dimensions. For the example of a 9.0-m antenna
receiving analog FM-TV, we could reduce the antenna diameter by a factor of 4 to 2.25
m (approximately 7 ft 6 inch diameter). This is the smallest size of antenna used by home
satellite TV systems operating in C band.
The above examples show how the link budget can be used to study different com-
binations of system parameters. Most satellite link analyses do not yield the wanted re-
sult at the first try, and the designer or analyst must use the link budget to adjust system
parameters until an acceptable result is achieved. More examples of link budgets are
included later in this chapter.
4.5 SATELLITE SYSTEMS USING SMALL

EARTH STATIONS
There are many applications in which satellites carry only one or two telephone or data
channels, or a direct-broadcast TV signal and use small, low-cost earth stations. In these
cases, earth stations costing a few hundred or a few thousand dollars are needed. There
are only two parameters in the equation for received power that we can adjust at the satel-
lite to allow us to use a small receiving antenna: satellite transmitted power and satellite
antenna beamwidth.
In domestic satellite systems, narrow beams can be used for transmitting from the
spacecraft to provide coverage over only the region that the system is designed to serve.
Because the dimensions of the antennas that can be mounted on most spacecraft are lim-
ited, the coverage zone cannot be made arbitrarily small. The earth’s disk subtends an
angle of about 17 when viewed from geostationary orbit, and can be illuminated with a
microwave horn having an aperture a few wavelengths in diameter. At 4 GHz, to obtain
a 4 spot beam, a dish 1.4 m in diameter is needed. As the frequency is increased, the di-
ameter of the spacecraft antenna in wavelengths is increased for a given dish diameter,
making it feasible to use more directive beams. However, unless a switched or multiple
beam system is used, the single transmit (or receive) beam must cover the whole region
that the domestic satellite serves. The topic of spacecraft antennas is explored in more
detail in Chapter 3.
As an example, consider the problem of providing service to the 48 contiguous states
of the United States, known as Conus, as illustrated in Figure 4.8; the constraints then
become apparent.
Viewed from a geostationary orbit, at a longitude of around 100 W, the continen-
tal United States subtends an angle of about 3 in the latitude plane (N–S) and 6 in the
longitude plane (E–W). Regardless of the frequency used, an aperture antenna to pro-
duce a single beam with 3-dB beamwidths of 6 by 3° has dimensions approximately
13 by 26, and a gain of 32 dB. For a receiving earth station at the edge of the cover-
age zone, the gain of the satellite antenna in that direction is typically 3 dB lower, or
c04.qxd 28/08/02 19:37 Page 118
Equator
FIGURE 4.8 Domestic satellite link to the Continental United States viewed from geosta-
tionary orbit, a beam approximately 6° 3° is required for coverage of the 48 contiguous
states. Additional beams may be provided for Alaska and Hawaii.
29 dB. For the GEO satellite system shown in Figure 4.8, the received carrier level and
CN ratio can readily be calculated for an earth station with a 3-m antenna at the edge
of the coverage zone, using a transponder with an output power of 5 W at 4 GHz, trans-
mitting a single carrier. Ignoring all losses, and using a receiving antenna gain of 40.0 dB
(63% efficiency)
Pr 7.0 29.0 196.5 40.0 120.5 dBW (4.27)
The noise power at the input to a low noise receiver with noise bandwidth of 30 MHz and
system noise temperature of 100 K is
N kTsBn 228.6 20.0 74.8 133.8 dBW (4.28)
Thus for this system the CN is 13.3 dB. This is some 3.8 dB above an FM threshold of
9.5 dB and provides an adequate margin for an operational system.
These figures are typical of those used in U.S. domestic satellite systems designed
for distribution of television programs to cable TV networks and broadcast TV stations
using a single C-band transponder for each video signal and FM with analog video
signals.
Direct Broadcast TV
Direct broadcast satellite television (DBS-TV) originally started in Europe in the 1980s
using analog FM transmission in Ku band. It achieved a reasonable measure of suc-
cess, due in part to the much slower introduction of cable TV systems in Europe than
c04.qxd 28/08/02 19:37 Page 119
SIDEBAR
FM demodulators exhibit a threshold effect which The threshold is at a CN of 9 dB for a nominal 30-
causes impulsive noise to appear at the output of MHz RF bandwidth, where threshold is defined as an
the demodulator at low CN ratios. When noise in extra loss of 1 dB in SN ratio as the CN ratio is
the receiver approaches the carrier magnitude, sud- reduced.
den rapid changes in phase of the FM signal can oc-
cur. The FM demodulator produces an output volt-
age proportional to the rate of change of phase (i.e.,
S/N at demodulator output (dB)

the instantaneous frequency) and interprets the
rapid phase changes caused by noise as large volt-
age changes. The voltage spikes at the FM demod- 50
ulator output occur at random time intervals, with
increasing numbers of spikes as the CN ratio falls
toward the threshold. In a telephone circuit, the
noise pulses sound like clicks or crackles; they can 40
easily be seen on an oscilloscope or counted by a
pulse counter. FM
threshold
In FM-TV systems, noise spikes that appear close
to threshold cause black and white or colored dots 30
across an otherwise good quality TV picture7. (In 0 10 20 30
satellite TV jargon, these noise spikes are called C/N ratio at demodulator (dB)
sparklies.) The interested reader should refer to a text FIGURE 4.9 Characteristics of a typical FM
on communication system theory for an analysis of threshold extension demodulator for TV recep-
threshold effects in FM demodulators5,6. Figure 4.9 tion. The FM threshold is at CN 9 dB,
shows the characteristics of a typical FM threshold where the SN at the demodulator output has
extension demodulator of a satellite FM-TV receiver. fallen 1 dB below the straight line.
occurred in the United States. In the 1990s, digital transmission became possible, and
several systems were developed in the United States in the 12.2 to 12.7 GHz band
allocated to DBS-TV services. In the United States, DIRECTV, a system built by a con-
sortium led by Hughes, has been very successful and had over 10 million customers by
year-end 2000, offering two hundred television and audio channels. Another DBS-TV
provider in the United States, Echostar, offered similar services and had 4 million cus-
tomers in year 2000. TV and audio channels are available from DBS-TV providers in
a mixture of subscription packages, much like cable TV companies offer, and as pay
per view for individual movies and special events. In rural areas of the United States,
DBS-TV offers hundreds of television channels in place of the three or four terrestrial
broadcasting stations that are typically available. In city areas, DBS-TV offers an
alternative to cable television at a similar cost. The development of low cost Ku-band
antennas and receivers, and high speed digital integrated circuits specifically for DBS
television, has made DBS-TV practical. Chapter 11 gives more detail on the design of
DBS-TV systems.
The 12.2- to 12.7-GHz band was set aside for exclusive use by DBS-TV satellites
in geostationary orbit so that high-power transponders could be used on specially de-
signed DBS-TV satellites. Typical transponder output levels are 100 to 200 W with a
flux density at the earth’s surface of up to 100 dBW/m2. The satellites can carry
16 transponders, with a typical total transmitted RF power of 2.6 kW, higher than for
other commercial satellites. DBS-TV satellites are large and heavy, generally use a three-
axis stabilized design, and have a large area of solar cells to generate the power required
c04.qxd 28/08/02 19:37 Page 120
by the transponders. Typical mass for a DBS-TV satellite is 6800 kg at launch, among
the largest commercial GEO satellites.
The flux density at the earth surface produced by a DBS-TV 160-W transponder is
typically in the range 105 to 115 dBW/m2, which allows small receiving antennas
(dishes) to be used for DBS-TV reception, with diameters in the range 0.45–0.75 m. The
small dish required for DBS-TV reception played a critical part in the acceptance and suc-
cess of DBS-TV in the United States. Previously, DBS-TV reception of cable television
signals was only possible at C band and Ku band with 2.0- to 3.5-m dishes. The local
governments of many cities and towns refused to permit these large dishes in residential
areas, although they became popular in rural areas and an estimated 4 M systems were
sold in the 1980s. Congress passed laws in the 1990s that prevented local governments
from restricting the use of antennas less than 1 m in diameter, opening up a large market
for Ku-band DBS-TV services.
The high flux density created by powerful transponders makes sharing of the DBS-
TV frequency bands impossible, so the 12-GHz DBS band, known as the Broadcast Satel-
lite Service (BSS) band, is allocated exclusively for television broadcasting. The small
home receiving antenna has a wide beam, typically 4 for a 0.45 m (18 inch) dish, which
forces wide spacing of DBS-TV satellites to avoid interference by the signals from adja-
cent DBS-TV satellites. A 9 spacing in the GEO arc has been adopted in the United
States, which restricts the number of DBS-TV satellites that can be placed in geostationary
orbit to serve the United States. In the 1990s the U.S. FCC successfully auctioned spec-
trum and orbital locations for DBS-TV satellites, raising hundreds of millions of dollars
from companies that foresaw a profitable commercial venture.
The DBS-TV system must provide a received signal power at the small receiving
antenna that has an adequate CN margin in clear sky conditions. Heavy rain will cause
attenuation that exceeds the link margin, so occasional outages will be experienced, es-
pecially during the summer months when thunderstorms and heavy rain are more frequent.
The CN margins used in DBS-TV systems are usually quite small to avoid the need for
a large receiving antenna. The selection of a CN margin is a design trade-off between
the outage level that customers can be expected to tolerate, the maximum allowable di-
ameter of the receiving dish antenna, and the power output from the satellite transpon-
ders. Typical designs with receiving antennas in the 0.5 to 0.75 m range and 100–200 W
satellite transponders yield rain attenuation margins of 3 to 8 dB depending on the location
of the receiving terminal within the satellite antenna coverage, and outage times totaling
10 to 40 h per year. These link margins are much lower than those found in high capac-
ity communication systems. Availability of the link has been exchanged for a smaller earth
terminal antenna and lower equipment costs to the user.
In discussing Ku-band rain attenuation, we will use statistics that are representative
of many locations in the central and eastern parts of the United States, where typical path
attenuation in rain exceeds 3 dB for 0.2% (15 h) and 6 dB for 0.01% (52 min) of an av-
erage year. Such attenuation levels at the given time percentage are typical of many tem-
perate latitude locations such as the west coast of the United States, central United States,
Virginia, and other states north of Virginia on the east coast of the United States, Europe,
Chile, Uruguay, and New Zealand. (See Chapter 8 for details of rain attenuation proba-
bilities and distributions.) DIRECTV claims that receiving systems designed for their DBS-
TV transmissions have an average annual availability exceeding 99.7%, which corresponds
to an outage time of 0.3% of the year, or about 25 h. For much of the United States, this
corresponds to rain attenuation in the slant path of 3 dB and requires a link margin of
5.7 dB when allowance is made for the increase in antenna noise temperature that
accompanies 3 dB of rain attenuation.
c04.qxd 28/08/02 19:37 Page 121
TABLE 4.5 Link Budget for Ku-Band DBS-TV Receiver
DBS-TV terminal received signal power

Transponder output power, 160 W 22.0 dBW
Antenna beam on-axis gain 34.3 dB
Path loss at 12.2 GHz, 38,000-m path 205.7 dB
Receiving antenna gain, on axis 33.5 dB
Edge of beam loss 3.0 dB
Clear sky atmospheric loss 0.4 dB
Miscellaneous losses 0.4 dB
Received power, C 119.7 dBW
DBS-TV terminal receiver noise power
Boltzmann’s constant, k 228.6 dBW/K/Hz
System noise temperature, clear sky, 145 K 21.6 dBK
Receiver noise bandwidth, 20 MHz 73.0 dBHz
Noise power, N 134.0 dBW
DBS-TV terminal C/N in clear sky
Clear sky overall C/N 14.3 dB
Link margin over 8.6-dB threshold 5.7 dB
Link availability throughout United States Better than 99.7%
A representative link budget for a GEO DBS-TV system serving the United States
is shown in Table 4.5. The threshold CN value is set at 8.6 dB, corresponding to a sys-
tem using QPSK with an implementation margin of 0.8 dB, half rate forward error cor-
rection that produces 6 dB of coding gain, and a maximum BER of 105. This requires a
clear sky CN ratio in the DBS-TV receiver of 8.6 5.7 14.3 dB. The link budget in
Table 4.5 shows how the required clear sky CN is achieved for a receiver located on the
3 dB contour of the satellite antenna beam. A receiver located in the center of this beam
would have a clear sky CN 3 dB higher, and a corresponding fade margin of 8.7 dB, suf-
ficient to ensure only a few outages each year.
In Table 4.5 a transponder output power of 160 W is used, with no backoff because
a single QPSK signal is transmitted. The satellite antenna gain is 34.3 dB on axis, corre-
sponding to a high efficiency antenna with a beam that is shaped to cover the land mass
of the United States. The beam is approximately 5.5 wide in the E–W direction and 2.5
in the N–S direction. The resulting coverage zone, taking account of the earth’s curvature,
is approximately 4000 km E–W and 2000 km N–S. A maximum path length of 38,000
km is used in this example. The receiving antenna is a high efficiency design with a front-
fed offset parabolic reflector 0.45 m in diameter and a circularly polarized feed. The off-
set design ensures that the feed system does not block the aperture of the antenna, which
increases its efficiency. The gain of this antenna is 33.4 dB at 12.0 GHz with an aperture
efficiency of 67%. The receiver is located at the 3dB contour of the transmitting an-
tenna, and miscellaneous losses of 0.4dB for clear sky attenuation at 12 GHz and 0.4 dB
for receive antenna mispointing and other losses are allowed. The result is a received car-
rier power of 119.7 dBW in clear sky conditions.
The noise power budget of the link is based on a receiver noise bandwidth of 20 MHz.
The IF filters in the receiver must be designed to match the symbol rate of the transmitted
signal, and to approximate a root raised cosine (RRC) transfer function. (See Chapter 5
for details of digital transmission techniques.) The noise bandwidth of all RRC filters is
always equal to the symbol rate of the digital transmission. In the DBS-TV system
c04.qxd 28/08/02 19:37 Page 122
described in Table 4.5, a QPSK signal with a symbol rate of 20 Msps is assumed, which
results in a receiver noise bandwidth of 20 MHz. The 20-Msps QPSK transmission delivers
a bit rate of 40 Mbps, but the half rate FEC coding reduces the data rate to 20 Mbps. A
20-Mbps data stream can carry three live compressed digital video signals using MPEG
2 encoding, or up to 10 prerecorded and processed video signals. The antenna noise
temperature is set at 35 K in clear sky conditions, and a 12-GHz LNA with 110 K noise
temperature is used. The result is a noise power of 134.0 dBW referred to the input of
the LNA, and a clear sky CN ratio of 14.3 dB. This is a worst case result for clear sky
conditions, since most of the DBS-TV receivers will lie inside the 3 dB contour of the
satellite beam.
DBS-TV receivers can be used outside the 3 dB contour of the satellite beam, but
will have a lower link margin and consequently more outages per year, if heavy rain occurs
frequently. For example, a receiver on the 5 dB contour of the satellite beam will have
a link margin of 3.7 dB, which would allow about 2 dB of rain attenuation before the
CN reaches threshold. If the user is in a relatively dry area, for example, central Canada,
the performance of the receiving system may be quite acceptable. The CN ratio in the
home receiver will fall when rain is in the path between the satellite and the receiving an-
tenna. Much of the reduction in CN ratio is caused by an increase in the sky noise tem-
perature. The following calculations show how the system noise temperature and (CN)dn
are determined when rain attenuation is present.
The first step is to determine the total path attenuation, A in dB, which is the sum
of the clear sky path attenuation due to atmospheric gaseous absorption, Aca and attenuation
due to rain, Arain.
A Aca Arain dB (4.29)
The sky noise temperature resulting from a path attenuation Atotal dB is found from the
output noise model of Section 4.3 using an assumed medium temperature of 270 K for
the rain.
Tsky 270 11 10A10 2 K (4.30)
The antenna noise temperature may be assumed to be equal to the sky noise tempera-
ture, although in practice not all of the incident noise energy from the sky is output
by the antenna, and a coupling coefficient, c, of 90 to 95% is often used when cal-
culating antenna noise temperature in rain. Thus antenna noise temperature may be
calculated as
TA hc Tsky K (4.31)
Almost all satellite receivers use a high gain LNA as the first element in the receiver
front end. This makes the contribution of all later parts of the receiver to the system noise
temperature negligible. System noise temperature is then given by
Ts rain TLNA TA K (4.32)
In Eq. (4.32), the LNA is assumed to be placed right at the feed horn so that there is no
waveguide or coaxial run between the feed horn of the antenna and the LNA. We will as-
sume that there are no feed losses. The increase in noise power, Nrain dB, caused by the
increase in sky noise temperature is given by
¢Nrain 10 log10 c d 10 log10 c d dB

kTs rain Bn Ts rain
(4.33)
kTsca Bn Tsca
where Tsca is the system noise temperature in clear sky conditions
c04.qxd 28/08/02 19:37 Page 123
The received power is reduced by the attenuation caused by the rain in the slant
path, so in rain the value of carrier power is Crain where
Crain Cca Arain dB (4.34)
The resulting (CN)dn rain value when rain intersects the downlink is given by
1CN2 dn rain 1CN2 dn ca Arain ¢Nrain dB (4.35)
where (CN)dn ca is the downlink CN ratio in clear sky conditions.
If a linear (bent pipe) transponder is used, the (CN)up ratio must be combined with
(CN)dn rain to yield the overall (CN)0 ratio for the link. Many digital systems use regenera-
tive transponders that provide constant output power regardless of uplink attenuation pro-
vided that the received CN ratio at the satellite is above the threshold of the onboard
processing demodulator. In this case the value of (CN)dn rain will be used as the overall
(CN)0 value in rain for the link.
EXAMPLE 4.5.1
In the example of a DBS-TV system in Table 4.5, a link margin of 5.7 dB is available before the
(CN)0 threshold of 8.6 dB is reached. We must calculate the increase in system noise temperature
that results from 3-dB rain attenuation in the downlink path to determine the increase in noise power
and thus the value of (CN)dn rain. This example shows how the link margin can be distributed between
rain attenuation and an increase in receiver system noise power caused by an increase in sky noise.
The clear sky attenuation is given in Table 4.5 as 0.4 dB. Thus total excess attenuation is
3.4 dB, and the sky noise temperature in rain will be, from Eq. (4.30)
Tsky rain 270 11 103.410 2 147 K
In clear sky conditions the sky noise temperature is Tca [0.95 (270 (1 100.04))] 23 K.
The sky temperature has increased from 23 K in clear sky conditions to 147 K when 3 dB rain
attenuation occurs in the downlink. We must calculate the new system noise temperature when rain
is present in the slant path, remembering that the antenna noise temperature has two parts: a fixed
part due to noise from the ground entering the antenna’s sidelobes, and a variable part due to sky
temperature. The LNA of the system in Table 4.5 has a noise temperature of 110 K and the clear
sky system noise temperature is Tca 145 K. The antenna temperature in clear sky conditions is
made up of a fixed portion of 12 K and a sky noise portion of 23 K. If we assume 95% coupling
of sky noise temperature to antenna noise temperature, the system noise temperature, given by
Eq. (4.32), increases to
Ts rain TLNA TA 110 12 147 269 K
The increase in receiver noise power referred to the receiver input is given by Eq. 4.33
¢Nrain 10 log10 3Ts rainTs ca 4 10 log10 3 269145 4 2.7 dB
From Eq. (4.35)
1CN2 dn rain 14.3 3.0 2.7 8.6 dB
Thus the link margin required to withstand 3 dB of rain attenuation in the downlink path is
5.7 dB, and this will guarantee a link availability exceeding 99.7% of an average year for the ma-
jority of the DBS-TV system customers in the United States. In the south eastern United States, in
states such as Florida and Louisiana where very heavy rain occurs more often than in other parts
of the United States, link availability may be slightly less than 99.7%. Shaping of the satellite beam
to direct more power to these parts of the United States helps to reduce the number of outages ex-
perienced in that region. Receivers located within the 1 dB contour of the satellite antenna beam
have shorter path lengths giving 2.3 dB higher CN than the receiver used in the example shown
in Table 4.5, so they have a link margin of 8 dB.
c04.qxd 28/08/02 19:37 Page 124
The calculation of the availability of these receivers requires some care, because we cannot
just add the extra 2.3 dB of link margin to the rain attenuation. Antenna noise increases with every
decibel of extra rain attenuation, reducing the received power level, C, and increasing the system
noise power N. An iterative (trial and error) procedure must be used to find the combination of
reduction in C and increase in N that leads to an additional 2.3-dB degradation in the CN value.
We will guess that increasing the rain attenuation from 3 to 4.5 dB leads to an increase in
noise power of 0.8 dB. With 4.5-dB rain attenuation, the system noise temperature is
Ts rain 110 12 0.95 270 11 0.3552 287 K
The increase in noise power from the clear sky condition is
¢N 10 log10 12871452 3.0 dB
Hence the decrease in CN for 4.5 dB of rain attenuation is 4.5 3.0 7.5 dB, and we are
a little below our new margin of 8.0 dB. Increasing the rain attenuation by a further 0.4 dB to 4.9 dB
and repeating the calculations gives a new system noise temperature in rain of 295 K. The corre-
sponding increase in noise power is 3.1 dB giving a reduction in CN of 8.0 dB, equal to the link
margin. A rain attenuation margin of 4.9 dB at Ku band would give an availability of 99.88% or
better over the central region of the United States.
This example demonstrates that the increase in noise temperature of a low noise DBS-TV
Ku-band receiving system is a significant factor when rain attenuation is present in the downlink
path. Rain attenuation alone cannot be equated to link margin.
4.6 UPLINK DESIGN
The uplink design is easier than the downlink in many cases, since an accurately speci-
fied carrier power must be presented at the satellite transponder and it is often feasible to
use much higher power transmitters at earth stations than can be used on a satellite. How-
ever, VSAT systems use earth stations with small antennas and transmitter powers below
5 W, giving low uplink EIRP. Satellite telephone handsets are restricted to transmitting at
power levels below 1 W because of the risk of EM radiation hazards. In mobile systems
the uplink from the satellite telephone is usually the link with the lowest CN ratio.
The cost of transmitters tends to be high compared with the cost of receiving equip-
ment in satellite communication systems. The major growth in satellite communications
has been in point-to-multipoint transmission, as in cable TV distribution and direct broad-
cast satellite television. One high-power transmit earth station provides service via a DBS
satellite to many low-cost receive-only stations, and the high cost of the transmitting station
is only a small part of the total network cost.
The satellite transponder is a quasilinear amplifier and the received carrier level deter-
mines the output level. Where a traveling wave tube is used as the output high-power am-
plifier (HPA) in the transponder, as is often the case, and FDMA is employed, the HPA must
be run with a predetermined backoff to avoid intermodulation products appearing at the out-
put. The output backoff is typically 1 to 3 dB when more than one signal is present in the
transponder, and is determined by the uplink carrier power level received at the spacecraft.
Accurate control of the power transmitted by the earth station is therefore essential, which is
easily achieved in a fixed network of earth stations. Where a very large number of earth sta-
tions access a single transponder using FDMA, such as in some VSAT networks and Intelsat
satellites, transponder output backoff of 5 to 7 dB may be required to maintain intermodulation
products at a sufficiently low level10. Even with a single access to the transponder (i.e., only
one carrier present) some backoff is normally applied to avoid the PM-AM conversion that
occurs when modulated signals are transmitted through nonlinear devices9.
c04.qxd 28/08/02 19:37 Page 125
4.6 UPLINK DESIGN 125
Earth station transmitter power is set by the power level required at the input to the
transponder. This can be done in one of two ways. Either a specific flux density is re-
quired at the satellite, or a specific power level is required at the input to the transponder.
Early Intelsat C-band satellites required high flux densities to saturate their transponders,
in the range 73.7 to 67.5 dBW/m2, depending on the transponder gain setting8. This
is a high flux density which requires a large earth station and a powerful transmitter gen-
erating up to 3 kW. Domestic GEO satellites operating into North America generally re-
quire lower flux densities allowing the use of smaller earth station antennas. At C band,
a typical uplink earth station transmits 100 W with a 9-meter antenna, giving a flux density
at the satellite of 100 W/m2.
Although flux density at the satellite is a convenient way to determine earth station
transmit EIRP requirements, analysis of the uplink requires calculation of the power level at
the input to the transponder so that the uplink CN ratio can be found. The link equation is
used to make this calculation, using either a specified transponder CN ratio or a required
transponder output power level. When a CN ratio is specified for the transponder, the cal-
culation of required transmit power is straightforward. Let (CN)up be the specified CN ratio
in the transponder, measured in a noise bandwidth Bn Hz. The bandwidth Bn Hz is the band-
width of the band-pass filter in the IF stage of the earth station receiver for which the uplink
signal is intended. Even if Bn is much less than the transponder bandwidth, it is important
that the uplink CN ratio be calculated in the bandwidth of the receiver, not the bandwidth
of the transponder. The noise power referred to the transponder input is Nxp W where
Nxp k Txp Bn dBW (4.36)
where Txp is the system noise temperature of the transponder in dBK and Bn is in units of
dBHz.
The power received at the input to the transponder is Prxpwhere
Prxp Pt Gt Gr Lp Lup dBW (4.37)
where PtGt is the uplink earth station EIRP in dBW, Gr is the satellite antenna gain in dB
in the direction of the uplink earth station and Lp is the path loss in dB. The factor Lup dB
accounts for all uplink losses other than path loss. The value of (CN)up at the LNA input
of the satellite receiver is given by
CN 10 log10 3Pr 1k Ts Bn 2 4 Prxp Nxp dB (4.38)
The earth station transmitter output power Pt is calculated from Eq. (4.23) using the
given value of CN in Eq. (4.38) and the noise power Nxp calculated from Eq. (4.36). Note
that the received power at the transponder input is also given by
Prxp N CN dBW (4.39)
The earth station transmitter output power Pt can also be calculated from the output power
of the transponder and transponder gain when these parameters are known and a bent pipe
transponder is used. In general
Prxp Psat BO0 Gxp dBW (4.40)
where Psat is the saturated power output of the transponder in dBW, BO0 is the output
backoff in dB, and Gxp is the gain of the transponder in dB.
With small-diameter earth stations, a higher power earth station transmitter is required
to achieve a similar satellite EIRP. This has the disadvantage that the interference level at
adjacent satellites rises, since the small earth station antenna inevitably has a wider beam.
Thus it is not always possible to trade off transmitter power against uplink antenna size.
c04.qxd 28/08/02 19:37 Page 126
G(θ ) dB
40
G(θ ) = 29 − 25 log θ
20 G(θ ) = 32 − 25 log
FIGURE 4.10 ITU-R

0
specifications on the side-
lobe envelope of transmit
−20 −10 0 10 20 antenna patterns for 2°
Angle off axis, θ (degrees) GEO satellite spacing.
There is a specification for transmit station antenna patterns, designed to minimize inter-
ference from adjacent uplinks. It is the uplink interference problem that determines satellite
spacing and limits the capacity of the geostationary orbit in any frequency band. Figure 4.10
shows the ITU-R (formerly CCIR) specification, G() (32 25 log10 ) dB, where is
in degrees, for 1 degree off axis, for satellites spaced by 3.
To increase the capacity of the crowded geostationary orbit arc south of the
United States, the FCC introduced new regulations in 1983 requiring better control of
6-GHz earth station antenna transmit patterns so that intersatellite spacing could be re-
duced to 2. The same specification has now been adopted by the ITU-R for the entire
geostationary arc. The requirement is for the transmit antenna pattern to lie below
G() 29 25 log10 dB in the range 1

7 from the antenna boresight and
G() 32 25 log10 dB beyond 711. The required antenna pattern envelope is
shown in Figure 4.10.
At frequencies above 10 GHz, for example, 14.6 GHz and 30 GHz, propagation dis-
turbances in the form of fading in rain cause the received power level at the satellite to
fall. This lowers the uplink CN ratio in the transponder, which lowers the overall (CN)0
ratio in the earth station receiver when a linear (bent pipe) transponder is used on the satel-
lite. Uplink power control (UPC) can be used to combat uplink rain attenuation. The trans-
mitting earth station monitors a beacon signal from the satellite, and watches for reductions
in power indicating rain fading on the downlink. Automatic monitoring and control of
transmitted uplink power is used in 14-GHz uplink earth stations to maintain the uplink
CN ratio in the satellite transponder during periods of rain attenuation. New generations
of Ka-band satellites employ uplink power level detection at the satellite. A control link
to each uplink earth station closes the loop.
Since the downlink is always at a different frequency from the uplink, a downlink
attenuation of A dB must be scaled to estimate uplink attenuation. The scaling factor used
is typically ( fup fdown)a where a is typically between 2.0 and 2.4. For example, an uplink
station transmitting at 14.0 GHz to a Ku-band satellite monitors the satellite beacon at
11.45 GHz. The uplink attenuation is therefore given by
Aup Adown 1 fup fdown 2 a dB (4.41)
where Aup is the estimated uplink rain attenuation and Adown is the measured downlink rain
attenuation. For a value of a 2.2 and ( fup fdown) 1.222 the factor ( fup fdown)a is 1.56.
c04.qxd 28/08/02 19:37 Page 127
4.7 DESIGN FOR SPECIFIED C/N: COMBINING C/N AND C/I VALUES IN SATELLITE LINKS 127
Hence a downlink rain attenuation of 3 dB would give an estimated uplink attenuation of

4.7 dB. This uplink attenuation value applies only to rain and does not include gaseous
attenuation or scintillation, which require different scaling ratios.
Uplink power control cannot be applied until a certain amount of attenuation has built
up in the link. This is typically around 2 dB for the downlink due to measurement inaccu-
racies, corresponding to about 3 dB for a Ka-band uplink. As rain begins to affect the link
between the earth station and satellite, the uplink CN ratio in the transponder will fall until
UPC starts to operate in the earth station transmitter. The transponder CN ratio will then
remain relatively constant until the UPC system reaches the maximum available transmit
power. Further attenuation on the uplink will cause the CN ratio in the transponder to fall.
EXAMPLE 4.6.1
A transponder of a Ku-band satellite has a linear gain of 127 dB and a nominal output power at sat-
uration of 5 W. The satellite’s 14-GHz receiving antenna has a gain of 26 dB on axis, and the beam
covers western Europe.
Calculate the power output of an uplink transmitter that gives an output power of 1 W from
the satellite transponder at a frequency of 14.45 GHz when the earth station antenna has a gain of
50 dB and there is a 1.5-dB loss in the waveguide run between the transmitter and antenna. Assume
that the atmosphere introduces a loss of 0.5 dB under clear sky conditions and that the earth sta-
tion is located on the 2 dB contour of the satellite’s receiving antenna. If rain in the path causes
attenuation of 7 dB for 0.01% of the year, what output power rating is required for the transmitter
to guarantee that a 1-W output can be obtained from the satellite transponder for 99.99% of the year
if uplink power control is used?
The input power required by the transponder is simply the output power minus the transpon-
der gain, so
Pin 0 dBW 127 dB 127 dBW
The uplink power budget is given by Eq. (4.11)
Pr EIRP Gr Lp Lat Lta Lra dBW
Rearranging and putting in the appropriate losses
Pt Pr Gt Gr Lp Lta Lat Lpt dBW
where Lta is the waveguide loss, Lat is the atmospheric loss, and Lpt is the pointing loss (antenna pat-
tern loss). Then assuming a path length of 38,500 km
Pt 127.0 50 26 207.2 1.5 0.5 2.0 dBW
That is
Pt 7.2 dBW or 5.2 W
If we provide an extra 7 dB of output power to compensate for fading on the path due to
rain, the transmitter output power will be
Pt rain 7.2 7 14.2 dBW or 26.3 W
4.7 DESIGN FOR SPECIFIED C/N: COMBINING

C/N AND C/I VALUES IN SATELLITE LINKS
The BER or SN ratio in the baseband channel of an earth station receiver is determined
by the ratio of the carrier power to the noise power in the IF amplifier at the input to the
demodulator. The noise present in the IF amplifier comes from many sources. So far in
c04.qxd 28/08/02 19:37 Page 128
our analysis of uplinks and downlinks we have considered only the receiver thermal noise
and noise radiated by atmospheric gases and rain in the slant path. When a complete satel-
lite link is engineered, the noise in the earth station IF amplifier will have contributions
from the receiver itself, the receiving antenna, sky noise, the satellite transponder from
which it receives the signal, and adjacent satellites and terrestrial transmitters which share
the same frequency band.
In the first edition of this text, a method was presented for adding these noise con-
tributions together in a hypothetical reference circuit. This was the standard technique used
in CCIR (now ITU-R) Recommendations at the time the first edition of this book was writ-
ten. More recently, addition of CN and CI (carrier-to-interference) ratios has been widely
used, and is easier to apply than the reference circuit approach, although both methods will
lead to the same result. The latter method using CN values is presented here.
When more than one CN ratio is present in the link, we can add the individual
CN ratios reciprocally to obtain an overall CN ratio, which we will denote here as
(CN)0. The overall (CN)0 ratio is what would be measured in the earth station at the
output of the IF amplifier
1C N2 0 1 31 1CN2 1 1 1CN2 2 1 1CN2 3 p 4 (4.42)
This is sometimes referred to as the reciprocal CN formula. The CN values must be
linear ratios, NOT decibel values. Since the noise power in the individual CN ratios is ref-
erenced to the carrier power at that point, all the C values in Eq. (4.42) are the same. Expanding
the formula by cross multiplying gives the overall (CN)0 as a power ratio, not in decibels
1C N2 0 1 1N1C N2 C N3 C p 2 C 1N1 N2 N3 p 2 (4.43)
In decibel units:
1C N2 0 C dBW 10 log10 1N1 N2 N3 p W2 dB (4.44)
Note that (CN)dn cannot be measured at the receiving earth station. The satellite always
transmits noise as well as signal, so a CN ratio measurement at the receiver will always
yield (CN)0, the combination of transponder and earth station CN ratios.
To calculate the performance of a satellite link we must therefore determine the
uplink (CN)up ratio in the transponder and the downlink (CN)dn in the earth station
receiver. We must also consider whether there is any interference present, either in the
satellite receiver or the earth station receiver. One case of importance is where the transpon-
der is operated in a FDMA mode and intermodulation products (IM) are generated by the
transponder’s nonlinear input–output characteristic. If the IM power level in the transpon-
der is known, a CI value can be found and included in the calculation of (CN)0 ratio.
Interference from adjacent satellites is likely whenever small receiving antennas are used,
as with VSATs (very small aperture terminals) and DBS-TV receivers.
Since CN values are usually calculated from power and noise budgets, their val-
ues are typically in decibels. There are some useful rules of thumb for estimating (CN)0
from two CN values:
• If the CN values are equal, as in the example above, (CN)0 is 3 dB lower than
either value.
• If one CN value is 10 dB smaller than the other value, (CN)0 is 0.4 dB lower than
the smaller of the CN values.
• If one CN value is 20 dB or more greater than the other CN value, the overall
(CN)0 is equal to the smaller of the two CN values within the accuracy of deci-
bel calculations (0.1 dB).
c04.qxd 28/08/02 19:37 Page 129
EXAMPLE 4.7.1
Thermal noise in an earth station receiver results in a (CN)dn ratio of 20.0 dB. A signal is received
from a bent pipe transponder with a carrier to noise ratio (CN)up 20.0 dB. What is the value of
overall (CN)0 at the earth station? If the transponder introduces intermodulation products with (CI)
ratio 24 dB, what is the overall (CN)0 ratio at the receiving earth station?
Using Eq. (4.42) and noting that (CN) 20.0 dB corresponds to a (CN) ratio of 100
1CN2 0 c d c d 50 or 17.0 dB
1 1
1 1CN2 up 1 1CN2 dn 0.01 0.01
The intermodulation (CI) value of 24.0 dB corresponds to a ratio of 0.004. The overall (CN)0
value is then
1CN2 0 c d 41.7 or 16.2 dB

1

0.01 0.01 0.004
Overall (C/N)0 with Uplink and

Downlink Attenuation
Most satellite links are designed with link margins to allow for attenuation that may oc-
cur in the link or increases in noise power caused by interference. (Interference is almost
always treated as though it were white noise, regardless of whether the interfering signal
actually has a uniform spectral power distribution or Gaussian statistics. When the inter-
ference has known characteristics, such as a depolarized cochannel or a jamming signal,
cancellation techniques can be used to reduce the level of interference.)
The effect of a change in the uplink CN ratio has a different impact on overall
(CN)0 depending on the operating mode and gain of the transponder.
There are three different transponder types or operating modes:
Linear transponder: Pout Pin Gxp dBW
Nonlinear transponder: Pout Pin Gxp ¢G dBW
Regenerative transponder: Pout constant dBW
where Pin is the power delivered by the satellite’s receiving antenna to the input of the
transponder, Pout is the power delivered by the transponder HPA to the input of the satel-
lite’s transmitting antenna, Gxp is the gain of the transponder, and all parameters are in
decibel units.
The parameter G is dependent on Pin and accounts for the loss of gain caused by
the nonlinear saturation characteristics of a transponder which is driven hard to obtain
close to its maximum power output—the gain is effectively falling as the input power
level increases. (See Chapter 6 for a detailed discussion of intermodulation effects and
nonlinearity in transponders.)
The maximum output power from a transponder is called the saturated output power
and is the nominal transponder power output rating that is usually quoted. The transponder
input–output characteristic is highly nonlinear when operated at this output power level.
When a transponder is operated close to its saturated output power level, digital waveforms
are changed, resulting in intersymbol interference (ISI), and FDMA operation results in the
generation of intermodulation products by multiplication of the individual signals. Transpon-
ders are usually operated with output backoff, to make the characteristic more nearly linear.
The exact amount of output backoff required in any given application depends on the spe-
cific characteristics of the transponder and the signals it carries. Typical values of output
backoff are 1 dB for a single FM or PSK carrier to 3 dB for FDMA operation with several
c04.qxd 28/08/02 19:37 Page 130
carriers. The corresponding input backoff values might be 3 dB and 5 dB, but the individ-
ual transponder characteristic must be known to make an accurate assessment. For conven-
ience in this text, we will frequently assume linear transponder operation when calculating
the overall (CN)0 ratio, even if this may not, in fact, be the case.
Uplink and Downlink Attenuation in Rain

Rain attenuation affects uplinks and downlinks differently. We usually assume that rain
attenuation is occurring on either the uplink or the downlink, but not on both at the same
time. This is usually true for earth stations that are well separated geographically, but not
if they are close together (
20 km). Heavy rain occurs with a somewhat random geo-
graphic distribution for less than 1% of the time, so the probability of significant attenu-
ation occurring on both the uplink and downlink simultaneously is small. In the following
analysis of uplink and downlink attenuation effects, it will be assumed that one link is
attenuated and the other is operating in clear air.
Uplink Attenuation and (C/N)up

The transponder receiver noise temperature does not change significantly when rain is
present in the uplink path to the satellite. The satellite receiving antenna beam is always
sufficiently wide that it “sees” a large area of the (warm) earth’s surface and local noise
temperature variations are insignificant. The noise temperature of the earth seen by a GEO
satellite varies from a maximum of 270 K for a satellite antenna beam over Africa and
northern Europe, to a minimum of 250 K over the Pacific Ocean. The corresponding sys-
tem noise temperature for the transponders on a GEO satellite is in the range 400 to 500 K.
There is effectively no increase in uplink noise power when heavy rain is present in the
link between an earth station and a satellite because the satellite antenna beam sees the
tops of cumulonimbus clouds above the rain, which are always colder than 270 K, instead
of the earth’s surface.
Rain attenuation on the uplink path to the satellite reduces the power at the satel-
lite receiver input, and thus reduces (CN)up in direct proportion to the attenuation on the
slant path. If the transponder is operating in a linear mode, the output power will be re-
duced by the same amount, which will cause (CN)dn to fall by an amount equal to the
attenuation on the uplink. When both (CN)up and (CN)dn are reduced by Aup dB, the
value of (CN)0 is reduced by exactly the same amount, Aup dB. Hence for the case of a
linear transponder and rain attenuation in the uplink of Aup dB
1C N2 0 uplink rain 1CN2 0 clear air Aup dB Linear transponder (4.45)
If the transponder is nonlinear, the reduction in input power caused by uplink attenuation
of Aup dB results in a smaller reduction in output power, by an amount G.
1C N2 0 uplink rain 1C N2 0 clear air Aup ¢G dB Nonlinear transponder (4.46)
If the transponder is digital and regenerative, or incorporates an Automatic Gain Control
(AGC) system to maintain a constant output power level
1CN2 0 uplink rain 1CN2 0 clear air dB Regenerative transponder or AGC (4.47)
The above equation will hold only if the received signal is above threshold and the BER
of the recovered digital signal in the transponder is small. If the signal falls below threshold,
the uplink will contribute significantly to the BER of the digital signal at the receiving
earth station.
c04.qxd 28/08/02 19:37 Page 131
Downlink Attenuation and (C/N)dn

The earth station receiver noise temperature can change very significantly when rain is
present in the downlink path from the satellite. The sky noise temperature can increase to
close to the physical temperature of the individual raindrops, particularly in very heavy
rain. A reasonable temperature to assume for temperate latitudes in a variety of rainfall
rates is 270 K, although values above 290 K have been observed in the tropics. An increase
in sky noise temperature to 270 K will increase the receiving antenna temperature markedly
above its clear air value. See Section 4.5 for an illustration of this effect. The result is that
the received power level, C, is reduced and the noise power, N, in the receiver increases.
The result for downlink CN is given by Eq. (4.48)
1CN2 dn rain 1CN2 dn clear air Arain ¢Nrain dB (4.48)
The overall CN is then given by
1CN2 0 1 31 1CN2 dn rain 1 1C N2 up 4 dB (4.49)
As noted earlier, unless we are making a loop-back test, we will assume that the value
of (CN)up is for clear air, and remains constant regardless of the attenuation on the
downlink.
System Design for Specific Performance

A typical two-way satellite communication link consists of four separate paths: an out-
bound uplink path from one terminal to the satellite and an outbound downlink to the
second terminal; and an inbound uplink from the second terminal to the satellite and an
inbound downlink to the first terminal. The links in the two directions are independent
and can be designed separately, unless they share a single transponder using FDMA. A
broadcast link, like the DBS-TV system described earlier in this chapter, is a one-way
system, with just one uplink and one downlink.
Satellite Communication Link

Design Procedure
The design procedure for a one-way satellite communication link can be summarized by
the following 10 steps. The return link design follows the same procedure.
1. Determine the frequency band in which the system must operate. Comparative
designs may be required to help make the selection.
2. Determine the communications parameters of the satellite. Estimate any values that
are not known.
3. Determine the parameters of the transmitting and receiving earth stations.
4. Start at the transmitting earth station. Establish an uplink budget and a transponder
noise power budget to find (CN)up in the transponder.
5. Find the output power of the transponder based on transponder gain or output
backoff.
6. Establish a downlink power and noise budget for the receiving earth station.
Calculate (CN)dn and (CN)0 for a station at the edge of the coverage zone
(worst case).
7. Calculate SN or BER in the baseband channel. Find the link margins.
c04.qxd 28/08/02 19:37 Page 132
8. Evaluate the result and compare with the specification requirements. Change pa-
rameters of the system as required to obtain acceptable (CN)0 or SN or BER
values. This may require several trial designs.
9. Determine the propagation conditions under which the link must operate. Calculate
outage times for the uplinks and downlinks.
10. Redesign the system by changing some parameters if the link margins are inadequate.
Check that all parameters are reasonable, and that the design can be implemented
within the expected budget.
4.8 SYSTEM DESIGN EXAMPLES
The following sample system designs demonstrate how the ideas developed in this chap-
ter can be applied to the design of satellite communication systems.
TABLE 4.6 System and Satellite Specification
Ku-band satellite parameters

Geostationary at 73° W longitude, 28 Ku-band transponders
Total RF output power 2.24 kW
Antenna gain, on axis (transmit and receive) 31 dB
Receive system noise temperature 500 K
Transponder saturated output power: Ku band 80 W
Transponder bandwidth: Ku band 54 MHz
Signal Compressed digital video signals with transmitted symbol rate of 43.2 Msps
Minimum permitted overall (CN)0 in receiver 9.5 dB
Transmitting Ku-band earth station
Antenna diameter 5m
Aperture efficiency 68%
Uplink frequency 14.15 GHz
Required CN in Ku-band transponder 30 dB
Transponder HPA output backoff 1 dB
Miscellaneous uplink losses 0.3 dB
Location: 2 dB contour of satellite receiving antenna
Receiving Ku-band earth station
Receiver IF noise bandwidth 43.2 MHz
Antenna noise temperature 30 K
LNA noise temperature 110 K
Required overall (CN)0 in clear air 17 dB
Miscellaneous downlink losses 0.2 dB
Location: 3 dB contour of satellite transmitting antenna
Rain attenuation and propagation factors
Ku-band clear air attenuation
Uplink 14.15 GHz 0.7 dB
Downlink 11.45 GHz 0.5 dB
Rain attenuation
Uplink 0.01% of year 6.0 dB
Downlink 0.01% of year 5.0 dB
c04.qxd 28/08/02 19:37 Page 133
4.8 SYSTEM DESIGN EXAMPLES 133
System Design Example 4.8.1

This example examines the design of a satellite communication link using a Ku-band geo-
stationary satellite with bent pipe transponders to distribute digital TV signals from an
earth station to many receiving stations throughout the United States. The design requires
that an overall CN ratio of 9.5 dB be met in the TV receiver to ensure that the video
signal on the TV screen is held to an acceptable level. The uplink transmitter power and
the receiving antenna gain and diameter are determined for each system. The available
link margins for each of the systems are found and the performance of the systems is an-
alyzed when rain attenuation occurs in the satellite–earth paths. The advantages and
disadvantages of implementing uplink power control are considered.
In this example, the satellite is located at 73 W. However, for international regis-
tration of this satellite location, the location would be denoted as 287 E. The link budg-
ets developed in the examples below use decibel notation throughout. The satellite and
earth stations are specified in Table 4.6, and Figure 4.11 shows an illustration of the satellite
television distribution system.
Ku-Band Uplink Design

We must find the uplink transmitter power required to achieve (CN)up 30 dB in
clear air atmospheric conditions. We will first find the noise power in the transponder for
43.2 MHz bandwidth, and then add 30 dB to find the transponder input power level.
Uplink Noise Power Budget
Ts 500 K 27.0 dBK
B 43.2 MHz 76.4 dBHz
N transponder noise power 125.2 dBW
The received power level at the transponder input must be 30 dB greater than the noise power.
Pr power at transponder input 95.2 dBW
Ku-band satellite
Uplink
Earth Ku-band
station uplink
Ku-band
downlinks
Program material
from studios
To cable TV
network
Cable TV receive stations

FIGURE 4.11 Satellite television distribution system.
c04.qxd 28/08/02 19:37 Page 134
The uplink antenna has a diameter of 5 m and an aperture efficiency of 68%. At 14.15 GHz
the wavelength is 2.120 cm 0.0212 m. The antenna gain is
Gt 10 log 30.68 1pDl2 2 4 55.7 dB
The free space path loss is Lp 10 log [(4R)2] 207.2 dB
Uplink Power Budget
Pt Earth station transmitter power Pt dBW
Gt Earth station antenna gain 55.7 dB
Gr Satellite antenna gain 31.0 dB
Lp Free space path loss 207.2 dB
Lant ES on 2 dB contour 2.0 dB
Pr Received power at transponder Pt 123.5 dB
The required power at the transponder input to meet the (CN)up 30 dB objective is
95.2 dBW. Hence
Pt 123.5 dB 95.2 dBW
Pt 28.3 dBW or 675 W
This is a relatively high transmit power so we would probably want to increase the
transmitting antenna diameter to increase its gain, allowing a reduction in transmit
power.
Ku-Band Downlink Design

The first step is to calculate the downlink (CN)dn that will provide (CN)0 17 dB when
(CN)up 30 dB. From Eq. (4.43)
1 1C N2 dn 1 1C N2 0 1 1CN2 up 1not in dB2
Thus
1 1CN2 dn 1 50 1 1000 0.019
1C N2 dn 52.6 1 17.2 dB
We must find the required receiver input power to give (CN)dn 17.2 dB and then find
the receiving antenna gain, Gr.
Downlink Noise Power Budget
Ts 30 110 K 140 K 21.5 dBK
Bn 43.2 MHz 76.4 dBHz
N transponder noise power 130.7 dBW
The power level at the earth station receiver input must be 17.2 dB greater than the
noise power in clear air.
Pr power at earth station receiver input 130.7 dBW 17.2 dB 113.5 dBW
We need to calculate the path loss at 11.45 GHz. At 14.15 GHz path loss was 207.2
dB. At 11.45 GHz path loss is
Lp 207.2 20 log10 114.1511.452 205.4 dB
c04.qxd 28/08/02 19:37 Page 135
The transponder is operated with 1 dB output backoff, so the output power is 1 dB below
80 W (80 W 1 19.0 dBW)
Pt 19 dBW 1 dB 18 dBW
Downlink Power Budget

Pt Satellite transponder output power 18.0 dBW
Gt Satellite antenna gain 31.0 dB
Gr Earth station antenna gain Gr dB
Lp Free space path loss 205.4 dB
La ES on 3 dB contour of satellite antenna 3.0 dB
Pr Received power at transponder Gr 160.2 dB
The required power into the earth station receiver to meet the (CN)dn 17.2 dB objec-
tive is Pr 120.1 dBW. Hence the receiving antenna must have a gain Gr where
Gr 160.2 dB 113.5 dBW
Gr 46.7 dB or 46,774 as a ratio
The earth station antenna diameter, D, is calculated from the formula for antenna gain, G,
with a circular aperture
Gr 0.65 1pDl2 2 46,774
At 11.45 GHz, the wavelength is 2.62 cm 0.0262 m. Evaluating the above equation to
find D gives the required receiving antenna diameter as D 2.14 m.
Rain Effects at Ku Band

Uplink Under conditions of heavy rain, the Ku-band path to the satellite station suffers
an attenuation of 6 dB for 0.01% of the year. We must find the uplink attenuation margin
and decide whether uplink power control would improve system performance at Ku band.
The uplink CN was 30 dB in clear air. With 6 dB uplink path attenuation, the CN
in the transponder falls to 24 dB, and assuming a linear transponder characteristic and no
uplink power control, the transponder output power falls to 18 8 12 dBW. The down-
link CN falls by 6 dB from 17.2 dB to 11.2 dB, and the overall (CN)0 falls by 6 dB to
11 dB. With the minimum overall CN set at 9.5 dB, the additional margin for uplink atten-
uation is 1.5 dB. Hence the link margin available on the uplink is 7.5 dB without uplink
power control. This is an adequate uplink rain attenuation margin for many parts of the United
States, and would typically lead to rain outages of less than 1 h total time per year.
Uplink power control (UPC) could be implemented so that the earth station trans-
mitter output power is increased when the uplink attenuation is estimated to have reached
3 dB. This would hold the value of overall (CN)0 in the receiver at 14 dB. If the UPC
system has a dynamic range of 6 dB, the uplink rain attenuation margin is increased to
12 dB and the maximum Ku-band transmitter power is increased to 34.3 dBW (2690 W).
Rain attenuation can exceed 12 dB at 14 GHz for a few minutes at a time in very heavy
thunderstorms, but there would only be a handful of such occurrences in an average year.
UPC definitely improves the ability of the uplink to resist rain attenuation, but at the ex-
pense of a considerably more powerful, and expensive, uplink transmitter. The extra ex-
pense can be justified in a television distribution system with many receiving stations.
There is also an increased risk that the additional power radiated by the uplink station
c04.qxd 28/08/02 19:37 Page 136
when UPC is active will cause interference at an unacceptable level into other satellite
links using the same frequencies. It would be advisable to increase the earth station an-
tenna diameter to increase its gain, and thus reduce the maximum transmit power required.
Downlink Attenuation and Sky Noise Increase The 11.45-GHz path be-
tween the satellite and the receive station suffers rain attenuation exceeding 5 dB for 0.01%
of the year. Assuming 100% coupling of sky noise into antenna noise, and 0.5-dB clear
air gaseous attenuation, calculate the overall CN under these conditions. Assume that the
uplink station is operating in clear air. We must calculate the available downlink fade
margin.
We need to find the sky noise temperature that results from a total excess path
attenuation of 5.5 dB (clear air attenuation plus rain attenuation); this is the new antenna
temperature in rain, because we assumed 100% coupling between sky noise temperature
and antenna temperature. We must evaluate the change in received power and increase in
system noise temperature in order to calculate the change in CN ratio for the downlink.
In clear air, the atmospheric attenuation on the downlink is 0.5 dB. The correspon-
ding sky noise temperature is 270(1 100.05) 29 K, which leads to the antenna tem-
perature of 30 K given in the Ku-band system specification. When the rain causes 5-dB
attenuation, the total path attenuation from the atmosphere and the rain is 5.5 dB. The
corresponding sky noise temperature is given by
Tsky rain T0 11 G2 where G 10A10 0.282
Tsky rain 270 11 0.2822 194 K
Thus the antenna temperature has increased from 30 K in clear air to 194 K in rain. The
system noise temperature in rain, Ts rain, is increased from the clear air value of 140 K
(30 K sky noise temperature plus 110 K LNA temperature)
Ts rain 194 110 304 K or 24.8 dBK
The increase in noise power is
¢N 10 log 1304 1402 3.4 dB
The signal is attenuated by 5 dB in the rain, so the total reduction in downlink CN ratio
is 8.4 dB, which yields a new value
1C N2 dn rain 17.2 8.4 8.8 dB
The overall CN is then found by combining the clear air uplink (CN)up of 30 dB with
the rain faded downlink (CN)dn rain of 8.8 dB, giving
1CN2 0 rain 8.8 dB
The overall (CN)0 is below the minimum acceptable value of 9.5 dB. The downlink link
margin is
Downlink fade margin 1CN2 dn 1C N2 min 17.2 9.5 7.7 dB
Since downlink rain attenuation of 5 dB causes the overall (CN)0 to go below the
minimum permitted value of 9.5 dB, we should recalculate the maximum attenuation that
the downlink can sustain. This involves an iterative process, since changing the attenua-
tion changes both C and N values in (CN)dn. At an attenuation level of 5 dB, the increase
in noise power is 3.4 dB, so a starting guess would be that decreasing the attenuation by
0.3 dB will decrease the noise power by 0.2 dB. The rain attenuation will then be a little
less than 5 dB.
c04.qxd 28/08/02 19:37 Page 137
Recalculating (CN)dn for a rain attenuation value of 4.7 dB gives

Tsky rain T0 11 G2 where G 10A10 0.339
Tsky rain 27011 0.3392 178 K
¢N 10 log 12881402 3.1 dB
1C N2 dn rain 17.2 4.7 3.1 9.4 dB
1CN2 0 rain 9.36 9.4 dB
The result is close enough to the required value of (CN)0 min 9.5 dB to conclude that
we can tolerate about 4.7 dB of rain attenuation on the downlink.
If better availability is required—less outage time—the diameter of the receiving
antenna can be increased. For example, if the receiving antenna diameter is increased to
2.4 m, (about 8 ft) the increase in antenna gain is 20 log10(2.402.14) 1.0 dB, which
increases the downlink margin to 8.7 dB. Repeating the iterative calculation outlined above,
the corresponding rain attenuation on the downlink is 5.5 dB with a noise power increase
of 3.2 dB. The downlink CN with 5.5-dB rain attenuation is 17.2 8.7 9.5 dB, and
the overall (CN)0 9.5 dB.
The extra antenna gain now ensures that the link meets the required specification,
which will keep outages to a total of about 50 min in an average year in the eastern United
States. However, an increase in antenna diameter will reduce the beamwidth of the an-
tenna and may require an upgrade in the tracking requirements. With a fixed pointing an-
tenna, diurnal motion of the satellite may cause a variation in received signal strength as
the satellite moves through the antenna beam.
Summary of Ku-Band Link Performance

The Ku-band link with a 2.4 m earth station antenna will suffer rain outages because at-
tenuation exceeding 5.5 dB will occur occasionally on the downlinks, affecting individual
customers. Outages will rarely occur on the uplink. With uplink power control (UPC) and
a more powerful transmitter, uplink outages can be restricted to a few minutes per year.
A 2.4-m receiving antenna is needed to ensure that the Ku-band downlink will be out for
no more than 0.01% of an average year with the given attenuation statistics. The threshold
value for overall CN was set at 9.5 dB because we can use QPSK and half rate error cor-
rection coding to obtain an equivalent (CN) ratio of about 15.5 dB without coding.
Allowing a 1 dB implementation margin (see Chapter 5), the BER on the downlink will
remain below 107 except when very heavy rain affects the downlink. In clear sky condi-
tions there will be no errors on the link. The 43.2 Msps QPSK signal with half rate FEC
can deliver a data rate of 43.2 Mbps, which can support seven MPEG-2 video channels.
The video distribution system described here is designed to deliver multiple video
channels to cable TV stations with low risk of outages. Direct broadcast satellite televi-
sion delivers video signals directly to the customer’s location using a much smaller 0.5-m
receiving antenna. The smaller antenna can be used because the DBS-TV satellites trans-
mit at a higher power level (160 W), the symbol rate is lower (20 Mbps) and availability
of the signal at the receiving antenna is guaranteed for only 99.7% of the year.
System Design Example 4.8.2

Personal Communication System Using
Low Earth Orbit Satellites
Low earth orbit (LEO) satellite systems are designed to provide personal communication
service similar to a cellular telephone, but over a much wider area. LEO satellite systems
c04.qxd 28/08/02 19:37 Page 138
SIDEBAR
Low earth orbit satellite communication systems next as one satellite goes below the horizon and
use a large number of satellites in multiple orbital another comes into view. The satellites are all iden-
planes at altitudes between 700 and 1400 km. The tical, so the link design is based on a single link be-
satellites can communicate with a limited portion tween one user and the gateway station via one
of the earth’s surface because of the low orbit alti- satellite. Examples of such LEO satellite systems
tude, and appear to an observer on the earth to fly are Iridium and Globalstar. An example of a
across the sky in a few minutes. Communication is medium earth orbit (MEO) satellite system is ICO
maintained between the user and a gateway station Global, which uses a smaller number of satellites at
by switching the channels from one satellite to the an altitude of about 10,000 km.
can cover sparsely populated regions of a country, or the world, where there are no ter-
restrial cellular telephone systems. The user has a handset similar to a cellular telephone
handset that provides two-way voice communications through a gateway station, usually
to a conventional telephone in a home or office connected to the public switched tele-
phone network (PTSN). Satellite telephones can equally well connect to another satellite
handset, or to a terrestrial cellular telephone.
Most LEO satellite systems operate in L band, in the 1500- and 1600-MHz bands,
and in the lower part of S band around 2460 MHz, frequency bands that are allocated for
mobile satellite communications. Some LEO systems use intersatellite links so a user can
connect to any point in the world without an intermediate return to earth. However, the
signals invariably pass through a gateway station at each end of the link to facilitate con-
trol of the call and to ensure that users can be charged for using the system. Connections
between the gateway earth stations and the satellites use S-band, C-band, Ku-band, or
Ka-band frequencies, depending on the system requirements. Only a small portion of the
radio spectrum at L band is allocated to LEO and MEO satellite systems, so L-band fre-
quencies are reserved for the critical links between the user and the satellite.
A handoff process is required for LEO satellites similar to that used in cellular tele-
phone networks, but the handoff between satellites should not be apparent to the user.
Most LEO satellites have multiple beam antennas, and the beam pattern moves across the
earth’s surface at the speed of the satellite—typically about 7.7 km/s or 17,200 mph. A
single beam is typically 500 km in diameter, so an individual user is in any one beam for
less than a minute. The system provides automatic switching from beam to beam within
the same satellite antenna coverage, much like a cellular telephone system switches users
from cell to cell, which nearly always requires a change in the link frequencies, but as
with satellite to satellite handoffs, the process must be transparent to the user.
The example below analyzes the links between a user and a gateway station. LEO
satellite systems employ digital transmission so that advantage can be taken of forward
error correction coding (FEC) and speech compression techniques. The bit rate of digital
voice in an LEO satellite link is typically 4800 bps, requiring powerful compression al-
gorithms. The low bit rate allows more signals to be sent in the available transponder
bandwidth and also helps maintain the CN ratio in the receivers. When FEC is applied
to a digital bit stream, carrier to noise ratios down to 5 dB can be used. The low bit rate
and operation of the receivers at low CN ratios are essential to make personal commu-
nication via an LEO satellite possible.
The link between the gateway station and the mobile terminal is defined as the out-
bound link, and the link from the mobile terminal to the gateway is the inbound link. Note
that there are four satellite paths, just as in all other two-way satellite communication systems:
c04.qxd 28/08/02 19:37 Page 139
LEO satellite
Ku-band Multiple
uplink and L-band
downlink uplink and
downlink
beams
Gateway
Users
Earth station
Satellite
To PSTN telephone
FIGURE 4.12 Two-way personal communication system using L-band LEO

satellite links.
outbound uplink, outbound downlink, inbound uplink, inbound downlink. Each has its own
unique frequency, and in most LEO satellite systems, one of the links will be weaker than
the other three links and will thus limit the system performance. One objective in the example
that follows is to identify the weakest path and to then attempt to improve that part of the
system. Figure 4.12 illustrates the two-way link between the gateway station and the handset.
Note that separate transponders are used for the inbound and outbound paths.
In this example, the mobile terminals transmit to a transponder on the satellite using
frequency division multiple access (FDMA) and single channel per carrier (SCPC) techniques.
FDMA and SPSC are discussed in Chapter 6. However, the principle is simple: each trans-
mitter is allocated its own frequency, just like broadcast stations. The available frequencies
are shared among active users on demand, as in a cellular telephone system, so a call begins
with a start-up sequence that establishes communication between the mobile terminal and the
local gateway station via the nearest LEO satellite. The gateway station then allocates fre-
quencies for the call. At the end of the call, the frequencies are released and become available
for another user. This is called demand assignment (DA), and the multiple access technique is
identified by the acronym SCPC-FDMA-DA, or, alternatively, SCPC-FDMA-DAMA where
DAMA stands for demand assignment multiple access. A common set of control channels at
preassigned frequencies enables call setup and teardown.
The link from the gateway station via the satellite to the mobile terminal uses time
division multiplexing (TDM). A TDM signal consists of a sequence of packets with ad-
dresses that repeat every 20 to 100 ms. The addresses identify which terminal should
receive each packet. The TDM bit stream rate must exceed the total bit rate of all active
terminals in a two-way telephone system so that there is sufficient capacity available for
each terminal within the TDM bit stream.
In this example we begin by assuming that 50 active users share one common TDM
channel. We will also assume that the gateway earth station operates at Ku band to and
from the satellite, and that the satellite employs a linear transponder (bent pipe) rather
than having onboard processing. The parameters of the satellite transponder, the mobile
c04.qxd 28/08/02 19:37 Page 140
TABLE 4.7 LEO Satellite Personal Communication System

Parameters
Satellite parameters
Saturated output power 10 W
Uplink frequency for mobile terminal 1650 MHz
Downlink frequency for mobile terminal 1550 MHz
Antenna gain 1650 MHz uplink (one beam) 23 dB
Antenna gain 1550 MHz downlink (one beam) 23 dB
Uplink frequency for gateway station 14 GHz
Downlink frequency for gateway station 11.5 GHz
Antenna gain 14 GHz uplink 3 dB
Antenna gain 11.5 GHz downlink 3 dB
Satellite receiver system noise temperature 500 K
Maximum range to edge of coverage zone 2200 km
Mobile terminal parameters
Transmitter output power 0.5 W
Antenna gain (transmit and receive) 0 dB
Receiver system noise temperature 300 K
Transmit bit rate 4800 bps
Receive bit rate 96 kbps
Required maximum bit error rate 104
Gateway station parameters
Transmitter output power (maximum per transponder) 10 W
Antenna gain (transmit, 14.0 GHz) 55 dB
Antenna gain (receive, 11.5 GHz) 53.5 dB
Receive system noise temperature (clear air) 140 K
Transmit bit rate (before FEC encoder) 300 kbps
Receive bit rate (after FEC decoder) 4800 bps
Required maximum bit error rate 104
terminal, and the gateway station are given in Table 4.7. The table gives the maximum
path length for any satellite–earth link. It is left as an exercise for the reader to determine
a suitable combination of orbital altitude and minimum elevation angle for the LEO system.
The user’s transmitter and receiver is called a mobile terminal in this example. It
could be a handheld device like a cellular telephone, sometimes called a satellite telephone
or handset, or the terminal could be mounted in a vehicle.
The satellite has multiple L band beams serving different parts of its instantaneous
coverage zone because a single beam from an LEO satellite serving different parts of its
instantaneous coverage zone would have both low gain and limited capacity. For an antenna
with a gain of 23 dB, G 200 and the beamwidth is 3 dB where
u3 dB 133,000 2002 12 12.8°
The use of a multiple beam antenna on the satellite increases the antenna gain to-
ward the mobile terminal, which increases the CN ratio of the signals in the mobile ter-
minal and gateway station receivers. The Ku-band antennas that link the satellite to the
gateway station have broad beams and low gain. The CN on these links is high through
the use of a relatively large antenna and a high transmitter power at the gateway earth
c04.qxd 28/08/02 19:37 Page 141
station, allowing the use of small and simple Ku-band antennas on the satellite. Figures
10.14, 10.15, and 10.16 in Chapter 10 illustrate satellite systems of this type.
The antenna gain at the mobile terminal is low, with a value of 0 dB used for calculation,
because the antenna coverage of the terminal must be very broad. If the terminal is a satellite
telephone, an omnidirectional antenna allows the user to move around freely. If the mobile
terminal antenna gain were to be increased, its beam would be correspondingly narrower, and
the user would have to point the handset at the satellite. In an LEO satellite system, the user
does not know which satellite is being used nor where it is in the sky, so requiring the user
to point the handset antenna at the satellite is not a feasible option. When the mobile terminal
is mounted in a vehicle with the antenna on the roof, pointing the antenna at the satellite is
not possible unless a sophisticated (and expensive) steered antenna is used.
In this example, we will begin by assuming that there are 50 users sharing a single
transponder on the satellite, and that one transponder serves one of the L-band beams
within the LEO satellite coverage, operating within a given set of frequencies. A large
number of users can share an LEO satellite through the provision of many transponders,
each of which is connected to one of the individual beams in the multiple L-band antenna
coverage of the satellite. The signal received by a mobile terminal from the gateway is a
TDM sequence of packets carrying 50 digital voice channels, each at 4800 bps. The bit
rate of the TDM signal would be 240 kbps if it carried only the voice signals, but will be
higher in practice because additional bits must be sent with each packet; a TDM bit rate
of 300 kbps is used in this example. Individual mobile terminals pull off their assigned
packets from within the TDM stream and ignore the rest. Initially, the links will be analyzed
without forward error correction.
All digital links are designed with ideal (Nyquist) filters which have noise bandwidth,
Bn Hz, equal to the symbol rate of the digital signal in symbols per second. In this example,
handheld transceivers send and receive binary phase shift keyed (BPSK) modulation. The
maximum permitted bit error rate of the digital signal of 104 leads to a SN ratio in the
speech channel of 34 dB. (SN 14Pe, where Pe is the BER—see Chapter 5.)
Inbound Link: Mobile Terminal to

Gateway Station
Each terminal transmits a BPSK signal at 4800 bps at an allocated frequency. The satel-
lite transponder shifts all received L-band signals in frequency before retransmission at
Ku band to the gateway station, and also amplifies the signals with a linear transponder.
At the gateway station, the antenna and RF receiver are connected to many identical IF
receivers tuned to the individual frequencies of the handheld transmitters. Each IF receiver
has a noise bandwidth of 4800 Hz, set by a square root raised cosine filter with 0.5,
giving an occupied channel bandwidth of 7.2 kHz (see Chapter 5 for details of the design
of digital links).
At the receiving end of the link, the CN at the input to the BPSK demodulator
must be high enough to provide an acceptable bit error rate. Here, we require a maximum
BER of 104 which provides a minimum SN of 34 dB in the speech channel. In Chap-
ter 5 it is shown that the theoretical CN required to achieve a bit error rate of 104 with
BPSK modulation is 8.4 dB. In a practical digital communication system, we always need
a higher CN than theory suggests because we do not have ideal Nyquist filters, and other
parts of the system are also not ideal, so we must add an implementation margin to account
for the nonideal nature of the system. In this example, the implementation margin is set
at 0.6 dB; we need a minimum CN 9.0 dB to meet the BER and SN specifications.
We can now design the satellite link to achieve the minimum CN.
c04.qxd 28/08/02 19:37 Page 142
Mobile Terminal to Satellite Link

We will establish power and noise link budgets for each of the four paths, beginning with
the uplink from the mobile terminal to the satellite.
The received power at the output of the uplink antenna on the satellite from
Eq. (4.11) is
Pr EIRP Gr Lp Lm dBW
where EIRP is the product of transmitter output power and transmitting antenna gain, PtGt
in dBW, Gr is the satellite receive antenna gain, Lp is the path loss of the link, and Lm
accounts for all other losses. The noise power, N, at the input to the satellite receiving
system from Eq. (4.13) is
N Pn kTsBn watts
k Ts Bn dBW
Path loss Lp is found from Eq. (4.12)
Lp 34pRl4 2 or 20 log10 14pR/l2 dB
where R is the distance in meters between the transmitting and receiving antennas in the
link and is the wavelength in meters.
The uplink frequency is 1650 MHz, giving 0.1818 m. The maximum range is
2200 km so maximum path loss is
Lp 20 log10 14p 2.2 106 0.18182 163.6 dB
We will assume that there are miscellaneous losses in the 1550 MHz link of 0.5 dB, caused
by polarization misalignments, gaseous absorption in the atmosphere, etc. The calculation of
the CN ratio is made for the worst case of an earth station located on the 3 dB contour
of the satellite antenna beam, so a 3 dB reduction in satellite antenna gain is applied, making
the value of Lm 3.5 dB. We can now set out the link power and noise budgets for clear
line of sight conditions, when there is no attenuation caused by obstructions in the path.
Uplink Power Budget

Parameter Symbol Value Units
EIRP of handheld unit PtGt 3 dBW
Gain of receiving antenna Gr 23 dB
Path loss at 1650 MHz Lp 163.6 dB
Miscellaneous losses Lm 3.5 dB
Received power at satellite Pr 147.1 dBW
Transponder Noise Power Budget

Boltzmann’s constant k 228.6 dBW/K/Hz
System noise temperature Ts 27.0 dBK
Noise bandwidth Bn 36.8 dBHz
Noise power N 164.8 dBW
The inbound uplink CN ratio in the transponder can now be calculated from the
power and noise budgets:
1C N2 up Pr N 147.1 dBW 1164.8 dBW2 17.7 dB
c04.qxd 28/08/02 19:37 Page 143
Note that this is the lowest CN ratio that should occur in the transponder in clear
air conditions, since the calculation was made for a mobile terminal at the longest range
from the satellite and at the edge of a satellite antenna beam. The mobile terminal an-
tenna gain has also been set to its minimum value of 0 dB. If the satellite were directly
overhead the range would be 1000 km instead of 2200 km, making the path loss lower
by 6.8 dB, and the miscellaneous losses would be 3 dB lower at the center of the satel-
lite antenna beam, making the power received at the transponder 10.8 dB greater, and
then (CN)up 28.5 dB. However, we cannot use this figure for the system design, oth-
erwise there would be only one user who could make calls, and then only for a brief mo-
ment as the satellite passes directly overhead. We must ensure that all users within the
satellite’s coverage zone have adequate CN ratios in their links for successful
communication.
Satellite to Gateway Station Link

The next step in calculating the CN ratio for the inbound link is to calculate (CN)dn
in the gateway receiver. We are operating the transponder in FDMA, so the individual
mobile terminal signals must share the output power of the transponder. We will assume
that 50 active terminal signals share the 1 MHz transponder bandwidth and that 3 dB
backoff is used at the transponder output to obtain quasi-linear operation of the transpon-
der HPA (remembering that we have assumed linear transponder operation in this ex-
ample). The transponder output power is therefore 10 dBW 3 dB 7 dBW (5 W).
The 5-W transponder output power must be shared equally between the 50 signals in
the transponder, giving 0.1 W 10 dBW per signal at the transponder output for the
downlink to the gateway station.
We can now establish a link budget for a single channel downlink from the satel-
lite to the gateway station. We will use the same worst-case conditions as for the uplink—
maximum path length and minimum satellite antenna gain, with miscellaneous losses of
3.5 dB, including the edge of satellite beam effect.
Downlink Power Budget

EIRP per channel PtGt 10.0 dBW
Gain of receiving antenna Gr 53.5 dB
Path loss at 11.5 GHz Lp 180.5 dB
Gateway Station Noise Power Budget

The CN ratio in the 4.8-kHz noise bandwidth of a gateway station IF receiver is
given by:
1CN2 dn Pr N 140.5 1170.32 29.8 dB
c04.qxd 28/08/02 19:37 Page 144
(CN)dn for the inbound downlink is higher than (CN)up for the inbound uplink because
of the high gain of the gateway station antenna. Because the gain of the antenna is high,
53.5 dB, which corresponds to an antenna diameter of 5 m and an aperture efficiency of
60%, its beamwidth is narrow, about 0.4, and the gateway station must track the satellite
as it crosses the sky.
The overall (CN)0 at the gateway is calculated by combining the uplink CN and
downlink CN values using Eq. (4.43), since both the transponder and the gateway sta-
tion receiver add noise to the signal. The values used in the formula are ratios, that is,
CN values are not in decibels.
1 1CN2 0 1 1C N2 up 1 1CN2 dn
For the inbound uplink, (CN)up 17.7 dB 1 58.9 as a ratio. For the inbound
downlink, (CN)dn 29.8 dB 1 955.0 as a ratio.
1CN2 0 1 11 58.9 1955.02 55.5 or 17.4 dB
The overall CN ratio of 17.4 dB at the gateway station receiver guarantees that
with BPSK and a bit rate of 4800 bps there will be extremely few bit errors and the
SN of the speech channel will be set by quantization noise in the analog to digital
converters. The maximum permitted BER is 104, which occurs with (CN)0 9.0
dB. We therefore have an inbound link margin of (17.4 9.0) 8.4 dB. However, we
must calculate the individual link margins for the uplink and downlink in order to
be able to use the margins for fading analysis. This will be done at the end of the
example.
Outbound Link
The outbound link from the gateway station to the mobile terminal sends a continuous
300 kbps TDM bit stream using BPSK modulation and a separate transponder with 1 MHz
bandwidth. The bit stream is a series of packets addressed to all 50 active terminals. The
noise bandwidth of the terminal receiver is 300 kHz, assuming ideal Nyquist filters. The
outbound uplink and downlink CN values are calculated in exactly the same way as for
the inbound link, and the power and noise budgets are combined to give CN ratios directly
from a single table.
At the uplink frequency of 14 GHz, clear air atmospheric attenuation of 1.0 dB is
included in the miscellaneous losses, together with the usual 3 dB loss for the user at the
edge of the satellite antenna beam.
Uplink CN Budget
Gateway station EIRP PtGt 65.0 dBW
Gain of receiving antenna Gr 3.0 dB
Path loss at 14.0 GHz Lp 182.2 dB
Uplink CN (CN)up 118.2 (146.8) 28.6 dB
c04.qxd 28/08/02 19:37 Page 145
Downlink C/N Budget

The satellite transponder carrying the single 300 kbps TDM outbound signal can be op-
erated close to saturation because there is only one signal in the transponder, thus elimi-
nating intermodulation problems. We will allow 1.0 dB backoff at the transponder output
to avoid saturating the transponder, giving a transmitted power Pt 9.0 dBW. Miscella-
neous losses on the downlink are 0.5 dB atmospheric loss and 3 dB for the edge of the
antenna beam.

EIRP of satellite PtGt 32.0 dBW
Gain of receiving antenna Gr 0 dB
Path loss at 1550 MHz Lp 163.1 dB
Received power at mobile Pr 134.6 dBW
Downlink CN (CN)dn 134.6 (149.0) 14.4 dB
Combining the CN values for the uplink and downlink gives the overall (CN)0
ratio at the mobile terminal receiver. Converting the CN values from decibels gives
(CN)up 28.6 dB 724.4, (CN)dn 14.4 dB 27.5 Hence, the overall (CN)0 for
the outbound link is
1C N2 0 1 31 1C N2 up 1 1CN2 dn 4 1 30.00139 0.03644 26.5 or 14.2 dB
Note that the downlink CN ratio is so much lower than the uplink CN ratio that the
overall CN ratio is almost equal to the downlink CN ratio.
The clear air (CN)0 value is 5.2 dB above the minimum allowed for BER 104
on the outbound link, leaving a 5.2 dB margin for blockage by buildings, the user’s head,
multipath effects, the ionosphere, or vegetative shadowing on the downlink. The link mar-
gins for the outbound link are much lower than for the inbound link, and it is therefore
the weakest part of the system. Attenuation exceeding 5.2 dB in the downlink from the
satellite to the mobile terminal will cause the BER to exceed 104 and the SN in the
speech channel will fall below 30 dB. A SN ratio of 30 dB in a speech channel is re-
garded as the minimum acceptable value for intelligible communication. Because of the
very steep characteristics of the BER vs CN ratio curve for BPSK, the speech channel
will be unusable if downlink attenuation exceeds 5.2 dB.
The link margins are quite small for a mobile system in which the line of sight be-
tween the satellite and the user can easily be blocked by trees, or by the user’s body. It
is the link between the mobile terminal and the satellite that sets the overall CN value
for both the inbound and the outbound links, but there is little room to change the sys-
tem parameters to yield higher margins. When the mobile terminal is a satellite telephone
handset, the transmitter power is limited by FCC regulations to ensure that there is no
short-term biological hazard to the user when the handset is transmitting. (See Chapter
10 for further details on radiation limits for portable equipment.) The power from the
satellite is limited by the transponder HPA output power and the low gain of the hand-
set antenna. However, a higher gain antenna would have a narrower beam and would
have to track the satellite automatically—a smart antenna could be built to do this, but
c04.qxd 28/08/02 19:37 Page 146
the small size of most mobile telephone handsets limits the available improvement to no
more than 3 or 4 dB.
In the next section, we will see that the performance of the link can be improved
by using forward error correction coding to lower the CN value at which the BER is
104, thus increasing the margins available on the links.
Optimizing System Performance

The preceding calculations show that the LEO satellite system can support two-way digital
speech with 50 active users per transponder, and provides a link margin of 8.4 dB in the in-
bound link and a margin of 5.3 dB in the outbound link. The RF bandwidth used by the in-
bound and outbound links is found from the symbol rates and the values of the filters (see
Chapter 5). For the outbound link using 0.5, the symbol rate is 300 kbaud, giving
Boutbound 300 11 a2 450 kHz
For the inbound link using 0.5, the symbol rate for one speech channel is 4800 baud,
giving
Binbound 4.8 11 a2 7.2 kHz
The inbound channels access the satellite transponder using SCPC-FDMA, so the
RF signals are distributed across the transponder bandwidth. We must space the channels
more than 7.2 kHz apart in the transponder so that the narrow band-pass filters in the gate-
way station receiver can extract each speech channel without interference from the adja-
cent channels. If we use a 10 kHz channel spacing, there will be a frequency gap, called
a guard band of 2.8 kHz between each channel, which will ensure minimal interference
from adjacent channels. With 50 channels sharing one transponder, the total bandwidth
occupied in the inbound link transponder will be 500 kHz.
Neither the inbound nor the outbound transponder bandwidth is fully utilized; in fact
only half of the available 1 MHz is used in each case. However, we cannot add additional
speech channels to the system because the CN values are already low, indicating that the
system is power limited with the given link margins. We can incorporate FEC coding, how-
ever, which lowers the CN threshold for the minimum BER. Half rate convolutional cod-
ing would be a good choice in this system because the threshold CN value can be much
lower. This allows a wider noise bandwidth to be used and thus better utilization of the
available transponder bandwidth. Using constraint length eight and soft decision decoding,
the CN ratio for BER 104 can be lowered to 3.5 dB; alternatively Turbo coding could
be used. (See Chapter 7 for details of FEC techniques.) However, the bit rate of the signal
is now doubled, since a half rate code adds as many coding bits as there are data bits in the
bit stream. The new outbound bit rate with FEC is 600 kbps, and each inbound speech chan-
nel has a bit rate of 9600 bps. The corresponding RF bandwidths for 0.5 Nyquist fil-
ters are 900 kHz outbound and 14.4 kHz per channel inbound. With 50 active users, the RF
signal bandwidths are within the available 1 MHz bandwidth of the satellite transponders.
Lowering the threshold value of CN for the maximum permitted BER of 104 im-
proves the link margins by a factor called coding gain. Coding gain is typically quoted as
the difference between the CN value required for a given BER without coding and the
CN required to obtain the same BER with coding. In this example, the coding gain is
8.4 3.5 4.9 dB. However, the coding gain cannot simply be added to the system mar-
gin, because the half rate FEC code doubles the bit rate of the signals and also doubles
the noise bandwidth of the filters in the receivers. Thus noise power increases by 3 dB in
every link receiver when FEC is added, and the CN values all fall by 3 dB. This results
c04.qxd 28/08/02 19:37 Page 147
in the overall values of (CN)0 for both the inbound and the outbound links falling by
3 dB. There is no need to recalculate all the link noise budgets and CN values, since all
of the values change by the same amount. (This is one advantage of using decibels for
link calculations.) With half rate FEC added to the system, the new CN values are all
3 dB lower than for the system without FEC.
Inbound Link
(CN)up 14.7 dB
(CN)dn 26.8 dB
(CN)0 14.4 dB
Outbound Link
(CN)up 25.6 dB
(CN)dn 11.4 dB
(CN)0 11.2 dB
The new link margins with a threshold overall (CN)0 of 3.5 dB are: inbound 10.9 dB,
outbound 7.7 dB. Although the improvement over the earlier values without FEC is only
2 dB, the increased link margins are valuable, so FEC is invariably used in satellite per-
sonal communication systems, as it is in almost all digital wireless applications, whether
satellite or terrestrial. FEC can be implemented by inserting a coding IC in the terminal,
and identical ICs in the gateway station, in the baseband bit streams. The 2 dB advantage
that FEC brings to the system’s link margins cannot easily be obtained any other way.
Link Margins with FEC

Rain attenuation affects the Ku-band links between the gateway station and the satellite,
and blockage affects the link between the mobile terminal and the satellite. Individual link
margins must be calculated to determine the amount of fading or blockage that can be tol-
erated in each link. Excessive rain attenuation in the Ku-band links could cause the links
to fail, which affects all 50 users. We must therefore ensure that the Ku-band link mar-
gins are sufficiently large to make a rain outage unlikely. Blockage of the line of sight to
a mobile terminal may cause that one terminal to lose its link, but this is less serious than
losing all 50 links simultaneously. The margin available for overcoming blockage should
be as large as possible, but is set by the system design and cannot be improved beyond
the values given above without a reduction in the number of users in the system.
It should be remembered that all the calculations are for a worst case: the user is at
the edge of the satellite coverage zone where the satellite is at maximum range, and also
at the edge of one of the satellite’s multiple L-band beams. Most of the users have higher
CN ratios on their links than the calculated worst case values most of the time, allow-
ing greater margins for blockage of the path to the handset. In commercial satellite sys-
tem design, the fact that most users are not at the edge of the coverage zone most of the
time is used in developing a “coverage advantage” factor that increases the average link
margin available per user and thus optimizes traffic capacity.
Rain Attenuation at Ku Band

Rain causes attenuation at Ku band, as discussed in Chapter 8. We must calculate the
rain attenuation margins for the inbound downlink and the outbound uplink and deter-
mine the probability of an outage. The link margin is the number of decibels by which
c04.qxd 28/08/02 19:37 Page 148
the CN ratio on an uplink or a downlink can be reduced before the overall (CN)0 for
that link falls to the threshold value. We will use 3.5 dB as the threshold value for over-
all CN in each case, assuming that half rate FEC is used. We will also assume that clear
sky conditions prevail on the uplink when extreme attenuation occurs on the downlink,
and vice versa.
For the inbound Ku-band downlink, using half rate FEC, the clear air CN ratio is
26.8 dB (ratio 478.6) and the L-band clear uplink CN ratio is 14.7 dB (ratio 29.5). With
a threshold at 3.5 dB (ratio 2.24), the minimum downlink CN will be given by (using
ratios, not dB)
1 1C N2 dn min 1 1CN2 0 1 1CN2 up
12.24 1 478.6 0.444
Hence the minimum permitted value for (CN)dn 2.25 1 3.5 dB. The downlink mar-
gin is 26.8 3.5 23.3 dB. Rain attenuation at 11.5 GHz very rarely exceeds this value
in the United States, so for a U.S. system, the Ku-band downlink margin is adequate.
The Ku-band uplink has a clear sky (CN)up ratio of 25.6 dB, but attenuation of the
uplink signal causes a reduction in received power at the transponder input. If the satel-
lite transponders are linear (bent pipe), the output power will fall when the input power
is reduced by uplink rain attenuation. Because the transponder is operated close to satu-
ration, there will not be a one-to-one correspondence in the changes in power level at the
input and output, but exact analysis is beyond the scope of this example; a linear rela-
tionship will be assumed here. The transponder nonlinearity actually increases the uplink
rain attenuation margin, because the output signal from the satellite will fall less than the
input signal to the satellite, so the results that follow represent a pessimistic estimate of
the margin available. A regenerative repeater always transmits at constant output power
and is very desirable in a digital system. It avoids the difficulty of attenuation on the up-
link causing a reduction in transponder output power.
Applying the same analysis as used for the Ku-band downlink, with (CN)dn 11.4
dB (ratio 13.8) in clear sky conditions and (CN)0 min 3.5 dB (ratio 2.24)
1 1C N2 up min 1/1CN2 0 1 1CN2 dn
12.24 1 13.8 0.374
Thus the minimum (CN)up ratio is 10 log(10.374) 4.3 dB, ignoring the effects of
coupling between input and output power in the transponder. When the latter effect is con-
sidered with a linear transponder characteristic, the limit is set by the (CN)dn ratio falling
to 3.5 dB. This will occur with 11.4 3.5 7.9 dB uplink attenuation, which is the
limiting value.
Uplink power control (UPC) can be used to prevent the input power level of the
transponder from falling when rain affects the uplink. It would be straightforward to use
uplink power control in this case. Attenuation on the downlink at 11.5 GHz is measured
using the satellite beacon, scaled to 14.4 GHz, and used to set the gateway station trans-
mit power level. If a 10 dB dynamic range of UPC is available, and attenuation is allowed
to reach 2 dB at 14.4 GHz before the UPC comes in, the downlink CN ratio will not fall
below 9.4 dB in an uplink fade, leaving a downlink margin of 5.9 dB during uplink rain
fades. This gives the 14.4 GHz uplink a rain attenuation margin of 12 dB, which would
maintain the link for better than 99.99% of a year throughout the United States. The open-
loop UPC system discussed here would probably have a margin of error of at least 1 dB
in estimating uplink attenuation under low fading conditions. Uncertainties in identifying
the propagation mechanism that is causing the fading and the difficulty of accurately setting
the clear sky baseline for the signal make greater accuracy unlikely.
c04.qxd 28/08/02 19:37 Page 149
If two mobile terminals are located within the same satellite beam coverage, and
are therefore operating through the same gateway earth station, the assumption of nonsi-
multaneous outage of the two links would not be valid. Such situations are expected to
be rare occurrences.
The gateway station would typically be sited in a dry region, such as Wyoming or
Idaho in the United States, to minimize the number and severity of rain attenuation events.
Thus rain attenuation at Ku band can be overcome by a large link margin for the down-
link and implementation of uplink power control in the uplink, and by intelligent siting
of the station. All 50 channels can be guaranteed to be unaffected by rain at the gateway
station.
Path Blockage at L-Band

Trees, buildings, and people are the most likely causes of blockage that affect the per-
formance of the mobile terminal at L band. Blockage by buildings is too severe to allow
the L-band link to operate, and most LEO satellite telephones will not work indoors. Some
systems like Iridium incorporate a cellular telephone into the handset. The cellular tele-
phone is used in preference to the satellite phone to reduce loading on the LEO satellite
system, and also whenever the satellite signal is unavailable, such as indoors. Paging op-
tions have been designed into some mobile satellite systems which permit users to be
alerted that there is an incoming call. The user still has to run outdoors to be able to re-
ceive the call, and this has evidently been a factor deterring the use of satellite telephone
by business people.
The link margins for the L-band links are calculated in the same way as the Ku-
band margins. Repeating the calculations with a minimum overall CN ratio of 3.5 dB
and no rain attenuation in the Ku-band links gives
L-band uplink margin 11.4 dB
L-band downlink margin 7.8 dB
The downlink from the satellite to the mobile terminal is therefore the most vulnerable
of the links, and cannot be made robust without reducing the number of users per
transponder. However, the value of 7.8 dB for the downlink margin is a worst case value
and most of the users will have a margin several decibels higher. A margin of 7.8 dB
can be exceeded by attenuation through a stand of trees. For example, if the user is in
a vehicle traveling along a road cut through a forest, and the satellite has a low eleva-
tion angle, the 7.8 dB attenuation margin may be exceeded from time to time, causing
repeated break up of the downlink signal. Transmission protocols and signal buffering
can be designed to reduce the impact of this type of intermittent loss of signal. Multi-
path effects when the satellite is at a low elevation angle can also cause variations in
signal level leading to lower performance and occasional outages in extreme cases, such
as an overwater path.
Summary of L-Band Mobile PCS System

Performance
The personal communication system in this example uses a network of low earth orbit
satellites to link a user anywhere in the system’s coverage zone to a gateway station, and
then to the public switched telephone network or another mobile terminal. The user’s ter-
minal operates in L band and is similar to a cellular telephone, with a low gain, omnidi-
rectional antenna. The transmissions are digital, with BPSK modulation, and use speech
c04.qxd 28/08/02 19:37 Page 150
compression to achieve a bit rate of 4.8 kbps per speech channel. There are a maximum
of 50 users in each of the satellite’s 50 L-band beams, giving a nominal satellite capacity
of 2500 users.
The inbound link from the user to the gateway station has a margin of 11.4 dB for
tree shadowing on the uplink to the satellite. The downlink from the satellite has a block-
age margin of 7.8 dB. The Ku-band links between the satellite and the gateway station
have large margins, and uplink power control is used to prevent uplink rain attenuation at
14 GHz from adversely affecting the downlinks to the mobile terminals. TDM is used on
the outbound link with half rate FEC coding and a bit rate of 600 kbps. SCPC-FDMA-
DAMA is used on the inbound links, with a channel bit rate of 9.6 kbps after FEC coding
is applied.
4.9 SUMMARY
This chapter has set out the procedures for calcula- used to find the overall (CN)0 ratio for different
tion of received power from a satellite and noise systems.
power in a receiver. Together, these figures give the Fading of both uplink and downlink simultane-
CN ratio for the receiving system. The specification ously is unlikely for 64 and 1411 GHz systems and
of a system will always require a minimum CN in can safely be ignored when computing link statistics.
the receiver, below which the link is considered in- At 3020 GHz the possibility cannot be ignored and
operable. The design of a link to achieve that mini- the joint effect has to be calculated. No attempt has
mum CN requires repeated application of the link been made to derive an optimization procedure for
and noise power equations to give CN for clear air the design of the “best” system within a frequency
conditions with acceptable bandwidth and antenna di- band and CN ratio specification. There are too many
mensions. When a linear (bent pipe) transponder is variables in the system, including the cost of anten-
used, the clear air value of CN ratio for the uplink nas, receivers, and other components, to produce a
and downlink must be combined to give the overall single optimization procedure. Iterative techniques
(CN)0 ratio in the earth station receiver. Once clear must be used to find a set of parameters for the earth
air performance has been calculated, the effect of rain stations and satellite that provide the performance re-
on the slant paths must be determined and the prop- quired from the satellite communication system. The
agation path statistics need to be studied to determine designer of a satellite communication system may
how much margin is required to meet worst-case con- have to go through several trial design procedures and
ditions. Examples are presented throughout this chap- compare the resulting systems to determine which one
ter showing how link power and noise budgets are best suits the particular application.
REFERENCES
1. S. SILVER, ed., Microwave Antenna Theory and Design, 6. H. TAUB and D. L. SCHILLING, Principles of Communi-
Vol. 12, MIT Radiation Lab Series, 1947. (Republished cation Systems, McGraw-Hill, New York, 1971.
by Peter Perigrinus, Stevenage, Herts, UK, 1984.) 7. M. L. GUSTAFSON, “The KLM Sky Eye 10 Receiver,”
2. W. L. STUTZMAN and G. A. THIELE, Antenna Theory and Satellite Television, 2, 66–70, November 1984.
Design, John Wiley & Sons, New York, 1981. 8. J. DICKS and M. BROWN, JR., “INTELSAT IV-A Trans-
3. J. D. KRAUS, Radio Astronomy, Cygnus-Quasar Books, mission System Design,” Comsat Technical Review. 5,
Powell, OH, 1982. (Originally published by McGraw- 73–103, 1975.
Hill, New York, 1966.) 9. I. A. GLOVER and P. M. GRANT, Digital Communica-
4. H. L. KRAUSS, C. W. BOSTIAN, and F. H. RAAB, Solid tions, Prentice-Hall, Europe, 1998.
State Radio Engineering, John Wiley & Sons, New 10. WALTER L. MORGAN and GARY D. GORDON, Com-
York, 1980. munications Satellite Handbook, John Wiley & Sons,
5. K. S. SHANMUGAM, Digital and Analog Communica- New York, 1989.
tion Systems, John Wiley & Sons, New York, 1979, 11. G. MARAL and M. BOUSQUET, Satellite Communication
pp. 356–360. Systems, John Wiley & Sons, New York, 1986.
c04.qxd 28/08/02 19:37 Page 151
PROBLEMS 151
PROBLEMS
1. A C-band earth station has an antenna with a trans- of 50 kbps and requires a receiver with a noise
mit gain of 54 dB. The transmitter output power is bandwidth of 50 kHz per channel. The earth sta-
set to 100 W at a frequency of 6.100 GHz. The sig- tions used to receive the voice signals have anten-
nal is received by a satellite at a distance of 37,500 nas with a gain of 40 dB (1 m diameter) and a re-
km by an antenna with a gain of 26 dB. The signal ceiver with Tsystem 150 K in clear air, and IF noise
is then routed to a transponder with a noise temper- bandwidth of 50 kHz.
ature of 500 K, a bandwidth of 36 MHz, and a gain a. Calculate the power transmitted by the satellite in
of 110 dB. one voice channel.
a. Calculate the path loss at 6.1 GHz. b. Calculate the CN in clear air for an earth station
b. Calculate the power at the output port (sometimes receiving one BPSK voice signal.
called the output waveguide flange) of the satellite c. What is the margin over a coded BPSK threshold
antenna, in dBW. of 6 dB?
c. Calculate the noise power at the transponder in-
4. Geostationary satellites use L, C, Ku, and Ka
put, in dBW, in a bandwidth of 36 MHz.
bands. The path length from an earth station to the
d. Calculate the CN ratio, in dB, in the transponder. GEO satellite is 38,500 km. For this range, calculate
e. Calculate the carrier power, in dBW and in W, at the path loss in decibels for the following frequen-
the transponder output. cies:
2. The satellite in Problem 1 above serves the 48 con- a. 1.6 GHz, 1.5 GHz b. 6.2 GHz, 4.0 GHz
tiguous states of the United States. The antenna on c. 14.2 GHz, 12.0 GHz d. 30.0 GHz, 20.0 GHz
the satellite transmits at a frequency of 3875 GHz to 5. Low earth orbit satellites use mainly L band, with
an earth station at a distance of 39,000 km. The an- ranges varying from 1000 to 2500 km. Calculate the
tenna has a 6 E–W beamwidth and a 3 N–S maximum and minimum path loss from earth to a
beamwidth. The receiving earth station has an antenna satellite, in dB, for the uplink frequency of 1.6 GHz,
with a gain of 53 dB and a system noise temperature and the downlink frequency of 1.5 GHz.
of 100 K and is located at the edge of the coverage
zone of the satellite antenna. (Assume antenna gain 6. A geostationary satellite carries a transponder with
is 3 dB lower than in the center of the beam.) a 20 W transmitter at 4 GHz. The transmitter is op-
Ignore your result for transponder output power in erated at an output power of 10 W and drives an an-
Problem 1 above. Assume the transponder carrier tenna with a gain of 30 dB. An earth station is at the
power is 10 W at the input port of the transmit an- center of the coverage zone of the satellite, at a range
tenna on the satellite. of 38,500 km. Using decibels for all calculations,
find:
a. Calculate the gain of the satellite antenna in the
direction of the receiving earth station. [Use the a. The flux density at the earth station in dBWm2.
approximate formula G 33,000(product of b. The power received by an antenna with a gain of
beamwidths).] 39 dB, in dBW.
b. Calculate the carrier power received by the earth c. The EIRP of the transponder in dBW.
station, in dBW. 7. An LEO satellite has a multibeam antenna with a
c. Calculate the noise power of the earth station in gain of 18 dB in each beam. A transponder with trans-
36 MHz bandwidth. mitter output power of 0.5 W at 2.5 GHz is connected
d. Hence find the CN in dB for the earth station. to one antenna beam. An earth station is located at
the edge of the coverage zone of this beam, where the
3. A 1411 GHz satellite communication link has a
received power is 3 dB below that at the center of the
transponder with a bandwidth of 52 MHz which is
beam, and at a range of 2000 km from the satellite.
operated at an output power level of 20 W. The satel-
Using decibels for all calculations, find:
lite transmit antenna gain at 11 GHz is 30 dB toward
a particular earth station. Path loss to this station is a. The power received by an antenna with a gain of
206 dB, including clear air atmospheric loss. 1 dB, in dBW.
The transponder is used in FDMA mode to send b. The noise power of the earth station receiver for a
500 BPSK voice channels with half rate FEC noise temperature of 260 K and an RF channel band-
coding. Each coded BPSK signal has a symbol rate width of 20 kHz.
c04.qxd 28/08/02 19:37 Page 152
c. The CN ratio in dB for the LEO signal at the Antenna gain, on axis, C band and
receiver output. Ku band (transmit and receive) 31 dB
8. A satellite in GEO orbit is at a distance of 39,000 Receive system noise temperature
km from an earth station. The required flux density at (C band and Ku band) 500 K
the satellite to saturate one transponder at a frequency Transponder saturated output power:
of 14.3 GHz is 90.0 dBW/m2. The earth station has a C band 40 W
transmitting antenna with a gain of 52 dB at 14.3 GHz. Transponder bandwidth: C band 36 MHz
Find:
Transponder saturated output
a. The EIRP of the earth station.
power: Ku band 80 W
b. The output power of the earth station transmitter.
9. A 12-GHz earth station receiving system has an
antenna with a noise temperature of 50 K, a LNA with SIGNALS
a noise temperature of 100 K and a gain of 40 dB, FM-TV analog signal to be received in a bandwidth
and a mixer with a noise temperature of 1000 K. Find of 27 MHz.
the system noise temperature. Multiplexed digital TV signals transmitted as
10. A geostationary satellite carries a C-band transpon- QPSK with symbol rate 27 Msps using half rate FEC
der which transmits 20 W into an antenna with an on- with coding gain 5.5 dB
axis gain of 30 dB. An earth station is in the center Minimum permitted CN overall 9.5 dB
of the antenna beam from the satellite, at a distance
12.1 Design a transmitting earth station to provide a
of 38,000 km. For a frequency of 4.0 GHz:
clear air CN of 26 dB in a C-band transponder at a
a. Calculate the incident flux density at the earth frequency of 6.285 GHz. Use an uplink antenna with a
station in watts per square meter and in dBW/m2. diameter of 9 m and an aperture efficiency of 68%, and
b. The earth station has an antenna with a circular find the uplink transmitter power to achieve the required
aperture 2 m in diameter and an aperture efficiency CN. The uplink station is located on the 2 dB contour
of 65%. Calculate the received power level in W and of the satellite footprint. Allow 0.5 dB for clear air
in dBW at the antenna output port. atmospheric attenuation and other losses.
c. Calculate the on-axis gain of the antenna in dB. 12.2 Design a C-band receiving earth station to pro-
d. Calculate the free space path loss between the vide an overall clear air CN of 13 dB in a 27 MHz IF
satellite and the earth station. noise bandwidth at a carrier frequency of 4.06 GHz.
Calculate the power received, Pr, at the earth sta- The antenna noise temperature is 20 K and the LNA
tion using the link equation: noise temperature is 55 K. You may assume a high
gain LNA and ignore the noise generated in other
Pr PtGtGrLp
parts of the receiver. The C-band satellite transpon-
where PtGt is the EIRP of the satellite transponder der is operated with 1 dB output backoff. Clear air
and Lp is the path loss. atmospheric attenuation on the downlink and other
Make your calculation in dB units and give your losses total 0.5 dB. Determine the diameter of the re-
answer in dBW. ceiving antenna, assuming an aperture efficiency of
11. Repeat parts (a) through (d) of Problem 10 for a 65%. The receiving terminal is located on the 3 dB
Ka-band transponder transmitting at a frequency of contour of the satellite footprint.
20.0 GHz. Reminder: Overall CN includes the effect of noise
radiated by the satellite transponder.
12. This sequence of questions requires you to de-
sign a communication link through a geostationary 12.3 a. Under conditions of heavy rain, the C-band
satellite to meet a CN and link margin specification. path from the transmitting station suffers an attenua-
tion of 2 dB. Calculate the overall CN at the earth
Use these constants:
station in a bandwidth of 27 MHz under these con-
Boltzmann’s constant k 228.6 dBW/K/Hz ditions, and find the uplink link margin.
Path length to satellite 38,500 km Reminder: The uplink margin is the number of dB of
attenuation that can occur on the uplink before the re-
SATELLITE ceiver overall CN reaches the limit of 9.5 dB.
Geostationary at 73 W longitude. b. Under conditions of heavy rain, the C-band path
24 C-band transponders, 28 Ku-band transponders to the receive station suffers an attenuation of 1.5 dB.
3.2 kW RF power output Assuming 100% coupling of sky noise into antenna
c04.qxd 28/08/02 19:37 Page 153
PROBLEMS 153
noise, and 0.3 dB clear air gaseous attenuation, cal- b. Draw up a link budget for the downlink from the
culate the overall CN under these conditions, and satellite to an earth station on the 3 dB contour of the
find the downlink margin. satellite antenna beam. Assume that the satellite trans-
Hint: You need to find the sky noise temperature mits at a power level of 180 W. Include a clear air at-
that results from a total excess path attenuation of mospheric loss of 0.5 dB and miscellaneous losses of
1.8 dB (clear air attenuation plus rain attenuation); 0.2 dB in your downlink power budget.
this is the antenna temperature. Then compute the c. The receiving terminal has a system noise tem-
new CN in rain, using the new Tsystem and received perature of 110 K in clear air. Draw up a noise power
power values. budget for the receiver using the receiver’s noise
12.4 Design a transmitting earth station to provide bandwidth.
a clear air CN of 30 dB in a Ku-band transponder d. Calculate the clear air CN ratio for the receiver
at a frequency of 14.15 GHz. Use an uplink antenna with a noise bandwidth of 20 MHz.
with a diameter of 5 m and an aperture efficiency of The minimum permissible CN ratio is 10.0 dB.
68%, and find the uplink transmitter power required What is the clear air link margin?
to achieve the required CN. The uplink station is lo- e. For 0.3% of the time at the receiving location,
cated on the 2 dB contour of the satellite footprint. heavy rain causes 2 dB excess path attenuation and
Allow 1.0 dB on the uplink for miscellaneous and the system noise temperature of the receiver increases
clear air losses. to 210 K. Calculate the CN under these rain condi-
12.5 Design a Ku-band receiving earth station to tions, and the link margin above the CN threshold
provide an overall clear air CN of 17 dB in a 27 of 10.0 dB.
MHz IF noise bandwidth at a carrier frequency of f. Many of the DBS-TV system customers live inside
11.45 GHz. The antenna noise temperature is 30 K the 2 dB contour of the satellite beam. Calculate the
and the LNA noise temperature is 110 K. You may clear air link margin and 0.3% time link margin for
assume a high gain LNA and ignore the noise gener- a receiver located on the 2 dB contour of the satellite
ated in other parts of the receiver. Determine the di- footprint.
ameter of the receiving antenna. The receiving ter- g. An uplink station for the DBS-TV satellite de-
minal is located on the 3 dB contour of the satellite scribed in Question 1 is located in Utah, and transmits
footprint, and clear air attenuation on the path and digital TV signals to 16 transponders on the satellite
other losses total 0.8 dB. using QPSK with three-quarter rate forward error cor-
12.6 a. Under conditions of heavy rain, the Ku- rection. The transmit earth station has a circular aper-
band path to the satellite station suffers an attenua- ture antenna with diameter of 6 m and an aperture
tion of 6 dB. Calculate the overall CN at the earth efficiency of 65%. Each transponder operates at a dif-
station in a bandwidth of 27 MHz under these con- ferent carrier frequency in the 17 GHz band, and the
ditions, and find the uplink link margin. RF channel noise bandwidth is 20 MHz. The noise
b. Under conditions of heavy rain, the Ku-band path temperature of the satellite receiver is 500 K (the
to the receive station suffers an attenuation of 5 dB. satellite always looks toward the “hot” earth).
Assuming 100% coupling of sky noise into antenna Use these values in the remaining parts of this
noise, and 0.3 dB clear air attenuation, calculate the question.
overall CN under these conditions, and find the Calculate the uplink path loss and the uplink an-
downlink margin. tenna gain at 17.5 GHz.
13. A direct broadcast television (DBS-TV) satel- h. The gain of the receiving antenna on the satellite
lite is in geostationary orbit at 100 west longitude. in the direction of Utah is 31 dB.
It carries 16 transponders, each with a saturated out- Draw up a clear air uplink budget for the link from
put power of 200 W and a bandwidth of 25 MHz. The the earth station to a single transponder on the satel-
antenna on the satellite has a gain (on axis) of 34 dB. lite using a transmit power of Pt watts, and atmos-
The receiving terminals all use antennas with a cir- pheric and other losses of 1.0 dB.
cular aperture with a diameter of 18 inches and an i. Calculate the noise power at the input to the satel-
aperture efficiency of 65%. The noise bandwidth of lite receiver in a noise bandwidth of 20 MHz. Hence,
the digital TV receiver is 20 MHz. find the uplink transmitter power required to achieve
Use a distance to the GEO satellite of 38,500 km a CN of 28 dB in the satellite transponder.
in your calculations. j. The gain of the satellite transponder must be set to
a. Calculate the free space path loss and the receiv- amplify the received signal at the transponder input
ing terminal antenna gain at 12.2 GHz. to an output level of 180 W. Calculate the gain of the
c04.qxd 28/08/02 19:37 Page 154
transponder in decibels. (Ignore the change in fre- HANDHELD TRANSCEIVER

quency in the transponder.) PARAMETERS
When designing RF equipment, a common rule to
avoid oscillation is to make the amplification at any
given frequency no higher than 60 dB. How would Antenna gain (transmit and receive) G 0 dB
you design a bent pipe DBS-TV transponder to Receiver system noise temperature Ts 300 K
provide the end to end gain that you calculated? Receiver system noise bandwidth Bn 100 kHz
k. The minimum permissible CN in the transponder
is 16.0 dB.
HUB STATION PARAMETERS
Calculate the clear air link margin for the uplink.
l. Ignore the result you calculated for the downlink Maximum transmit power Pt 100 W
CN in Problem 1, and use a value of 15 dB in this Receiver system noise temperature
question. Convert the clear air uplink and downlink (clear air) Ts 250 K
CN values to power ratios, and then find the over- Antenna gain at 29 GHz (transmit) Gt 54 dB
all CN, in dB, in the earth station receiving
Antenna gain at 19 GHz (receive) Gr 52 dB
terminal. Use the following formula (where CN
values are ratios, not in dB) and give your answer
in decibels: CONSTANTS
1 1CN2 overall 1 1C N2 up 1 1CN2 down Boltzmann’s constant k 1.38 1023 J/K
228.6 dBW/K/Hz
14. This is a multipart question. All the questions are
14.1. Preliminary calculations
about the satellite communications system described
below. a. Calculate the path loss, in dB, for a 2000 km path
at 1.6, 2.5, 19, and 29 GHz.
DESCRIPTION OF SYSTEM b. Calculate the noise power, in dBW, for the receiver
in the transponder and for the receivers at the hub sta-
A satellite communication system consists of 50 LEO
tion and the handheld unit, in a single voice channel
satellites in 750 km orbits, several hubs stations op-
bandwidth of 10 kHz. (Note: Use the bandwidth of
erating in Ka band, and many handheld transceivers
one speech channel, 10 kHz for all the calculations,
operating in L band. The handheld units transmit to
not 2 MHz.)
transponders at 1600 MHz and receive from transpon-
ders at 2500 MHz. The system uses digital speech c. The satellite has broad coverage antennas at L band
compressed into a transmission channel (RF) band- and Ka band with half-power beamwidths of 120. Es-
width of 16 kHz. Channels are spaced 20 kHz apart timate the gain, in dB, of the antennas at each frequency.
to allow a guard band between channels. 14.2. C/N Ratios
The parameters of the system are given below. Use the values you obtained in Problem 1 above, for
(You may not need all of these.) path loss, antenna gain, and noise power in this ques-
tion. Calculate CN values for stations located at the
SYSTEM VALUES edge of the coverage zone of the satellite, where the
satellite antenna gain is 3 dB below its maximum value,
Uplink frequency for handheld
and the range to the satellite is 2000 km. Take care to
transceiver 1600 MHz
use the correct path loss and receiver noise power val-
Downlink frequency for handheld ues for each frequency. Give your answers in decibels.
transceiver 2500 MHz a. Calculate the CN in the satellite transponder for
Uplink frequency for hub station 29 GHz the signal transmitted by one handheld transceiver lo-
Downlink frequency for hub station 19 GHz cated at the edge of the coverage zone (satellite an-
tenna gain 3 dB below maximum) and at maximum
Maximum range to edge of
range from the satellite (2000 km).
coverage zone 2000 km
b. Calculate the CN in the satellite transponder for
SATELLITE TRANSPONDER the signal transmitted by a hub station, using its full
output power.
Maximum output power Pt 20 W
c. Calculate the CN in the hub station receiver for
Transponder bandwidth 2 MHz the signal transmitted by a satellite transponder using
Transponder input noise temperature Ts 500 K its full output power.
c04.qxd 28/08/02 19:37 Page 155
PROBLEMS 155
d. Calculate the CN in the receiver of the handheld d. The communication system described needs two
unit for the signal transmitted by a satellite transpon- transponders to permit two-way voice communication
der using its full output power. between the hub station and the many transceivers.
e. Calculate the overall CN ratios at the hub station Based on your answers in Problem 3, find the gain of
and at the handheld receiver. each transponder from input port to output port.
(Note: The transponder gain does not include antenna
14.3. Trade-off studies
gain.)
The link between the hub station and the satellite op-
erating at Ka band uses a high gain antenna at the hub 14.4. Costs
station and achieves a high CN. The transceiver op- For any satellite system to be viable, the communi-
erating in L band uses a low gain, omnidirectional an- cation capacity must be sold at a price that is attrac-
tenna with low gain, which results in low CN. For tive to customers. This question looks at the cost of
satisfactory operation under all weather conditions, the system over its lifetime and calculates a minimum
the Ka-band links should have a minimum CN of 20 cost per voice circuit.
dB in clear air, and the L-band links should have a a. Each LEO satellite carries 20 transponders. What
minimum CN of 10 dB. is the total number of speech channels that the
The CN of the handheld transceivers can be satellite can support when fully loaded? How many
improved by using a multiple beam L-band antenna telephone circuits (it takes two channels to make a
on the satellite, with higher gain and narrower telephone circuit)?
beamwidth per beam. The high CN of the hub sta- b. Each LEO satellite costs $40 M in orbit and the
tion links can be traded for increased capacity. The LEO system costs $100 M per year to run. The ex-
hub station and transponder transmitter power can be pected lifetime of the satellites is 10 years, and the
shared among a group of voice channels. system requires a total of 10 spare satellites to be
a. Determine the minimum gain required by the L- launched over the 10-year period. Calculate the cost
band antennas on the spacecraft to achieve a CN of of operating the system for a 10-year period. Add a
10 dB at each L-band frequency. Using the higher 27% factor to cover interest payments and dividends,
of the two values, find the 3 dB beamwidth of one of and calculate the 10-year cost of the entire system.
the multiple beams. Estimate the number of beams c. Calculate the cost per minute per voice circuit as-
that will be needed to serve the coverage zone of a suming that each satellite can be loaded to an aver-
single 120 beamwidth antenna. age of 20% of its capacity over its lifetime.
b. Find the excess CN available on the Ka-band links d. Write two paragraphs discussing the cost of the
between the hub station and the satellite. By trading system and the cost of a voice circuit.
carrier power for capacity, find the number of channels
What price per minute would you set for a satellite
that the Ka-band links can carry with CN 20 dB.
voice circuit?
If the channel spacing is 20 kHz, can all of these chan-
nels fit into a 2-MHz bandwidth transponder? Would you expect customers to be willing to pay this
c. Determine whether the transponders are power amount for a satellite telephone connection?
limited or bandwidth limited. Give reasons for your How does the cost compare to terrestrial cellular
answer. telephone charges?
c05.qxd 17/07/2002 19:11 Page 156
CHAPTER 5
MODULATION AND MULTIPLEXING
TECHNIQUES FOR SATELLITE LINKS
Communications satellites are used to carry telephone, video, and data signals, and can
use both analog and digital modulation techniques. When GEO satellites were first used
for communications in the 1960s and 1970s, the signals were almost exclusively analog.
The advent of satellite communications made possible the transmission of wide bandwidth
signals between continents. For the first time, video signals could be sent between North
America, Europe, and Asia. Thousands of telephone channels could be multiplexed through
one transponder and sent across the United States or across the Atlantic or Pacific oceans.
The modulation and multiplexing techniques that were used at this time were analog,
adapted from the technology developed for microwave links in the previous two decades.
Frequency modulation (FM) was the modulation of choice and frequency division multi-
plexing (FDM) was used to combine hundreds or thousands of telephone channels onto
a single microwave carrier. Regional domestic and international satellite systems were
developed to exploit the high capacity and bandwidth that satellites offered.
In the 1980s, optical fibers came into widespread use, and GEO satellites were no
longer used for telephony within the United States. Long-distance telephone links using
optical fibers are digital, so all telephone signals sent via optical fiber have to be in dig-
ital form. At the same time, telephone exchanges became large digital computers instead
of large banks of mechanical switches. The change to digital voice signals made it easier
for long-distance communication carriers to mix digital data and telephone traffic and send
it through the same optical fibers and telephone exchanges. This forced telephone signals
to be converted to digital form at the telephone exchange, and rendered all the analog mul-
tiplexing methods obsolete. FDM has all but disappeared as a way to combine analog tele-
phone signals, replaced by time division multiplexing of digital voice signals. In the first
edition of Satellite Communications, the FM-FDM multiplexing techniques were covered
in detail. This material is included in the second edition in Appendix B, because there are
parts of the world where the older analog technology is still in use, and some readers may
want to refer to that material.
The distribution of television program material in North America and much of the
rest of the world is by satellite. Satellites are particularly useful for distributing the same
signal to hundreds or thousand of receivers (point to multipoint) and many of the world’s
GEO satellite transponders are used for video distribution to cable television systems.
Television signals continue to be transmitted via GEO satellites using frequency modula-
tion and analog video signals in 2000, but it seems probable that eventually all distribu-
tion of video signals will be digital. Digital television has emerged in the form of direct
broadcast satellite television (DBS-TV) and high definition television (HDTV) is also dig-
ital. Frequency modulation for the transmission of video signals, and also for single chan-
nel per carrier voice signals, is included in this chapter. FM is the last remaining form of
analog modulation used in satellite systems.
156
c05.qxd 17/07/2002 19:11 Page 157
5.1 FREQUENCY MODULATION 157
Analog multiplexing in the form of FDM has virtually disappeared, but frequency
division multiple access, FDMA, remains one of the major ways in which transponder ca-
pacity is shared among users. FDMA divides up the frequency band in the transponder
into channels, which are allocated to different signals on a fixed or on-demand basis. Most
SCPC systems employ FDMA, for example, and FDMA is widely used in VSAT systems.
Multiple access techniques are discussed in Chapter 6: this chapter concentrates on
the modulation methods and the digital multiplexing techniques that are used for voice
signals. The T-system and ITU digital hierarchies are described briefly in this chapter, and
since most signals are now transmitted digitally, analog signals must first be converted to
digital form and that process is also described. Once an analog signal is in digital form,
it can be transmitted over any digital communication link, multiplexed with other digital
signals, and sent very long distances without degradation. One major advantage of digital
transmission systems over analog is that error free transmission is possible. In a digital
telephone system, error free transmission means that no noise is injected into the base-
band channel, regardless of the transmission distance, so a telephone call over a distance
of 10,000 km has the same quality as a call over a distance of 10 km.
The design of digital communication links requires a different approach from the
design of analog links, although achieving optimum performance in a digital communi-
cation system requires a link with linear characteristics—an analog circuit rather than a
digital circuit. The techniques for the design of digital communication systems are cov-
ered in some detail because a good understanding of these methods is essential in setting
CN ratios correctly for a digital satellite link, and for estimating the bit error rate that
can be expected. The topic of bit error rates for BPSK and QPSK is covered briefly in
this chapter, and there is also a description of the compression techniques used for dig-
ital voice and video signals. For a more extensive treatment, the reader is referred to any
of several excellent texts on communication theory and communication systems1–4.
5.1 FREQUENCY MODULATION
Frequency modulation is used widely in analog radio communication systems. FM broad-

casting is the most familiar, and FM is also used in analog radio communication systems
such as AMPS cellular telephones. Video signal distribution for the cable TV industry is
now the only significant remaining analog FM satellite service. Digital signals may be
transmitted using a frequency modulation system called frequency shift keying (FSK). FSK
is rarely used in satellite links; phase shift keying (PSK) is preferred. In general PSK gives
lower BER for a given CN ratio than FSK.
Frequency modulation is used in analog satellite systems because it can provide an
improvement in the baseband SN ratio relative to the CN ratio in the IF part of the re-
ceiver. We saw in Chapter 4 that satellite links typically have receiver (CN)0 ratios be-
tween 5 and 25 dB. Most analog systems, including TV, try to maintain a SN ratio of
50 dB in the baseband channel. The use of FM can provide the improvement that can
deliver baseband SN ratios of 50 dB when the CN is 15 or 20 dB.
In general, we have a baseband SN ratio given by
SN 1CN2 0 FM improvement dB (5.1)
where (CN)0 is the overall carrier-to-noise ratio in the earth station receiver at the input to
the FM demodulator. The main disadvantage of using FM in this way is that we need wide-
band FM (WBFM) to achieve the improvement factor, and wideband FM, as its name implies,
uses up a lot of precious RF bandwidth. There is a general principle in communication theory
c05.qxd 17/07/2002 19:11 Page 158
158 CHAPTER 5 MODULATION AND MULTIPLEXING TECHNIQUES FOR SATELLITE LINKS
that you can trade bandwidth for power: in this case trading occupied RF bandwidth for carrier
power at the earth station input. As a result, a NTSC (National Television Standards Com-
mittee) video signal that occupies 4.2 MHz bandwidth at baseband becomes an RF signal
with a bandwidth of 30 MHz in a satellite transponder.
In an FM signal the information is carried by changes in the carrier frequency—called
the frequency deviation, or just deviation. The amplitude of the FM wave is constant, so an
FM signal is sometimes called a constant power waveform since the power in the waveform
doesn’t change with the modulation. The frequency of an FM wave varies directly in pro-
portion to the voltage of the baseband modulation signal. Thus the instantaneous frequency,
fi, of an FM wave is given by the linear relationship
fi fc ¢f fc kf m1t2 (5.2)
where fc is the carrier frequency, f is the frequency deviation, m(t) is the modulating volt-
age, and kf is the modulator constant (modulator sensitivity) in hertz/volt. The modulator
constant can have any value, although values around 10 MHz/V are common in analog
satellite communication systems used for video signal distribution.
Recovering the information signal from an FM wave is conceptually very simple—
we need a frequency to voltage converter with a constant K 1kf volts/hertz. Then the
recovered baseband signal is v(t) where
v1t2 ¢f 1 kf kf m1t2 1 kf m1t2 (5.3)
Thus an FM radio link can have unity gain between the transmitter baseband and the re-
ceiver baseband. An FM link can also transmit a DC (direct current) level, since voltage
is represented by frequency deviation of the carrier. Both of these features are valuable
for telemetry links.
Waveform Equation for FM

We must find the phase angle of the FM waveform because it is conventional to write the
equation of the wave in the form VFM(t) A cos(t (t)) A cos(2 fc (t)). This
expresses the frequency deviation in terms of phase variations. (The units of t (t) in
the term cos(t (t)) must be radians or degrees, since the argument of a cosine must
be an angle.)
After a time t, the phase angle, (t), of the FM wave modulated by an information
signal m(t) is
t
£1t2 32pf 2pk m 1t2 4 dt f
0
c f 0 radians (5.4)
where 0 is the starting phase angle at t 0. Hence the equation of the FM signal is given
by
t
VFM 1t2 A cos vct 1 32pk m1t2 4dt f 2
0
f 0 volts (5.5)
In general, we cannot evaluate the spectrum of the FM wave from Eq. (5.5) unless
we know an analytical expression for m(t).
The spectrum that is generated with any modulating waveform in an FM system is
much more complex than for amplitude modulation (AM) because FM is inherently a non-
linear modulation technique. A sine wave test tone at frequency fm applied to an ideal FM
modulator generates many sidebands at all possible integer multiples of the test tone
c05.qxd 17/07/2002 19:11 Page 159
frequency. Thus instead of the familiar pair of upper and lower sidebands in AM, fre-
quency modulation with a sine wave modulating signal generates an infinite number of
sidebands at frequencies fc fm, fc 2fm, fc 3fm … In practice, real modulators and
transmitters have finite bandwidths, so a finite range of sidebands is actually present in
any transmitted FM signal. The magnitudes of all the sidebands depend on both the am-
plitude and the frequency of the modulating signal, and to make matters worse, if the mod-
ulating signal is a sinusoid, the magnitudes of the sidebands can be found only from Tables
of Bessel functions.
If two sine waves at angular frequencies 1 and 2 are simultaneously applied to a
wideband FM modulator, the spectrum of the resulting FM wave contains not only all the
sidebands for each sine wave, but also all the sidebands from every possible cross-
modulation product between the waves—frequency terms of the form c m1 n2.
Clearly, any real modulating signal which contains many frequencies will generate an FM
spectrum that is continuous, rather than having identifiable sidebands.
Bandwidth of FM Signals: Carson’s Rule

Most texts on communication theory expand the FM waveform equation using a sinu-
soidal test tone as the baseband modulating signal. The expansion of the FM waveform
equation leads to an infinite number of harmonic terms of the form Jn() cos(c nm)t
where m is the frequency of the test tone. The magnitudes of the terms are given by
Bessel function coefficients Jn(). There is, fortunately, a much more simple and useful
way to determine the bandwidth required for an FM signal: Carson’s rule1.
Carson’s rule is empirical, but it works well and is the way that almost every com-
munications engineer working with an FM system actually determines the bandwidth of
FM signals.
Carson’s rule states that the bandwidth required to transmit an FM signal is given
by
B 21 ¢fpk fmax 2 Hz (5.6)
where fpk is the peak frequency deviation and fmax is the highest frequency present in the
modulating signal.
Students often have difficulty in understanding frequency modulation, particularly
when Bessel functions are involved. In practice, Bessel functions are not needed to
understand FM.
Frequency modulation is very straightforward if you remember two facts:
1. The frequency deviation of the carrier is directly proportional to the modulating
signal voltage.
2. The bandwidth required to transmit an FM signal is found from Carson’s rule.
Baseband SN Ratio for FM Signals

Chapter 4 described how to calculate the overall (CN)0 ratio for an earth station receiver.
The calculation was made using the noise bandwidth of the narrow band-pass filter in the
IF amplifier of the receiver, at the output of the IF amplifier and immediately before the
FM demodulator input. For an FM receiver, the value of overall (CN)0 ratio calculated
for the earth station receiver is the CN ratio used here to calculate baseband SN val-
ues. The carrier power is constant in any FM wave, which is simply a sine wave of varying
frequency. Assuming a 1 ohm impedance level, a sine wave with amplitude A volts has a
c05.qxd 17/07/2002 19:11 Page 160
power C A22 watts. For a receiver with IF noise bandwidth Bn Hz, the noise power at
the filter output, assuming white noise at the input, is N0 B watts, where N0 is the single
sided noise power spectral density.
Hence, the CN ratio at the output of the IF amplifier, measured in a noise band-
width Bn Hz is
C N 1A222 1N0 B2 A2 12N0 Bn 2 (5.7)
Information is carried in the deviation of the carrier frequency, f, by making f
proportional to the modulating signal voltage with a modulator constant 1K.
If we assume sine wave modulation at its maximum permissible magnitude, which is
the usual way to calculate SN ratios, we will create a peak frequency deviation fpeak. The
correct demodulator frequency to voltage conversion constant is K 1kf to provide unity
end-to-end gain. Achieving unity gain is useful in any communication system because it
helps to establish defined signal levels at the input and output of the communication link.
The signal voltage at the output of the FM demodulator is proportional to the fre-
quency deviation. However, to analyze signal to noise ratios we must work in terms of
signal power. For analysis purposes, we always assume an impedance of 1 ohm, and
convert a sine wave signal V cost volts to an equivalent power of P watts using P 21
V 21 12V 2 watts. Let us assume that the signal is a test tone sine wave of maximum
permitted amplitude. The baseband signal power, S, is given by
S 12 1K¢fpeak 2 2 1K¢frms 2 2 watts (5.8)
One characteristic of all FM demodulators operating above threshold is that they
suppress the noise at the demodulator output in proportion to the signal strength. Another
important characteristic of FM demodulators is that the output noise power spectral density
(NPSD) is proportional to the square of the baseband frequency, as illustrated in Figure 5.1.
Thus the baseband noise power at the demodulator output must be obtained by inte-
grating the baseband noise power spectral density, N0, over the baseband bandwidth—from
the lower edge frequency f1 to the upper edge frequency f2, as shown in Figure 5.1. Hence
the noise power present in the baseband at the output of the demodulator is given by4
f2
Nout 2N0 c d f 2d f 2N0 c d 1 f 2
f 13 2 3
K 2 K 2 3
(5.9)
A f1 A
NPSD
NPSD ∝ f 2
0 f max Frequency, f
FIGURE 5.1 Noise power spectral density at the output of an FM
demodulator.
c05.qxd 17/07/2002 19:11 Page 161
We will consider only the case where the baseband extends almost to DC, giving f1 0
Hz, so ( f 32
f 31)3 f 323 ( fmax)33, where fmax is the upper end of the baseband fre-
quency range.
Combining the expressions for signal power and noise power from Eqs. (5.8) and (5.9)
1SN2 out 1CN2 32 1Bfmax 2 1¢fpeak fmax 2 2 (5.10)
In decibel form, the baseband SN ratio for an FM receiver with test tone modulation is
1SN2 out C N 10 log10 1BRFfmax 2 20 log10 1 ¢fpeak fmax 2 1.8 dB (5.11)
where
BRF IF bandwidth of receiver RF BW of FM signal from Carson’s rule
¢fpeak peak frequency deviation at transmitter
fmax maximum frequency of baseband signal receiver baseband bandwidth
and the factor of 1.8 dB is equivalent to the numerical ratio 32.
The part of the Eq. (5.11) that follows CN is the SN improvement factor of Eq.
(5.1). Equation (5.11) gives an important result for FM transmission. Although Eq. (5.11)
was derived for a test tone signal, it is applied to any FM link to determine the baseband
SN ratio for the particular parameters of the link. The SN ratio calculated from Eq.
(5.11) therefore serves as a reference value for the quality of the FM link.
The baseband signal-to-noise ratio can be increased well above the CN ratio of
the received FM carrier. In satellite communication systems the CN ratio is always low;
too low to be converted directly to a baseband signal-to-noise ratio of an output signal.
The SN improvement factor can be made large by using a large deviation ratio, D, i.e.,
D W 1, where D is defined as
D ¢fpeak fmax (5.12)
However, a large peak frequency deviation ratio results in a wide RF bandwidth.
Carson’s rule (Eq. 5.6) shows that BRF 2(fpeak fmax) 2fmax (D 1). So using a
large deviation ratio to secure a large SN improvement causes the occupied RF band-
width to become much greater than the baseband bandwidth. This is spectrally inefficient,
but necessary in almost all analog FM satellite links.
Pre-emphasis and de-emphasis

Pre-emphasis and de-emphasis are used with all FM transmissions because it is possible
to reduce the noise power at the output of the de-emphasis circuit in the receiver and thus
to improve the baseband SN ratio. To understand pre-emphasis and de-emphasis, we must
start with the de-emphasis process in the baseband channel of the FM receiver. Equa-
tion (5.9) shows that the noise power spectral density at the output of an FM demodula-
tor is proportional to the square of baseband frequency. This is illustrated in Figure 5.1,
which shows baseband NPSD at the output of an FM demodulator plotted against base-
band frequency. At the high frequency end of the baseband there is much more noise
power than at lower frequencies; the noise power is concentrated into the higher frequen-
cies of the baseband. De-emphasis flattens the noise power curve and therefore reduces
baseband noise, improving the baseband SN ratio.
White noise across the bandwidth of the IF section of the FM receiver (ahead of the
demodulator) has equal power at all frequencies—a flat noise power spectral density (NPSD).
At the output of the FM demodulator noise power is proportional to the square of the baseband
c05.qxd 17/07/2002 19:11 Page 162
frequency. Thus the rms noise voltage Vn, which is equal to the square root of the noise
power, is proportional to the baseband frequency fb, as illustrated in Figure 5.2a
Vn aVn rms fb (5.13)
where Vn rms is the rms noise voltage at the input to the FM demodulator and a is a con-
stant across the IF bandwidth. The function of a de-emphasis circuit is to flatten the NPSD
by adding a circuit with a transfer function Gde( f ), proportional to 1fb after the demodulator
Gde 1 f 2 b fb (5.14)
where b is another constant. When the FM demodulator noise output is processed by the
de-emphasis circuit the output of the circuit is a voltage Vno where
Vno aVn rms fb b fb abVn rms (5.15)
The value ab is a constant. If it is less than one, we can add amplification of the signal
at baseband to compensate. The NPSD at the output of the FM demodulator followed by
a de-emphasis circuit is now independent of baseband frequency instead of following a
square law relationship.
The required gain characteristic of the de-emphasis circuit can be implemented quite
easily with a simple R–C circuit consisting of a series resistor–capacitor circuit wired as
a low pass filter. The circuit has constant gain up to its corner frequency, fd, and then has
the required gain proportional to 1fb, as shown in Figure 5.2b. The de-emphasis circuit
is placed immediately after the FM demodulator as illustrated in Figure 5.2c.
Pre-emphasis
The de-emphasis circuit in the FM receiver has the characteristics of a low pass filter,
which would cut all the high frequency content of the signal. We must add a comple-
mentary circuit in the baseband of the transmitter to counteract the effect of the de-
emphasis circuit in the receiver. This is called a pre-emphasis circuit, and it has a transfer
function, Gpe( f ), that is proportional to frequency
Gpe 1 f 2
1
f (5.16)
b b
When a signal is transmitted through the link, the signal is first pre-emphasized at the
transmitter and then de-emphasized in the receiver. The transfer function of the system,
G( f ), is therefore
G1 f 2 Gpe 1 f 2 Gde 1 f 2
1 b
fb 1 (5.17)
b fb
Thus the signal is unaffected by the process of pre- and de-emphasis, while the noise
power in the receiver baseband is significantly reduced.
A pre-emphasis circuit consists of two resistors and one capacitor in a high pass
configuration, with the same corner frequency fd that is used in the de-emphasis circuit.
The pre-emphasis circuit provides a voltage gain that is proportional to baseband fre-
quency above fd, up to the top of the baseband, as illustrated in Figure 5.2d. Although the
transfer functions of the pre- and de-emphasis circuits do not exactly match Eqs. (5.14)
and (5.16), the low frequency regions where the circuit transfer functions have constant
gain is where noise power from the FM demodulator is smallest, so the departure of the
circuit characteristics from the ideal has little effect on the improvement in baseband SN
ratio achieved with de-emphasis.
c05.qxd 17/07/2002 19:11 Page 163
Noise voltage at
detector output
Detector Characteristic desired

alone after filtering
Modulating frequency, fm
fd
(a)
Transfer function (dB)
fm
fd
(b)
FM De-emphasis
demodulator filter
(c)
Transfer function (dB)
FIGURE 5.2 Pre-

emphasis and de-
emphasis. (a) Noise
voltage at FM detector
output. (b) De-emphasis
filter characteristic.
(c) Location of de-
fm emphasis filter.
fd (d) Pre-emphasis filter
(d) characteristic.
c05.qxd 17/07/2002 19:11 Page 164
The baseband SN ratio is increased by a factor P dB with de-emphasis in the re-
ceiver. The value of P depends on the nature of the signal, the baseband bandwidth, and
the circuits used. P is different for voice, and video signals; typical values are 9 dB for
NTSC video and 5–10 dB for analog FM voice transmission. With de-emphasis in the
receiver, Eq. (5.11) becomes
1SN2 U C N 10 log10 1BRFfmax 2 20 log10 1¢fpk fmax 2 1.8 P dB (5.18)
The suffix U appended to the (SN) ratio indicates that this value of SN ratio is
unweighted. Weighting factors are discussed in the following paragraphs.
Pre-emphasis amplifies the high frequencies in the signal more than the low fre-
quencies, thus adding extra power to the baseband signal. With most voice and video
signals, the energy in the baseband spectrum at the higher frequencies is low, and little
extra power is added to the baseband signal by the pre-emphasis circuit. This is fortunate,
because an increase in the overall power in the baseband signal would require a reduc-
tion in the frequency deviation to keep within the allotted RF bandwidth. That would
reduce the FM demodulator SN improvement and negate part of the overall SN ratio
improvement obtained by adding de-emphasis.
5.2 ANALOG FM TRANSMISSION

BY SATELLITE
For the first 20 years of satellite communications, most signals were in analog form,
and were transmitted using frequency modulation. The traffic sent over satellite chan-
nels was primarily voice (telephony) with some video signal distribution. Voice chan-
nels were multiplexed using frequency division multiplexing (FDM), a technique that
had been employed in terrestrial telephone circuits since the 1920s. FDM has largely
disappeared from both terrestrial and satellite circuits as digital transmission over opti-
cal fibers has replaced analog transmission over microwave and coaxial cable links.
FDM remains in use in some countries however, so information relating to FDM te-
lephony and satellite transmissions using FDM/FM is included in Appendix B. The in-
terested reader will find a description of FDM/FM transmission techniques used in In-
telsat satellites, and a method for calculation of transponder capacity using FDM/FM,
in Appendix B.
Prior to the availability of satellites for video signal distribution, hundreds of mi-
crowave links were needed to deliver network television signals throughout the United States.
SIDEBAR
The distribution of video signals to cable TV com- can send the same signal to thousands of cable serv-
panies via geostationary satellites gained importance ice providers across the entirety of North America.
in the 1970s, and led to the expansion of television There are more than 10,000 individual cable tele-
networks in the United States. Television program vision systems in the United States, each serving a
material is generated at many different locations by large or small geographic area. Apart from local
major networks, and also by smaller companies, at TV networks which are received over the air with
sporting events and from electronic newsgathering VHF and UHF antennas, virtually all television
vehicles. These signals are distributed very effec- program material is sent to the cable stations by
tively via satellite, because a single uplink station satellite.
c05.qxd 17/07/2002 19:11 Page 165
5.2 ANALOG FM TRANSMISSION BY SATELLITE 165
A substantial investment was needed to create a national TV network, and only three national
networks had been created prior to 1970: ABC, CBS, and NBC. The lower cost and wide
distribution of video signals by satellite led to the creation of many new networks and TV
program channels, which initially were available through cable television systems. The
development of direct broadcast satellite television and digital cable systems has encour-
aged the creation of more television program channels, with over 200 available in the United
States alone.
Video distribution of signals for cable TV providers used analog frequency modu-
lation from the mid-1970s through the mid-1990s, at C band and Ku band, using a full
transponder per video signal. The typical cost of leasing a C-band or Ku-band transpon-
der in the mid-1990s was $1 M per year. Digital transmission was introduced to take ad-
vantage of video signal compression techniques, notably MPEG 2, which can reduce the
bit rate of a full motion video signal to 6.2 Mbps. Several such digital compressed video
signals can fit into one transponder, saving the distribution companies millions of dollars
each year. As a result, there has been rapid conversion to digital video transmission, and
analog FM transmissions continue (in 2000, in the United States) largely because some
cable TV receiving stations are not yet equipped to handle digital signals. Digital trans-
mission techniques are discussed in the second half of this chapter, and digital satellite
systems are reviewed in Chapter 10.
Television signals have traditionally been generated and distributed in analog form.
High definition television (HDTV) is digital, and the signals transmitted by direct broad-
cast satellites are also digital. It is clear that digital transmission will quickly overtake the
remaining analog services, but the distribution of analog video signals via geostationary
satellites using frequency modulation remains in place. For that reason, a discussion of
analog television by satellite is included here.
Television Signals
While a number of television transmission standards exist worldwide, the two in most
common use are the North American and Japanese 525 line60 Hz NTSC system and the
European 625 line50 Hz PAL system, respectively. In this text we will emphasize the
NTSC system.
The video signal of a monochrome (black and white) TV transmission carries an
analog representation of the brightness (i.e., the amount of white light) in the picture along
a series of horizontal scanning lines. This is called the luminance signal. Along with the
luminance signal, synchronization pulses are transmitted so that the TV receiver can recre-
ate the scanning process of the camera. Historically, monochrome TV developed before
color, and in the United States color TV was designed so that the color information could
be added to monochrome transmissions without degrading the performance of existing
black-and-white receivers. Any color may be created by an appropriate combination of
red, green, and blue light. Color TV could be transmitted by transmitting the color com-
ponents of each picture separately, but this scheme would require excessive bandwidth.
Instead, three linear combinations of the three components are transmitted and the component
values themselves are recovered at the receiver.
The TV camera generates voltage levels corresponding to the red, green, and blue
light at each point in the picture. We will identify these voltage levels by the letters R, G,
and B. A monochrome receiver responds to the amount of white light at a point in the
picture; this is the luminance, Y, and is related to the color voltage levels by
Y 0.30R 0.59G 0.11B (5.19)
c05.qxd 17/07/2002 19:11 Page 166
The luminance signal is transmitted so that monochrome receivers can receive a color im-
age in black and white. For color reconstruction, two other independent linear combina-
tions of R, G, and B must be transmitted along with Y so that all of the color components
can be recovered. These are called the I and Q signals, given by
I 0.60R
0.29G
0.32B (5.20)
Q 0.21R
0.52G 0.31B (5.21)
The letters I and Q stand for in-phase and quadrature, and together the I and Q signals
(when decoded with the luminance signal) carry the chrominance information about the
color at each point in the picture.
The I and Q signals modulate a color (or chrominance) subcarrier in such a way
that the amplitude of the resulting chrominance signal determines the saturation (de-
gree of purity) of the color at a point and the phase of the chrominance signal deter-
mines the hue (perceived shade) of the color. From the amplitude and phase of the
chrominance signal, a TV receiver determines the shade of the color and the amount
of white light to add. From the luminance signal it determines how bright the color
should be.
In terrestrial broadcasting the luminance (Y) signal, filtered to occupy the band from
0 to 4.2 MHz, modulates the video carrier with a vestigial sideband (VSB) modulator. The
upper sideband is transmitted in full; the lower sideband is partially removed. The result-
ing VSB signal is all that needs to be transmitted for the video portion of monochrome
television.
The chrominance information is transmitted by a color subcarrier at 3.579545 MHz
(hereafter abbreviated as 3.58 MHz). This value was chosen because it places the chromi-
nance signal at a relatively empty part of the luminance spectrum and minimizes color in-
terference with black-and-white reception3. Both the I and Q signals modulate the color
subcarrier through double-balanced mixers to generate double sideband suppressed car-
rier (DSBSC) signals. The subcarrier is phase shifted by 90 before it enters the Q mod-
ulator. Thus, both I and Q components may be recovered at the receiver. Figure 5.3 shows
the spectrum of the baseband video signal.
The baseband audio signal extends from 50 Hz to 15 kHz. It frequency modulates
an audio subcarrier and the resulting FM waveform is added to the video baseband sig-
nal. This leads to the composite TV signal of Figure 5.3b; it consists of the baseband video
signal below an FM modulated audio subcarrier. In U.S. domestic systems transmitting
analog television signals using FM, an audio subcarrier frequency of 6.8 MHz is standard;
6.2 MHz is also used.
In terrestrial broadcasting, the audio and video signals are combined and shifted
in frequency to an appropriate part of the VHF or UHF band for transmission. The ra-
diated signal is a complex combination of FM (the sound), VSB (the luminance), and
quadrature DSBSC (the chrominance) signals occupying a 6-MHz bandwidth. For satel-
lite transmission the baseband video signal (luminance and chrominance) frequency
modulates a video carrier and two stereo audio signals frequency modulate two audio
carriers.
The details of the video modulation depend on the transponder bandwidth available.
Typical values for network TV are a peak deviation fpk of 10.75 MHz and a maximum
video modulating frequency, fmax, of 4.2 MHz. By Carson’s rule (Eq. 5.6) this requires a
transponder bandwidth of 29.9 MHz. FM television signals are often overdeviated, trading
the larger improvement in video (SN) that results for the smaller degradation in picture
quality associated with truncating some of the sidebands5.
c05.qxd 17/07/2002 19:11 Page 167
Color subcarrier
Luminance signal at 3.579545 MHz
Chrominance
signal
f
0 2 4 6
Baseband frequency, MHz
(a)
Audio subcarrier
Video at 6.8 MHz
Audio
f
0 2 4 6
Baseband frequency, MHz
(b)
FIGURE 5.3 Spectra of baseband TV signals (a) Baseband video signal. (b) The
composite (video plus audio) TV signal as transmitted by U.S. domestic satellites.
A TV signal from a satellite is quite different from a broadcast TV signal. Con-

verters (sometimes called set-top boxes) that allow reception of satellite television trans-
mission on conventional home TV receivers must demodulate the incoming FM signals,
recover the baseband video and audio channels, and send separate video and audio sig-
nals to the TV receiver. Alternatively, the baseband TV signal can be remodulated into the
VSB format of broadcast TV transmissions using a locally generated carrier (usually VHF
TV channel 3 or 4).
SN Ratios for FM Video Transmission

The SN ratio in the baseband channel of an analog FM satellite TV receiver is given by
Eq. (5.18) with one additional factor, Q.
1S N2 W C N 10 log10 1BRFfmax 2 20 log10 1¢fpk fmax 2 1.8 P Q dB (5.22)
The factor Q, called a subjective improvement or weighting factor, combines two effects
that improve the apparent quality of a television picture when transmitted by FM. The
suffix W appended to the SN ratio indicates that a subjective weighting has been applied.
This reflects the nature of the video signal and the eye’s response to the on-screen dis-
play. Video signals usually vary between 0 V for maximum brightness of the TV screen,
and 1.0 V for full black. The calculation of SN ratio is based on sine wave signals, which
c05.qxd 17/07/2002 19:11 Page 168
gives a pessimistic result for the video signal. The second effect relates to the nature of
noise from a FM receiver when it appears on a TV screen. Because the FM demodulator
has low baseband noise power at low frequencies, the appearance of FM noise on a TV
screen is less annoying than white noise. The SN ratio given by Eq. (5.18) has to be
increased by the subjective factor Q to account for these effects. A typical value for Q in
a NTSC TV receiver is 8 dB.
EXAMPLE 5.2.1
A standard NTSC signal has a baseband video bandwidth of 4.2 MHz and is transmitted over
the satellite link in an RF bandwidth of 30 MHz using frequency modulation and standard pre-
emphasis and de-emphasis. At the receiving earth station the CN ratio in clear air conditions
is 15 dB. Calculate the baseband SN ratio for the video signal. Assume a de-emphasis im-
provement of 9 dB and a subjective improvement factor of 8 dB in the baseband signal to noise
ratio.
The first step in this calculation is to use Carson’s rule (Eq. 5.6) to find the peak frequency
deviation.
BRF 21¢fpk fmax 2 Hz
Hence
¢fpk BRF2
fmax
15
4.2 10.8 MHz
Now we can evaluate Eq. (5.22) to find the baseband video SN ratio
SN CN 10 log10 1BRFfmax 2 20 log10 1¢fpkfmax 2 1.8 P Q dB
15 10 log10 1304.22 20 log10 110.84.22 1.8 9 8 dB
15 8.5 8.2 18.8 dB 50.5 dB
The FM improvement here is 35.5 dB including the 8 dB subjective weighting. The un-
weighted improvement is 27.5 dB. Figure 4.9 in Chapter 4 shows the unweighted (SN )out versus
(CN )in of an FM demodulator designed for FM satellite TV using the NTSC standard. The un-
weighted SN improvement is the difference between (SN )out and (CN )in when the input signal
is above the 8-dB threshold. Reference to Figure 4.9 shows an unweighted improvement value of
28 dB. These values are typical of cable TV industry systems using analog FM transmission of
video signals.
A baseband video SN ratio of 50 dB is considered to produce a very good quality signal.
The ratings for video signals generally follows these standards:
SN 55 dB No perceptible noise, studio quality signal
SN 50 dB Very good quality signal, noise is just perceptible in background
SN 45 dB Good quality signal, some noise is visible but not annoying
SN 40 dB Poor quality signal, with a lot of noise visible
SN 35 dB Bottom limit for picture quality, high level of noise, very annoying
The cable TV industry uses video SN 45 dB as the lower limit for video signals delivered by
cable to customers’ homes. Satellite links delivering video signals to cable head-ends need to pro-
vide a baseband SN ratio that is well above 45 dB to allow for degradation in the distribution
network.
FM Threshold
All FM demodulators exhibit a threshold input CN value below which the SN im-
provement degrades. The degradation is caused by noise spikes that start to appear at the
c05.qxd 17/07/2002 19:11 Page 169
output of a conventional FM demodulator when the input CN ratio is below about 13 dB.
Analog FM satellite TV became economically feasible only with the development of
threshold extension demodulators that lowered the FM threshold to 9 dB, or lower. On a
TV screen, the noise spikes appear as sparklies, which are bright white dots and bars. A
very simplified explanation of FM threshold follows.
When the magnitude of the rms noise voltage in the IF amplifier approaches one-
quarter of the magnitude of the signal, occasional peaks of the noise will be equal to
or larger than the signal because of the random fluctuations of the noise voltage. As the
carrier power is reduced and noise becomes relatively stronger, the probability that an
instantaneous noise voltage exceeds the signal magnitude becomes more likely. If the
phase of the noise, which is also random, is 180° from the phase of the signal, small
changes in phase angle of either the signal or the noise around 180° cause the sum of
the signal and noise to execute very rapid phase changes. The FM demodulator responds
to changes in frequency, which are rates of change of phase. The larger the frequency
deviation, and therefore the larger the rate at which the signal phase angle is changing,
the bigger the noise output from the FM demodulator. Very rapid changes of phase of the
signal at the FM demodulator input cause very large noise spikes at the demodulator
output. The noise spikes are heard as clicks in an FM radio, and seen as sparklies when
viewing FM video.
Threshold for an FM demodulator is defined as the value of (CN)in at which there
is a loss of 1 dB in SN improvement. This point is at (CN)in 8.5 dB in Figure 4.9.
The rate at which noise spikes appear at the demodulator output increases rapidly as the
CN ratio at the demodulator input is reduced. For a conventional FM demodulator,
threshold occurs around CN 13 dB. One of the key developments in the early 1980s
that made home satellite TV systems possible was the threshold extension FM demodu-
lator. These demodulators have a threshold 4 to 5 dB lower than conventional FM
demodulators, which translates directly to the possibility of operating at a CN ratio
4 or 5 dB lower than could be achieved with a conventional demodulator. For infor-
mation on threshold extension techniques in FM demodulators, the reader is referred to
reference 6.
SCPC FM Links
FM is not widely used for voice links now, having been superseded by digital modula-
tions. However, analog single channel per carrier (SCPC) links have been quite widely
used for voice and have also been used for data transmission by the amateur satellite
community.
In an SCPC system, each earth station transmits at an assigned RF frequency. In
demand access (DA) systems, the frequency is allocated to the station at the time it
requests a channel. Single channel voice systems work well with wideband FM, and
data transmission can be achieved with transmitters designed for voice applications.
SCPC-FM voice systems obtain improvement in SN ratio in exactly the same way as
FM-TV. There is no subjective or weighting improvement though, unless companding
is used (Companding is a contraction of compressing and expanding, a process that de-
creases the dynamic range of the input signal with a nonlinear amplifier (compression)
at the transmitter and compensates with a complimentary nonlinear amplifier (expander)
at the receiver. The reduced dynamic range of the signal allows it to be transmitted at
a higher average level, which improves the signal-to-noise ratio at the receiver. The
-law and A-law compression circuits used in digital telephony are typical companding
operations.)
c05.qxd 17/07/2002 19:11 Page 170
For a linear system without companding, baseband SN ratio is found from Eq. (5.18)
1SN2 FM voice C N 10 log10 1BRFfmax 2 20 log10 1 ¢fpk fmax 2 1.8 P dB
Pre-emphasis and de-emphasis can be applied to FM voice signals with a typical subjec-
tive improvement of 7 or 8 dB. Threshold for narrow band threshold extension FM de-
modulators can be as low as 5 dB. Example 5.2.2 illustrates the performance of a typical
FM satellite voice link. The baseband bandwidth of 3.4 kHz corresponds to a telephone
quality voice channel. The RF channel bandwidth of 45 kHz is also typical.
EXAMPLE 5.2.2
A SCPC-FM satellite link has an RF channel bandwidth of 45 kHz and a baseband maximum fre-
quency of 3.4 kHz. De-emphasis provides a subjective improvement in baseband SN ratio of 7 dB.
Calculate the baseband SN ratio for the voice channel for a receiver CN ratio of 13 dB. If the
FM demodulator has an FM threshold at 6 dB, what is the link margin for this system?
We must first find the peak frequency deviation for the FM wave. Using the same approach
as in Example 5.2.1
¢fpk BRF2
fmax
22.5
3.4 19.1 kHz
The baseband voice channel SN is given by Eq. (5.18)
1SN2 FM voice CN 10 log10 1BRFfmax 2 20 log10 1¢fpk fmax 2 1.8 P dB
13 10 log10 145 3.42 20 log10 119.23.42 1.8 P dB
13 11.2 15.0 8.8 48 dB
This is an acceptable voice channel SN ratio. If the satellite link suffers attenuation in the
slant path, for example from heavy rain, the system is considered operational until the CN value
falls to the FM threshold value, at 6 dB. The link margin is therefore 13
6 7 dB. With CN
6 dB, the FM demodulator is operating at threshold, so the voice channel SN ratio is reduced by
7 1 8 dB, to 40 dB. This is an acceptable lower value for a voice channel. Noise would be
heard, but not at an annoying level. The 1 dB additional loss in SN ratio is because the FM
demodulator is at threshold.
Data Transmission Using Analog FM Channels

Data signals can be sent over an analog FM channel in much the same way as voice signals.
Section 5.3 discusses the transmission of digital signals over radio links. It is shown that the
key to successful transmission of data signals is to create the correct baseband waveform at
the output of the radio receiver, which also means creating the correct waveform at the input
to the transmitter. The received waveform is sampled and thresholded to yield a digital signal
identical to that transmitted, so there are no bit errors.
The baseband waveforms used in digital radio transmissions are called zero ISI wave-
forms (see Section 5.3 for a description), so the waveform at the transmitter input is an
analog signal, and must be sent through a linear transmission system to avoid distortion
of the waveform, which could lead to bit errors.
The theoretical baseband CN ratio of a binary digital waveform must be at least
10 dB to keep bit errors at an acceptable level. An implementation margin of between 0.5
and 2 dB has to be added to the theoretical figure to account for waveform distortion and
nonideal filters, etc., giving a typical minimum required SN ratio of about 12 dB. The
RF channel bandwidth for FM voice transmissions in the VHF and UHF amateur bands
c05.qxd 17/07/2002 19:11 Page 171
SIDEBAR
Transmitters designed to carry FM voice signals can waveform to drive the FM modulator of a voice chan-
be used to carry digital waveforms. Since an FM nel transmitter. Bit rates of 4.8 and 9.6 kbps can be
transmitter–receiver link is essentially linear, it is sent over an RF channel with 15 kHz bandwidth.
ideal for carrying zero ISI waveforms, and can op- Some of the amateur satellites operate in the VHF
erate at relatively low CN ratios. This technique has and UHF amateur bands and employ store-and-for-
been developed very successfully by the amateur ward techniques. Data received at the satellite are
satellite radio community for the relay of data sig- stored on board in memory until another station re-
nals over LEO amateur satellites8. Data signals are quests a download. All amateur radio transmissions
generated by a personal computer, and then sent to are required by law to be sent in the clear, not en-
a terminal node controller (TNC). The TNC packe- crypted, and can be downloaded by any suitably
tizes the data and creates the required baseband equipped earth station.
is only 15 kHz, so narrow band FM must be used. This results in the CN ratio at the
output of the FM receiver being numerically similar to the IF CN ratio—there is little
FM improvement. However, if the received signal from the satellite is above the FM thresh-
old of the receiver, the baseband SN will be high enough to ensure a small number of
errors in the recovered data stream.
The RF signal created by the FM transmitter is classified as FSK, but the waveform
is an example of continuous phase FSK (CPFSK). CPFSK changes the phase of the RF
waveform continuously, rather than abruptly as in classic FSK. The choice of the peak
frequency deviation in a CPFSK signal is related to the bit rate, and some choices lead to
another form of digital modulation, minimum shift keying (MSK). CPFSK and MSK are
both digital modulations that have good spectral efficiency and low error rates for a given
CN ratio. Example 5.2.3 shows how a low data rate digital satellite link using CPFSK
might be implemented.
EXAMPLE 5.2.3
Let’s examine a system that sends a digital data signal at a bit rate of 9.6 kbps in a nominal RF
channel bandwidth of 15 kHz. With near-ideal Nyquist filters, which can be implemented with digital
signal processing (DSP) techniques in the TNC, the bandwidth of the baseband digital waveform
can be restricted to 0.5 bit rate. For a 9.6 kbps data stream, the baseband bandwidth is 4.8 kHz.
If we use a peak frequency deviation of 3.6 kHz, the resulting RF signal bandwidth can be
found from Carson’s rule (Eq. 5.06)
BRF 2 1¢fpk fmax 2

2 13.6 4.82 16.8 kHz
Sending this RF signal through band-pass filters designed for 15 kHz FM analog transmissions re-
sults in little distortion of the recovered waveform. We can find the FM improvement for the base-
band channel waveform by applying Eq. (5.11), assuming no de-emphasis for the moment
1SN2 out CN 10 log10 1BRFfmax 2 20 log10 1¢fpeakfmax 2 1.8 dB

CN 6.2
2.5 1.8 CN 5.5 dB
Thus, if the CN ratio for the signal from the satellite is 10 dB, the baseband waveform has SN
15.5 dB, which is sufficient to ensure no bit errors in the recovered digital signal, even allowing for
an implementation margin to cover distortion of the waveform and nonideal filters.
c05.qxd 17/07/2002 19:11 Page 172
5.3 DIGITAL TRANSMISSION
Digital modulation is the obvious choice for satellite transmission of signals that originate
in digital form and that are used by digital devices. Virtually all signals sent via satel-
lites are now digital. Familiar examples are data transmissions to and from the Internet,
communications between remote terminals and computers, digital telephony, and TV
signals in digital form, such as HDTV and DBS-TV. Digital transmission lends itself
naturally to time division multiplexing (TDM) and time division multiple access (TDMA).
Analog signals that are transmitted digitally can share channels with digital data, since
all digital signals are handled in the same way, and their content is immaterial. Thus a
digital satellite link can carry a mix of telephone and data signals that varies with traf-
fic demand.
This section contains a review of digital transmission techniques. All digital links
are designed in much the same way, using a specific symbol rate, and specific filters
that minimize intersymbol interference (ISI). A symbol in a baseband link is a pulse of
current or voltage. In a satellite link, a symbol is almost always a phase state (BPSK
and QPSK) or a phase and amplitude state (QAM). Digital links are designed for a
specific symbol rate, but one symbol can carry more than one bit. It is important to dis-
tinguish between symbols and bits. They are easily confused because with binary mod-
ulations, such as BPSK, the symbol rate and bit rate are the same. Symbol rates are
given in baud or in symbols per second, abbreviated to sps. We will use symbols per
second in the analysis that follows because the difference between symbols per second
and bits per second is then more obvious, with the comment that the baud as the unit
of transmission rate equal to symbols per second is still in widespread use. The name
of the unit, the baud, is derived from the name Baudot, who was an early French pio-
neer of the telegraph.
Baseband Digital Signals

We will represent baseband digital signals as serially transmitted logical ones and zeroes.
While in computer circuitry a logical zero may be represented by a low voltage (nomi-
nally zero) and a logical one may be represented by a high voltage (e.g., 5 V), this arrange-
ment is inconvenient for transmission over any significant distance and is not used. To
understand why, imagine a transmission line carrying a bit stream encoded this way and
containing approximately equal numbers of ones and zeroes. About half the time the line
voltage will be 5 V and about half the time it will be 0 V; hence the line signal will have
a 2.5 V DC component. All circuits that carry this signal must have a frequency response
that extends to DC, and this is difficult to achieve since many communication circuits con-
tain transformers. To avoid this problem, digital modulators usually accept their input in
a polar non-return-to-zero (NRZ) format: logical ones and zeroes are transmitted as plus
or minus a stated value. Thus a logical one might be transmitted as 1 V and a logical
zero might be transmitted as
1 V. Zero volts is not transmitted except as a transient
value. Throughout this text we will assume a polar NRZ format for data signals unless
we explicitly state otherwise.
Baseband Transmission of Digital Data

Satellite links always carry RF signals, which requires that data be modulated onto a radio
frequency carrier for transmission. However, to provide a better understanding of the way in
which digital transmission systems are designed, we will begin by examining the case of a
c05.qxd 17/07/2002 19:11 Page 173
5.3 DIGITAL TRANSMISSION 173
baseband data link. In a baseband link, the frequency response of the link is assumed to extend
from DC to an upper limit fmax, where fmax is equal to the bandwidth of the link, B Hz. Data
is transmitted in the form of polar pulses; in a binary system, the pulses have amplitudes V
and
V volts, where V can take any value. As mentioned earlier, the average number of V
and
V pulses is made equal so that the average DC voltage on the transmission line is zero.
We begin the analysis of baseband links by determining the conditions under which
intersymbol interference can be minimized. The numerical results for bandwidth and
symbol rate in this section do not correspond to the requirements for a satellite link; they
refer only to a baseband link using a transmission line.
A random sequence of rectangular binary pulses has a power spectral density
sin1pf Ts 2 2
G1 f 2 Ts c d (5.23)
pf Ts
where Ts is the duration of the pulse7. This spectrum is illustrated in Figure 5.4d.
The familiar sin xx (also called sinc x) spectrum shows that energy exists at all
frequencies; to retain the rectangular pulse shape would require an infinite transmission
bandwidth. Practical communication systems are always bandwidth limited. Not only is
infinite bandwidth not available, interference considerations in radio links dictate that a
communication system should use the smallest possible bandwidth, and this is usually one
of the design criteria of a communication system.
In any digital communication system, a symbol is defined by the rate at which
information is sent over the link, in the form of pulses at baseband, or changes in phase
angle of a carrier, for example, in a PSK system. Our discussion of digital transmission
systems will be based on symbols, rather than bits, because we often want to send more
than one bit per symbol in an RF system to conserve bandwidth. Popular modulations that
transmit more than one bit per symbol are QPSK (two bits/symbol) and QAM (up to 10
bits/symbol). QPSK is widely used on satellite links, including direct broadcast satellite
television. High-speed modems designed for telephone lines use QAM to send a high bit
rate in a small bandwidth (e.g., 28.8 kbps in 4 kHz bandwidth).
Nyquist’s criterion for zero ISI, which forms the design basis for every digital trans-
mission system, is based on the use of square root raised cosine (RRC) filters and a specified
symbol rate on the link. If the transmission is binary, the symbol is a bit, and the symbol
has two states. When two bits are sent per symbol, the symbols have four possible states
and the system is denoted as quaternary, hence the Q in QPSK. If a symbol represents more
than one bit, the system is known generically as m-ary, with one symbol having m states.
QAM is a modulation that combines the four phase states of QPSK with multiple
pulse amplitudes as in m-ASK. For example, 256-QAM is a modulation in which each
symbol represents 8 bits and has 256 possible states. There are four phase states and pulses
can have 26 64 possible amplitude levels.
If we take the random pulse train shown in Figure 5.4a and band limit it by pass-
ing the pulses through a low pass filter, the pulse shape will be altered. As an example,
consider the effect of passing the rectangular pulse train through a single RC section, rep-
resenting a very simple low pass filter. The resulting waveform, shown in Figure 5.4b, has
been delayed and pulses are smeared in time—the decaying pulse from one transition ex-
tends into the next pulse interval. Where the pulse pattern is 10 or 01, the amplitude of
the second pulse at the sampling instant shown in Figure 5.4 has been reduced by the
presence of a delayed portion of the preceding pulse. This is called intersymbol interfer-
ence (ISI) and is likely to occur whenever a digital signal is passed through a band-limiting
filter. When noise is added to the waveform, ISI increases the likelihood that the receiver
will detect a bit incorrectly, causing a bit error. In a baseband system, ISI can be avoided
c05.qxd 17/07/2002 19:11 Page 174
v1(t )
0 t
(a )
v2(t )
0 t
sampling
Ts 2Ts 3Ts 4Ts 5Ts instants
(b )
log |H (f )|
R
v1(t ) C v2(t )
f
f0
Ts = RC
1
f0 =
2πRC
(c )
Before filtering
After filter
f
−4/Ts −3/Ts −2/Ts −1/Ts 0 1/Ts 2/Ts 3/Ts 4/Ts
(d )
FIGURE 5.4 Illustration of the effect of low pass filtering on an NRZ signal. (a) Random
NRZ polar pulse train. (b) Waveform output from an RC filter with Ts RC. (c) RC filter and
its transfer function 0H(f )0. (d) Spectrum of bandlimited NRZ pulse train.
by an appropriate choice of low pass filter. Nyquist9 proposed a technique that can theo-
retically produce zero ISI, now known as the Nyquist criterion. The objective is to create
in the receiver a pulse that resembles the sin xx shape, crossing the axis at intervals of
Ts, where Ts, is the symbol period. The receiver samples the incoming wave at intervals
of Ts, as shown in Figure 5.5, so that at the instant one pulse is sampled, the “tails” from
all preceding pulses have zero value. Thus previous pulses cause zero intersymbol inter-
ference (zero ISI) at each sampling instant.
Filters that produce the required zero ISI waveform in the receiver can be realized
in several ways. The baseband transfer function proposed by Nyquist was the raised cosine
c05.qxd 17/07/2002 19:11 Page 175
Data 1 1 1 0
Impulse
data
stream
Ts t
Individual
pulse
waveforms
t
Vt (t )
Baseband
signal
transmitted
t
Vr (t )
Noisy
received
signal
t
Sampling
instants
Ts t
Recovered
NRZ
pulses
t
Data 1 1 1 0
FIGURE 5.5 Transmission and reception of baseband zero-ISI pulses.
function, VNQ( f ), which has a normalized two-sided frequency characteristic given by
VNQ 1 f 2 1 for 0 f 0 611

a2
Rs
2
VNQ 1 f 2 cos2 e c 0 f 0
11
a2 d f
p Rs
2aRs 2
for 11
a2 f 11 a2
Rs Rs
2 2
VNQ 1 f 2 0 for 0 f 0 7 11 a2
Rs
(5.24)
2
where 0 1 and Rs 1Ts is the symbol rate in symbols per second. (Some texts
use the symbol r instead of .) The entire communication link must have this transfer
c05.qxd 17/07/2002 19:11 Page 176
function to ensure zero ISI. The pulse shape generated at the output of the link is vNQ(t),
the required zero ISI waveform, when the filter input is driven by an impulse, (t). The
waveform vNQ (t) is obtained as the inverse Fourier transform of the output from the Nyquist
raised cosine transfer function, which is simply the spectrum of the input pulse multiplied
by the frequency response of the system.
vNQ 1t2 F
1 3VNQ 1 f 2 S1 f 2 4 (5.25)
where F
1[ ] indicates the inverse Fourier transform and S( f ) is the spectrum of the input
pulse. If we use an impulse (t) as the input signal, the input signal spectrum is S( f ) 1,
which is referred to as a flat spectrum, and then
vNQ 1t2 F
1 3VNQ 1 f 2 4 (5.26)
The raised cosine transfer function with VNQ( f ) given by Eq. (5.24) has an impulse
response vNQ(t) given by
cos 1paRst2 sin 1pRs t2
vNQ 1t2 c d c d
11
2aRs t2
(5.27)
pRs t
Figure 5.6 shows the shape of several raised cosine function characteristics for val-
ues of between 0 and 1, and the corresponding waveforms generated by the impulse
response of these filters. Note that the raised cosine function shown in Figure 5.6 is a
voltage transfer function, and that all functions have a value VNQ( f ) 0.5 at a baseband
frequency f Rs2. The case of 0 in Eq. (5.24) yields a rectangular function with a
bandwidth of Rs2. This is the minimum bandwidth through which a signal with a sym-
bol rate Rs can be transmitted while still satisfying the zero ISI condition. Such a function
is not realizable in practice, since we cannot have an infinitely rapid attenuation slope at
one frequency. In fact, none of the Nyquist raised cosine functions can be created in
practice. The requirement in Eq. (5.24) that there be zero transmission above a frequency
Rs2 (1 ) cannot be met with any real circuit. Consequently, all digital transmis-
sion systems can only approximate the ideal Nyquist transfer function, and will therefore
always generate some intersymbol interference. However, the basis for the design of all
digital links is the ideal zero ISI Nyquist criterion. Real filters that give the link a trans-
fer function that approximates the ideal Nyquist zero ISI transfer function will minimize
ISI and maximize symbol rate.
The reader is reminded again that these results apply to a baseband link, not to a
satellite link.
Implementation of a link with a Nyquist transfer function requires specific parts of
the link to have different transfer functions. There are three separate parts to any com-
munication system: the transmitter, the transmission link, and the receiver. These three
parts are in series, so the overall system transfer function is the product of the three in-
dividual transfer functions. We will denote their transfer functions as Ht ( f ) for the trans-
mitter, L( f ) for the transmission link, and Hr( f ) for the receiver. We want the output of
the receiver to be a zero ISI waveform, which we achieve by creating a zero ISI spectrum
VNQ( f ) at the receiver output.
The spectrum of the waveform at the output of the receiver is given by
Vr 1 f 2 S1 f 2 Ht 1 f 2 L1 f 2 Hr 1 f 2 (5.28)
where S( f ) is the spectrum of the signal at the input of the transmitter. We will specify
that S( f ) 1, corresponding to an input consisting of delta functions (t) or (t) rep-
resenting logical ones and zeroes. We will also specify that the transfer function of the
c05.qxd 17/07/2002 19:11 Page 177
Vr (f )
1
α = 0.5
α=0
α=1
0 f
0 Rs /2 3Rs /4 R
Frequency
(a)
vr (t )
α=0
Voltage
α = 0.5
α=1
t
0 Ts 2Ts 3Ts
Time
(b)
FIGURE 5.6 Raised cosine filter frequency characteristic and impulse
response. (a) Raised cosine function transfer characteristics. (b) Corre-
sponding impulse responses.
link must be flat, such that 0L1 f 2 0 1 and that the phase response of the link is linear
with frequency. With these conditions in place, we want the end-to-end transfer function
of the link to be a Nyquist zero ISI raised cosine transfer function.
Hence
Vr 1 f 2 S1 f 2 Ht 1 f 2 L1 f 2 Hr 1 f 2 1 Ht 1 f 2 1 Hr 1 f 2 VNQ 1 f 2 (5.29)
or
VNQ 1 f 2 Ht 1 f 2 Hr 1 f 2 (5.30)
Equation (5.30) is an important result in the design of digital communication links.
It states that the transfer functions of the transmitter and receiver when multiplied
together must equal the Nyquist raised cosine transfer function. One obvious way to
c05.qxd 17/07/2002 19:11 Page 178
achieve this result is to make the transfer functions of the transmitter and receiver
identical, so that
Ht 1 f 2 Hr 1 f 2 1VNQ 1 f 2 (5.31)
A filter with a transfer function equal to the square root of a raised cosine function is
called a square root raised cosine filter or often just a root raised cosine (RRC) filter.
RRC filters are used as the basis for the design of most digital communication links, even
though no such filters actually exist. Real filters (such as Butterworth, Chebychev, Elliptic
function) can only approximate the RRC filter’s transfer function.
Using identical filters in the transmitter and receiver satisfies another desired crite-
rion in communication systems: the matched filter criterion1,4. A communication link
achieves the best possible SN ratio at the receiver when the spectrum of the signal at the
receiver input is a replica of the transfer function of the receiver. The easiest way to achieve
this condition is with identical (matched) filters in the transmitter and receiver. The
resulting structure for a baseband digital communication link is shown in Figure 5.7, to-
gether with the corresponding waveforms and spectra. Because the RRC filters in the trans-
mitter and receiver have zero transmission above the frequency fmax Rs2 (1 ),
Transmitter Receiver
v 1( t ) v 2( t ) v 3( t ) v 4( t )
NRZ Link Zero ISI

data waveform
Impulse Square root Square root
generator raised cosine raised cosine
filter filter
(a )
v 1( t ) NRZ pulse v1(f ) NRZ spectrum
t f
0 Ts 1/Ts 2/Ts
v 2( t ) Impulse v2(f ) Impulse spectrum
t f
0
v 4( t ) Receiver O/P v3(f ) Transmitter O/P
zero ISI spectrum
t f
Td Td + 2Ts Td + 4Ts 1/T
s
Time Frequency
(b ) (c )
FIGURE 5.7 Waveforms and spectra in a baseband data system with raised
cosine filters. (a) System block diagram. (b) Waveforms. (c) Spectra.
c05.qxd 17/07/2002 19:11 Page 179
the requirement that the link transfer function have a flat magnitude and linear phase ex-
tends only up to fmax. Similarly, the delta functions at the input can be narrow rectangu-
lar pulses with a sin xx spectrum that has a first null well above fmax. A useful rule of
thumb is that a pulse of duration Ts 110 fmax is sufficiently short that its spectrum
approximates to a delta function.
Band-pass Transmission of Digital Data

In a radio frequency communication system that transmits digital data, a parameter of the
RF wave must be varied, or modulated, to carry the baseband information. The most popular
choice of modulation for a digital satellite communication system is phase shift keying
(PSK), as described in the following section. Band-pass (or radio frequency) transmission
of digital data differs from baseband transmission only because modulation of an RF wave
is required: the receiver demodulates the modulated RF wave to recover the baseband data
stream. Thus intersymbol interference will occur at the receiver due to band limiting of the
modulated waveform, unless filters that satisfy the Nyquist criterion are used.
An additional constraint usually exists with radio communication systems. The band-
width occupied by any radio transmission is specified to avoid interference with other
transmissions at adjacent frequencies. The output of a transmitter must have a carefully
controlled spectrum that reduces out-of-band signals to a low level. Figure 5.8 shows the
Modulator Transmitter Receiver
v 1( t ) v 2( t ) v 3( t ) v 4( t ) v 5( t ) v 6( t ) v 7( t )
Eq.
Link
BPF Equalizer LPF
x
fc sin x fc
(a )
v 1( t ) NRZ pulse v2(f ) PSK spectrum
t f
0 Ts fc − Rs fc fc + Rs
v 2( t ) PSK wave v3(f ) BPF O/P spectrum
Rs = 1/Ts
t f
fc − Rs fc fc + Rs
v 7( t ) Receiver O/P v4(f )
Transmitted PSK
zero ISI spectrum
t f
Td Td + 2Ts fc − Rs fc fc + Rs
Td + 4Ts
(b ) (c )
FIGURE 5.8 Waveforms and spectra in a PSK data system with raised cosine filters and
xsin x equalization. (a) Block diagram of one channel of a QPSK system. (Equivalent to a
BPSK system.) (b) Waveforms. (c) Spectra.
c05.qxd 17/07/2002 19:11 Page 180
spectrum of a binary PSK (BPSK) signal generated from a random train of binary digits.
The slow decay of the spectrum beyond fc Rs results from the sudden phase reversals
of the PSK waveform.
The Nyquist filters used in a radio link must be band-pass filters, centered at the
carrier frequency of the RF signal. The single-sided transfer function of a band-pass RRC
filter is identical to the two sided baseband frequency response of the equivalent baseband
filter, with its center frequency shifted from 0 Hz to the carrier frequency fc Hz. Thus an
important difference between baseband RRC filters and band-pass (RF) RRC filters is that
the RF version has a bandwidth twice that of the baseband filter. In radio transmitters and
receivers, RRC band-pass filters are always implemented in the intermediate frequency
(IF) section of the transmitter or receiver, rather than the much higher radio frequency
section.
At the transmitter, the RRC filter limits the bandwidth of the transmitted baseband
signal to Bocc Rs(1 ), where Bocc is called the occupied bandwidth of the transmitted
signal, and Rs is the symbol rate. (Some authors call this the absolute bandwidth of the
signal1.) At the receiver, the RRC filter limits the noise that can reach the receiver output
to a noise bandwidth of Bn Rs. Note the very important distinction here: the signal oc-
cupies a bandwidth Bocc Hz, but the noise bandwidth of every band-pass RRC filter is Rs
Hz. It is common in electronic circuits to think in terms of 3 dB bandwidth, but the analy-
sis of radio communication systems requires knowledge of two different bandwidths: the
total bandwidth required to carry the radio signal, and the noise bandwidth of the receiver.
The total bandwidth, which is the occupied bandwidth, Bocc, when Nyquist filters are used
in the transmitter, defines the RF spectrum required to transmit all of the signal energy.
It is often referred to as channel bandwidth, because we can think in terms of establishing
a radio channel with bandwidth Bocc.
Hence for the case of a satellite communication system with a symbol rate Rs sps,
the required RRC filters are identical band-pass filters which have the following bandwidths:
At the transmitter, the RRC filter creates a signal with an occupied bandwidth
Bocc Rs 11 a2 Hz (5.32)
At the receiver, the noise bandwidth of the RRC filter is
Bn Rs Hz (5.33)
Every RRC band-pass filter, regardless of the value of the roll-off factor (also
given the symbol r in some texts), has a noise bandwidth equal to the symbol rate of
the link. This is the most important design features of digital radio links, and must be
observed whenever noise power is calculated in a digital radio receiver. Every RRC
band-pass filter in a digital radio transmitter generates a radio signal with a bandwidth
Rs(1 ) Hz, which is always greater than the numerical value of the symbol rate since
0.
As stated above, the band-pass filters in the transmitter and receiver are usually
implemented at an intermediate frequency fIF. The frequency response of the band-pass
RRC filters therefore extends from a lower limit fIF R2s 11 a2 to an upper limit fIF
2 11 a2 . The corresponding frequency range of the RF signal is the occupied bandwidth
Rs
Bocc, centered on the RF carrier frequency fc
Bocc c fc 11 a2 d c fc 11 a2 d
Rs Rs
(5.34)
2 2
Throughout the radio link, every RF and IF component must have a flat frequency
response (and linear phase characteristics) over the bandwidth occupied by the signal. If
c05.qxd 17/07/2002 19:11 Page 181
any part of the link does not meet the frequency response requirement, an equalizer must
be added in series with that part to force the required frequency response.
EXAMPLE 5.3.1
A satellite link has an RF channel with a bandwidth 1.0 MHz. The transmitter and receiver have
RRC filters with 0.5. What is correct symbol rate (pulse rate) for this link?
The relationship between symbol rate and bandwidth is given by Eq. (5.32)
Bocc Rs 11 a2 Hz
106 Rs 11 0.52 1.5 Rs
Rs 1061.5 666.7 ksps
EXAMPLE 5.3.2
A Ku-band satellite uplink has a carrier frequency of 14.125 MHz and carries a symbol stream at
Rs 16 Msps. The transmitter and receiver have RRC filters with 0.25.
What is bandwidth occupied by RF signal, and what is the frequency range of the transmit-
ted RF signal?
From Eq. (5.32)
Bocc Rs 11 0.252 1.25 Rs
1.25 16 106 20 MHz
The RF signal occupies the frequency range given by Eq. (5.34)
11 a2 fRF fc 11 a2
Rs Rs
fc
2 2
Hence, the frequency band occupied by the uplink signal extends from
14.125 0.01 14.115 GHz to 14.125 0.01 14.135 GHz
In many data transmission systems the baseband waveform at the transmitter input
has an NRZ format. A link with a Nyquist transfer function will produce a zero ISI wave-
form at the receiver output only when driven by an impulse, as shown by Eqs. (5.28) and
(5.29). If the transmitter is driven by an NRZ waveform with a symbol rate Rs, the spectrum
of the driving pulse has a shape
sin1pTs f 2
S1 f 2 (5.35)
pTs f
and the spectrum of the output of the RRC filter will be
sin1pTs f 2
Vt 1 f 2 1VNQ 1 f 2 c d
pTs f
To obtain zero ISI at the receiver, we must force the spectrum of the signal from
the transmitter to be 1VNQ 1 f 2, which can be achieved by using an equalizer in the
transmitter with a transfer function given by Et 1 f 2 c d . Then the output
pTs f
sin1pTs f 2
spectrum from the transmitter is given by
Vt 1 f 2 S1 f 2 Et 1 f 2 Ht 1 f 2
sin1pTs f 2
c d c d 1VNQ 1 f 2 1VNQ 1 f 2
pTs f
sin1pTs f 2
(5.36)
pTs f
c05.qxd 17/07/2002 19:11 Page 182
SIDEBAR
Adaptive equalizers are used in all high-speed tele- prove its shape. A training sequence is required in the
phone line modems because the frequency character- signal to allow the transversal equalizer to adjust to
istics change each time a new number is dialed. The the received signal. In a dial-up modem, the training
equalizer is implemented as a transversal equalizer, sequence is sent at the time the connection is first
which operates in the time domain rather than the fre- made, with the assumption that the line characteris-
quency domain1. A transversal equalizer works to im- tics will not change during the subsequent data trans-
prove the pulse shape at the output of the receiver so fer session. In a mobile radio system the equalizer
that the pulses more closely resemble the ideal zero must continuously adapt to the changing propagation
ISI shape. The received pulses are sampled repeat- path. For more detail on transversal equalizers, the
edly within the pulse period, and then weighted and reader should refer to a text on communications
delayed samples are added to the original pulse to im- system theory10.
The arrangement is illustrated in Figure 5.8a. The raised cosine filters have zero
transmission beyond frequencies defined by f fc fmax, where fmax Rs2 (1 ).
The xsin x equalizer operates only within the central lobe of the sin xx function. At
f 1Ts the xsin x function goes to infinity, so the RRC filter parameter must be less
than 1 for this system to work. In practice, RF filters with raised cosine shaping use
values between 0.2 and 0.5. The maximum gain of the equalizer for 0.5 is given
by G 20 log10 [0.75sin(0.75)] 10.5 dB, at the edge of the equalizer band. Trans-
fer functions are given in voltage terms, hence we must use 20 log ( ) to obtain the gain
of the equalizer in decibels.
The bandwidth of the link must at least equal the bandwidth occupied by the sig-
nal, otherwise the spectrum of the received signal will be altered by transmission over the
link and we will not have a zero ISI raised cosine waveform at the output of the receiver.
Thus the transfer function of the link, L( f ), must be such that 0 L( f )0 1 and £( f ) kf,
where £( f ) is the phase characteristic of the link with frequency, and k is a constant, over
the bandwidth occupied by the signal, Rs(1 ) Hz. If this condition is not met, either
in magnitude or phase, we can insert an equalizer in series with the link to force the re-
quired condition. The equalizer can be at the transmitter or the receiver. In systems where
the characteristics of the link vary with time, as in a mobile link which suffers from mul-
tipath interference, for example, the link equalizer can be adaptive. Multipath is less of a
problem for mobile satellite links than for cellular telephone systems, where adaptive
equalizers are commonly used, because the satellites are used only when their elevation
angle is at least 5.
The discussion of filter characteristics and signal spectra thus far has ignored the phase
response of the filters and the resulting phase spectra of the waveforms. It can readily be
shown7 that the phase response of all filters and equalizers must be linear with frequency
for the zero ISI condition to be met. Achieving a linear phase response throughout a com-
munication system can be difficult in practice, and may require the use of phase equalizers.
Transmission of QPSK Signals through

a Bandlimited Channel
Figure 5.8 shows a block diagram of a typical BPSK link. The same transmitter and re-
ceiver system, and the same satellite link, can be used to send QPSK signals, since QPSK
is nothing more than two BPSK signals generated from carriers that are in phase quadra-
ture, which can share a common link bandwidth. (Any pair of signals that is orthogonal
c05.qxd 17/07/2002 19:11 Page 183
can be separated by suitable signal processing in a radio receiver. Signals in frequency

bands that do not overlap are orthogonal, and signals in phase quadrature are orthogonal.
CDMA signals using ideal codes are also orthogonal.) The input equalizer with a xsin x
transfer function Et( f ) is placed after the PSK modulator because typically PSK modula-
tors are binary state devices which need to be driven by NRZ digital signals with voltage
levels V. The input to the PSK modulator must therefore be a V signal, with V
data l and
V data 0 (or vice versa), created from the logic levels of the NRZ wave-
form that carries data bits (usually 5 V and 0 V, or 3.3 V and 0 V).
The waveform at the output of the receiver, v7(t), must be a zero ISI waveform. This
can be achieved only if we can make the transfer function of the entire link, from input
terminal to output terminal, equal to VNQ( f ), the Nyquist raised cosine transfer function.
The condition for zero ISI in the link is therefore
Vi 1 f 2 Ht 1 f 2 Et 1 f 2 L1 f 2 Hr 1 f 2 GLPF 1 f 2 VNQ 1 f 2 (5.37)
where
V i( f ) sin xx, the spectrum of the NRZ data pulses at the system input
Ht( f ) Hr( f ), the transfer function of the band-pass RRC filters in the
transmitter and receiver
E t( f ) xsin x, the equalizer for the input NRZ waveform
L( f ) is the transfer function of the link, equalized if necessary to make it linear
GLPF( f ) is the transfer function of the baseband low pass filter following the
demodulator.
In order to satisfy the conditions for Eq. (5.37) to be true within the limits of the
bandwidth occupied by the signals in the link, Rs(1 ), we must ensure that
Vi 1 f 2 Et 1 f 2 1
L1 f 2 1
GLPF 1 f 2 1
Ht 1 f 2 Hr 1 f 2 VNQ 1 f 2 (5.38)
QPSK (quadrature phase shift keying) is the most popular choice of modulation
technique for use in satellite communication links carrying digital data. It will be
described in more detail in Section 5.4, but basically a digital data stream is taken
two bits at a time and used to generate one of four possible phase states of the trans-
mitted carrier. If the data rate is Rb, bits/s, the symbol rate for the QPSK carrier is
Rs Rb2 sps.
In order to recover the symbol stream with zero ISI, we must shape the transmit-
ted spectrum such that after demodulation a single symbol creates a zero ISI waveform
at the output of the demodulator. Then, sampling of the symbol stream can be achieved
with zero intersymbol interference. In practice, a QPSK system has two demodulators,
one for each pair of symbols (phase states) in the QPSK carrier. We shall consider only
one channel in looking at ISI.
Figure 5.8a shows a typical arrangement for one-half of a QPSK transmit–receive
link. The other half is identical except that the carrier used for modulation and demodu-
lation is shifted in phase by 90°. Since the carriers in the two channels have a 90° phase
difference, the channels are identified as in-phase (I) and quadrature (Q). The data pre-
sented to the QPSK modulators is in NRZ format and causes a jump in carrier phase at
each symbol transition, as seen in Figure 5.8b. The input data rate to the demodulator is
Rs Rb2 sps, giving the QPSK spectrum shown in Figure 5.9a.
c05.qxd 17/07/2002 19:11 Page 184
Normalized Power Spectral Density (dBW/Hz)

0
−10
−20
−30
−1.0 0 +1.0
Normalized frequency [f − fc ]
Rs
(a)
α = 0.45
Relative gain (dB)
−10
−20
−0.8 −0.4 0 0.4 0.8

Rs
(b )
0
FIGURE 5.9 (a) The fre-
quency spectrum of an
Relative gain (dB)
unfiltered QPSK signal with

−10 carrier frequency fc and
α = 0.45
equalized by symbol rate Rs. Only the
x central lobe of the spectrum
sin x is shown. (b) Transfer
function of RRC band-pass
−20 filter with 0.45. (c) The
frequency spectrum of a
−0.8 −0.4 0 0.4 0.8 QPSK signal with carrier
frequency fc and symbol
rate Rs after equalization by
Rs
xsin x and processing by
(c ) an RRC filter with 0.45.
c05.qxd 17/07/2002 19:11 Page 185
The central lobe of the unfiltered QPSK spectrum extends from ( fc

Rs) to ( fc
Rs), giving a band occupancy of 2Rs. The spectrum must be narrowed for transmission via
a radio channel, and this is achieved by use of a band-pass filter meeting the zero ISI
criterion, for example, a square root raised cosine filter.
The band-pass square root raised cosine filter is a transformation of the low pass
square root raised cosine filter and has a response 0 H( f )0 12 at frequencies ( fc Rs2)
and ( fc Rs2). The frequencies at which H( f ) falls to zero are determined by the roll-
off factor in Eq. (5.24). Matched filter operation of the link requires that the raised co-
sine filter response be split between the transmit end and the receive end of the link.
Thus a square root raised cosine response filter is required after the modulator in the
transmitter and before the demodulator in the receiver. Finally, because we are using
NRZ pulses rather than impulses, we need an xsin x equalizer with x [ fc
f Rs2]
to equalize the spectrum from the modulator.
EXAMPLE 5.3.3
A satellite transponder has a bandwidth of 36 MHz. Earth stations use RRC filters with 0.4.
What is the maximum bit rate that can be sent through this transponder with
(i) BPSK
(ii) QPSK?
The maximum symbol rate for an RF link is given by Eq. (5.32)
Rs B 11 a2 36 M1.4 25.7 Msps
The corresponding bit rates for BPSK and QPSK are
(i) BPSK Rb Rs 25.7 Mbps
(ii) QPSK Rb 2 Rs 51.4 Mbps
EXAMPLE 5.3.3
A data stream at 240 Mbps is to be sent via a satellite using QPSK. The receiver IF frequency is
240 MHz. Find the RF bandwidth needed to transmit the QPSK signal when raised cosine filters
with 0.45 are used.
The 240 Mbps signal is divided into two 120-Msps symbol streams and applied to I and Q
channel modulators fed by an IF carrier, with a 90° phase difference. The resulting spectrum from
each modulator has a width of 240 MHz between zeros of the central lobe of the PSK spectrum.
The I and Q signals are added and applied to an xsin x equalizer with x [ fc
fRs] extending
to 87 MHz from the carrier. The maximum gain of the function xsin x at 87 MHz from the
carrier is 9.5 dB. Figure 5.10a shows a block diagram of one-half of the transmit end of the QPSK
link.
The equalized spectrum is applied to the square root raised cosine filter. The response of this
filter is 3 dB down at fc Rs2, that is, at fc 60 MHz. In practice, one filter combining the square
root raised cosine and xsin x responses is used. The ideal combined response of this single filter
is shown in Figure 5.9c. Thus in the IF amplifier of the receiver, the signal spectrum is 6 dB down
at 180 and 300 MHz. The band-pass filter will have zero response for f fc
(Rs2)(1 ) and
f fc (Rs2)(1 ), that is, below 153 MHz and above 327 MHz. Figure 5.10b shows the trans-
mitted QPSK spectrum centered of the IF carrier.
If we examine the spectrum of the QPSK signal at the receiver, we find that the 3-dB band-
width is 120 MHz and the total frequency band containing all of the signal energy is 174 MHz. A
typical satellite transponder for such a signal would have a 3 dB bandwidth of 140 MHz. Beyond
c05.qxd 17/07/2002 19:11 Page 186
X Square root
sin X raised cosine QPSK
PSK modulator Equalizer filter signal
at RF
Eq.
NRZ
symbols Upconverter
120 Mbps
240 MHz Orthogonal

IF carrier channel
(a)
Normalized Power Spectral Density (dBW/Hz)
−10
−20
140 MHz
transponder channel bandwidth
−30 f
120 180 240 300 360
Frequency (MHz)
(b)
FIGURE 5.10 A QPSK data system with an NRZ format and symbol rate Rs 120 Msps sym-
bol rate. IF frequency is 240 MHz. (a) Transmit end block diagram for model of 120 Msps sym-
bol rate QPSK link. Only one channel is shown. (b) Transmitted QPSK spectrum after processing
by square root raised cosine filter and xsin x equalization. Filter roll-off factor is 0.45.
140 MHz the spectrum of the QPSK signal would be attenuated by the transponder filter, leading
to some spectral distortion of the receiver signal and consequent ISI in the demodulated waveform.
However, the energy contained in the QPSK spectrum beyond 70 MHz from the carrier is small,
and the ISI caused by the transponder filter is minimal.
Practical filters invariably cause some ISI because it is impossible to realize the
square root raised cosine characteristic exactly. Typically, with a high-speed digital link
operating at 120 Msps, up to 2 dB of additional carrier power must be provided to achieve
a 10
8 BER, compared to the theoretical power needed for this error rate, in a carefully
filtered QPSK link. The extra power is called an implementation margin. Implementation
margin accounts for the nonideal nature of real communication systems. Timing jitter lead-
ing to incorrect sampling instants, real filters rather than Nyquist RRC filters, and phase
and amplitude distortions all contribute to ISI levels that are higher than theory predicts.
Implementation margin covers all of these effects.
c05.qxd 17/07/2002 19:11 Page 187
5.4 DIGITAL MODULATION AND DEMODULATION 187
5.4 DIGITAL MODULATION

AND DEMODULATION
In this section we will review methods for digital transmission used on satellite links. We
will not attempt to summarize the extensive literature of digital communications in gen-
eral; the interested reader should refer to the references that deal specifically with digital
communication systems (e.g., [1, 10–13, 17]).
Terminology
While any parameter of a carrier waveform—amplitude, frequency, or phase—may be dig-
itally modulated, phase modulation is almost universally used for satellites. For historical
reasons, digital phase modulation is called phase shift keying, abbreviated PSK. An
m-phase PSK modulator puts the phase of a carrier into one of m states according to the
value of a modulating voltage. Two-state or biphase PSK is usually called BPSK, and
four-state or quadriphase PSK is termed QPSK. Other numbers of states and some com-
binations of amplitude and phase modulation are possible and are employed in terrestrial
links, but historically, satellite users have been reluctant to adopt anything besides BPSK
or QPSK. An important reason for this is the higher values of CN ratios required for
acceptable bit error rates. Several of the next generation of Ka-band satellites designed
for Internet access will use 16-QAM on the link between the satellite and hub station,
where higher values of CN can be maintained18,19. Baseband processing transponders are
used to convert the 16-QAM signals on the hub–satellite link to QPSK for the links to the
customer earth stations. The 16-QAM signal does not pass through a nonlinear transpon-
der and intermodulation problems are avoided.
Any type of PSK can be direct or differential, depending on whether it is the state
of the modulating voltage or the change in state of the modulating voltage that determines
the transmitted phase. Whether direct or differential, a PSK modulator causes the phase
of a carrier waveform to go to one of a finite set of values. The transition time plus the
time spent at the desired phase constitute a fixed time interval called the symbol period;
the transmitted waveform during the interval is called a symbol. The set of all symbols
for a particular modulation type is called its alphabet. Thus BPSK has a two-symbol
alphabet and QPSK has a four-symbol alphabet.
In the digital modulation process, a stream of incoming bits determines which symbol
of the m available in the alphabet will be transmitted. Mathematically, Nb bits are required to
specify which of m possible symbols is being transmitted where Nb and m are related by
Nb log2 m (5.39)
As defined by this equation, Nb is the number of bits per symbol for the m-PSK modulation
scheme. Standard practice is to make m a power of 2 so that Nb will be an integer.
Modulation and Coding

The boundary between digital modulation and digital encoding is not well defined. In
encoding for forward error correction (FEC), redundant bits are added to an incoming
bit stream so that errors in transmission may be detected and corrected at the other end
of the link. When the redundant bits are added at baseband and the composite (informa-
tion bits plus redundant bits) bit stream is used to phase modulate a carrier and produce
the transmitted symbols, then the division between modulation and encoding is obvious.
But the modulator itself may be designed to add redundant bits during the modulation
c05.qxd 17/07/2002 19:11 Page 188
process, as in trellis coding, for example, making encoding and modulation inseparable15,16.
In this section we will ignore encoding for FEC and concentrate strictly on the modula-
tion process for turning an incoming bit stream into RF symbols. We will assume that any
FEC encoding is done ahead of the modulator by methods to be presented later.
It is unfortunate that differential phase modulation is frequently called differential en-
coding, since it is a characteristic of the modulation and demodulation equipment and plays
no role in coding as it is usually understood. Differential encoding would more properly be
called differential modulation, and we will discuss it after we have presented direct modulation.
Bit and Symbol Error Rates

The figure of merit for a digital radio link is its bit error rate (BER), which should strictly
be called the bit error probability (Pb). Mathematically, this is the probability that a bit
sent over the link will be received incorrectly (i.e., that a 1 will be read as a 0 or vice
versa) or, alternatively, the fraction of a large number of transmitted bits that will be re-
ceived incorrectly. Like a probability, it is usually stated as a single number, for example,
1 10
4 or 0.0001. The BER plays the same role as an indicator of quality in a digital
communication system that the signal-to-noise ratio plays in an analog link. It is impor-
tant to remember that, despite its name, bit error rate does not depend on the speed of a
digital transmission. It is simply the likelihood that a single bit error will occur within N
received bits, independent of the rate of transmission.
Physically a bit error occurs because a symbol error has occurred. At some point in
the link noise has corrupted the transmitted symbol so that the decision circuitry at the
receiver cannot identify it correctly. For example, the carrier phase may have been trans-
mitted as 90° but additive noise may have changed the received carrier phase to
90°.
If one symbol carries Nb bits and if differential modulation is not used, then a single sym-
bol error may cause 1, 2, … , Nb bit errors. With differential modulation, an error in one
symbol will cause the symbol that follows to be misinterpreted, and the number of bit
errors per symbol error may exceed Nb, the number of bits per symbol.
The mapping of pairs of data bits to the phase angles of a QPSK signal is usually
done using Gray coding. Gray coding ensures that adjacent phase states differ by only a
single bit, rather than two bits. Noise must cause a phase error in excess of 135° to cause
a two bit error in a QPSK signal. Noise that causes a phase error of more than 45° but
less than 135° will cause a single bit error. It is therefore much more likely that a single
bit error will occur than a two bit error when Gray coding is used. As a result, it is usual
to assume that for QPSK signals the probability of a symbol error, Pe is equal to the prob-
ability of a bit error, Pb. The curve relating BER to CN is so steep that the coefficient
applied to Pb has very little influence on the bit error rate calculation, so it is usually set
to unity in practice. (See Figure 5.13.)
Symbol errors arise from thermal noise, from external interference, and from inter-
symbol interference. If only thermal noise is considered, then the symbol error rate or
symbol error probability (Pe) may be calculated unambiguously from EsN0, the energy
per symbol in joules divided by the noise power density in W/Hz, measured in the IF
bandwidth at the demodulator input. The higher the value of EsN0 the lower will be the
BER. The value of EsN0 may be determined from the input value of CN ratio, expressed
as a ratio (i.e., not in decibels).
Assume that C watts of carrier power are transmitted during one symbol interval Ts
seconds, where Ts 1Rs and Rs is the symbol rate on the link. The energy received dur-
ing that symbol period is Es, where
Es C Ts C Rs (5.40)
c05.qxd 17/07/2002 19:11 Page 189
The noise power spectral density N0 is the received noise power N (in watts) divided
by the noise bandwidth Bn (in hertz) at the demodulator input
N0 NBn (5.41)
Combining the last two equations we have
EsN0 CBn N0 Rs (5.42)
The square root raised cosine (RRC) filter discussed in the preceding section has a
noise bandwidth Bn Hz numerically equal to the symbol rate Rs in symbols per second.
Thus a receiver designed with filters of this type to achieve zero ISI also has BnTs 1
and EsN0 CN. Practical filters such as Butterworth or Chebychev come close to the
shape of the square root raised cosine filter, giving BnTs products close to unity.
Binary Phase Shift Keying (BPSK)

In binary phase shift keying, an incoming bipolar bit stream u(t) sets the phase of a
carrier to 90° (2 rad). Thus, if ui is the ith bit, then the transmitted carrier v(t), is
given by
v1t2 Vc cos1vct uip22 (5.43)
where Vc is an arbitrary amplitude frequently set to 1, and c is the carrier frequency in
rad/s. Since ui must have a value of 1, a logical one (baseband voltage V) is transmitted
by setting the phase to
2 rad and a logical zero (baseband voltage
V) is transmit-
ted with a 2 phase. Using trigonometric identities, we may rewrite Eq. (5.43) as
v1t2 uiVc sin1vct2 (5.44)
and we see that BPSK resembles amplitude modulation in which the modulating signal
has a value 1 or
1 only. This causes the BPSK waveform to have a constant ampli-
tude and an envelope AM detector cannot demodulate it.
To recover the data bits ui the receiver uses a correlation detector, which is the
equivalent of a matched filter. A correlation detector multiplies the received signal by a
replica of the transmitted signal, integrates the result over the symbol period, and sam-
ples the output of the integrator at the end of the period1. The practical implementation
of a correlation detector for BPSK signals is shown in Figure 5.11. The replica of the
transmitted signal is a locally generated carrier that creates a coherent (in phase) sine wave
at the carrier frequency. The BPSK detector of Figure 5.11 multiplies the received signal
by the local carrier in a mixer, and then uses a low pass filter, rather than an integrator to
recover the detected waveform. Ideally, the recovered time waveform at the low pass filter
IF BPSK multiplier Zero ISI waveform

signal
Sample Symbol
LPF
gate recovery
Output
symbols
sync sync
Carrier
recovery IF oscillator Bit clock
FIGURE 5.11 A coherent BPSK detector and symbol recovery circuit.
c05.qxd 17/07/2002 19:11 Page 190
output is a zero ISI waveform, which has a maximum value in the center of the received
symbol period.
Multiplying the BPSK signal by a coherent carrier, followed by low pass filtering
is the key to recovering the baseband signal. The BPSK signal is defined in Eq. (5.44),
without shaping by the Nyquist filters of the transmitter and receiver. When multiplied by
a locally generated, in phase carrier sin t, the multiplier output is v0(t) where
v0 1t2 uiVc sin vct sin vt
uiVc sin2 vct uiVc 12 11
cos 2vct2 (5.45)
The low pass filter removes the double frequency term cos2 t leaving an output signal vr(t)
vr 1t2 12 uiVc (5.46)
Thus the BPSK demodulator has recovered the modulating signal ui. The magnitude fac-
tor Vc is removed in the IF stages of the receiver by limiting the magnitude of the received
signal, exactly in the same way that a broadcast FM receiver limits a received FM signal
prior to demodulation.
At the center of each symbol interval, the output of the demodulator is sampled and
a decision circuit decides whether vr(t) is greater or less than zero volts (i.e., whether the
sample is a positive or a negative voltage, and thus determines whether ui was a 1 and
represented a data one or whether it was a
1 and represented a data zero.
The reference carrier that drives the multiplier (mixer) of the BPSK demodulator
shown in Figure 5.11 must be derived from the received signal. This is accomplished with
a carrier recovery circuit. One such circuit applies the BPSK signal to both inputs of a
multiplier (mixer) to obtain the square of the BPSK signal. This has the effect of strip-
ping off the modulation, but returns a double frequency signal component. The BPSK sig-
nal given by Eq. (5.44) is squared to give
3v1t2 4 2 1ui 2 2V 2c sin2 vct (5.47)
Since ui is either 1 or
1, (ui) is always 1. We will ignore the
2
V 2c
term since it is easy
to limit an AC (alternating current) waveform to a predetermined magnitude. Expanding
the sin2 ct term gives an output from the squarer circuit of
v0 1t2 12 31
cos 2 vct4 12
12 cos 2 vct (5.48)
We can extract the double frequency term with a high pass filter and then divide it
by two, which is easily accomplished with a phase locked loop (PLL), to give a reference
carrier at the correct frequency, and with the correct phase angle. (The PLL provides a
90° phase shift that converts cos t to sin t.) Other techniques for carrier recovery, such
as the Costas loop, are also used in coherent PSK receivers10,17.
Most carrier recovery loops for BPSK have a 180° phase ambiguity; that is, when
the loop is locked the phase of the recovered carrier may differ by 180° from the correct
value. This has the effect of interchanging logical ones and zeroes and causes the de-
modulated bit stream to be the complement of what was transmitted. With QPSK, the loop
can lock up in four different states, with offsets of 90°, 180°, and
90°. There are sev-
eral ways to eliminate the ambiguity; one is to use differential encoding in which adja-
cent symbols have the same phase if the modulating voltage is a logical 1 and are 180°
out of phase if the modulating voltage is a logical 0. This may be realized by a binary
phase shifter that toggles between 0° phase shift and 180° phase shift each time the
modulating bit is a 0. Incoming logical 1 values have no effect.
Differential modulation is more error prone than direct modulation, since an error
on a single bit in a differential system will cause one or more subsequent bits to be
c05.qxd 17/07/2002 19:11 Page 191
interpreted incorrectly. See Section 5.6 of reference 11 for a detailed analysis of errors in
differential PSK. Most practical satellite systems avoid differential encoding and check
the status of the recovered carrier phase periodically by transmitting a known (unique)
word. Logic in the receiver looks for the unique word. If the unique word is received cor-
rectly, then the recovered carrier phase is correct. If the unique word is the complement
of the known word (1s and 0s interchanged), than the recovered carrier phase is off by
180° (in a BPSK receiver) and the demodulated data stream should be complemented be-
fore it is sent to the end user.
Probability of a Symbol Error

The received signal will always be accompanied by noise. The noise power in the receiver
is calculated using the techniques discussed in Chapter 4, which provide a value of CN
ratio at the input to the demodulator. The CN ratio expresses noise as a power ratio rel-
ative to the received carrier. The noise in the receiver is assumed to be additive white
Gaussian noise (AWGN), leading to a description of the satellite link as an AWGN channel.
The assumption that the noise is additive, white, and has a Gaussian voltage distribution
is necessary to simplify the analysis of error probability. In satellite links, the assumption
is usually valid provided the noise is thermal in origin (e.g., from the receiver front end
and transponder). If the noise is actually interference from another communication link,
the assumption may not be valid, but often AWGN conditions are assumed for want of a
better way to analyze interference.
For the analysis of error probabilities, we need to work with voltages. The noise
voltage at the output of the demodulator is given by n0(t). At the sample instant, we will
assume a noise voltage n0 volts at the demodulator output. The decision circuit will make
an error if noise changes the sign of v(t) at the sample instant. This is illustrated in Fig-
ure 5.12, where the received signal has sample voltages of V and
V volts. At the sam-
pling instant, the output of the demodulator v0 will be the sum of the signal sample V
and the noise sample n0.
v0 V n0 (5.49)
There are two possible ways that an error can occur, depending whether a V or a
V signal was transmitted. If a signal V was sent and n0
V (i.e., the magnitude of the
noise sample is negative and larger than V) the sum of the signal and noise will be less than
zero, giving a symbol error. If the transmitted signal was
V and noise at the sampling in-
stant was greater than V, an error will occur because the sum is greater than zero volts.
We can calculate the probability that an error will happen, and thus the symbol er-
ror rate Pe by the following argument. At the correct decision time, the amplitude of the
signal will be V, where V is the peak magnitude of the waveform at the output of the
demodulator. The Gaussian distribution is symmetrical about zero volts, so the probabil-
ity of an error occurring is
Pe 12 P1output is 7 0 when
V was sent2 12 P1output is 6 0 when V was sent2
P1output is 7 0 when
V was sent2
P1n0 7 V2 (5.50)
Thus the probability of an error occurring in the transmission of symbols reduces
to the simple condition that, at the instant the receiver output waveform is sampled, the
noise voltage at the receiver output is larger in the wrong direction than the sample value
of the signal. Since the noise is defined to be Gaussian, we can find the probability that
the noise exceeds a given value V volts. The probability that the sampled value of the
c05.qxd 17/07/2002 19:11 Page 192
P (vo )
Decision circuit output Decision circuit output
is “0” when vo < 0 volts is “1” when vo > 0 volts
ui = +1
“1” transmitted
n (t ) < (−V )
causes error
2 ∞e−u 2du
area =
πx∫
vo
0 V
Demod output voltage vo = n (t ) + Vui
P (vo )
ui = −1
“0” transmitted
n (t ) > (+V ) causes error
2 ∞e−u 2du
∫
πx
vo
−V 0
Demod output voltage vo = n (t ) + Vui
FIGURE 5.12 Illustration of errors in a binary decision circuit caused by additive
Gaussian noise. The threshold is at zero volts.
AWG noise voltage exceeds a value V is given by the integral of the PDF of the noise,
from V to infinity1.

P1n0 7 V 2 exp c d dl
s l2
(5.51)
12p V
2s2
where
is the rms noise voltage.
The integral in Eq. (5.51) cannot be solved analytically. Numerical or approximate so-
lutions must be used, and one such expression is known as the Q function, Q(z). An alter-
native form is the complementary error function, erfc(x), which is closely related to Q(z).
Fortunately, there are relatively simple approximate expressions available for these functions
when the probability of an error is small—which is usually the case for a workable digital
link, where we expect bit errors to occur infrequently. The condition for probability of er-
ror to be small is that V W
. The sampled signal must be much larger than the rms noise
at the receiver output. A value of V 3
makes the following approximation valid1.

exp c d dl 1 3z 12p4 exp1

z222
1 l2
Q1z2 (5.52)
12p z 2
The probability that n0 exceeds V at the sample instant is given by
P1n0 7 V2 Q3Vs4 (5.53)
c05.qxd 17/07/2002 19:11 Page 193
The complementary error function erfc(x) can also be used to find the probability
of an error with a Gaussian noise voltage. The probability that n0 exceeds V at the sample
instant is given by
P1n0 7 V2 12 erfc3V 1s 122 4 (5.54)
Appendix C gives the relationship of the complementary error function erfc(x)
to Q(z), and tables of values for both the complimentary error function and the Q
function.
Note that the Q function and erfc function are defined for a normalized rms value
of the Gaussian variable of one. When applied to symbol error analysis, the Q function
has an rms noise voltage of one volt. Errors occur whenever noise at the sample instant
exceeds either V or V, depending on the transmitted symbol. The threshold for an error
is therefore given by a signal to noise voltage ratio of V1 V.
The probability of error for each transmitted data state is given by the Q(z) or erfc(x)
functions in Eqs. (5.52) and (5.54). We should ensure that we send equal numbers of data 1
and data 0 states to make the probability of an error occurring in the 1 states the same as
the probability of an error occurring in the 0 states. This usually requires a randomizer or
scrambler to be inserted in the data stream at the transmitter to prevent the occurrence of
long strings of data 1s or 0s which would violate the required condition. The scrambler
also helps the symbol clock in the receiver to stay synchronized by providing frequent
phase transitions in the received signal.
One symbol lasts for Ts seconds. The power in the symbol waveform is V 22R where
R is the input resistance of the decision circuit. We will assume a resistance R 1 ohm,
as is commonly done in the analysis of communication signals. We will assume a con-
stant amplitude V for the carrier waveform (ignoring the effects of the Nyquist RRC filters
on pulse shape), so the energy per symbol, Es, is given by
Es 12 V 2 Ts joules (5.55)
Assuming that we have a matched filter receiver, the sampled signal voltage at the de-
modulator output is V volts where
V 112EsTs 2 volts (5.56)
The noise power at the demodulator output is N
2R
2 watts, relative to a resist-
ance of 1 ohm. The noise is assumed to be white and therefore has a constant NPSD, N0
watts/Hz, across the noise bandwidth Bn Hz of the receiver. In a receiver with ideal RRC
filters,
Bn 1 Ts.
Hence the single sided noise power spectral density is given by
N0 N Bn s2Bn s2 Ts joules (5.57)
Hence
s 11N0 Ts 2 volts rms (5.58)
Combining Eqs. (5.56) and (5.58) yields
V 1s122 2 32EsTs 12 TsN0 4 2 3EsN0 4 (5.58)

The probability that a symbol error occurs is therefore
Pe 12 erfc3 11Es N0 2 4 Q3 112Es N0 2 4 (5.59)

c05.qxd 17/07/2002 19:11 Page 194
BPSK Bit Error Rate

For BPSK a bit and a symbol are the same, so Eq. (5.59) can be written as
Pb 12 erfc c d Qc d
Eb 2Eb
(5.60)
A N0 A N0
The analysis in Chapter 4 provides methods by which the carrier-to-noise ratio in an
earth station receiver or satellite transponder can be calculated for any satellite link. The re-
sults are in terms of the ratio of carrier power to noise power at the input to a demodulator,
with the ratio CN usually given in decibels. We need to relate the CN ratio for a receiver
to the EsN0 ratio that provides us with a way to calculate the probability of a symbol error.
In a receiver with ideal RRC filters, regardless of the value of the roll-off factor ,
the noise bandwidth of the filter is equal to the symbol rate, which is the reciprocal of the
symbol period: Bn Rs 1Ts, or
BnTs 1 (5.61)
A result from matched filter theory states that the energy per symbol is the carrier power
times the symbol duration if the transfer function of the receiving filter matches the
spectrum of the received signal. A correctly designed digital radio link with RRC filters
meets this criterion, so we have
Es C Ts joules (5.62)
and the single sided noise power spectral density N0 W/Hz is just the noise power N watts
divided by the noise bandwidth Bn in hertz
N0 NBn W/Hz (5.63)
Hence for the ideal conditions specified above where Ts Bn 1
EsN0 CTsBnN CN (5.64)
Applying the result of Eq. (5.64) we find that the bit error rate for a BPSK signal
in an ideal RRC filtered link is
Pe BPSK 12 erfc c d Qc d
C 2C
(5.65)
AN AN
Note that the CN value used in Eq. (5.65) is a linear power ratio, not a decibel value.
Using decibel CN ratios in BER equations is a frequent source of error for beginning
communications engineers.
Since coherent detection is the most efficient way of demodulating direct BPSK,
Eq. (5.65) is the relation normally used to determine EbN0, and hence the CN ratio that
a satellite link must maintain to meet a specified bit error rate requirement.
QPSK Bit Error Rate

QPSK is simply two BPSK links operating on the same radio channel with their carriers
in phase quadrature. The BER for each BPSK link is identical, and given by Eq. (5.65).
When the bit stream at the transmitter is split into two to drive the I and Q channel of a
QPSK transmitter, the symbol rate on the link is halved. But the error rate remains the
same as if the signal had been sent as a BPSK transmission at twice the symbol rate, with
the same transmitter power. This is because BER is probability of error per bit, and the
probability of a bit error is the same for all bits in a QPSK system, regardless of which
c05.qxd 17/07/2002 19:11 Page 195
channel (I or Q) they travel through. So QPSK ends up with the same BER as BPSK when
considered in terms of EbN0.
The total energy per symbol of a QPSK signal is therefore twice that of either of
the constituent BPSK signals, or a single BPSK signal sent over the same link with the
same EbN0 ratio. Hence
Es QPSK 2 Eb BPSK (5.66)
and therefore to obtain the same error rate for a QPSK signal that we can achieve with a
BPSK signal in an RF channel with a noise bandwidth Bn Hz, we require
1CN2 QPSK 2 1C N2 BPSK (5.67)
Thus for QPSK, transmitted at a rate Rb bits/second in a channel with noise band-
width Bn Hz
Pe QPSK 12 erfc c d Qc d Qc d
Eb 2Eb C
(5.68)
A N0 A N0 AN
The analysis of the system performance of a radio link using BPSK is always car-
ried out in terms of CN ratio, not EbN0. Almost all communication system textbooks
leave the BER results for radio links in terms of EbN0, and state that the BER perform-
ance for BPSK and QPSK are the same. However, when BER is considered as a function
of CN ratio, BPSK and QPSK do not have the same BER. It takes twice as much trans-
mitter power to deliver two BPSK data streams as to deliver one. Therefore, QPSK re-
quires 3 dB more CN ratio than BPSK to achieve the same error rate when transmitting
at twice the bit rate in the same channel bandwidth. If a link is operated with QPSK, the
CN required for a given error rate is 3 dB higher than when the same link is operated
with BPSK. The advantage of QPSK is that it can send twice as many bits per second rel-
ative to BPSK using a channel with a specified bandwidth. This advantage of QPSK can
only be exploited if the CN at the receiver is sufficiently high.
EXAMPLE 5.4.1
A satellite link achieves a CN ratio in the receiver under clear air conditions of 14.0 dB. (14.0 dB
power ratio of 25.) The receiver has a RRC filter with a noise bandwidth of 1.0 MHz and a roll-off
factor of 0.3, with ideal correlation detection BPSK and QPSK demodulators.
What are the bit rate, symbol rate, occupied (absolute) bandwidth of the link, and BER when
the link is operated:
(i) with BPSK modulation and
(ii) with QPSK modulation?
If rain attenuation on the link causes the received signal to be attenuated by 3 dB, what
are the new BER values for BPSK and QPSK modulations? Assume that ideal RRC filters are
used.
For all radio links using band-pass RRC filters, the symbol rate is equal to the noise band-
width of the RRC filter. Thus, for both BPSK and QPSK:
The symbol rate for this link is
Rs Bn 1 Msps
The occupied bandwidth of the RF signal is
Bocc Rs 11 r2 1.3 Rs 1.3 MHz
Probability of error can be found from Eqs. (5.65) and (5.68) using the tables in Appendix C.
c05.qxd 17/07/2002 19:11 Page 196
(i) BPSK
The bit rate Rb symbol rate Rs 1 Mbps.
BER in clear air Pe BPSK Q3 12CN2 4 Q1 112 252 4 Q17.072 7.8 10
11
(ii) QPSK
The bit rate Rb 2 symbol rate Rs 2 Mbps.
BER in clear air Pe QPSK Q 3 1CN2 4 Q1 1252 Q152 2.8 10
7
When rain attenuation reduces the received signal by 3 dB, the receiver CN 11 dB. (Power ratio
12.59.) The resulting BER values are:
(i) BPSK
BER in rain Pe BPSK Q3 112CN2 4 Q1 12 12.592 Q15.022 2.8 10
7
(ii) QPSK
BER in rain Pe QPSK Q3 11CN2 4 Q1 112.592 Q13.552 2.2 10
4
All the BERs are acceptable except the last value for QPSK. With a bit rate of 2 Mbps, and
a BER of 2.2 10
4 there are hundreds of errors occurring every second. Forward error correc-
tion would be needed in the QPSK link to maintain an acceptable BER.
Using QPSK rather than BPSK over a link with a noise bandwidth defined by the RRC fil-
ter in the receiver doubles the bit rate. However, the CN ratio must be higher to sustain an ac-
ceptable error rate. A decision on whether to implement BPSK or QPSK on a given link will rest
on the CN values that can be maintained, and the length of time for which the CN ratio might
fall to levels at which an unacceptably high BER results.
Figure 5.13 shows bit error rate for an ideal system with RRC filters having a fixed
bandwidth B Hz, carrying BPSK and QPSK signals. The QPSK system carries twice as
10−2
QPSK
Probability of bit error (BER)
10−4
NCFSK
DBPSK
BPSK
10−6
10−8
0 5 10 15
Receiver C/N in dB
FIGURE 5.13 Bit error rate as a function of CN for a link with ideal RRC filters and no
intersymbol interference or timing jitter. The curves shown are for BPSK and QPSK with
coherent detection (BPSK, QPSK), differential BPSK (DBPSK), and noncoherent FSK (NCFSK).
The implementation margin in each case is 0 dB.
c05.qxd 17/07/2002 19:11 Page 197
much information (twice the number of bits carried by the BPSK system) but needs that
extra 3 dB of CN ratio to achieve the same BER as BPSK. In a practical BPSK or QPSK
radio link, which must have real filters, and inevitably suffers phase jitter in the carrier
recovery circuit and timing jitter in the bit clock when the CN ratio is low, the ideal
results shown in Figure 5.13 cannot be achieved.
An implementation margin must be added to the CN ratio to account for the dif-
ference between a real system and the ideal system for which the results of Figure 5.13
apply. In low bit rate systems, such as SCPC channels in VSAT systems and LEO mobile
satellite links, implementation margins as low as 0.5 dB have been reported in the litera-
ture. For high bit rate systems carrying multi-megabit per second QPSK data streams, im-
plementation margins as high as 2 dB may be required. Hence BER for practical BPSK
and QPSK satellite links is calculated from the following relationships
1C N2 eff 1CN2 0
Implementation margin dB (5.69)
1C N2 eff ratio 101CN2 eff10 as a ratio (5.70)
BERBPSK 12 erfc c d Qc d
C 2C
(5.71)
A N eff ratio A N eff ratio
BERQPSK 12 erfc c d Qc d
C C
(5.72)
A 2N eff ratio A N eff ratio
EXAMPLE 5.4.2
A satellite link uses a bandwidth of 10 MHz in a 52 MHz wide Ku-band transponder. The transmitter
and receiver have RRC band-pass filters with roll-off factor 0.25. The overall (CN)0 ratio for
the carrier in the receiver is 16.0 dB in clear air, falling below 13.0 dB for 0.1% of an average year.
The transmitter and the receiver have both BPSK and QPSK modulators and demodulators. The im-
plementation margin for the BPSK demodulator is 0.8 dB and for the QPSK demodulator is 1.2 dB.
Determine the bit rate that can be sent through the link with BPSK, and with QPSK. Find
the bit error rate for each modulation in clear air conditions and for the 0.1% of the year conditions.
Which modulation would you recommend for this system?
The symbol rate for the link is 101.25 8.0 Msps.
With BPSK the bit rate equals the symbol rate, so Rb BPSK Rs 8.0 Mbps.
With QPSK the bit rate equals twice the symbol rate, so Rb QPSK 2Rs 16.0 Mbps.
For the link using BPSK, the BER is found from Eq. (5.71),
BERBPSK 12 erfc3 11CN2 eff ratio 4 Q3 112CN2 eff ratio 4
In clear air (CN)eff 16.0
0.8 15.2 dB, hence (CN)eff ratio 101.52 33.1 and
BERBPSK 12 erfc3 133.1 4 12 erfc3 5.754
Using the erfc(x) table in Appendix C and interpolating between the values of erfc(x) for x 5.7
and x 5.8, the BER can be estimated as
BERBPSK 12 5 10
15 2.5 10
15 0
With BPSK, the link delivers 8 106 bits per second. With a BER of 2.5 10
15 a bit error will
occur, on average, once every 1(8 106 2.5 10
15) 5 107 s. This is a time of about
112 years, so in reality there will be no bit errors on this link. Anytime that x 5.5 in the erfc(x)
expression [or z 8 in the Q(z) function] the bit error rate on the link is effectively zero.
For 0.1% of the year (CN)eff 13.0
0.8 12.2 dB, (CN)eff ratio 101.22 16.60, and
BERBPSK 12 erfc3 116.6 4 12 erfc3 4.074
Using the erfc(x) table in Appendix C and interpolating between the values of erfc(x) for x 4.0
and x 4.1, the BER will exceed 5 10
9 for 0.1% of an average year when BPSK is used as
c05.qxd 17/07/2002 19:11 Page 198
the modulation on this link. At a bit rate of 8 Mbps, a bit error occurs, on average, every 25 s with
a BER of 5 10
9.
When the modulation on the link is changed to QPSK, the bit error rate will increase, as
indicated by Eq. (5.72)
BERQPSK 12 erfc3 11C 2N2 eff ratio 4 Q3 11CN2 eff ratio 4
In clear air (CN)eff 16.0
1.2 14.8 dB, and (CN)eff ratio 101.48 30.20
BERQPSK 12 erfc3 1130.2 22 12 erfc33.894
Using the erfc(x) table in Appendix C, the BER can be estimated as
BERQPSK 12 4 10
8 2 10
8
With QPSK, the link delivers 16 106 bits per second. With CN 16.0 dB there is a bit error,
on average, once every 1(16 106 2 10
8) 3.12 s.
For 0.1% of the year (CN)eff 13.0
1.2 11.8 dB and (CN)eff ratio 101.18 15.14
BERQPSK 12 erfc3 1115.1422 4 12 erfc32.75 4
Using the erfc(x) table in Appendix C, BER will exceed 5 10
5 for 0.1% of an average year when
QPSK is used as the modulation on this link. At a bit rate of 16 Mbps, there are 800 bit errors every
second, on average, when CN 13.0 dB.
What is an acceptable bit error rate depends on the particular application. For financial trans-
actions, a BER of 10
12 is typically required. Satellite systems don’t often guarantee such a low er-
ror rate, so some form of error detection is needed, with retransmission of any data that are found
to be in error. For general applications bit error rates of 10
8 to 10
6 are acceptable. Digital voice
transmission can withstand occasional error rates as high as 10
4, which typically lead to baseband
SN ratios of 34 dB.
In this example, BPSK modulation has BER 10
8 for 99.9% of an average year, and would
therefore be a satisfactory choice. QPSK can deliver an acceptable error rate in clear air conditions,
2 10
8, but when the CN starts to fall the BER increases quickly, so that with 3-dB attenuation
in the link the BER is 5 10
5. This is not adequate for general applications, but would suffice
for digital voice links with a requirement that SN 30 dB for 99.9% of the year. The obvious ad-
vantage of using QPSK is that twice as many voice channels can be carried by a 16-Mbps bit stream
modulated with QPSK, compared to BPSK modulation which can carry only 8 Mbps, within the
available channel bandwidth (Bocc) of 10 MHz.
Generation of Quadrature Phase Shift Keying

(QPSK) Signals
In QPSK the phase, , of the carrier is set by the modulator to one of four possible val-
ues. We may write the result as
v1t2 V121cos vct
f2 (5.73)
where takes on the values 4, 34, 54, and 74 rad. The factor 12 is for our
later convenience. Using trigonometric identities to expand Eq. (5.73) we obtain
v1t2 V12 cos vct cos f V12 sin vct sin f (5.74)
The first term is a BPSK signal in phase with the carrier, and is called the I channel. The
second term is a BPSK signal in quadrature with the carrier and is called the Q channel.
Thus a QPSK waveform may be generated by combining two BPSK waveforms in quad-
rature. We may write the result as
v1t2 uiV cos vct uqV sin vct (5.75)
where ui, represents a binary data stream modulating the I channel and uq represents a bi-
nary data stream modulating the Q channel. For both of these signals a logical 1 corresponds
c05.qxd 17/07/2002 19:11 Page 199
TABLE 5.1 The Relationship

between the Modulating Bit
Streams ui, uq, and the Phase
Angle of the Modulated
QPSK Waveform
ui uq
1 1 4

1 1 34

1
1 54
1
1 74
to ui or uq 1, and a logical 0 corresponds to ui or uq

1. The relationship between
ui , uq, and is given by
ui 12 cos f (5.76)
uq 12 sin f (5.77)
and is summarized in Table 5.1. Note that is conveniently visualized as the phase an-
gle of a phasor whose in-phase component is ui, and whose quadrature component is uq.
See Figure 5.14.
The bits ui and uq are selected alternately from the input bit stream. For example,
ui may represent the odd-number bits and uq the even. In this case one binary data channel
enters the QPSK modulator and the outgoing symbol rate is equal to half of the incoming
bit rate.
QPSK modulators and demodulators are basically dual-channel BPSK modulators
and demodulators. One channel processes the ui bits and uses the reference carrier; the
other processes the uq bits and uses a 90° phase shifted version of the reference. Figures
5.15 and 5.16 show generalized block diagrams of a QPSK modulator and demodulator.
More detailed information is available in reference 13.
QPSK Variants
We noted that QPSK may be visualized as the sum of two BPSK signals whose carri-
ers are in phase quadrature. In conventional QPSK the bits ui and uq that modulate
these carriers both make step changes at the same time. If the bit changes are stag-
gered so that ui makes step changes at the beginning of each symbol period and uq
makes step changes at the midpoint of each symbol period, the result is called stagg-
uq
−1, 1 1, 1
φ = 3π /4 φ = π /4
ui
FIGURE 5.14 Phasor diagram showing the phase of a

−1, −1 −1, 1 QPSK waveform modulated with all possible pair combi-
φ = 5π /4 φ = 7π /4 nations of the bits (ui, uq). Each phasor is formed by
summing a ui and a uq signal.
c05.qxd 17/07/2002 19:11 Page 200
I channel
Multiplier
LPF
NRZ data
input
Bit Σ
select
QPSK output
LPF
Multiplier
Q channel
90° phase
shifter π /2
IF oscillator
FIGURE 5.15 QPSK modulator. The bit select block sends alternate bits to
the I andQ channels.
ered QPSK (SQPSK) or offset QPSK (OQPSK). If instead of steps the bits make si-
nusoidal transitions between their allowed values of 1 and
1, the result is minimum
shift keying (MSK) or fast frequency shift keying (FFSK). These modulation systems
produce spectra that are slightly different from conventional QPSK and that would
appear to have some advantages over QPSK for satellite transmission. While they have
received considerable academic attention, OQPSK, SQPSK, MSK, and FFSK have not
been widely adopted for commercial GEO satellite applications. The prevailing atti-
tude in the industry seems to be that any theoretical advantages that they might have
over conventional QPSK either vanish when these techniques are used over a real
transponder or else are so slight as not to justify the added expense that their
I channel
Multipliers
IF LPF
QPSK
input
splitter logic
to
sampler
LPF
Carrier Q channel
recovery
π /2 90° phase
shifter
IF oscillator
FIGURE 5.16 QPSK demodulator. The logic block outputs two bits
for each received QPSK symbol.
c05.qxd 17/07/2002 19:11 Page 201
5.5 DIGITAL TRANSMISSION OF ANALOG SIGNALS 201
implementation would require. For additional information the reader should consult
references 2, 13, and 14.
5.5 DIGITAL TRANSMISSION OF

ANALOG SIGNALS
The previous sections have discussed techniques for transmitting and receiving digital in-
formation via satellite. Now we will turn our attention to the problem of putting analog voice
and television signals into digital form for transmission, and recovering the analog signals
after reception. While the material presented is generally applicable to all analog signals,
we will emphasize baseband voice channels because virtually all telephone channels are now
digital. Digital television transmission is covered in Chapter 11.
Sampling and Quantizing

The basic processes in digital transmission of analog information are sampling, quantiz-
ing, and encoding. The principles underlying sampling are routinely presented in beginning
courses in communications theory, and we will not reproduce them here. See references 1
and 3 for details. The sampling theorem states that a signal may be reconstructed without
error from regularly spaced samples taken at a rate fs samples/second, which is at least
twice the maximum frequency fmax present in the signal. Instead of transmitting the con-
tinuous analog signal, we may transmit the samples. For example, voice signals on satel-
lite links are normally filtered at baseband to limit their spectra to the range 300 to 3400
Hz. Thus, one voice channel could be transmitted with samples taken at least 6800 times
per second or, as it is usually expressed, with a minimum sampling frequency of 6800 Hz.
Common telephone system practice is to use a sampling frequency of 8000 Hz. While trans-
mitting the samples requires more bandwidth than transmitting the original waveform, the
time between samples of one signal may be used to transmit samples of other signals. This
is time division multiplexing (TDM), and we will discuss it later in this chapter.
The samples to which the sampling theorem refers are analog pulses whose ampli-
tudes are equal to that of the original waveform at the time of sampling. The original
waveform may be reconstructed without error by passing the samples through an ideal
low pass filter whose transfer function is appropriate to the sampling pulse shape. A com-
munications system that samples an input waveform and transmits analog pulses is said
to use pulse amplitude modulation (PAM). Figure 5.17 sketches this process.
Analog pulses are subject to amplitude distortion, and they are also incompatible with
conventional baseband digital signals in which pulses take on only one of two possible val-
ues. Hence pulse amplitude modulation is not used in communication links. Instead, the ana-
log samples are quantized—resolved into one of a finite number of possible values—and
the quantized values are binarily encoded and transmitted digitally. Thus each sample is con-
verted into a digital word that represents the quantization value closest to the original ana-
log sample. Quantization may be uniform or nonuniform depending on whether or not the
quantized voltage levels are uniformly or nonuniformly spaced. At the receiver a digital-to-
analog (DA) converter converts each incoming digital word back into an analog sample;
these analog samples are filtered and the original input waveform is reconstructed.
A communications system that transmits digitally encoded quantized values is called
a pulse code modulation (PCM) system, a rather antiquated name for a process that is not
really a modulation or a code. When the PCM process is applied to voice and TV signals,
they are often referred to simply as digital voice and digital TV.
c05.qxd 17/07/2002 19:11 Page 202
v v v
Input
signal Samples Transmitted pulses
t t t
(a) Basic waveforms
PAM
Input signal Recovered
BPF Analog LPF
speech speech
300-3400 Hz gate 0-3400 Hz
signal signal
PCM
A/D signal D/A
converter converter
Sampling
pulse
generator
(b)
FIGURE 5.17 Pulse amplitude modulation (PAM). (a) Basic waveforms. (b) Block diagram
of a PAM communications system. Dashed lines show the additional components that
would convert it to a PCM system.
The standard word used in digital telephone systems has 8 bits, so with sampling
at 8 kHz, the bit rate of a digital telephone (PCM) channel is 64 kbps. Twenty-four tele-
phone channels are often transmitted on a single twisted pair telephone line using a sys-
tem known as T1. The T1 system (described later in this chapter) creates a frame of 24
PCM words, one from each of the 24 voice channels, and adds a single framing bit to
mark the end of the frame. This gives a 193-bit frame, and frames are transmitted at the
rate of 8000 frames per second. The resulting bit rate is 1.544 Mbps, which has become
a digital transmission standard in North America, regardless of whether digital voice or
digital data are being transmitted.
The quantization process, illustrated in Figure 5.18, prevents exact reconstruction
of the digitized waveform. (The sampling theorem requires that analog rather than quan-
tized samples be transmitted.) The error introduced is called quantization error; and a per-
son listening to a reconstructed speech signal perceives the quantization error as an added
noise called quantization noise.
A uniform quantizer (Figure 5.19) operates with L levels spaced A volts apart. The
input signal is amplitude limited to lie between
(L2) and (L2). The quantizer de-
termines in which level an incoming sample falls and puts out the identification number
of that level. This identification number is the digital word that represents the sample.
Transmitting L levels requires N bits where
N log2 L (5.78)
or
L 2N (5.79)
The levels are normally numbered 0 through L 1. Thus an 8-bit PCM (N 8) system
would quantize its incoming samples into one of 256 (L 2N 256) levels numbered
0 through 255. Samples of the analog signal are transmitted as binary words ranging
c05.qxd 17/07/2002 19:11 Page 203
v v
t t
(a) (b )
t
(c )
FIGURE 5.18 The quantizing process. (a) The input waveform and the quanti-
zation levels. (b) Quantized samples. (c) Quantized pulses. The pulse amplitude
is encoded digitally for PCM transmission.
v
4∆
Level 7 Signals with 3∆ < v ≤ 4∆ are transmitted as 3.5∆, encoded as
111
3∆
Level 6 Signals with 2∆ < v ≤ 3∆ are transmitted as 2.5∆, encoded as
110
2∆
Level 5 Signals with ∆ < v ≤ 2∆ are transmitted as 1.5∆, encoded as
101
∆
Level 4 Signals with 0 < v ≤ ∆ are transmitted as 0.5∆, encoded as 100
0
Level 3 Signals with −∆ < v ≤ 0 are transmitted as −0.5∆, encoded as
011
−∆
Level 2 Signals with −2∆ < v ≤ ∆ are transmitted as −1.5∆, encoded as
010
−2∆
Level 1 Signals with −3∆ < v ≤ −2∆ are transmitted as −2.5∆, encoded
as 001
−3∆
Level 0 Signals with −4∆ ≤ v ≤ −3∆ are transmitted as −3.5∆, encoded
as 000
−4∆
FIGURE 5.19 Levels and encoding for a uniform 3-bit (8 level) PCM quantizer.
c05.qxd 17/07/2002 19:11 Page 204
from 00000000 (decimal 0) through 11111111 (decimal 255). If the input signal ampli-
tude is uniformly distributed with an rms value of Vrms, the signal-to-noise ratio of the re-
constructed analog signal (assuming that only quantization noise is present) is given by1
1SN2 121Vrms ¢ 2 2 (5.80)
For uniform quantization and a signal input that has equal probability of any volt-
age level, the quantization noise added to the recovered analog signal gives a baseband
signal to noise ratio of (SN)Q where
1SN2 Q 6N dB (5.81)
Thus a standard digital telephone channel using an 8-bit word and uniform quanti-
zation will have an average quantization SN ratio of 48 dB, using linear quantization.
Nonuniform Quantization: Compression

and Expansion
Uniform quantization introduces more noise when a signal is small and one quantization
interval is large in comparison with the signal than it does when the signal is large and
one quantization interval is insignificant. The problem is most apparent with the quiet
talker. The quiet talker produces a voice signal with 30 dB lower power than the design
level of the telephone system. As a voltage ratio relative to 1 V,
30 dB is 0.0314 V. Thus
a telephone system in which the nominal power level is 0 dBm, and the impedance is
600 ohms (the standard values) has a nominal voltage range of 0.775 V. With an 8-bit
word and 255 quantization levels, the step size is 2 0.775255 6.1 mV. The quiet
talker produces a voltage level of 31.4 Vrms, with an equivalent peak sine wave level of
44.4 mV. Thus the quiet talker uses only the lowest 15 steps of the digital quantizer,
equivalent to using a 4-bit quantizer. A 4-bit quantizer gives a quantization baseband SN
ratio of 24 dB, so the quiet talker is producing signals that have, at best, a SN ratio of
24 dB rather than 48 dB.
Improved noise performance can be obtained using nonuniform quantization in
which the size of the quantization intervals increases in proportion to the signal value be-
ing quantized. The same effect can be obtained from a uniform quantizer if the input sig-
nal is compressed before quantization. The distortion introduced by the compressor must
be removed at the receiver by an expander. The transfer functions of the compressor and
expander are complementary, that is, their product is a constant and the amplitude distri-
bution of a signal that has passed through both a compressor and an expander is unchanged.
Compression at the transmitter followed by expanding at the receiver is called companding.
Companding was first employed on terrestrial telephone systems using analog com-
pressors that had logarithmic transfer functions. These were the so-called -law and A-law
compressors. Later developments in digital technology allowed digital implementation of
the compression and expansion functions and permitted the sampling, compression, quan-
tization, and encoding operations to be combined into one piece of equipment called a
coder. Companding with 8-bit digital voice channels leads to a subjective improvement in
baseband SN ratio which is typically taken to be 15 dB. Thus a companded digital
telephone link gives a baseband quantization SN ratio of 63 dB in the absence of bit
errors. Figure 5.20 shows the compression characteristic of a typical analog compression
circuit. The entire process of converting an analog voice signal to a 64-kbps digital bit
stream, and converting a 64 kbps digital voice channel back to an analog voice signal is
now done in a single integrated circuit. The telephone wire from a telephone subscriber
(the twisted pair of the last mile) is taken to a digitizing IC as close to the customer’s
c05.qxd 17/07/2002 19:11 Page 205
Vout Vout
Vin Vin
Compressor Expander
FIGURE 5.20 The transfer characteristic of a compressor and an
expander.
premises as possible. The digital side of the IC connects to a four wire circuit which has
separate go and return pairs. Telephone exchanges (now called switches) are digital com-
puters that cannot handle analog signals. All telephone voice signals must be converted
to digital form before they can be handled by a switch, so the conversion takes place close
to the customer.
The companding process improves the perceived SN ratio in the baseband chan-
nel for the quiet talker by increasing the number of steps in the quantizer at small signal
levels. However, with a fixed number of steps (typically 255 in an 8-bit system) the steps
must be larger for large signals. This increases the quantization noise that is present with
large signals, and therefore lowers the SN ratio for the loud talker. The effect of com-
panding is therefore to even out the impact of quantization noise over the dynamic range
of the baseband signal. When baseband SN ratio is calculated from signal and noise pow-
ers taking companding into account, the SN ratio is relatively constant with signal level
across the whole baseband. However, this is not what the listener perceives. When pre-
sented to a human ear, loud sounds appear to mask the increase in quantization noise at
high signal levels, and the perceived SN is much better than the calculated values would
indicate. With an equiprobable signal reaching the highest permitted signal level, the base-
band SN of a companded PCM speech channel appears to be about 15 dB higher than
the calculated value of 48 dB. This is called a subjective improvement in SN ratio, because
the effect depends on the physiological characteristics of the human ear, not on calculations
of signal power and noise power.
The compact disc (CD) used for sound recordings is another example of a digital
audio system. The CD is intended to reproduce music with high quality and therefore re-
quires a much better quantization SN ratio in the analog sound output than a telephone
channel. When a CD is recorded, a 16-bit linear quantizer is used, giving a baseband quan-
tization SN of 96 dB. The dynamic range of the human ear is about 120 dB, but most
sound reproduction systems have thermal noise SN ratios of less than 100 dB, so the
quantization noise in a CD is inaudible. Linear encoding is more accurate than com-
panding, where a match is needed between the compressor and the expander. Consequently
companding is not used when high quality sound reproduction is required. CDs are
recorded in stereo using 44 kHz sampling and 16-bit words, giving a bit rate of 1.408 Mbps.
The actual bit rate of the recorded material on the CD is much higher because error de-
tection and correction coding is applied to the digital data stream before it is recorded.
c05.qxd 17/07/2002 19:11 Page 206
Digital telephone signals are transmitted at 64 kbps or lower rates, because this is suffi-
cient to achieve speaker recognition and intelligible speech. There is no attempt to make
a telephone channel a hi-fi sound system because this would require the transmission of
a much higher bit rate, and therefore fewer channels per megabit.
Signal-to-Noise Ratio in Digital Voice Systems

Thermal noise causes bit errors in digital communication links, as discussed in Section
5.4. In a PCM system, the digital data are converted back to a baseband analog signal at
the receiver. We need to know the signal-to-noise ratio that corresponds to a given prob-
ability of a bit error occurring in the digital data at the receiver. The analysis is straight-
forward when only one bit error occurs in each PCM word; provided the BER is below
10
4 and we have 7 or 8 bits per word, the likelihood of two bit errors occurring in one
word is very small. We will assume this to be the case in the analysis that follows.
When a bit is in error in a PCM word, the recovered sample of the baseband ana-
log signal will be at the wrong level. This adds an impulse of amplitude Vn and duration
Ts, the period of one sample, to the true analog signal. The bit that is in error may be lo-
cated in any position in the PCM word. If the least significant bit is in error, Vn is small
and equal to , the analog to digital converter step size; if it is the most significant bit
that is in error, Vn will be large and equal to 2N
1 .
The resulting average signal-to-noise ratio in the baseband analog channel caused
by random bit errors is (SN)t, the subscript t standing for thermal noise, which is assumed
to be the cause of the random bit errors, where1
1SN2 t
3L2
(5.82)
1 41L2
12Pb
We can combine thermal noise from Eq. (5.82) with quantization noise from Eq.
(5.81) to find the overall PCM output signal-to-noise ratio
1SN2 PCM
22 N
(5.83)
1 4 Pb 22 N
When Pb is small, for example, less than 10
4, the quantization noise will dominate
and (SN)PCM 22N. For N 8 bits, this gives SN 48 dB. When Pb is larger, thermal
noise dominates; for example, with Pb 10
4 and N 8 and (SN) 14Pb 34 dB.
Figure 5.21 shows the transition from thermal noise to quantization noise as the predom-
inant noise source as the probability of a bit error decreases, for PCM systems using 7
and 8 bit words, with linear quantization. Clearly, for BERs below 10
6, quantization noise
is dominant. However, when the bit error rate is around 10
5 there is a bit error roughly
once per second in a 64-kbps PCM voice link. The bit error will be heard as a click, so
the calculation of the thermal noise SN is not meaningful. Since most PCM links oper-
ate with BERs below 10
6 most of the time, it is worthwhile using nonlinear encoding
(companding) to improve the baseband (SN) by reducing quantization noise.
The curves for baseband SN in a digital voice channel shown in Figure 5.21 by
the solid lines are for a system with linear encoding. When nonlinear encoding is used,
higher perceived baseband SN ratios are achieved as shown by the dotted lines provided
that thermal noise is not present. As soon as the probability of bit errors exceeds 10
6,
thermal noise becomes the dominant source of noise in the channel and the subjective im-
provement obtained by companding disappears. Figure 5.21 is plotted on a baseline of
BER. Comparison with Figure 5.13 in Section 5.4 shows that BER rises very quickly
when the CN of a digital radio system falls.
c05.qxd 17/07/2002 19:11 Page 207
60
With companding
Baseband S/N ratio (dB)
50
Without companding
40
FIGURE 5.21 Baseband

SN in digital speech system
30 using 8-bit PCM. Signal values
10−4 10−5 10−6 10−7 10−8
between zero and maximum
Bit error rate are equally probable.
For example, a QPSK satellite link with an overall receiver (CN)0 17 dB and an
implementation margin of 1 dB gives a BER of 10
10. In a speech channel using 64 kbps
bit rate, a BER of 10
10 gives one bit error every 2 days; the channel is essentially error
free and quantization noise will set the channel SN ratio at a subjective value of 63 dB
with -law or A-law companding. If the receiver CN ratio falls by 3 dB, due to rain at-
tenuation in the path, for example, the BER will increase to 4 10
6, giving a thermal
noise SN ratio 14Pb 48 dB. This is much lower than the companded quantization
SN ratio of 63 dB. The link therefore operates in one of two regimes. For about 99% of
the time on a typical satellite communication link, when clear air conditions prevail, there
are no bit errors and the baseband SN ratio will be 63 dB. For the remaining 1% of the
time when attenuation occurs on the link, the baseband SN will be below 63 dB, but will
only fall below 40 dB for a very small fraction of the time when the BER exceeds 4
10
4. The techniques described in Chapter 4 are used to design satellite communication
links so that bit error rates can be maintained above 4 10
4 for all but a very small
percentage of the time.
Mobile satellite systems are based on narrow bandwidth channels, and send voice
signals at bit rates as low as 2.4 kbps. Typical bit rates for satellite telephone channels are
around 4.8 kbps14. The reduction in bit rate from a standard 64-kbps digital voice chan-
nel is achieved using speech compression techniques. Analysis of speech waveforms shows
that vowel sounds tend to have frequencies that are usually below 500 Hz in speech, and
last for 50 ms or longer. Consonant have a much wider spectral content but last for shorter
periods. There are also gaps between words and sentences. This allows considerable room
to compress speech by transmitting what amounts to a spectral analysis of the waveform,
rather than the waveform itself. At the simplest level, for example, a 50 ms vowel sound
at a fundamental frequency of 200 Hz can be described as a sine wave at 200 Hz with a
specific amplitude, with a number of harmonics of specified amplitude, lasting for a duration
of T milliseconds. This information can be sent as 10 words of 8 bits giving a total of
80 bits, instead of 400 words of 8 bits that result from 8-kHz sampling of the waveform.
At the receiving end of the link, a vocoder reconstructs the signal from the information
sent over the link.
The above example is oversimplified, but it illustrates that a considerable reduction
in the transmission rate of digital speech signals is possible. The most widely used speech
compression technique is linear predictive encoding (LPE). Speech compression is
c05.qxd 17/07/2002 19:11 Page 208
achieved by the use of digital signal processing (DSP) integrated circuits. The DSP IC
contains a fast microprocessor and a large memory, and may be executing many millions
of operations a second. The analog speech waveform is sampled and digitized, typically
at 64 kbps for telephony, and then processed by the DSP compression IC to reduce the
bit rate. The low bit rate signal is transmitted to a decoding IC in the receiver which re-
generates the 64 kbps bit stream and thus recovers the original speech waveform. The
challenge in developing speech compression systems is to provide a natural sounding voice
at the receive end of the link which works well in any language. Once the compression
technique is reduced to a pair of ICs, albeit very complex and fast devices, the integrated
circuits can be manufactured in large quantities at reasonable prices. Systems designers
can simply insert a pair of DSP devices in the transmitter and receiver to reduce the bit
rate on the radio link.
Digital Television
Digital television is rapidly replacing analog TV, with direct broadcast satellite television,
digital cable TV, and high definition television (HDTV) all using digital transmission. A
television signal is made up of two separate parts: the video signal which creates the pic-
ture at the receiver, and the audio part that carries the accompanying sound signals. The
video and audio signals are digitized separately, and then multiplexed into a serial bit
stream for transmission. The audio signal is typically digitized with more bits than a tele-
phone channel to provide sound of good quality. The video signal bandwidth for a TV
signal is large: 4.2 MHz for the NTSC system and 5.25 MHz for the PAL (phase alter-
nate line) system. A NTSC video signal sampled at 10 MHz and encoded with 8-bit words
produces a bit stream at 80 Mbps. Transmission of this digital TV signal requires a band-
width of 50 MHz using QPSK and RRC filters with 0.25. This is a large RF bandwidth,
which would require the whole of one 54-MHz Ku-band transponder, making transmission
of full bandwidth digital TV very expensive.
To overcome the high bit rates required when an NTSC or PAL video signal is sam-
pled and digitized, compression techniques have been developed that reduce the bit rate
by a factor of 12 or more. The most important video compression techniques is known as
MPEG 2, where MPEG stands for the Motion Pictures Experts Group, an industry stan-
dards body. MPEG compression techniques are used in high definition television, digital
videodiscs, and direct broadcast satellite TV. The MPEG 2 system divides the picture into
8 by 8 pixel blocks and takes a discrete cosine transform (DCT) of each block. Only the
significant coefficients of the DCT are then transmitted, which greatly reduces the bit rate
to send each 8 8 pixel block. MPEG 2 also uses frame-to-frame comparison of the
video signal to determine which blocks within each frame need to be transmitted because
a change has occurred between frames. Compression factors of up to 75 can be achieved
with full motion video using MPEG 2 techniques. If degradation of the picture is allowed,
even higher compression ratios are possible, but definition is degraded and motion be-
comes blurred. Similar compression techniques are used for the transmission of pictures
over the Internet using JPEG (Joint Picture Experts Group) standards. The human eye is
less sensitive to noise and distortion in a video signal than the human ear is to distortions
in audio signals. This allows video and picture transmissions to use fewer bits and higher
quantization noise levels than audio transmissions. One major disadvantage of MPEG 2
compression is that it introduces a delay of about 1.5 s in the transmission of the TV sig-
nal. The delay is of no significance in recorded material, but is excessive for two-way live
transmissions.
c05.qxd 17/07/2002 19:11 Page 209
5.6 TIME DIVISION MULTIPLEXING 209
The digital video standard used in direct broadcast satellite television uses an aver-
age transmission rate of 6.2 Mbps for live video and 1.6 Mbps for prerecorded material.
This allows up to 10 video signals to be transmitted by a single 20 MHz bandwidth
transponder. For more details of direct broadcast satellite television see Chapter 11.
5.6 TIME DIVISION MULTIPLEXING
In time division multiplexing (TDM) a group of signals take turns using a channel. This
contrasts to frequency division multiplexing, presented earlier, where the signals occupy
the channel at the same time but on different frequencies. Since digital signals are pre-
cisely timed and consist of groups of short pulses with relatively long intervals between
them, TDM is the natural way for combining digital signals for transmission.
TDM Terminology: The U.S. T1

24-Channel System
In this section we will describe the U.S. Telephone T1 24-channel TDM system and use it
to introduce the terminology of time division multiplexing. While T1 was developed for ter-
restrial microwave circuits, it is also used on digital satellite links. We present it here as a
convenient vehicle for explaining TDM operation. In pure TDMA systems, the multiplexing
blends into the multiple access process.
A TDM system transmits a digital word from each channel in turn. Each word is a
group of bits that identify the quantization interval of the current sample. The words are
organized into frames. One frame contains one word from each channel plus some syn-
chronizing information that serves to identify the start of the frame. A frame is then a se-
ries of bits numbered sequentially from zero that carry synchronizing information plus the
quantized values of one sample from each channel. The bits within a frame are grouped
into slots. A slot contains all the bits from a common source. The slots within a frame are
numbered sequentially from zero. In the T1 system illustrated in Figure 5.22, there are 25
slots, numbered 0 through 24. Slot 0 contains a single bit and carries synchronization in-
formation. Slots 1 through 24 each contain 8 bits and carry telephone channels 1 through
24. Thus a T1 frame contains 1 (8 24) 193 bits. Standard digital telephone (PCM)
systems sample at an 8-kHz rate. Said another way, they transmit one sample of each
channel every 125 s. This is the frame interval; the frame rate is the reciprocal of the
frame interval and is always equal to the sampling rate, or an integer multiple of the sam-
pling rate. The T1 system must transmit 193 bits in the 125-s frame interval; hence its
bit rate is 193 bits divided by 125 s or 1.544 megabits per second (Mbps).
Frame synchronization is established and maintained by transmitting a known bit
pattern in slot 0. This bit pattern constitutes what is called the frame alignment word
(FAW ). When frames are transmitted sequentially, slot 0 of one frame is at the end of
the previous frame. The bits of the FAW are therefore both start of frame (SOF) and
Slots 0 1 2 3 4 24
1 bit 8 bits 8 bits 8 bits 8 bits 8 bits Time FIGURE 5.22 Slot
organization of one T1
193 bits
frame.
c05.qxd 17/07/2002 19:11 Page 210
end of frame (EOF) markers. There are several different FAWs that can be used in T1
systems. One T1 FAW is 100011011100; it contains 12 bits and requires 12 frames for
transmission. The group of frames that transmit the FAW make up a superframe. Thus
in the first frame of a superframe, slot 0 contains a 1. In the second frame slot 0 con-
tains a 0, and so on.
At this point let us summarize the operation of a T1 digital multiplexer. It receives
digitized voice signals from 24 telephone channels, which, for now, we will assume are
perfectly synchronized with each other and have exactly the same bit rate. The 8-bit word
samples from each channel flow into buffers and wait for the multiplexer to read them
out. The multiplexer reads them out and inserts them into outgoing frames as follows. The
first frame is transmitted by sending a 1(the first bit of the FAW), then one 8-bit word
from channel 1, then one 8-bit word from channel 2, and so forth through one 8-bit word
from channel 24. Then a new group of samples flows into the buffers. The multiplexer
forms frame 2 by sending a 0 (the second bit of the FAW) followed by the words from
each channel. A third group of samples enters and the process continues. When the buffers
have been filled and emptied 12 times, one superframe has been sent and the multiplexer
begins a new FAW.
At the receiving end of the link a demultiplexer must sort out the bits in each frame
and route the appropriate words to each outgoing channel. It must also keep track of the
number of the frame (within the superframe) that it is receiving. We may visualize the de-
multiplexing process by assuming that the incoming bits are clocked serially into a shift
register. At the instant the last bit has entered the register, its contents match the bits in
the frame of Figure 5.22. The multiplexer then does a parallel transfer of the bits for each
channel into their own individual registers for subsequent serial transmission over their
separate paths. At this point the digital channels have been demultiplexed.
The frame alignment bits go into a shift register, which, at the completion of the
superframe, should contain the frame alignment word. If it does not, then the multi-
plexer and demultiplexer are out of sync. When this occurs the demultiplexer seeks to re-
gain alignment; the process is called reframing. In it the demultiplexer looks at candidate
frame alignment bits until it finds one that is going through the requisite 100011011100
100011011100 100011011100 pattern. Obviously there is a trade-off between the num-
ber of frame alignment bits and the time required for refraining. If the entire FAW is trans-
mitted within each frame, then refraining time is much shorter than when the FAW is
transmitted with one bit per frame. Typical refraining times are about 50 ms,14 which is
sufficiently short not to cause significant degradation of speech in the 24 channels when
misalignment occurs.
Along with the information carried in the 24 channels must go the signaling infor-
mation necessary to route, initiate, and terminate the data channels. In the T1 system this
information is transmitted by “robbing” the least significant bit from slots 6 and 12 and
using these to form signaling channels A and B, respectively. Thus, channels 6 and 12 are
actually carried by a form of 7-bit PCM and channels A and B convey signaling infor-
mation at an 86 1.333 kbps rate.
The T1 system was developed by AT&T in the 1960s to expand the capacity of
twisted pair telephone cables between their telephone exchanges. A single twisted pair
could carry 24 voice channels using a T1 system instead of one baseband voice chan-
nel. This allowed a 24-fold expansion of channel capacity between exchanges without
the need to lay new cables. The development of many thousands of T1 links in the
1960s and 1970s provided AT&T with a capability to transmit any digital signal at 1.544
Mbps. As a result, 1.544 Mbps became a standard transmission rate for any digital sig-
nal, not just voice. When used for data, a 1.544 Mbps link is officially called a DS-1,
c05.qxd 17/07/2002 19:11 Page 211
5.6 TIME DIVISION MULTIPLEXING 211
TABLE 5.2 North American T-Carrier Digital

Transmission Standards
System
designation Bit rate (Mbps) Voice channels Digital signal
T1 1.544 24 DS-1
T1C 3.152 48 DS-1C
T1D 3.152 48 DS-1C
T1G 6.443 96 DS-2
T2 6.443 96 DS-2
T3 44.736 672 DS-3
T4 274.176 4,032 DS-4
T5 560.160 8,064 DS-5
rather than a T1 system. However, with the growth of the Internet and the widespread
use of high-speed digital links for Internet access, T1 and fractional T1 (12 T1 and 14 T1)
links are in common use.
The capacity of T1 and higher rate links used for digital speech transmission can
be increased by the use of a compression technique known as differential PCM (DPCM).
In DPCM the difference between adjacent 8-bit samples of the speech waveform are
transmitted, rather than the samples themselves. An improved form of DPCM, adaptive
differential PCM (ADPCM) can reduce the transmission rate for digital speech to 32 or
16 kbps. This can double or quadruple the capacity of a T1 link to 48 or 96 speech
channels17.
A DS-0 link runs at 64 kbps and there is a heirarchy of bit rates from DS-0 through
DS-5, at 560.160 Mbps which forms the North American digital transmission standard1.
Table 5.2 shows the T-system heirachy for voice and data circuits. Although the T-system
was developed for terrestrial circuits, it is often important for a satellite system to be able
to carry data between two terrestrial circuits, so it must be able to operate at the standard
T-n rates. There are very few satellite transponders that can carry a T-4 or T-5 signal at
274 or 560 Mbps.
Other TDM Systems

Outside the United States, most countries use digital transmission standards recommended
by the ITU-T (formerly CCITT), the international body that coordinates telephone system
standards. The ITU-T has recommended a standard 30 voice channel system with a bit
rate of 2.048 Mbps. The system has 32 channels running at 64 kbps each, with two channels
reserved for signaling and synchronization. For details on their slot and bit organization
the reader should consult reference [1].
The rapid development of optical fiber systems allowed very high-speed digital sig-
nals to be transmitted over long distances. Optical fibers have very large bandwidths com-
pared to radio links, and transmission has been demonstrated at rates exceeding 10 Gbps.
The most widely used transmission rate seems to be OC-48, which transmits digital data
at 2488 Mbps. Since optical fiber cables never contain a single fiber, but more usually
several hundred fibers, the capacity of an optical cable system is very large. The largest
and heaviest GEO satellites in orbit in 2000 could carry less than the equivalent of four
OC-28 circuits. For point-to-point communication, satellites cannot compete with optical
fibers; that is why the majority of earnings from commercial operation of satellites comes
c05.qxd 17/07/2002 19:11 Page 212
TABLE 5.3 Fiber-optic Cable Transmission Standards
System Bit rate (Mbps) 64-kbps Voice channel capacity
OC-1 51.84 672

OC-3 155.52 2,016
OC-9 466.56 6,048
OC-12 622.08 8,064
OC-18 933.12 12,096
OC-24 1,244.16 16,128
OC-36 1,866.24 24,192
OC-48 2,488.32 32,256
from broadcasting or point to multipoint transmission. Table 5.3 shows the hierarchy of
digital signals transmitted using fiber-optic cables. Interconnection of satellite links and
terrestrial circuits is feasible at rates up to OC-3, especially with the later generation of
Ka-band GEO satellites.
Channel Synchronization in TDM

Our explanation of the TI system made the tacit assumption that all 24 incoming PCM
channels were synchronized with each other and running at the same bit rate. This condi-
tion would hold if the voice channels had reached the originating earth station in analog
form and had been digitized by modulators running on a common clock. But if the chan-
nels came into the station in digital form, their synchronization would not be guaranteed.
They may be resynchronized for TDM transmission by a technique called pulse stuffing1,13.
In pulse stuffing the incoming words for each channel flow into an elastic buffer.
There is one such buffer per channel, and each buffer can hold several words. The mul-
tiplexer reads words out of the buffer slightly faster than they come in. Periodically the
multiplexer will go to the buffer and find less than a full word remaining. When that hap-
pens it inserts a dummy word called a stuff word into the frame in place of the word it
would have taken from the buffer. At the same time it places a message on the signaling
channel that states that a stuff word has been inserted. When the demultiplexer at the other
end of the link receives the message it ignores the stuff word. When it is time for the next
frame to be sent the buffer will have more than a full word waiting for transmission.
Pulse stuffing is normal on satellite links that transmit digital signals between dif-
ferent terrestrial TDM systems, because the TDM systems will not be synchronized. The
satellite link is run at a bit rate slightly higher than either of the terrestrial TDM systems
that it joins. Stuffing bits and words are added to the satellite data stream as needed to
fill empty bit and word spaces.
5.7 SUMMARY
Frequency modulation is used for analog signal trans- FM demodulator is characterized by a threshold.
missions in satellite links, primarily for video distri- Provided that a satellite link’s overall carrier-to-noise
bution to older receiving equipment at cable TV head ratio (CN) is above this threshold, the signal-to-
ends. Wideband FM is used to obtain an improvement noise ratio of the demodulated signal be signifi-
in SN ratio relative to the received CN ratio. An cantly greater than the incoming (CN). Additional
c05.qxd 17/07/2002 19:11 Page 213
REFERENCES 213
improvement in baseband SN ratio may be obtained links are designed to meet bit error rate requirements
through pre-emphasis and de-emphasis. De-emphasis in the same way as analog links are designed to deliver
decreases the noise power output of an FM demodu- minimum SN values.
lator; pre-emphasis distorts the transmitted baseband Analog signals must be digitized for transmission
signal before transmission to compensate for the de- over a digital link. This involves sampling the signal
emphasis at the downlink earth station. at a rate that is at least twice the highest frequency
In analog television (TV) transmission by satel- present and converting the sample values to digital
lite, the baseband video signal and one or two audio words. Standard practice with telephone channels is
subcarriers constitute a composite video signal that to use nonuniform quantization with a sampling rate
frequency modulates an uplink carrier. This system of 8 kHz, and to transmit 8-bit words giving a serial
requires a very wide bandwidth for transmission (usu- transmit bit rate of 64 kbps. This system for trans-
ally either a full transponder or a half transponder), mitting digital speech is often known by the old name
but it provides the FM improvement necessary to of PCM, for pulse code modulation. Adaptive differ-
achieve required (SN) values. ential PCM (ADPCM) allows the speech to be trans-
Digital modulation is obviously the modulation of mitted at 32 or 16 kbps. Linear predictive encoding
choice for transmitting digital data. Digitized analog can further reduce the bit rate of digital speech sig-
signals may conveniently share a channel with digi- nals to 4.8 kbps and is used in some LEO satellite
tal data, allowing a link to carry a varying mix of telephone systems. Television signals are transmitted
voice and data traffic. While baseband digital signals digitally using compression techniques such as
are often visualized as rectangular voltage pulses, MPEG 2 to reduce the bit rate to a manageable level.
careful pulse shaping is required to prevent inter- Digital signals from different channels are inter-
symbol interference (ISI) and to permit reasonably leaved for transmission through time division multi-
distortionless transmission through the limited band plexing (TDM). Data, or digitized samples of analog
width of a transponder. Square root raised cosine signals from each channel, are transmitted in turn.
(RRC) filters are used in the design of digital radio The time interval in which one sample or word from
links to create the required zero ISI waveforms at the each channel is sent is called a frame. Channels
receiver. The RRC filter does not exist, so real filters are identified by their position in the frame and
which approximate the RRC filter shape must be individual frames are identified by the presence of
used, with consequent nonzero ISI at the receiver out- synchronization bits that repeat with a known
put. Equalizers must also be used in digital transmit- pattern. Important TDM standards are the North
ters and receivers to compensate for changes in signal American T-system developed by AT&T for the
spectra that can cause ISI. transmission of digital telephony on terrestrial cir-
The common digital modulation schemes used on cuits, and now adopted for digital data transmission
digital satellite links are binary phase shift keying using fractions and multiples of the 1.544 Mbps T-1
(BPSK) and quadrature phase shift keying (QPSK). rate, the ITU-T system with 30 digital speech chan-
In these modulations, an incoming data stream sets nels and two signaling channels, using a bit rate of
the phase of a sinusoidal carrier to one of two (BPSK) 2.048 Mbps, and the optical fiber transmission stan-
or four (QPSK) values. The performance of a BPSK dards OC-1 through OC-48 with bit rates from 51.84
or QPSK link is described by its bit error rate. Digital to 2,488.32 Mbps.
REFERENCES
1. L. W. COUCH, Digital and Analog Communication Sys- Proceedings of the IEEE, 72, 1483 – 1505 (November
tems, Prentice-Hall, Englewood Cliffs, NJ, 6th Ed., 1984).
1998. 6. FTHENAKIS, EMMANUEL, Manual of Satellite Communi-
2. SIMON S. HAYKIN, Digital Communications, John Wiley cations, McGraw-Hill, New York, 1984.
& Sons, New York, 1988. 7. K. S. SHAMNUGAM, Digital and Analog Communication
3. FERREL G. STREMLER, Introduction to Communication Systems, John Wiley & Sons, New York, 1979.
Systems, Addison-Wesley, Reading, MA, 3rd Ed., 8. DAVIDOFF, ed., Amateur Satellite Handbook, American Ra-
1984. dio Relay League, 225 Main Street, Newington, CT, 2000.
4. H. TAUB and D. L. SCHILLING, Principles of Communi- 9. H. NYQUIST, “Certain Topics in Telegraph Transmission
cations Systems, McGraw-Hill, New York, 1971. Theory,” AIEE Transactions, 47, (April 1928).
5. W. H. BRAUN and J. E. KEIGLER, “RCA Satellite Net- 10. LEON W. COUCH, Modern Communication Systems,
works: High Technology and Low User Cost,” Prentice-Hall, Englewood Cliffs, NJ, 1995.
c05.qxd 17/07/2002 19:11 Page 214
11. W. C. LINDSEY and M. K. SIMON, Telecommunication 15. B. HONARY and G. MARKARIAN, Trellis Decoding of
Systems Engineering, Prentice-Hall, Englewood Cliffs, Block Codes, Kluwer, Dordrecht, 1984.
NJ, 1973. 16. E. BIGLIERI, D. DIVSALAR, P. J. MCLANE, and P. SIMON,
12. V. K. BHARGAVA, D. HACCOUN, R. MATYAS, and N. Introduction to Trellis Coded Modulation and Applica-
NUSPL, Digital Communications by Satellite, John tions, Macmillan, New York, 1991.
Wiley & Sons, New York, 1981. 17. I. A. GLOVER and P. M. GRANT, Digital Communica-
13. J. J. SPILKER, Jr., Digital Communications by Satellite, tions, Prentice-Hall, Englewood Cliffs, NJ, 1998.
Prentice-Hall, Englewood Cliffs, NJ, 1977. 18. www.hns.com/spacway
14. www.globalstar.com 19. www.cyberstar.com
PROBLEMS
1. A C-band satellite link sends a single NTSC-TV sig- 3. In Problem 2, two NTSC video signals are trans-
nal through a 36-MHz transponder on a C-band GEO mitted as FM carriers in a bandwidth of 36 MHz. Each
satellite. The NTSC video signal is modulated onto the FM carrier occupies a bandwidth of 16 MHz. A dig-
carrier using wideband frequency modulation, and the ital T1 carrier with a bandwidth of 2.0 MHz can be
bandwidth of the transmitted RF signal is 32 MHz. sent through the same transponder by using a gap be-
The baseband bandwidth of the TV signal is 4.2 MHz. tween the two FM carriers. Some of the transponder
a. Calculate the peak frequency deviation of the FM power must be devoted to the T1 carrier, with the re-
carrier using Carson’s rule. sult that the FM-TV carriers have reduced CN at the
b. Calculate the unweighted FM improvement factor earth station and lower video SN at baseband. This
for the video signal. question asks you to determine the reduction in
video SN. You will need to solve Problem 2 before
c. The overall CN in an earth station receiving the
attempting this problem.
FM-TV transmission is 17 dB. What is the unweighted
video SN ratio at baseband? a. The power at the output of the transponder must be
shared between the three RF signals in proportion to
d. De-emphasis and weighting factors improve the
bandwidth occupied by each signal. For convenience,
quality of the CN ratio by a subjective factor of
assume that the transponder radiates a total power of
17 dB. What is the weighted SN of the baseband
20 W. Calculate the power allocated to each signal
video signal?
when only two FM-TV signals are transmitted, and
2. When overall CN is sufficiently high, it is possi- when all three signals are transmitted.
ble to transmits two FM-TV signals in one 36-MHz
b. Using the results from part (a) above, determine
transponder. The signal-to-noise ratio improvement is
the reduction in CN of the FM-TV signals. Hence
reduced when two TV signals are transmitted rather
find the reduction in baseband video SN and the new
than one because the frequency deviation must be re-
value of unweighted video SN ratio, based on results
duced. Two NTSC FM-TV signals are transmitted
from part (d) of Problem 2.
through a 36-MHz bandwidth transponder. The band-
width of each signal is 16 MHz. c. What is the overall CN ratio at the earth station
receiver for the T1 carrier?
a. Calculate the peak frequency deviation of the FM
signal using Carson’s rule. 4. An NTSC-TV video signal with a baseband band-
b. Calculate the unweighted SN in the baseband width of 4.2 MHz is modulated by FM onto an RF
video bandwidth of 4.2 MHz for an overall CN ra- carrier with a peak frequency deviation of 10 MHz.
tio in the earth station receiver of (CN)0. a. What is the bandwidth of the FM signal according
c. What value must (CN)0 have to ensure that the to Carson’s rule?
unweighted (SN) of the video signal is 33 dB? b. The video signal in part (a) above is transmitted
d. Use the value of (CN)0 you found in part (c) through the transponder to another earth station where
above to find the baseband video SN ratio in clear the CN in the receiver is 19.8 dB in clear air in the
air conditions. The de-emphasis and subjective bandwidth you calculated in part (a) above. If subjec-
weighting factors for the video signal total 17 dB. If tive improvement factors for this video signal are P
the value of (CN)0 at the earth station receiver falls 8 dB, Q 7 dB, what is the baseband video SN?
by 4 dB because of rain in the downlink path, what c. The transponder is reconfigured to carry two video
is the weighted baseband video SN? How would you signals, each occupying 18 MHz, which are received
rate the quality of the video signal? with CN 18.0 dB. If only the frequency deviation
c05.qxd 17/07/2002 19:11 Page 215
PROBLEMS 215
and CN change for the two FM signals, compared 10. a. A 36 MHz bandwidth transponder is used to
to the case in part (b) above, what is the SN of each carry digital signals. A 20 MHz bandwidth in the
baseband video signal? transponder is occupied by a QPSK signal generated
5. A satellite telemetry link operating in S band uses by a transmitter with ideal Nyquist filters with pa-
frequency modulation to transmit the value of an ana- rameter 0.25. What is the symbol rate of the
log voltage on the satellite to a receiving earth sta- QPSK signal in Msps? What is the bit rate of the
tion. The voltage has a range from
1.0 to 1.0 V, QPSK signal?
and a maximum frequency of 1000 Hz. The FM mod- b. Under clear air conditions, the overall (CN)0 ra-
ulator on the satellite has a constant of 10,000 Hz per tio in the earth station receiver is 18.0 dB. If the QPSK
volt. At the receiving earth station the CN ratio of demodulator has an implementation margin of 1.5 dB,
this signal is 10 dB measured in the Carson’s rule what is the bit error rate of the baseband digital sig-
bandwidth, and is 3 dB above the FM threshold of nal in clear air conditions? How often does a bit er-
the FM demodulator. ror occur. (Give your answer in days, hours, minutes,
a. What is the Carson’s rule bandwidth for the FM or seconds, as appropriate.)
signal? c. Under rain conditions, the overall (CN)0 ratio of
b. What is the baseband SN ratio at the earth sta- the QPSK signal in part (a) above falls to 14.3 dB at
tion receiver output for the recovered analog signal? a receiving station. What bit error rate would you ex-
pect in the recovered bit stream? How often does a
6. A satellite link has an RF channel with a band-
bit error occur?
width 2.0 MHz. The transmitter and receiver have
RRC filters with 0.5. What is correct symbol rate 11. A satellite communication system is built as a
(pulse rate) for this link? star network with one large hub station and many re-
mote small earth stations. The system operates at Ka
7. A Ku-band satellite uplink has a carrier frequency
band using the K9 geostationary satellite, and carries
of 14.125 MHz and carries a symbol stream at Rs
digital signals which may be voice, data, or com-
16 Msps. The transmitter and receiver have ideal RRC
pressed video. The K9 satellite has transponders with
filters with 0.25. What is bandwidth occupied by
a bandwidth of 60 MHz that can be operated in ei-
RF signal, and what is the frequency range of the
ther of two modes: as a bent pipe or with a 40 Msps
transmitted RF signal?
QPSK baseband processor.
8. A T1 data transmission system transmits data at The outbound link from the hub to the remote sta-
1.544 Mbps over a GEO satellite link. At the receiv- tions has an uplink from the hub station to the satel-
ing terminal the clear air value of overall (CN)0 is lite that is the input of transponder 1. Signals from
16.0 dB. The modulation used on the link is BPSK the hub are transmitted using a single TDM carrier
and the implementation margin of the BPSK demod- and QPSK modulation with a symbol rate of 40 Msps.
ulator is 0.5 dB. In the initial system design the remote earth stations
a. Find the BER at the receiver output and the average use receivers capable of receiving 40-Msps QPSK
time between errors. signals. The transponder is operated in bent pipe
b. Rain affects the downlink from the satellite and mode with sufficient back-off to make it linear. The
the overall CN ratio in the receiver falls by 6.0 dB hub transmitter operates at an output power of 100
to 10.0 dB. What is the bit error rate now? W, which gives CN 30 dB in the transponder in
clear air, measured in the correct noise bandwidth of
9. A satellite data transmission system transmits data
a 40 Msps QPSK receiver equipped with RRC filters
from two T1 carriers as a single 3.088-Mbps bit
having 0.4. In clear air conditions, the resulting
stream using QPSK. The symbol rate on the link is
CN of the earth station receiver, ignoring noise trans-
1.544 Msps. The satellite link uses ideal RRC filters
mitted by the satellite, is 20 dB. The receiver has a
with 0.25. At the receiving terminal the clear air
QPSK demodulator with an implementation margin
value of overall (CN)0 is 16.0 dB and the imple-
of 1.0 dB. For the purposes of this question you may
mentation margin of the QPSK demodulator is 1 dB.
assume that all transmitters and receivers in the net-
a. What is the bandwidth occupied by this signal, and work and on the K9 satellite have ideal RRC filters.
the noise bandwidth of the receiver for this signal?
a. Find the overall CN in the earth station receiver
b. Find the BER at the receiver output and the aver- in clear air conditions and estimate the bit error rate
age time between errors. for the recovered data signal, assuming that FEC is
c. Rain affects the downlink from the satellite and the not used. What is the correct noise bandwidth for the
overall CN ratio in the receiver falls by 6.0 dB to earth station receiver that receives the QPSK signal,
10.0 dB. What is the bit error rate now? and what is the bit rate of the link?
c05.qxd 17/07/2002 19:11 Page 216
b. An uplink fade occurs which causes an attenuation For the case when FEC is used, determine the overall
of 10 dB between the hub station and the satellite. CN ratio and the expected BER during a rain fade
The transponder is operated in bent pipe mode. Find that causes the CN ratio of the received signal to fall
the overall CN in the remote earth station receiver by 5 dB but which does not attenuate the interfering
and estimate the BER of the recovered data. signal.
c. An uplink fade occurs which causes an attenuation d. If the extra bandwidth to implement half rate FEC
of 10 dB between the hub station and the satellite. is available at the satellite, would you recommend that
Transponder 1 is switched to operate with the 40 Msps FEC be used in this case? Give reasons for your
baseband processor. The QPSK demodulator in answer.
transponder 1 has an implementation margin of 1 dB. e. What are the advantages and disadvantages of us-
Find the overall CN in the remote earth station ing forward error correction in satellite links? Illus-
receiver and estimate the BER. trate your answer using the above example of a high
d. Rainfall statistics for the location of the hub sta- data rate signal sent to a small earth terminal.
tion show that attenuation at the uplink frequency will 13. A T1 data transmission system transmits data at
exceed 20 dB for 0.01% of an average year. If the hub 1.544 Mbps over a GEO satellite link. At the receiv-
station uses uplink power control to mitigate the ef- ing terminal the clear air value of overall (CN)0 is
fects of uplink rain attenuation, determine the maxi- 16.0 dB. The modulation used on the link is BPSK
mum uplink transmitter power (in watts) that must be and the implementation margin of the BPSK demod-
transmitted to ensure that the link BER does not ex- ulator is 0.5 dB.
ceed 10
6 at the remote earth station receiver output
a. Find the BER at the receiver output, and the time
for 99.99% of an average year when:
that elapses, on average, between bit errors.
(i) A linear bent pipe transponder is used b. Rain affects the downlink from the satellite and
(ii) A 40-Msps QPSK baseband processing transpon- the overall CN ratio in the receiver falls by 6.0 dB
der is used. to 10.0 dB. What is the bit error rate now? What is
e. Discuss the value of UPC at the hub station trans- the average time between bit errors?
mitter in this application. Would you recommend a c. The modulation method is changed to QPSK and
linear transponder or a baseband processing transpon- the bit rate is increased to 2 T1 3.088 Mbps, and
der be used on the K9 satellite? Give reasons for your the symbol rate on the link is 1.544 Msps (Mbaud).
answer. What is the bit error rate now? What is the average
12. A Ku-band VSAT station receives a TDM data time between bit errors?
stream at 1.544 Mbps from a GEO satellite. The mod- d. Rain affects the downlink from the satellite and
ulation is QPSK and under clear air conditions the the overall CN ratio in the receiver falls by 6.0 dB
downlink CN in the VSAT receiver is 20 dB (ig- to 10.0 dB. What is the bit error rate and average time
noring noise from the satellite). The CN in the satel- between errors now for the QPSK link?
lite transponder is 30 dB. A nearby terrestrial LOS e. Changing the modulation method to QPSK and in-
link causes interference with the VSAT such that the creasing the bit rate to 3.088 Mbps is likely to lead
carrier-to-interference ratio CI in the VSAT receiver to an unacceptably high bit error rate when the satel-
is 19.6 dB. All CN and CI values are quoted for lite downlink was affected by rain because the re-
the optimum noise bandwidth of the VSAT receiver. ceiver (CN)0 will fall by 6 dB. We could retain a bit
The receiver uses ideal RRC filters with 0.4 and rate on the link of 1.544 Mbps when using QPSK by
its QPSK demodulator has an implementation margin changing the transmitter and receiver RRC filters to
of 1 dB. operate at a symbol rate Rs 1.5442 0.772 Msps.
a. What is the symbol rate of the QPSK signal and What is the bit error rate and average time between
the noise bandwidth of the VSAT receiver? errors now for the QPSK link?
b. What is the clear air overall CN ratio in the VSAT f. Rain affects the downlink from the satellite in part
receiver, assuming that the interference can be con- (e) above, and the overall CN ratio in the receiver
sidered AWGN? What BER would you expect at the falls by 6.0 dB to 10.0 dB. What is the bit error rate
data output of the VSAT receiver assuming no FEC now for the QPSK link?
is applied to the signal? 14. The baseband average SN ratio for a probabil-
c. The system is redesigned and half rate FEC is added ity of bit error Pe (BER) with N-bit PCM is given by
to the signal so that the bit rate at the transmitter is
SN 10 log10 c d dB
doubled, but transmitter power is not increased. In the 22N
receiver, the FEC decoder has a coding gain of 6 dB. 1 4Pe 22N
c05.qxd 17/07/2002 19:11 Page 217
PROBLEMS 217
a. The effective CN in a digital receiver with QPSK (ii) The weighted SN ratio after pre-emphasis and
modulation is 15.6 dB under clear air conditions. subjective improvements are added,
What is the baseband SN for 8-bit PCM coded (iii) The link margin for the downlink given an FM
speech? threshold at 8.5 dB.
b. In moderate rain conditions the effective CN falls c. Heavy rain affects the uplink to the satellite causing
to 13.6 dB. What is the baseband SN for the 8-bit the CN in the transponder to fall to 18 dB. Assuming
PCM signal? linear bent pipe operation of the transponder, find:
c. In heavy rain, the effective CN falls to 11.6 dB. (i) The overall (CN)0 ratio in the earth station re-
What is the baseband SN for the 8-bit PCM signal?
ceiver and
d. The minimum acceptable baseband SN in a
(ii) The video SN ratio. Is this an acceptable SN
speech channel is usually set at 30 dB. What is the
for viewing a television picture?
corresponding minimum allowable effective CN for
a QPSK link carrying 8-bit PCM coded speech? d. Heavy rain affects the downlink from the satellite
causing 4 dB of rain attenuation. The uplink is oper-
15. Direct Broadcast Satellite TV In this question ating in clear air conditions. Assuming a medium noise
you are asked to analyze the performance of the TV temperature of 270 K in rain and 100% coupling of
system using frequency modulation. sky noise into antenna noise temperature, find:
The uplink station delivers a signal to the satellite
(i) The new value for (CN)dn in the earth station
which conforms to the following specification:
receiver,
Transponder and satellite characteristics
(ii) The corresponding overall (CN)0 ratio in the
Transponder bandwidth 25 MHz earth station receiver,
(CN)up in 20-MHz noise bandwidth 24 dB (iii) The video SN ratio. Is this an acceptable qual-
Saturated output power 200 W ity television picture?
Downlink frequency 12.5 GHz e. Draw a block diagram for the earth station receiver,
Downlink antenna gain, on axis 39.0 dB showing only the parts that relate to reception and
output of the NTSC video signal. Your block diagram
Atmospheric clear air loss 0.4 dB
must show the center frequency, gain, and bandwidth
All other losses 0.5 dB of each block, as appropriate.
Receive station parameters Do not specify filters with Q exceeding 50.
Antenna diameter 18 inches 16. This problem examines the design and performance
Aperture efficiency 70% of a digital satellite communication link using a geo-
Antenna noise temperature (clear air) 40 K stationary satellite with bent pipe transponders, used to
distribute digital TV signals from one central (hub) earth
Receiver noise temperature 90 K station to many receiving stations throughout the United
a. The uplink master station transmits an NTSC States. The link uses QPSK digital transmission at 20
video signal with a baseband bandwidth of 4.2 MHz Msps with half rate forward error correction. The half
to one transponder on the satellite using FM. The rate FEC gives a coding gain of 5.5 dB.
transponder is operated with 1 dB of output back-off The design requires that an overall CN ratio of
and the FM signal occupies a bandwidth of 24 MHz. 9.5 dB be met in the earth station receiver to ensure
For an earth station with a high gain LNA, at a dis- that noise in the video signal on the TV screen is held
tance of 38,000 km from the satellite, on the
4 dB to an acceptable level. The uplink transmitter power
contour of the satellite antenna beam. and the receiving antenna gain and diameter must be
Find: determined. The available link margins for each of
(i) The peak frequency deviation of the FM signal. the systems must be found and the performance of
the system analyzed when rain attenuation occurs in
(ii) The power at the input to the earth station LNA. the satellite–earth paths.
(iii) The downlink (CN)dn in a noise bandwidth of The system is specified in Table P.16.
24 MHz.
a. Uplink design
(iv) The overall (CN)0 in the earth station receiver.
Find the uplink transmitter power to achieve the re-
b. For the FM video signal in part (a) above: quired (CN)up 30 dB in the transponder in clear
(i) The unweighted video SN ratio at the baseband air atmospheric conditions. Find the noise power in
output of the receiver, the transponder for a noise bandwidth of 20 MHz,
c05.qxd 17/07/2002 19:11 Page 218
TABLE P.16 System and Satellite Specification
Ku-band satellite parameters

Total RF output power 3.2 kW
Antenna gain, on axis, Ku band 31 dB
(transmit and receive)
Receive system noise temperature 500 K
Transponder saturated output power: Ku band 80 W
Earth station receiver IF noise bandwidth 20 MHz
Minimum permitted overall C/N in receiver 9.5 dB
Transponder HPA output backoff 1 dB
Transmitting Ku-band earth station
Antenna diameter 5m
Required C/N in Ku-band transponder 30 dB
Miscellaneous uplink losses 0.3 dB
Location:
2 dB contour of satellite receiving antenna
Receiving Ku-band earth station
Receiver IF bandwidth 20 MHz
Antenna noise temperature 30 K
LNA noise temperature 110 K
Required overall (C/N)0 in clear air 17 dB
Miscellaneous downlink losses 0.2 dB
Location:
3 dB contour of satellite transmitting antenna
Rain attenuation and propagation factors at Ku-band
Clear air attenuation
Uplink 14.15 GHz 0.7 dB
Downlink 11.45 GHz 0.5 dB
Rain attenuation
Uplink 0.01% of year 6.0 dB
Downlink 0.01% of year 5.0 dB
and then add 30 dB to find the transponder input Calculate the path loss at the downlink frequency
power level. Calculate the earth station transmit an- of 11.15 GHz. Don’t forget the downlink losses. The
tenna gain, and the path loss at 14.15 GHz. Generate transponder is operated with 1 dB output backoff.
an uplink power budget and find the required power Find the transponder output power and then generate
at the transponder input to meet the (CN)up 30 dB a downlink power budget.
objective in the transponder. Don’t forget the various Hence find the receiving antenna gain Gr and di-
uplink losses. ameter for a frequency of 11.45 GHz. This diameter
b. Downlink design antenna will provide the required (CN)0 in the earth
station receiver under clear air conditions.
Assume a high gain LNA and ignore the noise gen-
erated in other parts of the receiver. Calculate the c. Rain effects
downlink (CN)dn to give overall (CN)0 17 dB A practical system must continue to operate under
when (CN)up 30 dB. Hence find the receiver in- adverse weather conditions, so we need a margin for
put power to give the required (CN)dn using a value rain attenuation and increase in sky noise tempera-
of receiving antenna gain Gr. ture during rain. In the following section you will
c05.qxd 17/07/2002 19:11 Page 219
PROBLEMS 219
determine the margins available on the uplink and Is the downlink link margin acceptable? If not, cal-
downlink to combat rain attenuation and increase in culate the gain and diameter of an earth station an-
sky noise temperature. tenna that will ensure an overall CN value that does
d. Uplink rain attenuation meet the specification.
Under conditions of heavy rain, the Ku band path to f. Summarize your design for the Ku-band earth
the satellite suffers an attenuation of 6 dB for 0.01% station and uplink and downlink. Compare the earth
of the year. We must find the uplink attenuation mar- station receiving antenna diameter for the Ku-band
gin and decide whether uplink power control would system with the antenna for a similar C-band
improve system performance at Ku band. system.
The uplink CN was 30 dB in clear air. With 6-dB If the Ku-band antenna is larger (and therefore has
uplink path attenuation, the CN in the transponder a much higher gain) explain why.
falls to 24 dB. (Rain on the earth has no effect on the 17. A satellite communication system uses a single
satellite transponder system noise temperature.) As- 54 MHz bandwidth Ku-band transponder to carry 400
sume linear transponder characteristic and no uplink two-way telephone conversations (800 RF channels)
power control. Find the transponder output power using analog modulation with single channel per car-
with 6 dB of rain attenuation in the uplink. rier frequency modulation (SCPC-FM). The parame-
Hence find the overall (CN)0 in an uplink rain ters of any one channel are:
fade of 6 dB, and the link margin available on the up-
Voice channel bandwidth 100–3400 Hz
link. Is this an adequate uplink margin, given the rain
attenuation for most of the United States? RF channel bandwidth 45 kHz
e. Downlink attenuation and increase in sky noise RF channel spacing 65 kHz
in rain Downlink path loss (incl. atmos. loss) 206.5 dB
The 11.45 GHz path between the satellite and the re-
Satellite downlink antenna gain (on axis) 29 dB
ceive station suffers rain attenuation exceeding 5 dB
for 0.01% of the year. Assuming 100% coupling of Demodulator FM threshold 5 dB
sky noise into antenna noise, and 0.5 dB clear air The transponder has a saturated power output of
gaseous attenuation, calculate the overall CN under 40 W, but is run with 3-dB output backoff to achieve
these conditions. Assume that the uplink station is op- near-linear operation. The uplink stations which trans-
erating in clear air. Calculate the available downlink mit the SCPC-FM signals to the transponder achieve
fade margin. (CN)up 25 dB in the 45 kHz channel noise band-
Find the sky noise temperature that results from width of the earth station receiver. The system noise
a total excess path attenuation of 5.5 dB (clear air temperature of the receiving earth station is 120 K in
attenuation plus rain attenuation); this is the new an- clear air.
tenna temperature in rain, because we assumed a. Calculate the power per RF channel at the transpon-
100% coupling between sky noise temperature and der output.
antenna temperature. Evaluate the change in re-
b. Calculate the gain of the antenna at a receiving
ceived power and increase in system noise temper-
earth station that is located on the
3 dB contour
ature in order to calculate the change in CN ratio
of the satellite footprint which will provide an over-
for the downlink.
all CN 10 dB in a receiver for a single RF chan-
In clear air, the atmospheric attenuation on the
nel with a noise bandwidth of 45 kHz, in clear air
downlink is 0.5 dB. The corresponding sky noise
conditions.
temperature is approximately 0.5 7 35 K, which
leads to the antenna temperature of 30 K given in c. Calculate the diameter of the receiving antenna
the Ku-band system specification. When the rain with a circular aperture having 65% aperture effi-
causes 5 dB attenuation, the total path attenuation ciency at a frequency of 11.5 GHz.
from the atmosphere and the rain is 5.5 dB. The sky d. The receiver applies a de-emphasis weighting of
noise will be much higher in rain. Find the increase 6 dB to the recovered voice signal and a psophome-
in noise power caused by the increase in sky tem- tric weighting of 2.5 dB.
perature. Calculate the weighted SN at the baseband out-
Hence find the new (CN)dn rain value with 5.5 dB put of the receiver.
attenuation on the downlink path. Find the overall e. Comment on the performance of the system. Is the
CN by combining the clear air uplink (CN)up of 30 SN adequate in clear air? If the downlink fades by
dB with the rain faded downlink (CN)dn rain to give 5 dB because of rain, what is the SN at baseband?
overall (CN)0 in rain. Is this acceptable for voice communications?
c05.qxd 17/07/2002 19:11 Page 220
18. In Problem 5, an analog voltage was transmitted quantization signal-to-noise ratio of the recovered
from a satellite to earth using frequency modulation. analog signal.
The signal could have been sent digitally using a dig- b. Find the BER for the recovered bit stream at the
ital to analog converter and PSK modulation. This output of the BPSK demodulator, and hence calcu-
problem compares the performance of the digital link late the average SN ratio in the analog voltage due
to the analog link of Problem 5. to bit errors.
The digital link is allocated an RF bandwidth of c. Solve Problem 5 for the FM version of this link.
25 kHz, and uses BPSK modulation. At the receiving Which link has the better performance? What changes
terminal, the CN ratio is 10 dB. The link has ideal should be made to the link with the poorer perform-
RRC filters with 0.25 and the BPSK demodula- ance to make the SN ratios approximately equal for
tor has an implementation margin of 0.5 dB. the FM and BPSK links? If an RF bandwidth of 50
a. The analog voltage is sampled at 2.5 kHz and kHz could be used for the BPSK signal, would the
converted to a series of digital words with an ana- addition of half rate forward error correction with a
log to digital converter. Determine the maximum coding gain of 6 dB improve the performance of the
number of bits in each word and the average BPSK link?
c06.qxd 09/08/02 14:18 Page 221
CHAPTER 6
MULTIPLE ACCESS
6.1 INTRODUCTION
Satellites are always built with the intention that many users will share the bandwidth al-
located to the satellite, allowing many separate communication links to be established
through the satellite’s transponders. In a large geostationary satellite operated by an
international carrier, the satellite can carry tens of thousands of simultaneous telephone
conversations between people in many different parts of the world. A domestic satellite
carrying television program material can broadcast its signals to thousands of cable tele-
vision companies serving millions of homes. The ability of the satellite to carry many
signals at the same time is known as multiple access. Multiple access allows the com-
munication capacity of the satellite to be shared among a large number of earth stations,
and to accommodate the different mixes of communication traffic that are transmitted by
the earth stations.
The basic form of multiple access employed by all communications satellites is the
use of many transponders. A large GEO satellite may have a communication bandwidth
of over 2000 MHz within an allocated spectrum of 500 MHz. Through frequency reuse
with multiple antenna beams and orthogonal polarization, the spectrum can be reused
several times over—as many as seven times in the case of INTELSAT IX satellites. The
frequency spectrum used by the satellite is divided into smaller bandwidths which are al-
located to transponders, allowing separate communication links to be established via the
satellite on the basis of transmit frequency. Transponder bandwidths of 36, 54, and 72 MHz
have been commonly employed on GEO satellites. The individual transponders may carry
one signal, a single analog television program, for example, or hundreds of signals, as
with mobile satellite telephone systems. The use of multiple transponders to divide up a
frequency band is not considered as multiple access, although the reason for their use is
to make it easier for different earth stations to share the available frequency spectrum
efficiently.
The signals that earth stations transmit to a satellite may differ widely in their
character—voice, video, data, facsimile—but they can be sent through the same satellite
using multiple access and multiplexing techniques1,2. Multiplexing is the process of com-
bining a number of signals into a single signal, so that it can be processed by a single am-
plifier or transmitted over a single radio channel. Multiplexing can be done at baseband
or at a radio frequency. The corresponding technique that recovers the individual signals
is called demultiplexing. Multiplexing is a key feature of all commercial long-distance
communication systems, and is part of the multiple access capability of all satellite
communication systems.
The designer of a satellite communication system must make decisions about the
form of multiple access to be used. The multiple access technique will influence the ca-
pacity and flexibility of the satellite communication system, its cost, and its ability to earn
revenue. The basic problem in any multiple access system is how to permit a changing
group of earth stations to share a satellite in such a way that satellite communication
221
c06.qxd 09/08/02 14:18 Page 222
222 CHAPTER 6 MULTIPLE ACCESS
Guard band
Ch 1 Ch 2 Ch 3 Ch 4
Transponder bandwidth Frequency

FDMA Guard time
Ch 1 Ch 2 Ch 3 Ch 4
Frame period Time
TDMA
Power
Station N
FIGURE 6.1 Multiple access
Station 4
techniques: FDMA, TDMA, and
Station 3 CDMA. Note that in the direct
Station 2
Station 1 sequence form of CDMA shown
Transponder bandwidth Frequency here, all the channels overlap in
CDMA both time and frequency.
capacity is maximized, bandwidth is used efficiently, flexibility is maintained, and cost to

the user is minimized while revenue to the operator is maximized. The multiple access
system should also allow for changing patterns of traffic over the 10 or 15 years of the
expected lifetime of the satellite. Usually, all of these requirements cannot be satisfied at
the same time and some may have to be traded off against others.
There are three basic multiple access techniques, illustrated in Figure 6.1. In freq-
uency division multiple access (FDMA) all users share the satellite at the same time, but
each user transmits at a unique allocated frequency. This approach to sharing the frequency
spectrum is familiar to us all, as it is the way that radio broadcasting has always shared
the air waves. Each radio station is allocated a frequency and a bandwidth, and transmits
its signals within that piece of the frequency spectrum. FDMA can be used with analog
or digital signals. In time division multiple access (TDMA) each user is allocated a unique
time slot at the satellite so that signals pass through the transponder sequentially. Because
TDMA causes delays in transmission, it is used only with digital signals. In code division
multiple access (CDMA) all users transmit to the satellite on the same frequency and at
the same time. The earth stations transmit orthogonally coded spread spectrum signals
that can be separated at the receiving earth station by correlation with the transmitted code.
CDMA is inherently a digital technique. In each of the multiple access techniques, some
unique property of the signal (frequency, time, or code) is used to label the transmission
such that the wanted signal can be recovered at the receiving terminal in the presence of
all other signals.
The distinction between multiplexing and multiple access is sometimes blurred. Mul-
tiplexing applies to signals that are generated at one location, whereas multiple access
refers to signals from a number of different geographical locations. For example, an earth
station might use time division multiplexing (TDM) to create a high-speed digital data
stream from many digital speech channels delivered to that earth station, and then mod-
ulate the data stream onto an RF carrier and transmit the carrier to the satellite. At the
satellite, the carrier can share a transponder using time division multiple access (TDMA)
or frequency division multiple access (FDMA) with other carriers from earth stations any-
where within the satellite’s coverage zone. The resulting signal is called TDM-TDMA or
TDM-FDMA. Note the distinction between TDM and TDMA: signals at one earth sta-
tion are combined by multiplexing, and then share a satellite transponder with signals
from other earth stations by multiple access.
c06.qxd 09/08/02 14:18 Page 223
6.2 FREQUENCY DIVISION MULTIPLE ACCESS (FDMA) 223
In all three of the classic multiple access techniques, some resource is shared. If the
proportion allocated to each earth station is fixed in advance, the system is called fixed
access (FA) or preassigned access (PA). If the resource is allocated as needed depending
on changing traffic conditions, the multiple access technique is called demand access (DA).
Demand access blurs some of the distinctions between FDMA and TDMA, since stations
in an FDMA-DA system transmit only when they have traffic. Demand access with FDMA
is widely used in VSAT systems, where earth stations may have traffic to send only in-
termittently 3. Fixed assignment would be wasteful of transponder capacity, so demand as-
signment is used. Similarly, a group of earth stations may access part of the bandwidth
of a transponder using TDMA, while other TDMA groups of earth stations share differ-
ent sections of the transponder bandwidth. This approach has been used in both VSAT
and mobile satellite systems. (See Chapter 9.) Demand assignment can also be used with
CDMA to reduce the number of signals in the transponder at any one time. The Globalstar
LEO mobile satellite system uses CDMA with demand assignment4.
Systems which combine both FDMA and TDMA techniques are sometimes called
hybrid multiple access schemes or multifrequency TDMA (MF-TDMA). In the sections
that follow, we will first discuss FDMA, TDMA, and CDMA as fixed assignment schemes,
and then cover demand access and hybrid multiple access.
6.2 FREQUENCY DIVISION MULTIPLE

ACCESS (FDMA)
Frequency division multiple access was the first multiple access technique used in satel-
lite communication systems. When satellite communications began in the 1960s, most of
the traffic carried by satellites was telephony. All signals were analog, and analog multi-
plexing was used at earth stations to combine large numbers of telephone channels into a
single baseband signal that could be modulated onto a single RF carrier. Individual tele-
phone channels can be shifted in frequency from baseband to a higher frequency so that
they can be stacked into a group of channels using frequency division multiplexing (FDM).
The process begins by limiting individual telephone channels to the frequency range
300–3400 Hz, and then frequency shifting 12 channels to the frequency range 60–108
kHz with 4-kHz spacing between channels by generating single sideband suppressed car-
rier signals with 12 carrier frequencies spaced 4 kHz apart. The 12 channels occupying
60–108 kHz are known as a basic group. Five basic groups can be frequency shifted to
the range 60–300 kHz to make a 60-channel supergroup occupying a baseband bandwidth
of 240 kHz. Supergroups can be stacked in the baseband to make up single signals that
consist of 300, 600, 900, or 1800 multiplexed telephone channels.
The analog FDM technique is now obsolete in the United States and many other
countries, but it was the primary method of multiplexing telephone channels for trans-
mission over terrestrial cable or microwave links for about 50 years. Early satellite systems
used FDM to multiplex up to 1800 telephone channels into a wide baseband occupying
up to 8 MHz, which was modulated onto an RF carrier using frequency modulation (FM).
Appendix B describes the process of voice multiplexing using FDM techniques, and
provides some details of satellite transmission of analog signals using FDM-FM-FDMA.
The FDM-FM RF carrier was transmitted to the satellite, where it shared a
transponder with other carriers using FDMA. The technique is known as FDM-FM-
FDMA, and was the preferred method for the transmission of telephone channels over
Intelsat satellites for more than 20 years. The main advantage of FDMA is that filters
can be used to separate signals. Filter technology was well understood when satellite
c06.qxd 09/08/02 14:18 Page 224
Transponder 1
10 10 10
USA USA Chile
3714 3726 3738

Carrier frequency in MHz
Transponder 2
20 10
USA Chile
3759 3771
Carrier frequency in MHz
FIGURE 6.2 Frequency plan for two C-band transponders using FDMA. The triangles are
symbols representing the bandwidth occupied by the signals, not power spectral densities.
The places and figures within the triangles are the transmitting station location and carrier
RF bandwidth. Frequencies shown are for the downlink from the satellite.
communications began, and microwave filters were used in earth stations to separate
the FDMA signals within a given transponder. In a fixed assignment system, each trans-
mitting earth station was allocated a frequency and bandwidth for each group of sig-
nals it wished to send.
Figure 6.2 shows a typical fixed assignment FDMA plan for two C-band transponders.
The triangles represent RF carriers with the transmitting earth station and RF bandwidth
shown inside the triangle. Frequencies shown are for the downlink from the satellite. Within
each transmission, signals (primarily telephone channels) for different destinations are mul-
tiplexed using FDM. Typical Intelsat FDM carriers with a bandwidth of 10 MHz carried
132 to 252 telephone channels. If a small group of channels is intended for a given earth
station, the entire carrier must be received and demultiplexed to recover those channels.
Channels sent by the same carrier but intended for other earth stations are discarded. The
36 MHz transponder bandwidth can be used to send one or two television signals instead
of hundreds of telephone channels.
The use of microwave filters to separate channels made the fixed assignment ap-
proach to FDMA very inflexible. Changing the frequency assignment or bandwidth of any
one transmitting earth station required retuning of the microwave filters at several receiving
earth stations. The fixed assignment FDM-FM-FDMA scheme illustrated in Figure 6.2
also makes inefficient use of transponder bandwidth and satellite capacity.
As an example, suppose an earth station in the west of the United States uses a Pa-
cific Ocean GEO satellite to send telephone channels to earth stations in Korea, Japan,
and Chile. The time difference between North America and the Pacific Rim countries
means that the channels will be busy for only a few hours per day, and at a different time
of day than the U.S.–Chile links. With fixed assignment, the frequencies and satellite ca-
pacity cannot be reallocated between routes, so much of the satellite capacity remains idle.
Estimates of average loading of Intelsat satellites using fixed assignment are typically
around 15%. It is not possible to achieve 100% loading of satellites used for international
c06.qxd 09/08/02 14:18 Page 225
1 2 3 4 5 6 7 8 9
40 kHz digital speech channels 10 kHz guard bands

FIGURE 6.3 Illustration of a Ku-band transponder bandwidth filled with a large number of
FDMA-SCPC digital speech channels. RF bandwidth of each channel is 40 kHz with 10 kHz
guard bands between channels.
traffic, or even for domestic traffic in many cases. Demand assignment and single channel
per carrier (SCPC) techniques allow higher loadings and therefore give satellite operators
increased revenue. There has been a steady move away from fixed assignment systems as
a result.
Every earth station that operates in an FDMA network must have a separate IF
receiver for each of the carriers that it wishes to receive. SCPC systems can have a
very large number of carriers in one transponder; as a result, FDMA earth stations tend
to have a very large number of IF receivers and demultiplexers which select individ-
ual carriers using narrowband IF filters. Figure 6.3 shows how the intermediate
frequency bandwidth of a receiving earth station could be configured to receive 1000
digital speech channels, each with a bandwidth of 40 kHz from a 54 MHz wide Ku-
band transponder. The 10-kHz frequency spaces between the channels are called guard
bands. Guard bands are essential in FDMA systems to allow the filters in the receivers
to select individual channels without excessive interference from adjacent channels.
All filters have a roll-off characteristic, which describes how rapidly a filter can change
from near zero attenuation in its pass band to high attenuation in the stop band. Typ-
ically, guard bands of 10 to 25% of the channel bandwidth are needed to minimize
adjacent channel interference.
FDM-FM-FDMA was a telephone transmission technique well suited to analog
telephone signals. Telephony has largely become digital, and frequency division
multiplexing has been replaced by time division multiplexing. Digital speech is now
used throughout telephone systems, so multiple telephone channels are always trans-
mitted as a high-speed digital signal. The T1 or DS-1 carriers are examples of lower
speed digital multiplexed carriers. Optical networks carry OC–48 digital signals at rates
up to 2.7 Gbps and beyond. Appendix B discusses the techniques used in FDM-FM-
FDMA systems, and the calculation of system capacity using this multiple access
scheme. Apart from the analog nature of FDM-FM-FDMA which has rendered it ob-
solete, it is a rather inflexible way to allocate satellite transponder capacity and is not
easily adapted to demand access.
FDMA is widely used as a method of sharing the bandwidth of satellite transpon-
ders. Ideally, a satellite would carry a very large number of transponders, each of which
could be allocated to a single RF carrier. In the case of telephony, each transponder would
have a bandwidth exactly matched to the RF spectrum of the transmitted telephone chan-
nel, with tight filtering to ensure that each signal can be separated from adjacent signals.
This approach is impractical: thousands of transponders would be needed and the satel-
lite could be used only for telephony. The builders and operators of satellites have
c06.qxd 09/08/02 14:18 Page 226
historically shown a strong preference for wideband transponders that can carry any type
of traffic—the bent pipe transponder that can carry voice, video, or data as the market-
place demands. As a result, transponders have always had wide bandwidths, with band-
widths of 36, 54, and 72 MHz commonly employed. When an earth station has a carrier
that occupies less than the transponder bandwidth, FDMA can be used to allow that carrier
to share the transponder with other carriers.
Allocating a wideband transponder to a single narrow bandwidth signal is clearly
wasteful, so FDMA is a widely used technique. When an earth station sends one signal
on a carrier, the FDMA access technique is called single channel per carrier (SCPC).
Thus a system in which a large number of small earth stations, such as mobile telephones,
access a single transponder using FDMA is called a single channel per carrier frequency
division multiple access scheme. Not surprisingly, this lengthy descriptor is abbreviated
to SCPC-FDMA. Hybrid multiple access schemes can use time division multiplexing of
baseband channels which are then modulated onto a single carrier. A number of earth sta-
tions can share a transponder using frequency division multiple access, giving a system
known as TDM-SCPC-FDMA. Note that the sequence of abbreviations is baseband mul-
tiplexing technique first, then multiple access technique next. TDM-SCPC-FDMA is used
by VSAT networks in which the earth stations carry more than one baseband signal.
FDMA has a disadvantage in satellite communications systems when the satellite
transponder has a nonlinear characteristic. Most satellite transponders use high-power am-
plifiers which are driven close to saturation, causing nonlinear operation. A transponder
using a traveling wave tube amplifier (TWTA) is more prone to nonlinearity than one with
a solid state high-power amplifier (SSHPA). Equalization at the transmitting station, in
the form of predistortion of the transmitted signal, can sometimes be employed to linearize
the transponder when fixed assignment is used. Linearization of solid-state and TWT HPAs
on the satellite is also possible. Nonlinearity of the transponder HPA causes a reduction
in the overall (CN)0 ratio at the receiving earth station when FDMA is used because in-
termodulation (IM) products are generated in the transponder. Some of the IM products
will be within the transponder bandwidth and will cause interference. The IM products
are treated as though they were thermal noise, adding to the total noise in the receiver of
the receiving earth station.
Intermodulation
Intermodulation products are generated whenever more than one signal is carried by a
nonlinear device. Sometimes filtering can be used to remove the IM products, but if they
are within the bandwidth of the transponder they cannot be filtered out. The saturation
characteristic of a transponder can be modeled by a cubic curve to illustrate the genera-
tion of third-order intermodulation. Third-order IM is important because third-order IM
products often have frequencies close to the signals that generate the intermodulation, and
are therefore likely to be within the transponder bandwidth.
To illustrate the generation of third-order intermodulation products, we will model
the nonlinear characteristic of the transponder HPA with a cubic voltage relationship and
apply two unmodulated carriers at frequencies f1 and f2 at the input of the amplifier
Vout AVin b1Vin 2 3 (6.1)
where A W b.
The amplifier input signal is
V1 cos v1t V2 cos v2t (6.2)
c06.qxd 09/08/02 14:18 Page 227
The amplifier output signal is

Vout AVin b1Vin 2 3
AV1 cos v1t AV2 cos v2t b1V1 cos v1t V2 cos v2t2 3 (6.3)
linear term cubic term
The linear term simply amplifies the input signal by a voltage gain A. The cubic
term, which will be denoted as V3out, can be expanded as
V3out 1V1 cos v1t V2 cos v2t2 3
b3V 13 cos3 v1t V23 cos3 v2t
2V22 cos2 v2t V2 cos v2t 2V22 cos2 v2t V1 cos v1t4 (6.4)
The first two terms contain frequencies f1, f2, 3f1, and 3f2. The triple frequency compo-
nents can be removed from the amplifier output with band-pass filters. The second two
terms generate the third-order IM frequency components.
We can expand the cosine squared terms using the trig identity cos2 x 12 [cos2x 1].
Hence the IM terms of interest become
VIM bV 21 V2 3cos v2t 1cos 2v1t 12 4
bV 22 V1 3cos v1t 1cos 2v2t 12 4
bV 21 V2 3cos v2t cos 2v1t cos v2t4
bV 22 V1 3cos v1t cos 2v2t cos v1t4 (6.5)
The terms at frequencies f1 and f2 add to the wanted output of the amplifier, so the third-
order intermodulation products are generated by the f1 2f2 and f2 2f1 terms.
Using another trig identity
cos x cos y cos1x y2 cos1x y2
The output of the amplifier contains IM frequency components given by
¿ bV 21 V2 3cos 12v1t v2t2 cos 12v1t v2t2 4
VIM
bV 22 V1 3cos 12v2t v1t2 cos 12v2t v1t2 4 (6.6)
We can filter out the sum terms in Eq. (6.6), but the difference terms, with frequencies
2f1 f2 and 2f2 f1 may fall within the transponder bandwidth. These two terms are
known as the third-order intermodulation products of the high-power amplifier, because
they are the only ones likely to be present at the output of a transponder which incorpo-
rates a narrow bandpass filter at its output. Thus the third-order intermodulation products
that are of concern are given by V3IM where
V3IM bV 21V2 cos 12v1t v2t2 bV22V1 cos 12v2t v1t2 (6.7)
The magnitude of the IM products depends on the parameter b, which describes the
nonlinearity of the transponder, and the magnitude of the signals. The wanted signals at
the transponder output, at frequencies f1 and f2, have magnitudes AV1 and AV2. The wanted
output from the amplifier is
Vout AV1 cos v1t AV2 cos v2t
The total power of the wanted output from the HPA, referenced to a 1 ohm load, is therefore
Pout 12 A2V 21 12 A2V 22 A2 1P1 P2 2 W (6.8)
where P1 and P2 are the power levels of the wanted signals. The power of the IM products
at the output of the HPA is
PIM 2 1 12 b2V 61 12 b2V 62 2 b2 1P 31 P 32 2 W (6.9)
c06.qxd 09/08/02 14:18 Page 228
It can be seen that IM products increase in proportion to the cubes of the signal powers,
with power levels that depend on the ratio (bA)2. The greater the nonlinearity of the
amplifier (larger bA ratio), the larger the IM products.
Intermodulation Example
Consider the case of a 36-MHz bandwidth C-band transponder which has an output spec-
trum for downlink signals in the frequency range 3705–3741 MHz. The transponder car-
ries two unmodulated carriers at 3718 and 3728 MHz with equal magnitudes at the input
to the HPA. Using Eq. (6.7), the output of the HPA will contain additional frequency com-
ponents at frequencies
f31 12 3718 37282 3708 MHz
f32 12 3728 37182 3738 MHz
Both of the IM frequencies are within the transponder bandwidth and will therefore
be present in an earth station receiver that is set to the frequency of this transponder. The
magnitude of the IM products will depend on the ratio bA, a measure of the nonlinearity
of the HPA, and on the actual level of the two signals in the transponder.
Now consider the case where the two signals carry modulation which spreads the
signal energy into a bandwidth of 8 MHz around each carrier. Carrier 1 has frequencies
3714 to 3722 MHz and carrier 2 has frequencies 3726 to 3734 MHz. Denoting the band
of frequencies occupied by the signals as fnlo to fnhi, the intermodulation products cover
the frequency bands
12 f1lo f2hi 2 to 12 f1hi f2lo 2 and 12 f2lo f1hi 2 to 12 f2hi f1lo 2.
The IM products are spread over bandwidths (2B1 B2) and (2B2 B1).
Hence the third-order IM products for this example cover these frequencies:
3706 3730 MHz and 3716 3740 MHz with bandwidths of 24 MHz.
The location of the 8 MHz wide signals and 24 MHz wide IM products is illustrated in
Figure 6.4. The intermodulation products now interfere with both signals, and also cover
the empty frequency space in the transponder.
Third-order IM products grow rapidly as the output of the transponder increases to-
ward saturation. Equation (6.9) shows that IM power increases as the cube of signal power:
in decibel units, every 10 dB increase in signal power causes a 30 dB increase in IM prod-
uct power. Consequently, the easiest way to reduce IM problems is to reduce the level of
the signals in the HPA. The output power of an operating transponder is related to its
Carrier 1 Carrier 2
IM products IM products
3705 3714 3722 3726 3734 3741

Frequency in MHz
FIGURE 6.4 Intermodulation between two C-band carriers in a
transponder with third-order nonlinearity.
c06.qxd 09/08/02 14:18 Page 229
saturated output power by output backoff. Backoff is measured in decibel units, so a

transponder with a 50 W rated (saturated) output power operating with an output power
of 25 W has output backoff of 17 dBW 14 dBW 3 dB. Intermodulation products are
reduced by 9 dB when 3 dB backoff is applied, so any nonlinear transponder carrying
more than one signal will usually have some backoff applied. Since a transponder is an
amplifier, the output power level is controlled by the input power, and there is a saturated
input power level corresponding to the saturated output level. When the transponder is op-
erated with output backoff, the power level at its input is reduced by the input backoff.
Because the transponder characteristics are not linear, input backoff is always larger than
output backoff. Figure 6.5 illustrates the operating point and input and output backoff for
a transponder with a nonlinear TWTA. The nonlinearity of the transponder causes the in-
put and output backoff values to be unequal. In the example shown in Figure 6.5, the
transponder saturates at an input power of 100 dBW. The transponder is operated at an
input power of 102.2 dBW, giving an input backoff of 2.2 dB. The corresponding out-
put backoff is 1.0 dB, giving an output power of 16 dBW (40 W), 10 W below the saturated
output power of 50 W (17 dBW).
Note that the TWTA has slightly different characteristics when operated with a single
carrier and multiple carriers. The generation of intermodulation products when multiple
carriers are present robs the wanted output of some of the transponder output power. For
the nonlinearity shown in Figure 6.5, the reduction in output power is 0.6 dB at saturation.
In the example above, both carriers had equal power. If the powers are unequal, the
weaker signal may be swamped by intermod products from the stronger carrier. This can
be seen from Eq. (6.9); the IM products that tend to affect Carrier 1 have voltages
proportional to the square of the voltage of Carrier 2.
Pout
20
Outback backoff Single

Saturated carrier
output
17 dBW
Multicarrier
15
Operating
point
Transponder Input backoff
output
power
dBW
Input
saturation
10
Pin
−105 −100
Transponder input power dBW
FIGURE 6.5 Typical input–ouput characteristic of a transponder using a traveling wave
tube amplifier (TWTA). To maintain quasi-linear operation and minimize intermodulation
problems with multiple carriers, the transponder must be operated below its saturated
output power level. In this example, the saturated output power from the transponder is
50 W (17 dBW) with a single carrier. The saturated output power is slightly lower with
multiple carriers because some output power is converted to intermodulation products.
c06.qxd 09/08/02 14:18 Page 230
Operation of a nonlinear transponder with multiple carriers requires careful bal-

ancing of the power levels of each carrier so that intermodulation products are evenly
spread across the transponder’s bandwidth. Judicious spacing of the carriers can be used
to place the highest intermods in gaps between carriers. The process is known as loading
the transponder. Sophisticated computer programs are used by satellite operators to opti-
mize the backoff level of a transponder such that intermodulation is minimized while
output power is maximized. When a very large number of carriers access a transponder
using FDMA, as might happen with a network of VSAT stations or a transponder used
with mobile satellite telephones, the transponder must operate in a quasi-linear region of
its characteristics. Quasi-linear means almost linear, either by equalization or by the
application of a large output backoff.
Earth station HPAs can also cause intermodulation if they carry multiple carriers
and operate close to saturation. In large earth stations where multiple carriers are more
likely to be transmitted, the HPA is often rated at a much higher level than the expected
transmit power. This allows substantial backoff to be used, keeping the amplifier in its
linear region.
In the above analysis of third-order intermodulation, only two carriers were con-
sidered. If there are three (or more) carriers present in a nonlinear transponder, inter-
modulation products at frequencies such as f1 f2 f3 can be generated that are likely
to be within the transponder bandwidth. When many carriers are present, as with a
transponder carrying narrowband SCPC signals, there will be a very large number of IM
products, making quasi-linear operation essential.
Calculation of C/N with Intermodulation

Intermodulation between carriers in a nonlinear transponder adds unwanted products into
the transponder bandwidth that are treated as though the interference were Gaussian noise.
For wideband carriers, the behavior of the IM products will be noiselike; with narrowband
carriers, the assumption may not be accurate, but is applied because of the difficulty of
determining the exact nature of the IM products.
The output backoff of a transponder reduces the output power level of all carriers,
which therefore reduces the (CN) ratio in the transponder. The transponder CN ratio
appears as (CN)up in the calculation of the overall (CN)0 ratio in the earth station receiver.
IM noise in the transponder is defined by another CN ratio, (CN)IM, which enters the
overall (CN)0 ratio through the reciprocal formula (using linear CN power ratios)
1C N2 0 1 31 1CN2 up 1 1CN2 dn 1 1CN2 IM 4 (6.10)
Techniques for the calculation of (CN)IM are beyond the scope of this text. Full knowl-
edge of the transponder nonlinearity and the signals carried by the transponder is required
to permit (CN)IM to be calculated5,6.
There is an optimum output backoff for any nonlinear transponder operating in
FDMA mode. Figure 6.6 illustrates the effect of the HPA operating point on each CN
ratio in Eq. (6.10) when the operating point is set by the power transmitted by the uplink
earth station. The uplink (CN)up ratio increases linearly as the transponder input power
is increased, leading to a corresponding nonlinear increase in transponder output power.
As the nonlinear region of the transponder is reached, the downlink (CN)dn ratio increases
less rapidly than (CN)up because the nonlinear transponder is going into saturation.
Intermodulation products start to appear as the nonlinear region is approached, increas-
ing rapidly as saturation is reached. With a third-order model for nonlinearity, the inter-
modulation products increase in power at three times the rate at which the input power to
c06.qxd 09/08/02 14:18 Page 231
C/N
FIGURE 6.6 CN ratios for a link
30 using the nonlinear transponder illus-
(C/N)IM trated in Figure 6.5. Overall (CN)0 at
the receiving earth station is the com-
bination of the three CN ratios shown
25
(C/N)up in this figure. As the power level at the
input of the transponder is increased,
C/N ratio in dB
(CN)up in the transponder increases

20 linearly, but (CN)dn in the earth station
receiver increases less rapidly as the
transponder saturates. Third-order in-
(C/N)dn termodulation products are generated
15 in the transponder as it saturates,
causing overall (CN)0 to peak at an
(C/N)o input level of 104 dBW. This is the
optimum operating point for this
transponder. The dashed lines show
−110 −105 −100 Pin CN ratios for a transponder that does
Transponder input power dBW not saturate.
the transponder is increases, causing (CN)IM to decrease rapidly as saturation is ap-

proached. When all three CN ratios are combined through Eq. (6.10), the overall (CN)0
ratio in the receiving earth station receiver has a maximum value at an input power level
of 104 dBW in the example in Figure 6.6. This is the optimum operating point for the
transponder. The optimum operating point may be many decibels below the saturated
output level of the transponder under some conditions5.
VSAT networks and mobile satellite telephones often use single channel per carrier
(SCPC) FDMA to share transponder bandwidth. Because the carriers are narrowband, in
the 10 to 128 kHz range typically, a 36 or 54 MHz transponder may carry many hundreds
of carriers simultaneously. The balance between the power levels of the carriers may not
be maintained, especially in a system with mobile transmitters that can be subject to fad-
ing. The transponder must operate in a linear mode for such systems to be feasible, either
by the use of a linear transponder or by applying large output backoff to force operation
of the transponder into its linear region.
EXAMPLE 6.2.1 Power Sharing in FDMA

Three identical large earth stations with 500 W saturated output power transmitters access a 36 MHz
bandwidth transponder using FDMA. The transponder saturated output power is 40 W and it is op-
erated with 3 dB output backoff when FDMA is used. The gain of the transponder is 105 dB in its
linear range. The bandwidths of the earth station signals are
Station A: 15 MHz
Station B: 10 MHz
Station C: 5 MHz
Find the power level at the output of the transponder, and at the input to the transponder, in
dBW, for each earth station signal, assuming that the transponder is operating in its linear region
with 3 dB output backoff. Each earth station must transmit 250 W to achieve an output power of
20 W from the transponder. Find the transmit power for each earth station when the transponder is
operated with FDMA by the three earth stations.
c06.qxd 09/08/02 14:18 Page 232
The output power of the transponder must be shared between the three signals in proportion
to their bandwidths. The output backoff of 3 dB means that the output power from the transponder
is Pt where
Pt 10 log10 40 3 16 3 dBW
13 dBW 20 W
The total bandwidth used is 15 10 5 30 MHz. The output power must be shared in
proportion to bandwidth used, so the transponder output power allocated to each earth station’s
signal is
Station A: B 15 MHz Pt 1530 20 W 10.0 W 10 dBW
Station B: B 10 MHz Pt 1030 20 W 6.67 W 8.2 dBW
Station C: B 5 MHz Pt 1530 20 W 3.33 W 5.2 dBW
The transponder gain is 105 dB, in its linear range, so for linear operation the transponder
input power for each earth station signal is
Station A: Pin 10.0 105 95.0 dBW
Station B: Pin 8.2 105 96.8 dBW
Station C: Pin 5.2 105 99.8 dBW
The EIRP at each earth station must be set to give the correct input power at the input to the
transponder. A single earth station must transmit 250 W 24 dBW to achieve a transponder out-
put of 20 W. For the transponder output power levels of each signal calculated above, the earth sta-
tion transmitter powers are
Station A: Pt 24.0 3.0 21.0 dBW 126 W
Station B: Pt 24.0 4.8 19.2 dBW 83 W
Station C: Pt 24.0 7.8 16.2 dBW 42 W
EXAMPLE 6.2.2 Channel Capacity with Demand Access FDMA

A large number of satellite telephones can access a single transponder on an LEO satellite using
FDMA-DA. Data transmitted from the satellite on initial access by the telephone is used to set the
transmit frequency and output power of the satellite telephone. The telephones transmit BPSK sig-
nals in L band with an occupied bandwidth of 12 kHz and an output power level between 0.05 and
0.5 W, such that the power level at the input to the transponder is always 144 dBW for any up-
link signal. The resulting CN ratio in clear air conditions for any one signal in the transponder is
16 dB. The transponder has a bandwidth of 1.0 MHz, a gain of 134 dB, and a maximum permitted
output power of 5 W. The center frequencies of the telephone transmitters are spaced 16 kHz apart
to provide a 4 kHz guard band between each signal.
Determine the maximum number of satellite telephones which can simultaneously access the
transponder. Is the transponder power or bandwidth limited? If the transponder is power limited,
what change could be made to increase the number of signals the transponder carries? What effect
would the change have on overall (CN)0 for the link?
If the transponder is bandwidth limited, the maximum number of signals, Nmax, that it could
carry is the available bandwidth divided by the signal bandwidth plus the guard band width
Nmax 1000 kHz16 kHz 62
The value of Nmax must be rounded down to the next lowest integer because we cannot send
fractional signals.
The power level of each signal at the input to the transponder is 144 dBW. The gain of the
transponder is 134 dB, so the output power for each signal is
Pt 144 134 10 dBW 0.1 W
c06.qxd 09/08/02 14:18 Page 233
6.3 TIME DIVISION MULTIPLE ACCESS (TDMA) 233
If we have 62 signals, each at a power level of 0.1 W, the total power at the output of the
transponder is 6.2 W. This exceeds the maximum permitted output power of the transponder, which
was set at 5.0 W. Hence the maximum number of satellite telephones that can simultaneously access
the transponder is 50, and the transponder is power limited.
We can increase the number of signals in the transponder to 62, which is the maximum pos-
sible number of telephones that can share the transponder at the same time because of the band-
width limit, by reducing the input power level by 10 log10 (6250) 0.9 dB. Then the output power
from the transponder, per signal, is
Pt 144.8 134 10.9 dBW 0.081 W
We can now transmit 62 signals from the transponder with a total output power level of 62
0.081 5.0 W, which meets the power limitation for the transponder.
The CN ratio in the transponder will be reduced by 0.9 dB because the input signal is 0.9
dB weaker. Hence (CN)up 16 0.9 15.1 dB. The transponder now transmits 0.9 dB less
power per signal, which will reduce the (CN)dn ratio at the receiving earth station by 0.9 dB. Hence
the overall (CN)0 ratio for the link will be reduced by 0.9 dB when the number of satellite tele-
phones sharing the transponder is increased from 50 to 62.
6.3 TIME DIVISION MULTIPLE ACCESS (TDMA)
In time division multiple access a number of earth stations take turns transmitting bursts
of RF signals through a transponder. Since all practical TDMA systems are digital, TDMA
has all the advantages over FDMA that digital signals have over analog. TDMA systems,
because the signals are digital and can be divided by time, are easily reconfigured for
changing traffic demands, are resistant to noise and interference, and can readily handle
mixed voice, video, and data traffic. One major advantage of TDMA when using the en-
tire bandwidth of a transponder is that only one signal is present in the transponder at one
time, thus overcoming many of the problems caused by nonlinear transponders operating
with FDMA. However, using all of the transponder bandwidth requires every earth sta-
tion to transmit at a high bit rate, which requires high transmitter power, and TDMA is
not well suited to narrowband signals from small earth stations. Nonlinearity in the
transponder can cause an increase in intersymbol interference with digital carriers;
equalizers can be used at the receiving earth stations to mitigate the effect.
Many of the concepts developed in Chapter 5 for time division multiplexing (TDM)
also apply to TDMA. The difference between TDM and TDMA is that TDM is a base-
band technique used at one location (for example, a transmitting earth station) to multi-
plex several digital bit streams into a single higher speed digital signal. Groups of bits are
taken from each of the bit streams and formed into baseband packets or frames that also
contain synchronization and identification bits. At a receiving earth station, the high-speed
bit stream must first be recovered using the techniques discussed in Chapter 5, which re-
quires demodulation of the RF carrier, generation of a bit clock, sampling of the received
waveform, and recovery of the bits. The synchronization bits or words in the packets or
frames must then be found so that the high-speed bit stream can be split into its original
lower speed signals. The clock frequency for the bit stream is fixed, and the frame length
is usually constant. Packet lengths can vary, however, which is the main difference be-
tween frames and packets. The entire process requires considerable storage of bits so that
the original signals can be rebuilt, leading to delays in transmission. In a GEO satellite
system, the largest delay is always the transmission time to the satellite and back to earth,
typically 240 ms. The transmission delay is unavoidable, but any additional delays should
be minimized.
c06.qxd 09/08/02 14:18 Page 234
Satellite
TDMA stream
from satellite
One frame
Bursts from each

earth station
Incoming bit
streams Earth terminals
FIGURE 6.7 Illustration of TDMA with three earth stations. Transmitting earth stations
must time their burst transmissions so that they arrive at the satellite in the correct se-
quence. The signal transmitted by the satellite is a continuous sequence of bursts separated
by short guard times.
TDMA is an RF multiple access technique that allows a single transponder to be

shared in time between RF carriers from different earth stations. In a TDMA system, the
RF carrier from each earth station sharing a transponder is sent as a burst at a specific
time. At the satellite, bursts from different earth stations arrive sequentially, so the transpon-
der carries a near continuous signal made up of a sequence of short bursts coming from
different earth stations. The principle of TDMA is illustrated in Figure 6.7.
The burst transmission is assembled at a transmitting earth station so that it will cor-
rectly fit into the TDMA frame at the satellite. The frame has a length from 125 s to many
milliseconds, and the burst from the earth station must be transmitted at the correct time to
arrive at the satellite in the correct position within the TDMA frame. This requires syn-
chronization of all the earth stations in a TDMA network, adding considerable complexity
to the equipment at the transmitting station. Each station must know exactly when to trans-
mit, typically within a microsecond, so that the RF bursts arriving at the satellite from dif-
ferent earth stations do not overlap. (A time overlap of two RF signals is called a collision
and results in data in both signals being lost. Collisions must not be allowed to occur in a
TDMA system.)
A receiving earth station must synchronize its receiver to each of the sequential
bursts in the TDMA signal and recover the transmission from each uplink earth station.
The uplink transmissions are then broken down to extract the data bits, which are stored
and reassembled into their original bit streams for onward transmission. The individual
transmissions from different uplink earth stations are usually sent using BPSK or QPSK,
and will inevitably have small differences in carrier and clock frequencies, and different
carrier phases. The receiving earth station must synchronize its PSK demodulator to each
burst of signal within a few microseconds, and then synchronize its bit clock in the next
few microseconds so that a bit stream can be recovered. In high-speed TDMA systems,
operating at 120 Mbps, for example, these are demanding requirements.
Bits, Symbols, and Channels

A potential source of confusion in the discussion of TDMA systems is that QPSK (or pos-
sibly QAM) modulation is typically used by transmitting earth stations, and data rates can
c06.qxd 09/08/02 14:18 Page 235
then be described either by bit rate or symbol rate. Both bit rates and symbol rates need
to be used in the discussion of digital radio transmission and TDMA systems, so the reader
must be clear on the distinction between a bit and a symbol.
A bit is the fundamental unit in digital transmission. Data are generated by termi-
nals (e.g., a personal computer) as bits, or by conversion of an analog speech or video
signal to digital form as a serial bit stream. The bit stream is described by its bit rate, in
bits per second, bps, thousands of bits per second, kbps, or millions of bits per second,
Mbps. (Note that the k and M prefixes are in units of 103 and 106, not the binary digital
version of 1024 and 1,048,576.) The bit stream must be modulated onto an RF carrier for
transmission to the satellite. Phase shift keying is invariably used as the modulation tech-
nique. In binary phase shift keying, BPSK, the logical data states of the bits, 1 and 0, are
converted into two opposite phase states of the RF carrier, say 0° and 180°. In quadrat-
ure phase shift keying, QPSK, two bits at a time are converted into one of four phase states
of the RF carrier (see Chapter 5 for details of QPSK). The state of the RF carrier is called
a symbol, and the symbol rate is in units of bauds, or symbols per second. For BPSK, bit
rate and baud rate are the same. For QPSK, the baud rate (symbol rate) is one half the bit
rate. The importance of symbol rate in any digital radio system is that it is the symbol
rate, not the bit rate, that determines the bandwidth of the RF signal, and consequently
the bandwidth of the filters in the receiver.
Occasionally, QAM modulation may be used on a satellite link. In QAM, carrier
symbols are generated from the four phase states of QPSK, but can also have different
amplitudes. This allows one symbol to convey more than 2 bits, which reduces the RF
bandwidth required for a given bit rate. However, an increased (CN)0 ratio is required in
the receiver to recover bits from QAM signals, so QAM can only be used in satellite links
with higher than usual (CN)0 ratios.
TDMA Frame Structure

A TDMA frame contains the signals transmitted by all of the earth station in a TDMA
network. It has a fixed length, and is built up from the burst transmissions of each earth
station, with guard times between each burst. The frame exists only in the satellite transpon-
der and on the downlinks from the satellite to the receiving earth stations. Figure 6.8 shows
a simplified diagram of a TDMA frame for four transmitting earth stations. Each station
transmits a preamble that contains synchronization and other data essential to the opera-
tion of the network before sending data. The earth station’s transmission is followed by
a guard time to avoid possible overlap of the following transmission. In GEO satellite sys-
tems, frame lengths of 125 s up to 20 ms have been used, although 2 ms has been widely
used by stations using Intelsat satellites. Earth stations must be able to join the network,
add their bursts to the TDMA frame in the correct time sequence, and leave the network
Next frame
Frame period T, µs
Stn 1 Stn 2 Stn 3 Stn 4 Stn 1
Guard time
FIGURE 6.8 TDMA
Traffic: N bits
frame with four transmit-
Preamble ting earth stations.
c06.qxd 09/08/02 14:18 Page 236
without disrupting its operation. They must also be able to track changes in the timing of
the frame caused by motion of the satellite toward or away from the earth station. Each
earth station must also be able to extract the data bits and other information from burst
transmissions of other earth stations in the TDMA network. The transmitted bursts must
contain synchronization and identification information that help receiving earth stations
to extract the required information without error.
These goals are achieved by dividing TDMA transmissions into two parts: a pre-
amble containing all the synchronization and identification data, and a group of traffic bits.
Synchronization of the TDMA network is achieved with the portion of the preamble trans-
mitted by each earth station that contains carrier and bit clock synchronization waveforms.
In some systems, a separate reference burst may be transmitted by one of the stations,
designated as the master station. A reference burst is a preamble followed by no traffic
bits. Traffic bits are the revenue producing portion of each frame, and the preamble and
reference bursts represent overhead. The smaller the overhead, the more efficient the
TDMA system, but the greater the difficulty of acquiring and maintaining network syn-
chronization. The preamble of each station’s burst transmission requires a fixed trans-
mission time. A longer frame contains proportionally less preamble time than a short
frame, so more revenue producing data bits can be carried in a long frame. Early TDMA
systems were designed around 125 s frames, to match the sample rate of digital speech
in telephone systems, in exactly the same way that T1 24-channel systems operate. A dig-
ital telephone channel generates one 8-bit digital word every 125 s (8 kHz sampling rate),
so a 125-s frame transmits one word from each speech channel. However, it is more ef-
ficient to lengthen the frame to 2 ms or longer so that the proportion of overhead to mes-
sage transmission time is reduced. It must be remembered that a longer frame requires
multiple 8-bit words when transmitting digital speech. For example, in a time period of
2 ms, a digital terrestrial channel will deliver sixteen 8-bit words, so a 2-ms TDMA frame
requires sixteen 8-bit words from each terrestrial channel in each transmitted burst.
Figure 6.9 shows a typical TDMA frame with 2.0 ms duration used by some earth
stations operating in TDMA through Intelsat satellites. All of the blocks at the start of the
CBTR UW TTY SC VOW VOW Digital speech channels
M satellite channels
1 2 3 4 5 6 7 M
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Sixteen 8-bit samples form each satellite channel

FIGURE 6.9 Structure of an Intelsat traffic data burst. A satellite channel is a block of six-
teen 8-bit samples from one terrestrial speech channel. Other blocks in the traffic burst are
used to synchronize the PSK demodulator, the bit clock, and the frame clock in the receiver
(CBTR, UW) and to provide communication links between earth stations (TTY, SC, and VOW).
CBTR, carrier and bit timing recovery; UW, unique word; TTY, teletype; SC, satellite channel;
VOW, voice order wire.
c06.qxd 09/08/02 14:18 Page 237
frame, labeled CBTR through VOW, are preamble. Speech channel data transmission be-
gins with satellite channel 1 and continues as a serial bit stream through channel M. A
satellite channel is made up of the number of bits delivered to the earth station by a sin-
gle digital speech channel during one frame period. The frame can equally well send dig-
ital data of any form as a serial stream of bits occupying the space taken up by M satellite
channels.
For the specific case of digital speech channels using serial transmission at a rate
rsp, the number of speech channels, n, that can be transmitted in a TDMA frame shared
equally by N earth stations can be calculated from the duration of the frame, Tframe in sec-
onds, the guard time and preamble length, tg and tpre, in seconds, and the transmitted bit
rate of the TDMA system, Rb. The time, Td, available in each station burst for transmission
of data bits is
Td 3Tframe N1tg tpre 2 4 N seconds (6.11)
In 1 s, the total number of bits, Cb, transmitted by each earth station is
Cb 3Tframe N1tg tpre 2 4 RbTframe (6.12)
Since each digital speech channel requires a continuous bit rate of rsp bps, the number of
speech channels that can be carried by each earth station is given by n where
n 3Tframe N1tg tpre 2 4

Rb
(6.13)
Tframe rsp
EXAMPLE 6.3.1 TDMA in a Fixed Station Network

A TDMA network of five earth stations shares a single transponder equally. The frame duration is
2.0 ms, the preamble time per station is 20 s, and guard bands of 5 s are used between bursts.
Transmission bursts are QPSK at 30 Mbaud.
Calculate the number of 64 kbps voice channels that each TDMA earth station can
transmit. If the earth stations send data rather than digital speech, what is the transmission rate
of each earth station in Mbps? What is the efficiency of the TDMA system expressed as
Efficiency 100% Message bits sentMaximum possible number of bits that could be sent?
Using Eq. (6.11) we can find the data burst duration for each earth station, Td, in
microseconds
Td 32000 5 15 202 4 5 375 ms
A burst transmission rate of 30 Mbaud is 30 million symbols per second, and QPSK symbols
carry 2 bits. Hence the transmitted bit rate in each burst is Rb 2 30 Mbps 60 Mbps.
The capacity of each earth station in bits per second is Cb where
Cb 375 60 1062000 11.25 Mbps
Since these bits are sent as 500 bursts each second, the number of message bits per burst is
11.25 106500 225,000 kbits.
The number of 64 kbps digital speech channels that can be carried by one earth station is
n 11,250,00064,000 1751.7812
We have to discard the fractional channel since we can send only whole channels. The
0.781 fraction of a speech channel represents 50,000 bits per second that cannot be transmitted
by each earth station, or 100 bits per burst. At a bit rate of 60 Mbps, 100 bits is equivalent to
1.67 s, so the guard time per burst would increase from 5 to 6.67 s when 175 speech chan-
nels are sent.
c06.qxd 09/08/02 14:18 Page 238
It is always a good idea to check the answer to any problem against a reference
value. If the earth stations transmitted without any guard times or preambles, the maximum
number of 64-kbps speech channels that could be sent by each of the five earth sta-
tions using a transmission burst rate of 60 Mbps would be n Rb(N rs) 60
106(5 64 103) 187.5. The calculated capacity per station of 175 channels is a
little lower than the maximum number of 187 channels; the difference—12 channels in
this case—is the number of speech channels lost by the need to include guard times and
preambles in the TDMA frame structure.
If we send data rather than digital speech, each earth station can transmit 11.25
Mbps. Overhead accounts for the equivalent of 0.75 Mbps, leading to an efficiency of
Efficiency 100% 11.2512.0 93.75%
Example 6.3.1 shows that there are diminishing returns to be obtained by increas-
ing the duration of the frame beyond 2.0 ms in this system. There is a 6.25% loss of po-
tential data transmission time when the 2.0-ms frame length of Example 6.3.1 is used with
the other parameters of this TDMA system. Doubling the frame length to 4.0 ms would
reduce the loss to about 3%, but would add additional delay, and increase the complex-
ity of the earth station equipment. In this case, a 2.0 ms frame period appears to be a good
compromise between delay, complexity, and efficiency.
Reference Burst and Preamble

Figure 6.9 shows the baseband content of a typical TDMA burst. The segments marked
CBTR and UW contain the carrier recovery waveform, bit clock synchronization,
unique word, and station identifier. The remaining blocks are part of the preamble
and will be described below. CBTR is the portion of the burst from a given earth sta-
tion that enables a receiving earth station to recover the remainder of the burst. CBTR
stands for carrier and bit timing recovery. Carrier recovery is required at the receiver
of any radio link in which coherent phase shift keying is the modulation technique.
A local carrier must be generated in the IF portion of the receiver from the received
BPSK or QPSK signal. This is typically achieved with some form of phase locked loop
(PLL), which must lock up quickly in a TDMA system. The local carrier drives the
multipliers in the demodulator to achieve coherent demodulation of the PSK signal.
(See Chapter 5 for details of the modulation and demodulation process in PSK systems.)
Once carrier phase lock is achieved, the demodulator will produce a baseband wave-
form corresponding to the bits that were modulated onto the carrier at the transmitting
earth station.
In a TDMA system, the bursts of RF signal received sequentially from different
earth stations do not have the same carrier frequency, phase, or bit rate. The differences
will be small, but sufficient to require the receiver to relock to each new carrier, and to
resynchronize the bit clock. The CBTR contains a sequence of predetermined signals that
ensure rapid lock to the carrier and fast synchronization of the bit clock. The carrier re-
covery portion of the CBTR sequence may consist of unmodulated carrier followed by a
number of symbols that follow a specific pattern, or the entire CBTR burst may be mod-
ulated. The first part of the CBTR burst is used to obtain lockup of the PLL, and the re-
maining portion is used for bit clock synchronization. In Figure 6.9, the CBTR burst has
a duration of 176 symbols with a transmission rate of 30 Msps, giving a burst duration of
5.86 s. Within this very short time period the carrier recovery circuit must achieve pre-
cise lock on the received signal and the bit clock must be brought into synchronization.
Carrier phase lock and bit synchronization must be achieved within the CBTR burst time
c06.qxd 09/08/02 14:18 Page 239
even when the carrier-to-noise ratio is very low, perhaps as low as 5 dB in VSAT links
employing half rate forward error correction to improve the BER of the received data.
Typical TDMA systems using QPSK send the first 48 or 50 symbols of the CBTR
burst as all ones or all zeroes on both the I and Q phases of the carrier, which correspond
to sending an unmodulated carrier, followed by a sequence of ones and zeros in both chan-
nels7,8. The CBTR burst is followed by a unique word (UW) of typically 20 to 48 bits
which serves several purposes. It acts as a transmit station identifier, a start of frame (SOF)
or burst marker, and as a carrier phase ambiguity detector.
Unique Word
The received bit sequence at the demodulator output is continuously run through a cor-
relator which looks for the appearance of the unique word in the bit stream. Figure 6.10
illustrates a simplified unique word correlator. The correlator is looking for a match be-
tween the incoming bit pattern and one of four stored sequences, corresponding to the
correct UW sequence and three variants in which the I and Q bits of the sequence are
inverted due to phase ambiguity in the carrier recovery circuit. QPSK carrier recovery can
result in ambiguity if the carrier recovery circuit locks up in the incorrect phase, which is
possible in many QPSK demodulators. When this happens, one or both of the I and Q bit
streams is inverted. A known bit sequence is required in the received signal for ambigu-
ity resolution. The pattern of ones and zeroes in the CBTR sequence and the unique word
allow the receiver to check for phase ambiguity and to invert the appropriate bit stream
(I, Q, or both) if ambiguity is found. In the correlator diagram shown in Figure 6.10, when
the correct UW sequence is present in the correlator in the correct position (i.e., at the
correct time within the incoming signal burst) the correlator output will maximize and
trigger the threshold detector.
When a new RF burst is received, the carrier recovery circuit locks up the local
carrier PLL, and the bit clock then synchronizes to the bit rate of the received signal.
Input from Shift register

receiver
Summer
Σ
Output to
threshold
Bit clock
circuit
0 1 0 1 1 1 Modulo 2 multipliers
(exclusive OR gates)
Unique word memory
FIGURE 6.10 Unique word correlator. The example shown here has a 6 bit unique word
for illustration—practical satellite systems use unique words of 24–48 bits. The bits stream
from the receiver output is clocked into the shift register serially. When the contents of the
shift register match the stored unique word the output of the summer is a maximum and
exceeds the threshold, marking the end of the unique word. This provides a time marker for
the remainder of the earth station’s transmission.
c06.qxd 09/08/02 14:18 Page 240
Bits then begin to flow into the correlator, which detects one of the four possible forms
of the UW and sets logical inverters that invert the appropriate bits, if necessary. The
resulting bit stream after the end of the UW is then output correctly and can be used
by the receiver. The time at which the threshold detector is triggered marks a known
point in the TDMA burst—the end of the UW sequence. This time is critical, because
all subsequent bits from the demodulator will be demultiplexed based on a count that
begins when the UW is detected. If the UW is detected at the wrong time, the recov-
ered data in the entire burst will be scrambled and the burst is lost. The UW and the
correlator circuits must therefore be designed to ensure that the UW is detected cor-
rectly in every burst with a very low probability of a timing error. An incorrectly
detected UW is known as a miss, and the probability that a miss occurs can be calcu-
lated from the bit error probability (BER) of the recovered bit stream and the length
of the UW.
The length of the UW and the correlator design are important factors in the de-
sign of a TDMA system. The detection of the UW occurs when there is a match be-
tween the UW output by the demodulator and the stored sequence. However, if the CN
ratio of the RF signal is low, there will be bit errors present in the UW, so precise cor-
relation cannot be guaranteed. A long UW allows a predetermined number of bits to be
in error while still ensuring that correct timing of the bit sequence is achieved. The prob-
ability that the specific bit sequence of a UW appears within the bit stream of traffic
data must also be low, so that the probability of a false alarm is small. A false alarm
occurs when a UW is inadvertently detected in the traffic data, causing the timing of
the burst to be reset. There are several ways that false alarms can be prevented, but the
use of a long UW reduces the likelihood of a false alarm occurring and is therefore
desirable. Once the time position of a UW within the TDMA frame is determined, a
window can be placed over the UW so that the correlator is operated only during a pe-
riod slightly longer than the UW duration. This will greatly reduce the chances of a
false alarm.
As an example of the probability of a missed detection, a 24 bit UW with a bit
error rate of 103 can be recovered with a miss probability of 1010 when three of
the bits in the UW are allowed to be in error, corresponding to a detection threshold
of 21249.
Although a miss probability of 1010 may appear sufficiently small at first sight, a
large number of bits is lost whenever a UW is detected incorrectly. The TDMA system
with five earth stations in Example 6.3.1 used a 2 ms frame with a traffic data time of
375 s per station. At a bit rate of 60 Mbps there are 22,500 traffic bits in each burst.
With a probability of missing the UW of 1010, the average BER of the traffic data re-
sulting from one missed UW is 2.25 108. This BER is too high to be acceptable in
many applications, so a longer UW would have to be used to lower the probability of a
missed detection. For further details of unique word detection and TDMA burst design,
the reader is referred to one of the references9–11.
There must be as many UW correlators as there are TDMA uplink stations in the
network when the UW is also used as a station identifier. If there is a large number of
stations it is simpler to use a single UW and a separate station identifier word. The bit
stream for each received burst is tagged with a station identifier so that the demultiplexers
in the earth station can route the data accordingly.
The remaining segments of the preamble provide the receiving earth station with bit
streams that are used in the management of the TDMA system. Referring to Figure 6.9,
there are groups of bits identified as TTY, SC, VOW. Sixteen bits (eight QPSK symbols)
in each burst are allocated to a teletype link (TTY) between the earth stations, and a
c06.qxd 09/08/02 14:18 Page 241
further 16 bits to a service channel (SC). There are two 64-bit voice order wire (VOW)
segments that are used in digital voice links between the earth stations. The TDMA burst
shown in Figure 6.9 is designed for use in a network of large earth stations in which
there are personnel at the earth stations or network control centers. The teletype and voice
channels within the preamble of each burst provide a closed communication network be-
tween the control stations that is used to manage the TDMA system. In a VSAT network
using TDMA, for example, there would be minimal communication requirement between
terminals for management purposes, so a VSAT TDMA burst would omit most of these
segments.
The final part of the TDMA burst carries the traffic data. In Example 6.3.1 which
used a 2.0 ms frame, speech data could be carried in satellite channels as sixteen 8-bit
words from each terrestrial channel, giving a satellite channel 128 bits in each burst.
The bits associated with each satellite channel are demultiplexed by counting the 128
bits of each satellite channel, beginning at the end of the preamble. Once again, the bit
timing established by the unique word becomes all important in extracting the satellite
channels correctly.
There are many different formats for preambles in TDMA systems, depending
on the design of the particular system. Figure 6.9 illustrates a TDMA preamble structure
designed for a large fixed network with high-speed bit streams. A network of mobile
earth terminals using TDMA would have different requirements and would require
a different preamble structure. However, the earlier segments that control carrier and
bit timing recovery, phase ambiguity removal, and station identification must always
be present.
A large earth station carrying high-speed data must link into a terrestrial data net-
work to deliver received bits to customers, and to receive incoming data for transmission
over the satellite link. The satellite link connects two earth stations which may be thou-
sands of kilometers apart, each interconnecting to its own high-speed terrestrial network.
The individual terrestrial networks are not synchronized, and will therefore inevitably run
at slightly different rates.
This makes the interconnection process difficult, because a data stream delivering
bits at 1.00001 Mbps cannot be connected directly to another data stream running at
0.99999 Mbps. Twenty bits would have to be discarded every second if a direct con-
nection were made, which upsets the customer who is paying to transfer the data. A
mechanism must be developed that allows for a difference in bit rates at the two ends
of the link. The usual solution is to run the bit clock of the satellite link slightly faster
than the fastest of the terrestrial link clocks, and to allow additional bit slots for stuffing
bits. Stuffing bits are inserted when there are no data bits available from the source
because of the difference in bit rates. At the receiving end of the link, the stuffing bits
are removed and the received data stream is retimed to match the outgoing terrestrial
data channels.
Guard Times
Guard times must be provided between bursts from each earth station so that collisions
are avoided. Earth stations must transmit their bursts at precisely the correct instant so
that the burst arrives at the satellite in the correct position within the TDMA frame. This
requires burst transmission timing to microsecond accuracy and tracking of the position
of the burst within the TDMA frame by the transmitting earth station. Long guard times
make it easier for the earth stations to avoid collisions, but waste time that could be used
to send revenue-earning traffic data. Typical guard times in high-speed satellite networks
c06.qxd 09/08/02 14:18 Page 242
appear to be in the range 1 to 5 s. The transmission time between an earth station and
a GEO satellite is about 120 ms. If a 2-ms frame time is used, there are typically 60
bursts between the earth station and the satellite at any time. The bursts must arrive at
the correct time to mesh between the bursts that arrive from other earth stations. If the
satellite range from the earth station increases by 300 m, due to eccentricity or inclina-
tion of a GEO satellite’s orbit, or E–W drift in the orbital plane, the transmission delay
increases by 1 s. Thus earth stations must monitor the guard times before and after their
bursts to ensure that transmission timing is correct. In LEO satellite networks that use
TDMA, range to the satellite is changing continuously and much larger guard times are
allowed.
Synchronization in TDMA Networks

Earth stations operating in a TDMA network must transmit their RF bursts at precisely
controlled times such that bursts from each of the earth stations arrive at the satellite in
the correct sequence12. This poses two problems: how to start up a new earth station that
is joining the TDMA network, and how to maintain the correct burst timing. If the satel-
lite is in low earth orbit, or if it is a GEO satellite with a rapidly changing range, each
earth station will perceive a different carrier frequency and frame rate, and even a different
frame length. It is usual for the bit rate of transmitted bursts to be an integer multiple of
the frame rate, which means that different earth stations must transmit at slightly different
bit rates.
Maintaining synchronization with the TDMA frame is easier than initial synchro-
nization when an earth station joins a TDMA network. One station is typically desig-
nated as the master station, and may generate a reference burst to mark the start of the
frame. Each of the stations within the network has a time slot within the frame, and must
maintain its transmissions within that time slot. There are guard times at each end of
each station’s burst, which define the accuracy that the burst timing must achieve. If the
guard times are 2 s, each earth station in the network must keep its bursts timed to
within 1 s.
This is usually done by monitoring the TDMA frame at the transmitting station and
adjusting the burst timing to keep the transmitted burst in the correct time slot in the frame.
The start of the reference burst, or the start of the master station’s preamble, marks the
start of transmit frame, SOTF, which is the master timing mark for all transmissions. All
earth stations in the TDMA network synchronize their clock timing with the SOTF marker.
When an earth station monitors its own transmissions to maintain the correct burst tim-
ing, this is called satellite loop-back synchronization. The TDMA frame is established at
the satellite, so an earth station receiving the frame must subtract the transmission delay
from the satellite to the earth station to obtain the SOTF timing at the satellite. It must
then transmit its bursts ahead of the SOTF by the same delay time so that the bursts arrive
at the correct instant at the satellite. Knowledge of the range of the satellite from the earth
station is crucial in calculating delay times. The range can be calculated from the orbital
elements of the satellite, which can be determined by a control station that repeatedly
measures range to the satellite.
There are several ways that earth stations can enter a TDMA network. In fixed
networks, the precise time at which an earth station should transmit can be calculated.
Provided the calculation is accurate, the earth station can transmit a reference burst (no
traffic) timed to fall in the center of its time slot. When the frames containing the refer-
ence burst arrive back at the earth station, the actual position of the burst can be checked
c06.qxd 09/08/02 14:18 Page 243
and corrections to the timing made if necessary. The station can then transmit traffic bursts
with the correct timing.
In TDMA networks which lack sophisticated timing control, an earth station wish-
ing to join the network can transmit a CDMA sequence at a low level, at an arbitrary time.
The CDMA sequence will inevitably collide with another earth station’s traffic burst,
causing minor interference. The transmitting earth station can use a correlator to com-
press the CDMA sequence into a single timing pulse using exactly the same process as
is used to find a unique word. However, in this case the CDMA sequence is overwritten
by interference from the traffic burst with which it collided. Given a suitably long
sequence, coding gain can overcome the interference and a pulse will appear at the
correlator output marking the end of the CDMA sequence. Alternatively, the transmitting
station can use a shorter sequence and step the timing of the CDMA burst until it falls
in the empty slot allocated to that station. The position of the pulse within the TDMA
frame gives the transmitting earth station the timing information needed to transmit its
bursts at the correct time.
If the signal transmitted by the satellite cannot be monitored by the transmitting
earth station, cooperative synchronization must be used instead. This situation arises when
a satellite has multiple beams, or when satellite switched TDMA is used. A TDMA burst
received in one beam can be retransmitted by the satellite in another beam that does not
cover the transmitting station. A control station is required that can monitor the timing of
each of the earth station’s bursts as they arrive at the satellite and send out instructions to
the earth stations when changes in timing are needed. In the Intelsat TDMA system, the
control station determines a delay time, DN, for each earth station that gives the time be-
tween the start of a receive frame and the start of a transmit frame at that earth station.
The correct transmit time is then determined by the position of earth station’s burst within
the transmit frame. If transmitting earth stations fall out of sync, the control station must
send a do not transmit (DNTX) code to the station to tell it to stop transmitting because
serious loss of data will occur to other users of the network when a station sends its bursts
at the incorrect time.
In satellite switched and multiple beam satellite systems, the cooperating control
station must provide information to a new earth station that wishes to join the network.
The same techniques described above can be used, but an earth station within the receiv-
ing beam must determine the timing of test transmissions and send that information to the
transmitting station.
The availability of a global GPS time standard with better than 1-s accuracy has
made some of these tasks easier. (See Chapter 12 for details of the GPS system.)
Transmitter Power in TDMA Networks

TDMA works well in fixed networks carrying high-speed data streams. Transponders can
be more heavily loaded because less backoff is needed with TDMA; only one RF signal
is present in the transponder at any time and there is no third-order intermodulation, so
backoff is needed only to keep PM–AM conversion at an acceptable level. Burst lengths
can be made variable to accommodate stations that have different bit rates. High uplink
transmitter power is required because every station must transmit bursts at a high bit rate
and high bit rate signals occupy a wide bandwidth. Maintaining an adequate CN ratio in
the transponder forces the uplink earth station to use a high-power transmitter. For small
earth stations such as VSATs and satellite phones, this is a major disadvantage compared
to SCPC-FDMA.
c06.qxd 09/08/02 14:18 Page 244
EXAMPLE 6.3.2 TDMA in a VSAT Network

As an example, consider a typical VSAT earth station in the United States which is part of a TDMA
network using a 54 MHz bandwidth transponder on a domestic Ku-band GEO satellite. The VSAT
earth station has a 1 m antenna that transmits a single 64 kbps signal at 14 GHz. Let’s assume that
the TDMA network uses QPSK modulation and that all transmitters have a symbol rate of 30 Mbaud.
We will set (CN)up at 20 dB, and then calculate the required uplink transmit power. The following
system parameters will be used:
Earth station antenna gain 41.5 dB, satellite antenna gain (on axis) 32.0 dB, edge of
beam loss 3 dB, path loss at 14 GHz 207.1dB, receiver noise bandwidth 30 MHz, transponder
noise temperature 500 K, atmospheric and other losses 1.0 dB.
The uplink power and noise budgets are
Earth station transmit power Pt dBW
Earth station antenna gain at 11 GHz 41.5 dB
Satellite antenna 32.0 dB
Edge of beam loss 3.0 dB
Other losses 1.0 dB
Path loss at 11 GHz 207.1 dB
Power at transponder input Pt 137.6 dBW
Boltzmann’s constant 228.6 dBW/K/Hz
Transponder noise bandwidth 74.8 dBHz
Transponder noise temperature 27.0 dBK
Transponder input noise power 126.8 dBW
We require (CN)up PrkTsBn 20 dB; hence Pt 137.6 126.8 20 dBW and Pt 30.8
dBW or Pt 1200 W.
Now consider the same earth station transmitting the same 64-kbps signal in a SCPC-FDMA
VSAT network using QPSK with a symbol rate of 32 kbaud and a receiver noise bandwidth of
32 kHz. The uplink power budget is unchanged, but the noise power in the transponder, measured
in a bandwidth of 32 kHz is 156.5 dBW.
To achieve (CN)up 20 dB in the transponder now requires an uplink transmitter power of
Pt 20 137.6 156.5 1.1 dBW 1.3 W.
The above example illustrates a key problem with TDMA for any small earth sta-
tion: uplink transmit power. No one is going to equip a 1-m VSAT station with a 1200-W
transmitter. Apart from the excessive cost, FCC regulation in the United States do not allow
small VSAT stations to transmit more than 2 W to limit interference to adjacent satellites.
If we change the multiple access technique for just two earth stations, so that each
transmits a burst of QPSK signal at 64 kbaud for half the time, the uplink transmitter
power requirement is doubled to 4.1 dBW or 2.6 W. This makes wideband TDMA an un-
likely choice in VSAT networks, and limits the number of stations that can share a TDMA
frame in a low earth orbit satellite telephone system. The Iridium LEO system was de-
signed to use a hybrid TDMA–FDMA multiple access scheme at L band to combine a
small number of digital telephone transmissions into a 50-kbps QPSK signal. Similar tech-
niques are used in some VSAT networks.
EXAMPLE 6.3.3 TDMA in a Fixed Earth Station Network

In Example 6.2.1, three identical large earth stations shared a single 36-MHz bandwidth transpon-
der using FDMA. The three earth stations transmitted signals with powers and bandwidths given
c06.qxd 09/08/02 14:18 Page 245
by
Station A: B 15 MHz Pt 125 W 21.0 dBW
Station B: B 10 MHz Pt 83 W 19.2 dBW
Station C: B 5 MHz Pt 42 W 16.2 dBW
The transponder total power output was 16 dBW with 3-dB output backoff and 105-dB
transponder gain.
The three earth station accesses to the transponder are changed to TDMA, with a frame length
of 1.0 ms, a preamble time of 10 s, and a guard time of 2 s. There is no reference burst in the
TDMA frame. The signals are transmitted using QPSK, and within the earth stations the bit rates
of the signals are
Station A: Rb 15.0 Mbps
Station B: Rb 10.0 Mbps
Station C: Rb 5.0 Mbps
Calculate the burst duration and symbol rate for each earth station, and the earth station trans-
mitter output power required if the transponder output backoff is set at 1.0 dB and the gain of the
transponder with this output backoff is 104 dB. Compare the uplink (CN) ratios in the transponder
for FDMA and TDMA operation given that station A’s transmission has a (CN)up ratio of 34 dB
when the transponder is operated in FDMA.
The transponder must carry a total bit rate of 15 10 5 30 Mbps within the 1.0-ms
frames. Thus each frame carries 30 Mbps 0.001 s 30 kb. The three preamble and guard times
take up 3 (10 2) 36 s in each frame, leaving 1000 36 964 s for transmission of
data. Hence the burst bit rate is
Rb burst 30 kb964 ms 31.12 Mbps.
Since we are using QPSK for the transmissions, the burst symbol rate on the link is
Rs burst 31.12 Mbps2 15.56 Msps
Each of the stations must transmit at the same burst rate of 15.56 Msps. The burst lengths
can be calculated from the time available in each frame for data transmission and the number of
bits each station must send in a 1 ms TDMA frame. The time available for data transmission is 964
s, which must be shared in proportion to the number of bits each station sends in a frame. The
number of bits in a frame is 30,000, so the sharing of bits and times within a frame is given below
Station A: Nb 15,000 bits TA 482.0 ms
Station B: Rb 10,000 bits TB 321.3 ms
Station C: Rb 5,000 bits TC 160.7 ms
We can easily check to see if these results are correct. Each earth station must have the stated
average bit rate, so if we multiply the burst duration for each earth station by the burst bit rate for
the transponder, 31.12 Mbps, we must have the correct number of bits/frame for each station.
Station A: TA 482.0 ms Nb 482.0 ms 31.12 Mbps 15,000
Station B: TB 321.3 ms Nb 321.3 ms 31.12 Mbps 10,000
Station C: TC 160.7 ms Nb 160.7 ms 31.12 Mbps 5,000
Each station must transmit at the same symbol rate of 15.56 Msps, regardless of the number of
bits sent per frame. In the previous FDMA example, a transponder output power of 20 W 13 dBW
was achieved with a total earth station power of 250 W 24 dBW and a transponder gain of 105 dB.
With TDMA, we are using a 1 dB transponder output backoff, and a transponder gain of 104 dB, so
the transponder output power is now 16 1 15 dBW, an increase of 2 dB, and we have lost 1 dB
of gain in the transponder. This requires an earth station output power, from each earth station, of
Pt es 24 2 1 27 W 500 W
c06.qxd 09/08/02 14:18 Page 246
TDMA requires a substantial increase in earth station transmitter power, relative to FDMA,
for a low capacity earth station which joins high-capacity stations in a TDMA system. In this ex-
ample, a station that was transmitting 42 W when the transponder was operated in FDMA must now
transmit 500 W when the transponder is operated in FDMA.
The uplink (CN)up for station A’s 15 MHz signal was 34.0 dB when the transponder was
operated in FDMA. With QPSK and a burst rate of 15.56 Msps, the noise bandwidth of the earth
station receiver, assuming ideal RRC filters, will be 15.56 MHz. The output of station A has been
increased from 21 to 24 dBW, so the input power level at the transponder will also have increased
by 3 dB. Hence the uplink (CN)up ratio in the transponder for station A signals is
1CN2 up 34.0 10 log10 115.0 15.562 3 36.8 dB
Since all earth stations transmit at the same power level and with the same burst rate, and all the sig-
nals have the same noise bandwidth, the (CN)up ratio for each of the signals in the transponder is
identical, at 36.8 dB. This is 2.8 dB higher than for the FDMA operation, but at the expense of a
large increase in the total uplink power from the three earth station transmitters.
Satellite Switched TDMA

One advantage that TDMA has when used with a baseband processing transponder is satel-
lite switched TDMA. Instead of using a single antenna beam to maintain continuous com-
munication with its entire coverage zone, the satellite has a number of narrow antenna
beams that can be used sequentially to cover the zone. A narrow antenna beam has a higher
gain than a broad beam, which increases the satellite EIRP and therefore increases the ca-
pacity of the downlink. Uplink signals received by the satellite are demodulated to recover
the bit streams, which are structured as a sequence of packets addressed to different
receiving earth stations. The satellite creates TDMA frames of data that contain packets
addressed to specific earth stations, and switches its transmit beam to the direction of the
receiving earth station as the packets are transmitted. Note that control of the TDMA net-
work timing could now be on board the satellite, rather than at a master earth station.
In the above example, the VSAT earth station could transmit data to a baseband pro-
cessing satellite using SCPC-FDMA to permit the use of a small antenna and low power
transmitter. The satellite could then use satellite switched TDMA to send that data to mul-
tiple earth stations, creating a mesh VSAT network. It is difficult to create a mesh VSAT
network using SCPC-FDMA. Baseband processors are considered in more detail in the
next section.
6.4 ONBOARD PROCESSING
The discussion of multiple access so far has assumed the use of a bent pipe transponder,
which simply amplifies a signal received from earth and retransmits it back to earth at a
different frequency. The advantage of a bent pipe transponder is flexibility. The transpon-
der can be used for any combination of signals that will fit within its bandwidth. The
disadvantage of the bent pipe transponder is that it is not well suited to uplinks from small
earth stations, especially uplinks operating in Ka band. Consider a link between a small
transmitting earth station and a large hub station via a bent pipe GEO satellite transpon-
der. There will usually be a small rain fade margin on the uplink from the transmitting
station because of its low EIRP. When rain affects the uplink, the CN ratio in the transpon-
der will fall. The overall CN ratio in the hub station receiver cannot be greater than the
CN ratio in the transponder, so the bit error rate at the hub station will increase quickly
as rain affects the uplink. The only available solution is to use forward error correction
c06.qxd 09/08/02 14:18 Page 247
6.4 ONBOARD PROCESSING 247
coding on the link, which lowers the data throughput but is actually needed for less than
5% of the time.
The problem of uplink attenuation in rain is most severe for 3020 GHz uplinks
with small margins. Outages are likely to be frequent unless a large rain fade margin is
included in the uplink power budget. Onboard processing or a baseband processing
transponder can overcome this problem by separating the uplink and downlink signals
and their CN ratios. The baseband processing transponder can also have different mod-
ulation schemes on the uplink and downlink to improve spectral efficiency, and can
dynamically apply forward error control to only those links affected by rain attenuation.
All LEO satellites providing mobile telephone service use onboard processing, and Ka-
band satellites providing Internet access to individual users also use onboard processing.
Baseband Processing Transponders

A baseband processing transponder has a receiver and transmitter similar to those found
in an earth station. The received signal from the uplink is converted to an intermediate
frequency and demodulated to recover the baseband signal, which is then processed and
reassembled. The baseband signal is modulated onto a carrier at a downlink frequency
and transmitted back to earth. The signals are invariably digital, although that is not a re-
quirement of a baseband processing transponder. The advantage of this process is that up-
link and downlink signal formats need not be the same, and that different forms of error
correction can be applied to the uplink and downlink. The CN ratios of the uplink and
downlink are not tied together through the reciprocal formula (Eq. 6.10). If the CN on
the uplink is low, because of an uplink rain fade, for example, bit errors will be present
in the recovered data in the transponder. The BER will depend only on the uplink CN
ratio. If the CN on the downlink is high, as is usually the case for all star networks work-
ing with a large hub earth station, no additional bit errors will occur on the downlink.
Separation of the uplink and downlink signals allows different modulation methods
to be used, as well as flexible error correction codes. In star networks, the CN ratio of
the uplink and downlink between the satellite and the hub station is usually high because
of the large antenna gain and high transmit power of the hub station. The high CN ratio
can be traded for a high level modulation such as 16-QAM, which reduces the bandwidth
required for the uplink and increases the spectral efficiency of the communication system.
16-QAM sends four bits per symbol, and requires only half the bandwidth of QPSK.
As an example, consider a system in which half rate forward error correction and
QPSK is used on the uplink from a small earth terminal. For a message data rate of
Rd bits per second, the transmitted bit rate will be Rb 2 Rd bps. An RF bandwidth of
Rb2 (1 up) Rd (1 up) Hz will be required on the uplink, where up is the roll-
off parameter of the (assumed) RRC filter in the transmitter. On the downlink to the hub
station, where the CN ratio is high, 16-QAM can be used without forward error correc-
tion. The RF bandwidth required for the downlink will be Rd4 (1 dn), where dn
is the roll-off parameter of the RRC filter in the transponder transmitter section. If we
assume the same roll-off parameter, , in both the uplink and the downlink transmitters
and receivers, the downlink requires only one-fourth of the uplink bandwidth to send the
same number of bits. The fourfold reduction in downlink bandwidth represents a consid-
erable improvement in the spectral efficiency of the satellite system.
Star networks with small uplink earth stations are widely used for VSAT systems
and satellite links that carry Internet traffic. One hub station linked to an Internet Service
Provider (ISP) can connect to hundreds of user earth stations located at the users’ homes.
This is a major growth area for Ka-band GEO satellites, a number of which are planned
c06.qxd 09/08/02 14:18 Page 248
for launch in the 2002–2005 time frame. The user terminal is typically a 0.5- to 0.8-m
antenna with a low power transmitter, 0.5–1 W output, that can send data to the satellite
at rates up to 400 kbps. The uplink operates in the 30 GHz band with a small rain fade
margin, resulting in low CN at the transponder input and requiring the use of error cor-
rection on the uplink during rain fades. It may be possible to remove all or most of the
error correction when the uplink is operating in clear air, allowing higher data speeds ex-
cept during rain fades. Rain fading occurs for less than 5% of the time, and affects only
a small number of users at any instant. Under most conditions, only a small number of
users will suffer severe rain fading at the same time, so most of the links to the satellite
will not need heavy error correction. This flexibility can increase the capacity of the satel-
lite links by a factor of two relative to a bent pipe transponder, at the cost of considerably
increased complexity in both the satellite and the user terminal.
Satellite Switched TDMA with

Onboard Processing
Baseband processing is essential in satellites using satellite switched TDMA, because data
packets must be routed to different antenna beams based on the address of the destination
earth station. The data in such systems is always sent in packets which contain a header
and a traffic section. The header contains the address of the originating station and the
address of the destination earth station. When satellite switched TDMA is used, the
transponder must extract the destination information and use it to select the correct down-
link beam for that packet. The satellite is operating much like a router in a terrestrial data
transmission system. Switched beam operation of an uplink from a small earth station is
more difficult to achieve because it requires synchronization of the earth station transmit
time with the satellite beam pointing sequence, in much the same way that a TDMA up-
link operates. However, the uplink can operate in a small bandwidth which overcomes the
chief disadvantage of classic TDMA—the requirement for high burst rate transmissions
and high transmit power.
Satellite switched TDMA can greatly increase the throughput of a transponder. Con-
sider, for example, a satellite providing Internet access to individual users in the United
States. The uplink and downlink beams at the satellite must provide coverage over an area
approximately 6° by 3°, as seen from the satellite. Antenna gain and beamwidth are re-
lated by the approximate relationship G 33,000(product of beamwidths in degrees).
This limits the maximum achievable satellite antenna gain to approximately 32.5 dB.
A satellite with switched beam capability can have much narrower beams with
higher gain than a satellite with a single fixed beam. The limitation on gain is the diameter
of the antenna, which must fit inside the launch vehicle shroud. For launchers available
in 2000, this limit is about 3.5 m. At 20 GHz, the uplink frequency for Ka band, an
antenna with a circular aperture of diameter D 3.5 m and aperture efficiency of A
65% has a gain G A (D)2 55.4 dB, and its beamwidth is approximately 75
D degrees 0.32°. The corresponding downlink antenna for 30 GHz that has a
beamwidth of 0.32° and a gain of 55.4 dB has a diameter of 2.33 m. The switched beam
satellite has an antenna gain almost 23 dB higher than the single beam satellite, which
can be traded directly for reduction in uplink or downlink transmit power, and uplink or
downlink data rate. However, the satellite must generate at least 170 beams to cover all
of the United States with 0.32° beams, with a consequent increase in satellite antenna
complexity.
Satellite switched TDMA and multiple beam antennas are a feature of most of the
proposed Ka-band Internet access satellites13,14. The Astrolink satellites, for example, have
c06.qxd 09/08/02 14:18 Page 249
6.5 DEMAND ACCESS MULTIPLE ACCESS (DAMA) 249
105 spot beams for links to small user terminals. The satellite uplink (30 GHz) antenna
has a diameter of 2.5 m and the downlink antenna has a diameter of 3.25 m. There are
five spot beams for links to hub stations; the large antennas used by the hub stations al-
low a lower gain antenna with a broader beam to be used on the satellite.
Coverage of the United States with multiple beams is not always provided uniformly.
Differences in population densities and the frequency of heavy rainfall make it advanta-
geous to provide more system capacity to metropolitan areas, and also to provide higher
link margins to areas with more frequent heavy rainfall, such as Florida and the south-
eastern states. In the most sophisticated of large GEO satellites, a steerable phased array
antenna can be used, with control of beam pointing from the ground via the satellite’s
telemetry and command link. The antenna beams can then be moved to provide coverage
of areas with highest demand for traffic. The growth of the terrestrial optical fiber net-
work will eventually fulfill the need for high-speed access to the Internet. Where direct
access to an ISP is available via optical fiber, the transmission rate is likely to be higher
and the cost to the user is likely to be lower. As the fiber network spreads through met-
ropolitan areas, an Internet access satellite can concentrate its service on less well popu-
lated and rural areas. A steerable beam antenna allows the geographical capacity of the
satellite to be reconfigured throughout its lifetime.
6.5 DEMAND ACCESS MULTIPLE

ACCESS (DAMA)
Demand access can be used in any satellite communication link where traffic from an
earth station is intermittent. An example is an LEO satellite system providing links to mo-
bile telephones. Telephone voice users communicate at random times, for periods rang-
ing from less than a minute to several minutes. As a percentage of total time, the use of
an individual telephone may be as little as 1%. If each user were allocated a fixed chan-
nel, the utilization of the entire system might be as low as 1%, especially at night when
demand for telephone channels is small. Demand access allows a satellite channel to be
allocated to a user on demand, rather than continuously, which greatly increases the num-
ber of simultaneous users who can be served by the system. The two-way telephone chan-
nel may be a pair of frequency slots in a DA-SCPC system, a pair of time slots in a TDM
or TDMA system, or any combination or FDMA, TDM, and TDMA. Most SCPC-FDMA
systems use demand access to ensure that the available bandwidth in a transponder is used
as fully as possible.
In the early days of satellite communication, the equipment required to allocate
channels on demand, either in frequency or time, was large and expensive. The growth of
cellular telephone systems has led to the development of low cost, highly integrated con-
trollers and frequency synthesizers that make demand access feasible. Cellular telephone
systems use demand access and techniques similar to those used by satellite systems in
the allocation of channels to users. The major difference between a cellular system and a
satellite system is that in a cellular system the controller is at the base station to which
the user is connected by a single hop radio link. In a satellite communication system, there
is always a two hop link via the satellite to a controller at the hub earth station. Controllers
are not placed on the satellites largely because of the difficulties in determining which
links are in use, and who will be charged for the connection. As a result, all connections
pass through a controlling earth station that can determine whether to permit the requested
connection to be made, and who should be charged. In international satellite communi-
cation systems issues such as landing rights require the owner of the system to ensure
c06.qxd 09/08/02 14:18 Page 250
that communication can take place only between users in preauthorized countries and
zones. The presence of the signals from all destinations at a central earth station also al-
lows security agencies the option of monitoring any traffic deemed to be contrary to the
national interest.
Demand access systems require two different types of channel: a common signal-
ing channel (CSC) and a communication channel. A user wishing to enter the communi-
cation network first calls the controlling earth station using the CSC, and the controller
then allocates a pair of channels to that user. The CSC is usually operated in random
access mode (see Section 6.6) because the demand for use of the CSC is relatively low,
messages are short, and the CSC is therefore lightly loaded, a requirement for any DA
link. Packet transmission techniques are widely used in demand access systems because
of the need for addresses to determine the source and destination of signals. Section 6.7
discusses the design of packets for use in satellite communication systems.
Bent pipe transponders are often used in demand access mode, allowing any con-
figuration of FDMA channels to be adopted. There seem to be few standards for demand
access systems in the satellite communication industry, with each network using a differ-
ent proprietary configuration. Figure 6.11 shows a typical 54 MHz bandwidth Ku band
transponder frequency plan for the inbound channels of a VSAT network using frequency
division multiple access with single channel per carrier and demand access (FDMA-SCPC-
DA) on the inbound link. The individual outbound RF channels are 45 kHz wide, to ac-
commodate the occupied bandwidth of 64-kbps bit streams transmitted using QPSK and
RRC filters with 0.4. A guard band of 15 kHz is allowed between each RF channel,
so one RF channel requires a total bandwidth of 60 kHz. A 54 MHz bandwidth transpon-
der can accommodate 900 of these 60 kHz channels, but it is unlikely that all are used at
the same time. Many VSAT systems are power limited, preventing the full use of the
transponder bandwidth, and the statistics of demand access systems ensure that the
likelihood of all the channels being used at one time is small. Considerable backoff is
required in a bent pipe transponder with large numbers of FDMA channels, as discussed
earlier in this chapter.
The outbound link of this particular VSAT network is a continuous TDM bit stream
transmitted through a separate transponder. A second transponder is used to allow for the
differences in transponder gain needed for the inbound and outbound channels of the VSAT
system. In VSAT systems, the inbound and outbound channels are usually symmetric,
offering the same data rate in opposite directions. Internet access systems are often
64 kbps QPSK channels
CSC 2 3 4 5 6 898 899 CSC
60 kHz channel spacing 15 kHz guard bands

between channels
FIGURE 6.11 Frequency plan for a 54-MHz transponder carrying 900 demand access
channels. Each channel has an occupied RF bandwidth of 45 kHz and carries one 64-kbps
signal. Channel 1 and channel 900 are common signaling channels (CSC) used by the
demand assignment system to set up access to the other 898 channels.
c06.qxd 09/08/02 14:18 Page 251
asymmetric, because requests for information can be short but the resulting replies may
be lengthy. The packet length of the TDM signal in the outbound direction may be fixed,
which suits a symmetrical network, or variable, which better suits an Internet channel
capable of downloading large files or video from the Internet.
The common signaling channels (CSC) shown in the inbound transponder in Figure
6.11 are located at the ends of the transponder bandwidth. When a VSAT earth station wants
to access the satellite, it transmits a control packet to the satellite on the CSC frequency and
waits for a reply. The control packet is received by the hub earth station and decoded. The
control packet contains the address of a terrestrial or satellite destination for the call, DA,
the address of the station requesting the connection, RA, any other relevant data (such as a
character, CP, to indicate that this is a control packet with no traffic data), and a cyclic
redundancy check (CRC) that is used in the receiver to check for errors in the packet. The
control station records both origination and destination station addresses and measures the
duration of the connection in order to generate billing data. In a true demand access sys-
tem, the control station allocates the VSAT an uplink frequency and a time slot of specified
duration in the outbound TDM frame. If the hub station has a large volume of data to send
to a particular VSAT station, it can allocate a longer time slot in the TDM frame to that
station. This is important in Internet access systems where a large file of video or other
multimedia data may have to be sent. The timeslots usually come in multiples of a fixed
minimum duration so that clock rates and buffer sizes are compatible. If the system becomes
busy and many stations are requesting large files, throughput to any one station will slow
down toward the standard minimum rate, exactly as in a terrestrial Internet server.
The outbound link transmits a continuous bit stream so that receivers can maintain
carrier phase and bit clock synchronization. The data is organized into a sequence of pack-
ets, addressed to the receiving stations, and organized into a frame. One frame contains
one packet for each earth station, as illustrated in Figure 6.12. In many TDM systems
there is always a time slot with addresses and other information for delivery of data to
each earth station in the network, but there may be no data to send. In this case, the packet
is assigned a special character to indicate no traffic.
Once an inbound frequency and an outbound time slot are allocated to the VSAT
station, the connection can be completed and data transfer or voice communication can
begin. The hub earth station designed to receive FDMA-SCPC signals has multiple re-
ceivers operating at different center frequencies and allocated to the many transponders
on the satellite.
The block diagram of a receiver for one transponder with a bandwidth of 54 MHz
is illustrated in Figure 6.13. The receiver amplifies and down-converts the received sig-
nal to an intermediate frequency of 700 MHz and then to a second IF at 70 MHz. Indi-
vidual FDMA-SCPC channels within the band 43 to 97 MHz are down-converted to a
standard IF frequency of 2 MHz in this example using local oscillators with frequencies
41 MHz through 95 MHz in steps of 60 kHz. There are a total of 900 such 2-MHz IF
receivers to cover all the frequency slots in the transponder bandwidth.
End of TDM frame Next frame
C E
A Data A Data A Data A Data R O A Data
896 897 898 899 C F 2
FIGURE 6.12 End of a TDM frame outbound to VSAT stations. A, station

address; CRC, cyclic redundancy check; EOF, end of frame marker.
c06.qxd 09/08/02 14:18 Page 252
Downlink at f c GHz
LNA Image rejection 700 MHz 700 MHz 70 MHz

BPF Mixer BPF IF amplifier Mixer BPF
First LO Second LO
f c − 700 MHz 630 MHz
70 MHz D B
IF amplifier Ch 1 data
D B
Ch 2 data
D B
Ch N data
Channel selection
filters 2 MHz 2 MHz PSK Bit recovery
Third LOs Mixers BPFs IF amplifiers demodulators circuits
spaced 60 kHz
FIGURE 6.13 Receiver for FDMA-SCPC hub station. Channel selection is by bandpass
filters in the 70 MHz IF section of the receiver and the third local oscillator. Each channel
selection filter and third local oscillator is at a different frequency, spaced in 60 kHz incre-
ments across the IF band. D, demodulator; B, bit recovery circuit.
The hub earth station transmits a continuous TDM signal to all VSAT stations in
the network. The symbol rate and the bandwidth of the TDM signal depends on the max-
imum bandwidth that the VSAT receiver can use, based on noise considerations. If the
network is symmetrical and uses all 900 possible FDMA-SCPC channels at a bit rate of
64 kbps per channel, the TDM signal must have a bit rate of 900 64 kbps 57.6 Mbps,
ignoring packet overhead. This bit rate is likely to be much too high for a VSAT station,
resulting in low CN ratio in the receiver. The transponder can be partitioned to carry
multiple groups of TDM signals with lower bit rates better suited to VSAT receivers.
EXAMPLE 6.5.1 FDMA-SCPC-DA

A VSAT network consists of 250 Ku-band VSAT earth stations sharing one inbound and one outbound
transponder on a GEO satellite. The transponder bandwidth is 54 MHz. The transmit data bit rate for
the VSAT stations is 64 kbps. Statistics for the VSAT network show that each VSAT generates an av-
erage data bit rate of 5 kbps with random time of arrival of data. The average outbound data rate per
VSAT station is 20 kbps. The inbound data link operates in FDMA-SCPC using demand access, QPSK
modulation with 0.5 RRC filters, and half rate forward error correction. The outbound data link
uses a single continuous TDM stream, QPSK modulation, and a 16-bit CRC word in each packet. De-
termine the bit rate and bandwidths of the VSAT and hub station receivers, suggest a frame and packet
size for the TDM link, an SCPC-FDMA frequency plan, and a demand access method.
Inbound Link: VSAT to Hub
The data rate at the VSAT station averages 5 kbps, but arrives in the form of variable length mes-
sages with random arrival times. The data are transmitted from the VSAT station on a half rate FEC
c06.qxd 09/08/02 14:18 Page 253
bit stream at 128 kbps carrying 64 kbps of data, as a QPSK symbol stream at 64 kbaud. The link
has 0.5 RRC filters, so the occupied bandwidth of the QPSK signal is Rs (1 ) 96 kHz.
The IF receivers in the hub earth station will have filters with a noise bandwidth of 64 kHz (equal
to the symbol rate of the signal) and total bandwidth of 96 kHz. Allowing a 20% guard band between
RF channels requires a carrier-to-carrier spacing of 115 kHz.
The maximum number of channels that can be carried by the inbound transponder is
54,000115 469. Several of these channels would have to be designated as common signaling
channels, so the maximum usable communication channels would be about 460. If the transponder
is power limited, fewer channels can be carried.
Each VSAT channel transmits data at 64 kbps, but data arrives at an average rate of 5 kbps.
Assuming the data does not require transmission in real time, a buffer can store data until a total
of 120 kb has been collected. The data are then packetized (see Section 6.7) and overhead bits are
added for addresses, CRC, etc. The VSAT station can then send a request for a channel to the hub
station and the data can be downloaded in about 2 s. The average time to collect 120 kb at the VSAT
is 24 s. This scenario favors a SCPC-FDMA multiple access scheme using a CSC to obtain access
to the satellite once every 24 s. If we assume that it takes 2 s to establish the connection and 2 s to
transmit the packets, a total of 4 s of satellite transmission is used every 24 s, giving a VSAT sta-
tion loading of 16.6%. This means that (in theory) six VSAT stations can share the same RF chan-
nel. In practice, it is impossible to load the channels to 100% of their capacity because data arrives
at random time intervals causing temporary overload when a large volume of data arrives at the
same time.
If we assume a 66% load factor for the inbound link, we can share one inbound channel
between four VSAT stations, which requires a total bandwidth of 544 13.5 MHz in the transpon-
der. Demand access has clearly achieved considerable savings in bandwidth and power in this case.
Outbound Link: Hub to VSAT
The hub station transmits a continuous TDM frame consisting of many sequential packets addressed
to each VSAT station. There are 250 VSAT stations with an average outbound data rate of 20 kbps
each. If we apply half rate FEC to the outbound data stream and use QPSK modulation, we will
transmit at 20 kbaud per station. With 250 stations, the average outbound symbol rate is 5 Mbaud,
and the outbound data rate is 5 Mbps because we are using QPSK modulation with half rate FEC.
The occupied bandwidth of the signal with 0.5 RRC filters is 7.5 MHz, and the VSAT receiver
noise bandwidth will be 5.0 MHz.
Let us assume a frame length of 5 s. On average each VSAT will receive 100 kbps during a
5-s period, but the statistics of the traffic suggest that there will be megabits of data for some sta-
tions and none for others in any given time period. Within the 5-s frame, if the frame time is di-
vided equally, each earth station could receive data for a period of 20 ms, and would receive 10,000
data bits at 5 Mbps. If we assume a 5% overhead allowance for the packet, there will actually be
9500 data bits delivered during the 20-ms period, plus 500 overhead bits. If there are no data bits
to be delivered to a given station, only the overhead portion of the packet, 500 bits, needs to be
transmitted. That allows other stations to use the spare time in the frame to send additional data at
a much higher than average rate.
Demand access is most valuable when the traffic mix changes a great deal. The multiple ac-
cess system described here was designed to meet the needs of the average data rates transmitted on
the inbound and outbound links. If many of the stations are inactive, the other stations can have in-
creased data rates. For example, suppose only 50 of the VSAT stations are active. Each VSAT sta-
tion can transmit at its maximum data rate of 64 kbps, and will deliver data as fast as the terminal
can supply it. To increase the inbound data rate above 64 kbps requires a wider channel bandwidth,
and a receiver in the hub station with a wider IF filter. Alternatively, the VSAT station could trans-
mit two carriers. The limitation on inbound data rate is likely to be VSAT EIRP and the resulting
uplink CN ratio in the transponder. SCPC-FDMA does not offer as much flexibility to change data
rates as TDM.
On the outbound link with only 50 VSAT stations active, the packet length for the active sta-
tions can be increased by a factor of five. Short packets must still be sent to all stations to main-
tain synchronization of the VSAT receiver, but need consist only of a control packet, which is 500
c06.qxd 09/08/02 14:18 Page 254
bits in length. The outbound link has a bit rate of 5 Mbps. Each VSAT station receives a control
packet of 500 bits every 5 s if there is no data to be sent to that station. Hence the 200 inactive sta-
tions receive an average bit rate of 100 bps, and the 50 active stations can share the remaining bits.
The bit rate for the 50 active stations is 5 Mbps 200 100 bps 4.98 Mbps, an average rate
of 99,600 bps for each active station. The normal packet length is 20 ms and delivers 10,000 bits.
The packet can now be extended to almost 100 ms for each of the active stations. The actual packet
length is 100 ms 80 s 99.92 ms to allow for the time required in each frame to send a sin-
gle control packet to each inactive VSAT. Thus the active stations receive 499,600 bits in each packet,
of which 500 are control bits, giving a data rate of 99.82 kbps. TDM frames which offer variable
packet length can easily accommodate a widely changing mixture of data rates delivered to each
VSAT station. A field within the packet header can tell the VSAT station how many data bits are
in the packet, allowing great variability.
6.6 RANDOM ACCESS (RA)
Random Access is a widely used satellite multiple access technique where the traffic
density from individual users is low. For example, VSAT terminals and satellite mobile
telephones often require communication capacity infrequently. These users can share
transponder space without any central control or allocation of time or frequency, provided
the average activity level is sufficiently low. In a true random access network, a user trans-
mits packets whenever they are available. The packet has a destination address, and a
source address. All stations receive the packet and the station with the correct address
stores the data contained in the packet. All other stations ignore the packet, unless it is
designated as a broadcast packet with information for all stations. In satellite communi-
cation systems, the network is more usually a star configuration, with a single hub and
many small earth stations or portable terminals. Inbound packets are received by the hub
earth station and forwarded to their destinations. Early work on random access techniques
for radio channels was done at the University of Hawaii, where the system was called
Aloha and was known by the generic term packet radio.
Random access cannot be used when traffic density exceeds 18%, and therefore
makes inefficient use of the bandwidth available in the transponder. Although there is a
saving in transmission time because no call set up is required, the low throughput and
poor spectral efficiency has restricted random access use in satellite systems to cases where
traffic bursts are short and highly intermittent. In general, it is used on single SCPC-FDMA
channels, rather than on whole transponders. The common signaling channel, described
in the previous section, is an example of an SCPC-FDMA random access channel within
a transponder that can successfully use random access because it is lightly loaded.
6.7 PACKET RADIO SYSTEMS

AND PROTOCOLS
Data transmission between computers or terminals requires agreed methods by which con-
nections are established and data is transferred. When we make a telephone call, there are
conventions and etiquette which define how a telephone connection is established and
when each person speaks. For example, you decide to call your friend John Doe. You lift
the telephone handset and hear dial tone. The telephone system is telling you that it is
ready for you to dial a number. You dial the telephone number of your friend and wait to
hear a ringing tone. The telephone system is telling you that it is trying to attract the
c06.qxd 09/08/02 14:18 Page 255
6.7 PACKET RADIO SYSTEMS AND PROTOCOLS 255
attention of your friend. If your friend answers, you expect to hear “Hello, this is John
Doe.” You might reply, “Hello, this is Bill Smith. How are you?” If the person who an-
swers just says “Hello,” you cannot be sure that you have reached John Doe, and you
might have to say “Hello, this is Bill Smith, is that John Doe?” If you dialed a wrong
number and have connected to the wrong person, you might say “Sorry, I must have di-
aled a wrong number.” Then you would put the phone down and try again. If you weren’t
sure whether the answerer said “John Doe” or “John Roe” because of noise on the line,
you would ask for a repeat transmission.
The process of creating a telephone connection to a friend makes use of signals pro-
vided by the telephone system and a set of procedures based in human etiquette and in-
telligence. Humans can readily determine whether they have reached the correct person,
and how to proceed if the call does not go through correctly. A data transmission system
lacks the intelligence of a human, and cannot readily adapt to changing responses in the
way that humans can. Data transmissions make use of packets and protocols to ensure
that automatic connection and transfer of data can be achieved reliably without human in-
tervention. One dictionary definition of protocol is, “The code of ceremonial forms and
courtesies of precedence, etc., accepted as proper and correct in official dealings, as be-
tween heads of state or diplomatic officials”15. It is this sense of the word protocol that is
used to describe the rules by which two data terminals can connect to each other through
a communication system and then transfer data.
The creation of protocols for data transmission is a very large subject, with many
books and papers devoted to the design and performance of different schemes. In this text,
only the briefest summary of the subject is included. The widespread ownership of per-
sonal computers and growth of the Internet helped spur development of efficient and pow-
erful protocols like TCPIP. The International Standards Organization (ISO) has created
a seven layer model for machine to machine communication known as the open systems
interconnection (OSI) which separates the functions of different parts of the system. The
ISO-OSI model is shown in Figure 6.14. Although the model is widely quoted as de-
scribing the structure of data communication systems, it rarely seems possible to identify
seven separate layers within any given system. The lowest layer, the physical layer, is the
one with which this text is concerned—the transport of bits from one point to another.
Regardless of the method of transportation, the ISO model assumes that bits are carried
across the physical layer, in both directions, possibly arriving with some errors. The
Application
Presentation
Session
Transport
Network
Data link control
Physical FIGURE 6.14 ISO-OSI seven layer model

for digital communications links.
c06.qxd 09/08/02 14:18 Page 256
remaining six layers of the model are embedded in the hardware and software of the ter-
minals at each ends of the link. The second layer of the model provides error detection
and correction, either in hardware or software, and the remaining layers are responsible
for organizing the data transfer, from making connections to billing for the service.
Terrestrial data communication has evolved through a series of protocols, beginning
with a standard by IBM, known as HDL, through X.25 to ATM (asynchronous transfer mode).
ATM uses a 53-byte packet and was designed to be transmitted over fiber-optic line networks
using the DS-3 44.736 Mbps transmission rate according to an IEE standard (P802.6). Digital
cellular radio systems, although using three different standards, are compatible, and any
cellular telephone designed for TDMA can work with any provider’s TDMA system. Satel-
lite communication systems have not evolved a set of standards, and many systems use
proprietary protocols. For example, the Iridium, Globalstar, and ICO Global LEO/MEO satel-
lite communication systems all use digital transmission with different protocols. Handsets
designed for use with the Iridium system cannot communicate with Globalstar or ICO Global
satellites. There is now intense interest in designing GEO satellite systems that are compat-
ible with the ATM protocol so that wide band satellite links can be connected directly to ter-
restrial data networks16, 20 –23. Generically, this is known as wideband by satellite. Typically,
the long delay time inherent in transmission via a GEO satellite creates problems when in-
terfacing to a terrestrial protocol designed for much shorter delays. Special interfaces are
needed at the earth stations that allow the protocol to be adapted for satellite use.
Data cannot be transmitted as a continuous bit stream in most cases, because ad-
ditional bits for addresses, error control, and other additional information that is not part
of the message data must be inserted into the bit stream. Data are sent in packets, usu-
ally with an agreed length and content, using a structure that is very similar to the TDMA
frame described in Section 6.3. Each packet typically consists of a header, which con-
tains address and control information, a block of message data, and a closing section
with error control bits and an end of packet flag. One protocol and packet design that
has been widely accepted for use in amateur satellite systems is AX.25. The AX.25 pro-
tocol is based on X.25, a protocol developed for terrestrial data communications, and is
used by amateur radio operators in a terrestrial data communication network. The pro-
tocol was adapted for use in amateur radio LEO satellites with VHF and UHF transpon-
ders operating in a store and forward mode. Several of these satellites were built and
orbited, providing a method for amateur radio operators to send messages by satellite17.
Figure 6.15 shows the structure of the AX.25 packet. All packets begin and end with
a unique word, called a flag, 01111110, which is not allowed to appear in any other part
of the packet. The flag marks the end of one packet and the start of the next packet, so that
the receiving data terminal can extract the packet contents correctly. The general format of
the packet contents is a header, followed by message bits, followed by a cyclic redundancy
check (CRC). The header contains addresses, in the form of amateur radio call signs, for
the sender and intended recipient, and control information that helps the receiving station
identify the contents of the packets. The control bits, for example, tell the receiving ter-
minal how long the packet is, and define whether this is a broadcast packet, intended to be
viewed by all receiving stations, or a packet for a specific recipient. Control bits also spec-
ify the type of packet—some packets contain no message bits and are sent to convey sys-
tem information. The CRC allows the receiver to check whether the packet was received
correctly, and to call for a retransmission if an error is detected. The interested reader should
refer to reference 17 for further details of the amateur radio satellite system.
All data transmission system must have some form of protocol, and data is almost al-
ways sent in packet form. Thus whenever multiple access techniques are discussed and dig-
ital data are transmitted, it can be assumed that some form of packet transmission is used.
c06.qxd 09/08/02 14:19 Page 257
6.8 CODE DIVISION MULTIPLE ACCESS (CDMA) 257
Flag Address Control FCS Flag
8 bits 112–560 bits 8 bits 16 bits 8 bits
AX.25 Un-numbered frame
Flag Address Control PID Message FCS Flag
8 bits 112–560 bits 8 bits 8 bits N × 8 bits 16 bits 8 bits
AX.25 Information frame
Flag = 01111110 FCS = Frame check sequence

FIGURE 6.15 AX.25 packets. The unnumbered frame (packet) is used only for control,
and contains no message data. All packets (frames) start and end with Flag, a unique word
with six ones. Data messages are processed to ensure that six ones do not appear as a
string anywhere except Flag. The value of N can be anywhere from 1 to 256. FCS; Frame
control sequence.
6.8 CODE DIVISION MULTIPLE

ACCESS (CDMA)
Code division multiple access is a scheme in which a number of users can occupy all
of the transponder bandwidth all of the time. CDMA signals are encoded such that in-
formation from an individual transmitter can be recovered by a receiving station that
knows the code being used, in the presence of all the other CDMA signals in the same
bandwidth. This provides a decentralized satellite network, as only the pairs of earth
stations that are communicating need to coordinate their transmissions. Subject to
transponder power limitations and the practical constraints of the codes in use, stations
with traffic can access a transponder on demand without coordinating their frequency
(as in FDMA) or their time of transmission (as in TDMA) with any central authority.
Each receiving station is allocated a CDMA code; any transmitting station that wants
to send data to that earth station must use the correct code. CDMA codes are typically
16 bits to many thousands of bits in length, and the bits of a CDMA code are called
chips to distinguish them from the message bits of a data transmission. The CDMA chip
sequence modulates the data bits of the original message, and the chip rate is always
much greater than the data rate. This greatly increases the speed of the digital trans-
mission, widening its spectrum in proportion to the length of the chip sequence. As a
result, CDMA is also known as spread spectrum. Direct sequence spread spectrum (DS-
SS) is the only type currently used in satellite communication; frequency hopping spread
spectrum (FH-SS) is used in the Bluetooth system for multiple access in short range lo-
cal area wireless networks18.
CDMA was originally developed for military communication systems, where its
purpose was to spread the energy of a data transmission across a wide bandwidth to make
detection of the signal more difficult (called low probability of intercept). Spreading the
energy in a signal across a wide bandwidth can make the noise power spectral density
(NPSD) in the receiver larger than the power spectral density (PSD) of the received sig-
nal. The signal is then said to be buried in the noise, a common feature of DS-SS signals,
c06.qxd 09/08/02 14:19 Page 258
and the signal is much harder to detect than a signal with a PSD greater than the receiver’s
NPSD. The correlation process that recovers the original data bits from a DS-SS spread
spectrum signal is also resistant to jamming, the deliberate transmission of a radio signal
at the same frequency to blot out someone else’s transmission. Both of these attributes
are valuable in tactical military communication systems.
CDMA has become popular in cellular telephone systems where it is used to en-
hance cell capacity. However, it has not been widely adopted by satellite communication
systems because it usually proves to be less efficient, in terms of capacity, than FDMA
and TDMA. The Globalstar LEO satellite system was designed to use CDMA for multi-
ple access by satellite telephones; the advantage of CDMA in this application is soft hand-
off. More details of the Globalstar system can be found with reference 4.
The GPS navigation system uses DS-SS CDMA for the transmission of signals that
permit precise location of a receiver in three dimensions. Up to 12 GPS satellites could
be visible to a receiver close to the earth’s surface at any one time. CDMA is used to share
a single RF channel in the receiver between all of the GPS satellite transmissions. Chap-
ter 12 gives details of the GPS signal structure and describes the process of data recov-
ery from the DS-SS satellite signals.
Spread Spectrum Transmission and Reception

This discussion of CDMA for satellite communications will be restricted to direct se-
quence systems, since that is the only form of spread spectrum that has been used by
commercial satellite systems to date. The spreading codes used in DS-SS CDMA sys-
tems are designed to have good autocorrelation properties and low cross-correlation. Var-
ious codes have been developed specifically for this purpose, such as Gold and Kasami
codes1.
The DS-SS codes will all be treated as pseudonoise (PN) sequences in this discus-
sion. Pseudonoise refers to the spectrum of code, which appears to be a random sequence
of bits (or chips) with a flat, noiselike spectrum. The generation of a DS-SS signal is
illustrated in Figure 6.16. We will begin by assuming that the system uses baseband signals.
Most DS-SS systems generate spread spectrum signals using BPSK modulated versions
of the data stream, but it is easier to see how a DS-SS system operates if the signals are
first considered at baseband. In Figure 6.16, a bit stream containing traffic data at a rate
v v
Modulator
1 1 1 1 1 1 1 1
t t
−1 −1 −1 −1 −1 −1 −1 −1
Incoming bit stream Rb bps Outgoing spread bit stream NRb bps
1 1 1 1 1 1 1 1
t
−1 −1 −1 −1 −1 −1
Spreading PN sequence NRb bps

FIGURE 6.16 The basic principle of a direct sequence spread spectrum (CDMA) system.
Each incoming message data bit is multiplied by the same PN sequence. In this example
the message sequence is 1 1 and the PN sequence is 1 1 1 1 1 1 1.
c06.qxd 09/08/02 14:19 Page 259
Rb, converted to have levels of 1 and 1 V corresponding to the logical states 1 and 0,
is multiplied by a PN sequence, also with levels 1 and 1 V, at a rate M Rb chips
per second. Each data bit results in the transmission of a complete PN sequence of length
M chips.
In the example shown in Figure 6.16, the seven chip spreading code sequence is
1110100, which is converted to 1 1 1 1 1 1 1. The spreading sequence mul-
tiplies the data sequence 0 1, represented as 1 1, leading to the transmitted sequence
1 1 1 1 1 1 1 1 1 1 1 1 1 1 shown at the right in Figure 6.16.
Recovery of the original data stream of bits from the DS-SS signal is achieved by multi-
plying the received signal by the same PN code that was used to generate it. The process
is illustrated in Figures 6.17 and 6.18.
Shift Shift
register
Demodulated
signal input b1 b 2 b 3 b4 b 5 b 6 b 7
− − + − + + + Phase shifters
Adder Output vo = −b1 − b 2 + b 3 − b4 + b 5 + b 6 + b 7
Bits clocked Bits clocked

in Register contents Output in Register contents Output
0 1 1 1 1 1 1 1 vo = 1 8 1 1 1 −1 1 −1 −1 vo = 1
1 −1 1 1 1 1 1 1 vo = 3 9 1 1 1 1 −1 1 −1 vo = 3
2 −1 −1 1 1 1 1 1 vo = 5 10 1 1 1 1 1 −1 1 vo = 1
3 −1 −1 −1 1 1 1 1 vo = 3 11 −1 1 1 1 1 1 −1 vo = 1
4 1 −1 −1 −1 1 1 1 vo = 3 12 1 −1 1 1 1 1 1 vo = 3
5 −1 1 −1 −1 −1 1 1 vo = 1 13 −1 1 −1 1 1 1 1 vo = 1
6 1 −1 1 −1 −1 −1 1 vo = −1 14 −1 −1 1 −1 1 1 1 vo = 7
7 1 1 −1 1 −1 −1 −1 vo = −7
FIGURE 6.17 Data bit recovery using an IF correlator (matched filter). In this example the
PN sequence is seven bits long for illustration. The CDMA chips from the receiver are
clocked into the shift register serially and the shift register contents passed through phase
shifters and added. The phase shifters convert 1 chips to 1 when the correct code is in
the shift register such that all the voltages add to a maximum when the received sequence is
correct. This figure shows the shift register contents and adder output for the chip sequence
in Figure 6.16. Note that a high spurious output of 5 occurs at the third clock step, indicating
that the seven bit sequence used here for illustration has poor autocorrelation properties.
c06.qxd 09/08/02 14:19 Page 260
v v
Modulator
1 1 1 1 1 1 1
t t
−1 −1 −1 −1 −1 −1
Incoming spread bit stream Recovered bit stream
1 1 1 1 1 1 1 1
t
−1 −1 −1 −1 −1 −1
Despreading PN sequence
FIGURE 6.18 A baseband correlator for dispreading CDMA signals. The original bit
stream is recovered by multiplying the received signal by a synchronized copy of the PN
sequence that was used in the transmitter.
At a CDMA receiver which knows the seven bit code, there will be a correlator that
has the code stored as multiplier settings. Figure 6.17 illustrates the correlation process.
Received chips are clocked into a shift register of length equal to the code sequence—
seven stages in this case. The word in the shift register is identified as b1 b2 . . . b7. At
each clock cycle the seven chip word with chip values bi in the correlator shift register
are multiplied by 1 or 1, corresponding to the chips in the code sequence, by the
blocks marked phase shifters in Figure 6.17. Note that received chips are clocked into the
correlator from the left, so the code sequence appears reversed (written from right to left)
in the phase shifters in Figure 6.17. The outputs of the phase shifters are added to give
the output word v0 p1b1 p2b2 p3b3 p4b4 p5b5 p6b6 p7b7. The value of v0
will be 7 or 7 when the correct code sequence exactly fills the seven stages of the
shift register.
Figure 6.17 shows one process by which the code sequence can be detected. The
shift register is originally filled with 1 chips, giving an output from the adder of
v0 1. The sequence generated in Figure 6.16 is clocked in from the left. The adder
output yields values of v0 1 5 3 3 1 1 as the chip sequence moves into
the register. On the next clock cycle, when all seven bits of the sequence are in the seven
stage register, the output of the adder is 7. A threshold detector after the adder with
a threshold level of 6 would detect the threshold crossing and output a logical 0 for
the first data bit. As the next 7 bits are clocked through the shift register, the output of
the adder fluctuates between 1 and 5, reaching v0 7 or 7 when the code se-
quence fills the shift register. Note that the seven bit sequence used for illustration here
would be a poor choice for a CDMA system because of the high spurious output from
the adder at the third clock step. PN sequences used in CDMA systems are required to
have good autocorrelation and good cross-correlation properties to minimize false
threshold crossings. Multiple shift registers like the one shown in Figure 6.17 can be
operated in parallel with each input delayed by an increment of one bit. If there are N
shift registers, N possible code positions are tested at each chip clock cycle giving faster
code acquisition.
When the code sequence is long, the multistage shift register detector shown in Fig-
ure 6.17 becomes unwieldy. Chip-by-chip multiplication is used instead. The multiplier
has inputs bi and pi, where bi is the received chip and pi is a stored PN sequence chip.
c06.qxd 09/08/02 14:19 Page 261
The output of the multiplier is then integrated over the duration of the code sequence to
yield a value N or N where N is the number of chips in the PN sequence. The process
is illustrated in Figure 6.18 for the seven chip code of Figure 6.16. The incoming signal
at the left side of the figure is multiplied by the despreading code sequence to give the
output at the right of the figure. In practice, a low pass filter is used instead of an integrator
to avoid the need to synchronize discharge (dumping) of the integrator contents when a
bit is detected. If the correct code is present in the input signal, the output of the multiplier
is 1 or 1 as each chip is present in the multiplier. In a practical DS-SS CDMA system,
there will be several other CDMA signals present at the correlator input as well as the
wanted code. The integrated value of the multiplier output for cross-correlation of any
other code with the stored code will yield a value 1 or 1 if the codes have ideal cross-
correlation properties. Given a sufficiently large value for M (i.e., a long PN code
sequence), it is possible to detect the bits of the wanted code in the presence of a large
number of unwanted CDMA signals.
The seven chip code used to illustrate DS-SS code correlation is not a good code
for a CDMA system. It exhibits poor autocorrelation because at one position in the
shift register the adder output is 5, too close to the peak value of 7. Noise added
to the incoming code could easily push the correlator output above the threshold caus-
ing a detected bit error. Ideally, we would like to have CDMA codes of length M chips
that have autocorrelation values of 1 or 1 everywhere except when the code is
aligned correctly, when the value should be M. When a different code is clocked into
the shift register in Figure 6.17, the cross-correlation should be 1 or 1 on all clock
cycles. Very few known codes have these ideal properties. There are Barker codes with
sequence lengths up to 23 chips that meet these requirement, but practical DS-CDMA
systems normally use longer codes with nonideal correlation properties. For example,
GPS satellites use 1023 chip Gold codes for the CA (course acquisition) code se-
quence that are built up from maximal length sequences which are easy to generate
with a shift register.
Practical CDMA systems use BPSK waveforms and correlate the received sig-
nals at IF rather than baseband. The shift register shown in Figure 6.17 is typically a
single stage multiplier, as shown in Figure 6.18, and the incoming signal and the PN
sequence are BPSK waveforms with 0° or 180° phase shifts. Multiplication of two
identical, cophased BPSK waveforms yields an output of 1. If the phase of the in-
put waveform is reversed (indicating that the original data bit was a 0 rather than a 1)
the output is 1. Coherent phase detection is required so that the IF waveforms can
be added in phase, but the correlation principle is the same. The main difficulty in DS-
SS CDMA receivers is that the received signal is buried in the noise, so the usual
techniques for carrier recovery cannot be used. Baseband correlation is rarely used in
DS-SS CDMA systems because the signals entering the correlator have CN ratios less
than 0 dB (negative CN ), so the signals always look like noise. A complete DS-SS
receiver and correlator for the GPS CA DS-SS signal is described in Chapter 12. The
GPS CA code sequence is 1023 chips in length, so chip-by-chip multiplication is used
in the receiver.
The bandwidth occupied by the original data signal with bit rate Rb, if transmitted
using BPSK and 0.4 square root raised cosine (RRC) filters, would be 1.4 Rb Hz.
The spread spectrum signal occupies a bandwidth M 1.4 Rb Hz, and must be received
through an IF RRC filter with a noise bandwidth of M Rb Hz. Suppose that a BPSK
receiver with the appropriate RRC filter with noise bandwidth Rb Hz receives the BPSK
signal with CN 11 dB. If we do not change the power level of the original BPSK sig-
nal by the process of spreading it into a bandwidth of 1.4 M Rb Hz, the CN in the spread
c06.qxd 09/08/02 14:19 Page 262
spectrum receiver will be 11 10 log10 M dB. If M is large (e.g., 1023 as in the GPS CA
code) the CDMA signal will have a CN ratio in the receiver much less than 0 dB;
19.1 dB for the GPS CA signal example. The despreading process using a correlator
to recover the original signal adds a processing gain equal to the numerical value M to
the (CN)SS ratio of the received spread spectrum signal. Hence the signal-to-noise (SN)out
ratio in the spread spectrum receiver after the correlator is given by
1SN2 out 1C N2 SS 10 log10 M (6.14)
This SN ratio must be sufficiently high for the receiver to recover the bits of the trans-
mitted signal with a reasonable bit error rate. For example, if a BER no larger than 106
is required, (SN)out must be greater than 11.0 dB, allowing a 0.4-dB implementation mar-
gin with no forward error correction.
DS-SS CDMA Capacity

In a DS-SS CDMA system where there are a number of CDMA signals present at the in-
put to each receiver, it is usual to treat the unwanted (interfering) CDMA signals as noise.
If a receiver has an input containing Q signals, each at a power level C watts, and the
receiver thermal noise power is Nt watts, the (CN)in ratio for the wanted signal is
approximately
1CN2 in 10 log10 3C 1Nt 1Q 12 C2 4 dB (6.15)
where [Nt (Q 1) C) watts is the total noise at the receiver input. The term (Q 1)
C I watts is the power of the Q 1 interfering CDMA signals. (Note that Nt and C must
be added in watts, not decibel units.) The correlator in the receiver adds a processing gain of
10 log10 M dB to the input CN, as seen in Eq. (6.14), and outputs a correlated signal with
a signal-to-noise ratio (SN)out. Hence the output SN ratio for the bit stream in the receiver
is given by
1SN2 out 10 log10 3C 1Nt 1Q 12 C2 4 10 log10 M dB (6.16)
If Q is a large number, it is probable that [Nt (Q 1) C] (Q 1) C watts,
and then Eq. (6.16) reduces to
1SN2 out 10 log10 31 1Q 12 4 10 log10 M 10 log3M 1Q 12 4 dB (6.17)
If Q is also large such that M W 1 then
1SN2 out 10 log10 1MQ2 dB (6.18)
Examination of Eq. (6.18) shows that M, the number of chips in the spreading code
must be 10 times larger than Q if the output SN ratio is to be greater than 10 dB, and
that the system capacity is independent of the thermal noise power in the receiver. The
bit rate of each signal is given by
Rb RcN B 3N 11 a2 4 (6.19)
where Rc is the chip rate and B is the transponder bandwidth. The total bit rate for the
transponder is given by M Rb B M[Q (1 )]. If M must be 10 times larger
than Q to allow demodulation of the spread signal without many bit errors, the total bit
rate through the transponder in bits per Hertz using CDMA will be numerically less
than one-tenth of the bandwidth in hertz. This results in poor utilization of the RF band-
width when CDMA is used, compared to FDMA or TDMA, as the following example
demonstrates.
c06.qxd 09/08/02 14:19 Page 263
EXAMPLE 6.8.1 CDMA in a Fixed Earth Station Network

A DS-SS CDMA system has a number of earth stations sharing a single 54 MHz bandwidth Ka-band
transponder. Each station has a different 1023 bit PN sequence which is used to spread the traffic
bits into a bandwidth of 45 MHz. The transmitters and receivers use RRC filters with 0.5 and
the chip rate is 30 Mcps. Determine the number of earth stations that can be supported by the CDMA
system if the correlated output SN 12 dB.
Equation (6.17) gives
1SN2 12 dB 10 log3 M 1Q 12 4 30.9 10 log 1Q 12

Hence
10 log 1Q 12 18.9 dB
Q 77 1 76
Each of the carriers has a bit rate of 30 Mbps1023 29.33 kHz, so the transponder carries a to-
tal bit rate of 77 29.33 kbps 2.258 Mbps. A 54-MHz bandwidth transponder operated in FDMA
or TDMA would have a much higher capacity.
The capacity of the system can be improved by adding half rate forward error (FEC) control
to the baseband signal to reduce the SN required for detection of the bits in the receiver. If the
FEC system has a coding gain of 6 dB, we can use SN 12 6 6 dB. Using Eq. (6.18),
because we now know M W Q
6 dB 10 log 1MQ2
gives Q M4 255 channels. The data bit rate of each channel (before application of half
rate FEC) is now 14.66 kbps and the total throughput of the transponder is 255 14.66 kbps
3.74 Mbps. This is still well below the capacity of a FDMA or TDMA system.
We can conclude that CDMA is useful in commercial systems only where efficient use of
satellite capacity is not important, or where the ease with which stations can leave and join the net-
work outweighs the loss of efficiency, or where power limitations in the transponder ensure that it
cannot be heavily loaded.
EXAMPLE 6.8.2 CDMA in an LEO Satellite Network

An LEO satellite communication systems uses direct sequence CDMA as the multiple access method
for groups of terminals within each of its multiple antenna beams. The terminals generate and re-
ceive compressed digital voice signals with a bit rate of 9.6 kbps. The signals are transmitted and
received at a chip rate of 5.0 Mbps as BPSK modulated DS-CDMA. In the absence of any other
CDMA signals, the input power level at the receiver input is 146.0 dBW for one CDMA signal,
and the noise temperature of the receiving system is 300 K. The satellite transmits 31 simultaneous
CDMA signals. Find the SN ratio for the 9.6-kbps BPSK signal after despreading, and estimate
the BER of the data signal, given a system implementation margin of 1 dB. If two of the multiple
beams from the satellite overlap, so that a second group of 31 DS-CDMA signals is present at the
receiver, find the BER of the wanted signal.
The thermal noise power in the receiver is Nt kTsBn. For the chip rate of 5.0 Mbps with
BPSK and ideal RRC filters, Bn 5.0 MHz. Hence
Nt 228.6 24.8 67.0 136.8 dBW 2.09 1014 W

There are 30 interfering CDMA signals overlaid in the 5-MHz bandwidth of the receiver fil-
ter. The total interfering power is
I 30 Pr 146.0 14.8 dB 131.2 dBW 7.59 1014 W

The carrier-to-noise plus interference ratio must be calculated in watts, not dBW, because we
cannot add noise and interference in decibel units, only in watts. The carrier-to-noise ratio in the
c06.qxd 09/08/02 14:19 Page 264
receiver for the wanted CDMA signal is

C 1Nt I2 2.51 1015 12.09 1014 7.59 1014 2
2.5196.8 0.0259
15.9 dB
The carrier power is well below the noise plus interference power, so the wanted carrier is
hidden below the noise and interference. This is called a low probability of intercept signal. CDMA
was first used by military radio communication systems because detection of a signal which is below
the noise floor is difficult.
The coding gain, Gc, for the CDMA receiver is given by the chip rate divided by the bit rate
Gc RcRb 5.0 Mcps9.6 kbps 520.8 27.2 dB
Hence, after correlation of the wanted code (despreading), the SN ratio of the 9600-bps
BPSK signal is
SN 15.9 27.2 11.3 dB
With an implementation margin of 1 dB, the effective SN is 10.3 dB 10.7 as a power ra-
tio. For BPSK, the BER is
Pe 12 erfc3 11CN2 eff ratio 4 12 erfc33.274 2 106
If a second group of 31 signals is present at the receiver from an overlapping satellite beam,
there will be additional interference which lowers the C(N I) ratio. The interfering power from
31 signals is
I 31 Pr 146.0 14.9 dB 131.1 dBW 7.76 1014 W
Hence the new C(Nt I) ratio is
C 1Nt I 2 2.51 1015 12.09 1014 7.59 1014 7.76 1014 2
2.51174.4 0.0144
18.4 dB
After correlation of the wanted code the SN ratio of the 9600 BPSK signal is
SN 18.4 27.2 8.8 dB
With an implementation margin of 1 dB, the effective SN is 7.8 dB 6.02 as a power ratio.
For BPSK the BER is
Pe 12 erfc3 11CN2 eff ratio 4 12 erfc32.454 3 104
We would need to add forward error correction to the baseband signal to improve the bit er-
ror rate. To achieve a BER of 106 in this case, a coding gain of about 3 dB would be adequate.
With half rate convolutional coding, a coding gain of 5.5 to 6 dB is typical, which would provide
a margin of 3 dB over a BER of 106 and a baseband data rate of 4.8 kbps. This bit rate will sup-
port a single digital speech channel with LPC linear predictive coding compression.
The advantage of overlapping beams in a mobile satellite system is that the wanted signal
can be transmitted by both satellites (using different CDMA codes) and blockage of one beam by
an obstruction on the ground does not cause loss of the signal if the other beam can still be re-
ceived. The wanted signal from both satellites can be combined at baseband using a rake receiver,
which improves the BER. With optimum combining of the same baseband signal, the BER will be
the same as for a single beam with 31 users.
EXAMPLE 6.8.3 GPS

The Global Positioning System (GPS) uses direct sequence CDMA for both the CA and P code
transmissions. The design and operation of GPS is discussed in detail in Chapter 12, from which
this example of a direct sequence CDMA system is drawn.
c06.qxd 09/08/02 14:19 Page 265
The CA code transmissions from GPS satellites are 1023-bit PN sequences, formed into
64 Gold codes. At any given time, no more than 10 GPS satellites are visible, so interference with
the wanted signal is limited to no more than nine overlaid CDMA signals. In this example we
will assume for simplicity that the signals are all received with equal power. There are variations
in the transmitted power between satellites, and a satellite close to the horizon has a longer path
length, so there will be variations in the received power level of individual satellite signals in
practice.
The received power level for a typical CA signal is given by the downlink budget in the
table below, assuming 0 dB receiving antenna gain. The CA code is transmitted at a bit rate of
1.023 Mbps using BPSK modulation. The receiver noise bandwidth is assumed to be 2 MHz.
Satellite EIRP (dBW) 26.8

Path loss (dB) 186.8
Receive antenna gain (dB) 0
Pr (dBW) 160.0
The interference from nine satellite spread spectrum signals of equal power for the CA code
is given by
I 160.0 9.5 150.5 dBW 8.91 1016 W
The thermal noise power in a noise bandwidth of 2 MHz for a system noise temperature of
273 K is kTs Bn where
Nt 141.2 dBW 7.59 1015 W
The noise and interference powers must be added in watts, not in decibels:
Nt I 8.48 1015 W 140.7 dBW
The nine interfering satellites have lowered the CN ratio for the wanted signal by 0.7 dB,
a relatively small decrease. In a GPS CA code receiver, thermal noise is the dominant factor, not
interference from other satellites as in Example 6.8.2.
The CN for one CA code signal with nine interfering signals is
C 1Nt I2 160.0 1140.72 19.3 dB
The theoretical coding gain for a 1023-chip code sequence is 10 log10 1023 30.1 dB. Hence
the SN ratio of the correlated CA signal, assuming ideal correlation, is
SN 19.3 30.1 10.8 dB
The GPS CA code signal recovered from the correlator is a 1-kbps polar binary waveform
with amplitude V volts. The primary use of this signal in a commercial GPS receiver is to deter-
mine the time of arrival of the CA code sequences, as it is the time that code sequences arrive
from each of the GPS satellites that provides the pseudorange information from which the position
of the receiver is calculated. The CA code also provides vital navigation message data that is needed
in the position calculation. (See Chapter 12 for details.)
Navigation data are modulated onto the 1-kbps signal at 50 Hz, so there are 20 V or 20 V
samples in succession for each 1 or 0 data bit of the navigation message. Allowing an implemen-
tation margin of 1 dB in recovery of the signal, the effective SN after the correlator is 9.8 dB which
will give a BER around 6 106. Averaging 20 samples will improve the error rate by a factor of
120, giving a BER close to 106. At a 50-bps data rate, errors in the navigation message will rarely
occur. (Theoretically, there will be a bit error in the navigation message once every 2 days. The
entire message repeats every 12.5 min, and vital data repeats every 30 s. The error is quickly
overwritten by new data.)
In this example we see that GPS satellites make excellent use of CDMA as a way to obtain
a high-speed bit stream from which timing information can be obtained, an essential ingredient for
c06.qxd 09/08/02 14:19 Page 266
any time of arrival position location system, and also a low speed data stream that provides the nav-
igation information for the solution of the location problem.
The example here uses nine interfering satellites with the same power level. Most GPS
receivers select the four strongest satellite signals to use in the position location solution. A more
realistic scenario would have four satellites at the maximum receive power level and the remainder
at a lower level, since GPS satellites orbit in constellations of four, with one constellation always
visible, to improve the accuracy of position location measurements. Thus we should expect less than
0.7 dB degradation in CN due to interference by other satellites’ CDMA signals, and the proba-
bility of a bit error in the navigation message then becomes very small.
Interestingly, the eastern European satellite navigation system known to the Western world
as GLONASS uses FDMA for multiple access. There are a maximum of 64 individual CA signals
in the GPS system, from a possible 64 GPS satellites, each of which could be allocated its own RF
frequency within a 1 MHz band. If all the parameters of the GPS system were held the same, but
multiple accesses were changed from DS-CDMA to FDMA, a multichannel receiver could receive
four (or more) 1 kbps BPSK data streams with CN ratio of 10.8 dB. This is essentially how the
GLONASS system works, with accuracy comparable to that achieved in the GPS system and a much
simpler receiver.
6.9 SUMMARY
Multiple access is the process by which a large tion transmits one traffic burst per frame. Frames
number of earth stations interconnect their links and individual traffic bursts are identified by stan-
through a satellite. In frequency division multiple dardized bit sequences called unique words. One of
access (FDMA), stations are separated by fre- the major technical problems in implementing
quency, while in time division multiple access TDMA is synchronization. Once synchronization is
(TDMA), they are separated in time. In code divi- acquired, it must be maintained dynamically to
sion multiple access (CDMA), stations use spread- compensate for orbital motion of the spacecraft.
spectrum transmissions with orthogonal codes to TDMA is often combined with FDMA, so that a
share a transponder without interference. Multiple small number of earth stations share a TDMA frame
access may be preassigned or demand (DAMA), de- forming one FDMA access to a transponder. This is
pending on whether or not it responds to changing called MF-TDMA.
traffic loads. In code division multiple access (CDMA) sta-
Frequency division multiple access is the most tions transmit at the same time and in the same fre-
widely used multiple access scheme. In it each earth quency bands using spread-spectrum (SS) techniques.
station is assigned frequency bands for its uplink CDMA avoids the centralized network control re-
transmissions. Because of the TWT backoff required for synchronization in TDMA, but tends to
quired to reduce intermodulation distortion with achieve rather poor spectral efficiency. The Global-
bent pipe transponders, the spectral efficiency (i.e., star LEO satellite system was designed to use
the number of channels that can be carried per CDMA, with the advantage that an earth station can
megahertz of bandwidth) degrades with the number receive the same signal from more than one satellite
of stations that access a transponder. FDMA is at the same time, allowing soft handoff between
widely used with VSAT earth stations and SCPC satellites.
systems where the uplink from the earth station is Random access is used in systems that have low
at a low power level. traffic requirements and can tolerate less than 18%
In time division multiple access (TDMA), earth utilization of the RF channels. The advantage of ran-
stations transmit in turn. Since only one carrier is dom access is that no central network control is
present at a time, no TWT backoff is required and needed.
thus full transponder EIRP is available. TDMA per- Digital links between computers require protocols
formance does not degrade with the number of ac- to ensure efficient transfer of data, and invariably use
cesses. TDMA transmissions are organized into some form of packet communication. Satellite sys-
frames; a frame contains one or two reference bursts tems have tended to use proprietary protocols, with
that synchronize the network and identify the frame the result that different satellite systems are not
and a series of traffic bursts. Each participating sta- compatible.
c06.qxd 09/08/02 14:19 Page 267
PROBLEMS 267
REFERENCES
1. L. W. COUCH, Digital and Analog Communication Sys- 12. J. J. SPILKER, Jr., Digital Communications by Satellite,
tems, Prentice-Hall, Englewood Cliffs, NJ, 6th Ed., 1998. Prentice-Hall, Englewood Cliffs, NJ, 1977.
2. FERREL G. STREMLER, Introduction to Communication 13. Astrolink URL.
Systems, Addison-Wesley, Reading, MA, 3rd Ed., 1984. 14. Spaceway URL.
3. J. L. EVERETT, (ed.) VSATs, Peter Peregrinus, IEE UK, 15. Webster’s New World Dictionary, 2nd Ed., William
1992. Collins World Publishing Company, 1980.
4. www.globalstar.com 16. IAN F. AKYILDIZ and SEONG-HO JEONG, “Satellite ATM
5. K. MIYA, ed., Satellite Communications Technology, Networks: A Survey,” IEEE Communications Maga-
KDD Engineering and Consulting, Tokyo, Japan, 1981. zine, Vol. 35, No. 7, pp 30–39, July 1997.
6. TRI T. HA, Digital Satellite Communication, McGraw- 17. DAVIDOFF, Amateur Satellite Handbook, American
Hill, New York, 1990. Radio Relay League. AARL, Newton, CT.
7. Intelsat TDMAIDSI System Specification (TDMAIDSI 18. www.Bluetooth.com
Traffic Terminals), (BG-42-65E Rev. 2), Intelsat, 19. www.globalstar.com
Washington, DC, June 23, 1983. 20. “Broadband via Satellite,” IEEE Communications Mag-
8. Technical Requirements for Inmarsat Standard-A Ship azine, Special Issue, July 1997.
Earth Stations (Issue 2), International Maritime Satel- 21. “Broadband Satellite Network Performance,” IEEE
lite Organization, London, UK, February 1983. Communications Magazine, Special Issue, March
9. ROBERT M. GAGLIARDI, Satellite Communications, 1999.
Lifetime Learning Publications, Belmont, CA, 1984. 22. C. K. TOH and V. O. K. LI, “Satellite ATM Network
10. K. FEHER, Digital Communications: Satellite Earth Sta- Architectures,” IEEE Network, September/October,
tion Engineering, Prentice-Hall, Englewood Cliffs, NJ, 1999.
1983. 23. B. G. EVANS and R. TAFAZOLLI, “Future Multimedia
11. V. K. BHARGAVA, D. HACCOUN, R. MATYAS, and P. NUSPL, Communications via Satellite,” International Journal of
Digital Communications by Satellite, John Wiley & Satellite Communications, 14, 467–474, 1996.
Sons, New York, 1981.
PROBLEMS
1. You are designing an FDM-FM-FDMA analog Problems 2 through 5 all involve a satellite and
link that will occupy 36 MHz of an INTELSAT VI earth stations with the same specifications.
transponder. The uplink and downlink center fre- Five earth stations share one transponder of a 64
quencies of the occupied band are 5985.5 MHz and GHz satellite. The satellite and earth station charac-
3760.5 MHz. The distance from the satellite to your teristics are given below:
earth station is 40,000 km. The saturation uplink flux
Satellite Transponder BW 36 MHz
density for your uplink is 75 dBW/m2 and the satel-
lite’s GT is 11.6 dBK1. At saturation the Transponder gain 90 dB (max)
transponder EIRP for your downlink is 29 dBW and Input noise temp. 550 K
the earth station’s GT is 41 dBK1. The transponder Saturated output power 20 W (max)
is linear in that the EIRP in dBW is BO dB below the
4-GHz antenna gain 20.0 dB
saturation value when the uplink flux density is
backed off BO dB below saturation. The intermodu- 6-GHz antenna gain 22.0 dB
lation carrier-to-noise ratio, (CN), in dB, is related Earth station 4-GHz antenna gain 60.0 dB
to the backoff BO in dB by 6-GHz antenna gain 63.0 dB
1CN2 I 7.86 0.714 BO Receive system temp. 100 K
Path loss At 4 GHz, Lp 196 dB
In other words, at saturation the value of (CN)I
is 7.86 dB. Find the maximum overall carrier-to-noise At 6 GHz, Lp 200 dB
ratio (CN), in dB that this link can achieve. What 2. The stations all operate in a TDMA mode.
backoff must be used to achieve it? (When you need Speech signals are sampled at 8 kHz, using
a frequency in your calculations, use the uplink or 8 bits/sample. The sampled signals (PCM) are then
downlink center frequency as appropriate.) Make multiplexed into 40-Mbps streams at each station,
your calculations for beam center. using QPSK.
c06.qxd 09/08/02 14:19 Page 268
a. Find the bit rate for each PCM signal. earth station in every frame. It takes 1 s to reposi-
b. The number of speech signals (as PCM) that could tion the satellite antenna beam from one earth station
be sent by each earth station, as a single access, with to another. Only the downlink antenna beam is
no overhead (i.e., no header or CRC, etc.). This is a switched; the uplink uses a common zone beam. The
TDM data stream. frame length to be used is 1000 s, with a 148-bit
c. The shortest frame time for any TDMA scheme. preamble and 2-s guard times between transmis-
sions arriving at the satellite The extra antenna gain
3. Assume that the TDMA system uses a 125-s at the satellite is traded for an increase in the data rate
frame time. Find the number of channels that each by using 16-QAM on the downlink. Other parame-
earth station can send within the TDMA frame when: ters of the system are unchanged.
a. No time is lost in overheads, preambles, and the a. Find the number of earth stations that can share
like. the transponder.
b. A 5-s preamble is added to the beginning of each b. Find the total data throughput of the transponder
earth station’s transmission. after all preamble bits have been removed.
c. A 5-s preamble is added to each station’s trans-
8. An LEO satellite system transmits compressed
mission and 2-s guard band is allowed between
digital voice signals to handheld terminals (sat-
every transmission.
phones). The satphones work in groups of 10. The
4. A 750-s frame time is used instead of a 125-s inbound bit stream from the satphone to the satellite
frame. Find the new channel capacities of the earth is at 10 kbps. The data are sent as a BPSK signal.
stations for the cases in Problem 3 above. The outbound bit stream from the satellite is at a bit
5. Find the earth station transmitter power and re- rate of 100 kbps, and consists of packets addressed
ceived (CN) when the system is operated: to each of 10 satphones. This signal is sent using
a. In TDMA with the transponder saturated by each QPSK, and all 10 satphones receive the 100 kbps bit
earth station in turn. stream.
The system operates in L band where rain fading
b. In FDMA with 3-dB input and output backoff.
can be ignored, but blockage from buildings and trees
6. A digital communication system uses a satellite is a significant factor. The satellite uses onboard pro-
transponder with a bandwidth of 54 MHz. Several cessing and multibeam antennas. The links use square
earth stations share the transponder using QPSK root raised cosine (RRC) filters with 0.5. In this
modulation using either FDMA or TDMA. Standard question we will be concerned only with the links
message data rates used in the system are 80 kbps and between the satellite and the satphones, and ideal
2.0 Mbps. The transmitters and receivers in the sys- RRC filters will be assumed.
tem all use ideal RRC filters with 0.25, and a. What is the noise bandwidth of the narrowest band-
FDMA channels in the satellite are separated by 100- pass filter in: (i) the satphone receiver and (ii) the
kHz guard bands. When TDMA is used, the TDMA satellite receiver for the inbound link?
frame is 125 s in length, and a 2-s guard time is
b. What is the occupied RF bandwidth of the radio
required between each access. A preamble of 148 bits
signals of: (i) the inbound link (phone to satellite) and
must be sent by each earth station at the start of each
(ii) the outbound link (satellite to phone)?
transmitted data burst.
c. The inbound link has clear air (CN)0 18.0 dB
a. What is the symbol rate for the 80-kbps and 2.0- and the BPSK demodulator on the satellite has an im-
Mbps QPSK signals sent using FDMA? plementation margin of 0.5 dB. What is the clear air
b. What is the symbol rate of each earth station’s BER in the baseband of the satellite receiver?
transmitted data burst when TDMA is used? d. What is the available fade margin [for (CN)0 on
c. Calculate the number of earth stations that can be the uplink to the satellite] if the inbound link operat-
served by the transponder when 80-kbps channels are ing threshold is set at BER 104?
sent using (i) FDMA and (ii) TDMA. e. The outbound link has clear air (CN)0 18.0 dB
d. Calculate the number of earth stations that can be and the QPSK demodulator in the satellite phone has
served by the transponder when 2.0-Mbps channels an implementation margin of 0.8 dB. What is the clear
are sent using (i) FDMA and (ii) TDMA. air BER?
7. The capacity of the TDMA system described in f. What is the available fade margin [for the over-
Problem 6 can be increased substantially by using all (CN)0 on the downlink to the satphone] if the
satellite switched TDMA. In a group of earth stations, outbound link operating threshold is set at BER
each station sends a 2.0-Mbps signal to every other 105?
c06.qxd 09/08/02 14:19 Page 269
PROBLEMS 269
9. A Ka-band satellite broadcasts digital television The stations share the transponder using FDMA,
signals over the United States. The nominal bit rate with 51-kHz guard bands between the edges of the
of the signal is 28 Mbps. The digital signal can con- RF signals. The RRC filters used in the VSAT trans-
vey up to 10 prerecorded NTSC video signals. QPSK mitters and the hub station receivers have a roll-off
modulation is used, and error mitigation techniques factor 0.4. To minimize intermodulation between
are employed that provide an effective coding gain of signals, the transponder is operated with 3-dB output
6 dB. [Coding gain of 6 dB means that when the back off.
(CN)0 value of the received signal is X dB, the BER a. Calculate the RF bandwidth occupied by each
corresponds to CN (X 6) dB.] VSAT transmission.
The QPSK demodulator in the receiver has an im-
b. Calculate the maximum number of VSAT stations
plementation margin of 1.6 dB. The transmitters and
that can be included in the network if the transpon-
receivers use ideal RRC filters with 0.25.
der is bandwidth limited.
a. What is the occupied bandwidth of the RF TV
c. Calculate the clear air CN ratio for a received sig-
signal?
nal at the hub station, and the link margin, if the num-
b. What is the symbol rate of the transmitted QPSK
ber of VSAT stations in the network is increased to
signal, and the noise bandwidth of the earth terminal
the number you calculated in (b) above. Remember
receiver?
that the power available from the transponder is fixed.
c. The minimum permitted BER after error mitiga- Adding more stations to the network lowers the power
tion in the receiver is 106. What is the minimum per- per channel at the transponder output.
mitted (CN)0 for the digital TV receiver?
d. The Ka-band link suffers rain attenuation that re- Part 2
duces (CN)0 in the receiver by 7 dB for 0.1% of the The VSAT network described in Part 1 is modified to
year. If the BER is 106 under the 0.1% year condi- be operate with TDMA on the VSAT uplinks instead
tions, what is the clear air (CN)0 value? of FDMA. There are 100 VSAT stations in the net-
e. A new coding algorithm is developed that provides work.
a coding gain of 7 dB with a bit rate that increases to The TDMA frame has a duration of 2 ms and is
30 Mbps. Assuming that the RRC filters in the sys- made up of 100 bursts from the 100 VSAT stations.
tem can be changed to match the new symbol rate, There is a preamble of 100 symbols at the start of
does implementation of the new coding algorithm im- each VSAT station burst, and each burst is separated
prove the system performance? If so, what is the new from the next burst by a guard time of 1.0 s.
(CN)0 margin? a. There are 100 VSAT station RF bursts in each
10. This problem is about multiple access techniques frame of 2.0 ms, and 100 guard times of 1.0 s. What
in the inbound link of a VSAT network. This set of is the duration of each station’s burst?
questions compares the operation of a Ku-band satel- b. Each VSAT station must deliver 128 kbps of data,
lite transponder in FDMA, in TDMA, and in FDMA- in the form of 128 k symbols, every second. How
RA. There are three parts to the problem. many data symbols are there in each RF burst, and
Part 1 what is the total number of symbols per burst after
accounting for the 100 symbol preamble at the be-
100 VSAT stations in a star network share one 54-
ginning of each burst? Hence find the burst rate for
MHz transponder using FDMA. Each VSAT station
the VSAT transmissions in symbols per second.
has a solid-state transmitter with an output power of
1 W and an EIRP of 41 dBW from a 1.1-m diame- c. If all the VSAT stations, and the hub receiver, have
ter antenna. The transmitted data signals have a bit RRC filters with roll-off factor 0.4, what is the
rate of 128 kbps and are transmitted using QPSK RF bandwidth occupied in the transponder?
modulation and half rate FEC, giving a symbol rate If the symbol rate of transmissions were increased
of 128 ksps. At the hub station, the overall CN ra- until all 54-MHz bandwidth of the transponder were
tio for each signal received from a VSAT station is filled, what is the maximum number of VSAT stations
16 dB in clear air. in the network?
The (CN)up ratio for one channel in satellite d. The transponder can be operated with 1-dB out-
transponder is 19.0 dB, and the (CN)dn ratio for one put backoff when TDMA is used, and the implemen-
channel in the hub receiver is 19.0 dB. The threshold tation margin of the hub receiver is 1.5 dB. The EIRP
CN ratio in any hub station receiver for BER 106 of the VSAT stations must be increased because the
is 9.0 dB. This includes the receiver implementation noise bandwidth of the hub receiver has increased. By
margin of 0.5 dB. comparing the symbol rate with 100 FDMA VSAT
c06.qxd 09/08/02 14:19 Page 270
stations in Problem 1 with the TDMA symbol rate for Uplink frequency for transponder 2 28.1 GHz
100 VSAT stations in part (b) above, estimate the Downlink frequency for transponder 2 21.6 GHz
decibel increase in EIRP required from each VSAT
Range to satellite (all stations) 38,000 km
transmitter.
Comment on the feasibility of transmitting this Satellite transponders
power level from a VSAT station. Saturated output power 30 W
Part 3 Transponder bandwidth 54 MHz
The FDMA system described in Part 1 is used with Transponder input noise temperature 500 K
random access to serve a very large number of VSAT Antenna gain, on axis, regional beam 33 dB
stations. All the parameters of Part 1 are the same,
Antenna gain, on axis, switched
except that each station has a small amount of data
spot beam 48 dB
to send at varying intervals of time. The average mes-
sage data rate for each VSAT station is 5.0 kbps and VSAT station parameters
the maximum permitted channel loading is 12%. Transmitter output power 1.0 W
a. How many VSAT stations can share each RF Transmit frequency 28.2 GHz
frequency? Receive frequency 21.7 GHz
b. What is the maximum number of VSAT stations
Antenna diameter 0.5 m
in the network when the number of RF channels is
the value you calculated in Part 1(c)? Aperture efficiency 65%
11. This problem examines the use of a Ka-band satel- Receiver system noise temperature
lite to provide connection to the Internet from a small (clear air) 250 K
two-way terminal. The problem is in three parts. The Receiver system noise bandwidth TBD
first part establishes the design of the communications Hub station parameters
links and terminals. The second part examines the ca-
Maximum transmit power 100 W
pacity of the satellite. The third part looks at changes
that must be made to support portable terminals. Transmit frequency 28.1 GHz
Part 1 Communication links Receive frequency 21.6 GHz
Description of the satellite communication system Receiver system noise temperature
A Ka-band GEO satellite is located at longitude (clear air) 250 K
100° W. Star networks can be built with a single hub Antenna diameter 4.0 m
station, two transponders on the GEO satellite, and a Aperture efficiency 65%
number of earth stations, identified here as VSATs.
Receiver system noise bandwidth TBD
The major parameters of the system components are
given below. You may not need all of these parame- Atmospheric losses and miscellaneous losses
ters to answer the questions, and additional parame- In clear air at 28 GHz 2.0 dB
ters are given in the individual questions. In clear air at 21 GHz 2.0 dB
The Ka-band satellite serves the United States.
Constants: Boltzmann’s constant, k, 1.38 1023
Coverage of the 48 contiguous states is achieved by
JK 228.6 dBW/K/Hz
a regional beam, but the satellite also carries an ad-
vanced antenna system with satellite switched spot Part 1 Problems
beams that allow data packets to be transmitted to CN ratios in clear air conditions
small earth stations with a high EIRP. This allows Make all calculations for the worst case of a VSAT
high-speed data transmission in the outbound link. station that is located on the 3 dB contour of the
The system is designed primarily to support In- satellite antenna beam (regional or spot), and for a hub
ternet access via satellite, with highly asymmetrical station on the 2 dB contour of the regional beam.
links. Requests for access to the Internet are made by The spot beam is used only for transmissions at 21
users at a low data rate through the satellite’s region GHz from the satellite to the customers’ earth stations.
beam. Replies from the Internet can be received at a All other links use the satellite’s regional beam.
high data rate using the satellite’s spot beam.
a. Calculate the free space path loss for a 38,000-km
System values path at 28.2 and 21.7 GHz.
Uplink frequency for transponder 1 28.2 GHz b. Calculate the gains of the hub and VSAT antennas
Downlink frequency for transponder 1 21.7 GHz at frequencies of 28.2 and 21.7 GHz.
c06.qxd 09/08/02 14:19 Page 271
PROBLEMS 271
c. Calculate the inbound overall CN in the hub sta- link is required to achieve a 99.7% availability for a
tion receiver in a noise bandwidth of 128 kHz when typical VSAT station for which slant path attenuation
the VSAT has a transmitter output power of 1 W and exceeds 4 dB at 21.7 GHz and 7 dB at 28.2 GHz, for
accesses the regional beam on the satellite. Make the 0.3% of an average year.
overall CN calculation for a single QPSK signal The link is declared unavailable if the BER ex-
which is transmitted by transponder 1 at an output ceeds 106 in the data stream supplied to the cus-
power of 1 W. tomer, or output by the hub station.
d. Calculate the outbound overall CN in transponder Begin your analysis by assuming that 20 active
2 with a hub station transmit power of 1 W. Make your VSAT stations share the output power of transponder
calculation in a receiver noise bandwidth of 1 MHz, 1 equally at all times using QPSK-SCPC-FDMA.
for a single QPSK signal, with the output power of Half rate FEC coding is used in the inbound and the
transponder 2 set at 1 W and the spot beam of the outbound link and provides a coding gain of 5 dB at
satellite transmitting to the customers’ terminals. a BER of 106 in the recovered data stream. The im-
Estimate the beamwidth of the spot beam from the plementation margin of the QPSK demodulators in
satellite. Using a map of the United States, estimate the hub receiver is 0.5 dB, and in the VSAT receiver
the minimum number of spot beam positions required the implementation margin is 0.8 dB. Assume that
to serve the entire United States. there are always 20 active VSAT stations receiving
Part 2 System performance data from the outbound link in packet form, using
TDM and a single QPSK carrier. Assume linear
Connection to the Internet is achieved by the follow-
operation of the transponders, but include the effect
ing procedure.
of increased sky noise when rain is present on the
The customer sends a connection request, in the
uplink.
form of a data packet, to the hub station via the satel-
Transponder 1 (inbound, SCPC-FDMA) is oper-
lite and its regional beam. The hub station decodes
ated with 2-dB output backoff.
the request and notes the location of the station. The
connection between the Internet and the hub is es- Transponder 2 (outbound, TDM) is operated with
tablished, through an Internet Service Provider (ISP) 1-dB back off.
and the public switched telephone network (PSTN). Part 2 Problems
A response from the ISP is sent back to the customer a. Determine the clear air overall CN required on
using the satellite’s spot beam. Since the packet from the inbound uplink and downlink required for one
the customer contains the VSAT station location, the VSAT transmission to meet the 99.7% availability cri-
hub station can send instructions to the satellite to terion, and the corresponding clear air CN in the hub
point the spot beam in the correct direction when station receiver with
transmitting packets to the customer. Note that with
(i) rain in the inbound uplink
a linear transponder (bent pipe) on the satellite, the
beam pointing instructions must be sent to the satel- (ii) rain in the inbound downlink. Remember to in-
lite at the same time as the data packet. clude the effect of increased sky noise.
The links between the ISP and the customer in this b. Using the results you obtained in Part 1, and Part
system are highly asymmetric. The customer can send 2 problem (a), determine the maximum data rate for
only short requests at a low data rate. The ISP can the VSAT request packets to meet the 99.7% avail-
dump data to the customer at a high data rate, mainly ability criterion with access to the transponder
because of the high EIRP of the satellite’s spot beam through the satellite’s regional beam, with 20 active
transmissions. This mode of operation suits applica- VSATs at any time.
tions where the customer is browsing the Internet for c. Determine the clear air overall CN in the VSAT
information, or is requesting large files or video station receiver for an outbound data rate of 1 Mbps
frames from the Internet. It works less well for send- using QPSK-TDM to meet the 99.9% availability
ing files from the customer to the Internet—as is done criterion, for
with e-mail, for example. In this problem you are
(iii) rain in the outbound uplink
asked to design a VSAT network based on your re-
sults from Part 1. (iv) rain in the outbound downlink. Remember to in-
Ka-band links are subject to high attenuation in clude the effect of increased sky noise.
rain. The outbound link is required to achieve a 99.9% d. Using the results you obtained in Part 1, determine
availability for a typical VSAT station for which slant the maximum data rate that can be supplied to each
path attenuation exceeds 7 dB at 21.7 GHz and 12 dB VSAT station with 20 active stations in the network
at 28.2 GHz, for 0.1% of an average year. The inbound at the same time, for the 99.9% availability criterion.
c06.qxd 09/08/02 14:19 Page 272
Note that for the small percentages of time used here, and portable receivers using the conditions in Part 1
you may assume that rain never occurs simultane- in clear air conditions. Don’t forget to allow an extra
ously in both the uplink and downlink. 1 dB of loss to account for antenna mispointing.
e. If your results from parts (b) and (d) above show that c. Assume that 10 active stations share each transpon-
either transponder 1 or 2 is not bandwidth limited, it is der. Determine the maximum data rates that customers
possible to optimize the system to transmit at higher bit can achieve on the inbound and the outbound links
rates. Redesign the VSAT and hub stations to increase with 99.7% availability of the inbound and outbound
the bit rates in either the inbound link, the outbound links.
link, or both links, within the limits that the VSAT d. Transponders 3 and 4 can be switched into base-
antenna diameter cannot exceed 1 m, and the transmit band processing mode. In this mode, the incoming
power cannot exceed 2 W. The hub station antenna QPSK signal is demodulated to baseband, the data
diameter cannot exceed 5 m and the transmit power bits are recovered and then remodulated onto a car-
cannot exceed 200 W. You might also consider whether rier for transmission as a new QPSK signal. This
the number of simultaneous users can be increased. The allows the transponder to transmit at its rated out-
satellite is leased and cannot be changed, except that put power at all times despite uplink attenuation.
the gain of the transponders can be adjusted to suit the The bit error rate for the link is then the sum of the
earth stations used in the network. BERs on the uplink and the downlink. Rework your
Part 3 Portable terminals solution to part (c) above using baseband proces-
The one advantage of radio systems over wired com- sors for both inbound and outbound links and de-
munications links is portability. This question asks termine the new data rates for the inbound and out-
you to design a portable Ka-band terminal which can bound links.
be used to connect to the Internet (provided the cus- e. Draw a block diagram of transponder 3 when used
tomer has a clear view of the southern sky). The crit- in its baseband processing mode. Your block diagram
ical element in a portable communications link is the should include all the filters, amplifiers, mixers,
antenna. A large antenna provides a high data rate, oscillators, modulators and demodulators, and all
but is cumbersome and must be pointed accurately at other important blocks. Label each filter and ampli-
Ka-band frequencies. A small antenna is easier to set fier with a center frequency and bandwidth, and in-
up, but cannot provide a high data rate. Let’s assume dicate the gain of each amplifier. Label all oscillators
that the dimensions of the antenna are limited to the with their frequencies.
dimensions of a typical laptop computer—0.25 m f. Comment on the performance of the fixed and
0.2 m—with an aperture efficiency of 25%, and that portable Ka-band Internet link system. If the
some method is provided that helps the customer transponders on the GEO satellite cost $1.5 M per
point the antenna beam toward the satellite so that year each to lease, and the service provider’s costs to
there is no more than 1-dB loss of gain due to an- support the customer base that shares these transpon-
tenna mispointing. ders are $ 0.5 M per year, what would you expect to
Because the portable terminals cannot achieve the have to charge the customer for access to the Inter-
same CN ratios as the fixed terminals, separate net when using the fixed terminal and the portable
transponders will be needed to service the portables. terminal? You can establish a charging structure made
For convenience, we will call these transponders 3 up of a monthly fee plus a per minute access charge.
(inbound) and 4 (outbound) and use the same fre- Assume that you can achieve a continuous level of
quencies as transponders 1 and 2. The ability of the activity of 20 fixed or 10 portable terminals for 12 h
system to operate during rain fades on the outbound per day.
link is relaxed with an availability of 99.7% required Each user can be assumed to connect to the In-
in each direction. ternet for 15 min once each day, but is active (in the
a. Calculate the gain and the beamwidth of the sense of data transfer over the satellite) for 1 min per
portable antenna at frequencies of 28.2 and 21.7 GHz. day. How do the data rates and the charges you pro-
b. Using your results from Part 1, find the inbound pose for the portable Internet access service compare
and outbound overall CN ratios in the hub station to typical charges for cable modem service?
c07.qxd 09/08/02 14:41 Page 273
CHAPTER 7
ERROR CONTROL FOR DIGITAL
SATELLITE LINKS
7.1 ERROR DETECTION AND CORRECTION
The transmission of information over a satellite communication system always results in

some degradation in the quality of the information. In analog links the degradation takes
the form of a decrease in signal-to-noise ratio. We saw in Chapter 5 that by using wide-
band FM we can trade bandwidth for power and achieve a good baseband signal-to-noise
ratio with a low carrier-to-noise ratio in the RF signal. In digital links we measure degra-
dation of the information content of a signal in terms of the bit error rate. By using phase
shift keying, usually coherent QPSK, we can again trade bandwidth for signal power and
achieve good bit error rates with low carrier-to-noise ratios.
A fundamental difference between analog and digital signals is that we can improve
the quality of a digital signal by the use of error correction techniques. No such technique
is available for analog signals since once the information is contaminated by noise, it is
extremely difficult to remove the noise, as we cannot in general distinguish between the
signal and the noise electronically. (There are techniques that attempt to distinguish be-
tween signal and noise in television pictures, by using the correlation properties of the
picture. They have been used successfully to enhance the quality of images of the moon
and other planets obtained by the Voyager and similar space probes. However, the time
taken to process the picture and the computer power needed make such techniques im-
practical for regular TV and voice transmissions.) In a digital system, we can add extra
redundant bits to our data stream, which can tell us when an error occurs in the data and
can also point to the particular bit or bits that have been corrupted. Systems that can only
detect errors use error detection. Systems that can detect and correct errors use forward
error correction (FEC). Systems which have only error detection must make a decision
about what action to take when an error is detected. The options are to ignore the error,
to flag the error, to send a block of information again, or to estimate the error and replace
the corrupted data. Which option is selected depends on the nature of the signal that is
transmitted. Collectively, these techniques are known as error control. Error control may
be implemented at the earth station as a permanent part of the satellite communication
link, or it may be applied by the end user. In advanced digital satellite communication
links the FEC may be switched in and out on demand, depending on the measured bit
error rate or CN ratio at a terminal.
Some confusion surrounds the term coding, since it is applied to several different
processes, not all of which are concerned with error detection and correction. In the pop-
ular sense, coding is used to describe the rearrangement of information to prevent unau-
thorized use. This process is known technically as encryption. It is widely used on both
analog and digital signals that are sent by cable and radio links. Digital signals are much
more amenable to encryption, which can be achieved by convolving the data bits with a
long bit sequence to destroy the intelligibility of the baseband data. To recover the
273
c07.qxd 09/08/02 14:41 Page 274
274 CHAPTER 7 ERROR CONTROL FOR DIGITAL SATELLITE LINKS
information, the bit sequence used in the encryption process must be known to the re-
cipient; this information is contained in the key to the code, which may be changed at
frequent intervals to maintain good security. We will not be concerned any further, in
this chapter, with encryption.
Coding is also a name applied to many processes that change data from one form
to another. For example, pulse code modulation (PCM) changes analog data into binary
words for transmission over a digital link. It is fundamental to the transmission of voice
and video signals by digital techniques, and uses a device commonly called a codec, short
for coder–decoder. The term coding is also applied to devices that scramble a digital data
stream to prevent the occurrence of long strings of 1s or 0s.
Throughout this chapter we shall use the term coding to refer to error detection or
error correction. This implies that additional (redundant) bits are added to the data stream
to form an error-detecting or error-correcting code. It is possible, in theory, to generate
codes that can detect or correct every error in a given data stream. In practice, there is a
trade-off between the number of redundant bits added to the information data bits and the
rate at which information is sent over the link. The efficiency of a coding scheme is a
measure of the number of redundant bits that must be added to detect or correct a given
number of errors. In some FEC systems the number of redundant bits is equal to the num-
ber of data bits, resulting in a halving of the data rate for a given channel transmission
rate. This is called half rate FEC. The loss of communication capacity is traded for a guar-
anteed low error rate. This technique is widely used in VSAT systems where the links to
and from the small antenna terminal have low CN ratios. By adding half rate FEC to the
data streams using convolutional coding, the reduction in BER at the baseband output of
the receiver is roughly equivalent to a 3-dB improvement in CN ratio. A 3-dB improve-
ment in CN ratio can be obtained by increasing the antenna diameter of the VSAT by
41%, but is an expensive and unwieldy option compared to inserting a half rate FEC in-
tegrated circuit into the terminal’s bit stream. Consequently, all satellite terminals which
tend to have low CN ratios (VSATs, satellite telephones, DBS-TV terminals) make use
of forward error correction to improve the bit error rate on the links to and from the small
terminal.
The operator of a digital satellite communication link has an option of providing
FEC as part of the link, or of providing only a basic transmission channel. At 64 GHz
and on many 1411 GHz links, the channel is provided without error correction or de-
tection. A minimum BER is guaranteed by the operator for a specified percentage of time,
based on the link design and projected performance. The user is then free to add error de-
tection or FEC to the data sent to and received from the link. If digital speech is sent, er-
ror detection or correction is rarely applied. With digital data, some measures must be
taken to guard against error, and the user will normally provide the necessary equipment.
Financial transactions and records are required to be transmitted with a BER of 1012.
Few communications links guarantee such low error rates and a customer sending finan-
cial data over a satellite link must use an automatic repeat request (ARQ) protocol to
ensure that the probability of an undetected bit error becomes extremely low.
Links operating at frequencies above 10 GHz are subject to increases in BER dur-
ing propagation disturbances. The link will be designed with a margin of a few deci-
bels so that the BER falls below an acceptable level, typically 106, for only a small
percentage of any month or year. The total time for which the margin is exceeded by
propagation effects will be less than 0.5% of any month in a well-designed system. Dur-
ing the remaining 99.5% of the month, the CN ratio of the received signal will be well
above threshold, and very low BER will result. There may, in fact, be no errors for long
periods of time and billions of bits can be transmitted with complete accuracy. Under
c07.qxd 09/08/02 14:41 Page 275
7.2 CHANNEL CAPACITY 275
these conditions, FEC and error-detection systems do nothing for the communication
system. However, unless we can detect a falling CN ratio directly, or an increase in
BER, coding may have to be applied all the time to be certain it is available when the
CN ratio approaches threshold. To that extent, coding is an insurance against the
possibility of bit errors; for most of the time it is unnecessary, but when it is needed, it
proves invaluable.
Common carriers, who supply communication links to users on a dial-up or leased
basis, do not generally apply FEC to their links, nor do they define the protocols to be
used. (A protocol defines an operating procedure in a link.) These are user-supplied serv-
ices and must be defined by the user for the data to be sent. In such cases, the error de-
tection and correction equipment will be located at the customers’ premises, whereas the
earth station may be a long distance away and accessed via terrestrial data links. The sit-
uation may be very different in a single-user network such as a direct broadcast television
system, where the coding and format of the transmitted data are specified by the company
that operates the uplink and satellites. A similar situation can arise in carefully controlled
systems such as a military communication system where the link operator specifies the
user’s earth station and operating parameters in detail.
7.2 CHANNEL CAPACITY
In any communication system operating with a noisy channel, there is an upper limit on
the information capacity of the channel. Shannon1 examined channel capacity in mathe-
matical terms, and his work led to significant developments in information theory and
coding.
For an additive white Gaussian noise channel, the capacity H is given by
H B log2 31 P 1N0 B2 4 bps (7.1)
where B is the channel bandwidth in hertz, P is the received power in watts, and N0 is the
single sided noise power spectral density in watts per hertz.
Equation (7.1) is commonly known as the Shannon–Hartley law. We can rewrite
Eq. (7.1) specifically for a digital communication link by putting H 1Tb where Tb is
the bit duration in seconds. The energy per bit is E joules, giving
E PTb PH (7.2)
Then substituting EbN0 PHN0 in Eq. (7.1) yields
log2 a1 b
H Eb H
(7.3)
B N0 B
The ratio HB is the spectral efficiency of the communication link, the ratio of the
bit rate to the bandwidth of the channel. Figure 7.1 shows the ratio log2(HB) plotted
against EbN0 in decibels for the case when H B and the link operates at a bit rate H
bps. Regardless of the bandwidth used, the EbN0 cannot go below 1.6 dB (ln 2) if we
are to operate at capacity. This is known as the Shannon bound. It sets a lower theoreti-
cal limit on the EbN0 we can use in any communication link, regardless of the modula-
tion or coding schemes. A link operating with H B is said to be power limited because
it does not use its bandwidth efficiently. Recently developed powerful FEC schemes such
as turbo codes2 allow links to operate down to CN ratios of 0 dB at acceptable bit error
rates. This is 1.6 dB above the Shannon limit, leaving only a little room for further
improvement.
c07.qxd 09/08/02 14:41 Page 276
log2 H
( )
B (NEb0 )
−6 −4 −2 0
0 (dB)
−0.5
−1.0
FIGURE 7.1 Relationship

−1.5 between HB and Eb N0 for
power limited case and low
−1.6 dB EbN0 ratio.
Figure 7.2 shows the case for a link with H B, operating at capacity. In this case,
we can increase the capacity for a given bandwidth without limit, but only by providing
large EbN0 ratios, implying high transmitter power. When H B, the link is said to be
bandwidth limited, implying that we could increase capacity by using the available trans-
mitter power in a wider bandwidth. Practical links using PSK do not achieve capacities
close to the Shannon theoretical capacity H. Shannon’s theory assumes essentially zero
bit errors; to achieve a bit error rate of 106, a QPSK link requires a theoretical EbN0 of
60
Eb
N 0 40
(dB)
20
0 FIGURE 7.2 Relationship

0 2 4 6 between HB and EbN0 for
bandwidth limited case high
log2 H( )
B EbN0 ratio.
c07.qxd 09/08/02 14:41 Page 277
7.3 ERROR CONTROL CODING 277
10.6 dB with a spectral efficiency of 2 bits/Hz. Equation (7.3) predicts a bound of EbN0
1.77 dB for this case. What coding, in particular FEC, can do for us is to improve the link
performance under conditions of low CN ratio, such as during periods of rain attenuation,
so that the BER of the link does not rise excessively. This takes us closer to the Shannon
capacity in the region of low CN ratios while not increasing excessively the bandwidth
required for transmission. A theoretical spectral efficiency of 8 bits/Hz can be achieved
with EbN0 15.0 dB4. A satellite link could use 256-QAM modulation to achieve this
spectral efficiency, but would need CN of 30.3 dB (EbN0 21.3 dB) to give a bit er-
ror rate of 106. This is well below the theoretical efficiency for the Shannon bound, even
when FEC is applied to the data stream.
7.3 ERROR CONTROL CODING
Error detection coding is a technique for adding redundant bits to a data stream in such
a way that one or more errors in the data stream can be detected. One redundant bit is
added for every N data bits; this allows a single error within that block of N bits to be de-
tected, whatever the number of bits in the block. A simple example of an error detecting
code system that has been in use for many years is the single bit parity applied to the 7-bit
ASCII code3. The ASCII code is widely used for transmission of data over telephone lines
and radio links.
The 7-bit ASCII code consists of 128 characters that have internationally agreed in-
terpretation and represent the alphabet in uppercase and lowercase letters, numerals 0 to
9, and a set of useful commands, symbols, and punctuation marks. An eighth bit, the par-
ity bit, is used for detection of error in the 7 data bits of the character. For example, in a
system using even parity, the parity bit is 0 when the sum of the data bits is even, and 1
when the sum is odd. Thus the sum of the data bits plus the parity bit is always made
even, or 0 in modulo-2 arithmetic. Figure 7.3 shows an example of even and odd parity
coding. In odd parity, the sum of the data bits plus the parity bit is always odd, or 1 in
modulo-2. Errors in the 7 data bits, or the parity bit, are detected at the receiving end of
the link by checking the 8 received bits of each character for conformity with the parity
Data bits Parity bit Sum (Modulo-2)
Even parity 0101101 0 0

Odd parity 0101101 1 1
(a)
Received codeword Sum of bits Error detected?
One error 01010010 1 Yes

Two errors 01010110 0 No
Three errors 11010110 1 Yes
(b)
FIGURE 7.3 (a) Example of even and odd parity for a 7-bit ASCII word.
(b) Example of error detection in a 7-bit ASCII word with even parity. Error
bits are underlined.
c07.qxd 09/08/02 14:41 Page 278
rule. In modulo-2 arithmetic, 0 0 0, 0 1, 1 0 1, and 1 1 0. This is the

exclusive OR function of digital logic. Similarly, 0 0 0, 0 1 1 and 1 0 0,
and 1 1 1. All the codes that we will be considering are binary, and modulo-2 arithmetic
will be used throughout this chapter.
Suppose we have a system using even parity, which transmits the character A in
ASCII code, as illustrated in Figure 7.3. At the receiving end of the link we check the 8
bits by modulo-2 addition. If the sum is 0, we conclude that the character is correct. If
the sum is 1, we detect an error. Should 2 bits of the character be corrupted, the modulo-
2 sum is 0, so we cannot detect this condition. We can easily calculate the improvement
in error rate (assuming that we discard corrupted characters) that results from adding a
parity bit to a 7-bit word.
For example, let the probability of a single-bit error occurring in the link be p, and
let us suppose that p is not greater than 101. The probability Pe(k) of k bits being in error
in a block of n bits is given by the binomial probability function
Pe 1k2 a b pk 11 p2 nk
n
(7.4)
k
where p is the probability of a single-bit error occurring, and
a b
n n!
(7.5)
k k!1n k2!
For example, with single parity and the 7-bit ASCII characters, we have one parity
bit, which allows us to detect one error, and 7 data bits. Two errors cannot be detected,
although three can. The most likely error that goes undetected is a 2-bit error. The proba-
bility that there are four or more errors in the 8-bit word is much lower than the probabil-
ity of 2-bit errors, provided the BER is no higher than 102, so when a single parity bit is
used, the probability of an undetected error occurring in an ASCII word is approximately
Pwc where
Pwc Pe 122 a b p2c 11 pc 2 6

8
(7.6)
2
where pc, is the single bit error probability for the 8-bit word (i.e., the BER on the link).
When pc is small, 11 pc 2 6 1 and
Pwc a b p2c 28 p2c

8
(7.7)
2
If we had not used parity, the probability Pwu of a single error in the 7-bit word with
bit error probability pc is
Pwu a b pu 11 pu 2 7 7 pu
7
(7.8)
1
Thus the improvement in rate of undetected errors for the ASCII words is approximately
4p, provided pc pu.
EXAMPLE 7.3.1
A data link transmits 7-bit ASCII words at a bit rate of 1 Mbps, with a single parity bit. The prob-
ability of a bit error on the link is p 103. Find the probability of an undetected error when
uncoded data is transmitted and when a single parity bit is added to each 7-bit word. What is the
c07.qxd 09/08/02 14:41 Page 279
7.3 ERROR CONTROL CODING 279
probability of an undetected bit error when the BER on the link is 106? How many undetected
character errors would be present if a 500-page textbook were transmitted this link, using single
parity?
Using the result in Eq. (7.8), the probability of error for uncoded 7-bit words is 7 103. If
we add a single parity bit, the probability of an undetected error is given by the result of Eq. (7.7)
as Pwc 28p2c 28 106. With a BER of 106, the undetected character error rate is 7 106
for no parity and 28 1012 with single bit parity. Based on 2000 characters per page (including
spaces) and 500 pages in the book, there are one million characters in the text. Without error de-
tection, there would be seven typos in the text caused by transmission over the link. With single bit
parity error detection, the text could be transmitted 35,700 times before a single undetected char-
acter error occurred. At a bit rate of 106 bps, the entire book can be transmitted in 7 s without par-
ity and 8 s with single bit parity. The single undetected error—which will appear in the text as a
typo—would occur after 79 h of transmission when single bit parity is used.
The above example illustrates how powerful even a single parity bit can be in the
detection of bit errors in a link with low BER. Some caution is needed in making the as-
sumption that the bit error rate is the same for uncoded and coded transmissions. When
we added the single parity bit to a 7-bit ASCII character, the transmission bit rate went
up from 7 bits per character to 8 bits per character. The increase in bit rate will result in
an increase in BER because the channel bandwidth must be increased and this increases
the noise power in the receiver and reduces the CN ratio, so not all the expected decrease
in character error rate will be achieved in practice.
Example 7.3.1 above for parity error detection is one case of block error detection.
We have transmitted our data as blocks, in this case 8-bit blocks consisting of 7 data bits
and 1 redundant parity bit. In general, we can transmit n bits in a block, made up of k
message bits plus r parity bits. There are two ways in which the blocks of data can be
formed. In a packet transmission system (see Chapter 6), the block is usually long, typi-
cally 64 to 2048 bits, and checksum or cyclic redundancy check (CRC) bits are added to
the block that generate an error condition at the receiver if there are errors in the block.
Alternatively, the blocks may be members of a set of n-bit codewords. Coding schemes
in which the message bits appear at the beginning of the codeword, followed by the parity
bits, are called systematic block codes.
Linear and Cyclic Block Codes

Linear block codes are codes in which there are 2n possible codewords containing k mes-
sage bits and (n k) redundant check bits. In a systematic linear block code the first k
bits of the codeword are the message and the remaining (n k) bits are the parity bits,
or possibly the other way round with the parity bits first. A codeword with n bits of which
k bits are message data is written as (n, k). Early work on single error correction linear
block codes was done by Hamming at Bell Labs in the 1940s3,5. A subset of linear block
codes called binary cyclic codes has been developed for which implementation of error-
correction logic is relatively easy. The codes can be generated using shift registers, and
error detection and correction can also be achieved with shift registers and some addi-
tional logic gates. A large number of binary cyclic codes have been found, many of which
have been named after the people who first proposed them. The best known are the
Bose–Chaudhuri–Hocquenghem codes (BCH codes), which were independently proposed
by three workers at about the same time in 1959–19606,7. Other examples of block codes
in widespread use are the Reed–Solomon codes, used on compact discs and DBS-TV sig-
nals, and the Golay codes. Most of the block codes were developed in the late 1950s for
c07.qxd 09/08/02 14:41 Page 280
error detection and correction with early computer memories, which were prone to cause
data errors in the recording and recovery of data. Subsequently, the codes were applied
to digital transmission of data.
The general form of a linear block codeword C is
C DG (7.9)
where D is the k-bit message and G is a generator matrix that creates the parity check
bits from the data bits. There are 2k valid codewords within a set of 2n codewords
generated using Eq. (7.9). For example, if we create a (7, 4) linear block code that has
four message bits and three parity bits, there are a total of 27 128 codewords in the
set, but only 24 16 codewords are valid. If we receive an invalid codeword we know
that an error has occurred on transmission, although we do not necessarily know how
many bits have been corrupted, or which of the message bits are wrong. One way to
correct errors in received codewords is to compare the received codeword with the
valid set of codewords and select the correct codeword that is closest to the received
codeword.
Some linear block codes are better for error detection or correction than others.
There are some basic theorems that define the capabilities of linear block codes in terms
of the weight, distance, and minimum distance. The weight, w, of a codeword C is the
number of nonzero components of C. The distance, d, between two codewords C1 and C2
is the number of components by which they differ. The minimum distance of a block code
is the smallest distance between any pairs of codewords in the entire code. With a single
error detection code, the number of errors that can be detected in a code with minimum
distance dmin is (dmin 1). The number of errors that can be corrected is (dmin 1)2,
rounded to the next lowest integer if the number is fractional. Thus it is always easier to
detect errors than to correct them with linear block codes, a feature which is exploited in
the Reed–Solomon codes used in CDs and the digital video transmissions of DBS-TV
systems.
If we are very clever about the generator matrix, G, that we use to generate the code,
we may be able to determine which bit is in error when we receive a corrupted codeword.
There are no rules for forming generator matrices, so “good” codes are found by inspira-
tion, and a lot of trial and error, using weight, distance, and minimum distance for guid-
ance. A good error detection code is one that detects as many bit errors as possible in a
codeword of given length, preferably when the bit errors are sequential (a burst error),
and a good error correction code corrects as many errors as possible, even when they are
burst errors. Table 7.1 shows some examples of burst error correcting block codes.
Golay Codes
The Golay code and the single error-correcting Hamming codes are examples of perf-
ect codes, in which all possible patterns of a given number of errors are corrected8. The
Golay code is a (23, 12) cyclic code with a minimum distance of 7 that corrects all pat-
terns of 3 or fewer bit errors. A closely related form of the Golay code has one overall
parity check bit added to form a (24, 12) code with a minimum distance of 8. The (24,
12) code will detect all patterns of 7-bit errors and correct all patterns of 3-bit errors.
It also has the advantage of a coding rate of one-half; rate one-half coding is easier to
implement than other rates because the 2:1 ratio between message bits and code bits
simplifies clock synchronization between the input data stream and the output coded
stream.
c07.qxd 09/08/02 14:41 Page 281
7.4 PERFORMANCE OF BLOCK ERROR CORRECTION CODES 281
TABLE 7.1 Examples of Burst Error Correction

Block Codes
Burst error
Code No. of parity correcting
(n, k) check bits ability q Code rate
7, 3 4 2 3/7
15, 9 6 3 9/15
511, 499 12 4 499/511
1023, 1010 13 4 1010/1023
131, 119 12 5 119/131
290, 277 13 5 277/290
34, 22 12 6 22/34
169, 155 14 6 155/169
103, 88 15 7 88/103
96, 79 17 8 38/56
59, 39 20 10 39/59
Source: Reprinted from K. S. Shamnugam, Digital and Ana-

log Communication Systems, John Wiley & Sons, New York,
1979, Table 9.5, p. 475, used with permission.
7.4 PERFORMANCE OF BLOCK ERROR

CORRECTION CODES
Calculation of the improvement of bit error rate with block encoding requires a compar-
ison of the uncoded error rate to that obtained after correction of blocks of encoded data.
For the perfect codes given in Section 7.3, the improvement can be determined exactly;
when other codes are used, only an upper and lower bound on the error rate after coding
can be established.
Figure 7.4 shows a comparison of the performance of a number of cyclic codes
when implemented in a coherent PSK link9. The curves are for an ideal link with no mo-
dem implementation margin and show error probability as a function of EbN0 at the de-
modulator input. The (127, 64) code corrects 10 errors, the (1023, 688) code corrects 36
errors. The (127, 113) BCH code is a double-error-correcting code that has been speci-
fied for use in 120-Mbps transmission using the Intelsat TDMA system9. Also shown are
BER curves for half rate convolutional coding with constraint length 7 and half rate turbo
coding with 18 iterations.
Note that by the use of FEC, we can reduce the EbN0 needed to achieve a 106 er-
ror rate from a theoretical 10.6 dB without coding to 7 dB with BCH (127, 64) coding.
The reduction that can be made in EbN0 by adding FEC to the link while maintaining
the same BER is called coding gain. Coding gain varies with BER, tending to be great-
est at low BER, and sometimes becoming negative when EbN0 becomes small. In Fig-
ure 7.4, for example, coding gain for the (127, 64) BCH code is negative when EbN0 is
less than 7 dB. When the BER exceeds 2 103, errors are no longer corrected and ad-
ditional errors are inserted into the bit stream. Convolutional codes perform better at low
EbN0 ratios, and are therefore widely used on satellite links.
Coding gain must be used with care in satellite links, because to obtain coding gain
we must increase the bit rate on the link, which results in a lower EbN0 because the bandwidth
c07.qxd 09/08/02 14:41 Page 282
BCH
(1023, 688)
10−2
Probability of bit error (BER)
10−4
Uncoded
Rate 1/2
BPSK
convolutional
10−6 code, soft
decision, K = 7
BCH
Rate 1/2 (127, 64)
turbo code,
18 iterations
10−8
0 5 10 15
Eb /No in dB
FIGURE 7.4 Probability of bit error in a BPSK link with coherent demodulation. Note that
these are theoretical curves for EbN0. To apply these results to a practical communication
system, an implementation margin must be added to the EbN0 values to obtain CN values.
For QPSK links, add 3 dB to the BPSK CN ratios.
of the receiver has to be increased, resulting in lower CN ratios and thus a higher BER.
This is particularly true when we apply half rate FEC to a link. If we want to achieve
the full coding gain (about 6 dB for many convolutional codes) we can send only half as
many bits. If we double the bandwidth of the link in order to send the same data rate, the
CN ratio will fall by 3 dB and the effective increase in CN will be only 3 dB instead
of 6 dB.
7.5 CONVOLUTIONAL CODES
Convolutional codes provide better error correction performance than block codes on
satellite links with low CN ratios in the receiver, and have simpler decoding structures.
Figure 7.4 shows that the BCH (1023, 688) block code and the constraint length 7 con-
volutional code have near equal correction capability when the BER is below 106. How-
ever, as the BER increases, the convolutional code performs better. Decoding cyclic codes
with large block lengths is difficult, and the choice of forward error correction code is
often based on the availability of coding and decoding devices as well as performance.
Convolutional codes are generated by a tapped shift register and two or more modulo-
2 adders wired in a feedback network. The name is given because the output is the con-
volution of the incoming bit stream and the bit sequence that represents the impulse
response of the shift register and its feedback network10. Figure 7.5 illustrates an exam-
ple of a simple convolutional encoder using a three stage shift register10–13. As each
c07.qxd 09/08/02 14:41 Page 283
7.5 CONVOLUTIONAL CODES 283
Switch
Input Output
3 stage
shift register
01010101 0011100111000110
Bit clock
+
FIGURE 7.5 A simple convolutional encoder. The switch changes state at each input
bit clock transition.
incoming information bit propagates through the shift register, it influences several out-
going bits, spreading the information content of each data bit among several adjacent bits.
The output data stream at the right of Figure 7.5 is different from the input stream at the
left; the data bits are not present in the coder output as in a block code. An error in any
one output bit can be overcome at the receiver without any information being lost. The
process is somewhat like forming an image from a hologram, where information is dis-
tributed more or less uniformly over a two-dimensional field. The image can be recon-
structed from only a portion of the field, although resolution is lost if a significant part of
the hologram is discarded.
The state of a convolution encoder is defined by the shift register contents that will
remain after the next input bits are clocked in. If the shift register is K bits long and in-
put bits enter in groups of k, then the encoder has 2Kk states. Putting in a group of k in-
put bits causes the encoder to change states; a change of state is called a state transition.
From a given state, a convolutional encoder can go to only 2k other states (although one
of these k options may be to remain in the starting state). Each state transition is associ-
ated with a unique sequence of input bits and is accompanied by a unique sequence of
output bits. The quantity K, which measures the length of the shift register, is called the
span or the constraint length of the encoder. If v output bits are transmitted for every k
input bits, then the encoding rate is kv. A ratio kv 12 is widely used, giving half rate
coding.
A decoder for convolutional codes keeps track of the encoder’s state transitions and
reconstructs the input bit stream. Transmission errors are detected because they corre-
spond to a sequence of transitions that could not have been transmitted. When an error is
detected, the decoder begins to construct and keep track of all the possible tracks (se-
quences of state transitions) that the encoder might be transmitting. At some point, which
depends on its speed and memory, the decoder selects the most probable track and puts
out the input bit sequence corresponding to that track. The other tracks that it had been
carrying are discarded. One of the best algorithms used for decoding is named for A. J.
Viterbi (see reference 14), and for this reason convolutional codes are sometimes called
Viterbi codes (Figure 7.6).
c07.qxd 09/08/02 14:41 Page 284
10−2
K=3
K=4
10−3
K=5
Bit error rate
K=6
K=7
10−4
K=8
10−5
4 5 6 7
Eb /No in dB
FIGURE 7.6 Performance of rate one-half convolutional codes with constraint lengths from
3 to 8 and 32-bit paths. (Reprinted with permission from Jerrold A. Heller and Irwin Mark
Jacobs, “Viterbi Decoding for Satellite and Space Communications,” IEEE Transactions on
Communications, COM-19, 835–848 (October 1981). Copyright © 1981 IEEE17).
Convolutional or linear block coding and digital modulation can be combined us-
ing trellis coding. Trellis coding employs a high level modulation such as 8-PSK or 16-PSK
and allows only certain sequences of modulation states to be transmitted. The receiver can
detect errors in the received data strings if an invalid sequence of modulation states is re-
ceived3,15,16. The advantage of trellis coding is that coding gain can be achieved with a
smaller increase in bit rate than with conventional convolution encoding. However, trel-
lis coding does not seem to have been adopted very widely. Recently, turbo codes2 have
been developed that allow links to operate down to CN ratios of 2 dB at acceptable bit
error rates. The decoding algorithms for turbo codes require powerful DSP integrated
circuits, but these are now becoming available.
7.6 IMPLEMENTATION OF ERROR DETECTION

ON SATELLITE LINKS
Error detection is invariably a user-defined service, forming part of the operating protocol
of a communication system in which the earth–satellite–earth segment may only be a part.
It allows the user to send and receive data with a greatly reduced probability of error, and
c07.qxd 09/08/02 14:41 Page 285
7.6 IMPLEMENTATION OF ERROR DETECTION ON SATELLITE LINKS 285
a very high probability that uncorrected errors are identified and located within a block of
data, so that the existence of an error is known even if the exact bit or word in error can-
not be determined. The penalty for the user is a reduced transmission rate, just as in FEC.
However, the transmission rate is reduced only when errors occur, so there is no bit rate
penalty under normal conditions when no errors occur, unlike FEC. Implementation of
error correction by use of error detection and retransmission requires the use of protocols.
A protocol is an agreed-upon set of actions that define how each end of the data link
proceeds so that data are transmitted in an accurate and ordered fashion through the link.
Error detection is readily accomplished using the coding techniques described in the
previous sections. In many communications systems it is not sufficient simply to detect
an error; it must be corrected. However, some systems such as speech and picture trans-
mission may simply count errors to determine whether the link is above an operating
threshold. Excessive error counts may cause the link to automatically shut down. When
a succession of pictures is transmitted, for example, weather maps, the previous value of
a pixel can be used when a new value is in error. When an error-detection code or CRC
is used and the error must be corrected, a retransmission of the block of data containing
the error must be made so that the correct data are acquired at the receiving terminal. If
an error is detected in the block, a not acknowledge (NAK) signal is sent back to the trans-
mit end of the link, which triggers a retransmission of the erroneous block of data. This
is called an automatic repeat request (ARQ) system.
In terrestrial data communication systems, it is common practice for the receiving
terminal to send an acknowledge (ACK) signal to the transmitting end whenever it re-
ceives an error-free block of data. Such protocols works well on terrestrial data links with
relatively low data rates and short time delays, but the long transmission delays in satel-
lite communication systems make it highly undesirable to send ACK signals for every er-
ror free block that is received, so Internet access via satellites must use a different access
protocol. Often a satellite terminal receives data using a terrestrial protocol and generates
the acknowledgements, then transmits the data over the satellite link using a different pro-
tocol. This is called spoofing.
There are three basic techniques for retransmission requests, depending on the type
of link used. In a one-way, simplex link, the ACK or NAK signal must travel on the same
path as the data, so the transmitter must stop after each block and wait for the receiver to
send back an NAK or ACK before it retransmits the last data block or sends the next one.
With a one-way delay of 240 ms on a GEO satellite link, the data rate of such a system
will be very low and is suitable only for links in which data are generated slowly, as when
someone is typing on a terminal keyboard. Satellite links usually establish two-way com-
munication (duplex channel) by the use of SCPC-FDM, or TDM, as discussed in Chap-
ter 6. The ACK and NAK signals can be sent on the return channel while data are sent
on the go channel, However, if the data rate is high, the acknowledgment will arrive long
after the block to which it relates was transmitted.
In a stop-and-wait system, the transmitting end sends a block of data and waits
for the acknowledgment to arrive on the return channel. The delay is the same as in the
simplex case, but implementation is simpler. Figure 7.7 shows an example of a stop-
and-wait sequence. In a continuous transmission system using the go-back-N technique,
data are sent in blocks continuously and held in a buffer at the receive end of the link.
Each data block is checked for errors as it arrives, and the appropriate ACK or NAK
is sent back to the transmitting end, with the block number appended. When a NAK(N)
is received, the transmit end goes back to block N and retransmits all subsequent blocks,
as illustrated in Figure 7.8. This requires the transmitter on a satellite link to hold at least
480 ms of data, to allow time for the data to reach the receive end and be checked for
c07.qxd 09/08/02 14:41 Page 286
Retransmitted
Transmitted blocks block
1 2 2 3
Time
K
K
K
NA
AC
AC
ACK
1 2 2 3
Time
Received blocks
Transmission
delay, t d Errors detected
in this block
FIGURE 7.7 Stop-and-wait ARQ system.
errors and for the acknowledgment to be sent back to the transmit end. Since there is a
delay in transmission only when a NAK is received, the throughput on this system is much
greater than with the stop-and-wait method. Throughput is the ratio of the number of bits
sent in a given time to the number that could theoretically be sent over an ideal link.
If sufficient buffering is provided at both ends of the link, only the corrupted block
need be retransmitted. This system is called selective repeat ARQ. In Figure 7.8, block 2
is corrupted, and blocks 4, 5, 6, and 7 are transmitted before the NAK message is re-
ceived. At this point, we could transmit block 3 only if blocks 4, 5, 6, and 7 are stored at
the receive end. On receiving a correct version of block 3, the receive buffer substitutes
it for the corrupted version and releases the data for retransmission. In systems handling
data rates of megabits per second, the buffer requirements for continuous transmission
systems become quite large because of the delay on the link via a geostationary satellite.
The Internet protocol TCP/IP (transmission control protocol/internet protocol) uses se-
lective repeat of blocks that contain errors. The relatively short delays on terrestrial paths
used for Internet traffic allow the use of selective repeat ARQ. The TCP/IP protocol
cannot be used over a geostationary link without modification because it was designed for
terrestrial links with short delays. The protocol times out before repeat transmissions and
replies are received when a GEO satellite link is used.
Transmitted blocks Retransmitted blocks
1 2 3 4 5 2 3 4
Time
ACK NAK NAK NAK NAK
ACK
0 1 2 3 4 5 2
Time
td Received blocks
Transmission
delay Errors detected
in this block
FIGURE 7.8 Example of a go-back-N-blocks ARQ system. N is three blocks in this
example.
c07.qxd 09/08/02 14:41 Page 287
7.6 IMPLEMENTATION OF ERROR DETECTION ON SATELLITE LINKS 287
Reference 19 contains a good survey of error-detection techniques for use in satel-

lite communication systems and of the various ARQ systems that can be implemented.
Some hybrid ARQ systems are described that combine FEC with retransmission
of blocks when uncorrected errors are detected. This combines the error-correction
properties of the FEC code for a limited number of errors with the error-detection
properties of the same code when too many errors are present for all of them to be
corrected20.
EXAMPLE 7.6.1
Calculate the frequency of retransmission, throughput, and buffer requirements of a satellite link
capable of carrying data at rates of (a) 24 kbps and (b) 1 Mbps when a block length of 127 bits is
used and the one-way path delay is 240 ms, for a bit error rate of 104 and a double error detect-
ing code (127, 120), using the following ARQ schemes:
1. Stop and wait.
2. Continuous transmission with transmit buffer only (go-back-N ).
3. Continuous transmission with buffers at both ends of the link (selective repeat).
Comment on the three ARQ schemes and their suitability for a satellite communication link.
For a 127-bit code block, the probability of one or two errors is given by Eq. (7.4)
Pe 1k2 a b pk 11 p2 nk
n
k
where k 2, n 127, and p is the probability of a single bit error, which is 104 in this example.
The probability of one or two errors being detected in a block of 127 bits is
P 1one or two errors2 127 0.9875 104 127 1262 0.9876 108
0.01262
Thus, on average, 1 in every 79 received blocks has a detectable error.
1. Stop-and-wait.
We send 127 bits and wait for an acknowledgement, which takes 0.485 s at 24 kbps and
0.4801 s at 1 Mbps. We therefore send only about two blocks each second at either data rate,
since the time is dominated by waiting for an acknowledgement. After 79 blocks have been
sent, on average, we detect an error; that is after about 39.5 s. The average data rate on the
link is approximately 254 bps for both transmission bit rates.
2. Go-back-N.
a. The time required to send 79 blocks of 127 bits at 24 kbps is 0.418 s.
The 79th block has a detected error: while the NAK signal goes back to the send end to
request a retransmission of block 79, a further 91 blocks arrive. These are discarded when
the retransmission arrives 0.48 s later, starting at block 79. This slows the throughput by about
54%, to about 11.2 kbps. The buffer at the transmit end must hold 0.48 s worth of bits, about
11,600 bits.
b. At a bit rate of 1 Mbps, the time to send 79 blocks is 0.01 s. We then wait 0.48 s for the
retransmission. The average data rate is then 20.46 kbps. The transmit buffer must hold
490 kbits of data.
3. Selective repeat system.
a. The only time lost in a selective repeat system is in the retransmission of blocks which
have errors. On average, one block in every 79 has to be retransmitted, so the rate effi-
ciency of the system is 7879, or 98.73%. At 24 kbps, the average data rate is 23.70 kbps,
and the buffer needed is 11,600 bits.
b. At 1 Mbps, the average data rate is 987.3 kbps, and the buffer must hold 490 kbit.
c07.qxd 09/08/02 14:41 Page 288
7.7 CONCATENATED CODING

AND INTERLEAVING
Sophisticated error correction and detection systems are used on some satellite links to
overcome burst errors and the effects of low CN ratios. Burst errors can occur when
the signal is temporarily blocked, or the CN ratio has become unusually low and the
BER is very high. Convolutional codes are limited in the number of sequential bit errors
that can be corrected, so long strings of incorrect bits will not be corrected by the FEC
decoder in the receiver. Burst errors can be overcome by using interleaving. Interleav-
ing is usually applied ahead of error control coding where a single coding operation is
used. When two separate error control coding operations are placed in series, the process
is called concatenation. Two coding operations can be applied in sequence to a data
stream to decrease the probability of undetected errors occurring because of low CN
ratio on a satellite link. Concatenated coding can achieve coding gains up to 9 dB3. With
concatenated coding, the interleaver may be placed between the two stages of error con-
trol coding. This is the strategy used in audio compact disc recordings and also in direct
broadcast satellite TV signals using the Digital Video Broadcast Standard for satellite
systems (DVB-S)18.
The purpose of interleaving is to spread out the errors that occurred in a burst and
thus to make it easier for an error correction system to recover the original data. Fig-
ure 7.9 shows a simple interleaver with 5 rows and 5 columns. Bits are read in to the rows
and read out by columns, which spreads out the bits in the resulting bit stream. Inter-
leaving can be illustrated more easily using letters of the alphabet rather than binary data.
Consider the following message: The cat sat on the table. Let us suppose that we trans-
mit the message with single bit error detection coding, and that two and three bit burst
errors occurs so that this message, without interleaving, is received as: The *** sat ** the
*able. where the * indicates an error. Because the English language is highly redundant,
and only certain letter combination make valid English words, we can guess at the mes-
sage based on what we received, although we might guess dog or cup instead of cat, or
cable instead of table. Now consider what happens when we use the 5 5 interleaver
shown in Figure 7.9 at the transmitter.
The message is read into the interleaver by row, as shown. The message is read out
by column to give: Tattaht hbe oel sn eca t. When this message is sent over the satellite
T h e c T * e c
a t s a a * s a
t o n t * n
t h e t t h * t
a b l e . * b l * .
(a ) (b )
FIGURE 7.9 (a) A 5 5 interleaver at a transmitter. Letters are used for il-
lustration purposes. (b) A 5 5 de-interleaver at the receiver. Letters are
used for illustration purposes.
c07.qxd 09/08/02 14:41 Page 289
7.7 CONCATENATED CODING AND INTERLEAVING 289
link, it is received with bit errors in the same positions as previously. We receive the fol-
lowing message: Tatt*** hbe **l sn *ca t. The message is read into the de-interleaver at
the receiver by column, and read out by row. The message at the receiver is then: T*e ca*
sat *n th* t*bl*. We could correct the message with a single character error correction
FEC code, or we could make an improved guess at the true message based on the received
message which now has no burst errors. Since there are no burst errors here, a single (char-
acter) error correction code would recover the original message correctly. The high level
of redundancy in the English language ensures that the message: T*e ca* sat *n th* t*bl*.
can be read with little error, although we cannot be certain whether ca* is cat, cam, can,
cap, or car. However, the rest of the message is unambiguous after de-interleaving. There
is no such redundancy in a stream of bits. Ones and zeroes are equally likely; indeed the
bit stream is usually scrambled to make certain that this is the case, so we cannot guess
at the correct bit when an error occurs.
Interleavers used in digital communication systems work on the same principle as
the simple example shown in Figure 7.9 (but using binary data) to break up burst errors
and make possible correction of the received data. A key element in the digital transmission
of analog data is that analog interpolation can be used to reconstruct an analog waveform
containing single word errors. In the example above using letters, interpolation between
the correctly received characters is much easier when the errors occur singly. We can
make a much better guess at the message received after interleaving than when there is
no interleaving. The same principle holds true for binary transmission of analog data.
Interpolation of an analog waveform is simple when only one point (a digital word) is
missing, as illustrated in Figure 7.10.
This approach is used very successfully on audio compact discs and in digital video
broadcast transmissions. Compact discs store audio data bits as changes in the reflectiv-
ity of the disc. The bits are about a micron long, so a scratch or a speck of dust on the
surface of the CD will cause a lengthy burst error. A long interleaver is used to spread out
the burst error and this makes reconstruction of the analog waveform possible for burst
errors as long as several thousand bits.
The interleavers used in satellite transmission of digital video signals are generally
shorter than those used on a CD, typically 12 7. Figure 7.11 shows the coding and de-
coding structure of a typical DBS-TV signal. In this system, the digital signal is first coded
using a 188204 rate Reed–Solomon code, then interleaved with a 12 7 interleaver, and
finally encoded with a rate three-quarters convolutional code with constraint length 7. At
the receiver, the convolutional code is used to correct some errors, which improves the
BER, but when the CN ratio in the receiver is low not all the errors will be corrected.
The signal is de-interleaved, which spreads out any uncorrected errors so that the proba-
bility of a burst error longer than a couple of bits becomes small. The Reed–Solomon
Vr (t )
Errors Straight line
interpolation
X X X X FIGURE 7.10 Illustration of

interpolation to fill in missing
t
X = error detected data in an analog waveform.
c07.qxd 09/08/02 14:41 Page 290
Input
Reed–Solomon Interleaver Convolutional
encoder encoder
Link
Convolutional De-interleaver Reed–Solomon

decoder decoder
Output
FIGURE 7.11 Coding and decoding structure for data in a DBS-TV system.
code is then used to detect the location of the remaining errors and to flag video words
which have errors. The video decoder looks at the digital words on each side of a word
that has been flagged with an error and replaces the flagged word with a new word that
is found by interpolation between the two neighboring words, as illustrated in Figure 7.10.
The resulting error in the analog video waveform is small. Thus the Reed–Solomon cod-
ing, which has a high code rate and uses only error detection properties allows errors in
the video signal to be corrected. Overall, the combined coding of the DVB signal increases
the bit rate by 45% when rate three-quarter convolutional encoding and 188204
Reed–Solomon coding is employed. The coding gain is typically 5.5 dB with BER 106
at the output of the convolutional decoder, and very few errors are left in the analog wave-
form after Reed–Solomon decoding and analog interpolation.
One of the features of the DBS-TV systems is that the software in the customers’ re-
ceivers can be modified via the satellite. Packets that are tagged as software can be sent to
all receivers, and changes made to the signal processing. This allows the system to change
the coding and decoding methods to implement improved error control strategies. One
DBS-TV system in the United States changed from rate two-thirds to rate three-quarters
convolutional encoding, freeing up additional bits for program material and other uses.
7.8 TURBO CODES
Turbo codes are the most powerful available codes for forward error correction (FEC).
The turbo code was first proposed by C. Berrou and R. Pyndiah from the Ecole Nationale
Superieure des Telecommunications de Bretagne (ENST), in France21,23. Turbo codes have
been demonstrated that can achieve a BER of 106 with a received signal at EbN0 of 0.7
dB, an improvement in coding gain of 1 dB over the most powerful concatenated and
interleaved convolutional codes22,24.
The basic form of turbo code generator uses two component codes, separated by an
interleaver. Message bits are read into an interleaver by row and then simultaneously read
out by rows and by columns into two separate encoders that use either block coding or
convolutional coding. One encoder is driven by the row message bits and the other by the
column message bits from the interleaver, so that an entirely different bit sequence is ap-
plied to each encoder, but both encoders are sending the same message bits. Since the row
output of the interleaver is the original data stream, one encoder has an input which is the
original message bit sequence and the other encoder input is an interleaved version of the
message bits. The outputs of the two encoders are combined by multiplying the 2-bit se-
quences (modulo two). Alternatively, the two outputs can be added and sent sequentially.
Turbo codes based on convolutional codes are usually known as CTC (convolutional turbo
codes) and those based on block codes as BTC (block turbo codes).
c07.qxd 09/08/02 14:41 Page 291
7.8 TURBO CODES 291
At the receiver, the incoming symbol stream is sampled to create a soft input to the
two decoders. A soft decoder creates a digital word from each sample using an ADC so
that information about the magnitude as well as the state of the received symbol is re-
tained. Recovered bits are given a weighting in the decoding process according to the con-
fidence level from the sampling process. For example, in a BPSK receiver, we expect to
recover symbols with sample magnitude V or V volts. We have greater confidence that
a symbol with a magnitude close to the expected value of V or V volts is correct than
for a received symbol with a magnitude close to zero volts because we can guess that a
significant amount of noise has been added to a received symbol with a value close to
zero volts. The symbol with a value close to zero volts is much more likely to be in er-
ror. The magnitude of the received symbols is retained through the decoding process,
which is one of the strengths in turbo codes. The process is known as soft input soft output
(SISO) decoding.
The soft input symbol stream is input to an interleaver of the same size as the in-
terleaver used in the transmitter and then read out by row and by column into two SISO
decoders corresponding to the two encoders at the transmitter. The outputs of the decoders
are two versions of the original message data, one of which was interleaved and the other
direct. The outputs from the decoders are compared, using the soft output values to ap-
ply confidence levels to the decoded bits. The decoding process is then repeated to ob-
tain a better estimate of the original transmitted data. The characteristics of the encoding
schemes and the decoding methods is such that repeated processing of the interleaver out-
put through the decoders reduces the number of errors remaining in the recovered data,
thus improving the BER. The disadvantage of this approach is that many iterations in the
decoders creates latency in the bit stream and the soft decoder must run at a clock rate
many times higher than the bit rate of the message data. The overall performance of turbo
codes can be improved by adding an outer layer of Reed–Solomon encoding, just as with
concatenated convolutional encoding.
Much research has been carried out to optimize the two codes used by a turbo en-
coder and the soft decoding process at the receiver. The aim is to achieve the highest cod-
ing gain with the smallest latency and the least complex decoder. With lower speed bit
streams such as compressed digital speech, it is possible to use fast DSP integrated circuits
to perform the decoding process, making turbo coding attractive for cellular telephone sys-
tems. The additional coding gain achieved by turbo codes and the possibility of operating
at CN ratios as low as 0 dB makes turbo coding a good choice for a fading radio chan-
nel. With higher bit rates, field programmable gate arrays (FPGAs) are needed to perform
parallel processing. Commercial turbo code products in the form of coder and decoder
boards became available in the late 1990s, and are now available as single chip codecs.
The Jet Propulsion Lab (JPL) in the United States has developed turbo codes for use
on its deep space research spacecraft24. The signals from spacecraft at interplanetary dis-
tances are very weak, so powerful error correction coding at very low received CN ratios
is essential. The JPL turbo code system uses a 16-bit block code and sends an uncoded
version of the message bits as well as the usual two encoded versions, and iterates in the
decoder using the three received versions of the message. JPL reports that the turbo code
implemented for the Voyager spacecraft uses a 16,384 bit interleaver and 10 iterations in
the decoder giving a half rate code with a BER of 105 at a received SN ratio of 0.7 dB.
Even more powerful codes are planned for the Cassini spacecraft with BER of 106 at an
input CN of 0 dB using a rate 16 code. Given Shannon’s limit of 1.6 dB for error free
recovery of data, turbo codes are approaching this theoretical performance limit.
Reference 24 contains a listing of web sites worldwide with information about turbo
code development, and a list of commercially available turbo code products.
c07.qxd 09/08/02 14:41 Page 292
7.9 SUMMARY
The transmission of data over a satellite communica- ms) in a satellite link makes simple stop-and-wait
tion link is likely to result in some errors occurring systems unattractive for real-time data transmission.
in the received data, for at least a small percentage of Throughput can be increased by providing data stor-
time, because of noise added by the transmission sys- age at both ends of the link and using continuous
tem. Many links guarantee only 106 bit error rate transmission in which corrupted data blocks are re-
and may not achieve this accuracy during periods of transmitted by interleaving them with subsequent
rain or other propagation disturbances. Bit errors con- data block transmissions.
tribute to the baseband (SN) ratio when digital Forward error correction (FEC) provides a means
speech is sent, but it is rarely necessary to correct bit of both detecting and correcting errors at the receiv-
errors in speech; the listener can make such correc- ing terminal without retransmission of data. FEC
tions because there is a lot of redundancy in speech codes add redundant parity check bits to the data bits
and for error rates up to 104 speech remains intelli- in a way that allows errors to be located within a code-
gible. When data are sent over a link, the receiving word. In general, twice as many errors can be detected
terminal does not know in advance what form the data by a FEC code as can be corrected. FEC has the ad-
take and can only detect or correct errors if extra, re- vantage over error detection that a single unit at each
dundant bits are added to the transmitted data. end of the link (a codec) can insert and remove the
Coding of data provides a means of detecting er- FEC code and make corrections as required. FEC is
rors at the receiving terminal. Error detecting codes used on all satellite links where the CN ratio at the
allow the presence of one or more errors in a block receiver is likely to be low. This includes satellite tele-
of data bits to be detected. Error correcting codes phones, VSAT terminals, and direct broadcast satel-
allow the receiving terminal equipment to locate and lite television. All of these systems make use of the
correct a limited number of errors in a block of data. coding gain from rate one-half or rate three-quarters
When error detection is employed, a retransmission FEC to achieve low BER.
scheme is often needed so that the data block can Interference tends to cause burst errors in which
be sent again when it is found to be in error. Re- many sequential bits are corrupted. Special burst-error
transmission schemes use ARQ (automatic repeat re- correction codes are available with the capability of
quest) techniques and are easiest to apply in packet correcting errors in a number of adjacent bits. Scram-
switched data networks, where data are not trans- bling and interleaving of data bits are other ways in
mitted in real time. The long round-trip delay (480 which the effect of burst errors can be reduced.
REFERENCES
1. C. E. SHANNON, “A Mathematic Theory of Communi- 9. K. FEHER, Digital Communications: Satellite Earth Sta-
cations,” Bell System Technical Journal, Part 1, tion Engineering, Prentice-Hall, Englewood Cliffs, NJ,
379–423; Part 11, 623–656 (1948). 1983.
2. G. DRURY, M. GARIK, and K. PICHAVANO, Turbo Codes: 10. T. MARATANI, H. SAITHOH, K. KOGA, Y. MIZUNO, and
Principles and Applications, Kluwer, Dordrecht, June Y. J. S. SNYDER, “Application of FEC Coding to the
2000. Intelsat TDMA Systems,” Proceedings of the 4th Inter-
3. L. W. COUCH, Digital and Analog Communication Sys- national Conference on Digital Satellite Communica-
tems, Prentice-Hall, Englewood Cliffs, NJ, 5th Ed., tions, Montreal, October 1978.
1993. 11. MARLIN P. RISTENBATT, “Alternatives in Digital Com-
4. K. S. SHAMNUGAM, Digital and Analog Communication munication,” Proceedings of the IEEE, 61, 703–721,
Systems, John Wiley & Sons, New York, 1979. June 1973. (Reprinted in reference 11, pp. 212–230.)
5. S. LIN and D. J. COSTELLO, Jr., Error Control Coding: 12. HARRY L. VANTREES, ed., Satellite Communications,
Fundamentals and Applications, Prentice-Hall, IEEE Press, New York, 1979.
Englewood Cliffs, NJ, 1983. 13. J. J. SPILKER, Jr., Digital Communications by Satellite,
6. A. HOCQUENGHEM, “Codes Corecteurs d’Erreurs,” Prentice-Hall, Englewood Cliffs, NJ, 1977.
Chiffres, 2, 147–156 (1959). 14. G. DAVID FORNEY, Jr., “The Viterbi Algorithm,”
7. R. C. BOSE and D. K. RAY-CHAUDHURI, “On a Class of Proceedings of the IEEE, 61, 268–278, March 1973.
Error Correcting Binary Group Codes,” Information (Reprinted in reference 11, pp. 286–296.)
Control, 3, 68–79, March 1960. 15. G. UNGERBOECK, “Trellis-Coded Modulation with Re-
8. W. W. PETERSON and E. J. WELDON, Jr., Error Correct- dundant Signal Sets,” Parts I and II, IEEE Communica-
ing Codes, 2nd Ed., MIT Press, Cambridge, MA, 1970. tions Magazine, 25, 5–21, February 1987.
c07.qxd 09/08/02 14:41 Page 293
PROBLEMS 293
16. S. LIN, T. KASAMI, T. FUJIWARA, and M. FOSSORIER, 20. S. LIN and P. S. YU, “A Hybrid ARQ System with Par-
Trellises and Trellis Based Decoding Algorithms for ity Retransmission for Error Control of Satellite Chan-
Linear Block Codes, Kluwer, Dordrecht, April 1998. nels,” IEEE Transactions on Communications, COM-
17. JERROLD A. HELLER and IRWIN MARK JACOBS, “Viterbi 30, 689–694, July 1982.
Decoding for Satellite and Space Communications,” 21. C. BERROU, A. GLAVIEUX, and P. THITIMAJSHIMA, “Near
IEEE Transactions on Communications, COM-19, Shannon Limit Error-Correcting Coding and Decoding:
835–848, October 1971. (Reprinted in reference 11, Turbo Codes,” ICC, pp. 1064–1070, 1993.
pp. 273–286.) 22. SERGIO BENEDETTO and GUIDO MONTORSI, “Per-
18. G. DRURY, G. MARHAVIAN, and K. PICHAVANO, Coding formance of Continuous and Blockwise Decoded Turbo
and Modulation for Digital TV, Kluwer, Dordrecht, Codes,” Communications Letters, Vol. 1, No. 3, May
November 2000. 1997, pp. 77–79.
19. S. LIN, D. J. COSTELLO, Jr., and M. J. MILLER, 23. R. PYNDIAH, “Near Optimum Decoding of Product
“Automatic-Repeat-Request Error Control Schemes,” Codes: Block Turbo Codes,” IEEE Transactions on
IEEE Communications Magazine, 20, 5–7, December Communications, August 1998.
1984. 24. http://www331. jpl.nasa.gov/public/JPLtcodes.html
PROBLEMS
1. Alphanumeric characters are transmitted as 7-bit Mbps when coding is added. What is the data rate
ASCII words, with a single parity bit added, over a with half rate FEC applied to the data?
link with a transmission rate of 9.6 kbps. c. Repeat the calculation of part (b) above when the
a. How many characters are transmitted each second? bit rate on the link is increased to 3.088 Mbps with
b. If a typical page of text contains 500 words with no increase in transmitter power.
an average of five characters per word and a space 4. The analysis of a 56-kbps data link shows that it
between words, how long does it take to transmit a suffers burst errors that corrupt several adjacent bits.
page? The statistics for burst errors on this link are given in
c. If the bit error rate on the link is 105, how many Table P.4.
characters per page are detected as having errors? a. Using Table P.4, select a burst error correcting code
How many undetected errors are there? that will reduce the probability of an uncorrected burst
d. On average how many pages can be transmitted error below 1010.
before (i) a detected error occurs or (ii) an undetected b. Calculate the data rate for messages sent over the
error occurs? link using the code you selected.
e. If the BER increases to 103, how many detected c. Estimate the average bit error rate for the coded
and undetected errors are there in a page of the text? transmission.
2. A (6, 3) block code has a minimum distance of 5. A satellite link carries packet data at a rate of
two. 256 kbps. The data are sent in 255-bit blocks using
a. How many errors can be detected in a codeword?
b. How many errors can be corrected in a codeword? TABLE P.4 Statistics for Burst
Errors on a Link in Problem 4
3. A QPSK data link carries a bit stream at 1.544
Mbps and has an overall (CN)0 ratio of 16 dB in the
No. adjacent bits Probability of
receiver at the VSAT earth station in clear air. The Corrupted occurrence
QPSK demodulator at the VSAT station has an im-
plementation margin of 1.0 dB. For 0.1% of the year 2 4 102
rain attenuation causes 3.0 dB reduction in the re- 3 2 103
ceiver (CN)0 ratio. For 0.01% of the year rain at- 4 3 104
tenuation causes 6.0 dB reduction in the receiver 5 1 106
(CN)0 ratio. 6 2 109
a. Calculate the BER in clear air, and the BER ex- 7 5 1011
ceeded for 0.1% and 0.01% of the year. 8 1 1012
b. Repeat the calculation when data are transmitted 9 3 1014
using half rate FEC with a coding gain of 5.5 dB (at 10 2 1017
all BERs). The bit rate on the link remains at 1.544
c07.qxd 09/08/02 14:41 Page 294
a (255, 247) code that can detect three errors. The 6. Repeat Problem 5 using a block length of 1024
probability of a single bit error, p, varies from 106 bits and a (1024, 923) code that can detect 22 er-
under good conditions to 103 under poor conditions. rors in a block. The (1024, 923) code can correct
The one-way link delay is 250 ms. 10 errors. Find the average number of blocks that
a. If no error detection is used, what is the message can be transmitted before an uncorrected error oc-
data rate for the link? curs when the BER is 103. Repeat the analysis for
b. For a link BER of 106, find the probability of de- a BER of 102.
tecting an error in a block of 255 bits when error Note: The probability of an unlikely event (11 or more
detection is applied. Hence find how often an error is errors in a block of 1024 data bits with Pb 103 in
detected. this case) can be calculated from the Poisson distri-
c. Estimate the probability that a block of 255 bits bution more easily than from the binomial distribu-
contains an undetected error when the link BER is tion. The Poisson distribution is given by
103 and error detection is applied.
lkel
d. Find the message data throughput when the link P1x k2
BER is 106 and a stop-and-wait ARQ system is used, k!
assuming one retransmission always corrects the where NPb. N is the block length, and k is the
block. number of bits in error.
c08.qxd 12/08/02 09:54 Page 295
CHAPTER 8
PROPAGATION EFFECTS AND THEIR
IMPACT ON SATELLITE–EARTH LINKS
Communications system design requires the development of a link budget between the
transmitter and the receiver that provides an adequate signal level at the receiver’s de-
modulator to achieve the required level of performance and availability. The perform-
ance of a link is usually defined for time percentages in excess of 99% over periods of
at least a month and is, for digital systems, determined by the bit error rate (BER) that
provides the minimum level of service. For analog systems, the CN at the demodula-
tor input that provides the minimum signal quality required defines the performance
level for that link. The availability of a link is usually defined for low outage time per-
centages (typically between 0.04 and 0.5% of a year, or between 0.2 and 2.5% of the
worst month, for satellite systems) and is, for digital systems, specified by the BER at
which an outage is declared for the link. For analog systems, the CN at the demodulator
input at which no usable signal can be demodulated defines the limit of availability.
Figure 8.1 illustrates the concept of performance and availability for a digital system
with BER as the determinant.
The link budget was covered in Chapter 4, as was link margin: the difference
in power level between clear sky conditions (essentially the performance level)
and that which exists at the threshold of the demodulator when the link is under im-
paired conditions (the availability level). Actually, there are two margins to consider
in a link budget: (1) the margin between the “clear sky” level and the performance
threshold; and (2) the margin between the performance threshold and the availability
threshold. Figure 8.2 illustrates these two concepts of margin for a typical digital Ku-
band downlink (11 GHz) located in the Mid-Atlantic region of the United States. As
can be seen, the attenuation experienced on the link varies with time percentage, grad-
ually falling through the performance threshold and then the availability threshold. It
is the link designer’s task to ensure that loss of signal occurs for no longer than the
time permitted for that service. The development of an accurate link budget, which
includes losses due to the passage of the signal through the atmosphere, is therefore
critical.
The key equation in the development of the link power budget in Chapter 4 was
Eq. (4.11), repeated here in modified form as Eq. (8.1).
Pr EIRP Gr Lp La dBW (8.1)
The complete volume(s) of the ITU material can be obtained from

International Telecommunications Union
Sales and Marketing Division
Places des Nation—CH-1211 Geneva 20
Switzerland
Email: sales@itu.int
http://www.itu.int/publications
295
c08.qxd 12/08/02 09:54 Page 296
10−10
Performance
10−8 threshold
Degraded performance region

BER
Availability
10−6 threshold
10−4
10−2
0.001 0.01 0.1 1.0 10.0 100%
Time percentage BER level measured
FIGURE 8.1 Schematic of the bit error rate (BER) statistic for a typical communications
link. A link is normally designed to provide a given performance specification for a very high
percentage of the time. In this example, a BER of 108 is the performance required for 99.9%
of the time. The time period over which the statistics are taken is usually a year or a month.
Atmospheric constituents (gases, clouds, rain, etc.) will cause the BER in clear sky conditions
to degrade. At some point, the BER will reach the level at which an outage is declared. This
point defines the availability specification. In this example, a BER of 106 is the availability
threshold and it must be met, in this example, for a minimum of 0.01% of the time.
10
Availability
8 threshold
Attenuation, dB

Performance
6 threshold
2
0.001 0.01 0.1 1.0 10.0 100%
Time percentage attenuation level measured
FIGURE 8.2 Schematic of the loss statistics encountered by a signal on transmission through
the atmosphere for a typical Ku-band communications link. In most communications links, an
allowance in power margin is built into the link so that the received signal is above the thresh-
old for satisfactory demodulation and decoding. This power margin is commonly referred to as
the fade margin since the signal, on occasion, appears to fade below the level established in
clear sky conditions. In the schematic above, the link experiences an equivalent fade of about 6
dB before it reaches the performance threshold level established for the link (see Figure 8.1). A
further fade of 2 dB, making a total reduction in signal level of 8 dB, takes the link below the
availability level established for the link (see Figure 8.1). The relationship between power level,
fade margin, and BER, will depend on the modulation used. It will also depend on the amount
of channel coding used. In the example above, no inner (FEC) or outer (Reed–Solomon, inter-
leaved) coding has been assumed for the link and the modulation is QPSK. For most heavily
coded links, the difference between good performance and an outage (a change on the order
of 2 to 3 decades of BER) will occur for a change in signal level of less than 1 dB.
296
c08.qxd 12/08/02 09:54 Page 297
8.1 INTRODUCTION 297
This equation indicates how the received power, Pr, in dBW depends on the transmitter
EIRP (the Effective Isotropic Radiated Power, which is a combination of the output
amplifier power, the gain of the transmitting antenna, and the losses associated with that
antenna system), the receiving antenna gain, Gr, (which includes, in this case, all losses
associated with the receiving antenna), the path loss, Lp, (given by 20 log10 [4R], with
being the wavelength of the signal and R the distance between the transmitting antenna
and the receiving antenna) and the attenuation contribution due to the atmosphere, La. Of
the terms on the right hand side of equation (8.1), the only one that is not essentially con-
stant with time is the atmospheric loss, La. The component La, usually referred to as prop-
agation loss, determines the margin required by the communications link to meet both the
performance and availability specifications.
8.1 INTRODUCTION
There are many phenomena that lead to signal loss on transmission through the earth’s
atmosphere. These include: Atmospheric Absorption (gaseous effects); Cloud Attenuation
(aerosol and ice particle effects); Tropospheric Scintillation (refractive effects); Faraday
Rotation (an ionospheric effect); Ionospheric Scintillation (a second ionospheric effect);
Rain Attenuation; and Rain and Ice Crystal Depolarization. Rain attenuation is by far the
most important of these losses for frequencies above 10 GHz, because it can cause the
largest attenuation and is usually, therefore, the limiting factor in Ku and Ka band satel-
lite link design. Raindrops absorb and scatter electromagnetic waves. In Ku and Ka bands,
rain attenuation is almost entirely caused by absorption. At Ka band, there is a small con-
tribution from scattering by large raindrops. The various propagation loss mechanisms are
illustrated in Fig. 8.3. We will discuss each of these loss mechanisms briefly; for a detailed
treatment the reader should refer to references 1 and 2.
Figures 8.1 and 8.2 introduced the concept of a time varying BER (or excess link
attenuation). Fig. 8.3 indicates where each of the loss mechanisms can be found along the
slant path to the satellite. It is also very useful to develop an appreciation for the various
time percentages over which each of the propagation loss mechanisms is significant.
Figure 8.4 illustrates this schematically, using the same curves from Fig. 8.1.
Signal loss—i.e. attenuation—affects all radio systems; those that employ orthog-
onal polarizations to transmit two different channels on a common, or partially over-
lapping, frequency band may also experience degradations caused by depolarization.
This is the conversion of energy from the wanted (i.e., the co-polarized) channel into
the unwanted (i.e., the cross-polarized) channel. Under ideal conditions, depolarization
will not occur. When depolarization does occur, it can cause co-channel interference
and cross-talk between dual-polarized satellite links. Rain is a primary cause of
depolarization.
Both attenuation and depolarization come from interactions between the propagat-
ing electromagnetic waves and whatever is in the atmosphere at the time. The atmos-
pheric constituents may include free electrons, ions, neutral atoms, molecules, and
hydrometeors (an arcane term that conveniently describes any falling particle in the at-
mosphere that contains water: raindrops, snowflakes, sleet, hail, ice-crystals, graupel,
etc.); many of these come in a wide variety of sizes. Their interaction with radio waves
depends strongly on frequency, and effects that dominate 30 GHz propagation, for ex-
ample, may be negligible at 4 GHz. The converse is also true. With one major exception
(ionospheric effects) almost all propagation effects become more severe as the frequency
increases.
c08.qxd 12/08/02 09:54 Page 298
298 CHAPTER 8 PROPAGATION EFFECTS AND THEIR IMPACT ON SATELLITE–EARTH LINKS
Sun
Satellite
Ionosphere
Depolarization
Rain cloud
Attenuation
Gases
Sky noise Melting

emission Rain layer
Refractive
effects
VSAT
FIGURE 8.3 Illustration of the various propagation loss mechanisms on a typical
earth–space path. The earth terminal (in this example a very small aperture terminal or
VSAT) is directed toward a satellite. Refractive effects (causing tropospheric scintillation);
gases; a rain cloud, melting layer, and rain, all exist in the path and cause signal loss. The
absorptive effects of the atmospheric constituents cause an increase in sky noise to be ob-
served by the VSAT receiver. While atmospheric gases and tropospheric scintillation do not
cause signal depolarization, collections of nonsymmetrical ice crystals and rain particles can
depolarize the transmissions through them. Above the lower (neutral) atmosphere is the ion-
osphere, which begins at about 40 km and extends well above 600 km. The ionosphere can
cause the electric vector of signals passing through it to rotate away from their original po-
larization direction, hence causing signal depolarization. At certain times of the day, year, and
11-year sunspot cycle, the ionosphere can cause the amplitude and phase of signals passing
through it to change rapidly, i.e., to scintillate, about a general mean level. The ionosphere
has its principal impact on signals at frequencies well below 10 GHz while the other effects
noted in the figure above become increasingly strong as the frequency of the signal goes
above 10 GHz. Finally, if the sun (a very “hot” microwave and millimeter wave source of in-
coherent energy) is in the VSAT beam, an increased noise contribution results which may
cause the C/N to drop below the demodulator threshold. Note: The above picture is not
drawn to scale. Most rainstorms occur below 10 km altitude and the ionosphere is not nor-
mally present below 40 km, and extends to more than 1000 km above the earth.
8.2 QUANTIFYING ATTENUATION

AND DEPOLARIZATION
Attenuation, A, is the decibel difference between the power received, Pr, at a given time
t and the power received under ideal propagation conditions (often referred to as “clear
sky” conditions). With all values in decibel units, we have
A1t2 Prclearsky Pr 1t2 (8.2)
Attenuation, A(t), on satellite communications links operating at C, Ku, and Ka-band is
primarily caused by absorption of the signal in rain. On most satellite links above 10 GHz,
c08.qxd 12/08/02 09:54 Page 299
8.2 QUANTIFYING ATTENUATION AND DEPOLARIZATION 299
Tropospheric scintillation and gaseous attenuation
Cloud attenuation and effects of melting layer
Heavy thunderstorm rain
10−10
Performance
10−8 threshold

BER
Availability
10−6 threshold
10−4
10−2
0.001 0.01 0.1 1.0 10.0 100%
Time percentage BER level measured
FIGURE 8.4 Approximate range of annual time percentages that various atmospheric im-
pairments affect a link (after Figure 2 of reference 3 © John Wiley & Sons, Inc. Reprinted
with permission). Tropospheric scintillation (a refractive effect in the lower atmosphere) and
gaseous attenuation are pervasive phenomena that occur all of the time, but at different lev-
els of impact depending on the climate, elevation angle, and time percentage of interest.
Clouds exist at various time percentages, depending on the climate, but are generally pres-
ent for at least 30% of the time in most locations. As the concentration of the frozen particles
in the cloud increases, many will start to fall and will melt on reaching the 0°C isotherm. This
will lead to enhanced attenuation in the melting layer. Drizzle rain will fall when the water va-
por concentration reaches saturation levels. Such rain is usually stratiform and falls for be-
tween 1 and 10% of the time, depending on the climate. During hot periods, convective rain
will fall, often in the form of thunderstorms. Heavy thunderstorms account for the highest
rainfall rates, and hence the highest path attenuations encountered, but they exist for only
small time percentages in a year. Not shown in the above figure are ionospheric effects,
which have a diurnal, seasonal, and 11-year cyclical impact, again depending on where the
earth station is and the precise earth–space path used.
rain attenuation limits the availability of the system and, to develop an adequate link mar-
gin, the rain attenuation to be expected for a given time percentage needs to be calculated.
This can be a complicated process, but there are basically three steps: (a) determine the
rainfall rate for the time percentage of interest; (b) calculate the specific attenuation of
the signal at this rainfall rate in dB/km; and (c) find the effective length of the path over
which this specific attenuation applies. The difficult part of this process is part (c) because
rain falls in two broad categories: stratiform rain and convective rain. These two separate
atmospheric mechanisms have different effects on satellite paths. Stratiform rain is gen-
erated in cloud layers containing ice, and results in widespread rain or snow at rainfall
rates of less than 10 mm per hour. Convective rain is generated by vertical air currents
that can be very powerful, leading to thunderstorms and high rainfall rates. Convective
rain is very important for satellite communication systems because it is the major cause
c08.qxd 12/08/02 09:54 Page 300
Path to the
Stratiform satellite
rain event
Melting level
Rain
Ground level
(a )
Convective
rain event
Melting level
Path to the
Rain satellite
Ground level
(b)
FIGURE 8.5 (a) Stratiform rain situation. In this case, a widespread system of stratiform
rain—that is rain that appears to be stratified horizontally—completely covers the path to the
satellite from the ground up to the point where the rain temperature is 0°C. This level is
called the melting level because, above it, the precipitation is frozen and consists of snow
and ice crystal particles. Frozen precipitation causes negligible attenuation. In general, the
signal path in stratiform rain will exit the rain through the top of the rain structure. (b) Con-
vective rain situation. In this case, a tall column of convective rain enters the satellite-to-
ground path. In some cases the storm will be in front of the earth station; in others, behind
it. Convective storms normally occur in the summer, thus the melting level is much higher
than in winter. In many cases, the melting level is not well defined, as the strong convective
activity inside the storm will push the liquid rain well above the melting level height. Except
for paths with very high elevation angles (70°), the signal path in convective rain will most
often exit from the side of a convective rainstorm.
of link outages. Stratiform rain consists of a generally constant rainfall rate over a very
large area while convective rain is generally confined to a narrow, but tall, column of rain.
Figure 8.5 illustrates the two rain processes and Figure 8.6 gives the concept of the path
attenuation calculation procedure for both rain types.
Stratiform rain occurs typically ahead of a warm front in an area of low pressure.
Large areas of cloud exist in which ice crystals are sufficiently large to slowly fall and
join other ice crystals to form snowflakes, which fall more quickly as their size increases.
If there is a high concentration of moisture in the clouds, in the form of ice, large
snowflakes may form. The snow falls until it reaches the melting layer. The melting layer
is simply the region of the atmosphere where the temperature transitions from below 0°C
c08.qxd 12/08/02 09:54 Page 301
Path to the
satellite
Melting level
hr
θ
Ground level
(a )
Path to the
satellite
Melting level
Height
adjustment
Effective factor
path length
box
Horizontal
adjustment
factor
Ground level
(b)
FIGURE 8.6 (a) Stratiform rain attenuation calculation procedure. In the case of stratiform
rain, the rainfall rate along the path can be considered to be uniform and the path com-
pletely immersed in the rain. The effective path through the rain—the path over which the
rain may be considered to be uniform—is therefore the same as the physical path length in
stratiform rain. The path attenuation A is therefore the specific attenuation (i.e., dB attenua-
tion per km) multiplied by the physical path length in the rain (i.e., hrsin ). (b) Convective
rain attenuation calculation procedure. In the case of convective rain, the melting level and
elevation angle are used to develop two adjustment factors: a height adjustment factor and
a horizontal adjustment factor. Once these factors have been used, a smaller box is created
inside which it may be assumed that the rainfall rate is uniform. The length of the path that
exists inside this box is the effective path length and it is this that is used to multiply the
specific attenuation with. In this case, the path exits through the top of the effective path
length box. In other cases, it may exit through the side.
to above 0°C. Snow falling into air at a temperature greater than 0°C melts and forms
raindrops. If the air at the earth’s surface is below 0°C the snow does not melt, but con-
tinues to the ground. The stratiform cloud mechanisms that generate snow result in low
rainfall rates, always less than 10 mm per hour, and widespread (stratiform) rain or snow.
This leads to generally constant attenuation of the slant-path signals over the entire path
length from the ground to the melting layer.
EXAMPLE 8.2.1
An earth station at sea level communicates at an elevation angle of 35° with a GEO satellite. The
melting level height of the stratiform rain is 3 km. Find (a) the physical pathlength through the rain;
(b) find the path attenuation if the specific attenuation is 2 dB/km.
c08.qxd 12/08/02 09:54 Page 302
Answer
(a) The vertical height, hr, of the rain is the difference between the melting level height (3 km)
and the height of the earth station (0 km, since it is at sea level), which gives hr 3 km.
Since the elevation angle is 35°, the physical pathlength, L, through the rain is given by L
hr(sin 35) 3(sin 35) 5.23 km.
(b) The rain is stratiform and so it can be considered to be uniform over the path. The specific
attenuation is therefore uniform along the path through the rain. If the specific attenuation is
2 dB/km, the path attenuation, A is given by A (2 dB/km) (5.23 km) 10.46 dB
10.5 dB
Convective rainstorms are very complex, and have both horizontal and vertical struc-
ture. A convective cell becomes established when a mass of warm moist air is pushed up
into colder air at a higher altitude. Adiabatic expansion of the air mass occurs, which cools
the air. When the air is cooled below its dew point, it condenses forming clouds, and drops
of water start to fall under gravity. The falling drops collide and coalesce with other drops
to make larger drops, leading to a drop size distribution. The maximum size of stable rain-
drops is about 6 mm—larger drops (sometimes exceeding 10 mm in average diameter)
are unstable and quickly up break into a collection of smaller drops under wind shear con-
ditions. Large raindrops fall quickly, with terminal velocities up to 8 or 9 m/s. If the falling
drops encounter supercooled water as they fall, hailstones can form. Hailstones can exceed
the 10 mm diameter limit of raindrops, and may reach golf ball size in severe thunder-
storms in the Great Plains. The accretion process can occur in an updraft as well as for
falling drops, and in a vigorous thunderstorm, updraft velocities can exceed 100 mph.
Since cold air is denser than warm air, once an updraft dies away at the top of a thun-
derstorm, cold air tends to flow downward, and can create a streamer, a narrow region of
intense rain and cold air. Streamers can be a few hundred meters wide or a kilometer wide.
At the surface, the streamer is observed as a microburst, which has strong wind shear as
the vertical down flow of cold air hits the ground and spills out in all directions. We are
all familiar with microbursts. Shortly before heavy rain falls there is often a cold wind,
followed by a downpour. The cold wind we feel is the outflow of cold air as it hits the
earth’s surface. The effect of convective rain on a satellite slant path depends on the an-
gle at which the path intersects a streamer. Streamers are rarely vertical, so if a slant path
SIDEBAR
Microbursts are dangerous for aircraft flying close to and tends to slow the rate of descent. The natural re-
the ground, especially when taking off and landing. action of the pilot is to reduce engine power to main-
Several serious accidents to passenger aircraft in the tain a constant rate of descent on the 3° glide slope.
1980s were attributed to the wind shear associated However, as soon as the aircraft passes the center of
with microbursts, and extensive research was carried the microburst, the wind direction is opposite, and is
out to develop ways to detect microbursts and wind now a tailwind, which reduces the speed of the aircraft
shear. Networks of anemometers, which measure relative to the air and increases the rate of descent of
wind speed, can be deployed around an airfield to de- the aircraft. If the engine power has been cut, the air-
tect wind shear, and terminal Doppler radar can be plane may sink into the ground before the engines
used for the same purpose. have developed enough power to keep the airplane
An aircraft on final approach to a runway is fly- aloft. Wind shear detection equipment at airfields
ing slowly and descending on a 3° glide slope. If the and improved pilot awareness of the dangers of
aircraft encounters a microburst, it first experiences a microbursts has reduced the incidence of accidents
headwind, which increases its speed relative to the air caused by microbursts.
c08.qxd 12/08/02 09:54 Page 303
6
Height above surface (km)
4 Heavy rain
Medium rain
2 18° slant path
Light rain
0
0 4 8 12
Range (km)
FIGURE 8.7 Example of an RHI scan through a rain storm. Radar reflectivity contours in a
rainstorm on June 15, 1986, measured with an S-band radar in Blacksburg, Virginia. The con-
tours represent light, medium, and heavy rain in a narrow vertical column. The radar and a
receiving station were collocated at the (0,0) point. Note the narrow extent of heavy rain in a
sloping column, and the effect on the slant path to the satellite at an elevation angle of 18°.
The statistical rain height Hi for Blacksburg is 4.1 km. In this example, rain is present up to an
altitude of 5.6 km above sea level.
is parallel to a streamer it will suffer very heavy attenuation if the streamer envelops the
path. If the slant path cuts across the streamer, the path length within the heavy rain may
be quite short, leading to relatively little attenuation despite the high rainfall rate.
Figure 8.7 shows an examples of a convective rain cell observed with an S-band
radar at Virginia Tech’s satellite tracking station. The radar was used to make vertical scans
across the slant path to a satellite (known to radar people as an RHI scan, for range-height
indicator, a WWII radar display mode). The complex shape of the storm cell requires the
use of artificial “adjustment factors” to convert the physical path through the rainstorm to
an effective path length over which the rain may be considered to be uniform. As well as
causing significant attenuation, rain and ice crystals can cause depolarization.
Depolarization is more difficult to quantify than attenuation. All signals have a po-
larization orientation that is defined by the electric field vector of the signal. (See Figure 8.8.)
In general, signals are never purely polarized; the direction of the electric field will never
be perfectly oriented or constant. Successful orthogonal polarization frequency sharing—
usually called dual-polarization frequency reuse—requires that there be sufficient isolation
between two orthogonal polarization states to permit the separation of the wanted polar-
ization (the copolarized signal) from the unwanted polarization (the cross-polarized signal)
at the receiving antenna5. The difference between the copolarized and the cross-polarized
signal energy will determine the cross-polarization discrimination at the receiver, the XPD,
and hence the level of interference between two orthogonally polarized signals.
To illustrate the process by which depolarization is measured; imagine a dual-
polarized antenna transmitting orthogonally polarized signals. We will call the two
polarizations V (for vertical) and H (for horizontal) for convenience, although there
c08.qxd 12/08/02 09:54 Page 304
FIGURE 8.8 Orthogonally polarized waveguide horn antennas. The polarization of an elec-
tromagnetic wave is defined by the orientation of the electric vector. In the example above,
two waveguide horns, excited in the TE10 mode, are radiating in the same direction. The top
horn is oriented such that the electric vector is vertically polarized; the bottom horn is turned
on its side compared with the top horn and so the electric vector is horizontally polarized.
The arrows indicate the electric field vector. Since the electric polarization vectors are ori-
ented 90° with respect to each other in the two horns, the transmitted signals are considered
to be orthogonally polarized. Orthogonally polarized signals do not interfere with each other,
even if they are at exactly the same frequency, provided they are “purely” polarized (i.e.,
there is no component of the signal present in the other, orthogonal, polarization). In all
cases, however, the transmitted signals are not purely polarized, due to antenna imperfec-
tions, so a component exists in the unwanted polarization. In addition, some of the energy in
one polarization can “cross” over to the other polarization due to asymmetric particles (e.g.,
large, oblate raindrops) existing in the propagation path. This cross-polarized energy can
give rise to interference between the two, mutually orthogonal polarizations. The degree of
cross-polarization to be expected along a given path is predicted using cross-polarization
models that are usually based on the rain attenuation along the path.
are infinitely many orthogonal polarization pairs. Let the complex phasor amplitudes of
the transmitted electric field vectors with polarization V and H be a and b, respectively,
as shown in Figure 8.9. The transmitting antenna is excited so that a and b are equal.
If the transmission medium between the transmitting and receiving antennas were
clear air, phasor a would give rise to a V polarization wave of amplitude ac at the receiving
antenna and phasor b would cause an H polarization wave of amplitude bc. The subscript
c stands for copolarized; these fields have the same polarization sense as their transmitted
counterparts. (See Figure 8.10.)
If asymmetrical rain or ice crystal particles exist in the transmission medium, some
of the energy in a will couple into a small (cross-polarized) H polarized field component
c08.qxd 12/08/02 09:54 Page 305
V
Direction of propagation
H
FIGURE 8.9 Fields excited by a dual-polarized antenna. The field radiated by the V horn
has the vertically polarized electric field vector indicated by a and the field radiated by the H
horn has the horizontally polarized field vector indicated by b. In most antenna systems, one
horn is used to radiate both polarizations simultaneously rather than two. This permits the
single feed horn to be located at the prime focus of the antenna to generate the best far field
pattern5. The two polarization senses to be transmitted are excited in separate parts of the
transmitter and are then coupled together via an ortho-mode transducer into a single wave-
guide section. This waveguide section, which can support both polarizations simultaneously,
is then used to couple the signals into a waveguide horn that is capable of radiating both
polarizations senses equally.
whose amplitude at the receiving antenna is ax, and b will give rise to a small (cross-polarized)
V polarized component bx. An ideal receiving system that introduces no cross-polarization
will have a V channel output (ac bx) and an H channel output (bc ax). The unwanted
bx term represents interference with the wanted signal ac and the unwanted ax term is in-
terference with the wanted signal bc. This interference will cause cross talk on an analog
link and increase the BER on a digital link. This generation of unwanted cross-polarized
components is called depolarization.
Vertical Vertical
a ac
At transmitting At receiving
antenna antenna
bx
ax
bc
b
Horizontal Horizontal
FIGURE 8.10 Illustration of signal depolarization in the transmission path. The transmitted
fields a and b produce copolarized components ac and bc at the receiving antenna. The trans-
mission medium in this instance is not clear sky, nor is the transmitting antenna perfectly po-
larized, and the anisotropy of the transmission medium and imperfections in the transmitting
antenna induce cross-polarized components of the transmitted signal to be received. These
cross-polarized components at the receiving antenna are ax and bx. With perfect antennas
and in the absence of depolarization ax and bx would be zero.
c08.qxd 12/08/02 09:54 Page 306
The measure of depolarization that is most useful for analyzing communications

systems is the cross-polarization isolation, XPI. In terms of the complex phasor field
amplitudes, it is given by Eq. (8.3) for the V polarized channel and by Eq. (8.4) for the
H polarized channel.
XPIV ac bx (8.3)
XPIH bc ax (8.4)
The XPI values are commonly expressed in decibels; for example,
XPIV 20 log10 0ac bx 0 dB (8.5)
Physically, the XPI is the decibel ratio of wanted power to unwanted power in the
same channel. The larger the XPI value, the less interference there is and the better the
communications channel will perform. XPI is difficult to measure. It requires the simul-
taneous transmission of signals at the same frequency in both polarization senses. The
COMSTAR series of satellites had a beacon that rapidly switched between two orthogo-
nal polarization senses, thus permitting the measurement of XPI [e.g., 6, 7]. More recently,
the ACTS8 and OLYMPUS9 satellites also incorporated a switched beacon to permit XPI
measurements. Most propagation experiments are much simpler than this and measure
simultaneously the wanted (the copolarized) and the orthogonal, unwanted (the cross-
polarized) signals that are received from a satellite beacon that transmits in only one po-
larization. In this case (referring to Figure 8.10) the experiment would measure (say) sig-
nals ac and ax that are derived from a singly polarized signal a that is transmitted from
the satellite. Measuring received signals bc and bx simultaneously from a singly polarized
signal b would provide the same result. This process allows the cross-polarization dis-
crimination, XPD to be derived
XPDV ac ax (8.6)
or in decibels
XPDV 20 log10 0ac ax 0 dB (8.7)
In most transmission situations encountered in practice, the values calculated for
XPI and XPD are the same10 and they are sometimes simply called the “isolation.” In prac-
tice, real antennas do not transmit polarization pairs that are exactly orthogonal, nor does
the isolation remain the same over the 3-dB beamwidth of the antenna. Receiving anten-
nas can also introduce cross-polarization. There is therefore a residual XPD component
present even in clear sky conditions. This must be accounted for in the link budget of a
dual-polarized, frequency reuse system. The residual XPD on axis is normally better for
linearly polarized antennas (30 to 35 dB) than for circularly polarized antennas (27
to 30 dB). These values represent antennas carefully designed for dual-polarized opera-
tion; inexpensive antennas will typically exhibit about 20 dB XPD for linear or circular
polarizations.
8.3 PROPAGATION EFFECTS THAT ARE NOT

ASSOCIATED WITH HYDROMETEORS
In this section we will discuss propagation effects that are not associated with raindrops
or ice crystals: atmospheric absorption, cloud attenuation, refractive effects that include
tropospheric scintillation and low angle fading/multipath effects, Faraday rotation, and
ionospheric scintillation.
c08.qxd 12/08/02 09:54 Page 307
8.3 PROPAGATION EFFECTS THAT ARE NOT ASSOCIATED WITH HYDROMETEORS 307
Atmospheric Absorption
At microwave frequencies and above, electromagnetic waves interact with molecules in
the atmosphere to cause signal attenuation. At certain frequencies, resonant absorption oc-
curs and severe attenuation can result. Figure 8.11 (from Figure 6 of reference 11) shows
these resonant absorption peaks on a zenith path (that is, a path at an elevation angle of
103
a
2
102
2
Zenith attenuation (dB)
10
2
B
A
2
10−1
10−2 3 5 2 5
10 102 2 3.5
Frequency (GHz)
FIGURE 8.11 Total zenith attenuation due to atmospheric gases calculated from 3 to 350
GHz (from Figure 6 of reference 11 © ITU, reproduced with permission). The two curves rep-
resent the gaseous attenuation that would be observed looking straight up from sea level
(i.e., on a zenith path) right through the neutral atmosphere on a satellite–earth path. Curve
A is for a dry atmosphere (i.e., no water vapor present) while curve B is for a standard at-
mosphere. A standard atmosphere consists of a surface pressure of 1013 hPa [a hPa has the
same numerical value as the old pressure unit of millibars], a surface temperature of 15°C,
and a surface relative humidity of 7.5 mg/m3. Curve A shows only the resonant absorption
peaks of the oxygen molecules (a broad peak at 60 GHz and a narrow peak at 118.75 GHz).
Curve B includes the resonant absorption peaks due to the water vapor molecule at 22.235,
183.31, and 325.153 GHz. The shaded portion “a” indicates a range of values since there are
many individual resonant absorption lines in this frequency region.
c08.qxd 12/08/02 09:54 Page 308
90°) from a sea level location right through the neutral atmosphere. Neutral means that
no ionization is present.
The first absorption band in Figure 8.11 is that due to water vapor at 22.235 GHz.
The K-band sets of frequencies are on both sides of this absorption band, which has led
to the terminology of Ku band (signifying frequencies under the absorption band) and Ka
band (signifying frequencies above the absorption band). It is common to specify a satel-
lite frequency band by the uplink frequency. From Figure 8.11, it can be seen that gaseous
absorption accounts for less than 1 dB on most paths below 100 GHz that lie outside the
absorption bands. However, in many new systems that employ very small system mar-
gins, it is important to account for the gaseous losses along the anticipated path. New pre-
diction procedures that attempt to account for all attenuating phenomena along the path
(e.g., [12]) include gaseous absorption.
Cloud Attenuation
Once considered to be largely irrelevant for satellite communications paths, clouds have be-
come an important factor for some Ka-Band paths and all V-Band (5040 GHz) systems. The
difficulty with modeling cloud attenuation is that clouds are of many types and can exist at
many levels, each type having a different probability of occurrence. The water droplet con-
centrations in each cloud will also vary, and clouds made up of ice crystals cause little at-
tenuation. Two models have been proposed13,14, both of which have similar accuracy. Typical
values of cloud attenuation for water-filled clouds are between 1 and 2 dB at frequencies
around 30 GHz on paths at elevation angles of close to 30° in temperate latitudes. In warmer
climates, where clouds are generally thicker in extent and have a greater probability of oc-
currence than temperate latitudes, cloud attenuation is expected to be higher. As with most
propagation effects, the lower the elevation angle, the higher the cloud attenuation.
Tropospheric Scintillation
and Low Angle Fading
The atmosphere close to the ground, sometimes called the boundary layer, is rarely still.
Energy from the sun warms the surface of the earth and the resultant convective activity
agitates the boundary layer. This agitation results in turbulent mixing of different parts of
the boundary layer, causing small-scale variations in refractive index. Figure 8.12 illus-
trates the process.
Stratified Layers (calm conditions) Turbulent Mixing (convective conditions)

(a ) (b)
FIGURE 8.12 Schematic of stratified and turbulent conditions in the boundary layer of the
atmosphere. In (a), the air is calm and the lower atmosphere next to the earth’s surface (the
boundary layer) forms into layers. Each layer has a slightly different refractive index, decreasing
in general with height. In (b), the earth’s surface has become heated by energy from the sun
and the resultant convective activity has mixed the formerly stratified layers into “bubbles” that
have different refractive indices. The turbulent mixing of the lower atmosphere will cause rela-
tively rapid fluctuations in a signal passing through it, which are called scintillations.
c08.qxd 12/08/02 09:55 Page 309
8.3 PROPAGATION EFFECTS THAT ARE NOT ASSOCIATED WITH HYDROMETEORS 309
When a signal encounters a turbulent atmosphere, the rapid variation in refractive

index along the path will lead to fluctuations in the received signal level. These fluctua-
tions are generally about a fairly constant mean signal level and are called scintillations.
Because the bulk of the fluctuations are caused within 4 km of the earth’s surface, they
are referred to as tropospheric scintillations. Tropospheric scintillations occur in many
weather conditions, as can be seen in Figure 8.13.
Tropospheric scintillation does not cause depolarization. The magnitude of the scin-
tillations becomes generally larger as the frequency increases, the path elevation angle
reduces, and the climate becomes warmer and more humid. Prediction models exist to
calculate this phenomenon with good accuracy15. On paths below 10° elevation angle, tro-
pospheric scintillation can be performance limiting; below 5° elevation angle it can become
availability limiting.
When the elevation angle falls below 10°, a second propagation effect becomes
noticeable: low angle fading. Low angle fading is the same phenomenon as multipath
fading on terrestrial paths. A signal transmitted from a satellite arrives at the earth sta-
tion receiving antenna via different paths with different phase shifts. On combination,
0 0
Copolar (dB)
Copolar (dB)
−2 −2
−4 −4
−6 −6
−8 28/07/76 −8 19/04/76
−10 −10
11.00 11.10 11.20 11.30 16.35 16.45 16.55 17.05
Time GMT Time GMT
(a ) (d )
0 0
−2
Copolar (dB)
Copolar (dB)
−2 −4
−4 −6
−8
−6 −10
−8 06/07/76
−12
19/07/76
−10 −14
11.30 11.40 11.50 12.00 18.30 18.40 18.50 19.00
Time GMT Time GMT
(b) (e )
0 0
Copolar (dB)
Copolar (dB)
−2 −2
−4 −4
−6 −6
−8 20/01/76 −8 03/05/76
−10 −10
07.25 07.35 07.45 07.55 09.00 09.10 09.20 09.30
Time GMT Time GMT
(c ) (f )
FIGURE 8.13 Scintillations observed under a variety of weather conditions on a 30-GHz
downlink from ATS-6 (from reference 16). Scintillations with various amplitudes can be ob-
served under different weather conditions. Two of the data sets were taken in clear weather,
two in cloud conditions, and two during rain, as follows: (a) clear-weather copolar signal with
low scintillation; (b) clear-weather copolar signal with high scintillation; (c) copolar scintilla-
tion in cloud; (d) copolar scintillation in cloud; (e) copolar scintillation and attenuation in
rain; (f ) copolar scintillation and attenuation in rain. Note the difference in scintillation ampli-
tude under what are apparently similar weather conditions along the path.
c08.qxd 12/08/02 09:55 Page 310
the resultant waveform may be enhanced or attenuated from the normal clear sky level.
Signal enhancement has been observed to exceed 8 dB on a 3.3° path at 11.198 GHz17,
while cancellation can cause complete link dropout. The mechanism for low angle fad-
ing has been interpreted as atmospheric multipath and also as the “defocusing and fo-
cusing” of the incoming signal. Both explanations have merit: the received signal is
made up of components that have arrived via different paths (i.e., multipath), but the
mechanism for developing the different paths is one of refraction rather than reflec-
tion at the atmospheric layer boundaries. Low angle fading is only significant in very
still air on very low elevation angle paths. It is normally not considered for satellite
paths when the elevation angle is above 10°. Note that the multipath effect referred to
here is occurring in the atmosphere, and is therefore different from multipath effects
in terrestrial radio links which are caused by reflections from the ground, buildings,
trees, etc.
Faraday Rotation in the Atmosphere

The ionosphere is that portion of the earth’s atmosphere that contains large numbers of
electrons and ions. At its lowest, it reaches down to close to 40 km above the earth; there
is no distinct upper boundary, but it exists well above 600 km above the earth. The ion-
osphere completely dominates radio propagation below about 40 MHz, but its effects on
the frequencies used by most communications satellites are minor.
Electrically, the ionosphere is an inhomogeneous and anisotropic plasma and an ex-
act analysis of wave propagation through it is extremely difficult. For a given frequency
and direction of propagation with respect to the earth’s magnetic field, there exist two
characteristic polarizations. Waves with these polarizations, called characteristic waves,
propagate with their polarization unchanged. Any wave entering the ionosphere can be re-
solved into two components with the characteristic polarizations. The phase shift and at-
tenuation experienced by the characteristic waves can be calculated at any point along the
propagation path, and the total field can be computed as the vector sum of the fields of
the characteristic waves. This total field can be interpreted as an attenuated and depolar-
ized version of the wave that entered the ionosphere. Thus, when a linearly polarized (LP)
satellite path signal reaches the ionosphere, it excites waves with the two characteristic
polarizations. These travel at different velocities, and when they leave the ionosphere their
relative phase is different from when they entered. The wave that leaves the ionosphere
has a different polarization from the LP wave that was transmitted. This is called Farad-
ay rotation, and its effect is essentially the same as if the field vector of the transmitted
LP wave had been rotated by an angle . For a path length through the ionosphere of Z
meters, the rotation angle is given by
2.36 104
f a
f2
b ZNB0 cos u dz rad (8.8)
Here, is the angle between the geomagnetic field and the direction of propagation, N is
the electron density in electrons/cubic meter, B0 is the geomagnetic flux density in Tes-
las, and f is the operating frequency in Hz. The rotation angle varies inversely with f 2.
Table 8.1 gives the value of and some other parameters with frequency15. The polar-
izations of an earth station antenna can be adjusted to compensate for the Faraday rota-
tion observed under average conditions. However, the rotation of the uplink will be in an
opposite sense to that on the downlink and so, to compensate in both directions at the
same time, a feed will be required that is able to rotate the relevant sections in opposite
TABLE 8.1 Estimated Ionospheric Effects for Elevation Angles of About 30° One-Way Traversal (from TABLE 1 of [15] © ITU,
c08.qxd 12/08/02 09:55 Page 311
reproduced with permission)
Frequency
Effect Dependence 0.1 GHz 0.25 GHz 0.5 GHz 1 GHz 3 GHz 10 GHz
2
Faraday Rotation 1f 30 rotations 4.8 rotations 1.2 rotations 108° 12° 1.1°
Propagation Delay 1f2 25 s 4 s 1 s 0.25 s 0.028 s 0.0025 s
Refraction 1f2 1° 0.16° 2.4 0.6 4.2
0.36
Variation in the Direction

of arrival (r.m.s.) 1f2 20 3.2 48
12
1.32
0.12
Absorption (auroral and/or

polar cap) 1f2 5 dB 0.8 dB 0.2 dB 0.05 dB 6 103 dB 5 104 dB
Absorption (mid-latitude) 1/f2 1 dB 0.16 dB 0.04 dB 0.01 dB 0.001 dB 104 dB
Dispersion 1f3 0.4 ps/Hz 0.026 ps/Hz 0.0032 ps/Hz 0.0004 ps/Hz 1.5 105 ps/Hz 4 107 ps/Hz
Scintillation See Rec. ITU- See Rec. ITU- See Rec. ITU- See Rec. ITU- 20 dB peak- 10 dB peak- 4 dB peak-
R P.531 R P.531 R P.531 R P.531 to-peak to-peak to-peak
NOTES:
(a) The estimated values are based upon a total electron content (TEC) of 1018 electrons/m2, which is a high value of TEC encountered at low latitudes in day-time with
high solar activity.
(b) The scintillation values in the last entry are maximum values observed near the geomagnetic equator during the early nighttime hours (local time) at equinox
under conditions of high sunspot number.
311
c08.qxd 12/08/02 09:55 Page 312
directions. The XPD that results when the polarization angle of an LP wave changes by
an amount is given by
XPD 20 log10 1cot ¢f2 (8.9)
Hence, a 6° change from average conditions would reduce the XPD on the link to about
19.6 dB.
Ionospheric Scintillations
Energy from the sun causes the ionosphere to “grow” during the day, increasing the total
electron content (TEC) by two orders of magnitude, or more. The TEC is the total num-
ber of electrons that would exist in a vertical column of area 1 m2 from the surface of the
earth all the way through the earth’s atmosphere. Typical values of TEC range from 1018
during the day to 1016 during the night. It is the rapid change in TEC from the daytime
value to the nighttime value, which occurs at local sunset in the ionosphere, that gives
rise to irregularities in the ionosphere. The irregularities cause the signal to vary rapidly
in amplitude and phase, which leads to rapid signal fluctuations that are called ionospheric
scintillations. The magnitude of the ionospheric scintillations varies with time of day,
month in the year, and year in the 11-year sunspot cycle. The greatest scintillation effects
are observed just after local sunset in the equinox periods during the sunspot maximum
years. The effects are also worst within about 20° of the geomagnetic equator and over
the poles. The length of the cycles averages at around 11 years, but has been as short as
9.5 years and as long as 12.518. Solar sunspot cycle 22 was from 1986.8 to 1996.4.
8.4 RAIN AND ICE EFFECTS
At frequencies above 10 GHz, rain is the dominant propagation phenomenon on satellite

links. Many experiments have been conducted on geostationary satellite links, using ex-
perimental satellites such as SIRIO, OTS, and CTS (Hermes) at Ku Band and ATS-6,
OLYMPUS, and ACTS at Ka Band. One experimental satellite, Italsat, also allowed 5040
GHz (V Band) experiments to be conducted in Europe. References 1, 2, 8, and 9 provide
detailed results and explanations of all the propagation phenomena.
Characterizing Rain
Most farmers, hydrologists, and city planners need to know how much total rain will fall
in a given period: that is, the rain accumulation. Indeed, most weather forecasts are given
in terms of how much precipitation will fall (or accumulate) over a given region. Rain ac-
cumulation, unfortunately, is of little use to satellite link designers, since it is the rate at
which the rain is falling that is important: that is, the rainfall rate. Rainfall rate is meas-
ured by a rain gauge, the most common of which is a tipping bucket rain gauge. This is
fairly accurate between rainfall rates of 10 to 100 mm/h. Peak values of 100 to 150 mm/h
may be expected for short periods during summer thunderstorms in the mid-Atlantic region
of the United States. Higher rainfall rates are observed in tropical regions.
The long-term behavior of rainfall rate is described by a cumulative probability dis-
tribution or by a cumulative distribution function (cdf ). The cdf for rainfall rate is com-
monly referred to as an exceedance curve. This gives the percentage of time (usually the
percentage of 1 year) that the rainfall rate exceeds a given value. Climate related param-
eters tend to be very variable. Rain accumulation can vary significantly from year-to-year,
c08.qxd 12/08/02 09:55 Page 313
8.4 RAIN AND ICE EFFECTS 313
10.0000
3-year average
1979
1980
1981
1.0000
Percentage of time rain rate is exceeded
0.1000
0.0100
0.0010
0.0001
0 20 40 60 80 100 120 140 160 180 200
Rain rate (mm/h)
FIGURE 8.14 Typical rainfall rate cumulative probability distributions or “exceedance”
curves. These sets were measured at Virginia Tech, Blacksburg, United States as part of a
3-year experiment with the Italian satellite, SIRIO. The 1979 data indicate a relatively dry
year, while those of 1981 indicate a relatively wet year. Despite this, a single, rare thunder-
storm in 1979 produced much higher rainfall rates than those observed in 1981 at low time
percentages. The availability level the link has to operate at will determine what rainfall rate
is of most importance and it will also give a range over which the design must cope. For
example, if 0.01% was the availability requirement, in 1979 the rainfall rate for this time per-
centage was 38 mm/h while in 1981 it was 58 mm/h. This shows the value of long-term sta-
tistics so that one year’s data do not bias the link design.
as can the exceedance curves, particularly at the low time percentages of interest to satel-
lite link designers.
Three annual exceedance curves taken from an experiment performed at Blacks-
burg, Virginia, are shown in Figure 8.14 and it can be seen that the rainfall rate at the
0.01% point varies between 38 and 58 mm/h over the 3 years. The cumulative attenua-
tion curves for the 3 years showed similar trends. Depending on the elevation angle of the
link, this can make a significant difference in the attenuation measured at the same time
percentage in each of the years. For this reason, link designers prefer to use values aver-
aged over many years of measurements for their propagation models. There have been
two approaches to developing these long-term statistics: rain climate maps and exceedance
contour maps.
c08.qxd 12/08/02 09:55 Page 314
Rain Climate Maps

Rain climate maps were the first approach to developing long-term rainfall rate statis-
tics that could be used for both propagation predictions and interference calculations.
These maps divide the world into regions where the average rainfall rate statistics are
the same, within a margin of about 10%. An example of a rain climate map is shown
in Figure 8.15 for the Americas19. The rainfall rate statistics for the 15 rain climate
165° 135° 105° 75° 45° 15°

75° 75° N
A A
E
G
C C E
G
60° 60°
B C
B
E F Latitude
45° D K F 45° in
Latitude
B E K
degrees N
H
M
30° E M
30°
N
E
N
15° 15°
P
N
0° 0° Equator
P
N E
15° N 15°
D C
E
30° 30°
K
E Latitude
D
45° 45° in
D degrees S
A A
60° 60° S
165° 135° 105° 75° 45° 15°
Longitude in degrees W
FIGURE 8.15 Rain climatic zones for the Americas (from Figure 1 of reference 19 © ITU,
reproduced with permission).
c08.qxd 12/08/02 09:55 Page 315
8.4 RAIN AND ICE EFFECTS 315
TABLE 8.2 Rainfall Rate Intensities for the Rain Climatic Zones (From TABLE 1 in [19] ©
ITU, reproduced with permission)
Percentage
of Time (%) A B C D E F G H J K L M N P Q
10 0.1 0.5 0.7 2.1 0.6 1.7 3 2 8 1.5 2 4 5 12 24

0.3 0.8 2 2.8 4.5 2.4 4.5 7 4 13 4.2 7 11 15 34 49
0.1 2 3 5 8 6 8 12 10 20 12 15 22 35 65 72
0.03 5 6 9 13 12 15 20 18 28 23 33 40 65 105 96
0.01 8 12 15 19 22 28 30 32 35 42 60 63 95 145 115
0.003 14 21 26 29 41 54 45 55 45 70 105 95 140 200 142
0.001 22 32 42 42 70 78 65 83 55 100 150 120 180 250 170
NOTE: See Figure 8.12 for the Rain Climatic Zones in N. and S. America
regions worldwide are given in Table 8.2. (Note that not all of these regions exist in the
Americas). The ease with which the tables and rain climate maps can be used is offset
by the clear inaccuracies that occur when large parts of the earth are given the same cli-
mate classification. The step-changes across the climate boundaries are also arbitrary
and are not supported by measured data. Wherever possible, it is always best to use
measured rainfall rate data as the attenuation prediction model input whenever these
data exist.
Rainfall Rate Exceedance Contour Maps

In an effort to overcome the inaccuracies of the rain climate maps, the ITU developed
a set of comprehensive rainfall rate exceedance curves for the whole world. An ini-
tial set for the Americas is shown in Figure 8.16. These maps are updated periodi-
cally and the latest set can be found on the ITU web site20. An example is shown in
Figure 8.1722.
Raindrop Distributions
Rain attenuation and depolarization occur because individual raindrops absorb energy from
radio waves. The drops absorb some of the incident energy and some is scattered. The size
and shape of raindrops have been measured2. The most common mathematical description
of the distribution of raindrop sizes is exponential and of the form
N1D2 N0e1DDm2 mm1 m3 (8.10)
where Dm is the median drop diameter and N(D) dD is the number of drops per cubic me-
ter with diameters between D and D dD mm. The rainfall rate R is related to N(D) and
also to the terminal velocity V(D) of the falling drops in meters per second with diameter
D by23
R 0.6 103p D V1D2N1D2dD mm/h

3
(8.11)
The details of scattering and absorption by a single raindrop, and the summation over the
drop population that the calculation of path attenuation from the drop size distribution
requires are beyond the scope of this text.
c08.qxd 12/08/02 09:55 Page 316
165° 150° 135° 120° 105° 90° 75° 60° 45° 30°
75° 75° N
10
60° 15 60°
15
10 20
30
45° 40 45°
50
60
70
80
30° 0
9 0 30°
1 0
Latitude
0
12 100
15° 0 15°
14 110
110
130
70 0
0° 0°
8
90
110
15° 15°
100
90
30° 80 30°
20 70
30 60
50
40
45° 45°
30
60° 60° S
165° 150° 135° 120° 105° 90° 75° 60° 45° 30°
Longitude in degrees W
FIGURE 8.16 Rainfall rate exceedance contours for the Americas (from reference 21 © ITU,
reproduced with permission). This was the first of a set of three rainfall rate exceedance con-
tours that were developed for the world. In this version, the contours only existed over land.
The latest versions20 include data over all of the surface of the earth (see Figure 8.17).
SIDEBAR
The first measurements of raindrop size distributions been absorbed by the flour, with dimensions that
were made by Laws and Parsons33 in 1944 with a corresponded to the raindrops that hit the tray in that
very ingenious experiment. The experiment was de- particular storm. From their flour measurements,
signed to find the sizes of raindrops in typical rain- Laws and Parsons derived an empirical exponential
storms. A baking pan (typically 1 m 0.5 m) was relationship between the rain rate and the drop size
filled with flour and placed out in the rain for a distribution that is still used today. The experiment
minute. The pan of flour was then baked in an oven, is repeated occasionally—by schoolchildren study-
and the loose flour sifted out. What remained were ing earth sciences—as a demonstration of rainfall
pellets of baked flour where raindrops had fallen and characteristics.
316
c08.qxd 12/08/02 09:57 Page 317
8.5 PREDICTION OF RAIN ATTENUATION 317
90°
80°
70°
60°
Latitude in degrees N
50°
40°
30°
20°
10°
120
0°
300° 320° 340° 0° 20° 40° 60° 80°
Longitude in degrees E
FIGURE 8.17 Rain intensity (mm/h) exceeded for 0.01% of the average year (Figure 2 of
reference 22). This map provides rainfall rate contours for the Northern Hemisphere between
longitudes 300° E and 80° E (Europe, North Africa, the Middle East, and parts of Russia,
India, and China).
8.5 PREDICTION OF RAIN ATTENUATION
Attenuation by rain can be predicted accurately if the rain can be precisely described all
the way along the path. Path attenuation is essentially an integral of all the individual in-
crements of rain attenuation caused by the drops encountered along the path. This is the
physical approach to predicting rain attenuation. Unfortunately, rain cannot be described
accurately along the path without extensive meteorological databases, which do not exist
in most regions of the world. Most prediction models therefore resort to semiempirical
approaches, which calculate an effective path length through the rain, Leff, over which the
rainfall rate is assumed to be constant. This constant rainfall rate leads to a constant specific
attenuation, R, and the path attenuation, A, is simply given by
A specific attenuation effective path length in rain
gRLeff dB (8.12)
The semiempirical approach is based on two premises: (1) Rainfall rate measured at a
point on the surface of the earth is statistically related (over a period of at least a year) to the
attenuation encountered along the path to a satellite from that same point; (2) The actual
c08.qxd 10/09/02 09:44 Page 318
length of the path through the rain medium can be adjusted such that an “effective” path length
is developed over which the rain can be considered to be homogeneous (see Figure 8.6).
The estimation of attenuation on the slant path to a satellite is essential to the process
of establishing a margin in the link budget that ensures the required availability of the link
is met. Over a period of many years, several attenuation models have been developed that
have been widely used. These include the Crane model35, the simple attenuation model
(SAM)36, the Dissanayake, Haidara, Allnutt (DAH) model4 and several models published
by the CCIR and ITU-R15,21,25. The ITU-R model, based on the DAH model15, is discussed
in detail here, because it provides the most accurate statistical estimate of attenuation on
slant paths, worldwide, at the time of writing. Appendix D discusses the simple attenua-
tion model (SAM), developed at Virginia Tech by Warren Stutzman and Keith Dishman.
The SAM model is less accurate than the ITU-R model, but allows the user to quickly
obtain an estimate of the slant-path attenuation at any frequency and rain rate.
A power law equation describes the relationship between point rainfall rate R and
specific attenuation, R, the attenuation measured over 1 km24
gR k1R0.01 2 a dB/km (8.13)
In Eq. (8.13), the suffix 0.01 to R denotes the rainfall rate measured for 0.01% of the
average year, a typical input time percentage for most models. Equation (8.13) holds
for all values of rainfall rate, however. The parameters k and are frequency dependent.
Table 8.3 gives values for k and for frequencies between 4 and 50 GHz25.
TABLE 8.3 Regression Coefficients for Estimating

Specific Attenuation (from TABLE 1 in [25] © ITU,
reproduced with permission)
Frequency
(GHz) kH kV H V
4 0.000650 0.000591 1.121 1.075

6 0.00175 0.00155 1.308 1.265
8 0.00454 0.00395 1.327 1.310
10 0.0101 0.00887 1.276 1.264
12 0.0188 0.0168 1.217 1.200
20 0.0751 0.0691 1.099 1.065
30 0.187 0.167 1.021 1.000
40 0.350 0.310 0.939 0.929
50 0.536 0.479 0.873 0.868
NOTES:
(1) The suffices V and H refer to vertical and horizontal polarization,
respectively
(2) Values of k and at frequencies other than those in the table can
be obtained by interpolation using a logarithmic scale for frequency,
a logarithmic scale for k, and a linear scale for .
(3) Values have been tested and found to be accurate up to a frequency
of 40 GHz; values between 40 and 50 GHz are expected to be accu-
rate but have not yet been tested.
(4) For linear and circular polarization, and for all path geometries, the
coefficients in equation (8.13) can be calculated using the values in
the above table and the following equations25
k 3 kH kV 1kH kV 2 cos2 u cos 2t4 2
a 3 kHaH kVaV 1 kHaH kVaV 2 cos2 u cos 2t4 2k
where is the path elevation angle and is the polarization tilt
angle relative to the horizontal ( 45° for circular polarization).
c08.qxd 12/08/02 09:57 Page 319
EXAMPLE 8.5.1 What is the specific attenuation at 10 GHz if the rainfall rate is 40 mm/h
and linear vertical polarization is used?
Answer From Table 8.3, kV 0.00887 and V 1.264 at a frequency of 10 GHz. Using
Eq. (8.13), we therefore have
Specific attenuation gR 0.00887 1402 1.264 0.9396 0.94 dB/km
If the rainfall rate were constant along the path, as it generally is in light, stratiform
rain (see Figure 8.5), then calculating the total attenuation for a given rainfall rate would
be simple. The physical path length through the rain, L, would be the same as the effective
path length and the total attenuation, A, is given by
A gR physical path length in rain gRL dB (8.14)
On short terrestrial paths (5 km, although this varies with rainfall rate: the lower the
rainfall rate, the longer the path), the path length through relatively constant rain can be taken
as the distance between the transmitting and receiving antennas. The path through the rain is
also at almost the same height along the whole path. This is not the case with satellite paths
where the signal follows a slanting path path through the atmosphere, and encounters rain of
different types and intensities on the way. Rain can take more than 10 min to fall from a
height of 5 km (the approximate upper limit of liquid water in a severe thunderstorm) to the
ground. If there are updrafts present, as is always the case in convective rain, it can take even
longer. There is therefore no instantaneous relationship between attenuation measured along
a path to a satellite and the rainfall rate measured at the earth station site. However, there is
a strong statistical relationship between the long-term cumulative statistics of rainfall rate and
the long-term statistics of slant-path attenuation. Many models of rain attenuation use
equiprobable values of rainfall rate and path attenuation to determine the cumulative statis-
tics of attenuation from those of rainfall rate. Figure 8.18 illustrates the procedure for find-
ing equiprobable values of rainfall rate and path attenuation.
The assumption that point rainfall rate on the ground is statistically related (over a pe-
riod of at least a year) to the attenuation observed on a satellite path to that same point has
been validated in many experiments worldwide. Since the path encounters highly variable
drop sizes and rainfall rates, the physical length L used in Eq. (8.14) has usually to be re-
placed by an effective path length Leff. We therefore find that the total path attenuation, A, for
a given satellite link is given by Eq. (8.12), which is repeated below for completeness
A gR effective path length in rain gR Leff dB (8.15)
The procedure by which the effective path length is calculated uses the statistical height
of rain (i.e., the melting level height), the height of the earth station above mean sea level,
and the elevation angle. See Figure 8.19 and earlier Figures 8.5 and 8.6.
In Figure 8.19, the rain is shown as filling the complete slant path up to the melt-
ing layer. This is a correct assumption in stratiform rain, which exists over large areas and
has a relatively constant rain rate along the slant path. It is rarely correct when convec-
tive rain is present. The rain rate and drop distributions are not constant, and the path may
not pass through the top of the rain cell. Figure 8.20 illustrates the problem.
The ITU-R procedure for predicting slant-path rain attenuation for GEO satellite paths
is contained in Section 2.2.1.1 of Rec. 61815. It uses a semiempirical approach to the pre-
diction of rain attenuation. Rather than attempt to predict attenuation by inputting rainfall
rate at every time percentage, it inputs only the rainfall rate measured (or predicted) for
0.01% of a year. It then extrapolates from this time percentage to other time percentages.
While this “one size fits all” approach is nonphysical, it removes the inherent inaccuracies
c08.qxd 12/08/02 09:57 Page 320
Percentage of time rain exceeds R (mm/h)
100
10 FIGURE 8.18 Cumulative sta-

tistics of rainfall rate and path at-
1 tenuation illustrating equiproba-
P ble procedures. For a given time
0.1 percentage, P, the rainfall rate is
read off the rainfall rate statistics
and the path attenuation is read
0.01
off the path attenuation statistics.
If the data for the two parameters
0.001 R (mm/h) have been taken over a long
0 R (P )
enough period (at least a year;
longer periods in multiples of
years), R(P) and A(P) are strongly
Percentage of time attenuation exceeds A (dB) related. Some models use the full
rainfall rate statistics to develop
100 path attenuation statistics. Others
use one time percentage to relate
10 the two statistics (e.g., the 0.01%
point) and develop the second
set of statistics from that single
1
point. The disadvantage of this
P
approach (i.e., it is nonphysical)
0.1 is outweighed by the improved
accuracy obtained by extrapolat-
0.01 ing to both low and high time
percentages, where the rainfall
0.001 A (dB) rate measurements are some-
0 A (P ) what suspect.
To satellite
He
El
H0
Earth Sea
station level
FIGURE 8.19 Geometry of a satellite path through rain. The height of the melting layer,
shown as He here, is normally considered to be the highest point at which rain attenuation
occurs. The rain fills the volume between the melting layer height and the ground. The
height of the earth station above mean sea level is given by H0.
c08.qxd 12/08/02 09:57 Page 321
(a ) ( b)
FIGURE 8.20 Example of different path length geometries. In both cases, a similar rainstorm
exists in the slant path. In case A, the path to the satellite exits through the side of the storm
cell while in case B it exits through the top (in a similar geometry as in Figure 8.5). The only
difference between the two paths is the elevation angle to the satellite. To develop an effective
path length to use in Eq. (8.12), use is made of both a “vertical adjustment factor” and a “hori-
zontal adjustment factor” to account for the possibility of either case A or case B occurring.
of using very low rainfall rates for time percentages of 0.1% (and higher) or very high rain-
fall rates for time percentages of 0.001%. The ITU-R is seeking to change this one size
fits all approach. The current procedure (early 2002) is reproduced below (the equation and
figure numbers have been changed to correspond with those in this chapter).
The following procedure provides estimates of the long-term statistics of the slant-
path rain attenuation at a given location for frequencies up to 55 GHz. The follow-
ing parameters are required:
R0.01: point rainfall rate for the location for 0.01% of an average year (mm/h)
hS: height above mean sea level of the earth station (km)
: elevation angle (degrees)
: latitude of the earth station (degrees)
f: frequency (GHz).
Re: effective radius of the earth (8500 km)
The geometry is illustrated in Figure 8.21.
A D
B
hR′
C
Ls
(hR′ − hs )
θ
hs
LG
A: frozen precipitation
B: rain height
C: liquid precipitation
D: Earth–space path
FIGURE 8.21 Schematic presentation of an earth–space path giving the parameters to be
input into the ITU-R rain attenuation prediction procedure (Figure 1 of reference 15 © ITU,
reproduced with permission).
c08.qxd 12/08/02 09:57 Page 322
Step 1: Calculate the rain height, hR , which is equivalent to h0 as given in

Recommendation ITU-R P.839.
Step 2: For 5° compute the slant-path length, LS, below the rain height from:
1hR¿ hS 2
LS km (8.16)
sin u
For 5°, the following formula is used:
21hR¿ hS 2
21hR¿ hS 2 12
LS km (8.17)
asin u 2
b sin u
Re
Step 3: Calculate the horizontal projection, LG , of the slant-path length from:
LG LS cos u km (8.18)
Step 4: Obtain the rainfall rate, R0.01, exceeded for 0.01% of an average year
(with an integration time of 1 min). If this long-term statistic cannot be
obtained from local data sources, an estimate can be obtained from the
maps of rainfall rate given in Recommendation ITU-R P.837.
Step 5: Obtain the specific attenuation, R, using the frequency-dependent
coefficients given in Recommendation ITU-R P.838 and the rainfall rate,
R0.01, determined from Step 4, by using:
gR k1R0.01 2 a dB/km (8.19)
Step 6: Calculate the horizontal reduction factor, r0.01, for 0.01% of the time:
1
r0.01 (8.20)
2
LGgR 2L G
1 0.78 0.3811 e
A f
Step 7: Calculate the vertical adjustment factor, v0.01, for 0.01% of the time:
hR¿ hS
z tan1a b deg
LGr0.01
LGr0.01
For z 7 u LR km
cos u
1hR¿ hS 2
Else LR km
sin u
If 0f 0 6 36 x 36 0f 0 deg
Else x 0 deg
1
v0.01
1LRgR
1 1sin 1u2 a31 11 e1u11x22 2 0.45b
f2
Step 8: The effective path length is
LE LRv0.01 km (8.21)
Step 9: The predicted attenuation exceeded for 0.01% of an average year is
obtained from:
A0.01 gRLE dB (8.22)
Step 10: The estimated attenuation to be exceeded for other percentages of an
average year, in the range 0.001% to 5%, is determined from the
attenuation to be exceeded for 0.01% for an average year:
If p 1% or 0f 0 36°: b0
If p 6 1% and 0f 0 6 36° and u 25°: b 0.005 1 0f 0 362
c08.qxd 12/08/02 09:57 Page 323
Otherwise: b 0.005 1 0f 0 362 1.8 4.25 sin 1u2

10.6550.033 ln 1 p2 0.045 ln 1A0.012 b 11p2 sin 1u22
Ap A0.01 a b
p
dB (8.23)
0.01
This method provides an estimate of the long-term statistics of attenuation due to
rain. When comparing measured statistics with the prediction, allowance should be
given for the rather large year-to-year variability in rainfall rate statistics (see Rec-
ommendation ITU-R P.678).
EXAMPLE 8.5.2
A Ku-band satellite is to be used in a video broadcasting system. The uplink will be from Miami,
Florida, where the studios of the company are located. Since the uplink will be used to feed more
than a million home receivers, the uplink availability must be 99.99% in the average year. The ques-
tion is therefore: what is the rain attenuation on the Miami uplink path for 0.01% of the average year?
The information on the link is as follows:

Polarization Vertical
Coefficients for calculating specific attenuation at 17.80 GHz kV 0.0510
V 1.0927
Rain climate regions for Miami Climate region M
Elevation angle 45°
Height of rain hR Assume 4 km in Miami
Height of Miami earth station site a.m.s.l. 0.05 km
(above mean sea level)
Answer
Step 1: We already know the rain height (given as 4 km).
Step 2: Find LS, the slant-path length below the rain height.
1hR¿ hS 2
LS , thus LS 5.5861 km
sin u
(Note: Keep all the significant figures at present.)
Step 3: Find LG, the horizontal projection of the slant-path length.
LG LS cos u, thus LG 3.95 km
Step 4: Find R0.01, the rainfall rate for 0.01% of an average year (mm/h).
From the Rain Climatic Zone information (Table 8.2) we have R0.01 63 mm/h.
Step 5: Find R, the specific attenuation, along the path for Miami for the rainfall rate encoun-
tered at 0.01% of an average year.
gR k1R0.01 2 a, giving gR 4.7175 dB/km
Step 6: Find r0.01, the horizontal reduction factor for Miami.
1
r0.01
0.3811 e2LG 2
LGgR
1 0.78
A f
Thus r0.01 0.7051 for Miami.
Step 7: Calculate the vertical adjustment factor, v0.01, for Miami.
To do this we need some intermediate parameters.
c08.qxd 12/08/02 09:57 Page 324
Part (a): Calculate

, where
hR¿ hs
z tan1 a b
LGr0.01
Thus,
54.81230 for Miami. This is greater than the elevation angle, , which
is 45°.
Part (b): Find LR, an intermediate parameter in calculating the effective path length.
Since
,
LGr0.01
LR
cos u
giving LR 3.9388 km.
Part (c): Find , the second intermediate parameter for calculating effective path length.
x 36 0f 0
where is the latitude of the site. Thus 36 25 11.0 for Miami.
Finally, calculate v0.01, from
1
v0.01
1LRgR
1 1sin1u2 a31 a1 e1u11x22 b 0.45b
f2
This gives v0.01 1.0332 for Miami.
Step 8: Calculate LE, the effective path length for Miami.
LE LRv0.01
which gives LE 4.0696 for Miami.
Step 9: Calculate A0.01, the predicted attenuation exceeded for 0.01% of an average year along the
path in Miami.
A0.01 gRLE
and this gives
A0.01 19.1983 dB for Miami 1 19.2 dB
The rain attenuation on the uplink from Miami for 0.01% of the average year will be 19.2 dB, which
is the answer to the question posed. This value, however, pertains to a fixed link that does not change
significantly with time. Such a situation would not apply to non geostationary orbit (NGSO) satel-
lite systems. A double-probabilistic approach is required for estimating the statistical impact of rain
attenuation on NGSO paths: the probability that attenuation will occur for a given elevation angle
and the probability that the satellite will be at that elevation angle. The first approach is documented
in ITU-R Rec. 61815 and is abstracted below.
Calculation of Long-Term Statistics for

NGSO Systems
For non-GSO systems, where the elevation angle is varying, the link availability for
a single satellite can be calculated in the following way:
a) calculate the minimum and maximum elevation angles at which the system will
be expected to operate;
b) divide the operational range of angles into small increments (e.g., 5° wide);
c) calculate the percentage of time that the satellite is visible as a function of
elevation angle in each increment);
c08.qxd 12/08/02 09:57 Page 325
d) for a given propagation impairment level, find the time percentage that the level
is exceeded for each elevation angle increment;
e) for each elevation angle increment, multiply the results of c) and d) and divide by
100, giving the time percentage that the impairment level is exceeded at this
elevation angle;
f) sum the time percentage values obtained in e) to arrive at the total system time
percentage that the impairment level is exceeded.
In the case of multi-visibility satellite constellations employing satellite path
diversity (i.e., switching to the least impaired path), an approximate calculation
can be made assuming that the spacecraft with the highest elevation angle is
being used.
Scaling Attenuation with Elevation Angle

and Frequency
Experience has shown that, if long-term attenuation data already exist at a site, it is more
accurate to scale measured results to another frequency or another elevation angle, instead
of predicting the path attenuation at the new frequency and/or elevation angle from rainfall
rate data. Two fairly simple (and surprisingly accurate) rules of thumb exist for scaling
over small changes in frequency and elevation angle:
(i) For a uniform rainfall rate environment (i.e., stratiform rain) and assuming a “flat
earth,” path attenuation in decibels scales with the path length through the rain (i.e.,
it follows a cosecant law);
(ii) Between about 10 and 50 GHz, attenuation in decibels scales as the square of the
frequency. These two laws are expanded below.
Cosecant Law
The attenuation in decibels at the same frequency at elevation angles El1 and El2 from the
same site are approximately related by
A1El1 2 cosecant1El1 2

A1El2 2 cosecant1El2 2
(8.24)
This formula breaks down when the elevation angle is low (10°) where its implicit flat
earth and uniform rainfall rate assumptions fail to hold.
EXAMPLE 8.5.3
A 12-GHz direct broadcast satellite link was found to experience 4 dB of rain attenuation at
an elevation angle of 45° for 0.01% of the time in an average year. What would be the rain
attenuation measured at the same time percentage for the same site if the elevation angle
were 10°?
Answer Let suffix 1 in Eq. (8.24) refer to the new elevation angle (i.e., 10°) and suffix 2 to the
old elevation angle. Thus,
A110°2 3cosecant110°2 cosecant145°2 4 A145°2
3 5.75881.4142 4 4 dB
16.2883 1 16.3 dB
The impact of elevation angle on a given link is clear from this example.
c08.qxd 12/08/02 09:57 Page 326
Squared Frequency Scaling Law

If A( f1) and A( f2) are the attenuations that would be measured on the same path at
frequencies f1 and f2 GHz, they are approximately related by
A1 f1 2 1 f1 2 2

A1 f2 2 1 f2 2 2
(8.25)
This formula relates the long-term statistics (i.e., annual statistics). It should not be used
for short-term frequency scaling (i.e., from second to second) on a link or for frequencies
that are close to any resonant absorption line.
EXAMPLE 8.5.4
A user measures rain attenuation statistics along a satellite link as 6 dB for 0.01% of a year when
using a carrier frequency of 10.7 GHz. The satellite operator wants to move the user from the cur-
rent transponder to a new one, which would change the carrier frequency to 11.4 GHz. What would
be the new rain attenuation value, all other link parameters remaining the same?
Answer Let suffix 1 in Eq. (8.25) refer to the new frequency (i.e., 11.4 GHz) and suffix 2 refer
to the old frequency (i.e., 10.7 GHz). Thus,
A111.42 3 111.42 2 110.72 2 4 A110.72
3 129.9600114.4900 4 6 dB
6.8107 1 6.8 dB
A more accurate form of frequency scaling can be found in ITU-R Rec. 61815, and is summarized
below.
ITU-R Long-Term Frequency Scaling

of Rain Attenuation
If A1 and A2 are the equiprobable values of rain attenuation, in dB, at frequencies f1 and
f2 (GHz), respectively, the attenuation at frequency f2 can be found from that at frequency
f1 from
A2 A1 1f2 f1 2 1H 1f1,f2, A12 (8.26)
where
f2
f1 f 2 4 (8.27)
1 10 f 2
H1f1,f2, A1 2 1.12 10 3 1f2 f1 2 0.5 1f1A1 2 0.55 (8.28)
8.6 PREDICTION OF XPD
Any particle that has spherical symmetry will cause no depolarization of an incident sig-
nal. Rain in the atmosphere starts as very small droplets. The surface tension within these
droplets is so strong that they retain their spherical shapes. As the drops collide, they co-
alesce into larger drops. The larger the drop, the more likely it is to distort out of a spher-
ical shape due to wind effects. In convective events, particularly severe thunderstorms, the
drops can become relatively large (many millimeters in average diameter) and so they will
distort into ellipsoidal forms, generally flattening out in the horizontal axis. Figure 8.22
illustrates the process.
c08.qxd 12/08/02 09:57 Page 327
8.6 PREDICTION OF XPD 327
Very small droplet: wind effects cannot overcome surface tension

and droplet remains perfectly spherical in shape.
Many droplets have collided and coalesced into larger drop. As drop
falls, pressure on underside overcomes surface tension forces and
drop begins to distort from spherical shape, taking on elliptical profile
in vertical axis. Viewed from below, drop is still essentially circular in
cross section.
Raindrop has grown in size through further collisions with other

raindrops. Larger mass causes raindrop to fall at higher velocity and
increased wind forces on underside cause drop to become strongly
ellipsoidal in vertical profile. This symmetrical, ellipsoid shape will
only exist in very still air and, even then, vertical motion will cause
drop to start to oscillate, forming oblate and prolate spheroids
alternately.
Large, oscillating drops will distort into shapes with no true axes of
symmetry. Drops will hollow out underneath due to drop’s fall and,
since such large raindrops generally form as result of severe
convective activity, turbulent air motion will cause large, oscillating
drops to break up. Smaller drops resulting from this breakup then
cause additional raindrop formation to occur.
FIGURE 8.22 Schematic of the shape of an individual raindrop from formation to maturity.
If all of the ellipsoidal drops in a rainstorm were aligned, then waves propagating
with their electrical field vectors parallel to the raindrops’ minor axes (for all practical
purposes, vertically polarized waves) would experience the minimum attenuation for that
rainfall rate, and waves propagating with their electric field vectors parallel to the major
axes (i.e., horizontally polarized waves) would experience the maximum attenuation. In
these two special cases, no depolarization would occur. The difference between the at-
tenuations experienced by waves with horizontal and vertical polarization is small—rarely
greater than a decibel. It is called the differential attenuation. In a like manner, waves with
horizontal and vertical polarization can experience differential phase shift as they pass
through an anisotropic medium. At frequencies below about 10 GHz, differential phase
shift is the more important phenomenon. At frequencies above about 30 GHz, differential
attenuation is more important. Between 10 and 30 GHz, either differential phase or dif-
ferential attenuation will be the major effect, depending on the elevation angle of the link
and the climate31.
Imagine now the case of a wave whose linear polarization is intermediate between
horizontal and vertical. We can resolve this wave into its vertically polarized and hori-
zontally polarized components as in Figure 8.23. These components propagate through
the rain with their polarizations unchanged, but the horizontal component is attenuated
more than the vertical component. If at any point we recombine the vertical and horizontal
components to reconstruct the wave, we find that its polarization has rotated toward the
vertical and a cross-polarized component is now present. This process is a simplification
of a complicated problem in electromagnetic wave scattering. For details of the process,
the reader should consult the extensive publications of T. Oguchi, the pioneer researcher
in the field (e.g., [26]).
Depolarization, while it is dependent to a great extent on the volume of rain that
is present in the path, the shape of the raindrops in the path and the orientation of their
major and minor axes also significantly affect it. The orientation will have two independent
features: one that is due to the rain medium, and is referred to as the canting angle; and
one that is due to the path geometry, and is referred to as the tilt angle.
c08.qxd 12/08/02 09:57 Page 328
Vertical
E iV Ei
Er
E rV
θ
Horizontal
E rH E iH
FIGURE 8.23 A simplified explanation of rain depolarization based on a drop with an ellip-
tical cross section. An incident electromagnetic wave with electric field vector Ei strikes a
raindrop. We resolve it into a horizontal component E iH and a vertical component E iV. The
horizontal component is attenuated more than the vertical component because it encounters
more water. Thus, when we recombine the horizontal and vertical field components E rH and
E rV that arrive at the receiver, we find that the received wave Er has had its polarization
rotated toward the vertical by the angle .
Canting Angle
Falling raindrops orient themselves so as to minimize the aerodynamic forces. In steady
fall, the minor axis of the drop is parallel to the net wind force and so their major axis is
horizontal when the raindrop is falling in still air. Under windy conditions, the aerody-
namic force will have two components: one due to the raindrop fall velocity (i.e., verti-
cal) and one due to the prevailing wind direction (i.e., horizontal). The resultant of these
two forces will lead to the raindrops’ major axis being canted out of the usual horizontal
orientation. The prevailing wind speed lessens with altitude, becoming zero at the ground.
The raindrop orientation will therefore vary with altitude. Since the horizontal wind di-
rection with respect to the path varies, the net horizontal component measured over a long
interval will be close to zero. The canting angle will therefore have a mean of zero. In
any given rainstorm, however, the canting angle will have a finite probability of being
nonzero, thus leading to enhanced depolarization for horizontal or vertical polarized waves
over short time intervals. Figure 8.24 illustrates the canting angle process schematically.
Tilt Angle
The tilt angle refers to the angle between the local horizontal (or vertical) and the actual
orientation of the electric field vector of the transmitted signal. The orientation of the elec-
tric field vector transmitted by a geostationary satellite is referenced to the equator at the
subsatellite point. Horizontal polarization is parallel to the equator and vertical polariza-
tion is perpendicular to the equator. An earth station that lies on the same longitude as the
GEO satellite (say, to the north) would receive signals polarized in the local vertical di-
rection if the satellite is transmitting a vertically polarized signal. If the location of the
earth station is moved either east or west from the longitude of the GEO satellite, the
vertically polarized signal transmitted by the satellite is now received out of the local ver-
tical at the earth station. That is, the polarization vector would appear to be tilted away
from the original orientation. The process is illustrated in Figure 8.25.
c08.qxd 12/08/02 09:57 Page 329
Canting
angle Raindrop
Local horizontal
Vertical wind
force due to Net
drop velocity wind force
Horizontal wind force due

to prevailing wind
FIGURE 8.24 Illustration of canting angle. The resultant of the prevailing wind force and the
wind force due to the raindrop fall velocity leads to a net wind force that is out of the vertical di-
rection. The raindrop, which has already distorted into an ellipsoid due to the wind force induced
by the drop velocity, will now orient itself to minimize drag forces. This means that the raindrop
will cant out of the horizontal and orient its minor axis to be parallel to the net wind force.
N
Earth as viewed
from GEO
S Equator
(a )
Orientation of local GEO arc as viewed from

electric field vector of earth station
vertically polarized
transmissions from
GEO satellite
Local horizontal
(b )
FIGURE 8.25 Schematic of tilt angle. In (a) above, S, is the subsatellite point of a GEO
satellite. Transmissions from the satellite will be horizontally polarized if they are parallel to
the equatorial plane. Vertically polarized transmissions will be orthogonal to the equatorial
plane. If an earth station were on the satellite longitude (here shown by the broken line SN)
it would receive the polarization vector in the orientation transmitted—although the polariza-
tion would be undefined at the subsatellite point. In (b) above, the earth station is not on the
equator. The arc shows how the GEO orbit would look from the earth station. In this in-
stance, the satellite is transmitting a vertically polarized signal. The orientation of the verti-
cally polarized transmissions may not be received at the local vertical, however. The local
orientation will depend on where the satellite is located on the GEO arc as seen by the earth
station. The polarization vector may therefore be tilted out of the transmitted orientation by
virtue of the link geometry. The polarization will only be vertical (or horizontal) at the earth
station site to a GEO satellite if the azimuth to the satellite is 0° or 180° from true north.
329
c08.qxd 12/08/02 09:57 Page 330
A simple equation that gives the tilt angle with respect to the horizontal, assuming
the transmissions from a GEO satellite are polarized in the north–south direction, is27
t arctan1tan Lesin b2 degrees (8.29)
where Le is the earth station latitude (positive for the Northern Hemisphere and negative
for the Southern Hemisphere) and is the satellite longitude minus the earth station lon-
gitude (i.e., L s Le), with longitude expressed in degrees east.
EXAMPLE 8.6.1
What is the perceived polarization tilt angle at an earth station located at 52° N, 1° E, for vertically
polarized signals transmitted from a GEO satellite located at 60° E?
Answer: Using Eq. (8.29)

t arctan1tan 52 sin360 14 2 arctan11.27990.85722 arctan11.49322 56.19°
The ITU-R XPD prediction method15 is based on the attenuation measured (or pre-
dicted) at the frequency of interest, plus additional terms to take account of the
canting angle distribution, the tilt angle, and ice crystal depolarization28. The step-
by-step procedure is summarized below.
Step 1: Calculate the frequency-dependent term:
Cf 30 log f for 8 f 35 GHz (8.30)
where f is the frequency in GHz.
Step 2: Calculate the rain attenuation dependent term:
CA V1 f 2 log Ap (8.31a)
where Ap is the rain attenuation in decibels exceeded for the required per-
centage of time, p, for the path in question, commonly called the copolar at-
tenuation or CPA;
V1 f 2 12.8f 0.19 for 8 f 20 GHz (8.31b)
V1 f 2 22.6 for 20 6 f 35 GHz (8.31c)
Step 3: Calculate the polarization improvement factor:
Ct 10 log31 0.48411 cos 4t2 4 (8.32)
where is the tilt angle. The improvement factor C 0 for 45° and
reaches a maximum value of 15 dB for 0° or 90°. The value 45°
corresponds to circular polarization.
Step 4: Calculate the canting angle dependent term:
Cu 40 log1cos u2 for u 60° (8.33)
where is the elevation angle of the link.
Step 5: Calculate the canting angle dependent term:
Cs 0.0052 s2 (8.34)
where is the effective standard deviation of the raindrop canting angle
distribution, expressed in degrees. The value of is 0°, 5°, 10°, and 15°
for 1, 0.1, 0.01, and 0.001% of the time, respectively.
Step 6: Calculate rain XPD not exceeded for p% of the time:
XPDrain Cf CA Ct Cu Cs dB (8.35)
c08.qxd 12/08/02 09:57 Page 331
8.6 PREDICTION OF XPD 331
Step 7: Calculate the ice crystal dependent term:

Cice XPDrain 10.3 0.1 log p2 2 dB (8.36)
Step 8: Calculate the XPD not exceed for p% of the time, including the effects of ice
crystals:
XPDp XPDrain Cice dB (8.37)
The rain attenuation below 8 GHz is fairly low and so the attenuation-dependent
XPD prediction method does not provide accurate results. To calculate XPD for fre-
quencies below 8 GHz, it is best to calculate the XPD at 8 GHz and then scale in
frequency to the required frequency using15
f2 11 0.48411 cos 4t2 2
XPD2 XPD1 20 log c d for 4 f1, f2 30 GHz (8.38)
f1 11 0.48411 cos 4t1 2
Unpublished results from the Italsat experiment29 appear to show that it is possible to
predict XPD from 35 GHz up to 50 GHz by amending Eq. (8.30) and changing the val-
ues of V( f ) in Eq. (8.31a) to
Cf 26 log f (8.39a)
V1 f 2 20 (8.39b)
EXAMPLE 8.6.2
What is the value of XPD at 0.01% of the time for a 12-GHz link that experiences 7-dB attenuation
for this period of time? The elevation angle is 30°. Calculate the XPD for tilt angles of 20° and 0°.
Answer Using the step-by-step procedure we have:

Step 1: Cf 30 log f 32.3754
Step 2: V( f ) 12.8 f 0.19 12.8 1.6034 20.5236
CA V( f ) log Ap 20.5236 log (7) 17.3445
Step 3: Tilt angle of 20°
C 10 log [1 0.484 (1 cos 4 )] 10 log[1 0.484(1 cos 80)] 3.6456
Tilt angle of 0°
C 10 log[1 0.484 (1 cos 4 )] 10 log[1 0.484(1 cos 0)] 14.9485
Step 4: C 40 log(cos ) 40 log(cos 30) 40 log(0.8660) 2.4988
Step 5: C 0.0052 2 0.0052 102 0.52
Step 6: XPDrain Cf CA C C C
For 20° 32.3754 17.3445 3.6456 2.4988 0.52 21.6953 21.7 dB
For 0° 32.3754 17.3445 14.9485 2.4988 0.52 32.9982 33.0 dB
Step 7: Cice XPDrain (0.3 0.1 log p)2
For 20° 21.6953 (0.3 0.1 log p)2 21.6953 (0.3 0.1 log 0.01)2
21.6953 (0.3 0.2)2 1.0848
For 0° 32.9982 (0.3 0.1 log p)2 32.9982 (0.3 0.1 log 0.01)2
32.9982 (0.3 0.2)2 1.6499
Step 8: XPDp XPDrain Cice
For 20° 21.6953 1.0848 20.6105 20.6 dB
For 0° 32.9982 1.6499 31.3483 31.3 dB
Note: The single, best way to reduce depolarization is to operate with polarization senses that are
linear vertical or horizontal as perceived by the receiving antenna. This can be seen from the very
different results calculated in the above example when the tilt angle was 0° (i.e., the signal is being
received in linear, horizontal polarization) compared with those when the tilt angle is 20°.
c08.qxd 12/08/02 09:57 Page 332
Ice Crystal Depolarization

The calculation procedure for ice crystal depolarization incorporated in the calculation of
XPD has been found to have wide variations in accuracy. At high elevation angles and at
frequencies below 10 GHz, the procedure tends to agree with measured data. That is, ice
crystal depolarization occurs only in severe thunderstorms and so it is a rare occurrence.
However, on low elevation angle paths, the contribution due to ice crystals has been ob-
served to occur for quite high time percentages. At frequencies above 30 GHz, it is ex-
pected that ice crystal depolarization will be a significant effect, particularly at elevation
angles below 30°.
Rain Effects on Antenna Noise

At frequencies below about 50 GHz, rain attenuation is mostly caused by absorption rather
than by scattering of the signal energy out of the path. Any absorber with a physical tem-
perature greater than absolute zero (0 kelvins) will act as a black body radiator. At fre-
quencies below 300 GHz, the radiation is in the form of white Gaussian noise with a noise
power given by kTB, where T is the equivalent noise temperature of the absorber. Rain-
drops are absorbers at microwave frequencies and, when the raindrops fall through the
antenna beam, some of their isotropically radiated thermal energy will be detected by the
receiver (see Chapter 4). Rain will therefore cause not only signal attenuation and depo-
larization; it will also cause an increase in sky temperature, which, in turn, will increase
the overall system noise temperature. The impact of the increase in sky noise temperature
can be high for low noise receiving systems at Ku band, as illustrated in the examples in
Absorbing medium (e.g., a

rain cloud) with effective
temperature Tm kelvins
Temperature radiated
by medium, Tr , is
Tr = (1 − σ )Tm kelvins
FIGURE 8.26 Schematic of the additional radiated sky temperature due to absorption in
rain. The added temperature received by the antenna due to radiation from the “hot” rain-
storm will cause an additional component to be added to the system noise temperature. This
additional component is similar to the noise temperature contribution from a lossy feed. In
Chapter 4, in the analysis of system noise temperature, a noise temperature contribution due
to signal loss, Tl, was calculated using a “gain” component Gl, where Gl was a linear value.
For example, when the component at a physical temperature of 280 K caused a loss of 2 dB
(which 1/1.58 0.63 of the original value), Tl Tp(1 Gl) 280(1 0.63) 103.6 K. The
parameter Gl is identical to , i.e. a loss of 2 dB is the same as a fractional transmission of
0.63 of the original signal.
c08.qxd 12/08/02 09:57 Page 333
8.7 PROPAGATION IMPAIRMENT COUNTERMEASURES 333
Chapter 4. Rain attenuation in the 1 to 3 dB range can cause the system noise level to
increase by 1 to 3 dB, leading to a reduction in CN ratio (in dB) in rain, which is twice
the rain attenuation value.
The increase in antenna noise temperature due to rain, Tb, may be estimated by
Tb 280 11 eA4.34 2 K (8.40)
where A is the rain attenuation in decibels and the value 280 K is an effective tempera-
ture of the rain medium in kelvins. Values between 273 and 290 K may be used, depending
on whether the climate is cold or tropical.
An alternative approach is to treat the rain as a passive attenuator with a fractional
transmission coefficient of . If the rain totally attenuates the signal, 0; if the rain
medium is completely transparent and no attenuation takes place, 1. Figure 8.26
illustrates the process.
EXAMPLE 8.6.3
What is the additional noise temperature contribution of an antenna compared with that in clear sky
when there is 4 dB of rain attenuation in the path? You may assume that the rain medium is at a
temperature equivalent to 285 K.
Answer An attenuation of 4 dB causes the signal to be reduced by a factor of 2.5119. The

fractional transmission coefficient, , would therefore be 12.5119 0.3981. (Another way of look-
ing at this is to say that only 39.81% of the original signal power is being received during the 4-dB
rain event.) The additional sky temperature radiated would therefore be 285 (1 0.3981) 171.5395
K 171.5 K. Note that, if the system noise temperature had been 200 K, the effective system noise
temperature is now 200 171.5 371.5 K. In other words, the signal power has decreased by
4 dB and the noise power has increase by 2.7 dB. A 4-dB rain attenuation has thus led to a 6.7 dB
reduction in CN. This is somewhat simplistic, since the receiving antenna efficiency is not 100%,
and it therefore does not accept all of the radiation that is incident upon it. However, the enhanced
sky noise contribution received by the antenna during rain conditions will be close to that radiated
by the rainstorm. Careful attention must be paid in the system design to allow for enhanced sky
noise contributions as well as signal degradations when developing link budgets. Put another way,
the key in link budget calculations is to find the change in carrier-to-noise, CN, rather than just
the change in carrier power, C.
8.7 PROPAGATION IMPAIRMENT

COUNTERMEASURES
Attenuation
Many research groups have investigated the use of fade countermeasures. Fade mitigation
has been shown30 to fall into three main classes:
• Power control (i.e., varying the EIRP of the signal to enhance CN)
• Signal processing (i.e., changing the parameters of a signal to improve BER)
• Diversity (i.e., choosing a different path or time to take advantage of decorrelated
fading)
Interestingly, the three main classes of fade mitigation affect a link differently and are
complementary in nature30. For satellite systems that use frequencies at Ka band and above,
c08.qxd 12/08/02 09:57 Page 334
all three classes of fade mitigation techniques might be required for high availability links.
We will look briefly at each technique.
Power Control
In adaptive power control, the transmitter power is adjusted to compensate for changes in
signal attenuation along the path. At its simplest, it is like automatic gain control in a re-
ceiver, which adjusts to fluctuations in the received energy so as to hold the receiver out-
put constant. Many satellite links are operated such that the uplink is the critical portion
of the connection; that is, the first part of the overall connection that will drop out in a
rain fade is the uplink. The overall availability (and performance) of the connection is
therefore enhanced if the uplink operates with an increased EIRP in rain. This is referred
to as up-link power control (ULPC).
ULPC can operate closed loop, where the signal power is detected at the satellite
and a control signal sent back to the earth station to adjust the power, or open loop,
where the fade on the downlink signal is used to predict the likely fade level occurring
on the uplink. Closed-loop operation is always more accurate but is more expensive to
implement; hence most ULPC systems are, at present, open loop.
Open loop ULPC becomes more difficult the further apart the downlink and uplink
frequencies are. It becomes even more difficult at Ka band when the downlink (20 GHz)
and uplink (30 GHz) frequencies are on either side of the 22-GHz water vapor absorption
line. The ratio of 30-GHz attenuation to 20-GHz attenuation is less than 1 for 20-GHz
attenuation values of less than 1.0 dB, since cloud attenuation (i.e., essentially water vapor
absorption) is higher at 20 GHz than at 30 GHz due to the proximity of the 22-GHz water
absorption line to the 20-GHz downlink. Figure 8.27 gives the average 30 :20 GHz
attenuation ratios, with uplink attenuation as parameter. Note that the long-term 30 :20 GHz
4
(30/20) GHz attenuation ratio
2 6 10 14 18 22
30 GHz attenuation, dB
FIGURE 8.27 Instantaneous 30:20 GHz attenuation scaling ratio with 20-GHz attenuation as
parameter. The solid curve above is a prediction of the scaling ratio that takes into account
both rain attenuation and tropospheric scintillation. The pair of broken curves are the bounds
of individual instantaneous measurements of the uplink and downlink attenuation values.
The large range of scaling ratios shows that great care must be taken in developing open
loop ULPC algorithms that use only a measure of the amplitude of the downlink frequency.
c08.qxd 12/08/02 09:57 Page 335
attenuation scaling ratio does not become established until the uplink attenuation is above
7 dB. Another major consideration is power flux density variations at the satellite. If many
earth stations are operating under rain fade conditions with the same satellite, as could
happen in a VSAT network with many hundreds of thousands of earth stations,
implementing ULPC can lead to significant received power fluctuations at the satellite,
and this has capacity implications. Some of the advanced Ka-band satellites with multi-
ple switched beams can also implement downlink power control, if sufficient bandwidth
and power are available.
Signal Processing
The move from very large earth stations (e.g., the Intelsat Standard A) to a multiplicity
of small earth stations has been accompanied by a shift in the median traffic stream. It is
rare to find a non-video or non-Internet network distribution link via a satellite at a rate
of more than 2 Mbit/s. The need to make small traffic streams economic by using VSATs
has led to the introduction of onboard processing (OBP) techniques. This process typi-
cally translates the digital carriers arriving at the satellite to baseband for processing and
onward transmission back to earth. The process is generically called MCDDD (multi-
carrier demodulation, demultiplexing, and decoding). The OBP process is carried out at
baseband and allows each individual traffic packet to be switched to the correct output
port of the satellite antenna for transmission down to earth following recoding and re-
multiplexing. By detecting the signal level of each packet on arrival at the OBP, not only
can most bit errors be removed but the transmitting earth station can also be alerted if the
energy level of the received packet has fallen, so that ULPC can be used at the earth sta-
tion to correct the signal level (within the power level range of the ULPC system). The
use of OBP separates the uplink from the downlink and each part of the link can be treated
separately in developing a link budget.
Diversity
Many diversity schemes have been proposed, but few have been implemented as yet due
to the cost. If OBP techniques are being used on the satellite, a form of time diversity can
be used. In this approach, additional slots in the TDMA frame can be assigned to the rain-
affected link so that the same signal can be sent at a slower rate, essentially lowering the
bandwidth and raising the CN. The FEC rate could also be changed in the OBP payload.
If the satellite operates in a number of frequency bands (e.g., C band and Ku band), a rain
affected Ku-band link could be switched to C band, which is not attenuated significantly
by rain. To be able to do this, spare C-band capacity must be held in reserve on the satel-
lite so that it can be used when required. Similarly, each Ku-band earth station would need
to have a dual-band antenna and receiving system so that they could switch between the
two bands. The added cost has not justified this approach to date. However, the V-band
systems in design at present may find it economic to include a low capacity Ka-band or
Ku-band payload to use in those traffic streams that are the highest priority. Of all the
diversity schemes, that of site diversity appears to offer the most significant gain in
availability.
Site diversity is a technique whereby two, or more, earth stations are located suffi-
ciently far apart to ensure that the rain impairments observed at each of the stations are
generally uncorrelated. More exactly, it is the paths through the rain that are uncorrelated
and so the technique is more accurately described as path diversity. The earth stations are
c08.qxd 12/08/02 09:57 Page 336
connected together so that any one earth station can be used to support the traffic stream
while the other(s) is (are) suffering a rain fade.
If we assume that there are two earth stations, identified by suffixes 1 and 2, which
are operated in a site diversity mode, then the joint attenuation AJ(t) is defined by
AJ 1t2 minimum3A1 1t2, A2 1t2 4 dB (8.41)
The average single-site attenuation AS(t) is the mean of A1(t) and A2(t), namely
AS 1t2 3A1 1t2 A2 1t2 4 2 dB (8.42)
An ideal system that monitors the received downlink signals at both sites and always se-
lects the stronger of the two experiences an attenuation of AJ(t) and the diversity system
would perform better than either site alone. How much better is measured by two statistical
quantities, diversity gain and diversity improvement.
Diversity gain, GD(P), is the decibel difference between the average single-site at-
tenuation AS(P) equaled or exceeded P% of the time and the joint attenuation AJ(P) equaled
or exceeded P% of the time.
GD 1P2 AS 1P2 AJ 1P2 dB (8.43)
Diversity improvement ID(A) is the ratio between the percentage of time PS that the av-
erage single-site attenuation AS exceeds A dB to the percentage of time PJ that the joint
attenuation AJ exceeds A dB.
PS 1A2
ID 1A2
PJ 1A2
(8.44)
Diversity gain determines system margin, and it is the measure of diversity system
performance that we will use here. In addition, diversity gain has been shown to be sta-
ble from year to year and, as such, is a reliable statistic to use in system design. Diver-
sity improvement, on the other hand, is extremely variable from year to year. Figure 8.28
illustrates these two concepts.
The first, and still the best, diversity gain model is that due to Hodge32, which has
been adapted by the ITU-R in Rec. 61815. Hodge developed the diversity gain model through
an iterative analysis of diversity data available. Intuitively, he assumed site separation was
the key element. In this, he has been proved correct. The procedure is abstracted below.
Step 1: Calculate the gain contributed by the spatial separation of the two earth stations
from
Gd a11 ebd 2 (8.45)
where d is the separation (km) between the two sites,
a 0.78A 1.94 11 e0.11A 2
b 0.5911 e0.1A 2
and A path attenuation (dB) for a single site.
Step 2: Calculate the frequency-dependent gain from:
Gf e0.025 f (8.46)
where f frequency (GHz).
Step 3: Calculate the gain term dependent on elevation angle from:
G u 1 0.006 u (8.47)
where path elevation angle (degrees).
c08.qxd 12/08/02 09:57 Page 337
Percentage of time attenuation exceeds A
100
P S (A )
ID(A) =
P J (A )
P S (A )
P J (A )
10
G D (P ) = A S (P ) − A J (P )
1.0 G D (P )
P
Average single-site
Joint
0.1 A (dB)
A
A J (P ) AS(P )
FIGURE 8.28 Illustration of diversity gain and diversity improvement (diversity advantage).
At a given percentage of time, P, the diversity gain GD(P) is the decibel difference between
the average single-site attenuation exceeded AS(P) and the joint attenuation exceeded AJ(P).
At a given attenuation, A, the diversity improvement ID(A) is the ratio of the percentage of
time PS(A) that the single-site attenuation exceeds A to the percentage of time PJ(A) that the
joint attenuation exceeds A.
Step 4: Calculate the baseline-dependent term from the expression:

Gc 1 0.002 c (8.48)
where angle (degrees) made by the azimuth of the propagation path with
respect to the baseline between sites, chosen such that 90°.
Step 5: Compute the net diversity gain, G, as the product:
G Gd Gf Gu Gc dB (8.49)
The use of a site diversity system is very expensive if traditional approaches are used.
That is, two large earth stations connected together via a very high-speed terrestrial link.
It has only been used operationally to date by the gateway stations of the Iridium net-
work. These gateway earth stations operate in Ka band and are single-point failures for
the network. As such, the expense of a diversity setup was well justified. Another proposed
approach to site diversity has been to use wide area diversity33, in which a multitude of
VSATs are linked via routers to a metropolitan area network.
Depolarization
Depolarization compensation is a technique whereby the feed system of the antenna is
adjusted in such a way as to correct for depolarization in the path. Alternatively, the
c08.qxd 12/08/02 09:57 Page 338
orthogonal channels may be cross-coupled in the receiver and, provided good samples
of the signal in each channel can be obtained, the interfering (i.e., depolarized) signal
may be removed by subtracting the correct amount of signal. Few earth stations have
implemented depolarization compensation, as it is an expensive undertaking. Those
earth stations that have implemented depolarization compensation have done so at
C band.
At C band, differential phase is the primary cause of depolarization. As the fre-
quency increases, differential attenuation becomes an increasingly significant cause of de-
polarization until, at V band, differential attenuation dominates completely. For this rea-
son, the amount of rain depolarization observed for each dB of rain attenuation on
commercial communications satellite links is largest at C band, reducing monotonically
to V band. Most Ka-band systems for direct-to-home (DTH) Internet services have rain
margins of less than 10 dB. At this attenuation level, depolarization effects are not
significant.
8.8 SUMMARY
The design of radio systems includes a link margin With the exception of ionospheric effects, propa-
that is intended to provide for changes in the received gation phenomena are weather dependent. To over-
signal level due to both equipment effects and ran- come this problem, statistical models are used. Long-
dom changes in the environment between the trans- term measurements of rainfall rate are statistically
mitter and the receiver. The link margin permits the related to long-term path attenuation measurements
communications system to operate with both the re- when taken over the same period and at the same site.
quired performance, a measure of the service quality In this case, long-term is at least 1 year so that all of
required for a significant fraction of the time, and the seasons normally experienced in a given year may
availability, a measure of the time period when us- be included.
able service is provided. Developing an adequate link The prediction of rain attenuation has taken two
margin is critical to the acceptance of the service. distinct paths: one uses measured data and develops
However, each additional dB of link margin that is an empirical model to predict the phenomenon on a
provided comes with a cost associated with it. A lot worldwide basis; the other attempts to model the
of care, therefore, goes into developing an accurate physics of the process. Statistical models of rain de-
estimate of the likely impairments on any given link polarization, tropospheric scintillation, gaseous
that would cause the performance and availability of losses, cloud attenuation, low angle fading, and re-
the service to fall below acceptable levels. A key to lated propagation effects have been developed. Most
this estimate is an understanding of the propagation of these models provide usable predictions for fre-
effects along the path between the satellite and the quencies between 4 and 50 GHz, but care must be
earth station. taken when predictions for unusual path geometries
Propagation effects cause two principal phenomena (e.g., 5°) or severe climates (e.g., tropical regions)
to be observed at the receiving terminal: a change in are required.
the wanted signal level, which is referred to as signal More recently, the impact of individual rain
attenuation or fading; and a change in the unwanted fades—their occurrence statistics, duration of indi-
signal level, which is referred to as depolarization or vidual events, time between fades of the same level—
cross-polarization. Attenuation and depolarization ef- has become important for developing user perception
fects are a function of the signal frequency, the at- for direct-to-home (DTH) services. Counter-measures
mospheric conditions, and the path geometry. In gen- to rain fades may take many forms—for example, in-
eral, the higher the frequency, the warmer and wetter creasing the TDMA frame allocation, changing the
the weather, and the lower the operating elevation an- modulation index, changing the power level, chang-
gle of the earth station, the worse the propagation ef- ing the frequency—and it is likely that some of
fects are. The only time this is not true is for ionos- them will be included in the Ka-band DTH services
pheric effects, where the effects on commercial satellite planned for the first decade of the twenty-first
systems are only of significance at C band or below. century.
c08.qxd 12/08/02 09:57 Page 339
REFERENCES 339
REFERENCES
1. L. J. IPPOLITO, Propagation Effects Handbook for Satel- 14. A. W. DISSANAYAKE, J. E. ALLNUTT, and F. HAIDARA,
lite Systems Design, ITT Industries, Advanced Engi- “Cloud Attenuation Modeling for SHF and EHF Appli-
neering and Sciences, Ashburn, Virginia. Developed for cations,” Invited paper, Special issue of International
NASA as an update of Reference Publication 1082(04), Journal of Satellite Communications, revised and ac-
National Aeronautics and Space Administration, Wash- cepted, January 2000.
ington, DC, September 2000. 15. Rec. ITU-R P.618-6, Propagation Data and Prediction
2. J. E. ALLNUTT, Satellite-to-Ground Radiowave Propa- Methods Required for the Design of Earth–Space
gation, ISBN 0 86341 157 6, Peter Peregrinus, 1989 Telecommunications Systems, 1999.
[2nd Ed. in preparation]. 16. T. PRATT, and D. J. BROWNING, “Co-polar Attenuation
3. J. E. ALLNUTT, “Refraction and Attenuation in the Tro- and Radiometer Measurements at 30 GHz for a Slant
posphere,” Wiley Encyclopedia of Electrical and Elec- Path to Central England,” Proceedings ATS-6 Meeting
tronics Engineering, John G. Webster, ed., ISBN 0 471 (ESA SP-131), ESTEC, Noordwijk, The Netherlands,
13946 7, Vol. 18, pp. 379–388, 1999. 15–20, October 1977.
4. A. W. DISSANAYAKE, J. E. ALLNUTT, and F. HAIDARA, 17. E. C. JOHNSTON, D. L. BRYANT, D. MAITI, and J. E.
“A Prediction Model that Combines Rain Attenua- ALLNUTT, “Results of Low Elevation Angle 11 GHz
tion and Other Propagation Impairments along Satellite Beacon Measurements at Goonhilly,” IEE Con-
Earth–Satellite Paths,” IEEE Transactions on Antennas ference Publication No. 333, ICAP 91, pp. 366–369,
and Propagation, Vol. 45, No. 10, October 1997, April 1991.
pp. 1546–1558. 18. K. MURSULA, and T. ULICH, “A New Method to Deter-
5. M. A. B. TERADA, “Reflector Antennas,” Wiley Ency- mine the Solar Cycle Length,” Geophysic Research Let-
clopedia of Electrical and Electronics Engineering, ters, 25, 1837–1840, 1998.
JOHN G. WEBSTER ed., ISBN 0 471 13946 7, Vol. 18, 19. Rec. ITU-R PN.837-1, Characteristics of Precipitation
pp. 360–379, 1999. for Propagation Modelling, 1994.
6. D. C. COX and H. W. ARNOLD, “Comparison of Mea- 20. The general web site for the ITU is http://www.itu.int
sured Cross-Polarization Isolation and Discrimination For ITU model software see http://www.itu.int/brsg/sg3/
for Rain and Ice on a 19 GHz Space–Earth Path,” Radio databanks/tropospheric.html
Science, 19, 617–628, March–April 1984. 21. Report 564 of the Recommendations and Reports of the
7. P. N. KUMAR, “Depolarization of 19 GHz Signals,” CCIR, Volume V, Propagation in Non-Ionized Media,
Comsat Technical Review, 12, 271–293, Fall 1982. International Telecommunications Union (ITU), Geneva,
8. ACTS experimental results, presented in a series of pro- Switzerland, 1982.
ceedings; e.g., Proceedings of the Twenty-third NASA 22. Rec. ITU-R P.837-2, Characteristics of Precipitation for
Propagation Experimenters Meeting (NAPEX XXIII) Propagation Modelling, 1999.
and the Advanced Communications Technology Satel- 23. M. P. M. HALL, Effects of the Troposphere on Radio
lite (ACTS) Propagation Studies Workshop, Nasser Communication, Peter Peregrinus, Ltd., Stevenage, UK,
Golshan and Christian Ho, eds., Falls Church, Virginia, 1979.
June 2–4, 1999. For updated information, see the NASA 24. R. L. OLSEN, D. V. ROGERS, and D. B. HODGE, “The aRb
web site. Relation in the Calculation of Rain Attenuation,” IEEE
9. OLYMPUS Propagation Experiment (OPEX), Second Transactions on Antennas and Propagation AP-26,
Workshop of the Olympus Propagation Experimenters, 318–329, March 1978.
WPP-083, Volumes 1 (Attenuation Measurement and Pre- 25. Rec. ITU-R P. 838, Specific Attenuation Model for Rain
diction), 2 (Depolarization), 3 (Radiometry and Meteoro- for Use in Prediction Methods, 1992.
logical Measurements), 4 (Radar), and 5 (Data Process- 26. T. OGUCHI, “Electromagnetic Wave Propagation and Scat-
ing), Noordwijk, The Netherlands, 8–10 November 1994. tering in Rain and Other Hydrometeors,” Proceedings of
10. P. A. WATSON, and M. ARABI, “Cross-Polarization the IEEE, 71, 1029–1078, September 1983.
Isolation and Discrimination,” Electronics Letters, 9, 27. J. E. ALLNUTT, and D. V. ROGERS, “System Implications
November 1973, 516–519. of 1411 GHz Path Depolarization. Part II: Reducing
11. Rec. ITU-R P.676-3, Attenuation by Atmospheric Gases, the Impairments,” International Journal of Satellite
1997. Communications, 4, 13–17, 1986.
12. A. W. DISSANAYAKE, J. E. ALLNUTT, and F. HAIDARA, 28. C. W. BOSTIAN, and J. E. ALLNUTT, “Ice Crystal Depo-
“A Prediction Model that Combines Rain Atten- larization on Satellite-to-Earth Microwave Radio Paths,”
uation and Other Propagation Impairments along Proceedings of the IEE, 126, 951–960, 1979.
Earth–Satellite Paths,” IEEE Transactions on Anten- 29. F. BARBALISCIA, A. PARABONI, and C. RIVA, Private com-
nas and Propagation, Vol. 45, No. 10, pp. 1546–1558, munication, December 1999.
October 1997. 30. L. CASTANET, J. LEMORTON, and M. BOUSQUET, “Fade
13. E. SALONEN, and S. UPPALA, “New Prediction Method Mitigation Techniques for New SatCom Services at Ku-
of Cloud Attenuation,” Electronics Letters, Vol. 27, Band and Above: A Review,” COST 255 First Interna-
No. 12, pp. 1106–1108, 1991. tional Workshop on Radiowave Propagation Modelling
c08.qxd 12/08/02 09:57 Page 340
for SatCom Services at Ku-Band and Above, WPP-146, 33. J. O. LAWS and D. A. PARSONS, “The Relation of Rain-
243–251, October 1998. Size to 10Intensity,” Trans. Amer. Geophys. Union, Vol.
31. (a) D. V. ROGERS and J. E. ALLNUTT, “System Implica- 24, pp. 432–460, 1943.
tions of 1411 GHz Path Depolarization. Part I: Pre- 34. J. E. ALLNUTT, and B. ARBESSER-RASTBURG, “Low El-
dicting the Impairments,” International Journal of evation Angle Propagation Modeling Considerations for
Satellite Communications, Vol. 4, No. 1, pp. 1–12, 1986. the INTELSAT Business Service,” ICAP 85, IEE Conf.
(b) J. E. ALLNUTT and D. V. ROGERS, “System Implica- Publ. 248, 62–66, 1985.
tions of 1411 GHz Path Depolarization. Part II: Re- 35. R. K. CRANE, “Prediction of Attenuation by Rain,” IEEE
ducing the Impairments,” International Journal of Satel- Transactions on Communications, Vol. COM-28, No. 9,
lite Communications, Vol. 4, No. 1, pp. 13–18, 1986. pp. 1717–1735, September 1980.
32. D. B. HODGE, “An Empirical Relationship between Path 36. W. L. STUTZMAN and W. K. DISHMAN, “A Simple
Diversity and Gain,” IEEE Transactions on Antennas Model for the Estimation of Rain Induced Attenuation
and Propagation, AP-24, 250–251, March 1976; D. B. along Earth–Space Paths at Millimeter Wavelengths,”
HODGE, “An Improved Model for Diversity Gain on Radio Science, 17, 1465–1476, November–December
Earth–Space Paths,” Radio Science, 17, 1393–1399, 1982.
November–December 1982.
PROBLEMS
1. a. For a typical satellite-to-ground link, give one c. If the elevation angle of the transmitting earth station
propagation impairment that has less impact on the is 50°, it is 200 m above mean sea level, and the height
average signal level as the frequency increases from of the rain hR is 4.5 km, what are the predicted values
1 to 30 GHz and give one propagation impairment of rain attenuation measured for time percentages 0.01
that has a greater impact on the average signal level and 0.001% of a year for the uplink transmitter?
as the frequency increases from 1 to 30 GHz. 3. A simple procedure to calculate the effect of
b. Ionospheric scintillation effects are variable in changing the elevation angle is to use the cosecant
both time and location. (i) What causes the time vari- law [see Eq. (8.24)]. To calculate the effect of chang-
ability? (ii) Over what periodic interval of time do we ing the frequency, either a simple formula may be
expect to observe the worst effects repeating on a used [see Eq. (8.25)] or a more complicated proce-
given satellite link? (iii) In any given year, when dure may be used [see Eq. (8.26) et seq.]. A DTH
would the worst ionospheric effects be observed? satellite ISP consortium is investigating whether it
(iv) On any given day, when would the worst ionos- should change location and frequency for its uplink
pheric effects be observed? (v) Within what latitude transmitter. Using their currently allocated frequency
range on the earth’s surface are these effects the worst (30 GHz) on the uplink, they have observed an at-
for geostationary satellite links? tenuation of 30 dB at an elevation angle of 50° at the
2. A DTH (direct-to-home) satellite ISP (Internet time percentage of interest to their service. They have
service provider) consortium is designing a system to been notified that they would be permitted to use an
provide digital multimedia service to CONUS (CON- uplink frequency at 18 GHz, but they would have to
tinental United States) coverage. They are investigat- relocate their earth station. The rain climate is the
ing where in the United States to locate the uplink same as their current earth station, but the elevation
transmitter for their service. The uplink to the Inter- angle would now be 15° instead of 50°.
net/Multimedia satellite will be at 30 GHz and the a. Using the simple frequency scaling formula [Eq.
downlink to the user terminals will be at 20 GHz. (8.25)], what is the attenuation that would be observed
a. Using Table 8.2 and Figure 8.15, determine the ap- at a frequency of 18 GHz if an attenuation of 30 dB
proximate region within CONUS where you would was observed at 30 GHz (both at an elevation angle
expect to obtain the lowest outage time due to rain of 50°)?
for the uplink transmitter in the Internet/Multimedia b. If the elevation angle is reduced from 50° to 15°,
satellite system. what attenuation would be observed at a frequency of
b. For time percentages of 0.01 and 0.001% of a year, 18 GHz, using the value you calculated for 18 GHz
what are the specific attenuations observed at the site for an elevation angle of 50° in part (a) above as input?
you have selected in part (a), assuming horizontal po- c. Based on the answer you obtained in part (b),
larization is employed? (Hint: you will need to refer would you recommend moving the uplink transmitter
to Table 8.3 for this part of the problem.) and changing the uplink frequency to 18 GHz?
c08.qxd 12/08/02 09:57 Page 341
PROBLEMS 341
d. Would your answer change if you used the more d. Is the tropospheric scintillation an availability-
complicated frequency scaling formula in Eq. (8.26) limiting phenomenon when acting in isolation from
instead of the frequency-squared law in Eq. (8.25)? other effects?
4. A cable TV network (CATV) downlink signal to a e. Is rain attenuation a performance-limiting phe-
cable head-end terminal is at approximately 12.5 GHz. nomenon, when acting in isolation from other effects?
The attenuation measured on the CATV downlink for f. Is rain attenuation an availability-limiting phe-
0.01% of the time in the location of interest is 12 dB. nomenon when acting in isolation from other effects?
The elevation angle is 20°. The downlink is using dual- g. When combining tropospheric scintillation and
polarization frequency reuse to permit all of the chan- rain attenuation effects for this DTH terminal, does
nels to be sent through one satellite. In order to meet the terminal meet both the performance and the avail-
the QOS (quality of service) guarantees established ability specifications?
for the terminal, the XPD must be no lower than 15 dB.
h. If your answer is “no” to either the performance
a. What is the predicted XPD for 0.01% of the time or availability questions in part (g) above, what ad-
if the downlink signal is linearly polarized with a tilt ditional CN margin is required to meet the per-
angle of 0°? formance and availability specifications?
b. What is the predicted XPD for 0.01% of the time
6. An earth station complex is being designed to pro-
if the downlink signal is circularly polarized?
vide high availability access at Ka band to a satellite
c. Is the XPD minimum of 15 dB for the QOS in- system used in a DTH Internet/Multimedia service of-
terference criterion met in either of these cases? fering. The earth station must meet an annual 99.99%
5. Tropospheric scintillation is not an absorptive ef- availability level, that is, the maximum outage in any
fect and so will not lead to an increase in receiver noise year must not exceed 0.01% of a year. The maximum
temperature, as rain attenuation will. There will be a EIRP from the earth station on the 30-GHz uplink is
slight increase in noise temperature due to enhanced limited by interference considerations, resulting in a
humidity levels that lead to tropospheric scintillation, maximum rain fade margin of 30 dB on the uplink.
but we will ignore that aspect for this question. For other reasons, the downlink rain fade margin
A direct-to-home (DTH) Ku-band receiver has the (which includes the CN reduction due to noise
following design specifications: temperature increase) at a frequency of 20 GHz is
System noise temperature in clear sky: 100 K limited to a maximum of 9 dB. The earth station com-
Clear sky CN: 11 dB plex is situated in a region where the rain attenuation
Performance margin: 99% of a year with CN at 30 GHz is 50 dB for 0.01% of a year. The ap-
10 dB proximate frequency scaling formula may be used to
Availability margin: 99.9% of a year with CN find the equivalent 20 GHz fade level for 0.01% of a
6 dB year. As can be seen, a single earth station will not
The climate in which the DTH receiver is to operate be able to meet the availability requirements in this
has been predicted to have the following tropospheric location. To overcome the rain attenuation, consider-
scintillation and rain attenuation statistics: ation is being given to operating two earth stations as
a site diversity pair. The earth stations will operate at
Annual Scintillation Rain attenuation an elevation angle of 40°. The baseline between the
%-age time fade level fade level earth stations is 90° to the satellite azimuth.
a. Calculate the diversity gain achievable at 30 GHz
10% 0.5 dB 0 dB and 20 GHz for site separations of 2, 3, 5, 10, 15, and
1% 1.5 dB 1 dB 20 km in the above diversity system.
0.1% 2.5 dB 3 dB b. If an additional margin of 3 dB must be allowed to
0.01% 3.5 dB 10 dB provide a hysteresis band to assist switching between
the earth stations on the 30 GHz uplink (i.e., the 50-
a. What is the reduction in CN due only to tropos- dB rain margin is increased to 53 dB), which is the
pheric scintillations at the four percentage times? first site separation calculated in part (a) that will per-
b. What is the reduction in CN due only to rain at- mit the 30-GHz uplink availability target to be met
tenuation at the four percentage times (the effective (going from the lowest separation to the highest)?
temperature of the rain medium is 280 K)? c. If an additional margin of 0.5 dB must be allowed
c. Is the tropospheric scintillation a performance- for inefficiencies in the diversity combining on the
limiting phenomenon, when acting in isolation from 20-GHz downlinks between the diversity pair of earth
other effects? stations (i.e., the 9-dB rain margin is increased to
c08.qxd 12/08/02 09:57 Page 342
9.5 dB), which is the first site separation calculated outage of 0.04% of a year is permitted. The 0.04%
in part (a) that will permit the 20-GHz downlink can be split equally, 0.02% on the uplink and 0.02%
availability target to be met (going from the lowest on the downlink. Thailand has a much more severe
separation to the highest)? rain climate than the west coast of the United States
7. The attenuation that would be observed on a satel- and it may make the overall link design easier if a
lite path may be estimated by measuring the increase nonsymmetrical split is made of the total outage time.
in antenna noise temperature and calculating the as- If the earth station in Thailand is in climate M and
sociated attenuation A from Eq. (8.40). Instruments the earth station on the western seaboard of the United
that do this are called radiometers. They are useful States is in Climate D (see Table 8.2); the uplink fre-
for attenuation values between 0 and about 10 dB, but quency is 12 GHz in both cases and the downlink
not very good for larger values of inferred attenua- frequency is 10 GHz in both cases; linear, vertical
tion. (The effect is called radiometer saturation.) polarization is used both ways; the elevation angle is
Using Eq. (8.40), show why this occurs. 5° at both the Thailand and the U.S. earth stations,
8. Geostationary communications satellites that use find the following:
linear transponders have both advantages and disad- a. How should the 0.04% outage time be divided up
vantages when compared with satellites that use on- between the 12-GHz Thailand uplink and the 10-GHz
board processing (OBP). One of the advantages of a U.S. downlink on the Thailand–U.S. one-way link so
linear transponder is its flexibility in usage—it can that the same rain attenuation is experienced on the
easily accommodate analog and digital traffic with Thailand uplink as it is on the U.S. downlink?
varying capacity streams, provided they fit within the
transponder bandwidth and power limitations. An- b. How should the 0.04% outage time be divided up
other advantage is the ability to apportion outage to between the 12-GHz U.S. uplink and the 10-GHz
different parts of a given link. A typical two-way link Thailand downlink on the U.S.–Thailand one-way
is designed one way at a time. That is, for example, link so that the same rain attenuation is experienced
in a service between the United States and Thailand, on the U.S. uplink as it is on the Thailand downlink?
the link from Thailand to the United States is designed [Note: for both parts (a) and (b) in the above ques-
separately from the return link from the United States tion, the answer should be rounded to increments of
to Thailand. For high availability services, a one-way 0.005%, e.g., 0.03%:0.01% or 0.035%:0.005%.]
c09.qxd 12/08/02 09:59 Page 343
CHAPTER 9
VSAT SYSTEMS
The acronym VSAT stands for very small aperture terminal and, like many technical
terms, its precise meaning has changed over the years. The first earth station antennas
used in commercial satellite communications systems were very large and expensive1,
with typical aperture diameters of 30 m. These large antennas operated in C band (64
GHz). With the rapid expansion of satellite telecommunications worldwide, there was a
need to make access to the satellite more affordable. This came about in two ways: a sig-
nificant increase in the transmit power capabilities of satellites and the move to frequency
bands above C band. Both led to a rapid decrease in the size and cost of the earth sta-
tion antenna.
9.1 INTRODUCTION
Most VSAT systems operate in Ku band, with earth station antenna diameters of 1 to
2 m and transmitter powers of 1 or 2 W. The earth stations are usually organized in a
star network, in which the earth stations connect to a central hub station via a GEO
satellite. Data rates on the links are from a few thousand bits per second up to 256
kbps, depending on the traffic requirements. VSAT systems are used to link businesses
and stores to a central computer system so that sales transactions can be completed
more rapidly than by using a telephone line and modem, and so that a central office
can rapidly distribute and collect information from a large number of locations in a
region or country.
In the 1990s there was rapid growth of VSAT networks in the United States. Busi-
nesses adopted VSAT networks for the transmission of data, as an alternative to the ter-
restrial telephone and data systems then available. The next decade is expected to see
growth of VSAT networks operating in Ka band as new 3020 GHz GEO satellites be-
come available. These networks may operate directly to the home for Internet connections
and delivery of multimedia material. Few VSAT systems are used just for voice traffic,
although the data rates are well matched to digital voice bit rates. For this reason, voice
over IP (VOIP) has become a growth segment in VSAT operations.
Large antennas are usually implemented using a symmetrical configuration, for ease
of construction, with the feed on the boresight axis. The feed can either be in front of the
antenna (a front-fed design) or behind the antenna, as in Cassegrain or Gregorian designs.
Further, these different approaches may be axially symmetric or offset (see Section 9.6). A
common break point in the design of antennas is at a main reflector diameter of about 100
wavelengths. If the diameter is larger than this, the additional cost of a Cassegrain or Gre-
gorian design is more than outweighed by the superior off-axis and polarization performance
achievable, and also the increased gain (up to 1 dB) that can be achieved by shaping the re-
flectors. Cassegrain and Gregorian antennas require a subreflector with a minimum diame-
ter of 10 wavelengths. If the main reflector is less than 100 wavelengths in diameter, the
subreflector becomes an appreciable fraction of the main reflector diameter and causes
343
c09.qxd 12/08/02 09:59 Page 344
344 CHAPTER 9 VSAT SYSTEMS
TABLE 9.1 Summary of Characteristics for the Intelsat VSAT

IBS Antennas
C-band antenna standard F1 H4 H3 H2
G/T (4 GHz), dB/K 22.7 22.1 18.3 15.1

Typical antenna diameter, m 3.5–5.0 3.5–3.8 2.4 1.8
Voltage axial ratio
(circular polarization): 1.09 1.09 1.3 1.3
XPD isolation value, dB: 27.3 dB 27.3 dB 17.7 dB 17.7 dB
Ku-band antenna standard E1 K3 K2
G/T (11 GHz), dB/K 25.0 23.3 19.8

Typical antenna diameter, m 2.4–3.5 1.8 1.2
Voltage axial ratio
(linear polarization): 31.6 20.0 20.0
XPD isolation value, dB: 30.0 dB 26.0 dB 26.0 dB
Source: Intelsat Earth Station Standards (IESS) 207 (C-Band) and 208 (Ku-Band)2.
significant blockage and scattering problems. Earth stations with antenna aperture diame-
ters less than 100 wavelengths were called small aperture terminals and, as the size reduced,
the term VSAT was coined and then USAT (ultra small aperture terminal). Table 9.1 gives
a summary of the standard antennas used in the Intelsat IBS (Intelsat business service). These
are typical sizes of VSAT antennas used in international and domestic services. Figure 9.1
shows a typical VSAT antenna on the roof of a commercial building.
FIGURE 9.1 A typical VSAT antenna.

c09.qxd 12/08/02 09:59 Page 345
9.2 OVERVIEW OF VSAT SYSTEMS 345
The standard VSAT antennas in Table 9.1 are not as small as the Ku-band direct-
to-home (DTH) antennas used for direct broadcast satellite television reception, which
are typically 0.5–0.8 m in diameter. DBS-TV satellites use very powerful transpon-
ders, typically 160–240 W compared to 20 to 50 W of Ku-band satellites used for
VSAT service. VSAT antennas are also much larger than the ultimate USAT—a hand-
held satellite telephone as used in Iridium, Globalstar, New ICO, and other mobile
satellite service (MSS) systems, which have an omnidirectional antenna. While the size
of the VSAT antenna is a key factor in making the service both economically attrac-
tive to the user and environmentally acceptable to the community, it places severe re-
strictions on the end-to-end system design. A careful balance has to be drawn between
satellite transponder loading, transmitted power (both up and down), VSAT antenna off-
axis emission (for interference considerations), clear sky performance, and—especially
for Ku-band frequencies and above—availability during impaired propagation conditions
(i.e., during rain).
9.2 OVERVIEW OF VSAT SYSTEMS
The underlying concept behind most VSAT systems is to bring telecommunications serv-
ice directly to the end user without any intermediate distribution hierarchy. Historically,
traffic from individual users was bundled together into ever-larger groups and carried over
trunk transmission lines via terrestrial microwave systems, satellite systems, or optical
fiber cables, before being divided up (demultiplexed) into smaller traffic streams and re-
distributed to the users at the far end. This is still the most economical transmission ar-
chitecture for point-to-point communications when the services are being brought into
areas with relatively high concentrations of users. Such conditions do not always apply,
however, and VSAT networks take advantage of the wide area broadcast capabilities of
GEO satellites.
In many regions of the world, the potential users are either widely distributed or the
existing telecommunications infrastructure lacks the capacity to expand quickly to meet
the demand for new users. This situation applies to most developing countries and, in
many cases, network implementations have been adopted that “leapfrog” the conventional
evolution of telecommunications systems. Geostationary satellites allied to microwave cel-
lular technologies have been used to bypass completely the traditional expansion of ana-
log telephony. One such solution is wireless local loop (WLL) coupled with VSAT
distribution architectures. Figure 9.2 illustrates the concept schematically3.
The VSAT/WLL concept usually has an optimum range of user densities where the
economics are most favorable. This is illustrated in Figure 9.3. The information in Figure
9.3 is only approximate and can be affected by local topography (e.g., a large mountain
range, which would favor satellite delivery), the availability of optical fibers in the coun-
try’s telecommunications network, or significant transportation routes such as a major rail
system, which allows a lower cost optical fiber to be laid alongside the railroad tracks or
right-of-way.
VSAT networks allow multimedia traffic to be brought directly to the end user, but
generally handle only small traffic streams (sometimes as little as the equivalent of one
voice circuit). The traffic stream is also usually intermittent in nature: the user accesses
the satellite in a demand assigned multiple access (DAMA) mode whenever a message is
to be sent and receives a short reply in due course. This is typical in a point of sale (POS)
VSAT system that is used to transmit credit card information at a gas pump or a store reg-
ister. Information about the sale and the customer’s credit is sent to a central computer
c09.qxd 12/08/02 09:59 Page 346
Farm
Wireless
Public
Payphone
Satellite VSAT & WLL Village

Gateway Terminal
Backhaul Local Loop
(VSAT technology) (WLL technology)
FIGURE 9.2 Schematic of a VSAT/WLL communications network (from Figure 2-5 of refer-
ence 3). The geostationary satellite is used to link a large number of VSATs with the main
switching center in a large city. Each VSAT acts as the link to the local switching center in
the village or rural community, with the final mile of the telephony link being carried over
a wireless local loop.
facility, and an authorization or denial is received in response. The interaction between

the VSAT and the main hub earth station in the POS transaction is completely automatic
and transparent to the user, the customer in this case.
Most VSAT networks do not generate enough traffic to justify a dedicated satellite.
Many do not even have enough traffic at any given instant to fill one satellite transponder.
For this reason, most VSAT networks are designed around the use of leased transponders,
~0 Users/km2 ~10 Users/km2 ~100 Users/km2 ~1000 Users/km2
User Density in number of users per square kilometer
Uneconomic: VSAT/WLL: Fiber/microwave FS:

Requires Appears the best Traditional terrestrial fixed
large subsidy for technological service appears the best
any implementation implementation technological implementation
FIGURE 9.3 Approximate economic break points in the implementation choices for serv-
ing new regions with different population densities. Physical distances, major transporta-
tion routes, and geographic barriers, as well as the individual country’s demographics and
political influences, can alter the break points.
c09.qxd 12/08/02 09:59 Page 347
9.3 NETWORK ARCHITECTURES 347
in the case of a large network, or a fractional transponder lease for a medium to small
network. The network architecture selected will depend on a number of factors that are
discussed in the following section.
9.3 NETWORK ARCHITECTURES
There are three basic implementations of any telecommunications service: one-way; two-
way; and split-two-way (sometimes referred to as split-IP, when referring to Internet traffic,
since the outbound and inbound channels are routed over different systems). The two-way
implementation is further divided into two basic network architectures: Star and Mesh.
We will look first at the three basic implementations.
One-Way Implementation
This is the mode of a satellite used in the broadcast satellite service (BSS). The intro-
duction of digital technology allows the provider and user much greater flexibility in
the operation of a broadcast network. By means of proprietary software in the user ter-
minals, different parts of the downlink can be accessed by different subscribers according
to the programs ordered from the supplier (and paid for by the user). This form of chan-
nel selection is called narrowcasting. There can be many narrowcasting groups within
a larger broadcasting area. Figure 9.4 gives a schematic of this one-way (broadcast)
application.
Split-Two-Way (Split IP) Implementation

This implementation is used when there is no normal return channel as, for example, with
Ku-band broadcast satellite service (BSS) systems that carry Internet traffic7. The relatively
Narrowcasting
Group
Broadcasting
Master Coverage Area
Station
FIGURE 9.4 Schematic of a broadcast satellite service coverage region in which smaller, nar-
rowcasting groups exist within the broader coverage area (from Figure 2-1 of reference 3). The
master control station sends encoded signals within the broadcast stream that enables certain
users to have access to particular channel groupings according to the subscriber’s choice.
c09.qxd 12/08/02 09:59 Page 348
high capacity downlink stream is not complemented by an uplink capability from the user
terminal. If the BSS downlink is used as the download channel from an Internet service
provider, the only option the user has for a return link is via another telecommunications
channel, such as a standard telephone line. The Internet protocol (IP) is therefore split
between a satellite downlink (outbound) channel and a terrestrial telephone (inbound, or
return) channel; hence the term split IP for this implementation. The advantage of this
approach is that the VSAT terminal does not require a transmit capability, which signifi-
cantly reduces its cost and complexity. The disadvantage is that the telephone line
connection must usually be through a modem, with a bit rate generally restricted to 56 kbps
or less.
Two-Way Implementation
In this case, a return link is designed into the service so that two-way communications
can be set up over the same satellite, from the hub to the user and from the user back to
the hub. The VSAT/WLL implementation illustrated in Figure 9.2 is a two-way service
between the hub (in this case the satellite gateway) and any VSAT terminal. The archi-
tecture selected is the key to the economics of two-way connections: it can be either Mesh
or Star. These two architectures are illustrated in Figures 9.5a,b, with the topology as
viewed by the satellite shown in Figures 9.6a,b.
Initially, the most common VSAT architectures were Star networks since the very
low receive GT (gain-to-noise temperature ratio) of the VSATs, coupled with their limited
transmit EIRP, was compensated for by using a large hub with high GT and EIRP. The
Satellite Satellite
Master Control
Station (the hub) VSATS
VSATS
(a ) (b)
FIGURE 9.5 (a) Illustration of a Star VSAT network. In this network architecture, all of the traffic is
routed via the master control station, or hub. If a VSAT wishes to communicate with another VSAT,
they have to go via the hub, thus necessitating a “double hop” link via the satellite. Since all of the
traffic radiates at one time or another from the hub, this architecture is referred to as a Star network.
(b) Illustration of a Mesh VSAT network. In this network architecture, each of the VSATs has the ability
to communicate directly with any of the other VSATs. Since the traffic can go to or from any VSAT,
this architecture is referred to as a Mesh network. It will still be necessary to have network control and
the duties of the hub can either be handled by one of the VSATs or the master control station func-
tions can be shared among the VSATs.
c09.qxd 12/08/02 09:59 Page 349
9.4 ACCESS CONTROL PROTOCOLS 349
VSAT
VSAT
VSAT Satellite Hub
VSAT
VSAT
(a )
VSAT VSAT
VSAT VSAT
VSAT Satellite VSAT
VSAT VSAT
VSAT VSAT
(b)
FIGURE 9.6 (a) Topology of a Star VSAT network viewed from the satellite’s perspective.
Note how the VSAT communications links are routed via the satellite to the hub in all cases.
(b) Topology of a Mesh VSAT network from the satellite’s perspective. Note how all of the
VSATs communicate directly to each other via the satellite without passing through a larger
master control station (hub).
cost of the hub was therefore quite high and, at least for the smaller VSAT networks,
somewhat prohibitive. This led to the concept of a shared hub, where several networks
operate through one main hub. The difficulty with this approach for large countries with
widely dispersed communities is that the host computers for the small VSAT networks
are rarely close to the hub. A high-speed terrestrial data link is required between the host
computers of the networks and the hub, which increases the cost of the network. Rather
than have one large hub for all of the VSAT networks sharing the same satellite, the over-
all network evolved to allow each subnetwork to have its own hub as soon as the eco-
nomics made it attractive. In this way, the host computer of each VSAT network can be
co-located with its own hub, thus eliminating the cost of the interconnection between the
hub earth station and the computer controlling the service offered through the VSAT net-
work. Whether the hub is shared or dedicated on the one hand or the VSAT is connected
to a single user or a local area network (LAN) with multiple users sharing access through
an Ethernet connection on the other, in every case there will need to be an access control
protocol.
9.4 ACCESS CONTROL PROTOCOLS
The International Standards Organization (essentially a standards committee of the United

Nations) has specified the open systems interconnection (ISO/OSI) that mandates a seven-
layer model for a data communication system, as shown in Figure 9.7.
c09.qxd 12/08/02 09:59 Page 350
User 1 User 2
Application Application
Presentation Presentation
Session Session
Transport Transport
Network Network
Link Link
Physical Physical
FIGURE 9.7 The ISO-OSI seven-layer “stack” for interconnecting data terminals. In this
example, user one and user two are conducting a two-way communications session with
each other. Each user interacts with their local device (e.g., a computer keyboard/visual dis-
play unit) at the application layer of the ISO-OSI stack. Their transaction is then routed via
the various layers, with suitable conversions, etc., until the content is ready to be transmit-
ted via the physical layer.
A satellite communications link occupies primarily the physical layer, which is where
bits are carried between the terminals. A VSAT network must have terminal controllers at
each end of the link and these occupy the network and link layers, the two layers above
the physical layer. The network control center typically controls the system and is re-
sponsible for the remaining layers. Unfortunately, few communications systems conform
in an easily identifiable way to the seven layers of the ISO-OSI model. (For example, the
IP protocol stack of five layers simply puts the first three layers of the ISO/OSI stack into
one layer). It is, however, very useful as a conceptual model which identifies functions
that must be performed somewhere in every data communication network. Most data com-
munication networks use some form of packet transmission, in which blocks of data are
tagged with an address, error control parity bits, and other useful information before trans-
mission. The receiving end of a link checks arriving packets for errors, and then sends an
acknowledgement signal (ACK) that the packet was received correctly, or a not acknowl-
edge signal (NAK) that tells the transmit end to resend a particular packet because the
packet had an error. Some systems do not send acknowledgements, only NAK signals to
request a retransmission of a packet with an error, since this speeds up data transmission.
This is the error control method used in the Internet protocol TCP/IP. Generically, such
systems are known as automatic repeat request (ARQ) systems. Chapter 7 discusses packet
transmission systems and the problem of error detection and correction in packet networks
using satellite links.
The ISO-OSI stack was initially developed for terrestrial communications systems.
For this reason, the protocols that implement the functions of each layer were designed
for use in terrestrial circuits with low delay and low bit error rate (BER), that is, very high
performance levels. These are key points when trying to use such protocols over satel-
lites, particularly those in geostationary earth orbit (GEO). Many of the early protocols
had a connection time-out of a few milliseconds. If no reply was received from the recipient
c09.qxd 12/08/02 09:59 Page 351
in this interval, transmissions ceased. Similarly, an errored signal received from the source
or an intervening node would trigger an automatic error recovery sequence. For exam-
ple, the X.25 and X.75 packet systems use an ARQ approach, which, on detecting an
error in a packet, immediately requests a retransmission and halts further transmissions
until the corrected packet is received. Frame relay and ATM (asynchronous transfer mode)
systems flag the error but continue the flow of information (continuous transmission
ARQ). In both cases, the errored transmission must be corrected and suitable buffers at
the receiver end (or intermediate node) used to restore the packets in their original or-
der. The more errors that occur in the link, necessitating many retransmissions of pack-
ets, the slower the effective data throughput rate of the link becomes. The potential for
delay and (propagation induced) errors are therefore critical design elements in digital
VSAT connections.
Delay Considerations
A typical slant range to a GEO satellite is 39,000 km. The one-way delay over such a
GEO link (earth station to satellite to earth station) is 2 (rangevelocity) 260 ms.
The one-way delay in a typical 4000-km transcontinental link via fiber-optic cable is
a little over 13 ms. Neither example includes processing delay (e.g., source coding
and/or compression, channel coding, baseband processing in the switching elements,
frame length) which can add several tens of milliseconds or even over a hundred
milliseconds.
The time out element of a protocol is often referred to as the window of the con-
nection. As long as the window is “open,” communications can continue without inter-
ruption. Figure 9.8 illustrates a continuous transmission ARQ system that has a 60-ms
Packets received by User 2
A B C D
Time
10 ms ACK ACK ACK

transmission A B C
delay ACK
ACK window
A window
Time
Time B
A2
A1
A B C D
Time
Packets transmitted by User 1
FIGURE 9.8 Illustration of a communications link with a 10-ms one-way delay and a 60-ms
window. In this example, a packet or frame is sent at instant A1 from user 1 to user 2. User 2
receives the transmission without error and sends an acknowledgement back, which is re-
ceived at instant A2, 20 ms after the initial transmission from user 1. This is well within the
time window of 60 ms. The time window rolls forward after each successful acknowledge-
ment. Thus the transmission from user 1 at instant B1 is received by user 2, and the ac-
knowledgement received by user 2 at instant B2, within the new rolling time window of 60
ms. Each packet or frame is successfully received in this example.
c09.qxd 12/08/02 09:59 Page 352
Packets received by user 2
A B C D
Time
ACK
C
ACK A ACK
ACK B ACK B
window window A
Time
Time A2
A1 (outside
ACK window A)
A B C D
Time
Packets transmitted by User 1
FIGURE 9.9 Illustration of a communications link with a 260-ms one-way delay and a
60-ms window. In this example, a packet or frame is sent at instant A1 from user 1 to
user 2. User 2 receives the transmission without error and sends an acknowledgement
back, which is received at instant A2, 260 ms after the initial transmission from user 1.
Unfortunately, instant A2 is well after the rolling window time-out of 60 ms. Transmis-
sions from user 1 are automatically shut down by the protocol when the time-out of
60 ms is exceeded. Ignoring processing delays in this example, user 1 is only transmit-
ting for 60 ms in every 260 ms, thus drastically lowering the throughput. Again, no
propagation errors are assumed to occur.
window with a 10-ms one-way delay and Figure 9.9 illustrates a link with a 60-ms window
and a 260-ms one-way delay.
Clearly, satellite systems have to operate satisfactorily, and seamlessly (i.e., the user
has no idea whether the link is terrestrial or via a satellite), with existing terrestrial net-
works or their utility is severely compromised. This is particularly true for GEO systems
and there are two ways to make terrestrial protocols work with a satellite link. First, the
protocols can be changed so that the time-out window is well in excess of 260 ms; sec-
ond, the satellite element of the packet network can be configured to exist as a separate
subnetwork within the global packet network. In practice, both solutions are adopted.
Figure 9.10 illustrates the concept4.
The VSAT and hub “protocol” equipment act as processing buffers to separate
the satellite (VSAT) network from the terrestrial network. This is sometimes known as
spoofing because the terrestrial part of the system uses a conventional protocol and is
unaware of the VSAT network’s existence. The electronic processing and emulation
permit traffic to flow seamlessly between two very different networks without opera-
tor intervention. In essence, this is the interface through which the VSAT user is con-
nected to the VSAT network via the physical layer (see Figure 9.11). Once the user’s
traffic has moved from the terrestrial network through the interface and is inside the
VSAT network, the packet header is reorganized, with the appropriate routing and ad-
dress of the traffic attached, so that the information can pass successfully over the satel-
lite network to the correct recipient. Network management of the VSAT system, which
includes congestion control, is also carried out in this element of the VSAT network,
termed the network kernel. In addition, all of the necessary protocol conversions are
carried out so that the packet or frame can successfully pass over a satellite connection
with a long delay.
A typical data link layer protocol (layer 2 in the ISO-OSI stack) that is used in a
low delay, terrestrial link employs modulo-8 operation. That is, the protocol will transmit
c09.qxd 12/08/02 09:59 Page 353
User’s User’s
higher higher
layers layers
Gateway Gateway
Layer 3 Protocol Protocol

Layer 3
Layer 2 Interface Network Network Interface
kernel kernel Layer 2
Physical Physical
Layer 1 Layer 1
interface interface
User VSAT Hubstation User

terminal terminal
VSAT network
Terrestrial link
Satellite link
FIGURE 9.10 Protocol architecture of a Star VSAT network (Figure 2.2.1 of reference 4
©
ITU, reproduced with permission). VSAT networks are normally maintained as independent,
private networks, with the packetization handled at the user interface units of the VSAT
terminals. The satellite access protocol (with a larger time-out window) is handled in the
VSAT/hub network kernel, which also handles packet addressing, congestion control, packet
routing and switching, and network management functions. Protocol conversion and, if
necessary, emulation is handled by the gateway equipment.
only 7 unacknowledged frames before it stops transmissions; this leads to the low
throughput demonstrated in Figure 9.9, particularly for GEO satellite links. The high
level data link control (HDLC) protocol used in layer 2 for satellite systems therefore
usually employs a modulo-128 operation. That is, 127 frames may be sent without re-
ceiving any acknowledgements before the protocol shuts down transmissions. Moving
from modulo-8 to modulo-128 operation significantly increases the “window” size per-
mitted for the link layer control. The concept, called protocol emulation, is demonstrated
in Figure 9.114.
Another critical function performed in the VSAT interface/kernel sections is to
respond to polling activity from the terrestrial packet networks. It is normal for packet
networks to poll users to see if there are packets to be sent. The interface/kernel
elements in the VSAT network respond to the polling signals of the terrestrial net-
work immediately, thus avoiding the long delay that would occur if the polling signal
had to be passed over the satellite link. Negative acknowledgements are made to the
polling signals until a request to send data is received over the satellite link. Given that
the correct protocols have been inserted at ISO-OSI layer 2 within the VSAT system,
and the management functions have been carried out (i.e., polling, switching, routing,
addressing, and flow control) so that the link can operate successfully at a protocol
level, there still remains the major part of the system design question to answer: how
is the physical connection to be established over the satellite? To answer this question
we must move from protocol design/emulation to transmission engineering. First,
we will cover some of the basic techniques involved in developing a transmission
design.
c09.qxd 12/08/02 09:59 Page 354
Low throughput at layer-2

modulo-8 operation
A B
Protocol emulators
inserted at both ends
of VSAT network
A B
B′ A′
VSAT Hub
Mimic B with Optimized for satellite Mimic A with

modulo-8 modulo-128 operation modulo-8
FIGURE 9.11 Schematic of protocol emulation to permit a VSAT network to operate
seamlessly with a terrestrial network (Figure 2.2.2 of reference 4 ©ITU, reproduced with
permission). In the modulo-8 operation shown in the top part of the figure, the VSAT net-
work simply passes on the traffic over the satellite without any change to the protocols in
the link layer (layer 2). This results in extraordinarily low throughput for GEO systems. In
the lower part of the figure, the bottom two layers of the ISO-OSI stack are formed inside
the VSAT network and the modulo-8 operation is changed to modulo-128. The two-layer
stack also emulates the other side of the VSAT network so that terrestrial network A be-
lieves it is linked directly with terrestrial network B. That is, both terrestrial networks are
mimicked at the VSAT interface/kernel.
9.5 BASIC TECHNIQUES
Satellite link design has been covered extensively in Chapters 4, 5, and 6. However, some
of the major elements that concern VSAT systems design will be reviewed below. These
involve: selecting an appropriate multiple access scheme, evaluating signal formats, and
establishing effective coding and interference practices.
c09.qxd 12/08/02 09:59 Page 355
9.5 BASIC TECHNIQUES 355
Multiple Access Selection

As set out in Chapter 6, there are three fundamental multiple access schemes: frequency
division multiple access (FDMA), time division multiple access (TDMA), and code divi-
sion multiple access (CDMA). Within TDMA, there are two broad subdivisions of access:
those that are closely controlled in time and access ability and those (like ALOHA and
other Ethernet-like connections) that are loosely controlled in time and access ability.
Multiple access schemes that do not closely control time, frequency, and/or code are sig-
nificantly less efficient than those that do5,6. Pure ALOHA, which is a random access
scheme, has a maximum throughput of 18.4%5. By combining some aspects of TDMA
with the random access of ALOHA, slot reservation ALOHA can have an efficiency ex-
ceeding 60%5,6. Slot reservation is akin to a controlled access TDMA scheme with a very
large frame.
The intended application and the potential interference environment often determine
the choice between FDMA, TDMA, and CDMA for VSAT networks, with economics also
playing a major part.
FDMA generally offers the lowest costs for entry-level VSAT systems from the
user’s perspective since the receiver bandwidth and terminal transmit power required
are the lowest. These systems carry thin route traffic, typically the equivalent of one
digital voice channel at 64 kbit/s. The occupied bandwidth of an RF channel carrying
a digital signal with a symbol rate Rs and using error control coding with a code rate
Rc is given by
B Rs 11 a2 Rc Hz (9.1)
where is the roll-off factor of the root raised cosine (RRC) filters in the link. For example,
in a link using QPSK modulation where two bits of information are carried by each trans-
mitted symbol, a message information rate of 64 kbit/s results in a transmission symbol
rate of Rs 32 ksps. If the message data bits are encoded with one-half rate FEC, code
rate Rc 12, the occupied bandwidth required for a 64 kbit/s signal is
Bocc 32,000 11 a2 12 Hz 64 11 a2 kHz (9.2)
Typical values of for satellite links lie between 0.25 and 0.5, with the higher value
being easier, and thus cheaper, to realize when conventional analog filters are used in the
transmitter and receiver. If an 0.5 RRC filter is used, the occupied RF bandwidth of
this signal is
Bocc 64 11 0.52 96 kHz (9.3)
A VSAT that is required to transmit a 64 kbit/s stream of data using QPSK modulation,
using half rate FEC and a raised cosine filter with a roll-off factor of 0.5, therefore needs
an RF channel bandwidth of 96 kHz and has a receiver noise bandwidth of 64 kHz. Note
that the roll-off of the RRC filter, while adding additional spectrum requirements, does
not alter the noise bandwidth; all RF and IF RRC filters used in digital radio links have
a noise bandwidth in hertz equal to the symbol rate in symbols per second. In practice, a
guard band will have to be added between FDMA channels so that adjacent signals do
not overlap in frequency at the satellite, and to allow the filters in the receiver that extract
individual channels to roll off between channels. Most VSATs will operate unattended for
most of the time and will be exposed to all weathers. The frequency of the fundamental
oscillator may therefore drift. For this reason, fairly large guard bands need to be designed
in, which generally leads to a spectrum allocation at the satellite around 120 kHz for each
c09.qxd 12/08/02 09:59 Page 356
Satellite
36 MHz B/W
Ku-band
downlink
Other RF channel
VSAT channels from VSAT #N
FEC
IF receiver decoder
Receiver
64-kbps
QPSK terrestrial
Hub
demodulator channel
station
Other
VSAT channels
FIGURE 9.12 Schematic of a 64-kbit/s equivalent voice channel accessing a satellite using
FDMA. The 64-kbit/s information rate is contained in a bandwidth of 96 kHz when transmitted
to the satellite. The bandwidth of the satellite transponder (from frequency f1 to frequency
f2) is divided up, or channelized, into increments of 96 kHz so that a large number of VSATs
can access the transponder at the same time. Each of the 96-kHz channels requires a certain
amount of spectrum on either side to guard against drift in frequency, poor VSAT filtering,
etc. The 96-kHz channels plus the guard bands on either side add up to a channel allocation
of about 120 kHz per VSAT. From a spectrum allocation viewpoint, therefore, a typical
36-MHz satellite transponder would permit the simultaneous access of 300 VSATs, each of
which is transmitting the equivalent of a 64-kbit/s voice channel. Because each VSAT uses a
single channel continuously on the uplink, it is often referred to as SCPC (single channel per
carrier)-FDMA.
64-kbit/s (voice) channel of the type described above. This situation is illustrated in
Figure 9.12.
In Figure 9.12, the 64-kbit/s data stream, equivalent to a PCM digital voice channel,
could be derived from a point of sale device, for example, a credit card reader, an Internet
access request, a digitized analog voice channel, etc., all of which require onward transmis-
sion over a VSAT network. The 64-kbit/s equivalent voice channel shown in Figure 9.12 is
the output of the terrestrial–satellite interface equipment, after the required emulations and
protocol conversions have taken place prior to transmission over a satellite network. This
channel from the VSAT to the hub via the satellite is called the inbound or inroute channel.
The RF transmission to the satellite from the VSAT will have a frequency that falls
within the bandwidth of a specific transponder on the satellite. If the transponder oper-
ates in a bent pipe mode, with no onboard processing, the satellite will retransmit the mul-
tiple VSAT channels on the downlink with exactly the same channelization as on the
uplink. Thus, in the example used in Figure 9.12, a transponder with 36-MHz bandwidth
would transmit 300 channels to the hub station or other VSATs in the network. If the
VSAT network is being operated in a Mesh mode, each VSAT must have a frequency syn-
thesizer that allows it to select any of the 300 possible downlink channels, and the network
must have a control channel that tells each terminal at which frequencies it should receive
and transmit.
c09.qxd 12/08/02 09:59 Page 357
It is more usual, however, for an FDMA VSAT system to operate in a Star mode,
illustrated in Figure 9.5a. The hub earth station is therefore designed to receive all 300
downlink channels. The digital signal in each channel is recovered and the address infor-
mation read off so that the hub can forward the information to the intended user. If the
required end user is external to the VSAT network (i.e., in the terrestrial network) the in-
formation is passed through the hub interface equipment and on to the public switched
telephone network (PSTN). If the required end user is within the VSAT network, or a re-
sponse has been received through the interface equipment from the terrestrial PSTN, all
of the information to be transmitted back to the various VSAT terminals is reassembled
at the hub into a return channel.
The return link from the hub to the satellite, and from thence to the individual VSAT
terminals, is not normally sent as a multitude of narrowband FDMA channels. In most
cases, the return channel from the hub to the VSAT terminals, called the outbound or out-
route channel, is a single, wideband stream in a time division multiplexing (TDM) for-
mat. In the TDM stream, the separate, low data rate, narrowband signals for the individual
VSATs are assembled in a predetermined format so that each of the VSATs can extract
the required information destined for that VSAT. Figure 9.13 illustrates the TDM down-
link concept used in FDMA VSAT Star networks. Note here the important difference be-
tween TDM and TDMA, which are often confused. Time division multiplexing, TDM, is
not a multiple access technique. Digital signals from various sources are assembled into
a single, high-speed data stream at one point, such as the hub of a VSAT network, and
then transmitted as a single continuous stream. In TDMA, several sources, such as earth
stations, transmit in sequence so that a sequence of RF signals is assembled at the satellite.
In the FDMA Star VSAT network example shown in Figures 9.12 and 9.13, the
VSAT network is quite large. Separate transponders are used for the inbound and the out-
bound channels. In many VSAT networks, the total instantaneous capacity required does
not justify two separate transponders for the inbound and outbound signals. Such a case,
using a shared transponder, is illustrated in Figure 9.14.
In designing FDMA links, care should be taken to allocate the correct transmit power
per channel in calculating the link budget so that the power spectral density is the same
for every channel. For example, if a 54-MHz transponder is operated at an output power
of 54 W, the power spectral density at the transponder output is 1 W per MHz. A single
inbound 120-kHz downlink channel transmitted by the satellite will therefore have a trans-
mit power level of (120 kHz54 MHz) 54 W 120 mW. This transmit power level is
multiplied by the gain of the antenna in the direction of the hub (less feed and other losses)
to give the effective isotropic radiated power, or EIRP, per 120-kHz channel. Operating a
transponder in an FDMA mode also requires careful power balancing to ensure linear
operation, which usually requires the output amplifier to be backed off.
FDMA operation of a transponder with a large number of simultaneous RF channels
requires significant output backoff in the transponder to obtain near-linear operation. The
nonlinearity of the transponder at output levels close to saturation causes the generation of
third-order intermodulation products that degrade the CN ratio in the SCPC channels. Am-
plifier output backoff values of between 3 and 7 dB can be found in the VSAT literature.
The value required in any particular case depends on the transponder nonlinearity charac-
teristic, the number of RF channels carried by the transponder, and the extent to which the
power spectral densities of each RF channel in the transponder is matched. Backoff at the
transponder output lowers the channel EIRP, and therefore degrades the downlink CN ra-
tio. Automated transponder loading plans are used to optimize the backoff at the transpon-
der output such that the overall CN of the link is maximized. The problem of optimizing
transponder backoff becomes particularly difficult when the bandwidth is split between the
c09.qxd 12/08/02 09:59 Page 358
“Outbound” TDM stream from hub

via satellite
f1′ 36 MHz Satellite transponder f2′
Downlink “outbound”
TDM stream from hub, via
satellite, to each VSAT terminal
Combined channel Transmission

rate ~20 Mbit/s bandwidth ~36 MHz
Demodulation and
decoding
Demultiplexing combined
channel into individual
equivalent 64 kbit/s channels
Terrestrial channel to
Pick off required 64 kbit/s 64 kbit/s User equipment
signal that is intended for this Terrestrial/VSAT
equivalent
VSAT from network interface
voice channel
demultiplexed channel stream
FIGURE 9.13 Schematic of the TDM downlink “outbound” channel from the hub, via the
satellite, to the individual VSAT terminals. The 300 individual, narrowband, “inbound” chan-
nels received at the hub from the VSATs are sent back to the VSATs in a single, wideband,
“outbound” TDM stream at a combined transmission rate of 20 Mbit/s. Each VSAT receives
the downlink TDM stream and then demodulates and decodes it (i.e., changes the modulated
band-pass signal into a baseband line code and removes the FEC). The line code is then
passed through a demultiplexer which is used to extract the required part of the stream that
contains the equivalent 64-kbit/s voice channel destined for that VSAT terminal. Carrier recov-
ery and bit recovery circuits are used in the receiver in order to be able to identify the exact
position of the required VSAT channel in time. The bandwidth of the satellite transponder
(from frequency f1’ to frequency f2’) is fully occupied in this example.
inbound and outbound directions. The end-to-end gain setting of the transponder controls
the backoff; it may be impossible to optimize the gain for both directions at the same time.
A TDM downlink format is sometimes paired with a TDMA uplink plan, particu-
larly for some advanced multimedia services that operate in Ka band (3020 GHz). Up-
link FDMA formats are not as bandwidth efficient as TDMA. On the other hand, a VSAT
c09.qxd 12/08/02 09:59 Page 359
Inbound narrowband VSAT channels Outbound wideband TDM stream
36-MHz satellite transponder

FIGURE 9.14 Illustration of a VSAT network frequency assignment in which the inbound
and outbound channels share the same transponder in the satellite. In the example here, 18
MHz of spectrum is allocated for each side of the system connection. On the uplink to the
satellite, the collection of FDMA narrowband channels transmitted by the VSATs coexists in
the same transponder with the wideband TDM stream transmitted up by the hub. On the
downlink from the satellite, the hub receives the collection of individual narrowband chan-
nels while the wideband TDM downlink stream is received by each VSAT. The precise fre-
quency assignment can vary to suit the capacity of the VSAT network.
that uses TDMA on the uplink is required to transmit at the full burst rate of the TDMA
scheme, and must therefore have a much more powerful transmitter than an SCPC-FDMA
VSAT. If the average traffic for an individual VSAT is only one equivalent voice circuit
(64 kbit/s), having to transmit at 5 Mbit/s, say, instead of 64 kbit/s can pose major diffi-
culties. The VSAT transmit power must be increased by a factor of 500064 78 to main-
tain the same uplink CN, since the earth station receiver must have a bandwidth that is
wider by the same factor. VSAT economics and bandwidth efficiency trade-offs have led
to a hybrid TDMA-FDMA approach called MF-TDMA (multifrequency TDMA). This is
illustrated in Figure 9.15.
In the MF-TDMA example shown in Figure 9.15, each of the VSATs has to transmit
at a burst rate that is approximately five times the normal single VSAT single-channel rate.
If each VSAT transmits at a message data rate of 64 kbit/s and there are five VSATs sharing
Inbound, downlink TDM stream to hub

Inbound, uplink MF-
TDMA VSAT bursts
Hub
A B C D E
FIGURE 9.15 Example of a multifrequency TDMA (MF-TDMA) scheme. In this particular case,
five VSAT terminals (A, B, C, D, and E) share the same frequency assignment; that is, they all
transmit at the same frequency. However, they each have a unique time slot in the TDMA frame
when they transmit, so that they do not interfere with each other. The bursts from each VSAT
are timed to arrive at the satellite in the correct sequence for onward transmission to the hub.
c09.qxd 12/08/02 09:59 Page 360
the same frequency, the minimum burst rate would be 5 64 kbit/s 320 kbit/s. However,
time gaps (the equivalent of frequency guard bands in an FDMA scheme) have to be left in
between each of the individual payloads within the TDMA frame to avoid overlaps due to
incorrect clock timing. The satellite transponder plan would look very much like Figure 9.14
with the exception that each of the single channel, narrowband, inbound frequency slots would
actually be allocated to a number of VSATs sharing a small TDMA frame transmitted at the
same frequency. In the same way that the hub used in the FDMA scheme detects all of the
individual inbound VSAT frequencies and then bundles the outbound return traffic into
one wideband TDM stream, the hub in the MF-TDMA scheme detects each of the inbound
MF-TDMA VSAT signals and bundles the outbound traffic into a wideband TDM stream.
CDMA schemes were originally employed in VSAT systems for encryption purposes
in military applications because they have a very low probability of intercept. However, unless
there is a severe interference environment, as in some terrestrial microwave cellular radio sys-
tems, CDMA schemes are not normally selected because they are, in general, less bandwidth
efficient than FDMA or TDMA. TDMA, in particular MF-TDMA for narrowband appli-
cations, is normally more bandwidth efficient than FDMA. Each VSAT operating in a CDMA
mode transmits with the same frequency and at the same time, and relies on the orthogonal
coding employed in a direct sequence, or frequency-hopping, spread spectrum application, to
provide complete mutual separation of the individual communications signals. CDMA comes
into its own for VSAT satellite applications when off-axis emissions from the earth terminals
are likely to cause interference into another satellite. This is illustrated in Figure 9.16.
2°
2° USAT(2)
WSAT
Geostationary orbit arc:
satellites at 2° spacing
USAT(1)
Beamwidth of
VSAT
VSAT
FIGURE 9.16 Illustration of how a VSAT can cause interference to other satellite systems.
In this example, the VSAT is transmitting to a wanted satellite (WSAT) but, because the an-
tenna of the VSAT is small, its beam will illuminate two other adjacent, unwanted satellites
(USATs) that are 2° away in the geostationary arc. In a like manner, signals from USAT
(1) and USAT (2) can be received by the VSAT, thus causing the potential for interference if
the frequencies and polarizations used are the same. Off-axis emission is closely specified
by the ITU-R and is a key element in uplink power control design. When LEO constellations
are sharing the same frequency bands as GEO systems, the use of CDMA may confer some
advantages for coordination purposes at the expense of system capacity.
c09.qxd 12/08/02 09:59 Page 361
In Figure 9.16, if the wanted satellite (WSAT), unwanted satellite one (USAT 1),
and unwanted satellite two (USAT 2), all use the same frequencies and polarizations, the
use of CDMA would prevent interference between the systems if orthogonal CDMA codes
are used. There would, however, be an increase in the noise received since each CDMA
channel would appear as a noiselike signal to every other CDMA channel. Thus, each ad-
ditional CDMA signal will incrementally reduce the CN of the channel. There is no hard
and fast CN threshold for CDMA links, as there is for TDMA and FDMA links, when
the received signal characteristics become unusable. This soft threshold allows for some
design flexibility, but care has to be taken to avoid undue errors induced by excess self-
interference from the VSAT signals. Different satellite multiple access schemes can be
used with a range of VSAT network access schemes, offering considerable flexibility to
the system designer. Figure 9.17 illustrates this schematically.
In Figure 9.17, the satellite transponder is accessed using any of the three satellite
multiple access modes: FDMA, TDMA, or CDMA. In addition to the multiple access
scheme used, the VSAT network requires some form of satellite network access control.
V
B
P
VSAT
Host F H
computer E B
P E V
Hub station
B
P
VSAT
Satellite capacity
access protocol
(FDMA, TDMA,
CDMA)
Satellite network access protocol

(satellite-efficient access protocol such as MF-TDMA/TDM
plus addressing, switching, routing, flow control, congestion
control)
Customer’s data protocol
FIGURE 9.17 The different layers of protocols used in VSAT networks (after Figure 3-1 of
reference 3). The host computer sends traffic for the VSAT network to the front-end processor
(FEP) at the hub station (the master control station of the VSAT network). The FEP passes the
traffic to the hub baseband equipment (HBE) to be formatted for transfer over the satellite link
via the selected satellite access protocol. The satellite then passes the outbound (or outroute)
traffic on the downlink to the VSATs. The VSAT baseband processor (VBP) then extracts the
relevant traffic for the user and forwards it after any necessary protocol conversion, etc.
c09.qxd 12/08/02 09:59 Page 362
The satellite network access control ensures that the most efficient access protocol is used,
for example, MF-FDMA/TDM or SCPC FDMA, for that particular satellite VSAT net-
work. Not shown in this figure are the protocol emulators that act as the interface between
the terrestrial and satellite networks, or the satellite access control that monitors switch-
ing, flow and congestion control, addressing, etc. (see Figure 9.10). Most of the signal
formats for satellite multiple accesses have been discussed in detail in Chapter 6. The sub-
section below reviews the generic case for digital satellite multiple access using TDMA
schemes.
Signal Formats
The VSAT uplink signal, the inbound or inroute channel, in an MF-TDMA multiple access
format must contain sufficient information for the intended receiver to acquire the carrier
frequency, lock onto the incoming data so that the timing of the bit stream can be ob-
tained, and then identify the start of the payload transmission. This generic procedure is
shown in Figure 9.18. More details of various TDMA/packet signal formats can be found
in Chapter 6.
Modulation, Coding, and Interference Issues

Modulation and channel coding are key considerations in determining the efficient and
error-free transfer of information over a communications channel. They also have an
impact on the potential for interference to another system and from another system. A
modulation that has a large number of bits per symbol (e.g., 64 QAM) will occupy a
relatively small bandwidth but it will require relatively high amplifier linearity and a
high CN ratio in the receiver. It is also more susceptible to interference than modula-
tions with fewer bits per symbol. High-index modulations require significantly more
margin than low-index modulations. In choosing the most appropriate modulation and
channel coding for a VSAT system, ease of implementation is also a major factor since
VSATs are very cost-sensitive. Faced with these trade-off decisions, the most common
forms of modulation used in VSAT systems are QPSK and, when spectrum efficiency
is less important, BPSK. In an ideal QPSK system with ideal RRC filters and no chan-
nel coding, a value of EbN0 of 10.6 dB will provide a BER of 106, corresponding to
a receiver overall CN ratio of 13.6 dB, ignoring any implementation margin. The CN
requirement can be significantly reduced if channel coding is applied. Coding aspects
are covered extensively in Chapter 7. Only those aspects that touch on VSAT systems
will be reviewed below.
End of previous
frame
CR BTR UW Payload
FIGURE 9.18 Generic sequence for the start of a burst from a VSAT inbound signal. When
the burst is received at the hub, the first part of the packet enables the carrier recovery (CR)
to occur, followed by the bit timing recovery (BTR). The unique word (UW) identifies the start
of the payload in the new frame.
c09.qxd 12/08/02 09:59 Page 363
Channel Coding Channel coding can take the form of a block code or a convolu-
tional code. Convolutional coding is a process where the encoding and decoding process
is applied to a group of bits in sequence rather than a bit at a time, as in a block code.
The number of bits in the encoding sequence, k, is called the constraint length of the con-
volutional code. In the decoding process, k bits are used to evaluate the value of each bit
transmitted. Since the encoding process is applied to the signal prior to transmission and
is used to detect and correct for bit errors, it is called a forward error correcting (FEC)
code. In a like manner, a block FEC code is applied to the channel prior to transmission.
Convolutional and block codes can be used together on a channel. One example is a chan-
nel that first has an inner convolution code applied to the bit sequence and then has an
outer interleaved code such as a Reed–Solomon code applied. Reed–Solomon codes com-
bine good error detection capability with high code rates. This form of concatenated cod-
ing is used extensively in many communications systems (see for example, the coding
used in direct broadcast satellite television discussed in Chapter 11), since the interleaved
coding will counter burst errors while the convolutional FEC coding will counter individual
bit errors. Direct broadcast satellite television is one example of such a coding approach,
and the recording of music on CDs is another. The encoding and decoding procedure is
illustrated schematically in Figure 9.19.
For VSAT systems that have small traffic streams, excess processing delay can add
significantly to the end-to-end link delay. This is very important for GEO systems and for
LEO/MEO systems with satellites that have large onboard processing capabilities. The
processing delay due to first interleaving a signal and then de-interleaving it adds a fixed
amount of overhead, as well as requiring buffering at both ends of the transmission link.
For this reason, Reed–Solomon outer codes are not normally added to signals that have
information rates below about 256 kbit/s, even though the lower EbN0 value for a given
BER performance is so significant (see Figure 9.20). For links that have no real require-
ment for instant response times and multimedia interactivity, but require the best BER
performance for a given EbN0 (typical of most Internet links), Reed–Solomon codes
are a very practical way of reducing the power requirements for a given link and BER
specification.
Communications
Input bit channel Output bit
stream FEC Reed–Solomon Reed–Solomon FEC stream
encoder interleaver de-interleaver decoder
Outer code
Inner code
FIGURE 9.19 Schematic of the encoding and decoding process when an “inner” and
“outer” code are applied to a communications signal. The Reed–Solomon interleaved block
code is applied after the FEC (either block or convolutional) on the encoding side. The re-
verse occurs on the decoding side. While it might look like the FEC code is outside the
Reed–Solomon code, it is the time they are applied that counts. Since the FEC is applied first
in the encoding process, it is then wrapped by the Reed–Solomon code, which becomes the
outer wrapping of the doubly coded signal.
c09.qxd 12/08/02 09:59 Page 364
10−1
10−2
Coherent PSK w/o FEC
10−3
Viterbi R = 1/2, K = 7
BER
10−4
Sequential R = 1/2
10−5
Viterbi R = 1/2,
K = 7 + Reed–Solomon
10−6
FIGURE 9.20 BER vs Eb /N0 perform-
ance of coherent QPSK for various
10−7 types of codes (from Figure 3.3.5 of
0 2 4 6 8 10 12 14 reference 4. ©ITU, reproduced with
Eb/No (dB) permission).
Interference Interference between systems operating with similar characteristics

(i.e., frequency bands, polarizations, and services) is usually the subject of intense de-
bate, particularly when a new system seeks to operate close to an existing system, in
terms of orbital separation or antenna beam directions. The interaction between opera-
tors seeking to ensure that no harmful interference is caused by, or to, their respective
systems is called coordination. The coordination process is the subject of extensive reg-
ulation by the ITU and national frequency management authorities [e.g., the Federal
Communications Commission (FCC) in the United States]. The key aspect in such co-
ordination exercises lies in determining the power radiated by the interfering station in
the direction of the interfered-with station. The calculation of the received interference
power will have four elements:
• The output power of the interfering station’s transmit amplifier
• The transmit gain of the interfering station’s antenna in the direction of the interfered-
with station
• The receive gain of the interfered-with station’s antenna in the direction of the
interfering transmissions
• The path loss between the two stations.
The interference geometry is illustrated in Figure 9.21.
Recommendation ITU-R S.7288 mandates the maximum permissible levels of off-
axis EIRP density from a VSAT transmitting in the 14-GHz band (Ku band). The relevant
part of the Recommendation is abstracted below.
“VSAT earth stations operating in the 14 GHz frequency band used by the fixed
satellite service must be designed in such a manner that at any angle specified
below, off the main lobe axis of an earth station antenna, the maximum EIRP in any
c09.qxd 12/08/02 09:59 Page 365
WSAT
Gain, Gw (dB), in
direction of
USAT wanted satellite
Gain of antenna of interfered-

with satellite, Gs, (dB),
towards the VSAT
Path to satellite which

will have a fixed path loss Main lobe
and variable loss due to and first
propagation impairments sidelobes of
VSAT
antenna
Gain, Gu (dB), in direction of
interfered-with satellite
VSAT with HPA

power of P (dBW)
FIGURE 9.21 Illustration of the interference geometry between a VSAT and a satellite of
another system. The EIRP of the VSAT toward the interfered-with satellite [P(dBW) Gu(dB)]
is the interference power from the VSAT into the interfered-with satellite. To develop the
interference link budget, the gain of the interfered-with satellite in the direction of the VSAT,
Gs(dB), would be used, plus any additional effects along the path (such as site shielding, if
used, expected rain effects for given time percentages, etc.)
direction within 3° of the geostationary-satellite orbit should not exceed the

following values:
Angle off-axis Maximum EIRP in any 40-kHz band

2.5° 7° 33 25 log dBW
7° 9.2° 12 dBW
9.2° 48° 36 25 log dBW
48° 6 dBW
In addition, the cross-polarized component in any direction degrees from the
antenna main-lobe axis should not exceed the following limits:

2.5° 7° 23 25 log dBW
7° 9.2° 2 dBW”
There are two important notes contained in the footnote to the above Recommendation.
Since the Recommendation was developed for 3° satellite spacing in GEO, the first note
indicates that the off-axis limits may need to be reduced by up to 8 dB where the satellite
spacing is 2°. The second note pertains to CDMA VSAT systems. When there are N VSATs
expected to transmit simultaneously on the same frequency, the maximum permitted EIRP
values should be decreased by 10 log N.
The rapid increase expected in satellite delivery of Internet-like traffic direct to
homes and offices has led to a multitude of proposed new constellations of satellite systems,
c09.qxd 12/08/02 09:59 Page 366
with the result that interference aspects have received a lot of study within the ITU and
ETSI (European Telecommunications Standardization Institute). The off-axis limits have
been tightened under a new proposal before ETSI, which is expected to be reflected in a
new version of ITU-R S.728 above. The new proposals are directed toward the use of Ka-
band satellites with VSAT apertures well below 1 m. The new limits proposed to ETSI
are given below.
“The maximum EIRP in any 40 kHz band within the nominal bandwidth of the copo-
larized component in any direction degrees from the antenna main beam axis
shall not exceed the following limits under clear-sky conditions, within 3 of the
geostationary orbit plane.
1.8° 7.0° 19 25 log 10 log N dBW
7.0° 9.2° 2 10 log N dBW
9.2° 48° 22 25 log 10 log N dBW
48° 10 10 log N dBW”
For CDMA systems, N is the number of VSAT earth stations transmitting simultane-
ously on the same frequency. For TDMA and FDMA systems, N 1. The proposed
recommendation by ETSI is for satellites that are spaced 2° apart in GEO. The values
in both cases (CDMA and TDMA/FDMA) may be relaxed for directions more than 3°
away from the geostationary plane since VSAT antenna patterns are normally opti-
mized for the GEO arc. During rain fade conditions, the values in the above equations
may be exceeded through the application of uplink power control (ULPC) at affected
stations, in order to overcome the rain attenuation. The effective off-axis emission lev-
els received by adjacent satellite systems are not expected to vary substantially from
that during the clear sky condition if the uplink power control system is designed
and operated properly. Uplink power control systems have the potential to be very
inaccurate9, particularly those that rely on open-loop control10,11. Carefully controlled
experiments at Ku band12 and Ka band13 have shown that there is an irreducible error
on the order of
1.2 dB for open-loop ULPC and, even with closed-loop ULPC, where
the power level is measured at the satellite, there can be time-delay constraints that
will limit the accuracy14.
9.6 VSAT EARTH STATION ENGINEERING
Antennas
The key element in a VSAT system is the earth station antenna used at the VSAT earth
stations. The small size and low transmit power of a VSAT station are the factors that
keep the price of the earth station at a level that makes a VSAT network an economic al-
ternative to a terrestrial data network using telephone lines and modems. Large antennas
are usually implemented using a symmetrical configuration, for ease of construction, with
the feed on the boresight axis. The feed can either be in front of the antenna (a front-fed
design) or behind the antenna, as in a Cassegrain or Gregorian design. Figure 9.22 shows
these three antenna configurations.
The front-fed antenna has a paraboloidal reflector with a feed at its focus. The feed
is often a scalar horn, a circular waveguide horn with a wide flare angle and corrugations
on the internal surface. A scalar horn has a circularly symmetric pattern leading to equal
beamwidths in all planes, and good aperture efficiency. When only a portion of the
c09.qxd 12/08/02 09:59 Page 367
9.6 VSAT EARTH STATION ENGINEERING 367
Symmetric
F F
F
Offset
F F
F
Front fed Cassegrain Gregorian

FIGURE 9.22 Configuration of front-fed, Cassegrain, and Gregorian antennas. The top
three configurations are axially symmetric while the bottom three are offset-fed designs that
reduce the aperture blockage considerably.
paraboloid is used, an offset reflector results. An offset reflector fed by a scalar horn is
used as the preferred configuration for most DBS-TV receiving antennas.
The basic design of the Cassegrain antenna has a paraboloidal main reflector and
a hyperboloidal subreflector. One focus of the subreflector is coincident with the fo-
cus of the main reflector; the feed is at the other focus of the subreflector. The surface
profile of both reflectors can be modified by redistributing energy across the aperture
so as to increase the aperture efficiency and reduce blockage caused by the subreflec-
tor. This is called a shaped reflector Cassegrain antenna. The Cassegrain configura-
tion is used widely for large earth station antennas. It is preferred over the Gregorian
configuration because the subreflector is closer to the main reflector and therefore eas-
ier to support. However, where there is a severe off-axis interference environment, an
offset-fed Gregorian design is the best to use. Gregorian antennas are also occasion-
ally used on DBS-TV satellites with a phased array feed to create complex coverage
regions on earth.
Transmitters and Receivers

Historically, large earth stations are assembled as discrete elements. On the receive side,
the antenna and feed components are connected by waveguide to the front-end, low noise
amplifier (LNA). Behind the LNA, a mixer/down-converter changes the signal from radio
frequency (RF) to an intermediate frequency (IF). After filtering and amplification, the
IF signal is demodulated, demultiplexed, and decoded, and the baseband signal forwarded
to the user. The transmit side is the mirror image of the receive side with the signal in-
put at baseband and the output at RF, with the LNA receiver replaced by a high-power
amplifier (HPA) transmitter. This design of earth station is typical of a hub station used
in a VSAT network. Much of this discrete component design has changed with the
c09.qxd 12/08/02 09:59 Page 368
ODU
Building roof
IFL
IDU
FIGURE 9.23 Schematic of a VSAT user setup. The VSAT outdoor unit (ODU) is located
where it will have a clear line of sight to the satellite and is free from casual blockage
by people and/or equipment moving in front of it. The interfacility link (IFL) carries the
electronic signal between the ODU and the indoor unit (IDU) as well as power cables
for the ODU and control signals from the IDU. The IDU is normally housed in a desktop
computer at the user’s workstation and consists of the baseband processor units and
interface equipment (e.g., computer screen and keyboard). The IDU will also house the
modem and multiplexer/demultilexer (mux/demux) units if these are not already housed
in the ODU.
Antenna
Feed
LNC DEM To
IFL Baseband data
processor terminal
(BBP) equipment
HPC MOD
Outdoor unit (ODU) Indoor unit (IDU)

FIGURE 9.24 Schematic of the typical configuration of a VSAT earth station (after Figure
4.1.1 of reference 4 ©ITU, reproduced with permission). The low noise converter (LNC)
takes the received RF signal and, after amplification, mixes it down to IF for passing over
the interfacility link (IFL) to the IDU. In the IDU, the demodulator extracts the information
signal from the carrier and passes it at baseband to the baseband processor. The data ter-
minal equipment then provides the application layer for the user to interact with the infor-
mation input. On the transmit operation, the user inputs data via the terminal equipment
to the baseband processor and from there to the modulator. The modulator places the
information on a carrier at IF and this is sent via the interfacility link to the high-power
converter (HPC) for upconversion to RF, amplification, and transmission via the antenna
to the satellite.
c09.qxd 12/08/02 09:59 Page 369
9.6 VSAT EARTH STATION ENGINEERING 369
introduction of digital receivers and the need to develop cheap, mass-produced VSAT
terminals.
The VSAT earth station can now be considered as two basic components: an outdoor
unit (ODU) and an indoor unit (IDU). This is illustrated in Figure 9.23
The ODU and IDU units are broken down further in Figure 9.24. Figure 9.24 gives
a typical configuration of a VSAT that has the modulator/demodulator (modem) equipment
located in the IDU. In some cases, the low noise block (LNB) or low noise converter (LNC)
and the HPB (high-power block) or HPC (high-power converter) house the complete RF
output and RF input stages of the transmitter and receiver, the up-converters and down-
converters and, in many cases, the modem. With the mass production of VSATs, the LNBs
and LNCs are being developed on application specific integrated circuits (ASICs), very
often grown as a single microwave monolithic integrated circuit (MMIC).
The hub station has a more complex design than the VSAT station since it not only
has to handle all of the inbound and outbound traffic, rather than just “thin route” traffic of
a single or a few users, but it also has to manage the network control functions. Figure 9.25
gives the general layout of the hub earth station.
Outbound
TDM channels
UC
Transmit Line
Outbound
PCE inter-
HPA modulators
face
Control Bus
IF Interface
equip-
ment
Hub to
antenna LNA Inbound host
deodulators Receive com-
PCE puters
DC
Inbound
MF-TDMA
channels
HUB
control
interface
FIGURE 9.25 Schematic of a typical hub master control station (after Figure 4.2.1 of refer-
ence 4. ©ITU, reproduced with permission). The line interface equipment handles the terres-
trial ports to the host computer. The control bus via the hub control interface allows all of
the transmit, receive, and switching functions to be carried out. The transmit processing and
control equipment (PCE) prepares the TDM stream for the outbound link to the VSATs. This
stream passes through the IF interface (the equivalent of the interfacility link of the VSAT in
Figure 9.23) to the up-converter (UC) that mixes the IF to RF. The high-power amplifier (HPA)
amplifies the TDM stream and the antenna transmits the signal. On the receive side, the an-
tenna passes the individual inbound MF-TDMA signals to the low noise amplifier (LNA) for
amplification prior to down-conversion (DC), demodulation, and so on to the user.
c09.qxd 12/08/02 09:59 Page 370
As far as possible, all of the equipment used is purchased as commercial off-the-

shelf (COTS) to reduce costs. A critical part in the economics of satellite access is the
antenna. All satellite systems require controlled access to the satellite and it is usual to
specify the frequency tolerances allowed, the range of power flux densities acceptable at
the satellite antenna, and the polarization purity of the transmission as a minimum on the
transmit side. On the receive side, a minimum antenna G/T will be set for a given eleva-
tion angle. Antenna tracking tolerances will also be specified (if needed). To keep costs
to a minimum, most satellite systems develop antenna standards (see Table 9.1) and, for
a given antenna purchased, performance and availability will be guaranteed by the space
segment provider for that antenna standard and for the power levels specified. If a non-
standard antenna is used, however, the satellite system owner will require on-site testing
of the complete earth station system to ensure compliance with specifications. This would
be extremely expensive for a VSAT user.
9.7 CALCULATION OF LINK MARGINS

FOR A VSAT STAR NETWORK
The minimum allowed carrier to noise, (CN)0, for a typical inbound VSAT link is 6.0
dB, with BPSK modulation and half rate FEC encoding, giving a BER of 106. This is
called the threshold (CN)0 value, and will vary depending on the modulation and FEC
methods used on the link. Ignoring system errors, such as mispointing the antennas or
satellite problems, the clear sky CN ratio on a given inbound link can be reduced to
the threshold (CN)0 by either a rain fade on the uplink or a rain fade on the downlink.
In a like manner, there are two rain fade margins for the outbound link: from the hub
to the satellite (uplink) and from the satellite to the VSAT (downlink). The entire two-
way system drops below the performance minimum when any one of the low margin
links drops below threshold (i.e., the design margins are exceeded). We want a link
failure to be much less likely on the satellite–hub links (either uplink or downlink), be-
cause such a failure affects every VSAT in the network. Lack of performance or avail-
ability on a VSAT–satellite link (either uplink or downlink) affects only that individual
VSAT connection.
The link fade margin is found by using the reciprocal formula (CN) ratio with
the (CN)0 value set to its threshold value. The procedure for the calculation of link
margin is different for networks using linear, bent pipe, transponders and networks that
are connected via a satellite that uses onboard processing (OBP). A bent pipe transpon-
der simply translates the uplink frequency to the downlink frequency, without any re-
generation of the signal. A rain fade on the uplink will therefore be reflected in reduced
output power of the satellite on the downlink. When OBP is present, the transmitted
power from the satellite is always held constant, irrespective of any fade on the in-
coming, uplink signal. An uplink fade, therefore, causes a worse BER to be received in
the transponder on that uplink, which will add errors to the signal to be transmitted
on the downlink from the satellite. Hence with OBP, bit error rates on the uplink and
downlink add as follows
1BER2 overall 1BER2 uplink 1BER2 downlink (9.4)
With a bent pipe transponder, link margins are calculated differently for the up-
link downlink, forward (outbound) and return (inbound) links. On the downlinks, the
c09.qxd 12/08/02 09:59 Page 371
9.7 CALCULATION OF LINK MARGINS FOR A VSAT STAR NETWORK 371
reciprocal (CN) formula is used to find the lowest value of (CN)down [ (CN)down min]
that, combined with (CN)up, gives the threshold value for (CN)0, which we have called
(CN)threshold in the equation below.
1 1 1

1C N2 threshold 1C N2 up 1CN2 down min
(9.5)
or
1 1 1

1C N2 down min 1C N2 threshold 1CN2 up
(9.6)
Rain attenuation on the uplink will cause the power received at the transponder to
fall, causing the output power from the transponder to be reduced. On an inbound link
with many VSAT stations operating in SCPC-FDMA, rain fading on one VSAT uplink
will not significantly affect the total power received by the satellite, so the transponder
operating point will not change and the transponder output power for that link will fall
linearly with uplink attenuation. Thus (CN)down at the hub station receiver will fall as
(CN)up fades, dB for dB, and
1CN2 0 fade 1CN2 0 clear sky AdB uplink fade (9.7)
On an outbound link (hub to VSAT), using TDM, rain attenuation on the uplink
from the hub will affect the operating point of the satellite transponder, because only
one signal (that of the hub station) is present. If the input to output characteristics of
the satellite transponder are known, the change in the received uplink power, which
causes a change in the operating point of the transponder, can be taken into account in
calculating the resulting downlink (CN) ratio. The downlink (CN) ratio will be higher
for a nonlinear transponder because, when the transponder is operated at the beginning
of its nonlinear region (as most are) to maximize transponder output power, a 1-dB re-
duction in input power will cause less than a 1-dB change in output power. In general,
the difference is less than 0.5 dB so, if the nonlinear characteristics of the transponder
are not known, assuming the transponder to be truly linear (1 dB for 1 dB) between
input and output will result in no more than 0.5-dB error in the calculation of uplink
attenuation margin.
The downlink attenuation margin in most VSAT networks tends to be small, and
is usually the limiting factor in the network design. Uplink power control (ULPC or
UPC) at the hub station can maintain a relatively constant power level at the input
to the transponder over the operating range of the ULPC system, which in turn in-
creases the fade margin by the available range of the ULPC system. Most ULPC sys-
tems operate in an open-loop mode and so care must be taken not to overdrive the
satellite transponder by increasing the power too much above the true rain fade level
by incorrect operation of the ULPC system. For this reason, most ULPC systems do
not add power to the uplink until at least a 1- to 2-dB rain fade has been confirmed on
the downlink, and even then, the addition of power on the uplink is always lagging
behind the true rain fade level so as to avoid both transponder saturation problems
and interference into other systems. Once the ULPC system has reached its maxi-
mum level of added power, any further increase in rain fade level will cause the overall
(CN) ratio in the VSAT receiver to fall. Figure 9.26 in Example 9.1 shows typical
(CN)0 behavior in a VSAT receiver with and without ULPC at the hub station of a
Star network.
c09.qxd 12/08/02 09:59 Page 372
(C/N )0
Outbound link (dB)
14
12 2-dB uplink attenuation

threshold for starting
ULPC adjustments
10
7-dB ULPC range
Up to 7-dB of
uplink power
8 control applied
(C/N )0 threshold for BER = 10−6

6
4
No uplink
power control
0
0 2 4 6 8 10 12
Inferred uplink attenuation (dB)
FIGURE 9.26 Outbound link performance with rain attenuation on the uplink, with and
without ULPC. Due to the error in predicting the uplink attenuation from a measure of the at-
tenuation on the downlink, the up-link power control (ULPC) does not begin to operate until
the inferred uplink attenuation reaches 2 dB. From that point on, for every additional inferred
dB of uplink attenuation, a dB of ULPC is added until the ULPC maximum is reached. The
ULPC maximum is 7 dB in this case.
9.8 SYSTEM DESIGN PROCEDURE:

EXAMPLE 9.1
Few VSAT system designs start from scratch. In most cases, either a satellite already ex-
ists and a VSAT network has to be designed around its capabilities or a VSAT network is
already in place and the opportunity occurs to upgrade it with a new satellite launch. In
both cases, the VSAT system needs to be optimized carefully, balancing the transmission
requirements of the inbound and outbound links so that the space segment capacity is used
efficiently. In the design example below, Ku-band satellite transponders are to be used in
a VSAT system. The general characteristics of the hub and VSATs are known. The design
example shows the steps needed to develop a link budget, determine capacity informa-
tion, and carry out trade-off calculations to optimize the links for a Star VSAT network.
The design of VSAT networks is dominated by the low gain and small transmitter power
of the VSAT station. The links must operate with low CN ratios at the VSAT receivers
and with low fade margins.
c09.qxd 12/08/02 09:59 Page 373
9.8 SYSTEM DESIGN PROCEDURE: EXAMPLE 9.1 373
Description of System
The satellite communications system consists of a single hub earth station, a geostationary
satellite operating with a linear transponder, and a number of VSAT stations. The system
uses two Ku-band transponders on the GEO satellite. Each VSAT station sends and receives
a 64-kbit/s data stream to and from the hub. Digital data are sent to the hub from the VSATs
by the inbound link via one transponder at a message bit rate of 64 kbit/s using binary phase
shift keying (BPSK) and half rate forward error correction (FEC) coding, giving a trans-
mitted bit rate of 128 kbps. The occupied RF bandwidth of each VSAT channel is 160 kHz,
corresponding to ideal RRC filters with 0.25. Multiple access for the inbound link is
by SCPC-FDMA with RF channels spaced 200 kHz apart to allow a 40-kHz guard band
between channels. Data from the hub station to the VSATs (the outbound link) are sent as
a continuous TDM stream of packets using a second transponder and BPSK with half rate
FEC. The VSAT antenna has a diameter of 1 m and a saturated output power of 2 W.
System Parameters
System values
Uplink frequency for transponder 1 14.02 GHz
Downlink frequency for transponder 1 11.72 GHz
Uplink frequency for transponder 2 14.08 GHz
Downlink frequency for transponder 2 11.78 GHz
Range to satellite (all stations) 38,500 km
Satellite transponders
Maximum output power 20 W
Transponder input noise temperature 500 K
Antenna gain (maximum, transmit and receive) 34 dB
VSAT station parameters

Antenna gain: transmit 41.5 dB
receive 40.0 dB
Receiver system noise temperature (clear air) 150 K
Receiver system noise bandwidth TBD
Hub station parameters

Maximum transmit power 200 W
Receiver system noise temperature (clear air) 150 K
Antenna gain at 14 GHz (transmit) 50 dB
Antenna gain at 11.7 GHz (receive) 48.5 dB
Receiver system noise bandwidth 128 kHz
Atmospheric losses
In clear air at 14 GHz 0.7 dB
In clear air at 11.7 GHz 0.5 dB
Constants
Boltzmann’s constant, k, 1.38 1023 J/K S 228.6 dBW/K/Hz
c09.qxd 12/08/02 09:59 Page 374
For this VSAT network, the requirement is to determine an appropriate transponder

loading (i.e., the number of channels passing through the transponders on the inbound and
outbound links) and determine whether the link is power limited or bandwidth limited. It
is also necessary to find whether the link is balanced (i.e., the same number of inbound
and outbound channels) so that the most efficient use is made of the transponders.
The best approach in most design exercises is to develop the basic link calculations
so that the operation of the system is understood and then to perform a trade-off analy-
sis. There are four RF links in a two-way VSAT system. The inbound link from the VSAT
to the hub station has an uplink to transponder 1 on the satellite and a downlink to the
hub station. The outbound link from the hub station to the VSAT has an uplink to transpon-
der 2 on the satellite and a downlink to the VSAT. Each of the four links has its own CN
ratio, which is calculated from a separate link budget. The uplink and downlink CN
values on the inbound link are combined to give an overall inbound (CN)0 in the hub
station IF receiver, and the uplink and downlink CN values on the outbound link are
combined to give an overall outbound (CN)0 in the VSAT receiver. Once the CN val-
ues are known, bit error ratios for the inbound and outbound links are calculated from the
inbound and outbound CN values.
Preliminary Calculations
All link budgets require knowledge of the free space path loss between the earth station
and the satellite and the noise powers in the operating bandwidths. The free space path
loss, Lp, for a range of 38,500 km is calculated as follows.
Free Space Path Loss From Chapter 4, the path loss, Lp, is given by
Lp 20 log14pRl2 dB (9.8)
where R is the range (38,500 km) and is the wavelength, in meters. The path losses at
11.7 GHz ( 0.02564 m) and 14.0 GHz ( 0.02143 m) are therefore Lp(14.0) 207.1
dB and Lp(11.7) 205.5 dB. (Note that it is usual to state dB values only to one decimal
place. If the value is to be used in additional calculations, higher accuracy may be re-
tained until the full answer is known. In this case Lp(14.0) 207.073 S 207.1 dB and
Lp(11.7) 205.515 S 205.5 dB.)
Noise Powers
Noise Power in Transponder 1, Inbound SCPC FDMA Channels The inbound
VSAT links that pass through transponder 1 have a message data rate of 64 kbit/s with
half rate FEC encoding, giving a transmitted bit rate of 128 kbps, with BPSK modula-
tion. From Section 9.5, the noise bandwidth is therefore 128 kHz, since BPSK has one
bit per symbol. From Chapter 4, noise power, Np, is given by
Np kTsBn W (9.9)
with Ts 500 K S 27 dBK, Bn 128 kHz S 51.1 dBHz, and k 228.6 dBW/K/Hz. This
gives the noise power at transponder 1 input as Ntr1 228.6 27 51.1 150.5 dBW.
Noise Power in the Hub Station Receiver, Inbound SCPC FDMA Channels
The inbound VSAT signals arrive at the hub station after being retransmitted (i.e.,
transponded) by the satellite. The noise bandwidth is still 128 kHz, since each VSAT channel
is received by a separate IF receiver at the hub station. The system noise temperature of
the hub station receiver is Ts 150 K S 21.8 dBK.
c09.qxd 12/08/02 09:59 Page 375
Hence
Nhub 228.6 21.8 51.1 155.7 dBW (9.10)
Noise Power in Transponder 2, Outbound TDM Channels The outbound TDM

bit stream from the hub station to the VSATs passes through transponder 2. Not all of the
VSATs will be transmitting and receiving simultaneously and so the TDM stream should
be sized for the likely average activity rate of the network. No information was provided
on the network size so, for this example, we will assume a starting value noise bandwidth
of 1 MHz in the VSAT receiver as an initial value for the first set of calculations. This
corresponds to a BPSK signal with a baseband data rate of 500 kbit/s and half-rate FEC
encoding. The system noise temperature of satellite transponder 2 is Ts 500 K S 27 dBK.
Hence, for the 1-MHz noise bandwidth of the VSAT receiver, the noise power at the input
of transponder 2 is
Ntr2 228.6 27 60 141.6 dBW (9.11)
Noise Power in the VSAT Receivers, Outbound TDM Channel Each of the
VSATs receives the outbound TDM stream from the hub station in a noise bandwidth of
1 MHz. The system noise temperature of the VSAT receiver is Ts 150 K S 21.8 dBK.
Hence
NVSAT 228.6 21.8 60 146.8 dBW (9.12)
Link C/N Ratios

We will now look at the inbound links from the VSATs to the hub. Each VSAT transmits
a 128-kbit/s stream made up of 64 kbit/s message bits encoded with half rate FEC and
modulated with BPSK. The bit streams from the many VSATs are received via the satel-
lite at the hub station using separate IF receivers for each VSAT channel. Each of these
separate channels has a noise bandwidth of 128 kHz, numerically equal to the symbol rate
of the BPSK signal.
We need to set a minimum allowable CN and an implementation margin (a small
factor allowed to compensate for practice vs theory), so that the link margin can be
calculated. We could begin by setting a threshold for the overall CN ratio of 10 dB at
the demodulator input. This assures that the link will provide a BER in excess of 106. A
minimum CN ratio of 10 dB is pessimistic, however, for most digital systems. A more
realistic approach for BPSK with a transmission bit rate of 128 kbit/s is to expect a cod-
ing gain of 5.5 dB with half rate forward error correction and an implementation margin
of 0.5 dB. Given that the theoretical minimum CN ratio for BER 106 using BPSK
modulation is 10.6 dB, the threshold (CN)0 ratio will be 10.6 dB implementation mar-
gin coding gain 10.6 0.5 5.5 5.6 dB. For the TDM wideband downlink data
stream, a higher implementation margin is expected, 1.0 dB, which leads to a (CN)0
threshold 6.1 dB.
The frequency used for this VSAT network is Ku band. At Ku band and above, rain
can cause appreciable signal loss for small time percentages in any given year. Since rain
is an absorbing medium, there will also be an increase in sky noise when rain is present
in the path. Depending on the rain climate at the VSAT stations and the availability re-
quired, a link margin will be needed, over and above the implementation margin required
by the demodulator, so that the VSAT system can operate normally under rain conditions.
The difference between the threshold (CN)0 ratio and the clear air (CN)0 ratio is the
link margin. For this VSAT network, we will use a clear sky (CN)0 overall of 14.0 dB,
c09.qxd 12/08/02 09:59 Page 376
giving a link margin of 8.4 dB on the inbound link and a link margin of 7.9 dB on the
outbound link.
Inbound Links
The next step in the design is to develop the link power budgets and to evaluate the CN
ratios of the inbound and outbound links. A worst-case design is adopted, with the VSAT
station on the 3 dB contour of the satellite beam, and losses included in the link budget
to account for earth station antenna mispointing, polarization mismatch, etc. If the link
works successfully under worst-case conditions, it will always work for any VSAT station
under those conditions.
VSAT Uplink Power Budget We will assume that the VSAT is located at the satel-
lite beam’s edge-of-coverage, the 3 dB contour of the satellite’s uplink (receiving)
antenna pattern. The power received at the satellite from a single VSAT, in dB, is given
by
Pr Pt Gt Gr Lp Losses (9.13)
where Pt is the transmit power (2 W), Gt is the gain of the VSAT antenna (41.5 dB),
Gr is the gain of the satellite receive antenna (34 dB), Lp is the free space path loss at
14 GHz (207.1 dB), and the losses include the satellite antenna edge of beam loss
(3 dB), clear air uplink atmospheric loss due to gases (0.7 dB), and miscellaneous
losses of 0.5 dB to account for antenna mispointing, feed loss, etc. The link budget is
shown below.
Pt (VSAT transmit power) 2 W 3.0 dBW
Gt (VSAT transmit antenna gain at 14 GHz) 41.5 dB
Gr (satellite receive antenna gain at 14 GHz) 34.0 dB
Lp (free space path loss at 14 GHz) 207.1 dB
Leoc (edge of coverage loss of satellite antenna) 3.0 dB
Latmos (atmospheric gaseous loss) 0.7 dB
Lmisc (miscellaneous losses) 0.5 dB
Pr (received power at satellite transponder input) 132.8 dBW
Uplink Inbound C/N in Transponder 1 Each VSAT is transmitting to a sepa-

rate receiver in the hub station with a noise bandwidth of 128 kHz. At the input to the
transponder receiver, the received power of each VSAT carrier signal is 132.8 dBW and
the noise power in the receiver for each VSAT channel is 150.5 dBW. The uplink CN
ratio in a noise bandwidth of 128 kHz is therefore
1C N2 trans1 132.8 1150.52 17.7 dB 5 S ratio of 58.886
Downlink Inbound VSAT Channel Power Budget at the Hub At pres-

ent, we do not know the number of uplink (inbound) VSAT channels that will be sharing
the same 54-MHz transponder on the satellite. The total power available in the transpon-
der is 20 W. For this design example we will begin by assuming that each of the inbound
VSAT signals is retransmitted by transponder 1 at an output power level of 1.0 W, and
then develop a VSAT channel downlink power budget using the relevant values of Pt, Gt,
Gr, Lp, Leoc, Latmos, and Lmisc. There will be one significant difference, however. It is normal
c09.qxd 12/08/02 09:59 Page 377
to locate the hub earth station well within the satellite antenna coverage. Thus, only 1 dB
is assumed for the edge of coverage loss, Leoc.
Pt (satellite transmit power) 1 W per channel 0.0 dBW
Gt (satellite transmit antenna gain at 11.7 GHz) 34.0 dB
Gr (hub station receive antenna gain at 11.7 GHz) 48.5 dB
Lp (free space path loss at 11.7 GHz) 205.5 dB
Leoc (1 dB contour of satellite antenna) 1.0 dB
Pr (received power at the hub earth station receiver input) 125.0 dBW
Downlink C/N at the Hub from Transponder 1 on the Inbound Links

The inbound streams from the VSATs, transmitted down by transponder 1 in the satellite,
are each received at the hub in a noise bandwidth of 128 kHz. The VSAT carrier power
in the hub is 125.0 dBW and the noise power in the hub station receiver in a noise band-
width of 128 kHz is 155.7 dBW. The CN is therefore
1CN2 hub inbound 125.0 1155.72 30.7 dB 5 S ratio of 1,174.96
Overall Inbound C/N at the Hub The overall CN of the inbound VSAT chan-
nels is calculated by the reciprocal formula where the CN values are power ratios and
not their dB values. Namely:
1 3 1C N2 overall 4 1 3 1CN2 hub inbound 4 1 3 1C N2 trans1 4
1 3 111752 4 1 3 158.882 4
0.0008511 0.0169837
0.0178348
giving 1CN2 overall 110.01783482 56.0702824 S 17.5 dB in clear sky
Note how the VSAT uplink CN ratio of 17.7 dB dominates the overall inbound CN ratio
in the hub station receiver because of the high CN ratio on the downlink to the hub station.
Analysis of the Inbound Side of the VSAT System The saturated output
power of the satellite transponder was given as 20 W. The VSAT streams are accessing
the satellite in an SCPC-FDMA format, prior to amplification in transponder 1. When
FDMA is used with a large number of simultaneous channels, the output amplifier must
be backed off to avoid the generation of unwanted intermodulation products. Typical output
backoff is 2 to 3 dB. With an output backoff of 2 dB, the 20-W (13 dBW) transponder
saturated output power is reduced to 11 dBW S 12.6 W. Each VSAT channel was as-
sumed to use 1 W of power on the downlink. Hence, the transponder can share 11 dBW
(12.6 W) among a maximum of 12.6 VSAT channels.
The calculation of the overall CN ratio on the inbound link from the VSATs showed
that the overall (CN)0 at the hub in clear sky was 17.5 dB. A 14-dB clear sky value was as-
sumed earlier to provide sufficient margin above the demodulator threshold to meet the avail-
ability requirements. There is therefore excess CN in the link between the VSAT and the
hub station and it is possible to share the downlink power between a much larger number of
VSATs and still keep the overall CN 14 dB in clear air. Using the reciprocal ratios of
CN values (NOT the dB values) and forcing the overall CN to be 14 dB, we can write
1 3 1C N2 overall 4 1 3 1CN2 hub inbound 4 1 3CN2 trans1 4
c09.qxd 12/08/02 09:59 Page 378
which yields
1 3 1252 4 1 3 1CN2 hub inbound 4 1 3C N2 trans1 4
The ratio (CN)trans1, the uplink CN on the inbound link from the VSAT, is fixed
by the transmitting parameters of the VSAT at 17.7 dB S a ratio of 58.88. The minimum
permissible downlink CN, [(CN)hub inbound], to maintain an overall (CN)0 of 14 dB is
3 1CN2 hub inbound 4 1 3 11 252 11 58.882 4 43.45 S 16.4 dB
However, earlier calculations had shown that the (CN)hub inbound was 30.7 dB. There is
therefore an excess CN of 30.7 16.4 14.3 dB S a ratio of 26.9 channels. We can
therefore increase the number of channels by a ratio of 26.9 from the 12.6 channels we
had before, when we assumed 1 W per channel, to a total of 26.9 12.6 339 chan-
nels. The power allocated to each VSAT downlink at the transponder output is now
(1 W)26.9 or (12.60 W)339 0.037 W/channel. (Remember that the 20-W output
amplifier of the transponder was backed off by 2 dB to 12.6 W.)
The next stage of the system calculation is to check whether the inbound link is
power or bandwidth limited. In most cases, power may be traded for bandwidth, giving
additional flexibility in the system design. Each VSAT transmission requires 200-kHz
RF bandwidth (160-kHz plus guard bands either side). The transponder bandwidth is 54
MHz. With 200 kHz per channel, the transponder can handle 54 MHz/200 kHz 270
channels. There is sufficient power available to carry 339 VSAT channels, but there is
only space for 270 VSAT channels. The inbound link is therefore considered to be band-
width limited with a maximum capacity of 270 channels and a clear air link margin of
8.4 dB.
Inbound Links with 270 Channels

Downlink VSAT Channel Power Budget at the Hub With 270 channels,
the power allocation per channel in the 20-W transponder is 20270 74.1 mW per chan-
nel 1 11.3 dBW. The received power will therefore drop from 125 dBW to 136.3
dBW. The noise power will remain the same, giving the CN of the inbound channels at
the hub as
1CN2 hub inbound 136.3 1155.7 dBW2 19.4 dB 5 S ratio of 876
Overall Inbound (C/N)0 at the Hub The overall (CN)0 of the inbound VSAT
channels with 270 channels arriving at transponder 1 is
1 3 1C N2 overall 4 1 3 1CN2 hub inbound 4 1 3 1CN2 trans1 4
1 3 1872 4 1 3 158.882 4
0.0115 0.0169837
0.0285
giving 1CN2 overall 1 10.02852 35.1197 1 15.5 dB
The overall CN ratio has dropped 2 dB from the case where 20 channels were assumed.
The relatively small change in CN (from 17.5 to 15.5 dB) for an 11.3-dB increase in ca-
pacity (20 channels to 270) shows that the dominant CN in the inbound link is the up-
link channels from the VSAT. The much larger number of channels (270) might, however,
require an additional backoff in the output power amplifier on the satellite to avoid inter-
modulation products for so many closely, and regularly, spaced carriers. This reduction in
output power for the inbound carriers might drop their CN to below that of the uplink
c09.qxd 12/08/02 09:59 Page 379
VSAT CN (i.e., below a value of 58.88). If that were the case, the link would be downlink
limited (satellite to hub) rather than uplink limited (VSAT to satellite) which would not
be a good design.
Outbound Links
The outbound, TDM link goes from the hub, via the satellite, to the VSAT stations. The
TDM link is a continuous bit stream at a bit rate Rb bits per second using BPSK modu-
lation and half rate FEC. The bit stream conveys data packets that are addressed to the
relevant VSATs sequentially. The CN ratio in the VSAT receiver must not be lower than
the 6.1-dB threshold that gives BER 106 with this modulation. We will initially de-
sign the outbound link for an overall (CN)0 10 dB in clear air in the VSAT receiver,
giving a link margin of 3.9 dB to overcome rain attenuation and noise temperature in-
crease (during rain) for the time percentages the availability requirements impose. In the
same way as for the inbound links, a link power budget is tabulated and the CN margins
evaluated to provide the necessary insights into the performance of the overall system.
Hub Uplink Power Budget
Pt (hub station transmit power) 100 W 20.0 dBW
Gt (hub transmit antenna gain at 14 GHz) 50.0 dB
Gr (satellite receive antenna gain at 14 GHz) 34.0 dB
Lp (free space path loss at 14 GHz) 207.1 dB
Leoc (edge of coverage loss of satellite antenna) 1.0 dB
Pr (received power at satellite transponder input) 105.3 dBW
Uplink C/N in Transponder 2 The hub station is transmitting to each VSAT

receiver in a noise bandwidth of 1 MHz. We calculated the received power of the TDM
stream earlier to be 105.3 dBW and the noise power in the satellite transponder for the
TDM stream to be 141.6 dBW. The uplink CN is therefore
1C N2 trans2 105.3 1141.62 36.3 dB 5 S ratio of 4265.86
Outbound Downlink Power Budget at the VSAT Transponder 2 carries a

single TDM–BPSK carrier, so we will assume 1-dB output backoff, giving an output power
of 12 dBW. This yields the power budget below.
Pt (satellite transmit power) 12.0 dBW
Gt (satellite transmit antenna gain at 11.7 GHz) 34.0 dB
Gr (VSAT receive antenna gain at 11.7 GHz) 40.0 dB
Lp (free space path loss at 11.7 GHz) 205.5 dB
Leoc (1 dB contour of satellite antenna) 3.0 dB
Pr (received power at the hub earth station receiver input) 123.5 dBW
Downlink C/N at the VSATs from Transponder 2 on the Outbound

Links The outbound TDM stream from the hub station, transmitted down by transponder
2 in the satellite, is received at each VSAT in a noise bandwidth of 1 MHz. We calculated
c09.qxd 12/08/02 09:59 Page 380
the received power of each VSAT carrier signal in the hub to be 123.5 dBW and the noise
power in the VSAT receiver as 146.8 dBW. The outbound downlink CN is therefore
1C N2 VSAT outbound 123.5 1146.82 23.3 dB 5 S ratio of 213.86
Overall Outbound C/N at the VSATs The overall CN of the outbound TDM
stream received by the VSAT receiver is calculated, as before, using the reciprocal for-
mula with the CN values as power ratios. Namely:
1 3 1C N2 overall 4 1 3 1CN2 VSAT outbound 4 1 3CN2 trans2 4
1 3 1213.82 4 1 3 14265.82 4
giving 1CN2 overall 203.6 S 23.1 dB in clear sky conditions
Analysis of the Outbound Side of the VSAT System The value of 10 dB

was assumed for the VSAT (CN)0 ratio in clear sky conditions to meet the availability
requirements at a BER of 106. As can be seen from the outbound calculations, a CN
value higher than 10 dB exists at the VSAT. This extra margin of CN can be used to pro-
vide a wider bandwidth (higher capacity) TDM stream or to increase the rain margin for
the same capacity. We will investigate the maximum bit rate that can be used on the out-
bound link with an overall (CN)0 of 10 dB in the VSAT receivers in clear air.
The CN for the outbound link is high from the hub to the satellite (36.3 dB) but
lower from the satellite to the VSAT (23.3 dB). To find the maximum bandwidth we can
use on both parts of the link, we need to find the reduction in CN ratio, (CN)reduction, that
will apply equally to both parts of the link and still provide an overall CN of 10 dB in clear
sky conditions. This works out to a CN reduction of 13.1 dB on each part of the link. That
is (36.3 13.1) 23.2 dB on the uplink and (23.3 13.1) 10.2 dB on the downlink.
A CN reduction of 13.1 dB amounts to a reduction ratio of 20.4. This means that the out-
bound stream from the hub to the VSAT now has a noise bandwidth 20.4 (the original
bandwidth) 20.4 1 MHz 20.4 MHz, and an occupied bandwidth of 30.6 MHz. The
equivalent bandwidth of the VSAT channels in the TDM stream is 128 kHz (ignoring any
overhead requirements for frame and bit recovery, addressing, etc.), so the maximum num-
ber of VSAT channels that can be served by the hub is (20.4 MHz)(128 kHz) 159. Since
there is not enough power to provide the required CN ratio if the 54-MHz transponder were
full to capacity in terms of bandwidth, the outbound link is power limited.
A summary of the system design and analysis so far is given in Table 9.2. Table 9.2
shows that the inbound link has excellent availability, better than 99.99% of an average
year for a network located in the eastern United States. The outbound link is much more
likely to fail, because of rain on the uplink or downlink. Availability is 99.55%, corre-
sponding to outages totaling about 40 h a year. The outages occur for approximately equal
times of 20 h on the uplink and downlink when uplink power control is not used. When
the uplink rain attenuation exceeds 3.6 dB, all VSAT terminals in the network have out-
bound BER 106. Using ULPC at the hub station significantly reduces outages caused
by uplink rain attenuation, limiting the outages to about 25 min in an average year.
System Analysis
It is now important to look at the total capacity (bit/s) that can be delivered from the
VSATs to the hub (inbound link) and from the hub to the VSATs (outbound link) to see
if there is a reasonable match. The lower of the two rates will determine the maximum
number of VSATs that can be simultaneously served.
c09.qxd 12/08/02 09:59 Page 381
TABLE 9.2 Performance of a VSAT Star Network with 1-m

Diameter VSAT Antennas and 2-W Transmit Amplifier
Link Inbound Outbound
Channel data rate 64 kbit/s 10.2 Mbit/s

Modulation BPSK BPSK
Multiple access SCPC-FDMA TDM
No. of 64 kbit/s channels 270 159
RF bandwidth occupied in transponder 54 MHz 30.6 MHz
Clear sky (C/N)up 17.7 dB 23.2 dB
Clear sky (C/N)down 30.7 dB 10.2 dB
(C/N)0 clear sky 17.5 dB 10.0 dB
(C/N)0 threshold 6.1 dB 6.4 dB
Uplink fade margin 10.3 dB 3.6 dB
Downlink fade margin 24.6 dB 3.7 dB
Uplink fade margin with 7-dB ULPC Not applicable 10.6 dB
Downlink rain attenuation margin 20.5 dB 1.8 dB
Link availability, eastern U.S. (no ULPC) 99.995% 99.55%
Link availability, eastern U.S. (7-dB ULPC) 99.995% 99.70%
The inbound link through transponder 1 was found to be bandwidth limited, with a
fully loaded capacity of 270 VSAT stations. At an information rate of 64 kbit/s per chan-
nel, the transmission rate is 17.28 Mbit/s. The outbound link (transponder 2) was found
to be power limited and could carry a maximum of 159 channels. With an equivalent
64-kbit/s data rate per channel, the outbound link can handle 10.176 Mbit/s. There is
clearly a large mismatch between the two sides of the links, which may lead to inefficient
use of the space segment.
The inbound and outbound links can be balanced if we increase the gain of the
VSAT antenna. For example, the CN of the receiving terminal can be increased by 3 dB
by increasing the diameter of the VSAT antenna by a factor of 12, from 1 m to 1.41 m
in diameter. We can maintain the same VSAT EIRP by reducing the transmitter power by
3 dB, from 2 to 1 W, which will keep the outbound link CN unaltered. The lower trans-
mitter power will reduce the cost of the transmitter sufficiently to compensate for the in-
creased antenna diameter. The new downlink (CN)0 ratio is 13.2 dB in clear air with 159
channels, which can be reduced to 10.2 dB with 318 channels. At 128 kbit/s per channel,
the TDM bit rate in the outbound channel is 40.7 Mbps. Allowing for overhead in the
packet transmissions and using = 0.4 RRC filters will restrict the data bit rate to around
34.56 Mbps, which matches the inbound data rate for 270 VSAT channels at 128 kbps.
The inbound link overall CN is controlled by fading of the uplink signal when rain
occurs between the VSAT and the satellite. The link margin for the inbound uplink is 8.4
dB and we can use almost this entire margin for rain attenuation on the uplink, because
the satellite input noise temperature is unaffected by rain at the earth’s surface. We can
assume a linear response from the transponder for a single VSAT signal, because the rain
does not affect the remaining stations, assuming wide geographic separation between ter-
minals. The inbound link’s overall (CN) ratio is dominated by the uplink CN from the
low power VSATs. Hence, each dB of rain attenuation on the inbound uplink causes
approximately 1 dB reduction in overall CN in the hub station receiver, so we have an
uplink rain attenuation margin of 8.4 dB for each VSAT on the inbound link.
c09.qxd 12/08/02 09:59 Page 382
The outbound link CN is controlled by the downlink from the satellite to the VSAT
station. When rain occurs in this link, the uplink CN at the transponder remains the same,
and the downlink CN fades. We have a fade margin of 4.1 dB on the outbound downlink,
which must be shared between rain attenuation and increase in VSAT receiver system
noise temperature. In this case, rain attenuation of 2.1 dB will cause the sky temperature
to increase from a clear air value of 40 K (based on 0.7-dB clear air attenuation) to 128 K.
The VSAT system noise temperature will increase by a similar amount, from 150 K in
clear air to 238 K in rain, resulting in a noise power increase of 2.0 dB. Thus the down-
link CN ratio will be reduced by 4.1 dB when downlink rain attenuation is 2.1 dB. The
annual availability of the VSAT terminal will depend on its location, but typically in the
eastern U.S. rain, rain attenuation of 2 dB is experienced 4 h in an average year at Ku
band. For locations with more frequent occurrence of heavy rain, a larger antenna can be
specified. A 2-m diameter antenna has 3-dB greater gain than a 1.41-m antenna, which
translates directly to a 3-dB increase in downlink fading margin.
The major parameters of the redesigned VSAT system are detailed in Table 9.3.
Table 9.3 shows that the redesigned VSAT network has a balanced capacity on each link
and can support 270 channels at 64 kbit/s each. Link availability is improved on the out-
bound links, with less than 160 min of outage caused by uplink rain attenuation in an av-
erage year, without ULPC. Rain on the downlink will cause individual VSAT stations to
suffer outages totaling 22 h in an average year. When 7 dB of ULPC is available at the
hub station, outages caused by uplink attenuation on the outbound link are reduced to less
than 20 min in an average year.
Figure 9.26 illustrates the effect of using ULPC at the hub station, based on typical
14-GHz rain attenuation in the eastern United States. Uplink power is increased once the
uplink attenuation exceeds 2 dB and then compensates for rain attenuation until the at-
tenuation exceeds 9 dB. Without ULPC, the (CN)0 for the outbound link reaches the
6.4-dB threshold with an uplink attenuation of 4.5 dB. With a ULPC margin of 7 dB, the
threshold (CN)0 is reached when the uplink rain attenuation exceeds 11.5 dB.
TABLE 9.3 Performance of an Improved VSAT Star Network with

1.4-m Diameter VSAT Antennas and 1-W Transmit Amplifier
Link Inbound Outbound
Channel data rate 64 kbit/s 17.3 Mbit/s

Modulation BPSK BPSK
Multiple access SCPC-FDMA TDM
No. of 64 kbit/s channels 270 270
RF bandwidth occupied in transponder 54 MHz 51.3 MHz
Clear sky (C/N)up 17.7 dB 20.9 dB
Clear sky (C/N)down 30.7 dB 10.9 dB
(C/N)0 clear sky 17.5 dB 10.5 dB
(C/N)0 threshold 6.1 dB 6.4 dB
Uplink fade margin 10.3 dB 4.5 dB
Downlink fade margin 24.6 dB 4.3 dB
Uplink fade margin with 7-dB ULPC Not applicable 11.5 dB
Downlink rain attenuation margin 20.5 dB 2.1 dB
Link availability, eastern U.S. (no ULPC) 99.995% 99.65%
Link availability, eastern U.S. (7-dB ULPC) 99.995% 99.80%
c09.qxd 12/08/02 09:59 Page 383
9.9 SOME NEW DEVELOPMENTS 383
In some VSAT service applications, however, a much larger outbound capacity may
be required than inbound, particularly if the link has been designed for Internet connec-
tions where the “download” channel will need a much higher throughput than the “upload”
request from the user. If the VSAT system considered above in the design example were
matched so that the same transmission rate is on each side (inbound and outbound),
the maximum capacity would be limited to 159 equivalent channels. As is the case for
most Star VSAT networks, the number of outbound channels is limited by the downlink
CN ratio. The downlink CN ratio can be increased by increasing the satellite transmit
power, increasing the VSAT antenna diameter, or reducing the rain margin. Additional
trade-off analyses can also be performed that consider other aspects of the overall network
operation.
9.9 SOME NEW DEVELOPMENTS
A number of significant technology enhancements are being planned for new VSAT sys-
tems. These are being driven in two main technical areas that seem to develop rapidly in
parallel: digital signal processing and microminiaturization. The first enables large amounts
of complex “transactions” to take place in microseconds, while the second enables the
product to be built into ever-smaller packages.
Perhaps the change that had the biggest impact on satellite developments over the
1990s was the shift from military spending being the major financial input for space ven-
tures to entrepreneurial commercial funding. With that major change has come market
driven economics that force the end product to be reliable and low-cost. Behind all of the
developments has been the surge in demand for communications capabilities that can bring
any stream of bits into any terminal, anywhere: the ultimate multimedia portable unit.
The term multimedia implies the capability of handling any traffic stream, whether
it be voice, video, fax, or data. The mixing of short voice packets with long data pack-
ets is now achieved through asynchronous transfer mode (ATM) techniques. The trans-
fer of the ATM packet or frame is typically handled through virtual containers or circuits
which exist at the time the connection is set up, but are torn down once the transaction
is completed, thus freeing the capacity to handle other traffic. The sizing of the virtual
container is set by the link requirement but must be within the overall capacity of the
channel: there would be no point in attempting to transfer the equivalent of 1 Mbit/s
down a 64-kbit/s channel. The access technology must also be capable of mixing the var-
ious virtual containers within the available channel capacity. Equally important will be
the ability of the intervening nodes to handle these virtual containers. One such node is
the satellite.
Most advanced satellites are designed to handle the vast increase in Internet-like
traffic and they take advantage of onboard processing (OBP) technology. OBP was once
considered to be overly expensive in both mass and power for commercial satellites but
is now the enabling technology of most new systems. The Iridium constellation was the
first commercial satellite system to employ OBP extensively although, in the civilian area,
INTELSAT VI was the first to use onboard, satellite switching between beams. In the Irid-
ium system, the uplink signals from the handheld units are mixed down to baseband within
the satellite payload, the header information stripped off the frames, and the traffic re-
assembled and routed to the appropriate output port, whether it be for downlinking to
another handheld unit, downlinking to the gateway (hub) earth station, or cross-linking to
another satellite. All of the new satellites being proposed for Internet service at Ka band
(3020 GHz) have OBP capabilities. This is not a trivial undertaking. Not only must the
c09.qxd 12/08/02 09:59 Page 384
OBP design allow for a large variety of traffic types and permit flexible connection between
any points in the coverage, it must also be incredibly reliable. Most satellites are designed
for at least a 10-year lifetime and, given the need to design, develop, and launch satellite
networks, system designers must attempt to predict the likely customer requirements 15
to 20 years hence. This is a daunting task given the typical Internet product lifetime of
about 2 years. Based on this, a linear transponder, for all its analog antiquity, will likely
remain the best satellite resource for the foreseeable future. Further, the simplicity of the
network architecture for GEO systems over those required for non-geostationary satellite
systems, argues for linear transponders on GEO satellites as the most cost effective solu-
tion for almost all commercial undertakings.
In the same way that the hub and VSAT interface units for transmission emulate ter-
restrial packet connections over the satellite, the virtual containers in terrestrial ATM con-
nections must be converted to satellite virtual packets (SVPs) for transmission through a
satellite OBP payload. The SVPs will be a common baseband element at the link layer of
the ISO/OSI stack. The SVP approach permits fast packet switching architectures to be
used and builds on the terrestrial protocols and applications that have already been de-
veloped for ATM systems. It also permits application specific integrated circuit (ASIC)
chips that are being employed for ATM networks to be used in the user terminals. MF-
TDMA uplink access protocols will be the typical multiple access technique used and
multi-carrier demodulation, demultiplexing, and decoding (MCDDD) of the incoming data
stream will precede the OBP payload on the satellite. Since the direct-to-home (DTH)
Internet services from these satellites will be at Ka band, they will suffer appreciable
degradations in rain. The OBP satellite payload will therefore have capabilities to moni-
tor the incoming bit stream and request changes in power or modulation to counteract the
rain impairments detected. These requests are sent to the individual VSATs or to the
network control station for onward routing to the user.
There is a large potential market for two-way Internet connections to the home by
satellite. To be successful, the service must offer high data rates in both directions, at a
cost that is acceptable to the customer.
9.10 SUMMARY
Very small aperture terminals (VSATs) have become DBS-TV would have been impossible without
a part of everyday life around the world. In highly MPEG 2. In a like manner, digital signal processing
developed regions, they act as links in the retail and the integration of many functions into one chip
chain, taking point of sale data from automated ter- set have made DTH terminals a practical reality.
minals (e.g., a gas pump) from the customer to the Whether the Internet DTH terminals will be a com-
credit card authorization center. In all regions of the mercial success remains to be seen, as many factors
world they have become the fastest growing segment have to be balanced in this endeavor. Some of these
of video distribution: the direct-to-home (DTH) re- factors are relatively straightforward to ascertain:
ceivers in the direct broadcasting service (DBS) us- link margins, capacity allocations, and trade-off be-
ing geostationary satellites. Soon, DTH terminals tween outbound and inbound links. Others are more
will be two-way links via satellites in the Global In- subjective, such as customer acceptance of outages
formation Infrastructure, better known as the GII or at Ka band and the ability of onboard processing
simply the Internet. The development, and more (OBP) payloads to adapt to changes in traffic mixes
importantly the universal acceptance, of VSATs re- over the lifetime of a particular satellite system.
quired a number of technological breakthroughs, the Nevertheless, VSATs will always be a major part
most important of which were digital compression of every satellite system, growing in importance as
techniques and very high-density integrated circuits. new enterprises seek to provide multimedia streams
It can be argued that the remarkable success of directly to customer premises.
c09.qxd 12/08/02 09:59 Page 385
PROBLEMS 385
REFERENCES
1. Satellite Communications: the First Quarter Century of 9. J. E. ALLNUTT, Satellite-to-Ground Radiowave Propa-
Service, DAVID W. E. REES, John Wiley & Sons, New gation, Peter Perigrinus Ltd. (for the IEE), 1989.
York, 1989 ISBN 0-471-62243-5. 10. L. CASTANET, J. LEMORTON, and M. BOUSQUET, “Fade
2. Intelsat Earth Station Standards are available from Mitigation Techniques for New SatCom Services at Ku-
Intelsat, 3400 International Drive, NW, Washington DC, Band and Above: A Review,” COST 255 First Interna-
20008-3098. They are also available on the web at tional Workshop on Radiowave Propagation Modelling
http://www.intelsat.com for SatCom Services at Ku-Band and Above, WPP-146,
3. Intelsat VSAT Handbook, September 1998. Available pp. 243–251, October 1998.
from Application Support and Training, Intelsat, 3400 11. H. VASSEUR, M. CZARNECKI, L. CASTANET, and M.
International Drive, NW, Washington DC, 20008-3098. BOUSQUET, “Performance Simulation of a Ka-Band
4. “VSAT Systems and Earth Stations,” Supplement No. 3 VSAT Videoconferencing System”, COST 255 First In-
to the Handbook on Satellite Communications, Interna- ternational Workshop on Radiowave Propagation Mod-
tional Telecommunications Union, Geneva, 1994 (for up- elling for SatCom Services at Ku-Band and Above
dates on this handbook, please refer to http://www.itu.int). WPP-146, pp. 227–234, October 1998.
5. D. RAYCHAUDHURI and K. JOSEPH, “Channel Access Pro- 12. INTEL-1474 Final Report, Demonstration of Advanced
tocols for Ku-Band VSAT Networks: A Comparative Networking Concepts, COMSAT Laboratories, February
Evaluation,” IEEE Communications Magazine, Vol. 26, 1997.
No. 5, pp. 34–44, May 1988. 13. A. W. DISSANAYAKE, “Application of Open-Loop Up-
6. Multiple Access Communications: Foundations for link Power Control in Ka-Band Satellite Links,”
Emerging Technologies, N. ABRAMSON, ed., IEEE Press, Proceedings of the IEEE, Vol. 85, No. 6, pp. 959–969,
New York, 1993. June 1997.
7. H. D. CLAUSEN, H. LINDER, and B. COLLINI-NOCKER, 14. D. G. SWEENEY and C. W. BOSTIAN, “Implementation
“Internet over Direct Broadcast Satellites,” IEEE Com- Adaptive Power Control as a 3020 GHz Fade Coun-
munications Magazine, Vol. 37, No. 6, pp. 146–151, termeasure,” IEEE Transactions on Antennas and Prop-
June 1999. agation, Vol. 47, No. 1, pp. 40–46, January 1999.
8. Recommendation ITU-R S.728, Maximum Permissible
Levels of Off-Axis E.I.R.P. Density from Very Small
Aperture Terminals (VSATs), 1992.
PROBLEMS
1. a. What does the acronym VSAT stand for? b. Give three examples of leapfrog technology.
b. What is the typical range, in meters, of the aper- A country with an emerging economy is seeking to
ture diameter for a VSAT operating with a Ku-band increase the communications capability in its interior,
satellite? which, for the present, lacks a significant terrestrial
c. As a direct broadcast satellite service (DBSS) oper- communications infrastructure. They plan to do this,
ator, you want to identify the appropriate receive an- in part, with a VSAT/WLL architecture. A typical
tenna to use in the home market. Calculate, and set down VSAT will handle a two-way T1 stream (1.544
in tabular form, the gain (in dB) and 1-dB beamwidth Mbit/s), which is capable of incorporating 24, 64
of the following antenna diameters: 0.1, 0.2, 0.3, 0.4, kbit/s digital voice/data channels.
0.5, 0.6, 0.7, 0.8, 0.9, and 1.0 m. Assume a frequency c. If a linear satellite transponder, SCPC approach is
of 12 GHz, an antenna efficiency of 55%, and that the used, what RF bandwidth will this VSAT T1 stream
1-dB beamwidth is half that of the 3-dB beamwidth. require on the satellite? Assume no FEC is used, a
d. If users are able to point their antennas to within root raised cosine filter roll-off factor 0.3, and

0.5° and require a minimum gain of 30 dB, what QPSK modulation is employed.
antenna diameter range is available to the users? d. If half rate FEC is used, what is the occupied satel-
e. Given this acceptable range of antenna diameters, lite bandwidth now?
which one of these antenna diameters would you e. What is the noise bandwidth in cases (a) and (b)
choose, stating your reasons? above?
2. a. Explain in your own words what “leapfrog f. If realistic guard bands on the satellite are assumed,
technology” is. and ignoring satellite power issues, how many T1
c09.qxd 12/08/02 09:59 Page 386
streams can be handled by a 72-MHz transponder in Assume the following:

cases (a) and (b) above? (Note: state your guard band • Altitude of satellites is 1400 km
requirements clearly.) • Minimum operational elevation angle is 40°
g. If more than 24 channels are required by the WLL • Delay introduced by bent pipe satellite is 0 ms
through the VSAT, how would you go about accom-
• Delay introduced by any switching element (ground
plishing this if (i) these are all voice channels; (ii)
or space) is 15 ms
these are a mix of voice and data channels; (iii) these
are all data channels? • All antennas track accurately (earth terminal and
ISL antennas)
3. a. Explain what Mesh and Star architectures are
• Gateway earth station located relatively close to
in a VSAT network.
SOHO terminals
b. Give two advantages and disadvantages of the
d. What is the propagation path delay from a termi-
Mesh and Star architectures.
nal on the ground to one of the NGSO satellites when
c. Give the three major types of multiple access the satellite is viewed at the lowest permissible ele-
schemes that are used in satellite systems. vation angle?
d. Which one of the three multiple access schemes e. If a typical trans-Atlantic link via Teledesic requires
identified in (c) is the closest to the ALOHA multi- three satellites in the link (up to the first satellite,
ple access scheme? across to the second, over to the third satellite and
e. What are the advantages and disadvantages of an down to the end user terminal from the third satellite)
MF-TDMA access scheme from a system perspective and the user terminals at both ends operate at 40° el-
(i.e., how does this access scheme affect the earth ter- evation angle, what is the total one-way signal delay
minal and the satellite payload design)? if the satellite to satellite to satellite path is 6000 km?
f. What does MCDDD mean and how will it affect f. What is the one-way signal delay for the Skybridge
the design of the satellite payload? system between the same two terminals in (e) above
g. Why has a TDM approach been adopted for most if the terrestrial link is 8000 km and the satellites are
downlink applications for digital VSAT and Internet at 40° el. angle to the gateways?
applications to small terminals? g. If no “spoofing” is performed at the satellite seg-
4. a. What do the symbols ACK and NAK mean when ment–earth segment interface, what is the minimum
applied to a packet switched communications system? window size required for the Teledesic system in (e)
b. What is meant by the term window when used with and the Skybridge system in (f), ignoring the length
packet communications systems? of the messages being sent?
c. What does the term spoofing mean when applied 5. When power and bandwidth issues have been op-
to the interface between dissimilar networks. timized, the fundamental limitation for most wireless
An Internet service provider (ISP) is evaluating a systems is nearly always interference. Interference can
couple of non-geostationary satellite orbit (NGSO) be caused deliberately, as in the jamming of an op-
constellations that are being designed to provide posing entity’s signals, or unintentionally, as in a mis-
global Internet access directly to small office/home pointed antenna (due to high wind) or an inadvertent
office (SOHO) terminals located on the SOHO prem- increase in amplifier power (due to operator error).
ises: Skybridge and Teledesic (see Table 10.7). Both These are usually classified as short-term interferers
of the NGSO systems have opted for orbital altitudes and there is usually no way to protect a system against
of approximately 1400 km, but there the similarities such interferers other than to clear them down. Of
end. Teledesic employs an onboard processing pay- more interest to commercial systems is the potential
load and an end-to-end system architecture using for long-term interference caused by nearby systems.
intersatellite links (ISLs) between the satellites to A Ku-band VSAT system is being designed for a
complete the network. Skybridge employs a bent pipe new service to be offered to two-way SOHO terminals
approach on the satellite with the long-distance com- close to a major urban center. One equation used to de-
ponent being carried over the terrestrial network, and termine the maximum EIRP permitted in any 40-kHz
just the end elements of the network employing the band at an angle off the main-beam axis between
satellites (similar to Figure 10.28). In the question be- 2.5° and 7° is given by 33 25 log dBW. This equa-
low, the information is for illustrative purposes only tion is generally used for satellite spacings of 3°.
for use in this example and does not reflect in any a. Using the above off-axis equation, what is the max-
way the true system parameters of either of the two imum off-axis EIRP permitted 3° from the antenna
systems. main beam axis?
c09.qxd 12/08/02 09:59 Page 387
PROBLEMS 387
There is a range of antennas being considered for the EIRP incorrectly may violate agreed interference
the SOHO 14-GHz uplink. Assuming an antenna limits.
efficiency of 55%, A Ka-band VSAT SOHO terminal employs a fixed
b. What is the on-axis gain value, in dB, for anten- increment of uplink power control under rain fading
nas having antenna aperture diameters of 1, 0.8, 0.6, conditions. In clear sky, the EIRP is at its nominal
0.4, and 0.2 m at a frequency of 14 GHz? level; under rain fade conditions, 7 dB of additional
c. What are the 3-dB beamwidths of the antennas in EIRP is switched in as a single-step command.
part (b) above? a. If the VSAT uplink EIRP in clear sky conditions
If we assume that the 6-dB beamwidth is 1.5 operates 3 dB below the agreed interference thresh-
the 3-dB beamwidth, the 10-dB beamwidth is 2 the old, what is the smallest rain fade level at which the
3-dB beamwidth, and reasonable interpolations can ULPC may be switched on to provide the fixed in-
be used for beamwidth values close to these, crement of 7 dB of additional EIRP without violat-
ing the interference limit? In this part of the problem,
d. What is the gain of each of the antennas in part
assume no error in measuring the rain fade level or
(b) 3° away from the main beam axis?
setting the EIRP level.
e. If there are three choices of output amplifier, 1, 0.5,
b. If the rain fade measurement accuracy is
0.5 dB,
and 0.1 W, and assuming no losses between the out-
what is the revised answer to part (a)?
put amplifier and the antenna, what is the EIRP value
on-axis for each of the antennas in part (b) using these c. If the rain fade measurement accuracy is
0.5 dB
three possible output amplifiers? and the EIRP level may only be set to an accuracy of

0.5 dB, what is the revised answer to part (a)?
f. What is the EIRP value of each of the antenna plus
d. Some ULPC systems require the uplink signal
amplifier combinations in part (e) 3° away from the
to be detected and measured on the satellite and a
main beam axis?
downlink data channel from the satellite contains the
g. Which of the antenna and amplifier combinations required ULPC information for the VSAT SOHO
above meet the off-axis limitations of part (a)? terminal. If the round-trip delay, which includes all
h. What additional EIRP allowance needs to be made of the propagation, processing, and routing delays, is
if the system must now meet the interference re- 2 s and the maximum rain fade is 1 dB/s, how would
quirements for satellites spaced 2° apart rather than this change your answers to parts (a), (b), and (c)?
3° apart as in parts (a) through (g)? e. Interference rules permit the EIRP commanded by
6. Many VSAT systems will operate close to accept- ULPC systems to exceed the long-term interference
able long-term interference limits, both in terms of limits for small time intervals. If (i) this interval is
the interference they can tolerate from similar nearby 60 s; (ii) we assume a worst-case scenario of the path
systems and the interference they cause to similar sys- attenuation going from a very large value to zero atten-
tems nearby. It is also very possible that many of these uation in zero seconds (which happens occasionally
VSAT systems will operate close to the performance due to intermittent accidental blockage of the antenna
and availability minima acceptable to the services be- aperture); (iii) the limit of the path attenuation meas-
ing offered. There are a number of techniques that uring equipment has been set so as to accommodate
may be used to increase the margin available in such the rain attenuation measurement accuracy, the ULPC
cases, one of them being uplink power control (ULPC level measurement accuracy, and the ULPC amount
or UPC). The amplifier power is increased during (7 dB), what is the minimum time constant of the path
those periods when rain attenuation occurs in the path attenuation measuring equipment that is required to
so that the received CN remains the same. Increasing enable the interference criteria to be met?
c10.qxd 28/08/02 19:38 Page 388
CHAPTER 10
LOW EARTH ORBIT AND
NON-GEOSTATIONARY
SATELLITE SYSTEMS
For 6 years after the launch of Sputnik 1 on the night of October 4/5 1957, all of the space
vehicles launched were either inserted into low earth orbit (LEO) or they were small sci-
entific research vehicles dispatched on missions to the earth’s moon or to the inner plan-
ets: Mars and Venus. The primary reason for the use of LEO was the generally small throw
mass of the launchers then in use. The throw mass of a launcher is the mass of the space-
craft that is ejected from the spent rocket. The throw mass includes both the payload for
the mission and the spacecraft bus system within which it will function. It also includes
any additional rocket motors and fuel that will be needed to boost the spacecraft into the
final trajectory and/or to maintain that trajectory. While the throw mass of the launchers
used by the former USSR in that period was significantly higher than that of the U.S.
launchers, political reasons dictated their capabilities be used in noncommercial areas. It
was therefore left to the United States to open up the geostationary (Clarke) orbit1 in 1963
with Syncom 1 and with it, the commercialization of space. The Clarke orbit, more usu-
ally called the geostationary earth orbit (GEO) or geostationary satellite orbit (GSO), is
a unique resource that has enabled the generation of many billions of dollars in revenues
per year from communications satellites and the associated launchers. Communications
satellites and their launchers were the only commercial space ventures in the twentieth
century that had any significant return on investment. This may change with the new gen-
eration of non-geostationary satellite orbit satellite constellations currently under devel-
opment, in deployment, and in operation for a variety of commercial ventures, although
the first ventures have unfortunately been conspicuous failures.
The terms low earth orbit (LEO) and medium earth orbit (MEO) are generally used
for specific orbit altitude ranges, for reasons that we will see later. LEO satellites are con-
fined between an upper orbit altitude of about 1500 km and a lower orbit altitude dictated
by atmospheric drag (generally around 500 km). MEO satellites have a lower orbit alti-
tude of around 1500 km and an upper bound set by the GEO altitude of around 36,000
km. Most MEO systems, however, orbit in the 10,000 to 15,000 km range.
LEO and MEO satellites—now generally referred to as non-geo-stationary orbit
(NGSO) satellites—have been used in a variety of roles. From an era in the late 1950s
when every launch made front-page news we have now become somewhat blasé about
satellites: they have become part of everyday life, much like computers and the Internet.
NGSO satellites brought us the first communications satellite (SCORE), the first pictures
of our cloud cover for weather forecasting (TIROS), the first navigation aids in space
(TRANSIT), the first live television pictures across oceans (TELSTAR), the first Geo-
graphic Information Systems pictures of the earth (SPOT), the first infrared, ultraviolet,
and X-ray view of the universe from outside the earth’s atmosphere and, of course, the
first manned missions (Vostok and Mercury). Each of these missions has been succeeded
388
c10.qxd 28/08/02 19:38 Page 389
by more complex satellites with more advanced capabilities: soon the International Space
Station (ISS) will assume scientific missions accompanied by one, or more, free-flying
modules and, for the first time, watchers on the earth with their naked eye in daytime may
be able to see a satellite as it passes overhead. Already, in early 2002, the ISS had an or-
bital mass in excess of 250,000 lb and dimensions close to a Boeing 747. As the satellite
missions became more complex, the requirements for the specific orbits became more pre-
cise. Some satellites have to be very close to the earth, some in highly elliptical orbits,
and yet others in orbits with a plane that matches the view angle to the sun. This chapter
reviews the different earth orbits available and what missions may use them to advantage.
10.1 INTRODUCTION
The geostationary orbit has been the preferred orbit for satellite communication systems
for 35 years, and is likely to continue to be the orbit that provides most of the revenue
for satellite system operators. The reason is simple: more bits can be sent per dollar of
capital investment when a satellite is in a geostationary orbit than in any other orbit. This
was realized quite early in the development of satellite communications, and Intelsat,
which was the first provider of commercial satellite systems, developed a series of geo-
stationary satellites, beginning in 1965 with Early Bird (INTELSAT I). Commercial and
national satellite systems followed in the 1970s and 1980s, all using GEO satellites. Direct-
to-home (DTH) satellite television broadcasting, one of the most successful applications
for satellite communication systems, also requires GEO satellites so that customers can
use small fixed dish antennas. In such a DBS-TV system, the major investment is in earth
stations, not in the satellite. Ten million earth stations bought for $250 each, for example,
cost $2.5 billion, well in excess of the cost of a cluster of GEO DBS-TV satellites.
There are some specialized applications that require non-geostationary satellites.
Surveillance of the earth’s surface, for both military data gathering and earth resources
applications, requires satellites in low earth orbit that cover the entire surface of the earth.
Satellites providing global navigation, such as the Global Positioning System (GPS) con-
stellation, must utilize orbits that place the satellites in widely spaced positions in the sky,
as seen by the receiver. Some of the satellites can be in GEO, but most must be in in-
clined orbits with an even distribution over the earth’s surface. GPS uses 24 satellites in
orbits with an altitude of 20,000 km and an inclination of 55°.
Mobile satellite communication systems demand an earth station with a low gain
antenna that has a near omnidirectional pattern. A GEO satellite used for communication
with a satellite telephone that is handheld, like a cellular telephone, requires a very large
antenna with hundreds of beams to achieve a very high gain. The high gain satellite an-
tenna is needed to compensate for the low gain of the antenna employed by the user’s
telephone handset. An alternative to a GEO satellite with a high gain antenna is a LEO
or MEO satellite constellation with a smaller multibeam antenna. Because the satellite is
not geostationary, a large number of satellites is required to maintain continuous cover-
age. The Iridium system used 66 satellites in LEO, for example, to provide continuous
global coverage.
Building, launching, and maintaining a constellation of communication satellites in
low earth orbit is expensive. When low earth orbit satellite constellations were first pro-
posed for mobile satellite services, the satellites were envisaged to be small, simple, and
low cost compared with GEO satellites. Early estimates for the cost of the Iridium sys-
tem, for example, were between $1 billion and $2 billion. As the development of the LEO
systems progressed, the satellites became more and more complex and their cost steadily
c10.qxd 28/08/02 19:38 Page 390
390 CHAPTER 10 LOW EARTH ORBIT AND NON-GEOSTATIONARY SATELLITE SYSTEMS
increased, becoming comparable to the cost of GEO satellites. The satellites used by the
ICO global system (now called New ICO), for example, are actually modified versions of
a large GEO satellite, the Hughes (Boeing) 601 design. Since any LEO or MEO system
requires many more satellites than a GEO system serving the same region, the cost of the
LEO or MEO system will exceed the cost of the equivalent GEO system.
The Iridium system, with 66 satellites in low earth orbit, eventually cost over $5
billion, compared with a typical cost of $250 million to launch and maintain a large GEO
satellite. Iridium failed as a commercial venture because the final cost greatly exceeded
initial projections, and the system was unable to attract a sufficient number of customers
quickly enough. Debt repayments on the high capital cost of the system came due before
the customer base had built to a large enough size to provide substantial revenue. How-
ever, analysis of the cost per bit transmitted through an Iridium satellite shows that it is
much higher than the cost per bit for a GEO satellite, and any LEO system must there-
fore be able to offer considerable advantages to its customers over that of an equivalent
GEO system if it is to succeed commercially. It remains to be seen whether the other non-
geostationary mobile satellite systems can succeed where Iridium failed. As the twenty-
first century starts, prospects do not look bright for NGSO systems with Iridium, Orbcomm,
Globalstar, and ICO filing for bankruptcy protection. ICO has emerged from the Chapter
11 filing as New ICO, but is still struggling with identifying its mission. Initially con-
ceived as a mobile satellite system provider, emphasis has moved to the provisioning of
Internet-like service to mobile customers as a preliminary to Teledesic.
This chapter discusses a number of applications and satellite systems that are not in
GEO orbit, beginning with those in simple, circular, equatorial orbits; moving through sim-
ple inclined orbits to those with high eccentricity; and then reviewing those that take ad-
vantage of specific attributes of their orbit for observations (sun synchronous orbits) or the
provisioning of navigation services through half-sidereal periodic orbits (GPS). The so-
called inclined orbit GEO satellites are not discussed in this chapter. These satellites, once
fully stationary GEO satellites, have had their in-orbit operational life extended by removing
station keeping in the N–S direction while maintaining E–W station keeping so that the
average subsatellite point remains nominally the same. Such inclined-orbit operation was
first started after an unusual run of launch vehicle failures in the 1986 time frame when
every single type of commercial satellite launcher failed (including the tragic loss of the
space shuttle Challenger). The up to 2-year hiatus in some satellite replenishment programs
forced inclined orbit operation of GEO satellites on all service providers. Currently, such
SIDEBAR
The first spacecraft launches relied on terrestrial radar could not be successful unless the designed orbit was
tracking and guidance commands transmitted from the achieved within the specified tolerance. In some cases,
ground. This, and the relatively crude control capabil- the mission was for a single spacecraft (such as a me-
ities of the rockets themselves, dictated relatively wide teorological satellite) while, in others, a constellation
error bounds for the intended orbit. Indeed, achieving of spacecraft would be required to achieve the mission
orbit in those early days—any orbit—was declared a goals. In all cases, careful analysis of the mission goals
success! Rapid advances in rocketry, which included led to the selection of a particular orbit altitude, ellip-
the ability for multiple restarts of high-energy upper ticity, and inclination and system architecture (number
stage engines, and the inclusion of sophisticated on- of satellites, number of planes, spacing of satellites
board guidance computers, quickly enabled spacecraft within the plane, connectivity, etc.). Quite often, tight
mission planners to design with some confidence or- launch windows were also dictated—specific time
bits that were mission-specific. That is, the mission periods when the launches had to be executed.
c10.qxd 28/08/02 19:38 Page 391
10.2 ORBIT CONSIDERATIONS 391
problems are avoided by designing the orbital maneuvering life (OML) to be many years
longer then the orbital design life (ODL). This aspect is discussed in Chapter 2.
In the sections that follow, we will examine the parameters that need to be deter-
mined in the selection of an orbit that will achieve given mission goals. Only earth orbit
missions are considered; spacecraft missions requiring escape velocity from the earth are
beyond the scope of this book.
10.2 ORBIT CONSIDERATIONS
Once in orbit, the motion of a satellite is determined by orbital mechanics, as discussed in

Chapter 2, Orbital Mechanics and Launchers. However, while the satellite moves in such
a way as to balance centrifugal and centripetal forces, the earth is also in motion beneath
it. As well as rotating once a sidereal day, the earth also moves around the sun; and the so-
lar system, with the sun at its center, is orbiting around the center of the home galaxy, the
Milky Way. There is therefore a complex relationship between the various motions of the
natural and artificial bodies. How many of these need to be considered simultaneously will
depend on the design goals of the satellite system. A satellite designed to observe the earth’s
surface will not need to know where the stars are at any particular time, but the location
of the local star, the sun, may be important if the satellite needs to use sunlight to illumi-
nate its coverage region on the surface of the earth. On the other hand, a satellite designed
to observe background thermal radiation levels of deep space in the infrared band will need
to know the position of each of the neighboring planets. Should the telescope of the satel-
lite inadvertently point toward one of these planets, the temperature viewed would not re-
flect that of the true background radiation level. In the sections that follow, we will review
all of the different NGSO orbits that have been used for scientific, military, and commer-
cial satellite missions. The simplest NGSO orbit is an equatorial orbit.
Equatorial Orbits
Equatorial orbits lie exactly in the plane of the geographical equator of the earth. That is,
the orbital path lies directly above the equator at all times. In order to take advantage of
the 0.45 km/s eastward rotational velocity of the earth, most satellites are launched toward
the east into a prograde orbit. A westerly directed orbit is called a retrograde orbit. A
satellite in an eastwardly directed equatorial orbit will have two periods: a real orbital
period that is referenced to inertial space (the galactic background) and an apparent or-
bital period that is referenced to a stationary observer on the surface of the earth. The real
orbital period, denoted here as T hours, is given by Eq. 2.6. The apparent orbital period
to the observer on the equator will be P hours where
P 124T2 124 T2 hours (10.1)
To be exact, 23.9344 h, one sidereal day, should be used in place of 24 h in Eq. (10.1).
Table 10.1 (from reference 2) illustrates the difference between P and T for a number of
orbital altitudes and elevation angles. It also shows the time the satellite is visible to the
observer, neglecting atmospheric refraction and assuming the satellite can be tracked down
to 0°, that is, right down to the horizon. Other implications of the observing time are con-
sidered in more detail in Section 10.3, Coverage and Frequency Considerations.
The plane of a satellite’s orbit must be in the plane of the equator for the satellite
to be in equatorial orbit. This can be achieved by launching the satellite in one of two
ways. The first launch method is to locate the launch site on the equator and to launch
c10.qxd 28/08/02 19:38 Page 392
TABLE 10.1 Orbital Periods and Observing Time
Orbital period
Orbital height Observing Time

(km) True (hours) Apparent (hours) (hours)
500 1.408 1.496 0.183

1,000 1.577 1.688 0.283
5,000 1.752 1.890 0.587
10,000 5.794 7.645 2.894
35,786 23.934
Source: From Table 1.1 in reference 2.
the spacecraft toward the east along the equatorial plane. The second method is to launch
the satellite into an inclined orbit and to execute a maneuver either during the launch tra-
jectory or when the satellite is in an inclined orbit that changes the plane of the initial or-
bit so that the final orbit is in the plane of the equator. Removing the inclination from the
orbit, so that the satellite orbits exactly over the equator, requires significant energy,
particularly if the launch site is well removed from the equator. The first two sites from
which orbital flights were made, Cape Canaveral in the United States and Baikonur in
Kazakhstan, were not close to the equator (approximately 28° N and 46° N, respectively).
In addition, the early launch vehicles lacked the ability to alter the trajectory significantly
during launch. The first artificial earth satellites were therefore placed into inclined orbits,
that is, the planes of the orbits were inclined to the equatorial plane.
Inclined Orbits
There are advantages and disadvantages to inclined orbits, depending on the mission goals
and the data recovery requirements. The greater the inclination of the orbit is, the larger
the surface area of the earth that the satellite will pass over at some time in its flight.
Figure 10.1 illustrates this for a LEO satellite.
In Figure 10.1b, the inclined orbit will take the spacecraft, at one time or another,
over the earth’s entire surface that lies approximately between the latitudes given by
the orbital inclination. For example, an orbit with an inclination of 30° will cover all re-
gions that lie approximately between latitudes 30° north and 30° south. The superior cov-
erage of the earth with an inclined orbit satellite is counterbalanced by the disadvantage
that the master control station (MCS) will not be able to communicate directly with the
satellite on every orbit as with an equatorial orbit satellite. A LEO satellite orbits the earth
with a period of 90 to 100 min and, for an inclined orbit satellite, the earth will have ro-
tated the master control station out of the path of the satellite on the next pass over the
same side of the earth. Depending on the quantity of data that need to be passed to the
MCS, or if real-time communications are required continuously, a system architecture that
employs multiple satellites will need to be considered.
The simplest, and lowest cost, solution to pass data between an inclined orbit satel-
lite and an MCS is to design the satellite to store the data acquired over many orbits (when
it is out of sight of the MCS) and then, when it passes within radio range of the MCS, to
dump the data rapidly to the MCS. This is called store-and-forward and it is one of the ca-
pabilities of some LEO systems, including Orbcomm satellites3. It was also the technique
used for the very first communications satellite, Project SCORE, in December 1958. In the
c10.qxd 28/08/02 19:38 Page 393
Coverage of LEO satellite
Equatorial LEO satellite
S
(a )
Coverage of inclined orbit

LEO satellite (shaded area)
Inclined orbit LEO satellite
S
(b )
FIGURE 10.1 (a) Coverage of an equatorial orbit LEO satellite. The LEO satellite is in an
equatorial LEO orbit and so it will only pass over the equator. The coverage of the equatorial
LEO satellite will therefore be limited to a swathe of the earth close to the equator, determined
by the height of the orbit and the beamwidth of the satellite’s antenna. In this example, the or-
bit is assumed to be circular and the antenna beamwidth has been ignored. (b) Coverage of an
inclined orbit LEO satellite. The LEO satellite is in an orbit that is inclined at approximately 40°
to the equator. The satellite will therefore pass over, at one time or another, all regions of the
earth between latitudes 40° N and 40° S of the equator. The coverage of the inclined orbit LEO
satellite will therefore be a swathe of the earth between about 40° of the equator, determined
by the height of the orbit and the beamwidth of the satellite’s antenna. In this example, the or-
bit is assumed to be circular and the antenna beamwidth has been ignored. Note: The higher
the orbit and the greater the inclination, the further the satellite’s total coverage will reach.
Orbcomm system, if a user on the ground is unable to establish contact via an Orbcomm
satellite to a gateway earth station (GES) in the Orbcomm system, a “GlobalGram®” may
be left stored within the satellite for later transmission to the GES when it comes into view
of the satellite. The downlink transmission rate must be high enough to enable all of the
stored messages in a LEO satellite to be sent to the MCS in the period when it is within
range of the satellite. If a continuous, real-time connection is required between a LEO satel-
lite and the MCS, there are only two approaches that can be used.
• The first approach is to locate control stations around the world so that the LEO
satellite is never out of sight of at least one of the control stations. Terrestrial or
GEO satellite connections are then established between the many control stations
and the MCS to bring the LEO data back to the MCS in real time.
c10.qxd 28/08/02 19:38 Page 394
• The second approach is to establish intersatellite links (ISLs) to relay the LEO data
traffic back to the MCS. The ISLs can either be set up among the LEO satellites in
the constellation, so that the LEO data traffic is relayed between the LEO satellites
in orbit via their ISLs, or the ISL link can be set up between the LEO satellite(s) and
one or more GEO satellites. The GEO satellite relays the LEO data traffic back to
the MCS directly, if it is within sight of the MCS, or via another GEO satellite. Irid-
ium4 and the early design of Teledesic5 adopted the former solution of LEO satel-
lites interlinked via ISLs within their own constellation. Skybridge6 and Globalstar7
chose a different approach. NASA uses the TDRSS satellites for shuttle missions
(see sidebar). Geostationary relay satellites have also been used for military recon-
naissance missions by at least the United States and Russia—transmitting onward
data received from LEO observation satellites—and a similar system was in its im-
plementation stage at the end of 1999 for some civilian earth observation missions.
Figure 10.2 illustrates the two concepts.
In both of the examples shown in Figure 10.2, the LEO satellite is in a circular orbit.
A circular orbit gives a constant dwell time over a given coverage region since the angu-
lar velocity of the satellite is the same at any point in the orbit. In many cases, mission
goals will dictate different dwell times for different parts of the orbit. A circular orbit will
not achieve this result. To accomplish variations in dwell time around the orbit, the satel-
lite must be in an elliptical orbit.
Elliptical Orbits
As noted in Chapter 2, an elliptical orbit will have a nonzero eccentricity. The orbit ec-
centricity, e, is determined by the lengths of the semimajor axis, a, and the semiminor
axis, b, of the orbit ellipse
e2 1 1b2 a2 2 (10.2)
Alternatively, if Ra is the distance between the center of the earth and the apogee point of
the orbit and Rp is the distance between the center of the earth and the perigee point, the
eccentricity is
e 1Ra Rp 2 1Ra Rp 2 (10.3)
Figure 10.3 illustrates the geometry of Eq. (10.3).
In Eqs. (10.2) and (10.3), if the orbit is exactly circular, a b and Ra Rp, and
the eccentricity reduces to zero. In general, no orbit is truly circular for a variety of reasons,
SIDEBAR
NASA built and operated a number of relay stations Communication with the Space Shuttle is main-
for the Mercury, Gemini, and Apollo programs. In none tained using the Tracking and Data Relay Satellite
of these missions was the manned spacecraft ever out System (TDRSS). Several TDRSS satellites in geo-
of real-time contact with the Manned Spaceflight Cen- stationary orbit relay data from the Shuttle to several
ter in Houston, United States (which acted as the MCS earth stations around the world that then send the data
in this case) except when the Apollo craft was behind to NASA’s MCS for manned space flight in Houston,
the moon or in the re-entry phase where ionized plasma Texas.
caused a radio blackout for all spacecraft.
c10.qxd 28/08/02 19:38 Page 395
Master
control
station
Inclined orbit
LEO satellite
Ground tracks of satellite

on successive orbits
(a )
Inclined orbit
LEO satellite
Master
control
station
GEO satellite
(b)
FIGURE 10.2 (a) Store-and-forward concept. In this LEO application, the satellite stores in-
formation it has gathered while orbiting the earth and, once within range of the master con-
trol station, it downloads the stored data. The (uplinked) data storage rate is usually low, a
few kbit/s at most, while the download is at a much higher rate due to the small time the
satellite has available when it is within range of the master control station. (b) Real-time data
transfer via a GEO satellite. In this approach, the LEO satellite can transfer data in real time
via the GEO satellite to the master control station whenever it can “see” the GEO satellite. If
there were a number of GEO satellites equipped with intersatellite links (ISLs) distributed
around the geostationary orbit, then the LEO satellite need never be out of real-time contact
with the master control station.
but eccentricity values of 103 or less can be considered to correspond to circular orbits
for all practical purposes. The eccentricity is another way of describing the variation in
the radius of the orbit. If Rav is the average radius of an orbit from the center of the earth,
then the variation, R, in the orbital radius, is given by8
¢ R eRav (10.4)
4
For a geostationary satellite (Rav 42,164.17 km) with an eccentricity of 10 , R will
be 4.2 km. For a LEO constellation with a circular orbit of approximately 800 km above
the earth, with each LEO satellite having an eccentricity of 104, R will be 0.7178 km
c10.qxd 28/08/02 19:38 Page 396
Satellite
Perigee Earth Apogee
Rp Ra
FIGURE 10.3 Schematic of an elliptical orbit illustrating ellipticity. The satellite orbits the
earth with a perigee distance from the center of the earth given by Rp and an apogee dis-
tance from the center of the earth given by Ra. Note that the perigee and apogee are always
exactly opposite each other in the orbit. This is true of any object in any orbit around any
other body.
(assuming the earth mean radius is 6378.137 km). If the orbit becomes less circular and
the eccentricity increases to 103, R increases to 7.178 km. If the LEO satellites in the
constellation pass over (under) each other, then the vertical separation must be sufficient
to prevent any likelihood of a collision between satellites. The average orbital altitude and
eccentricity of the orbit will determine the likelihood of a collision. One of the more fa-
mous orbits has an eccentricity 0.74. This is a special case of a highly elliptical orbit
(HEO) known as the Molniya orbit.
Molniya Orbit
The former Soviet Union had a difficult communications design problem. Much of the
landmass is in far northern latitudes. Archangel, the port on the White Sea, is close to lat-
itude 60° N; immense tracts of Siberia lie inside the Arctic Circle. To compound the prob-
lem further, the country was spread across 11 time zones: it was the largest country in the
world (and Russia still is). The signals from a geostationary satellite can reach well in-
side the Arctic Circle if operations at elevation angles below 5° are permitted, but a sin-
gle GEO satellite cannot reach that far north over 11 time zones simultaneously. A new
type of orbit was required to provide good communications coverage over the former
USSR. What transpired was the Molniya system.
The first Molniya satellite was launched in April 1965 and it gave its name to both
the system of satellites and to the unique orbit. The word Molniya means flash of light-
ning in Russian. The apogee of the Molniya orbit is at an altitude of 39,152 km and the
perigee is at an altitude of 500 km. The orbital period is 11 h and 38 min and the orbital
inclination is 62.9°. This combination of apogee, perigee, and inclination ensures that the
ground track of the Molniya orbit repeats every other orbit. That is, if the orbit passes ex-
actly over Moscow on orbit one, it will do so again on orbit three, five, seven, nine, and
so on. Figure 10.4 illustrates the orbit geometry.
Two Molniya orbits, with the planes of the orbits separated by 180°, will thus pro-
vide coverage over the extreme latitudes of Russia for 24 hours per day using two satel-
lites, correctly phased—one in each of the two Molniya orbits. When one of the satellites
c10.qxd 28/08/02 19:38 Page 397
GEO orbit
distance
N
FIGURE 10.4 Schematic of a Molniya orbit. In this example, the trajectory is configured to
have a large dwell time over the northern part of the orbit so that it can serve a country that
has most of its landmass in this region. This was the design adopted for the original Molniya
system of the former Soviet Union. Approximately 60% of this Molniya orbit, which stretches
more than 3000 km beyond the height of a GEO orbit, has good look angles for latitudes be-
tween 30° N and 90° N. This translates to more than 6 h of the 11 h 38 min orbital period.
is at its apogee over Russia in Molniya orbit 1, the other satellite will also be at its apogee
somewhere over North America in Molniya orbit 2. By the time the second satellite has
moved once more to its apogee in Molniya orbit 2, the earth will have revolved half a
turn under it and Russia will again be spread beneath it. Figure 10.5 illustrates the dual
Molniya orbit concept.
Satellite 2 Satellite 1
Molniya Molniya
orbit 2 orbit 1
FIGURE 10.5 Schematic of an operational Molniya system. Satellite 1 in Molniya orbit 1 is

providing service over Russia at close to its apogee while the second satellite is also close
to its apogee in Molniya orbit 2. Molniya orbits l and 2 are separated by 180° in their orbital
planes. By the time satellite 2 has moved around its orbit once and back to its apogee (a pe-
riod of about 12 hours), the earth will have rotated about 180° and the second satellite will
be over Russia.
c10.qxd 28/08/02 19:38 Page 398
The two-satellite, dual Molniya orbit requires that earth station operations be car-
ried out at elevation angles well below 30° for full 24 hours-per-day coverage of one
region. Note that, if four satellites are used, one in each of four Molniya orbits that
have planes 90° apart, high elevation angle service could be provided to any region at
high latitudes in the Northern or Southern Hemisphere. The planes of the four Mol-
niya orbits would be orthogonally distributed around the earth with one satellite in each
Molniya orbit, correctly phased in its own orbit to provide coverage from the apogee
sections of that orbit as the region rotates beneath the four-satellite constellation. Up
to eight Molniya satellites, in eight different Molniya orbital planes separated by 45°
and suitably phased around their orbit, have been used to provide continuous coverage
over Russia.
The Molniya orbit has another advantage for specific services intended for latitudes
well away from the equator. The orbit takes the satellite so far away from the equatorial
plane and the apogee is so distant from the earth that long dwell times with elevation an-
gles close to 90° can be achieved at high latitudes on the earth. This fact was used in a
proposal for a Molniya orbit to deliver mobile satellite service (MSS) to automobiles. A
view of the earth, with the plot of the orbit track, abstracted from this proposal9 is shown
in Figure 10.6.
The Molniya orbit, in addition to the long delay time associated with the commu-
nications range when at apogee and the lack of continuous 24-h contact with a single
spacecraft from a fixed coverage, also suffers from three drawbacks that increase the over-
all end-to-end costs. The first is the requirement to track the spacecraft. The second is the
need to switch communications to the other Molniya satellite—rather like a mobile radio
handoff situation—when the first goes out of coverage as the other comes into coverage.
Due to the wideband nature of the traffic and the large angular separation between suc-
cessive Molniya satellites as seen from one earth station, this requires two reflector an-
tennas at each site. At the end of the twentieth century, phased array antennas still could
not provide accurate coincident tracking of both transmit and receive beams simultane-
ously well away from the (unsteered) electrical boresight over bandwidths that exceed a
few percent of the carrier frequency and at a cost that commercial systems can accept.
The third drawback to a Molniya orbit is the radiation environment that the satellite has
to pass through four times a day—twice on ascent and twice on descent. While the first
two drawbacks may be less of an inhibition to commercial success in the long term with
direct-to-home services when relatively inexpensive, and efficient, phased array antennas
are available that track over the required range of look angles, the third drawback will
always be a major factor.
Radiation Effects
The effect of radiation on electronics in space is generally separated out into two main as-
pects10: total dose and single-event upsets. The total dose is simply the cumulative effect
of radiation over the lifetime of the electronics in space and is mainly due to trapped elec-
trons and protons in the Van Allen belts. (The Van Allen radiation belts are discussed later
in this section.) Eventually, the cumulative effect of radiation will degrade the performance
of the transistor junction/chip such that it cannot be relied on to generate the correct re-
sponses, etc. This is particularly harmful in the computer elements that control the opera-
tion of the satellite and the payload. Single-event upsets are caused by heavy ions ejected
from the sun, usually protons, impacting the circuitry at a critical point such that they de-
posit enough charge to induce an energy (bit) flip, that is, change an open circuit to a closed
circuit, create a logical one instead of a logical zero, etc. These single-event upsets are
c10.qxd 28/08/02 19:38 Page 399
(a )
160° 120° 80° 40° 0° 40° 80° 120° 160°

80°
60°
40°
20°
0°
20°
40°
60°
160° 120° 80° 40° 0° 40° 80° 120° 160°
(b)
FIGURE 10.6 View from above the Molniya orbit apogee showing the ground track8.
(a) View from the apogee point of a Molniya orbit positioned at almost 0° longitude when at
apogee. (b) Ground track of the Molniya orbit shown in Figure 10.6a. Note the two apogees
in the orbit, one over close to 0° longitude and the other at close to 180°. The apogee occurs
at a high latitude, from which the elevation angles are well above 70° over quite a large
region. With these high elevation angles, blockage of buildings would be minimized and thus
allow relatively high availability for an MSS system operating to automobiles in most cities.
This proposal8 was for a European MSS system, but the apogee could be phased to occur at
any longitude so that cities in high latitudes, but arbitrary longitude, could operate to an
MSS satellite in Molniya orbit.
c10.qxd 28/08/02 19:38 Page 400
SIDEBAR
The Van Allen radiation belts were named after the dis- The radiation belts consist of high-intensity protons
coverer of these belts, James A. Van Allen, who was and electrons that are temporarily trapped in the earth’s
the principal investigator for a radiation measurement magnetic field. While the trapped electrons can have
package on the United States’ first successful artificial energies up to 7 MeV (seven million electron volts),
earth satellite, Explorer 1, launched in January 1958. the trapped protons can have energies up to 500 MeV10.
more critical if the bit flip is permanent, that is, “latch-up” occurs in a set position from
which it cannot be changed.
The relative motion between the liquid core and the solid mantle and outer crust
above it generates the earth’s magnetic field. The magnetic field lines stretch out around
the earth as shown schematically in Figure 10.7. While generally symmetrical close to the
earth, the magnetic field lines of the earth become distorted further out from the earth due
to interaction with the energy flowing toward the earth from the sun. The boundary where
the solar atmosphere and the earth’s magnetic field meet far out in space is called the bow
shock, much like the pressure waves concentrating in front of the wing of an aircraft.
Since the earth’s magnetic and geographic poles are not coincident, the magnetic equator
(and magnetic latitudes) will be different from the geographic equator (and geographic
latitudes). The geomagnetic latitude can be computed from11:
f arcsin 3sin a sin 78.5° cos a cos 78.5° cos 169° b2 4 (10.5)
where is geographic latitude and is geographic longitude. North and east coordinates
are considered positive, and south and west coordinates negative.
Magnetic field lines

NM
FIGURE 10.7 Representation of the magnetic field lines that flow between the north and
south magnetic poles of the earth. The earth has a strong magnetic field due to having a
liquid core that is spinning at a different rate than the solidified outer shell. The magnetic
poles, however, are not coincident with the geomagnetic poles and so the magnetic equator
is not located in the same position as the geographical equator. Sometimes the geomagnetic
latitudes are referred to as dip latitudes since they will correspond to the dip in the magnetic
field at that point. NM is the north magnetic pole.
c10.qxd 28/08/02 19:38 Page 401
The electrons and protons become ensnared in the earth’s magnetic field when
their kinetic energy cannot overcome the trapping effect of magnetic lines of force at
the given point of encounter in space. Since the magnetic field strength decreases with
increase in altitude on a given radial from the center of the earth, only the electrons
are trapped in the higher reaches of the earth’s environs (10,000 km altitude above
the earth) since the field forces are relatively low at these altitudes. Both electrons and
the higher energy protons are trapped lower down in the earth’s atmosphere 200 to
10,000 km10, where the field is relatively more intense. The radiation levels induced
by the electrons and protons fluctuate wildly with latitude, longitude, altitude, and with
the sunspot cycle.
SIDEBAR
Sunspots are disturbances on the surface of the sun. cycle period is not constant. The period has been as
Sunspots appear to generate huge outflows of energy short as 9.5 years and as long as 12.5 years12. The first
from the sun and the amount of energy closely follows cycle that has been given an official number is the
the number of sunspots—or rather groups of sunspots— 1755–1766 period. The last full solar cycle of the twen-
which can be counted on the surface of the sun. The tieth century (1986.8–1996.4) was labeled Cycle 22. A
sunspot count, and hence the level of energy, varies with schematic of this cycle, showing the large variation in
a mean period of about 11 years, although the actual sunspot count that exists, is given in Figure 10.8. The
cycle spans a 22-year Hale cycle as the magnetic field turn of the twentieth century saw us in cycle 23 with
lines associated with the sunspot activity on the sun’s the two-to-four year period of peak activity starting in
surface reverse every 11 years. The 11-year sunspot the fall 1998 equinoctial period.
200
150
Sunspot Number
100
50
0
1986 1988 1990 1992 1994 1996
Year
FIGURE 10.8 The general variation of the sunspot number over solar cycle 22. The
smoothed sunspot number is averaged over several months. The fluctuations in the actual
sunspot number are shown about this smoothed average. Not only does the sunspot count
vary widely from month to month, it does so also from day to day. The higher the average
sunspot number is, the larger the variation in actual sunspot number count is in general.
Note the more rapid rise than decline in the average sunspot number count and the fairly
long period when the sunspot activity was very high. Because of the “flat” nature of the
sunspot maximum period (up to 4 years) it is usual to determine the sunspot periods from
their minima.
c10.qxd 28/08/02 19:38 Page 402
Outer Van Allen belt
Inner Van Allen belt N
~15,000 km
~1,500 km
FIGURE 10.9 Pictorial representation of the two Van Allen radiation belts. The above
schematic is a vertical “slice” through the radiation belts that exist around the equator. The
shaded areas are the regions in the two belts where the radiation is at a maximum. The two
principal NGSO regions lie “under” the first Van Allen radiation belt—the low earth orbit or
LEO region—and “between” the two radiation belts—the medium earth orbit or MEO
region—so as to avoid the highest radiation doses. However, radiation never falls to zero and
exists in all areas (see reference 10).
The variability of the sunspot cycle leads to large fluctuations in the radiation envi-
ronment in space. While there are large variations in the radiation environment with lati-
tude, height above the earth, and with orbital inclination, it is normally considered that there
are two main Van Allen radiation belts where the effect is more concentrated. The center of
the first belt is at a height of about 1500 km above the earth and the second at around 15,000
km, measured around the equator, although these distances are somewhat arbitrary and there
is some evidence that the outer belt may actually be two merged belts. The belts can be con-
sidered as doughnut-shaped, with the energy at its highest toward the center of the given
belt. Figure 10.9 illustrates the concept. The trapped electrons and protons travel northwards
and southwards along the magnetic field lines shown in Figure 10.7. They are reflected when
they are close to the magnetic poles10 and so, statistically, spend more of their time closer
to the equator than the poles; hence the Van Allen belts are positioned around the geomag-
netic equator. The closer to the center of the radiation belts a satellite is positioned and the
longer it is in space, the higher the total radiation dose becomes.
Total dosage for semiconductors that are fabricated using silicon is measured with
a unit called the krad(Si). A rad(Si) is a unit of energy absorbed by silicon from radiation
and it is equivalent to 0.01 J/kg10. Radiation in near-earth space is highly variable. It
changes both with height above the earth and with the inclination of the orbit with re-
spect to the equatorial plane. Since the radiation is concentrated at the equator, satellites
that are in equatorial orbits will receive a higher dosage than those that are in polar or-
bits will. In a like manner, as the orbital height moves from very close to the earth (300
km) outward for the first few thousand kilometers, the radiation dose will increase. Table
10.2 gives some typical examples of total radiation dosage for a LEO satellite designed
for a 10-year operational lifetime.
c10.qxd 28/08/02 19:38 Page 403
TABLE 10.2 Typical Total Doses for Various Orbits
Orbital height (km)
Orbital type (degrees) 800 1100 2000
Polar orbit (90°) 30 krad(Si) 100 krad(Si) 500 krad(Si)

Equatorial orbit (0°) 2000 krad(Si)
The data are based on a 10-year mission life using silicon-based electronics
in a satellite with a 2.5 mm thick aluminum skin.
Source: Data extracted from the text of reference 10.
Choosing an orbit that has a reduced level of radiation can therefore reduce the
potential for radiation damage. Where this is not possible, then either radiation hard-
ened (rad-hard) devices must be selected for the satellite or suitable shielding em-
ployed. Both are expensive options, the former because of the fabrication costs and the
latter because radiation shields can be heavy and are nonproductive elements of the
payload. Developing electronic devices that can withstand total radiation doses of
1 Mrad(Si) is possible with rad-hard technologies but newer techniques for approach-
ing these levels of radiation hardening of devices will be needed for constellations of
dozens of satellites. New, relatively cheap production processes have been shown to
provide consistent shielding to 100 krad(Si) total dosage10 and it is likely that such
techniques, plus local site shielding with aluminum strips, will be largely employed
for the foreseeable future. The same approach is being used for protection against sin-
gle-event upsets.
Sun Synchronous Orbit

A sun synchronous orbit is a special form of low earth orbit where the plane of the orbit
maintains a constant aspect angle with the direction to the sun. Some satellite missions
require a specific orbit with such a constant relation to the direction of the sunlight. One
example is an earth resources satellite that requires a large amount of direct sunlight to
illuminate the region below the satellite so that photographs can be taken. This satellite
SIDEBAR
With the ever-smaller integrated circuits being devel- thresholds of 37 MeV-cm2/mg and the incidence of
oped for flight operations, there is an increased like- heavy ions with a LET exceeding 37 MeV-cm2/mg is
lihood that the linear energy transfer (LET) that is very rare10. The potential for latch-up will also reduce
generated by the heavy ion collision will cause a sin- as technology advances permit devices to be used that
gle-event upset. The potential for an upset depends on employ a lower drive voltage. In addition, many of the
the LET generated and the threshold level at which silicon-on-insulator (SOI) and silicon-on-sapphire
the device will incur a single-event upset. Many space- (SOS) technologies have been found to be immune to
bound integrated circuits (ICs) have LET thresholds latch-up, as there are no parasitic paths10. This is en-
greater than 37 MeV-cm2/mg10, which means that couraging, as there are some satellites that will require
heavy ions with LETs of less than this amount will a unique type of orbit that cannot be selected from a
cause few single-event upsets. Fortunately, ICs can be radiation dosage perspective. One such orbit is the sun
manufactured relatively inexpensively with LET synchronous orbit.
c10.qxd 28/08/02 19:38 Page 404
Sunlight
N B
Sun
Earth
FIGURE 10.10 Examples of two sun synchronous orbits. In the illustration above, the earth
is viewed from above the North Pole, N, with the sunlight illuminating the left side of the
earth. Two sun synchronous orbits are shown. Orbit B is designed so that it will always have
one-half of the orbit with the sun almost directly behind it; orbit A is designed to be always
within sight of the sun—the so-called sunset–sunrise orbit since it will be orbiting the termi-
nator. The terminator is the line that divides night from day.
would be in orbit B in Figure 10.10. Another example of a satellite needing this same
orbit is a meteorological satellite, where images of the clouds and their directions of motion
are critical in developing forecasts. While communications satellites have returned the
greatest tangible investment returns for their owners, it is arguable that meteorological
satellites have led directly to huge savings in human life, as well as to less property dam-
age and farm animal destruction, in extreme weather situations. The early meteorological
satellites (e.g., TIROS) were in LEO sun synchronous orbits, but all recent meteorologi-
cal satellites are in GEO orbits to provide more instantaneous and continuous coverage.
Other satellites that employ sun synchronous orbits are surveillance satellites.
Some surveillance satellites use orbit B of Figure 10.10, so that the maximum
illumination is provided once per orbit. Others use orbit A of Figure 10.10. This particu-
lar “sunset–sunrise” orbit always has the satellite illuminated by the sun while the region
below it has the sun at almost grazing incidence. There are two advantages in this orbit.
First, the satellite need not have a large battery capacity for eclipse operations since it is
always illuminated. Second, since the shadows are so long in the region being surveyed,
changes in terrain or structures will be immediately obvious.
SIDEBAR
Before synthetic aperture radars were orbited, the sance aircraft for the first time at Peenemunde toward
sunset–sunrise orbit was used advantageously to de- the end of the Second World War. Similarly, the ill-
tect changes in terrain following natural disasters fated USSR moon rocket was detected on its launch
such as earthquakes. The long shadow of a German pad by a surveillance satellite using the shadow it
V2 rocket allowed it to be detected by a reconnais- cast.
c10.qxd 28/08/02 19:39 Page 405
Movement of
Alignment
earth around
error
sun
A
Plane of satellite’s
Sun
orbit around earth
FIGURE 10.11 Illustration of alignment changes of the orbital plane of a satellite due to
the movement of the earth around the sun. In the illustration above, a satellite has been
launched into an LEO orbit (position A) in which the sun is directly overhead on the sunward
side of the orbit. The view is from above the North Pole of the sun, orthogonal to the plane
of the earth’s orbit around the sun. When the earth has moved to position B, the plane of the
satellite’s orbit—which is fixed in inertial space—now has an alignment error with respect to
the sun. If it is essential that the orbital plane of the satellite always be in line with the direc-
tion to the sun on a day-to-day basis over a long period, then the plane of the satellite’s orbit
will need to change at the same rate that the alignment error is increasing. That is, the plane
of the orbit will need to ”precess” to match the movement of the earth around the sun.
Figure 10.11 illustrates how the sun synchronous orbit is achieved. If a satellite is
in a perfectly circular LEO orbit over the poles of the earth, a carefully timed launch
would put the orbit in such a position that the sun is directly behind the satellite on the
sunward side of the first orbit. This is position A of Figure 10.11. However, a short while
later, the earth will have moved in its orbit around the sun and the plane of the satellite’s
orbit (now in position B) will no longer be aligned with the direction of the sunlight. In
order to make the satellite’s orbital plane always keep pace with the apparent change in
position of the sun, it must be launched into a retrograde orbit. A retrograde orbit has a
velocity component in a westerly direction. In practice, a LEO satellite launched into an
orbit with an inclination of close to 98° to the equator (measured counter clockwise from
the equator looking east) will move the orbital plane in time to the earth’s movement
around the sun. Elliptical orbits with different retrograde inclinations (see the constella-
tion Ellipso in Section 10.5) will also yield sun synchronous orbits. The change (rotation)
in the orbital plane is called precession. A key advantage of a sun synchronous orbit is
that it will repeat its track every half day. It can therefore be used to make measurements
at given times of the day and night so that correlation exercises can be attempted.
A sun synchronous orbit will pass over almost all of the earth at one time or an-
other. Determining the instantaneous surface area of the planet seen by the satellite and
over which information is required—or to which communications is to be established—
is another issue. This portion of the earth’s surface is called the coverage area or coverage
region.
SIDEBAR
One example of a spacecraft in a sun synchronous or- and 2 P.M. local time equivalents over the same region
bit was the Mars Explorer spacecraft, which was put so that local heating effects and cooling effects could
into a sun synchronous orbit around Mars in 1998. be accurately tracked.
The orbit was used to measure temperature at 2 A.M.
c10.qxd 28/08/02 19:39 Page 406
10.3 COVERAGE AND FREQUENCY

CONSIDERATIONS
General Aspects
In some cases, the designer of a satellite system has few degrees of freedom in designing
a payload to provide optimum coverage. This occurs in some missions where a shared
spacecraft has to accommodate a number of payloads. Examples exist in the scientific
community when (generally) low-cost, multiple missions are being developed for a sin-
gle spacecraft. In other cases, more freedom exists in the design stage and the mission
planners can vary a range of parameters iteratively to arrive at an optimum coverage.
The mission goals will directly determine the coverage that has to be achieved by a
given satellite system. This in turn leads to the selection of orbit, payload technologies, etc.
For example, if a communications satellite system has to provide coverage of the European
Union (EU), there is a minimum altitude at which a single satellite can operate and still cover
all of the EU at once. If the coverage of the EU must be continuous, a GEO orbit can be se-
lected or a constellation of NGSO satellites can be designed to provide the necessary cover-
age overlap between successive satellites. The determination of coverage area, while initially
an exercise in simple geometry, is eventually heavily influenced by the available technology
both on the ground and in space, and other aspects such as the radiation environment. We
will consider first the geometrical aspects of determining an optimum coverage.
In Figure 10.12, a spacecraft orbits at distance rs from the center of the earth, C.
We will assume that the spacecraft is a communications satellite and that it needs to be
in contact with an earth station located at E. The elevation angle to the satellite is . Using
the sine rule we have
3rs sin 190 u2 4 3d sin 1g2 4 (10.6)
which yields
cos 1u2 3rs sin 1g2 4 d (10.7)
All three parameters in equation (10.7) have key inputs to the architecture of the satellite
system. The angle will yield the coverage area on the surface of the earth assuming the
satellite has a symmetrical coverage about nadir. The distance d will determine the free space
path loss along the propagation path, and will be a factor in the link budget design. The el-
evation angle will influence the GT ratio of the antenna, the blockage probability from
terrain and buildings near the antenna and the likely propagation impairments that will be
encountered along the path to the satellite. For systems that operate in frequency bands that
suffer significant degradations in rain, the elevation angle can be the critical design element
(see Chapter 8 for more details on propagation effects along earth-space paths).
Frequency band
Low earth orbit satellite systems providing data and voice service to mobile users tend to
use the lowest available RF frequency. The EIRP required by the satellite transponder to
establish a given CN ratio in the mobile receiver is proportional to the square of the RF
frequency of the downlink, as the analysis in the next paragraph shows. The power that
must be transmitted by a mobile transmitter is also proportional to RF frequency squared
when the mobile uses an omni-directional antenna. Since the cost of satellites increases
as the EIRP of the transponders increases, a lower RF frequency yields a lower cost sys-
tem. This is one reason why L-band is allocated for mobile satellite services.
c10.qxd 28/08/02 19:39 Page 407
10.3 COVERAGE AND FREQUENCY CONSIDERATIONS 407
Local horizon at
earth station, E, looking
toward satellite, S
rs
d
δ
θ
γ
E re C
FIGURE 10.12 Geometry for calculating coverage area. The satellite, earth station, and the
center of the earth are all in the same plane in this figure. SCE is the central angle, , and the
elevation angle, , is the angle between the local horizon and the satellite at the earth station
in the plane of the figure. The line SC joins the satellite and the center of the earth and cuts
the surface of the earth at point Z. To an observer at point Z, the satellite is at zenith. The
satellite is at a distance, d, from the earth station and at a distance rs from the center of the
earth. The radius of the earth is given by re, a good average value for which is 6370 km.
Consider a LEO satellite with a coverage zone on the earth’s surface that has an
area A m2. A transponder on the satellite with output power Pt watts drives an antenna
with a linear gain Gt to produce an EIRP from the satellite of PtGt watts. The average flux
density across the coverage zone is therefore
F PtGt A watts/m2 (10.8)
The value of the flux density is independent of frequency. The mobile receiver has an an-
tenna that is nearly omnidirectional, with a gain Gr, where Gr is typically less than 3 dB.
The effective receiving area of this antenna is given by
Ae Grl24p (10.9)
The received power at the mobile earth station is given by Pr F A, hence
PtGtGrl2
Pr watts (10.10)
4pA
Thus the received power at the mobile terminal with an omnidirectional antenna in-
creases as the square of the wavelength, or decreases as the square of the frequency. The
lower the RF frequency, the greater the received power for any given coverage zone. By
reciprocity, the same result will apply when the mobile terminal transmits with an omnidi-
rectional antenna. It therefore makes sense for mobile systems, which are forced to use
omnidirectional antennas so as to avoid having to steer a directional antenna, to use the low-
est possible RF frequency. That is why Orbcomm’s data relay LEO satellite system uses VHF
c10.qxd 28/08/02 19:39 Page 408
and UHF frequencies. Orbcomm satellites have a single transmit beam at the satellite that
serves the entire coverage zone. For the same reasons, L band is allocated for mobile satel-
lite service, but to achieve similar CN ratios with L-band links as Orbcomm satellites achieve
with VHF links, the L-band satellite must provide multiple beams from a high gain antenna.
One disadvantage of VHF and UHF frequency bands is a high noise power due to the
natural environment. For this reason, the antenna noise temperature for a system operating
at VHF or UHF will be much higher than the receiver noise temperature. Environmental
noise temperature falls with increasing frequency; by L band it is not a significant factor.
The worst possible choice of frequency for a mobile system is Ka band (or above).
A Ka-band mobile downlink operating at 20 GHz requires 22.5 dB more transmitted EIRP,
or receiving antenna gain, than the same system operating at 1.5 GHz. Occasionally pro-
posals are aired for Ka-band mobile systems, but such a system can only succeed with a
steered directional antenna on the mobile terminal. Conventional mechanically steered
Ka-band dishes are thousands of dollars more expensive than a simple whip antenna, and
self-steering phased arrays are even more costly.
It is worth noting in passing that it is the omnidirectional antenna of a mobile ter-
minal that drives up the cost of every transmitted bit in a mobile system. Suppose that a
fixed terminal with an antenna of gain Grx supports a bit rate of Rb bits per second. A mo-
bile terminal with a low gain antenna, gain Grm, and the same satellite EIRP and path loss
can support a much lower data rate of Rb (GrmGrx). For example, an 18-inch DBS-TV
antenna operating at 12.5 GHz has a gain of 33 dB. A satellite link using the DBS-TV
antenna can receive data at 2000 times the rate of a mobile terminal that has an omnidi-
rectional antenna with a gain of 0 dB, with the same overall CN value in the receiver.
Given equal costs for the space segment of the communications link, and a mobile sys-
tem operating in Ku band, a system operator must charge the mobile user 2000 times as
much per delivered bit compared with the DBS-TV terminal owner. Looked at another
way, for the same monthly fee, the DBS-TV customer can receive signals at 20 Mbps,
equivalent to several compressed digital television signals, while the mobile terminal cus-
tomer can receive only 10 kbps—a single voice channel.
Antenna gain is the system designer’s friend. Mobile systems will become much
more attractive economically when a self-steering, self-phasing, phased array is available
for mobile terminals which has even a moderate gain. A 10-dB increase in antenna gain
translates directly to a 10-fold increase in bit rates.
Elevation Angle Considerations

As we have seen in Chapter 8, rain attenuation can cause significant attenuation on a slant
path. At Ka Band (3020 GHz), even light rain can cause appreciable signal loss. Light
rain is usually stratified, so the higher the elevation angle the lower the rain attenuation
for a given rainfall rate. Figure 10.13 illustrates the geometry.
Most commercial satellite systems require that earth stations operate above certain
minimum elevation angles. For example, Intelsat requires that all earth stations using Intelsat
C-band (64 GHz) satellites operate above 5°, otherwise the earth station does not meet
Intelsat’s standard specification and must be qualified for operation on an individual basis.
To qualify an earth station on an individual basis is an expensive undertaking. At Ku band
(1411, 1412 GHz) the standard antennas in the Intelsat system are required to operate
above a minimum elevation angle of 10°. In creating the original Teledesic system archi-
tecture5, the overriding design input for the coverages was that no earth station should op-
erate at an elevation angle below 40°. This requirement, when coupled with an orbital height
of around 800 km, led to an unrealistic initial constellation of 840 operational satellites to
c10.qxd 28/08/02 19:39 Page 409
Paths to satellite
Height of melting layer
B C
Stratiform rain
A
Earth station
FIGURE 10.13 Illustration of the decrease in the path through rain as the elevation angle
to the satellite increases. Light rain is generally formed at the melting layer height in strati-
form clouds. The rain then falls fairly uniformly over a wide area. Freezing precipitation (hail,
ice crystals, dry snow) does not cause any appreciable attenuation to radio waves. Since the
rain is uniform, the attenuation per meter will be constant everywhere in the stratiform rain
shower and the total path attenuation will be given by the length of the signal path in the
rain. The higher elevation angle path (AC) will therefore suffer less attenuation than the
lower elevation angle path AB.
provide full global coverage. Most satellite systems now, whether for the mobile satellite
service (MSS) or the fixed satellite service (FSS) at frequencies above 10 GHz, tend to
limit the elevation angle of the user to no less than 10° so that reliable service can be
provided. Given a minimum elevation angle and an orbital height, the geometry setup in
Figure 10.12 can be used to develop a coverage area, assuming that the satellite has a sym-
metrical beam aimed at nadir. A plot similar to that shown in Figure 10.14 results.
Track of sub
satellite point along
surface of earth
Movement of coverage
area under satellite
Orbital path
of satellite
Earth
FIGURE 10.14 Illustration of coverage area under a satellite. In this example, an NGSO
satellite moves along a path over the earth with a nadir pointing antenna, that is, the an-
tenna has its electrical axis directed straight down toward the subsatellite point. The antenna
will have a finite usable beamwidth which will allow a given portion of the surface to be illu-
minated at the same time. This is shown as the shaded portion in the figure. Increasing the
altitude of the satellite’s orbit will increase the coverage area. Alternatively, the altitude can
remain fixed and the beamwidth increased in order to cover a larger surface area.
c10.qxd 28/08/02 19:39 Page 410
The shadowed area on the earth in Figure 10.14 is the maximum instantaneous cov-
erage on the surface of the earth that can be achieved from that satellite. The calculation
process for the instantaneous coverage has as input the minimum elevation angle a user can
tolerate and the orbital altitude selected. (Instantaneous coverage means that, if a snapshot
were taken and the motion of the satellite frozen at an instant in time, the region of the earth
covered by the satellite’s antenna at that particular time would be the instantaneous cover-
age from that satellite. The word instantaneous is used to separate out coverages that are de-
veloped by scanning or hopping satellite antenna beams. For the scanning or hopping beam
concepts, full coverage is obtained by moving elemental beams around the coverage area to
pick up traffic. A full, and instantaneous, coverage of the observed region is therefore not
obtained with scanning or hopping beams.) The instantaneous coverage from a satellite,
however, is not always served by one beam from the satellite antenna due to the lack of
available spectrum and a concomitant need for extensive frequency reuse. This is particu-
larly true for MSS systems which, like terrestrial microwave cellular systems15, have to divide
up their coverages into cells covered by separate beams in order to provide enough capacity
into a given cellular structure. Each cell, here a separate beam from the satellite antenna,
will have a portion of the spectrum allocated to it. The simplest spectrum reuse pattern is a
three-cell configuration. The spectrum is divided into three roughly equal portions and a
three-cell pattern is built up over the coverage area. Figure 10.15 illustrates the concept.
There are many other cell reuse patterns that are possible15.
Spectrum A
Spectrum B
Spectrum C
Instantaneous coverage
FIGURE 10.15 Illustration of a three-cell reuse pattern. The instantaneous coverage of the
satellite antenna is shown as the circle with a broken line. Within this coverage, individual
beams formed by the satellite antenna make a regular pattern that fills up the instantaneous
coverage. The spectrum that has been allocated to this satellite has been divided up into
three portions, called Spectrum A, Spectrum B, and Spectrum C. These different spectra are
indicated with different shadings above. None of the three spectra are adjacent to the same
spectral allocation. Note: In general, each of the individual beams will overlap their neigh-
bors for two reasons. First, by overlapping the individual beams, there are no “holes” in the
instantaneous coverage. Second, physics will prevent the beams going from full power to
zero power over a negligible distance. It is usual to develop coverages using the half-power
(3 dB down) point of the beams as the edge of coverage gain/power. There will therefore be
energy spilling over into adjacent beam coverages. This is why it is necessary to employ a
different spectral allocation in adjacent beams, unless a CDMA access technique is used.
c10.qxd 28/08/02 19:39 Page 411
TABLE 10.3 Frequency, Antenna, and Capacity Characteristics

of the Big LEOs
Parameter Iridium Globalstar New ICO
Mobile user link

Frequency (up/down) (GHz) 1.62135–1.6265 1.619–1.6215/ 1.980–2.010/
2.4835–2.4985 2.170–2.200
Maximum bandwidth (MHz) 5.15 11.35 30
Spot beams per satellite 48 16 163
Nominal capacity per satellite 1,110 2,400 4,500
(voice circuits)
Orbital altitude (km) 780 1,414 km 10,355 km
Source: [13, 14].
Number of Beams per Coverage

The very small spectrum allocation available for MSS systems (
50 MHz), and the many
competing systems that aim to provide mobile satellite services, place a number of con-
straints on the system design. Table 10.3 shows the spectrum, antenna, and resulting
capacity of the three major MSS systems: Iridium, Globalstar, and ICO-Global, known
generically as Big LEOs. ICO emerged from Chapter 11 bankruptcy protection in 2000
with a new name: New ICO. At the time of going to press, it was not clear how the New
ICO satellite payload design was being affected by the proposal to merge New ICO MSS
applications and Teledesic Internet-like services.
New ICO is an MSS system spun off from the International Maritime Satellite
Organization (Inmarsat). While it uses a medium earth orbit architecture, with 10 satellites
in two planes at an altitude of 10,355 km above the earth, New ICO is generally consid-
ered together with Iridium and Globalstar as one of three Big LEO MSS systems. Because
the New ICO satellites are so much further away from the earth than both Iridium and
Globalstar satellites, the New ICO satellites need to generate many more beams per satel-
lite within each instantaneous coverage in order to achieve sufficient capacity/km2 within
the coverage. Figure 10.16 presents a comparison of the Iridium and New ICO array of
multiple spot beams developed within their respective instantaneous satellite coverages.
The requirement placed on the MSS satellite antenna to generate multiple beams
within a given instantaneous coverage is a key driver in the payload technology. Tradi-
tional satellite antennas have evolved from simple, front-fed reflector antennas with one
feed horn, to offset-fed designs with more than a hundred feeds16. Such multiple feed horn
reflector antenna designs are necessarily large and heavy. The greater the number of in-
dividual beams to be generated, the heavier the reflector antenna and associated feed horns
and beam forming network. Depending on the precise spacecraft mission, there is a thresh-
old where the cost and complexity of a phased array antenna implementation will be less
than that of the equivalent reflector antenna.
A phased array antenna usually has a nonmechanically steered array of radiators.
The radiating elements can be passive devices (e.g., dipoles or feed horns) or active de-
vices (e.g., patch elements that include amplifiers). The steering of the beam is carried
out by varying the phase (and amplitude for full sidelobe control) of the signal in each
radiating element. For a passive device, the phase control is achieved in the feed matrix
placed between the high-power amplifier and the radiating antenna elements while, for
the active device, there is a phase shifter per element per beam. In many cases, it is possible
c10.qxd 28/08/02 19:39 Page 412
+20°
+10°
0°
−10°
−20°
+30° +20° +10° 0° −10° −20°
(a )
90
60
30
Latitude
−30
−60
−90
−180 0 180
Longitude
(b)
FIGURE 10.16 (a) User spot beams developed by an Iridium satellite (Figure 11 of refer-
ence 13 © 1997 IEEE, reproduced with permission). The satellite covers about 40° of the
earth’s surface from its orbital height of 800 km, which translates into about 4000-km diameter
main coverage. This coverage is divided up into 48 spot beams. Each of the spot beams has
the same beamwidth: the curvature of the earth has caused the outer spot beams to appear
elliptical. The boundary of each spot beam denotes the 3-dB-down point of that spot beam.
(b) User spot beams developed by an ICO-Global satellite (Figure 20 of reference 13 © 1997
IEEE, reproduced with permission). The satellite covers about 110° of the earth’s surface from
its orbital height of 10,355 km, which translates into about 12,000-km diameter main coverage.
This coverage is divided up into 163 spot beams. Each of the spot beams has the same
beamwidth: the curvature of the earth has caused the outer spot beams to appear elliptical.
The boundary of each spot beam denotes the 3-dB-down point of that spot beam. Note that
the spot beam size of the ICO-Global satellite is similar to that of the Iridium satellite.
to include the amplifier as part of the active phased array radiating element. This partic-
ular phased array concept is referred to as a direct radiating phased array. Figure 10.17
illustrates the two phased array approaches. In either approach, the scan angle is often a
critical design limitation.
Off-Axis Scanning
The design of a point-to-point wireless communications system requires that the anten-
nas at either end be directed toward each other for maximum gain advantage. This was
the approach adopted for the fixed service (FS), the terrestrial microwave communications
c10.qxd 28/08/02 19:39 Page 413
Direction
of beam Phase front
of beam
A A A A A A A
φ φ φ φ φ φ φ
HPA
Input signal
(a )
Direction Phase front

of beam of beam
Direct radiating
element
G, G, G, G, G, G, G,
φ φ φ φ φ φ φ
Input signal
(b )
FIGURE 10.17 Illustration of scan angle control mechanisms for phased array antennas.
(a) Passive phased array. (b) Direct radiating array. In (a), the high-power amplifier (HPA) has
had the power divided up among a number of different feed lines. Each feed line is acted on
by a variable phase-change () and a variable attenuator (A) device. The resultant output sig-
nal is then fed to a passive feed horn. The sum of the many phases and amplitudes gener-
ated by the feed horn cluster will develop the antenna coverage. In (b), the phase and ampli-
tude are controlled by the direct radiating device at the end of the feed line. The amplitude is
controlled by the gain of the radiating amplifier, G, and the phase, , can either be controlled
within the amplifier unit itself or by a phase element associated with the radiating device. To
develop a large number of beams, many signal lines will feed each element and a complex
phase front will be developed. Each beam direction will be given by the composite phase of
the associated phase front for that signal; each beam shape will be given by the number of
individual elements contributing to the development of that phase front.
service. If the transmitting antenna has to communicate with more than one receiving
antenna, and these antennas are located in different positions, a compromise must be
reached between the gain of the transmitting antenna toward the various receiving anten-
nas. In this case, most, if not all, of the receiving antennas will not be on the boresight
(main beam axis) of the transmitting antenna. Figure 10.18 illustrates the problem.
c10.qxd 28/08/02 19:39 Page 414
Main lobe of Boresight of

transmit antenna transmit antenna
Receive Transmit
antenna antenna
(a )
Main lobe of
transmit antenna
Set of receive
antennas
Required total
scan angle
Transmit
antenna
Boresight of
transmit antenna
(b )
FIGURE 10.18 (a) Point-to-point line-of-sight terrestrial communications link. The transmit
antenna illuminates the receive antenna along its electrical boresight, providing the
maximum gain for the link. The receive antenna has its electrical boresight directed toward
the transmit antenna (not shown explicitly here). The transmit and receive gains are
therefore maximized. (b) Point-to-multipoint line-of-sight terrestrial communications link. In
this plan view of a point-to-multipoint system, one example of which is called LMDS—Local
Multipoint Distribution System, the transmit antenna has to cover a number of receive
antennas spread over a large total scan angle. There are two main options available to the
link designer: use a single, wide-angle beam to cover the receive antennas (as has been
illustrated here); or set up several different transmit antennas, each directed toward a given
receive antenna. In this latter concept, the function of the transmit antenna can be divided
up among a small group of antennas that provide 360° coverage, as in the sectored antenna
approach of cellular systems.
Exactly the same design compromise illustrated in Figure 10.18b faces satellite sys-
tem designers who have to provide coverage over a large instantaneous area from a sin-
gle satellite. A satellite is a prime example of a point-to-multipoint system. There are two
basic input geometrical parameters that are used in the initial design phase of a satellite
antenna: the orbital height and the instantaneous coverage requirements for a single satel-
lite. Figure 10.19 presents the three main design options for orbital altitude: LEO, MEO,
or GEO and Table 10.4 lists some scan angle requirements for various satellite altitudes,
with atmospheric refraction ignored.
c10.qxd 28/08/02 19:39 Page 415
Total scan angle
LEO MEO GEO
FIGURE 10.19 Schematic of the total scan angles for LEO, MEO, and GEO satellites. The
further away the satellite is from the earth, the smaller the satellite antenna scan angle
needed to provide an instantaneous coverage out to a given user elevation angle minimum.
In the above figure, the satellites are all in an equatorial orbit. The view is from above the
earth with one side of the earth in sunlight and the other in darkness. The terminator is
directly under all three satellites.
A fixed antenna with a parabolic reflector is able to scan its main beam away from
the electrical boresight axis by repositioning the feed transversely from the prime focus.
However, the plane wave that is present in the aperture of a focused parabolic reflector
antenna becomes distorted when the feed horn is moved away from the focus, resulting
in an effect known as coma. Coma causes a reduction in antenna gain, an increase in
sidelobe levels, and an increase in cross-polarization. The reduction in gain and polariza-
tion purity can be held to relatively small values if the focal length, f, of the antenna is
long with respect to the antenna diameter, D, and the off-axis scan angle is small. A value
of fD 1 is generally taken as the required design goal, and is usually implemented by
employing a double reflector configuration such as the Cassegrain antenna. Cassegrain
antennas have a large equivalent fD ratio while being mechanically compact. GEO satel-
lite antenna designs that scan over the full earth coverage (8.7°, which corresponds to
TABLE 10.4 Scan Angle and Latitude/Longitude Ranges

for Different Satellite Altitudes
LEO MEO GEO

Orbit
and orbital height 750 km 1,800 km 10,000 km 14,000 km 35,786 km
Scan angle 57.2° 47.1° 21.5° 17.1° 8.25°

Latitude/longitude 12.8° 22.9° 48.5° 52.9° 61.8°
range
The minimum elevation angle to the user terminal for the data in this table is 20°. The
coverage is assumed to be a cone of revolution around a nadir pointing direction. Note
that, even though the scan angle is smaller for satellites at higher altitudes, the latitude
(and longitude) coverage increases with altitude. Thus, for a given scan angle, instanta-
neous coverage increases with satellite altitude.
Source: Some of the data have been extracted from references 7 and 13.
c10.qxd 28/08/02 19:39 Page 416
0° elevation angle at the extreme earth station locations) have been successfully imple-
mented using reflector antenna technology. However, as the scan angle requirement
increases (due to the satellite moving from GEO to a lower orbit) and the number of in-
dividual beams that are needed within the instantaneous coverage grows large, phased
array antennas prove to be the best match to the system design17. Phased array antennas
are used on the Iridium, Globalstar, and New ICO satellites, while Cassegrain or Gregorian
(another double reflector configuration) antennas are used on many large GEO satellites.
A number of factors influence the coverage of a phased array antenna from a given
satellite. While the phase of the signal in the radiating elements determines the steering
of the beam, it is usual to have the main beam axis (generally normal to the surface of
the antenna array panel) directed at nadir. This will lead to the edge-of-coverage users suf-
fering two loss components that are larger than a signal transmitted in the nadir direction.
First, they will be further away from the satellite, and so will suffer a free space path loss
that is increased with respect to a nadir user. For the LEO example in Table 10.4 that is
at an altitude of 750 km, the 57.2° scan angle leads to a slant range of 1681 km at the
20° edge of coverage. The difference in path loss between the range at nadir (750 km)
and edge of coverage (1681 km) is 7.0 dB. The 1800-km orbit LEO satellite has a 5.5-
dB path loss difference between nadir and edge of coverage. For mobile satellite systems
that must operate with this rapid variation in path loss as the satellite passes by the user,
power control is employed to offset changes in perceived power level at both the satellite
and the earth terminal (the handset). The second higher loss component experienced by
edge of coverage users is that the satellite antenna will incur a scan loss as it attempts to
direct energy away from the main beam axis (the boresight directed at nadir) out to the
edge-of-coverage user. Figure 10.20 illustrates the change in path loss with scan angle.
Satellite
Total instantaneous
coverage angle (scan
d EOC angle)
dN
Boresight
directed at
nadir
Edge
Edge of
of coverage
coverage (EOC)
(EOC)
Instantaneous
coverage region
FIGURE 10.20 Illustration of path loss and scan angle loss evaluation for a phased array.
The phased array has its prime axis pointed at nadir. The energy received at nadir from the
satellite will be greater than that received at edge of coverage (EOC) for two reasons. First, the
path loss will be less since the nadir distance, dN, is less than the EOC distance, dEOC. Second,
there will be a scan loss associated with the antenna reaching out to cover the EOC region.
c10.qxd 28/08/02 19:39 Page 417
Satellite
Total
instantaneous
coverage angle
Total instantaneous
coverage angle (the
scan angle)
Scan angle,
φ, away
from nadir
Nadir
direction
3-dB
Individual beam formed beamwidth of
over edge of the individual
instantaneous coverage beam
Section through surface of earth
Required instantaneous
coverage
FIGURE 10.21 Illustration of the scan angle of an individual beam within an instantaneous
coverage. The instantaneous coverage is developed through many smaller beams spread
over the region to provide sufficient frequency reuse for the users in that area. Only one of
the small, individual beams is shown above as a shaded area on the right. This beam is
scanned to the edge of the instantaneous coverage. Note that, a user within the small, indi-
vidual beam will have to factor two components into the link budget: (1) the gain loss due to
not being at the center of the individual beam; and (2) the scan loss due to not being at the
nadir point of the instantaneous coverage region (here assumed to be the boresight of the
phased array antenna on the satellite).
Scan loss for a phased array antenna normally follows the relationship17
Scan loss 1cosine f2 k (10.11)
where is the scan angle off boresight and k is an empirical number between 1.2 and
1.5. Note that the negative sign in Eq. (10.11), which was not incorporated in the refer-
enced article, allows the sign on both sides of the equation to agree [see the example in
Eq. (10.12) below]. Figure 10.21 illustrates the geometry for Eq. (10.11).
A typical value to use for k is 1.317. For example, a LEO system that needs to scan
57.2° away from boresight will have a scan loss
Scan loss 1cosine 57.22 1.3 0.4507 1 3.5 dB (10.12)
Thus the scan loss is 3.5 dB for a beam transmitted to the edge of an instantaneous cov-
erage of 57.2°. The edge-of-coverage path will also suffer an additional path loss com-
pared with the nadir path of 7 dB for a LEO satellite orbiting at a height of 750 km. The
edge of coverage signal is therefore 10.5 dB below the nadir signal in this example. To
counteract these two loss components—scan loss and enhanced path loss—the phased
array boresight can be redirected toward the edge of coverage. The problem with this
approach is that at least three phased array panels are required on the spacecraft to
c10.qxd 28/08/02 19:39 Page 418
Bus section
330 inches
Solar
array panel
Battery and
radiator
Main mission 180 inches

antenna panel Communications
section
Gateway
antenna Cross-link
antenna
FIGURE 10.22 A sketch of an Iridium satellite (Figure 10 of reference 13 © IEEE 1997, re-
produced with permission, courtesy of Motorola). One of the three phased array antennas
is shown as the “main mission antenna panel” in the above figure. The instantaneous cov-
erage is developed using these three phased array antennas, much like a three-sector mi-
crowave cellular coverage within a cell. Iridium uses an FDMA/TDMA multiple access tech-
nique. Since communications can be established from two of the antennas into an area
that is on the “joint” between the two sector coverages, the signals must be accurately
controlled in time so that they arrive at all three antenna array panels at the same instant
and the TDMA bursts do not overlap in time. Not shown clearly in this figure are the four
ISL antennas that communicate with the other satellites in the constellation. Two of the
four antennas link “forwards” and “backwards” within the same orbital plane while the
other two link eastwards and westwards. Iridium ISL links are only possible with satellites
that are moving in the same general direction. That is, a north-going satellite cannot estab-
lish an ISL link to a south-going satellite in another plane.
illuminate the full instantaneous coverage. With three antenna array panels, the scan loss
is reduced by at least 1.5 dB17. This solution was adopted by Iridium. The three phased
array antenna panels can be seen clearly in Figure 10.224,17. In addition to the required
antenna scan angle, the height of the orbit is the other key geometrical parameter that
influences the design of a LEO constellation.
Determination of Optimum Orbital Altitude

The locations at the edge of coverage within the instantaneous coverage region normally
present the greatest problems in the design of a satellite service. It is at the edge of cov-
erage that the power flux density into the user terminal is at its lowest. Even if the user
is at the center of the individual beam that serves the edge of coverage (see Figure 10.21)
there are still the two additional factors that determine whether the link can provide ade-
quate service: the scan loss and the added free space path loss at edge of coverage when
compared with nadir. Minimizing the total additional loss in the transmission path to edge
of coverage is a design goal. If the orbital altitude is increased, the free space path loss
will increase, but the scan loss will decrease. For a LEO constellation of satellites with a
given large instantaneous coverage requirement, as the orbital altitude increases from a
minimum of 500 km, the scan loss will decrease faster than the path loss increases. There
will therefore be an optimum altitude for a LEO constellation, based on the number of
satellites per plane and whether more than one satellite must be in view to any given user
c10.qxd 28/08/02 19:39 Page 419
30 0.
−0.1
25
Minimum grazing angle (°) −0.2
20
Relative loss (dB)

−0.3
15 −0.4
−0.5
10
−0.6
5
−0.7
0 −0.8
1000 1400 1800
Satellite altitude (km)
FIGURE 10.23 Relative transmission loss and minimum grazing angle vs satellite altitude
for a constellation of LEO satellites with 10 satellites per plane and double coverage (Figure 1
from reference 18 reprinted with permission from Microwave Journal). In the calculations for
this figure, it was assumed that there would always be two satellites in view for any user.
With 10 satellites per plane, a minimum scan angle at the satellite is calculated and, from
this, the elevation angle for the edge-of-coverage user is found. This angle is referred to as
the minimum grazing angle in the figure. The scan loss free space path loss for edge of
coverage are normalized to an orbital height of 800 km. As the orbital height increases, the
scan loss free space path loss reaches a shallow minimum between about 1350 km and
1800 km altitude above the earth.
at all times. Figure 10.23 shows this trade-off for a constellation with 10 satellites per
plane and double coverage (i.e., two satellites always in view from all possible user sites).
Figure 10.24 shows a similar trade-off for a constellation with 15 satellites per plane with
the same double coverage requirement18.
40 0.4
Minimum grazing angle (°)
30 0.2
Relative loss (dB)
20 0
10 −0.2
0 −0.4
1000 1400 1800
Satellite altitude (km)
FIGURE 10.24 Relative transmission loss and minimum grazing angle vs satellite altitude
for a constellation of LEO satellites with 15 satellites per plane and double coverage (Figure 2
from reference 18 reprinted with permission from Microwave Journal). The calculations for
this figure are similar to those carried out for Figure 10.23, except for this constellation there
are 15 satellites per plane. With 15 satellites per plane, a minimum scan angle at the satellite
is calculated and, from this, the elevation angle for the edge-of-coverge user is found. This
angle is referred to as the minimum grazing angle in the figure. The scan loss free space
path loss for edge of coverage are normalized to an orbital height of 800 km. As the orbital
height increases, the scan loss free space path loss reaches a more pronounced minimum
than that in Figure 10.23. This time the minimum is between about 950 and 1300 km above
the earth rather than 1350 km and 1800 km altitude found for a constellation with one-third
the number of satellites per plane.
c10.qxd 28/08/02 19:39 Page 420
Figures 10.23 and 10.24 indicate that there are a number of iterative analyses that
can be performed to balance scan loss, free space path loss, number of satellites in view
at any user site, and orbital altitude. Once these “geometrical” analyses have been per-
formed, it is necessary to look at other factors. From the satellite hardware design aspect,
a critical factor is the radiation environment: the higher the LEO orbit altitude, the worse
the radiation environment becomes as it approaches the first main Van Allen radiation belt
at around 1500 km. Perhaps the most critical factor is the RF transmit power available
from the user’s handset. Battery and adaptive handset antenna technology may be able to
increase the available EIRP from the user’s phone, but the biological radiation limits im-
posed for safe usage will place an upper bound on handset EIRP.
Radiation Safety and Satellite Telephones

In the United States, the Federal Communications Commission (FCC) mandates strict lim-
its on radiated power levels throughout the spectrum. Their rules are usually issued in
dockets (see, e.g., reference 22, which provides the main FCC Web site and the dockets
for evaluating the environmental effects of radio frequency radiation). The Office of
Engineering of the FCC has also posted an RF Safety Program on its Web site23, which
provides guidance on the specific absorption rate (SAR) for wireless phones and devices.
Many of these guidelines have been developed through IEEE Committees (see e.g., ref-
erence 24), which have made many proposals to the American National Standards Insti-
tute (ANSI) on this issue. Safe exposure levels are given as 0.08 W/kg as averaged over
the whole body for the general population and 0.4 W/kg for occupational or controlled
exposure for professionals working in this area. These values do not provide enough in-
sight into handheld units held close to the head and many studies are still underway, some
of which are reported in reference 25. It is clear that handset power levels are well below
those that cause ionization damage to tissue. However, while the short-term effects of the
power levels used in handsets have been proven to be negligible, there have been insuffi-
cient studies at present to provide the long-term effects of such exposure, that is, over
more than 10 years of handset use. A number of international groups, in particular ETSI
in Europe, are collaborating on such studies.
Projected NGSO System Customer

Service Base
A single satellite in an NGSO system will not provide continuous 24-h coverage over a
given area. If a national or regional coverage is desired, a constellation of NGSO satel-
lites is required with orbits tailored to match the coverage. This was the approach adopted
for the Molniya system where a minimum of two satellites in two Molniya orbits could
provide continuous 24-h service. Most of the new NGSO systems have been aimed at
mobile users. For mobile users, the problem is to generate sufficient transmit power in
a handheld terminal without exceeding the limits for electromagnetic radiation from the
antenna into the head and body of the user. Low power transmissions from the handheld
unit requires either a satellite in low earth orbit or a very large antenna on a MEO or
GEO satellite. All three alternatives have been developed13. The driving forces behind
the decisions made in choosing a system architecture will be discussed in Section 10.5
and some typical systems will be reviewed in Section 10.6. We will now look at two
closely associated elements of an NGSO system—or any telecommunications system for
that matter—that can have significant implications on customer acceptance: delay and
throughput.
c10.qxd 28/08/02 19:39 Page 421
10.4 DELAY AND THROUGHPUT CONSIDERATIONS 421
10.4 DELAY AND THROUGHPUT

CONSIDERATIONS
Delay in a communications link is not normally a problem unless the interactions between
the users are very rapid—a few milliseconds apart in response time. Long delays, such as
those associated with manned missions to the moon, required the development of agreed
procedures, much like tactical military or police communications requires specific hand-
off code words such as “over” to signal the end of one user’s input. For most commercial
satellite links that are over long distances, particularly those with satellites in geostation-
ary orbit, the main problem was not delay, but echo. A mismatched transmission line will
always have a reflected signal. If the mismatch is large, a strong echo will return. Over a
GEO satellite link, the echo arrives back in the telephone headset about half a second af-
ter the speaker has spoken, and usually while the speaker is still speaking. This will in-
terrupt the speaker and the conversation becomes fragmented. The development of echo
suppressors and, even better, echo cancellers, solved the problem. Figure 10.25 illustrates
the one-way propagation time for a typical LEO, MEO, and GEO system.
Based on the calculations shown in Figure 10.25, the time delay for a signal passing
between LEO user 1 and LEO user 2 in the same instantaneous coverage is 5.4 ms (2.7 ms
up and 2.7 ms down) and the go and return (round-trip) delay between the two users is twice
this at 10.8 ms. It is rare, however, for a user to be immediately underneath a LEO satellite
and, for LEO satellites in higher orbits, the round-trip delays due to propagation time can
be more than double this. Globalstar, which has a maximum path length from the satellite
to the user of 2500 km, will have a maximum round-trip delay time of 33 ms. For GEO
users, the up and down (forward) link delay is typically 230 ms with the round-trip delay
460 ms. However, Figure 10.25 does not tell the whole story. Most MSS systems use voice
compression to reduce the bandwidth required for a single voice channel. The coded bit rates
for a single voice channel range from 2.4 kbit/s for Globalstar to 6.25 kbit/s for Iridium13.
The vocoders sample the incoming analog voice signal and produce excellent, low data rate
digital reproductions—but at a price in delay. The access scheme can also add additional
delay. If the channel is operated in a simplex fashion, i.e., you cannot send at the same time
as you are receiving, there can be a delay in response. The Iridium TDMA access mode
uses a time division duplexing (TDD) scheme. A TDD scheme allows transmissions to oc-
cur for a certain period (while receive functions are off) and then transmissions cease while
receive operations are in use. In the present Iridium TDMA access scheme, eight users share
a frequency assignment and, within this frequency channel, share a 45-ms transmit frame
and a 45-ms receive frame. There can therefore be up to 90 ms between transmission and
reception of specific parts of a message. On the Iridium satellite, the onboard processing
system translates the received signal to baseband, the header address information is read,
and the appropriate route selected for onward transmission. The baseband signal is then re-
formatted, up-converted to the RF band, and transmitted. All of this takes time. The forward
delay (ground-to-satellite plus satellite-to-ground) within the same instantaneous coverage
averaged 153 ms in the initial operational tests of Iridium. A transoceanic link delay using
intersatellite links averaged 253 ms—almost the same as for a GEO satellite link. Delay can
also have an adverse effect on the throughput of the signal, as noted in Chapter 9. If the pro-
tocol used in the link is not adapted for the particular delay environment, appreciable re-
duction in throughput will occur. Customer acceptance of a service has been found to be
driven by three prime factors: access ability (i.e., can the required connection be obtained
immediately on request?), availability (i.e., once connected, will the call be dropped?), and
performance (i.e., is the error rate low and the throughput high?). Pricing will attract cus-
tomers but it will not keep them for long if all three prime factors are not met.
c10.qxd 28/08/02 19:39 Page 422
GEO satellite: 35,786 km
One-way delay:
119.3 ms
MEO satellite: 10,355 km
One-way delay:
34.5 ms
LEO satellite: 800 km
One-way delay:
2.7 ms
FIGURE 10.25 One-way propagation delay for the three orbits: LEO, MEO, and GEO. The
one-way delay figures shown above have been calculated assuming the radio signal propa-
gates at the speed of light in a vacuum, i.e., 3 108 ms. That is, no account has been taken
of any delay due to the refractive index of the atmosphere not being unity. Also, no account
has been taken of any processing delay imposed on the signal from any source coding,
channel coding, modulation, or access scheme used.
As a final element in the discussion on delay, it is worth noting the challenges

that face system designers when intersatellite links (ISLs) are employed to relay signals
around a LEO constellation. It is a fairly straightforward matter to design an ISL to con-
nect two GEO satellites or a LEO satellite to a GEO satellite: the relative motions are
not that large. Consider now a LEO system attempting to establish connections across
the constellation. The connections will have to be both in plane (i.e., around the same
orbit plane of that particular ring of satellites) and across planes. When the satellites
are close to the equator, the orbital planes are at their furthest separation and the rate
of change between two LEO satellites traveling in the same direction is at a minimum.
As satellites move closer to the poles, the more rapidly they have to steer their ISL
antennas to maintain contact. In some operational modes, Iridium switches off the across-
plane ISL links when the spacecraft are above latitudes of about 60° 13. In no case, how-
ever, can Iridium maintain an ISL link between planes where the satellites are moving
c10.qxd 28/08/02 19:39 Page 423
10.5 SYSTEM CONSIDERATIONS 423
“Seam” across which microwave

ISLs are unable to track LEO
satellites in adjacent plane
FIGURE 10.26 Schematic of the ISL seam in the Iridium constellation. The Iridium satellites
are in an orbit that is close to polar (86.5° inclination). There are four ISL antennas on each
satellite that are used to communicate with adjacent satellites. The ISLs operate at 23 GHz
and use solid reflector tracking antennas. The inertial mass of the antennas combined with
the need to have a stable satellite platform for the normal communications mode to the earth
limits the rate of change of the tracking mechanism. Satellites across the seam are traveling
at a closing speed of about 36,000 mph (58,000 kmh) and it is likely that only lightweight
optical ISLs will be able to track at the angular rates of change required across an LEO seam.
in opposite directions. There will therefore be a seam in the constellation across which
no ISL links can operate. This is illustrated in Figure 10.26. The revised Teledesic sys-
tem (once 840 satellites, then 288 satellites, and now fewer than 200 satellites) has re-
portedly been designed to operate across the LEO “seam” and so it is likely that the
ISLs will be optical and not microwave. Optical ISL antennas are much smaller and
lighter than microwave ISL antennas and so impose less tracking restrictions due to in-
ertial forces when under acceleration. Whether or not to use ISLs; whether to design to
operate across the seam if ISLs are used; selecting an orbital height, number of satel-
lites visible at any instant, coverage region; etc.; all interact in the overall system de-
sign. We will now look at other system considerations that can affect the design of the
satellite network in other respects.
10.5 SYSTEM CONSIDERATIONS
There are four important factors that influence the design of any satellite communication
system: incremental growth, interim operations, (satellite) replenishment options, and end-
to-end system implementation.
c10.qxd 28/08/02 19:39 Page 424
Incremental Growth
The 1964 decision by the Interim Communications Satellite Committee (soon to become
The International Telecommunications Satellite Organization Intelsat a few years later) to
select a GEO satellite system rather than a 12-satellite MEO system that was supported
by major entities on both sides of the north Atlantic at that time was driven by incremental
growth plans as well as by launcher technology. The primary international traffic route
was across the Atlantic Ocean, followed (a long way second) by the Indian Ocean region,
and (an even longer way third, at that time) by the Pacific Ocean region. The system could
be grown incrementally with a GEO architecture. The first GEO satellite, Early Bird, was
placed over the Atlantic Ocean region in 1965. For the first decade of operations, new
satellites were launched into the Atlantic Ocean region to replace satellites that had been
operating there. The satellites being replaced were moved to the Indian Ocean region and
the satellites replaced in the Indian Ocean region were moved to the Pacific Ocean re-
gion. It was not until INTELSAT VII that Intelsat specifically designed a satellite for the
Pacific Ocean region from scratch. This approach to incremental growth served Intelsat
well. By comparison, the new LEO and MEO mobile service systems now in operation
require all of the satellites to be in operation before full operations can begin. However,
most of the LEO and MEO system operators developed interim operations plans where a
reduced number of satellites could provide useful service.
Interim Operations
Interim operations for LEO and MEO systems serve two functions: they can bring a serv-
ice on line gradually, introducing the technology to the market while teething problems
are sorted out; and they can act as fall back plans should multiple satellite failures occur
over a short period. Nearly all of the LEO and MEO systems undertook such interim op-
erations. Orbcomm began commercial operations with less than half of its 36-satellite con-
stellation in place, thus becoming the first commercial LEO system to establish a revenue
stream. Globalstar began with 32 out of the planned 48-satellite constellation and New
ICO plans to start operations with six out of the planned ten-satellite constellation. Irid-
ium, since it uses ISLs to complete the network, required all 66 satellites to be available
before beginning beta testing in November 1998. The technical planning for interim
operations includes relaxing the number of satellites visible to any user at any particular
time, which lowers the number of satellites required to complete the constellation. The
elevation angle minimum for users is also usually lowered, the gaps between operational
satellites in the same plane are made symmetrical, and the orbits adjusted if possible to
maximize coverage over those parts of the day when user service requests are highest.
Most LEO constellations have at least four satellites per plane and multiple spacecraft
launches are used in the constellation buildup. A Pegasus launch vehicle carries eight
Orbcomm satellites into orbit, A Delta II carries five Globalstar satellites, and a Proton
carries seven Iridium satellites into LEO. When a satellite fails in service, there is an in-
orbit spare to take its place. If more than one satellite fails in a plane, additional satellites
must be launched to replenish the system.
Replenishment Options
Launching five or more satellites to replace one failed satellite makes little economic
sense. As a result, the LEO service providers use smaller rockets to replenish their system.
Table 10.5 lists the primary and replenishment launchers used by the Big LEO systems
and Orbcomm.
c10.qxd 28/08/02 19:39 Page 425
10.6 OPERATIONAL NGSO CONSTELLATION DESIGNS 425
TABLE 10.5 Primary and Replenishment Launchers for the Big-LEO Systems
and Orbcomm
Primary launchers Secondary/replenishment launchers

LEO/MEO system (number per launch) (number per launch)
Iridium Delta II (5) and Proton (7) Long March (2)

Globalstar [note 1] Delta II (5) and Soyuz (4) [Note 2]
New ICO [note 3] Atlas IIAS (1), Proton (1), [Note 4]
Zenit (1), Delta III (1)
Orbcomm Pegasus (8) Taurus (2)
Notes:
1. Globalstar initially selected Zenit but canceled the launch services contract when the first
launch failed with 12 satellites aboard.
2. Globalstar has not selected a replenishment launcher.
3. The Zenit rockets for New ICO are to be launched from a floating platform. Sealaunch is a joint
venture of Boeing and Hughes.
4. New ICO satellites are so large, and the orbital altitude so high, that only one satellite is
launched per rocket. Any of the selected rockets could act as replenishment launch vehicles.
End-to-End System Implementation

A communications system can be part of a larger network (e.g., just providing the long-
distance portion of the connection) or it can provide the full end-to-end system imple-
mentation, from user to user. AT&T and Intelsat, when they were first set up, did not
provide end-to-end service: AT&T provided long-distance capacity for local telephone
companies and Intelsat provided satellite capacity for entities such as AT&T to carry their
international traffic. Neither company interacted directly with the end user. Indeed, specific
laws or protocols prevented this from happening.
The design of an NGSO system will be heavily influenced by the decision on whether
or not to provide service directly to the end user. It will also be impacted by the decision
on whether or not to include established telephone companies in the delivery of the serv-
ice. By their very nature, mobile satellite systems have committed to serve the end user
directly. However, different approaches have been taken with regard to including estab-
lished telephone companies. Two examples of organizations that took opposite decisions
are Globalstar and Iridium. Globalstar elected not to bypass existing telephone companies
while Iridium did. These decisions led to a very different architecture for the two systems,
which will be discussed in the next section.
10.6 OPERATIONAL NGSO

CONSTELLATION DESIGNS
Seven satellite constellation designs are reviewed briefly in the following discussion, four
MSS offerings with multiple beams, one with single beam coverage providing both two-
way services and one-way store-and-forward services, and two Internet-multimedia
satellite systems.
Ellipso
The Ellipso constellation drew from studies of the world’s population distribution and
the potential market for MSS users. Figure 10.27 (abstracted from data in reference 19)
c10.qxd 28/08/02 19:39 Page 426
30
Percentage of population 25
20
15
10
0
S S S S S S S S S N N N N N N N N N
90– 80– 70– 60– 50– 40– 30– 20– 10– 0– 10– 20– 30– 40– 50– 60– 70– 80–
80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90
Latitude range south (S) and north (N) in degrees
FIGURE 10.27 Percentage of the world’s population living in the given latitude ranges
(data extracted from reference 19 and reproduced with permission). The data in the figure
show that more than 85% of the world’s population lives in the northern hemisphere.
Designing a satellite system that spends most of its time in the northern hemisphere
would therefore cover the world’s population more efficiently.
shows that more than 85% of the world’s population lives north of the equator. Addi-
tional studies19 concluded that an equatorial constellation of MEO satellites could serve
the bulk of the world’s population. Ellipso therefore adopted an incremental approach to
their service offering. The first set of satellites would be in a circular equatorial orbit.
The second set would be in elliptical equatorial orbit, with the ellipticity of the orbit de-
signed to provide dwell times over the regions of greater demand. The third set of satel-
lites would be in sun synchronous 3-hour orbits inclined at 116.6° to provide coverage
over the highly industrialized northern hemisphere regions. The equatorial orbit groups
of the Ellipso system are called Concordia™ and the sun synchronous group is called
Borealis™. Details can be found in Table 10.6. The Ellipso spacecraft is based on the
Boeing GPS satellite bus and up to five satellites can be launched by a single rocket. No
onboard processing is performed; the signals received at the satellite are transponded
down to gateway earth stations for onward routing via the terrestrial PSTN or satellite
network. No ISLs are used.
Globalstar
In a similar manner to Ellipso, Globalstar elected to develop a constellation that was
aimed at the populous regions of the earth. The Globalstar orbital planes are therefore
inclined at 52° to the equator, thus ignoring the sparsely populated high-latitude regions.
To minimize the power requirements of the user handset, the constellation altitude was
lowered to just below the first Van Allen radiation belt. This increased the total number
of satellites needed to 48. No onboard processing or ISLs are used; the signals received
at the satellite are simply transponded down and the gateway earth stations process the
signals for the onward routing (see Figure 10.28). Like Ellipso, service over water is
c10.qxd 28/08/02 19:39 Page 427
TABLE 10.6 System Parameters of Five NGSO Constellations Aimed at Data and Voice Communications
System parameter Ellipso Globalstar New ICO Iridium Orbcomm
Number of planes 1S3S5 6 2 6 4S5

Satellites per plane 1 7 then 8 5 11 4 8 then
1 7 and 2 3 then 4 8 and 1 4
1 7, 2 3, 2 5
Total complement 23 48 10 66 36
Orbital inclination 3 at 0°, 2 at 116.6° 52° 45° 86.5° 4 at 45°, 1 at 72°
Orbit type 1 circular (0°), Circular Circular Circular Circular (45° and 72°)
2 elliptical (0°),
2 sun synchronous
Orbital height (km) 1 circular 8050, 1414 10,255 780 775
2 elliptical 6149–8050,
2 sun synchronous 633–7605
Spot beams per satellite 61 16 163 48 1
Satellite lifetime 5 to 7 years 7.5 years 12 years 5 to 7 years 5 to 7 years
427
c10.qxd 28/08/02 19:39 Page 428
Instantaneous coverage
regions of two satellites
User 1
User 2
Gateway Gateway
earth station PSTN connection (fiber-optic cable or GEO satellite) earth station
FIGURE 10.28 Schematic of end-to-end connection for satellites that have no onboard
processing or ISLs. User 1 is in a different instantaneous coverage region than that occupied
by user 2. The signal from user 1 is picked up by the gateway earth station and relayed to the
gateway earth station of user 2. The signal is then sent up to the satellite from the second
gateway earth station and then down to user 2. If user 1 or user 2 is using a fixed telephone (or
computer) the signal would simply pass over the regular PSTN circuits and not via the space
segment. Because the users must be in line-of-sight contact with a gateway earth station, no
maritime traffic can be picked unless the ship is close to land (and a gateway earth station).
restricted to coastal regions where the satellite is within radio range of a gateway earth
station.
New ICO
ICO Global is the company that was spun off from the International Maritime Satellite
Organization (Inmarsat); New ICO is the company that emerged from bankruptcy
protection in 2000. Inmarsat was initially set up solely for the purpose of providing re-
liable communications to maritime traffic. Later, Inmarsat also provided aeronautical
services, in addition to priority links for safety communications, whether on land or sea.
New ICO, although primarily aimed at the LMS market (Land Mobile Services), also
needed to provide capacity for maritime links. New ICO elected not to include ISLs in
their system architecture nor any significant onboard processing. Since a LEO constel-
lation would not provide maritime coverage without ISLs, a higher orbit was necessary.
If little onboard processing is used, traffic routing from mobile to mobile would have to
be carried out at the gateway earth stations (as it is for Ellipso and Globalstar) necessi-
tating a double-hop link. A double-hop link involves two uplinks and two downlinks. (A
double hop is used in Figure 10.28; two different earth–space links are used to complete
the connection.) A double-hop configuration is not feasible for a GEO constellation
since the overall delay would be completely unacceptable at about 1 s. New ICO there-
fore adopted a MEO constellation. An inclination of 45° is used, but since the orbit
altitude is so high, full global coverage is possible. Toward the end of 2000, it was learned
that New ICO was modifying the payload and service offerings to provide two-way
Internet-like connections in a joint venture with Teledesic. At the time of going to press,
it is unclear how this synergy will evolve.
Iridium
The genesis of Iridium was formed around the need to communicate from anywhere to
anywhere on the surface of the world, even where no telecommunications infrastructure
c10.qxd 28/08/02 19:39 Page 429
existed. The system therefore must be stand-alone. From this—and the need for a low
power handset—came the concept of first 77, and then 66, almost-polar orbiting LEO
satellites linked via ISLs. Each of the satellites in the constellation acts as a switching
node. Uplink signals are received and demodulated at the satellite using onboard pro-
cessing to recover individual data packets at baseband so that the header information
can be read. Using this information and links to the network control stations, the next
node for each packet is determined and the packet is reformatted with the next address.
The baseband data packet is then processed and up-converted for transmission either
to the ground at L band directly to another Iridium user or at 20 GHz to a gateway
earth station, or over one of the four ISL links (at 23 GHz) to the next satellite in the
chain. Onboard processing is needed to carry out the entire message routing and for-
matting functions.
Orbcomm
Many research organizations and businesses need to obtain data from locations that are
either inaccessible on a regular basis or are moving within areas without good cellular
telephone coverage. Examples are buoys measuring water characteristics in rivers and at
sea, and delivery trucks. Tracking of high value cargo on trucks is another application that
needs to send a short message to a central station at regular intervals. A GPS receiver on
the cargo determines its location and this information is sent with an ID number via an
Orbcomm satellite. If the truck carrying the cargo is hijacked, its route can be followed
and the truck intercepted.
Much of this information is neither required in real time nor does it need a high ca-
pacity link. Orbcomm developed their system around this requirement and have orbited a
constellation of satellites with both two-way data communications and store-and-forward
capabilities (see Table 10.6). The satellites are lightweight (40 kg) and simple in design
and execution3. A single beam is used to develop the instantaneous coverage and no on-
board processing is used.
A terminal that is within the coverage area of a satellite and a gateway station (which
includes almost all of the United States) can send short messages to the gateway station
in real time. The message length is limited to a few hundred bytes. A terminal that has
data waiting to be uploaded for store-and-forward listens for the passage of a satellite and
then uploads its data when the satellite is in view. The data, in the form of a packet with
the address of the intended recipient, are stored and transmitted to a gateway station for
onward transmission to the recipient when the satellite is within range of the gateway sta-
tion. Orbcomm satellites carry short messages, with a relatively high cost per transmitted
bit. The system is therefore most attractive to users who want to send a small number of
high value bits, such as requests for help in emergency situations or tracking information
for high value cargo.
None of the five NGSO constellations above were initially designed to carry traffic
at rates higher than 10 kbps. This is not adequate for Internet access, which has emerged
as a potentially important requirement in mobile systems. Two NGSO constellations that
addressed this market from the outset are Skybridge and Teledesic.
Skybridge
Skybridge evolved a similar approach to coverage as Globalstar, by selecting an inclined
orbit that covers the major population densities. Like Globalstar, Skybridge satellites
c10.qxd 28/08/02 19:39 Page 430
carry a nonprocessing payload and do not have intersatellite links; so all traffic is
transponded down to the gateway earth stations for processing and onward routing. How-
ever, Skybridge satellites are intended to carry wideband traffic and therefore use fre-
quencies above 10 GHz. They chose to employ the same Ku-band frequencies as the
FSS service in GEO uses: 12.75–14.5 GHz on the uplink and 10.7–12.75 GHz on the
downlink. To allow successful coordination with existing FSS GEO systems, they elected
to prevent any operations (up or down) whenever a satellite look angle is within 10° of
the GEO orbital plane. This requirement led to a relatively large number of satellites
(80 vs 48) for the constellation. The decision not to use ISLs also required a very large
number of gateway earth stations (on the order of 200). Skybridge also uses the con-
cept of a fixed earth cell (see Figure 10.29). More details of Skybridge can be found in
Table 10.7.
Teledesic
Teledesic started from the same precept as Iridium, but is designed for Internet-like
data traffic rather than voice communication. Any user can access any other user or ISP
(Internet service provider) independent of location and the existing telecommunications
Position 1
Track of sub
satellite point along
surface of earth
Position 2
Coverage does not move

under satellite: cell
Orbital path is a stationary coverage on
of satellite surface of earth
Position 3
Earth
FIGURE 10.29 Concept of a stationary cell. Unlike the coverage of the NGSO satellite
show in Figure 10.14, a stationary coverage (or “fixed earth cell”) of an NGSO satellite
does not move with the satellite. The phased array antenna on the satellite steers the
beam, while the satellite transits, to keep the coverage on the surface of the earth con-
stant. As the satellite moves between positions 1, 2, and 3, stationary coverage is main-
tained on the surface of the earth. Separate antennas are used for communications cover-
age and gateway links. In this way, a gateway need not necessarily be within a given
stationary coverage.
c10.qxd 28/08/02 19:39 Page 431
TABLE 10.7 System Parameters of Two

NGSO Constellations Aimed at Internet
Multimedia Communications
System parameter Skybridge Teledesic
Number of planes 20 12
Satellites per plane 4 24
Total complement 80 288
Orbital inclination 53° 90°
Orbit type circular circular
Orbital height (km) 1469 1400
Spot-beams per satellite 18 —
Satellite lifetime 7 years 7 years
infrastructure. The concept of Teledesic is to provide a complete worldwide data com-

munications system above the surface of the earth using satellites, instead of on the
earth’s surface using fiber-optic cables. This requirement dictated the use of wideband
data links, onboard processing, and ISL links. To avoid the necessity of coordinating
with existing systems, Teledesic chose to move their operations completely into Ka
band. As noted earlier, to reduce the impact of rain, Teledesic also limited the eleva-
tion angle at which users could access the satellites (the mask angle) to 40°. The ini-
tial Teledesic constellation had a complement of 840 satellites (22 planes with 40
operational satellites per plane) plus 40 spare satellites in orbit. The orbital altitude
was later moved up from 700 km to about 1400 km, which reduced the number of
planes to 12, with 24 operational satellites in each plane (see Table 10.7). The early
estimates of Teledesic’s system cost were between $9 B and $12 B, using 840 satel-
lites. Reduction in the number of satellites to 288 lowered the cost significantly, and
further reductions in the number of satellites seem likely to make the cost of creating
the system more acceptable.
Other companies are seeking to provide Internet access from satellites with lower
cost solutions than those of Teledesic and Skybridge. One company aiming to do so
for a “mere” $2.6 B is Virtual Geosatellite20. However, for a given bit rate, no pro-
posed system is lower in cost than a GEO alternative21. The geostationary earth orbit
has the unique characteristic of providing data transfer by satellite at the lowest cost
per bit. None of the currently scheduled or operating LEO and MEO satellite constel-
lations has been able to demonstrate a significant added value from the use of their
particular service when a commercial return on investment is required. There is a clear
military requirement for many of the new constellations—from anywhere to any-
where—without any intervening infrastructure, but the growth in terrestrial cellular sys-
tems and optical fiber links has removed much of the potential commercial demand
for these new services. At the turn of the twenty-first century, more than 90% of all
Internet traffic flowed through about 30 metropolitan areas. If these conurbations are
connected via optical fibers or through high-powered spot beam antennas from GEO,
the remaining traffic is what a LEO or MEO system would pick up. The same is true
for cellular telephony: what the major cities do not provide leaves very little traffic for
a high priced LEO or MEO alternative. Table 10.8 gives some early 2000 data on
Internet traffic centers.
c10.qxd 28/08/02 19:39 Page 432
TABLE 10.8 Internet Traffic Centers
1. London (18 terabits/s) 9. Washington, DC (4.0 terabits/s)

2. New York (13.2 terabits/s) 10. San Francisco (3.9 terabits/s)
3. Amsterdam (10.9 terabits/s) 11. Toronto (3.5 terabits/s)
4. Frankfurt (10.5 terabits/s) 12. Chicago (2.7 terabits/s)
5. Paris (9.7 terabits/s) 13. Seattle (2.6 terabits/s)
6. Brussels (6.2 terabits/s) 14. Vancouver (2.5 terabits/s)
7. Geneva (5.9 terabits/s) 15. Tokyo (2.4 terabits/s)
8. Stockholm (4.4 terabits/s)
Source: The data above were extracted from Telegeography at

http://www.telegeography.com. The data provide the Internet capacity of the top 15
major cities in terms of their Internet traffic, averaged over a day26.
EXAMPLE 10.6.1 System Design

A company wishes to develop a LEO constellation that provides continuous global coverage.
They are restricted to an orbital height of 750 km due to user terminal power, operating time
between battery charges, and satellite launcher capabilities. The following design data are
required:
• The length of the coverage arc on the surface of the earth within the instantaneous coverage;
• The gain of the satellite antenna if one beam is to illuminate this coverage;
• The number of satellites needed to complete one plane with a suitable overlap; and
• The number of satellites needed to complete a global system.
Length of Coverage Arc Figure 10.12 illustrates the geometry of a satellite and user
terminal. If the minimum elevation angle is set at 10°, we know rs re 750 km (the orbital
height), 10°, and re 6378 km (average radius of the earth). We need to find the central angle,
, (angle ECZ in Figure 10.12) which will allow us to find the length of half of the arc under the
coverage-arc EZ. Using the sine rule, we have
sin 1d2 re sin 1angle SEC2 rs (10.13)
The angle SEC 90° 100° and this yields 61.7859 61.79°
If d 61.79°, then g 180 100 61.79 18.21°
Arc EZ is therefore given by re (with in radians) 2027.1 km. The diameter of the instan-
taneous coverage region is therefore 2 2027 4054 km and the coverage angle measured at the
center of the earth is 36.42°.
(Note that this assumes the coverage is symmetrical about the nadir pointing direction SC in
Figure 10.12.) Alternatively, we could have derived the same result by noting that the circumfer-
ence of the earth is diameter of the earth 6378 2 40074 km. The fraction of this
illuminated by the satellite is (2)(2) (with in radians) 0.1012. The total coverage diameter
arc 40074 0.1012 4055 km.
Gain of Satellite Antenna The angle in Figure 10.12 is half of the antenna beamwidth.
The full angle of the antenna beamwidth at the satellite is therefore 2 61.79° 2 123.6°.
The gain of an antenna can be related to the 3-dB beamwidth using the approximate relation-
ship
Gain ratio 33,000 13-dB beamwidth in degrees2 2 G (10.14)
which gives G 33,000 1123.6°2 2 2.16 1 3.3 dB
c10.qxd 28/08/02 19:39 Page 433
Instantaneous
coverage arc = 4054 km
Angle γ of instantaneous
coverage arc at center of
earth = 36.42°
FIGURE 10.30 Coverage results from Example 10.6.1. The coverage of one satellite
orbiting at an altitude of 750 km is shown. The coverage arc has been calculated to be
4054 km when a minimium elevation angle of 10° is assumed. The angle at the center of
the earth that subtends the instantaneous coverage arc, , has been found to be 36.42°. To
complete the coverage around the earth using this satellite configuration would require
(360)(36.42) satellites. Rounding to whole number yields a minimum number of 10
satellites per plane.
Number of Satellites per Plane We now have the situation set up in Figure 10.30. Since
each of the satellites will cover 36.42360 0.1 of the earth’s circumference, we will need a min-
imum of 10 satellites in one plane.
Total Number of Satellites for a Global System By the same logic used above,
if 10 satellites are required to complete coverage around (say) the equator, 5 complete planes of
satellites will be needed to complete the full global coverage. (Remember that one plane of satel-
lites, if in a polar orbit, will have satellites on both hemispheres of the earth, some going north-
wards and some southwards. There will therefore be 10 “slices” around the earth made up of
5 planes of satellites.) The total minimum number of satellites needed is therefore 50. It should
be noted that this is an absolute minimum number. In addition to coverage gaps potentially ex-
isting, there will be a need to have spare satellites in orbit to take care of satellite failures. Other
architecture requirements can now be imposed. One could be the need for simultaneous cover-
age of any user by two satellites. The coverage of each satellite is unchanged, but the satellites
must be half the distance apart so that two satellites always are in view throughout the constellation.
A second architecture rule might be that no user is required to operate below 20° (rather than
10°). There are many other possible variations, for example, covering only the latitudes between
65° of the equator, inserting elliptical orbits to increase dwell time over a particular region. It
is thus easy to see why some constellations need many dozens of satellites to complete the full
architectural requirements.
c10.qxd 28/08/02 19:39 Page 434
10.7 SUMMARY
Launching satellites into a geostationary orbit is a Limited by radiation dosage limits, the telephone
complex task that was not achieved successfully un- handsets for the mobile satellite systems could radiate
til 6 years after the first satellite was orbited in 1957. only feeble amounts of power. To compensate for the
The quest for the geostationary orbit grew from a pa- low power and minimal GT in the handsets, either
per by Arthur C. Clarke that established this orbit as technologically adventurous antenna systems were re-
the prime location for communications satellites. If quired in geostationary orbit or equally adventurous
the satellite is geostationary, the earth station antenna (as it turned out, but this time in an economic sense)
need not be steerable, greatly reducing the cost of the constellations of non-geostationary satellites were
system. Non-geostationary satellites, however, con- required for a global mobile telephony system.
tinued to be launched for a huge range of missions. Satellite constellation design is a complex mix of
TIROS satellites photographed the weather over the coverage requirements, capacity, and connectivity.
earth in sun synchronous orbits, orbits that precessed The iterations in a design will also have to bear in
to match the rotation of the earth around the sun. mind the radiation environment in space caused by
TRANSIT satellites began the experiments into nav- the Van Allen radiation belts and the need to have both
igational aids that were space borne. MIDAS, the first an incremental design philosophy and replenishment
early warning satellite; IRAS, the first infrared as- options for the satellites that fail. From the design ex-
tronomy satellite; and Explorer satellites that probed amples shown, it is clear that geostationary satellites
the inner reaches of space above the earth continued will always provide cheaper costs on a per bit basis
the ever expanding list of experimental spacecraft than a satellite system in any other orbit until tech-
launched into non-geostationary orbits for research nology breakthroughs enable smart antenna design
purposes. Molniya, which means flash of lightning in for the user handsets. Even then, care must be taken
Russian, was a satellite series that established the first that terrestrial systems have not taken away the cus-
regional satellite system in 1965. The Molniya orbit tomer base. This is a very interesting period in the evo-
has found many uses for a range of other satellite sys- lution of telecommunications. The Internet, and by
tems in non-geostationary orbit. implication digital data communications, has become
In the 1990s, a whole series of proposals arose for the greatest growth area in the transfer of information
constellations of non-geostationary satellite commu- globally. The extent to which satellites in non-geo-
nications systems, some in low earth orbit and others stationary orbit can fit into the global information in-
in medium earth orbit. The thrust for moving the satel- frastructure successfully will determine whether they
lites “down” from geostationary orbit can be summed have any commercial future in this field. That they
up in one word: power (or, more strictly speaking, four have a future in navigation (GPS and the European
words: EIRP). The non-geostationary satellite systems system, Galileo) and geographic information systems
were aimed at the mobile user and so required that the is assured, but the revenue streams in these two cases
user terminal antennas be essentially nondirectional. have yet to match the investment requirements.
REFERENCES
1. A. C. CLARKE, “Satellite Communications Systems,” 10. J. M. BENEDETTO, “Economy Class Ion-Defying ICs in
Wireless World, pp. 305–308, 1945. Orbit,” IEEE Spectrum, Vol. 35, No. 3, pp. 36–41,
2. J. E. ALLNUTT, Satellite-to-Ground Radiowave Propa- March 1998.
gation, Peter Perigrinus, Ltd., UK, 1989. 11. ITU-R Rec. 435-5, Prediction of Sky-Wave Field
3. http://www.orbcomm.com Strength between 150 and 1600 kHz, 1986.
4. http://www.iridium.com 12. K. MURSALA and T. ULICH, “A New Method to Deter-
5. http://www.teledesic.com mine the Solar Cycle Length,” Geophys. Res. Lett. Vol.
6. http://www.skybridgesatellite.com 25, pp. 1837–1840, 1998.
7. http://www.globalstar.com 13. J. V. EVANS, “Satellite Systems for Personal Communi-
8. GARY D. MORGAN and WALTER L. MORGAN, Principles cations,” IEEE Antennas and Propagation Magazine,
of Communications Satellites, ISBN 0-471-55796-X, Vol. 39, No. 3, pp. 7–20, June 1997.
John Wiley & Sons, 1993. 14. F. J. DIETRICH, P. METZEN, and P. MONTE, “The Glob-
9. PROFESSOR WATSON, private communication, 1996 (then alstar Cellular Satellite System,” IEEE Transactions on
at York University, now at Bath University, both in the Antennas and Propagation, Vol. 46, No. 6, pp. 935–942,
UK). June 1998.
c10.qxd 28/08/02 19:39 Page 435
PROBLEMS 435
15. THEODORE S. RAPPAPORT, Wireless Communications: Microwave Journal, vol. 42, No. 5, pp. 282–290, May
Principles and Practice, ISBN 0-13-375536-3, Prentice- 1999.
Hall, Englewood Cliffs, NJ, 1996. 19. http://www.ellipso.com
16. M. A. B. TERADA, “Reflector Antennas,” Wiley Ency- 20. http://www.virtualgeo.com
clopedia of Electrical and Electronics Engineering, 21. http://www.hcisat.com
John G. Webster ed., ISBN 0-471-13946-7, Vol. 18, 22. http://www.fcc.gov/oet/dockets/
pp. 360–379, 1999. 23. http://www.fcc.gov/oet/rfsafety/
17. T. J. SCHUSS, J. UPTON, B. MYERS, T. SIKINA, A. 24. ANSI/IEEE C95.1-1992, “IEEE Standard for Safety
ROHWER, P. MAKRIDAKAS, R. FRANCOIS, L. WARLDLE, Levels with Respect to Human Exposure to Radio Fre-
and R. SMITH, “The IRIDIUM Main Mission Antenna quency Electromagnetic Fields, 3 kHz to 300 GHz”,
Concept,” IEEE Transactions on Anennas and Propa- copyright of the IEEE, 1992.
gation, Vol. 47, No. 3, pp. 416–424, March 1999. 25. K. R. FOSTER and J. E. MOULDER, “Are Mobile Phones
18. P. CHIAVACCI, “The Influence of Phased-Array An- Safe?,” IEEE Spectrum, pp. 23–28, August 2000.
tenna Systems on LEO Satellite Constellations,” 26. DR. FENG, Virginia Tech, private communications, 2000.
PROBLEMS
1. What is the preferred orbit [approximate orbital the time between successive passes when the satellite
height (maximum and minimum, or average if close is directly over the earth station) if the true orbital
to circular), approximate orbital inclination, range of period is the following?
subsatellite points on the equator if geostationary, and c. 2 h
approximate orbit eccentricity] for a satellite that d. 6 h
needs to do the following:
e. 12 h
a. Observe the polar ice caps at least every 2 h;
f. Did anything look strange at first sight with your
b. Observe the Falkland Islands (Islas Malvinas) with answer to (e) above? If yes, what was it and can you
an elevation angle to the satellite from the surface of explain the answer you arrived at for (e), which looks
the islands of at least 60° for 6 h every 24 h; counterintuitive at first sight?
c. Observe the development of tropical cyclones in Note: (1) For this question, do not use a sidereal day
the southern Pacific Ocean and hurricanes in the in developing your answers; assume 24 h for the
northern Pacific Ocean for 24 h each day; earth’s rotational period. (2) If you saw nothing
d. Observe swathes of the earth below the satellite il- strange at first sight in your answer to (e) you are
luminated by sunlight from directly behind the satel- either an exceptional orbital mechanic or not very
lite on every pass the satellite makes on the sunlit side curious!
of the earth; and 3. The International Space Station (ISS) has an ex-
e. Make observations of interstellar X-rays for 1 h perimental package that will be located in the free-
per day when more than 40,000 miles from the flying module (FFM). The experiment in the FFM
earth, but be able to relay information back to earth was located there to avoid unnecessary vibrational
once per orbit when less than 1000 miles above the tremors and thruster accelerations from the ISS im-
earth; pacting the gravity sensitive biological experiments
Note: The satellite in question need only do one of on board the FFM. The experiments require contin-
the above five missions at a time, not all of them at uous communications with a main research labora-
the same time. tory located in New Mexico, so that real-time
monitoring and adjustment to the experiments can be
2. Why is it optimum (in terms of launch energy re-
made. This requires huge quantities of data and high-
quirements) to do the following?
resolution video to be transmitted 24 h per day to
a. Launch a satellite toward the east New Mexico. An average, one-way data rate of OC-
b. Launch a satellite from the equator 192 (approximately 10 Gbit/s) is needed for the
A fully steerable earth station antenna is on the equa- data/video link to New Mexico. A return link of OC-
tor and it observes a satellite that is in a circular equa- 1 (51.84 Mbit/s) suffices for the uplink control path
torial orbit moving in an easterly direction. It can to the FFM.
observe the satellite down to the horizon, which is at a. What is your preliminary outline system design so-
an effective elevation angle of 0° in every direction. lution for the data link between the FFM and the earth
What is the apparent orbital period of the satellite (i.e., station in New Mexico? Give justifications for your
c10.qxd 28/08/02 19:39 Page 436
choice of system architecture and frequency bands high apogee and a relatively low perigee—but
(see the note at the end of this question regarding the orbital period of a Tundra orbit is one sidereal
frequency bands). day rather than about half that amount. Three satel-
b. If 8-phase PSK modulation with 34-rate FEC is lites are each in a different Tundra orbit, with the
employed for the OC-192 downlink from the FFM, planes of the three Tundra orbits 120° apart. By
approximately what instantaneous bandwidth will be phasing the position of each of the three satellites
required for the link? in its Tundra orbit, there are always two satellites
visible at high elevation over the coverage region.
c. If the instantaneous bandwidth of a given
The Tundra approach therefore needs three satel-
transponder is limited to about 10% of the carrier
lites as opposed to the single satellite needed in
frequency, what is the lowest downlink carrier fre-
the GEO approach, but the Tundra approach re-
quency that can be contemplated for this link in the
quires far fewer terrestrial repeaters to achieve full
FSS bands?
coverage in built-up areas that cause blockage to
d. If you selected a preliminary answer to part (a) be- GEO satellite signals. In the system level example
fore you looked at parts (b) and (c), did you change below, the costs given are for example only and do
your mind once you knew the tentative answers to not represent in any way the true costs of either
parts (b) and (c)? of the proponent broadcast radio systems noted
e. If occasional shuttle blockages of the transmission earlier.
path, or other small interruptions occur in the link, If programming and content support, and the unit
that lead to outage intervals of up to 5 min per day, cost of the car radio system, are assumed to be the
what is the maximum data storage required for the same for both system approaches; the cost of a GEO
transmission buffers? satellite is $200 M each, including launch; the cost
f. The OC-192 data link has been designed to of the Tundra satellites is $125 M each, including
have some redundancy in the information capacity. launch; and the average cost of a terrestrial repeater
If there is a redundancy margin of 0.5% in the is $25 k each; find:
data transmission requirement (that is, for every a. If no terrestrial repeaters are needed for either the
1000 bits actually sent, normally only 995 bits are GEO or Tundra approach, which system is less ex-
required so there is spare capacity to send 5 bits pensive to bring to operational status, a single GEO
more of information within those 1000 bits, if the satellite or three Tundra satellites?
need arises), will the 5 min per day of anticipated b. If it is necessary to have an in-orbit spare for the
outage be handled by the OC-192 link transmission GEO approach (i.e., two GEO satellites need to be
rate? launched prior to broadcast radio services being of-
Note: Four FSS bands are 64 GHz (C band), 1411 fered) but no in-orbit spare is required for the three
GHz (Ku band), 3020 GHz (Ka band), and 5040 Tundra satellites (since they provide a strong meas-
GHz (VQ Band). Remember that the uplink carrier ure of in-orbit redundancy by virtue of their orbit
frequency is given first in each paired band. placement), which system is now the cheapest to bring
4. The broadcast of digital radio signals from satel- to operational status? (Still no terrestrial repeaters are
lites has a number of advantages and disadvantages. required.)
Some of the advantages are: crisp, clear sound; unin- c. If the operational control center costs for a
terrupted reception regardless of distance from your GEO satellite are $6 M per year for one satellite
home location; and the ability to listen, commercial and $1 M per year for each additional satellite; the
free, to your favorite radio station wherever you travel operational control center costs for a Tundra
within the coverage region. The main disadvantage is satellite are $12 M per year for one satellite and $4
the possible blockage by tall structures. This last is- M per year for each additional satellite; and an
sue may prove to be critical in the success, or other- operational lifetime of 10 years is assumed in this
wise, of two different approaches to the provisioning part of the problem, which system is now the
of such services. cheaper to bring into service and operate over a 10-
In the United States, two companies (XM Radio year lifetime (assuming two GEO and three Tundra
and Sirius Satellite Radio) have approached the satellites)? (Still no terrestrial repeaters are
system design of a broadcast satellite radio system required.)
in two very different ways. One has adopted a GEO d. Using the assumptions in part (c), if the Tundra
approach (www.xmradio.com) and the other a Tundra approach requires 10 times fewer terrestrial repeaters
approach (www.siriusradio.com). The Tundra orbit than the GEO approach, at what number of terrestrial
is similar in concept to the Molniya orbit—a very repeaters for the Tundra approach are the total costs
c10.qxd 28/08/02 19:39 Page 437
PROBLEMS 437
of the two satellite broadcast radio systems approxi- erage region that access the satellite is 1% at any
mately equal to each other? given time of those who have signed up with that
e. Given the answer to (d), which approach might be satellite service. That is, for every 10,000 customers
the better to pursue? State your reasons, giving as signed up, 100 are using the satellite at any given
many parameters that might influence your choice time.
between the two system approaches. (i) How many instantaneous communications chan-
f. Using all of the assumptions in (d), which of the nels must the satellite be able to provide if the
two approaches would you select—giving your rea- coverage region contains a potential user population
sons—for a digital satellite radio broadcast service to of 120 million?
be offered in (i) Indonesia, (ii) Europe, (iii) the Pacific (ii) If the bandwidth allocation for the satellite
Islands, and (iv) South America. communications service permits only 6000 chan-
5. This question concerns the effect of radiation on nels to be provided without frequency reuse, how
electronic equipment on a spacecraft. many separate beams will be required for this cov-
erage region, ignoring interference between indi-
a. What are the two principal effects that radiation
vidual beams?
has on electronic equipment?
(iii) If interference requirements between individual
b. What particles principally cause these effects? beams dictate that no frequency may be reused by
c. What is the prime generator of these particles? adjacent beams, what is the approximate new mini-
d. What causes the generation of these particles to mum number of separate beams that are required,
fluctuate with time? ignoring for the moment the exact geometry of the
e. Is there a periodicity in these fluctuations? If yes, beams? (Note: For this part of the question, assume
what is the approximate period of these fluctua- that the bandwidth allocation has been divided into
tions? three separate bands to develop the frequency reuse
pattern.)
f. Is there a particular region, or are there particular
regions, in near-earth space where concentrations of (iv) If the physical size of the total coverage region
these particles are to be found? If yes, what is it (or remains the same between cases (ii) and (iii) above,
what are they)? what can you say about the size of the new individ-
ual instantaneous beams in part (iii) within the over-
g. What orbital inclination of a satellite circling
all coverage?
the earth will cause that satellite to receive the
highest radiation dosage compared with other in- (v) In addition, what can you say about the effect
clinations? this change in size has on the communications
capability and satellite connectivity complexity
h. Which of the radiation particles is the strongest
issues?
and therefore the hardest to protect against?
i. What are two ways to reduce unexpected perform- 7. a. Define the terms “Zenith” from the point of
ance or decreased lifetime in electronic equipment view of an earth station on the surface of the earth
exposed to radiation? and “Nadir” from the point of view of an earth or-
biting satellite.
6. The design of a communications satellite’s antenna
b. A geostationary satellite is required to provide
is fundamental to that satellite’s ability to perform its
communications coverage over the whole of the earth
assigned task. The portion of the earth’s surface illu-
(i.e., a global horn antenna). What is the maximum
minated via that antenna is called the coverage. This
off-axis angle that needs to be contained within the
question considers the definitions of coverage, fre-
antenna coverage, measured from nadir?
quency reuse, and capacity issues.
c. A LEO satellite constellation is being designed
a. What is the difference between total coverage and to provide telecommunications service on a global
instantaneous coverage for a satellite antenna illumi- basis at a downlink frequency of 12 GHz. An
nating the surface of the earth? evaluation is being carried out to determine what
b. What is the fundamental difference between a would be the best altitude for the system (which will
“hopping” beam and a “scanning” beam that might use circular orbits), what is the best mix of number
be employed on advanced communications satel- of beams vs complexity, etc. An initial determina-
lites? tion of the optimum orbital height is 1400 km. As-
c. A satellite needs to provide communications ca- sume the minimum operational elevation angle to
pability over a given coverage region. It has been any part of the coverage on the surface of the earth
predicted that the average number of users in the cov- is 20°.
c10.qxd 28/08/02 19:39 Page 438
(i) What is the scan angle from nadir to edge of selected polar orbit for their system. Find the
coverage? following:
(ii) What is the difference in path loss from nadir to a. What is the minimum number of satellites that can
edge of coverage? be used to provide continuous coverage around one
(iii) What is the scan loss at edge of coverage if the polar orbit plane assuming operations can continue
parameter k [see Eq. (10.11)] is 1.4? down to an elevation angle of 0°?
(iv) If the requirement is to provide the same power b. What is the minimum number of orbit planes that
density at the edge of coverage as exists at nadir, what need to be used to provide continuous coverage over
additional amount of power (in dB) is needed at the the entire globe?
edge of coverage? c. (i) Repeat (a) above for a minimum elevation angle
of 20°.
(v) In what ways do you think you could cope with
this large difference in power requirements between (ii) Repeat (a) above for a minimum elevation angle
nadir and edge-of-coverage power? of 40°.
(iii) Repeat (b) above for a minimum elevation angle
8. Historically, nearly every commercial service
of 20°.
of whatever nature has been developed incremen-
tally. Most major enterprises that exist today (iv) Repeat (b) above for a minimum elevation angle
started as small operations in a single location. of 40°.
Global satellite systems do not have that luxury: d. What are the advantages and disadvantages to the
they have to start global. The designers of the LEO service provider if they adopt the constellation in part
satellite system in Problem 7 above want to exam- (b) above?
ine how they could provide global coverage with e. What are the advantages and disadvantages to the
different levels of service so that they can have a service provider if they adopt the constellation in part
measure of incremental growth options. They have (c) (iii) and part (c) (iv) above?
c11.qxd 10/08/02 10:30 Page 439
CHAPTER 11
DIRECT BROADCAST SATELLITE
TELEVISION AND RADIO
Geostationary satellites have carried television program material almost since their in-
ception for commercial service in the late 1960s. The limited bandwidth of undersea
cables designed for voice communications prevented their use for video signals, so live
television signals could not be transmitted beyond the limits of any continent at that
time. AT&T engineered microwave links in the 1950s that allowed video signals to be
distributed throughout the United States, and other countries quickly followed suit to es-
tablish national television networks. The first time that a GEO satellite was used exten-
sively for video transmission was for the Tokyo Olympic Games in 1968, which were
broadcast live in the United States using a link through an early Intelsat satellite over the
Pacific Ocean.
The growth of cable TV (CATV) systems in the United States in the 1970s en-
couraged the use of domestic North American satellites for distribution of cable TV signals.
One transponder on a GEO satellite can send a video signal to thousands of independent
cable systems distributed throughout the country, an example of point to multipoint trans-
mission. Satellites are a very effective way to distribute wideband signals, and there was
rapid growth in the use of C-band transponders for video signals, using FM and one
transponder for each video signal1. Ku-band satellites and compressed digital video sig-
nals followed, making it possible to transmit several video signals through one transponder,
with corresponding savings in transmission costs. A large fraction of all the transponders
in most of the world’s domestic and regional GEO satellite systems are devoted to the
distribution of video signals.
Figure 11.1 shows the earth station complex at Virginia Tech, in Blacksburg, Virginia,
which is equipped to uplink analog and digital video signals for distribution to educa-
tional institutions. The large antenna in the left of the photograph is a 9-m C-band
Cassegrain antenna used to transmit analog video FM-TV C-band signals to a domestic
satellite. The antenna is equipped with two 3-kW C-band transmitters connected to or-
thogonal polarization ports on the antenna feed, allowing simultaneous transmission of
two video signals to two orthogonally polarized transponders on the satellite. The antenna
was used for many years to distribute graduate classes to 16 locations in and around the
Commonwealth of Virginia.
The smaller antenna in the middle of Figure 11.1 is a 5.5-m Ku-band Cassegrain
uplink antenna used to transmit multiple digital compressed video signals. The University
moved to compressed digital transmission at Ku band when the leasing price of C-band
transponders suddenly increased following the failure of a large domestic C-band satel-
lite. The antenna at the right of Figure 11.1 is a Simulasat antenna. The reflector is a par-
abolic torus antenna aligned with the GEO arc and has seven feeds. Each feed illuminates
a section of the reflector which approximates a paraboloid. The antenna is used by the
University’s campus cable TV network to receive video signals from seven GEO satellites.
Several smaller receive-only antennas can be seen in the background.
439
c11.qxd 10/08/02 10:30 Page 440
440 CHAPTER 11 DIRECT BROADCAST SATELLITE TELEVISION AND RADIO
FIGURE 11.1 The Virginia Tech earth station complex in Blacksburg, Virginia.
Video distribution and direct broadcast television (DBS-TV) have become a major
source of revenue for the satellite communications industry. At the end of 1998, the to-
tal revenue earned by all satellite communications entities worldwide was estimated at
$30 B2. Of this total, $17 B was estimated to have been earned by video distribution and
direct broadcast television. The rapid growth of digital DBS-TV may increase the per-
centage of the industry’s revenues that come from video services even further during the
first decade of the 2000s. Many satellite communication systems that were designed for
voice and data transmission have ended up distributing TV signals. The revenues available
from entertainment television have become a major driver in the satellite communications
industry.
In 2001, two direct broadcast satellite radio services began operation in the United
States using S-band frequencies. The satellites provide a wide range of radio program-
ming, aimed primarily at drivers of road vehicles. Repeaters are used in city areas to
overcome the problem of satellite visibility around tall buildings.
11.1 C-BAND AND KU-BAND HOME

SATELLITE TV
In the early 1980s, the development of low noise GaAsFET amplifiers for C-band, and
improved threshold extension demodulators for video signal receivers, allowed much
smaller diameter antennas to be used to receive C-band FM video signals distributed
through GEO satellites. A market rapidly developed in the United States for home satel-
lite TV systems using 3-m and 3.6-m dish antennas (10-ft and 12-ft diameter) and set-top
c11.qxd 10/08/02 10:30 Page 441
11.2 DIGITAL DBS TV 441
receivers that could receive the video signals from domestic GEO satellites. At that
time, the signals were not scrambled, so owners of satellite dishes could receive a wide
range of television programming free of charge. The cable TV industry in the United
States became concerned about the growth of home satellite TV receiving systems, and
tried to have Congress pass laws which would ban their use. Congress did not pass such
laws, but instead told the industry to scramble (encrypt) their signals and to charge cus-
tomers for the descrambling information, and then passed laws that made the unautho-
rized use of descrambling equipment illegal. An estimated four or five million C-band
and Ku-band FM satellite TV systems were sold in the United States by the time that
Ku-band direct broadcast satellite television arrived in the 1990s, using digital trans-
mission and 0.5-m dishes, and offering more channels than the earlier system at a
comparable price3,4.
DBS-TV originally started in Europe and the United States in the 1980s using ana-
log FM transmission in Ku band. Satellite TV was much more successful in Europe than
in the United States in the 1980s, possibly because there were fewer alternative sources
of TV programming in Europe. Most European countries offered only a handful of broad-
cast TV channels, and cable service has never been as widespread in Europe as in the
United States. Nevertheless, at least one European satellite based direct broadcast TV
system failed during the 1980s, and two satellites built for a U.S. company intending to
enter the DBS-TV field were sold to a European company. The market for DBS-TV
systems grew slowly in the 1980s, and then very rapidly after the introduction of high
capacity digital DBS-TV satellites in the 1990s.
Primestar developed an analog (FM) DBS-TV service in the United States using
transponders on medium power (50–90 W) Ku-band GEO satellites located at 85° west
longitude, and a receiving terminal with a 1-m dish. Primestar offered up to 40 TV
channels by subscription for a fee of about $40 per month, and based its business plan
on leasing the DBS-TV receiving system to customers rather than requiring outright
purchase.
11.2 DIGITAL DBS TV
In the 1990s, digital video transmission became feasible, and several systems were de-
veloped in the United States in the 12.2- to 12.7-GHz band allocated to DBS-TV services.
The development of low cost Ku-band antennas and receivers, and high-speed digital in-
tegrated circuits specifically for DBS television that incorporate QPSK demodulation, er-
ror control, decryption, and MPEG decoding made DBS-TV practical. The digital signal
processing is incorporated in a single integrated circuit that implements the digital video
standard used by all the DBS-TV systems, DVB-S. The large volumes in which DBS-TV
receivers have been manufactured have allowed the cost of a receiving system to be re-
duced steadily since the start of DBS-TV service. Figure 11.2 shows the rapid growth in
subscribers to DBS-TV systems and the cost of a typical home DBS-TV installation in
the United States during the 1990s.
Directv, a fully digital DBS-TV system owned by Hughes Electronics Corporation,
was developed by a consortium of companies led by Hughes, and began limited service
in 1994 with a single GEO satellite at 101° W longitude. The first satellite, called DBS-1,
was launched in December 1993, and was followed by two more satellites, DBS-2 and
DBS-3, in 1994 and 1995. A fourth satellite was added in 1999, and a fifth satellite was
launched in 2000 with a transmit antenna capable of providing spot beams, using loca-
tions of 101° W and 109° W 3.
c11.qxd 10/08/02 10:30 Page 442
15 1200
Number of DBS-TV subscribers Purchase price Subscribers
Cost of DBS-TV system in U.S. $

in United States (millions)
10 800
5 400
0 0
1995 1997 1999 2001
Year
FIGURE 11.2 Growth in subscribers to DBS-TV systems and fall in typical home DBS-TV
installation cost in the United States in the 1990s. By 2001, suppliers were offering complete
DBS-TV systems free to anyone who would sign a one year contract.
The Echostar Communications Corporation started service with its Dishnetwork in

March 1996 with a single satellite at 119° W. In 2001 there were six Echostar satellites
in orbit, at longitudes 61.5° W, 119° W, and 110° W. By 2001, Dishnetwork had 5 million
customers4. Table 11.1 summarizes the major parameters of two of the DBS-TV satellites
serving U.S. customers in 2001.
Directv has grown its customer base very rapidly with over 10 million customers
by year-end 2000. Directv transmits 200 TV and audio channels that are available in a
mixture of subscription packages, much like cable TV companies offer, with pay per
view for individual movies and special events. In rural areas, DBS-TV offers hundreds
of television channels in place of the three or four terrestrial broadcasting stations that
are typically available. In city areas, DBS-TV offers an alternative to cable television at
a similar cost. In 1999, Directv acquired Primestar, which provided an additional orbital
slot at 85° W.
All of the U.S. DBS-TV satellites use digital video transmission, as do several of
the European satellites. The main European DBS-TV provider is SES (Société Européenne
de Satellites), based in Luxembourg, which had eight Astra DBS-TV satellites in orbit in
1999. Two further satellites were due to be launched in 2001. The Astra DBS-TV satel-
lites are some of the largest GEO satellites in orbit. Astra 2A, built by the Hughes Space
and Communications Co., had an on-orbit weight of 7335 kg when launched in 1998, with
a life expectancy of 15 years. Astra 2D is a Hughes (Boeing) 601 high-power satellite,
due for launch in 20016.
The 12.2- to 12.7-GHz band was set aside for exclusive use by DBS-TV satellites
in geostationary orbit so that high-power transponders could be used on specially designed
DBS-TV satellites. Typical transponder output levels are 100 to 240 W with flux density
at the earth’s surface up to 105 dBW/m2. The satellites can carry up to 32 transponders,
giving a total transmitted RF power up to 3.2 kW, higher than for any other commercial
satellite in 1999. DBS-TV satellites are typically large and heavy, generally use a three-axis
c11.qxd 10/08/02 10:30 Page 443
TABLE 11.1 DBS-1R and Echostar 6 Satellite Specifications
DBS-TV satellite DBS-1R Echostar 6
Location in geostationary orbit 101° W longitude 119° W longitude

Launch date July 14, 2000 July 13, 2000
Satellite manufacturer Hughes Space and Space Systems Loral
Communications Inc.
Type designation HS 601-HP SSL 1300
Frequency band Ku band (12.2–12.7 GHz) Ku band (12.2–12.7 GHz)
Transponders 16 active (4 spare) 32
Output power 200 W (dual 100 W TWTAs) 125 W (can be paired)
Solar power system Two wings, GAAlAs cells Two wings
Beginning-of-life output power 8.7 kW 11.27 kW
End-of-life output power 7.7 kW
Batteries 27-cell NiH 350-Ahr NiH
Propulsion Liquid apogee motor 110 lbf Bipropellant
Station keeping thrusters
N–S (xenon ion) 4 106 lbf (0.17 N)
E–W (bipropellant) 4 5 lbf (10 N)
N–S (bipropellant) 4 2 lbf (22 N)
Dimensions in orbit
Length over solar arrays 86 ft (26 m) 102.1 ft (31.1 m)
Width over antennas 23 ft (7.0 m) 28.4 ft (8.66m)
Dimensions, stowed
Height 13 ft 3 in (4 m)
Width 11 ft 9 in (2.7 3.6 m)
Mass at launch 3446 kg (7581 lb)
Mass in orbit (beginning of life) 2304 kg (5069 lb) 8157 lb (3700 kg)
Antennas
Transmit (2) 2.72 m (2) 2.39 m
Receive 1.32 m 1.19 m
stabilized design, and have large solar sails to generate the DC power required by the
transponders. Figure 11.3 shows Echostar 6, a large GEO three-axis stabilized DBS-TV
satellite built for Echostar by Space Systems Loral.
The flux density at the earth surface produced by medium- and high-power transpon-
ders used on DBS satellites is in the range 105 to 115 dBW/m2, which allows small
receiving antennas (dishes) to be used for DBS-TV reception, with diameters in the range
FIGURE 11.3 Echostar 5 DBS-TV satellite. (Photo courtesy of EchoStar Communications

Corporation.)
c11.qxd 10/08/02 10:30 Page 444
0.45–0.9 m. The small dish required for DBS-TV reception played a critical part in the
acceptance and success of DBS-TV in the United States. Previously, DBS-TV reception
of cable television signals was possible at C band and Ku band with 2.5- to 3.5-m dishes.
The local governments of many cities and towns refused to permit these large dishes in
residential areas, although they became popular in rural areas.
The U.S. Congress passed laws in 1997 that prevented local governments from re-
stricting the use of antennas less than 1 m in diameter, opening up a large market for Ku-
band DBS-TV services. European DBS systems use similar size dishes, but lower
transponder power on the satellite. The EIRP values are similar, but western Europe can
be covered by a much narrower antenna beam from a GEO satellite than the United States,
allowing higher gain antennas to be used on the European DBS satellites. The combination
of a higher gain satellite antenna and lower transponder output power produces similar
flux densities at the earth’s surface, and the receiving terminals can therefore use antennas
with similar diameters.
The small receiving antenna has a wide beam, typically 4° for a 0.45-m dish, which
forces wide spacing of DBS-TV satellites to avoid interference at the receiving antenna
by the signals from adjacent DBS-TV satellites. A 9° spacing in the GEO arc has been
adopted by the United States, which restricts the number of DBS-TV satellites that can
be placed in geostationary orbit to serve the United States. In the 1990s the U.S. FCC
successfully auctioned spectrum and orbital locations for DBS-TV satellites, raising hun-
dreds of millions of dollars from companies that saw a profitable commercial venture.
The first entrant into the high-power DBS-TV field, Directv, spent about $1 B to
develop their system and needed about two million customers to break even. That num-
ber was quickly passed and the Wall Street Journal described Directv as “one of the most
successful business ventures of the century.” Directv uses four satellites in two pairs spaced
half a degree apart at a nominal orbital location of 100° W. The early DBS-TV satellites
served the entire United States from this GEO location, using relatively broad beams. In
2000, new legislation allowed DBS satellite operators to compete with cable television
companies by supplying news from local TV stations to specific regions. This made spot
beams serving only a part of the United States very desirable, and later generations of
Directv DBS-TV satellites incorporate large transmit antennas that can generate spot beams
on centers of population in the United States.
Each DBS-TV satellite carries up to 32 high-power transponders covering part of
the 12.2- to 12.7-GHz broadcast satellite band (BSS), and the satellites at each orbit lo-
cation transmit in opposite hands of circular polarization (CP). Signals with opposite hands
of circular polarization are orthogonal, and a suitably designed earth station antenna can
separate two signals with opposite hands of circular polarization. In the simplest DBS-
TV receiving terminal, an electronically controlled polarizer is used immediately behind
the antenna feed. The polarizer can be set to receive LHCP (left-hand circular polariza-
tion) or RHCP (right-hand circular polarization) by changing the voltage supplied to the
low noise block converter (LNB) unit at the antenna. Typically, a supply voltage of 7 V
will cause the antenna to receive one polarization and reject the other. Increasing the
voltage above 14 V causes the antenna to switch polarizations. The polarizer converts the
circularly received polarization signal to a linearly polarized signal in a section of wave-
guide, and a linear probe in the waveguide converts the signals to currents that drive the
LNA input.
More complex receivers use an orthogonal mode transducer (OMT) with two LNBs
so that both hands of circular polarization can be received at the same time by using two
LNBs and two receivers. The dual-polarization receiving system is needed when more
than one channel must be received at a time. A dual-channel DBS-TV receiver with a
c11.qxd 10/08/02 10:30 Page 445
dual-polarized antenna allows two TV channels to be viewed by different people in the

same household at the same time. In 2000, Directv charged an additional $5.00 above the
basic monthly subscription fee for this grade of service, rather than requiring the payment
of two separate subscriptions.
Customers wanting to receive signals from more than one orbital location need an an-
tenna with two feeds. Reception from two satellites spaced 9° apart in GEO can be achieved
with a larger antenna, 0.45 0.6 m (18” 24”) that produces two beams separated by the
appropriate angle. The two feeds can be seen in the photograph of a Dishnetwork antenna
in Figure 11.4b.
DBS-TV receiving antennas are typically an offset parabolic reflector design with
the feed below the antenna aperture. The offset feed design eliminates blockage of the
aperture by the feed which occurs in symmetrical reflector antenna designs, and improves
the aperture efficiency of the antenna, and therefore increases its gain. Offset fed para-
bolic reflectors have a beam squint effect in the plane of symmetry when operated in op-
posite hands of circular polarizations. For the 0.45 m diameter antenna widely used for
DBS-TV reception in the United States, the LHCP and RHCP beams are squinted about
0.25° from the antenna’s boresight. The 3-dB beamwidth of the antenna is around 4°, so
the squint effect does not cause significant loss of gain.
(a) (b)
FIGURE 11.4 DBS-TV receiving antennas. (a) Directv antenna mounted on the wall of a
house (b) Dishnetwork antenna mounted on a post. Note the two feeds to allow reception
from two satellites at separate locations in the geostationary orbit.
c11.qxd 10/08/02 10:30 Page 446
12.2–12.7 GHz
DBS-TV signal
Image rejection 900–1400 MHz Low noise

LNA BPF Mixer IF amplifier block converter
mounted on
antenna feed
Coaxial cable
Local oscillator to set top
Ku-band
11.3 GHz receiver
antenna
900–1400 MHz 70 MHz QPSK Baseband

IF amplifier Mixer IF amplifier demod amplifier
Input
Select Tuned BPF Tuned LO
select Frequency
polarization Microprocessor
transponder synthesizer
Video
D/A
Audio
Inner De-interleaver Outer Digital MPEG 2 Analog output
decoder decoder demux decoder to TV set
FIGURE 11.5 Block diagram of a DBS-TV receiver.
A Directv receiving antenna mounted on the wall of a house is shown in Figure 11.4a,
and a Dishnetwork antenna mounted on a post is shown in Figure 11.4b.
Figure 11. 5 shows a block diagram of a DBS-TV receiver. The entire front end of
the receiver is located at the antenna feed in the form of an LNB to minimize loss of sig-
nal and hence to maintain the lowest possible system noise temperature. The electronic
polarizer is switched by changing the voltage supplied to the LNB via the cable that in-
terconnects the antenna and set-top receiver. The entire 12.2–12.7 GHz band is down-
converted by the LNB to the 900–1400 MHz band, where cable losses are much lower
than at Ku band. The down-converter consists of a dielectric resonator local oscillator and
mixer, followed by an IF amplifier and band-pass filter. The high gain LNB can drive
100 m of coaxial cable without any reduction in signal quality. Where longer cable runs
are needed, amplifiers for the 900–1400 MHz band can be used to boost the signal strength.
The set-top box accepts the entire 500-MHz band and separates out the individual transpon-
der frequencies. Any one of these frequencies (and the corresponding polarization) can be
selected on demand by the user.
The user enters a desired channel number into the set-top box using an IR remote
control, for example, channel 362, which is converted via a stored look-up table in the re-
ceiver to an RF channel frequency and polarization. The signal from the required transpon-
der is then selected by the receiver by setting the correct polarization at the antenna and
tuning the set-top local oscillator to the appropriate IF channel frequency. The QPSK sig-
nal is then demodulated. The result is a multiplexed bit stream, typically at a bit rate up
to 40 Mbps, which contains the bits for channel 362 and several other video signals. The
bit stream is encrypted and contains error control coding bits and data bits. The bit stream
is processed to correct and detect errors, de-interleaved, and decrypted. A digital demul-
tiplexer then extracts the bits for the wanted channel, 362 in this example, sends them to
c11.qxd 10/08/02 10:30 Page 447
11.3 DBS-TV SYSTEM DESIGN 447
a MPEG 2 decoder, and finally generates analog video and audio signals with D/A
converters to drive the TV set.
The look-up table in the receiver that relates channel numbers to frequencies,
polarizations, and instructions for the TDM demultiplexer is downloaded from the satel-
lite on a regular schedule. This allows the service provider to change the transponder which
carries a particular signal, and to alter the mix of signals on a given transponder as required,
without the customer being aware of the changes.
The satellite is also used to address individual receivers and to load another look-
up table that specifies which channels the user is authorized to receive. If the user fails to
pay his or her bills to the service provider, the receiver will eventually be instructed to
show only a message that it has been disconnected for failure to make timely payment for
the service. This process involves a smart card, which identifies each receiving system,
and enables decryption of the satellite signals. The high level of protection applied to the
DBS-TV signals is intended to prevent unauthorized reception by users who have not paid
monthly fees. Hackers have reportedly broken the encryption system of Directv from time
to time by reprogramming smart cards, but Directv retaliated in early 2001, disabling the
pirated cards and shutting down nonpaying viewers5.
Pay per view channels are handled differently from broadcast channels. A customer
wishing to buy a movie or a sporting event selects the desired channel and authorizes the
system to make a charge. The Directv receive terminals have no uplink capability, and
must therefore use terrestrial telephone circuits to send charging information to the central
office of the TV service provider. The cost of the pay per view event is then added to the
customer’s monthly bill. This requires a connection between the DBS-TV receiver and
the PSTN at the customer’s premises. The receiver dials a toll free number, and down-
loads a record of the charges for that customer, and any other information that the receiver
is programmed to deliver to the service provider. Such information might include the
pattern of channels that the customer selects and watches, which is valuable data for
advertisers.
Using the satellite to convey instructions to the receiver brings some notable ad-
vantages to the DBS-TV customer. A customer who wishes to change the level of service
he or she receives need make only a single phone call to obtain a service upgrade. The
customer owns the DBS-TV receiving equipment and is responsible for its maintenance,
so there are no service calls by satellite TV providers.
11.3 DBS-TV SYSTEM DESIGN
The DBS-TV system must provide a received signal power at the small receiving
antenna that provides an adequate CN margin in clear air. Heavy rain will cause at-
tenuation that exceeds the link margin, so occasional outages will be experienced, es-
pecially during the summer months when thunderstorms and heavy rain are more
frequent. The CN margins used in DBS-TV systems are small, to avoid the need for
a large receiving antenna.
The selection of a CN margin is a design trade-off between the outage level that
customers can be expected to tolerate, the maximum allowable diameter of the receiving
dish antenna, and the power output from the satellite transponders. Typical designs with
receiving antennas in the 0.45 to 0.9 m range and 100 to 250 W satellite transponders
yield rain attenuation margins of 3 to 8 dB and outage times totaling 5 to 40 h per year
depending on the receiver’s location. However, since most customers don’t watch TV for
24 h per day, they will not be aware of all the outages. Unfortunately, thunderstorms tend
c11.qxd 10/08/02 10:30 Page 448
G
E
C C G
60° 60°
B B C
Latitude
E F
−4 dB K −4 dB
D F
E H
−6 dB B
−4 dB −6 dB
M
30° 30°
E M
N −2 dB
E
N
P
150° 120° 90° 60° 30°

Longitude
FIGURE 11.6 Coverage pattern of DBS-TV satellite with rain region contours. The solid
lines show ITU-R rain zones for the United States. The dashed lines show the satellite an-
tenna beam contours in decibels relative to maximum gain. Note that the satellite antenna
has maximum gain in the southeast of the United States where heavy rain is most frequent.
to occur more often in the late afternoon and evening, resulting in more outages during
prime viewing time.
The Ku-band transmit beam from the satellite carrying the DBS-TV signals is shaped
to deliver more power to those areas that suffer the highest occurrence of heavy rain, such
as the states in the southeastern part of the United States. This creates a larger link mar-
gin in those areas and helps to keep outages to an acceptable level. Figure 11.6 shows
satellite antenna EIRP contours over the United States for the Conus beam of a typical
DBS-TV satellite located at 101° W, with U.S. rain zones superimposed. The transmit an-
tennas on the DBS-TV satellites have diameters up to 107 inches (2.71 m) which gives a
spot beam gain of 49.3 dB. Multiple spot beams are used to provide local TV program-
ming to selected cities and conurbations, while the Conus beam provides service through-
out the contiguous 48 states. The high gain of the spot beam allows the local program
services to be transmitted at a lower transponder output power level, and also permits fre-
quency reuse by spatial beam separation. The dual Gregorian reflector system of the trans-
mitting antenna on the satellite is fed by a complex feed structure that produces the Conus
beam contours shown in Figure 11.6. The 6 dB contour of the beam is approximately
5.5° degrees wide in the E–W direction and 2.5° wide in the N–S direction, correspon-
ding to a 3 dB contour that is 4.0° 1.8°. The estimated gain of the Conus beam is
36.5 dB. The coverage zone within the 6 dB contour of the Conus beam, taking account
of the earth’s curvature, is approximately 4000 km E–W and 2000 km N–S.
Florida, Alabama, and Louisiana, for example, are in rain zones M and N of the
United States that has a rainfall rate of 50 mm per hour for five times the number of hours
per year that this rain rate occurs in Washington, D.C., and much of the eastern portion
of the United States. At Ku band, a rain rate of 50 mm/h will cause about 6-dB attenua-
tion on a typical DBS-TV slant path, sufficient to ensure an outage of the DBS-TV sig-
nal. This rainfall rate is exceeded for about 5 h per year in Florida and 1 h per year in
Washington D.C. The central and western parts of the United States have rainfall rates of
50 mm per hour for much less than 1 h per year, and therefore do not need such large
link margins. See Chapter 8 for detailed maps of rainfall rates and the occurrence of heavy
c11.qxd 10/08/02 10:30 Page 449
11.4 DBS-TV LINK BUDGET 449
rain across the United States, and for the techniques used to convert rainfall statistics into
attenuation data that can be used to calculate outage times. Note that much of the DBS-
TV coverage of the United States shown in Figure 11.6 lies inside the 6 dB contour of
the satellite antenna beam, giving beam loss values between 4 and 6 dB, so the usual
edge of beam loss of 3 db cannot be applied here. The regions within the 4dB to 6
dB contours are those that do not have frequent heavy rainfall.
Some manufacturers of DBS-TV receiving systems offer larger dishes for customers
living in high rainfall zones. Increasing the antenna diameter from 18 inches to 24 inches,
for example, increases its gain by 2.5 dB. This increase in antenna gain adds directly to
the rain fade margin of the receiver, and lowers the outage time in heavy rain.
11.4 DBS-TV LINK BUDGET
In this discussion, rain attenuation statistics at Ku band will be used that are representa-
tive of many locations in the central and eastern parts of the United States, where typical
path attenuation in rain exceeds 3 dB for 0.2% (15 h) and 6 dB for 0.01% (52 min) of an
average year. The distribution of the fades is random, with some long fades in the heav-
iest of thunderstorms and numerous shorter fades in brief periods of heavy rain. Directv
claims that receiving systems using 0.45 m (18 inch) diameter antennas designed for their
DBS-TV transmissions have an availability exceeding 99.7%, which is an outage time of
0.3% of the year, a total of about 25 h in an average year. For much of the United States,
this corresponds to a rain attenuation in the slant path of 3 dB and requires a link margin
of 5.7 dB when allowance is made for the increase in antenna noise temperature that
accompanies 3 dB of rain attenuation.
A representative link budget for a GEO DBS-TV system serving the United States
is shown in Table 11.2. The path length of 38,500 km is the maximum expected path
length for a receiver in the United States and a satellite at longitude 101° W. The threshold
CN value is set at 8.6 dB, corresponding to a system using QPSK with an implementa-
tion margin of 0.8 dB, forward error correction coding that produces 6 dB of coding gain,
TABLE 11.2 Link Budget for Ku-Band DBS-TV Receiver
Downlink power budget

Transponder output power, 160 W 22.0 dBW
Antenna beam on-axis gain (Conus coverage) 36.5 dB
Path loss at 12.2 GHz, 38,500-km path 205.9 dB
Receiving antenna gain, on axis 33.5 dB
Beam contour loss 3.0 dB
Miscellaneous and gaseous attenuation losses 0.8 dB
Received power, C 117.7 dBW
Noise power budget
Boltzmann’s constant, k 228.6 dBW/K/Hz
System noise temperature, clear air, 143 K 21.6 dBK
Receiver noise bandwidth, 20 MHz 73.0 dBHz
Noise power, N 134.0 dBW
C/N in clear air 16.3 dB
Link margin over 8.6-dB threshold 7.7 dB
Link availability throughout U.S. Better than 99.7%
c11.qxd 10/08/02 10:30 Page 450
and a maximum BER of 106. This requires a clear air CN ratio in the DBS-TV receiver
of 8.6 5.7 14.3 dB in clear air. The link budget in Table 11.2 shows that a link mar-
gin of 7.7 dB is achieved for a receiver located on the 3 dB contour of the satellite an-
tenna beam. Earth stations close to the 6 dB contour of the satellite beam have a link
margin of 4.7 dB. A receiver located in the SE states of the United States, within the 2dB
contour of the satellite beam, has a link margin of 8.7 dB.
In the link budget shown in Table 11.2, the transponder output power is 160 W, cor-
responding to a transponder with a saturated output power of 200 W and 1 dB output
backoff. A small amount of output backoff is normally used to avoid excessive AM-PM
and PM-AM conversion in the transponder. The receiving antenna is a high efficiency de-
sign with an offset parabolic reflector 0.45 m in diameter and a circularly polarized feed.
The offset design ensures that the feed system does not block the aperture of the antenna,
which increases its efficiency. The gain of this antenna is 33.5 dB at 12.2 GHz assuming
an aperture efficiency of 66%.
The receiver in Table 11.2 is located at the 3dB contour of the transmitting an-
tenna. Miscellaneous losses of 0.4 dB for gaseous attenuation at 12 GHz and 0.4 dB for
receive antenna mispointing and other losses are included in the link power budget. The
result is a received carrier power of 117.7 dBW in clear air conditions.
The noise power budget of the link in Table 11.2 is based on a receiver noise band-
width of 20 MHz, an antenna noise temperature of 35 K in clear air, and a 12-GHz LNA
with a noise temperature of 110 K. The result is a noise power of 134.0 dBW in a noise
bandwidth of 20 MHz referred to the input of the LNA, and a clear air CN ratio of
14.3 dB. The noise bandwidth of a digital receiver is set by the bandpass filters in the fi-
nal IF stage of the receiver, immediately before the demodulator. The filter must be de-
signed to match the symbol rate of the transmitted signal, and has a root raised cosine
(RRC) transfer function. (See Chapter 5 for details of digital transmission techniques.)
The noise bandwidth of all RRC filters is always equal to the symbol rate of the digital
transmission. In the DBS-TV system described in Table 11.2, a QPSK signal with a sym-
bol rate of 20 Msps is assumed, which results in a receiver noise bandwidth of 20 MHz.
However, with MPEG-2 encoding of the video signals, the data rate is not constant. The
VBS-S digital video standard is designed to allow for variable bit rates, and the figure of
20 Msps is probably a maximum value. Complex coding schemes are used on DBS-TV
digital transmissions, making use of Reed–Solomon block encoding, interleaving, and an
inner layer of convolutional encoding, as discussed in Chapter 7 and illustrated in Fig-
ure 7.11. The error mitigation scheme allows 6 dB of coding gain to be achieved with a
code rate greater than half rate FEC, which allows more of the bits in the 40-Mbps bit
stream to be allocated to data and fewer to parity bits. Data rates of 23 to 27 Mbps are
reported for the data stream of some DBS-TV systems, with rate three-quarter inner con-
volutional coding. The overall code rate with 188204 outer Reed–Solomon coding and
rate three-quarter inner convolutional coding is 0.69, which provides a message data rate
of 27 Mbps with a coded bit rate of 39.1 Mbps.
11.5 ERROR CONTROL IN DIGITAL DBS-TV
Digital DBS-TV transmissions typically use a transmitted symbol rate of 20 Msps using
QPSK, which gives a bit rate of 40 Mbps. Error correction coding and control bits oc-
cupy 13 to 17 Mb of the bit stream, leaving 23 to 27 Mb for digital TV data. A 23-Mbps
data stream can carry three live compressed digital video signals using MPEG 2 encod-
ing, or up to 10 prerecorded and processed video signals. Prerecorded material, which
c11.qxd 10/08/02 10:30 Page 451
11.5 ERROR CONTROL IN DIGITAL DBS-TV 451
comprises the majority of programming on satellite TV channels, is heavily processed to

reduce its bit rate as far as 1.6 Mbps. When prerecorded material such as movies is dig-
itized and processed through MPEG 2 compression, the data rate can be reduced to
1.6 Mbps, but results in digital artifacts appearing in the picture, especially when there is
a lot of motion in the scene. A digital artifact appears as a freezing of the entire picture
for a fraction of a second, caused by overloading of the MPEG 2 processing, or as a block
or pixel of the wrong color. The artifacts can be removed one by one by a digital artist
who works on the recorded material to paint out the effects. The final result is a movie
or show that can be recorded in digital compressed form at an average rate of 1.6 Mbps
for replay over the satellite. Live program material with a lot of motion in the picture can
cause the bit rate of an MPEG 2 coded signal to increase above the average value of
6.2 Mbps. Mixing prerecorded and live material in a single transponder helps even out
the bursty nature of live material.
Error control in the digital video standard is achieved in a similar way to compact
disc recordings. The compressed digital video signal bit stream is first split into blocks of
bits and encoded with a Reed–Solomon linear block code. The coded bit stream is then
interleaved (see Chapter 7 for details of interleaving) and encoded again with a convolu-
tional code. The double layer of error control coding is called a concatenated code. At the
receiver, the recovered bit stream is first decoded with a Viterbi decoding algorithm to re-
move the convolutional coding, and a limited number of errors in the bit stream are cor-
rected. The corrected bit stream is then de-interleaved, and a Reed–Solomon decoding
algorithm is applied. Figure 11.7 shows a block diagram of the coding and decoding
operations performed on the bit stream.
The coding process used for digital video and audio bit streams relies for much of
its error correction capability on the fact that the end signal delivered to the user is ana-
log. Bit errors in a digital video or audio signal result in the wrong voltage occurring
when the errored word is converted to a voltage by the receiver’s digital-to-analog con-
verter (DAC). If the system knows that a particular word is in error (i.e., the coding
scheme detected an error but was unable to correct it), that word can be flagged and the
error can be removed by interpolation of the analog waveform. When a word is known
to contain a bit error, it is replaced by a new word which is calculated to have a value
midway between the two adjacent words in the bit stream, as illustrated in Figure 7.10
in Chapter 7. The result is an interpolated value in the analog voltage waveform output
by the DAC. The interleaving process ensures that most bit errors are single errors, in-
creasing the probability that words contain only a single error and that the error will be
detected.
Data
input (204, 188) Rate 3 /4
Reed–Solomon Interleaver Convolutional
encoder encoder
Link
Convolutional Reed–Solomon
De-interleaver
decoder decoder
Data
output
FIGURE 11.7 Block diagram of the coding and decoding operations

in a DBS-TV signal.
c11.qxd 10/08/02 10:30 Page 452
All error correcting codes are able to detect at least twice as many errors as they
can correct. The use of analog waveform interpolation allows the video (and CD) error
correction process to take advantage of the error detection capabilities of the outer layer
of Reed–Solomon (R-S) coding to correct errors in the analog waveform without using
error correction coding. If error correction coding were used instead of analog interpola-
tion, at least twice as many parity bits would be needed in the bit stream, which would
reduce the number of data bits drastically. The outer layer of Reed–Solomon linear block
coding has good error detection properties, and the decoding process can be implemented
in real time in a high-speed microprocessor. This is an important consideration in a real-
time system, like TV broadcasting. Decoding of the error control codes must be possible
in real time in a low cost processor, and this is one of the driving factors in the selection
of the concatenated code used for CDs and for the video broadcast standard. Bit errors
that are not corrected or detected by the error control coding process result in individual
words that are in error.
In an uncompressed digital television picture, a single word error would result in a
pixel on the TV screen that is the wrong color. In a digital compressed TV signal, each
word influences many pixels on the screen, and a word error may result in a block of the
wrong color. MPEG 2 achieves a compression ratio of 8 to 10 with live video, so a sin-
gle word error in an MPEG 2 encoded bit stream might be expected to cause errors in 8
or 10 pixels on the TV screen, or even more pixels if error propagation occurs. However,
all digital compression schemes are designed to minimize the effect of errors in incom-
ing words, so the impact of single word errors is usually confined to a small square on
the screen which has the wrong information. As the bit error rate of the recovered bit
stream in the receiver increases, the impact of the word errors becomes more severe, and
larger blocks of the TV picture are corrupted. The receiving system is able to recognize
the high error rate and will blank the screen until an acceptable error rate is restored. Thus
a rain fade on a DBS-TV link which goes below the receiver threshold is characterized
by the initial appearance of small squares of incorrect color on the TV screen, followed
by larger block errors, and then a blank screen. When the rain intensity eases as the storm
moves through the slant path, the signal will return above threshold and the picture will
reappear on the TV screen.
The user of a DBS-TV system is usually aware of a thunderstorm or very heavy
rain in the locality when the signal goes below threshold and the TV screen goes blank.
This seems to make loss of the TV picture more acceptable to users, and most DBS-TV
customers appear to be satisfied with a nominal availability of 99.7%. The actual avail-
ability is undoubtedly higher than 99.7% for most of the customers in the United States,
and few complaints seem to arise from the loss of signal in heavy rain.
11.6 MASTER CONTROL STATION

AND UPLINK
Direct broadcast television satellites are relay devices that provide a very large coverage
area serving millions of customers. The many signals that are broadcast by the satellites
are collected at a master control station and uplinked to the satellites by a group of large
antennas with fade margins sufficient to overcome any expected rain fade. The video
and audio signals that are uplinked to the DBS-TV satellites are available in prerecorded
form on video tape or disc, or are collected from other satellites or fiber-optic lines. This
is a large operation which requires substantial resources and a sizable labor force.
c11.qxd 10/08/02 10:30 Page 453
11.6 MASTER CONTROL STATION AND UPLINK 453
FIGURE 11.8 Dishnetwork uplink earth station. (Photo courtesy of EchoStar

Communications Corporation.)
Figure 11.8 shows the master control station of the Dishnetwork system operated
by Echostar Communications Corporation; the location is in U.S. rain zone B2, provid-
ing a low probability of heavy rain. The statistics of region B2 show that a rain rate of
50 mm/h is exceeded for only 5 min in a typical year. Directv’s uplink earth station is
located in Colorado, within the U.S. B2 rain zone. The major European uplink station
for DBS-TV, operated by SES, is located in Luxemborg, which is also in the European
rain zone B2.
The uplink station must transmit hundreds of signals to the DBS-TV satellites 24 h
a day, 365 days a year. Most of the signals are prerecorded, either from satellite feeds
which are used to distribute new video and audio program material, or from archived ma-
terial. The uplink stations have hundreds of tape and video disc players, all under computer
control, which supply the video and audio signals for each channel. The signals are mainly
stored in digital form, allowing direct multiplexing into bit streams for the individual
transponders. Analog signals must be digitized and compressed before being multiplexed
with other signals into the bit streams that are sent to each transponder. More details of
the uplink centers operated by DBS-TV companies can be found from their web sites3–5.
A simplified block diagram of the transmitting equipment at an uplink station is
shown in Figure 11.9. One uplink antenna will typically transmit up to 16 RF channels
to one DBS-TV satellite. Each RF signal is a QPSK modulated Ku-band carrier with a
symbol rate of up to 20 Msps, occupying a bandwidth up to 27 MHz. The encoded, com-
pressed, and multiplexed bit stream drives a video exciter which generates QPSK modu-
lation of an intermediate frequency carrier, typically at 70 MHz. The 70 MHz signal is
upconverted to the transponder input frequency in the transmitter, which contains a trav-
eling wave tube high-power amplifier (HPA). The HPA is usually rated at a much higher
power than its normal operating output power level, which provides sufficient output
c11.qxd 10/08/02 10:30 Page 454
QPSK IF amplifier Upconverter LPA HPA Multiplexer

modulator
M MUX
17 GHz
Other RF uplink
70 MHz LO RF LO
signals antenna
MPEG 2 Digital Reed–Solomon Convolutional

encoder multiplexer encoder Interleaver encoder
MPEG
ADC MUX Coder I Coder
2
Analog video
and audio Other digital
signals signals
FIGURE 11.9 Simplified block diagram of a DBS-TV uplink earth station.
backoff of the HPA to ensure linear operation. The signals from any number of HPAs are
multiplexed together in microwave combiners and sent to the antenna feed for transmission
to the satellite.
11.7 INSTALLATION OF DBS-TV ANTENNAS
Installation of a home satellite TV system offers an interesting challenge to home own-

ers who do not have much knowledge of microwave antennas and satellite communica-
tion systems. A DBS-TV system antenna with a diameter of 0.45 m (18 inches) has a
beamwidth of 4°, and needs to be pointed to an accuracy of 0.5° for optimum recep-
tion of the satellite signal. The problem is to provide a simple method for pointing the
antenna in azimuth and elevation within about 2° so that a signal can be received and
peaked. Directv and Dishnetwork offer an installation kit and book of instructions that
make the process quite easy.
The antenna is mounted onto a 2-inch diameter tube with a swiveling clamp. The
lower end of the tube has a mounting bracket that is bolted to any convenient surface that
provides a clear view of the southern sky, and the tube is set vertical using a plumb line
or level. The antenna can then be rotated in azimuth. Elevation angle is set by rotating the
dish about a horizontal axis provided by a bolt that is part of the swiveling clamp, and
using a simple angle scale marked on the mounting. When the dish is set to the correct
azimuth and elevation angles the bolts in the clamp are tightened down and the antenna
is permanently set to the correct look angles.
Directv provide an azimuth and elevation look angle calculator in the set-up menu
of their receiver, and also a signal strength meter with both numerical and audio outputs.
The azimuth and elevation look angles can also be found from tables and maps in the
DBS-TV receiving system installation guides, and by using software that can be down-
loaded from web sites. The Directv on-screen calculator provides elevation and azimuth
angles for the Directv satellites at longitude 101° W based on the zip code of the user, or
the latitude and longitude of the earth station. The azimuth angle is given relative to mag-
netic north, and the installation kit includes a small compass that allows the user to set
an approximate azimuth angle. The elevation angle can be set within 1° by careful adjustment
c11.qxd 10/08/02 10:30 Page 455
11.8 SATELLITE RADIO BROADCASTING 455
of the clamp, which ensures that the satellite will be within the elevation plane 3 dB
beamwidth of the antenna when the azimuth angle is correct.
The procedure used to find the satellite is quite simple. The antenna is rotated in
azimuth until a tone is heard from the TV receiver, indicating that a signal is being
received. The antenna azimuth angle is adjusted to maximize the loudness of the tone
and the signal strength value (a number between 0 and 100) shown on the screen, which
ensures that the satellite is within the azimuth 3 dB beamwidth of the antenna. (The pro-
cedure typically requires two people because the TV set is rarely close to the antenna
installation point.) Once the satellite signal has been acquired, the azimuth and elevation
angles are adjusted alternately to maximize the signal strength, and the clamp is tight-
ened down to hold the antenna at the correct angles. When the above procedure is fol-
lowed with care, the antenna can be set to the correct azimuth and elevation angles in a
few minutes.
11.8 SATELLITE RADIO BROADCASTING
In the United States in 2001, two companies commenced transmission of digital radio sig-
nals from satellites, each offering 50 radio channels for a monthly subscription of about
$10–13. Generically, the system is called Satellite Digital Audio Radio Service (SDARS)8.
The target audience is in automobiles and other road vehicles, which is where most radio
listening occurs in the United States. A vehicle equipped with an SDARS receiver can re-
ceive the same program anywhere in North America, a selling point that the companies
hope will make their systems financially successful. Although subscription television, both
satellite and cable, has been very successful in the United States, SDARS is the first at-
tempt to create a subscription radio service—in contrast to terrestrial radio broadcasting
which has always been free to the listener, supported by advertising revenue. SDARS
vehicle radios have initial pricing in the $300–$500 range.
The SDARS satellites have high-power transponders to compensate for the low
gain omnidirectional antenna on the vehicle, and both systems use terrestrial repeaters
in large cities to augment the satellite signal when blockage occurs by tall buildings.
XM Satellite Radio Inc.9 based in Washington D.C. uses two satellites in GEO at 85° W
and 115° W longitudes, appropriately named “Rock” and “Roll.” Each satellite transmits
in a separate 3.7 MHz wide band in the frequency ranges 2332.5–2336.5 MHz and
2341–2345 MHz. Sirius Satellite Radio Inc.10 based in New York city, has three satel-
lites equally spaced in a 24 h polar elliptical orbit centered at a longitude of 100° W with
its apogee over North America. The satellites are above the horizon for listeners in the
United States for approximately 16 h in each 24 h, with two of the three satellites trans-
mitting in separate 4.2 MHz wide bands in the frequency ranges 2320–2324 MHz and
2328.5–2332.5 MHz. The highly elliptical orbit of the Sirius satellites can provide a
higher elevation angle than a GEO satellite, which is desirable in cities to minimize
blockage by tall buildings, but requires a handoff between satellites. Terrestrial repeaters
operate in the same frequency bands, in the gaps between the two satellite downlink fre-
quencies. Table 11.3 provides some details of the SDARS systems. XM repeaters receive
their signals directly from the SDARS satellites; Sirius repeaters are fed by relay via a
Ku-band GEO satellite.
Because of the high probability of the satellite signals being blocked by buildings in
a city and trees in rural areas, both systems utilize time diversity to overcome short inter-
ruptions in signal. The transmissions from the two satellites, and from the terrestrial re-
peaters, are delayed by varying amounts up to 5 s. The satellite radio receiver delays the
c11.qxd 10/08/02 10:30 Page 456
TABLE 11.3 U.S. Satellite Digital Audio Radio Service
Parameter XM Satellite Radio Inc. Sirius Satellite Radio Inc.
Number of satellites Two in GEO at Three in highly elliptical

85° W and 115° W 24-h orbit at 100° W
Downlink frequencies 2332.5–2336.5 MHz 2320.0–2324.0 MHz
2341.0–2345.0 MHz 2328.5–2332.5 MHz
Uplink frequencies 7050–7075 MHz 7060–7072.5 MHz
Terrestrial repeaters 1500 in 70 cities 105 in 46 cities
Total number of audio 100 100
channels
Transmission rate 4.0 Mbps 4.4 Mbps
before FEC
Satellite downlink modulation TDM-QPSK TDM-QPSK
signals to achieve a common timing and then selects or combines the signals to achieve
the best SN ratio. Signal transmission formats from the satellites are very similar to those
used in DBS-TV: TDM-QPSK modulation is used to send multiple signals as a high-speed
digital data stream, and concatenated Reed–Solomon outer layer and half rate convolutional
inner layer coding is used for error control.
11.9 SUMMARY
Satellite broadcasting of television has become a ma- interleaving. The recovery of a good quality analog
jor part of the satellite communications industry. In signal from a bit stream with bit errors relies on de-
1999, DBS-TV and video distribution earned more interleaving to spread out bursty errors, correction of
than half the revenues of the satellite communications some errors by the inner layer of convolutional cod-
industry, worldwide. Most DBS-TV and distribution ing, and detection of remaining, uncorrected, errors
of video signals is now digital, and Directv and by the outer layer of Reed–Solomon linear block cod-
Echostar in the United States have been major success ing. Detected errors that remain after the decoding of
stories with a total of 15 million customers by the end the digital signal typically result in single word errors.
of 2000. DBS-TV systems operate with small anten- Delivery of a bit stream through a direct broadcast
nas and low cost receiving systems, and offer a very satellite can be adapted to serve Internet users who re-
large number of video and audio channels, making quire the download of large blocks of data. The de-
them attractive to customers. The link budget for a typ- velopment of DBS-TV satellites with spot beams for
ical DBS-TV signal shows that the link margin is in local news broadcasts has provided the higher satel-
the 4 to 8 dB range, which yields a better than 99.7% lite antenna gain needed for uplinks from 0.5-m dishes.
availability in the United States. Shaping of the trans- Both Directv and Dishnetwork offer Internet access
mitted beam from the satellite provides higher clear terminals with uplink capability, although the service
air CN ratios in regions where heavy rainfall occurs is not available outside the spot beam coverage areas,
most often, such as the southeast of the United States. where the connection to the ISP must be made through
Digital DBS-TV signals are transmitted as a a terrestrial telephone link and the PSTN.
20-Msps QPSK signal occupying about 27 MHz of Satellite radio broadcasting commenced in 2001
transponder bandwidth. The 40-Mbps signal has a from three Sirius satellites in elliptical orbits and two
data rate between 23 and 27 Mbps with the remain- XM satellites in GEO. The signals are transmitted in
ing bits used for error control and system operation. S band at 2.3 GHz and are aimed primarily at auto-
DBS-TV digital signals make extensive use of error mobiles, which is where most people listen to the ra-
correction and error detection techniques in the form dio. Repeaters are used in city areas to overcome
of a double layer of error control coding with signal blockage by tall buildings.
c11.qxd 10/08/02 10:30 Page 457
REFERENCES 457
REFERENCES
1. LEON W. COUCH, Modern Communications Systems: 3. www.directv.com
Principles and Applications, Prentice Hall, Englewood 4. www.dishnetwork.com
Cliffs, NJ, 1995. 5. www.slashdot.com
2. WTEC Panel Report on Global Satellite Communica- 6. www.ses.com.lux
tions Technology and Systems, International Technol- 7. www.ssloral.com
ogy Research Institute, Baltimore, MD, December 8. DAVID H. LAYER: “Digital Radio Takes to the Road,”
1998, ISBN 1-883712-51-3. Available from NTIS as IEEE Spectrum, Vol. 38, No. 7, pp. 40–46. July 2001.
NTIS report PB99-117954. Available online at 9. www.xmradio.com
http://itri.loyola.edu 10. www.siriusradio.com
c12.qxd 10/08/02 10:31 Page 458
CHAPTER 12
SATELLITE NAVIGATION AND THE
GLOBAL POSITIONING SYSTEM
12.1 INTRODUCTION
The Global Positioning Satellite System (GPS) has revolutionized navigation and position
location. It is now the primary means of navigation for most ships and aircraft and is
widely used in surveying and many other applications. The GPS system, originally called
NAVSTAR, was developed as a military navigation system for guiding missiles, ships,
and aircraft to their targets. GPS satellites transmit L-band signals that are modulated by
several codes. The CA (coarse acquisition) code was made available to the public in the
mid-1980s. The secure high accuracy P code allows authorized users (mainly military) to
achieve positioning accuracy of 3 m. This was the accuracy that the military users wanted
for targeting smart bombs and cruise missiles, but such accuracies are also useful for auto-
landing aircraft in fog and for docking ships in bad weather.
The first commercial use of GPS was in surveying, but by 1990 several companies
had produced low-cost, handheld GPS receivers for general position location and naviga-
tion. Increased sales and larger volume production quickly brought down the price of a
GPS receiver, and the market expanded rapidly. GPS receivers are now a consumer product,
and will soon be found in every car and cellular telephone.
The GPS system has been successful because it provides a direct readout of the pres-
ent position of a GPS receiver with a typical accuracy of 30 m. There are other position
location systems, such as LORAN, (a contraction of long range navigation) that can also
provide direct readout of position, but not with the accuracy and reliability of GPS. The
success of GPS is an excellent example of what satellites do best: broadcasting. An
unlimited number of GPS receivers can operate simultaneously because all that a GPS
receiver has to do to locate itself is to receive signals from four GPS satellites.
The GPS space segment consists of 24 satellites in medium earth orbit (MEO) at a
nominal altitude of 20,200 km with an orbital inclination of 55º. The satellites are clus-
tered in groups of four, called constellations, with each constellation separated by 60º in
longitude. The orbital period is approximately one-half a sidereal day (11 h 58 min) so
the same satellites appear in the same position in the sky twice each day. The satellites
carry station-keeping fuel and are maintained in the required orbits by occasional station-
keeping maneuvers, just like GEO satellites. The orbits of the 24 GPS satellites ensure
that at any time, anywhere in the world, a GPS receiver can pick up signals from at least
four satellites. Up to 10 satellites may be visible at some times, and more than four satel-
lites are visible nearly all of the time. Replacement satellites are launched as needed, so
there may be more than 24 operational GPS satellites at any given time.
Figure 12.1 shows a GPS satellite. The satellites weigh 1877 kg at launch and have
a design lifetime of 10 years. In 2000, there were 30 GPS satellites in orbit, some of which
were spares. Because GPS is an integral part of the defense of the United States, spare
GPS satellites are kept in orbit and more spares are ready for immediate launch. The GPS
458
c12.qxd 10/08/02 10:31 Page 459
FIGURE 12.1 GPS block II-F satellite.
system is operated by the U.S. Air Force from the GPS master control station (MCS) at
Falcon Air Force Base in Colorado Springs, Colorado. The MCS and a series of subsidiary
control stations around the globe continuously monitor all GPS satellites as they come
into view and determine the orbit of each satellite. The MCS and other stations calculate
ephemeris data for each satellite, atomic clock error, and numerous other parameters
needed for the navigation message. The data are then transmitted to the satellite using a
secure S-band link and used to update onboard data. There are five GPS monitor stations
located in Hawaii, Colorado Springs, Ascension Island in the Atlantic Ocean, Diego Garcia
in the Indian Ocean and Kwajalein in the Pacific Ocean1. The monitor stations have pre-
cise cesium time standards and make continuous measurements of range to all visible
satellites. These measurements are performed every 1.5 s, and used to provide updates for
the navigation messages.
The position of a GPS receiver is found by trilateration, which is one of the sim-
plest and most accurate methods of locating an unknown position. In trilateration, the dis-
tance of the unknown point from three known points is measured. The intersection of the
arcs corresponding to three distances defines the unknown point relative to the known
points, since three measurements can be used to solve three equations to give the latitude,
longitude, and elevation of the receiver. The distance between a transmitter and a receiver
can be found by measuring the time it takes for a pulse of RF energy to travel between
c12.qxd 10/08/02 10:31 Page 460
460 CHAPTER 12 SATELLITE NAVIGATION AND THE GLOBAL POSITIONING SYSTEM
the two. The distance is calculated using the velocity of electromagnetic waves in free
space, which is assumed to be equal the velocity of light, c, with c 299,792,458 m/s.
Time can be measured electronically more accurately than any other parameter by the use
of atomic clocks, and this is how the GPS position location system can achieve a meas-
urement accuracy of 1 m in a distance of 20,000 km. To achieve a position location accuracy
of 1 m, timing measurements must have an accuracy better than 3 ns. This is possible
with modern digital circuitry and a great deal of averaging.
Each satellite carries several high accuracy atomic clocks and radiates a sequence
of bits that starts at a precisely known time. A GPS receiver contains a clock that is syn-
chronized in turn to the clock on each satellite that it is receiving. The receiver measures
the time delay of the arrival of the bit sequence, which is proportional to the distance be-
tween the satellite and the GPS receiver. When the distance of a GPS receiver from three
satellites has been measured, the remaining piece of information that is required is the
position of each satellite. This is calculated in the GPS receiver using the ephemeris for
the satellite orbits that are broadcast by each satellite in its navigation message. Since the
time at which the transmitted bit sequence started is known at the receiver, the position
of the satellite at that time can be calculated from its orbital data. Making the calculation
for four satellites provides the receiver with sufficient information to determine its posi-
tion with very good accuracy. Four satellites, rather than three, are needed because the
clock in the receiver is not inherently accurate enough. The fourth distance measurement
provides information from which clock errors in the receiver can be corrected and the
receiver clock synchronized to GPS time with an accuracy better than 100 ns.
GPS satellites transmit two signals at different frequencies, known as L1 and L2.
The L2 signal is modulated with a 10.23 Mbps pseudorandom (PN) bit sequence called
the P code that is used by military positioning systems. The P code is transmitted in an
encrypted form known as the Y code, which restricts the use of the P code to authorized
users.
The L1 frequency carrier is modulated by a 1.023 Mbps PN sequence called the
CA code that is available for public use, and also carries the P code as a quadrature mod-
ulation. The higher bit rate of the P code provides better measurement accuracy than the
1.023 Mbps CA code. CA stands for coarse acquisition and P stands for precise. GPS
systems using the secure Y code require the CA code as an intermediate step in making
distance measurements with high accuracy. The accuracy of CA code receivers was
deliberately degraded some of the time by a process called selective availability (SA).
SA causes variations in the CA code satellite transmissions that result in less accurate
calculation of position. SA was discontinued in May 2000, but can be reinstituted if the
President of the United States declares a National Emergency.
The GPS system provides two categories of service. The precise positioning serv-
ice (PPS) receivers track both P code and CA code on L1 and L2 frequencies. The PPS
is used mainly by military users, since the P code is encrypted into the Y code before
transmission and requires decryption equipment in the receiver. Standard positioning serv-
ice (SPS) receivers track the CA code on L1. This is the service that is used by the gen-
eral public. The P(Y) and CA codes transmitted by each satellite create direct sequence
spread spectrum signals which occupy the same frequency bands. Both the CA codes
and the P codes are publicly available, but the P code cannot be recovered in a GPS re-
ceiver without a knowledge of the Y code decryption algorithm. In this discussion we will
concentrate on the CA code and its use in position location.
The former USSR built and operated a global navigation system that is very simi-
lar to GPS, known in the West as Glonass for global navigation satellite system. Almost
everything about Glonass is similar to GPS except the multiple access technique. Glonass
c12.qxd 10/08/02 10:31 Page 461
12.2 RADIO AND SATELLITE NAVIGATION 461
uses FDMA, with a different transmit frequency at each satellite. The equivalents of the
P code and CA code can be transmitted by Glonass satellites in RF bandwidths of 20
kHz and 2 kHz, so 100 satellites can be accommodated in a bandwidth of 2 MHz. An
FDMA receiver with 100 channels is simpler than a CDMA receiver. A frequency syn-
thesizer that can be tuned to the unique frequency of each satellite is required, rather than
the digital correlators that recover the GPS signals in a CDMA receiver.
The European Union is considering building a similar satellite navigation system
called Galileo, scheduled for operation by 2008, to provide precise navigation signals
without dependence on the United States.
12.2 RADIO AND SATELLITE NAVIGATION
Prior to the development of radio, navigation was by compass and landmarks on land, and
by the sun and stars at sea. Neither technique provides high accuracy, and shipwrecks
caused by inaccurate navigation and foggy weather were a common occurrence. On land,
people often got lost in wilderness areas (and still do). Pilots of light aircraft, relying
solely on a map and landmarks, would get lost and run out of fuel before they found some-
where to land. With a GPS receiver and a map, it is impossible to get lost. GPS receivers
are very popular with airplane pilots, owners of sea-going boats, and wilderness hikers.
The development of aircraft that could fly above the clouds, and particularly the
building of large numbers of bomber aircraft in the 1930s, made radio navigation essen-
tial. Military thinking after WWI, and during WWII, placed high reliance on the ability
of bomber aircraft to win a war by destroying the weapon manufacturing capability of the
enemy. During WWII, the allies sent 1000 bomber aircraft at a time to targets in Germany,
causing immense destruction to many cities. The philosophy of mass destruction contin-
ued after WWII with the development of nuclear bombs, intercontinental ballistic missiles
(ICBMs), and cruise missiles. However, bomber aircraft, ICBMs, and cruise missiles must
find their targets, so accurate navigation is an essential part of each of these weapon sys-
tems. This demand for accurate targeting of airborne weapons led to the development of
GPS. However, the majority of GPS users are now civilian, and the worldwide market for
GPS equipment is projected to be worth $25 B by 2005.
Commercial aircraft fly on federal airways using VOR (VHF omni range) beacons.
The airways are 8 miles wide to allow for the angular accuracy of VOR measurements,
which is better than 4°. GPS will eventually replace VOR navigation, allowing aircraft to
fly directly from point of origin to destination, but the system of VOR beacons in the
United States is likely to remain for many years as a backup to GPS.
GPS can provide a single navigation system with better accuracy and reliability than
all earlier radio navigation aids. It can provide navigation of aircraft directly between air-
ports, instead of indirectly via airways, while providing absolute position readout of lat-
itude and longitude. Differential GPS can be used instead of ILS to provide the required
straight line in the sky for an instrument approach to a runway, and can be linked to an
autopilot to provide automatic landing of aircraft in zero visibility conditions. Ships can
safely navigate and dock in treacherous waters in bad weather by using differential GPS.
Eventually, GPS will replace all other means of navigation, although some may be re-
tained as backup systems in case of failure of the GPS receiver(s) or jamming of the
signals.
GPS was preceded by an earlier satellite navigation system called Transit, built for
the U.S. Navy for ship navigation, which achieved much lower accuracy and became
obsolete when GPS was introduced. Transit satellites were in low earth orbits and the
c12.qxd 10/08/02 10:31 Page 462
SIDEBAR
The earliest radio navigation systems, developed in three radio transmitters that transmit at the same fre-
the 1930s and 1940s, were simple transmitting sta- quency, or very close frequencies. In the earlier forms,
tions operating in the mf (AM) radio band. A spe- the phase of the radio wave was used, but later sys-
cial antenna and receiver can make an indicator tems like LORAN time the arrivals of pulse trans-
needle point at the transmitter, so that an aircraft or missions. The receiver compares the phase or time of
a ship can home in on the beacon. The beacons used arrival of the radio signals from two transmitters and
by aircraft are called non-directional beacons calculates the difference in distance between the two
(NDB) and the receiver is called an automatic di- transmitters. A line with a constant difference in the
rection finder (ADF). Most commercial aircraft distance between two points is a hyperbola with the
carry ADF receivers, which are still required for in- two transmitters at the foci. A third transmitter pro-
strument approach procedures at some airports. vides two more hyperbolas, and their intersection
NDBs and an ADF receiver were the main means of locates the receiver, hence the name hyperbolic
radio navigation for aircraft in the 1920s and 1930s, navigator.
but have several serious disadvantages. If a strong LORAN uses pulse transmissions in the 100–500
wind is blowing across the path that the aircraft is kHz RF band, and can provide reliable navigation
taking to the NDB, it may fly a curved path instead with accuracy of a fraction of a mile at ranges of hun-
of a straight line. dreds of miles from the transmitters. The U.S. Coast
The NDB was largely superseded by the VOR bea- Guard built LORAN stations along the coastline of
con in the 1940s. VOR stands for VHF omni range. the United States to provide navigation assistance for
A VOR transmitter generates a rotating VHF radio ships in coastal waters where the danger of running
beam and also radiates a continuous sine wave signal aground is greatest. LORAN is steadily being
that is phase referenced to the time that the rotating replaced by GPS, and the Coast Guard now has dif-
beam sweeps through north. A VOR receiver on the ferential GPS systems in place to help ships navigate
aircraft synchronizes to the reference signal and in estuaries and rivers.
measures the angle of the beam relative to north at Instrument landing systems (ILS) are essential
the time it is received. With two VOR transmitters and when aircraft must land in conditions of poor visibil-
a map showing their locations, the aircraft can deter- ity. An ILS installation at an airport provides two
mine its position. radio beams that allow the aircraft to fly an approach
Many VOR stations have DME (distance measur- along a straight line to the runway threshold. The
ing equipment). The aircraft DME equipment trans- localizer, a VHF transmitter and antenna at the end
mits a pair of pulses and measures the time for the of the runway, provides two modulated beams in
round-trip to the VOR and back, which provides a the horizontal plane. A vertical needle on a course
measurement of range to the VOR. With knowledge deviation indicator (CDI) in the aircraft cockpit
of range and angle to a VOR, navigation is possible shows the aircraft’s lateral position relative to a line
using a single VOR-DME station. leading to the runway threshold. A glide slope trans-
WWII aircraft needed to navigate to targets over mitter at the side of the runway transmits another
enemy territory where there were no VORs available. radio beam which points upward at about 3°. A hor-
Hyperbolic navigators were developed in Germany, izontal needle on the CDI shows the position of the
Britain, and the United States during WWII to pro- aircraft relative to the glide slope, with a sensitivity
vide radio navigation at longer ranges than can be of 10 ft as the aircraft approaches the ground. The
achieved in the VHF band by using frequencies pilot of the aircraft tries to keep both the CDI nee-
between 100 kHz and 2 MHz. The frequencies can dles centered, so that the aircraft flies a straight line
propagate round the earth’s curvature making long- down the glide slope and arrives at a height of 50 feet
range navigation possible. A hyperbolic navigator has above the runway threshold.
system used the Doppler shift observed at the receiver when a beacon signal was trans-
mitted by the satellite. Because of the high velocity of LEO satellites—about 7.5 km/s—
their signals are significantly shifted up in frequency when the satellite appears over the
horizon with a component of velocity toward the receiver. The Doppler shift falls to zero
as the satellite passes the observer, and then becomes negative as the satellite flies away.
c12.qxd 10/08/02 10:31 Page 463
12.3 GPS POSITION LOCATION PRINCIPLES 463
Observation of the Doppler shift with time, which may need to be as long as 10 min, and
a knowledge of the satellite orbit, allows calculation of the receiver’s position.
There was never a sufficient number of Transit satellites to provide continuous
position data, and the long time required to obtain an accurate position fix was a disad-
vantage. A similar system called SARSAT, for search and rescue satellite, is used to find
emergency locator transmitters (ELTs) on aircraft that have crashed. Most general avia-
tion aircraft carry an ELT, which turns on at a frequency of 121.5 MHz when subjected
to high G forces, as might be experienced if the aircraft crashes. Certain LEO satellites
carry 121.5-MHz receivers that relay the signals to earth stations at rescue coordination
centers. If an aircraft ELT turns on, a SARSAT satellite will eventually fly by and relay
a Doppler shifted signal to the rescue station.
Analysis of the Doppler shift over the observation period provides information about
the location of the ELT, but with an accuracy of only 1 or 2 km. Almost 97% of ELT lo-
cations turn out to be false alarms—the ELT was dropped or accidentally turned on. It
seems probable that GPS and cellular phones or satellite phones will eventually replace
the SARSAT system.
12.3 GPS POSITION LOCATION PRINCIPLES
The basic requirement of a satellite navigation system like GPS is that there must be four
satellites transmitting suitably coded signals from known positions. Three satellites are
required to provide the three distance measurements, and the fourth to remove receiver
clock error. Figure 12.2 shows the general arrangement of position location with GPS.
The three satellites provide distance information when the GPS receiver makes three
measurements of range, Ri, from the receiver to three known points. Each distance Ri
can be thought of as the radius of a sphere with a GPS satellite at its center. The receiver
lies at the intersection of three such spheres, with a satellite at the center of each sphere.
Locally, at the receiver, the spheres will appear to be planes since the radii of the spheres
are very large. A basic principle of geometry is that the intersection of three planes com-
pletely defines a point. Thus three satellites, through measurement of their distances to
the receiver, define the receiver location close to the earth’s surface. There is another
point in outer space where the three spheres intersect, but it is easily eliminated in the
calculation process.
S2
S3
S1
S4
FIGURE 12.2 General

arrangement of position
locations with GPS. The aircraft
must receive signals from four
GPS satellites to be able to
determine its position.
c12.qxd 10/08/02 10:31 Page 464
Although the principles by which GPS locates a receiver are very simple, requiring
only the accurate measurement of three ranges to three satellites, implementing the
measurement with the required accuracy is quite complex. We will look first at the way
in which range is measured in a GPS receiver and then consider how to make the meas-
urements. Range is calculated from the time delay incurred by the satellite signal in trav-
eling from the satellite to the GPS receiver, using the known velocity of EM waves in free
space. To measure the time delay, we must know the precise instant at which the signal
was transmitted, and we must have a clock in the receiver that is synchronized to the clock
on the satellite.
GPS satellites each carry four atomic clocks which are calibrated against time
standards in the GPS control stations around the world. The result is GPS time, a time
standard that is available in every GPS satellite. The accuracy of an atomic clock is typ-
ically 1 part in 1011. However, it is too expensive to include an atomic clock in most
GPS receivers, so a standard crystal oscillator with an accuracy of 1 in 105 or 1 in 106
is used instead. The receiver clock is allowed to have an offset relative to the GPS satel-
lite clocks, so when a time delay measurement is made, the measurement will have an
error caused by the clock offset. For example, suppose the receiver clock has an offset
of 10 ms relative to GPS time. All distance measurements will then have an error of
3000 km. Clearly, we must have a way to remove the time error from the receiver clock
before we can make accurate position measurements. CA code receivers can synchro-
nize their internal clocks to GPS time within 170 ns, corresponding to a distance meas-
urement uncertainty of 50 m. Repeated measurements and integration improve the
position location error to well below 50 m.
It is surprisingly easy to remove the clock error, and this removal is one of the
strengths of GPS. All that is needed is a time measurement from a fourth satellite. We
need three time measurements to define the location of the receiver in the three un-
known coordinates x, y, and z. When we add a fourth time measurement we can solve
the basic position location equations for a fourth unknown, the receiver clock offset
error . Thus the four unknowns in the calculation of the location of the receiver are
x, y, z, and .
Position Location in GPS

First, we will define the coordinates of the GPS receiver and the GPS satellites in a rec-
tangular coordinate system with its origin at the center of the earth. This is called the
earth centered earth fixed (ECEF) coordinate system, and is part of the WGS-84 de-
scription of the earth. WGS-84 is an internationally agreed description of the earth’s
shape and parameters, derived from observations in many countries4. GPS receivers use
the WGS-84 parameters to calculate the orbits of the GPS satellites with the accuracy
required for precise measurement of the range to the satellites. The Z-axis of the coor-
dinate system is directed through the earth’s North Pole and the X- and Y-axes are in the
equatorial plane. The X-axis passes through the Greenwich meridian—the line of zero
longitude on the earth’s surface, and the Y-axis passes through the 90° east meridian. The
ECEF coordinate system rotates with the earth. The receiver coordinates are (Ux, Uy, Uz),
and the four satellites have coordinates (Xi, Yi, Zi), where i 1, 2, 3, 4. There may be
more than four satellite signals available, but we use only four signals in a position cal-
culation. The measured distance to satellite i is called a psuedorange, PRi, because it
uses the internal clock of the receiver to make a timing measurement that includes er-
rors caused by receiver clock offset. The geometry of a GPS measurement is illustrated
in Figure 12.3.
c12.qxd 10/08/02 10:31 Page 465
12.3 GPS POSITION LOCATION PRINCIPLES 465
S2
S1
S3
R1 R2
R3 FIGURE 12.3 Position location

by measurement of the distance
to three satellites. The GPS re-
ceiver is located at point X,
where three spheres with radii R1,
R2, and R3 intersect. The centers
of the spheres are the three GPS
satellites S1, S2, and S3. If the
X distances R1, R2, and R3 are
measured, the location of the
point X can be uniquely defined.
Psuedorange, denoted as PRi, is measured from the propagation time delay Ti

between the satellite (number i) and the GPS receiver, assuming that EM waves travel
with velocity c.
PRi Ti c (12.1)
The distance R between two points A and B in a rectangular coordinate system is
given by
R 2 1xA xB 2 2 1yA yB 2 2 1zA zB 2 2 (12.2)
The equations which relate pseudorange to time delay are called ranging equations:
1X1 Ux 2 2 1Y1 Uy 2 2 1Z1 Uz 2 2 1PR1 tc2 2
1X2 Ux 2 2 1Y2 Uy 2 2 1Z2 Uz 2 2 1PR2 tc2 2
1X3 Ux 2 2 1Y3 Uy 2 2 1Z3 Uz 2 2 1PR3 tc2 2
1X4 Ux 2 2 1Y4 Uy 2 2 1Z4 Uz 2 2 1PR4 tc2 2 (12.3)
where is receiver clock error (offset, or bias).
The position of the satellite at the instant it sent the timing signal (which is actu-
ally the start of a long sequence of bits) is obtained from ephemeris data transmitted along
with the timing signals. Each satellite sends out a data stream that includes ephemeris data
for itself and the adjacent satellites. The receiver calculates the coordinates of the satellite
relative to the center of the earth, (Xi, Yi, Zi), and then solves the four ranging equations
for the four unknowns using standard numerical techniques for the solution of nonlinear
simultaneous equations. (The equations are nonlinear because of the squared terms.)
The four unknowns are the location of the GPS receiver, (Ux, Uy, Uz), relative to
the center of the earth and the clock offset —called clock bias in GPS terminology.
The receiver position is then referenced to the surface of the earth, and can be displayed
in latitude, longitude, and elevation. Typical accuracy for a low-cost GPS receiver us-
ing the GPS CA code is 30 m defined as a 2DRMS error. The term DRMS means the
distance root mean square error of the measured position relative to the true position of
c12.qxd 10/08/02 10:31 Page 466
the receiver. If the measurement errors are Gaussian distributed, as is often the case,
68% of the measured position results will be within a distance of 1DRMS from the true
location and 95% of the results will be within 2DRMS of the true location. Accuracy
in GPS measurements is usually defined in terms of 2DRMS, in the horizontal or vertical
plane.
In practice, the error surface that encloses 68 or 95% of all measurements is not a
circle but an ellipse, and the error in any dimension is affected by several dilution of
precision (DOP) factors. DOP is discussed later in this chapter. The U.S. Department of
Defense has the ability to degrade the position measurement accuracy of CA code re-
ceivers by applying selective availability (SA). SA exists to allow the accuracy of CA
code receivers to be degraded in the event of a national emergency (i.e., enemy action)
affecting the United States and was applied to the CA signals most of the time until May
2000. SA was switched off on May 1, 2000, and will not be used again unless the secu-
rity of the United States is threatened. With SA off, the accuracy of GPS position meas-
urements with the CA code increased dramatically, particularly in the vertical dimension.
Variation in elevation readout of a typical CA code receiver with SA on could be as large
as 200 m. With SA off, the variation may be as small as 10 m.
Selective availability and atmospheric propagation effects (tropospheric and ionos-
pheric) all cause errors in the timing measurements made by a GPS receiver, leading to
position location errors. The atmosphere and the ionosphere introduce timing errors be-
cause the propagation velocity of the GPS signals deviates from the assumed free space
value. The errors can be largely removed if a number of GPS reference stations are built
at precisely known locations. The stations observe the GPS signals and compute the current
error in position as calculated from GPS data. This information can then be broadcast to
all GPS users as a set of corrections to be applied to GPS measurements. The system is
called a wide area augmentation system (WAAS).
A network of 24 WAAS stations built in North America for the U.S. Federal Avia-
tion Administration (FAA) provides aircraft with improved position measurement accu-
racy. Using WAAS, accuracies of a few meters can be obtained with CA code receivers.
In the event of a national emergency, WAAS would be switched off to prevent enemies
using GPS for accurate targeting of weapons. WAAS also includes an integrity monitor-
ing system to ensure that the GPS signals used by aircraft do not contain errors which
could cause false readings. WAAS is required to send a warning of possible errors within
5.6 s if a problem is detected with any GPS satellite signal.
Similarly, a single reference station at a known location—for example, an airport—
can determine the local measurement error in GPS and broadcast this information to GPS
users so that greater accuracy can be obtained with a CA code receiver. This is one (sim-
ple) form of differential GPS (DGPS). More complex forms of differential GPS use a ref-
erence station which transmits the signals received from each GPS satellite so that phase
comparisons can be made by the receiver. With lengthy integration times and a sophisti-
cated phase comparison receiver, differential GPS accuracies of 1 cm can be obtained.
With DGPS, the receiver computes its position relative to the reference station rather than
in latitude and longitude. Differential GPS is used when a vehicle needs to be positioned
accurately with respect to a fixed point, such as an aircraft with respect to a runway or a
ship with respect to a berth.
GPS Time
The clock bias value which is found as part of the position location calculation process
can be added to the GPS receiver clock time to yield a time measurement that is synchronized
c12.qxd 10/08/02 10:31 Page 467
12.4 GPS RECEIVERS AND CODES 467
to the GPS time standard. The crystal oscillator used in the GPS receiver is highly stable
over a period of a few seconds, but will have a frequency which changes with tempera-
ture and with time. Temperature changes cause the quartz crystal that is the frequency
determining element of a crystal oscillator to expand or contract, and this changes the
oscillator frequency.
Crystals also age, which causes the frequency to change with time. The changes are
very small, but sufficient to cause errors in the clock time at the receiver when the clock
is not synchronized to a satellite. Calculating the clock bias by solving ranging equations
allows the receiver clock time to be updated every second or two so that the GPS receiver
time readout is identical to GPS time.
Every GPS receiver is automatically synchronized to every other GPS receiver any-
where in the world through GPS time. This makes every GPS receiver a super clock, which
knows time more accurately than any other time standard. Prior to the widespread use of
GPS receivers, standard time transmissions were broadcast by the U.S. National Institute
of Science and Technology (NIST, formerly the Bureau of Standards). The broadcasts
were made in the HF (shortwave) band, and could be received throughout the United
States. However, the HF signals propagate over long distance by reflection from the ion-
osphere, which introduces an uncertain delay into the time of arrival of the signal. The
time standard provided by GPS is typically accurate to better than 170 ns, and has been
used to synchronize electric power generators across the United States, for scientific ap-
plications that require synchronized clocks in different locations, and as a long-term
frequency standard.
The time standard on board each GPS satellite consists of two cesium clocks plus
two rubidium clocks (atomic clocks). An atomic clock uses the fundamental resonance of
the cesium or rubidium molecule as a frequency reference to lock a crystal oscillator. In
the GPS satellites, the master oscillator is at 10.23 MHz; all code rates, the L1, and the
L2 RF frequencies are multiples or submultiples of 10.23 MHz. The atomic clocks are up-
dated by the controlling ground stations to keep them within 1 s of Universal Time
Coordinated (UTC), and the navigation message broadcast by each satellite contains
information about its current clock errors relative to GPS time. (UTC is a worldwide time
standard. Greenwich Mean Time (GMT) is equal to UTC.)
12.4 GPS RECEIVERS AND CODES
GPS satellites transmit using pseudorandom sequence (PN) codes. All satellites transmit
a CA code at the same carrier frequency, 1575.42 MHz, called L1, using BPSK modu-
lation. The L1 frequency is 154 times the master clock frequency of 10.23 MHz. The CA
code has a clock rate of 1.023 MHz and the CA code sequence has 1023 bits, so the PN
sequence lasts exactly 1.0 ms. The exact values of the frequencies are about 0.005 Hz
lower than stated here to allow for relativistic effects caused by the high velocity of the
satellites in their orbits (3.865 km/s). (GPS measurements are one of the few examples
where relativistic effects must be taken into account, because the clocks are mounted on
platforms moving at very high speeds.)
The P code is transmitted using BPSK modulation at the L2 carrier frequency of
1227.6 MHz (120 10.23 MHz), and is also transmitted with BPSK modulation on the
L1 carrier frequency, in phase quadrature with the CA code BPSK modulation. Figure
12.4 shows the way in which the L1 and L2 signals are generated on board a GPS satellite.
The CA and P code transmissions from all GPS satellites are overlaid in the L1
and L2 frequency bands, making GPS a direct sequence spread spectrum (DS-SS) system
c12.qxd 10/08/02 10:31 Page 468
Q
π/2
L1
1575 MHz
I
Σ
L1 + P
+ C/A
C/A
+ NAV
code
1.023 C/A + NAV L1 output
Mbps
Navigation
message multipliers
50 bps
P + NAV
P
code 10.23
Mbps L2 output
L2 + P
I + NAV
L2
1227 MHz
FIGURE 12.4 Signal generation in a GPS satellite.
(see Chapter 6 for details of spread spectrum techniques). The receiver separates sig-
nals from individual GPS satellites using knowledge of the unique CA code that is
allocated to each satellite. At most, 12 GPS satellites can be seen by a receiver at any
one time, so the coding gain in the spread spectrum receiver must be sufficient to over-
come the interference created by 11 unwanted signals while recovering the twelfth
wanted signal.
The CA Code

The CA codes transmitted by GPS satellites are all 1023 bit Gold codes. GPS CA
Gold codes are formed from two 1023 bit m-sequences, called G1 and G2, by multi-
plying together the G1 and G2 sequences with different time offsets. An m-sequence is
a maximum length pseudorandom (PN) sequence, which is easy to generate with a shift
register and feedback taps. A shift register with n stages can generate a PN sequence
2n 1 bits in length. The bit pattern is set by the feedback taps and combining logic
of the shift register. The PN sequences G1 and G2 are both generated by 10-bit shift
registers and are therefore both 1023 bits long. The clock rate for the CA code is 1.023
MHz, so each sequence lasts 1.0 ms. Figure 12.5 shows a generator diagram for the
CA code.
The CA code for a particular satellite is created with an algorithm that includes the
identification number of the GPS satellite, thus creating a unique code for each satellite.
The satellite with ID number i has a CA code sequence Ci(t)
Ci 1t2 G11t2 G21t 10iTc 2 (12.4 )
where Tc clock period for the CA code.
There are 64 Gold sequences available for satellites numbered 1 through 64. A total
of 100 Gold sequences can be created using the algorithm in Eq. (12.4), but not all the
c12.qxd 10/08/02 10:31 Page 469
12.4 GPS RECEIVERS AND CODES 469
G1
3 10
10 bit G1 shift register
X1 Reset
to 1
+
Phase C/A code
selector
÷10 +
S1 S2
G2
2 3 6 8 9 10 1023
decode
1 kHz clock
10.23 MHz X1
Master clock + ÷20
50 Hz clock
10 bit G2 shift register
FIGURE 12.5 C/A code generator.
sequences have sufficiently low cross-correlation properties, and reference 4 states that
only 37 are actually used in the GPS system. Low cross-correlation of the sequences is a
requirement because the GPS receiver can pick up signals from as many as 12 satellites
at the same time.
A correlator in the receiver looks for one of the sequences and must reject all other
sequences that are present. Two CA code sequences with zero cross-correlation would
achieve a rejection ratio of 1023, but the 64 available CA code sequences will not all
have zero cross-correlation. The selected group of 37 are the sequences with the lowest
levels of cross-correlation among the available set of 100 Gold code sequences. They also
have low autocorrelation time sidelobes, another requirement of direct sequence spread
spectrum systems.
The CA code sequence length of 1.000 ms gives range ambiguity of 300 km, since
the code travels at a velocity of approximately 3 108 m/s and therefore has a length in
space of 3 105 m. The entire CA code sequence repeats in space every 300 km, leading
to ambiguity of position only if the GPS receiver is in outer space. The ambiguity is eas-
ily resolved if the receiver knows roughly where it is; just knowing that the receiver is lo-
cated close to the earth’s surface is usually sufficient. The user can enter the approximate
location of the GPS receiver when it is first switched on to help resolve any ambiguities
quickly.
Figure 12.6 shows a simplified block diagram of a CA code GPS receiver. The an-
tenna is typically a circularly polarized patch antenna with an LNA mounted on the printed
circuit board. A conventional superhet receiver is used to generate an IF signal in a band-
width of about 2 MHz, which is sampled using I and Q sampling techniques and processed
digitally. The digital portion of the receiver includes a CA code generator, a correlator,
c12.qxd 10/08/02 10:31 Page 470
Receiving
antenna
Digital signal
LNA Down-converter IF amplifier A/D converter processing
A/D DSP
Timing
Clocks Navigation data
message
Local oscillator
Microprocessor
Display
FIGURE 12.6 Simplified GPS receiver.
and a microprocessor that makes the timing measurements and calculates the receiver’s
position. Most GPS receivers make use of a 12-channel IC chip set that can be purchased
for about $25.00 (Year 2000 prices).
12.5 SATELLITE SIGNAL ACQUISITION
The GPS receiver must find the starting time of the unique CA code for each of four satel-
lites. This is done by correlating the received signal with stored CA codes, as in any di-
rect sequence spread spectrum system. (See Chapter 6 for details of this process.) Usually,
the receiver will automatically select the four strongest signals and correlate to those. If
the geometry of the strongest satellites is poor, that is, the satellites are close together and
have pseudoranges that are nearly equal, the receiver may also use several weaker signals.
If the receiver is making a cold start, with no information about the current position of
GPS satellites, or its own location, it must search all 37 possible CA codes until it can
correlate with one. Once correlation is obtained, the data stream (called the navigation mes-
sage) from that satellite can be read by the receiver. The data stream contains information
about the adjacent satellites, so once one signal is correlated, the receiver no longer needs
to search through all the other 36 possible codes to find the next satellite; it can go directly
to the correct code. Searching all 36 CA codes of 1023 bits for correlation can be a slow
process. In the worst case, 36 codes might have to be searched before a correlation could
be obtained. However, available satellites in 2000 all had numbers between 13 and 45 5, so,
on average, 16 codes might have to be searched before correlation is successful.
A direct sequence spread spectrum receiver locks to a given code by matching the lo-
cally generated code to the code received from the wanted satellite. Since the start time
of the code transmitted by the satellite is not known when the receiver commences the lock-
ing process, an arbitrary start point must be selected. The locally generated code is
compared to the received code, bit by bit, through all 1023 bits of the sequence, until
either lock is found, or the receiver concludes that this is not the correct code for the
satellite signal it is receiving.
If the starting time for the locally generated code was not selected correctly, corre-
lation will not be obtained immediately. (This will occur with a probability of 99.9% when
c12.qxd 10/08/02 10:31 Page 471
12.5 SATELLITE SIGNAL ACQUISITION 471
the timing of the locally generated sequence is selected at random.) The locally generated
code is then moved forward one bit in time, and correlation is attempted again. The process
is continued 1023 times until all possible starting times for the locally generated code
have been tried. If the satellite with that particular CA code is not visible, no correlation
will occur and lock will not be achieved. It takes a minimum of 1 s to search all 1023 bit
positions of a 1023 bit CA code, so in a typical case, it will take at least 15 s to acquire
the first satellite. Many receivers search for a given CA code several times before mov-
ing to the next code, so several minutes may elapse before the correct CA code is found,
given no other information. Once one CA code is found, the remaining satellites can then
be acquired in a few seconds because their IDs are known from the data transmitted in
the navigation message of each satellite.
Although it takes only 20 s on average to lock to the CA code of one satellite, the re-
ceiver must find the Doppler frequency offset for at least one satellite before correlation can
occur. The receiver bandwidth is matched to the bandwidth of the CA code. The theoreti-
cal noise bandwidth of the CA code receiver is 1.023 MHz and the velocity of the satellites
is 3.865 km/s. The angle between the spacecraft velocity vector and a receiver on earth is
76.1° when a GPS satellite is at the horizon, so the maximum velocity component toward a
receiver is vr 928 m/s, giving a maximum Doppler shift in the L1 signal of vr 4.872
kHz, ignoring the effect of earth rotation. Allowing the satellite to reach an elevation angle
of 5° before it is used for a position measurement limits the Doppler shift that must be ac-
commodated by the receiver to 4 kHz. From a cold start, the receiver must try eight Doppler
frequency shifts of up to 4 kHz in 1-kHz steps when searching for the signal from a satel-
lite. This can increase the acquisition time of the first satellite to several minutes. Figure 12.7
illustrates the search process. There are eight possible Doppler shifts for each signal, and
1023 possible code positions, giving 8184 possible signal states that must be searched.
Once any of the GPS satellites has been acquired, the navigation message provides
sufficient information about the adjacent satellites for the remaining visible satellites to be
acquired quickly. The receiver may need to search in Doppler shift because the position of
+4 kHz
Doppler shift
Signal
0
BIF = 1 kHz
−4 kHz
0 1 µs 1 ms
C/A code timing

FIGURE 12.7 Code synchronization and Doppler tracking matrix.
c12.qxd 10/08/02 10:31 Page 472
the receiver relative to the satellites is not known, but their CA codes are. The GPS re-
ceiver retains the information from the navigation message when switched off, and it also
runs its internal clock. When next switched on, the receiver will assume that its position is
close to its last known position when it was switched off, calculate which satellites should
be visible, and search for those first. This greatly speeds up the acquisition process. If the
receiver has been moved a large distance while turned off, a cold start may be needed.
The correlation process described above assumes that each satellite is acquired se-
quentially. Some lower cost GPS receivers use sequential acquisition of the satellites, and
also make timing measurements sequentially, one satellite at a time. More sophisticated
receivers have parallel correlators which can search for and acquire satellites in parallel.
Twelve parallel correlators guarantee that all visible GPS satellites will be acquired, and
start-up time is much shorter than with sequential acquisition. Accuracy is also better with
parallel processing of the signals.
Integrity monitoring of the GPS position measurement is possible by using a fifth
satellite to recalculate the receiver position. With five satellite signals there are five pos-
sible ways to select four pseudoranges to use in the ranging equations, leading to five cal-
culations of position. If there is disagreement between the results, one bad measurement
can be eliminated. If more than one result disagrees with the others, the integrity of the
measurements is compromised. GPS receivers used for navigation of aircraft in instru-
ment meteorological conditions (IMC, in the clouds) and for instrument landings are
required to have integrity monitoring to guard against receiver or satellite failures and
interference with or jamming of GPS signals.
The P code for the ith satellite is generated in a similar way to the CA code. The
algorithm is
Pi 1t2 X11t2 X21t iTc 2 (12.5)
where Tc is the period of the X1 sequence, which contains 15,345,000 bits and repeats
every 1.5 s. The X2 sequence is 37 bits longer. The P code repeats after 266.4 days, but
is changed every 7 days for security reasons. The long length of the P code sequence
makes the distance measurements unambiguous. P code sequences cannot be acquired eas-
ily because they do not repeat, a deliberate feature to prevent unauthorized users from op-
erating high accuracy GPS receivers. The CA code provides information to authorized
users on the starting time of the P code; this is contained in the navigation message as an
encrypted handover word. If the current feedback tap settings for the P code generators
are known, and the handover word is decrypted, the receiver can start the local X Code
generators close to the correct point in the P code sequence. This allows rapid acquisition
of the P code, and is the origin of the name coarse acquisition for the CA code.
12.6 GPS NAVIGATION MESSAGE
A key feature of the GPS CA code is the navigation message. The navigation message
contains a large amount of information that is used by GPS receivers to optimize the
acquisition of satellite signals and calculate position. The navigation message is sent at
50 bps by BPSK modulation of the CA and P codes. Effectively, 20 CA code sequences
form one navigation message bit. The phase of the 20 sequences is inverted between the 1
and 0 bits of the message by modulo-2 addition of the navigation message data to the CA
and P code sequences. The navigation signal is extracted by a 50-bps BPSK demodulator
that follows the CA or P code correlator. The narrow bandwidth of the navigation message
ensures a high SN ratio at the demodulator input and correspondingly low probability of
c12.qxd 10/08/02 10:31 Page 473
12.7 GPS SIGNAL LEVELS 473
TABLE 12.1 GPS Navigation Message: Subframe Details
Header Telemetry message: health of satellite, handover word

Subframe 1 Satellite clock correction data. Age of transmitted data
Subframe 2 and 3 Ephemeris for this satellite
Subframe 4 Almanac data for satellites 25 and higher. Ionospheric model data
Subframe 5 Almanac data for satellites 1–24. Health data for satellite 1–24
Note: Subframes 1, 2, and 3 repeat all data every 30 s. Subframes 4 and 5 repeat every 30 s, but
transmission of the full data set requires 25 subframes over a period of 12.5 min.
bit errors in the navigation message. Satellites with elevation angles above 10° will typi-
cally give a SN ratio of greater than 17 dB at the output of the correlator.
The complete navigation message is 1500 bits, sent as a 30-s frame with 5 sub-
frames. However, some information is contained in a sequence of frames, and the com-
plete data set requires 12.5 min for transmission. The most important elements of the
message are repeated in every frame. The subframes contain the satellite’s clock time data,
orbital ephemeris for the satellite and its neighbors, and various correction factors. Details
of the subframes are given in Table 12.1.
The calculation of position in a GPS receiver requires very accurate knowledge of the
location of the satellite at the time that the measurements of pseudoranges are made. If the
pseudorange is measured to an accuracy of 2.4 m, we must know the satellite position to an
even greater accuracy, and that requires very accurate calculation of the GPS satellite orbits.
By comparison, the orbit of a communication satellite does not need to be known to the
same level of accuracy. As described in Chapter 2, the GPS system uses modified WGS-84
data to define the earth’s radius, Kepler’s constant, and the earth’s rotational rate. Data on
the speed of EM waves is taken from the International Astronomical Union. The WGS-84
data set also includes a very detailed description of the earth’s gravitational field, which
is essential for precise location of the satellites in their orbits. All of these parameters and
corrections are stored in every GPS receiver, and used in calculating position.
12.7 GPS SIGNAL LEVELS
GPS receiver antennas have low gain because they must be omnidirectional. We will assume
a worst-case gain of G 0 dB, corresponding to an isotropic antenna. In practice, G 0
dB in many directions, but may fall to 0 dB in some directions. The omnidirectional antenna
picks up radiated noise from the environment, making the antenna temperature close to 273
K. LNA temperatures can be as low as 25 K, so a system noise temperature of 273 K will
be used as a typical value. Typical GPS antennas are circularly polarized patches or quadrafi-
lar helices that have carefully shaped patterns that cut off quickly below 10° elevation to min-
imize noise pick up from the ground. The LNA is mounted directly below or behind the
antenna to avoid the increase in noise temperature caused by lossy antenna cables.
GPS satellites have an array of helical antennas that provide gain toward the earth,
and 10 W transmitters, leading to EIRP values in the range 19 to 27 dBW. The CA code
transmitted by the satellite is a direct sequence spread spectrum signal, so the CN ratio
in the CA code’s RF bandwidth will be less than 0 dB. This is typical of systems that
use direct sequence spread spectrum signals. The low CN ratio of the spread spectrum
signal is converted to a usable SN by correlation of the code sequences, which adds a
despreading (processing) gain to the CN ratio. The theoretical processing gain of a direct
sequence spread spectrum signal is equal to the ratio of the chip rate to the bit rate in the
c12.qxd 10/08/02 10:31 Page 474
spreading sequence, but losses in the correlation process always make practical gains a
little lower. For the CA code transmitted at 1.023 Mbps and a 1-ms correlation time, the
theoretical processing gain is 1023, or 30.1 dB. The corresponding processing gain for
the P code is 40.1 dB.
The GPS receiver can pick up signals from up to 10 satellites at the same time. The
RF energy from the satellite spread spectrum transmissions adds to the noise in the receiver
as an interference term, I. For simplicity, in the following analysis we will assume that
there are 10 GPS satellites visible, that there are 9 interfering satellites generating random
signals (noise) out of which the receiver must extract the 10th signal, and that all the re-
ceived signals are of equal strength. The signals from interfering satellites are treated as
random noise because the Gold codes that they transmit have very low cross-correlation
with the code from the wanted satellite. Noise has zero cross-correlation with the wanted
signal, and the Gold codes used by GPS satellites are chosen because they closely approxi-
mate noise.
Nine interfering GPS satellites represents a worst case; in practice the number of
visible satellites varies between four and ten, and the signal strengths also vary depend-
ing on the elevation angle of the satellite and the antenna pattern at the receiver. The worst
case is actually when a weak signal from a satellite at a low elevation angle must be ex-
tracted from stronger signals from satellites at higher elevation angles. GPS receivers au-
tomatically select the strongest signals for processing so that the worst case can be avoided,
but if the sky is partially blocked by obstructions, a weak signal may have to be used.
Table 12.2 shows the downlink signal power budget for the L1 and L2 carriers. A
receiving antenna gain of 0 dB is assumed.
The interference from nine CA code spread spectrum signals of equal power is
given by the sum of the received power (in watts) from each satellite
I 9 1016 W 1 150.5 dBW
The thermal noise power, N, in a noise bandwidth of 2 MHz for a noise temperature of
273 K is kTBn watts, where
N 141.2 dBW 1 7.59 1015 W
The noise and interference powers must be added in watts, not in decibels:
N I 8.49 1015 W 1 140.7 dBW
Hence the worst case CN for one CA code signal in this scenario is
C 1N I2 160.0 1140.72 19.3 dB (12.6)
Similar analysis yields the C(N I) values for the two P code signals, as shown
in Table 12.3. Note that thermal noise is the major factor in setting C(N I), since in
TABLE 12.2 Link Budget for L1 and L2 Carriers
Carrier L1 (1575 MHz) Carrier L2 (1227 MHz)
Code CA code P code P code
EIRP (dBW) 26.8 23.8 19.7

Path loss (dB) 186.8 186.8 185.7
Receive antenna gain (dB) 0 0 0
Pr (dBW) 160.0 163.0 166.0
c12.qxd 10/08/02 10:31 Page 475
12.8 TIMING ACCURACY 475
TABLE 12.3 CN, Noise, and Interference Budget for L1 and

L2 Carriers
Carrier L1 (1575 MHz) Carrier L2 (1227 MHz)
Code CA code P code P code
Ts (dBK) 24.4 24.4 24.4

Bn (dBHz) 63.0 73.0 73.0
N (dBW) 141.2 131.2 131.2 (thermal)
I (dBW) 150.5 153.0 156.0 (nine satellites)
N I (dBW) 140.7 131.1 131.2
Pr (dB) 160.0 160.3 166.0
C(N I) (dB) 19.3 31.9 34.8
Gproc (dB) 30.1 40.1 40.1
SN (dB) 10.7 8.2 5.3
the worst case of interference caused by nine visible satellites, all received at maximum
power, the interference power level is 9.3 dB below the thermal noise power. The CN
ratio at the receiver is 0.7 dB lower when the interference from the nine visible satellites
in included. A more realistic scenario would have four satellites at the maximum receive
power level and the remainder at a lower level, since GPS satellites orbit in constellations
of four, with one constellation always visible, to improve the accuracy of position loca-
tion measurements. Thus we would expect less than 0.7 dB degradation in the CN ratio
due to interference by other satellites’ CDMA signals for almost all of the time.
The SN at the correlator output is 10.7 dB for the CA code and 8.2 dB for the
L1 P code, using the CN values in Table 12.3 and the theoretical processing gains for
each code with no losses in the correlation and filtering of the signals. Historically, the
earlier generations of GPS satellites have had transmitter EIRPs up to 3 dB higher than
indicated by Table 12.3. Receiving antennas with gain greater than 0 dB also help to in-
crease the CN ratio, so CA code SN ratio can be up to 6 dB higher than the specifi-
cation value of 10.7 dB for the CA code.
The navigation message has a 50-bps bit rate, and each bit extends over 20 CA
code correlation periods. The CA code correlator output is passed through a 50 Hz band-
width filter which integrates the 20 pulses from the correlator to give a single message
bit, in the form of a 50 bps BPSK signal. The SN ratio of the BPSK message signal will
theoretically be 13 dB higher than the SN at the correlator output, at 23.3 dB. However,
the correlation and filtering processes are not perfect and an implementation margin of
several dB must be allowed. Nevertheless, the SN ratio of the BPSK signal will be above
20 dB in most cases, guaranteeing error free detection of the navigation message.
12.8 TIMING ACCURACY
The position location process requires an accurate measurement of the time of arrival of
the code sequence at the receiver. The output of the CA code correlator is a 1 s wide
pulse that repeats every millisecond. The accuracy with which a timing measurement can
be made on a single pulse is given by the approximate relationship6
1
dt
3Bn 1S N4
seconds (12.7)
c12.qxd 10/08/02 10:31 Page 476
where t is the rms timing error, Bn is the noise bandwidth of the RF channel, and SN
ratio is the signal-to-noise power ratio (not in dB) for the pulse in the noise bandwidth
Bn.
The SN ratio after the correlator is
SN CN Gp losses (12.8)
where Gp is the correlator processing gain. For the CA code Gp 1023 30.1 dB and
SN 19.3 30.1 dB losses
11.7 dB losses
If we assume the specification value for SN of 11.7 dB and losses of 1.7 dB,
SN 10 dB, a power ratio of 10. The theoretical noise bandwidth of the correlator is
Bn 1 MHz (IF noise bandwidth) so
dt 1 3106 1104 s 0.316 ms (12.9)
A typical GPS receiver will update the display no more than twice a second, so the pulses
from the correlator can be averaged over a period of half a second, which will decrease
the rms error by 1500 22.4 to an rms value of 14 ns, assuming randomly distributed
errors. The 14 ns rms timing error translates to an rms distance error of 4.2 m. However,
four distance measurements are needed to obtain a position measurement, so with no
other errors accounted for, the basic position measurement accuracy of the CA code re-
ceiver is about 8.4 m 14.2 142 measured as an rms value. A higher CN ratio in the
receiver will improve the accuracy, but other errors, discussed later in Section 12.10, will
lower the accuracy.
The accuracy achieved by commercial CA code GPS receivers was better than ex-
pected by the designers of the GPS system. Military strategists became concerned that
CA code GPS receivers could be used to target weapons against the United States with
considerable accuracy. The U.S. Department of Defense (DOD) introduced selective
availability (SA), a scheme to deliberately degrade the accuracy of CA code receivers
by varying some of the parameters of the GPS satellites. Selective availability was
switched off on May 1, 2000, and will be turned on only if the security of the United
States is threatened.
12.9 GPS RECEIVER OPERATION
A CA code GPS receiver must be able to correlate signals from at least four satellites,
calculate time delays, read the navigation message, calculate the orbits of the GPS
satellites, and calculate position from pseudoranges. The key to accurate position deter-
mination is accuracy in the timing of the arrival of the Gold code sequences from each
satellite in view. All GPS receivers use a microprocessor to make the required calculations
and to control the display of data. There are many different ways that this can be done,
depending on the application for which the receiver is intended. The tasks of the micro-
processor will not be considered here—it is assumed that once accurate timing data is
available and the navigation message read that the microprocessor can complete its
required tasks.
Most CA code GPS receivers use an IC chip set that contains 12 parallel correla-
tors. This allows the receiver to process signals from up to 12 satellites at the same time,
which helps keep all the signals synchronized. Some simpler receivers use a single
c12.qxd 10/08/02 10:31 Page 477
12.9 GPS RECEIVER OPERATION 477
correlator and process four satellite signals sequentially, with consequent lower accuracy.
The received GPS signals are converted to a suitable IF frequency in the front end of the
receiver, and then processed to recover the CA codes. In more recent GPS receivers,
much or all of the signal processing is done digitally using DSP techniques.
The explanation of the signal processing techniques used in GPS receivers that fol-
lows is based on block diagrams that can be implemented in analog or digital form. The
blocks presented in this discussion are those that would be found in an analog receiver.
Most GPS receivers implement them using digital signal processing techniques (DSP).
We will start the analysis by considering the signal received from the satellite at the out-
put of the IF stage of the receiver.
The IF signal in the GPS receiver will consist of the sum of a number (up to 12) of
signals from visible GPS satellites. The IF carrier signal has several BPSK modulations
applied to it by the satellite, and when received on earth has been Doppler shifted by satel-
lite and earth motion. The IF signal from N GPS satellites in view is
s1t2 a 5AiCi 1t2Di 1t2 sin 3 1vi vd 2t fi 1li 2 fi 4 6

N
(12.10)
i1
where Ai is the amplitude of the received signal.

Ci(t) is the Gold code modulation
Di(t) is the navigation message modulation
i is the IF frequency of the received carrier
d is the Doppler shift of the received signal
i(li) is the phase shift along the path
i is the phase angle of the transmitted signal
The key to successful measurements in a GPS CA code receiver is to generate a
signal in the receiver that is identical to the signal received from satellite i, but without
the navigation data that is modulated onto the transmitted signal. When the correct signal
is generated in the receiver it has the correct CA code for satellite i, the code has the
correct starting delay, and the correct Doppler shift has been applied. The locally gener-
ated signal is multiplied by the received signal, which contains several other signals from
visible GPS satellites, and the output is integrated over the CA code length of 1 ms. The
result is a constant output for a period of 20 ms, corresponding to the duration of a nav-
igation data bit. The precise matching of the locally generated signals to the received sig-
nals from four visible GPS satellites ensures that the local receiver’s chip clocks and CA
code generators are exactly in sync with the received signals. When this condition is
achieved, the start time of each CA code sequence and the corresponding chip clock tran-
sition provide the high accuracy time marker that makes GPS time delay measurement
possible.
The receiver must measure i(li) in Eq. (12.10) as a time delay in order to obtain
the pseudorange for each of the N satellites in view, and it must recover the Ci(t) modulation
by correlation. The Di(t) modulation contains the navigation message as a 50 bps BPSK
modulation of the Ci(t) signal. Both the C(t) and D(t) signals are modulated onto the car-
rier of the satellite signal by binary phase shift keying (see Chapter 5) and therefore have
values 1. Demodulation of BPSK signals requires a locally generated carrier which is
locked to the phase of the received carrier, and recovery of the data signal requires a bit
clock that is locked to the bit rate of the received signal.
The wanted signal is buried below the receiver noise and CDMA interference. We
must multiply the signal and noise by the wanted CA code sequence to despread the
c12.qxd 10/08/02 10:31 Page 478
signal and to bring it above the noise. The nominal bandwidth of the signal is 1 kHz af-
ter the correlator, since the 1023 bit sequence of the CA code repeats every millisec-
ond. However, the IF carrier can be shifted in frequency by up to 4 kHz because of
Doppler effects. The receiver must therefore first search in Doppler frequency space—
eight 1 kHz frequency offset steps—until a signal is found. This is done as part of the
signal acquisition process by incrementing the frequency of the locally generated carrier
in 1 kHz steps.
Part of a typical receiver structure for the GPS CA code is shown in Figure 12.8.
The function of the non-coherent delay lock loop is to set the frequency of the volt-
age controled oscillator (VCO) in the receiver to match CA code rate of the received
signal, and to align the received chip transitions correctly. GPS satellites generate all
their signals from a master clock, which means that there is phase coherence between
the chips, the codes and the RF frequencies of all GPS signals from a particular satel-
lite. The delay lock loop shown in Figure 12.8 takes advantage of the coherent nature
of the GPS CA signals, so that the VCO becomes both a time reference for the CA
code signals and also the chip clock. The PN code generator in Figure 12.8 must be
set to the correct code, and its start time must also be set correctly, for the loop to lock.
When the IF CA code in the receiver is correctly generated and has the correct
frequency and timing, it will exactly match the received CA code at the input to the
delay lock loop.
The delay lock loop has three paths: punctual, early (half chip ahead), late (half chip
behind). The delay lock loop steers the chip clock so that the punctual output can be used
to drive the CA code generator. The CA code chip rate is generated by the VCO. The
incremental process of trial and error which eventually finds the correct sequence and tim-
ing was described above. The early-late channels in the delay lock loop generate output
Punctual
BPF
Output
D i sin(ω IFt )
C(t − τ ) B = 50 Hz
Early
Envelope +
BPF detector
C(t − τ − Tc /2) Σ
Envelope −
BPF detector
Late
Input C(t − τ + Tc /2)

D i C(t − τ ) sin(ω IFt + φ )
PN code VCO Loop
generator (chip clock) filter
B = 1 kHz
Code select
FIGURE 12.8 Noncoherent code lock loop and navigation message recovery.
VCO, voltage controlled oscillator.
c12.qxd 10/08/02 10:31 Page 479
12.9 GPS RECEIVER OPERATION 479
signals which steer the phase of the VCO so that the navigation message is recovered
correctly.
The locally generated carrier that is used to demodulate the C(t) signal must be
Doppler shifted to match the Doppler offset of the received signal, and modulated with
the correct CA code sequence, starting at the correct time. The correct Doppler shift,
code sequence, and start time are all unknown when the receiver is first switched on. The
signal is buried below the noise, so it is not possible to determine the correct parameters
by direct analysis of the received signal. The receiver must therefore be designed to search
all possible Doppler shifts, code sequences, and code start times until an output is ob-
tained from the correlator indicating that a satellite signal has been found. Once one GPS
satellite signal has been found, information contained in the navigation message can be
used to steer the receiver to the parameters needed to acquire the other visible satellites.
If the receiver is turned off and then turned on again, the microprocessor memory has the
last known satellite configuration stored, and can derive expected signal parameters by
allowing for the time for which the receiver was off.
The output of the CA code correlator with Doppler corrected IF frequency for the
satellite signal with code number M is
x1t2 Am R 1tm t2Dm 1t2 sin3vm 1t2 fm 1lm 2 fm 4 n1t2 (12.11)
where R(m ) is the autocorrelation function of the wanted code number M, and n(t)
is the output from cross-correlation with all other codes.
The time shift (tm ) to the correlation peak is the wanted measurement that pro-
vides the pseudorange to the satellite. The output of the correlator is a despread signal at
baseband, which is modulated with the 50 bps navigation message. With the CA code
removed by the correlation process, it is a straightforward process to demodulate the nav-
igation message D. Passing this signal through a narrow bandwidth bandpass filter
improves the SN ratio and ensures that the message is recovered without errors. The IF
carrier is recovered with a special type of phased locked loop (PLL) called a Costas loop.
A Costas loop compensates for the arbitrary phase of the received signal.
The despread IF carrier is BPSK modulated by the navigation message Dm(t)
y1t2 AmR 1tm t2 Dm 1t2 sin 3vm 1t2 fm 1lm 2 fm 4 n1t2 (12.12)
The IF carrier signal is limited to remove any amplitude variations, which sets Am 1. Then
y¿1t2 R 1tm t2 Dm 1t2 sin3vm 1t2 f¿m 4 n1t2 (12.13)
The navigation message D(t) is recovered by multiplying the IF signal y (t) by
sin[m(t) m] and low pass filtering to obtain the 50 bps signal. The reference carrier
for the BPSK demodulator can be derived from the output of the Costas loop. The
demodulated message signal is z(t) where
z 1t2 R 1tm t2Dm 1t2 n¿1t2 (12.14)
Provided that the correlation peak of z(t) crosses the threshold and n (t) doesn’t, we can
recover the data message Dm(t) correctly. If everything works correctly in the receiver, the
SN of the signal y (t) is at least 17 dB, so there will be no bit errors. Even if a bit error
occurs in the navigation message, it is removed when the next message is received about
30 s later.
Figure 12.9 shows a Costas loop which is often used as the demodulator for low
speed BPSK signals such as the 50 bps GPS navigation message. The loop has an I channel
and a Q channel driven by a VCO. The VCO frequency is set by the sum of the outputs
c12.qxd 10/08/02 10:31 Page 480
Navigation
1
2 D(t ) cos φ message
LPF
B = 50 Hz
Input sin(ω IFt + φ )
D i sin(ω IFt + φ )
VCO Loop
ω IF filter
π /2 1
sin(2 φ )
8
cos(ω IFt + φ )
LPF 1
2 D(t ) sin φ
FIGURE 12.9 Costas loop. LPF, low pass filter.
from the I and Q channel detectors, which steers the VCO phase such that the I channel
is in phase with the signal. The I channel output is then (ideally) a zero ISI waveform
which can be integrated and sampled to recover the navigation message bits.
12.10 GPS C/A CODE ACCURACY
The major sources of error in a GPS receiver that calculates its position are:
Satellite clock and ephemeris errors
Selective availability (when switched on)
Ionospheric delay (advance)
Tropospheric delay
Receiver noise
Multipath
The accuracy that can be achieved with a GPS CA code receiver can be found by
using a range error budget. The figures in square brackets [ ] are for the case when selective
availability (SA) is turned off. Typical values of range error are given in Table 12.4. All
values are in meters (m). Note that a value of 2.4 m error is assigned to receiver noise.
The value calculated in Section 12.8 was 4.2 m, for a worst-case received signal strength.
The range error introduced by the ionosphere and the troposphere can be partially
removed by receiving identical signals at two different carrier frequencies. This technique
is used by high precision P code receivers. The P code signal is transmitted on the L1 car-
rier at 1575.42 MHz, in phase quadrature with the CA code signal. The P code is also
transmitted on the L2 carrier at 1227.60 MHz. Algorithms are used in the P code receiver
to calculate the net delay of the signal caused by the ionosphere and the atmosphere, and
to then remove the errors from the calculated ranges. CA code receivers use a standard
atmosphere and ionosphere and assume a constant delay at a given elevation angle. Vari-
ations in the density of the atmosphere with atmospheric pressure changes, and in the free
c12.qxd 10/08/02 10:31 Page 481
12.10 GPS CA CODE ACCURACY 481
TABLE 12.4 Range Error for C/A Code

Measurements (m)
Satellite clock error 3.5

Ephemeris errors 4.3
Selective availability 32.0 [0]
Ionospheric delay 6.4
Tropospheric delay 2.0
Receiver noise 2.4
Multipath 3.0
RMS range error 33.4 m with SA
[RMS range error 9.5 m without SA]
Brackets indicate SA off.
electron content of the ionosphere, lead to departure from the standard values and hence
to errors in the pseudorange calculation. There are plans to transmit the CA code at a
third and a fourth L-band frequency from later GPS satellites to provide improved accu-
racy with CA code receivers8.
The range error shown in Table 12.4 is for one satellite–earth path, for the pseudo-
range that is calculated from the timing measurements using the receiver clock. How-
ever, four pseudorange measurements are needed to make a position determination. Thus
the position location output of the GPS receiver combines four path errors, which are
not necessarily equal because of the geometry of the satellites in the sky and the dif-
ferent signal strengths at the receiver input. Receiver position is calculated in (x, y, z)
coordinates, and the errors in x, y, and z depend on the elevation angle of satellites, the
satellite geometry, and the other parameters in the error budget. The calculated position
will have different levels of error in the x, y, and z directions. To account for these
differences several dilution of precision factors (DOP) are defined. A DOP factor mul-
tiplies the basic position measurement error to give a larger error caused by the particular
DOP effect.
Dilution of Precision: HDOP, VDOP, and GDOP

Horizontal dilution of precision is one of the most important DOP factors for most GPS
users. It provides an error metric for the x and y directions, in the horizontal plane. A typ-
ical HDOP value is 1.5, and it is often the smallest of the DOPs. Horizontal measurement
error for a CA code receiver is typically 14.3 m with SA off (1DRMS) and 50 m with
SA on (1DRMS). GPS practice uses 2DRMS as the quantifier for accuracy in position
determination giving a 2DRMS accuracy of 28.6 m with SA off. The 2DRMS accuracy
figure means that 95% of all measurements yield a position within 28.6 m of the true
location of the GPS receiver, in this example.
There are many DOP factors in GPS. The more important ones are horizontal dilu-
tion of precision, HDOP, vertical dilution of precision, VDOP, and geometric dilution of
precision, GDOP. Other DOPs include position dilution of precision, PDOP, and time di-
lution of precision, TDOP. In general, VDOP and GDOP are most likely to degrade the
accuracy of GPS position measurements. VDOP accounts for loss of accuracy in the
vertical direction caused by the angles at which the satellites being used for the position
measurement are seen in the sky. If the satellites are all close to the horizon, the angles
c12.qxd 10/08/02 10:31 Page 482
between the satellites and the receiver are all similar and VDOP can be large. In the worst
possible case, if all the satellites were at the horizon, it would be impossible to make an
accurate measurement in the vertical direction. A change in range to at least one satellite
must occur when the receiver is moved, otherwise the receiver cannot detect that change.
If all the satellites are at the horizon, no range change occurs for vertical movement of
the receiver and consequently vertical accuracy is very poor. Similarly, if all the satellites
were clustered directly overhead, HDOP would be large.
VDOP is important in aircraft position measurements, where height above the ground
is a critical factor, especially when landing. CA code receivers suffer from significant
VDOP and cannot provide sufficient vertical accuracy for automated landing of aircraft.
CA code GPS receivers cannot guarantee sufficient vertical accuracy unless operated in
a differential GPS mode.
The GPS satellites are configured in orbit to minimize the probability that a DOP
can become large, by arranging the orbits to provide clusters of four satellites with suitable
spacings in the sky. However, if the receiver’s view of the sky is restricted, for example,
by buildings, the geometry for the position calculation may not be ideal and GDOP can
become large. This causes all the other DOP values to increase. Aircraft, and ships at sea
always have a clear view of the sky, but automobiles often do not. CA code receivers
may revert to two-dimensional measurements (x and y) using three satellites when the sky
is obstructed.
12.11 DIFFERENTIAL GPS
The accuracy of GPS measurements can be increased considerably by using differential

GPS (DGPS) techniques. There are several forms of DGPS, all of which are intended to
increase the accuracy of a basic GPS position measurement, and to remove the effects of
selective availability. A second, fixed, GPS receiver at a reference station is always required
in a differential GPS system. In the simplest forms of DGPS, a second GPS receiver at a
known position continuously calculates its position using the GPS CA code. The calcu-
lated (x, y, z) location is compared to the known location of the station and the differences
in x, y, and z are sent by a radio telemetry link to the first GPS receiver. The accuracy of
the CA code position measurement can be increased from 100 m to about 10 m, with
SA in effect, but this technique works well only if the two stations are close together and
use the same four satellites for the position calculation.
In a more sophisticated form of differential GPS, the monitoring station at a known
location measures the error in pseudorange to each satellite that is visible at its location,
and telemeters the error values to users in that area. This allows other GPS users to select
which satellites they want to observe, and extends the area over which the DGPS system
can operate. The accuracy of a CA code measurement can be increased to 5 m for re-
ceivers within 10 km of the reference station and to 10 m for receivers within 500 km of
the reference station.
The most accurate forms of differential GPS use the relative phase of the many
signals in the GPS transmissions to increase the accuracy of the timing measurements.
Suppose that you could count the number of cycles of the 1575 MHz L1 carrier wave be-
tween a satellite and a GPS receiver, and that the GPS satellites are stationary for the length
of time it takes to make the count at two separate locations. The wavelength of the L1
carrier is 0.19043 m, so movement of the receiver by 0.01 m directly away from the satel-
lite would change the phase angle of the received wave by 18.9°. If the total number of
cycles between the satellite and the receiver is known, and fractional cycles are measured
c12.qxd 10/08/02 10:31 Page 483
12.11 DIFFERENTIAL GPS 483
with a phase resolution of 20°, the true distance to the satellite can be found to 0.01 m
accuracy. In principle, measurements which compare the phase angle of the received L1
carriers from several GPS satellites could therefore be used to detect receiver movements
at the centimeter level. This is called differential phase or kinematic DGPS.
The obvious difficulty is that we cannot count the number of cycles of the L1 car-
rier between the satellite and the receiver. However, we can make phase measurements
and time of arrival comparisons for various GPS signals at two different locations and
resolve motion between the two locations. If one of the receivers is a fixed reference sta-
tion, it is then possible to locate the second GPS receiver very accurately with respect to
that fixed location.
This technique is valuable in land surveying, for example, where a reference station
can be set up at a known location, such as the corner of a plot of land, and the position
of the plot boundary relative to that point can be measured. The same technique can be
used to find the position of an aircraft relative to an airport runway so that a precision ap-
proach path can be established.
The difficulty with DGPS phase comparison measurements is that the L1 carrier
has cycles which repeat every 0.19043 m, and one cycle is identical to the next. This cre-
ates range ambiguity which must be resolved by reference to the wavelengths of other
signals. The 10.23 MHz P code transmission of the L1 carrier has a P code chip length
in space of 29.326 m, which is 154 cycles of the L1 carrier. The ambiguity of the carrier
waveform can be resolved within the 29.326 m length of a P code chip by comparison of
the time of arrival of a particular cycle of the L1 carrier with the time since the start of
the P code chip. Similar ambiguity resolution for the 29-m P code chips is possible using
the length of the CA code chip and the CA code sequence. The length of a CA code
chip at 1.023 MHz is 293.255 m, and the length of a CA code sequence is 293.255 km.
When ambiguity resolution is applied using all of these waveforms, very small movements
of the receiver can be detected and ambiguity out to 293 km can be removed. Aircraft
flight paths have been tracked to an accuracy of 2 cm over distances of tens of kilometers
using phase comparison DGPS techniques.
This explanation of kinematic differential GPS is oversimplified, because the satel-
lites are moving and measurements over a considerable time are required to resolve
ambiguity to the centimeter level. The P code can be used for real-time differential meas-
urements without knowledge of P code itself, because only a comparison of the time of
arrival of the code bits is required. Selective availability is not applied to the P code, so
differential measurements made with the P code cannot be affected by SA.
In the Wide Area Augmentation System (WAAS) developed by the FAA for aircraft
flying in North America, 24 WAAS receive stations continuously monitor their position
as calculated from the CA codes of all visible satellites in the GPS system. The stations
also use the P code transmissions to make accurate differential measurements of the
pseudorange to each visible satellite. The actual position of the WAAS stations is known
very accurately from prior survey data, so each WAAS station can calculate the error in
the pseudorange to each visible satellite. The 24 WAAS stations send their data to a central
station with an uplink to a GEO satellite. The central station validates the data, combines
all the information, and sends a sequence of pseudorange correction data to all GPS users
via the satellite. The central station also determines whether any of the data is in error,
and sends a warning signal called an integrity message to instruct aircraft not to use the
GPS system, or a particular satellite, because the data are not reliable. This is an essential
part of the FAA strategy for using GPS as the primary means of aircraft navigation. If the
aircraft is relying on GPS information alone to determine its position, that information
must have a very high reliability.
c12.qxd 10/08/02 10:31 Page 484
The WAAS GEO satellite transmits signals which are in a similar format to the L1
signals transmitted by a GPS satellite. A conventional GPS receiver with suitable software
can extract the pseudorange error values from the WAAS satellite transmission and obtain
markedly improved accuracy in its position determination. Thus no hardware changes are
needed to convert a GPS receiver to use WAAS data. The GEO satellite can also be used
to augment GPS satellites for position measurements, since it radiates the same signal for-
mat. The calculation of pseudorange error from the P code sequence, rather than (x, y, z)
position data error from the CA code, significantly increases the accuracy of the WAAS
DGPS system.
Eventually, it seems probable that local area augmentation systems (LAAS) using
differential GPS will be established at many airports to replace or augment existing ILS
precision approach systems. Advanced LAAS DGPS systems have been demonstrated to
achieve better than 1-m accuracy in three dimensions, with update rates sufficiently fast
to control a passenger aircraft. This is sufficient to allow DGPS position data to be cou-
pled to the aircraft autopilot so that blind landings can be made automatically in zero vis-
ibility conditions. Several demonstrations of autoland using DGPS were made in the late
1990s using Boeing 737 and 757 aircraft.
Aircraft used by overnight delivery companies will likely be fitted with GPS blind
landing systems first, since cargo aircraft are subject to fewer restrictions than passenger
aircraft and overnight delivery is subject to delays when airports are closed by low visi-
bility weather. Typically, a good autoland system fitted to a large aircraft can achieve more
consistent landings than a skilled pilot, so autoland may eventually become as common
for landings as autopilot use is for en route operation. Weather may eventually be less of
a factor in causing delays to passenger aircraft arrivals and departures.
12.12 SUMMARY
The Global Positioning System has revolutionized yields a unique solution for the position of the
navigation and become a consumer product with unknown point.
many applications. A system that was originally con- The apparent distance between each satellite and
ceived for targeting nuclear weapons has become a the receiver is determined by measuring the time-of-
worldwide commodity. Eventually GPS receivers will flight of a signal transmitted by the satellite, using
be present in all aircraft, ships, and automobiles so real-time clocks on the satellite and at the receiver.
that navigation will become a simple task, and no one Electromagnetic waves travel at 3 108 m/s, or 300 m
with a GPS receiver (and the ability to read a map) per microsecond. To achieve position location
will ever be lost. GPS receivers will be embedded in accuracy of a few meters, the time-of-flight of GPS
all cellular telephones so that the user will know his signals must be determined to one-tenth of a mi-
or her location, and an emergency call will always crosecond or better accuracy.
contain location information. GPS may eventually re- GPS satellites have four atomic clocks with
place all existing aircraft and ship radio navigation extreme accuracy, but most commercial GPS
systems and become the sole means of navigation, receivers use low-cost quartz crystal controlled clocks
permitting aircraft to select their own routes between that are not accurate to a microsecond. The signals
airports and to be landed automatically in zero visi- from a fourth GPS satellite are used to calculate the
bility conditions under the control of a local area dif- offset error in the clock of the GPS receiver, which
ferential GPS system. allows the receiver to track time with an accuracy bet-
The principles of GPS position location are sim- ter than one-tenth of a microsecond. Thus four GPS
ple, but its execution with the required accuracy is satellites are needed to make an accurate three-
complex. Position location with GPS is based on dimensional position measurement. The distances
trilateration; measurement of the distances from an between the GPS satellites and the receiver are known
unknown point to three known points (GPS satellites) as pseudoranges, because they may be in error by
c12.qxd 10/08/02 10:31 Page 485
PROBLEMS 485
thousands of kilometers. The receiver clock error data and also on the L2 carrier at a frequency of 1227.60
provided by the fourth psuedorange measurement MHz. Receivers that can use both L1 and L2 trans-
allows correction of pseudoranges to true ranges. The missions are able to remove some timing errors in-
satellites all transmit navigation messages that con- troduced by atmospheric and ionospheric delays.
tain the ephemeris for each GPS satellite; these data The accuracy achievable with a CA code GPS
are used by the GPS receiver to calculate satellite or- receiver is typically 30 m, but can be degraded to 200
bital positions. m by selective availability (SA), a DOD applied ef-
GPS satellites use direct sequence spread spec- fect that deliberately reduces the accuracy of the CA
trum techniques, transmitting two pseudorandom code position measurements. SA was introduced as
sequences known as CA and P codes at two fre- a security measure to prevent CA code receivers
quencies. Each satellite is assigned a unique set of from achieving very high accuracy, but was removed
codes which are used by the receiver to make time- on May 1, 2000. Differential GPS makes use of a
of-flight measurements and also to identify the fixed GPS receiver in a known location to calculate
satellites. Commercial GPS position measurement re- errors in the GPS position measurement, or errors in
ceivers use the course acquisition (CA) code trans- the pseudoranges to all visible satellites, which are
mitted at a chip rate of 1.023 Mbps by GPS satellites then sent over a radio link to other GPS receivers.
on the L1 carrier at a frequency of 1575.42 MHz. The The errors do not vary greatly over a wide area,
CA code is never changed. Authorized users, pri- allowing CA code GPS receivers to achieve far
marily military, use the P code which has a chip rate greater accuracy than is possible with a single re-
of 10.23 Mbps to provide greater positioning accu- ceiver. Differential GPS is the basis of the U.S.
racy. The P code is encrypted and changes once a FAA’s Wide Area Augmentation System (WAAS)
week. P code signals are transmitted on the L1 car- that will improve the accuracy of CA code receivers
rier, in phase quadrature with the CA code signal, in aircraft to 10 m.
REFERENCES
1. B. W. PARKINSON, and J. J. SPILKER, The Global Posi- 4. G. STRANG, and K. BORRE, Linear Algebra, Geodesy, and
tioning System-Theory and Applications, American In- GPS, Wellesley-Cambridge Press, Wellesley, MA, 1997.
stitute of Aeronautics and Astronautics, Washington, DC, 5. Aviation Week, Aerospace Source Book, p. 162, January
1996. 17, 2000.
2. E. D. KAPLAN, Understanding GPS Principles and Ap- 6. M. I. SKOLNIK, Introduction to Radar Systems, 2nd Ed.,
plications, Artech House, Norwood, MA, 1997. McGraw-Hill, New York, NY, 1981.
3. B. H. HOFFMAN-WELLWENHOF, H. LICHTENEGGER and 7. B. CLARKE, Aviation Application of GPS, McGraw-Hill,
J. COLLINS, GPS Theory and Practice, Springer Verlag, New York, NY, 1996.
New York, NY, 1992. 8. Aviation Week, January 28, 2002.
PROBLEMS
1. Find the exact altitude of a GPS satellite that has 3. An observer at the geographical North Pole has a
an orbital period equal to precisely one half of a side- GPS receiver. At an instant in time, four GPS satel-
real day. Use a value of mean earth radius re lites all have the same range from the observer, and
6378.14 km and a sidereal day length of 23 h 56 min the GPS receiver records a measured delay time for
4.1 s. the CA signal of 0.17097528 s for each satellite.
The four satellites’ coordinates are calculated to be
2. Find the maximum Doppler shift of the L1 signal (0, 13280.5, 23002.5), (0, 13280.5, 23002.5),
frequency for a GPS satellite at an altitude of (13280.5, 0, 23002.5), (13280.5, 0, 23002.5), where
20,200 km when the satellite has an elevation angle all distances are in km. Assuming an earth radius of
of 10°. 6378.0 km at the North Pole, so that the observer’s
Hint: Maximum Doppler shift occurs when the coordinates are (0, 0, 6378), determine the clock off-
observer is in the plane of the satellite orbit. Find the set error in the GPS receiver. [Use Eqs. (12.1) and
velocity of the satellite and the component of velocity (12.3), and take the velocity of light in free space to
toward the observer. be 2.99792458 108 m/s.]
c12.qxd 10/08/02 10:31 Page 486
4. Accurate position location using GPS requires pre- The corresponding measured delay times for the CA
cise knowledge of the speed of light. In most appli- code sequences from the satellites are:
cations, we use a velocity of light of 3.0 108 m/s. 1, 0.12102731 s; 2, 0.12102731 s; 3, 0.12102731 s;
Solve Problem 3 above and then recalculate the clock 4, 0.11738995 s.
offset using c 3 10 m/s instead of the more pre- Find the clock offset in the GPS receiver, and de-
cise value given in Problem 3. What is the error in the termine the radius of the earth at the South Pole. Use
clock offset? What is the difference in the ranges to a value for the velocity of light in free space c
the satellites when the approximate value for c 3 2.99792458 108 m/s, and work your solution to a
10 m/s is used? Discuss the corresponding position er- precision of 1 m. You will need to solve two simul-
ror due to the approximation. Why is it essential to taneous nonlinear equations from the set in Eq. (12.3)
use the exact value of the velocity of EM waves? in which the unknowns are the clock offset and the
5. A CA code GPS receiver is located at the geo- value of zp. Start with an estimated value zp 6378
graphic South Pole, coordinates (0, 0, zp). Four GPS km, and then solve the two simultaneous equations.
satellites are used to determine the radius of the earth This will give two unequal values for the clock off-
at the South Pole. At the instant of time that the meas- set. Use iteration of the value of zp to find the correct
urement is made, the satellites have the following values for clock offset and earth radius at the South
coordinates: Pole.
1, (0, 13280.500, 23002.500); 2, (0, 13280.500,
23002.500);
3, (13280.500, 0, 23002.500); 4, (0, 0, 26561.000).
appA.qxd 10/08/02 10:24 Page 487
APPENDIX A
DECIBELS IN COMMUNICATIONS ENGINEERING
Most readers of this book will be familiar with the practice of expressing power ratios in
decibels, abbreviated dB. The dB ratio A of two power levels p1 and p2 is given by
A 10 log10 a b dB
p1
(A.1)
p2
provided that p1 and p2 are expressed in the same units. Although the decibel is formally
defined only for a power ratio, p1 and p2 and A can also be expressed in terms of many
combinations of voltage, current, resistance, electric field strength, magnetic field strength,
and so on.
It is common practice in communications engineering to use decibels and the
mathematical properties of the logarithm to transform multiplicative equations to additive
equations, to manipulate the additive equations into particularly convenient forms, and to
define new logarithmic units with dB in their names for some of the quantities that appear.
When first presented, this practice is confusing to many people, and we hope to clarify it
here.
Consider the simple voltage divider circuit with resistors RS and RL shown in
Figure A.1. The rms voltages across the source and the load resistance are VS and
VL, respectively; the rms power supplied by the source is pS W and the rms power
delivered to the load is pL W. From elementary circuit theory, these quantities are
related by
V 2S
pS (A.2)
RS RL
V 2L
pL (A.3)
RL
VS RL
VL (A.4)
RL RS
pLRL
pL (A.5)
RL RS
letting
RL RS RT (A.6)
then
pS RL
pL (A.7)
RT
Equation (A.7) is a multiplicative equation, and pS and pL must have the same units.
Expressed another way, whatever units we substitute in for pS, pL will have the same units.
The reader should keep this in mind for what comes next.
487
appA.qxd 10/08/02 10:24 Page 488
488 APPENDIX A
RS
+ +
VS RL VL
− −
FIGURE A.1 Voltage divider

circuit used as illustration in
Appendix A.
Solving for the ratio pLpS and expressing the result in decibels, we have
10 log10 a b 10 log10 a b
pL RL
(A.8)
pS RT
Invoking the properties of logarithms we may rewrite this as
10 log10 1 pL 2 10 log10 1 pS 2 10 log10 a b

RL
(A.9)
RT
Without affecting the correctness of this equation or of its predecessor we may divide
pL and pS by 1 W. Expressed in the form of Eq. (A.9), the result is
10 log10 a b 10 log10 a b 10 log10 a b

pL pS RL
(A.10)
1W 1W RT
The first term above is the decibel ratio of pL to 1 W. This is defined as pL expressed
in units of decibels above 1 W, or pL in dBW. If we represent this quantity as PL
then
PL 1dBW2 10 log10 a b
pL
(A.11)
1W
Likewise the source power in dBW is PS where
PS 1dBW2 10 log10 a b
pS
(A.12)
1W
Substituting PL and PS into Eq. (A.10) yields
PL 1dBW2 PS 1dBW2 10 log10 a b

RL
(A.13)
RT
If we had expressed the powers in Eq. (A.9) in milliwatts and then divided both of
them by 1 mW, the only effect would have been to express PL and PS in decibels above
1 mW or dBm
PL 1dBm2 10 log10 a b
pL
(A.14)
1 mW
PS 1dBm2 10 log10 a b
pS
(A.15)
1 mW
appA.qxd 10/08/02 10:24 Page 489
DECIBELS IN COMMUNICATIONS ENGINEERING 489
and Eq. (A.9) would have become
PS 1dBm2 PL 1dBm2 10 log10 a b

RL
(A.16)
RT
Equations (A.13) and (A.16) differ only in the logarithmic power units that appear on their
left-hand sides. These equations would be true so long as PL and PS were both expressed
in the same logarithmic units. This happens because units that cancel by division in mul-
tiplicative equations like Eq. (A.7) cancel by addition or subtraction in additive decibel
equations like Eqs. (A.13) and (A.16). We can use this to write a general form for both
these equations:
PS PL 10 log10 a b dB
RL
(A.17)
RT
The “dB” after the equation means that the quantities involved must be expressed in a
consistent set of logarithmic units.
Equation (A.17) is typical of many of the equations used in this text that contain a
mixture of decibel and nondecibel units. We can change it to one involving only decibel
quantities if we divide both the resistances by 1 ohm () and transform the ratio on the
right-hand side to a difference
PS PL 10 log10 a b 10 log10 a b
RL RT
(A.18)
1 1
Now let us define a new unit for our own use called the dB for decibels above 1 ohm.
This is a dubious but expedient use of the term decibel! Thus
RL 1in dB2 10 log10 a b

RL
(A.19)
1
We can also express RT in dB by taking its log. Thus in these new units Eq. (A.16)
becomes
PL 1dBm2 PS 1dBm2 RL 1dB2 RT 1dB2 (A.20)
Likewise we could have expressed the resistance in kilohms (k) and then divided all of
the resistance terms by 1 k, inventing a new unit that we will call the dBk. The result
could have been either
PL 1dBm2 PS 1dBm2 RL 1dBk2 RT 1dBk2 (A.21)
or
PL 1dBW2 PS 1dBW2 RL 1dBk2 RT 1dBk2 (A.22)
The units in these two equations cancel by addition and subtraction rather than by
multiplication and division. Hence so long as both powers are in the one common deci-
bel unit and so long as both resistances are in another common decibel unit, we may write
a general form of these equations
PL PS RL RT dB (A.23)
A common alternative is to rearrange Eqs. (A.21) and (A.22) so that input quanti-
ties and output quantities are on opposite sides of the equation. For example, if the usual
problem is to find the load power, then Eq. (A.21) would most usefully be expressed
PL 1dBW2 PS 1dBW2 RL 1dB2 RT 1dB2 (A.24)
appA.qxd 10/08/02 10:24 Page 490
490 APPENDIX A
Readers seeing expressions like Eq. (A.24) for the first time object to its apparent
addition of dBW and dB, since they have learned by sad experience that quantities with
different units should not be added. But logarithmic units are different: quantities with
different logarithmic or decibel units are added; the test of correctness is whether or not
the units cancel by addition or subtraction. Thus in Eq. (A.24) the dB units cancel in
the subtraction of RT in dB from RL in dB and the dBWs cancel because they appear
on both sides of the equal sign.
The approach we have followed in this text is to use decibel units where the prac-
tice is common (principally for power) and to call the reader’s attention to other common
db units at the point of first introduction. Besides the dBW and the dBm, a common power
unit is the dBp, which is power expressed in dB above 1 picowatt.
We must emphasize again that decibel units of resistance are irregular and that we
have introduced them here only as a teaching tool.
appB.qxd 10/08/02 10:25 Page 491
APPENDIX B
FDM/FM/FDMA ANALOG TELEPHONE

TRANSMISSION
While digital modulation has many inherent advantages over analog frequency modula-
tion (FM) for telephone signals, much of the early investment in the Intelsat network was
for FDM/FM telephone systems, and these are still widely used outside the United States
where telephone systems are less well developed and the transition to an all digital net-
work has not yet occurred. This Appendix will discuss analog multiplex FDM/FM system
design and characteristics in detail.
Satellite FDM/FM analog telephone links resemble the terrestrial microwave point-
to-point links that carried most long-distance telephone traffic between 1950 and 1985.
Figure B.1 sketches a typical system. In it a multiplexer takes the baseband signals from
many individual telephone conversations, translates them to adjacent channels in the RF
spectrum, and combines them. Essentially, the multiplexer stacks the individual channels in
non overlapping spectral bands. The resulting composite FDM signal frequency modulates
an IF carrier (usually at 70 MHz) to create an FM (frequency modulation) multiplex signal.
The IF carrier is converted to the appropriate uplink frequency, amplified, and transmitted
to the satellite. At the satellite the signal is amplified, down-converted to the downlink band,
and retransmitted. At the receiving earth station the downlink signal is amplified and down-
converted to IF. The frequency modulated IF signal drives an FM demodulator, which
recovers multiplex signals with the voice channels stacked in frequency. Then a demultip-
lexer uses product detectors and filters to translate each channel back to baseband.
Baseband Voice Signals

A baseband voice signal is the voltage generated by an individual telephone set. While its
detailed characteristics depend on the speaker, the telephone system in the United States
assumes a flat spectrum extending from 300 to 3100 Hz. ITU-T recommends a baseband
bandwidth of 300 to 3400 Hz, but some designs assume a 0–3000 Hz spectrum. We will
assume the ITU-T baseband spectrum of 300 to 3400 Hz.
Schematically the spectrum of a baseband voice signal is often represented by the
triangle shown in Figure B.2; in a normal spectrum the peak of the triangle is to the
reader’s right and in an inverted spectrum (one in which the order of frequencies has been
reversed) the peak is to the reader’s left. The spectrum is not really triangular; this is just
a convenient symbol.
The amplitude of a voice signal in a communications link depends on where and how
it is measured. In telephone engineering practice, signal powers are expressed in terms of
transmission levels, their decibel levels with respect to a reference point. At the reference
point, the signal power in dBm is indicated by the unit dBm0; the 0 stands for the zero trans-
mission level point or the test point. Thus, a 2 dBm0 signal is one that produces an average
power of 2 dBm at the reference point. A suitable power meter placed at the 5 dB trans-
mission level point would measure the absolute power in the 2 dBm0 signal as 7 dBm.
491
appB.qxd 10/08/02 10:25 Page 492
f
70-MHz To
Incoming FM signal
voice Frequency upconverter
Multiplexer
channels Composite modulator Bandwidth = BIF and
FDM transmitter
signal
Each channel 0 to f max
has bandwidth
b Hz 70-MHz IF
carrier
(a )
70-MHz Outgoing
From FM signal voice
IF Demodulator Demultiplexer
channels
amplifier Composite
(C/N)i (S/N)wc
FDM
signal
(S/N)o
(b )
FIGURE B.1 Transmitting and receiving ends of a typical FDM system. (a) At uplink earth
station. (b) At downlink earth station.
Frequency
(a )
FIGURE B.2 Representa-

tions of the spectrum of a
telephone system baseband
Frequency
voice signal. (a) Normal spec-
(b) trum. (b) Inverted spectrum.
492
appB.qxd 10/08/02 10:25 Page 493
FDM/FM/FDMA ANALOG TELEPHONE TRANSMISSION 493
When telephone engineering began, the test point was accessible and meters could
be connected to it. The Bell System standardized the transmission level at the outgoing
side of the toll transmission switch as 2 dB below the test level point or 2 dBm0 (2
dB with respect to the zero test level point). With later switchboards, the test point lost
its accessibility and disappeared. But the 2 dB standard transmission level at the toll
transmission switch remained, and transmission levels are defined from this reference
exactly as if the zero test level point still existed.
Under Bell System standards that prevailed for many years, the long-term aver-
age power carried by a single voice channel in a telephone system was taken to be 18
dBm0 (ITU-R assumes 15 dBm0). The peak instantaneous power in the channel
is about 18 dB higher or 0 dBm0. Thus, telephone equipment is often adjusted by
applying a 1-kHz tone at 0 dBm0 to the system to simulate peak power on one
channel. This is called the test tone. We will return to it later in our discussion of
multiplexing.
The reader should be aware that the original Bell and ITU-R values of 18 dBm0
and 15 dBm0 for the average power level in a single telephone channel are very con-
servative and that many carriers use other values in some applications. In general, the
number of voice channels that a transponder can carry varies inversely with the average
power level per channel1.
Voice Signal Multiplexing

The process of shifting analog voice channels in frequency and combining them for trans-
mission is called frequency division multiplexing (FDM). The procedure is hierarchical;
individual channels are combined into groups; the groups are combined into larger groups;
the larger groups are combined into still larger groups, and so on. The names of the groups
and their internal channel arrangements vary between administrations and countries. In
this section we will use terminology largely drawn from reference 4.
The first step in voice channel multiplexing is to combine the baseband signals into
a basic group (often called simply a group) extending from 60 to 108 kHz. The channels
are stacked one above the other at 4 kHz intervals. The stacking is done by double-sideband
suppressed-carrier (DSBSC) amplitude modulating each voice channel onto an appropri-
ate carrier, filtering out the upper sideband, and saving and summing the lower sidebands.
The result is a single-sideband suppressed-carrier (SSBSQ) signal. See reference 2 for a
detailed explanation of SSBSC and DSBSC techniques. Figure B.3 illustrates the process.
The carrier frequency in kHz of the nth channel is given by 112 4n; thus channel 1 is
at the top of the spectrum and channel 12 is at the bottom. Since each channel occupies
only 3.1 kHz, there are 0.9 kHz guard bands between channels. These prevent interfer-
ence and simplify the filtering process when the baseband signals are recovered at the re-
ceiver. Selecting the lower sideband in the modulation process inverts the spectra of the
channels, but they will be inverted again and put back in the right order at the receiver.
While a single basic group could be transmitted by itself, most satellite and terres-
trial microwave links carry significantly more channels. Terrestrial systems use a rigid hi-
erarchy of channel combinations extending from the 12 channel basic group through the
60 channel basic supergroup to 600 channel basic mastergroups and beyond. The largest
named combination in the Intelsat hierarchy is the basic supergroup, five groups stacked
in a 240 kHz band. The stacking is done by SSBSC modulating the individual groups onto
appropriate carriers and summing the resulting lower sidebands. Figure B.4 illustrates gen-
eration of a 312 to 552 kHz supergroup. The carriers are spaced 48 kHz apart; that for
group 5 is highest at 612 kHz. The spectra of the individual groups are inverted when the
appB.qxd 10/08/02 10:26 Page 494
494 APPENDIX B
Incoming
channels 104–108
1 kHz
BPF
108 kHz
Multiplexed
100–104 signal
2 kHz +
BPF
• •
104 kHz
• •
• •
60–64
12 kHz
BPF
64 kHz
(a )
12 11 10 2 1
60 64 68 72 100 104 108

Frequency (kHz)
(b)
FIGURE B.3 Multiplexing 12 telephone channels to form a basic group. (a) The basic
hardware. (b) The spectrum of the multiplex signal showing individual channel spectra.
basic supergroup is formed. But the individual channel spectra in the groups are them-
selves inverted; the second inversion puts them back in the original order and the indi-
vidual channel signals in the supergroup are frequency-shifted versions of the original
baseband signals. Supergroups are normally separated by 12 to 14 kHz guard bands.
Since transponder bandwidth in a satellite system is usually limited, the arrange-
ment of channels in the Intelsat system follows a more flexible format than that used by
terrestrial microwave links, In INTELSAT V, for example, earth stations are allowed 12,
24, 36, 48, 60, 72, 96, 132, 192, 252, 312, 432, 492, 552, 612, 792, 972, 1092, 1332, or
1872 voice channels. All of these numbers are divisible by 12 (the number of channels in
a basic group), but the channel arrangement varies with the number of channels in order
to make efficient use of the transponder. For example, 132 channels are multiplexed by
combining a basic group in the 12 to 60 kHz band with one basic supergroup from 60 to
300 kHz and a second basic supergroup from 312 to 552 kHz. Table 5.13 summarizes the
appB.qxd 10/08/02 10:26 Page 495
Incoming 12-channel
basic groups
312–360
1 kHz
BPF
420 kHz
360–408
2 kHz
BPF
468 kHz
Outgoing
408–456 supergroup
3 kHz +
BPF
• •
516 kHz
• •
• •
504–552
5 kHz
BPF
612 kHz
FIGURE B.4 Schematic hardware for multiplexing five 12-channel basic groups to form a
basic supergroup.
channel combinations available on the INTELSAT IV through VI spacecraft and lists some
of the associated system specifications. The baseband spectrum between 0 and 12 kHz is
reserved for an intrasystem channel called the order wire, which carries housekeeping in-
formation, and for the energy-dispersal signal, if used.
The amplitude of a multiplexed telephone signal is a random function of time whose
characteristics depend on N, the number of channels. For N greater than or equal to 24,
the signal amplitude is usually represented by a Gaussian probability distribution with
zero mean and rms value a. The probability p(v) that the instantaneous voltage has the
value v is given by
1
ev 2s
2 2
p1v2 (B.1)
s12p
For design purposes, the rms value is usually taken as that which will result in 1 mW
(0 dBm) total power at the impedance level of the system.
Usually the amplitude of a multiplexed voice signal with N 24 is hard limited to
lie between 3.16 times the rms value. This causes clipping for something less than 0.2%
of the time and introduces negligible distortion. Thus it is common practice to equate the
peak value of the signal to 3.16 times the rms values. For N 24 the multiplexed spec-
trum is more “peaky” and a much higher peak-to-rms ratio may be used.
appB.qxd 10/08/02 10:26 Page 496
496 APPENDIX B
Frequency Modulation with Multiplexed

Telephone Signals
The signal-to-noise ratio described by Eq. (5.9) in Chapter 5 exists at the output of the
FM demodulator and describes the ratio of the total power in the multiplexed telephone
channels to the total thermal noise power. Let us now consider the FM detector output
signal-to-noise ratio for a single telephone channel located at the high-frequency end of
a multiplex signal. This is the (SN) at the output of the demultiplexer. If the channel
bandwidth is b Hz, then the noise power output of interest is that between fmax b and
fmax. Hence the term f2 – f1 in Eq. (5.10) becomes fmax ( fmax – b).
Since fmax W b, ( fmax b)3 f max
3
(1 3bfmax) f max
3
3bf 2max and Eq. (5.10)
becomes
3BIF 1¢frms 2 2 BIF 1¢frms 2 2
1SN2 wc 1C N2 0 c d 1C N2 0 c dc 2 d
13bf 2max 2
(B.3)
b f max
In decibel form, Eq. (B.3) is
1S N2 wc 1C N2 0 10 log10 1BIFb2 20 log10 1¢frms fmax 2 (B.4)
where the subscript wc means worst channel. See Figure B.5. Note that b is the band-
width of one baseband voice channel (nominally 3100 Hz), and (CN)0 is the overall
carrier-to-noise ratio of the link.
For the narrowband voice channel indicated by the shaded region, the noise power
N may be calculated by multiplying the noise spectral density at fm by b. This leads to
Eq. (B.3). For wideband channels the noise power must be calculated by integrating
from f1 to fm.
Equation (B.3) describes the ratio of signal power to thermal noise power in a tele-
phone channel at the upper end of a multiplexed baseband signal. But the frequency response
of neither the human ear nor a telephone receiver is flat, and a telephone listener will respond
differently to noise in different parts of the audio spectrum. Some of the noise that is pres-
ent in bandwidth b Hz will be unnoticed, and the effective signal-to-noise ratio will be higher
than that given by Eq. (B.3) by a weighting factor. Its value depends on the frequency response
of the telephone receiver and of the user’s ear. ITU-T and common satellite practice use
Power spectral density of noise at

demodulator output
Proportional to f 2
Frequency, f
f1 fm − b fm f2
FIGURE B.5 Noise power spectral density at the output of an FM demodulator.

appB.qxd 10/08/02 10:26 Page 497
psophometric weighting. We will adopt the latter and use the symbol p for the psophometric
weighting factor. The numerical value of p is 1.78; this corresponds to 2.5 dB4.
Noise at the high-frequency end of the input spectrum to an FM detector is
demodulated with greater output than noise at the low end, as discussed in Chapter 5.
The rising noise at the detector output can be suppressed using a de-emphasis filter with
the proper characteristics in the receiver, and the complimentary pre-emphasis filter in
the transmitter. The pre-emphasis and de-emphasis filters have inverse characteristics for
frequencies above a specified frequency f1. In practice the characteristics shown in
Figure 5.1 in Chapter 5 need to be maintained only up to the highest baseband frequency
present.
The use of pre-emphasis in the transmitter and de-emphasis in the receiver improves
the overall (SN) of the demodulated signal. The degree of improvement depends on the
filters and the modulating waveforms used. In ITU-R standard satellite and microwave links,
pre-emphasis improves the output signal-to-noise ratio of a telephone system by a factor of
2.5 (4 dB) over that given by Eq. (B.3)5. Other values apply to SCPC and TV transmission.
Generally the more nonuniform the modulation spectral density, the larger the
pre-emphasis improvement. Reference 4 provides several circuits for pre-emphasis and
de-emphasis filters.
Pre-emphasis improvement and the improvement due to psophometric weighting are
independent of each other; hence the right side of Eq. (B.3) may be multiplied by p and
by w to yield the psophometrically weighted signal-to-noise ratio on the worst multiplexed
telephone channel at the output of an FM link using pre-emphasis and having an overall
carrier-to-noise ratio of (CN)0.
BIF 1 ¢frms 2 2
1SN2 wc 1CN2 0 c dc 2 d pw (B.5)
b f max
In decibel form, Eq. (B.5) is
1SN2 wc 1C N2 0 10 log10 1BIFb2 20 log10 1¢frms fmax 2 P W (B.6)
where P is 2.5 dB and W is 4 dB.
Bandwidth Calculation for FDM/FM

Telephone Signals
We derived Eq. (B.6) and its predecessors assuming a sinusoidally modulated waveform
with an rms frequency deviation frms and requiring a transmission or IF bandwidth BRF
Hz. In this section we will relate these quantities to the number of channels N carried by
a multiplexed telephone signal and to the available transponder bandwidth.
For link performance calculations, the rms frequency deviation frms that should be
used is the rms test-tone deviation. This is the rms carrier deviation that a single 1kHz
0dBm sine wave called the test tone would produce when supplied to the modulator in-
put, and it represents a standardized test signal in one telephone channel. Putting this an-
other way, the transmitter is designed and adjusted to produce this rms carrier frequency
deviation when the modulator input signal is a standard 1-kHz 0-dBm test tone. The rms
test-tone deviation is related to the rms deviation that a multiplexed telephone signal will
cause by the loading factor, l. For N voice channels, l is given by reference 5 as
20 log10 l L 15 10 log10 1N2 for N 7 240 (B.7)
20 log10 l L 1 4 log10 1N2 for 12
N
240 (B.8)
The product frms is called the rms multicarrier deviation.
appB.qxd 10/08/02 10:26 Page 498
498 APPENDIX B
The ratio of the peak frequency deviation fpk to the rms multicarrier deviation
l frms is given by the peak factor, g. For a large number of channels (typically N
24), g is taken as 3.16 (corresponding to 10 dB) and for small numbers of channels (typ-
ically N 24), a value of 6.5 (18.8 dB) may be used5. If necessary, the incoming voice
signals may be amplitude limited to force a true peak-to-rms ratio of 3.16 on the multi-
plex signal. Thus for an analog FDM/FM telephone link
¢fpk lg¢frms (B.9)
The use of a high peak-to-mean ratio results in a low rms frequency deviation and consequently
a low (SN) ratio improvement factor for an average baseband level. Amplitude limiting will
cause distortion on large signal peaks, but it allows a higher average frequency deviation and
a greater (SN) ratio improvement. In SCPC systems where the peak-to-mean ratio of the
single telephone signal is large, companding is often used to reduce the peak-to-mean ratio
of the signal before it is applied to an FM modulator.
The 3.16 factor used in calculating the peak deviation of the carrier frequency repre-
sents the 0.1% extreme of a Gaussian distribution of signal voltages. For 0.2% of the time,
the signal voltage will exceed 3.16 times the rms value, assuming a Gaussian probability
distribution. Because we have to restrict the bandwidth of our RF signal to avoid interfer-
ence with adjacent channels, the FM modulator at the transmitter will have to be preceded
by a limiter that prevents large peaks of signal from overdeviating the carrier frequency. This
distorts the multiplexed signal, but for N 24 the effects of limiting are small.
The maximum modulating frequency, fmax, depends on the multiplexing scheme used,
that is, on the number of channels multiplexed and how they are organized into basic
groups, basic supergroups, and so on. When the standards of a satellite system are estab-
lished, fmax is tabulated for the allowed values of N. Table B.1 contains the fmax values for
INTELSAT IV through VI. If fmax is not known or if a new satellite link is being designed,
a good estimate to use for fmax in kHz is 4.2 N 5.
For a minimum required worst channel (SN)wc ratio, (typically about 50 dB) and
an overall (CN)0 fixed by the link power budget, a satellite systems engineer may trade
off values of N, BRF, and fpk. The number of channels, N, determines BRF through Carson’s
rule by
BRF 2 1lg¢ frms fmax 2 (B.10)
where l and fmax depend on N. The minimum (SN)wc and BRF in turn determine the required
value of ( frmsfmax) in Eq. (B.5). A satisfactory solution requires that BRF not exceed the
allocated transponder bandwidth and that the rms test-tone deviation frms be achievable
by the modulator. After discussing minimum (SN) ratio requirements, we will present
an example calculation that illustrates the interdependence of all the variables involved.
Telephone Performance Specifications

While U.S. engineers tend to think of system performance requirements in terms of deci-
bel signal-to-noise ratios, international practice often expresses system specifications in
terms of absolute channel noise levels measured in picowatts (psophometrically weighted),
abbreviated pWp, or in dB above a 1-pWp reference level, abbreviated dBp. (Unfortu-
nately the dBp abbreviation is used both for weighted and for unweighted picowatts.)
Picowatts are particularly useful when noise power contributions from several sources
must be combined. Decibel powers cannot be added directly.
To convert between picowatts and dBp and milliwatts and dBm, it is necessary first
to remember that a psophometric weighting filter reduces the power level of white noise
appB.qxd 10/08/02 10:26 Page 499
by 2.5 dB and second that 0 dBp (unweighted) corresponds to 90 dBm (unweighted).
If P is an absolute power level to be expressed in different units, then
P in dBp 1unweighted2 10 log10 1P in pWp2 2.5 (B.11)
and
P in dBm 1unweighted2 P in dBp 90 (B.12)
Assuming a standard 0-dBm signal level, then
1SN2 unweighted P1dBm2 87.5 10 log10 1P in pWp2 (B.13)
1SN2 weighted 90 10 log10 1P in pWp2 (B.14)
Thus a 7500-pWp channel noise level corresponds to a weighted (SN) of 51.25 dB
and an unweighted (SN) of 48.75 dB. Typical satellite link designs allow 7500 to 10,000
pWp total thermal noise for the space segment (up and down links including the inter-
modulation noise added at the spacecraft). The Intelsat specification for the INTELSAT
IV, IV-A, and V space segments is 8000 pWp 3.
Practical Examples
In this section we will apply the equations that describe FDM/FM analog telephone trans-
mission to several examples involving the INTELSAT V spacecraft. The numbers
describing the satellite are taken from Table B.1.
EXAMPLE B.1
An INTELSAT V transponder using a global beam achieves a (CN)0 ratio of 17.8 dB in clear air
at an earth station. The transponder carries 972 channels on a single carrier; the FDM/FM signal
fully occupies a 36-MHz bandwidth in the transponder. If the weighted (SN) on the top baseband
channel is 51.0 dB, find the rms test-tone deviation and the rms multicarrier deviation that must be
used. Compare these with the tabulated values. If the weighted (SN) on the top baseband channel
is 51.0 dB, find the rms test-tone deviation and the rms multicarrier deviation that must be used.
Compare these with the tabulated values.
First we will illustrate the procedure to follow if the multiplexing scheme is not known. Es-
timating fmax as 4200 N 4.082 MHz and using standard pre-emphasis and psophometric weight-
ing values of 4.0 and 2.5 dB, substituting into Eq. (B.6) we have
36 106
51.0 17.8 10 log10 a b 20 log10 a b 6.5
¢frms
3.1 103
4.082 106
Solving for frms
51.0 17.8 40.6 6.5 13.9 20 log10 a b

¢frms
4.082 106
Hence frms 778 kHz is the rms test-tone deviation.
Under the loading rule of Eq. (B.7), L 15 10 log10 (972) 14.88 and l
10(14.8820) 5.55. Thus the rms multicarrier deviation is frms 5.55 778 kHz 4.32 MHz.
We may check this answer by computing the Carson’s rule bandwidth BRF.
BRF 2 (3.16 4.32 MHz 4.082 MHz) 35.5 MHz, which is close to the 36 MHz allowed
by the transponder bandwidth and the original design.
These are slightly different from the published values because the true value of fmax (deter-
mined by the multiplexing hierarchy) is 4.028 MHz. Using this value of fmax we find frms to be
813 kHz and the occupied bandwidth is 36.6 MHz. The published value of frms is 802 kHz; this
leads to an occupied bandwidth of 36.2 MHz and a weighted (SN) of 50.9 dB.
500
TABLE B.1(a) INTELSAT IV-A, V, V-A, and VI Transmission Parameters (High-Density FDM/FM Carriers)
appB.qxd 10/08/02 10:26 Page 500
Carrier-to-total
noise temperature
Deviation (rms) ratio at
Carrier capacity Top baseband Allocated satellite Occupied for 0-dBm0 Multichannel operating point Carrier-to-noise
(number of frequency BW unit bandwidth test tone rms deviation (8000 200 pW 0p) ratio in occupied
channels) (kHz) (MHz) (MHz) (kHz) (kHz) (dBWK) BW (dB)
n fm bs b0 fr fmc (CT) (CN)
72 300 2.5 2.25 125 261 141.7 23.4

192 804 5.0 4.5 180 459 136.3 25.8
252 1052 7.5 6.75 260 733 137.1 23.2
312 1300 10.0 9.0 320 1005 137.1 22.0
492 2044 15.0 13.5 377 1488 134.4 22.9
612 2540 20.0 17.8 454 1996 134.2 21.9
792 3284 20.0 18.0 356 1784 129.9 26.2
792 3284 25.0 22.4 499 2494 132.8 22.3
972 4028 25.0 22.5 410 2274 129.4 25.7
1332 5884 36.0 36.0 591 3834 129.3 23.8
Source: (Reprinted with permission of the International Telecommunications Satellite Organization from Standard A Performance Characteristics of Earth Stations in
the INTELSAT IV, IV-A, and V Systems Having a G T of 40.7 dB/K (BG-28-72E Rev. 1), Intelsat, Washington, DC, December 15, 1982.)
TABLE B.1(b) INTELSAT IV-A, V, V-A, and VI Transmission Parameters (Regular FDM/FM Carriers)
Carrier-to-total
noise temperature
appB.qxd 10/08/02 10:26 Page 501
12 60 1.25 1.125 109 159 154.7 13.4

24 108 2.5 2.00 164 275 153.0 12.7
36 156 2.5 2.25 168 307 150.0 15.1
48 204 2.5 2.25 151 292 146.7 18.4
60 252 2.5 2.25 136 276 144.0 21.1
60 252 5.0 4.0 270 546 149.9 12.7
72 300 5.0 4.5 294 616 149.1 13.0
96 408 5.0 4.5 263 584 145.5 16.6
132 552 5.0 4.4 223 529 141.4 20.7
96 408 7.5 5.9 360 799 148.2 12.7
132 552 7.5 6.75 376 891 145.9 14.4
192 804 7.5 6.4 297 758 140.6 19.9
132 552 10.0 7.5 430 1020 147.1 12.7
192 804 10.0 9.0 457 1167 144.4 14.7
252 1052 10.0 8.5 358 1009 139.9 19.4
252 1052 15.0 12.4 577 1627 144.1 13.6
312 1300 15.0 13.5 546 1716 141.7 15.6
372 1548 15.0 13.5 480 1645 138.9 18.4
432 1796 15.0 13.0 401 1479 136.2 21.2
432 1796 17.5 15.75 517 1919 138.5 18.2
(continued )
501
502
TABLE B.1(b) (continued )
appB.qxd 10/08/02 10:26 Page 502
Carrier-to-total
noise temperature
432 1796 20.0 18.0 616 2279 139.9 16.1

492 2044 20.0 18.0 558 2200 137.8 18.2
552 2292 20.0 18.0 508 2121 136.0 20.0
432 1796 25.0 20.7 729 2688 141.4 14.1
492 2044 25.0 22.5 738 2911 140.3 14.8
552 2292 25.0 22.5 678 2833 138.5 16.6
612 2540 25.0 22.5 626 2755 136.9 18.1
792 3284 36.0 32.4 816 4085 137.0 16.5
972 4028 36.0 32.4 694 3849 133.8 19.7
972 4028 36.0 36.0 802 4417 135.2 17.8
1092 4892 36.0 36.0 701 4118 132.4 20.7
Source: (Reprinted with permission of the International Telecommunications Satellite Organization from Standard A Performance Characteristics of Earth Stations in
the INTELSAT IV, IV-A, and V Systems Having a G/T of 40.7 dB/K (BG-28-72E Rev. 1), Intelsat, Washington, DC, December 15, 1982.)
appB.qxd 10/08/02 10:26 Page 503
REFERENCES 503
EXAMPLE B.2
A single carrier that will occupy (when modulated) 9 MHz of an INTELSAT V transponder can
produce a (CN)0 of 14.7 dB in clear air at a standard earth station using the satellite’s global beam.
Assuming an 8000-pWp space segment noise allocation, how many telephone channels can the
transponder carry?
This example illustrates the kind of analysis a systems engineer would perform to determine
telephone channel allocations for a proposed spacecraft. It requires an iterative solution.
First, by Eq. (B.14), 8000 pWp corresponds to a weighted (SN) of 51.0 dB. Substituting the
value for (CN)0 14.7 dB, the 3.1-kHz channel bandwidth, and an IF bandwidth equal to the
9-MHz occupied bandwidth into Eq. (B.6), we obtain
9 106
51.0 14.7 10 log10 a b 20 log10 a b 6.5
¢frms
3.1 10 3
fmax
Solving for ( frmsfmax)
51.0 14.7 34.6 6.5 4.8 20 log10 a b

¢frms
fmax
or
a b 0.57
¢frms
fmax
For N channels fmax in Hz is approximately 4200N. Putting all these numbers, a g value of 3.16, and
l for N 240 from Eq. (B.9) into Eq. (B.10), we obtain
9 106 2 510 114 log10 N220 3.16 0.57 4200N 4200N6
This reduces to
1071.43 N 31.61 1010.2 log10 N2 14
Substituting a few values and using trial and error to match N to the result from the above equa-
tion, we find that N 191.2 solves the equation. The tabulated value for INTELSAT V is 192. Solv-
ing this problem required a preliminary assumption that N 240. Suppose instead we had assumed
N 240 when getting l from Eq. (B.7). The equation to be solved for N then would have become
9 106 2 510 11510 log10 N220 3.16 0.57 4200N 4200N6
1071.43 N 30.320 10 10.5 log10 N 2 1 4
The solution to this equation is approximately N 195.6 and it violates the hypothesis that N 240.
At this point in the process the incorrect initial assumption becomes apparent.
REFERENCES
1. W. H. BRAUN and J. E. KEIGLER, “RCA Satellite Networks: Having a GT of 40.7 dBK, Publication (BG-28-72E
High Technology and Low User Cost,” Proceedings of the Rev. 1), Intelsat, Washington, DC, December 15, 1982.
IEEE, 72, 1483–1505 (November 1984). 4. Recommendations and Reports of the CCIR, 1978, Vol.
2. H. L. KRAUSS, C. W. BOSTIAN, and F. H. RAAB, Solid IV, International Telecommunication Union, Geneva,
State Radio Engineering, John Wiley & Sons, New York, Switzerland, 1978.
1980. 5. H. L. VAN TREES, Satellite Communications, IEEE Press,
3. Standard A Performance Characteristics of Earth New York, 1979.
Stations in the INTELSAT IV, IV-A, and V Systems
appC.qxd 28/08/02 19:41 Page 504
APPENDIX C
COMPLEMENTARY ERROR FUNCTION erfc(x)

AND Q FUNCTION Q(z)
Equivalence Formulas and Tables of Values

The complementary error function erfc(x) and the Q function Q(z) both give the area un-
der the tail of a Gaussian distribution. The parameters x and z define the lower limit of
integration of the Gaussian function, with an upper limit of infinity. The functions are im-
portant in digital communications because they define the probability that additive white
Gaussian noise with a normalized rms value of 1 volt exceeds a threshold set at x or z
volts, giving the probability of a symbol error due to noise (see Chapter 5 for details). A
useful approximation to erfc(x) for x 1.5 is1
exp 1u2 2
erfc1x2
1pu
where u x 12 and is the rms value of the Gaussian variable.
An approximation for Q(z) with 1 for z 3 is2
1
ez 2
2
Q1z2
12pz
The equivalence between erfc(x) and Q(z) is
erfc1x2 2Q1 12z2
Q1z2 12 erfc a b
x
12
REFERENCES
1. SIMON HAYKIN, Digital Communications, John Wiley & 2. LEON W. COUCH, Digital and Analog Communication
Sons, New York, 1988. Systems, Macmillan, New York, 1990.
504
appC.qxd 28/08/02 19:41 Page 505
TABLE OF Q FUNCTION Q(z) 505
Table of Q Function Q (z)
z Q(z) z Q(z)
0 0.5 5.0 2.872 E7

2.0 2.280 E2 5.1 1.701 E7
2.1 1.791 E2 5.2 9.981 E8
2.2 1.394 E2 5.3 5.799 E8
2.3 1.075 E2 5.4 3.372 E8
2.4 8.220 E3 5.5 1.902 E8
2.5 6.227 E3 5.6 1.073 E8
2.6 4.674 E3 5.7 6.000 E9
2.7 3.476 E3 5.8 3.320 E9
2.8 2.562 E3 5.9 1.820 E9
2.9 1.871 E3 6.0 9.979 E10
3.0 1.354 E3 6.1 5.310 E10
3.1 9.702 E4 6.2 2.827 E10
3.2 6.889 E4 6.3 1.490 E10
3.3 4.847 E4 6.4 7.778 E11
3.4 3.378 E4 6.5 4.021 E11
3.5 2.332 E4 6.6 2.058 E11
3.6 1.595 E4 6.7 1.043 E12
3.7 1.081 E4 6.8 5.236 E12
3.8 7.252 E5 6.9 2.603 E12
3.9 4.821 E5 7.0 1.281 E12
4.0 3.174 E5 7.1 6.244 E13
4.1 2.070 E5 7.2 3.014 E13
4.2 1.337 E5 7.3 1.440 E13
4.3 8.558 E6 7.4 6.816 E14
4.4 5.423 E6 7.5 3.194 E14
4.5 3.404 E6 7.6 1.482 E14
4.6 2.117 E6 7.7 6.810 E15
4.7 1.303 E6 7.8 3.098 E15
4.8 7.948 E7 7.9 2.396 E15
4.9 4.800 E7 8.0 6.226 E16
appC.qxd 28/08/02 19:41 Page 506
506 APPENDIX C
Table of Function erfc(x)
x erfc(x) x erfc(x)
0.05
0.0 0.94363
1.00000 2.0 5.167 E3
0.10 0.88754 2.1 3.267 E3
0.15 0.83200 2.2 2.029 E3
0.20 0.77730 2.3 1.237 E3
0.25 0.72367 2.4 7.408 E4
0.30 0.67137 2.5 4.357 E4
0.35 0.62062 2.6 2.515 E4
0.40 0.57161 2.7 1.426 E4
0.45 0.52452 2.8 7.932 E5
0.50 0.47950 2.9 4.331 E5
0.55 0.43668 3.0 2.321 E5
0.60 0.39614 3.1 1.220 E5
0.65 0.35797 3.2 6.297 E6
0.70 0.32220 3.3 3.187 E6
0.75 0.28884 3.4 1.583 E6
0.80 0.25790 3.5 7.713 E7
0.85 0.22933 3.6 3.687 E7
0.90 0.20309 3.7 1.729 E7
0.95 0.17911 3.8 7.951 E8
1.00 0.15730 3.9 3.587 E8
1.05 0.13776 4.0 1.587 E8
1.10 0.11979 4.1 6.889 E9
1.15 0.10388 4.2 2.932 E9
1.20 0.08969 4.3 1.224 E9
1.25 0.07710 4.4 5.012 E10
1.30 0.06599 4.5 2.013 E10
1.35 0.05624 4.6 7.925 E11
1.40 0.04771 4.7 3.060 E11
1.45 0.04030 4.8 1.159 E11
1.50 0.03389 4.9 4.303 E12
1.55 0.02838 5.0 1.567 E12
1.60 0.02363 5.1 5.596 E13
1.65 0.01962 5.2 1.959 E13
1.70 0.01621 5.3 6.727 E14
1.75 0.01333 5.4 2.265 E14
1.80 0.01091 5.5 7.476 E15
1.85 0.00889 5.6 2.420 E15
1.90 0.00721 5.7 7.680 E16
5.8 2.390 E16
1.95 0.00582 5.9 7.291 E17
appD.qxd 10/08/02 10:27 Page 507
APPENDIX D
THE SIMPLE ATTENUATION MODEL
The first edition of Satellite Communications included an attenuation model developed by

Warren Stutzman and Keith Dishman of Virginia Tech that has proved remarkably useful
and accurate over the years1. The model provides an estimate of the slant-path attenua-
tion that can be expected to occur for a given percentage of a year based on published
statistics for the occurrence of heavy rain. The simple attenuation model (SAM) has the
advantage over the ITU-R model in Chapter 8 that it is easy to find the attenuation that
is likely to be exceeded for any rain rate, slant-path elevation angle, and frequency between
2.9 and 180 GHz. It may be less accurate than the ITU-R model in a statistical sense, but
is nevertheless a very useful tool for the analysis of slant-path attenuation.
The model uses an empirically derived relationship for the specific attenuation A1
on a slant path with a given rain rate R mm/h
A1 aRb dB/km (D.1)
Specific attenuation is defined as the attenuation that is caused by rain over a distance of
1 km. The values of a and b are empirically determined constants established from the
analysis of slant-path propagation measurements of attenuation and rain rate taken over a
period of several years. The Satellite Communications Group at Virginia made slant-path
measurements with a series of experimental satellites in the 1970s and 1980s. It was from
these measurements and the work of others in the same field that the SAM was devel-
oped. The values of a and b are found from the following relationships
a 4.21 105 f 2.42 2.9 f 54 GHz (D.2)
4.09 102 f 0.669 54 f 180 GHz
b 1.41 f 0.0779 8.5 f 25 GHz (D.3)
2.63 f 0.2.72 25 f 164 GHz
Note that the values of f used in Eqs. (D.2) and (D.3) must be in GHz.
The attenuation on a slant path through rain depends on two factors: the specific at-
tenuation A1 given by Eq. (D.1) and the path length L through the rain. Hence the total
attenuation on the path, A dB, is given by
A A1L aRbL dB (D.4)
where L is in km. On terrestrial paths, L is usually known and evaluation of Eq. D.4 is
straightforward. On a slant path, the value of L is generally highly variable and unknown.
Raindrops take a surprisingly long time to fall from the upper part of a storm to the
ground, and usually do not fall vertically. A rain gauge located close to a receiving ter-
minal, or under the slant path, may measure a rain rate that differs considerably from the
rain rate in the slant path at high altitudes. A radar pointed along the slant path can re-
veal the intensity and extent of the rain, and has been used successfully to estimate slant
path attenuation2, but for statistical estimation purposes we must use an effective path
length, Leff.
507
appD.qxd 10/08/02 10:27 Page 508
508 APPENDIX D
The concept of a statistical equivalent path length is valid because rainfall rate and
slant-path attenuation show the same statistical behavior over a period of 1 year. Compari-
son of exceedance curves for rain rate at a receiving site and slant-path attenuation, as
illustrated in Figure D.3 and in Figure 8.14 in Chapter 8, for example, shows that the shapes
of the curves are very similar. We can define a path average rain rate Rp(t) such that

L1t2
aR1y,t2 dy aRp 1t2 L1t2 dB
b b
A1t2 (D.5)
0
where y is along the slant path and A(t) and L(t) are the attenuation along the path and
the path length, which are varying with time.
All the rain that passes through the propagation path should eventually reach the
ground. Hence the exceedance curve for the path average rain rate should be the same as
the exceedance curve for the ground (point) rain rate. Therefore the attenuation value A(P)
equaled or exceeded for P percent of the time should be proportional to the value of ground
rain rate R(P) equaled or exceeded for the same P percent of time. The proportionality
factor between A(P) and a[R(P)]b is the effective path length, Leff. Thus we can find the
statistical path attenuation for any rain rate as
A1P2 a3R1P2 4 bLeff dB (D.6)
The effective path length can be found from measured slant-path attenuation and rain
rate by identifying and tabulating corresponding values of A(P) and R(P) with P as a pa-
rameter. However, it is much more convenient to calculate the effective path length from an
effective rain height which can be expressed by an equation. Effective rain height is a fic-
titious altitude at which all rain suddenly ceases. In stratiform rain, with R 10 mm/h, the
SAM model calculates effective rain height He km and then sets the path length in rain as
He H0
L
sin 1EL2
km (D.7)
where H0 is the height of the earth station above mean sea level. This concept is illus-
trated in Figure D.1.
Temperature decreases with altitude in the lower atmosphere, eventually reaching
0oC, the zero degree isotherm, at a height Hi km. Above this level, liquid water is assumed
To satellite
He
El
H0
Earth Sea
station level
FIGURE D.1 Geometry for the calculation of rain attenuation on a slant path in the SAM
model. H0 is the earth station’s height above mean sea level and He is the rain height. The
total propagation path length in rain is L.
appD.qxd 10/08/02 10:27 Page 509
THE SIMPLE ATTENUATION MODEL 509
to be absent. Frozen water, in the form of snow or ice pellets, may exist above the zero de-
gree isotherm, but causes much less attenuation than rain and is ignored in the SAM model.
(In reality, water is often present above an altitude Hi in convective storms, carried upward
by strong updrafts.) Heavy rain is always associated with convective storms, which occur
primarily during the spring, summer, and fall months in temperate climates. The zero degree
isotherm height Hi depends on the latitude of the earth station and is estimated by
Hi 4.8 km 0 ¶e 0 30° (D.8)
7.8 0.1 0 ¶e 0 km 0 ¶e 0 7 30°
where e is the latitude of the earth station. Note that in winter, if it snows, the zero de-
gree isotherm is at ground level, so the value of Hi is a typical value for use in calculat-
ing slant-path attenuation, not a real value.
In the SAM model, there are two rain models which have different effective rain
heights. In stratiform rain, for which R 10 mm/h, the rain height is constant and equal
to Hi, and the effective path length is equal to L in Eq. (D.7). In convective rainstorms,
when R 10 mm/h, the effective rain height depends on the rain rate because strong
storms push rain higher into the atmosphere, lengthening the slant path. However, the rain
rate is not uniform with altitude, so the model creates an effective path length that may
be longer or shorter than the value found in Eq. (D.7).
Based on empirical data, the following expressions for effective rain height, He, were
derived
He Hi km R 10 mm/h
He Hi 10 log10 1R
102 km R 7 10 mm/h (D.9)
In convective rain, when R 10 mm/h, a modified value of effective path length must be used
1 exp c gbloge a b L cos 1El2 d

R
Leff µ ∂ km R 7 10 mm/h
10
(D.10)
gbloge a b cos 1El2
R
10
where the empirical constant 122.
The procedure for determining the expected attenuation exceeded on a given slant
path using the SAM model is simple. Here are the steps.
1. Determine the location of the earth station, its latitude ¶e, and the elevation angle
to the satellite, El. (See Chapter 2 for the calculation of El.)
2. Locate the rain zone of the earth station using Figure D.2 (reproduced from Figure
8.15 in Chapter 8).
3. Go to Table D.1 (reproduced from Table 8.2 in Chapter 8) and find the average rain
rates exceeded for the required percentages of a year.
4. Calculate the specific attenuation, A1, for the frequencies used by the satellite from
Eq. D.4 using the appropriate values of a and b from Eqs. (D.2) and (D.3).
5. Find the height of the zero degree isotherm, Hi, from Eq. (D.8) and the effective
rain height, He, from Eq. (D.9).
6. Calculate the effective path length, Leff, from Eq. (D.7) if R 10 mm/h, or from
Eq. (D.10) if R 10 mm/h.
7. Finally, calculate the average attenuation exceeded for the specified percentage of
time from Eq. (D.4): A A1 Leff dB.
appD.qxd 10/08/02 10:27 Page 510
510 APPENDIX D
165° 135° 105° 75° 45° 15°
A A
E
G
C C E
G
60° 60°
B C
B
Latitude
E F in
D K F degrees
B E K N
H
M
30° E M
30°
N
E
N
P
N
0° 0° Equator
P
N E
N
D C
E
30° 30°
K
E Latitude
D in
D degrees
S
A A
60° 60°
165° 135° 105° 75° 45° 15°
Longitude in Degrees W
FIGURE D.2 Rain climate zones for the Americas. (From Figure 1 of reference 19
of Chapter 8, © ITU. Reproduced with permission.)
Remember that the value of slant-path attenuation calculated with any statistical
model is an estimated value of attenuation that will be exceeded under the specified con-
ditions in an average year. There are no average years, so considerable fluctuation about
the estimated value should be expected, especially around the 0.01% of a year level, as
illustrated in Figure D.3. This is one of the factors that makes the design of satellite links
difficult: the models provide only estimates of slant-path attenuation, but designs have to
be based on these estimates with just a small margin added. In the first year of operation
of a given satellite link, rain attenuation might exceed the estimates by several decibels,
causing the link to be out of service for longer than the expected percentage of the year.
appD.qxd 10/08/02 10:27 Page 511
THE SIMPLE ATTENUATION MODEL 511
TABLE D.1 Rainfall Rate Intensities for the Rain Climatic Zones
Percentage
of Time (%) A B C D E F G H J K L M N P Q
10 0.1 0.5 0.7 2.1 0.6 1.7 3 2 8 1.5 2 4 5 12 24

0.3 0.8 2 2.8 4.5 2.4 4.5 7 4 13 4.2 7 11 15 34 49
0.1 2 3 5 8 6 8 12 10 20 12 15 22 35 65 72
0.03 5 6 9 13 12 15 20 18 28 23 33 40 65 105 96
0.01 8 12 15 19 22 28 30 32 35 42 60 63 95 145 115
0.003 14 21 26 29 41 54 45 55 45 70 105 95 140 200 142
0.001 22 32 42 42 70 78 65 83 55 100 150 120 180 250 170
Note: See Figure D.2 for the Rain climate zones.

Source: From Table 1 in Reference 19 of Chapter 8.
The owner of the link may believe that the design is incorrect, and complain that the per-
formance of the link is below specification. It takes several years of operation to determine
the average attenuation exceeded for a given percentage of a year, and thus to find out
whether the design is correct after all. Some owners are unwilling to wait that long, so
the designer of the link may be tempted to include a larger than necessary rain attenua-
tion margin to guarantee that the performance specification is met every year.
EXAMPLE D.1
The SAM model is used here to estimate the rain attenuation that will be exceeded for 0.01% of an
average year on a 45o elevation angle slant path from an earth station in Blacksburg, Virginia, to a
satellite operating at a frequency of 28 GHz. The latitude of the earth station is 37.229o N with an
elevation H0 0.640 km. From Figure D.2, the earth station is in ITU region K, and Table D.1
gives a rainfall rate of 42 mm/h exceeded for 0.01% of an average year.
From Eqs. (D.2) and (D.3), the calculated values of the coefficients a and b at a frequency
of 28 GHz are (independent of the rainfall rate)
a 4.21 105 f 2.42 0.134
b 2.63 f 0.272 1.062
The latitude of the earth station is 37.229, so Eq. (D.8) gives the average zero degree isotherm
height as
Hi 7.8 0.1 37.229 4.08 km
For 0.01% of the year the average rain rate exceeded is 42 mm/h, which is greater than 10 mm/h,
so Eq. (D.9) gives
He Hi log10 1R
102 4.08 0.62 4.70 km
The effective path length exceeded for 0.01% of an average year can then be found from Eqs. (D.7)
and (D.10)
He H0 4.70 0.64
L 5.74 km
sin 1EL2 sin 45°
1 exp c gbloge a b L cos 1EL2 d

R
Leff µ ∂
10
gbloge a b cos 1EL2

R
10
gbloge 1R
102 cos1EL2 1
22 1.062 loge 14.22 cos 45° 0.0490
Leff 11 e0.049 e 5.74 2
0.049 5.0 km
appD.qxd 10/08/02 10:27 Page 512
512 APPENDIX D
10.0000
11.7 GHz
19.04 GHz
28.56 GHz
1.0000
Percentage of time attenuation is exceeded
0.1000
0.0100
0.0010
0.0001
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Attenuation (dB)
FIGURE D.3 Cumulative attenuation distribution for the calender year 1978 as measured at
the authors’ earth station on satellite downlinks at three frequencies. The 11.7-GHz path was
at 33° elevation; the other two were at 45°.
The attenuation exceeded on the slant path is then given by Eq. (D.4)
A aRbLeff 0.134 421.062 5.0 35.5 dB
Comparison with measured data for the author’s earth station in the year 1978 in Figure D.3 shows the
measured value to be 32 dB. A variation of a few decibels can be expected when comparing attenuation
measured for a specific year to the statistical value.
The SAM model works well for higher rain rates and attenuation levels. At low rain rates it
tends to underestimate attenuation, partly because it does not take account of attenuation occurring
in clouds and above the zero degree isotherm.
REFERENCES
1. W. L. STUTZMAN and W. K. DISHMAN, “A Simple Model Transactions of Antennas and Propagation, AP-26,
for the Estimation of Rain-Induced Attenuation along 318–329, March 1978.
Earth–Space Paths at Millimeter Wavelengths,” Radio 3. Rec. ITU-R PN.837-1, Characteristics of Precipitation
Science, 17, 1465–1476, November–December 1982. for Propagation Modelling, 1994.
2. R. L. OLSEN, D. V. ROGERS, and D. B. HODGE, “The aRb
Relation in the Calculation of Rain Attenuation,” IEEE
glo.qxd 10/08/02 10:29 Page 513
GLOSSARY OF TERMS
AND ACRONYMS
Definitions of Common Packet AFC Automatic Frequency Control

Network Abbreviations AKM Apogee Kick Motor
The acronyms set out below have come from many ALOHA Shared random packet access channel
documents. However, a large number have been AM Amplitude Modulation
abstracted from two sources: (a) a collection put AM-PM (Unwanted conversion of) Amplitude
together by Dr. Gory Fairhurst and Dr. Tim Modulation to Phase Modulation
Spracklen, and presented at a spring 1996 VSAT AMI Alternate Mark Inversion coding (bipolar)
course given at the University of York, England; (in high-speed PL)
and (b) from references 3 and 4 in Chapter 9.
AOCS Attitude and Orbit Control System
10B2 10-MHz baseband LAN using thin coaxial
AOR Atlantic Ocean Region
cable (of IEEE 802)
ANSI American National Standards Institute
10B5 10-MHz baseband LAN using thick coaxial
cable (of IEEE 802) ARP Address Resolution Protocol (translated IP
1 HA) (cf. RARP)
10BF 10-MHz baseband LAN using fiber (with
ARQ Automatic Repeat reQuest (i.e., frame
repeaters) (of IEEE 802)
retransmission procedure)
10BT 10-MHz baseband LAN using twisted pair
ASI Adjacent Satellite Interference
cable (of IEEE 802)
ASCII American Standard for the Computer
802 Family of network standards specified by the
Interchange of Information
IEEE
ASIC Application Specific Integrated Circuit
802.3 A LAN standard based on Ethernet (see
802, 10BX) ASK Amplitude Shift Keying
802.11 A wireless LAN standard for the 2.4-GHz ATM Asynchronous Transfer Mode (of B-ISDN)
ISM band AUI Automatic Unit Interface (standard interface
A Address field (usually follows the flag at the to an LAN MAU)
frame header) AWGN Additive White Gaussian Noise
ABM Asynchronous Balanced Mode (a CO-DL Abort 8-bit sequence, which ends an invalid
protocol in HDLC) HDLC frame (cf. Flag, idle)
AC Alternating Current Az Azimuth
ACK (i) Sequence number indicating correct BBP Baseband Processing
reception in a CO protocol bps Bits Per Second (unit of clock, utilization,
ACK (ii) A flag in TCP indicating that an throughput); also bits/s
acknowledgment is present B Byte (group of 8 bits) (sometimes known as an
ACTS Advanced Communications Technology octet)
Satellite BCC Block Checksum Character in asynchronous
ADC, A/D Analog to Digital Converter (converts PL (cf. CRC)
analog signal to digital stream) BCH Bose–Chaudhuri–Hocquenghem (family of
ADF Automatic Direction Finder (radio error correcting codes)
navigation device) BER Bit Error Ratio (or Rate) in digital circuit
ADPCM Adaptive Differential Pulse Code (cf. SNR in an analog circuit)
Modulation BFN Beam Forming Network
513
glo.qxd 10/08/02 10:29 Page 514
514 GLOSSARY OF TERMS AND ACRONYMS
BISYNC An IBM protocol using one-by-one CI0 Carrier power to Interference noise power
frame acknowledgement (BSC) (per Hz) ratio
B-ISDN Broadband Integrated Services Digital CIM Carrier power to Intermodulation noise
Network power ratio
BPF Bandpass filter CN Carrier power to thermal Noise power ratio
BPSK Binary Phase Shift Keying CN0 Carrier power to thermal Noise power (per
BSC An IBM protocol using one-by-one frame Hz) ratio
acknowledgement (BISYNC) CPFSK Continuous Phase Frequency Shift Keying
BSS Broadcast Satellite Service CT Carrier power-to-noise Temperature ratio
BTC Block Turbo Codes CTC Convolutional Turbo Codes
BTR Bit Timing Recovery CSMA/CD Carrier Sense Multiple Access with
CA Course Acquisition (one of the codes Collision detection (LAN)
transmitted by GPS satellites) DA Demand Assignment (allocation of resource
CATV CAble Television (Originally Community for communications)
Antenna Television) DAC, DA Digital-to-Analog converter (converts
C & M Control and Monitoring digital stream to analog signal)
C band Radio frequency band 4–8 GHz DAMA Demand Assigned Multiple Access (for
CBTR Carrier and Burst Timing Recovery (used efficient use of resource)
in preamble of TDMA frames) DBS-TV Direct Broadcast Satellite Television
CCIR An international committee of the ITU on dBHz Decibel-Hertz [10 log10(bandwidth in Hz)]
radio (now called the ITU-R) dBK Decibel-Kelvins [10 log10(noise temperature
CCITT An international standards committee of in K)]
the ITU (now called the ITU-T) dBW Decibel-Watts [10 log10(power in W)]
CCMF Centralized Control and monitoring dBm Decibel-milliwatts [10 log10 (power in mW)]
Functions DC Direct Current
CD Compact Disc DC (Frequency) Down Converter
CDI Course Deviation Indicator (radio navigation DCE Data Circuit-terminating Equipment
display) (modem side of a PL)
CDC Control error Delay Channel DGPS Differential GPS
CDM Code Division Multiplexing (form of DISC Disconnect request frame in HDLC
spread spectrum multiplexing) (CO-DL) (cf. FIN in TCP)
CDMA Code Division Multiple Access (common DL Data Link layer (Layer 2 of OSI) (e.g.,
form of spread spectrum access) Ethernet, HDLC)
CL Connection-Less (datagram) (e.g., UI, LLC1, DM Disconnected Mode frame in HDLC (rejects
Ethernet, IP, UDP) a request)
CMC Convolve Multiply Convolve (a process DME Distance Measuring Equipment (radio
used in Acousto-Optical MCDs) navigation aid)
CO Connection-Oriented (e.g., HDLC ABM, TCP) DNS Domain Name Service (converts IP name
COTS Commercial-Off-The Shelf to/from IP address)
CPU Central Processor Unit (or microprocessor) DNTX Do Not Transmit
CR Carrier Recovery DOD U.S. Department Of Defense
CRC Cyclic Redundancy Check (check for bit DOP Dilution of Precision (GPS accuracy
errors) parameter)
CSC Common Signaling Channel DPLL Digital Phase Locked Loop clock recovery
CUG Closed User Group (common VSAT private for synchronous circuits
user groups) DSAP Destination Service Access Point (part of
CI Carrier power to Interference noise power IEEE LLC frame)
ratio DSBSC Double-SideBand Suppressed Carrier
glo.qxd 10/08/02 10:29 Page 515
GLOSSARY OF TERMS AND ACRONYMS 515
DS-CDMA Direct Sequence Code Division FDM-FM-FDMA Frequency Division

Multiple Access Multiplexed Frequency Modulation Frequency
DS-SS Direct Sequence Spread Spectrum (same Division Multiple Access
as DS-CDMA) FDMA Frequency Division Multiple Access
DS-n Hierachy for digital transmission rates FEC Forward Error Correction
DSP Digital Signal Processing FEP Front End Processor
DTE Data Terminal Equipment (computer side FET Field Effect Transistor
of a PL) FFSK Fast Frequency Shift Keying (similar
DTH Direct-To-Home (VSAT terminal in a modulation to MSK)
private home)
FIFO First-In First-Out (i.e., a queue or buffer)
DVB-S Digital Video Broadcast standard for
FIN TCP flag bit indicating the end of a TCP
Satellite systems
connection
Demux Demultiplexer (separates previously
FM Frequency Modulation
multiplexed circuits) (cf. Mux)
FPGA Field Programmable Gate Array
E1 European data rate and framing standard for
circuits at 2.048 Mbit/s FPLMTS Future Public Land Mobile
Telecommunications System
EEPROM Electrically Erasable Programmable
Read Only Memory FR Frame Relay (connection-oriented datagram
EIRP (e.i.r.p.) Equivalent Isotropically Radiated circuit)
Power (combines power and antenna gain) FRMR Frame Reject frame in CO-DL HDLC
El Elevation (indicates link error)
ELT Emergency Locator Transmitters FSK Frequency Shift Keying
ELV Expendable Launch Vehicle FSS Fixed Satellite Service
EM Electromagnetic FTP File Transfer Protocol (of IP stack) (also
EbN0 Energy per bit over thermal Noise power generic term)
(per Hz) ratio Flag (i) The 8-bit delimiter between HDLC
ENST Ecole Nationale Supérieure des frames (cf. SYN, abort, idle)
Telecommunications de Bretagne Flag (ii) A bit which indicates a binary state
EODL End Of Design Life (e.g., status flag, SYN, FIN)
EOF End Of Frame G universal gravitational constant (6.672 1011
Nm2/kg2)
EOML End Of Maneuvering Life
GaAsFET Gallium Arsenide Field Effect
ES Earth Station
Transistor (Low noise RF transistor)
ES End System (i.e., User’s computer) supporting
all OSI layers GEO Geostationary Earth Orbit (GSO)
ESA European Space Agency GES Gateway Earth Station
ETSI European Telecommunications GHz GigaHertz (units of 109 Hz)
Standardization Institute GMSK Gaussian Minimum Shift Keying (MSK
EU European Union using Gaussian shaped pulses)
E–W East–West (station-keeping maneuver) GLONASS GLObal NAvigation Satellite System
F Flag (cf. Flag) (Russian Federation equivalent of GPS)
FAW Frame Alignment Word GPS Global Positioning System
FCC Federal Communications Commission GSM Global System for Mobile communications
(ETSI standard)
FCS Frame Check Sequence (identical to CRC)
(cf. BCC) GSO Geostationary Satellite Orbit (GEO)
FDDI Fiber Distributed Data Interface (Ring GTO Geostationary Transfer Orbit
LAN operating at 100 Mbit/s) GT Gain-to-noise Temperature ratio (of a
FDM Frequency Division Multiplexing (a method receiving system)
of combining signals at different frequencies into HA Hardware Address in IP and Ethernet (PL
a single signal) address of an interface)
glo.qxd 10/08/02 10:29 Page 516
HBE Hub Baseband Equipment IP (i) Internetworking Protocol (CL network

HCI Hub Control interface protocol of IP suite)
HDTV High Definition Television IP (ii) Suite of Internet protocols using IP
network protocol
HDB3 High Density Bipolar 3 coding (in high-
speed PL) (cf. AMI) IPX Popular proprietary network protocol
(Novell) (cf. IP)
HDLC High Level Data Link Control (family of
DL protocols) IS Intermediate System (e.g., router, switch)
supporting up to OSI NL
HEO Highly Elliptical Orbit
ISI InterSymbol Interference
HF High Frequency
ISL InterSatellite Link
HP Horizontally Polarized
ISDN Integrated Services Digital Network (all
HPA High-Power Amplifier
digital public network)
HPC High-Power amplifier and (frequency)
ISO International Standards Organization
Converter
(standards committee of the UN)
Hub The central earth station of a VSAT STAR
ISS International Space Station
network
ITU International Telecommunications Union
I Information frame in HDLC (carries data in a
CO-DL) (cf. UI) ITU-R Radiocommunication sector of the ITU
(formerly CCIR)
I In-phase component of a signal
ITU-T Telecommunications Standardization
IC Integrated Circuit
Sector of the ITU (formerly CCITT)
ICBM InterContinental Ballistic Missile
JPEG Joint Picture Experts Group
ICO Intermediate Circular Orbit
JPL Jet Propulsion Lab
Idle Sequence of digital 1’s indicating no frames
kbps Kilobits/second (1000 bps) of clock,
in HDLC (cf. Flag, abort)
utilization, and throughput (also kbit/s)
IM InterModulation
kHz kiloHertz (units of 103 Hz)
IO Input or Output of a computer interface
kB KiloByte (1000 bytes) (210 bytes) (cf. B for
ICMP Internetworking Control and Monitoring Byte)
Protocol (of IP suite)
K band Radio frequency band 16–24 GHz
IDU Indoor Unit (of a VSAT earth station)
Ka band Radio frequency band 24–36 GHz
IEE Institution of Electrical Engineers (UK)
Ku band Radio frequency band 8–16 GHz
IEEE Institute of Electrical and Electronic
L band Radio frequency band 1–2 GHz
Engineers (USA)
LAAS Local Area Augmentation System
IF Intermediate Frequency (between baseband
(enhanced GPS)
and RF)
LAN Local Area Network (e.g., Ethernet)
IFL Interfacility Link (cable/fiber link between
(cf. WAN, MAN)
ODU and IDU)
LAP Link Access Protocol (defined by X.25)
IFRB International Frequency Registration
(see also LAPB)
Board
LAPB Link Access Protocol Balanced (defined by
IHL Internet Header length (length of the PCI in
X.25) (see also LAP)
an IP datagram)
LAPD Link Access protocol on the D (i.e., data)
ILS Instrument landing system
Channel
IM Intermodulation (two or more frequencies
LC Line Code (the digital stream after the AD
creating unwanted products)
used to drive the modulator)
IMC Instrument Meteorological Conditions (in
LEO Low Earth Orbit
the clouds)
LET Linear Energy Transfer
Inbound Channel establishing a connection from
a STAR VSAT toward the hub LHCP Left-Hand Circular Polarization
Inroute Channel establishing a connection from a LIE Line Interface Equipment
STAR VSAT toward the hub LIM Terrestrial Line Interface Module
glo.qxd 10/08/02 10:29 Page 517
LLC Logical Link Control (DL of IEEE stack) msb Most significant bit (leftmost bit of a byte or
(similar to HDLC) word)
LLC1 CL-DL protocol using UI frames based on MSK Minimum Shift Keying (form of FSK)
HDLC (IEEE 802 suite) MSM Microwave Switching Matrix (used for
LMDS Local Multipoint Distribution System interconnecting links at RF or IF)
LO Local Oscillator (see VCXO) MSS (i) Maximum Segment Size (largest allowed
LORAN LOng RAnge Navigation payload at TCP layer)
LNA Low Noise Amplifier (used at the front end MSS (ii) Mobile Satellite Service
of an earth station receiver) MTBF Mean Time Before Failure
LNB Low Noise Block (front end receiving MTU Maximum Transfer Unit (max. size of DL
equipment with LNA, VCXO, etc.) frame in IP) (cf. MSS)
LNC Low Noise amplifier and (frequency) Mux Multiplexer (combines circuits into a single
Converter bit stream) (cf. Demux)
LP Linearly Polarized NACK Negative ACK in a CO protocol
LPC Linear Predictive Coder (retransmission request) (e.g., REJ)
LPE Linear Predictive Encoding (used in speech NASA National Aeronautical and Space Agency
and video compression) (U.S. space agency)
LPF Low Pass Filter NBFM NarrowBand Frequency Modulation
LRE Low bit Rate Encoding NCC Network Control Center
lsb Least significant bit (rightmost bit of a byte NDB NonDirectional Beacon (radio navigation
or word) aid)
LT Line Termination NF Noise Figure
LU Logical Unit (of SNA) N-ISDN Narrowband ISDN based on circuits at
MAC Medium Access Control (lower part of DL 64 kbps (64 kbits/s)
in IEEE 802 stack) NIST National Institute of Science and
MAN Metropolitan Area Network (or regional Technology (formerly the U.S. Bureau of
network) (cf. LAN, WAN) Standards)
MAU Medium Attachment Unit (LAN transceiver NL Network Layer (layer 3 of the OSI stack) (see
electronics (PL) IP, X.25)
MBONE Multicast IP backbone network (used Non-GSO Non-Geostationary Satellite Orbit
for multimedia networking) NPSD Noise Power Spectral Density
MCD MultiCarrier Demodulator NRZ Non Return to Zero
MCDD MultiCarrier Demodulator and N–S North–South (station keeping maneuver)
Demultiplexer NRZ Non Return to Zero coding within the bit
MCDDD MultiCarrier Demodulation, period (1 high, 0 low) (PL)
Demultiplexing, and Decoding NT Network termination
MCPC Multiple Circuits Per Carrier (medium-to- NTSC National Television Standards Committee
high capacity link) (Established color TV standards in the United
MCS Master Control Station States)
MEO Medium Earth Orbit OBP OnBoard Processing (used on advanced
MF-TDMA MultiFrequency Time Division satellites to connect links, etc.)
Multiple Access OC-n Hierachy for fiber-optic bit rates
MHz MegaHertz (units of 106 Hz) ODU Outdoor Unit (of a VSAT terminal)
MMIC Microwave Monolithic Integrated Circuit OEM Original Equipment Manufacturer
MODEM Modulator/Demodulator converts OMT Orthomode Transducer (separates
digital signal from/to line code (cf. LC) modes/polarizations in antenna feeds)
MPEG Moving Picture coding Expert Group OQPSK Offset keyed QPSK (QPSK with a half
(created video bit rate compression) symbol time shift between I and Q signals)
glo.qxd 10/08/02 10:29 Page 518
OSI (i) Open Systems Interconnection (reference RA Random Assignment (also called ALOHA)
model) RARP Reverse Resolution Protocol (translates
OSI (ii) Suite of protocols defined using the OSI HA 1 IP address) (cf. ARP)
reference model (cf. IP) REJ REJect request frame (Go-Back-N ARQ in
Outbound Channel establishing a connection HDLC) (cf. SREJ)
from the hub to a STAR VSAT RF Radio Frequency
Outroute Channel establishing a connection from RG58U Type of flexible coaxial cable used for
the hub to a STAR VSAT IEEE 802.3 10B2 LANs
PABX Private Automatic Branch Exchange RHCP Right-Hand Circular Polarization
Packet Information block identified by a label at RLV Reusable Launch Vehicle
layer 3 of ISO-OSI stack RNR Receiver Not Ready HDLC supervisory
PAD Packet Assembler/Disassembler (interface frame (CO-DL) (positive ACK)
between packet and nonpacket networks) RRC Root Raised Cosine (Frequency response of
PAL Phase Alternate Line (color TV standard zero-ISI filter)
used in Europe and elsewhere) RS Reed–Solomon (block error detecting and
PBX Private Branch Exchange correcting code)
PCE Processing and Control Equipment (part of RS-232 V.24 serial interface using 25-pin D-
the hub BBP equipment) connector (cf. RS-449)
PCI Protocol Control Information (header) (of RS-449 Differential synchronous serial interface
OSI reference model) using 37-pin D-connector)
PCM Pulse Code Modulation RTT Round-Trip Time (time to receive a response
PCN Personal Communications Network from remote system)
PCS Personal Communications System RZ Return to Zero coding within the bit period
(i.e., 0 between bits) (in PL)
PDN Public Data Network
S band Radio frequency band 2–4 GHz
PDU Protocol Data Unit (of OSI) (e.g., a packet
at NL, or frame at DL) SA Selective Availability (applied to GPS signals
until May 2000)
PFD Power Flux Density (power per m2 at a
given plane) SABM Set Asynchronous Balanced Mode in
HDLC (modulo-8, CO-DL)
PISO Parallel-in Serial-out shift register used in
serial transmitter (PL) SABME SABM Extended frame in HDLC
(modulo-128, CO-DL)
ping Network testing application protocol
(ICMP/IP) SAM Simple Attenuation Model
PLMN Public Land Mobile Network SAP Service Access Point (of OSI) (e.g., port in
UDP or TCP)
PL Physical Layer (layer 1 of OSI stack) (e.g.,
RS-232, RS-449) SAR Specific Absorption Rate
SARSAT Search And Rescue SATellite
PLL Phase Locked Loop clock recovery at PL
(see also DPLL) SAW Surface Acoustic Wave
PN PseudoNoise SC Service Channel
POS Point Of Sale SC SpaceCraft (satellite)
POTS Plain Old Telephone Service SCADA Supervisory Control And Data
Acquisition
PSK Phase Shift Keying
SCC (i) Subnetwork Control Center
PSTN Public Switched Telephony Network
(ii) Satellite Control Center (controls satellite
PVC Permanent Virtual Circuit payload when in orbit)
Q Quadrature phase component of a signal SCPC Single Channel Per Carrier
QAM Quadrature Amplitude Modulation (QPSK (nonmultiplexed, low capacity FDMA)
and ASK combined) SCPC-FDMA-DA Single Channel Per Carrier
Q band Radio frequency band 36–46 GHz Frequency Division Multiple Access
QPSK Quadrature phase Shift Keying Demand Access
glo.qxd 10/08/02 10:29 Page 519
SCU Subnetwork Control Unit SYN (i) Synchronization character (8 bits)

SDH Synchronous Digital multiplexing Hierarchy delimiting a block (cf. Flag)
(European, see SONET) SYN (ii)TCP flag bit indicating the start of a TCP
SDLC Synchronous Data Link Control (An IBM connection)
version of HDLC (usually a polled link) allowing TCP Transmission Control Protocol (CO transport
multiple unacknowledged frames) protocol of IP suite)
SDU Service Data Unit (of OSI reference model) TCP/IP Transmission Control Protocol–Internet
(cf. PDU) Protocol
SER Symbol Error Rate TDD Time Division Duplexing
SES Société Européenne de Satellites TDM Time Division Multiplexing (sharing a link
in time)
SIPO Serial-In Parallel-Out shift register used in
serial receiver equipment TDM-SCPC-FDMA Time Division Multiplexing
Single Channel Per Carrier Frequency Division
SISO Soft Input–Soft Output (used in error
Multiple Access
correction decoders)
TDMA Time Division Multiple Access (sharing a
SLIP Serial Line IP (DL protocol using resource in time)
asynchronous communications)
TDRSS Tracking and Data Relay Satellite System
SNA Proprietary suite of protocols used by IBM
TE Terminal Equipment
(cf. IP)
TEC Total Electron Content
SNAP SubNetwork Access Protocol supports IP
over LLC (IEEE 802.3) telnet Remote log-in and terminal emulation
application protocol (TCP/IP)
SNR (SN) Signal-to-Noise Ratio (analog
quantity measured in dB) (cf. BER) THz TerraHertz (units of 1012 Hz)
TNC Terminal Node Controller
SMPT Simple Mail Transfer Protocol (TCP/IP
e-mail application) T1 U.S. data rate and framing standard for
circuits of 1.544 Mbps (Mbits/s)
SOHO Small Office/Home Office (relatively low
capacity user) T-n U.S. heirachy for digital transmission rates
SONET Synchronous Optical Network; U.S. ToS Type of Service in IP (indicates type of data
standard similar to SDH being transported)
SOTF Start Of Transmit Frame TR Token Ring (IEEE 802.5) LAN protocol using
a ring architecture
S-PCN Satellite Personal Communications
TTC Telemetry, Tracking, and Control
Network
TTC&M Telemetry, Tracking, Control, &
sps Symbols per second
Monitoring
SREJ Selective Reject request frame (in HDLC
TTL (i) Time To Live in IP (number of hops
CO-DL) (cf. REJ)
before packet is discarded)
SSAP Source Service Access Point (part of IEEE
TTL (ii) Transistor–Transistor Logic family of
LLC DL frame)
integrated circuits
SSB Single SideBand modulation (a narrowband TST Time domain, Space domain, Time domain
form of AM) switching
SSBSC Single-SideBand Suppressed-Carrier TTY TeleTYpe
modulation
TV TeleVision
SS-FDMA Satellite Switched FDMA
TVRO TeleVision Receive Only
SSMA Spread Spectrum Multiple Access (see TWTA Travelling Wave Tube Amplifier
also CDMA)
TX-PCE Transmit Processing and Control
SSPA Solid-State Power Amplifier Equipment
SS-TDMA Satellite Switched TDMA UA Unnumbered Acknowledgement frame in
SSTO Single Stage To Orbit CO-DL HDLC
STDM Statistical Time Division Multiplexing (cf. UART Universal Asynchronous Receiver
TDM or packet mode) Transmitter (device)
glo.qxd 10/08/02 10:29 Page 520
UC (Frequency) Up-Converter WWII World War II

UDP Universal Datagram Protocol (CL transport XPD Cross-Polarization Discrimination (a
protocol of IP suite) measure of signal polarization)
UHF Radio frequency band 300 MHz–1 GHz XPI Decibel ratio of wanted power to unwanted
UI Unnumbered Information CL frame in HDLC power
(cf. LLC1) X.25 Recommendation (of ITU-T) for packet-
ULPC Uplink Power Control (one switched communications (specifies interface
countermeasure against signal fade) (UPC) between DTE and DCE for terminals operating
in the packet mode and connecting PDNs by
UMTS Universal Mobile Telecommunications
dedicated circuit)
System
UNIX A common “open” operating system used
by many computers ITU-R Standards Applicable to
UPC Uplink Power Control (one countermeasure VSAT Systems
against signal fade) (ULPC) Existing Standards and Recommendations
UPT Universal Personal Telecommunications • Recommendation ITU-R S.725, Technical
USART Universal Synchronous/Asynchronous Characteristics for Very Small Aperture Ter-
Receiver Transmitter minals (VSATs), 1992. [Earth station RF
UT Universal Time (Equal to GMT) aspects]
UTP Unshielded Twisted Pair cabling (used in • Recommendation ITU-R S.726-1, Maximum
10BT LANs) Permissible Level of Spurious Emissions from
UW Unique Word Very Small Aperture Terminals (VSATs),
1992–1993. [Permitted emitted power in given
V band Radio frequency band 46–56 GHz bands]
VCO Voltage Controlled Oscillator (used with • Recommendation ITU-R S.727, Cross-Polar-
mixer to change freq.) ization Isolation from Very Small Aperture
VCXO Voltage Controlled Crystal Oscillator Terminals (VSATs), 1992. [Cross-polarized
VHF Radio frequency band 30–300 MHz power to be used in interference calculations]
VOIP Voice Over Internet Protocol • Recommendation ITU-R S.728, Maximum
VOR VHF Omnidirection Range beacon (radio Permissible Levels of Off-Axis E.I.R.P. Density
navigation beacon) from Very Small Aperture Terminals (VSATs),
1992. [Power levels away from the main
VOW Voice Order Wire (Station–Station voice
beam axis to be used in interference calcula-
link)
tions]
VP Vertically Polarized
• Recommendation ITU-R S.729, Control and
VSAT Very Small Aperture Terminal Monitoring Function of Very Small Aperture
VSB Vestigial Sideband (form of AM used in TV Terminals (VSATs), 1992. [Functions that the
broadcasting) VSAT should be able to control in its operation]
V.24 Recommendation (of ITU-T) for serial Draft Standards and Recommendations Being
communications (RS-232) Considered
VT Virginia Tech
• Draft New Recommendation, Connection of
WAAS Wide Area Augmentation System
VSAT Systems with Packet-Switched Public
(enhanced GPS)
Data Networks (PSDNs) Based on ITU-T Rec-
WAN Wide Area Network (e.g., the Internet) ommendation X.25, 1995. [ITU-R standard on
(cf. LAN, MAN) the interconnection of VSATs with Public
WARC World Administrative Radio Conference Switched Networks]
WBFM WideBand Frequency Modulation Standards and Recommendations under
WGS-84 World Geodetic System 1984 Preparation
WLL Wireless Local Loop • Connection of Private VSAT Networks to the
WWI World War I Public ISDN.
glo.qxd 10/08/02 10:29 Page 521
European Technical Standards • ETS 300 160 Control and Monitoring at a Very
(ETS) for VSAT Systems Small Aperture Terminal (VSAT).
• ETS 300 161 Centralized Control and Moni-
Existing Standards toring for VSAT Networks.
• ETS 300 157 Receive-Only Very Small Aperture • ETS 300 194 The Interconnection of VSAT Sys-
Terminals (VSATs) Used for Data Distribution tems to Packet Switched Public Data Networks
Operating in the 1112 GHz Frequency Bands. (PSPDNs).
• ETS 300 159 Transmit/Receive Very Small Aper-
ture Terminals (VSATs) Used for Data Commu- Physical Constants
nications Operating in the Fixed Satellite Service
a Radius of GEO (42,164.17 km)
(FSS) 111214 GHz Frequency Bands.
Kepler’s constant (3.986004418 105
• ETS 300 333 Receive-Only Very Small Aper- km3/s2)
ture Terminals (VSATs) Used for Data Distri- re Mean earth radius (6378.137 km)
bution Operating in the 4 GHz Band. Me Mass of the earth (5.98 1024 kg)
• ETS 300 332 Transmit/Receive Very Small Aper- G Universal Gravitational constant (6.672
ture Terminals (VSATs) Used for Data Commu- 1011 Nm2/kg2)
nications Operating in the Fixed Satellite Ser- k Boltzmann’s constant (1.39 1023 J/K
vice (FSS) 6 GHz and 4 GHz Frequency Bands. 228.6 dBW/K/Hz)
ind.qxd 28/08/02 19:33 Page 522
INDEX
A Agila 2, 11 beamwidth and gain, 115

A-law, 169 Allocations, frequency, 97 boresight, 101, 366
ACK see Acknowledgement American Mobile Satellite Cassegrain, 367
signal Corp., 7 cell structure, 410
AKM see Apogee Kick Motor Ameristar, 8 coma, 415
ALOHA, 355 Amos 1, 12 contour, 3 dB, 121
AMCS-1, 7 Amplifier DBS-TV, 442, 446
AMPS see analog mobile back-off, 77, 115, 124, 357 deployment, 86
phone service front end, 107 dual polarized, 445
AOCS see attitude and orbit IF, 107 edge of coverage loss, 117
control system noiseless, 105 efficiency, 82
ARQ see automatic quasilinear, 124 fD ratio, 415
repeat request RF, 107 feed, 107
ASCII code, 277 Analog front-fed, 367
AT&T, 425 FM transmission, gain, 80, 82, 101
ATM see asynchronous satellite, 164 gain, non-circular
transfer mode mobile phone service, 157 coverage, 117
ATS-6, 86, 309 signals, digital transmission global, receive horn, 58
AWGN see added white of, 201 global, transmit horn, 58
Gaussian noise TV transmission (from Gregorian, 367
Absolute bandwidth, 180 satellite), 115 hopping beam, 410
Absorption video, 439 horn, 80
atmospheric, 297 Angle horn, scalar, 366
resonant, 307 azimuth, 31 inflatable, 85
water vapor, 308 canting, 327 isotropic, 101
Adaptor, booster, 58 central, 33 main lobe, 414
Acceleration, 17 elevation, 31 mispointing loss, 115
Access Control Protocols, 349 minimum grazing, 419 multiple beam, 221, 248
Access, random, 254 phased array scan, 412 noise, 332
Accuracy tilt, 327 nulls, 80
CA code, 480, 481 Angular offset fed, 121, 367
GPS, 464 distance, 30 omnidirectional, 80
timing (GPS), 475, 476 momentum, 20 parabolic torus, 439
2DRMS, 464 velocity, 25 paraboloid, 80
AceS Asia Cellular, 9 Anomaly pattern, 80, 125
Acknowledgment signal, eccentric, 25 phased array, 411
286, 350 mean, 26 radiating elements, 411
Across plane ISL seam, 422 true, 25 reflector, 80
Active device, 411 Antenna satellite, 59, 80
Acquisition aperture, 81 satellite, gain, 432
sequential, 472 array, 80 scan loss, 416
Added white Gaussian noise, 191 axially symmetrical, 367 scanning beam, 410
Adjustment factor beam, hemi, 81 sidelobe level (mask,
height, 301 beam, spot, 75 ITU-R), 125
horizontal, 301 beam, zone, 75 Simulasat, 439
Afristar, 8 beamwidth, 82 smart, PCS, handset, 145
522
ind.qxd 28/08/02 19:33 Page 523
INDEX 523
solid angle, 101 Axially symmetric Bandpass

spot beam, 248, 412 antenna, 367 filter, 75
steerable phased array, 249 Axis transmission of digital
telemetry and command semimajor, 23 data, 179
bicones, 58 semiminor, 23 Bandwidth
wire, 80 Atmospheric absolute, 180
Aperture attenuation, 103 channel, 180
antenna, 80 drag, 17 FDM/FM, 497
(antenna) efficiency, 82 loss, 103 limited, 276
effective, 102 multipath, 310 noise, 115, 355
Apogee, 22, 23 Atmosphere occupied, 180, 355
Apogee kick motor, 44, neutral, 298 occupied, equation, 180
48, 63 standard, 307 Base, customer service, 420
Apollo, 394 Attenuation, Baseband
Arab satellite, 9 atmospheric, 103 digital signals, 172
Arabsat, 9 cloud, 297, 308 onboard processing
Ariane -5 -40 -42P -44P -42L differential, 327 transponder, 59
-44LP, 45 gaseous, 307 noise power, FM, 160
Ariane 44, 5 rain, 15, 116, 297, processor (VSAT), 368
Ariane 5, 46 507–510 SN ratio, FM signals, 159
Aries, first point of, 27 rain, at Ku-band, 147 transmission, digital
Arthur C. Clarke, 1, 4 rain, prediction, 317 signals, 172
ANIK IA, 4 scaling (frequency and Basic transmission
ANIK E1, E2, 12 elevation), 126, 325 theory, 100
Apparent orbital period, 391 specific, 297, 318 Bathtub curve (failure
Apstar-V, 6 total path, 122 rates), 89
Arc zenith, 307 Batteries, 71
coverage, 432 Attitude and orbit control Nickel-Hydrogen, 72
jets, 63 system, 57, 60 capacity, 72
Argument of perigee west, 28 sensor, 65 Baud (rate), 172, 235
Ascending node, 28 Automatic repeat request, Baudot, 172
Ascension, right, 27, 28 287, 350 Beamwidth
Asia Broadcasting and go-back-N, 287 1 dB, 37
Communications, 9 selective repeat, 287 3 dB, 82, 306
Asia satellite, 9 stop-and-wait, 287 Bearing and power transfer
Asiasat, 9 Availability assembly, 58
Asiastar, 8 link, 295 Beginning of life, 71
Astra DBS-TV satellites, 442 threshold, 296 Bell System, 3
IH, 5 Azimuth angle, 31 Bent pipe (transponder), 59, 78,
1A, 1B, etc., 11 113, 370
K, 14 B Big LEOs, 411
Astrolink, 248 BAPTA see bearing and power Binary phase shift keying, 180,
Asynchronous transfer mode transfer assembly 187, 362
(ATM), 256 BER see bit error rate Bit error rate, 100, 188, 296
Atlas II, IIAS, 45 BER vs. Eb No, 364 vs. CN, 196
Atlas IIAS, 425 BPSK see binary phase shift Bit
Atlas IIAS, IIIA, IIIB, V, 46 keying flip (radiation effect), 398
Atmospheric absorption, 297 BS-3N, 9 redundant (FEC), 187
Attenuation BTR see bit timing recovery timing recovery,
downlink, rain, 130 Back-off, output power, 77, 115, (TDMA), 362
rain, 15, 116, 297, 317, 124, 129, 357 Bipropellant (maneuvering
507–510 Background, galactic, 24 fuel), 63
rain margin, 120 Baikonur (Cosmodrome), 392 Black body radiation, 105
uplink, rain, 130 Bandlimited channel, 182 Block code, 280
ind.qxd 28/08/02 19:33 Page 524
524 INDEX
Blockage Cable TV, 439 Climate parameters (rainfall

path, L-band, 140 Canting angle, 327 rate), 314
PCS, body, 145 Capacity, channel, 275 Clock
PCS, trees, 145 Carrier recovery atomic, 467
Bluetooth, 257 BPSK, 190 bias, 465
Boltzmann’s constant, 115 TDMA, 362 offset, 464
Booster adapter, 58 Carrier (signal), 75 Cloud attenuation, 297, 308
BorealisTM, Ellipso orbit, 426 Carson’s rule, FM signals, 159 Coarse acquisition code, see
Boresight, antenna, 101, 366 Cassegrain antenna, 102, 367 code CA
Box, station-keeping, 41 Cape Canaveral, 3, 44, 392 Code
Brahe, Tycho, 22 Carrier-to-Noise, 15, 100, 362 ASCII, 277
Brasilsat, 9 overall, 100, 113 BCH, 280, 283
Broadcast negative, 261 binary cyclic, 280
satellite radio, direct, 440 reciprocal formula, 128 block, 280
satellite television, direct, 72, vs. BER, 196 block, systematic, 280
118, 439 Cartesian coordinate system, 20 CA, 264, 460, 467–470
Broadcasting satellite system, 9 Catalyst (hydrazine convolutional, 284, 450
Bsat-1A, -1B, 9 propellant), 63 distance, 281
Budget Celestial mechanics, 17 efficiency, 274
link, 113, 295 Cell Golay, 281
noise power, 134 fixed earth, 430 Gold, 468
power, 134 structure (antenna), 410 Kasami, 258
Bulges, equatorial, 39 Cells, solar, 59, 71 locally generated, 471
Burst error, 281 Cellular telephony, 157 outer, 296
Central angle (satellite P, 460, 467, 472
C coverage), 33 Reed-Solomon, 291
CA code see coarse acquisition Centrifugal force, 18 spreading, see CDMA
code and code Centripetal force, 18 trellis, 285
CBTR, 237 Challenger, 44 turbo, 275, 285, 292
CCIR now called ITU-R Chandra (X-Ray telescope), 44 weight, 281
CDMA, 14, 258, 262, 355 Channel Code division multiple access
capacity, 262 capacity, 275 see CDMA
DS-SS, 262 capacity, FDM/FM, 232, Coding, 273
frequency hopping, 257 488–493 channel, 363
Gold code, 258, 266 coding, 363 concatenated, 290
Kasimi code, 258 I and Q, 183 gain, 146, 282
spreading codes, 258 synchronization, TDM, 212 Coefficients, regression, 318
CN see carrier-to-noise Channelization, of Coherent detector, 189
CN margin transponder, 76 Cold start (CDMA
DBS-TV, 447 Characteristic waves correlation), 470
CN ratio (ionosphere), 310 Columbia, 7
DBS-TV, 450 Chinese Long March rocket, 47 Coma, antenna, 415
GPS receiver, 475 Circular polarization Combining CN and CI, 127
overall, 113, 123 DBS-TV, 444 Command (satellite control), 70
rules of thumb, 128 depolarization, 297 Communications satellite
CONUS see continental Circuit act, 3, 4
United States carrier recovery, 190 Communications
CP see circular polarization hypothetical reference, 128 subsystem, 59, 72
CRC see cyclic redundancy check Circumscribed circle (orbit), 25 Companding, 204
C-band, 3 Clarke, Arthur C., 1, 4 Complementary error function
GEO link budget example, Clarke orbit, 388 (erfc), 192
clear air, 114 Clear Compressed digital video, 439
GEO link budget example, air (or sky), 116 Compression and
worst case, 114 sky (level), 295 expansion, 204
ind.qxd 28/08/02 19:33 Page 525
INDEX 525
Compression, speech, 207 Cumulative distribution Demodulator

Comsat, 3, 7 function, 312 BPSK, coherent, 190
ConcordiaTM, Ellipso Cumulative probability FM, threshold, 119, 169
orbit, 426 distribution, 312 QPSK, 200
Constant Customer service base, 420 Demultiplexing, 221
Boltzmann’s, 105, 115 Cyclic redundancy check Delta II, 45, 425
Kepler’s, 19 (CRC), 256 Delta III, 425
gravitational, 19 Delta III, IV, 46
Constraint length (convolutional D Demodulator, FM, 160
code), 284 DAMA see demand assigned Density, flux, 101, 120, 125
Continental United States multiple access Depolarization
(coverage), 118 DA-SCPC see demand ice crystal, 297, 332
Contour, antenna, 3 dB, 121 assigned rain, 297
Control DBS see direct broadcast Descending node (orbit), 28
attitude, satellite, 64 satellite Determination, orbit, 42
packet, 251 DBS-TV see Direct Broadcast Design
power, uplink, 126, 371 Satellite Television combining CN and CI, 127
system, orbit, 66 Direct Broadcast Satellite example, PCS, LEO
system, thermal, 59 Television, 6, 15, 389 system, 137
Control structure (satellite Congress, law, 120 examples, system, 132
maneuvers) DBS-1 R, 443 life, end of, 38
command word, 70 DBS-TV link budget, 449 uplink, 124
control code, 70 DME see distance measuring to specified CI ratio, 127
execute, 70 equipment to specified CN ratio, 127
quality (components), 87 DOP see dilution of precision system, to specific
Controller, terminal node (data DSBSC see double sideband performance, 131
packet), 171 suppressed carrier Despun
Conus beam, 448 DS-SS see direct antenna, 61
Convective rain, 298 sequence–spread spectrum shelf, 58
Convolutional encoding, 450 DS-SS CDMA, 262 Deutsche Telekom
Coordinate system DTH see direct to home Geschaftsbereich, 9
cartesian, 20 DVB-S see digital video Detector
geocentric, 20, 27 standard coherent, 189
orbital plane, 21 Damper, nutation, 58 correlation, 189
polar, 21 Data transmission, analog FM Deviation see frequency
rectangular, 27 channels, 170 deviation
Co-polarized, 297 Date, Julian, 28 Deviation
Correlation detector, 189 Day frequency (FM), 158
Cosines, law of, 34 Julian, 28 over (FM TV), 166
Costas loop, 479 sidereal, 25 ratio, FM, 161
Coverage mean solar, 25 Device
arc, 432 Decibels, 487–490 active, 411
(area and region), 406 Declination, 27 passive, 411
central angle, 33 Decoding Differential
edge of satellite antenna soft input, soft output attenuation, 327
beam, 115 (SISO), 293 modulation, PSK, 187, 190
instantaneous, 410, 415 Viterbi, 284 PCM, 211
stationary, 430 Defocusing, 310 phase, 327
Criterion, Nyquist, 174, 179 Degraded performance, 296 Digital
Cross-correlation, 469 Delay lock loop, 478 carriers, US standards, 211
Cross polarization, 297 Delay, propagation, compressed TV, 452
Cross polarization 311, 421 demodulation, 187
discrimination (XPD), 303 Demand assigned multiple heirarchy, 211
isolation (XPI), 305 access, 139, 249, 345 modulation, 187
ind.qxd 28/08/02 19:33 Page 526
526 INDEX
Digital (Continued) E Elliptical orbit, 22, 394

transmission, 172 EIRP see effective isotropic Embratel, 9
transmission of analog radiated power Emergency locator
signals, 201 ECEF coordinate system, 464 transmitter, 463
video standard, 441 ERFC (erfc) complementary Emissions, off-axis
voice, SN, 206 error function (antenna), 360
Dilution of precision (GPS function, 192 Emphasis
errors) table, 504 De-, 161, 163
DOP, 466 EIRP see effective isotropic Pre-, 161, 163
GDOP, 481 radiated power End of
HDOP, 481 ELV see expendable launch design life, 38, 390
VDOP, 482 vehicle life, 71
Dip latitude (magnetic), 400 ENST, 292 maneuvering life, 38, 390
Direct EOC see edge of coverage Energy, solar radiation, 71
broadcast satellite EODL see end of design life European Space Agency, 44
radio, 440 EOML see end of Equalizer, 78
broadcast satellite maneuvering life training sequence, 182
television, 72, 118, 439 ESA see European Space transversal, 182
insertion launch, Agency Equation, waveform, FM, 158
(satellite), 48, 49 Early Bird, 1, 389, 424 Equator
PSK modulation, 187 Earth geographic, 400
radiating phased array, 412 average radius, 20 geomagnetic, 400
sequence–spread core, 400 Equatorial
spectrum, 261 mantle, 400 bulges, 39
to home (service station, 15 orbit, 391
delivery), 389 station, simplified, 107 plane, 2
Directv, 7, 119, 441, 444 station, for satellite systems, Equiprobable values
receiving antenna, 446 small, 117 (statistical), 319
Discrete cosine transform, 208 East-West Equivalent noise source,
Dishman, 507 maneuver, 41 105, 109
Dishnetwork, 442 station keeping, 63, 67 output, 110
antenna, 445 Eccentric anomaly, 25 Error
Distance Eccentricity, 22, 24, 394 burst, 281
angular (orbit), 30 Echo I and II, 3 function, complementary, 192
Hamming (coding), 280 Echostar, 6, 7, 443 quantization, 202
measuring equipment Eclipse, solar, 52, 71 sampling, 201
(navigation), 462 Ecliptic, 40 Error control, 273
minimum (coding), 281 Edge of coverage loss, 117, 416 DBS-TV, 451
Distortion, group delay, 78 Effective isotropic radiated Error correction, 273
Domestic (satellites), 125 power, 101, 296 DBS-TV, 452
Doppler shift, 48, 462, 471 uplink, 123 forward, 187
Double-hop links, 348 Effective pathlength, 301, Error detection, 273
Double conversion see 507, 508 DBS-TV, 452
superheterodyne Efficiency, aperture, 102 Error rate
Double sideband suppressed Electrical noise, 105 bit, 188
carrier, 166 Elements BPSK, 194
Downlink, 100 orbital, 20, 29 QPSK, 194
CN, 131 radiating, antenna, 411 symbol, 188
CN budget, PCS, 145 Elevation angle, 31 Example, NGSO system
design, 112 considerations of, 408 design, 432
Drift, orbital, 39, 42, 57 minimum, 37, 408 Expendable launch vehicle
Duration, solar eclipse scaling of attenuation (ELV), 43
(satellite), 71 with, 325 Explorer 1, 3
Dwell time, 394 Ellipso, 425 Explorer, Mars, 405
ind.qxd 28/08/02 19:33 Page 527
INDEX 527
Eutelsat, 10 elliptic function, 178 Frequency Division Multiple

Examples, system design, 132 matched, 178 Access see FDMA
Exceedance curves, 312 roll-off (RRC), 177 Frequency Division
First point of Aries, 27 Multiplexing see FDM
F Fixed Front-fed antenna, 367
FAW see frame alignment word access, 223 Function, transfer, FM
FCC see Federal assignment, 224 pre/de-emphasis, 163
Communications earth cell, 430
Commission power sharing, 231 G
fD ratio, antenna, 415 satellite service, 409 GaAsFET see gallium arsenide
FDM see multiplexing, service, 412 field effect transistor
frequency division Flag (start of frame, etc), 256 GEO see geostationary
FDM/FM, 491 Flux density, 101, 120, 125 earth orbit
FDMA, 77, 124, 139, 157, 222, Focus, prime, 305 GEO/LMDS, 15
232, 355 Force, 17 GEO orbit requirements, 25
FDMA-SCPC-DA, 139, centrifugal, 18 GPS see global
250, 252 centripetal, 18 positioning system
FDM-FM-FDMA, 223, 225 gravitational, 18 GT, 105, 112, 116, 349
FEC see forward error in-plane, 39 Gain
correction out-of-plane, 40 antenna, 101, 297, 432
FFSK see fast frequency Formula, reciprocal CN, 128 coding, 146
shift keying Forward error correction, 122, Galactic background, 24
FM-FDM see Appendix B 187, 273, 296 Galaxy -5 -6 -9 -1RR, 8
FM see frequency modulation margin, PCS, 147 Galileo, 461
FM Fourier transform, Nyquist ISI Gallium arsenide field effect
Carson’s rule, 159 filter, 176 transistor, 106
companding, 169, 204 Fractional transmission Garuda 1, 9, 11
demodulator, 160 coefficient, 332 Gateway (earth station,
improvement, 157, 161 Frame hub), 138, 139, 143
satellite TV, 441 alignment word (TDM), 209 Gaussian distribution, 191
threshold example, 168 basic TDMA, 362 GE Americom, 7
waveform equation, 158 reframing, 210 Gemini, 394
FS see fixed service relay, 351 Generation of QPSK
FSK see frequency shift keying TDMA, Iridium, 421 signals, 198
FSS see fixed satellite service Free fall, 18 Geographic equator, 400
Fade margin, 296 Frequency Geomagnetic equator, 400
Fading, low angle, 309 allocations, 97 Geostationary
Failure rates (bath tub band, 406 earth orbit, 1, 4, 35
curve), 89 deviation, FM, 158 transfer orbit, 46, 48
Faraday rotation, 297, 310 deviation, peak (Carson’s Geosynchronous, 35, 37
Fast frequency shift keying, 200 rule), 159 graveyard orbit, 39
Federal Communications deviation, rms Global
Commission 74, 120, 444 multicarrier, 497 beam, 80
Feed highest (Carson’s rule), 159 transmit horn (antenna), 58
loss, 122 intermediate (IF), 367 receive horn (antenna), 58
matrix, antenna, 411 L1, 460 Global positioning system, 3, 6,
phased array, 83 L2, 460 264, 389, 458
Fiber, optical, 156 modulation, 15, 157 differential, 461, 466,
Field, gravitational, 38 modulation threshold, 118 482–484
Figure, noise, 111 (see also modulation) kinematic, 483
Filter radio (RF), 16, 367 time, 464, 466
bandpass, 75 reuse, 221, 303, 410 Globalstar, 5, 13, 223, 258, 411,
Butterworth, 178 scaling of attenuation, 325 424, 426
Chebychev, 178 shift keying, 157 Glonass, 460
ind.qxd 28/08/02 19:33 Page 528
528 INDEX
Gold code see code and CDMA Hot bird (DBS-TV), 10 Injection, low side
Grazing angle, minimum, 419 Hub station, 99, 138 (mixing), 107
Graveyard, geosynchronous Hughes, 48 Inmarsat see international
orbit, 39 Hughes Electronics maritime satellite
Gravitational Corporation, 441 organization
constant, 19 Hybrid multiple access, 223 Inner code, 363
field, 38 Hydrazine, 62 In-route signal, 356
force, 18 Mono-methyl, 63 Instantaneous
Gravity gradient boom (attitude Hydrometeors, 297 coverage, 410, 415
control), 67 Hypergolic (propellant), 63 Instrument landing
Greenwich Hypothetical reference system, 461
meridian, 31 circuit, 128 Intelsat see international
observatory, 28 telecommunications
Gregorian antenna, 367 I satellite organization
Grid reference, 30 ICO global, 390 Integrity monitoring, 466,
Ground track, 395 ICO, New, 390, 428 472, 483
Group delay distortion, 78 IDU see Indoor Unit INTELSAT K, 5
Growth, incremental, 424 IF see intermediate frequency INTELSAT I, 4
Gstar 4, 7 IFL see inter-facility link INTELSAT V, 5
Guard bands, 355, 493 ILS see instrument landing INTELSAT VI, 5, 58
Guiana space center system INTELSAT VII, 5, 424
(Kourou), 44 IM see intermodulation INTELSAT VIII, 5
ISI see interference, INTELSAT IX, 5
H intersymbol Inter-facility link, 368
HDLC see high level data link ISL see inter-satellite link Interference
control ISO-OSI model, 255 level, 125
HDTV see high definition TV ISS see international space inter-symbol (ISI), 129,
HEO see highly elliptical orbit station 170, 172
HPA see high power amplifier ITU see international intersymbol, zero ISI, 175
Hale cycle (see also sunspot telecommunications union off-axis, 126, 364
cycle), 400 Ice crystal depolarization, 332 Interim operations, 424
Hamming distance Impact, micrometeor, 71 Interleaved code, 296
(coding), 280 Implementation margin, 121, Interleaving, 290
Handset 141, 186, 197 DBS-TV, 451
radiation safety limits, 420 Improvement Intermediate frequency, 106,
PCS, 138 FM, 157, 161 107, 180
Height adjustment factor, 301 subjective, 205 Intermodulation, 75
Hemi-beam, satellite Indoor unit (VSAT), 368 products, 77, 124, 128
antenna, 80 Information third order, 226–230, 357
High definition TV, 16, 156 (see also Shannon), 275 International maritime satellite
Highly elliptical orbit, 4, 455 theory, 275 organization, 5, 10, 84
High level data link control, 353 throughput, 420 International
High power amplifier, 78, 367 Inbound (inroute) telecommunications
equalization, 230 signal, 138, 356 satellite organization, 3, 10,
linearization, 226 Inclination, 28 73, 425
nonlinearity, 226 moon’s orbit, 40 International space
quasi-linear, 230 earth’s spin axis, 40 station, 17, 389
Home satellite TV, 440 sun’s equatorial plane, 40 International
Hopping beam (antenna), 410 Inclined orbit operation, 37, telecommunications
Horizontal adjustment 66, 390 union, 3, 125
factor, 301 Incremental growth, 424 Internet
Horn Inertial space, 25 satellites, 431
antenna (scalar), 366 Infra red sensor (attitude traffic centers, 432
corrugated, 82 control), 65 Interpolation, 291, 451
ind.qxd 28/08/02 19:33 Page 529
INDEX 529
Inter-satellite link, 394, LPC see linear predictive inbound, PCS, 141
422, 426 encoding margin, 116, 295
across plane, 422 L band, 7 margin calculation, VSAT star
Inter-symbol interference, 129 L-star, 9 network, 370
Ionosphere, 480 Land mobile service, 428 margin, with FEC, PCS, 146
Ionospheric Latch up (permanent bit outbound, PCS, 144
scintillation, 297, 312 flip), 400 performance, 295
Iridium, 5, 13, 389, 394, 411, Latitude, 30 Lobe, main, antenna, 414
418, 428 dip angle (magnetic), 400 Local oscillator (LO), 78, 107
Isotropic antenna, 101 Launch Location, primary, Indian, 31
direct insertion into Long March rocket, 47, 425
J orbit, 48 Longitude, 30
JCSat, 10 rockets, next generation, 46 Look angle determination, 30
JPEG see joint picture experts vehicle price, 46 Look angles, 31
group vehicle selection factors, 47 Loral Skynet, 8
Jets, arc, 63 Law Loss
Joint picture expert A-, 169 edge of coverage, 416
group, 208 of cosines, 34 feed, 122
Julian day, 28 -, 169 mechanisms, 298
Julian date, 28 Laws and Parsons, 316 miscellaneous, 116, 143
Juno I, 3 Laws of motion, 17 path, 103, 125, 297
Length, constraint propagation, 116, 297
K (convolutional code), 284 scan, antenna, 416
Ka band, 5, 79 Level, waveguide, 111
Kasami code see code, Kasami interference, 125 Low
Kazakhstan, 42, 392 side lobe, 125 angle fading, 309
Kennedy space flight Life, earth orbit, 4, 5, 99, 388
center, 44 beginning of, 71 noise amplifier, 104, 107,
Keplerian orbit, 38 end of, 71 122, 367
Kepler’s constant, 19 Lifetime, noise block, 107, 369
Kepler’s laws, 22 design, 38 noise block converter, 446
Kopernicus -1 -2, 9 maneuvering, 38 side injection (mixer), 107
Koreasat, 11 operational, 37 Luminance, TV, 165
Korea telecom, 45 Limited, power, 72
Kourou (Guiana space Linear M
center), 44 energy transfer MEO see medium earth orbit
krad(Si) (space radiation (radiation), 403 m-sequence see PN sequence
level), 402 polarization, 297, 310, 326 MF-TDMA see multi-frequency
Ku band, 4 predictive encoding, 207 TDMA
downlink, system design Linearity (of transponder), 77 MPEG-2, 16, 117, 165, 208,
example, 134 Link 447, 451
rain attenuation, 120, 147 availability, 295 MSAT 1, 12
rain effects, 135 budget, 113, 295 MSK see minimum shift
uplink, system design, budget equation, 103 keying, 200
example, 133 budget example, C-band, MSS see Mobile Satellite
earth coverage, 115 Service
L budget, example, Ku-band MTBF see mean time between
LEO see low earth orbit DBS TV, 121 failures
LEOs, big, 411 CN ratios, VSAT star Maneuvers, station keeping, 60
LMS see land mobile service network, 375 Maps,
LNA see low noise amplifier combining CN and CI rain climatic, 314
LNB see low noise block ratios, 127 rainfall exceedance
(converter) design, satellite, 96 contour, 316
LP see linear polarization equation, 100, 103 Mabuhay Philippines, 11
ind.qxd 28/08/02 19:33 Page 530
530 INDEX
Maneuver, orbital, Monitoring system (on board), 68 contributions, 128

east-west and north-south, 41 Monochrome (TV), 165 electrical, 105
Maneuvering life, end of, 38 Morse code, 15 figure, 111
Margin Motion, figure, standard
fade, 75, 296 laws of, 17 temperature, 111
implementation, 121, 141, planetary, 22 power, baseband, FM, 160
186, 197 Mu-law (-law), 169 power budget, DBS-TV, 450
link, 295 Multi-frequency TDMA, 359 power spectral density, 105,
link, with FEC, 146, 147 Multipath, atmospheric, 310 160, 161
system, 103 Multiple access, 221 source, equivalent, 106, 109
Mars Explorer, 405 demand assigned, 139 source, equivalent
Mascons see mass hybrid, 223 output, 110
concentrations Multiplexing, 156, 221, suppression, FM
Mass concentrations, 39 493–495 demodulator, 160
Master control station, 392 analog, 156 temperature, 105
DBS-TV, 452 frequency, 156 temperature increase (due to
Master group, 493 frequency division, 156, rain), 116, 122
Mean time between failure, 90 223, 493 temperature, sky, 116
Mean temperature, system, 105
anomaly, 26 N thermal, 129
solar day, 2 NACK see negative white, 161
Medium earth orbit, 4, 5, 16, acknowledgement Noiseless amplifier, 105
99, 388 NASA, 88 Non-GEO rain attenuation
Melting layer, 298, 300 NGSO see non-geostationary prediction, 324
Mercury (capsule), 388, 394 satellite orbit Non-geostationary satellite
Mesh network, 348 NGSO system design, 432 orbit (NGSO) systems, 427
Microburst, rain, 302 N2H2 see hydrazine Non return to zero (line code),
Micrometeor (impacts), 71 NRZ see non return to zero 172, 174
Microwave source, hot (sun), 53 NS 513, 11 QPSK channel, 186
Minimum NTSC (TV standard, US), North-south
elevation angle, 37 165, 168 maneuver, 41
grazing angle, 419 Nadir, 31 station keeping, 63
shift keying, 200 Nahuel 1, 11 Number of NGSO satellites
Misalignment, polarization, 142 NahuelSat, 11 per plane, 433
Mixer, 107 Narrowcasting, 347 total, 433
Mixing, turbulent, 308 Navigation message, 460, 470, Nutation, 60
Mobile 472–473 damper, 58
satellite service, 409, 425 Navstar GPS, 14 Nyquist criterion, 174, 179
terminal, 140 Negative acknowledgement
Modulation, 156 signal, 286, 350 O
cross product (FM), 159 Neutral atmosphere, 298 OBP see on board processing
differential, 190 Next generation launchers, 46 ODU see outdoor unit
frequency, 491 New ICO, 390, 411, 428 OMT see orthogonal (ortho)
16-QAM, 247 New Skies company, 5, 11 mode transducer
sine wave, 160 Newton, 17, 18, 23 Observatory, Greenwich, 28
Modulo-8, 354 Newtonian equations, 17 Occupied bandwidth, 180
Modulo-128, 354 Nickel-hydrogen batteries, 72 Off-axis
Molniya, 4, 67, 396 Node, emission, 360
orbit, 396 ascending, 28 interference, 364
Momentum descending, 28 scanning, 412
angular, 20 Noise Offset-fed antenna, 367
wheels, 60 bandwidth, 115 Omnidirectional antenna, 80
wheel, unload (dump bandwidth and symbol rate, On board processing, 76,
energy), 64 RRC filter, 121 247–248, 370
ind.qxd 28/08/02 19:33 Page 531
INDEX 531
Operational lifetime, 37 Outbound (outroute) optimizing, system

Operations, interim, 424 signal, 138, 357 performance, PCS, 146
Optical fiber, 156 Outdoor unit (VSAT), 368 threshold, 296
Options, replenishment, 424 Outer code, 296, 363 Perigee, 22
Optimizing system performance, Output power time of, 26
PCS, 146 back-off, 77, 115, 124, west, argument of, 28
Optimum orbital altitude, 418 129, 357 Perihelion, 25
Orbcomm, 13, 393, 407, saturated, 115, 129 Period
424, 429 Out-route signal, 356 anomalistic, 38
Orbit Over deviation GEO satellite, 24
apogee, 22, 23 (FM TV), 166 orbital, 23
determination, 42 Overall CN ratio, 113, 123 symbol, 187
eccentricity, 22 Personal communication
elliptical, 22, 30 P service, 137
equatorial, 391 P code, 460, 467, 472 design example, LEO
geostationary transfer, 46 PAM see pulse amplitude system, 137
graveyard, modulation inbound (inroute) link, 141
geosynchronous, 39 PAS 1, 8 outbound (outroute)
highly elliptical, 455 PCM see pulse code link, 144
inclined, 37 modulation Perturbations, orbit, 38
osculating, 38 PCS see personal Phase
perigee, 22, 23 communication service channel, in, 183
perturbations, 38 PLL see phase lock loop continuous, 171
precession, 40 PM-AM conversion, 124 equalizers, 182
stable, 17 PN see pseudo noise front, 412
Orbit raising, slow, 48 PSK see phase shift keying lock loop, 239
Orbital PSTN see public switched quadrature, 183
altitude, optimum, 418 telephone network Phase shift keying, 157, 173
drift, 39 PT Pasifik Satelit, 11 binary, 180, 187, 189, 362
elements, 20, 38 Packet continuous, 171
elements (determination radio, 254 generation, QPSK, 198
of), 68 header, 256 minimum, 171
maneuvers, 41 Palapa C1, 11 QPSK variants, 199
period, 23 PanAmSat, 8 quaternary, 116, 121, 187,
period, apparent, 391 Parity, 277 198, 296
plane, 23 single, 278 Phased array
radius, 19 Passive antenna, 411
slot spacing, 74 device, 411 direct radiating, 412
slots, 84 phased array, 412 scan angle, 412
stability, spinner, 60 Path Pitch axis (satellite), 63
stability, three axis, 60 attenuation, total, 122 Plane,
velocity, 19 blockage, 149 equatorial, 27
Orthogonal polarization, 221 loss, 103, 125, 297 across seam, ISL, 422
Ortho-mode transducer Path length, orbital, 28
(OMT), 305, 444 effective, 298, 301, number of (NGSO) satellites
Oscillator, local (LO), 78 507, 508 per, 433
Osculating orbit, 38 physical, 319 Planetary motion, 22
Orthogonal Pay per view, 447 Points, stable and
mode transducer, Peenemunde, 404 unstable, 39
(OMT), 305, 444 Pegasus, 424, 425 Polarization
polarizations, 297 Performance circular, 297, 326, 444
Outage, 100, 113, 120 degraded, 296 linear, 297, 310, 326
sun transit, 53 design to, specific, 131 misalignment, 142
DBS-TV, 447 link, 295 orthogonal, 72
ind.qxd 28/08/02 19:33 Page 532
532 INDEX
Power Pulse Rain

carrier, uplink, 123 amplitude modulation, 201 accumulation, 312
control, uplink, 126, 148, 371 code modulation, 201, 356 added noise temperature, 111
limited, 72, 275 Pyrotechnic (deployment), 86 attenuation, 15, 116, 297,
noise, spectral density, 105 317, 507–510
output, back-off, 77, 115, Q attenuation at Ku-band, 147
124, 129, 357 Q factor see subjective attenuation margin, 120
output, saturated, 115, 129 improvement factor attenuation prediction, 317
system, 59, 71 Q function (Gaussian), 192 attenuation statistics,
thermonuclear, 71 tables, 505 DBS-TV, 449
Precession, orbit, 40, 405 QAM see quadrature amplitude climatic maps, 314
Precise code see P code modulation convective, 298
Prediction QPSK see quaternary phase effects, Ku-band, 135
rain attenuation, GEO, 317 shift keying exceedance contour
rain attenuation, NGSO Quadrature amplitude maps, 316
satellites, 324 modulation, 173, 235 height, 508
site diversity gain, 336 Qualification, space, 87 microburst, 302
XPD, 326 Quality stratiform, 298
Preflight testing, satellite, 88 assurance, 87 streamer, 302
Pressure, solar, 60 control, 87 Raindrops
Price, launch vehicles, 46 Quantization error, 202 absorption, 297
Prime focus, 305 Quantizing, 201 distribution, 315
Primestar, 441, 442 Quaternary phase shift scatter, 297
Probability of symbol error, 191 keying, 116, 121, 198, 296 shape, 327
Products, intermodulation, 77, spectrum of, 184 size, 302
124, 128 Rainfall rate, 312
Prograde orbit, 391 R Rain gauge, tipping
Project Score, 3, 388 RF see radio frequency bucket, 312
Propagation impairment RHI see range height Random access, 254
countermeasures indicator (radar) Range height indicator, 303
attenuation, 333 RLV see reusable Range ambiguity, 483
depolarization, 337 launch vehicle Ranging
diversity, 335 RRC see root raised (orbital ephemeris), 59
power control, 334 cosine (filter) tones (range
signal processing, 335 RZ see return to zero (line code) determination), 69
site diversity, 336 Rad-hard see radiation hardened Ratios CN and CI,
Propagation loss, 116, 297 Radar, S-band, 303 combining, 127
Propellant tanks, 58 Radiating elements, receive power, 102
Protocol, 255 antenna, 411 Receiver
spoofing, 286, 352 Radiation command, 58
stack, 350 belts (Van Allen), 398, DBS-TV, 446
window, 351 402, 426 GPS, 476–480
X.25, 256 black body, 105 Reciprocal CN formula, 128
Protocols, access control, 349 dose, 403 Recovery, carrier
Proton rocket, 42, 425 effects, 398 (TDMA), 362
Proton M, 46 hardened, 403 Redundancy (spare), 75,
Primary location, hazards, EM, 124 87, 90
Indian Ocean, 31 linear energy transfer, 403 parallel, 91
Atlantic Ocean, 32 safety, handset, 420 ring, 91
Project Score, 3 Radio frequency (RF), 16, series, 91
Psuedorange (GPS), 465 107, 367 Redundant bits (coding), 274
Pseudonoise (sequence), 467 Radius, Reed-Solomon
Public switched telephone average earth, 20 code, 296, 363
network, 426, 428 orbital, 19, 22 DBS-TV, 450
ind.qxd 28/08/02 19:33 Page 533
INDEX 533
Reference Sampling, 201 Sidereal day, 25

circuit, hypothetical, 128 Satellite Signal
grid, 30 antenna gain, 432 polar, 172
station, 482 axes, defined, 63 bipolar, 172
Reflector digital radio service, 445 Signal-to-noise, 15, 100
shaped, 367 domestic, 125 in digital voice systems, 206
offset parabolic, 445 electrical model, 87 Simple attenuation model
Refractive effects, 308 GPS, 458 see SAM
Reframing (TDM), 210 link design, 96 Sinc function see (sin x)x
Regression coefficients, 318 mechanical model, 97 Sine wave modulation, 160
Reliability, 87, 88 number, per plane, 433 Single
Repeater, 15 prototypes, 87 channel per carrier, 169,
Replenishment options, subsystems, 57 225, 359
satellite, 424 systems using small earth event upset (radiation), 398
Resonant absorption, 307 stations, 117 parity, 278
Retrograde orbit, 391 telephone, 140 stage to orbit, 43
Return to zero (line code), 172 thermal model, 87 Sino satellite, 11
Reusable launch vehicle, 43 tool kit, (STK), 35 SinoSat 1, 11
Reuse virtual packet, 384 (Sin x)x, 173
frequency, polarization, 72 Satellite orbital velocity, 19 SIRIO satellite, 313
frequency, spatial, 72 Saturated output power Sirius 1, 12
Right ascension, 27 (HPA), 115, 129 Sirius Satellite Radio Inc, 455
Roll axis (satellite), 63 Scalar antenna horn, 366 Sky noise temperature, 332
Root raised cosine filter, 121, Scaling attenuation, 126, increase in rain, 116,
173, 355 127, 325 122, 136
roll-off factor, 177 Scan angle, Skybridge, 394, 429, 431
Rotation, Faraday, 297, 310 phased array, 412 Slots, orbital, GEO satellite, 84
Rules of thumb, CN, 128 total, 414, 415 Small earth stations, satellite
Scan loss, antenna, 416 systems, for, 117
S Scanning beam (antenna), Smart card, 447
SA see selective availability 78, 410 Societe Europenne des
SAM (simple attenuation Scanning, off-axis, 412 Satellites, 5, 11, 442
model), 507 Scintillation, Solar
SARSAT, 463 ionospheric, 297, 312 cells, 59
SBS-4, 5, 6, 8 tropospheric, 297 day, mean, 25
SCPC see single channel per Score see Project Score eclipse, 71
carrier Seam, ISL, across plane, 422 eclipse duration, 71
SCPC-FDMA, 226 Selection factors, launch pressure, 60
SDARS see satellite digital vehicles, 47 sail, 70
radio service Selective availability, 460, sunspot cycle, 400
SES see Societe Europenne des 466, 476 Solid state high power amplifier,
Satellites Semilatus rectum, 22 78, 226
SISO (soft input soft output) Semimajor axis, 23 Source, hot microwave
see decoding Semiminor axis, 23 (sun), 53
SN see signal-to-noise Service, customer, base, 420 Soyuz rocket, 425
SSHPA see solid state high Shake and bake tests, 88 SS-Loral see space systems-
power amplifier Shaped reflector, 367 loral
SSPA see solid state power Shannon, 275 Spacecom satellite
amplifier bound, 275 communication, 12
SSTO see single stage to orbit -Hartley law, 275 Spacelab, 44
STK see satellite tool kit Shear, wind, 302 Space
STS see space transportation Shift, Doppler, 49 flight center, Kennedy, 44
system Shuttle, space, 29, 394 center, Guiana (Kourou), 44
Safety, radiation handset, 420 Sidebands, FM, 159 communications corp., 12
ind.qxd 28/08/02 19:33 Page 534
534 INDEX
Space (Continued) Sub-reflector, 344 TNC see terminal node

frontier, 17 Sub-satellite point, 31 controller
inertial, 25 Subsystems, satellite, 57 TTC&M see telemetry, tracking,
qualification, 87 Sunset-sunrise orbit, 404 control and monitoring
shuttle, 29 Sunspot (cycle), 401 TV
transportation system, 43 Sun synchronous orbit, chrominance, 165
wings, 17 390, 403 color, 166
Space systems-loral, 6 Sun transit outage, 53 color subcarrier, 167
Sparklies see Demodulator, FM Superbird A, B, 11 DBS, 118
threshold Super group, 493 DSBSC, 166
Specific Superheterodyne, 107, 108 hue, 166
attenuation, 297, 318 Swedish space corporation, 12 monochrome, 165
performance, design to, 131 Symbol, 235 saturation, color, 166
Spectrum, QPSK, 184 bits per, 187 set-top boxes, 167
Speech compression, 207 error probability, 191 VSB, 166
Spin (spun) up, 62 error rate, 188 TWTA see traveling wave tube
Spin stabilized, 71 period, 187 amplifier
Split two-way (split-IP), 348 rate, 172, 235 Taurus rocket, 425
Spoofing, protocol, 286, 352 recovery circuit, 189 Teledesic, 394, 423, 430, 431
Spot beam, 80 Synchronous orbit, sun, Telemetry, tracking, control, and
antenna, 412 390, 403 monitoring, 42, 68
DBS-TV, 448 Synchronization, channel Telenor satellite, 12
multiple, 80 (TDM), 212 Telesat Canada, 12
SPOT satellite, 388 Syncom (first GEO Telurometer (distance measuring
Spread spectrum see CDMA satellite), 48 device), 69
Spreading codes see CDMA System, Telstar I and II, 3, 388
Sputnik, 1 design examples, 131 Temperature
Sputnik 1, 388 design, NGSO sky noise, 332
Stable constellation, 432 increase due to rain, 116,
orbit, 17 design, procedure, 131 122, 131
point (in the orbit), 39 design, to specific Tempo 2, 7
Stabilized performance, 131 Terminal node controller, 171
Spin(ner), 60 noise temperature, 116 Test
Three-axis, 60 performance, optimizing, point, 492
Staging (rockets), 42 PCS, 145 preflight (satellite), 88
Standard Systematic block code, 280 shake and bake, 88
atmosphere, 207 Switch tone, 492
noise temperature, 111 beam, 78 tone (FM), 158
Star network (VSAT), matrix transponder, 75 visibility, 36
343, 348 Thermal
Stationary coverage T noise, 129
(NGSO), 430 T carrier standards, USA, 211 noise, sun, 53
Station-keeping TCP/IP, 287 problems, 87
box, 41 TDD see time division Thin-route traffic, 355
maneuvers, 60 duplexing Third order
thrusters, 60 TDM see time division intermodulation, 226, 357
Statistics, long-term, 313 multiplexing Thor 1, 2, 12
Store-and-forward, 171 TDM-FDMA, 222 Three axis stabilized
Stratified layers, 308 TDM-SCPC-FDMA, 226 (satellite), 63
Stratiform rain, 298, 409 TDMA see time division Threshold
Streamer, rain, 302 multiple access FM, 118
Stutzman, 507 TDRSS satellite, 7, 44, 394 FM, extension, 169
Subjective improvement, 205 TIROS satellite, 388, 404 Throughput, information, 421
(weighting) factor, 167 TMI communications, 12 Throw weight/mass, 47, 388
ind.qxd 28/08/02 19:33 Page 535
INDEX 535
Thrusters digital, through bandlimited Uplink

arc jet, 63 channel, 182 CN, 131
ion, 63 error free, 157 CN budget, PCS, 144
orbit maneuvering, 63 level, 491 carrier power, 123
Tilt angle, 327 (link) equation, 100 design, 123
Time theory, basic, 100 Ku-band, uplink, system
universal, 28 Transponder, 59, 72, 75 design example, 133
zulu, 28 backoff, 357 power control, 126, 135,
GMT, 467 bandwidth, 221, 226 148, 371
UTC, 467 baseband processing, 247
Time division bent pipe, 226 V
duplexing, 421 channelization, 76 V2 rocket, 404
multiplexing, 209, 222, DBS-TV, 119 VOR beacons, 461
226, 357 HPA, 221 VOW see voice order wire
Time of perigee, 26 linear, 130 VSAT see very small aperture
Time division multiple access, linearity, 77 terminal
77, 223–246, 355 receive noise temperature VSAT hub station, 343
burst duration, 243, 245 change, 130 VSAT star network, 343
burst transmission, 234 switch matrix, 75 VSAT/WLL, 15, 346
efficiency, 237 types, 129 VSB see vestigial sideband
frame, 234 Transversal equalizer Van Allen radiation belts , 398,
frame length, 234 (ISI), 182 402, 426
guard bands, 225 Traveling wave tube amplifier, Vandenburg air force
guard times, 235 78, 124, 226, 229 base, 44
multifrequency, 223 Trellis code, 285 Velocity
preamble, 236 Trilateration, 460 angular, 25
reference burst, 236 Troposphere, 480 satellite, 19
satellite switched, 246, 248 Tropospheric scintillation, of light, 460
synchronization, 222 297, 308 Very small aperture terminal,
transmitter power, 243 True anomaly, 25 107, 124, 298, 343
Tipping bucket rain Turbo code, 275, 285, 292 Vestigial sideband, 166
gauge, 312 Turbulent mixing, 308 Virginia Tech, 439
Titan III, 45 Turksat 1B, 1C, 12 Virtual Geosatellite, 431
Titan III, IV, 46 Turk Telekom, 12 Visibility test, 36
Total Tycho Brahe, 22 Viterbi decoding
dosage (radiation), 398 Tyuratam, 42 algorithm, 284, 451
path attenuation, 122 Voice order wire, 240
Tracking (of satellite), 68 U Vostok, 388
Traffic centers, Internet, 432 ULPC see uplink power control
Training sequence UPC see ULPC W
(equalizer), 182 USAT see ultra small aperture WAAS see wide area
Transducer, orthomode, 305 terminal augmentation system
Transfer UT see universal time, 28 WAC see world administrative
function, FM, pre/de- UW see unique word conference (formerly
emphasis, 163 Ultra small aperture WARC)
orbit, geostationary, 46 terminal, 344 WARC see world administrative
TRANSIT (satellite), 388, 461 Unique word (TDMA), 191, radio conference
Transmission 239, 362 WBFM see wide band FM
coefficient, fractional, 332 correlator, 240 WESTAR I, 4
data, using analog FM Universal time, 28 WLL see wireless local loop
channels, 170 Unload, momentum wheel Water vapor absorption, 308
digital, 172 energy, 64 Waveform, equation,
digital, of analog Unstable point (in the FM, 158
channels, 201 orbit), 39 Waves, characteristic, 310
ind.qxd 28/08/02 19:33 Page 536
536 INDEX
Weighting WorldSpace corporation, 8 Z

psophometric, 497 World radio conferences, 15 z-axis intercept (satellite,
subjective improvement antenna reference), 63
factor, 167 X Zenit rocket, 425
Wide area augmentation system, XPD see cross polarization Zenith, 32
466, 483 discrimination Zenith attenuation, 307
Wideband FM, 15, 157 XPD prediction, 326 Zero-ISI waveform, 170
Wind shear, 302 XPI see cross polarization Zero
Window, protocol, 351 isolation ISI filter, 175
Wireless local loop, 15, 345 X.25, X.75, 256, 351 non return to (line code),
Word X-33, X-34 (SSTO), 43 172, 174
frame alignment, 209 return to (line code), 172
unique, 191 Y Zone beam, 80
World Administrative Radio Y code, 460
Conference, 74 Yaw axis, 63
This Wiley Student Edition is part of a continuing
program of paperbound textbooks especially designed
for students in developing countries at a reduced price.
Second Edition
THIS BOOK IS FOR SALE ONLY IN THE COUNTRY TO
Pratt
WHICH IT IS FIRST CONSIGNED BY WILEY INDIA
PVT. LTD. AND SHOULD NOT BE
Bostian
RE-EXPORTED. Allnutt
SATELLITE
FOR SALE ONLY IN :
SATELLITE COMMUNICATIONS
INDIA, BANGLADESH, NEPAL, PAKISTAN, SRI
LANKA AND BHUTAN
Wiley Acing the GATE Series

A Complete Solution for all GATE Preparation Needs
AS PER LATEST GATE SYLLABUS
COMMUNICATIONS
Timothy Pratt
Charles Bostian
Jeremy Allnutt
www.wileyindia.com
SPECIAL INDIA EDITION

The content of this book may have been
modiﬁed to suit Indian context.
ISBN978-93-88991-29-2
eISBN: 938899129-X
SECOND
EDITION
Wiley India Pvt. Ltd.
Customer Care +91 120 6291100
csupport@wiley.com
www.wileyindia.com
9 789388 991292
www.wiley.com

Timothy Pratt, Charles W. Bostian, Jeremy E. Allnutt - Satellite Communications-Wiley Indian (2003)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Timothy Pratt, Charles W. Bostian, Jeremy E. Allnutt - Satellite Communications-Wiley Indian (2003)

Uploaded by

Copyright:

Available Formats

This Wiley Student Edition is part of a continuing

program of paperbound textbooks especially designed

Wiley Acing the GATE Series

SPECIAL INDIA EDITION

John Wiley & Sons

Authorized India Edition

ISBN: 978-93-88991-29-2 (ebk)

ABOUT THE AUTHORS

Charles W. Bostian is Clayton Ayre Professor of Electrical and Computer Engineering at

Jeremy E. Allnutt is a professor in the Electrical and Computer Engineering Department

2. Orbital Mechanics and Launchers 17

4. Satellite Link Design 96

4.4 Design of Downlinks 112

5. Modulation and Multiplexing Techniques for Satellite Links 156

Example 6.3.1 TDMA in a Fixed Station Network 237

Tropospheric Scintillation and Low Angle Fading 308

9. VSAT SYSTEMS 343

9.6 VSAT Earth Station Engineering 366

10. Low Earth Orbit and Non-Geostationary Satellite Systems 388

11. Direct Broadcast Satellite Television and Radio 439

12. Satellite Navigation and the Global Positioning System 458

Appendix A Decibels in Communications Engineering 487

Appendix B FDM/FM/FDMA Analog Telephone Transmission 491

Example B.2 503

Appendix C Complementary Error Function erfc(x) and Q Function Q(z) 504

Appendix D The Simple Attenuation Model 507

1.2 A BRIEF HISTORY OF SATELLITE COMMUNICATIONS 3

1.2 A BRIEF HISTORY OF SATELLITE

1.2 A BRIEF HISTORY OF SATELLITE COMMUNICATIONS 5

1.3 SATELLITE COMMUNICATIONS IN 2000

Organization Satellites Type Transponders Orbit location

American Mobile AMSC-1 Mobile land 16 L band 101° W

Organization Satellitesa Type Transponders Orbit location

GE-1A and GE-5 are

Organization Satellites Type Transponders Orbit location

Arab Satellite Arabsat 2A Broadcasting 22 C band, 12 Ku band 26° E

Organization Satellites Type Transponders Orbit location

Paris, France telecommunications

www.nahuelsat.com.ar Nahuel 2 Telecommunications 46 C band, 36 Ku band 81° W

Organization Satellites Type Transponders Orbit location

Space Communications Superbird A Broadcasting, 23 Ku band, 2 Ka band 158° E

There are 101 satellites listed in Table1.2.

Organization Satellites Type Transponders Constellation orbit

TABLE 1.4 Other Satellite Systems

System Satellites Type and lifetime Application Orbits

1.4 OVERVIEW OF SATELLITE COMMUNICATIONS 15

1.4 OVERVIEW OF SATELLITE

2.1 ORBITAL MECHANICS

Developing the Equations of the Orbit

18 CHAPTER 2 ORBITAL MECHANICS AND LAUNCHERS

The satellite has a mass, m,

2.1 ORBITAL MECHANICS 19

TABLE 2.1 Orbital Velocity, Height, and Period

Orbital height Orbital velocity Orbital period

Intelsat (GEO) 35,786.03 3.0747 23 56 4.1

20 CHAPTER 2 ORBITAL MECHANICS AND LAUNCHERS

2.1 ORBITAL MECHANICS 21

FIGURE 2.3 The orbital plane coor-

FIGURE 2.4 Polar coordinate system in the plane

22 CHAPTER 2 ORBITAL MECHANICS AND LAUNCHERS

Where 0 is a constant and e is the eccentricity of an ellipse whose semilatus rectum p is

Kepler’s Three Laws of Planetary Motion

2.1 ORBITAL MECHANICS 23

Describing the Orbit of a Satellite

Where 0 is a constant and e is the eccentricity of an ellipse whose semilatus rectum p is

h 12p2 T 1m12 2 1a32 2 (2.25)

T 2 14p2a3 2 m 4p2 16628.142 33.986004418 105 s2