69950171 Introduction to Radar Systems | Radar | Antenna (Radio)

Introduction to

RADAR SYSTEMS

>21

McC RAW-HILL

Introduction to

RADAR SYSTEMS
MERRILL
I.

SKOLNIK

Research Division
Electronic Communications, Inc.

McGRAW-HILL BOOK COMPANY,
New York
San Francisco

INC.

1962

Toronto

London

PRESTON

fcZUO^ 8381 ±u
^o7c5

^St<<=>

1 °t°S

<f

INTRODUCTION TO RADAR SYSTEMS
Copyright 1962 by the McGraw-Hill Book Company, Inc. Printed in the United States of America. All rights reserved. This book, or parts thereof, may not be reproduced in any form without permission of the publishers. Library of Congress Catalog Card Number 61-17675

©

57905

PREFACE
The subject matter of electrical engineering may be classified according to (1) components, (2) techniques, and (3) systems. Components are the basic building blocks that are combined, using the proper techniques, to yield a system. This book attempts to present a unified approach to the systems aspect of radar. Although the subject of radar systems is of particular interest to specialists in the radar field, it is also of interest to a much wider audience, especially the civilian and military users of radar, the electrical and mechanical components specialists whose devices make up a radar system, the operations analysts and systems engineers who must plan for employing radar as part of larger systems, as well as practicing engineers and scientists in related
fields.

This book originated in the notes for a graduate course in radar systems engineering taught for several years in the Graduate Evening Division of Northeastern University (while the author was a staff member at MIT Lincoln Laboratory) and, later, as an off-campus course at the Martin Co. for the Drexel Institute of Technology. Since most electrical engineering courses are usually concerned with either components or techniques, a course dealing with electronic systems (in this instance, radar systems) broadens the engineering background of the student by giving him the opportunity to apply the material learned from his components and techniques courses, as well as introducing him to the techniques, tools, and analytical procedures of the systems
engineer.

The book may be divided into four parts. Chapters 1 to 5 deal with subjects which are characteristic of radar per se and include a brief introduction and historical survey, the prediction of radar range performance, and discussions of the pulse, CW, FM-CW,
the subsystems and major components constituting a radar system, such as transmitters, modulators, duplexers, antennas, receivers, and indicators. The emphasis is on those aspects of components of interest to radar. Only brief consideration is given to the operating principles of components. Many books are available that can provide more detailed descriptions

MTI, pulse-doppler, conical-scan, and monopulse radars. The second part, Chapters 6 to 8, is concerned with

than

possible in the limited space allotted here. third part, Chapters 9 to 12, treats various topics of special importance to the radar systems engineer. These include the detection of signals in noise and the extraction of information from radar signals, both of which are based on modern
is

The

communication theory and random-noise theory. This is followed by the environmental factors influencing radar design, for example, propagation, clutter, weather, and interference. The last portion of the book deals with radar systems and their application. Several brief examples of radars are given in Chapter 13. The book concludes with a chapter on the application of radar to the detection of extraterrestrial objects such as planets, satellites, meteors, aurora, and the moon. Although mathematics is a valuable tool of the systems engineer, no special mathematical background is assumed here. Where mathematics is necessary, it is reviewed
briefly in the text.

To
and

its

attempt to treat thoroughly all aspects of a radar system, its component parts, analysis is an almost impossible task within a single volume, since the subject

vi

Preface

of radar encompasses almost all electrical engineering. Extensive references to the published literature are included for those desiring more detail. Radar has been used on the ground, on the sea, and in the air, and undoubtedly it will be used in space. The environment in which a specific radar operates will have an important influence on its design. Although an attempt is made to be as general as possible, when it is necessary to particularize the radar environment, a ground-based radar is assumed unless otherwise stated. The function of the radar systems engineer is to utilize' the available components and techniques to evolve a system that will operate in a particular environment and satisfy the objectives and requirements desired by the potential user. It is hoped that this book will serve to aid those involved in this process. Merrill I. Skolnik

CONTENTS
Preface

CHAPTER

1.

THE NATURE OF RADAR
Introduction
1

1.1

1.2
1.3

1.4
1.5

1.6

The Radar Equation 3 Radar Block Diagram and Operation Radar Frequencies 7 History of Radar Development 8 Applications of Radar 14
References
19

chapter

2.

THE RADAR EQUATION
Prediction of Range Performance Minimum Detectable Signal 21

20 20

2.1

2.2
2.3

Receiver Noise

23 25

2.4
2.5

Probability-density Functions

2.6 2.7
2.8

Signal-to-noise Ratio 29 Integration of Radar Pulses 35 Radar Cross Section of Targets Cross-section Fluctuations 50

40

2.9

2.10
2.11

2.12 2.13 2.14

Transmitter Power 56 Pulse Repetition Frequency and Range Ambiguities Antenna Parameters 58 System Losses 61 Propagation Effects 66 Summary 67 70 References

57

CHAPTER

3.

CW AND FREQUENCY-MODULATED RADAR
The Doppler CW Radar
Effect

72

3.1

72

3.2
3.3

73

3.4
3.5

Radar 86 Frequency-modulated 103 Airborne Doppler Navigation Multiple-frequency Radar 106

CW

CW

References

111

CHAPTER

4.

MTI

AND PULSE-DOPPLER RADAR
113

113

4.1

4.2
4.3

Moving-target-indication (MTI) Radar Delay Lines and Cancelers 119 140 Subclutter Visibility

4.4
4.5

MTI

Using Range Gates and
1

Filters

151

Pulse-doppler Radar

53

4.6 4.7 4.8 4.9

Noncoherent MTI 154 155 MTI from a Moving Platform—AMTI 157 Fluctuations Caused by Platform Motion Effect of Sidelobes on Pulse-doppler AMTI Radar
References

159

162

CHAPTER

5.

TRACKING RADAR
Tracking with Radar Sequential Lobing
Conical Scan
1

164

5.1

5.2
5.3

164 165
vii

66

V1U

Contents
5.4
5.5

Simultaneous Lobing or Monopulse
Target-reflection Characteristics

175 184

and Angular Accuracy

5.6
5.7 5.8

5.9

5.10

Tracking in Range 189 Tracking in Doppler 190 Acquisition 190 Examples of Tracking Radars Comparison of Trackers 195 196 References

192

CHAPTER

6.

RADAR TRANSMITTERS
Introduction 198

198

6.1

6.2
6.3

6.4
6.5

Magnetron Oscillator 199 Klystron Amplifier 216 Traveling-wave-tube Amplifier
Amplitron and Stabilitron 227 Grid-controlled Tubes 233 Comparison of Tubes 244 Modulators 248 References 255

225

6.6 6.7
6.8

CHAPTER

7.

ANTENNAS
Antenna Parameters 260 Antenna Radiation Pattern and Aperture Distribution Parabolic-reflector Antennas 269 Scanning-feed Reflector Antennas 277 Cassegrain Antenna 282 Lens Antennas 286 Array Antennas 294
320 Cosecant-squared Antenna Pattern
Pattern Synthesis
Effect of Effect

260 264

7.1

7.2
7.3

7.4
7.5

7.6
7.7

7.8 7.9

329 330

7.10
7.11

Broadband Signals on Antenna Patterns of Errors on Radiation Patterns 336
343

7.12 7.13

Radomes
References

Focused Antennas
349

347

CHAPTER

8.

RECEIVERS
356 Superheterodyne Receiver 357 Receiver Noise 361 Noise Figure 363 Effective Noise Temperature 365 Environmental Noise 366 RF Amplifiers 373 Crystal Mixers 385 IF Amplifiers 388
Displays
391

356

8.1

The Radar Receiver

8.2
8.3

8.4
8.5

8.6
8.7 8.8

8.9

8.10
8.11

Dup lexers
References

395 403

chapter

9.

DETECTION OF RADAR SIGNALS IN NOISE
Introduction 408 Matched-filter Receiver 409 Correlation Detection 418 Detection Criteria 422 Inverse Probability 427 Detector Characteristics 430

408

9.1

9.2
9.3

9.4
9.5

9.6 9.7 9.8 9.9

Performance of the Radar Operator Delay-line Integrators 445 Binary Integration 446 References 449

439

Contents
chapter
10.

IX

EXTRACTION OF INFORMATION FROM RADAR SIGNALS
Introduction

453

10.1

453

10.2
10.3

10.4
10.5

10.6
10.7 10.8 10.9

Phase and Amplitude Measurements 453 Review of Radar Measurements 455 Statistical Estimation of Parameters Likelihood Function 461 Theoretical Accuracy of Range and Doppler-velocity Measurements Uncertainty Relation 474 Angular Accuracy 476 Transmitted Waveform 482 Pulse Compression 493 498 References

462

CHAPTER

11.

PROPAGATION OF RADAR WAVES
Introduction
501 Propagation over a Plane Earth 501

501

11.1

11.2
11.3

The Round Earth
Refraction

506

11.4
11.5

506

Anomalous Propagation

11.6
11.7 11.8 11.9

509 Low-altitude Coverage 512 Radar Diffraction Screen 516 Attenuation by Atmospheric Gases Microwave-radiation Hazards 518 519 References

517

chapter

12.

CLUTTER, WEATHER,

AND INTERFERENCE

521

12.1

12.2
12.3

12.4
12.5

12.6
12.7 12.8
12.9

Introduction 521 Ground Clutter 522 Sea Clutter 527 Clutter Reduction 534 Meteorological Echoes 539 543 Attenuation by Precipitation Visibility of Targets in Weather Clutter Angels 551
Interference

545

554
559 567

12.10

ECM

and

ECCM

References

CHAPTER

13.

SYSTEMS ENGINEERING
Systems Engineering
570

AND DESIGN
572 582

570

13.1

13.2
13.3

Radar Parameter Selection 571 Example Aircraft-surveillance Radar

13.4
13.5

ASDE

13.6
13.7

579 Airborne Weather-avoidance Radar Bistatic Radar 585 Radar Beacons 594
References
601

CHAPTER

14.

RADAR DETECTION OF EXTRATERRESTRIAL OBJECTS
Introduction

603

14.1

603

14.2
14.3

14.4
14.5 14.6 14.7
14.8

624 Observation of Ionized Media Detection and Tracking of Earth Satellites and Space Vehicles 634 References

Radar Radar Radar Radar Radar Radar

604 Echoes from the Moon Echoes from the Planets 610 618 Detection of the Sun 619 Detection of Meteors 621 Observation of Auroras
628

Index

637

1

THE NATURE OF RADAR
1.1.

Introduction

is an electronic device for the detection and location of objects. It operates by transmitting a particular type of waveform, a pulse-modulated sine wave for example, and detects the nature of the echo signal. Radar is used to extend the capability of man's senses for observing his environment, especially the sense of vision. The value of radar lies not in being a substitute for the eye, but in doing what the eye cannot do. Radar cannot resolve detail as well as the eye, nor is it yet capable of recognizing the "color" of objects to the degree of sophistication of which the eye is capable. However, radar can be designed to see through those conditions impervious to normal human vision, such as darkness, haze, fog, rain, and snow. In addition, radar has the advantage ofbeing able to measure the distance or range to the object. This is probably its most important attribute.

Radar

Radar

Antenna

Fig.

1.1.

Block diagram of an elementary form of radar.

An elementary form of radar, shown in Fig. 1.1, consists of a transmitting antenna emitting electromagnetic radiation generated by an oscillator of some sort, a receiving antenna, and an energy-detecting device, or receiver. portion of the transmitted signal is intercepted by a reflecting object (target) and is reradiated in all directions. It is the energy reradiated in back direction that is of prime interest to the radar. The receiving antenna collects the returned energy and delivers it to a receiver, where it is

A

velocity.

processed to detect the presence of the target and to extract its location and relative The distance to the target is determined by measuring the time taken for the radar signal to travel to the target and back. The direction, or angular position, of the target may be determined from the direction of arrival of the reflected wavefront. The usual method of measuring the direction of arrival is with narrow antenna beams. If relative motion exists between target and radar, the shift in the carrier frequency of the reflected wave (doppler effect) is a measure of the target's relative (radial) velocity and may be used to distinguish moving targets from stationary objects. In radars which continuously track the movement of a target, a continuous indication of the rate of change of target position is also available. The name radar reflects the emphasis placed by the early experimenters on a device to Radar is a contraction of the detect the presence of a target and measure its range. words radio detection and ranging. It was first developed as a detection device to warn of the approach of hostile aircraft and for directing antiaircraft weapons. Although a well-designed modern radar can usually extract more information from the target
1

2

Introduction to Radar Systems

[Sec. 1.1

signal than merely range, the measurement of range is still one of radar's most important functions. There seem to be no other competitive techniques which can measure

range as well or as rapidly as can a radar. Radar was the code word officially adopted by the United States

Navy in November,

1940, as the designation for what had previously been called, among other things' radio echo equipment. The United States Army Signal Corps, which also did pioneer work in radar development, used the term radio position finding until it too adopted the

name radar in 1942. The following year radar was substituted by the British for their own term RDF. The origin of the R is obscure, but DF is supposed to stand for
which was purposely chosen to hide the fact that a range-measuring device was under development. Shortly after the term was coined, however, means were devised for also determining the angular position, so that almost immediately
direction finding,

RDF

usefulness as a code name. In France, radar was known as {detection electromagnetique), and in Germany it was called Funkmessgerat. It is
its

lost

some of

DEM
now

almost universally called radar. The most common radar waveform is a train of narrow pulses modulating a sinewave carrier. Although the pulse is normally rectangular in shape, it need not be, and could be one of many possible shapes. The distance, or range,f to the target is determined by measuring the time taken by the pulse to travel to the target and return. Since electromagnetic energy travels at the speed of light, the range R is

*
The

=


cAt

(l.D

velocity of light c is 3 x 10 8 m/sec, if R is measured in meters and A?, the time duration for the wave to travel out and back, is measured in seconds. One microsecond of round-trip travel time corresponds to a distance of 0.081 nautical mile, 0.093 statute mile, 1 64 yd, or 492 ft. The accepted unit of distance is the nautical mile
(n. mi.), which is equal to 6,076 ft, or 1,852 m. The radar range is also sometimes given in yards, especially for artillery or short-range missile fire control. In
is secondary to convenience, the radar radar mile is denned as 2,000 yd. The difference is less than 1 per cent. Once the transmitted pulse is emitted by the radar, a sufficient length of time must elapse to allow any echo signals to return and be detected before the next pulse may be transmitted. Therefore the rate at which the pulses may be transmitted is determined by the longest range at which targets are expected. If the pulse repetition frequency were too high, echo signals from some targets might arrive before the transmission of the next pulse, and ambiguities in measuring range might result. Echoes that arrive after the transmission of the next pulse are called second-time-around (or multipletime-around) echoes. Such an echo would appear to be at a much shorter range than the actual and could be misleading if it were not known to be a second-time-around echo. The range beyond which targets appear as second-time-around echoes is called the maximum unambiguous range and is

some

instances,

when measurement accuracy

mile

used as a unit of range. between it and the nautical mile
is

A

d Kunamb

=


c

(1.2)

where f r = pulse repetition frequency, in cycles per second. A plot of the maximum unambiguous range as a function of pulse repetition frequency is shown in Fig. 1.2.
t Range and distance to the target are used synonymously in radar parlance although, in artillery usage, range is the horizontal projection of the distance. For aircraft targets, slant range is sometimes used to represent the distance from radar to target, and ground ranee is used for the projection of the J slant range on the ground.

Sec. 1.2]

The Nature of Radar

3

Although most radars transmit a pulse-modulated waveform, there are a number of other suitable modulations that might be used to fulfill the functions of target detection and location. An example of a very important type of radar which does not use a altimeter predates the application Although the altimeter. pulsed carrier is the of radar and is not universally considered a radar, it nevertheless operates on the radar transmissions Even simple unmodulated principle with the ground as the target. have found application in radar. The most familiar is probably the radar speedometer, in widespread use by many highway police departments to enforce automobile speed

FM

FM

CW

limits.

A

radar employing an unmodulated
10,000

CW

transmission utilizes the doppler

TTTR

10

100

1,000

10,000

Pulse repetition frequency, cps

Fig. 1.2. Plot of

maximum unambiguous

range as a function of the pulse repetition frequency, based

onEq.

(1.2).

presence of moving targets. The doppler effect causes the signal reflected by a moving target to be shifted in frequency by an amount
effect to detect the

/*

=

where

d

vr

X
1.2.

= = =

y
target,

d- 3 )

doppler frequency, cps
relative velocity between radar and wavelength of carrier frequency,

m/sec

m
t,

The Radar Equation

power of the radar transmitter of Fig. 1.1 is denoted by P and if an omniis used, that is, one which radiates uniformly in all directions, the power density (power per unit area) at a distance R from the radar is equal to the trans2 mitter power divided by the surface area AttR of an imaginary sphere of radius R, or
If the

directional antenna

Power density from omnidirectional antenna

=

Pt

477R

2

(1.4)

Radars usually employ directive antennas, instead of omnidirectional antennas, to
channel most of the radiated power

P

t

into

some

particular direction.

The gain

G

t

of

4

Introduction to Radar Systems
is

[Sec. 1.2

a measure of the increased power radiated in the direction of the target as compared with the power that would have been radiated from an isotropic antenna. It may be denned [Eq. (7.6)] as the ratio of the maximum radiation intensity from the subject antenna to the radiation intensity from a lossless isotropic antenna with the

an antenna

same power
mitting gain

input.

The power

density at the target from an antenna with a trans-

G

t

is

Power density from

directive antenna

Pc *— =—
reradiates

(\ 5)

The

target intercepts a portion of the radiated
(1.6)].

power and

it

in the direction

of the radar [Eq.

Power reradiated

in target direction

=

PCn
*
'

(1.6)

the radar cross section of the target and has the dimensions of area. a characteristic of the target and is a measure of its size as seen by the radar. The power density in the echo signal at the radar receiving antenna is then
It is

The parameter a is

Power

density of echo signal at radar =4

7(47r/? 2 ) 2

——— \
)

.

.

(11) }
^

The radar antenna captures a portion of the echo power. of the receiving antenna is A r the echo power P received r
,

If the effective capture area

at the radar

is

p

Pt \j r A r <j G si a
t

(477/? )

2

2

This is the fundamental form of the radar equation. Note that the important antenna parameters are the transmitting gain and the receiving area. Antenna theory gives the relationship between antenna gain and effective area as

ArrA t

AttA t

where the subscripts r and t refer to the receiving and transmitting antennas, respectively. If a common antenna is used for both transmission and reception (as is usually the case), the reciprocity theorem of antenna theory states that G = G = G and A t = A r = A " r Using these relationships, Eq. (1.8) becomes
'.

t

PtK*
4ttA

or

Pr

= (4tt) 3 7Z^
t

P G 2tfo
/?
4

(1-106)

The maximum radar range i?max is the distance beyond which the target can no longer be detected. It occurs when the received echo signal P just equals the minimum r detectable signal Smin Therefore
.

R,

(_PjAbJ
\4TrA 2 Smin mi 7

J
'

(1.11a)

~

or

R„

P G2X2 a
t

(1.11ft)

.(AirfSram-

the TR switch reconnects the receiver to the antenna. Consider the box labeled "timer." in the upper right side of the figure. 1. (There are some cases. If the receiver were not disconnected and if the transmitter power were sufficiently large. The timer. however. For example.3. for instance. which is also called the trigger generator. Video amplifier LO Fig. Radar Block Diagram and Operation typical pulse radar using an oscillator such as the magnetron for the transmitter may be described with the aid of the block diagram shown in Fig. A typical radar used for the detection of conventional aircraft at ranges of 100 or 200 employ a peak power of the order of 1 to 10 Mw. Because of the implicit nature of relationships between the parameters that appear in the radar equation. A A . and a pulse repetition frequency of several hundred pulses per second. In practice.3. The modulated RF pulse generated by the transmitter travels along the transmission line to the antenna. and Eq. when anomalous propagation or subrefraction effects occur. from Eq. it is usually found that the observed maximum radar ranges are different from those predicted with the simple radar equation ( 1 1 1 a) or ( 1 1 1 b). a pulse width of several microseconds.1 la) would indicate a /H relationship. the receiver might be damaged. On the other hand. The above simplified versions of the radar equation do not adequately describe the performance of practical radars. common antenna is usually used for both transmitting and receiving. (1.116) are two forms of the radar equation which describe range performance. or the synchronizer. These timing pulses turn on the modulator which pulses the transmitter. the timer is of more modest proportions and only has to trigger the grid of a vacuum tube or miles might thyratron. .Sec.8) shows range independent of wavelength. Many important factors that affect range are not explicitly included. generates a series of narrow timing. 1. 1. pulses at the pulse repetition frequency. On the other hand. Although the timer and the modulator both are switches. Eq.) There are many reasons for the failure of the simple radar equation to correlate with actual performance. fast-acting switch called the transmit-receive (TR) switch disconnects the receiver during transmission. 2. where larger ranges might result. or trigger. Block diagram of a pulse radar.11a) and (1.3] The Nature of Radar 5 Equations (1. (1.1 lb) it might be thought that the range of a radar varies as A*. After passage of the transmitted signal. The modulator must be capable of switching the high-power transmitter and might be a rather large device. Actual ranges are often much smaller than predicted. where it is radiated into space. as discussed in Chap. they are shown as separate boxes in the block diagram since different considerations enter into their design. . one must be careful about making generalizations concerning radar performance on the basis of these equations alone. (1. 1. Duplexer The operation of a JL ATR Transmitter A Modulator Timer w» RF amplifier » Mixer A IF amplifier Det.3.

The radar receiver is usually of the superheterodyne type. Additional devices' might include a means for automatically compensating the receiver for changes in radar frequency possible. Monitoring devices (not shown) are usually employed to ensure that the radar is operating properly. and means for allowing the antenna to automatically track a moving target. Many variations are Furthermore. The block diagram of Fig.4. The PPI maps the target in angle and range on a polar display. where it belongs. 1. a traveling-wave tube. 1. RF amplifier and use or front end. In the absence of the A ATR ATR.4b).4a) and the plan position (deflection modulation) (b) PPI The A-scope displays the presentation displaying range vs. The RF amplifier shown as the first stage of the superheterodyne might be a low-noise parametric amplifier. Many microThe mixer signal to wave radar receivers do not have an the mixer as the -4r.6 Introduction to Radar Systems [Sec. If separate antennas are employed for transin. rotary joints in the transmission lines to allow movement of the antenna. information is shown. or a maser. cator. indicator. simple but important monitoring device is a directional coupler inserted in the transmission line to sample a fraction of the transmitted power. beam intensity (z axis) as the electron beam is made to sweep outward from the center with range. (a) A-scopepresentation most commqn forms of indicators using cathode-ray displaying amplitude vs. The and the are together called the duplexer. The two Fig. usually a cathode-ray tube A A (CRT). A A common form of radar antenna is a reflector with a parabolic shape fed from a . Target amplitude is used to modulate the electron. acts on reception to channel the received signal power into the receiver. The RF pulse modulation is extracted by the detector and amplified by the video amplifier to a level where it can operate the indicator. range tubes are the A-scope (Fig. (AFC) or gain (AGC). 1 . and local oscillator (LO) convert the RF an intermediate frequency (IF) since itis easier to build highgain narrowband amplifiers at the lower frequencies. 1. which has no effect during the transmission portion of the cycle. typical IF amplifier might have a center frequency of 30 or 60 Mc and a bandwidth of 1 or 2 Mc. and no angle angle (intensity modulation). 1. target amplitude (y axis) vs. first stage. the diagram is by no means complete since it does not include devices normally found in most radars. Timing signals are also supplied to the indi- Target positional information is obtained from the direction of the antenna and is used to properly display the coordinates of the target location. circuitry for discriminating between moving targets and stationary objects (MTI). The beam rotates in angle in response to the antenna position. a portion of the received power would be dissipated the transmitter rather than enter the receiver.3 portion of the radiated power is reflected by the target back to the radar and enters the receiver via the same antenna as used for transmitting. a duplexer may not be necessary if the isolation between the two separated antennas can be made sufficiently large. range (x axis). reflex klystron is commonly employed as the local oscillator. The (anti-transmitreceive) switch.3 is only one version of a radar. The output from the directional coupler may be used as a measure of the transmitted power or to test the fidelity of the transmitted waveform. TR ATR mitting and receiving. or PPI (Fig. many receiver circuits for reducing interfering or unwanted signals.

.5.4] The Nature of Radar point source. Broadcast band L *tter designo ions II L 1 Bond B Band 9 Band 10 Band 1 Band 12 Infr S C X Ku Ka 1 Aud o frequenci s Microwave region Video frequencies 1 1 30cps 300 cps 3kc 30kc 300 kc 3Mc 30 Mc 300 Mc 3Gc 30Gc 300 Gc 3. The early radar developers were forced to design their equipments to operate at the lower frequencies. The CH (Chain Home) radars employed by the British to provide early warning against air attack during World War II operated at a frequency in the vicinity of 25 Mc. The parabolic reflector focuses the energy into a narrow beam just as does an optical searchlight or an automobile headlamp.. and the lower the frequency. 1. Very few modern radars are found below 200 or above 35. both in terms of dollars and frequency allocations. probably out of habit and size. Early in the development of radar. The place of radar frequencies in the electromagnetic spectrum is shown in Fig. and take advantage of ionospheric reflections. Although higher transmitter powers are usually easier to achieve at the lower frequencies.4. They tend to group into separate bands for reasons of economy. 1. The radar region is shown extending from about 25 to 70. 1. Radar frequencies are not found over the entire frequency region. about 2 to 20 Mc.000 Mc. the broader will be the beamwidth for an aperture of a beamwidth can be obtained with a paraAt 25 Mc. for the rather compelling reason that suitable components were not available at higher frequencies. The antenna beamwidth is inversely proportional to the size of the antenna aperture (measured in wavelengths). a spread of more than 1 1 octaves. This is a very low radar frequency by modern standards. a letter code such as S.000 Mc.000 Gc Frequency Fig. An exception to this are radars that operate at high frequency (HF).5. .000 Mc. known as the microwave region. at 70. Radar Frequencies Conventional radars have been operated at frequencies extending from about 25 to 70. Radar frequencies and the electromagnetic spectrum. . Radar frequencies ''/////////.000 Mc. Although its original purpose was to guard military secrecy. etc. a 1 ft 1° bolic-reflector antenna approximately in diameter. the designations were carried over into peacetime use. These are not necessarily the limits since radars can be operated at frequencies outside either end of this range. L. 1. given For example.— * 1 Sec. Some of the various nomenclature employed to designate the various frequency regions is also shown. an antenna diameter of more than \ mile would be necessary to achieve the same beamwidth Considerations such as this stimulated the development of components and techniques at the higher radio frequencies. reasonable size are not suitable for most applications. was employed to designate radar frequency bands. The beam may be scanned in space by mechanically pointing the antenna. the poor angular accuracy and poor resolution which result with antennas of Wovelength 1km -* 100m 10m * «— 10cm VLF Very low frequency i *- <— LF Low — * «— MF—* *— HF High frequency VHF—> — UHF— Ultrahigh — SHF—> Super high frequency Cenfi metric <-- EHF^high Medium frequency frequency Very high frequency Extremely frequency frequency Millimetric My riomet ric waves Kilometric Hectometric Decometric Metric Deci metric woves waves waves waves waves waves waves Decimillimetric waves Band 4 Bond 5 Bond 6 Bond 7 '/////////. X.

5-40 Gc UHF L S C X K„ K K„ Millimeter >40 Gc The "band" method for designating frequency as adopted by the CCIR (Comite Consultatif International Radio) in 1953 is also shown in Fig. The more commonly used letter Although these are a convendesignations are indicated in Fig. For instance.5 the need for some convenient short nomenclature.5 are the audio frequencies.1 Frequency 300-1. These are taken to be the range of frequencies that may be displayed on a cathode-ray tube.000-8.000-4. In radar parlance.000 Mc 4. 1. Two other methods of naming frequency bands shown in Fig. they have no official status and there is not always general agreement as to the limits associated with each band.5 are based on frequency subdivisions and metric subdivisions.8 Introduction to Radar Systems [Sec. experimentally tested the theories of II. The video-frequency range is quite arbitrary. and highly directive antennas. The lower limit of microwaves is shown as 300 Mc since waveguide components and power klystron amplifiers are commercially available at The upper end of the microwave region is difficult to specify. The number of the exponent of 10 which expresses the upper frequency limit designates the band in question. 1 is .5 and in Table 1. The transition between the microwave region and the lumped- UHF constant region is not sharp. World War Maxwell and demonstrated the similarity between radio and light waves. rather than lumped-constant. The microwave region is that frequency region where distributed-constant.000-12.000 Mc 8. Their use is not very precise. Hertz showed that radio waves could be reflected by metallic and dielectric bodies. It is t Much of the material in this section concerning the early development of United States radar based on an unpublished report by Guerlac. 1.000 Mc.000 Mc 2. but this frequency.5. beyond the millimeter region. although it can be considered to extend even higher since frequencies of several thousand megacycles or more may be displayed on cathode-ray tubes. For example. the designation ultrahigh frequency (UHF) usually refers. circuits are employed. ient form of nomenclature.500 Mc 12. cavity resonators. to frequencies from about 300 to about 1.000-2. in practice. Also shown in Fig.000 Mc. in 1886. microwave techniques are more profitably replaced by optical techniques. Heinrich Hertz. 1 . The video frequencies are also indicated.5-18 Gc 18-26.000 Mc 1. It extends from zero frequency to the order of several megacycles in most radar and television applications. which may be defined as the range of frequencies audible to the normal human ear. 1. teristic of the microwave region is that the size of the components is comparable with the wavelength.5 Gc 26. Table Radar frequency band 1.5. 1. 1.1. and they define only general areas. History of Radar Development! full-fledged technology did not occur until Although the development of radar as a the basic principle of radar detection is almost as old as the subject of electromagnetism itself. the band extending from 3 x 108 to 3 x 109 is band 9. L or 5 band would be used to designate the UHF frequencies above 1. The frequency "band N" extends from 3 x 10* _1 to3 x lO^cps. Examples of distributed-constant The characdevices are waveguides.

if Although Marconi predicted and successfully demonstrated radio communication between continents. The first application of the pulse technique to the measurement of distance was in the basic scientific investigation by Breit and Tuve in 1925 for measuring the height of the ionosphere. In 1903 a German engineer by the name of Hiilsmeyer experimented with the detec- waves reflected from ships. In some of my tests I coming across a metallic object. Today. immediately reveal the presence and bearing of the other ship in fog or thick weather. 2 His methods were demonstrated before the German Navy. 13.Sec. but generated little interest. would be reflected back to a receiver screened from the local transmitter on the sending ship. later work in radio engineering was almost entirely at longer wavelengths. This type of radar originally was called wave-interference radar. A.5] interesting to note that although Hertz's experiments The Nature of Radar were performed with 9 relatively short wavelength radiation (66 cm). Hyland of the Naval Research Laboratory. The shorter wavelengths were not actively used to any great extent until the late thirties. or flutter. 4 However. One was the radar detection mentioned above. Marconi recognized the potentialities of short waves for radio detection and strongly urged their use in 1 922 for this application. Apparently Marconi's suggestion stimulated A. He also suggested that radio waves be used for the transfer of power from one point to the other without the use of wire or other transmission lines. such a radar is called a bistatic radar (Sec. by L. Successful pulse radar had to await the development of suitable components. and a better understanding of pulse receivers. electric waves can be completely reflected by conducting have noticed the effects of reflection and detection of these waves by metallic objects miles away. and his detection technique was dismissed on the grounds that it was little better than a visual observer. Taylor and L. he said 3 tion of radio : Hertz. observed in an ordinary television — CW A CW on weak stations when an aircraft passes overhead.6). and thereby. The wavelength was proposal was submitted for further work but was not accepted. The first experimental radar systems operated with and depended for detection upon the interference produced between the direct signal received from the transmitter and the doppler-frequency-shifted signal reflected by a moving target. which rays. first detection of aircraft using the wave-interference effect was It made in June. especially high-peak-power tubes. This effect is the same as the rhythmic flickering. It seems to me that it should be possible to design apparatus by means of which a ship could radiate or project a divergent beam of these rays in any desired direction. 5 m. Young of the Naval Research Laboratory to confirm experimentally the speculations concerning radio detection. the other was the suggestion that very short waves are capable of propagation well beyond the optical line of sight a phenomenon now known as tropospheric scatter. 1. such as another steamer or ship. C. H. The first experimental detections of aircraft used this radar principle rather than a monostatic (single-site) pulse radar because equipment receiver. The 1930. first As was shown by bodies. The state of technology at that time was not sufficiently adequate to obtain ranges of more than about a mile. he was apparently not successful in gaining support for some of his other ideas involving very short waves. He obtained a patent in 1904 in several countries for an obstacle detector and ship navigational device. more than a decade was to elapse before the detection of aircraft by pulse radar was demonstrated. In a speech delivered before the Institute of Radio Engineers. especially CW CW CW was readily available. In the autumn of 1922 they detected a wooden ship using a wave-interference radar with separated receiver and transmitter. 1 was made accidentally .

The problem of extracting target-position information from such radars was a difficult one and could not be readily solved with the techniques existing at that time. an effort was started at N RL in the spring of 934 to develop a pulse radar. Strange as it may now seem. less than three months after the start of the project. 1939. The first echoes at 200 Mc were received July 22. Ranges of 20 to 24 kiloyards were obtained on battleships and cruisers. The range was only 2\ miles. first The NRL to Guerlac. with a radar operating at a frequency of 28.5 while he was working with a direction-finding apparatus located in an aircraft on the ground. It was recognized that the limitations to obtaining adequate position information could be overcome with pulse transmission. in maneuvers held during January and February of 1939. This was never carried out. Nevertheless. 1936. 1935. The shortcomings were corrected. but he was not allowed to devote his full time CW 1 CW CW CW CW 1 to the effort.3 Mc and a pulse width of 5 ^asec. the necessary components did not exist. Hyland noted an increase in the received signal. The transmitter at a frequency of 33 Mc was located 2 miles away. for the purpose of obtaining some knowledge of distance and velocity.York. According first tests of the 60-Mc pulse radar were carried out in late December. 1 938. when several Bell Telephone Laboratories engineers reported the detection of aircraft during the course of other experiments. The range of the 200-Mc radar was limited by the transmitter. The XAF was tested aboard the battleship A'en. When aircraft passed through the beam. 1934. and Hyland 6 on a "System for Detecting Objects by Radio. These tests were "hopelessly unsuccessful and a grievous disappointment. Although the power delivered to the antenna was only 6 kw. The chief reason for this failure was attributed to the receiver's being designed for communications rather than for pulse reception. The work received low priority and was carried out principally by R. in the early days pulse radar encountered much skepticism. Early in wave-interference radar was demonstrated by NRL. It was realized by the NRL experimenters that higher radar frequencies were desired. By 1932 the equipment was demonstrated to detect aircraft at distances as great as 50 miles from the transmitter. Young. attempt with pulse radar at was at a frequency of 60 Mc. a 60-Mc The early wave-interference radars were useful only for detecting the presence of the target. A proposal was made by NRL in 1933 to employ a chain of transmitting and receiving stations along a line to be guarded. 28. lacking official encouragement and funds from the government. This radar was also the first to employ a duplexing system with a common antenna for both transmitting and receiving. 1 the CW NRL — — . and the first radar echoes obtained at using pulses occurred on Apr. especially for shipboard application. By early June the range was 25 miles. 5 The NRL work was disclosed in a patent filed and granted to Taylor. M. and met with considerable success. 934. Page. 1. a range of 50 miles the limit of the sweep was obtained by February. and the beam crossed an air lane from a nearby airfield. although it was fully supported by the NRL administration. where large antennas could not be tolerated. The range was only 10 to 12 miles. This occurred in January. The NRL work on aircraft detection with wave interference was kept classified until 1933. However. By October. but the work continued at a slow pace. however. In the spring of 1937 it was installed and tested on the destroyer Leary.10 Introduction to Radar Systems [Sec." No pulse echoes were observed on the cathode-ray tube. The success of the experiments at 28 Mc encouraged the NRL experimenters to develop a 200-Mc equipment." The type of radar described in this patent was a wave-interference radar. The development of higher-powered tubes by the Eitel-McCullough Corporation allowed an improved design of the 200-Mc radar known as X AF. This stimulated a more deliberate investigation by the NRL personnel. 1936. The limited ability of wave-interference radar to be anything more than a trip wire undoubtedly tempered what little official enthusiasm existed for radar. and early January.

1936.) But unfortunately.Watt concluded that this type of death ray required fantastically large amounts of power and could be regarded as not being practical at that time. the British were British interest in radar began about the possibility of producing a death ray using radio waves. he recommended that it would be more promising to investigate means for radio detection as opposed to radio destruction. it was possible to detect the beats to about an 8-mile range. was detected by an SCR-270. Watson. detection. the frequency-modulated aircraft radio altimeter was probably the first commercial application of the radar principle. The SCR-268 was used in conjunction with searchlights for radar fire control. Sec. when Sir Robert Watson-Watt was asked . This was necessary because of its poor angular accuracy. one of six Hawaii at the time. its range accuracy was superior to that obtained with optical methods. Instead. 1941 . by observing the beats between the echo signal and the directly received signal (wave interference). the spontaneous and independent development of pulse radar by several investigators in this country and abroad seems to make it difficult to assign sole credit to any one person for its origin. 1. By December of that year the Army tested its first pulse radar. Although it was not recognized as being a radar at the time. the British expended a large amount of effort on radar development. Nineteen of these an interest in radar during the Signal Corps work in pulse radar apparently resulted from a visit to in January. However.5 miles. The first operational radar early 1930s. available in 1 938. The transmitter and receiver were separated by about 5. The SCR-268 remained the standard fire-control equipment until January. In that same month the detection of an aircraft was carried out. CXAM. (The only available means for locating aircraft prior to World War II were sound locators whose maximum detection range under favorable conditions was about 20 miles. In 1939 the Army developed the SCR-270. The technique was similar to the first United States radar-detection experiments. well experienced in the military applications of radar. a former director of the Signal Corps Laboratories.Sec.) WatsonWatt was allowed to explore the possibilities of radio detection. The early developments of pulse radar were primarily concerned with military applications. 1935. By the time the United States entered the war. 3. using 6-Mc communication equipment. obtaining a range of 7 miles. 1944. a long-range radar for early warning.5] The Nature of Radar 11 orders were placed for a manufactured version called the radars were installed on major ships of the fleet by 1941. he issued two memoranda outlining the conditions necessary for an effective radar system. A modified SCR-270 was also the first radar to detect echoes from the moon in 1946. Blair. The first equipments were operated in aircraft as early as 1936 and utilized the same principle of operation as the FM-CW radar described in in In the case of the radio altimeter. when it was replaced by the SCR-584 microwave radar. The attack on Pearl Harbor in December. the target is the ground.3. In Britain the development of radar began later than in the United States. and in February. in early 1935. 9 " 12 But because they felt the nearness of war more acutely and were in a more vulnerable position with respect to air attack. The basic patent 8 describing the prototype of the SCR-268 was awarded to Colonel William R. 1 (There were also 16 SCR-268s assigned to units in Honolulu. 7 The United States Army Signal Corps also maintained The beginning of serious NRL used for antiaircraft fire control was the SCR-268. The claims contained in this patent apparently cover most of the basic ideas inherent in pulse-echo radio ranging and Although Colonel Blair's patent may legally make him the originator of pulse radar. When the aircraft receded from the receiver. The SCR-584 could control an antiaircraft battery without the necessity for searchlights or optical angle tracking. the true significance of the blips on the scope was not realized until after the bombs had fallen.

the CH radar stations began 24-hour duty. The success of microwave radar was by no means certain at the end of 1940. in that month. The former phenomenon was quickly exploited for the detection and location of surface ships and submarines. the first radar measurement of the height of aircraft above ground was made by measuring the elevaIn March. Most of the stations were operating by September. 1934. which continued until the end of the By June. increased to 90 miles and the frequency was raised to 25 Mc. an aircraft-interception radar (AI). by 1939. Also.12 Introduction to Radar Systems [Sec. primarily the very high frequency (VHF) band. and Russia apparently did not late thirties.5 an aircraft target. The Randell and Boot magnetron operated at a wavelength of 10 cm and produced a power output of about 1 kw. 1940. . 1938. mounted on an aircraft. they developed. under the administration of the Massachusetts Institute of Technology. 1935. were not that ground-based search radars such as guide fighter aircraft to a complete interception at night or in bad Consequently. The Al radar operated at a frequency of 200 Mc. ranges greater than 40 miles were obtained on bomber aircraft. for the detection and interception of hostile aircraft. In September of that year a British technical mission visited the United States to exchange information concerning the radar developments in the two countries. undertake the development of a microwave AI radar and a microwave antiaircraft The British technical mission demonstrated the cavity-magnetron fire-control radar. The exploration of the microwave region for radar application became the responsibility of the Radiation Laboratory. radar was developed independently in France and Germany during the middle and Other countries such as Japan. where techniques and components were more readily available. the range of detection had tion angle of arrival of the reflected signal. and plotted the track of the aircraft which flew Neville Chamberlain. During the development of the AI radar it was noted that radar could be used for the detection of ships from the air and also that the character of echoes from the ground was dependent on the nature of the terrain. In the same month. 14 Italy. The latter effect was not exploited initially. power tube developed by Randell and Boot and furnished design information so that it could be duplicated by United States manufacturers. 13 radar system was operated against aircraft in December. the British NRL CH war. (Chain Home) radar stations at a frequency of 25 Mc were successfully A series of demonstrated in April. The frequency was 12 Mc. Guerlac 1 states that this Mc attempt was not successful. The British realized the advantages to be gained from the better angular resolution possible at the microwave frequencies. 1. to Munich to confer with Hitler and Mussolini. Until the middle of 1940 the development of radar in Britain and the United States was carried out independently of one another. weather. In addition to the developments carried out in the United States and Great Britain. t Schooley points out that a 60at the Naval Research Laboratory. had demonstrated the pulse technique to measure range of experiments This was almost a year sooner than the successful with pulse radar. an improvement by a factor of 100 over anything previously achieved at centimeter wavelengths. the British Prime Minister at that time. but as indicated previously in this section.f By September. organized in November. but is now used for The British realized quite early CH sufficiently accurate to airborne mapping radars. enter the field of radar until they became aligned with either Germany or the Allied powers. 1936. The development of the magnetron was one of the most important contributions to the realization of microwave radar. 1937. They suggested that the United States especially for airborne and naval applications. Therefore the United States Service Laboratories chose to concentrate on the development of radars at the lower frequencies.

thus permitting better — — moving-target-indication (MTI) radars. microwave region. radar is still the only means of detecting and locating reflecting objects at long ranges and will continue to be used until a better substitute is the weapon. with the advent of Sputniks and intercontinental ballistic missiles. and the pace of radar development slowed considerably. The AN/CPS-6B. The AI radar was developed to the point where most ofthe functions of aiming and firing of weapons normally carried out by the pilot were taken over by the radar and computer. In the area of air defense most of the functions of recognizing and plotting aircraft tracks. the AN/FPS-3. During the thirties. most of the significant developments were carried out in the UHF trend. especially for long-range search radars. for the most part. Before this section on the history of radar development is closed. Inthe 1950s. Advances in crystal-mixer technology and in low-noise traveling-wave tubes improved the sensitivity of microwave receivers by an order of magnitude. The development of radar was sparked primarily by military needs. however.5] The Nature of Radar 13 At the close of World War II most of the scientists and engineers engaged in radar development returned to their normal peacetime pursuits. by the early fifties. there was a reversal ofthe upward frequency and a large amount of radar development was again carried out in the UHF region. radar has found many civilian applications. However. mention should be . and the AN/FPS-6 height finder were the primary radars used for long-range surveillance of aircraft in this country during that period of time. The need for accurate tracking arose mainly from the requirements of guided missiles. The further development of monopulse tracking radar also came about in this period. were carried out automatically by electronic digital computers such as those in the SAGE (Semiautomatic Ground Environment) system. found. The post-World War II radars were more accurate and of greater range capability than their wartime counterparts. Another advance during the fifties was the closer integration of the radar system to This was made possible primarily by the development of electronic computer techniques during this period. normally the function of an operator. However. especially in air and marine navigation. 1. Another component in which considerable advance has been made is the receiver. Radar technology is still in the process of growing. The high-power klystron amplifier was first developed not for radar application but for the linear accelerator at Stanford University. The parametric amplifier and the solid-state maser further improved receivers to the point where external noise and losses in the transmission lines are more important in determining receiver sensitivity than the device itself. During the forties. The advantage of the klystron amplifier over the magnetron the only other high-power tube used for microwave radar application up to that time is that klystrons are capable of greater power output than magnetrons and their stability is far better. One of the more important of these was the introduction of the high-power klystron amplifier. The integration of radar and weapon was even closer in the guided missile. This resulted in the development of radars with very high power transmitters and large antennas. In the late 1950s. Although it may leave much to be desired in many applications. the range required of radars was greatly increased over that required for aircraft detection. some new developments became available which increased the capability of radar. This is but one of many examples that illustrate how basic research can unpredictably contribute to the advancement of practical technology. The radars in operational use during the decade following the war were.Sec. The accuracy of tracking radars in the fifties was an order of magnitude better than those ofthe previous decade. radar development was restricted to frequencies at or lower. based on designs initiated during the war.

6 in nature. radar is used by ships. On the sea. especially in bad weather or with poor visibility. Applications of Radar Radar has been employed on the ground. Air-trafnc-control radar monitors air traffic in the vicinity of airports and en route between air terminals. or even higher. Airborne radar may be used to detect other aircraft. suggesting that the bat must make use of the frequency change to indicate the target distance. but rises to a maximum and then decays. Military Applications. or land vehicles. Commercial aircraft carry radar altimeters to determine their height above the ground and weather-avoidance radar to navigate around dangerous storms. 10. Shipboard radar may observe other will ships or aircraft. The ears of the bat act the same as an antenna to give the bat's radar directional properties. Radar has also been used as an aid in surveying over very large distances. Civilian Applications. especially tornadoes and hurricanes. 1. 17 The bat emits a series of ultrasonic pulses about 2 msec in width at a repetition frequency of the order of 10 to 20 cps under ordinary circumstances. depending upon the state of activity. In addition. on the average. as discussed in Sec. large and small. for navigation. just as do the FM radars described in Chap. In foul weather. 1. The chief use of radar outside of the military has been for navigation.14 Introduction to Radar Systems [Sec. however. and on the sea and undoubtedly be used in space. Even more remarkable is the fact that the bat's transmission is not a simple pulse but is more like a frequency-modulated pulse or an FM pulse-compression waveform. Bats have been observed to detect obstacles as close as 5 cm. Another interesting observation is that hundreds or even thousands of bats issue from caves in flight without apparent difficulty from mutual interference. Note that the length of a 2-msec ultrasonic pulse is 70 cm. in the air. One of the more important applications of radar is in the detection and tracking of weather disturbances. A large number of the civilian applications of radar mentioned above also apply to the military. radar is used with GCA (ground control of approach) systems to guide aircraft to a safe landing. The shape of the transmitted pulse is not exactly rectangular. the prf might be as high as 50 to 60 cps.9. In flight. Perhaps the application with which the reader has had most contact is the speedmeasuring radar used by many of the highway police.6. Examples of such radars are those in the DEW (Distant Early Warning) line for the detection of aircraft. or it may be used as a navigation aid to locate shore lines or buoys. for periods of time of the order of several seconds. The repetition frequency does vary. The frequency-modulated transmissions emitted by one species of bat start at a frequency of 78 kc and decay to 39 kc. both marine and air. Surveillance radars detect and locate hostile targets for the purpose of taking proper military action. the BMEWS (Ballistic Missile Early Warning System) radars for . ships. The porpoise and the bat are both known to use ultrasonic echo-locating principles similar to electromagnetic radar echo location or ^ ultrasonic sonar. The average frequency at the peak amplitude of the pulse is 48 kc. Ground-based radar has been applied chiefly to the detection and location of aircraft or space targets. 15 17 made of "radar" found The ordinary bat contains a built-in ultrasonic "radar" enabling him to fly through dark caves with impunity and find and catch insects on the wing for food. or it may be used for storm avoidance and navigation. A bat at rest might emit pulses at a rate of 5 to 10 cps. radar is used by the military for surveillance and for the control of weapons. 3. The nature of the vehicle that carries the radar and the environment in which it operates have a significant influence on its design. It is found experimentally that the intensity of the emissions is much reduced if the bat's head is pointed 45° or more from the normal (assuming that the ultrasonic receiver is observing the bat head on). especially radar navigation.

Fig. Portable surveying radar MRA-2 Tellurometer system.6. 1.7. Electronics Department. Inc. Heavy Military mm Fig.) . AN/FPS-24 long-range search radar. 1.) (Courtesy General Electric Co.. 15 (Courtesy Tellurometer.

AN/MPQ-10 mortar-detection radar. 1.9.Fig.8. (Courtesy Sperry Gyroscope Co. AN/SPG-49 Talos missile-tracking radars on board the U.) Fig.) . Galveston.S.S. 1. 16 (Courtesy Sperry Gyroscope Co.

6] detecting The Nature of Radar and tracking intercontinental . Fig. The Frescan radar shown in An example 7. radio astronomy.5) designed to be mounted in the nose of an aircraft. The use of radar measurement tool by research scientists has vastly increased our knowledge of meteorology. Examples of radars for the control of weapons include the acquisition radars and tracking radars of air defense systems such as those of Nike.7) is mounted on top. The antenna is fixed. 1. the exploration of interplanetary space.7). a small portable equipment capable of precisely measuring the distance RDR-1D system antenna. and the beam controlled electronically. 1.9. 1. aurora.6 to 1. meteors. shipboard surveillance radars. 1 1 1 is a three-dimensional (3-D) pencil-beam radar. pedestal.6) is The reflector. 12. used to guide the missile to the target. This large radar is contrasted with the Fig. is shown in Figure 1. ESAR (Fig. and feed horn weigh more than 135 tons. 7. The two smaller dish-shaped radars are the AN/SPW-2. The AN/FPS-24 (Fig. spoiler grid is shown in the upper half of the antenna to provide a cosecant-squared beam (Sec.9) for improved ground mapping. In addition. the AN/MPQ-10.) between two points (Sec. A beacon interrogating antenna (Sec. AI (airborne-interception) radar used to guide a fighter aircraft to its target. 13. 1. 1.8 shows two AN/SPG-49 missile-tracking radars mounted on board ship.7). 1. and other objects of the solar system.10 illustrates the 22-in. and the AEW Warning) radars. Figure 1. Their function is to automatically acquire and track targets for the Talos surface-to-air missile systems. The antenna beam is stabilized electronically to compensate for the pitch and roll of a ship at sea.12 and Sec. Radar can guide space vehicles and satellites and may be used for Scientific Applications. 13. while azimuth scanning is obtained by mechanical rotation of the antenna. 30-lbMRA-2 Tellurometer surveying radar (Fig. 17 of the SAGE ballistic missiles the long-range search radars (Airborne Early system.Sec.10) for the surveillance of aircraft. Some of the many varied shapes radars may take are illustrated in Figs. Fig. Its antenna is 120 ft wide and 36 ft high. and bombing as a radars. Elevation coverage is obtained with electronic frequency scanning (Sec. Examples. . homing radars on guided missiles.12. The sloping position of an electronically scanned array radar face of the building measures 50 by 50 is is ft. 1. A mortar-detection radar. airborne-weather-radar(Courtesy Bendix Radio.5). . the techniques and components developed for radar have been put to good use in such basic research as microwave spectroscopy.10. A 7. and radar astronomy. a large frequency-diversity radar (Sec. 3.-diameter antenna for the RDR-1D airborne weather radar system (Sec.7).

) 18 (Radar to the left rear is the AN/FPS-18 gap- .11. filler ESAR.S. electronically steered array radar. Galveston. 1.12.) {Courtesy Bendix Radio. Fig. 1. Frescan 3-D radar (Courtesy Hughes Aircraft Co.Fig.S. radar.) mounted on the masthead of the missile cruiser U.

. 21.. Schooley. 65. E. A. Am. and M. Proc." issued to W. pp. MIL-4. 10. 434-440. vol. V. Sir Robert "Three Steps to Victory. 9. 1958. pp. vol. Nov. British Patent 13.803.S. Acoust. Wireless World. A. issued to Christian Hulsmeyer.: "Echoes of Bats and Men. Wilkins. 6. : A Young. 370-377. England) and how TRE went about its business from 1935 to the end of World A War 12." Odhams Press. vol. Conn. Crawford. March. Aug. . 95-99. 37. Griffin. D. A. AugustSeptember. Rev. Ltd. A. Breit. 247-255. 1960.. September." issued to A. IRE. 1949. vol. New Haven.. II. 1933. 5. Wilkinson. 22. entitled "Hertzian-wave Projecting and Receiving Apparatus Adapted to Indicate or Give Warning of the Presence of a Metallic Body. 2. U. very 11. 1948. B. vol. "System for Detecting Objects by Radio. G." Yale University Press. 1946. R. R. 1934." The Dial Press.: Short Survey of Japanese Radar. L. pp. 28. D. Hyland. 1959. Eng. W. W..981. readable description of the history of radar development at TRE (Telecommunications Research Establishment. Radio Telegraphy. 8.. Phys. vol. Griffin. IRE Trans. 15. "Radar.S.: "One Story of Radar. Such as a Ship or a Train.: "OSRD Long History. Proc. Proc.. 1957 "The Pulse of Radar. 475-492. IRE. R. Vieweger. Inc. p. 14. Watson. November. 1946. . Elec. Division 14. pp. vol." vol. U. Origins of Radar: Background to the Awards of the Royal Commission. Patent 2.: The Story of Radar. L. R. A. Research (London). Mumford: Some results of a Study of Ultra-shortwave Transmission Phenomena. I. 4. vol. and 7..170. A. D. 554-575. 22. 1953. 1959. April. Sept..: Measurements of the Ultrasonic Cries of Bats.819. R. Patent 1. October. 555-561. London. and W. 1952. Griffin. Department of Commerce. Soc. 237. 1904. H. pp. p. S. Taylor. : 13. New York. R. pp. and pp. A. 1957. 3.Watt. "Object Locating System. Englund. no." Doubleday & Company. 4. C.The Nature of Radar 19 REFERENCES 1. 6. C." available from Office of Technical Services. 16. 10. New York.: "Listening in the Dark. pp. Guerlac. P.. 58.884." Marconi. 20. 27. J. Blair. 455^163. L. March. 405. 17." Cambridge University Press. October.: Pulse Radar History. IRE. Tuve: Test of the Existence of the Conducting Layer. H. 1922.: Radar in the Signal Corps. 1926. 1950. New York. in the Line of Projection of Such Waves.S. vol. U. G. F. Rowe.

1. (2. 1 Part of this discrepancy is due to the failure of Eq. If all those factors affecting radar range were known. The statistical nature of these several parameters does not allow the maximum radar range to be described by a single number. 2 minimum detectable signal. the simple radar equation does not predict the range performance of actual radar equipments to a satisfactory degree of accuracy. except for the target cross section a. 2 a . to make an accurate prediction of radar performance. if one is employed. the transmitted power must be large. This will be better appreciated as we proceed through the chapter and note the various factors that must be taken into account. the simple radar equation will be extended to include most of the important factors that influence radar range performance. collected with a large antenna aperture (also receiver must be sensitive to weak signals. Its specification must include a statement of the probability that the radar will detect a certain type of target at a particular range. the radiated energy must be concentrated into a where Rr. the received echo energy must be gain). The radar equation states that if long ranges are desired.1) = m = m narrow beam (high transmitting antenna gain). 1. the effort required to specify completely the effects of all radar parameters to the degree of accuracy required for range prediction is usually not economically justified. however. Unfortunately. A compromise is always necessary between what one would like to have and what one can actually get with reasonable effort. 1) but which have an effect on the radar performance are the meteorological conditions along the propagation path and the performance of the radar operator. Prediction of Range Performance The simple form of the radar equation derived in Sec. (2. 20 . Other statistical factors which do not appear explicitly in Eq. 1) to explicitly include the various losses that can occur throughout the system or the loss in performance usually experienced when electronic equipment is operated in the field rather than under laboratory-type conditions. in principle. In some cases the actual range might be only half that predicted. In this chapter. the quality of the prediction is a function of the amount of effort employed in determining the quantitative effects of the various parameters.= radar cross section. synonymous with high and the In practice. Another important factor that must be considered in the radar equation is the statistical or unpredictable nature of several of the parameters. watts t antenna gain Ae antenna effective aperture. But. watts Smin All the parameters are to some extent under the control of the radar designer.2 expressed the radar range R max in terms of radar and target parameters maximum P GA e a t Smin J transmitted power.: 2 THE RADAR EQUATION 2. The predicted values of radar range are usually optimistic. it would be possible. P — G= _(477) 2 (2. as is true for most endeavors. The minimum detectable signal Smin and the target cross section a are both statistical in nature and must be expressed in statistical terms.

but it cannot be so low that noise peaks cross the threshold and give a false indication of the presence of targets. The threshold level must be low if weak signals are to be detected. it would be more difficult to recognize its presence. Consider the output of a typical radar receiver as a function Threshold level Time Fig. One such if it is necessary to reproduce faithfully the shape of the input waveform. however. but it is possible to approach it with practical receiver circuits. This is called threshold detection. This is deliberate and is necessitated by brevity. Thus. If the signal were small. This is not the same as the concept of "impedance match" of circuit theory. This might represent one sweep of the video output displayed on an A-scope with the receiver gain turned all the way up to make the noise level visible. This is especially true if a human operator The makes the detection Detection is decision. The envelope has a fluctuating appearance caused by the random nature of noise.2] The Radar Equation 21 A complete and detailed discussion of all those factors that influence the prediction of is beyond the scope of a single chapter. the envelope would not generally exceed the threshold if noise alone were present. More detailed information will be found in some of the subsequent chapters or in the references listed at the end of the chapter. 1 is assumed to be from a matched-filter receiver (Sec.Sec. Typical envelope of the radar receiver output as a function of time.1 .1.2. if the threshold level were set sufficiently high. 2 . 2. a signal is assumed to be present. If a large signal is present such as at A in Fig. 2. technique is the least-square smoothing and prediction theory of Wiener. The voltage envelope of Fig. If the receiver output exceeds the threshold. based on establishing a threshold level at the output of the receiver. The weakest signal the receiver can detect is called the minimum detectable signal.2). Minimum ability Detectable Signal of a radar receiver to detect a weak echo signal is limited by the noise energy that occupies the same portion of the frequency spectrum as does the signal energy. The ideal matched-filter receiver cannot always be exactly realized in practice. it is greater than the surrounding noise peaks and can be recognized on the basis of its amplitude. 2. A nearly matched filter receiver for a radar transmitting a rectangular-shaped pulse is usually characterized by a bandwidth B approximately the reciprocal of the pulse width t. but would exceed it if a strong signal were present. For this reason many subjects appear to be treated only lightly. or The output of a matched-filter receiver is the cross correlation between the Br ph 1 Hence it does not preserve received waveform and the impulse response of the filter. 1). A matched filter is one designed to maximize the output peak signal to average noise (power) ratio. 9. of time (Fig. . The specification of the minimum detectable signal is sometimes difficult because of its statistical nature and because the criterion for deciding whether a target is present or not may not be too well defined. radar range will 2. 2. 2. Other receiver design techniques must be employed the shape of the input waveform.

(2.5) If the ratio of ouput (IF) signal-to-noise ratio (S minimum detectable signal Smin is that value of 5.01 1 .11 4 5 Double-tuned { Staggered triple Staggered quadruple Staggered quintuple 1 2 1 . E. .019 1.4b) noise figure may be interpreted. X Applies to a transitionally coupled double-tuned circuit or to a stagger-tuned circuit with two MIT New tuned circuits. usually at the output of the IF amplifier before the nonlinear Table 2. better noise figures occur at lower frequencies. L. No matter whether the noise is generated by a thermal mechanism or by some other mechanism. The receiver The available gain A". Equation (2.22 1.04 1 1 1 1.3.. Lawson and G. 1. r- _S IN t t S IN (2. 2. . p.16 1.048 1. noise out of ideal receiver at std temp T (2. Gaussian 065 t J. Rearranging Eq. as a measure of signal-to-noise-ratio degradation as the signal passes through the receiver. The noise N is measured over the linear portion of the receiver inputoutput characteristic. the total noise at the output of the receiver may be considered to be equal to the thermal-noise power obtained from an ideal receiver multiplied by a factor called the noise figure. 8. vol. A discussion of the additional noise sources given in Sec. =-- kT B n Ga N— — 2 noise out of practical receiver a .. ceivers.6) this kTQ B n For many radar applications only assumption is satisfactory.4a) may be rewritten as bandwidth Ga is . McGraw-Hill Book Company.. 1950. Uhlenbeck (eds. noise figure is shown to depend upon the configuration of the first few input stages and the frequency of operation.12 1. 8. kToK is the input noise Bn is that of the IF amplifier in most rethe ratio of the signal out SB to the signal in S and t in an ideal receiver.): "Threshold Signals. the input signal may be expressed as <j The _ kT B n F n S N o (2. However. of stages 1 Type of receiver coupling circuit Single-tuned Ratio of noise bandwidth to 3-db bandwidth 1 57 2 3 1. In Chap. according to the Institute of Radio Engineers definition." Radiation Laboratory Series..4a) N = Ga = noise output from receiver available gain The temperature T is taken to be 290°K. 177. The noise figure Fn of a receiver is defined by the equation important except to know is in nonideal receivers Fn where . . . second detector.3 that it exists. (2.46). 24 Introduction to Radar Systems [Sec. In general. corresponding to the minimum /N ) m m necessary for detection.1. York. then = kT B »'«1»j This assumes that the input receiver noise is / \Ay v </min m in . 24. Inc. it is strictly applicable when the receiver .. . therefore. Comparison of Noise Bandwidth and 3-db Bandwidth! No.14 1.

: Sec.4. With relatively noisy receivers. receiver noise. 2. but one cannot predict exactly what will occur for any particular event. If the event E occurs of the event E is the ratio mjn. and the probability of drawing any spade is 13/52 to l. the temperature seen by the receiver may be lower or higher than 290°K. 4 8 In this section we shall briefly review probability and the probability-density function and cite some examples. Phenomena of a random nature can be described with the aid of probability theory. the probability of drawing the ace of 1/13. range from x x to x 2 is found by integrating p{x) over the range of interest. Predictions concerning the average performance of chance events are possible by observing and classifying occurrences. 2. 8. (2.6). all of which are "" times out of a possible total of n. . is the effective noise temperature discussed in Sec. the probability ofdrawinganyaceis4/52 The scale of probability 1/4. When the receiver is connected to an antenna. or Probability (x x t Probabilities are < x < x 2) = p(x) dx to 1. the greater is its probability. It cannot be precisely predicted any more than one can predict the name of a card blindly drawn from a shuffled deck. (number of values total in range Ax at x)/Ax .9) sometimes expressed in per cent (0 to 100) rather than . equal segments of length Ax and count the number of times that x The probability-density function p{x) is then defined as . falls in each interval.6) into Eq. An ranges from impossible event is assigned the probability 0. (2. 1) results in the following form of the radar equation n4 __ maX PfiA e o (47rfkT B n F n (SjN <2 j. The intermediate probabilities are m = = assigned so that the more likely an event. Probability-density Functions The basic concepts of probability theory needed in solving noise problems may be found in any of several references.f An event which is certain is assigned the probability 1. Substituting Eq. the effect of an antenna temperature different from 290°K would hardly be noticed unless the temperature were high. ) mln Before continuing the discussion of the factors involved in the radar equation. Noise is a random phenomenon.5. lim Aa->o number of values =N (2.„ „.. (2. distance from a fixed reference point. p(x) = . For example. with low-noise receivers resulting from the use of the maser and the parametric amplifier. the An alternative description of effect of antenna temperature is important (Sec. it is necessary to digress and review briefly some topics in probability theory in order to describe the signal-to-noise ratio in statistical terms. However.4] The Radar Equation 25 input is at the standard temperature 290°K. The distance of x from the reference point might Divide the line into small represent the value of the noise current or the noise voltage. the probability equally likely. particular experiment in which there are n different possible outcomes. Consider a Probability is a measure of the likelihood of occurrence of an event.8) centered at x The probability that a particular measured value lies within the infinitesimal width dx The probability that the value of x lies within the finite is simply p{x) dx. 8. spades from a deck of 52 cards is 1/52. One of the more useful concepts of probability theory needed to analyze the detection of signals in noise is the probability-density function. Consider the variable x as representing a typical measured value of a random process such as a noise voltage or Imagine each x to define a point on a straight line corresponding to the current. especially useful when dealing with low-noise receivers.

4 definition. that is. or uniform. p(x). the probability-density function is positive. The meansquare value (m 2) of the current when multiplied by the resistance! gives the mean power. where k is a constant. The square root of the variance a is called the standard deviation and is the root-mean-square (rms) value of the a-c component. is The variance We shall consider three examples of probability-density functions. with k = 1 I2tt It also applies to the distribution of the round-off (quantizing) error in numerical com- The constant k may be found by applying Eq.13) The quantities m 1 and m 2 are sometimes called the first and second moments of the random variable x. = Ja k dx = 1 or k = b oo The average value of x is m 1_ 1 = t In noise theory it is | . with equal probability. . that is.12) and the mean-square value is 2 <x )av x 2 p(x) dx — CO (2. average. distribution describes the phase of a random sine wave relative to a particular origin of time. customary to take the resistance as ohm or the conductance as 1 mho. If x represents an electric voltage or current. the integral of the probability density over al 1 values of x must be By equal to unity.10) is variable function. If the random variable is a noise current. 2. m is the d-c comx ponent. the product of the variance and resistance gives the mean power of the a-c component. that is. that described by the probability- (<t>(x)) av = J— \ <f>(x)p(x)dx qo (2.10). and the Rayleigh.x dx ~b = 1 b a 2 J. the uniform Gaussian. It is the value read by a direct-current voltmeter or ammeter. The variance jx 2 is defined as ((x = a2 = - mi )\ v = v (X - mi ) 2 p(x) dx ~ =m 2 m\ = (x 2 ) av - {x)h (2. anywhere from to 2tt.26 Introduction to Radar Systems [S ec . The uniform probability-density (Fig.11) This follows from the definition of an average value and the The mean. the phase of the sine wave may be found. Since every measurement must yield some value. or probability-density function.14) QO the mean-square deviation of x about its mean and is sometimes called the second central moment.2a) is defined p(x)= ik [0 as fora<x<a for x < a and x > a + b putations. value of x is <x)av = m x =\ = m2 = •/ xp(x) dx (2. p(x)dx=l The average value of a density function. p(x) dx " (2. A rectangular. 2. The mean-square value of voltage times the conductance is also the mean power. is (2. <f>(x).

2. b 3 m9 — a m. It can be shown that condition normalizing satisfy the ij = J— CO xp(x) dx = x m2 = 2 x p(x) dx = x% + a2 ^2 =m —m = 2 1 a2 J— X (2. . 12 = standard deviation 2^3 a* (a) b Fig. (c) Rayleigh = . such as thermal noise or shot noise. Also. may be statistics. ab . a Gaussian representation is often more convenient to manipulate mathematically. and the parameters have been adjusted to ] is the exponential [ of Eq. The Gaussian density function has a represented by Gaussian bell-shaped appearance and is denned by p(x) = sjlna- exp -(x 2 x 2 ) (2. in noise theory. 2. provided that the contribution of any one quantity is The probability density of the .4] The Radar Equation 27 This result could have been determined by inspection.16) sum of a large number of independently distributed probability-density function no matter what the Gaussian approaches the quantities individual distributions may be. (d) Rayleigh (power). b . since probability density (Fig. (6) Gaussian.Sec.2. (2. x2 (voltage). mean-square. or 2 ™2 and the variance is = a +b [ Ja x — dx = . 2 a . 2 + . Examples of probability-density functions. 10). value is The second-moment. w (a) Uniform.2b) is one of the most important many sources of noise. The Gaussian. or normal. 2.15) 2o- where exp function.

The probability that the input-noise voltage will lie between x and x + dx is the d-c p(x) dx = . then no matter how high the threshold were set. Therefore the probability-density function for the output of a linear half-wave rectifier with a Gaussian noise input is < < P(y) dy = — 1 exp yjliraa 2a 2 a 2 _v — f. even though the electrons are emitted from the cathode with other than a Gaussian distribution. This is the central limit theorem.. which is exactly \. the probability of y is zero. The average value. As an example of the Gaussian density function. of the output y is (y)av = f 00 yp(y) dy 1 = -=- f°° y exp «'-<>° Jliraa Jo —v 1 f°° dy + -f2a"cr 2J-oo 2 \ y d(y) dy IT- ~ aa 2 2 ex P TT — 2 0=^ V2^ probability-density function is also of special interest to the radar systems engineer. The probability-density function of the input noise is assumed to be Gaussian with zero mean. It describes the noise output from a narrowband filter (such as the The Rayleigh . It is the reason that shot noise resulting from the impact of electrons upon the anode of a vacuum tube can be represented by a Gaussian distribution. If the noise at the input of the threshold detector were truly Gaussian. Another interesting property of the Gaussian distribution is that no matter how large a value x we may choose. or d-c component.4 not comparable with the resultant of all others. y 28 Introduction to Radar Systems [Sec. consider the problem of determining component at the output of a linear half-wave rectifier when the input is thermal noise. However. the probability diminishes rapidly with increasing x. 2. Since negative values of y are not permitted.dy + \d(y) dy 2 y > where d(y) erties : is the Dirac delta function n (impulse (imj function) and has the following propd(y) /•0 + e E = y 1 e ^ > 6(y) 1 dy = Note that/?(j) has both a continuous and a discrete part. there is always some finite probability of finding a greater value. there would always be a chance that it would be exceeded by noise and appear as a false alarm. and for all practical purposes the probability of obtaining an exceedingly high value of x is negligibly small and may be regarded as being almost impossible. — 1 exp —x— 2 V27ror 2a 2 dx — oo < x < is oo The output y of the half-wave rectifier for an input x y = ax x > between y and y dx when x > 0: j=0 x<0 will lie The probability that the rectifier output y > as the probability that x lies between x and x P(y) dy + dy is the same + = p(x) dx = = 1 exp ^iirao —f2a a i i —v 2 dy for x > The probability of obtaining y is the same as x 0.

17) \ <X >av/ The parameter x might represent a voltage. Most radar receivers used in conjunction with an operator viewing a CRT display meet this condition and may be considered envelope Either a square-law or a linear detector may be assumed since the effect on detectors. or average.5] The Radar Equation 29 IF filter in a superheterodyne receiver). with probability-density function given by p(v) dv = 1 exp y/2iry) —v2 2^o dv (2.5. where w represents power instead of voltage voltage squared. value of the If x 2 is replaced by w. Signal-to-noise Ratio In this section the results of statistical noise theory will be applied to obtain the signal-to-noise ratio at the output of the IF amplifier necessary to achieve a specified probability of detection without exceeding a specified probability of false alarm. (2. (assuming the resistance is 1 ohm). 2.Sec. 2.3.3). in turn. The Rayleigh density function is p(x) dx 2x =— — <X )av 2 / x2 2 exp I — \ I dx x > (2.19) In some cases.18). Another mathematical description of statistical phenomena is the probability in Fig. 9. but not so wide as to pass the high-frequency components at or near the intermediate frequency. The density function may be found from the distribution function 2.7). The noise entering the IF filter (the terms filter and amplifier are used interchangeably) is assumed to be Gaussian. as in Eq. one which rejects the carrier frequency but passes the modulation envelope.2c and for w The standard distribution function P(x). The second detector and video amplifier are assumed to form an envelope detector. and many kinds of clutter and weather echoes. that is. (2. and (x2 } av the mean. 2. x is shown deviation of the Rayleigh density is equal to the mean value (x 2 ) av in Eq. output signal-to-noise ratio thus obtained may be substituted into Eq. To extract the modulation envelope. the detection probability by assuming one instead of the other is small (Sec. 2. Envelope detector. Consider an IF amplifier with bandwidth BIF followed by a second detector and a video amplifier with bandwidth Bv (Fig.2d. is used in the radar equation.17) and w in Eq. (2. the video bandwidth must be wide enough to pass the low-frequency components generated by the second detector. defined as the probability that the value x is less than some specified value P(x)= J-oo p(x)dx or p(x) = —P(x) dx (2.18) \ w / where w is the average power. by differentiation. The IF omplifier Second detector Video omplifier (%) (By) Fig.6) to find the minimum detectable signal. which.20) . the distribution function may be easier to obtain from an experimental set of data than the density function. The video bandwidth Bv must be greater than 5 IF /2 in order to pass all the video modulation. the cross-section fluctuations of certain types of complex radar targets. A typical plot of the Rayleigh density function for in Fig. 2. (2.6). the Rayleigh density function is p(w) dw = — exp w I \ dw w > (2.

2.exp (filter — ) dR Equation lie (2.exp (- — 2 ) dR (2. T = ft . 11 A .24) gives the probability of a false alarm. If Gaussian noise were passed through a narrowband IF filter— one whose bandwidth is small compared with the mid-frequency— the probability density of the envelope of the noise voltage output is shown by Rice 9 to be p(R) dR = .4. where output.— I \ rp \ \ dR (2. The average time interval between crossings of the threshold by noise alone is defined as the false-alarm time\ 7}„.23) 2y> ! = exp (--^1 =Pfa \ 2wJ (2.Y-ao Um±YT JV* 1 v k =1 Tk is the time between crossings of the threshold VT by the noise envelope. (2. Since the probability of a false alarm is the probability that noise will cross the threshold.24) Whenever the voltage envelope exceeds the threshold. Eq. The false-alarm probability may also be defined as the duration of time the envelope is actually above the threshold to the total time it where Time Fig.30 Introduction to Radar Systems [Sec. Envelope of receiver output illustrating false alarms due to noise. t This definition differs from that given by Marcum. by definition.21) R is the amplitude of the envelope of the of the Rayleigh probability-density function. a target detection is considered to have occurred. (2.2 1 ) is a form The probability that the envelope of the and noise voltage will V2 between the values of Vx is * Probability (Vx <R< V = 2) ' \ . comparison of the two definitions is given by Hollis. 2.22) The probability that the noise voltage envelope will exceed the voltage threshold VT is Probability (VT <R< oo) = f °° JvT R R — exp . 10 who defines the false-alarm time to be the time in which the probability is J that a false alarm will not occur. or mean-square value of the noise voltage. when the slope of the crossing is positive.5 where p(v) dv is the probability of finding the noise voltage v between the values of v and v + dv. denoted P {iX . y> is the variance. and the mean value of v is taken to be zero.

(2.] Equating . the probability of a false alarm is 1. 2.4. for example.11 x 10~ 9 From Eq. 15 min Threshold-to-noise ratio VT/Z i// . as in Eq. 2.000 voltage.5.25) XT* where T. (2. . approximately the reciprocal of the bandwidth B. (2. Average time between false alarms as a function of the threshold level VT and receiver bandwidth B. the bandwidth of the IF amplifier were 1 and the average false-alarm time that could be tolerated were 15 min. 2.5. we get 1 Tfa expBif K (2. or A' Pfa = = kN l __ Wav <r*>.24) and (2. with V^j2\p as the abscissa.5] The Radar Equation 31 could have been above the threshold. d Fig. Eqs. which in the case of the envelope detector is B 1F [It does not matter in most applications whether the bandwidth is that defined by the half-power points or by noise considerations.3). A Mc .25).45 times the rms value of the noise 10.aB The average duration of a noise pulse is t k and Tk are defined in Fig. 2. y>„ is the mean-square noise voltage.26) 2y plot of Eq.26) is shown in Fig.24) the threshold voltage necessary to achieve this false-alarm time is 6. */av (2.b Sec. If. (2.

For a 1-Mc bandwidth. Errors in this approximation are 1 primarily because of the assumption that Br not serious in practice because of the exponential relationship between the threshold and & . Tu will be reduced by a factor n. a value of 10 log 10 ( Kf. These can be avoided only by good engineering design or by recognizing them as non-Gaussian noise and not ever. other sources of noise can enter the receiver and falsely cause an alarm to be excited. 2. An equivalent interpretation of the false-alarm probability is as follows. t . The formulation of the false-alarm probability given above is only approximate. Pia will be used in this text instead of n f How. Tr the pulse repetition the pulse repetition frequency. while a value of 10 log10 ( Kf. electric razors.77 db changes the false-alarm time by five orders of magnitude. the false-alarm probability will be reduced by the fraction of time the On the other hand. If n pulses are added \\n s the false-alarm probability is P fa together (integrated) so as to improve detection. exponential relationship between the false-alarm time 7>a and the threshold level VT results in the false-alarm time being sensitive to variations or instabilities in the threshold level. the number of independent decisions period. For example.. If the receiver were turned off (gated)for a fraction of time (as in atracking radar with a servo-controlled range gate or a radar which turns the receiver off during the time of transmission). intervals per pulse period t] = T jr r = r r.72 db results in a false-alarm time of 10. Thus far. signal. the data of Marcum 10 til . Next. Therefore the number of T{ll rfj T{Jr. where B is the bandwidth. so that instabilities which lower db results in the threshold slightly will not cause a flood of false alarms. On the average there will be one false decision out of n f possible decisions in the false-alarm The average number of possible decisions between false alarms is the falsetime rfa The number of decisions nf in time 7>a is equal to the number of range alarm number n . However.. the threshold level would probably be adjusted slightly above that computed by Eq..26). times the false-alarm time 7>„. that the false-alarm probability is the probability that a noise pulse will cross the threshold during an interval of time approximately equal to the reciprocal of the band6 width. consider a sine-wave signal of amplitude A to be present along with noise at the input to the IF . one independent channel. therefore. in theory. where t is the pulse width. In practice. the false-alarm probability due to Gaussian noise may be reduced to an insignificant level by the proper selection of the threshold. there are of the order of 10 noise pulses per second. the false-alarm probability will be increased accordingly. The specification of a tolerable false-alarm time usually follows from the requirements The desired by the customer and depends on the nature of the radar application. a receiver with only a noise input has been discussed. power-line surges./2y ) = Thus a change in the threshold 14.32 Introduction to Radar Systems [Sec. if the radar output consists of more than is not operative./2y ) = 12.000 hr. times the \jf number of pulse periods per second/.95 an average false-alarm time of 6 min. and/r = 1/7. of only 1 . microphonics. if the bandwidth were 1 Mc. (2. etc. possible decisions is n t \\TiA B as before.5 The reason for this quite small. -6 Hence the false-alarm probability of any one pulse must be small (< 10 ) if false-alarm The false-alarm probabilities of practical radars are is times greater than 1 sec are to be obtained. Such is the nature of Gaussian noise. the probability of false alarm. The probability of false alarm will then be Whenever appropriate. and Swerling34 pertaining to integration loss presented later in this chapter are given in terms of n f rather than P It has been shown that.26). Such sources of noise might be local-site noise due to ignition systems. these effects are usually not important since small changes in the probability of false alarm result in even smaller changes in the threshold level because of the exporeceiver nential relationship of Eq. In practice. Since t IjB. (2. is = f = ^ = = in the time Pfa = njn.

2. Although the receiver designer prefers to operate with voltages. threshold voltage VT \\f\ 2.Af — .5 is shown.l. Equation (2. A = and Eq. (2. an asymptotic expansion for / (Z) /„(Z) " « -L= JlirZ (l \ + —+••) 8Z / When the signal is absent.29) may be used to plot a family of curves relating the probability of detection to the threshold voltage and to the amplitude of the sine-wave signal. the probability- density function for noise alone. it is more convenient for the radar system engineer to employ power relationships. 2. -'('-" ^) 72^0 _ ' 2fo_ ^ 2 N/277(/l/ x/t/ o) X where the error function is 1 _ Vt~ A 4A z .21)] is plotted along with that for signal and The noise [Eq.exp y> X*±A)ljM) 2rp I dR (2. If VT \y\ is increased to reduce the probability of a false alarm. (2. 33 The frequency of the signal is the same as the IF midband frequency IF output of the envelope detector has a probability-density function given by 9 p s (R) The dR=* Wo exp (^ X*±^)JM) dR 2ip Q / (2 . R will exceed the predetermined threshold (\ VT The probability of detection Pa is therefore {" ->v t Pa = (" JvT Ps(R) dR = A .29) may be converted to power by replacing the signal-to-rms-noise-voltage ratio with the following: A _ fl signal amplitude _ V2(rms signal voltage) _ / \ ? signal power V power/ _ /2S^ rms noise voltage rms noise voltage noise \N 1 .28) \y) ' This cannot be evaluated by simple means. the probability of detection will be reduced also. 1+(VtSA 2/y> 2 -4) /Vu (2.29) defined as erfZ 2 L = -^ C u* du [ e~ graphic illustration of the process of threshold detection is shown in Fig. The probability that the signal will be detected (which is is the probability of detection) the same as the probability that the envelope ./ Sec. series approximation valid when RA/y) 1.6. and terms in A~ 3 and beyond can be neglected is 9 > — > (VT expi-^ . while the double-crosshatched area under the curve for noise alone represents the probability of a false alarm. (2.27) \ip / where I (Z) is the modified Bessel function of zero order and argument Z.21). A \R A\.27)] with Ajip\ 3. A = A = crosshatched area to the right of VT \\p\ under the curve for signal-plus-noise represents the probability of detection.27) reduces to Eq. defined by /o(Z)= For . The Radar Equation . Equation (2. The probability density for noise alone [Eq. (2.lo2^!7! is Z large. and numerical techniques or a series approximation must be used.5] filter.

1 J /%yy>K (^ 2 = 3) probability of a false alarm as a parameter. (2.998 0. ability of the false alarm and from Fig. desired false-alarm time For example.60 0.6)]. There are several interesting facts illustrated by Fig.7 This is the determines the signal-to-noise ratio signal-to-noise ratio that is used in the equation ..50 probability of detection. 2. Figure 2. The for minimum detectable signal [Eq. 5 i i 6 Wo" Fig.50 0.05 1 I L 16 18 10 12 14 (V/yl). the is C. 2..999 0.24)].999.99 0.9999 0. 2.7 db for 0.34 Introduction to Radar Systems Noise alone 0.30 0.80 0.il X 10~ 9 . Probability of detection for a sine and the probability of false alarm.1 db is required to yield a 0. is At first glance. 2.10 I0 false alarm 10 io I 'icr 12 0.6.3 0.7./2y by In (1/Pfa ) [from Using the above relationships. i 2*3 w$4<ZY///fr>>. (2.7 apply to a single radar pulse. suppose that the was 1 5 min and the IF bandwidth was 1 Mc.9995 0. 14.6 [Sec.7.90. signal-to-noise ratio. Both the false-alarm time and the detection probability are specified by the system requirements. The radar designer computes the prob- —*x. Ff.95 0. signal-to-noise ratios of Fig. Probability-density functions for noise alone and for signal-plus-noise. illustrating the process of threshold detection. wave in noise as a function of the signal-to-noise (power) .5 db for 0. 2.90 "D « O I" 0.2 0.98 .5 ~ 1 \ Threshold probability of detection plotted in Fig.7 as 0.5 We shall also replace Eq.7 indicates that a signal-to-noise ratio of 13.1 0. db 20 ratio Fig.995 0..4 -/ \ \ l/77-^Signol + noise a function of the signal-to-noise ratio with the 0. 2.70 0. it might seem that the signal-to-noise ratio required for detection higher than that dictated by 0. and 16.20 1 0. Thisgivesafalse-alarmprobabilityofl. 2.40 0.

This is rather small compared with that of a receiver designed to optimize the signal-to-noise ratio of a single pulse (which is of the order of a megacycle or so for radars with pulse widths in the vicinity of 1 //sec). All practical integration techniques employ some sort of storage device. For example. On the other hand. The discussion in this section is concerned primarily with integration performed by electronic devices in which detection is made automatically on the basis of a threshold (o crossing. or coherent. while integration after the detector is called postdetection. I year.05 sec. antenna scan rate. Many techniques might be employed for accomplishing integration. the signal-tonoise ratio would be increased to 15. One might be inclined to say that so long as the signal is greater than noise. integration.7 is that a change of only 3. 2. and antenna scan rate 5 rpm (30°/sec). 2. deg pulse repetition frequency.2 db.9. 2.) If.6] The Radar Equation 35 intuition.Sec. or noncoherent. is An example of a predetection integrator a narrowband IF filter with a bandwidth approximately equal to the reciprocal of the time on target. Perhaps the most common radar integration method is the cathode-ray-tube display combined with the integrating properties of the eye and brain of the radar operator. integration. a radar with a 1-Mc bandwidth requires a signal-to-noise ratio of 14. Another interesting effect to be noted from Fig. many pulses are usually returned from any particular target on each radar scan and can be used to improve detection. Integration of would be 16. If the false-alarm time were increased from 1 5 min to 24 hr. The number of pulses nB returned from a point target as the radar antenna scans through its beamwidth is nn = -4lr=Mr ( 2. the probability of detection. (The storage device in this instance is the inductance and capacitance constituting the narrowband resonant network. hence postdetection integration is not concerned with preserving RF phase.7 applies for a single pulse only.6. Predetection integration requires that the phase of the echo signal be preserved if full benefit is to be obtained from the summing process.50). the time on target were 0. Such reasoning may not be correct when the false-alarm probability is properly taken into account. If the false-alarm time were as high as (0. detection should be accomplished. However.50.4 db can mean the difference between reliable detection (0. even for a probability of detection of 0. 9.7 db for a 0. and the probability of false alarm as given in Fig. . the signal-to-noise ratio required for detection is not a sensitive function of the false-alarm time. Integration before the detector is called predetection. Radar Pulses The relationship between the signal-to-noise ratio.30) where dB r = = = antenna beamwidth.999) and marginal detection Also. rpm m Typical parameters for a ground-based search radar might be pulse repetition frequency 300 cps.90 probability of detection and a 15-min falsealarm time.8 and 9. phase information is destroyed by the second detector. A definite distinction must be made between these two cases. cps antenna scanning rate. Integration may be accomplished in the radar receiver either before the second detector (in the IF) or after the second detector (in the video). deg/sec f = 0. the required signal-to-noise ratio 2. These parameters result in 1 5 hits from a point target on each scan. the bandwidth of the IF predetection filter would be approximately 20 cps.5° beamwidth. The process of summing all the radar echo pulses for the purpose of improving detection is called integration. as discussed in Sees. 1.4 db. for example.

a number of similar integrators. the bandwidth of the should be about one-half the bandwidth of the predetection filter that same number of pulses. 2. To circumvent this. 2.36 Introduction to Radar Systems this [Sec. the resultant signal-to-noise ratio would be less than n times that of a single pulse. The design of the predetection integrator is further complicated if the target is in motion and produces a doppler-shifted echo that lies outside the passband of the integrator.pass filter made up of a resistor and a capacitor in the video portion of the receiver. all of the same signal-to-noise ratio.86. Postdetection integration is therefore preferred. were integrated by an ideal predetection integrator. and instability of the transmitter frequency might make it difficult to maintain the frequency of the echo signal within the narrowband IF filter. signal-to-noise (power) ratio would be exactly n times that of a single pulse. By coherent it is meant that the phase of the received signal must remain constant with respect to the phase of the transmitted signal. The simplest form of postdetection integrator might consist of a low. each tuned to a slightly different frequency. The integration loss is shown in Fig. The Q of an IF predetection filter would have to be large. Examples of the postdetection integrapredetection integration would be equal to n. The integrawhere integration loss in decibels is defined as L t (n) 10 log 10 [l/E^ri)].(„) = iMk n(S/N) H (2 .) number of pulses integrated value of signal-to-noise ratio of a single pulse required to produce given probability of detection (for n 1) (S/N)„ value of signal-to-noise ratio per pulse required to produce same probability of detection when n pulses are integrated The improvement in the signal-to-noise ratio when n pulses are integrated postdetection The improvement with ideal is nEt (ri) and is the integration-improvement factor. In general. or more precisely. Because of spectrum foldover produced by the second detector. even though the integrated signal-to-noise ratio may not be as great. it is easier to implement in most applications. The IF filter used ahead of a video postdetection integrator should be the matched filter designed for a single pulse. low-pass video filter consisting simply of a capacitor and a resistor than it is to obtain a narrowband IF filter. signals are expected.6 is For convenience. the resultant. If the same n pulses were integrated by an ideal postdetection device. If n pulses. In addition. postdetection integration not as efficient as predetection integration. These curves were derived from data given by Marcum. The comparison of predetection and postdetection integration may be briefly summarized by stating that although postdetection integration is not as efficient as predetection integration. This loss in integration efficiency is caused by the nonlinear action of the second detector. where n (SjN^ = = = = ( = = . can be used to cover the frequency region in which echo integration. postdetection integration is easier to implement than predetection low-pass filter integrates the It is an easier task to obtain a narrowband. which converts some of the signal energy to noise energy in the rectification process. the predetection integrator requires that the phase of the RF or IF carrier oscillations be maintained coherent over a time corresponding to the time on target.8o. tion-improvement factor (or the integration loss) is not a sensitive function of either the probability of detection or the probability of false alarm. The efficiency of postdetection integration relative to ideal predetection integration has been computed by Marcum 10 when all pulses are of equal amplitude. The integration efficiency may be defined as follows £. or integrated.3. tion-improvement factor l (n) nEt (ji) are shown in Fig. a comb filter. 2.

. the number of pulses integrated.8. (b) integration loss as a function of n. d false-alarm number.000 10. 2. 10 courtesy IRE Trans.000 10.) P = = .000 number of pulses integrated (postdetection) (a) I I I I III I I l ll I I ll l I l l l l l ll i l l l I l l i I 10 100 n = number of pulses (A) 1. (a) Integration-improvement factor.90^=10*s' - ^ - 10 — ~ s\^ * Jr y^ ^ •*" i ^ ^ JT J&r Zz^s /%?/ ' ^ s1 1 ^ / i -^ I — — [ i i i 1 il I I i i i i i i 1 1 i i i i i i i i 10 n. square-law detector. 100 1. . 2. P d nf and nf (After Marcum.6] 1.Sec.000 Fig.000 I The Radar Equation I I I I 37 I I l| I I I f I I M| i i i ii 1 1/| i i i i 1 1 if - 100 — ^0><Pd-O5o/ stg&^Pd =0. probability of detection.

7). For the desired probability of detection and probability of false alarm as computed above. where V = voltage amplitude of i'th pulse exp (— y) = attenuation factor per pulse Consider a train of n pulses. false-alarm t . Most practical integration techniques do not sum the echo Practical integrators such as the pulses with equal weight as assumed above. resonant circuit. . In a In a narrowband filter time constant. 2. . or number of pulses integrated n. An improvement It is factor proportional to n applies to the ideal pre- hardly ever achieved in practice. It may be used to determine the required signal-to-noise ratio per pulse (S/N) n at the output of the IF amplifier when n pulses are integrated. - Figure 2. and low-pass filter. the recirculating-delay-line integrator./ and the Kth pulse is attenuated by e~ in 1)v In a factor e~ y pulse 3 is attenuated by e~ period and is the pulse-repetition TjRC. if n pulses are integrated. For the specified average false-alarm time Tt&. = njPu = Tta B.38 Introduction to Radar Systems [Sec. RC = RC RLC = . when a large number of hits are integrated (small signal-to-noise ratio per pulse).33) \ refers to the pulse stored the longest.8cr. y niL/cy/R. is given a weight of unity. For the desired probability of detection Pd number of pulses integrated n. the the electrostatic storage tube apply an exponential weighting factor to the integrated RC RLC pulses. where 7 is the an low-pass filter y resonant circuit. where the nth pulse i Pulse 1. the difference between the postdetection and preThe slope of the postdetection integrationdetection integration is more pronounced. compute the false-alarm probability Pf a p{a nlTta.frf]. by applying the following procedure 1.8a are two straight lines representing the improvement that if the integration-improvement factor were equal to n and to «-. detection (large signal-to-noise ratio per pulse). the last pulse to be received. integrator.7) taking account of integration may be written 3. Also plotted in Fig. find the integration-improvement factor nEf(n) from Fig. the voltage out of the integrator is V= 2 K exp [-(n-i>] = 4 i (2. On the other hand.7 relates for a single pulse (n 1) the signal-to-noise ratio to the probability of detection and the probability of false alarm. the integration-improvement factor is not much different from that which would be obtained from a perfect predetection detection integrator. Divide the signal-to-noise ratio (S/N\ found from step 2 by the integrationimprovement factor nE (ri) to obtain the signal-to-noise ratio per pulse (S/N) n required at the output of the IF amplifier for the specified P d and Pf a The radar equation (2. o-4 maX = P GA anE {n) {^fkT B n F n {SIN\ t e t (2 3 „ Exponential Weighting. An improvement factor proportional to n h fits the experimental data found with an operator viewing a cathodeWhen only a small number of hits are integrated postray-tube display (Sec. If the operator performance were actually that specified by the n l curve. 2. 4.6 would be expected respectively. 2. and number n. enter Fig. proper implementation of automatic postdetection integration could offer an improvement in detection capability over that of an operator. 2. that is. or pulse-repetition period. where fr is the pulse repetition frequency and rj is the number of pulse . improvement curve approaches the slope of the n k curve for n large. receiver bandwidth B. 2. and n/TtaB. .: = = = intervals per radar sweep.7 to find the signal-to-noise ratio (S/AOi for single-pulse detection. pulse 2 is attenuated by ~ 2. 9. .

the antenna remains stationary until n pulses are transmitted and received.6] recirculating-delay-line integrator. 0. the exponential integrator with dumping is almost as efficient as the integrator with uniform weighting. 2.9.34) was derived by comparing the average signal-to-noise ratio for the exponential integrator to the average signal-to-noise ratio for the uniform integrator. e~ v is The Radar Equation y is 39 the attenuation around the loop (loop gain).8a. This corre- sponds to oscillation in the delay-line integrator. 2. Exponential weighting of the pulses results in less efficient integration than uniform weighting. 2. 2. = number of pulses and were used in a continuous-scan radar. Efficiency of an exponential integrator as a function of ny. rather than by comparing the probability of detections as was done in the case of the integration-improvement factor described by Fig. ("dumped") after n pulses or whether the integrator operates continuously An integrator with dumping might be used with a step-scan radar. or to a zero bandwidth and a vanishing . (2. noise in the continuous integrator with the continuous integrator. The efficiency given by Eq. For a dumped integrator the efficiency is 12 = tanh (ny/2) n tanh (y/2) (2 34) This is plotted in Fig. some targets might only be seen with half the number of hits. and whether the contents stored in the integrator and in the electrostatic storage tube. where n e~y is the attenuation factor per pulse (y is assumed small). An example of an integrator that dumps is an electrostatic storage tube that is erased whenever it is read. since there are few applications besides the step-scan radar where it is known beforehand when the integrator is ready to be dumped of its contents. As long as ny is less than unity.) The efficiency depends upon the number of pulses integrated. a factor describing the tube operation. When y = builds up to an "infinite" value and detection of signals is not possible. The weighting integration efficiency p is defined as the ratio of the integrationimprovement factor with exponential weighting to that with uniform weighting. The dumped integrator is not the general rule in practice. dumping. without (In the step-scan radar. but this is not optimum In the dumped integrator values of y -> uously.9. the integrator is operated continare best. In most cases.Sec. after which it is discontinuously stepped to the next position. Another example is a capacitor that is discharged on read-out. the weighting factor. (The optimum weighting function in a radar system would be one which duplicates the antenna scan envelope.) If dumping are erased ny Fig.

13.7 small output in the RC and RLC integrators. the diffracted field is the total field in the presence of the object. tanh (y/2) is replaced by of ny = 1. The maximum efficiency occurs for a value gain of a delay-line integrator. the energy scattered in other directions may also be important. and one may talk about scattering and diffraction interchangeably. when scattered equally in all directions. except that the larger the target size. „ . On the other hand. 2. 2. or in other terms. and hence the radar cross section. For most common types of radar object is illuminated by an electromagnetic wave. as with a bistatic or waveinterference radar. In the case of forward scattering in bistatic radar. 13 When an object scatters an electromagnetic wave.7. and terrain. and the optimum bandwidth of an integrating filter. the value of y that maximizes the efficiency may be determined. can be determined by solving Maxwell's equations with the proper boundary conditions applied. and to zero read-out in the electrostatic The efficiency of the continuous exponential-weighting integrator may be shown to be H l-exp(. 2. where the receiver is not at the same location as the transmitter (Sec.. the determination of the radar cross section with Maxwell's equations can be accomplished only for the most simple of shapes. where a is the radius of the sphere and A is the wavelength). however.9. (2. the scattered field -and the diffracted field could be quite different. Scattering and diffraction are variations of the same physical process. 2. The radar cross section of a simple sphere target is shown in Fig. the scattered field is defined as the difference between the total field in the presence of the object and the field that would exist if the object were absent (but with the sources unchanged). produces an echo at the radar equal to that from the " target.— . reflected : power — toward source/unit solid angle . after Lord Rayleigh. who. 40 tube. a portion of the incident absorbed as heat and the remainder is reradiated (scattered) in many different directions. In some cases. 15 The region where the size of the sphere is small compared with the wavelength {l-najl 1) is called the Rayleigh region. the larger the cross section is likely to be.6). If the number of pulses to be integrated is known beforehand. ships. Introduction to Radar Systems [Sec. and solutions valid over a large range of frequencies are not easy to obtain. : incident power density/477 — — = hmx 4mRr . or the loop y/2. in the early 1 870s. With radar backscatter. Lord Rayleigh was interested in the scattering of light by fields are the < microscopic . Radar Cross Section of Targets The radar cross section of a target is the area intercepting that amount of power which.36) r^ where R= distance between radar and target reflected field strength E = r Ef = strength of incident field targets such as aircraft. first studied scattering by small particles. The portion of the reradiated energy scattered or reflected in the back or rearward direction is of chief interest in radar. 14 Unfortunately. the scattered field. may be found. the cross section does not bear a simple relationship to the physical area.10 as a function of its circumference measured in wavelengths (27Tfl/A.257. A plot of this equation is also shown in Fig.H y) [n tanh (y/2)]* K ' For y small. the two same. In theory. In the present section we shall be concerned only with the backscatter When an is energy CW cross section.

a = radius. 12. 2. His work preceded the orginal electromagnetic echo experiments of Hertz by about fifteen years. and lowering the radar frequency to the point where rain or cloud echoes are negligibly small will not On the other hand. rather The Radar Equation 41 than in radar. The Rayleigh scattering region is of interest to the radar engineer because the cross sections of raindrops and other meteoroSince the cross logical particles fall within this region at the usual radar frequencies.001 0. 15 the method of calculation employed. as in a meteoro.7] particles. minimum greater than the optical value. rather than eliminate. 2.3 0. > .7). ) The behavior of the radar cross sections of other simple reflecting objects as a function of frequency is similar to that of the sphere.Sec. Radar cross section of the sphere.10.11 shows the experimentally measured backscatter (radar) cross section for a right-circular cone as a function of 5°. were desired to actually observe. that of the intermediate the the small cone 2/2. the higher radar frequencies (Sec.4 0. will depend will not be aspect-sensitive.0 I I 1 2 3 4 5 6 20 Circumference /wavelength 2-na/X Fig. or resonance.8 1. The radar is assumed to be in the same plane as the axis of ' of . if it seriously reduce the cross section of the larger desired targets. sions of the sphere are large compared with the wavelength {2-irajX 2 the between In tto cross section optical the approaches section 2-na\l. region. where the dimenFor large 1). aspect.7 db oscillatory with frequency within this region. and The diameter of the base of the large cone is 2A. The maximum value reached is 5. value. A = wavelength. would be preferred 10p n —— i i i i i i i -i —— i 0. its cross section The cross section of other objects. raindrop echoes. At the other extreme from the Rayleigh region is the optical region. logical or weather-observing radar. 13 1416 21 Since the sphere is a sphere no matter from what aspect it is viewed. Figure 2. the radar cross The cross section is optical and the Rayleigh region is the Mie. while the where the cross section goes to zero in the limit of infinite wavelength. each with an apex angle of 1 that cone A. 22 Three different sizes of cones are shown.5 0. -4 rain and clouds are essentially section of objects within the Rayleigh region varies as A The invisible to radars which operate at relatively long wavelengths (low frequencies). upon the direction as viewed by the radar. excluding the Rayleigh region. usual radar targets are much larger than raindrops or cloud particles. however. is 4 db below the (The theoretical values of the maxima and minima may vary according to optical.

and the polarization line is perpendicular to the plane containing the cone axis and the of sight (vertical polarization). The abscissa is the aspect angle. 2. (From IRE Trans. The angle 6 = 0° corresponds to viewing the base of the cone.7 £ -12 -16 60 80 100 . 2. 2. 120 deg 160 180 Aspect angle Fig. (1) Large cone (base diameter 2A). Experimentally measured backscatter cross section for a right-circular cone as a function a.) . 23 90 Angular orientation 8 Fig. and = 180°. {From Shostak and Angelakos.75A-diameter sphere. (2) intermediate cone (base diameter A). 22 of aspect. (3) small cone (base diameter A/2). The ordinate is the radar cross section relative to a sphere with a diameter of 4.) cone.12.11.346 Mc. Backscatter cross section of a long thin rod.42 Introduction to Radar Systems 12 [Sec. relative to a 4. the apex of the cone. ) T Measured Calculated 20 30 40 50 60 70 80 Peters.751 (Measurements were made at a frequency of 9.

maximum would be about 5 db below that shown. snow. The cross sections of some typical simple scattering objects for particular aspects are 13 These are valid if the dimensions of the object are large tabulated in Table 2. the plane of shown in Fig. a a = = tan * rz~ \6n o Incidence at angle to normal na 2 cot 2 4t7/4 2 „ /47TO .sir ») a radius of plate plate area Large flat plate of arbitrary shape Circular cylinder Normal Incidence at angle 6 to broadside A = sin 2 al cos 6 " (kL sin 6) a = ~ 1-n sin 2 6 L = radius cylinder length t Mentzer.Sec. 11 compared with the wavelength. Both theoretiaspect. orthogonal curvilinear coordinate directions In the case of the sphere. where a 77a the radius thus a The cross sections of simple scattering objects are of interest not only because of the insight they give to the scattering properties of more complex radar targets such as aircraft. The geometrical-optics cross section is a = TrRyRz where i?j and R 2 are the two principal radii of curvature with respect to two on the surface.7] The Radar Equation 2. 2 . . or digital aid of the with computed may be sections quency.2.13. Formulas for Radar Cross Sections of Scatterers of Large Characteristic DiMENSiONst Scatterer Aspect Radar cross a section Definition of symbols = 7ra 2 a 6 = = radius Axial a a = A2 tt.tan 4 107T cone half angle apex radius of curvature o Paraboloid Axial = A^l ° 2£o = = = = = Prolate spheroid Axial <r-— a l a b 0„ semimajor axis semiminor axis half angle of target Axial Circular plate . ships. 2. but it is more convenient to make cross-section measurements on scale . and surface objects. ice) and certain classes of space objects. and the radar frecities. R l =R2 = a. 2. Complex Targets. If the rod were of steel instead of silver. the methods of applied to compute the radar cross section. 6 Jl\-j. When the radius of curvature of the reflecting surface geometrical optics may be is large compared with the wavelength. The radar cross section of the ogive 23 In both Figs. Target cross is . 2. = they may be measured experimentally. The target cross section can be measured with full-scale targets. Table 2.12 and 2. the first cal and measured data are shown.13a is plotted in.136.2. The radar cross section of complex targets such as ships. and terrain are complicated functions of the viewing aspect computers. 2. Figure 23 2 is polarization is perpendicular to the line of sight but is in the same plane as the longi- tudinal axis of the object (horizontal polarization). aircraft.Fig. 1 43 a plot of the backscatter cross section of a long thin rod as a function of The rod is 39A long and A/4 in diameter and is made of silver. but they are characteristic of such important targets as meteorological objects (rain.

The phases and amplitudes of the individual signals might add to give a large total cross section. The relative phases and amplitudes of the echo signals from 31. The theoretical computation of target cross section was pioneered and developed by Siegel and associates at the University of Michigan Radiation Laboratory. 24 A complex target may be considered as comprising a large number of independent objects that scatter energy in all directions. Peters. 2. 23 IRE Trans. total cancellation. Most radar cross-section information concerning complex targets is obtained in this manner. Consider the scattering from a relatively "simple" complex target consisting of two . changed and cause a scintillating echo. (From the individual scattering objects as measured at the radar receiver determine the total cross section.44 Introduction to Radar Systems [Sec.13. The energy scattered in the direction of the radar is of prime interest.7 models at the proper scaled frequency.0 Measured : Points calculated by optics Calculated curve due to traveling waves Calculated maxima due to traveling waves 30 40 [6) 50 60 Angular orientation Fig. or the relationships with one another might result in total In general. the behavior is somewhere between total reinforcement and If the separation between the individual scattering objects is large compared with the wavelength—and this is usually true for most radar applications— the phases of the individual signals at the radar receiver will vary as the viewing aspect is cancellation. 2. (a) Dimensional drawing of the ogive.) (6) backscatter cross section of the ogive.

modulation a produced and measurement running during the . 1 5. each with different scattering properties. of the order of 1 to 2 kc.14. Another restriction placed on / is that it be small compared with the Furthermore. 2. Sphere target f* " Sphere target Vt = Kyja where cos -^ V2 = Kjo cos AttR. 2. The resultant signal depends upon the phase of each echo signal as well as the amplitude.14). 2.sin /J 2 ')]) by trigonometry cos A+ B = 2 cos A+B B cos so that Vr = Kjo 2 cos 4nR cos „ I— cos 4"R = KJa. a World War II medium-range aircraft 25 is shown in Fig. Also. isotropic objects (such as spheres) separated a distance / (Fig. where c is the velocity of propagation With and r is the pulse duration. 2.7] The Radar Equation 45 By isoequal. The resultant voltage from the two objects is K is a constant Vr = Ksjo^cos.Sec. Radar Fig. The separation / is assumed to be less than ct/2. The echo signals from the two reflecting objects add vectorially. 2.37) where ar = 4<r = 2 1 cos or £r + cos (2. The radar wavelength was 10 cm. V A s (2. which includes the parameters involved in the radar equation. between the scatterers which affect the resultant cross section. two-engine bomber. tropic scattering is meant that the radar cross section of each object is independent of the viewing aspect. it is complicated enough to indicate the type of behavior to be expected with practical radar targets. interactions may occur scatterers.16. the Polar plots of crr /cr for various values of //A are shown in Fig. Lav — \R 2 cos sin 6 1 cos . The radar cross sections of actual targets are far more complicated in structure than Practical targets are composed of many individual the simple two-scatterer target. R2 as R. R x on distance R from radar to target. assumption. Although this is a rather simple example of a "complex" target. Geometry of the two- scatterer complex target. These data were obtained experimentally by mounting the aircraft on a turntable in surroundings free from other The propellers were reflecting objects and by observing with a nearby radar set. both scatterers are illuminated simultaneously by the pulse packet.38) The ratio aJaQ can be anything from a minimum of zero to a maximum of four times cross section of an individual scatterer. An example of the cross section as a function of aspect angle for a propeller-driven The aircraft is the B-26. The cross sections of the two targets are The RF voltage assumed equal and are designated a received at the radar from each target of cross section a is proportional to this .

The accuracy of theoretical cross sections is claimed to be from 2 to 10 db.7 2 3 4 -9°° -90° 3 4 -90° U-4A-J Fig. The maximum echo signal occurs in the vicinity of broadside. The theoretical data (solid curves) represent averages over a limited aspect angle. The cross section can change by as much as 1 5 db for a change in aspect of only |°. The frequency dependence is seen to be slight. Siegel's cross-section-computation technique lends itself quite well to the analysis of the relative contribution of various target components to the over-all cross . (£>) / = 2A 4A. Figure 2.17 compares the theoretical and experimental cross section of the B-47 bomber aircraft as a function of aspect angle and frequency. the fine structure is not included. 2. 2. 1 5. (a) / = A.46 Introduction to Radar Systems [Sec.38)]. (2. where the projected area of the aircraft is largest. Polar plots of <t t /ct for the two-scatterer (c) / = complex target [Eq. Experimental-measurement accuracies also are of the same order of magnitude.

600 Mc 20 40 60 80 100 120 140 160 180 Azimuth. Comparison of the theoretical and experimental cross section of the B-47 jet aircraft as a function of aspect and frequency as obtained by various investigators. Siegel.17.) University of Michigan 330Mc Evans Signal Laboratory !50Mc Ohio State University 195 Mc — Radiation.7] 35db The Radar Equation 47 Fig.) . Inc. University of Michigan. 2. Inc. u courtesy McGraw-Hill Book Company. {From Ridenour. 2. 2. (Courtesy K.16.Sec. deg Fig. Experimental cross section of the B-26 two-engine bomber at 10-cm wavelength as a function of azimuth angle.

691 0. but no single component dominates over the entire range of aspect angles. J 1 V5 \ 1 - . Cross section as a function of wavelength for the nose-on aspect of a large jet aircraft.19. Most . 1 2. 2.& / <b / 1 ^ § i / / / / / J\\ 7 A " \ /A\ Jf\\ \ \ M 10° to.18.719 0. section.19. 8 for a typical large manned jet aircraft. The variation of the cross section as a function of wavelength for the nose-on aspect is shown in Fig.71 m.723 Fig.^ s J Leading edge of stabilizer / A \ Trailing edge ~~ of stabilizer-^ 1 ^ 10' 1 \-rv \ ^. Siegel.) The effect of the various components is shown in Fig. University of Michigan) 0. Siegel. deg off nose-on aspect Fig. 2.7 I 1 X it A / 10' - . (Courtesy K.695 0. 0. University of Michigan. 2.-\ .48 Introduction to Radar Systems 10' I ! I [Sec.707 0.715 0.711 meters 0. (Courtesy K.703 A. 24 There are many significant contributors to be considered.699 0. 2. The radar cross sections presented above apply for horizontal polarization. Radar cross section of the components of a typical large manned jet aircraft at a wavelength of 0. yn // 1 1 \ \ 1 10"' 10" 1 T ' — ' ' L"*« Azimuth angle.

The radar scattering properties of a target for any polarization may be described by a 2 x 2 matrix of <r's corresponding to transmitting each of two orthogonal polariThis zations and receiving with the same polarization or the orthogonal polarization. and the minimum radius of curvature of the mean (unperturbed) surface is assumed large compared with the wavelength. absorbing materials have little effect on the radar cross section when (1) the radar wavelength is large compared with the target dimensions (Rayleigh scattering) 30 or (2) the target is observed by a forward scatter (bistatic) radar whose wavelength is small compared with the target dimensions.20. 2. 31 The measured radar cross section of a man has been reported 32 to be uency.20. Experimental cross sections for a large aircraft at approximately 75 polarization and aspect. either theoretical or experimental. electromagnetic energy (Sec. 1 db.495-1. Apparently considerable surface roughness can be It has been tolerated before a significant effect on the value of cross section is obtained. but in a statistically uniform and isotropic manner. the cross sections may be different.098-0.7] The Radar Equation 49 search radars whose prime targets are aircraft usually employ this type of polarization. ever. The surface slopes are assumed small.800 9. An If vertical or some other polarization is used.01 A without - causing a change in cross section of more than 0.05 0.. Most cross-section data. The "roughness" to which this statement applies are surface irregularities distributed at random. assume the target to have a smooth reflecting surface. example of the difference between horizontal and vertical polarization is shown in Fig. as follows Mc a. 13 26 " 28 is known as the polarization scattering matrix.120 2.22 i i i 50 . Siegel.033-2. 12.1 : Sec. 2. University of Michigan) Mc as a function of . 2. deq measured from nose-on 140 in 160 180 plane of wing Fig. reported 29 that the roughness depth of a sphere's surface can be as large as 0.88 0. rather than reflect. J „ 1 40 M In A "f It ii i ml = 301 | / \ j vliil IHl ill lJU ''HI i in / 1 / 1 T3 • \ /> ' i \ v/ \ J J In \K- ^~ i / 20 10 polarization polarizotion 1 1 i i 1 1 20 40 60 100 120 80 Aspect angle. m2 410 1.890 4.368-1.997 0. (Courtesy K.140-1.10). Radar cross sections can be considerably reduced by properly shaping the target or by Howcoatings that absorb.375 0.33 0..

however. The cross-section data presented in this section lead to the conclusion that it would not be appropriate to simply select a single value and expect it to have meaning in the computation of the radar equation without further qualification. Variations in the echo signal may be caused by meteorological conditions.95. 33 ) toward the cross sections of complex targets (the usual type of radar target) are quite Therefore. aspects of the aircraft.6 assumed that the echo from a particular target did not vary with time. However.8 The spread in cross-section values represents the variation with aspect and polarization. The probability-density function gives the probability of finding any particular value of target cross section between the values of a and a da. typical pulse-by-pulse record of the echo sensitive to aspect. 2. to properly account for target cross-section fluctuations.999 being typical). 2. sec —* from a Meteor jet aircraft flying Fig. that is. while the autocorrelation function describes the degree of correlation of the cross section with time or number of pulses.99. variations in the echo signal will result. or 0. Similar effects occur for propeller-driven aircraft such as the B-26 (Fig. especially in tracking radars. 2. a value of cross section that is exceeded some specified (large) fraction of time.21 Hay33 reports that analyses of records of this type show that the period of the fluctuation varies from several seconds at long ranges to a few tenths of a second at short ranges. of variations in the target cross section. It is usually not practical to obtain the experimental data necessary to compute the probability-density function One method of accounting + . Pulse-by-pulse record of the echo signals radar. for a fluctuating cross section in the radar equation is to lower bound. The degree of echo modulation for this target varies from 26 db to less than 10 db for different . as the target aspect changes relative to the radar. Cross-section Fluctuations The discussion of the minimum signal-to-noise ratio in Sec.21. Methods for dealing with the cross sections of complicated targets are discussed in the next section. 2. The minimum cross section of typical aircraft or missile targets generally occurs at or near the head-on aspect. The spectral density of the cross section (from which the autocorrelation function can be derived) is also sometimes of importance. the echo signal from a target in motion is almost never constant. the lobe structure of the signal received antenna pattern. In practice. The A from a Meteor aircraft (British two-engine jet fighter) flying toward a radar is represented in Fig.8. This procedure results in a conservative prediction of radar range and has the advantage of simplicity. {After Hay. For all practical purposes the value selected is a minimum and the target will always present a cross section greater than that selected. Curves of cross section as a function of aspect and a knowledge of the trajectory with respect to the radar are needed to obtain a true description of the dynamical variations of cross section. or equipment instabilities.50 Introduction to Radar Systems [Sec. 2. But the chief source of fluctuation is that Time. the probabilitydensity function and the correlation properties with time must be known for the particular target and type of trajectory. The fraction of time that the actual cross section exceeds the selected value would be close to unity (0. 0. select a 2.16). The fluctuation period also depends on radar wavelength.

p(<r) = ^exp(-^) #av V (2. dependent Case 3. In two of the four cases. The probability-density function for the cross section a is given by the Rayleigh .22 for n = 10 hits integrated. (2. the signal-tonoise ratio would have to be 16. or as one large reflector subject to fairly small changes in orientation.— ) <7>0 (2.8 db/pulse. a signal-to-noise ratio of 6. the number of independent scatterers must be number may be as few as four or five. Cross-section fluctuations of objects with dimensions large compared with a wavelength are also expected to approximately follow the Rayleigh probability-density function.95. the fluctuations are assumed to be more rapid and uncorrected pulse to pulse. The probability-density function assumed in cases 3 and 4 is more indicative of targets that can be represented as one large reflector together with other small reflectors. density function p(a) = — exp(. if the characteristics of the essentially infinite.40) Om l Case 4. but if the target cross section fluctuated with a Rayleigh distribution and were scan to scan uncorrelated (case 1). These typical situations bracket a wide range of practical cases. The fluctuation is pulse to pulse according to Eq. = . if the desired probability of detection were 0. Eq. A more economical method to assess the effects of a fluctuating cross section is to postulate a reasonable model for the fluctuations and to analyze it mathematically. The majority of radar targets are probably of this nature. Therefore. in theory. the nonfluctuating cross section will be called case 5. For purposes of comparison. the value of cross section to be substituted in the radar equation is the average cross section a av The signal-to-noise ratio needed to achieve a specified probability of detection without exceeding a specified false-alarm probability can be calculated for each model of target behavior.Sec. An echo fluctuation of this type will be referred to as scan-to-scan fluctuation.39) where cr av is the average cross section over all target fluctuations. When the detection probability is large.39). (2. In the other two cases. Case 2. the fluctuation is assumed to be independent from scan to scan case 1 but the probability-density function is given by . echoing areas. The Rayleigh probability-density function assumed in cases 1 and 2 target consisting of applies to a many independent fluctuating scatterers of approximately equal Although. 2. For example. In all the above cases. In this case. all four cases in which the target cross section is not constant require greater signal-tonoise ratio than the constant cross section of case 5. Swerling34 has calculated the detection probabilities for four different fluctuation models of cross section. A comparison of these five cases for a false-alarm number n f = 10®^ n/Pfa) is shown in Fig. in practice the . This assumption ignores the effect of the antenna beam shape on the echo amplitude. 2.40).8] The Radar Equation 51 and the autocorrelation function from which the over-all radar performance is determined. Most radar situations are of too complex a nature to warrant obtaining complete data.2 db/pulse is necessary if the target cross section were constant (case 5). it is assumed tjiat the fluctuations are completely correlated during a particular scan but are completely uncorrected from scan to scan. The four fluctuation models are as follows Case 1 The echo pulses received from a target on any one scan are of constant amplitude throughout the entire scan but are independent (uncorrelated) from scan to scan. This increase in signal-to-noise corresponds to a reductionin range by a factor of 3. as in The probability-density function for the target cross section is also given by but the fluctuations are more rapid than in case 1 and are taken to be infrom pulse to pulse instead of from scan to scan.28.

detection probability Pd and false-alarm probability Pt&. the pulse to pulse (cases 2 and 4). The data in these two figures. n/Pfa ).80 .05 0. a greater signal-to-noise ratio is required when the fluctuations are uncorrelated scan to scan (cases 1 and 3) than when the fluctuations are uncorrelated In fact. Figure 2. = = = case than for any of the four fluctuating cases." 1 ) 10 pulses integrated and false-alarm number n. 3. In the region where the signal-tonoise ratio required for a given detection probability is greater for the nonfluctuating I I I I I I I 0.22. The resultant that which would apply if detection were based upon a If n pulses are integrated. may be used to find the signal-to-noise ratio per pulse for any of the four fluctuating cases.24 were derived from his report. to the signal-to-noise ratio is signal-to-noise ratio (5/A') 1 1 and 2 or cases 3 found from step 1 above. From Fig. n f . the detection probability is lower (<0. along with the detection-probability curves of Fig.30. the actual performance of the radar might not measure up to the performance predicted as if the target cross section were constant. The curves presented in Figs.: 52 Introduction to Radar Systems [Sec.30 S 0. 2.30) than would normally be considered useful for radar application and is of little practical interest. The procedure is as follows 1. 2.60 - - 1-0. 2.23 and 2.98 0. db 25 30 n Fig. 0.02 0.7 corresponding to the desired value of as a function of Pd .23 determine the correction factor for either cases and 4 to be applied single pulse. more likely it will be for the fluctuations to average out. n. 2.3 0.01 - - I I i l i -10 10 15 20 5 Signal-to-noise ratio per pulse. This does not occur when the cross section is assumed to be constant throughout a particular scan of n pulses. and cases 2 and 4 will approach the nonfluctuating case. 2. 2.90 <£"0.10 0.20 0.95 / 1 - 0.8 target cross section are not properly taken into account.40 I o a. Swerling34 computed the detection-probability characteristics for fluctuating targets and signal-to-noise ratio. the larger the number of pulses integrated. (Adapted from Swerling. Find the signal-to-noise ratio from Fig.99 - 5 4 2 . 2.7.1 - 0. Comparison of detection probabilities for five different models of target fluctuation for 10 8 (n. the signal-to-noise ratio found in step 2 for a single pulse .70 2 0.22 also indicates that for probabilities of detection greater than about 0.

8] The Radar Equation 53 001 005 0.000 five cases Fig.95 0.1 0. the target cross section fluctuates. as compared with a nonfluctuating target.24.23.3 Probability of detection 0.2 0.9 0. 2.8 0. number of pulses integrated for the . single hit. 2.5 0.99 when Fig. » 1. Additional signal-to-noise ratio required to achieve a particular probability of detection. 2.Sec. 100 n 500 1. = 10 20 50 Number of pulses integrated.7 0. Integration-improvement factor as a function of the of target fluctuations considered.

20 0. shown in Fig. Effect of correlation between pulses on the detection probability.8 divided by the integration-improvement factor = nE^ri) from Fig.23 and 2.98 0. db -Jill 24 26 signal voltage Fig. 2. plete correlation The fluctuation models considered above assume either combetween the pulses in any particular scan (cases 1 and 3) or else complete independence between the pulses (cases 2 and 4).05 _ — / / i _ 10 Mean 12 14 18 20 22 16 signal-to-noise ratio per pulse. two pulses integrated (n 2).50 0.24 is in some cases greater than n. The signals are assumed to be correlated according to the correlation coefficient. 2.30 — — i A 1/^ 7 ! /? i i i 'y y ^^ y /- ^^~y — - 0. = — in Figs. 2. 9 (<>!<#* (2. it is likely that the pulses of a particular scan will lie within these two extremes and be partially correlated. The signal-to-noise per pulse will always be less than that of .999 0. 2. p correlation coefficient. square-law detector.95 — fluctuation (constant signal ' amplitude)-- J 1 c? c 0. 35 courtesy IRE Trans. Schwartz35 considered the effect of partial correlation on the addition of two fluctuating signals (n = 2).54 is Introduction to Radar Systems /<(«) [Sec. The parameters {SjN\ and nE {n) . Rayleigh fluctuation. the signal-to-noise ratio required for « = target.24 are essentially . are those substituted into the radar equation (2. A portion of Schwartz's results voltage correlation coefficient.60 H 1 1 £ °. 2. The false-alarm 2.995 0. These represent two extreme cases of fluctuations.80 o I ^ : 0.10 0. In general.25. is The power correlation coefficient is p 2 .41) where xlt x2 x lt x 2 °i> a \ = amplitudes of two successive pulses = mean values (here assumed zero) = variances of x and x ± 2 The two variances are assumed equal.40 0. for in those cases in \vhich the integration-improvement factor is greater 1 is larger than for a nonfluctuating than n. .24 in order { to find the signal-to-noise ratio per pulse.25 for several values of the signal probability is 10 -10 The results for .99 i i i i i It should also be noted that the data III! / No 0.90 0. the integration efficiency factor E^ri) > 1 . at least over the range of 10 6 to 10 10 Partial Correlation.) {After Schwartz.70 °.998 0. One is not getting something for nothing. independent of the false-alarm number. or in other words.32) along with cr av The integration-improvement factor in Fig. an ideal pre- detection integrator for reasonable values of Pd 0.

.0 on a would be high that it would also be observed on the next scan. the likelihood to-scan efficient with correlation cocorresponds to complete p (p = independence scan to scan). at least over the range The data in Fig. incoming (After Sponsler. The effects of the antenna beam shape and of nonuniform weighting of pulses by the post- A detection integrator are also taken into account. Sponsler37 has applied the theory of Markov chains to describe the observed scan-toscan correlation. 36 His analysis applies to a large family of probability-density functions of the signal fluctuations and for very general correlation properties of the signal fluctuations. S.2 0. more general treatment of fluctuating pulsed signals in the presence of noise has been given by Swerling. the likelihood is large that all the succeeding pulses would be below the mean. When the return echo consists of a train of more than two pulses. The greater the degree of correlation between the pulses. for outgoing trajectories triangles. measurements compared with theory. as might have been expected. the outcome of any particular event Instead. if the correlation between pulses were weak. 3 ' courtesy IRE .25 apply to the case of only two pulses. According to Schwartz. . itwould probably not be seen on the next scan. it would be likely that the below-average pulses would be counterbalanced by the above-average pulses and the combined signal-to-noise ratio would average to a value suitable for detection. by chance. pulse were below the mean value necessary for detection. 2. Scan-to-scan Correlation and the Markov It has been experimentally observed Process. if the pulses were highly correlated. Also shown is the detection probability for a nonfluctuating target signal. In the theory of Markov chains. Effect of scan-to-scan correlation Experimental on detection probability. preceding event but not on directly the outcome of on the dependent event is particular any of the other preceding events.6 0. or if the target is not seen on a particular scan. the greater must be the signal-to-noise ratio required per For if. It might also be trajectories. especially Trans. and open circles and triangles represent miss following miss. due to atmospheric effects. the first pulse in order to achieve a specified detection probability. it is expected that similar con- clusions will apply. 2. Solid circles and triangles represent detection following correlation detection. Solid lines are theoretical curves based on scan- particular scan. and altitude. . that is. to 10~ 10 the range for which computations were made. = and p = 1 On the other hand. 38 The theory of Markov chains is discussed in texts Sponsler presents in his paper some experimental data liX relating the blip-scan ratio at a particular range T with the transition probability p for a particular aircraft The data are attributed to P.26. 2. on probability theory. if the target were observed Fig. blip-scan ratio 1.4 0. The above are essentially speculations as to the cause of for no conclusive experimental proof has been offered to scan-to-scan substantiate that these are indeed the causes. correlations. Olmstead of the The ordinate is the transition Bell Telephone Laboratories and are shown in Fig.Sec. the outcome of any is not assumed to be independent of other events.8] partial correlation fall The Radar Equation 55 (completely uncorrelated) between the two extremes of p (completely correlated). the false-alarm probability does not significantly affect the from 10~ 5 general conclusions concerning the partially correlated pulses. Circles are the slow variations of target aspect or to the lobe structure of the antenna pattern.) when the radar beam just grazes the surface of the earth. that in some instances there may be correlation between the detection of targets from scan to 0.26.8 V. The scan-to-scan correlation might be due to scan. 2.

the average power is related to the peak power by Pav The ratio PAv lP r/Tr or rfT is t. while a CW radar which transmits continuously has a duty cycle of unity. The theory of Markov chains has also been applied by Sponsler37 to the cumulative detection probability of a radar in which the detection decision is made automatically. The figure seems to indicate that for this particular set of data the scan-to-scan correlation coefficient was appro ximately J. the integration efficiency E. It is defined as the power averaged over that carrierfrequency cycle which occurs at the maximum of the pulse of power. or the probability of detection upon a single scan. The signal-tonoise ratio {SjN) x depends on the desired probabilities of detection and false' "alarm. The abscissa is the blip-scan ratio. The target cross section a is not under the control of the radar designer.43) gives the range at which a target of cross section a would be detected with a probability Pd without exceeding a specified false-alarm . If a constant value of cross section is used.9. (2. defined as the probability that if a target is detected it on a particular scan. be detected on the next scan. A typical pulse radar for detection of aircraft might have a duty cycle of 0. will 2. The peak pulse power as used in the radar equation is not the instantaneous peak power of a sine wave. the effective receiving aperture A e and the receiver noise figure Fn The type of waveform transmitted and the receiver design determine B n r and. .26 are the theoretical curves that would have been obtained if the data followed a simple Markov process. Eq. if the application of the theory of Markov is valid. If the transmitted waveform is a train of rectangular pulses of width t and pulse-repetition period Tr ( Tr = 1 // r). it is sometimes more convenient to express the radar equation in terms of the energy E = P&w jfr contained in the r transmitted waveform D4 E T GA e onEi(n) (47rfkT F n (B n r)(SIN) 1 (2. Transmitter t Power The power P in the radar equation (2.{n). The curves are labeled with the values of the correlation coefficient p between successive pairs of observations.) The average radar power Pav is also of interest in radar and is defined as the average transmitted power over the pulse-repetition period.9 probability /Vi.1) is called by the radar engineer the peak power. (Peak power is usually equal to one-half the maximum instantaneous power. . If the transmitted waveform is not a rectangular pulse. Also shown in Fig.001 or less. with no knowledge assumed as to the previous scans. The blip-scan ratio may also be considered as the ratio of the number of times that a particular target is observed (as a "blip" on the scope) at a particular range to the number of times it could have been observed (scans).436) In this form. . we get Ri = P. .42) called the duty cycle of the radar.56 Introduction to Radar Systems [Sec. to some extent. 2. The important parameters affecting range are the total transmitted energy nE T the transmitting gain G. the range does not depend explicitly on either the wavelength or the pulse repetition frequency. = ^=p f tT r (2. v GA e on Ei (n) (4nfkT F n (B nr)(SIN)Jr The bandwidth and the pulse width are grouped together since the product of the two is usually of the order of unity in most pulse-radar applications. Writing the radar equation in terms of the average power rather than the peak power. 2.

The prf may be changed continuously within prescribed limits. 2. those for B and C are not. If the cross section fluctuates. The multiple-time-around echoes on the A-scope cannot be distinguished from proper target echoes actually C within the maximum unambiguous range. 2. and C in The pulse repetition frequency (prf) at Fig.27a. other schemes that might be employed to "mark" successive pulses so as to identify multiple-time-around echoes include changing the One method of pulse amplitude. pulse width.27c. or more specifically without exceeding a specified maximum false target indications. 2." Consider the three targets labeled A. echoes from multipletime-around targets will be spread over a finite range as shown in Fig. Target A is located within the maximum unambiguous . 2. 2. Multiple-time-around echoes.27b. the likelihood of obtaining Echo signals received target echoes from the wrong pulse transmission is increased. Instead of modulating the prf. phase.10] The Radar Equation 57 rate of probability PtB . or it may be changed discretely among several predetermined values. They can result in erroneous or confusing range measurements. Pulse Repetition Frequency and Range Ambiguities is determined primarily by the maximum range which targets are expected. range -R una mb [Eq.24. operate echoes is to from an unambiguous range target will appear at the same place on the A-scope on each sweep no matter whether the prf is modulated or not. after an interval exceeding the pulse-repetition period are called multiple-time-around echoes. such schemes are not so successful in practice as one pulse to pulse. target B is at a distance greater than iJ unam b DUt " ess tnan The appearance 2/? unamb . (1. and B and C are multiple-time-around targets. targets need only two separate repetition frequencies in order to be resolved. B.10. frequency. The nature of some multiple-time-around echoes may cause them to be labeled as "ghost. where A is within R unBLmb . B and C. (c) appearance of the three targets on the A-scope with a changing prf. average cr av .27. respectively." targets. or even "flying saucers. (a) Three targets A. correct. 2. while target is greater than 2i? unamb but less than 3J? unamb of the three targets on an A-scope is sketched in Fig. The number of separate pulse repetition Second-time frequencies will depend upon the degree of the multiple-time targets.2)] of the radar. distinguishing multiple-time-around echoes from unambiguous 39-42 The echo signal with a varying pulse repetition frequency.Sec. If the prf is made too high. However. (b) appearance of the three targets on the A-scope." or "angel.23 and 2. . and (S/N^ and E (n) modified according to ( a must be replaced by its Figs. 2. Only the range measured for target A is n L^ » "unamu n *J x £J1_&LJCJ1 i I i o t=o t=Vfr Time (or t = */f r t = Vfr range) -» A A (6) B C A_ A Range -» Range (c) —* Fig. or polarization of transmission from Generally.

the directive antenna channels the radiated energy into a beam to enhance the Almost The antenna gain G is a measure of power radiated in a particular direction by a directive antenna to the power which would have been radiated in the same direction by an omnidirectional antenna with 100 per cent efficiency. 2. The Millstone Hill radar operated at a prf of 30 cps. to antenna/47r Note that the antenna gain is a function of direction. Because of the inverse-fourth-power relationship in the radar equation. Beamwidths of typical pencil-beam may be of the order of a few degrees or less. since the echo-signal amplitude can fluctuate strongly for reasons other than a change in range. or simply the antenna pattern.3). 7. If it is greater than unity in some it must be less than unity in other directions. is a plot of antenna gain as a function of the direction of radiation. Range information is usually sacrificed in pulse-dopplerradar applications. for example. to adding one or more radars and operating them on a time-shared basis. or nearly so. When used as an AI radar.<f>) m\ = Power radiated per s- unit solid angle in -2 — azimuth 6 and elevation r <j> (2.28a) and the fan beam (Fig. the power gain of an antenna used for transenergy concentrated in the directon of the target.44) power delivered directions. however. . This follows from the conshall usually servation of energy. 2. the number of targets it must handle is small hence sufficient time is generally available to resolve any ambiguities that might exist. 14. [Sec. (A typical antenna pattern plotted as a function of one angular coordinate is shown in Fig. ambiguous-range radars is the pulsedoppler radar (Sec. Ambiguities may theoretically be resolved by observing the variation of the echo signal with time (range).000 times the unambiguous range of the basic pulse rate. Hence the gain and the effective area of a transmitting antenna are the same when the antenna is used for receiving. giving an unambiguous range of about 2.5).11 One of the fundamental limitations is the foldover of nearby targets nearby strong ground targets (clutter) can be quite large and can mask weak multiple-time-around targets appearing at the same place on the display. more time is required to process the data when resolving ambiguities.28Z>). The antenna beam pattern. But for the general search-radar application. the mission /-vfl is G(p.2. Also. 1 . the target-tracking radar for the control of .) Antenna beam shapes most commonly employed in radar are the pencil beam (Fig. A rather unique application involving the resolution of range ambiguity is that of the first radar detection of Venus (Sec.11. but by modulating the pulse transmissions it was possible to correctly resolve the ambiguities and measure a range almost 9. we mean When we speak of antenna gain in relation to the radar equation. the rate of change of the echo signal from a target at long range will be different from that of a target at short range. which states that the properties of an antenna are the same no matter whether it is used for transmission or reception. These techniques to resolve ambiguities are similar. operation with an ambiguous range is usually not warranted unless special circumstances make it necessary. On transmission. This is not always a practical technique. in principle. as. would that is. Antenna Parameters all radars use directive antennas for transmission and reception. the maximum gain G. unless otherwise specified. It will be recalled that this principle was used in the derivation of the radar equation in Sec. An example of the use of high-repetition-rate. One of the basic principles of antenna theory is that of reciprocity. 4. 58 Introduction to Radar Systems like. The pencil antennas beam is axially symmetric.700 nautical miles. More precisely. 2.. 2. Pencil beams are commonly used where it is necessary to measure continuously the angular position of a single target in both azimuth and elevation. 1 .

The coverage of a simple fan beam is usually inadequate for targets at high altitudes The simple fan-beam antenna radiates very little of its energy in close to the radar. On the other hand. Since the number of resolution cells which the fan-beam radar must search is considerably less than the number which the pencilbeam radar must search. Many long-range ground-based search radars use a fan-beam pattern narrow in azimuth and broad in elevation When ground-based search radars employing fan beams are used against airreflector surface craft targets. a fan-shaped pattern. as in the common obtained.200. the two are at odds with one another.| This is especially true if there is a large number of resolution cells to be searched. Although a narrow beam can. One method of generating a fan beam is with a parabolic reflector shaped to yield the proper ratio between the azimuth and elevation beamwidths (Fig. 1° were required to scan the samevolume. search a large sector or even a hemisphere. The number of resolution cells can be materially reduced if the narrow angular resolution cell of a pencil-beam radar is replaced by a beam in which one dimension is broad while the other dimension is narrow. Usually. it is not always desirable to do so.28.Sec. is no resolution in elevation Therefore no height information is available. and time). operational requirements place a restriction on the maximum scan time (time for the beam to return to the same point in space) so that the radar cannot dwell too long at any one radar resolution cell. a height-finding radar actually measures elevation angle rather than height. (6) fan-beam-antenna pattern.) If a fan-beam search radar 1 ° in azimuth and 45° in elevation were required to scan 360° in azimuth (complete circular coverage). The more slowly the radar antenna scans. the more pulses will be available On the other hand. the total number of angular resolution cells would be 360 x 45 16. radars vary from 1 to 60 rpm. The rate at which a fan-beam antenna may be scanned is a compromise between the rate at which target-position information is desired (data rate) and the ability to detect weak targets (probability of detection). 59 weapons or The pencil beam may be generated with a metallic shaped in the form of a paraboloid of revolution with the electromagnetic energy fed from a point source placed at the focus.6). doppler velocity. 1. 2. if a pencil-beam radar with a beamwidth of Fig. a slow scan for integration and the better the detection capability. the fan-beam radar can dwell longer in eachcell and more hits per target can be obtained. if necessary. (Strictly speaking. range. that is. 2. (a) Pencil-beam-antenna pattern. . = t The radar resolution cell is in general a five-dimensional space (two orthogonal-angle coordinates.11] The Radar Equation missile guidance. 5 or 6 rpm being typical. One method of achieving elevation-angle information for targets located by a fan-beam search radar is to employ an additional fan-beam radar with the narrow dimension in elevation instead of in azimuth. Scan rates of practical search rate means a longer time between looks at the target. Unfortunately. the scanning region might be considered as being divided into 360 angular resolution cells. height-finding radar.

106) gives Ideally. The echo signal is therefore independent of range for a conx esc r 2 2 stant-altitude target. The maximum gain of an antenna is related to its physical area A (aperture) by (2. The value of the constant. the gain as a function of elevation angle is given by G(<£) = G(<£ )-^i csC <p <f> for <f> <<fx<f> m 2 (2. but it is always much less than this with a single antenna because of practical difficulties. From <f> the antenna pattern is similar to a to <f> normal antenna pattern. i? 4 cscV =K K. In the airborne case. </>„ to <f> m = <f> = (f> .49) where / is the dimension of the antenna in the plane of the angle 6. but from the antenna gain varies as esc 2 <f>. and since = Rjh. The product of pA is the effective aperture A e A typical reflector antenna with a parabolic shape will produce a beamwidth approximately equal to . the received power becomes p = KJ* = K (2. The cross section a varies with the viewing aspect.46) J? K is a constant. The cosecant-squared antenna has the important property that the echo power Pr received from a target of constant cross section at constant altitude h is independent of the target's range R from the radar. <f> = . it is possible to modify the antenna pattern to radiate more energy at higher angles.11 this direction.48) G = where p X 4 P = antenna efficiency = wavelength of radiated energy -f- efficiency depends on the aperture illumination and the efficiency of the antenna feed.45) where G(<£) <A)> ^m = gain at elevation angle = angular limits between which beam follows esc = shape This applies to the airborne search radar observing ground targets as well as groundbased radars observing aircraft targets.9). 2. However. The antenna 6° = — ' (2. In the cosecant-squared antenna (Sec. The cosecant-squared pattern may also be generated with an array-type antenna. and X and / are measured in the same units.47) where K is a constant. The height h of the target is assumed constant.~^ 4 1 (2. (2. depends upon the distribution of energy (illumination) across the aperture. Substituting the gain of the cosecant-squared antenna [Eq. 7.45)] into the simple radar equation (1. the angle ^ is the depression angle. The gain of a typical cosecant-squared antenna used for groundbased search radar might be about 2 db less than if a fan beam were generated by the same aperture. the <f> p where <j> = Pk G\<j> 3 (47r) ) esc 4 4 <f>X 2 o csc </.60 Introduction to Radar Systems [Sec. In practice. . The cosecant-squared antenna may be generated by a distorted section of a parabola or by a true parabola with a properly designed set of multiple feed horns. the earth is not flat. upper limit m should be 90°. the power received from an antenna with a cosecant-squared pattern is not truly independent of range because of the simplifying assumptions made. One technique for accomplishing this is to employ a fan beam with a shape proportional to the square of the cosecant of the elevation angle. and the radiation pattern of any real antenna can be made to only approximate the desired cosecantsquared pattern. in this case taken to be 65.

The losses reduce the signal-to-noise ratio at the receiver output.3x2. collapsing loss. There is always some finite loss experienced in the in decibels losses The antenna. the to transmitter lines which connect the output of the radar lower At the 2. They may be of two kinds.01 0. Theoretical (one-way) attenuation of RF transmission lines.15 <£ 0.) predictdiscussed further.497 \l.34 . 49-2B. ^ 3.7 4. sum total can be significant. Waveguide Lines and Transmission Index R. depending upon whether or not they can be predicted with any degree of precision beforehand. Fig.4x1.12] 2. In addition to the losses in the transmission line itself.0 10 Frequency.372 x 0.40 1. used unity) are transmission Plumbing Loss.872 o o x 0. and their possible many are small.872 x 2. there (number less In this section.— Sec. in shown are lines transmission radar per 100 ft for is exceptionally frequencies the transmission line introduces little loss. The antenna beam-shape loss.90 xO. loss (number greater than unity) and efficiency other.0 ^0. Estimates of the latter type of loss must be made on the basis of considerable experience and experimental observations. .84 X1. ASESA. the reciprocal of the simply is One interchangeably. pulses (or integration efficiency) has already been mentioned in Sec.6 and need not be 10 i n i ii : 1 1 — i i i i iii| TTI \ 1.1 I l l l I ll I I I I I I 111 1 1 I I I I II 1. loss mechanisms in a complete radar system. They may be subject to Although the loss associated with any one factor may be variation and uncertainty. qigocycles 100 sizes are in inches Fig 2 29.F. than At the have to be taken into account. Losses not readily subject to calculation and which are less operator of lack fatigue or operator to and degradation able include those due to field prior motivation. 1 The Radar Equation 61 System Losses At the beginning of this chapter it was mentioned that one of the important factors omitted from the simple radar equation was the losses that occur throughout the radar system. Services Armed of (Data dimensions.622 \ 1. and losses in the microwave plumbing are examples of losses which can be calculated if the system configuration is known. unless its length may higher radar frequencies.29.122 x 0. from and are the inside Fittings. attenuation may not always be small and long. 2. 2.12. These losses are very real and cannot be ignored in any The loss due to the integration of many serious prediction of radar performance.

the amplitude of the returned pulses will be modulated by the beam shape.62 Introduction to Radar Systems loss if [Sec. in this case the duplexer. 8. the greater the isolation required from the duplexer on transmission. The precise value of the insertion loss depends to a large extent on the particular design. the maximum antenna gain may be substituted into the radar equation and a beam-shape loss introduced. Since the same transmission line is generally used for both receiving and transmission. the plumbing losses might be as follows .55fe ( A0) '. The latter is normally used.5 db 3. Generally. This assumption is made for convenience.l i — 2 (2. But unlikely that the target will always be in the direction corresponding to maximum gain.12 can occur at each connection or bend in the line and at the rotary Connector losses are usually small. The two methods are equivalent. The one-way-power (two-way. The antenna gain that appears in the radar equation in reality it is was assumed to be a constant equal to the maximum value.voltage) antenna pattern may be approximated by the Gaussian expression exp (— a 2 d 2 ). where 6 is the angle measured from the center of the beam. 2.000 Mc) radar.11). employing an average value of the antenna gain. because the Gaussian function does not represent the sidelobe radiation of the normal antenna. </" 2 _2 exp 2 -5. but similar results can be obtained with any other arbitrary time relationship between the train of radar pulses and the beam center.776/6%. Consider the train of radar pulses to be so oriented relative to the antenna pattern that one of the pulses is coincident with the beam center. It deviates considerably from the actual antenna pattern at angles too far from the center of the beam. In an 5-band (3. For a typical duplexer it might be of the order of 1 db (Sec.4 db 1 . an additional on its way to the receiver. The received signal suffers some attenuation as it passes through the unfired TR tube antenna joint poorly made. the larger will be the insertion loss on reception. maximum for each pulse. The echo signal power received by the radar when the pulse is transmitted and received from the beam center is denoted as Sv The total signal power represented by n pulses received with the Gaussian antenna pattern and integrated without further loss is /2 ? [~i S 1 1 + . for example. The total energy from a group of echo pulses radiated and collected by a practical antenna will be less than that which would have been received from an antenna with a rectangular pattern whose gain was equal to the maximum gain of the practical antenna. By insertion loss is meant the loss introduced when the component. or alternatively. a 2 is a constant equal to 2.50) UB . It is further assumed that a transmitted pulse and its received echo occur at essentially the same point of the antenna pattern. the loss to be inserted in the radar equation is twice the one-way loss.5 Beam-shape Loss. The duplexer also introduces loss when in the fired condition (arc loss) approximately 1 db is typical. Therefore it is incorrect to assume a constant value of gain equal to the Antenna beams are not rectangular hence the amplitudes of the echo pulses will vary as the shape of the antenna pattern.0 0. is inserted into the transmission line. This expression for the antenna pattern is valid in the vicinity of the center of the beam. The loss in received energy may be taken into account by . 100 ft of RG-l 13/U Al waveguide transmission Loss due to poor connections (estimate) Rotary-joint loss line (two-way) Duplexer loss Total plumbing loss db db 0.4 db 1. unless the antenna pattern is rectangular in shape. used. and 6B is the beamwidth measured between half-power points. but if the connection is it can contribute significant attenuation. If the antenna scans past the target.

the reduction in the signal-to-noise ratio of a sine. e 2 \ (2. the beam-shape loss would be increased by approximately 6 db over that given by Eq. Although a well-designed and engineered receiver will not limit the received signal under normal circumstances. Other analyses of bandpass limiters show that for small signal-to-noise ratios. . = 6B /nB where nB is the number of Making this substitution.96 db.53) °Be' 6 Ba = azimuth angle between target and antenna = azimuth half-power beamwidth 6 = elevation angle between target and antenna QBe = elevation half-power beamwidth Limiting Loss.52). only a fraction of a decibel for a large number of pulses will ratio (ratio of video limit level to rms noise level) is as large as 2 or 3. 2. For example.55fe (A0) /^] 2 1+2 2 k =l This may be put in another form by noting that A0 pulses received between the 3-db beamwidth. or to a pencil beam if the target passes through its center. scanning loss are similar in principle to those for computing beam-shape loss. If the antenna is stationary (searchlighting the target). The techniques for computing called the scanning loss.12] The Radar Equation 63 where A0 is the angular separation between pulses. the probability of detection be lowered.Sec. everything else being held constant. an additional loss. all lying uniformly 1. The ratio If the target passes at the outer involves the square because of the two-way transit. we have . The beam-shape loss in this case is simply Beam-shape where da e loss = exp 5 55 Zsl xo Ba e 2 _i_ ils. If this were not so. The beam-shape loss considered above was for a beam shaped in one plane only.wave signal imbedded in narrowband 43 However. the maximum signal received not correspond to the signal from the beam center. would have to be computed. The antenna scanning speed was assumed slow enough so that the gain on transmission is the same as the gain on reception. the loss is if we integrate 1 1 pulses. The beam-shape loss is reduced by the ratio of the square of the maximum antenna gain at which the pulses were transmitted divided by the square of the antenna gain at beam center. It If the applies to a fan beam. intensity-modulated CRT displays such as the PPI and the B-scope have limited 10 dynamic range and may limit. The beam-shape loss (number greater than unity) relative to a radar that integrates all n pulses with an antenna gain corresponding to that at beam center is ft Beam-shape loss = (n _ 1)/2 (2-51) 2 exp [-5. If the signal is limited in the receiver. (2. Beam-shape loss = { 1 + 2^2 n -i)n (2-52) exp(-5. 44 Results derived of the input noise. According to Marcum. Scanning loss is important for rapid-scan antennas or for very long range radars such as those designed to view extraterrestrial objects. target passes through will any other point of the pencil beam. by appropriately shaping the spectrum Gaussian noise is tt/4 (about 1 db). limiting results in a loss of provided the limiting integrated.55k 2/n|) between the 3-db beamwidth. the degradation can be made negligibly small. edge of the antenna 3-db beamwidth. the transmitted pulses and echo pulses all appear at the same place in the antenna beam.

The mathematical derivation of the collapsing loss may be carried out as suggested by Marcum.5 db and L/10) is 1. the collapsing loss L c (m. Nonideal Equipment.56b) For example. Another example of collapsing loss occurs if the video bandwidth is smaller than optimum since the effect is the same as integrating additional noise samples (unless range gating is used). L. 10 who has shown that the integration of m noise pulses.7 db. {2 . The noise added to the signal results in a degradation of the signal-to-noise ratio. or = = + + L c{m . Collapsing loss occurs in cathode-ray-tube displays which collapse range information. integration of The is equivalent to the n signal-plus-noise pulses each with signal-to-noise ratio {SjN)lR c collapsing ratio R c is defined as m+ . 2. The collapsing loss may be defined as where {SjN) m+n are is the signal-to-noise ratio per pulse required for detection when there m extra noise pulses integrated along with n signal-plus-noise pulses. such as the C-scope (plot of elevation angle vs.8 db.12 from an analysis of signal-to-noise ratios alone. accounted for by the collapsing loss. so that the collapsing loss is 1. A collapsing loss also results if the output of two or more radar receivers is combined and only one of the receivers contains signal while the other contains noise. 10 The collapsing loss is thus equal to the ratio of the integration loss L.8fc.64 Introduction to Radar Systems [Sec. in video mixing (where more than one radar output is superimposed on the same indicator) or polarization diversity. however. 2. nor should it be expected that any individual tube . not only with the noise energy contained within that range interval. However. transmitting tubes are not all uniform in quality. for m n pulses to the integration loss for n pulses. are not necessarily related to signal detectability with an operator or an electronic threshold detector.90 and nt = 108 From Fig. azimuth angle). and (S/N)„ is when no extra noise pulses are present. for example.55) number of extra noise pulses n number of signal-plus-noise pulses Mathematically. along with n the signal-to-noise ratio per pulse required signal-plus-noise pulses with signal-to-noise ratio per pulse (S/N). The echo signal from a particular range interval must compete in a collapsed-range display.(40) is 3. but with the noise energy from all other range intervals at the same elevation and azimuth. R< where m = ^^ n (2. realistic A analysis of detectability should be based on the statistics of signal and noise and include the probabilities of detection and false alarm. the collapsing loss is equivalent to the integration of m n signal pulses instead of «.n) noise pulses and the = R ^n \ Ii(m + n) . Collapsing Loss. as.n) = M^O is (Z56a) In terms of the integration-improvement factor It {n). It may happen that the radar integrates a number of unwanted noise samples along with the wanted signal-plus-noise pulses. The transmitter power introduced into the radar equation was assumed to be the output power (either peak or average). assume that 10 signal-plus-noise pulses are integrated along with 30 Pa = 0.

If the operator is fatigued or not handled may sufficiently motivated.1. with the application and the type of tube. as when he becomes overloaded or panicky. data could appear on the PPI at a rate of 300 bits/sec. a reduction of the order of 2 db might be used as an approximate value for system design purposes. or a loss of 7. Much higher losses than a these losses separately. the transmitted power may be less than the design value. The intolerable to an reduced not be the false-alarm time will the the receiver threshold should be increased depends upon the application and may be stability of the circuits. This mismatch in bandwidth can Another factor contributing to operator loss result in a loss in operator performance. Shannon sampling theorem. few decibels have sometimes been attributed to the operHowever.12] will The Radar Equation 65 remain at the same level of performance throughout its useful life.5-db loss in signal-to-noise ratio. for one reason or another. in the event of circuit instabilities— and these need only be which amount by level. This assumes a good operator observing a PPI presentation under good conditions.5. whenever possible. antenna the of revolution resolution cells on one By the revolution were 12 sec. (2. Two loss mechanisms sometimes blamed on the operator are unaccounted-for propagation effects and the It would seem better. other loss mechaator. The amount of loss expected with various types of nonmatched re- A ceivers is shown in Table 9. the operator efficiency is 0. An operator's capacity for searching a CRT display and recognizing radar the presence of target echoes is limited. an operator-efficiency factor approximately given by the following expression: Po is = 0.26)]. Under less favorable conditions. The reduction in power varies.57) 45 where y> is the single-scan probability of detection (blip-scan ratio). to consider losses due to field degradation. Because of the exponential relationship between the false-alarm time and the threshold level [Eq. a slight db. Thus. or design. When the about scan detection probability is 0. recalled that the usual detection criterion indicates the presence of a target It will be whenever the envelope of the signal crosses a threshold. but lacking a better number. therefore. the radar display of number total range.600. Operator Loss. it level can cause a significant change in the false-alarm time.175. this corresponds to a minimum bandwidth of 150 cps. Only a receiver "matched" in the communication-theory sense makes optimum use matched receiver may of the total signal energy contained in the target echo signal. it seems that in a large number of reported cases. so that. occurs when he resorts to guessing. To account for this. power. change in the threshold practice. nisms besides the operator were included in the operator loss.5 db. contained the capacity of the operator bandwidth. the operator efficiency might be somewhat less. the value of the transmitted power inserted into the radar equation should be less than the advertised. If. 2. the information bandwidth that can be adequately a PPI is many times on displayed be can information which rate at The be even less.7. singleonly 0. corresponding to a 1. introduced into the must be loss additional an thus be only approximated in practice. of course. radar system.7y> 2 (2. Based on both empirical and experimental results.— Sec. the efficiency is factor is not linear with detection probability y>. . In necessary to set the threshold level slightly higher than slight calculated. The information bandwidth of a human operator is of the order of 10 cps (20 bits/sec). which is far beyond the operator's ability to handle. for example. 1 A typical value of loss for a nonmatched receiver might be about Variations in receiver noise figure are also to be expected. The operator Even with y> = 1. the in 180 resolvable elements in azimuth and 20 elements one for If the time be would 3.

the inclusion of the above losses into the radar equation should give a realistic description of the performance of the by engineering personnel and experienced radar under normal conditions (ignoring anomalous propagation effects). the gates may be wider than optimum for practical reasons. apply. loose cable TR To minimize field A A Other Loss Factors. water in the transmission lines. the performance usually deteriorates even more than can be accounted for by the above losses. good estimate of the field degradation is difficult to obtain since it cannot be predicted and is dependent upon the particular radar design and the conditions under which it is operating. [Sec. 2. the spectrum and/or shape of the transmitted pulse. when a radar is operated under field conditions. there are very few radar applications which approximate free-space conditions. etc. In a radar with overlapping range gates. radars should be designed with built-in automatic performance-monitoring equipment. Radar characteristics that might be monitored include transmitter power.66 Introduction to Radar Systems Field Degradation. medium in which radar waves propagate can have a significant effect on radar performIn In most cases of practical interest. It may even equipment operated by professional engineers under adverse Factors which contribute to field degradation are poor tuning. and (3) lobe structure caused some all overshadow : by interference between the direct wave from radar to target and the wave which arrives at the target via reflection from the ground. 2. However. There are many causes of loss and inefficiency in a radar. tubes. The effects of non-free-space propagation on the radar are of three categories (1) attenuation of the radar wave as it propagates through the earth's atmosphere. but also for the purpose of keeping them to a minimum by careful radar design. Another factor that has a profound effect on the radar range performance is the propagation medium discussed briefly in the next section and in Chap. 4. Propagation Effects In analyzing radar performance it is convenient to assume that the radar and target are both located in free space. to field conditions. It is important to understand the origins of these losses. especially when the equipment is operated and maintained by inexperienced or unmotivated personnel.13. 11. weak connections. Although they may each be small. One of the few cases which might is a target at high altitude close to the radar with the surface of the earth nonreflecting at the frequency of radar operation. not only for better predictions of radar range. deterioration of receiver noise figure. to some extent. The MTI discrimination technique results in complete loss of sensitivity for certain values of target velocity relative to the radar. incorrect mixer-crystal current. The additional noise introduced by the nonoptimum gate width will result in some degradation. (2) refraction of the radar wave by the earth's atmosphere. These are called blind speeds.13 is When a radar system operated under laboratory conditions technicians. However. degradation. poor tube recovery. receiver noise figure. and the decay time of the TR tube. The blind-speed problem and the loss resulting therefrom are discussed in more detail in A Chap. the sum total can result in a significant reduction in radar performance. the earth's surface and the instances the propagation factors might be important enough to other factors that contribute to abnormal radar performance. Careful observation of performance-monitoring instruments and timely preventative maintenance can do much to keep radar performance up to design level. ance. . degradation of 3 db is sometimes assumed when no other information is available. radar designed to discriminate between moving targets and stationary objects (MTI radar) may introduce additional loss over a radar without this facility.

can see around the curvature of the earth beyond the limits of the geometrical line of sight. The attenuation factor exp (2ajR) must be included in is the radar equation if a is large or if the propagation path is long. In Sec. . the other path includes a reflection from the ground. and is a form of anomalous propagation. Typical values of the attenuation constant are given in Sec. The density gradient of the water vapor can also be such that the radar wave is bent upward. In fact. conditions may be quite favorable and large amounts of bending will be experienced. The gradient of density results in a bending of the radar waves in a manner analogous to the bending of light waves by an optical prism.2 increased at certain elevation angles it is shown that the radar range theoretically can be at the expense of zero coverage at range. or ducting. and the radar range will be less than it would normally be. In some cases.8.wavelength radiation is one of the factors which determine the upper limit X of usable radar frequencies. said to be line of sight is so severe that. Atmospheric attenuation is essentially negligible at the lower end of the radar frequency spectrum. and reduce the range found in this In addition to attenuation by atmospheric gases. Therefore the radar. the radar signal will suffer considerable attenuation near the region of the geometrical line of sight and beyond (Sec. The density of the atmosphere Refraction. but targets within the line of sight. where a the attenuation constant and R is the range. The amount of attenuation depends upon the frequency of operation as well as the gas constituting the medium. the process cannot be considered as standard radar practice except under special circumstances. 2. since the effects of anomalous propagation conditions can be quite pronounced and may be the predictions. but it may be band. Water vapor is the atmospheric component chiefly responsible for the bending of radar waves in the lower atmosphere. An approximate method of accounting for attenuation is to solve for the range as if attenuation were absent manner according to the amount of attenuation. because of the exponential relationship. is not uniform with altitude. with the result that the radar range will be considerably increased. 11. and the echo signal may be larger than if in free space or it may be smaller. free-space by as much as a factor of 2 over the other angles. consequently radar waves will normally be bent around the earth. the solution of the radar equation for range is not simple. the relatively large attenuation quite important at frequencies above of millimeter. The two-way attenuation of the radar signal in the atmosphere is exp (2a/?). The enter into the radar equation for the prediction of range were briefly considered . largest single factor contributing to inaccurate radar The presence of the earth's surface not only restricts the line of sight.14] The Radar Equation 67 The gases and water vapor constituting the earth's atmosphere Attenuation. The atmosphere is usually denser at the lower altitudes. normal radars may be limited to line-of-sight propagation or less. The attenuation of electromagnetic waves in the diffraction region beyond the 11. depending on their relative phase. This condition is called superrefraction.Sec. Summary . Lobing. These abnormal available. In this chapter. When this factor is included. The result is a loss of intensity over that experienced if in free space. 2. 1 1 .6). The two waves can interfere destructively or constructively. some of the more important factors that Prediction of Radar Range. it can also have a serious effect on the radar coverage for arrives over the direct path One paths separate via two target the arrive at Two waves from radar to target. effects on radar could be predicted if sufficient meteorological data were probably expensive and data is of proper type the obtaining of However.14. attenuate electromagnetic radiation. This is unfortunate. if powerful enough. for all practical purposes.

60)].60) (2. (S/N)^ and the propagation factors.61) B nr a sa 1 most radar applications 65A (2. is more appropriate to replace Pav /fr = Pkr by E„ the energy Other auxiliary relations useful when dealing with the radar equation are " "2? ~°Bfr (2.24. sec f = pulse repetition frequency. cps (S/N)i = signal-to-noise ratio required at receiver output (based on single-hit s maximum radar range. cps t = pulse width.6: (2.58) = G= A= p n = = £<(«) = integration efficiency (less than unity) L = system losses (greater than unity) not included in other parameters a = attenuation constant of propagation medium a = radar cross section of target.14 radar equation with the modifications indicated in this chapter.1).49) where nB co t = number of pulses received within half-power antenna beamwidth 6B = antenna rotation rate.59) P. 2. and A the signal-to-noise ratio per pulse and the integration-improvement factor be modified according to Figs.23 and 2. m Fn = noise figure k = Boltzmann's constant = 1. 2. nf for (2. fluctuating target cross section requires that the average value of a be substituted in the radar equation and that false alarm. antenna gain antenna aperture. 2 antenna efficiency number of hits integrated m m 2 r detection) In some applications it transmitted per pulse. rpm P = peak power Pfa = probability of false alarm r = average false-alarm time = linear antenna dimension m fa / Figure 2. depending mainly on the statistics of a. The maximum radar range is a statistical quantity. respectively. given the probabilities of detection The false-alarm probability follows from the specified average false-alarm time [Eq. becomes "max where i?max o4 = PnvGApanEjin) exp (2aK ma x) — — (2.30) 6ft> m (2. [Sec. If n pulses are integrated postdetection. 2. .48) G Jav ~P~t AirAp X2 = Tf = duty cycle r (2.8a.7 may be used to obtain the value of{SjN\. (2.— 68 Introduction to Radar Systems (2. the integrationimprovement factor may be obtained from Fig. n B n Tia.38 x 10" 23 joule/deg T = standard temperature = 290°K Bn = receiver noise bandwidth.

tracking radar.58) can be generally applied to a r is replaced by E t (the energy transmitted during the target observation time t) and if the integration-improvement factor is made equal to unity. The pattern antenna The range. 46-48 radar if Pav // Equation (2.50 1 0.40 _ . and case 2 represents a ratio for a non1 corresponds to the range at which the blip-scan pulse-to-pulse correlation. For example.14] The Radar Equation 69 The radar equation developed in this chapter applies primarily to a pulse radar.2 1. n described by Rayleigh fluctuation cross-section a represents 1 case section. or it may be readily modified to accommodate CW. with a large antenna at high microwave frequencies the Fresnel region canat 8. Examples of theoretical blip-scan curves.30.58) is the effect in propagation of account take properly difficult to anomalous propagation. Fig. MTI. a 1 20-ft-diameter antenna operating region 2 Fresnel the gain in antenna has a value of D /X of about 19 nautical miles.10 0.50. A figure of merit sometimes used to express the relative performance of radar is the transmitter radar performance figure. Case 5 is 10 hits integrated and n.80 - o 0. This is whether antenna the of scan and record on each number repeated many times until sufficient data are obtained to compute the average A .95 — Cose — 1 0. 13. R M = = = fluctuating target is 0.20 0. target the antenna to the focusing can be restored by through the focal plane is the same as that at infinity.0 10". practical method of measuring the Blip-scan Ratio and Detection Probability.70 - 2 0. If the target is in the Fresnel region [R antenna gain G the wavelength]. field is to fly an aircraft on a radial course the performance of search radars operating in or not the target is detected. 2.8 2. T 49 50 Generally. cross for a nonfluctuating Rayleigh cross-section fluctuation with statistics.6 1. It is in order to set an assumed be factors may safety reasonable although manner. CW < +D D = D = = - 0.9. where used as the basis for radar system design. However.000 Mc not always be ignored. an exact upper bound on its effects. 2. correlated scan to scan. target diameter. the and effective area normally associated with the far zone decrease.98 0. although FM-CW. of Perhaps the most important factor not explicitly included in Eq.99 1 0. pulse-doppler.30 0. it is Further consideration of the radar equation will be given in Chap.0.Sec. Fresnel region is much shorter than practical radar ranges and can usually be ignored.60 ° 0. defined as the ratio of the pulse power of the radar to the power of the minimum signal detectable by the receiver. where antenna.90 ~ 0. The radar range equation is strictly valid in the far zone (Fraunhofer region) of the 2 {Da a T) M. and X antenna diameter.05 i 0. (2.

J. p 592 McGraw-Hill Book Company. January. P. 1947. D. These were computed from Fig. July. 11.: "Extrapolation. Barasch. L. 1945. A.. July." John Wiley & Sons. Siegel. vol.: False Alarm Time in Pulse Radar. Ridenour.50." John Wiley & Sons. 1960. The abscissa is plotted in terms of R the range at which the detection . Research Inst. Mueller: Radar Back-scattering Cross Sections for Nonspherical IRE Trans. G. and T. Hollis. New York. and D. 46-156. /. W. 1959. Examples of theoretical blip-scan curves are shown in Fig. 38. 1956. V. F. R. 267-270. p.: "Statistical Theory of Signal Detection. 3. New York.: "Principles and Applications of Random Noise Theory. REFERENCES 1. with Engineering Applications.. H. Middleton: A Theoretical Comparison of the Visual. assuming that the radar integrates 10 hits and that the falsealarm number nf = 10 8 The three curves represent a nonfluctuating cross section (case 5). Cambridge. correlated scan to scan (case 1). 1197-1203. but it attempts to evaluate the performance of an actual radar equipment under somewhat controlled and realistic conditions. October. 1960 Parzen. M.vol 17 dd 940-971 rr J FK ' 1946. 13. Scharfman. IRE Trans vol. Kaplan: Scattering of Electomagnetic Waves by Spheres Michigan Eng. D. vol. 16. K. W. Mass. 2. P. It may be computed from the blip-scan data and the scan rate. S. B. vol. Root: "Introduction to Random Signals and Noise. 1189.. 17. 2. J. The plot of blip-scan ratio as a function of range should not be confused with the cumulative detection probability as a function of range. L. 19.. S. and W. 42. : ' : W. and a Rayleigh cross-section fluctuation with pulse-to-pulse correlation (case 2). Schultz. The experimentally found blip-scan ratio curve is subject to many limitations. N.14 of scans the target was seen at a particular range (blips) to the total number of times it could have been seen (scans). This is called the blip-scan ratio. 1949. Appl. and E. 12. 1956. vol. 25. F. 7. Gere. pp. W. November. and aspect. The aspects commonly considered are either head on or from the rear. IRE. H. IRE. " VV ' 18. Van Vleck." McGraw-Hill Book Company. E. and T. 20. V. Inc. vol. Phvs . a cross-section fluctuation described by Rayleigh statistics. Bell System Tech J vol 23 pp 282-332 1944. King. J. 9. Proc. ' . Proc. Rice. pp." John Wiley & Sons Inc J New York. H. AP-5. New York. 1352-1356." Pergamon Press. Harrington. O.." Pergamon Press. Rept. Wiener. AP-4. : Appl. Mentzer. 2255-20-T. N.. 1960. New York. Wu: "The Scattering and Diffraction of Waves. 1958. 1956. Aural." MIT Radiation Laboratory Series. 1957. 1954.. 51-53. Sleator: The Theoretical and Numerical Determination of the Radar Cross Section of a Prolate Spheroid. Proc IRE vol 37 pp. J. and Meter Reception of Pulsed Signals in the Presence of Noise. King. 145-267. I. Spherical Shell.. : ' ' ' 6. pp. 1954. C. 24. Mathur. Proc. 2. N. Inc New York. J. J. 15. July. vol. Marcum. Weil. L. Back-scattering Cross Section of a Thin. The latter is defined as the cumulative probability of detecting a particular target by the time it reaches a particular range. It is the probability per scan for a particular target at a particular range.. Inc. and T.22. Phys. altitude. IRE R. 770-777.. and Smoothing of Stationary Time Series. Rogers: Signal-to-noise Improvement through' Integration in a Storage Tube. 1950. T. R. Mathematical Analysis of Random Noise. A Statistical Theory of Target Detection by Pulsed Radar. April. Andreasen. B." Harvard University Press. 44. IRE Trans vol AP-4 nn Univ. 8... 1. May. and vol.: "Modern Probability Theory and Its Applications. These are the two easiest to provide in actual field experiments.: "Scattering and Diffraction of Radio Waves. H. Methods of Solving Noise Problems.: Scattering from Dielectric Coated Spheres in the Region of the First Resonance Targets. m . A. pp. pp. 1956. M.: The Measurement and Interpretation of Antenna Scattering. Interpolation. probability for a nonfluctuating cross section is 0. R. IT-6. 2. 266-275.: "Radar System Engineering. and F. 1949. 5. B. IRE. November. Dielectric.70 Introduction to Radar Systems [Sec. Inc. 4. 1955 14. Bendat.. Helstrom. 10. M. New York. Davenport.30. Mathematical Appendix Bennett. pp. 1958. July.. Trans. July. 609-638. pp. vol.

. Hiatt. 344-346. and vol. 43. 1953. A. Manasse. IT-4. 38. Angelakqs: Back-scatter from a Right-circular Cone.. H. R. J. IRE Trans.. vol. J. 1753-1762. pp. September. 23. Conv. V. R. 1958. Proc. vol.: Radar Target Classification by Polarization Properties. C.: End-fire Echo Area of Long... B. pp. February. A. AP-4. G. Senior. 13. 2591-1-H on Contract AF 19(604)-1949. vol. R. F.. 29. Siegel. 1957. AP-8. / 24. : : Effects of Signal Fluctuation on the Detection of Pulse Signals in Noise.The Radar Equation 21. 133-139. Thin Bodies. AST1A Document 1 17533. M. P. 8. 1958. Phys. P. T. Natl. 1957. M. vol. 720-727. and S. sec. Proc. 27. 1960. 269-308. Hu. Ridenour.. vol. pp. 48. 47. IRE. I. 1959. Elbinger. Influence of the Polarization of the Radiated Waves on Radar Detection. IRE Trans. Nesbeda. Inc. 48. Appl. Schwartz. Research. Proc. E. July. and R.. M. R. 12. and Pulse Doppler Radar. E. Proc. B. 1. vo'. ACR-10. Potter. Detection of Fluctuating Pulsed Signals in the Presence IRE Trans. 70. Siegel. : Radar Data Sequences. and H. Thaler: Detection Range Predictions for Pulse Doppler Radar. 1. May. vol. E. 1953. 25. September. fig. 8. 65-80. L. A of the Radar Cross Sections! of Aircraft and Missiles. Burgener. Schultz. vol.. IRE. and S. pp. vol. 1. 37-40. M.. Range Ambiguity Resolution in High PRF Radar. vol. First-order Markov Process Representation of Binary IRE Trans. R. 26. Meltzer. and R. 1956. G. Electronics Conf. IRE Trans. Cohn. Symposium Rept." MIT Radiation Laboratory Series. and D. McGill University. pp. Univ. 48. Proc. Lerner: Loss of Signal Detectability in Band-pass Limiters. Hiatt. Proc. Probability of Detection for Fluctuating Targets. W.. 42. 44. L. Proc. Radiation Lab. rend. 1956. June 22-25. 175-178. R. 156-1511 Sept. 56-64. P. A. Electronics Conf. 1956. 41. Conf. 3. 1290-1296. M. 271-281. S. IRE. IT-2. Jr. 34. Pircher. McGraw-Hill Book Company. R. I. D. pp. 8. 1960. 48. Siegel: Theoretical Method for the Calculation 22. pp.. 4. 4th Natl. "Radar Syjstem Engineering. and R. Radar Polarization Power Scattering Matrix. 49. Mar. 1635-1642.: "An Introduction to Probability Theory and Its Applications. Mich. N.. 1960.. vol. IT-3. IRE. 1960. pp. vol. 293-328. 1957. Crispin.. pp. IRE vol. 48. vol. 34-38. 30. New York. (IRE). June. : April. C: Optical Fresnel-zone Gain of a Rectangular Aperture. 744-760. Goodrich. 65-69. 1960. 50. pp. 71 B. Pulse. p. March. N.. 48. June. vol. October. IRE Intern. Jr. 82-83. pp. 1T-3. Natl. 2d ed. AP-6. Varela. D. Jr. vol. Trans. 1957. pp. pp. June 27-29. K. 1-2. M. IT-6. P. 239. L. 1947. Skillman. vol. corrections in Proc. 1958. pp. IRE Intern. 45. 1960. pp. C.. New York. 1755. I. IRE Trans.: Signal-to-noise Ratios in Band-pass Limiters.) 46. 476-481. pt. 1960. 1956. 36. Copeland. IRE. and D. pp. /. IRE... W. 105-113. Proc.. Univ.. J. Symposium on Radar Detection Theory. {Berkeley) Electronics Research Lab. 35. 44. 1960. IRE Trans. 248-252. Mooney Multiple High-PRF Ranging. John Sons. 32.. vol. J. December. and H. October. C. 27. Weston Study of Surface Roughness and Its Effect on the Back Scattering Cross Section of Spheres. G. H. Ming-Kuei: Fresnel Region Field Distributions of Circular Aperture Antennas. A. 1956. Record. vol. 24. Rept.: Statistical Theory of Target Detection by Pulsed Radar. January. Price. (Quoted with permission of the author. Conv. vol. 46.. July. Sponsler. IRE. pp. Inc. Polk. R. Bussgang. Hay. Cohn. Leger: Elimination of Ambiguities from High Pulse Repetition Rate Radars. 2008-2016. and H. Safran: Unified Analysis of Range Performance of CW... Military Electronics Conv. W. F. March. Symposium on Microwave Optics. 1958. J. vol. 931. pp. Swerling. 1960. K. P. Leger: Suppression of False Range Indications in High Repetition Rate Radars. / Hiatt. Proc. IT-6. Compt.8. 33. Its History and Present Status. 28.. February. 1959. 37. April. pp.: Far Field Scattering from Bodies of Revolution. Davenport. Siegel. vol. 47. S." vol.. Montreal. G.: The Operator Factor Concept. 1960. 66-71. : ONR A A January.. vol. 1960. Calif.: The Interpretation of the Radar Cross Section of an Aircraft Model. Peters. vol. Marcum. Swerling. of Noise. M.. IRE. Shosfak. Record. vol. 40. and K. pp. p. 1957. 1954. L. Wiley & : pp. Appl. Proc. : : j : : A 1960. A. 7. pp. September. 31. Elbinger. pp. . pt. Repjt. K. vol. W. King: Measurement of the Radar Cross Section of a Man. Office of Naval Research Contract N7onr-29529. July 26. pp. A.. R. Weil: Forward Scattering by Coated Objects Illuminated by Short Wavelength Radar. 48. and V. Feller. M. Weil: The Ineffectiveness of Absorbing Coatings on Conducting Objects Illuminated by Long Wavelength Radar. 1630-1635. IRE Trans. 39. May. Sci.. Graves.

the total angular excursion W <j> magnetic wave during its transit to and from the target is AnRfX radians.1) n The doppler frequency fd where yj.= — dj> 4ndR 4nvr . time is equal to a <f> <f> . mission and reception help segregate the weak echo from the strong leakage signal. but — the isolation is A feasible technique for when there is usually not sufficient. an apparent shift in frequency will If R is the distance from radar.. detecting the shift in the echo It is well known in the fields of optics and acoustics that if either the source of oscillation or the observer of the oscillation is in motion.2a) A C = transmitted frequency = velocity of propagation = 3 x 10 8 m/sec 72 . Extremely large isolations between antennas are not necessary when the doppler shift in frequency is used for detection since the presence of a portion of the transmitted signal in the receiver In most instances it is a necessity and is required for is not. the radar to between the radar and the target is 2R/L The distance R and the wavelength X are assumed to be measured in the same units. If the target A change in with respect to are continually changing. transmitted signal is made on the basis of differences in time. harmful. but the time that elapses between the transmission of the pulse and the receipt of the echo Separation of the echo signal and the is a measure of the distance to the target. (3. c = 2 = 2-^I° J!ir (3. separating the received signal from the transmitted signal relative motion between radar and target is based on recognizing the change in the echo-signal frequency caused by the doppler effect. after which the receiver is turned on to listen for the echo. This i s the doppler effect and is the basis of C result. A pulse radar transmits a relatively short burst of electromagnetic energy. R and the phase frequency m d given by doppler angular the This is frequency. eo d = 2„/d where f d vr = doppler frequency shift = relative (or radial) velocity of target with respect to radar shift is =-= T .3 CW AND FREQUENCY-MODULATED RADAR 3. is in motion. in principle.1. frequency. Since one wavelength corresponds to an made by the electroangular excursion of 2tt radians. The echo not only indicates that a target is present. it might be as little as 10 Separate antennas for transthat of the transmitted power sometimes even less. The radar transmitter may be operated continuously rather than pulsed if the strong transmitted signal can be separated from the weak echo. in the two-way path contained wavelengths A of number total the target. The Doppler Effect trans- A radar detects the presence of objects and locates their position in space by mitting electromagnetic energy and observing the returned echo. The received-echo-signal -18 power is considerably smaller than the transmitter power.

whether in a or a pulse radar (MTI) application. 103y ' n M .000 I ' = I | 90°). the early radar experimenters worked almost exclusively with continuous rather than pulsed transmissions (Sec. Historically. I I I !/ I I / T/l I >'1 II I) / \ . Mc Fig.000 1 i i i i i ii 100 1. 3. CW Radar Consider the simple CW radar as illustrated by the block diagram of Fig.000 target relative velocity. Two of the more important early applications of the radar principle were the proximity (VT) fuze and the FM-CW altimeter. either modulated or unmodulated.1.000 100. has had wide application. In addition to allowing the received signal to be separated from the transmitted signal. The proximity fuze was first employed in artillery projectiles during World War II and greatly enhanced the effectiveness of both field and antiaircraft artillery. 3.5).2a. and X — —J— 3. 1. The relative velocity may be written vr = v cos d. 3.s )/ 1 1 / / / - ~~- - 2 10 10 s\ \X\ 1 1 / 1 l/l 1 \A\ 1 1 l ill Mill 10. v r in knots. the doppler frequency is maximum. f . the radar provides a measurement of relative velocity which may be used to distinguish moving targets from stationary objects or clutter.2ft)] as a function of radar frequency and The type of radar which employs a continuous transmission. 73 in cycles per second. where v is the target speed and 6 is the angle made by the target trajectory and the line joining radar and target. /\ \ y\ \ i4- ~ - . Doppler frequency [Eq.Sec. Radar frequency. The doppler is zero when the trajectory is perpendicular to the radar line of sight (0 10. (3. is The CW radar also serves as a means CW CW 3. (3. in 1928.2] If/.000 - = 100 —/ - «///// //// / / 7/// / / / y$/ / / / // // // /V/ / / y / / / // / V / / / /// /yJ' JV /.2b) A plot of this equation is shown in Fig. When = 0. is CW f /a and Frequency-modulated Radar in centimeters. 1 of interest not only because of its many applications. The is transmitter generates a continuous (unmodulated) oscillation of frequency which . The first practical model of the FM-CW altimeter was developed by the Western Electric Company in 1938. although the principle of altitude determination using radio-wave reflections was known ten years CW CW earlier.1.2. 1. but its study for better understanding the nature and use of the doppler information contained in the echo signal.

when the received signal frequency is greater than the (transmitted signal frequency. by an in amount ± fd as given by Eq. is receiving antenna. the frequency from the transmitted frequency/.180 cps when A 10 cm. 2 The narrow-bandpass characteristic of the ear results in an effective increase in the signal-to-noise ratio of the echo signal. The sign of/7 is lost in this process. If audio detection were desired for those combinations of target velocity and transmitter frequency which do not result in audible doppler frequencies. Earphones are not only simple devices. The received echo signal at a frequency/. 3. but the ear acts as a selective bandpass filter with a passband of the order of 50 cps centered about the signal frequency. If exact knowledge of the doppler frequency is not necessary. (a) Simple CW radar block diagram.2b. earphones are especially attractive provided the doppler frequencies lie within the audio-frequency response of the ear. Sometimes both conditions cannot be met simultaneously and a com- promise is necessary. 3. AAAAAAA f«±f* M CW transmitter fo U Detector (mixer) Beat -frequency amplifier Indicator Frequency Fig. The minus sign applies if the distance is increasing (receding target). It might have a frequency-response characteristic similar to that of Fig. For example. the maximum doppler frequency produced by an aircraft with a speed of 600 knots is 6.74 Introduction to Radar Systems scattered. A portion of the radiated energy is intercepted by the target it and is some of in the direction of the radar.2). The upper cutoff frequency is selected to pass the highest doppler frequency expected.2.. the doppler signal could be heterodyned to = . ± fd enters the radar via the antenna and is heterodyned in the detector (mixer) with a portion of the transmitter signal f to produce a doppler beat note of frequency/. The indicator might be a pair of earphones or a frequency meter. 3. With subsonic aircraft targets and transmitter frequencies in the middle range of the microwave frequency region. (b) response characteristic of beat-frequency The low-frequency cutoff must be high enough to reject the d-c component caused by stationary targets. (3. amplifier. The purpose of the doppler amplifier is to eliminate echoes from stationary targets and to amplify the doppler echo signal to a level where it can operate an indicating device. [Sec. If the target received signal will be shifted in where it is collected by the motion with a velocity vr relative to the radar.2 radiated by the antenna. the doppler frequencies usually fall within the passband of the ear. applies if the distance between target The plus sign associated with the doppler frequency and radar is decreasing (closing target). that is. but yet it must be low enough to pass the smallest doppler frequency expected.

single antenna serves the purpose of transmission and reception in both the proximity fuze and the simple radar described above. The amount of isolation required depends on the transmitter power and the accompanying transmitter noise as well as the ruggedness and the sensitivity of the receiver. When the output from the doppler amplifier is of sufficient magnitude. 3. additional isolation is usually required between the transmitter and the receiver if the sensitivity is not to be degraded either by burnout or by excessive noise. which is the detection and location of reflecting objects by "radio" means. Except where the radar operates with relatively low transmitter power and insensitive receivers. 34 comparison of the radio-proximity-fuze block diagram (Fig. CW CW A and Frequency-modulated Radar 75 also be detected and measured by conventional frequency meters. it is not possible to eliminate A CW Antenna 7 Oscillating detector Amplifier fd Firing circuit Detonator f>±ft Fig.2] the audible range. usually one that counts cycles. For example. undesirable. However.3. usually a thyratron. radar block diagram (Fig. completely the transmitter leakage. An example of the radar principle is the radio proximity (VT) fuze. It may seem strange that the radio proximity fuze should be classified as a radar. the firing circuit. 3.3) with the between the two. single tube operating as an oscillating detector acts as both the The doppler frequency can CW A transmitter and the receiver in the radio proximity fuze. The additional noise introduced by the transmitter reduces the receiver sensitivity. These are (1) the maximum amount of power the receiver input circuitry can withstand before it is physically damaged or its sensitivity reduced (burnout) and (2) the amount of transmitter noise due to hum. There are two practical effects which limit the amount of transmitter leakage power which can be tolerated at the receiver. transmitter leakage is not always moderate amount of leakage entering the receiver along with the echo signal supplies the reference necessary for the detection of the doppler frequency shift. The echo signal is detected in the plate circuit of the oscillating detector. The amount of isolation needed in a long-range radar is more often determined by the noise that accompanies the transmitter leakage signal rather than by any damage caused by high power. 3. For example. microphonics stray pickup. As with the simple radar. Isolation between Transmitter and Receiver. If a leakage signal of sufficient magnitude were not present. as in the proximity fuze. but it fulfills the same basic function of a radar. used with great success during World War II for the fuzing of artillery projectiles. the fuze doppler amplifier has a frequency-response characteristic corresponding to the expected CW range of doppler frequencies. and instability which enters the receiver from the transmitter. a sample of the transmitted signal would have to be deliberately introduced into the receiver to provide the necessary A reference frequency. In principle. a single antenna may be employed since the necessary isolation between the transmitted and the received signals is achieved via separation in frequency as a result of the doppler effect.Sec. is triggered to initiate the detonation process. the isolation between transmitter and receiver must be 50 db. VT fuze block diagram. if the safe value of power which might be applied to a receiver were 10 mw and if the transmitter power were 1 kw. In practice. suppose the isolation between the transmitter and CW CW .2) further illustrates the similarity 3.

be the need for reducing the noise modulation accompanying the transmitter signal. Turnstile junctions 6 achieve isolations as high as 40 to 60 db. Isolations of the order of 80 db or more are possible with high-gain antennas. — — adjusted to cancel the transmitter signal that leaks into the receiver via the receiving antenna. when race. If complete elimination of the direct leakage signal at the receiver could be achieved. 10 . An additional 10 db of isolation or more may be obtained. 1 . When the antenna designer is restricted by the nature of the application. 7 is An important factor which limits the use of isolation devices with a common antenna the reflections produced in the transmission line by the antenna. circulator. The amount of isolation which can be readily achieved between the arms of practical hybrid junctions such as the magic-T. ratIn some instances. 5 It will be recalled (Sec. The more directive the antenna beam and the greater the spacing between antennas. The largest isolations are obtained with two antennas one for transmission.3) that the receiver of a pulsed radar is isolated and protected from the damaging effects of the transmitted pulse by a fast-acting switch called the TR. Separate antennas for transmitting and receiving might also be used. Isolation between transmitter and receiver might be obtained with a single antenna by using a hybrid junction. matched antenna with a voltage-standing-wave ratio a is \p\ = (a — 1)/(ct + 1). 70 db at K band. the greater will be the isolation.02. and there will always be some transmitted The reflection coefficient from a missignal reflected back toward the receiver. and as low as 20 db at L band. the VSWR must be less than 1 . One limitation of the hybrid junction is the 6-db loss in over-all performance which results from the inherent waste of half the transmitted power and half the received signal power. an isolation of perhaps 60 db or more might be achieved. The antenna can never be perfectly matched to free space. 8 Further isolation may be obtained by introducing a controlled sample of the transThe phase and amplitude of this signal are mitter signal directly into the receiver. Therefore. the other for reception physically separated from one another. the more stringent will. which short-circuits the receiver input during the transmission period. For example. if an isolation of 20 db is to be obtained. large isolations may not be possible.2 mw of leakage signal appeared at the receiver. 3.22. If the minimum detectable signal were 10 -13 watt (100 db below mw). 9 Another method of increasing isolation between separated antennas is with electromagnetic absorbing material or metallic baffles placed between the antennas. The use of orthogonal polarizations for transmitting and receiving is limited to short-range radars because of the relatively small amount of isolation that can be off the receiver during transmission with a TR-like device CW since the transmitter is obtained. the VSWR must be less than 1. or short-slot coupler is of the order of 20 to 30 db. If 40 db of isolation is required. The greater the desired radar range. or turnstile junction or with separate polarizations. extreme precision is exercised. since echoes it from nearby fixed targets might not entirely solve the isolation problem can also contain the noise components of the transmitted signal. the transmitter noise must receiver were such that 10 1 be at that least 1 10 db (preferably 130 or 140 db) below the transmitted carrier.76 Introduction to Radar Systems [Sec. typical isolations between transmitting and receiving antennas on missiles might be about 50 db at Xband. The lie transmitter noise of concern in doppler radar includes those noise components within the same frequency range as the doppler frequencies. Turning radar is not possible in a operated continuously. Ferrite isolation devices such as the circulator do not suffer the 6-db loss inherent in Practical devices have isolation of the order of 20 to 50 db. the hybrid junction. Both the loss in performance and the difficulty in obtaining large isolations have limited the application of the hybrid junction to short-range radars.

2] CW and Frequency-modulated Radar 77 Although the use of two antennas can provide a high degree of isolation. If a single antenna of area 2A were used for both transmission and reception. is in Transmitting ^ y CW c h ' 'o Oscillator Mixer ht 'o ' + 'if. CW doppler radar with nonzero IF receiver. where the doppler frequencies usually are found. these shortcomings may be overlooked in many applications. Block diagram of superheterodyne. The noise power produced by the flicker effect varies as l// a where a is approximately unity.2 some respects analogous to a superheterodyne receiver. the total antenna area is 2A. Such a receiver is simpler than one with a more conventional intermediate frequency since no IF amplifier or local oscillator is required. 3. > o-W Sideband filter Receiving fo+f f i f antenna ^ > 'o * W Receiver mi <er IF amplifier f. Thus. In addition to the loss of effective area. which is independent of frequency. a loss of a consequence.4 shows a block diagram of the radar whose receiver operates with a nonzero IF. the detector of the receiver can introduce a considerable amount of flicker noise. Separate antennas are shown for transmission and . effective aperture is The receiver of the simple CW radar of Fig. the reduction in sensitivity caused by the simple doppler receiver with zero IF cannot be tolerated. low-power applications this decrease in sensitivity might be tolerated since it can be compensated by a modest increase in antenna aperture and/or additional transmitter power. Receivers of this type are called homodyne receivers. But for maximum efficiency with radar. at the lower range of frequencies (audio or video region). the use of two separate antennas usually results in a somewhat more difficult mechanical mounting and scanning problem than does the single antenna. Flicker-effect noise occurs in semiconductor devices such as crystal detectors and cathodes of vacuum tubes. especially if large isolations are necessary and can be obtained in no other way. 3. The effects of flicker noise are overcome in the normal superheterodyne receiver by using an intermediate frequency high enough to render the flicker noise small compared with the normal receiver noise.1) shows that the single antenna will be capable of 6 db greater performance (received signal four times greater) than two separate antennas of equal total area. However. the simpler receiver is not as sensitive because of increased noise at the lower intermediate frequencies caused by flicker effect. Figure 3.Sec. For short-range. the radar equation (2. If the area of each of the two antennas is A. sometimes called sideband of the local oscillator is replaced by the leakage signal from the transmitter. or superheterodyne receivers with zero IF.4. This results from the inverse frequency dependence of flicker noise. This is in contrast to shot noise or thermal noise. 3. Nevertheless. resulting in reduced receiver sensitivity. 2d detector ft Doppler amplifier Indicator f*'* Fig. CW CW CW . 11 The function Intermediate-frequency Receiver.

Several factors tend to spread the CW .5. 3. If the waveform of the echo signal were known. 3. Since the output of the mixer consists of two sidebands on either side of the carrier plus higher harmonics. the reference frequency is affected by the CW same drift and the difference frequency (IF) remains unchanged. 3. 3.4) is that it be wide enough to pass the expected range of doppler frequencies. where/ and <5 are the frequency and duration of the sine wave. (Note that this is the same as the spectrum of a pulse of sine wave. the IF oscillator is easier to stabilize than either reception. In principle. the echo is not a pure sine wave of finite duration but is perturbed by fluctuations in cross section. Only the IF frequency oscillator need be kept stable in the configuration shown in Fig. Practical receivers can only approximate this characteristic.78 Introduction to Radar Systems [Sec.2. a narrowband filter one just wide enough to reduce the excess noise without eliminating a significant amount of signal energy might be used. 9. 3. provided the IF oscillator is stable. the reference signal could have been generated with a separate local oscillator. One of the requirements of the doppler-frequency amplifier in the simple radar vs (Fig. and /is the frequency variable over which the spectrum is plotted (Fig. its frequency spectrum would be a delta function (Fig. the local oscillator (or reference signal) is derived in this receiver from a portion of the transmitted signal mixed with a locally generated signal of frequency equal to that of the receiver IF. if some pains were taken to keep the oscillator frequency and the transmitter frequency stable. 3. the transmitter or a separate local oscillator. as well as its carrier frequency.56). This type of receiver is sometimes called a sideband superheterodyne. a narrowband filter selects one of the sidebands as the reference signal. The improvement in receiver sensitivity with an intermediate-frequency superheterodyne might be as much Frequency as 30 db over the simple receiver of Fig. the use of a wideband amplifier covering the Frequency expected doppler range will result in an increase in noise (A) and a lowering of the receiver sensitivity.2 Instead of the usual local oscillator found in the conventional superheterodyne receiver. as in the conventional superheterodyne receiver. The frequency spectrum of a finite-duration sine wave has a shape of the form [sin 7r(/-/ )^]M/-/ ). (a) Receiver Bandwidth.5a) and the receiver bandwidth would be infinitesimal.) In many instances.4.2. If the transmitter of the system in Fig. These must be known if an approximation to the bandwidth required for the narrowband doppler filter is to be obtained. the matched filter could be specified as HM'- CW ^A k^ — — signal energy over a finite frequency band.4 drifts slowly in frequency. 3. the only difference being the relative value of the duration <5. 3.2) or the IF amplifier of the sideband superheterodyne (Fig. Frequency spectrum of CW oscillation of (a) infinite quency of the doppler-shifted echo signal were known duration and (b) finite duration. respectively. Since it operates at a lower frequency than would a local oscillator. If the freFig. In most cases of practical interest the expected range of doppler frequencies will be much wider than the frequency spectrum occupied by the signal energy. 3. If the received waveform were a sine wave of infinite duration. Consef quently. The more normal situation is an echo signal which is a sine wave of finite rather than infinite duration. But a sine wave of infinite duration and an infinitesimal bandwidth cannot occur in nature. beforehand. outlined in Sec.

12 The spectrum produced by propeller modulation is more like that produced by a sine-wave signal and The frequency range of its harmonics rather than a broad. is further. Dl(2 s v v From the above relationships the approximate width of the spectrum is therefore feet. The constant k depends upon the manner in which the reflector is For practical reflector antennas. the spectrum may be further widened if the target cross section fluctuates. 3. scanning fluctuations. feet per second. Thus of finite duration and the bandwidth of the receiver must be of the order of on target 6J6B Although this is not an exact relation. but the spread in spectrum nevertheless exists. 7. An antenna consisting of an array of elements can be scanned by electrically controlling the phase shift in each element (Sec. k might vary from 60 to 80. when B is measured in degrees. reflector. independent of the transmitted frequency. it fluctuations widen the spectrum by modulating the echo signal. the target aspect were to change at target aspect of as little as J°. In addition to the spread of the received signal spectrum caused by the finite time on The target. The following is a qualitative argument to justify such a the signal the reciprocal of the time .88)2t>. the spread in the spectrum of the received signal due to the finite time on target would be equal to 18 cps. This is = moving . has been reported 9 that the aircraft cross section can change by 1 5 db for a change in If. It has sometimes been stated that the width of the frequency spectrum due to the finite time on target corresponds to the doppler frequency that would be produced by 9 Although this may be approxithe velocity of the periphery of the rotating antenna. might necessitate an increased receiver bandwidth if it were large compared with the spectral bandwidth of the transmitted signal. for some reason.2] CW a and Frequency-modulated Radar effects are 79 target accelerations. CW radar with an antenna beamwidth of 8B deg scanning at the rate of & target (duration of the received signal) The time on is 6 = s 6 B j6 s sec. the rate of 10°/sec (perhaps a deliberate target maneuver or flight instabilities). The width of the received signal spectrum (1/5) is equal to 6jdB . white-noise spectrum.Sec.2). mately true for a mechanically scanning antenna.fl 6 A approximately equal to the doppler frequency shift (fd 2vJX) of an object The similarity is at a speed equal to the peripheral velocity of the antenna. the echo A modulation of this amount signal would be modulated at a rate as high as 15 cps. Some of these spectrum-broadening Assume deg/sec. just as with a mechanically rotating antenna. and D in X where v is in is v 57.type = = 1 _ 6S 6B _ (0. The beamwidth of a kljD. where antenna such as a paraboloid of diameter D is given by 6B X is the wavelength. Perhaps the most significant reason for stating that no causal relation exists between the spread in frequency and the mechanical motion of the antenna is that the scanning beam need not be generated by mechanically moving a reflector antenna. The peripheral velocity of the antenna typical value. because of the finite time on target. A similar argument applies to a target flying through the beam of a stationary antenna.. still which tend to broaden the bandwidth considered below. No part of the antenna is physically in motion. it appears to be a fortuitous result with no physical significance. it is a good enough approximation for purposes of the present discussion. If the antenna beamwidth were 2° and if the scanning rate were 36°/sec (6 rpm). propeller modulation depends upon the shaft-rotation speed and the number of It is usually in the vicinity of 50 to 60 cps for World War II aircraft propeller blades. The echo signal from a propeller-driven aircraft can also contain modulation components at a frequency proportional to the propeller rotation.7). with 65 a illuminated. In a particular case. etc. 6 S in degrees per second. conclusion. claimed to be fortuitous.

the doppler frequency shift may not be known precisely. Therefore approximate methods must ordinarily be used to obtain the filter characteristic. The composite effect of the various spectrum-widening factors is difficult to predict. When the band in which the doppler frequencies are expected is known. a bank of narrowband filters spaced throughout the frequency range permits a measurement of frequency and improves the signal-to-noise ratio. the exact shape of the received waveform is not likely to be known. required receiver bandwidth is U= or If a r is in feet per 2a r At X = 2a r (llAfd ) X (3. of Since the time At in which the doppler a filter of bandwidth Af a is approximately 1 jAf d changes frequency by an amount Afd should not be less that the filter build-up time. X 2a r At X J A filter of bandwidth Afd will just accommodate this change in frequency (assuming all . In some instances.2. If the received waveform were exactly known. 3.6 is at IF. but not so wide as to introduce more noise than need If the target aircraft — — . It might permit the detection of propeller-driven aircraft passing on a tangential trajectory. the other band-widening factors to be negligible). then performs a 2g maneuver a moderate maneuver for a military fighter but a large maneuver for a commercial aircraft the receiver bandwidth must be approximately 20 cps when the transmitted wavelength is 10 cm. there is no assurance that a proper matched filter could be readily constructed. 9. When the doppler-shifted echo signal is known to lie somewhere within a relatively wide band of frequencies. The response time. If one effect were much larger than all the rest. or video portion of the receiver. In many cases. The change in velocity causes a change in doppler frequency Af d equal to . 2 Ay. all knowledge of the exact value of doppler velocity is lost. the increased bandwidth results in increased noise and reduced sensitivity. CW radar since it might mask the target's doppler signal or it might cause an erroneous measurement of doppler frequency shift. the receiver bandwidth could be approximated by the rms value of the individual bandwidths. The filter bank diagramed in Fig. These filters can be in either the RF. the receiver passband may be widened to include the entire range of expected doppler frequency. obviously the other factors could be neglected and the receiver characteristic determined by the dominant factor. and even if it were.2 engines. where ar is the acceleration of the target with respect to the radar. it would be possible. propeller modulation can be of advantage. a further widening of the received signal spectrum can occur. The bandwidth of each individual filter is wide enough to accept the signal energy. or build-up time. in theory. . IF.80 Introduction to Radar Systems This could be a potential source of difficulty in a [Sec. even though the doppler frequency If the target's relative velocity is shift is zero. not constant with time but is changing. In the event the various effects were all of equal magnitude.3 fl ) Afa = {^Y second per second and X in centimeters. Although the received echo signal will then fall somewhere within the receiver bandwidth. Furthermore. 3. The change in relative velocity Av r over a time At is equal to a r At. to compute the shape of the receiver characteristic which maximizes the signal-to-noise ratio by the matched-filter theory of Sec. However.

CW and Frequency-modulated Radar 81 The center frequencies of the filters are staggered to cover the entire range of doppler frequencies. (a) Block diagram of IF doppler filter bank. 5.2 No. IF amplifier Filter I— » Filter No. whether in the IF or video. The more niters used to cover the band.6. Therefore the sign of the doppler shift is lost with a video filter bank. 4 Det. The improvement is still in signal-to-noise ratio filter bank is not as good as can be obtained with an IF measure the magnitude of doppler frequency Filter preserved.Sec. the bank of doppler filters may be replaced by a single narrowband tunable filter which searches in frequency over the band of expected doppler After detecting and recognizing the signal. the less will be the maximum loss experienced.— Mixer Filter No. a frequency which lies to one side of the IF carrier appears. increases the complexity of the receiver.) One advantage of the foldover in the video is that only half the number of filters are required than in the IF filter bank. but the greater the probability of false alarm. 13 . When the system requirements permit a time sharing of the doppler frequency range. at same video frequency as one which lies an equal amount on the other side of the IF.2 instead of in the IF. with a video ability to filter 3. its search in frequency for additional signals. the maximum reduction in signal-to-noise ratio of a signal which lies midway between adjacent channels compared with the signal-to-noise ratio at midband is 3 db. and it cannot be directly determined whether the doppler frequency corresponds to an approaching or (The sign of the doppler may be determined in the video by other to a receding target. * Filter Na/2 Det. the frequencies until a signal is found.2] be. . (a) f\ h h U (b) Frequency Fig. measurement are accomplished by visual observation. the foldover. and detection reeds in which A bank of overlapping doppler filters. 1 — — — — Det. bank. 3 Det. There are many techniques which may be used to achieve narrowband IF filters. bank of narrowband filters may be used after the detector in the video of the simple A CW radar of Fig. If the filters are spaced with their half-power points overlapped. continue to programmed filter may be One of the techniques for accomplishing this is similar to the tracking speed gate described in Sec. means.7 or to the phase-locked filter. Mechanical filters and crystal A simple video filter bank may be obtained with vibrating filters are two possibilities. 3. 3. as described later. but the Because of No. (b) frequency-response characteristic of doppler filter bank. after detection. -*.

radians/sec <o d doppler angular frequency shift a constant phase shift. In channel A the signal is processed as in the simple radar of Fig. This also requires a relatively long time in which to make observations. This might be determined with CW separate rier. Although the doppler-frequency spectrum "folds over" in the video because of the action of the detector. 3. from The zero-doppler-frequency component has. Sign of the Radial Velocity. A relatively long observation time is necessary to.5) amplitude of transmitter signal a constant determined from the radar equation co angular frequency of transmitter. Furthermore. 3. moving targets are to be distinguished stationary objects. Spectra of received signals. in practice. In some applications of radar it is of interest to know whether the target is approaching or receding. borrowed from single-sideband communications. In the multiple-filter bank. known as the phasing method. if the echo frequency is greater than the carrier.2 If.7). a finite bandwidth due to the finite time on target.4) from a moving r target will be E = where kx E cos [(m ± m d)t + </>] (3.7.target signals In the wideband I F where a bank of filters is not it would be necessary to center a rejection band about the IF frequency in order to remove stationary targets.2b) fixed clutter concen- CW fo Frequency la) trated in a finite spectrum about zero frequency. shift. (b) (c) approaching The direction of target motion might be determined by measuring the doppler frequency as a function of time and observing whether the frequency is increasing or decreasing. The the target If the echo-signal frequency lies t! Frequency 16) motion may also be found from the change in amplitude of the received signal with time. variations in the amplitude of the echo signal from a complex target can be considerably greater than the change in amplitude due to the change in range.8. and therefore the direction of target motion. However. 3. removes the stationary. The clutter-rejection band of the doppler filter must be wide enough to accommodate this spread. may be found by splitting the received signal into two channels as shown in Fig. The received k. video filter characteristic of the simple also serves to remove the energy of radar (Fig. the zero-doppler-frequency component must be removed. 3. (a) No relative doppler target no . in any of the above techniques. which depends upon range of initial detection (f> The sign of the doppler frequency. the target is approaching (Fig.82 Introduction to Radar Systems [Sec. motion target. If the transmitter signal is given by t E =E the echo signal cos m t (3. it is possible to determine its sign from a technique. receding target.2. below the carreceding. Fig. the observation interval. clutter fluctuations. this is not always a satisfactory method. and equipment instabilities. filters located is on either side of the intermediate frequency. removal of those filters in the vicinity of the RF or IF carrier used. E = = = CW . within n Frequency (c) reliably detect a significant change. The low-frequency cutoff in the . 3. 3. since the echo signal does not vary rapidly with range except at short direction of target ranges.

9 for both transmission and reception.6) The other channel is similar. In channel B.» Sec. 3. 14 The direction of motor rotation is an indication of the direction of the target motion. determining the relative phase relationship between the two channels is to apply the outputs to a synchronous two-phase motor. although a single antenna can be employed if proper duplexing means are used. 3. two-phase motor.8a) the other hand. antenna >j CW transmitter y • 90" phase shift ' Mixer Receiving A Channel A Synchronous motor antenna s y • indicator Mixer Li Channel B Fig.8i>) The sign of a> d and the direction of the target's whether the output of channel Transmitting B leads or lags the output of channel A motion may be determined according to One method of .8. An electronic technique for measuring the relative phase between the two signals is shown in Fig. except for a 90° phase delay introduced signal. 3. 15 To simplify the description of operation. the outputs from the two channels are EA (+) On = fc 2 £o cos (m d t if + <j>) EB(+) = k 2 E cos (m a t <f> (3. Therefore a 90° phase shift is introduced between the . The received signal is divided into two channels (A and B) and fed into separate detectors. a separate antenna is shown in Fig. EA (~) = k 2 E cos (w d t - <j>) EB (-) = k 2 E cos ico d t - <f> - J (3. Measurement of doppler direction using synchronous. A portion of the transmitter signal is fed directly into the detector of channel A. This has been used in a rate-of-climb meter for vertical take-off aircraft to determine the velocity of the aircraft with respect to the ground during take-off and landing.7) approaching (positive doppler).9. 3.2] signal CW EA = and Frequency-modulated Radar 83 and a portion of the transmitter heterodyne in the detector (mixer) to yield a difference signal k2 E cos (±w d t + 4>) (3. the reference from the transmitter is delayed 90°. the targets are receding (negative doppler). The output of the channel B mixer is in the reference EB If the target is = k 2 E cos [±o> d t + <f> + y + +A (3.

called the downgate. The time taken in traveling from the radar to the target is R /c.9) where an arbitrary phase shift. 3. targets. Measurement of doppler direction as used for VTO aircraft rate-of-climb meter. as in the system of Fig. The sign of the phase shift determines the direction of motion.8. Derivation of Doppler Frequency Shift. Logue. a pulse will be generated (5) from the upgate to indicate a receding target. 3.) (After J A output (1) with the output of the inverting circuit (4) will indicate approaching.2 doppler beat notes in the two channels. The form of the transmitted signal is taken to be limiter sin (cv + <f>o) (3. The signal received at the stationary target is the same . since it is the argument of the sine factor and not the amplitude which is of importance in deriving the doppler frequency shift. To determine the sign of the 90° phase shift. but not receding. the two signals are first amplified and limited. Assume. which indicates direction as well as magnitude of the doppler frequency. Rate-of-climb meter of (a) ( 1 ) Limiter A T i r (4) Inversion of B » (2) Limiter 8 r (b) (5) Up gate [coincidence of —'*)] (3)and(1 (3) Differentiation of B (6) Down gate [coincidence of (4)and(1)] Fig. taking account of the time delays in the transit of the energy from the radar to the moving target and back.9* shows the limited waveforms (1 and 2). A similar comparison in a coincidence circuit. The effect of a moving target on the frequency and phase of the radar echo may be derived from simple considerations of the voltage waveforms. as well as all other waveforms considered here.84 Introduction to Radar Systems [Sec. In the rate-of-climb meter a receding target is one which is ascending. The amplitude of the transmitted signal waveform. The output from limiter . Figure 3. l& Electronics. The waveform from B is differentiated (3) and also inverted (4). 3.4(1) and the differentiated output of B (3) are compared in the coincidence circuit labeled the upgate.9. initially. Approaching targets produce no output from the upgate coincilimiter dence circuit. The pulses from the two coincidence circuits are counted and displayed on a zero-center-scale microammeter. is assumed to be unity. that the target is stationary at a distance R from the radar. If both signals 1 and 3 are positive. where c <f> is is the velocity of propagation.

11) If the target is in but will vary with time.16) Thus the moving-target echo is shifted in frequency by the amount ± co d and in phase by ^co d t as compared with the signal that would have been received from a stationary target [Eq. The high-voltage modulator needed to pulse a power tube cycle is is not found in a unity. If the velocity of the target with respect to the radar is v r and if the acceleration may be taken to be zero. Consequently.13) into Eq.12) gives the signal from the to a receding target.2] as that which was transmitted at the target is sin CW _ <j + is is and Frequency-modulated Radar jc in the past. as is techniques. while the plus sign applies Substituting Eq. where radar at a moving target as sin \co t-^±^(t c c Q + <f>o\ = sin w (l±^)f-^ (R ±Vo) + <£o (3.) (3. is not capable of determining range. When used for the detection of targets at Advantages and Limitations of radar is characterized by simpler equipment than a short and moderate ranges. CW radar. the distance will not be constant.11)]. whereas the typical pulse-radar receiver bandwidth is measured in terms of megacycles. ot _ ^S + ^ (3. motion with respect to the radar. the pulse radar of equivalent detection (range) capability. (3. R(t) =R T v r (t -g = . (3-13) R is the distance between radar and target at the time t t The minus sign associated with v r applies for a target approaching the radar. The receiver bandwidth of CW CW CW CW CW in terms of kilocycles or less. (3.15) C Since = 2a> v r /c. (3. the distance to the target .12) where the distance R(t) is a function of time. 3. the received echo signal becomes <»d)t sin (ft>o ± ~ -^2-2 T m a + t <£ (3.10) The echo earlier signal back at the radar same as the signal at the target a time RJc than Eq. 85 by the radar a time R Therefore the signal Wo ^ fa the = S i n (a. The difference between and pulse-radar techniques may be likened to the difference between radio and TV radar.14) The echo signal that signal signal received back at the radar at any instant of time is the same as the was at the target a time R{t)jc earlier.Sec.10) and sin («* -2 -^-° + A. 9 The simple radar is usually measured the conventional pulse radar. (3. Peak power is less in the CW radar since the duty is Electrical breakdown as a result of high peak power usually not a . the received echo from a moving target may be written sin «„ co d r — 2R(tj c J + <t>o\ = sin *('±*?)' 4>0 (3. however. Radar. The signal at the moving target may be written R(t) sin \cd ] + 4>o) (3.

Its ability to handle multiple targets can be increased by providing resolution in the doppler-frequency domain as with a bank of narrowband doppler filters. transmitters are smaller in size and weight than comparable pulse transmitters.86 Introduction to Radar Systems [Sec. The sharper or more distinct the timing mark. radar has found wide application. In spite of its limitations. 16 and as an aircraft navigation aid (Sec. CW CW CW CW duplexer recovery time. frequency. 17-19 3. The transmitter noise that finds its way into the receiver degrades the receiver sensitivity. The pulse radar has no similar limitation to its maximum range because the transmitter is not operative when the receiver is turned on. 3. The minimum range of a pulse radar depends on the extent of the pulse in space and the factor in power. that is. the more accurate the measurement of range and the broader the transmitted spectrum. The narrower the pulse. Small or zero relative velocities occur for targets whose paths are perpendicular to the radar beam. The simple radar is usually a single-target device.3. Since the radar uses the doppler frequency shift for detection. the broader will be the transmitted spectrum. In a typical application. it permits moving targets to be discriminated from stationary objects (clutter). The number of targets that the radar can resolve at any one time is equal to the number of doppler filters. Some sort of timing mark must be applied to a carrier if range is to be measured. is more complicated than the simple pulse radar. A radar can operate. for equivalent detection capability. called MTI radar. The power limitation is different from that in the pulse radar. Perhaps the application of closest concern to the reader is its use as a radar speed meter extensively CW CW CW CW CW CW CW A CW employed by traffic-enforcement agencies. in principle. Such a radar. tangent or crossing trajectories. But the more distinct the timing mark. either amplitude. the measurement of railroad-freight-car velocity to control humping operations. The timing mark permits the time of transmission and the time of return to be recognized. few of the other typical applications to which the radar has been applied include the detection of tornadoes. the transmitter can be 25 to 50 per cent as heavy as a corresponding pulse transmitter. the more accurate the measurement of the transit time. Its use in the radio proximity fuze and as a rate-of-climb meter for VTO (vertical take-off) aircraft has already been mentioned. A widely used technique to broaden the spectrum of radar is to frequency-modulate the carrier. or phase. A pulse radar also may be made to discriminate between moving targets and stationary objects by use of the doppler effect. 3. The average transmitter is of comparable magnitude in and pulse radars. The timing mark is CW CW . This limitation can be overcome by increasing the bandwidth of the transmitted signal as with frequency modulation or by transmitting two or more frequencies simultaneously. The spectrum of a transmission can be broadened by the application of modulation. even though the magnitude of the vector velocity might be large. since in a radar the maximum power is dependent upon the amount of isolation and the transmitter noise.3 equipment design as it might be with pulse radar. Perhaps one of the greatest shortcomings of the simple radar is its inability to obtain a measurement of range. Both the radar and the pulse MTI radar are blind to targets with zero or small relative velocities. This follows from the properties of the Fourier transform. Therefore a finite spectrum must inability of the simple The CW Radar CW radar CW •>( of necessity be transmitted if transit time or range is to be measured. An example of an amplitude modulation is the pulse radar. however. There is a practical limit to the amount of power that can usefully be employed with a radar. especially where the measurement of velocity is important. Frequency-modulated to measure range is related to the relatively narrow spectrum (bandwidth) of its transmitted waveform.4). against targets down to almost zero range.

10Z>. (a) Linear frequency modulation .10a. such as a crystal diode. (6) triangular (c) beat note of (6). the beat frequency due only to the target's range. necessarily be triangular. 2 -^fo (3. line Fig.10c for Fig. Fig 3. the transmitter frequency is changed as a function time.frequency as a function of region. The beat note is of constant frequency except at the turn-around the beat frequency is If the frequency is modulated at a rate/m over a range A/. a beat note/ element 6 in a nonlinear note (difference frequency) is a measure of the is no doppler frequency shift. an echo at a object reflecting there is a If in 3. in shown is time resulting beat. Solid curve represents frequency modulation dashed curve represents echo. the frequency cannot be continually changed in one Periodicity in the modulation is necessary. 3. echo signal is heterodyned with a portion of the transmitter signal If there will be produced.3] CW and Frequency-modulated Radar 87 The transit time is proportional to the difference in frethe changing frequency. Frequency-time relationships in FM-CW radar. solid by the The dashed line in the figure represents the signal will return after a time T = 2R/c. . Sec.. as in the triangular- CW frequency-modulation waveform shown in Fig. The modulation need not The shape. c c Thus the measurement of the beat frequency determines the range R.J7) radar. r is the beat r frequency is If the rate of change of the carrier frequency is/ the beat echo signal. The greater the of the measurement the accurate more frequency deviation in a given time interval. it can be sawtooth. where f target's range and/. If the =/ . the transit time and the greater will be the transmitted spectrum. fr=JoT= In any practical direction only. transmitter quency between the echo signal and the transmitter signal. 3. .10. Range and Doppler Measurement. 2±Zi <£•) z^z: Time transmitted signal. In the frequency-modulated CW radar (abbreviof time in a known ated FM-CW). sinusoidal. or some other triangular 3. modulation. Assume that the distance R. as shown with linearly increases frequency transmitter manner.

The doppler frequency shift causes the frequency-time plot of the echo signal to be shifted up or down (Fig. 88 Introduction to Radar Systems [Sec. and the doppler frequency shift/. \[fh (up) +/. Similarly. (a) Transmitted (solid curve) and echo (dashed curve) frequencies.3 . If/. it is decreased. the beat frequency/. On one portion of the frequency-modulation cycle. If this In the above. Ideally. (J>) beat frequency. the target is approaching the radar. If. portion of the cycle will be the difference between the beat frequency and an erroneous range measurement FM Transmitted signal Received signal 5sS> fr+fd ±±J- ^ [b) =£ ^ Time Fig. (down) is the sum of the two [Eq.+/* the average beat frequency. [Eq."> » ' ) —>— Amplifier —» Limiter > Frequency counter > Indicator Fig. 3. the isolation between transmitting and receiving antennas is made sufficiently large so as to reduce to a negligible level the transmitter leakage signal which arrives at the receiver via the coupling between antennas. while on the other portion.19a) (3. the target was assumed to be stationary. for example. by switching a frequency counter every half modulation cycle. Frequency-time relationships in radar when the received signal is shifted in frequency by the doppler effect. one-half the difference between the frequencies will yield the doppler frequency. for example. This assumes The range frequency/ may be extracted by measuring . 3. 3. on the decreasing portion. applicable. or up.19a)]. A portion of the transmitter signal acts as the reference signal required to produce the beat frequency. Transmitting A block diagram illustrating the principle of the FM-C W radar is shown in Fig. Block diagram of FM-CW radar.12a). (up) and/. (down)] =/. 3.196)].11. that is. 3. the beat frequency fb (up) produced during the increasing. a doppler frequency shift will be superimposed assumption is not on the FM range beat note results. It is introduced directly into the receiver via a cable or other direct connection. (3. the beat frequency (Fig. (3. The frequency of the amplitude-limited beat note is usually measured with a cycle-counting frequency meter calibrated in distance.— — .12. FM-CW due to the range/. /»("P)=/r-/i (3. 3 1 1 antenna J Receiving ?! < FM transmitter —< ' Referenc e sig nal antenna . The beat frequency is amplified and limited to remove any amplitude fluctuations.196) / 6 (down) =/. (down) are measured separately.126) is increased by the doppler shift.

(3. there will be a corresponding to each target. „. a single frequency corresponding to a single target may be singled out and continuously observed with a narrowband tunable filter. (3.18) to each. / This is certainly ous since a sinusoidal or almost sinusoidal frequency modulation is easier to obtain with The beat frequency obtained with practical equipments than are linear modulations.21) =V r sin 2nUt sin 27rfm -T)+£f Aim (t T) The received signal [Eq. interpretation of the measurements the inequality sign between f r and/j. (3. (3. 3. over a modulation cycle. When more than will contain one target is present within the view of the radar. or no application of the If the FM-CW radar is used for single targets only. the problem of resolving targets and measuring the range of each becomes more complicated. it advantageis not necessary to employ a linear modulation waveform. and the difference meter. an incorrect > occur with a high-speed target at may result./ But if the motion of the targets were to produce a doppler frequency shift. If it is not known that the roles of the meters are reversed because of a change in range. Assume the transmitted signal to be sinusoidally modulated with a voltage waveform vt = V sin t (2*/ ^ The voltage received from the vr target is + ^f sin 2vfJ\ Aim delayed by a time T = * (3. and Frequency-modulated Radar 89 If> on tne otner hand. modulation cycle as it is with linear over the sinusoidal modulation is not constant beat frequency measured average the shown that it may be However.18).23) . or if the mixer were not operated in its linear region.20)] are heterodyned in a mixer to give a difference-frequency signal of vb = kVt VT sin sin (tfm T) cos (^ Jm 2nfm (* - |) l//m . they must be separated from one another. This might be accomplished with a bank of narrowband filters. the averaging meter will measure doppler velocity. when substituted into Eq. the range to each component frequency target may be determined by measuring the individual frequency components and applying Eq. or if the frequencymodulation waveform were nonlinear. +2nf„T\ we may write (3. and consequently little radar in this mode of operation seems to have been made. If the system is linear. In principle. In many cases the advantages of the multiple-target FM-CW radar do not outweigh the practical difficulties inherent in its realization.y^ fd such as might fashort range. such as in the radio altimeter. To measure the individual frequencies.3] /r CW < . yields the correct value of FM target range. or alternatively.21)] and the transmitted signal [Eq. the mixer output more than one difference frequency. the roles of the averaging and the difference-frequency measurements are reversed. T «a -n-f and — Jm Vh t sin nfm T s» -n A/T Therefore the voltage waveform of the difference-frequency signal becomes = k V F [sin 2nf T + r it AfT cos (2w/m / - Ttfm T)\ (3 .22) where A: is a constant of proportionality. modulation.20) 2R/c and may be written (3.Sec. sin Trfm Since T< m T.

The large target cross section and the relatively short ranges required of altimeters permit low transmitter power and low antenna gain. To extract the doppler frequency. The reflex klystron offers the advantage that it can be At UHF frequencies (up to However.7rfm T) = t 77 AfTfm sin {2nfm t . MnM = fr (3. The average of the beat frequency over one-half a modulating cycle is fb = TnT 1 n A/T/m Sin (277/m/ ~ ^» T + *)dt = n A//m Tcos irfm T (3.nfm T + tt) (3. One of the major applications of the FM-CW radar principle has been as an altimeter on board aircraft to measure height above the earth. if the average frequency is measured. 1 or 2 Gc) the triode can supply the necessary transmitter than several gigacycles. (3. Backward-wave oscillators might also be used. or they may be mechanical-modulated by vibrating an internal reed assembly which varies the capacity across the straps of the anode cavity. argument of Eq. A f = 2Affm T=. and its inertia causes a rounding of the may be electronically frequency-modulated. \ ^^o \ \ to / / / / / / / / / / / / Af ' Time Fig.) the range. —* waveform. Since the relative motion. 3. the range frequency can be extracted as before. either the klystron (reflex oscillator or amplifier) or the magnetron may be used.23) with respect = . the effect of the doppler frequency shift may usually be neglected. the modulation waveform must have equal upsweep and downsweep time intervals.13. provided the average beat frequency is measured. FM Radar Equipment. at frequencies greater CW electronically frequency-modulated by changing the reflector voltage. and hence the doppler velocity between the aircraft and ground. a driving voltage with triangular waveform could produce a frequency modulation with the desired triangular shape.3 The frequency may be found by to time. is small. it can be shown that any reasonable-shape modulation waveform can be used to measure / / ' / J\ -£\ O\ \r> ^ \-33 \ » / / 1 ft \\ 1 N \ 3\ \« "SA \«a.24) The minus sign obtained from differentiation of the cosine is equivalent to a phase shift of 77 radians. /«.26) Although the above example assumed the modulation waveform to be sinusoidal. Example of a practical frequency-modulation (From - Capelli. 3. 20 21 If the target is in motion and the beat signal contains a component due to the doppler frequency shift. the reed does have mass. power.90 Introduction to Radar Systems differentiating the [Sec.25) Since fm T<4 and cos nfm T & b 1.f (* A/TX277/-J sin (2nfm . 22 IRE Trans. However. If the vibrating reed assembly had no mass. CW magnetrons .

the modulation is linear approximately 60 per cent of the time.200-4. Too little or too much reference signal lowers the sensitivity of the receiver. There is.000 0-20. 3. the receiver might consist of a crystal mixer followed by a low-frequency amplifier and a frequency-measuring device. An example of a practical frequency-modulation curve is shown in Fig. 22 The exact shape of the modulation waveform is not important so long as only a single target is within the view of the radar and if the beat frequency is averaged over a modulation cycle. Even though the transmitter noise may be considerably reduced in the direct reference signal. as was shown in Fig. Unfortunately. most practical devices which can be readily frequency-modulated by electronic means also produce undesirable amplitude modulation as a consequence. 8. typical FM radar altimeter. typical installation of a radar altimeter with separated receiving and transmitting The receivers employed in FM-CW radars CW discussed earlier in this chapter.000 Transmitter t 0-10.13. 3.11. The direct connection permits better control of the magnitude of the reference signal. a rounded turnover cannot seem to be avoided.8). Mechanical resonances in the vibrating reed will further distort the waveform. The function of the reference signal can be performed by the transmitter leakage. however. In practice. the ultimate performance will be determined by the unavoidable leakage signal and its noise components which find their way into the receiver by way of antenna coupling or by reflecThe noise which accompanies the leakage signal may be tions from nearby objects. The distortions in exist in do not The parameters of a 3. are given in Table AN/APN-22 Radar Altimeter t 4.: Sec. reduced by improving the isolation between the transmitting and the receiving antennas. and as a result. Characteristics of the Frequency Transmitter power Frequency excursion Af Modulation frequency fm Mc Antenna beamwidth Range Over land Over water Accuracy: 0-40 ft 40-20. 3. 23 are similar to those of the simple radar In its simplest form. that the achievement of a perfectly linear triangular frequency modulation is a difficult task. Although proper shaping of the driving voltage minimizes the effects of mechanical resonances and inertia. A A .000 ft ft ±2 ft ft ±5% Mechanically modulated CW magnetron Wimberly and Lane. the crystal mixer can be made to operate more efficiently.1. the AN/APN-22. The error in range due to the separation of transmitter and receiver can be more readily compensated for if the reference signal is introduced into the receiver via a known length of cable rather than an unknown length of leakage path. a practical limit to the amount of isolation which can be achieved.1. therefore. some distortions still occur. Another advantage of supplying the reference signal by a direct connection is that transmitter noise may be reduced with a balanced mixer (Sec. It seems.5 watts 70 Mc 120 cps 60° Table 3. just as too little or too much oscillator power degrades the sensitivity of a superheterodyne receiver. This is similar to the conventional crystal video receiver except for the presence of the reference signal necessary to extract the difference frequency and the range. better technique for introducing the reference signal in the receiver is by direct connection.3] CW and Frequency-modulated Radar 91 frequency modulation if a triangular driving voltage is applied. 21 the frequency-modulation waveform caused by mechanical inertia an oscillator which is electronically modulated.400 1.

either side of/ (?) and separated from/ (r) by the lower sideband the local-oscillator frequency IF The filter selects f (t) and the upper sideband.3 antennas on an aircraft might provide 65 to 70 db of isolation.14. output of the mixer consists of the varying transmitter frequency f (t) plus two sideband frequencies. and the doppler velocity frequency/.. The IF signal is amplified and applied to the balanced detector along with the local-oscillator signal/IF The output of the detector contains the beat frequency (range frequency and the doppler velocity frequency). The \ /&(/) •M transmitter ) Timing signal 1 f (t) fir 1 Mixer Local oscillator 1 1 f QU ) 1 1 <f (t >->if \ 4 Sideband filter Doppler Switched frequency counter vejocity foit) -fir ' ' \ J A Low-frequency ff f U-T) Receiver IF flF+f* Balanced detector \\ mixer amplifier amplifier \ ' (f. 3. The sideband superheterodyne receiver. but not the carrier or other sideband. . The local-oscillator frequency IF should be the same as the intermediate frequency used in the receiver. since the rate of change of altitude is usually small.{ t-D-fn (t Average frequency counter Range Fig. The selection of the local-oscillator frequency is a bit different from that in the usual superheterodyne receiver. transmitter noise can cause erroneous range information. 3ll4. where fb is composed of the range frequency/. In addition to a reduction in receiver sensitivity. whereas in the conventional superheterodyne the LO frequency is of the same order of magnitude as the RF signal. 3. In Fig. one on . — fw and rejects the carrier > . 3./ / 92 Introduction to Radar Systems [Sec. Only the averaging frequency counter need be used in an altimeter application. the output of the receiver mixer is an IF signal of frequency IF +/„. the other feeds a switched frequency counter to determine the doppler velocity (assuming r f fd). is more sensitive and stable and is preferred wherever its slightly more complex construction can be accepted.14.= fr. 24 Further isolation may be obtained by proper adjustment of the phase and amplitude of a direct signal to cancel the leakages. A portion of the frequency-modulated transmitter signal is applied to a mixer along with the oscillator signal. the output of the low-frequency amplifier is divided into two channels: one feeds an average-frequency counter to determine range. although more complex than the homodyne (zero-IF) receiver. The filtered sideband serves the function of the local oscillator. The sideband filter must have sufficient bandwidth to pass the modulation. When an echo signal is present. The sideband that is passed by the filter is modulated in the same fashion as the transmitted signal. which is amplified to a level where it can actuate the frequency-measuring circuits. Block diagram of FM-CW radar using sideband superheterodyne receiver. A block diagram of an FM-CW radar with a sideband superheterodyne receiver is shown in Fig.

portion of the transmitter signal is applied to a mixer. The local oscillator may be a reflex klystron or some other This local oscillator is oscillator whose frequency can be controlled electronically. the gain of the low-frequency amplifier should be made to increase at the rate of 12 db/octave. along with a portion of the local-oscillator signal. always be some lag in the local-oscillator frequency since there must be a difference between it and the transmitter if an error is to be discerned. and its frequency measured. The echo range. 3. the local oscillator is made to stay in step with the changing transmitted However. matic frequency control quite similar to that of auto- a conventional superheterodyne receiver. In essence. as is signal from an isolated target varies inversely as the fourth power of the the radar equation. where the echo large echo signals. will signals are weaker. 3. in the . The mixer output is amplified and applied to a frequency discriminator which generates a d-c voltage proportional to the difference between the transmitted and the local-oscillator frequencies. well known from control (STC) employed in conventional pulse radar (Sec.2). Its frequency is the carrier frequency in (AFC) A ± The difference frequency from the mixer is equal to the IF. This lag is not important since the two IF signals are combined in the balanced detector and are subject to the same error. This is amplified and detected in the balanced mixer. The discriminator voltage is used to correct the local-oscillator frequency so is The local-oscillator signal it in synchronism with the transmitted frequency. Therefore the frequency characteristic of the low-frequency amplifier in the FM-CW radar may be shaped to provide attenuation at the low frequencies corresponding to short ranges and Less attenuation is applied to the higher frequencies. the RF bandwidth necessary for precise range measurement is discarded after its purpose has been served. of operation is following superheterodyne.Sec. However. Block diagram of FM-CW radar with signal-following superheterodyne receiver. 8. Receiver mi er IF amplifier Balanced detector Low-frequency a mplifi er Frequency counter(s) Fig. 3. as to vary ^ y c FM transmitter - Modulator 1 l Mixer IF amplifier IF amplifier Local oscillator D-c amplifier Frequency discriminator > 1 1 s I ! .14. thus permitting relatively narrow IF bandwidths. 3. In both the signal-following superheterodyne and the sideband superheterodyne. while one at long range will result in a weak signal at high frequency. applied to the receiver mixer to produce the IF signal.3] CW Its principle and Frequency-modulated Radar 93 receiver Another example of the superheterodyne principle applied to the FM-CW radar This is known as the signalis shown in the block diagram of Fig. for constant target Amplifier response shaping is similar in function to sensitivity time cross section.15. With this as a criterion. The output of the amplifier would then be independent of the range.15. more like that of a conventional superheterodyne receiver than the oscillator of the the IF sideband superheterodyne of Fig. there signal in order to provide the proper reference signal at the receiver. A target at short range will generally result in a strong signal at low frequency.

the response is also cut off at the lowfrequency end.2 0. therefore. 3.16.3). The low-frequency-amplifier bandwidth must be sufficiently wide to encompass the expected range of beat frequencies. 6 a y% 12 » 18 c o £ 30 35 42 0. Since the frequency excursion A/is inversely proportional to range.16. Lowered gain at low altitudes also helps to reduce interference from unwanted reflections. altimeter. A/is varied to maintain the beat frequency constant. The value of the frequency excursion is then a measure of the altitude and may be substituted into Eq. 23 - 25 When a fixed-frequency excursion (or deviation) is used. The response at the upper end of the frequency characteristic is rapidly reduced for frequencies beyond that corresponding to maximum range.8 1. the technique of servo-controlling the frequency . the signal-to-noise ratio is reduced and the receiver sensitivity degraded. thus reducing the amount of noise with which the signal must compete. the beat frequency FM excursion is usually applied at all altitudes above a predetermined minimum. the echo signal from an extended target such as the ground varies inversely as the square (rather than the fourth power) of the range. to further reduce the extraneous noise entering the receiver. For extended targets. 12. 3 1 7 overcomes this limitation. If there is a minimum target range. A similar servomechanism technique may be used to maintain the aircraft at a fixed preset distance from the ground. Another method of processing the range or height information from an altimeter so as to reduce the noise output from the receiver and improve the sensitivity uses a narrowbandwidth low-frequency amplifier with a feedback loop to maintain the beat frequency constant.5 0. kc Fig. The beat-frequency amplifier need only be wide enough to pass the received signal energy. since the greater the range. the greater the echo area illuminated by the beam (Sec.94 Introduction to Radar Systems [Sec.3 altimeter. 3. as in the usual can vary over a considerable range of values. Instead of maintaining the frequency excursion A/ constant and obtaining a varying beat frequency.0 2 3 5 8 10 20 30 Frequency. A com- promise between the isolated (12-db slope) and extended (6-db slope) target echoes might be a characteristic with a.3 0. the radar is better operated .18). 3. (3. slope of 9 db/octave. Since the bandwidth is broader than need be to pass the signal energy. Frequency-response characteristic of low-frequency amplifier of typical altimeter. the low-frequency amplifier gain should increase 6 db/octave. but the attenuation of the low frequencies effects a reduction of low-frequency interfering noise. When used in the altimeter. A typical frequency-response characteristic with a slope of approximately 8 db/octave is shown in Fig.1 0. The system shown in the block diagram of Fig. The constant output produced by shaping the doppler-amplifier frequency-response characteristic is not only helpful in lowering the dynamic range requirements of the frequency-measuring device. The frequency excursion is maintained by a servomechanism to that value which permits the beat frequency to fall within the passband of the narrow filter.

but are important if the altimeter is part of a blind landing system. Above 200 ft the output of the potentiometer connected to the servomotor determines the range. 4 cN A/ where range (altitude). or step error. The discreteness of the frequency measurement The average number of cycles gives rise to an error called thtfixed error. As has been mentioned. and accurate. The beat-frequency amplifier would have to be designed so that its bandwidth could be changed in discrete steps corresponding to the frequency excursion employed. denotes time average. In addition.17. Above 200 ft the frequency excursion is made to vary inversely with altitude so as to maintain a constant beat frequency of approximately 6. 3..18) may be rewritten as over/. FM altimeter with servo control of transmitter frequency excursion. m = velocity of propagation. 95 low more normal manner with a and hence a varying beat frequency. the residual path-length error caused by the circuits and transmission lines. cps R= c . The theoretical accuracy with which distance can be measured depends upon the bandwidth of the transmitted signal and the ratio of signal energy to noise energy. Measurement Errors. ' * FM transmitter Modulator Servomotor and potentiometer Servo omplifier Balanced detector Low-trequency amplifier Limiter (narrowband) Frequency counter Height indicator Fig. a common form of frequency-measuring device is the cycle counter. N . The AN/APN-22 radar altimeter is operated in the conventional manner for altitudes from to 200 ft.000 ft. Below 200 ft the output of the frequency counter is a measure of range. of the beat frequency fb in one period of the modulation cycle/m is/Jfm where the bar Equation (3. Another technique that could be used to narrow the bandwidth beat-frequency amplifier without the need for a servomechanism is to employ discrete rather than continuous frequency excursions. The absolute accuracy of radar altimeters is usually of more Errors of 1 or 20 ft might not be of importance at low altitudes than at high altitudes significance when cruising at altitudes of 30. The total cycle count is a discrete number since the counter is unable to measure fractions of a cycle. which measures the number of cycles or half cycles of the beat during the modulation period.000 cps. measurement accuracy might be limited by such practical restrictions as the accuracy of the frequency-measuring device. errors caused by multiple reflections and transmitter leakage. 3. m/sec A/ = frequency excursion. The total charge per second (current) is indicated on a milliammeter calibrated in range or altitude. stable.3] at very CW altitudes in the and Frequency-modulated Radar fixed A/. 20 21 26 The principle of operation of a frequency counter is based upon generating a fixed amount of charge for every cycle or half cycle of the unknown frequency. and the frequency error due to the turn-around of the frequency modulation.Sec. . The widely used cycle-counting type of frequency meter is simple.

the error would be 35 ft. If the frequency excursion were one-tenth this value 2 8/? 38/? 4 8/? 5 8/? Range (6) Fig. the count of positive zero crossings would be 5. Under these conditions (that is. Large frequency excursions are necessary if the fixed error is to be small. but if the phase were shifted tt radians. A/ a function of altitude). that is.3 Since the output of the frequency counter is an integer. will change by v radians \f2nf T AttR\X changes by n radians this corresponds to a change in range of one-quarter of a wavelength.4 cycles in duration over the modulation period. a quarter wavelength is small compared with the fixed range error . If the phase relative to the modulation period were like that shown in Fig. The dependence on the phase of the beat frequency may be illustrated by considering a counter which counts only those zero crossings with positive slope (1 count per cycle). a relatively large error at low altitudes. same as (a) but shifted in phase by -n radians. Assume that the beat-frequency signal is 5. (a) 5. the count would be 6 another shift of v radians would change the count back to 5. the count will depend on the particular configuration of the counter. The phase of the beat-frequency signal.96 Introduction to Radar Systems [Sec.5 ft. The count measured by a frequency counter depends upon the phase of the beat frequency with respect to the time interval over which the measurement is made. (3. 200 The frequency excursion of the AN/APN-22 altimeter is 70 Mc for altitudes less than ft. according to Eq. the range will be an integral multiple of r/(4 A/) and will give rise to a quantization error equal to N 6R or = -£4 a*(ft ) = ^L A/(Mc) A/ (3.286) Note that the fixed error is independent of the range and carrier frequency and is a function of the frequency excursion only.28a) (3. However. 3. In addition. (7 Mc). the fixed error is a function of the altitude. At radar frequencies. (c) variation of counts with range. 1 8a. Variation of beat-frequency cycle count with phase.18. 3. It will be recalled that above 200 ft the frequency excursion is made to vary inversely with altitude in the AN/APN-22 by a servomechanism control system. The fixed error is 3. = .22). the fixed error expressed as a percentage will be constant. whether it counts threshold crossings or zero crossings or whether it is full wave (2 counts per cycle) or half wave (1 count per cycle). . 3.4 cycles of beat frequency.

resulting in a fixed error of a few feet. Most altimeters are designed with excursions of the order of 100 Mc. The AN/APN-22 employs the technique of varying the modulation frequency at a 10-cps rate. the necessity of averaging over an interval of time may increase the scan rate and prove to be an unacceptable restriction. For example. The fixed error is not present in a noncounting (continuous) frequency meter such as a frequency discriminator. The total phase shift therefore varies with time and permits the fixed error to be averaged. This technique not only eliminates the fixed to N N+ +/ . In the altimeter this condition is usually satisfied. the inaccuracy caused by the fixed error is usually of little or no operational importance unless A/is small. However. There are several techniques that can be used to make the fixed error small. causing the phase shift of the beat signal to vary. If a full. the fixed error might not be averaged out and could prove troublesome. One technique. In other applications of FM-CW radar. the discriminator circuit is more difficult to operate as a switched frequency meter than is the counter. At high altitudes. where f x is a 27 frequency small compared with the modulation frequency The phase shift /m associated with the beat signal will be 2irf T [from Eq.19.wave counter were used. (3.18c. A different method of employing a frequency discriminator to obtain both the range and doppler velocity is shown in Fig. Over smooth terrain such as airport runways or calm water. The count also changes by 1 every time the range changes by dR. Wobbling the transmitter phase results in a wobbling of the phase of the beat signal. Therefore the fixed error will jump back and forth between N and N + 1 cycles every time the range changes by one-quarter of a wavelength. The uncertainty in the count N is illustrated in Fig. The discriminator output is a voltage proportional to frequency and is continuous rather than discrete. waves on the water. FM FM . the count increases or decreases by 1. already mentioned. The discriminator might be used in those altimeters mentioned previously where the frequency deviation is controlled by a servomechanism which maintains the beat frequency constant.3] CW and Frequency-modulated Radar 97 dR caused by the discrete measurement of frequency. 3. The restriction caused by the limited bandwidth of the frequency discriminator may be overcome by transposing the relatively low beat frequency to a higher frequency in an effort to reduce the required percentage bandwidth. discriminator circuits with sufficient stability and linearity do not seem to be capable of as wide a range of frequency operation as the frequency counter and in the past have not been popular in altimeter systems. or turbulent air average out the fixed error provided the time constant of the indicating device is large compared with the time between fluctuations. such as a scanning search radar. The effects of the fixed error also can be reduced by wobbling the modulation frequency or the phase of the transmitter output. Every time the phase of the echo signal changes by w radians (or njl radians for a full-wave counter). The above methods assume that the radar application permits sufficient time for the averaging to take place./ / Sec. is to make the frequency excursion A/ large. Normal fluctuations in aircraft altitude due uneven terrain. but it averages the fixed error in a manner similar to the averaging accomplished by changes in altitude. the range uncertainty would lie between 2SR and 4dR.22)] plus 2-nfy. Also. Another method which has been employed to average the fixed error is to vary the phase of the reference signal ref by transposing to a frequency ref t . 3. The units of range are in increments of fixed error dR = c/(4 A/). if the frequency meter reads a count of 3. The indicating system is designed so that it does not respond to the 10-cps modulation directly. An average reading between and 1 will be obtained when displayed on a normal meter movement. the count would jump back and forth every eighth wavelength. 3. if there is reason to do so. when both range and doppler information are required.

3.19. Therefore the doppler velocity may be measured by averaging the discriminator output. 20 29 In this system (Fig. 3. Another technique for eliminating the fixed error is known as the double-modulated radar. 3.29) IF. 4> T) + <f> </>] = k sin O (3. The double system eliminates the fixed error and permits a smaller frequency deviation to be used than in the usual system. the signal leaving the IF amplifier gives Ja If this frequency O to obtain the frequency of ~2ndt~ Jw±Jd 27rA#m K •sin(27r/m t- <f> T) (3. Basically. The doppler velocity and the range must not vary appreciably over the averaging period if accurate measurements are to be obtained. ml ). the system up to the input of the limiter is quite similar to the sideband superheterodyne receiver discussed previously (Fig. 3./ 98 error. Range may be determined by extracting the a-c component in a narrowband filter centered about the frequency m and calibrating the output voltage f from this filter directly in range. 28 actually employs a triple-conversion receiver rather than the single-conversion receiver shown. The other is the a-c component at a frequency ml whose amplitude is proportional to target f range. and the other FM transmitter Modulator Mixer * Local Sideband filter Direct-current output proportional to Doppler velocity »" Mixer IF amplifier Limiter Frequency discriminator Selective amplifier at frequency-^ Amplitude of output at -frequency f m corresponds to range Fig. and an IF signal is extracted whose frequency is some multiple (including unity) of/m2 ± the doppler frequency fd Therefore. FM FM .22). The received signal is mixed with the reference signal. = k sin [2-rrf lv (t is + T) ± 2nf dt + nAfTcos (2nfm t .fd the doppler frequency.20) the transmitted signal is modulated at two FM ' and/m2 The modulating frequency fml is of low frequency and corresponds to the modulating frequency (/J in the usual system. just as in the system described in Fig. if this signal is amplified.19. FM >/ . and hence to target relative velocity. T <j> where/IF the frequency of the represents the various = = 2R/c. parameters are as denned in Eq.14). irf T m T.3 but it does not confuse the range and doppler velocity of fast. One component is d-c. while fmi might be a few kilocycles. Block diagram of FM-CW technique for eliminating the fixed Differentiating error. However. Introduction to Radar Systems [Sec. which is proportional to the doppler frequency shift (±fd).30) is applied to a frequency discriminator centered at the intermediate frequency IF . it is more complicated frequencies/*. the output voltage will consist of a steady component corresponding to the doppler frequency f d and an a-c component of frequency m with amplitude prof portional to range. 3. its output will contain two components.! . constant phase shifts introduced in the system. limited. (3. The system as described in Ref. The voltage from the IF amplifier preceding the limiter is vn. nearby targets 28 (when/i r) as does the combination of average and switched frequency counters. and applied to a frequency discriminator. while/m2 is a relatively high frequency (fm2 The frequency fml might be of the order of 100 cycles. There is no loss in generality in considering >f the simpler system.

3. The reflection of the transmitted signals at the antenna caused by impedance mismatch.3] CW and Frequency-modulated Radar 99 than the system of Fig. At short ranges the residual path error can also result in a The residual path error is the error caused significant error unless compensated for. It can be reduced by an attenuator introduced in the transmission line at low altitude or by a directional coupler or an isolator. 3.20 whose frequency excursions are only a fraction of that of the usual Modulator f m1 ± FM transmitter . between transmitter and receiver antennas. while the unwanted signals are shown by the broken arrows. 3. and the number of independent observations. Multipath signals also produce error. especially at high 4. the more accurate will be the range measurement. radar if there are uncontrolled variations Other errors might be introduced in the modulation frequency. the signal-to-noise ratio. In Chap. The wider the transmitted spectrum and the greater the signal-to-noise ratio. Figure 3. the factors which affect the accuracy of radar measurements are discussed. FM radar system must take more time or make more is observations or obtain higher signal-to-noise ratios if comparable accuracy to be achieved. . Target motion can cause an error in range equal to v r T where vr is the relative velocity and T is the observation time. This can limit the ultimate receiver sensitivity. Block diagram of double-modulated FM radar. 3. or frequency excursion. altitudes. Those systems such as described by FM Figs. Before leaving the subject of fixed error. 5. 22 The wanted signal is shown by the solid line.Sec. and it is mentioned that the accuracy with which range can -be measured is a function of the transmitted spectral FM width. The unwanted signals include 1. . it may be worthwhile to mention briefly the relation between transmitted bandwidth and accuracy.»fm Modulator f mZ Bandposs Mixer 90° phase shift H- filter Direct-current output proportional to doppler velocity 90° phase shift IF amplifier Frequency discriminator Amplifier Amplitude of output at frequency f to and limiter corresponds range m Bandposs -»| Mixer | «- filter Fig. 10.19 and 3. The double-bounce signal. Multipath reflections (reflections from unwanted targets) can also introduce errors 30 They can be reduced with into the FM-CW radar system and must be avoided. The standing-wave pattern on the cable feeding the reference signal to the receiver. causing a change in the impedance seen by the transmitter. by delays in the circuitry and transmission lines. CW 2.20. The leakage signal entering the receiver via coupling due to poor mixer match. This is usually important only at low altitudes.21 shows some of the unwanted signals that might occur in the FM altimeter. The interference due to power being reflected back to the transmitter. 3.19 and is more limited in both maximum range and minimum range than either the previous system or the usual system. in the transmitter frequency.

is frequency-modulated by a sine wave.fa) cos 3(2nfm . 3.3 highly directive antennas and in ground-based radars by lowering the height of the antenna to reduce the path difference between the direct and the reflected rays.<f>J .100 Introduction to Radar Systems [Sec.21.m + 2/ (£>) (3.<f>J . 22 IRE Trans.31) where f = m f = A/= carrier frequency modulation frequency frequency excursion (equal to twice the frequency deviation) may be written The difference frequency signal »d = UD) cos (Infy . 31 The received signal is not affected by a double-sideband cancellation network since a doppler-frequencyshifted signal is equivalent to a single-sideband modulation. 31 depending on the method employed for adjusting and maintaining the phase and amplitude of the cancellation signal to the correct values. the noise is usually of sufficient magnitude compared with the echo signal to require some means of minimizing the leakage that finds it way into the receiver.fa) cos 2(27rfm .Q) + 2J {D) sin (2irfd . Echoes from short-range targets—including the leakage signalmay be attenuated relative to the desired target echo from longer ranges by properly processing the difference-frequency signal obtained by heterodyning the transmitted ind CW AM and received If the signals. Unwanted signals in FM altimeter.<£J + 2/4 (Z>) cos {2irfa . AM CW The techniques described previously for reducTransmitter Receiver © t© f ing leakage in the radar apply equally well to the radar.32) 4> 1 t <f> t 2 t t t t t t <t> ) • • • 5 . sin \2irf t + ^f sin 2nfa •) 2f„. Although advances have been made in reducing the and FM noise generated by high-power transmitters. These remedies only relieve rather than eliminate completely the problem of multipath. (3. The ability of the FM-CW radar to measure range provides an additional basis for obtaining isolation.) The double-sideband noise components of the transmitter may be further canceled in the radar receiver with a simple cancellation network. Transmitter Leakage. the difference frequency obtained by heterodyning the returned signal with a portion of the transmitter signal may be expanded in a trigonometric series whose terms are the harmonics of the modulating frequency fm 8 20 Assume the form of the transmitted signal to be ' CW carrier .2J3(D) sin (27rfd . Fig.fa) cos 4 (2nfm . Separate antennas CW FM-CW and direct cancellation of the leakage signal are two techniques which give considerable isolation. 3. The sensitivity of FM-CW radar is limited by the noise accompanying the transmitter signal which leaks into the receiver. To obtain a cancellation of as much as 60 db (residual voltage one-thousandth the original) requires a closedloop servo system to automatically correct for changes in the leakage signal produced by antenna scanning and the like.) cos (2nfm .2J (D) cos (2nfd . (From Capelli. The degree of isolation that can be obtained by cancellation of the leakage signal might vary from 10 db 9 to 60 db.

the frequency excursion A/can be adjusted to obtain that value of D which places the maximum of the Bessel function at the target range. The J amplitude applies maximum response to signals at zero range in a radar that extracts the d-c doppler-frequency component. 3. Therefore. (3. the amplitude of the leakage signal at zero range may theoretically be made equal to zero.22. second. At greater ranges. etc. the higher will be the order of the Bessel function and the less will be the amount of microphonism-leakage feed through. it would enhance the leakage signal and reduce the target signal. and Frequency-modulated Radar and order 0.22). . 3.Se c 3. Although higher-order Bessel functions may reduce the zero-range response. where the target is expected. 1 .. or third harmonic).) In principle. fm 2'm Frequency 34. When only a single target is involved. dt This is a cosine wave at the doppler frequency with an amplitude proportional to J (D).3] . Each of these harmonics of fm is modulated by a doppler-frequency component with amplitude proportional to J„(D). respectively (AfIf J sin 2nfm R /c t distance to target at time = (distance that would have been <f> <f> The difference-frequency signal of Eq. they may also reduce the response at the desired target range if the target happens to fall at or near a range corresponding to a zero of the Bessel function. 3. The higher the number of the harmonic. Figure 3. a condition opposite to that desired. The product of the doppler-frequency factor times the nth harmonic factor is equivalent to a suppressed-carrier double-sideband modula. 4/S» Fig. (After J J Saunders* IRE Trans. 101 where J Jx J2 = Bessel functions of first kind D= R = c 2.23 shows a plot of several of the Bessel functions. any of the /„ components of the difference-frequency signal can be extracted in the radar.23) shows that if one of the modulationfrequency harmonics is extracted (such as the first. CW etc. This results from the property that Jn (x) behaves as x n for small x.32) consists of a doppler-frequency component of amplitude J (D) and a series of cosine waves of frequency fm 2fm 3fm etc. . Consider first the d-c term J (D) cos (2nf <£ ). . . if the J term were used. . the effect of the J Bessel function is to reduce the echo-signal amplitude in comparison with the echo at zero range (in addition to the normal range attenuation). The argument of the Bessel function is proportional to range. An examination of the Bessel functions (Fig. Spectrum of the difference-frequency signal obtained from an FM-CW radar sinusoidally modulated at a frequency/™ when the target motion produces a doppler frequency shift L. = fa = 2v f /c = doppler frequency shift v = relative velocity of target with respect to radar = phase shift approximately equal to angular distance 4Trf R jc m = phase shift approximately equal to 2-nfm R^c r r measured if target were stationary) velocity of propagation tion (Fig. FM-CW - D This is the range at which the leakage signal and its noise components (including microphony and vibration) are found.

The doppler-shifted echo is heterodyned with the transmitted signal to produce the beat-frequency signal of Eq. 2.102 Introduction to Radar Systems [Sec.31). (3. One of the harmonics of/m is selected The filter bandwidth is wide (in this case the third) by a filter centered at the harmonic. (3. the CW FM radar because the operate with the nth harmonic as compared with a practical The loss in transmitter leakage noise is suppressed by the nth-order Bessel function. 1 . radar using the third harmonic (/3 term) is shown in Fig. enough to pass both doppler-frequency sidebands.3 The technique of using higher-order Bessel functions has been applied to the type of block diagram of a doppler-navigation radar discussed in the next section. Eq. 3. ^> ^\ Circulator Directional coupler Transmitter f m Frequency modulation x3 frequency multiplier Mixer 3d harmonic filter Mixer Low-pass filter Doppler frequency Fig. 3. 3.23.24. Sinusoidally ponent). D= (A///J sin 2irfm Rt. 839 to be from 4 is reported component with the J Bessel operating signal energy when 3 CW .32).lc.24. (third) harmonic of/OT . modulated FM-CW radar extracting the third harmonic {Jz Bessel com- Since the total energy contained in the beat-frequency signal is distributed among all the harmonics.4 Fig. The transmitter is sinusoidally frequency-modulated at a frequency fm to generate the waveform given by A CW -J {D) 12 D -0. The filter output is mixed with the The doppler frequency is extracted by the low-pass filter. Plot of Bessel functions of order 0. extracting but one component wastes signal energy contained in the radar. other harmonics and results in a loss of signal as compared with an ideal to radar designed the superior in generally is signal-to-noise ratio However. 3. and 3 .

because of the periodicity of D. The frequency excursion of the modulation waveform can be adjusted by a servomechanism to maintain the maximum of the Bessel function at the aircraft's altitude. Doppler radar can navigation system is based on the the aircraft relative to the earth. where/ is the carrier frequency. including zero. the extraction of the higher-order modulation frequencies is not practical with a stationary target. the amplitudes of the modulationwhere or Jn (D) cos frequency harmonics are proportional to Jn {D) sin Att^RqIc AttRqIL Therefore the amplitude depends on the range to the target in RF wavelengths. The frequency translator is not needed in an airborne doppler navigator since the antenna beam is directed at a depression angle other than 90° and a doppler-shifted echo is produced by the motion of the aircraft. and c is A . An aircraft with a doppler radar whose antenna beam is directed at an angle y to the horizontal (Fig. doppler shift in the antenna path. Plot of / 3 (Z>) as a function of distance. 3. The nulls in the curve suggest that echoes from certain ranges can be suppressed if the modulation parameters are properly selected. This might be accomplished with a single-sideband generator (frequency translator) inserted between the directional coupler and the RF mixer of The frequency translation in the reference signal path is equivalent to a Fig. it single antenna with a circulator is shown in not necessary in many applications. the aircraft's present position CW can be computed by dead reckoning. Typically. altimeter. 3. shift in frequency is/ d = velocity. The curve is mirrored plot of J3 (D) as a function of distance is shown in Fig. doppler-radar principle. In order to use the properties of the Bessel function to obtain isolation in an FM-CW A A </> <f> = = <f> . Leakage introduced by the circulator and by reflections from the antenna are at close range and thus are attenuated by the Jz factor. {From Saunders* IRE Trans. The sine or the cosine terms can take any value between 1 and 1. such as in an altimeter. 3.24. one position to another without the need of navigation information transmitted to the One method of obtaining a self-contained aircraft aircraft from a ground station.4. is CW and Frequency-modulated Radar 103 Although two separate transmitting and receiving antennas may be used. the depression angle y might single antenna beam from a doppler radar measures be in the vicinity of 65 to 70°. the role of the doppler freintroduced by translating the reference frequency to some different value. + — Distance Fig. for a change in range corresponding to one RF wavelength.4] to 10 db. For this reason. The drift speed of drift angle the and true provide angle is the angle between the horizontal projection of the centerline of the aircraft (heading) and the horizontal component of the aircraft velocity vector (ground track). when the doppler frequency shift may be artificially 3. The (fjc) v cos y. 3.Sec. 3. If the target is stationary (zero doppler frequency).25. From the ground-speed and drift-angle measurements. the block diagram of Fig. 3.26a) will receive a doppler-shifted echo signal from the ground.) quency is essentially zero. Airborne Doppler Navigation 3240 An important requirement of aircraft flight is for a self-contained navigation system capable of operating anywhere over the surface of the earth under any conditions of It should provide the necessary data for piloting the aircraft from visibility or weather.25. v is the aircraft the velocity of propagation.24.

Fig. The angular displacement of the antenna from the aircraft . as might be obtained from a compass. the doppler frequency in the two forward beams will not be the same. cally disposed Roman god who looked both forward and backward at the same the two forward aircraft.104 Introduction to Radar Systems [Sec. initially that and two backward beams are symmetriIf the aircraft's velocity vector is about the axis of the same direction as the aircraft heading. To convert this vector velocity to a horizontal reference on the ground. 3. 3. the speed and direction of travel. that is. or alternatively.266. the antennas might be fixed relative to the aircraft and the ground velocity components calculated with a computer. A minimum of three noncoplanar beams are needed to determine the vector velocity. Doppler-navigation radar measures the vector velocity relative to the frame of reference of the antenna assembly.4 one component of aircraft velocity relative to the direction of propagation. The heading of the aircraft. (0) aircraft employing four doppler-navigation beams to obtain vector velocity. 3. (a) Aircraft also be known for proper navigation.26 with single doppler-navigation antenna beam at an angle y to the horizontal. This difference in frequency may be used to generate an error signal in a servomechanism which rotates the antennas until the doppler frethe not in quencies are equal. The vertical reference may be used either to stabilize the antenna beam system so as to align it with the horizontal. must A practical form of doppler-navigation radar might have four beams oriented as in A doppler-navigation radar with forward and rearward beams is called a Assume Janus system. indicating that the axis of the antennas is aligned with the ground track of the aircraft. the direction of the vertical must be determined by some auxiliary means. A=*- (/>) Fig. after the time.

However. 4. The two channels may be operated simultaneously or timed-shared." These are caused by the high prf commonly used with pulse-doppler radars when it is necessary to achieve unambiguous doppler measurements. Leakage between transmitter and receiver limits the sensitivity of the doppler-navigation radar. The use of the two rearward beams in conjunction with the two forward beams results in considerable vertical motion of the improvement in accuracy. it will not be detected. means CW CW W CW CW CW A radar. and the other aft and to the left. but not without some inconvenience or possible loss in over-all performance. radar would seem to be the ideal method of obtaining dopplerIn principle. and the magnitude of the doppler is a measure of the speed along the ground track. altitude holes exist at or near those altitudes where the echo time is an integral multiple of the pulse-repetition period. whereas a radar must usually employ two separate antennas in order to achieve the needed isolation. 3. However. Two beams give the two components of the aircraft velocity tangent to the surface of the earth. It eliminates the error introduced by aircraft and reduces the error caused by pitching movements of the antenna. Coherence is obtained on a relative basis in the process of comparing the signals received from the forward and backward direction. A third component. the leakage signal may be reduced relative to the signal from the ground by A CW . The Janus system can be operated incoherently by using the same transmitter to feed a pair of beams simultaneously. one beam is directed ahead and to the right of the ground track. pulse systems suffer from loss of coverage and/or sensitivity because of "altitude holes. By frequency-modulating the transis with a frequency-modulated mission. A forward-left and an aft-right are also fed by the transmitter as a second channel. Another method of achieving the necessary isolation in a doppler-navigation radar system. However. transmitted signal might be obtained with a low-power oscillator followed by an amplifier that is pulsed on and off at the In essence. Navigation may also be performed with only two antenna beams if some auxiliary is used to obtain a third coordinate. the navigation information. usually results in ambiguous range. The high prf. in practice. the doppler frequency is extracted. Typically. frequency resulting from the mixing operation is twice the doppler frequency. But more important. is needed and may be provided from some nondoppler source such as a barometric rate-of-climb meter. By heterodyning in a mixing element the echo signal The difference received in the fore and aft beams. a pulse-doppler radar may be considered as a "sampled CW" desired rate. The pulse-doppler mode of operation has the further advantage in that each beam can operate with a single antenna for both transmitter and receiver. Techniques exist for reducing the undesired effects of altitude holes.5.Sec. Thus. a pulse radar with ambiguous prf can result in If the transmitter is pulsed just when the ground echo arrives back at the lost targets. the radar is not adequate at long ranges. One method of eliminating the ill effects of leakage is by pulsing the transmitter on and turning the receiver off for the duration of the transmitted pulse in a manner similar to the pulse-doppler radar described in Sec. j ust as it does in any C radar. although it gives unambiguous doppler. The primary advantage of the two-beam system is a reduction in equipment. Changes in transmitter frequency affect the echo signals in the two directions equally and are therefore canceled when taking the difference frequency.4] CW and Frequency-modulated Radar 105 heading is the drift angle. the vertical velocity. stable transmitter frequency is not needed in this system as it is in the coherent system. the accuracy is not as good as with systems using three or four beams. radar. 33 The pulse-doppler system must be coherent from pulse to pulse if the doppler frequency shift is to be correctly measured.

= The region of unambiguous range may be extended considerably by transmitting two CW signals differing only slightly in frequency.) The sine-wave CW and returns to the radar 2R/c. However. 9 It will be shown that the measurement of range using two CW frequencies results in an unambiguous range separate which corresponds to a half wavelength at the difference frequency. This results in an error in the ground speed. are easier to conceive and understand by analysis in the time domain. thus lowering the signal-to-noise ratio. frequencies. where n is an integer.000 ft. Consequently.2).5 extracting a harmonic of the modulating frequency and taking advantage of one of the higher-order Bessel functions as described in the previous section. At radar frequencies this unambiguous range is much too small to be of practical interest. Another source of error is the mass movement of water caused by tides. which results in a doppler frequency 3. 3. 5 943 to be deis a direct application of - the principle that a range difference measurement. shift in addition to that caused by the aircraft's motion.33) gives the maximum unambiguous range as equal to A/2.27.5 per cent and drift angles to 0. The phase be used as a measure of the = range. The radar measurement of time delay or range is fundamentally the measurement of when time and direction remain constant (Sec. the . (The amplitudes of the signals are all taken to be unity since they do not influence the result. currents.000 ft the doppler navigator can reliably measure distances over land to an accuracy of at least 0.5°. If the transmitted and received signals are compared in a phase detector. (3. or winds. 38 At altitudes of 40. Multiple-frequency CW Radar 942 47 the variation of phase with frequency. Transmitted (a) and received (b) signal spectra in the 4t7 two-frequency CW radar.5. although it may be quite adequate for certain types of position-finding equipment at relatively low frequencies. It has been claimed that a doppler-navigation system based on this principle can provide 1 50 db of isolation. c Ac/> _ X -A^ (3. Specular reflection reduces the backscattered energy. One source of error overSvater is due to an increase in specular reflection of the incident beam. where c is the velocity of propagation. 3. single stationary target measurement is a phase- Frequency (a) Consider the problem of measuring the range R of a by using a radar radiating a single-frequency sine wave of the form sin 2nf t. The pulse radar and the FM-CW radar can be analyzed in these terms. but they 10. The measurement of the phase difference A</> is unambiguous only if A<£ does not exceed 2-n radians. the measurement of KA/-H range with multiple CW scribed in this section. the output is proportional to the phase difference between the two signal travels to the target after a time T= £lFrequency it) signals and is A</> = 27rf difference may therefore „ __ T 4Trf R/c. 38 Over water the accuracies are slightly worse. (A phase of 2t7ii + A<f> radians. the amount necessary to operate at altitudes of 50.33) 4Trf Fig. It also causes an apparent increase in the angle of depression by favoring the returns from the lower half of the incident beam. cannot be distinguished from a phase of A</> radians.106 Introduction To Radar Systems [Sec.) Substituting A<f> 2n into Eq.

fd2 doppler frequency shift associated with frequency 2 Since the two RF frequencies/ and/2 are approximately the same (that is. c (336b) The phase difference between these two components 2 ^ = Mf c fl )R Hence which is is r = the lM4-n-A/ (3 . 3.16) and may be written v 1R = = sin 2HA ±/„i)t v 9J? sin Mh ±/«)' ^^ T . (3.5] CW is and Frequency-modulated Radar that obtained 107 a unambiguous range can be made considerably greater than single frequency when only transmitted./ Sec.35b) where range to target at a particular time t t (range that would be measured if target were not moving) fdl doppler frequency shift associated with frequency/]. the amplitudes of all signals are set equal to unity. (3.36a) v 2D sin (±2tt// ^p> T 2-rrf^ = Ajr^fR.^2 T c 2TTfdl t +& + <f> (3. If the target carries a beacon or some other form of echo-signal augmentor.35a) 2nfd2 t 2 (3./2 A/ where A/ </) the doppler frequency shifts f dl and/ 2 are approximately equal to one another. 3. For convenience. the method may still be applied by measuring the phase difference between the two RF carrier signals. The two-frequency doppler frequency shift. The form of the doppler-shifted signals at each of the two frequencies/! and/ is similar to Eq. as with a stationary target. 3. A block diagram of a two-frequency CW radar is shown in Fig. 3.28. CW technique . is two continuous sine waves of by an amount A/. The voltage waveforms of the two components of the transmitted signal v 1T and v 2T may be written as to consist of The transmitted waveform assumed frequency/ and/2 (Fig. Therefore we may write fdl d2 d R = = = = =/ + signal and heterodynes each component with the corresponding transmitted waveform and extracts the two doppler.33).38) same as that of Eq.34Z>) where <^ and 2 are arbitrary (constant) phase angles. with A/substituted in place of/.27a) separated v 1T v 2T <f> = = sin (2-nfit sin (Infy + + fc) <£i) (3. The echo signal is shifted in frequency by the doppler effect (Fig.frequency components given below: The =f =f - receiver separates the two components of the echo received signal v 1B = = sin (±27rfa t - ^^ T is 2*fd t^ (3.34a) (3. the doppler frequency shift may be simulated by translating the echo frequency. as with a single-sideband modulator.27b). of the simple phase-measuring device. like that The equipment CW radar except for the addition of the second channel and a When for measuring range was described as using the the doppler frequency is zero.

if f A//2. 1 80 cps if the difference frequency is to fall outside the passband of the doppler filters. and unless the transmitted frequencies and their > harmonics are rejected by the insertion of narrow bandstop niters. (3. becomes ambiguous. fd = nautical miles. Eq. that is. If one echo is much stronger than the other.fd f is less < than one-half the difference between the two transA//2.i> is R unamb — 2 A/ (3. consider a two-frequency radar with/i = 3. A large difference in frequency between the two transmitted signals improves the accuracy of the range measurement since large A/means a proportionately large change Transmitter A^ Receiver Range indicator Fig. there is a limit to the value of A/. it is usually desirable to make the difference frequency greater than the expected range of doppler frequencies. If more than one target is present. 3. This phase difference may be used as a measure of the elapsed time.5 If the doppler frequency d mitted frequencies. On the other hand. and hence range. they may swamp the doppler signals from the target. When the two signals slip in phase by 1 cycle. Block diagram of the two-frequency CW radar for the measurement of range. since A<£ cannot be if the range is to remain unambiguous. each transmitted signal lies within the doppler-frequency d acceptance band of the other receiver. qualitative explanation of the operation of the two-frequency radar may be had by considering both carrier frequencies to be in phase at zero range.28.000 Mc {X = 10 cm). Since the insertion of rejection niters complicates the receiver and eliminates a portion of the doppler-frequency band. the system might be designed to measure the range A CW of this target and ignore the others. the relative phase between the two increases because of their difference in frequency. in A<f> for a given range. greater than 2n radians However. Therefore . the maximum expected doppler frequency Therefore the difference between the two frequencies must be greater than 6. 600 knots. 9 In addition to discrimination on the basis of . the two signals may be readily separated.39) A/ must be less than c/2/? unamb This relationship is plotted in Fig.39) gives the maximum unambiguous range of a pulse radar. 3. As they progress outward from the radar. 3. The two-frequency radar is essentially a single-target radar since only one phase difference can be measured at a time.29.180 cps. the echo signal becomes complicated and the meaning of the phase measurement is doubtful.R„ nail . Note that when A/ is replaced by the pulse repetition rate. The maximum unambiguous range in this case is approximately 13 If the is maximum target velocity is 6. the measurement of phase.108 Introduction to Radar Systems [Sec. The maximum unambiguous range . As an example.

000 <J — Nw ^v \. 3. a fourth frequency can be /j measurement. three or . if further accuracy is required.5] CW and Frequency-modulated Radar 109 amplitude.40) E= energy contained in received signal N = noise power per cycle of bandwidth If this is compared with the rms range error theoretically possible with the linear pulse-compression waveform whose spectrum occupies the same bandwidth A/. multiple targets can be discriminated on the basis of doppler. (3. or else a single tunable narrowband doppler filter may be time-shared over the entire band of doppler frequencies.Sec.5. The theoretical accuracy with which range can be measured with the two-frequency radar can be found from the methods described in Sec.000 3 7. Substituting the unambiguous range of Eq. 1 . 6R = °unam6 2 77(2£/A/ i (3. 3.000 s.29. O (/I — \^ — o 1. series of narrowband doppler niters can separate signals. where k is a factor of the order of or 20./i gives an ambiguous but accurate range measurement while the pair of frequencies/a.41) ) Both accurate and unambiguous range measurements can be made by transmitting more frequencies instead ofjust two. The selection of A/represents a compromise between the requirements of accuracy and ambiguity.000 4.39) into Eq.29. nautical miles Maximum unambiguous range vs.000 £ Q. FM CW 1! U. — _ — N. difference frequency in the two-frequency CW radar. 3. \ ^v _ — - Q. It can be shown that the theoretical rms range error is A CW SR where = 4nAf(2EIN f (3. the pair of frequencies/3. <x> 300 - c a> 200 100 10 ii 20 ii 40 60 80 100 II 200 1 1 1 1 l\ 1 400 600 1. if the three frequencies/!. . o 3 a- 700 _ — 500 400 ^V N./! are chosen close enough to resolve the ambiguities in the/3 Likewise. range. the error obtained with the two-frequency waveform is less by the factor 0. 1 0.000 Maximum unambiguous Fig. /2 and/3 are such that/3 — ft = k(f2 — /i).W>J _ II 1 1 1 1 1 II 1 1 1 1 M L ~ T3 o o 5.40) indicates that the greater the separation A/ between the two freHowever.40) gives the rms error as quencies.000 2. Equation (3. the less will be the rms error. For example. the frequency difference must not be too large if unambiguous measurements are to be made. (3.

modulated frequencies separated from the carrier by 10. or when A/= 1/T = c/2#unamb which is the same as derived phase. The phases of the returned signals at the master Since the master and the unit are compared with the phases of the outgoing signals. the range (or the angular measurement in the interferometer) will not correspond to either target.000 Mc.900. the measurement is to be unambiguous. The measurement of range by measuring the phase difference between separated frequencies is analogous to the measurement of angle by measuring the phase difference between widely spaced antennas. the spectrum and target resolution approach that obtained with a pulse or an . Although the multiple-frequency distance-measuring technique was described in terms of a C transmission. in the multiple-frequency closely spaced array or a reflector antenna. and 9. as might be used for satellite tracking or space surveillance. as in an interferometer antenna. Two signals of fail to make the proper measurements. the multiple-frequency rr A/T. range-measurement technique. radar technique has been applied to the accurate The multiple-frequency measurement of distance in surveying and in missile guidance. remote units are stationary. The function of the doppler frequency is provided by modulating the retransmitted signals at the remote unit in such a manner that a 1-kc beat frequency is obtained from the heterodyning . the entire aperture must be filled. and T is the will produce a waveform of the form (sin JVz)/sin z. there is no doppler frequency shift. The entire spectrum must be continuous if By analogy with the antenna targets are to be both resolved and unambiguous. It is capable of measuring line-of-site distances from 500 ft to 40 miles to within an accuracy of 1 part in 300. frequencies spaced A/ apart radar containing problem./2 .990. = previously. the accuracy is determined by the difference between Additional frequencies are added in between if the largest and the smallest frequency. The spacing between the individual antennas in the interferometer system corresponds to the separation between The minitrack frequencies in the multiple-frequency distance-measuring technique.000. The Tellurometer The consists of a master unit at one end of the line and a remote unit at the other end. 41 Both the interferometer antenna and the multiple-frequency CW radar are singleIf two echo signals are alike in all respects except for a difference in target devices. The accuracy of the interferometer antenna depends on the distance between the two If resolution is to be obtained. with four single-sideband W CW The 10-Mc ties.5 transmitted and its ambiguities resolved by the less accurate but unambiguous measurement obtained from the three frequencies/i. 9. it can be applied to the improvement of the range measurement with a long pulse (pulse compression). The interferometer antenna gives an accurate but ambiguous measurement of angle./3 As more frequencies are added. In general. system is an example of an interferometer in which angular ambiguities are resolved in a manner similar to that described. while the difference frequencies of 1 Mc. 9. and 10 kc permit the resolution of ambigui- The remote unit at the other end of the line receives the signals from the master unit and amplifies and retransmits them. where z The measurement is ambiguous when the transit time to the range R and back.000 of the distance ±2 in. as with a elements. Similarly. master unit transmits a carrier frequency of 3.110 Introduction to Radar Systems [Sec. appear the same as one signal whose phase is that of the vector sum of the two signals. 100 kc. FM-CW waveform. denominator sin z = 0. difference frequency provides the basic accuracy measurement. these systems will different phase will CW N . 3. The Tellurometer is the name given to a portable electronic surveying instrument which is based on this prin42-44 ciple.000 Mc. The ambiguities may be resolved by additional antennas spaced closer together.

F. Aural. The radar method of surveying permits long distances to be measured conveniently and accurately.. pp. Phys. 2. The AN/APN-22 Radio Altimeter. vol. IRE Trans. vol. March. no. K. vol. : 1956." MIT Radiation Laboratory Series. pp. 60. pp. 1946. Inc. Inc. vol. 1949. Court. pp. October. Heaton: Improved Radio Altimeter. : 25. 539-543. 138-140. pp. 21 Study of the Double Modulated F. December. Collette.M. pp. Dropkin: Nonquantized Frequency-modulated AltimANE-1. W.: Doppler Radar. 1955.. W. G. Radar. M.. October. vol. 9. 13. Greene. G... Hockfrequenztech. vol.. 32. B. 15. A. : McGraw-Hill Book Company. November-December. no. 21. . ANE-8. 888-893. 33. Wireless Engr. 6. 11. pp. New York. J. 12. P. 27. 1959. Kerr... vol. How Accurate Are Radar Speed Meters?. 10. 1954. Ann. Zurich. 130-133. 1955. vol. 48-49. Military Automation. pp. Inst.. K. Luck. 387-398. pp. 21. June. 330. November. 2. ANE-1. vol. E. J. Cacheris. Proximity Fuzes for Artillery. G. This provides isolation between transmitter and receiver and aids in the suppression of ground reflections which can cause errors in the measurement. vol. B. 1946. H. Buecks.: Microwave Vehicle-speed Indicator.. vol. 698-700. 6. vol. 2. New York. 16. Electronics. Electronics. no. vol. an der E. 19. Middleton: Theoretical Comparison of the Visual. Ridenour. February. and H.T. pp. 1955. W. 1955.. 17. Q. 1957. Elec. The phase of the 1-kc signals contains the same information as the phase of the multiple frequencies. no. IRE. 40^15. and J. 1959. 32. pp. March. vol. pp. Rate-of-climb Meter Uses Doppler Radar. 8-14. December.: Direction Sensitive Doppler Device. Gardner. (ed.): "Propagation of Short Radio Waves. E. Saunders. 104-109." MIT Radiation Laboratory Series. pp. Electronic Ind. C.:The Radio Proximity Fuse. Lyons: Receivers with Zero Intermediate Frequency. 19. 96-99. Meyer. IRE. L. pp. February. Labrousse: Un altimetre radioelectrique a modulation de frequence. G. H... pp. June. ANE-1.: Mar. Jr. : W. Selridge. Electronics. 1957. vol. IRE Trans. MTT-3. it can operate by day or night and can measure distances through underbrush and even small trees. Post-war Developments in Continuous-wave and Frequency-modulated Radar. especially over inaccessible terrain. 1954. 1954. vol. IRE Trans. Lane. 2. Proc. 30. April. 18. A 3. IRE Trans. J. S. : M. 1 8.. 335-336. 1946. and Meter Reception of Pulsed Signals in the Presence of Noise. radioHectricite. H. E. vol. Barlow. pp. Electronics.H. January. 1. 1. 85-87. C. Block. /. Saunders. no. 47.. 8. Unlike conventional optical surveying instruments. Proc. J. D. 43. Each unit weighs less than 30 lb. p. J. Goldberg: Applications of the Turnstile Junction. May 22. McGraw-Hill Book Company. F.: "Radar System Engineering. IRE. 17. IRE Trans. vol. H. 1961. vol. 24. 37. vol. 19. D. 1947. 12. no. vol. Inc. : 7. 10. G. A. 1954. AP-4. ANE-1. 1956. June. Appl.: Radio Altimeter. 46-50. Proc. vol. 7-19. and A. Eng. September. Kalmus.: Digital-counter Techniques Increase Doppler Uses. Bonner. Each Tellurometer unit radiates about 100 of power. pp. pp. Jr. Van Vleck. P. mw REFERENCES 1. 340-355. eter. and A. March. pp. New York. Roberts. F. 1947. 132-134. Wimberly. 10. 1955. no. 66. 940-971. Logue. 1959. Rept. 1959. p. "Frequency Modulated Radar. DOPLOC Uses Phase-locked Filter. G: The Long Quest. June 1.. 66-74. 6. 5." McGraw-Hill Book Company. K. M.: Control of Surface Currents by the Use of Channels. 2. A.. no. C. 6. June. and R. Keiser. no. no.: Radar Meter Helps Enforce Traffic Laws. 4.. 3-7. H. pp. and D. Barker. The antenna is a small paraboloid with crossed feeds to make the polarizations of the transmitted and received signals orthogonal to one another. 15-21. 2.. pp. 150-152. Capelli. Electronics.. Sandretto.: Rotating Wave Radar. vol. B. 26. T. pp. F. 23. and J. Verlag Leemann. IRE Trans. N. 22. M. J. Ismail. vol. 14.: A 1949. 1 3. June. 1954. CW and Frequency-modulated Radar 111 process at the receiver of the master unit. 1951. Brantley. 42. E.. 20. W. IRE Trans. A. Kalmus. P. 28. July. H.. M. Boxcar Radar. vol. Wireless World. Electronics.

19-27. F. Skolnik. 1958. pp. vol. pp. I. Franklin. 6. 6. A. R. Elec. 31. A. M. AP-3. Appendix. May. Ohio). M. Navigational Airborne Self-contained and Lightweight J. to Doppler Radar Sensors. 5.S. Hochfrequenztech. R. IRE Trans. Doppler Technique. vol. Tollefson. 695-696. Electronics Low Noise 39. 1956.: : A Witmer. .. Space/Aeronautics. IRE Trans.: An Analysis of Bistatic Radar. 44. 32. C. IRE.: Principles and Performance Analysis of Doppler Navigation Systems. May. vol J 44. 1947. A.« « 100-103. 755-760. 30. Condie. vol. R. Proc. K. IRE Trans. A A CW : : 46. pp. 45.: The AN/APN-96 Doppler Radar Set. 49. 1958. Tracking Systems. pp.112 Introduction to Radar Systems Precise New System of F. South African Inst. Radio. Conf. December. M. E. Aeronaut. 1948.. ANE-4. Aeronaut. J. Conv. Proc. 35. Natl. ANE-8. 1959. 202-211. Ohio). 1956. Studien iiber Radarsysteme mit Frequenzmodulation. R. M. 161-172. T. 778-807. . W. . Canton. U. 45 Varian. Measuring Systems. Fried. Inst. 10. L. R. Patent 2. IRE 44. pp. 1957.. 1957.: Tracking the Earth Satellite and Data Transmission by pp. . June. AN/APN-67. vol.: Application of Frequency Modulation Techniques 1959. McKay. Trans. Wadley. Conf. September. pt. pp. an der E. The Nature of Doppler Velocity Measurement. 176-196. May. pp. pp. ANE-4. A.: (Dayton. 43 Poling. 37. and 38. 1960. H. Trans. vol. Rent. pp. and Locating System. . pp. vol. vol. Verlag Leemann... 1960. 34. F.. 1957. pp. Bibby. pp. 1950. IRE Trans. 30-33. March. Equipment in Surveying. . 143-148. IRE Natl. Proc. W. 62-1. Feb. R. C. 1961. J.: The Design of Airborne Doppler Velocity : 34. T. 143-161. 1959. 71-79. Robinson.M. C. vol..: FM-CW Radar. also see discussion. Tellurometer Manual. Zurich. 103-112. pp. F. 185-192.615.. 1140-1145. pp. 29. T.S. IRE. December. P. IRE Trans. Electronics pp. Commerce Publ. 1957. IRE. J. Berger. G.: Application of Electronic Distance Measuring Trans. Ismail. June. Proc. K. pp. December. B. Woodyard: Object Detecting 36. Hastings. 1958. 197-201. vol.: Basic Design Considerations: Automatic Navigator vol. Dept. and J. M.. ANE-4. A. A.: The AN/APN-81 Doppler Navigation System. hngrs. discussion. April-July. Record. B. System. U. H. : Raydist: A Radio Navigation and ^ Tracking System. F. IRE. 41. vol. 263-267.T. Thompson.435. (Dayton. September. ANE-4. Glegg. 683-687. W. Bonnelle. Hansen. pp. 33. Proc. December. W. 1957. T. E. Kurt J.. vol.. Natl. „. M. Radar. 47. Mengel. vol. 13. D. ANE-4. 157-175. Moody. Tele-Tech . Brown. IRE McMahon. 28. 1957. October. vol. C. Sollenberger. 40. 133-144. pp. MIL-4. Electronic Principles of the Tellurometer. August. 42. vol. Berger. IRE Trans. J. N. Proc. Mitchell: 47..: Multipath Phase Errors in CW-FM 1955. W.H.

/ d = 2v r jX. Indeed. Typically. But by the end of World War II the techniques and components for extracting doppler information with pulse radar were developed. moving targets may be discerned by the change in target position from scan to scan. 113 . MTI usually refers to a radar in which the dopplerfrequency measurement is ambiguous but the range measurement is unambiguous. MTI radar can extract the moving-target echo from the clutter echo even . Ambiguous range means that multiple-time-around echoes are possible. Moving-target-indication shift in (MTI) Radar 1 The doppler frequency caused by a moving target targets just as in may be used in pulse radar But. if discerning moving targets were the only advantage to be gained with doppler information. a distinction is sometimes made between the MTI radar and the pulse-doppler radar. it to distinguish fixed from moving CW radar. The ordinary pulse radar which does not use doppler information does not have this capability. In the pulse-doppler radar the doppler measurement is usually unambiguous and the range may or may not be ambiguous. although they are both based on the same physical principle. In the postwar years they were improved upon. or those which enter the radar receiver via the antenna sidelobes. while ambiguous doppler implies that "blind speeds" fall within the range of expected target speeds. The early pulse radars did not make use of the doppler information inherent in the echo signal from moving targets Consequently. but the echo from a target moving with relative velocity vr will be shifted in frequency by an amount given by the doppler formula (3. The fixed targets are called clutter.4 MTI AND PULSE-DOPPLER RADAR 4. A pulse radar which makes use of the doppler information is known as an MTI radar. In practice. The distinction between the two radars arose historically. and it is usually still applied. an especially appropriate name since they tend to "clutter" the cathode-ray-tube display with unwanted information. which stands for moving-target indication. In many instances. it would not be worth the trouble to instrument the radar to extract doppler. but the term MTI will be used when it is necessary to refer to the entire class of pulse radars which employ doppler information. (The radar resolution cell in this instance is the volume illuminated by a from fixed pulse packet.) Echo signals from fixed targets are not shifted in frequency. It is also called pulse-doppler radar. they were sometimes of little value in regions where large clutter echoes existed. In this text the historical distinction between the two will be maintained.1. where X is the wavelength of the transmitted signal. might be asked. why should the pulse radar be complicated by extracting the doppler information if the target range only purpose is to distinguish fixed from moving targets ? Since a pulse radar measures and angular position on each scan. The fixed-target echoes with which the desired target echo must compete are those included within the same radar resolution cell as the target. Another characteristic feature of the MTI radar is its delay-line canceler used to detect the doppler frequency shift. and most modern search radars usually include some means of extracting the doppler information to detect moving targets in the presence of clutter.2). the difference between MTI and pulse-doppler radars is only a matter of nomenclature. But doppler permits the pulse radar to discern moving targets in the presence of fixed targets even when the echo signal targets is orders of magnitude greater.

2 The area MTI using scan-to-scan developed using storage-tube techniques. 4. doppler radars can detect moving targets even when the clutter echo is 70 to 90 db greater than the target echo. fully A CW CW CW W coherent If the MTI radar.114 if Introduction to Radar Systems [Sec. A type of MT) which does not use doppler information directly is called area MTI. and the necessary radar may be converted into a pulse radar as shown in In principle.2) .target echo. 4. 1 is that a small portion of the power that generates the transmitted pulses is diverted to the receiver to take the place signal does more than function as a replaceof the local oscillator. where A1 is the amplitude and/i the carrier frequency. (a) Simple CW radar. It acts as the coherent reference needed to detect the doppler frequency shift. la. receiver. this C ment for the local oscillator. 4. mapping operation to the next are displayed. Devices of this type have been successA V f <- cw oscillator ' 'Reference signal r t±A \\ 'J > J Receiver f« Indicator (a) Pulse modulator " 1^1 c Power amplifier CW oscillator ft V Reference signal 'j' 333. shown in Fig.16 by providing a power amplifier and a modulator to turn the amplifier on and The chief difference between the pulse radar off for the purpose of generating pulses. By coherent it is meant that the phase of the transmitted signal The reference signal is the distinguishing feature of is preserved in the reference signal. In the area MTI successive time-spaced "relief" maps of the observed area are subOnly those objects which have changed position from one tracted from scan to scan. (b) pulse radar using doppler information. radar such as was described in Sec. the reference signal Vre{ = A 2 sin 2irf and the doppler-shifted echo-signal voltage ^echo is t t (4. indicator. 4. the major emphasis cancellation will not be discussed further in this chapter.1) = A sin Mft±tet 4^R (4. It consists of a transmitter. will be on MTI radars using sweep-to.sweep cancellation or its equivalent. oscillator of Fig. However. Instead. is CW oscillator voltage represented as A1 is sin 2Trft t. lb and the one described in Chap.\ v (b) Indicator f* Fig.1 Some pulsethe clutter echo is 20 to 30 db greater than the moving. 4. the antennas.1. Fig. 3.2 is simple Description of Operation. .

4. while the waveform of Fig. 4.2*. If. range).2. fd has a value other than zero and the voltage corresponding to the difference frequency from the mixer [Eq.2c) and many pulses will be needed to extract the doppler information. The is more applicable to a radar whose primary function video signals shown in Fig. since they contain both positive and negative amplitudes. Only the low-frequency (difference-frequency) component from the mixer is of interest and is a voltage given by Kdlff Note that Eqs. 4./ Sec 4I J - MTI and Pulse-doppler Radar 115 where A2 Az t = amplitude of reference signal = amplitude of signal received from a target at a range R fd = doppler frequency shift = time c = velocity of propagation The reference signal and the target echo signal are heterodyned in the mixer stage of the receiver. (4. t to tion (4.1) to (4. ^j: (c) Fig. 4. (4.2c is more typical of aircraft-detection radar. An example of the output from the mixer when the doppler frequency is large fd compared with the reciprocal of the pulse width is shown in Fig. For stationary targets the doppler frequency shifty will be zero. (b) video pulse train for doppler frequency fd train for doppler frequency d 1/ T (c) < video pulse v .2 are called bipolar. Ambiguities in the measurement of doppler frequency can occur in the case of the discontinuous measurement of Fig.3) represent imposed. 4.3)] be a function of time.3) sine-wave carriers upon which the pulse modulais equal to the doppler frequency /^. will on the other hand. 4. hence K diff will not vary with time and may take on any constant value from A iy including zero.2b might be the detection of extraterrestrial targets such as ballistic missiles or man-made satellites.3) (Fig./d is small compared with the reciprocal of the pulse duration.2c. but not when the measurement is made on the basis of a single pulse. is = At sin {l^t -4 -^^j (4. single sweep on an A-scope might A . (a) RF echo pulse train . Moving targets may be distinguished from stationary targets by observing the video output on an A-scope (amplitude vs. The difference frequency +A — L- X >( Hfl (a) AA (/>) A/ W inn > 1/T . when the target is in motion relative to the radar. The doppler signal may be readily discerned from the information contained in a single pulse. 4. the pulses will be modulated with an amplitude given by Eq. The case illustrated in Fig. However.

[a-e) Successive sweeps on an MTI radar A-scope display (echo amplitude as a function time). look a water tower. The delay-line canceler acts as a filter to eliminate the . One method commonly employed to extract doppler information in a form suitable for display on the PPI scope is with a it is delay-line canceler (Fig. 4. not appropriate for display on the PPI. moving targets cannot be arrows. arrows indicate position of moving targets.116 Introduction to Radar Systems [Sec. this is not a targets from point targets by the stretching of the echo pulse.g. (/) superposition of many sweeps. 4.) formation can flying in Echoes from fixed targets (pulse-repetition intervals) are shown in Fig. The superposition of the successive A-scope sweeps is shown in Fig. with like point targets. Also. 3/ The moving targets prod uce. 4.. vary in amplitude from sweep to sweep at a rate corresponding to the doppler frequency.3. 4. 4=4. appear as in Fig. reliable means of discriminating moving from fixed targets since some fixed targets can of Fig. 43b to e. 4. time. (It may be possible to distinguish extended ground However. Although the butterfly effect is suitable for recognizing moving targets on an A-scope. some moving targets such as aircraft Successive A-scope sweeps look like extended targets. but echoes from moving targets. Delay-line > Receiver Subtractor circuit Full-wove rectifier To indicator r=VP rt Fig.3a. indicated by the two distinguished from fixed targets. a "butterfly" effect on the A-scope.4).1 This sweep shows several fixed targets and two moving targets On the basis of a single sweep. remain constant throughout. MTI receiver with delay-line canceler. e.

Sec. just as was the input.1] d-c MTI AND PULSE-DOPPLER RADAR 117 component of fixed targets and to pass the a-c components of moving targets. Pulse modulator » fl+fc TR Klystron amplifier tf+$*£ Mix Stalo f? Mix + S frfd IF amplifier Coho Reference signal Phase detector To delay-line canceler Fig. The echo signal is heterodyned with the stalo signal to produce the IF signal just as in the con- RF ventional superheterodyne receiver. for . The stalo. 4. The coho is a stable oscillator whose frequency is the same as the intermediate frequency used in the receiver. and the mixer in which they are combined plus any low-level amplification are called the receiver-exciter because of the dual role they serve in both the receiver and the transmitter. 4..1Z> is not necessarily the most typical. In Fig. further description of this type of MTI radar may be found in Ref. 4. the output of the coho/c is also mixed with the local-oscillator frequency/. One is a normal video channel. A The characteristic feature of coherent MTI radar is that the transmitted signal must be coherent (in phase) with the reference signal in the receiver.5. This is accomplished in the radar system diagramed in Fig. The fixed targets with unchanging amplitudes from pulse to pulse are canceled on subtraction.5. the coherent reference is supplied by an oscillator called the coho. 3. 4. it must be converted to unipotential voltages (unipolar video) by a full- wave rectifier.5 by generating the transmitted signal from the translation coho reference signal. the video signal experiences a time delay equal to one pulserepetition period (equal to the reciprocal of the pulse-repetition frequency). Block diagram of MTI radar with power amplifier transmitter.stable /ocal oscillator. 4.5. coho. The function of the stalo is to provide the necessary frequency from the IF to the transmitted (RF) frequency. which stands for coherent oscillator. Although the phase of the . The output of the subtraction circuit is bipolar video. In addition to providing the reference signal. The simple MTI radar shown in Fig. The video portion of the receiver is divided into two channels. The significant difference between this MTI configuration and that of Fig. and subtraction results in an uncanceled residue. The outputs from the two channels are subtracted from one another. Before bipolar video can intensity-modulate a PPI display.\b is the manner in which the reference signal is generated. the amplitudes of the moving-target echoes are not constant from pulse to pulse. 4. The local oscillator must also be a stable oscillator and is called stalo. In the other. However. 4. The block diagram of a more common MTI radar employing a power amplifier is shown in Fig.

The phase of the coho is locked to the phase of the transmitted pulse each time a pulse is generated. available at Before the development of the klystron amplifier. In an oscillator the phase of the RF bears no relationship from pulse to pulse. a coherent reference signal may be readily obtained with the power oscillator by readjusting the phase of the coho at the beginning of each sweep according to the phase of the transmitted pulse. 6. Block diagram of MTI radar with power oscillator transmitter. block diagram of an MTI radar (with a power oscillator) is shown in Fig. oscillations to "lock" in step applied to the coho and causes the phase of the coho with the phase of the IF reference pulse. These include the triode. The various arrangements may be classified according to .6. tetrode. 1 4 5 ' ' The type of MTI radar illustrated in Fig. Each of these has its advantages and disadvantages. klystron. 4. normal amplitude detector since its output is proportional to the phase difference between the two input signals. any stalo phase shift is [Sec.118 Introduction to Radar Systems phase of the transmitted signal. The phase detector differs from the are both fed into a mixer called the phase detector.6 has had wide appli- reference signals in the The two methods described above are not the only ones of obtaining coherent MTI. the only high-power transmitter microwave frequencies for radar application was the magnetron oscillator. cation. A V ' Magnetron oscillator 3 ulse Trigger modulator generator ' RF locking pulse t Mix Stalo Mix ' IF amplifier 1 r Coho IF locking pulse cvJ reference Phase detector signal To delay-line canceler Fig. coho until the next locking the phase of IF locking pulse is generated to relock the A A CW CW pulse comes along. which stands for waster-oscillator /?ower amplifier. The phase of the coho is then related to the phase of the transmitted pulse and may be used as the reference signal for echoes Upon the next transmission another received from that particular transmitted pulse. and the Amplitron. Any one of a number of transmitting-tube types might be used as the power amplifier.1 stalo influences the canceled on reception because the stalo that generates the transmitted signal also acts as the local The reference signal from the coho and the IF echo signal oscillator in the receiver. However.6. 4. 4. portion of the transmitted signal is mixed with the stalo output to produce an IF beat This IF pulse is signal whose phase is directly related to the phase of the transmitter. A transmitter which consists of a stable low-power oscillator followed by a power amplifier is sometimes called MOPA. which are more fully discussed in Chap. For this reason the reference signal cannot be generated by a continuously running oscillator. traveling-wave tube. 4.

2. The carrier frequency might be typically 5 or 30 Mc. magnesium. An attenuator might also be inserted in the direct channel to aid in equalizing 1 the gain. Delay times of this magnitude cannot be achieved with practical electromagnetic delay lines. and (3) the echo and reference signals are compared at RF or IF.7. but it is of the order of magnitude of 10~ 5 that of electromagnetic waves. Subtracter Full-wave 1 video rectifier Fig. The carrier frequency modulated by the bipolar video is divided between two channels. The signal suffers considerable attenuation in the delay line. hence acoustic delay lines can be of manageable proportions. but frequencies ranging from 5 to 60 Mc or higher have been used. the acoustic waves are converted back to electromagnetic waves. The radar output is not applied directly to the delay line as a video signal since it would be differentiated by the crystal transducers that convert the electromagnetic energy into acoustic energy. 4. The velocity of acoustic waves depends on the delay medium. an amplifier with similar delay characteristics is included in the direct (undelayed) channel. water-glycol mixtures. but water. A block diagram of a typical delay-line-cancellation network is shown in Fig. Automatic balancing circuits pulse Carrier oscillator + Delay line Amplifier Det.Sec. while in the other it proceeds undelayed. and it must be amplified in order to bring it back to its original level. After the delay. aluminum. This difficulty is circumvented by converting the electromagnetic waves into acoustic waves and accomplishing the PRF trigger system Trigger generotor '"' l """ | AGC . It was one of the first practical MTI filter techniques developed and is usually less complex than other possible techniques. This results in eight possible combinations. and glass have also been used in delay lines. Thus delay times as long as several milliseconds are required for typical ground-based surveillance radars. or vice versa (2) the locking takes place or IF.7.2] MTI AND PULSE-DOPPLER RADAR . Canceled DpIo p[ Modulator Bipolor Uir ect channe Attenuation Amplifier Det. The delay line must introduce a delay equal to the pulse-repetition interval. The bipolar video from the phase detector modulates a carrier before being applied to the delay line. but the frequency response and linearity of the two must be similar in order to effect good cancellation. 1 whether at RF 4. 4. Since the introduction of an amplifier into the delay channel can alter the phase of the delayed waveform and introduce a time delay. Delay Lines and Cancelers Although the simple delay-line canceler is limited in its ability to do all that might be desired of an MTI filter. 119 (1) the transmitter locks the oscillator. 4. delay in an acoustical delay line. In one channel the signal is delayed. Both liquids and solids have been used as the acoustic delay media. and vice versa. Block diagram of a delay-line canceler. The length of the electromagnetic delay path would have to be equal to twice the unambiguous range of the radar. it has been widely used. Good cancellation in a typical application might result in an uncanceled voltage residue of the order of 1 per cent or . The amplification of the direct-channel amplifier need not be as large as that of the delayed-channel amplifier. Mercury and fused quartz are the two media most widely employed for MTI radar application.

cell Delay medium ^Bonding material Fig. but they sometimes produce cells The end unwanted reflections which interfere with perfect cancellation. The lack of a good bonding material hampered the early development of the solid delay line. sturdier absorbing end cell may be made by soldering the transducer crystal to a solid material whose acoustic impedance matches that of the delay medium. An absorbing end cell will eliminate or reduce the reflections. only moving targets produce an output from the subtractor. A typical transistorized MTI canceler operating at a prf of 360 pps and producing a cancellation ratio of 36 db may be housed in a f-ft 3 cabinet and operate on only 8 watts of power. The discovery of a satisfactory bond using evaporated indium or various other cements made the solid delay line a practical device. One of the simplest acoustic delay lines consists of a straight cylindrical tube filled with liquid mercury. One absorbing cell used in the early delay lines consisted of backing the transducer with the same medium. „ cell -»f\~\ .8. but are not very rugged. crystal ' . the unwanted reflections will be sufficiently attenuated by the line and may not be troublesome.120 Introduction to Radar Systems [Sec.000 . The purpose of the bonding is to provide maximum transfer of acoustic energy between the transducer and the delay medium. but an absorbing end cell increases the insertion loss by 12 db (6 db per transducer). Reflecting cells are more efficient and easier to construct. Ideally. When the delay medium is solid rather than liquid. crystoK cr ' 1 X . a good acoustic match can be obtained with a backing of hard lead. 4. energy back to electromagnetic energy. 4. an additional problem is encountered in bonding the transducer to the delay medium. The transit time of acoustic waves in mercury at room temperature is approximately 1 7.«sec the line must be A . For mercury. similar transducer (the receiving crystal) at the output PPL A i •Transmitter Receiver c End . The box labeled "automatic balancing" detects any amplitude or timing differences and generates an automatic gain-control (AGC) error voltage to adjust the amplifier gain and a timing-control error voltage to adjust the repetition frequency of the trigger generator. 6 ' Delay-line Construction. Consequently.000 . The canceled bipolar video from the subtractor is rectified in a full-wave rectifier to obtain unipolar video signals for presentation on the 40 db. If the line is long enough and if the cancellation requirements are not too stringent. enclosing the transducers may be either absorbing or reflecting. 4. The electromagnetic energy is converted into acoustic energy by a piezoelectric transmitting crystal. acoustic delay lines of the line converts the acoustic is crystal transducer are relatively broadband devices. However. To produce a delay of 1 . Any residue after cancellation is due to differences in channel gain or to the prf not being equal to the reciprocal of the delay time. 7 1 * The basic elements of an acoustic delay line are outlined in Fig. the medium has a damping effect which broadens the bandwidth. For most applications the length of the line must be more than 1 . a pilot pulse is inserted in the canceler. The quartznormally a high-g device with an inherently small bandwidth.5 ^sec/in. In order to maintain the gain of the two channels constant.8. when the transducer is coupled to the delay medium. End K~\^^ _ . Mercury end cells of this type have been used in operating equipments.asec for the secondaries to be attenuated sufficiently. Basic elements of an acoustic delay line.2 The outputs from the delayed and undelayed channels are detected to remove the carrier and are then subtracted.

However. Another method of obtaining a more compact delay line is to make use of multiple reflections in a tank filled with liquid. it seems that the net saving in weight and space with the liquid tank is not significant as compared with the folded line.8 2. This is a manageable size in a ground-based radar. where the longitudinal A mode is preferred. Volumetric delay line line. The impedance mismatch of 36 db assumes perfectly reflecting end cells. (There are six reflecting surfaces or three corner reflectors in this particular folded line.. and the size and weight are less than with liquid. and solid delay lines are not only practical.9. Since the velocity of propagation is slower in the shear mode. This is in contrast to the liquid line. in length exclusive of end cells. greater beam spread also results in solids because of the greater velocity.) The signal makes 31 passes across the line. the longer delay path required in solid media is not a limitation since it is quite practical to construct a solid line to obtain multiple folded paths similar in cross section to that of the liquid-mercury tank of Fig. 4. using multiple reflections in a Thus the solid delay line permits comparable delay times in tank of mercury (also similar smaller packages. but in many respects they are superior to liquid lines.. Free-space attenuation in mercury Tubular attenuation Impedance mismatch 36 db 12. This technique has not proved to be too practical with liquid lines.000 /tsec delay is as follows loss of the crystals .7 db db The loss due to these three effects is 52 db. 4. 4. The alignment of the reflecting surfaces is a problem. rough surfaces cause additional loss. One of the disadvantages of either solid or liquid delay lines is the large insertion loss.9.9. and it has been difficult to obtain a leakproof construction. more compact configuration may be had by folding the line back on itself one or more times. total insertion loss . Solid delay lines were not used in the early MTI radars because of development difficulties.2] MTI AND PULSE-DOPPLER RADAR 121 57 in. Absorbing end cells cause an additional 12-db loss.) Thus the attenuation could be as much as 70 to 75 db. The acoustic signal may be reflected at the folds by two plane reflectors set at 45° with the path of the beam and 90° with each other. 4. (The crystal is made with only 14 facets since no reflections take place at one of the facets. Each fold in the line increases the insertion loss by about 1 to 3 db. it is preferred to the longitudinal mode. Solids are capable of supporting both the shear and the longitudinal mode of propagation. The insertion loss of a typical folded 8 : mercury delay line with 1. using multiple folded paths can be constructed with a shape similar to the tank of Fig. it is of importance to make the delay line as compact as possible. 4. The most suitable solid delay medium to multiple reflections in solid has been fused quartz. but in those applications where space is at a premium. The velocity of sound in solids is greater than that in mercury. Although the solid delay line delay lines). There is no leakage problem with the solid delay Fig.9. Although these difficulties might be overcome. These difficulties were surmounted. The attenuation in the tube is based on a smooth surface. This is sometimes called a 15 MS-31 design. consequently a slightly longer delay path is necessary for the same total delay time. a more suitable shape is the many-sided polygon as illustrated by the 15-sided polygon in Fig. as A shown in Fig. 4. 10.Sec. A further loss of 5 to 10 db occurs at the reflecting surfaces of the folded line.

there is usually a continuous background of unwanted responses which bear no relationship to the shape of the input pulse. Radiation from the transmitter sidelobes might be internally reflected over some path other than the main path and find its way to the receiving transducer. The secondary responses are similar in shape to the input pulse. The operating frequency is 1 5 Mc for both lines. Multiple reflections in 31-pass quartz delay line. is The delay path from the receiving crystal Input transducer Output transducer "Absorber material Fig. 4. These may be eliminated or reduced with a straight-line delay path of large cross section.10. Further reflections can result in secondaries at any odd number of delay times. One source of secondaries is the third-time-around signal caused by the reflections at the receiver crystal which travel back up the line and are again reflected by the transmitter crystal toward the receiver signals three times that of the original delay. Secondaries may also be produced by such processes as conversion from one mode of propagation to another (longitudinal to shear mode. The signal makes 31 1 1 . In addition to discrete secondaries. (Courtesy Bliley Electric Company. secondaries might also appear at the receiving transducer because of the sidelobe radiation from the transducer's diffraction pattern.10. and dispersion effects.122 Introduction to Radar Systems [Sec. 4.000-^sec fused-quartz delay line and a mercury delay line is shown in Table 4. . 4. and they may arrive at the receiver either before or after the main delay.) inhomogeneities within the medium. In the folded-path delay line using multiple internal reflections. where it may be detected by the main beam or by its sidelobes. A comparison of the characteristics of a . or vice versa). scattering by crystal.2 Another disadvantage of acoustic delay lines is the presence of unwanted secondary which may arise from a number of sources. The quartz line is a 15-sided polygon as in Fig.

the bandwidth The insertion loss of the solid quartz wider. On the other hand. Quartz-crystal transducers when used with solid lines are normally designed to generate shear waves (to reduce mode conversion).1. -100 -55 to +100 + 300 — 38 to +80 passes. The solid line less subject to The video signal [Eq.4) phase shift AnfRjc amplitude of video signal At a time t + T. lb Size. its higher coupling coefficient would result in a significantly lower over-all insertion loss. (4. The delay-line canceler acts as a filter which rejects the d-c component of clutter. the unwanted secondary responses generated in the solid delay line are usually greater than in the cylindrical mercury line. °C 13 t Arenberg.7 Mc as measured between the half-power points. and it operated at a carrier frequency of 1 5 Mc with a bandwidth of 6. An experimental delay line developed at the Bell Telephone Laboratories 15 using barium-titanate transducers with a fusedquartz delay medium resulted in a total midband insertion loss of 20 db. less line. mechanical shock and vibration as well as temperature variations.2] MT1 AND Table 4. db below main delay 3-db bandwidth. This may be readily accomplished by reflecting the longitudinal waves off a surface at a critical angle which completely converts the longitudinal waves to shear waves. which must be converted to shear waves. PULSE-DOPPLER 15 RADAR 123 Performance of l. The length of this line was 1 . If barium-titanate ceramic were used for the transducer elements rather than quartz crystals.000 ohms Secondaries. Most commercial delay lines have used quartz crystals as the transducer elements. but barium-titanate transducers normally generate longitudinal waves. line is The mercury delay than the mercury is line and its size and weight are less. the video voltage from the same target will be k = = V2 = k sin [iTrfit + T)- <f> ] (4. db. and the manufacturing of quartz lines may be slightly more involved than the manufacturing of mercury lines.Sec. 4.000 /^sec. into 1. The spurious responses were claimed to be as good as obtained with quartz-crystal transducers. in. Mc 25 8 Weight.3)] received from a particular target at a range i? is V1 =ksin(2 1rfd t-4^ where </>„ (4. Filter Characteristics of the Delay-line Canceler. 34 Temperature x 10 6 /°C Temperature range.5) . Other types of delay devices which might conceivably be used for MTI application are magnetic drums or disks and electrostatic storage tubes. The tendency in most modern MTI radars using delay-line cancelers is to employ the solid delay line rather than the liquid line. The quartz-crystal transducers constitute a significant portion of the total insertion loss (approximately 36 db out of a total of 45 db for a typical solid line using fused quartz or 65 db for a liquid-mercury line). where T = the pulse-repetition interval. 3 25 coefficient of delay. the filter also rejects energy in the vicinity of the pulse repetition frequency and its harmonics. db below main delay 40 50 6 1 50 55 Third-time-around signal. Because of its periodic nature. is is a cylinder.000-/<sec Delay Lines at Fused quartz 45 Met Mercury 65 Characteristic Insertion loss.

6)] consists of a cosine wave at the doppler frequency^ with an amplitude k sin irf T. f --Jd— - nfr (4. . tt. . time = Relative frequency response (visibility factor) of the single-delay-line canceler. . 2.7) where n The delay-line canceler not only eliminates the d-c component caused by clutter (n — 0).8). (4. 4. 3. (4.1 1 .8a) the «th blind speed. a Thus the amplitude of the canceled video output is a function of the doppler frequency shift and the pulse-repetition interval. . + *) 4>o (4. or when d Tin the amplitude factor of Eq. 124 Introduction to Radar Systems [Sec. and the relative velocity in knots. (4. 4. but unfortunately it also rejects any moving target whose doppler frequency happens to be the same as the prf or a multiple thereof. and hence the video voltage. If the The with CW CW first blind speed is to be greater than the maximum radial velocity expected from the target. f = pulse repetition frequency 1. Those relative target velocities which result in zero MTI response are called blind speeds and are given by nX = 0. 4. to the amplitude of the normal radar video k] is shown in Fig. the product Xf Thus the MTI radar must operate at long r must be large. measured in centimeters.8&) blind speeds are one of the limitations of pulse MTI radar which do not occur radar.6) is 0. 2.2 is the Everything else is assumed to remain essentially constant over the interval Tso that k same for both pulses. Vt Frequency Vr T -- Fig. f r in cps. without loss of generality. the pulse radar is blind to those targets whose radial velocity satisfies Eq. the blind speeds are v " =^Mi 102 (4. . delay l// r Blind Speeds.2tt. The output from the subtractor is v=h- k sin nf (l T cos **(. . It is also assumed. The relative frequency-response characteristic of the delay-line canceler [ratio of the amplitude of the output from the delay-line canceler. are equally divided between the delayed and the undelayed channels of the canceler. .. k sin (nf d T). . (4. The response of the single-delay-line canceler will be zero whenever the argument 7rf etc. It will be recalled that the radar was blind to targets with zero or near-zero radial velocity. . . . In addition. or prf. They are present in pulse radar because doppler is measured by discrete samples (pulses) at the prf rather than continuously.. The ordinate is sometimes called the visibility factor. The output from the canceler [Eq. that the gain through the delay-line canceler is unity. 1 1 . T IT where v„ is = -? 2 If A is n = 1.6) The normalizing factor £ multiplies each video voltage since it is assumed that the power.

2. Unfortunately. so that in systems applications which require good MTI performance.ADAR 125 wavelengths (low frequencies) or with high pulse repetition frequencies.000 Mc (X band). For example. to resolve range ambiguities by varying the pulse repetition frequency as described in Sec. Mc (UHF). Low radar frequencies have the disadvantage that antenna beamwidths. However. Furthermore.2] MTI AND PULSE-DOPPLER R. Since commercial jet aircraft can have speeds of the order of 600 knots. \.000 Mc (S band). (/? unamb 600 knots.000 Maximum unambiguous range. 1 \l 1 1 1 1 1 1 1 MIL - - X% 1. of 300 In practice. long-range MTI radars that operate in the region of L or S band or higher and are primarily designed for the detection of aircraft must usually operate with ambiguous doppler and blind speeds if they are to operate with unambiguous range. ^ ~~ ^%> N^\° \o x&\ \ \^> \cx v«\ --* \o X > ~E ~ \<?. Unambiguous doppler information (no blind speeds) as well as unambiguous range are simultaneously possible in a long-range ground-based search radar if the transmitted frequency is sufficiently low.12. and military aircraft even higher. It is also possible to reduce the effects of blind speeds with a staggered prf as described later. and 4 nautical miles at 10. the MTI performance will usually suffer.10. the necessity for resolving range ambiguities in this type of radar adds to its complexity and generally requires a longer time on target. It is possible. are wider than at the higher frequencies and would not be satisfactory in applications where angular accuracy or angular resolution is important. Plot of MTI radar blind speed as a function of maximum unambiguous range. with radar frequency as the parameter. the is primarily determined by the unambiguous range requirement. for a given-size antenna. 4. Fig.000 \?> X \ N. blind speeds in the MTI radar can be a serious limitation. 4. 1 \l 1 X'W M \ ^\. be easy to avoid. the maximum unambiguous range would be 1 30 nautical miles at a frequency = 0. the first blind speed might be placed outside the range of expected doppler frequencies if ambiguous range can be tolerated. 13 nautical miles at 3. The pulse repetition frequency cannot always be varied over wide limits since it In Fig.12. first blind speed v t is plotted as a function of the maximum unambiguous range If the first blind speed were cT/2). The presence of blind speeds within the doppler-frequency band reduces the detection capabilities of the radar.000 . Blind speeds can sometimes be traded for ambiguous range. or both. in principle. there are usually constraints other than blind speeds which determine Therefore blind speeds might not the wavelength and the pulse repetition frequency.Sec. a radar operating at a frequency of . NP\ \^ >y _ \ INlk 10 first \ ^ \c \ ^ X 100 i i ^v iS IN 1 \ 1 1 XI I is \ i iS INI 1 100 nautical miles 1. 4.

.2 cm with a pulse repetition frequency A first blind speed computed from Eq. (4.000 Mc (X band) for an antenna of the same size. It is assumed for purposes of comparison mi) is the same as the that the maximum sensitivity of the MTI radar (i. all unity. (4. If the target is assumed to be of constant cross section velocity with equal probability. greater than the echo from the target. if the echo from the cloud were 7 db 7 db below the maximum. (The problem of defining the minimum detectable signal is ignored here since it is the The relative performance between MTI and non-MTI radars which is of interest. receiver with a delay-line canceler occurs when sin {jrf d T) = _ doppler frequency shift is small. nor is the assumption that all doppler velocities are equally likely a good one in all circumstances. we may write . The probability that the signal will be detected by the MTI radar compared with the probability that it will be detected by the non-MTI radar is The Relative probability = */2 . the angular resolution is 1 00 times worse than it would be at 10. should be remembered that this applies to a target in the absence of clutter An MTI radar will maintain its performance in those situations . vx = = relative velocity first of target blind speed Equation (4. Voltage response at low doppler frequencies Maximum voltage where vr vt vr response d T_ tu^.e.) relative (voltage) response of the MTI than that corresponding to maximum radar when the target radial velocity response is is other Relative (voltage) response = Vk sin {jrfd T) (4. It is also assumed that the signal strength (power) received = by the MTI radar at maximum response is K times the minimum detectable signal. both would appear to be of equal strength at the output of the delay-line canceler..126 Introduction to Radar Systems [Sec. 1} It represents the fraction of time Eq. that MTI radars with the type of characteristics as illustrated by the above example do not eliminate as much of the storm clutter as might be desired.. No this simple analysis of the relative detection capability is only an approximation. thunderstorm moving at a relative velocity of 30 mph (26 knots) will produce a response In other words. consider (as does Ridenour 1 ) a a wavelength of X = 9. (4. it ("in the clear").9) expresses the ratio compared with maximum response. as in the double-delay-line canceler discussed later. assuming MTI.2 100 Mc and a prf of 200 cps can achieve an unambiguous range of 400 nautical miles with a first blind speed of 600 knots. 4. Response of Single-delay-line Canceler. The delay-line filter characteristic should have more attenuation in the vicinity of zero doppler frequency.10) is greater than It should be cautioned that values of doppler are equally likely.13 shows that the performance of the MTI radar is not as good as a radar without This is plotted in Fig. definition has been given for minimum detectable signal.6)]. when nfd T non-MTI radar sensitivity. it is and can have any radial of interest to ask what is the probability of the target being detected with the MTI radar as compared with the probability that it will be detected with a normal non-MTI radar. _.13. 4. fr = radar operating at 2.10) is equal to or greater than unity. But at 1 00 Mc. The maximum sensitivity of an MTI When the 1 [Eq. 4.000 cps. (4. . Although Fig. The between the response of a target moving with velocity As an example.sin^O/V*) 77-/2 (4 .8) is 180 knots. Thus we may conclude that this type of performance Experience has confirmed is not good enough to eliminate the slowly moving cloud.10) target will be detected by the MTI radar whenever Eq.

4. then the probability-density function for p(y) probability-density function for p(x) by p(y) related to the dy = p(x) dx = (4. 4. and a one-to-one correspondence between y and x. : observed as any other angle. then P(y) = (4.13) dy/dx .13. although it is a convenient one and probably as good an assumption as any if there is no prior knowledge of target behavior.12a) or p(y) — dy/dx np(x) (4. This is A Radar Fig. Target (Any angle is just as likely to be to be equally probable. for example.14. In this section the radial. It is assumed that the target velocity v is constant and that the target trajectory makes an angle 6 with the axis of the radar beam All values of the angle d are assumed as shown in Fig.11).12b) If each value of y corresponds to n values of x. that is.14. (4. concise nique involved is standard in probability theory. with the target "in the clear"). Geometry of target. 4. 4. Plot of Eq. It is The relative velocity is vr = desired to find the probability-density function for the relative velocity.velocity probability-density function for two different target assumptions will be derived. functional relationship between two random variables y =f{x).) v cos 0. 16 If there is a explanation is given. the probability that the reladvr tive velocity will lie between the values of vr and v r + . radar and is assumed single-valued. a problem in determining the change in the probThe techability-density function with a change in variable. by Bendat. compared with a non-MTI radar (all doppler frequencies assumed equally likely.2] MTI and effectiveness of the Pulse-doppler Radar is 127 where the normal non-MTI radar drastically reduced by the presence of large clutter echoes.Sec. The assumption used above Probability of Obtaining a Particular Radial Velocity. case 1 target with constant velocity. 10 K= ratio of 50 40 20 30 MTI echo signal power to non-MTI minimum detectable signal Relative detection probability of a single-delay-line canceler MTI as Fig. that all radial velocities are equally likely may not always be realistic.

v r) = p(v)p(v ) r (4.16.17) Because v r can never be greater than divided into two parts. 4.18) =- —— 1 within the limits of integration v min and v m& /""max . (4. Probability-density function of assuming constant velocity and all angles equally likely. A plot of this equation is shown in Fig. We wish to findp(vr (+0 and — 0). 4. 1 6) must be integrated over the variable v.18) and (4. 1 6) 0. but with the appropriate values of v mln and .15).14) ATI since 6 is equally likely over the range from of v r corresponds to two distinct values of to 2n. = — .6 again assumed to be equally to 2w.19) is shown in Fig.16.15) sin 6 The minus sign obtained on differentiation is ignored since probability-density functions must always be positive.128 Introduction to Radar Systems is [Sec. Fig. 4.4 0. the variable v r within the PiiPr) = "(^max 1 (l> 2 - I*?)-* dv ^min) ""min = When fc<«v) v r lies -(cosh-i ^22 - cosh" 1 ^) l 1 for < Vr < Vn (4. the When vr < i>min. the joint probability-density function for v and vr is p(v. and integration is v. (4. 4. Each value P(v r) = 2p(6) 1 1 dv r jdB vr ttv <v (4. If the relative velocities are distributed according to Eq. (4.13).2 In our example the probability-density function for the angle 6 given by 1*0) =r (4. while the target velocity is assumed to lie anywhere within the range v min to v max Therefore p{v) (t>max -1 tfmin) and p(vr) is given by Eq. (4. (4. v„ PlO>r) p(v. 1 5. so that 1 ).19). integral will never exceed v. CASE 2 TARGET VELOCITY BETWEEN D min AND When the target velocity v and the angle ttmax: are both random independent variables.19) A plot of the probability-density function as given by Eqs.15.v r )dv (4. (4. To find the probability-density function for vr Eq. - 7r(l>max — fminj J» r (v 2 -viy i dv= 7r(l) m ax — cosh ^min) -1 "max Vr for y min <V < r Umax (4. instead of uniformly as assumed in the derivation of Eq. is The angle likely over the interval .1 1 (with relative velocity instead of frequency as abscissa) and a curve similar to Fig. 4.11) or of Fig. 4. the over-all target-radar response characteristic for the simple-delay-line canceler may be obtained by multiplying the ordinate of Fig.

Note that the first blind speed of is increased several times over what it would be for a radar operating on only a single pulse repetition frequency. In the example of Fig. One method of obtaining the second pulse repetition frequency is to add to the MTI delay line in the cancellation network a short section of line that is switched in and out of the system periodically. the pulse-repetition interval of the transmitted signal must also be changed. Therefore. or the period might be alternated on every other pulse. 1 7. it is known as a staggered prf.16. The pulse repetition frequency might be switched every other scan or every time the antenna is scanned a half beamwidth. it is not always advisable if the possibility of large second-time-around clutter echoes exists.2] MTI AND PULSE-DOPPLER RADAR 129 Multiple and Staggered Pulse Repetition Frequencies. or some other convenient grouping of pulses.Sec. 4. pulse repetition frequencies are in the ratio of 5 :4. 4. every pulse. Zero response occurs only when the blind speeds of each prf coincide. Switching may be accomplished every scan. In addition to changing the length of the delay line. clutter echo will A second-time-around not cancel in the delay line when the prf's are staggered pulse to pulse. regions of low sensitivity might appear within the composite the composite response = . . assuming target velocities distributed v mBK . Instead of using two separate radars. Although switching of the repetition period on every alternate pulse may be convenient. 1 7. every half beamwidth. 4. 17 Fig. passband. When the switching is pulse to pulse. the blind speeds are coincident for 4/7\ Although the first blind speed may be extended by using 5/r2 more than one prf. Probability-density function of relative velocity. if one radar were "blind" to moving targets. 4. uniformly from v min to An example of the composite (average) response of an MTI radar operating with two The separate pulse repetition frequencies on a time-shared basis is shown in Fig. the same result can be obtained with one radar which time-shares its pulse repetition frequency between two or more different values {multiple prf 's). it would be unlikely that the other radar would be "blind" also. The blind speeds of two independent radars operating at the same frequency will be different if their pulse repetition frequencies are different.

130 Introduction to Radar Systems [Sec. Consider the simple block diagram in Fig. l/T^. which illustrates that portion of the radar which might be used to produce a pulse transmission staggered every other pulse. The period between the transmitted pulses is alternately T— e and T+ e as . The trigger pulses are alternately switched between an undelayed path and a short section of delay line of length e.17. which. in turn. Nevertheless. (6) same for/.18. Trigger to modulato Ganged switches-M Delay Bipolar video line PRF generator period T € Delay-line canceler delay T _Canceled "video Fig. (a) Frequency-response characteristic (visibility factor) for/. 4. The prf generator output is a steady train of pulses with a uniform interval T. These pulses trigger the modulator.2 Echoes might result that could be taken for those of a moving target. fires the transmitter. Frequency (i) Fig. one should carefully consider the effect of second-time-around echoes before specifying staggered prf 's. 4. = 1/Ti. Observation of second-time-around echoes over several scans will show whether the target is in motion or is stationary and may be ignored. (c) composite frequency response with 7i/r2 i. Means for generating a staggered prf. The reason second-time-around echoes do not cancel is described below. 4. 4.18.

I t Received echo at/J I I -J Received echo ot I ^ih- h^-i (A) -&*-i B I Fig. 4. Basic PR . in those periods where the transmitter trigger Therefore pulse was undelayed. 4. 4. In addition to the extra equipment required of the staggered-repetition-rate MTI radar. 4. uncanceled residue results. fixed-clutter target. as shown.32) is shown in Fig. a second-time-around target appears. Another method of generating a staggered prf using a recirculating trigger pulse (as described later in Fig. (/>) second-time-around clutter target. (a) Normal.-j T Received echo <«) h— r Basic > \ ' r- PRF H Transmitter trigger r+e- t-r-*-i| -J Jb L i -jfrh- -j. 1 8) appear from pulse to pulse at The delay line in the canceler is of the same time with respect to the trigger pulses. Timing relationships in staggered prf MTI radar. just as if the target were in motion. 1 1 T+t —H*— T-t-*)£ i Transmitter trigger i LJ I -r+e — i L Received echo at/3 T J 7^ ^ r +c\ r +t \. length T.20.2] MTI AND PULSE-DOPPLER R. the received echo would be undelayed. 4. . 4. The received pulses are delayed an amount e in addition to their transit time Tr This additional delay e may be readily accounted for in the timing circuits. Conversely.19. the received echo would be delayed by an amount e.ADAR 131 of Fig. .196. The target echo returns after a time In those periods where the transmitted pulse was delayed. the pulse-to-pulse echoes do not have Therefore an at the input to the delay-line canceler. the data rate (number of hits per scan) is effectively lowered because of the time sharing between the two frequencies.-\T ->U+cfr. shown in the timing relationships Tr . 4.19a. target echoes at the input to the delay line {B in Fig.h Sec. This is illustrated by When the same time relationship the timing relationships in Fig.

4. (a) Double-delay-line canceler. 4. 4. Double cancellation requires tion in response for moving targets. the output from the single-delay-line canceler V= is fcsimr/^Tcos 2-nL (+{)-*.21a). 4.21* has the same frequency-response characteristic as the double-delay-line canceler. earlier. 1 1) does not have as broad a clutter-rejection null as might be desired in the vicinity of d-c or at doppler frequencies corresponding to the prf and its harmonics.20) from the second canceler :eler the difference between \ Kand V (4.132 Introduction to Radar Systems Cancellation. signal f(t) is inserted into the adder along with the signal from the A . canceler. canceler. (b) three-pulse-comparison canceler. The clutter-rejection notches may be widened by passing the output of the delay-line PRF- Trigger generator [Switching Bipolcr Pulse amplifier video Modulator Delay I line _J Delay line T Undelayed channel Det. The operation of the device is as follows.6) subtracted from the canceler output V from the pulse period Tsec V= The resultant output k sin tt/jTcos is (4.22) rather than the sine-wave response of a single-delay-line Input € _r Deloy line /"=V/ r Delay line T-\/f r Output Input Delay line T-% Output Delay line T=Vf r -2 Fig. The two-delay-line configuration of Fig. in In the double-delay-line Fig.21. (4. [Sec. 4. 4.21) V-V = k The amplitude of _ 2 sin nfd T sin „ [2irf^t + T) - <£ ] the double canceler compared with a non-MTI radar has a sinesquared shape (Fig.20. This may be seen by inspection since the double-delay-line canceler is simply two single-delay-line cancelers in cascade.2 Double The frequency-response function of a single-delay-line canceler (Fig. more equipment and results in a slight reducNote that the double-delay-line frequency- response characteristic is the square of the single-delay-line characteristic. canceler through a second delay-line canceler (Fig. Canceled video Subtractor Det. Means for generating a staggered prf with a recirculating pulse. 4.

23. drift in the pulse amplitudes from their correct values results in a first-order lack of cancellation.2] MTI AND its PULSE-DOPPLER RADAR 133 preceding pulse period. (After White and Ruvin. w IRE Natl. any also a small quantity. Canonical-configuration comb filter. and if both are out of adjustment by a small amount. the other is still capable of canceling stationary clutter. The ideal MTI filter characteristic one which rejects the clutter spectrum without eliminating any moving targets. Shaded area represents clutter spectrum. This general configuration may be used to implement any 18 realizable filter.Sec. 4. Record. 18-20 The almost any desired frequency response have been described in the literature. in the three-pulse-comparison network. Fig. 4. is . the residue which is left is the product of two small quantities and is On the other hand. If the circuits drift out of adjustment because of aging of components or some other cause.23). basic technique employs a number of delay lines in cascade with feedback and feedforward paths (Fig.) Shaping the Frequency-response Characteristic. Frequency response of single-delay-line canceler (solid curve) and double-delay-line canceler (dashed curve). plus the signal The output of the adder is therefore 2f(t - + T) +f(t + IT) which is the same as the output f{t) from the double-delay-line canceler T) -fit -/(/+ + T) +fit + IT) This configuration is commonly called the three-pulse-comparison canceler. Cortv.22. parison canceler Another advantage of the double canceler over the three-pulse-comis that each section can be adjusted separately. it is called the canonical configuration. 4. The three-pulse-comparison canceler is equivalent to the delay-line canceler only so long as they are both in perfect adjustment. Fig. f{t) amplitude weighted by the factor —2.22. cancellation networks which constitute the double canceler drifts out of adjustment. but it is possible to obtain delay-line filters with frequency characteristics more suitable than the sine or the sine-squared The techniques for synthesizing delay-line filters with characteristic of Fig. the double-cancellation network will not If either one of the two deteriorate as rapidly as the three-pulse-comparison network. The ideal characteristic is not achievable in practice. with from two pulse periods previous. 4. and because of this property. 4.

Ideally the two are equivalent. White and Ruvin 18 configuration sections.) papers. (a) Three-pole Chebyshev lowpass filter characteristic with 0. 4.0 0.2 The canonical configuration is useful for conceptual purposes. . The weighting factors shown on the feedback paths apply to the characteristic of Fig. Natl.25. band as well as in the passband.24a. whether it is a Butterworth.134 Introduction to Radar Systems [Sec. 4. This type of filter characteristic may be obtained with a single delay line in cascade with a double delay line as shown in Fig. 4. example. 1 state that the canonical may be broken section having into cascaded no more than two oj (a) delay elements. 2 3 Angular frequency.24c.25. single-delay-line canceler (Fig. 4.5 1A (i>) -S 0.26).24). the three-pulse-comparison canceler (an example of the canonical configuration) gives poorer as.24. 18 19 Consider the frequency-response Chebyshev lowhaving 0. 1 " Proc.27 for various values of the feedback factor k. the three-pulse-comparison canceler.5 db ripple in the passband (Fig. 4.5-db ripple in the passband (6-rf)delay-line filter charac. characteristic of a three-pole filter pass Fig. (After White. 4. the case of the double canceler vs.22) plotted in Fig. An example of the use of these filter characterrejection istics periodic applied to the design of a delay-line filter is given in either of White's - teristics derived from (a). 4. 4. This type of configuration is sometimes preferred to the canonical configuration. Chebyshev.22) = exp (jlvjT) - 1 exptpTr/D-fc is (4. The three different delay-line-filter frequency-response characteristics shown in Fig. "0 Frequency 1/7" performance than the double canceler (an example of cascaded sections). Thus no feedback or feedforward path need span more than two delay elements. Con/. The synthesis technique described by White and Ruvin may be applied with any known low-pass filter characteristic. for 1. or Bessel filter or one of the filters based on the elliptic-function transformation which has equal ripple in -the W) Fig. Electronics.24/3 to d were derived from the low-pass filter characteristic of Fig. 4. Form of the delay-line filter required to achieve the characteristic of Fig. But if the delay lines are not in perfect adjustment. on Aeronaut. but it may not always be desirable to design a filter in this manner. Its frequency-response is characteristic 20 H(f) Equation (4. 4. Another example of a delay-line periodic is filter with adjustable frequency response the double-loop. 4.24c.

. it is also necessary that the delay line be more accurate than with video cancellation. = n/T = nfr n = 0. 2tt. In addition. limited the application of IF-canceler circuits.27. Hence a requirement of the IF delay-line canceler is that the intermediate frequency must be an integral multiple of the prf. 4. . (After Urkowitz.26. These two additional requirements imposed upon the IF canceler are not present with the video canceler and have. Trans. When there is no doppler frequency shift. 4. it is only necessary that the time delay be accurate to within a fraction of the pulse width (of the order of 1 per cent for 40-db cancellation). 20 IRE Unlike normal video cancellation. . w = . The two IF signals of amplitude VIF subtracted by the canceler are vi V. (4.) 1 — k = amplifier gain.26 with changing feedback factor k. .25) The output of the phase a voltage Vv =± kVlF sin [t7(/if ±/ )T] cos \2tt ±f + UiY±f )d it i (426) to The video voltage is at the doppler frequency /.SeC 4 2 ] - MTI AND PULSE-DOPPLER RADAR 135 Cancellation at IF. 1.Wht*o" WifRo' (4. tt. Frequency Fig. 4. the two signals reinforce rather than cancel. . lv T = . 2. (After Urkomtz. With the video canceler. to perform the delay and cancellation in the IF rather than the video portion of the radar receiver (Fig.) to within a fraction of the period of the intermediate frequency. fd 0. If the delay time were to vary by one-half of an IF period. single-delay-line canceler of Fig. 1 It is possible.24) One-half the difference between these two signals Vd = VlF sin [tt(/if ±fd )T~] cos 2H/if± /„)(' + detector is j) -^^2 _j^R^ (4. . . = ^if sin M/JFi/i)*..23) = Kp sin MflF ±fd )(t+T)is (4. and has an amplitude proportional sin7r(/ IP ±/. in principle.27) Fig. 20 IRE Trans. while in the IF canceler the delay time must be accurate .)r Canceled output (4. 4. cancellation at IF involves a residue which is a function of the intermediate frequency times the time delay (f T).27)] will not be zero unless Trf or/IF 0. single-delay-line canceler. in the past. Double-loop. Response shaping of doubleloop. but the amplitude of the canceled signal [Eq.28). 3.

The output direct channels.29. 4. of the phase detector is passed through a differentiating circuit. Reference signal Received J signal Delay line IF amplifier - ~L Subtracfor f . 4. The output of the discriminator is proportional to the amplitude differences between two successive pulses. that Similar considerations apply to performing the subtraction at the carrier frequency is. and the two outputs are compared in the phase detector. The block diagram of the phase-detector cancellation circuit is shown in Fig. . The block diagram of the two-mixer FM cancellation system is shown in Fig. 1 is an example relative The gain stability between the delayed and the undelayed channels must be maintained within close limits if perfect cancellation is to result. The differentiated output is proportional to the difference in frequency between the two signals and thus is a measure of the uncanceled amplitudes of two successive video pulses. The frequencies of the delayed and AM the undelayed signals are compared. Block diagram of IF delay-line canceler. In the undelayed channel the modulated signal centered about f Q is mixed oscillator of frequency^ in mixer 1 to obtain a modulated with the output of a stable signal centered about a carrier f fv The outputs from the two channels are heterodyned in mixer 2 to obtain a frequency-modulated signal centered about /j. and any differences in frequency correspond to (frequency-modulation) cancellation.2 in the video delay-line canceler. 4.6)]. canceler is that it is easier to keep in adjustment since gain The advantage of the canceler. CW + . amplified. variations between the two channels are not as important as with the FM FM 1 AM Bipolor video Canceled Direct from phase — FM oscillator ' Phase detector channel Differentiation circuit -* bipolar video detector Delay line Delay amplifier Fig. This disadvantage may be alleviated by converting the video amplitude variations into frequency variations. Thus the two channels of delay-line cancelers must be maintained in perfect gain adjustment. The video signal modulates the reactance-tube oscillator centered at frequency f One portion of the frequency-modulated signal is delayed. AM is applied to the PPI. is FM Delay-line Cancellation? The delay-line canceler of which Fig. This frequency-modulated signal is limited and applied to a discriminator to convert its frequency variations to amplitude variations. The output of delay-line canceler the differentiating circuit is of the same form as the output of the A full-wave rectifier inverts the negative pulses before the canceled video [Eq. An oscillator (such as a reactance-tube oscillator) is frequency-modulated according to the amplitude of the The frequency-modulated signal is divided between the delayed and the bipolar video. Phase detector " Canceled video from mixer r* Fig.30. 4. and fed into mixer 2.29. 4.136 Introduction to Radar Systems [Sec. FM delay-line canceler using phase detector. There are two types of cancellation.28. which is one of its chief limitations. by subtracting the delayed and the direct signals is before the carrier removed. (4. FM cancellation systems. This is called differences in amplitude. .7 sometimes called an AM (amplitude-modulation) delay-line canceler. 4. One is called phase-detector and the other is called two-mixer cancellation.

) FM single harmonic of the delay line used for double cancellation by operating at the fundamental and the crystal transducer. Stable CW oscillator Bipolar video /i from phase — FM oscillator Mixer No. 4. {After Solomon? 1958 Proc. An example of an Fig. the advantage of the adjustment necessary with the A FM canceler that does not require the continual M canceler in order to maintain maximum performance.31 and in the undelayed channel of Fig. This canceler is slightly different from that shown in Fig. 4. if double cancellation is desired. 4. 2 f\ Limiterond discriminator Conceled - bipolor video deteetor . third tronics. double-cancellation network 2 using a single delay line is In the diagram two delay lines are shown.2] MTI and Pulse-doppler Radar 137 The phase-detector canceler is the simpler of the two. Also. 4. thus eliminating a second reactance-tube deviable oscillator. FM shown in 30-Mc oscillator 1 90-Mc Bipolar video delay line Posfdelay amplifier 90 Mc Mixer No. FM canceler to attenuate fixed clutter signals is it is probably no better Both types are capable of reducing clutter by about 30 to 35 db.30 in that mixer 1 is in the delayed channel in Fig. 1 If the delay time and the pulse repetition period are not exactly equal. 4. 2 90 Mc 30 Mc Double 1 Mixer 45 Mc No. on Military Elec- The ability of the than the AM canceler.31 . FM delay-line canceler using two mixers. On one pass the delay line operates at the fundamental of the crystal transducer (30 Mc).30. 1 60 Mc m > FM oscillator 90 Mc Mixer 30 Mc Mo. 4.30.t %+f\ Mixer No.Sec. fo Delay line Delay amplifier Fig. phase-detector cancellation presents additional complexities.31. In general. IRE Conf. but it is not . it is not difficult to maintain a stable prf. but only a single line is used in practice. while on the second pass the crystals are driven at the third harmonic (90 Mc). However. 4. 2 Limiter and discriminator canceled video out 30-Mc delay line Postdelay amplifier 30 Mc Mixer No. \ 1 75 Mc 45-Mc oscillator Fig. but it requires high driving voltages for the phase detector if the system is to be broadband. Generation of the Pulse Repetition Frequency. perfect cancellation cannot be expected from the delay-line canceler. The two-mixer cancellation can be used directly in double-cancellation MTI.

0. Or alternatively.000 for a 1 °C change. This technique was used in several early MTI systems. so D3 + D z ^ D (4. Postdelay amplifier Def.5 per cent in fused quartz. 3 ' . D4 Trigger D3 generator ' —<— Di Pulse amplifier .D l 1 2 The delays in the trigger pulse amplifier D 3 can be made small. The velocity of sound waves in mercury will change by 1 part in 3. 32.28) For perfect cancellation. The delay time of a typical trigger generator such as a blocking t Special glass delay lines 1 can be obtained with temperature coefficients of time delay less than ppm/°C.j If a maximum temperature variation of 50° were anticipated. In the former method a stable oscillator generates the prf and the delay line is made of variable length so that it may be adjusted to match the pulse-repetition period. variable-length delay line may be constructed using a straight cylindrical tube filled with liquid and having a telescoping section. the difference in the time delays in the delayed and undelayed (4. 4.2 always easy to achieve a stable delay line.300 for a change in temperature of 1 °C. or T= D + D . J Fig. Use of the delay line to establish the prf. The pulse-repetition period and the delay time may be maintained equal to one another by either adjusting the length of the delay line to the pulse-repetition interval or by using the delay line to generate the prf. 4. 4.5 per cent in mercury and 0. and D 3 plus the delay D2 1 will usually be of the order of Z> 1. a condition which cannot be met in practice.30) The result is that the trigger-generator delay D4 must be zero. The rate at which the trigger pulse circulates about the delay line determines the prf. Delay line Bipolar video *1 Subtracter Canceled bipolar video Modulator f Direct channel 0% Det. . a single line may be used for both functions. The pulse-repetition period Tis determined by the total delay time around the loop which contains the delay line D lt pulse amplifier Z>3 and trigger generator Z> or 4 The delay trigger pulse . may be used to generate separately the prf and to provide the delay for MTI cancellations. One reason the delay time might change is that the velocity of propagation of the delay medium is a function of the temperature. A line itself may establish the pulse repetition frequency by circulating a around the delay line as illustrated by the block diagram of Fig. T = D + D3 + Dt t (4. Time delays introduced by the circuitry (other than the delay line) in the delayed channel of the canceler must be balanced with equal time delay in the undelayed channel. within the same environment.32.29) channel must also equal T. Two separate delay lines adjacent to one another enclosed. the total delay time would vary by 1. The total time delay may also be influenced by the rest of the circuitry involved in the delay-line canceler.138 Introduction to Radar Systems [Sec. The velocity change in fused quartz is 1 part in 1 0.

Use of generator. The position of the reflector may be made adjustable so as to vary the compensating delay. This . analogous to the half-silvered mirror of optics. Modulator »- Half-silvered" n. 4. and it must be designed so that. 4. Trigger generator —< Pulse amplifier Bipolar video Acoustic delay line > Postdelay amplifier Defector —*— Electrical delay line 1 Canceled bipolar video Modulator Direct chonnel Subtracfor Detector 1 Fig. PULSE-DOPPLER RADAR 139 One technique for compensating for the time delay in the trigger generator is to increase the time delay D with a length of 1 electrical delay line as shown in Fig. A blanking gate of duration T/2 may also be applied to the pulse amplifier for additional reliability. prf. The pulse repetition frequency can be generated with a second acoustic delay line similar to that used for cancellation except that the prf delay line should be slightly shorter (on the order of a microsecond) to compensate for the delay Z> in the trigger 4 generator. 4.34. Usually the recovery time constant of the trigger generator is made at least one-half the pulse-repetition period (Tj2). ' . Conceled bipolar video mirror Fig. short piece of electrical delay line to compensate for the time delay of the trieeer ' 5& additional delay in D 1 may also be obtained acoustically by adding to the delay second receiving crystal at the appropriate distance from the output end of the line (Fig. is placed in the mercury delay line just before the normal output crystal to reflect a portion of the acoustic energy into a second output crystal.33. 4. 4.2] oscillator MTI AND might be of the order of 0. The trigger generator must be self-starting.33. When the same delay line is used for both MTI cancellation and for generating the The line a some means must be had for distinguishing the trigger pulse from the echo signal. The electrical delay line may operate at either the carrier frequency or the video frequency. it will not trigger additional pulses until the next pulse interval. but it is usually inserted in the video because video delay lines are easier to achieve than carrier-frequency lines.34). PRF-*--. The use of an electrical delay line is a simple "Solution to the problem. Use of half-silvered mirror in mercury acoustic line to compensate for the additional time delay in the trigger generator. PRF- Trigger generator Pulse amplifier Acoustic delay lelay line Bipolar video Is 3 ' I Postdeloy amplifier Def. Another technique is to use a different carrier frequency for the trigger pulse than that used for the echo signal. once fired. The discrimination may be made on the basis of time selection or amplitude selection. 1 ^sec.Sec. This is made possible by the large bandwidth of delay lines. Subtractor Det. A 45° reflector. The two delay lines should be operated side by side so that any changes which take place in the thermal environment will affect both lines equally.

a residue at the output of the delay-line canceler may result from instabilities in the transmitting or receiving equipment or from changes in amplitude from pulse to pulse due to the shape of the antenna pattern. If l)st pulse fed into the line. 4. clutter echoes are not always stationary. Still another technique for equalizing the MTI delay time and the pulse-repetition period makes use of both the stable prf oscillator and a circulating signal in the delay This method has been called electronic frequency tracking. J Fig. . 4. The coincidence circuit compares the time of occurstable oscillator must be tunable.3. video > — Modulator -»- — ' Delay line Postdeloy amplifier —— Det. The advantages of this technique are that any drift in the delays of the cancellation loop are automatically compensated and there are no additional mechanical Its disadvantages are that the prf oscillator must remain stable with no jitter. Subclutter Visibility In the discussion of the delay-line canceler it was assumed that the echo signal from and did not vary in either amplitude or phase from pulse to pulse. the prf and the nth pulse (delayed) and the (n If the two pulses are not coincident in time. time difference is converted to a voltage which is used as an error signal to change the + + PRF t Turable. 4. In practice. a relatively large number of tubes or transistors is required as compared with other techniques. visibility of. rence of the nth pulse from the delay line with the (n l)st pulse (undelayed) are coincident. Two radars with the same subclutter visibility might not have the same ability to detect targets in clutter if the resolution cell of one (pulse width times beamwidth) is greater than the other and accepts a greater clutter . and parts. 4. This uncanceled residue might be mistaken for a moving-target signal. 30 db implies that a moving target can be detected in the presence of clutter even though the clutter echo power is 1 .3 arrangement straightforward. Generation of the prf with frequency tracking.35. In addition to the internal motion of clutter. the the MTI delay line are in synchronism. 1 Subtractor Canceled ^ — bipolar video *- Direct channel Det. caution should be exercised in applying it to describe the relative performance of two different MTI radars. which is subclutter defined as the gain in signal-to-clutter power ratio produced by the MTI. but it requires increased space and weight because of the necessity of two lines.000 times the target echo power. The line (Fig.35). . frequency of the prf oscillator in a direction which will bring the two pulses in coincidence. for example. Although the subclutter visibility is widely used as a measure of MTI radar stationary clutter was fixed A A performance. they may be in motion so as to produce an uncanceled residue at the output of the delay-line canceler. however. measure of the performance of an MTI radar is the subclutter visibility. stable PRF oscillator I Coincidence circuit .— ' 140 Introduction to Radar Systems is [Sec.

It is a common type of MTI radar and consists of a pulsed oscillator transmitter. <f> — <j> . does not equal the pulse-repetition interval if there are variations in pulse width or if the transmitter frequency changes during the pulse.) (4. given by Coho signal at time of phase lock . 4. 4. Coho in the signal to phase detector = V c sin (a> c t + <f>t — <t>s) (4. It will oscillator as the transmitter (Fig. = V c sin ((o CQ t + <f) t — ff> s ) (4. but one starts with greater clutter power because its resolution cell "sees" more clutter targets. the coho oscillator might drift in frequency from its initial value to a value co c Therefore the coho signal which is fed to the phase detector at the time of the first echo pulse is . not in clutter.1 1). 4. fixed target at range T = 2R/c after transmission. the MTI radar considered in this section is that illustrated by the block diagram of Fig. pulse amplitude. that is. is defined as the ratio of the (voltage) signal strength from a target traveling at a specified radial velocity to the signal strength from the same target when it is traveling at an is optimum radial velocity. Both the subclutter visibility and the cancellation ratio are usually expressed in decibels. such as the magnetron.3] MTI and Pulse-doppler Radar 141 signal power. or locking. The performance of MTI radar will deteriorate if the transmitter. The target-visibility factor.32) and (4. pulse which locks the phase of the coho to that of the transmitter. or coho frequencies will result in an uncanceled residue from fixed targets at the output of the delay-line canceler. which was mentioned earlier in this chapter (Fig. and a delay-line canceler to extract the doppler information from moving targets. the frequencies of each of . unless otherwise noted. if the time delay in the delay The effect of equipment stability on upon the particular configuration of the radar. If the transmitted angular frequency at the time a particular pulse is transmitted is w t and the phase is 4> the transmitted signal may be written as line .33) During the time T when the pulse travels to the target and back. The coho signal at time of phase lock is of frequency co c0 but with phase t s and is is the phase of the stalo.6).6. a portion .34). (o t =w + s (o c and the phase-detector output is a signal of constant amplitude sin co f T . or the echo drift in frequency. 22 Oscillator Stability. It may be denned as the ratio of a fixed-target signal voltage after MTI cancellation to the voltage of the same target without MTI cancellation. It is also possible for the MTI performance to degrade because of such seemingly unimportant things as vibrations caused by the blowers used to cool the stalo. Transmitted signal =V t sin (to t t + </>. signal from a. A change in either the transmitter.34) The output of the phase detector is a sine wave with argument equal arguments of Eqs.Sec. The stalo frequency at the time w and the IF echo signal at the input to the phase detector is IF echo signal where <f> s R arrives back at the radar receiver a time the echo is received is = KIF sin [(a> t — co s)t — co t T + <f> t — <f> s ] (4. or t. The cancellation ratio is sometimes used to describe the performance of the delay-linecanceler network. (4. stalo.35) In a perfect system. Equipment Stability. The cancellation ratio is a number less than 1.32) be recalled that in the MTI which uses an of the transmitted power provides the reference. the stalo. both radars might reduce the clutter power equally. Therefore Phase-detector output for pulse 1 to the difference = k sin [(co t — m — w )t — s c a> t T ] (4. The target-visibility factor applies only when the target MTI performance depends For purposes of discussion. 4.31) The echo s. On the succeeding pulse a time Hater.

If the oscillator frequencies are not the same from pulse to pulse. so that the ouput from the phase detector may be written as Phase-detector output for pulse 2 = k sin [(co' t — w' s — co' c )(t + T) — co' T^\ t (4.39) taken since it is assumed that a full-wave rectifier follows the In arriving at the above expression it was assumed that sin (Aco t/2) (** Aco t/2 and sin (Aco if t/2) s=a Aco if t/2. .40) where 2-nf co. + a/ )(t + T+ i) . Therefore the absolute value of the cosine factor can take on any value from to 1 depending upon the value of T Its average value is 2/7T.41) . The absolute value CR = Aco if (t + y . co IF = co t — w s It will be assumed that the coho frequency is not exactly tuned to the intermediate frequency. (4. then a>' = w t + Aco t a>' s — <t) c + A(o c The difference between the transmitter and stalo angular frequencies (o' c is the IF angular frequency. Although the argument of the sine factor is small. The output of the delay-line canceler is the difference between Eqs. .. Performing the integrations and a trigonometry.38) Aw IF Aco t Aco s Aw. to give the cancellation ratio (CR): — (co + Aco )T = T to = T + t t ] (4. 4. t . which are good assumptions for most radar applications. and it is likewise assumed that the sine may be replaced by its argument.39) is also a small quantity.36) where the primes denote the shifted values of oscillator frequencies.142 Introduction to Radar Systems [Sec.Aco )(t + / s c T) t Equation (4. stalo. (4. and coho frequencies from one pulse period to the next is denoted = a> s + Aco s and by Acoj. The difference will be denoted .35) and (4.3 the oscillators are assumed to have changed. For small errors.a/«(t+ ^) .37) t r and CR = - I CTo + r sin (Aco IF ( —mT t 1 fT +r sin ) dt tJTo where Aco tJt [Aw (r + T) — (m + Aa> )T„]\dt < t t . .A/i F T c (4. (4.A Wo (r +T+^j + Aco T t = 4 (A/. take T equal to its = average value T/2. .36). Acys and Aco c respectively. + A/ ) - A/ 4 - A/IF T | (4. . In most cases the pulse width is small compared with the pulse period and t/2 may be neglected with respect to T. If the change in transmitter. The cancellation ratio c is CR = 4 |1. ..37) is averaged over the pulse width r from divided by k.To t X cos J is "iF^o + y +A Wo (t + T + l) . Replacing the rapidly varying cosine term by its average value gives the following: delay-line canceler. the argument of the sine factor in Eq. the cancellation ratio becomes = + — — little CR = 2 sin i ^) + Aco. the argument of the cosine factor can be a relatively large quantity because the term 2co t T is large. Aco IF = a> t —w — s co c .5(A/. a stationary target will produce a nonzero output equal to Output of delay-line canceler = k sin (Aw IF — co^o) — k sin [(A«IF + Aeoj — Aw .Aco TQ - 2co t T (4. For purposes of illustration.

the stalo and the coho.48) w cl = stalo and coho frequencies on transmission of second pulse cl m' sl and <x>' k = = stalo and coho frequencies on reception of second pulse a constant .<)(t + T) + («j el . c <f> s <f> e ] V = + w )t + (4. the cancellation ratio is equal to 6 A/s r.t + <£ c ] (4. a time T later. If. The stability of the coho will be better than that of the stalo since it operates at a much lower frequency. may be written fl> V2 where = k sin a> sl and [Ki .5) leads to [(6<r s) 2 + (6er c ) 2 + (4a t f + (4crIP 2 ) ]ir (4. (neglecting amplitude factors) t is + The received RF signal from a target at a distance R = cT /2 V = sin + a>X* ~ W + & + sin [(eo. Similar considerations 5 cps pulse to pulse and less than 17 cps for a apply to the coho frequency drift.41) are not all of the same sign and the direction of the frequency drift can be either positive or negative. <f>c\ (4-44) On reception the stalo m' and s a>' . and a mistuned drifts in effect of frequency combined The coho frequency is difficult to predict because the terms of Eq. Only two oscillators are involved.45) on reception Vc = sin («# + 4> c ) (4. stalo. (4. transmitter. while the stability of an Z. For CR 20-db ratio. but the residue is equal to 4 &ft T.45) and (4.46) is Hence the output of the phase detector for the first echo pulse obtained by heterodyning Eqs. (4. the cancellation ratio were required to be —40 db and if the pulse 10~ 3 sec).47) The output of the phase detector for the next pulse.3] MTI and Pulse-doppler Radar 143 If all frequency drifts are zero except the stalo.66 cps. the drift must be less than than 1. It is further assumed that the power amplifier has The transmitted signal negligible effect on the phase of the transmitted signal.42) A similar analysis of the stability requirements for the MTI using a power amplifier an interesting result for the case of uniform frequency drifts. = — CRrms = (Fig. 4. Short-term stability of a commercial S-band stalo might be of the order of 7 to 10 cps.46) and is Pi the difference signal = k sin [(co s - w'. t. but with more sophisticated 8 cps.)t + (co c - m' c )t - (ca s + m^T^ (4.000 cps to pulse for example. the stalo frequency shift from pulse (T = — 30 db. If it is assumed that the drifts follow a Gaussian distribution about some mean value and if the standard deviations are designated a t where i stands for s. coho.(o'cl )(t + T) . usually at 30 or 60 Mc. c.-band stalo might be 4 to 23-24 The stability could be improved if desired. 4.41) it is seen that if the only frequency drift is that of the transmitter. The transmitted frequency is assumed to be the sum of the two. repetition frequency were 1.{w A + eI )r ] (4. equipment than a simple stalo oscillator. (4. If there were no frequency drifts in the oscillators. From Eq. then the rms cancellation ratio may be given by must be less . the cancellation ratio would be 4 A/" IF r. c and coho frequencies are assumed is to have drifted to new values The IF signal KIF The coho signal = sin [(co s is + w )(t ~ e To) - a>'.Sec. if the coho were mistuned. or IF.43) is r [(a..

£ "• . an uncanceled residue will result from the output of the delay-line The output of the phase detector for a fixed target (f canceler. Accuracy of Delay Time andPRF. c w.52) . although small. 0) is. or si - »(£ ^ + £ r .l - s = —T— T At Aft> S7.3). 4. m c Am ffl. Nevertheless.D 144 If it is Introduction to Radar Systems assumed that the seen that 0) s [Sec.0 otherwise where k F{t) = a constant = shape of echo signal (assumed rectangular) R = range f = transmitted frequency c = velocity of propagation t If the difference is assumed to be a pulse of constant amplitude and of width t. the output of the delay-line canceler will be F(t) V The average of F(t cancellation ratio is = k[F(t + At) - F(()] sin^^ c (4. f or CR = — T (4. c = —7— T At is The output of the * delay-line canceler V2 — V± . the drift amplifier. and ft>cl- o). the above crude analysis indicates a possible advantage of the power amplifier MTI as compared with the power oscillator MTI. (4. If the delay time and the pulse-repetition period are not exactly equal.[£ - r *(' + Thus a uniform frequency amplifier MTI. between the prf and the delay time is At.3 stalo and the coho drift at a uniform rate Aco s Am J At and Am J At.5i) + At) F(t) over a time interval t At is 2 At jr. d may may = V= 4 kF(t) sin ^^ c c c F(t) = l (4-50) 1. then it is — . In practice. The average the average value of V divided by the amplitude of the input signal — + k sin (ATrR f jc).£ ") . from Eq. drift seems to result in perfect cancellation for the power not be uniform and the phase variation introduced by the not always be neglected.

37). If the individual scatterers remain fixed from pulse to pulse. Although rectangular pulse shapes were assumed.3] MTI AND PULSE-DOPPLER RADAR 145 if the pulse width were 1 ^sec. Internal Fluctuation of Clutter. The echo at the radar receiver is the vector sum of the echo signals received from each of the individual scatterers (Fig.05 ^sec for —20 db. Likewise the stability of the oscillator which generates the prf must be held to within 1 part in 2 x 10 5 to obtain a —40-db cancellation ratio. rain. Vector summation of the contributions from the many independent scatterers constituting the clutter. must be less than 0.36. Echoes from trees. 0. that the number of individual scatterers making up the composite clutter signal is large. 4. Radar resolution cell (in angle and range) and clutter model consisting of many Fig. However. 7 — A Fig. Although clutter targets such as buildings. the difference between the prf and the delay time. vegetation. there are many types of clutter that cannot be considered as absolutely stationary.015 ^sec for —30 db. the relative phase as well as the amplitude from each scatterer influences the resultant composite signal. bare hills. At. ib radar. and that the radar cross section of any individual scatterer is . water or mountains produce echo signals that are constant in both phase and amplitude as a function of time.36). Because of its varied nature. sea. similar derivation can be made to determine the tolerance permitted in the pulse width. and these fluctuations can limit the performance of MTI towers. 4.005 /^sec for a cancellation ratio of —40 db. Hence the phase and amplitude of the new resultant echo signal will differ pulse to pulse. But any motion of the scatterers relative to the radar will result in different phase relationships at the radar receiver. for purposes of analysis.000-^sec delay line must not fluctuate more than 0.Sec. independent scatterers randomly distributed.37. If it can be assumed that the relative phases of the echo signals received from the individual scatterers are random. 4. 4. temperature of 20°C. 4. it can be shown that the residue area is independent of the pulse shape and depends only upon the maximum amplitude of the pulse. Therefore the temperature of a 1.05°C if a 40-db cancellation ratio is to be achieved. The temperature coefficient for fused quartz is approximately 10~ 4 part per degree centigrade at a frequency of 10 Mc and a For example. the resultant echo signal will also remain fixed. it is difficult to describe precisely the clutter echo signal. that is. and chaff fluctuate with time. most fluctuating clutter targets may be represented by a model consisting of many independent scatterers located within the resolution cell of the radar (Fig. and 0.

54) Power spectra of various clutter targets. Since the time waveform and the frequency spectrum are related by .) tributed target more discussion of the properties of clutter will be found in Chap.146 Introduction to Radar Systems [Sec. (3) sea echo. 4. This is called the clutter attenuation (abbreviated CA). (From Barlow.1 - A 0.5 echo power w is ) \4 0. function for a target which can be rep- 0.01 i i 1 ' 1 \ 1 10 15 Frequency.3 small compared with the total cross section. windy day (a =-. 4.05 i 2 which the fluctuations take place.02 - resented as one large reflector together with other small reflectors is 26 0. (2) sparsely wooded hills. The Rayleigh I.53) where w \ \5 is is the average power. An example might be trees on the side of a mountain.8 x 10 13 ).38.000 Mc. and the average power of the residue is a measure of the degree to which the clutter signal is attenuated by the delay-line canceler.38 are given in the caption. The experimentally measured power spectra of clutter signals may be approximated by W(f) = 2 \g(f)\ = 2 |g | exp [-({)*] (4.38.1. Clutter fluctuations give rise to an uncanceled output from the delay-line canceler which may be calculated in a manner similar to that described for equipment fluctuations.3 x 10"). cps Fig.U f-^l ! 1 1 probability-density function p{w) for the fluctuations in the clutter 0. while the echo from about the trees contributes the fluctuating The probability distribution portion. In addition to the fluctuating com- ponent of the clutter echo signal. 20 25 P(w) — exp — = 4w ex P i 2vv J ) w > (4. calm day (a = 3.9 x 10 "). (1) Heavily wooded hills. 12.55) where W(f) gif) = clutter-power spectrum as a function of frequency = Fourier transform of input waveform (clutter echo) f = radar carrier frequency a = a parameter dependent upon clutter Values of the parameter a which correspond to the clutter spectra in Fig. Examples of the power spectra of typical clutter are shown in Fig.41 x 10 16 ). These data 27 apply at a frequency of 1.2 p(w) = — exp (. (4) rain clouds (a = 2. that clutter IRE. Further (5) chaff (a = 1 x 10 16 ). The difference between the delayed and the undelayed waveforms is averaged.— w > (4. 4. 4. 1 distinguishes targets is Another property of clutter which it from normal radar is usually a diswhile aircraft are normally point targets. there usually a constant (d-c) component ~ 0. 20-mph wind blowing (a = 2. then the probability-density function for the envelope of the fluctuating echo signal may be represented by Rayleigh statistics. The echo from the mountain constitutes the constant portion. 27 Proc.

4.59) and (4.&\ = * 2/ sin ^-\ ^exp (. (4.Sec. 4. 4. The equivalent expression for the clutter attenuation provided by a double-delay-line canceler is 29 CA = 3 - — 4 exp [-(77/ //r ) 2/a] + exp [-4(7.39 28 for several values of the ratio f /fr and as a function of the parameter a.55) for the input g x (f) into the above gives f°° exp [-«(///„)*] df CA °° Jo 4 f Jo sin (*///. when multiplied by the frequencyresponse function of the delay-line canceler. CA) may be written as the ratio of the input power divided by the output 2 power \gi(f)?df \ gl (f)\ df (4-58) CA = T^ 2 \g2(f)\ df = -J^ 4 \ Jo gl (f)\* sin* (nf/fr )df Jo Substituting Eq. The frequencyresponse function of the delay line is exp (—j(o/f r).4. it is also possible to compute the clutter attenuation using the frequency spectrum instead of the time waveform. the clutter attenuation may be 1 written as CA = _ ^ exp — [-(nfJfM (4./ //r ) 2/«] (4.38.60) Equation (4.56) Jr Jr The Fourier transform of the input waveform gx (f). gzif) = gi(f)H(f) (4-57) The attenuation of abbreviated the clutter signal by the delay-line canceler (clutter attenuation. If the exponent in the denominator of Eq. (4. or CA *» — Wolf? (4. where f r is the pulse repetition frequency and is equal to the reciprocal of the time delay of the line.62) The frequency dependence of the clutter spectrum as given by Eqs.) 2 exp [-a(/// ) a ] df Integrating.59) was derived for a single-delay-line canceler.59) is small compared with unity.j Jr ^ ' (4. The delay-line canceler was shown in its barest essentials in Fig. yields the Fourier transform of the output waveform g2 (f). CA ^ 12^/o/X) 4 (4. The frequencyresponse function of the delay-line canceler is therefore H(f) =1- exp (.3] MTI and Pulse-doppler Radar 147 the Fourier transform. (4.59) Equation (4. the exponential term can be replaced by the first two terms of a series expansion with little loss of accuracy. 4. Also indicated on the figure are the values of a which correspond to the experimental data of Fig.61) For (nf lfryia< 1.59) is plotted in Fig.61) cannot be extended over too great a frequency range since the derivation does not take .

39. . into account frequency.63) because of the two-way It is further assumed in this analysis that the antenna elevation angle is zero. an uncanceled residue would result at the output of the delay. 4. in a radar carried on board a moving vehicle and will be considered later in Sec. ground-based radar. the echo signal will consist of the vector sum of the contributions from the many independent scatterers randomly distributed Scanning Fluctuations.8. may have considerably different reflecting properties at a wavelength of 1 cm. power gain as a function of the angle 8 (two-way enters as the square in Eq. (4. The 10 10' 10' 7 10' 8 a. the received echo power is equal to the sum of the average power scattered by each of the objects. to'' to 20 Clutter spectrum exponent Fig. (From Grisetti et al. the echoes within the radar resolution cell. the clutter echo to be the vector sum of from a number of independent scatterers included within the radar resolution cell. The gain transit of radar signals. Antenna motion may be due to a rotation and/or a translational movement of the radar platform. Trans. Effect of internal fluctuations on clutter attenuation. 4.) Assume. The power received from a stationary.63) the one-way antenna voltage gain). Fluctuations caused by motion of the radar platform are not present in a fixed.line canceler if the antenna is in motion. where the dimensions are comparable with the wavelength. 2S IRE.148 Introduction to Radar Systems [Sec. which is long compared with the dimensions.3 any variation of cross section of the individual scatterers as a function of leaves and branches of trees. 4.. Only the fluctuations in the echo caused by the rotational scanning of the antenna (scanning fluctuations) will be considered here. 50 cm. distributed clutter target will be proportional to J where G(6) is G 2(0) dd (4. Because the scatterers are considered to be independent. for example. Assuming the usual idealized clutter target. from those at a wavelength of. They are of importance. however. say. as before. Even if all the individual scatterers were fixed so that there were no internal clutter motion.

x (7(0) ~ =G exp / ^ 2.64) where A0 is the angular motion of the antenna between pulses is . 28 The residue left after cancellation may be divided into an amplitude component and a phase component. the oo since the angle oo to limits of integration do not extend from does not extend beyond 2-n. if an amplitude detector is used (as in the noncoherent radar described in Sec. (4.65) and evaluating the integrals gives where CA = -2S2.7760 2 \ gg-j is measured from the axis of the beam. however. are no longer illuminated. (4. and G is the maximum antenna gain. while others enter the beam and become illuminated. integration. which In the above the voltage difference G(6 follows from the definition of the derivative when A0 is small. By converting phase . Most scatterers remain within the beam. if R r Therefore. The rms value of each of these . the phase fluctuations are eliminated. two components is rjVl. In fact. leaving only the amplitude fluctuations. The clutter attenuation in this case is twice that of Eq.65) were chosen.776 (4. The vector r can be resolved into two components. If the amplitude fluctuations are . Let R be the rms value of the signal voltage R. (4. however. Substituting this into Eq. > eliminated with a limiter. one in the direction of R. while the rms phase fluctuation is approximately r /(R V2). only the phase fluctuations remain. Strictly speaking. [G(0 /: + A0) - G(0)] 2 dd (4.3] MTI and Pulse-doppler Radar 149 The scanning motion of the antenna causes the beam to shift to a slightly different azimuth on each pulse.Sec. The antenna pattern is assumed to be of Gaussian shape + — — + . for ease of vicinity of the main beam. 4. and let r be the rms value of the voltage residue r which remains after cancellation.6). Thus the rms amplitude fluctuation of the residue is r /V2.66) where nB = 0#/A0 is the number of hits included within the 3-db beamwidth B There is very little difference in the results if instead of a Gaussian antenna pattern.v/. 4. The clutter attenuation of Eq. B is the beamwidth included within the half-power points of the antenna pattern. the other at the right angles to R. a pattern of the form (sin 0)/0 is chosen.65 ) [G'(6)fde J — 00 Afl) G(d) was replaced by G'(0) A0. Some scatterers.65) is then Rl/r^. The result is that the total number of illuminated scatterers will be essentially the same from pulse to pulse but their relative distribution in space and their relative phase The resultant echo-signal voltage therefore varies from relations will be different.radians.66). The clutter attenuation then CA = 7^ J — oo — [G(0 f 00 /*0O G\6)d6 G 2 (6)d6 ~ 2 G(0)] dd + A0) - ~^t (A0) 2 ( 4 . pulse to pulse. and an uncanceled residue remains at the output of the delay-line The uncanceled residue from two successive pulses is proportional to canceler. the angular region of interest is only that region in the The limits in Eq. Integration over the entire range of values does not appreciably affect the final result if narrow beamwidths and reasonably low antenna sidelobe levels are assumed. (4.

or [Sec. The antenna is then shifted rapidly to the next angular position. 4.40. Even and odd antenna patterns may be obtained with two-feed - . is [G'(6 + A0) - G'(6)] A0 & G"(6)(A6f.65) will be small (ideally zero). Another technique for reducing the effects of scanning is based on radiating an antenna pattern which maximizes the clutter attenuation. This is also plotted in Fig. The denominator of Eq. where it again remains stationary during the observation time. Since the usual antenna pattern G(0) is an even function of 0.150 Introduction to Radar Systems twice that of Eq. This is called step scanning. (4. the clutter attenuation is CA 1. If either Antenna scanning fluctuations may be eliminated by holding the beam stationary at each angular sector for a period of time sufficient to obtain the number of pulses required for detection.67) A plot of this equation is shown in Fig.66). 4.388 (4. the derivative pattern G'(d) will be an odd function.3 fluctuations to amplitude fluctuations with the phase detector (as in the coherent MTI). The clutter attenuation may canceler through a second delay-line canceler (double cancellation). the clutter attenuation will be twice this value. 4. 5 10 of hits number 100 200 included within 3 db beamwidth 50 500 1. the residue will be considerably reduced. Clutter attenuation with a scanning antenna for single cancellation and double cancellation.000 Fig.40. 4.5. 70 be improved by passing the output of the delay-line The output of 60 -q50 C o 140 3 C <D °30 o20 10 ' I i I I i il J 20 I I I I l I I I I I II II 2 rig.40. and thus the clutter attenuation will be large. which results in the CA r (A0) 4 J— \ G 2 (0) dd (4. just as with single cancellation. Antenna pattern assumed the second canceler following: to be of Gaussian shape. 28 30 If an antenna pattern proportional to G'(6) is added to the normal antenna pattern G(6) before cancellation with G(6 + A0). (4. For a Gaussian beam shape the clutter attenuation is k^/11.68) [G"(d)J dd CO an amplitude or a phase detector is employed in the receiver.

MTI Using Range Gates and Filters In the previous discussion.Sec. including the nature of the clutter illuminated by the radar beam. Complete stabilization of the radiation patterns during the cancellation periods eliminates the residue caused by scanning. This effect is known as clutter foldover. but it still leaves much to be desired. output from each gate may be applied to a narrowband filter since the pulse shape need no longer be preserved for range resolution. as was seen in described. which maximizes the signal-toUnfortunately. antennas and proper antennas. In a pulse radar the clutter spectrum is reflected about each of the spectral lines (Fig. from the other range intervals which do not contain the target signal. The loss of the range information and the collapsing loss may be eliminated by first quantizing the range (time) into small intervals. The delay-line canceler is one form of filter used to approach the matched-filter The single-delay-line canceler is a poor approximation to the ideal characteristic. the additional complexity required to achieve special filter characteristics and the need for maintaining perfect adjustment in the delay lines if the theoretical performance is to be achieved in practice. economic or space considerations limit the degree to which the delay-line canceler can be made to approach the ideal matched filter. The video clutter spectrum. 9. resolution in an ordinary system because the duration of its impulse response is approximately the reciprocal of the bandwidth. 4. The ideal-filter characteristic would reject the maximum amount of clutter energy without significantly rejecting any doppler signals which fall outside the clutter spectrum. filter. their difference. Double cancellation is somewhat better than the single delay line. as in simultaneous lobing or monopulse The two feeds are displaced from the antenna axis to produce two adjacent beams. Range resolution is Once the radar return is quantized into range intervals. and the equipment stability. 4. independent of the antenna motion. the signal-to-noise ratio is reduced when narrowband filters are used without range gating because of the collapsing The collapsing loss is caused by additional noise that enters the filter loss (Sec. Although a simple narrowband filter might be used in an MTI radar to pass the doppler-frequency components of moving targets and reject the direct current due to A narrowband filter destroys the range clutter. synthesize and some practical compromise must usually be made. but it has the advantage of simplicity. the matched filter is sometimes difficult to noise ratio at the output. an odd pattern. but they are usually of the order of the pulse width. Furthermore. the MTI radar was assumed to use a delay-line canceler as the filter which rejects clutter echoes and passes only those doppler-frequency-shifted In this section a different filtering technique for signals returned from moving targets. is spread over a finite frequency range. The sum of the two gives an even pattern. 4.12). . An advantage of the delay-line canceler Its chief limitations are as a filter is that range information is preserved in the output.38. They may be designed to have a wide Hence only variety of frequency characteristics as described in a previous section. frequency spectrum depend upon several factors. The ideal filter is the matched filter (Sec.22) located at the pulse repetition frequency and its harmonics. it suffers from two major limitations. Delay-line cancelers are not limited to sin x or sin2 x characteristics. the antenna scanning. This process is called range gating. 2. be will echoes rejecting clutter The shape and magnitude of the Fig.2). Proper combination of the even and odd antenna patterns stabilizes the far-field amplitude and phase radiation patterns in space. 4. The width of the range gates depends upon the range accuracy desired and the complexity which can be tolerated. A collapsing loss does not take place since noise from the other range intervals is excluded. the established by gating.4] MTI and Pulse-doppler Radar 151 RF combining circuitry.4.

not only because of better clutter rejection. The output of the phase detector is sampled sequentially by the range gates. An echo from a moving target produces a series of pulses which vary in amplitude according to Range gate No. Following the doppler filter is a full-wave linear detector and an integrator (a lowfilter). The shape of the rejection band is determined primarily by the shape of the bandpass filter of Fig. 5. 4. depending on the characteristics of the clutter spectrum. Block diagram of MTI radar using range gates and filters. Following the threshold detector.— — 152 Introduction to Radar Systems [Sec. 4. but also because the threshold device eliminates many of the unwanted false alarms due to noise. constant circuit elements. the proper time. 3 " " „ To dato processing or display (Range L ->-j gate -*— Boxcar generator [No. The clutter-rejection filter is a bandpass filter whose bandwidth depends upon the extent of the clutter spectrum but is less than The doppler filters utilize lumpedfj2. whose purpose is to aid in the filtering and detection process by emphasizing the fundamental of the modulation frequency and eliminating harmonics of the pulse repetition frequency (Sec.41. 4. the doppler frequency. Fig. 4. pass . The range gate acts as a switch or a gate which opens and The range gates are activated once each pulse-repetition closes at interval.3). Only those signals which cross the threshold are reported as targets. t Boxcar generator Bandpass (Ooppler) filter Full-wave Low pass -> filter > - linear Threshold -* defector (integrator) Range gate No. 4. -go £=Vr Frequency Zfr 3/.42.41.41. 4. the outputs from each of the range channels must be properly combined for display on the PPI or A-scope or for any other appropriate indicating or data-processing device.42. Each range gate opens in sequence just long enough to sample the voltage of the video waveform corresponding to a different range clutter-rejection niters interval in space.4 A block diagram of the video of an MTI radar with multiple range gates followed by is shown in Fig. Frequency-response characteristic of an MTI using range gates and filters. The output of the range gates is stretched in a circuit called the boxcar generator. The frequency-response characteristic of the range-gated MTI might appear as in Fig. The CRT display from this type of MTI radar appears "cleaner" than the display from a normal MTI radar. where f r is the pulse repetition frequency. The output of the integrator is applied to a threshold-detection circuit. The lower cutoff frequency can be designed to be adjusted to different values. 2 — Boxcar generator -> Threshold Phase detector Range gate No.n Threshold Fig. The purpose of the detector is to convert the bipolar video to unipolar video. The output for a stationary target is a series of pulses of constant amplitude.

target of knowledge the ambiguities not only confuses ranges and occur when the blind called are These detected. CW CW CW CW advantage over as. the This is possible if the doppler frequency shift is at least suitable RF or IF filters. Pulse-doppler Radar 31 ~ 33 The pulse-doppler radar is a form of MTI radar usually. a this of However. one which of doppler frerange expected the within speeds and which has many blind canceler quencies. operating with high prf's to avoid doppler blind The presence of range speeds. that is. 4. are not present. for example. for purposes generally agreed upon. also because limitations the clutter filter characteristic and the clutter spectrum. may be limited in measure range under these conditions.5] MTI AND Pulse-doppler Radar 153 MTI radar using range gates and niters is usually more complex than an MTI with a The additional complexity is justified in those applications filter MTI are where good MTI performance and the flexibility of the range gates and between match better the from only not results performance The better MTI desired single-delay-line canceler. One other method should be mentioned of achieving coherent MTI. .5. such as maintaining the time delay of temperature changes. ability to The pulse-doppler radar. the performance of the pulse-doppler radar approaches that The pulse-doppler radar. but also creates intervals range. On the other hand. characteristics These attributes. 3. the of coverage reduce the range ambiguities can be resolved by transmitting at more than one prf. blind and ranges blind both range are ambiguous. when easily CW its doppler radar. A series of range gates and doppler 2. A klystron amplifier transmitter rather than a magnetron oscillator A relatively high pulse repetition frequency with ambiguous range but unambigu- ous doppler (no blind speeds within expected range of doppler frequencies) The above is not meant to define a pulse-doppler radar. characterized by one or more of the following: rejection niters rather than the delay-line 1. which further speeds. the other in those applications where range information is not obtained. nor should it be implied that a pulse-doppler radar. the pulse-doppler radar has CW radar receiver is always on. If the number of a single pulse is of cycles of the doppler frequency shift contained within the duration from clutter by separated be may targets moving from echoes returned sufficient. . the transmission. whereas the radar is reduced because of the blind spots in pulse-doppler the of capability detection will usually be range resulting from the high prf. When the prf must be so high that the number of range ambiguities is too large to be of the resolved. A radar with any one of these characteristics is necessarily a does not seem to be precise distinction between the MTI and the pulse-doppler radars radar will pulse-doppler text. signals or by leakage transmitter off during The pulse-doppler radar avoids this difficulty since its receiver is turned Even so. 4. and the can be ranges blind of effect The radar. may have to accept ambiguous range information. like the CW radar. but constant in spite peculiar to the delay-line canceler.Sec. in a homing missile. radar in that the detection performance is not limited by an advantage over the reflected from nearby clutter or from the radome. be will not target the where doppler and the the both If transmitter is turned on and the receiver is turned off. more complex than that of with a operate can doppler pulse the while receive and transmit for separate antennas nor the pulse-doppler radar seems to have a clear-cut Neither the single antenna. but not necessarily. there will exist alleviated. above the by characterized usually be assumed to be visibility) than is possible usually result in better MTI performance (better subclutter uses a delay-line with the type of MTI radar considered previously. Pulse-doppler-radar equipment often use two will radar that a except radar.

MTI radar which uses amplitude instead of phase fluc- tuations is called noncoherent (Fig. Amplitude fluctuations due to doppler produce a butterfly modulation similar to that in Fig. Amplitude limiting cannot be employed in the noncoherent MTI receiver. The advantage of the noncoherent MTI is its simplicity. If clutter were not present. the desired targets would not be detected. 14). A logarithmic gain characteristic not only provides protection from saturation. the transmitted pulse width is relatively wide and its spectrum is narrow. but in this case. The transmitter must be sufficiently stable over the pulse duration to prevent beats between overlapping ground clutter. The high speed of significantly greater 4.3. ThereMTI radar. 4. however.43. Except for the inclusion of means to extract the doppler amplitude component. The local oscillator of the noncoherent radar does not have to be as frequency-stable as in the coherent MTI. Noncoherent MTI The composite echo signal from a moving target and clutter fluctuates in both phase and amplitude.target detection is to take place. It is not usually applicable to aircraft targets. else the desired Fig. but this is not as severe a requirement as in the case of coherent radar. } TR Power oscillator Modulator which may The operation of this type of radar. 4.154 Introduction to Radar Systems [Sec. fore the IF amplifier must be linear. In these cases. depends upon that is a reference signal at the radar receiver coherent with the transmitter signal. The phase detector is not used since phase information is of no interest to the noncoherent radar. extraterrestrial targets results in doppler shifts that are usually than the spectral width of the transmitted signal. In these systems. the noncoherent MTI block diagram is similar to that of a conventional MTI pulse radar. Its chief limitation is that the target must be in the presence of relatively large clutter signals if moving. to provide a switch to disconnect the noncoherent MTI operation . The clutter serves the same function as does the reference signal in the coherent MTI. The detector following the IF amplifier is a conventional amplitude detector. Mixer LO It is also possible to use the amplitude fluctuations to recognize the doppler component IF amplifier produced by a moving target. 4. It is possible. 4. amplitude fluctuations are removed by the phase detector. The noncoherent MTI radar does not require an Amplitude internal coherent reference signal or a phase defector detector as does the coherent form of MTI. hence it is attractive for those applications where space and weight are limited.6 comparable with or greater than the spectral width of the transmitted signal. or if a large dynamic range is required. To cancellation circuits t. but it also tends to make the clutter fluctuations at its output more uniform with variations in the clutter input amplitude. The doppler component contained in the amplitude fluctuations may also be detected by applying the output of the amplitude detector to an A-scope. The coherent MTI and the pulse-doppler radar make use of the phase fluctuations in the echo signal to recognize the doppler component produced by a moving target. The output of the amplitude detector is followed by an MTI processor such as a delay-line canceler. Clutter echoes may not always be present over the range at which detection is desired. It has also been called externally coherent. it can be logarithmic. they ride on top of the clutter echoes. be called coherent MTI.6. but it can sometimes be applied to radars designed to detect extraterrestrial targets such as satellites or astronomical bodies (Chap. Block diagram of a noncoherent amplitude fluctuations would be lost.43).

the detection of a moving target in the presence of clutter is more difficult than if the radar were From the viewpoint of the radar the clutter appears to be in motion. as the Navy's seaward extension of the Early Warning line. The lower the altitude of the target aircraft. for example. the more likely that clutter will be present. and revert 4. change with time. the clutter rejection filters must be bandpass rather than low-pass as in the stationary case. 4.6 except for the manner in which the coho utilized. Coherent AMTI. It is quite similar to the ordinary coherent signal is MTI radar shown in Fig. different from the relative velocity between clutter and radar. The frequency of this oscillator is made to be proportional to the relative velocity between radar and clutter and may be controlled according to the position of the antenna with respect to the clutter. and stationary. irrespective of its application. The noncoherent technique is a relatively cheap form of MTI that might be used in applications where equipment simplicity is an important consideration and where only moderate MTI performance is needed. the more difficult it For instance. In principle. The closer the relative velocity of the clutter to that of the target. clutter velocity depends on the aircraft velocity and the direction of the clutter relative Since the clutter doppler frequency is not zero with a to the aircraft velocity vector. the converse may not be difficulty in seeing the ship in the presence considerable of sea clutter. MTI from a Moving Platform—AMTI When the radar itself is in motion. 4. 4.7. Another possible military application of AMTI is in long-range search radars installed in highaltitude aircraft for the purpose of detecting other aircraft. there is no clutter but there is more than one moving target. only slight difficulty (relatively speaking) in separating aircraft targets from sea clutter On the other since their doppler frequency shifts will normally be widely separated. as with a shipboard or airborne radar.44. The coherent MTI radar which was discussed in a previous section may be applied as an AMTI radar if the frequency of the coherent oscillator (coho) is shifted to compensate The block for the relative velocity of the radar platform with respect to the clutter. it is usually more difficult to achieve a good AMTI radar than it is to achieve a good MTI radar. The AMTI is of special interest as a radar technique. the target with the lowest doppler frequency can act as the reference signal and detection can take place. a radar on board a moving ship will experience will be to separate the two. The present discussion will be confined to the airborne MTI radar (commonly abbreviated AMTI). noncoherent MTI radar. However. any of the MTI techniques that have been discussed can be applied to the AMTI radar. A high-speed fighter aircraft might employ an AMTI radar during the search phase of an interception to seek out the hostile target in the presence of clutter. in In an airborne surveillance radar. in a to normal radar whenever sufficient clutter echoes are not present. since good AMTI performance is not always true. An airborne radar might experience easy to achieve and represents a challenge to the radar designer. moving radar.7] MTI and Pulse-doppler Radar 155 If. the doppler effect shifts the clutter echo signal just as any other target with the same Since the relative velocity between radar and target will usually be relative velocity. the problem is more difficult than with a stationary radar since the relative velocity of the clutter will.Sec. not all perform equally satisfactorily. the relative general. diagram of a coherent AMTI radar is shown in Fig. the clutter echo may be discriminated on the basis of doppler frequency. One of the sidebands of the heterodyned signals is selected by a narrowband filter and is used instead . hand. for their doppler frequencies may not be too different. and in general. However. called the The output of the coho is mixed with a signal from a tunable oscillator doppler-frequency oscillator.

156

Introduction to Radar Systems

[Sec. 4.7

of the coho as the reference for the phase detector. This signal is coherent with the transmitted signal but is shifted in frequency by an amount sufficient to compensate for the relative velocity of clutter. As the radar antenna beam scans in angle, the frequency of the doppler-compensation oscillator must be correspondingly changed since the relative clutter velocity changes with the direction of the antenna beam. Doppler compensation is possible if the antenna beamwidth is sufficiently small so that the patch of illuminated clutter returns an echo in which the contributions from the various scatterers constituting the
clutter

experience nearly identical doppler frequency shifts. However, when the antenna pattern is broad in elevation and the size of the illuminated patch is determined by the pulse width rather than the antenna beamwidth, the angle to the clutter
will

change as

^ V

TR
'

Power
oscillator

i

Mixer
'

Stalo

Mixer

"

IF
amplifier

Coho

Phase
detector

f c\U Sideband
filter

M
fc

V
fd

Mixer

Doppler frequency
oscillator

Control

fc-fd

\
To cancellation
circuits

Fig. 4.44. Block diagram of a coherent

AMTI

radar.

Therefore the effective doppler frequency of the ground appreciably over the range interval of interest and make the doppler compensation of the coho signal extremely difficult, limiting the usefulness of the coherent AMTI in some instances.
clutter

the pulse travels out in range.

may vary

Pulse-doppler AMTI. It was pointed out previously that the pulse-doppler radar is capable of good MTI performance; therefore, properly modified, it should also be one of the better forms of AMTI radar. The ground-clutter signal, shifted in frequency by the doppler effect, may be eliminated by a rejection filter centered at the doppler frequency in either the video or the IF. Since the clutter doppler frequency shift changes as the antenna scans, a tunable filter must automatically track the

changing

the ability of the pulse-doppler radar to eliminate clutter will be limited if the rejection filter cannot continually track the changing doppler frequency caused by a changing relative velocity. For this reason narrow pencil-beam antennas are preferred to broad fan beams. With a narrow pencil beam, changes in doppler occur as the antenna is scanned in angle, but with a broad fan bearn^ the doppler may change as the pulse sweeps across the clutter, traveling at the velocity of light. If the clutter echo changes in frequency too rapidly, a single broad clutter-

doppler.

As in the coherent AMTI,

might be used, with a resultant loss in detection capability. Noncoherent AMTI. The noncoherent MTI principle can also be applied to a radar on a moving platform. It is especially attractive for operation in aircraft, where space and weight must be kept to a minimum. The noncoherent AMTI is limited, as was its
rejection filter

Sec. 4.8]

MTI and

Pulse-doppler

Radar

157

ground-based counterpart, by the need for sufficient clutter signal to provide the reference upon which the doppler fluctuations may be detected.
4.8. Fluctuations

Caused by Platform Motion30 3436

radar were discussed, including and scanning fluctuations. in addition, there is another But AMTI radar. the performance of limit the These also serious source of fluctuation in the AMTI radar, caused by motion of the radar platform. Fluctuations due to platform motion are quite similar to antenna scanning fluctuations. In fact, scanning fluctuations are but a special case of platform motion. The patch of clutter which the radar illuminates is assumed to consist of a large number of independent scatterers randomly located within the resolution cell of the The echo signals from each of these scatterers add vectorially at the radar radar.

In Sec. 4.3, some of the limitations of the

MTI

internal fluctuations of the clutter,

equipment

instabilities,

However, if the radar beam moves between pulses, the distance to each of the scatterers changes. A change in distance results in a change in phase, and the vector addition of the echo signals from all the scatterers may not be the same from pulse to pulse. Not only will the resultant amplitude change from pulse to pulse because of relative phase differences between the individual scatterers, but it may differ because of the shape of the antenna pattern. Thus the clutter return, instead of being constant, will fluctuate from pulse to pulse, and an uncanceled residue will result at the output of the delay-line canceler. The uncanceled residue also can be analyzed as a spread in the clutter energy spectrum; hence the MTI with range gates and filters will likewise be adversely affected by platform motion. The clutter attenuation due to radar platform motion has been derived for the 34 and Dickey. 35 A more descriptive discussion than delay-line canceler by George 36 found in either of these two papers is given in Ridenour (Ref. 1 sec. 16.13). Urkowitz (stationary or reflector contains a that terrain include extended George's formulas to moving) which has much greater reflectivity than the area around it. Echo fluctuations due to the motion of the antenna on board the aircraft may be resolved into four components one component is due to the rotation of the antenna (scanning fluctuation), and the other three components are due to the motion of the aircraft in space. Dickey derived expressions for these four components of fluctuation, assuming three different types of antenna radiation patterns— a rectangular pattern, a Gaussian pattern of the form exp (— a2 2), and a pattern of the form (sin 6)16. The three components of
receiving antenna.
,
:

aircraft

motion are defined by a rectangular coordinate system. The z axis is located along the center of the antenna beam, and the x axis is horizontal. The y axis is, in general, not vertical, but falls in the same vertical plane as the z axis. The approximate clutter attenuation for the four components, assuming a Gaussian antenna pattern, aref

CA for rotation
1.388

CA for x axis ~ CA for y axis

y88A^_
(nvT6B sin a)
2

a

2

1.37(»Tsm a) 2
2
:

1.388(Afc)

(ttvcTt cos a sin

2
</>

tan

2
<f>)

CA for z axis—large enough to be neglected
either the phase component (limiting t These expressions apply to an MTI receiver sensitive to receiver with phase detector) or to the amplitude component (noncoherent MTI), but not to both.

158

Introduction to Radar Systems

[Sec. 4.8

where nB

motion and projection of antenna beam in the horizontal plane (beam points straight ahead at a = 0°; beam points perpendicular to aircraft at a 90°) elevation angle between horizontal plane and center of beam t = pulse width a = aperture of a uniformly illuminated antenna c = velocity of propagation These approximate expressions apply when the attenuation is large. More exact expressions may be found in Dickey's paper. 35
<x
<f>

= number of hits per 3-db beamwidth 6B = 3-db beamwidth X = wavelength v = target velocity T = pulse-repetition period = azimuth angle between direction of aircraft =

=

The clutter attenuation for rotation is the same as that for the scanning fluctuation declutter

rived in Sec. 4.3 for the stationary radar. The along the z axis is usually well attenu-

Direction of aircraft

motion

CA r
Fig. 4.45. Regions in which each component of ground-clutter residue is likely to be prominent. (After Dickey, 35 IRE Trans.)

be neglected. The total be obtained by adding contributions from each of the four components of motion. At any particular direction, some of the components have greater effect than others. Figure 4.45 indicates the regions where each component is likely to predominate. The x component is important in regions to the side. The y component becomes of importance along the ground track where the x
fluctuation

ated

and

may may

to zero. It is large, however, only where the depression angle is large, and it is important, therefore, at high altitudes. The scanning component, as was found previously, is dependent on the rate of rotation and is independent of the azimuth or elevation angle. This component may limit the AMTI performance at long range and along the ground track, where the x and y components both become small. At long range and along the ground track the z component also may be appreciable. Along

component goes

the ground track, except at extreme range, the pulse length contributes more to the fluctuation than the beamwidth, while at right angles to the ground track, the reverse
is true.

ditions are

Contours of constant clutter attenuation for a particular set of assumed conshown in Fig. 4.46. Clutter attenuation depends only slightly on the

antenna pattern. 35 Figure 4.46 shows that the clutter attenuation (and hence the AMTI performance) deteriorates when the antenna beam is perpendicular to the ground track or when the beam is pointing directly below the aircraft. Clutter fluctuations are least (greatest attenuation) along the ground track at relatively long ranges. The above example shows the difficulties involved in attempting to design an AMTI radar with 360° scanning coverage. In practice, the problem may be even more formidable than that of this example, since aircraft speeds can be considerably higher than the 250 knots assumed, and the pulse repetition frequency might be smaller than 2,000 cps, especially if a long, unambiguous range were desired. Both a higher aircraft velocity and a lower prf make the AMTI performance worse.

Sec. 4.9]
Direction
of

MTI AND
motion

PULSE-DOPPLER

RADAR

159

\Z <B>

Fig. 4.46. Typical ground-clutter attenuation for airborne MTI. 2,000 cps ground speed 250 knots altitude J /xsec prf

Beam
=

=

;

=

;

=

;

width = 3.0°; pulse length 20,000 ft; wavelength = 3.2 cm;

antenna rotation

=

12 rpm.

{From Dickey, 3 ' IRE Trans.)

4.9. Effect of Sidelobes

on Pulse-doppler

AMTI

Radar32

There will always be undesired sidelobe radiation from an antenna in directions other than the main beam. In an airborne radar the troublesome sidelobes are those which illuminate the ground. Although the sidelobe radiation may be small compared with that from the main beam, the relatively short distance to the ground plus the relatively large cross section of the ground at perpendicular incidence (Sec. 12.2) combine to give large clutter contributions from the sidelobes. Therefore the movingtarget signal must compete not only with the clutter illuminated by the main beam, but with clutter illuminated by the sidelobes. In this section the effect of the sidelobes on the pulse-doppler AMTI radar will be considered qualitatively. Similar considerations apply to other types of AMTI radars. The spectrum of the transmitted waveform of a pulse radar is depicted in Fig. 4.47. It consists of a series of spectral lines separated from one another in frequency by the pulse repetition frequency f The envelope of the spectral lines follows a (sin x)jx r shape about the transmitted frequency f The width of the envelope as measured between the first pair of zero crossings aboutyj, is equal to 2/t, where t is the pulse width. If both the target and radar were stationary and if there were no clutter echoes, the frequency spectrum of the echo signal would be the same as that of the transmitted However, the relative motion between radar and target as well as between signal.
. .

160

Introduction to Radar Systems

[Sec. 4.9

radar and clutter and the additional clutter signal received from the antenna sidelobes will substantially modify this idealized signal spectrum. The spectrum of the received signal for the pulse AMTI radar might appear as in Fig. 4.48. Only that portion of the spectrum in the vicinity of/ is shown. The shape of the clutter spectrum about each of the other spectral components spaced at intervals equal to the pulse repetition frequency is the same as that about/ The leakage of the
.

transmitter signal into the receiver produces the spike at a frequency^ and the spikes at/ nf Also in the r where n is an integer and r is the pulse repetition frequency.

±

,

f

vicinity of

f

is

the clutter energy

from the sidelobes which illuminate the ground

M

-2/r

Jjillix.
'o

Frequency

Fig. 4.47. Spectrum of pulse-radar transmitted waveform.

directly
is

beneath the

aircraft.

called the altitude return.

The echo from the ground directly beneath the aircraft The altitude return is not shifted in frequency since the

between radar and ground is essentially zero. Clutter to either have a relative-velocity component and hence some doppler frequency shift; consequently the clutter spectrum from the altitude return will be of finite width. The shape of the altitude-return spectrum will depend upon the
relative velocity

side of the perpendicular will

Transmitter to receiver leakage
Altitude return

Main-lobe clutter
-Target echo (head-on

lAAAAAAAAAAAAAAAAAAAAAAAAAl

Receiver noise

fo-fr

fo+fr

Frequency

Fig. 4.48. Portion of the received signal spectrum in the vicinity of the carrier frequency /„, for a pulse-doppler AMTI radar. (After Maguire, 32 Proc. Natl. Conf. on Aeronaut. Electronics.)

RF

variation of the clutter cross section as a function of antenna depression angle (Sec. The cross section of the clutter directly beneath the aircraft for a depression 12.2). angle of 90° can be quite large compared with that at small depression angles. The large cross section and the close range can result in considerable altitude return.

The
angle

clutter illuminated

by the antenna sidelobes in directions other than

directly

beneath the aircraft

made by

the

may have any relative velocity from +v to —v, depending on the antenna beam and the aircraft vector velocity (v is the aircraft

The clutter spectrum contributed by these sidelobes will extend 2vjX cps on either side of the transmitter frequency. The shape of the spectrum will depend upon the nature of the clutter illuminated and the shape of the antenna sidelobes. For purposes of illustration it is shown in Fig. 4.48 as a uniform spectrum.
velocity).

Sec. 4.9]

MTI AND
clutter directly illuminated

PULSE-DOPPLER RADAR
is

161
also

The ground
shown

in Fig. 4.48.

by The doppler frequency shift
fc

the

main beam of the antenna of the main-beam clutter is

— cos = 2v

,

<j>

(4.70)

where is the angle between the direction of the the antenna beam (Fig. 4.49). Also shown in
cf>

aircraft vector velocity

and the

axis of

this figure are the various sources

of

clutter signals.

The

finite

antenna beamwidth

results in

a

finite

associated with the

main-beam

clutter.

The spread
AcS

is

spread of the doppler frequency approximately

A/
where
A<f>

c

= — sin
A

<f>

=


A

-

sin

<f>

(4.71)

was

set
is

differentiation

equal to the antenna beamwidth 6B The negative sign produced on ignored. The maximum doppler spread occurs when the beam is
.

Radar
aircraft

-Simplified perimeter of elevation side-lobe pattern Direction of aircraft velocity

Target aircraft

Ground

clutter

Fig. 4.49. Sources of clutter signals.

zi {After Maguire, Proc. Natl. Conf. on Aeronaut. Electronics.)

For example, if the radar antenna beamwidth were 2° and the wavelength 0. 1 m(f = 3,000 Mc) and if the aircraft velocity were 400 knots, the doppler- frequency spread would be 144 cps. (The maximum doppler frequency in this example corresponds to 4,120 cps.) Equation (4.71) indicates that the spread in doppler will be small if the beamwidth B and the depression angle
perpendicular to the aircraft vector velocity.
cf>

are small.

The altitude return may be eliminated by turning the receiver off (gating) at that range corresponding to the altitude of the aircraft. Gating the altitude return has the disadvantage that targets at ranges corresponding to the aircraft altitude will also be eliminated from the receiver. A better method of suppressing the altitude return in the pulse radar is to eliminate the signal in the frequency domain, rather than in the The same rejection time domain, by inserting a rejection filter at the frequency f The clutter energy from filter will also suppress the transmitter-to-receiver leakage. the main lobe may also be suppressed by a rejection filter, but since the doppler frequency of this clutter component is not fixed, the rejection filter must be tunable and servocontrolled to track the main-lobe clutter as it changes because of scanning or because of changes in aircraft velocity. The position of the target echo in the frequency spectrum depends upon its velocity relative to that of the radar aircraft. If the target aircraft approaches the radar aircraft head on (from the forward sector), the doppler frequency shift of the target will be greater than the doppler shifts of the clutter echoes, as shown in Fig. 4.48. A filter can be used to exclude the clutter but pass the target echo. Similarly, if the targets are receding from one another along headings 180° apart, the target doppler frequency
.

162

Introduction to Radar Systems
again
lie

[Sec. 4.9

outside the clutter spectrum and may be readily separated from the In other situations where the radar may be closing on the target from the tail or from the side, the relative velocities may be small and the target doppler will lie within the clutter doppler spectrum. In such situations the target echo must compete with the clutter energy for recognition. large part of the clutter
shift will

clutter energy

by

filters.

A

energy may be removed with a bank of fixed narrowband filters covering the expected range of doppler frequencies. The bandwidth of each individual filter must be wide enough to accept the energy contained in the target echo signal. The width of the filter will depend upon the time on target, equipment fluctuations, and other effects which broaden the echo-signal spectrum as discussed previously. Each filter may be followed by range gates and integrators. The use of a parallel bank of niters lowers the sensitivity of detection somewhat, because the increased false-alarm rate of a filter bank as compared with a single filter must be compensated by increasing the threshold of
detection.
niters
is

However, the loss is small. The chief limitation of a fixed bank of doppler the additional equipment complexity. If only a few targets are expected,

narrowband doppler tracking filters might be used, one for each target. The filter must search through the expected doppler range before it can "lock on." If the radar receiver must search in both range and doppler to find its target, a relatively long search time might be required. In spite of its complexities and its shortcomings, a pulse-doppler radar is one of the
better techniques for

AMTI application.
REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

Ridenour, L. N.: "Radar System Engineering," MIT Radiation Laboratory Series, vol. 1, chap. 16, McGraw-Hill Book Company, Inc., New York, 1947. Solomon, K. A Double Delay and Subtraction Airborne Clutter Canceller, Proc. Conf. on Military Electronics (IRE), 1958, pp. 235-240. Eastwood, E., T. R. Blakemore, and B. J. Witt: Marconi Coherent MTI Radar on 50 Cms, Marconi Rev., vol. 19, 2d quarter, no. 121, pp. 53-60, 1956. Emslie, A. G.: Moving Target Indication on MEW, MIT Radiation Laboratory Rept. 1080, Feb. J r 19,1946. Tanter, H.: Radar Receiver with Elimination of Fixed-target Echoes, Elect. Commun., vol 31 pp. 235-248, December, 1954. Gager, C. Transistorized MTI Canceller, Airborne Instruments Laboratory Monograph from Pulse of Long Island (IRE), February, 1960. Chance, ,B., R. I. Hulsizer, E. F. MacNichol, and F. C. Williams (eds.): "Electronic Time Measurements," MIT Radiation Laboratory Series, vol. 20, chap. 12, McGraw-Hill Book Company, Inc., New York, 1949. Blackburn, J. F. (ed.): "Components Handbook," MIT Radiation Laboratory Series, vol. 17, chap. 7, McGraw-Hill Book Company, Inc., New York, 1948. Chance, B., F. C. Williams, V. W. Hughes, D. Sayre, and E. F. MacNichol, Jr. (eds.) "Waveforms," MIT Radiation Laboratory Series, vol. 19, chap. 23, McGraw-Hill Book Company, Inc., New r J> York, 1949. Emslie, A. G.: Moving Target Indication on MEW, MIT Radiation Laboratory Rept. 1080, J
:
:

:

>

Feb. 19, 1946.
11.

^

Arenberg, D. L.: Ultrasonic Solid Delay Lines,

J.

Acous. Soc. Am., vol. 20, pp. 1-28, January,

12.

13. 14. 15.

16.

17.

Huntington, H. B., A. G. Emslie, and V. W. Hughes: Ultrasonic Delay Lines, Pt. I, /. Franklin pp. 1-24, January, 1948; Emslie, A. G., H. B. Huntington, H. Shapiro, and A. E. Benfield, Pt. II, pp. 101-115, February, 1948. Arenberg, D. L.: Ultrasonic Delay Lines, IRE Natl. Conv. Record, 1954. Bliley Electric Co.: Bulletin 48, Erie, Pa., 1955. May, J. E., Jr. Low-loss 1000 Microsecond Ultrasonic Delay Lines, Proc. Natl. Electronics Conf., vol. 11, pp. 786-790, 1955. Bendat, J. S.: "Principles and Applications of Random Noise Theory," pp. 114-118 John Wiley & Sons, Inc., New York, 1958. Perlman, S. E.: Staggered Rep Rate Fills Radar Blind Spots, Electronics, vol. 31, no 47 pp
Inst., vol. 245,
:

82-85, Nov. 21, 1958.

VV

'

MTI AND
18.

PULSE-DOPPLER

RADAR

163

White, W. D., and A. E. Ruvin: Recent Advances in the Synthesis of Comb Filters, IRE Natl. Conv. Record, vol. 5, pt. 2, pp. 186-199, 1957. 19. White, W. D.: Synthesis of Comb Filters, Proc. Natl. Conf. on Aeronaut. Electronics, 1958, pp. 279-285. 20. Urkowitz, H.: Analysis and Synthesis of Delay Line Periodic Filters, IRE Trans., vol. CT-4, pp. 41-53, June, 1957. 21. McKee, D. A.: An MTI Cancellation System, MIT Lincoln Lab. Tech. Rept. 171, Jan. 8,

FM

1958.

Ruvin, A.: Blower Vibration and MTI, Airborne Instruments Laboratory Monograph from Pulse of Long Island (IRE), April, 1959. 23. Stephenson, J. G.: Designing Stable Tunable Microwave Oscillators, Electronics, vol. 28, pp. 184-187, March, 1955. 24. Dauksher, W. J.: Stable Local Oscillator for S-band Radar, Electronics, vol. 29, pp. 179-181,
22.

September, 1956.
25. Goldstein, H.:

The

Effect of Clutter Fluctuations

on MTI,

MIT Radiation Lab.

Rept. 700, Dec.

27, 1945.
26. Swerling, P.
:

Probability of Detection for Fluctuating Targets,

IRE Trans., vol. IT-6, pp.

269-308,

April, 1960.
27. 28. Grisetti,

29.

30. 31.

32.

33.

Barlow, E. J. Doppler Radar, Proc. IRE, vol. 37, pp. 340-355, April, 1949. R. S., M. M. Santa, and G. M. Kirkpatrick: Effect of Internal Fluctuations and Scanning on Clutter Attenuation in MTI Radar, IRE Trans., vol. ANE-2, no. 1, pp. 37^1, March, 1955. Kroszczynski. J. Efficiency of Attenuation of Constant Echoes in Simple and Double Cancellation Apparatus, Prace, Przem. Inst. Tele., vol. 8, no. 24, pp. 41-46, 1958. (Translated by Morris D. Friedman, MIT Lincoln Laboratory.) Anderson, D. B. A Microwave Technique to Reduce Platform Motion and Scanning Noise in Airborne Moving-target Radar, IRE WESCON Conv. Record, vol. 2, pt. 1, pp. 202-211, 1958. Sargent, R. S.: Moving Target Detection by Pulse Doppler Radar, Electronics, vol. 27, no. 9, pp. 138-141, September, 1954. Maguire, W. W.: Application of Pulsed Doppler Radar to Airborne Radar Systems, Proc. Natl. Conf. on Aeronaut. Electronics (Dayton, Ohio), pp. 291-295, 1958. Richardson, R. E.: Some Pulse-doppler Radar Design Considerations, MIT Lincoln Lab. Tech.
: : :

Rept. 154, Aug. 12, 1957. S. Fluctuations of Ground Clutter Return in Airborne Radar Equipment, Proc. IEE, vol. 99, pt. IV, no. 2, pp. 92-99, April, 1952. 35. Dickey, F. R., Jr.: Theoretical Performance of Airborne Moving Target Indicators, IRE Trans., PGAE-8, pp. 12-23, June, 1953. 36. Urkowitz, H.: An Extension to the Theory of the Performance of Airborne Moving-target Indicators, IRE Trans., vol. ANE-5, pp. 210-214, December, 1958. 37. Bauer, P.A. Low Temperature Coefficient Ultrasonic Solid Delay Lines, Solid State J., vol. 2, no. 12, pp. 23-29, December, 1961.
34.

George, T.

:

:

5
TRACKING RADAR
5.1.

Tracking with Radar

A tracking-radar system measures the coordinates of a target and provides data which
All or only part of the available radar data range, elevation angle, azimuth angle, and doppler frequency shift—may be used in predicting future position; that is, a radar might track in range, in angle, in doppler, or with any combination. Almost any radar can be

may be used to determine the target path and to predict its future position.

considered a tracking radar provided its output information is processed properly. But, in general, it is the method by which angle tracking is accomplished that distinguishes what is normally considered a tracking radar from any other radar. It is also necessary to distinguish between a continuous tracking radar and a track-while-scan

(TWS) radar. The former supplies continuous tracking data on a particular target, while the track-while-scan supplies sampled data on many targets. In general, the continuous tracking radar and the radar employ different types of equipment. The antenna beam in the continuous tracking radar is positioned in angle by a servomechanism actuated by an error signal. The various methods for generating the

TWS

error signal

may be classified as sequential lobing,

conical scan,
shift

and simultaneous

lobing

or monopuhe.

The range and doppler frequency

can also be continuously

tracked, if desired, by a servo-control loop actuated by an error signal generated in the radar receiver. The information available from a tracking radar may be presented on a

cathode-ray-tube (CRT) display for action by an operator, or it may be supplied to an automatic computer which determines the target path and calculates its probable
future course.

The tracking radar must first find its target before it can track. Some radars, such as the SCR-584, operate in a search mode in order to find the target before switching to a tracking mode. Although it is possible to use a single radar for both the search and the tracking functions, such a procedure usually results in certain operational limitations.
used in its tracking mode, it has no knowledge of other antenna pattern is a narrow pencil beam and if the search volume is large, a relatively long time might be required to find the target. Therefore many radar tracking systems employ a separate search radar to provide the information necessary to position the tracker on the target. search radar, when used for this purpose, is called an acquisition radar. In some applications, separate search and track radars may not be practical or even desirable. An example is the airborne interception (AI) radar, in which the angular search volume is not too large and is usually restricted to the forward sector only. In
is

Obviously,

when

the radar

potential targets.

Also,

if the

A

addition, there

is

usually

Even

in those applications

tracker usually will
target.

room to spare in an aircraft for two separate radars. where a separate radar supplies acquisition information, the have to perform some limited angular search in order to find the
little

The scanning fan-beam search radar can also provide tracking information to determine the path of the target and predict its future position. Each time the radar

beam

scans past the target, its coordinates are obtained. If the change in target coordinates from scan to scan is not too large, it is possible to reconstruct the track of
164

Sec. 5.2]
the target.

Tracking Radar

165

This is called track-while-scan. It provides tracking information on a sampled, basis rather than continuously. The simplest manifestation of a TWS radar may be had by providing the PPI-scope operator with a grease pencil to mark the target pips on the face of the scope. line joining those pips that correspond to the same target provides the target track. When the traffic is so dense that human operators cannot maintain pace with the information available from the radar, the target trajectory data may be processed automatically in a digital computer as is done in the SAGE air defense system. Whenever the term tracking radar is used in this book, it refers to the continuous tracker rather than to track-while-scan, unless otherwise
discrete, or

A

specified.

The chief use of the continuous tracking radar has been for the control of military weapons such as antiaircraft artillery and missile guidance. Tracking radars are used also for guidance in the launchings of satellites and space vehicles. Because of their
versatility, tracking radars often

have been used for general-purpose instrumentation

or as a research tool.
5.2. Sequential

Lobing

pencil

The antenna pattern commonly employed with tracking radars is the symmetrical beam in which the elevation and azimuth beamwidths are essentially equal. A

pencil-beam antenna has many advantages for tracking-radar applications. It provides high gain by concentrating the radiated power in the direction of the target. It reduces unwanted echoes from other targets and from the ground. The angular coordinates of the target can be determined more precisely than with a fan beam. However, a simple pencil-beam antenna is not suitable for tracking radars unless means are provided for determining the magnitude and direction of the target's angular position with respect to some reference direction, usually the axis of the antenna. The. difference between the target position and the reference direction is the angular error. The tracking radar attempts to position the antenna to make the angular error zero. When the angular error is zero, the target is located along the reference direction. One method of obtaining the direction and the magnitude of the angular error in one coordinate is by alternately switching the antenna beam between two positions (Fig. This is called lobe switching, sequential switching, or sequential lobing. Figure 5.1). 5.1a is a polar representation of the antenna beam (minus the sidelobes) in the two switched positions. A plot in rectangular coordinates is shown in Fig. 5.1Z>, and the error signal obtained from a target not on the switching axis (reference direction) is shown in Fig. 5.1c. The difference in amplitude between the voltages obtained in the two switched positions is a measure of the angular displacement of the target from the switching axis. The sign of the difference determines the direction the antenna must be moved in order to align the switching axis with the direction of the target. When the voltages in the two switched positions are equal, the target is on axis and its position may be determined from the axis direction. Two additional switching positions are needed to obtain the angular error in the orthogonal coordinate. Thus a two-dimensional sequentially lobing radar might consist of a cluster of four feed horns illuminating a single antenna, arranged so that the right-left, up-down sectors are covered by successive antenna positions. Both transmission and reception are accomplished at each position. cluster of five feeds might also be employed, with the central feed used for transmission while the outer four feeds are used for receiving. High-power RF switches are not needed since only the receiving beams, and not the transmitting beam, are stepped in the five-feed arrangement. One of the limitations of a simple unswitched nonscanning pencil-beam antenna is that the angle accuracy can be no better than the size of the antenna beamwidth. An important feature of sequential lobing (as well as the other tracking techniques to be

A

|

166

Introduction to Radar Systems

[Sec. 5.3

discussed) is that the target-position accuracy can be far better than that given by the antenna beamwidth. The accuracy depends on how well equality of the signals in the switched positions can be determined. The fundamental limitation to accuracy is system noise caused either by mechanical or electrical fluctuations. Sequential lobing, or lobe switching, was one of the first tracking-radar techniques to be employed. Early applications were in airborne-interception radar, where it provided directional information for homing on a target, and in ground-based antiaircraft fire-control radars such as the SCR-268. It is not used as often in modern tracking-radar applications as some of the other techniques to be described.

Time

Fig. 5.1. Lobe-switching antenna patterns and error signal (one dimension), (a) Polar representation of switched antenna patterns; (ft) rectangular representation; (c) error signal.

5.3.

Conical Scan

of the simultaneous lobing technique described in the previous an offset antenna beam rather than discontinuously step the beam between four discrete positions. This is known as conical scanning (Fig. 5.2). The angle between the axis of rotation (which is usually, but not always, the axis of the antenna reflector) and the axis of the antenna beam is called the squint angle. Consider a target at position A. The echo signal will be modulated at a frequency equal to the rotation frequency of the beam. The amplitude of the echo-signal modulation will depend upon the shape of the antenna pattern, the squint angle, and the angle between the target line of sight and the rotation axis. The phase of the modulation depends on the direction of the angle between the target and the rotation axis. The
section
is

A logical extension

to rotate continuously

conical-scan modulation isextracted fromtheecho signal and applied to a servo-control

system which continuallypositionstheantennaon the target.

[Note that two servos are

t The squint angle is also sometimes used to describe the angle between the two major lobe axes in a lobe-switching antenna (IRE Standards 54 IRE 12 S 1), but this use of the term is not employed here.

Sec. 5.3]

Tracking Radar

167

Both the rectangular required because the tracking problem is two-dimensional. is on target, as antenna the When used may be coordinates ] tracking polar (az-el) and axis coincide, and the in B of Fig. 5.2, the line of sight to the target and the rotation
.

conical-scan modulation is zero. block diagram of the angle-tracking portion of a typical conical-scan tracking radar The antenna is mounted so that it can be positioned in both in Fig. 5.3. shown is

A

Target axis

-

Beam
rotation

Radar
Fig. 5.2. Conical-scan tracking.

azimuth and elevation by separate motors, which might be either electric- or hydraulicThe antenna beam is offset by tilting either the feed or the reflector with driven. respect to one another. One of the simplest conical-scan antennas is a parabola with an offset rear feed If the feed maintains the plane of polarization rotated about the axis of the reflector. A rotating feed such as is used in the fixed as it rotates, it is called a nutating feed.
Transmitter

To rotary joint

on antenna /—
Ref. aen.x

Duplexer

Receiver with

AGC

Third detector

Error-signal
filter

smZw^t
cos Zirf s
t
'

^/1-j//"
Scan motor

Error
signal

\

n%p&31 U%s,&

,^<
/£•»

Elevation

Elevation-angle error detector

servo
amplifier

Elevat on

^

T

motor
\

Azimuth
servo
servo motor -r
amplifier

Azimuth-ongle
error detector

Fig. 5.3. Block diagram of conical-scan tracking radar.

SCR-584
small,

rotates a dipole

requires a rotary joint.

and thus rotates polarization. The latter type of feed The nutating feed requires a flexible joint. If the antenna is

it may be easier to rotate the dish, which is offset, rather than the feed, thus avoiding the problem of a rotary or flexible RF joint in the feed. A typical conical-scan The same motor that provides the conicalrotation speed might be 30 rps (1 ,800 rpm) scan rotation of the antenna beam also drives a two-phase reference generator with two
.

168

Introduction to Radar Systems

[Sec. 5.3

outputs 90° apart in phase.
tion and azimuth errors.

two rotary joints motion in azimuth; the
via

These two outputs serve as a reference to extract the elevaThe received echo signal is fed to the receiver from the antenna (not shown in the block diagram). One rotary joint permits
other, in elevation.

a conventional superheterodyne except for two features peculiar to the conical-scan tracking radar. One feature not found in other radar receivers is a means of extracting the conical-scan modulation, or error signal. This is accomplished after the second detector in the video portion of the receiver. In the block diagram this function is indicated as the third detector. The purpose of the low-pass errorreceiver
is

The

signal filter is to remove the harmonics of the conical-scan frequency, the prf, and the harmonics of the prf if they are present. The error signal is compared with the elevation and azimuth reference signals in the angle-error detectors, which are phase-sensitive 1-5 detectors. A phase-sensitive detector is a nonlinear device in which the input signal

mixed with the reference signal. The input and reference signals are of the same frequency. The output d-c voltage reverses polarity as the phase of the input signal changes through 180°. The magnitude of the d-c output from the angle-error detector is proportional to the error, and the sign (polarity)
(in this case the angle-error signal) is
is

an indication of the direction of the error. The angle-error-detector outputs are amplified and drive the antenna elevation and azimuth servo motors. When the antenna is directly on target, the error signal is zero. The angular position of the target may be determined from the elevation and azimuth of the antenna axis. The position can be read out by means of standard angle trans-

ducers such as synchros, potentiometers, or analog-to-digital-data converters. The difference between the phase-sensitive detector and phase detector is often one of actual operating conditions. 5 The phase detector measures the phase difference between two sinusoidal signals of the same frequency. In the phase-sensitive detector the output voltage reverses polarity as the phase of the input changes through 180°. Identical circuits can be used for phase measurement and for phase-sensitive detection. It is usually assumed that the amplitudes of the reference and the input signal are the same in the phase detector, while in the phase-sensitive detector, the reference is much larger than the input signal. Boxcar Generator. The purpose of the third detector and filter is to pass the modulation at the conical-scan frequency and to reject the pulse repetition frequency and its harmonics. In the early S-band version of the SCR-584, this was accomplished with a more or less conventional amplitude detector and filter. In the Z-band version and in most modern radars the filtering function is performed with a device called the boxcar generator. 6 The boxcar generator was also mentioned in the discussion of the MTI receiver using range-gated filters (Sec. 4.4). In essence, it clamps or stretches the video pulses of Fig. 5.4a in time so as to cover the entire pulse-repetition period (Fig. 5Ab). This is possible only in a range-gated receiver. (Tracking radars are normally operated with range gates.) The boxcar generator consists of an electric circuit that clamps the potential of a storage element, such as a capacitor, to the video-pulse amplitude each time the pulse is received. The capacitor maintains the potential of the pulse during the entire repetition period and is altered only when a new video pulse appears whose amplitude differs from the previous one. The boxcar generator eliminates the pulse repetition frequency and reduces its harmonics. It also has the practical advantage that the magnitude of the conical-scan modulation is amplified because pulse stretching puts more of the available energy at the modulation frequency. The pulse repetition frequency must be sufficiently large compared with the conical-scan frequency for proper boxcar filtering. If not, it may be necessary to provide additional filtering to attenuate undesired cross-modulation frequency components. Automatic Gain Control. 1 9 The echo-signal amplitude at the tracking-radar

s N. 5. (2) the conical-scan modulation (angle-error signal). would be lost if the receiver were to In the conical-scan tracking radar an AGC that maintains the d-c level saturate. 0-c amplifier AGC filter T Delay voltage V c of a tracking-radar receiver.Sec.5. It is shown later in this section (during the derivation of the error-signal voltage) that the d-c level of the receiver must be maintained constant if the angular error is to be linearly to prevent saturation One of the purposes of AGC in any receiver is the error signal The scanning modulation and related to the angle-error signal voltage. 5.5. Block diagram of the AGC portion portion of a tracking-radar receiver is shown in Fig. The three major causes of variation in amplitude are (1) the inverse-fourth-power relationship between the echo signal and range. and (3) amplitude fluctuations in the target cross section. (a) Pulse train with conical-scan modulation. same pulse train after passing through by large signals. Fig. ^. 5. filter and fed back to control the gain of the IF amplifier. will be greater and the signal the feedback the larger will be loop should pass all frequencies from direct current to just below the filter in the An AGC A AGC . The the gain reduction. Video amplifier Range gate and to angle-error detector boxcar ' ' . The function of the automatic gain control (AGC) is to maintain the d-c level of the receiver output constant and to smooth or eliminate as much of the noiselike amplitude fluctuations as possible without disturbing the extraction of the desired error signal at the conical-scan frequency. 5. Conical scan modulation mixer IF amplifier > 2d det. boxcar generator.4. constant results in an error signal that is a true indication of the angular pointing error.j* V K *^-- -») -Vtr (a) -Vtr (A) (Z>) Fig. example of the portion of the video-amplifier output is passed through a low-pass or smoothing The larger the video output. .3] receiver will not Tracking Radar 169 be constant but will vary with time.

If the range variation were 10 to 1 the contribution to the dynamic range would be 40 db." between the AGC AGC AGC A and azimuth angle-tracking loops. but the performance will not be as good. The error signal may be recovered from the voltage with a narrow bandpass filter centered at the scan-modulation frequency. In such cases the AGC can still perform satisfactorily since the loop gain is usually low for small signals. and in some applications it should be as little as 2°. The middle stages are usually the ones controlled since the first stage gain should remain high so as not to influence the noise figure of the mixer stage. An alternative AGC filter design would maintain the AGC loop gain up to frequencies much higher than the conical-scan frequency. 8 For this reason. and the output would be used to measure range in the normal manner. . The required dynamic range of the AGC will depend upon the variation in range over which targets are tracked and the variations expected in the target cross section. a filter with a sharp attenuation characteristic in the vicinity of the conicalscan frequency might not be desirable because of the relatively large amount of phase filter shift will be zero unless the feedback voltage exceeds a value Vc In the block diagram the feedback voltage and the voltage Vc are compared in the d-c amplifier. If the feedback voltage exceeds Vc the is operative. The voltage Vc is called the delay voltage. The terminology may be a bit misleading since the delay is not in time but in amplitude. 8 It is found in practice that the maximum gain variation which can be obtained with a single IF stage is of the order of 40 db. In conelevation ventional tracking-radar applications the phase change introduced by the feedback-loop should be less than 10°.3 conical-scan-modulation frequency. The loop gain of the filter measured at the conical-scan frequency should be low so that the error signal will not be affected by action. 5. The target cross section might also contribute another 40-db variation. any output which might appear from the receiver would be due to the failure of the AGC circuit to regulate completely. The error signal from the conical-scan tracker will be derived assuming that a properly designed eliminates all signal modulations except the conical-scan modulation. In many applications of AGC the delay voltage is actually zero. This is called undelayed AGC. AGC voltage since AGC AGC AGC Consider an echo pulse train with conical-scan modulation as shown in Fig. In this case. The scan modulation would be effectively suppressed in the output of the receiver. 5. Hence the dynamic range of operation required of the receiver AGC might be of the order of 90 db. If the responds to the conical-scan frequency. while if it is less.170 Introduction to Radar Systems [Sue. Thus the AGC will not regulate weak signals. the error signal might be lost. phase change of the error signal is equivalent to a rotation of the reference axes and introduces cross coupling. the error signal can be recovered from the . The effect is similar to having a delay voltage. it varies at the conical-scan frequency. Another 10 db ought to be allowed to account for variations in the other parameters of the radar equation. there is no action. The output of the feedback loop prespecified minimum . It is also best not to control the last IF stage since the maximum undistorted output of an amplifying stage is reduced when its gain is reduced by the application of a control voltage. The voltage will also contain any amplitude fluctuations that appear with the echo signal. or "cross talk. . The purpose of the delay voltage is to provide a reference for the constant output signal and permit receiver gain for weak signals. Cross talk affects the stability of the tracking and might result in an unwanted nutating motion of the antenna. If the delay voltage which it would introduce. Error Signal. Therefore two to three stages of the IF amplifier must be gain-controlled to accommodate the total dynamic range. or perhaps more. AGC AGC were zero.4a. The phase shift of this filter must be small if its phase characteristic is not to influence the error signal.

. 5. To K" = {exp 1-aXdl + d\)-}}I (2a\0 T ) Kn = ^ff^ I {2a 6 q e T ) l .1) will G(6) =G exp {-aW) (5. . . 5.6b).6) In Eq. R6 q R6 T are small. they may be related by .6a. <£ ) (5.4) The following relationship may be derived from expressions 00 for Bessel functions given by Whittaker and Watson11 exp (—x cos f) = I (x) + 2 ^ l n = n l ( x ) cos "V (5-5) where /„ is the nth-order Bessel function of imaginary coefficient.3) Substituting 6 2 as given by Eq. R6 g Rd T on a sphere of radius R (Fig. = angle between antenna-beam axis and target axis that is. (5.3] Tracking Radar is 171 linear The pulse repetition frequency f r. of antenna pattern Referring to Fig. 2. .716/6%. The angles 6. value of G(0) at 8 constant 2. into Eq. (5. Assuming a detector. Using the above relationship. Since the distances Rd. where/. 5. be expanded to obtain the various frequency components contained in the received signal. is antenna beamwidth measured between 3-db. . where 6 is measured in degrees and 6 B also in degrees. Eq.Sec.4) may be written as G(0 = exp i-a\d\ + 6%)-] I (2a\d T) + 2 2 I n (2a%6 T = i ) cos (2rrnfs t - n<j> ) \ (5. K' is determined = k = Equation llfor k/f r integer < < t k/f r + r \0 otherwise = 0. 9 T is the angle between the axis of rotation and the target axis <f> is the rotation angle of the conical scanner as measured from some arbitrary phase reference and cj> is the angle defined by the target and the reference axis. with G = gives G(0) = exp \_-a\d\ + 2 T )] exp [-2a%6 T cos (<f> - O }] (5. or half-power. 6 Q 6 T may also be defined by the lengths of arc Rd. and the pulse width is t. let </>. 1.6) 2irfs t has been substituted for simplify the algebra.3). is the conical-scan frequency. points. . . (5. the angle Qq is the squint angle denned by the antenna-beam axis and the axis of rotation. {R6f = (RdQf + (R6 T f + 2R*6 e T cos g (<f> 1. = = . the law of cosines to the angle <f> — cf> .2). The two-way-voltage (or one-way-power) antenna pattern may be approximated by the Gaussian function (5.2) where 6 a2 — G = maximum antenna gain. the video pulse train may be represented by the V{t) expression 10 (5. (5. First the expression for the antenna modulation factor G(t) will be derived.1) = K'G{t)Fk {t) where K' G(i) = = Fk (t) = constant determined by design of AGC (without by parameters of radar equation) modulation due to antenna pattern waveform representing unmodulated pulse train AGC.

6).7) Target Locus of antenna-beam center Fig. (b) head-on view of conical-scan antenna beam. Fourier series expansion of Fk (i) is The Fk (t)=frr where 1 + 2.172 Introduction to Radar Systems [Sec. The factor Fk {t) in Eq.* — +2 n= l n<f> ) (5. K m cos 2nmf r H)] (5. (a) Geometry and symbols for derivation of conical-scan error signal. the antenna scan-modulation factor becomes G(0 = K" 1 K n C° s (27771/.1) is the rectangular pulse train of unit amplitude. (5.8) K = 2 sin mTTTyr rmrfrT .6.3 Substituting the above in Eq. 5. (5. 5.

(5. the elevation-angle-error voltage for small error Elevation-error signal proportional to (5. An infinite number of sidebands ±nfs centered about the prf. Likewise. t In many cases.and azimuth-error signals are applied to their respective servo motors which position the antenna for zero-error signal.10) is fed into <f> . (5. r consists of four parts A d-c component of magnitude K'K"f T An infinite number of a-c components corresponding to the conical-scan frequency its 2. no :'K"frr(l \ + | K n cos (2rrnf st n<f> ) n=l K m cos2^mf (t-l) +J \ 2/ m=X t r 2 2 ^P {cos |>(m/ + r OD 00 If If n/. the number of harmonics would be finite.13) m Cjd T sin <f> The elevation. suitable low-pass filter permits only the direct current and the scan frequency f s the error signal becomes 1/(0 = K'K"fTr + 2K'frr exp [-a 2(0 2 + a fl 2 c r )]/i(2a M cos (2vft t - fa) (5..]}) (5-9) (5.3] Tracking Radar 173 Substituting Eqs. angle-error detectors (for example. One of the the horizontal and vertical projections of the target angle T are extracted.)f . (Vr is the magnitude of the reference-signal voltage. 5.1 1) as a function of d T /6 B and for various values of 6J6 B For small angular error. .)f . Fig. f s and 3. r >/. (5. harmonics mfr .7) and (5. directly proportional to the angular error if the error is small Thus the output of the azimuth angular detector is a voltage and if A" is maintained is constant by the AGC.8) into Eq.1) gives the video voltage. In the angle-error detectors. 5.7. the azimuth channel) is supplied with a reference Both signals are signal Vr cos 2-nfj. the phase-sensitive detector is designed so that the output amplitude is a function of the input signal only and is independent of the reference-signal amplitude. and that a to pass. Azimuth-error signal = 2 2K'fr rVr exp [-a (0* + 2 r )]/1 (2fl 2 e9 r) cos <j> (5. (5.nmf T r n^ ] + cos \2v{mf r n/.12) where C 1 is a constant. (5. 5 This does not affect the above analysis. Eq. (5.9) The 1 theoretical conical-scan-modulated echo signal as represented by Eq.) The output of the azimuth-error detector is a d-c voltage proportional to the magnitude of the a-c component of the error signal of Eq. is shown in Azimuth-error signal <** C$ T cos <f> (5.77/n/ T + r «&. The only component of interest is the modulaIf it can be assumed that/ tion at the conical-scan frequency /. and in a practical radar.1 1) approaches the following. which are phase-sensitive detectors.11) A plot of Eq. signal represented by Eq.10) times the magnitude of the or reference voltagej times the cosine of the phase difference between the two./r and its harmonics Each of these components has a different amplitude. derived from the reference generator mounted on the antenna scanner. while the other is supplied with Vr sin 2nf s t. harmonics nfs infinite An number of its a-c components corresponding to the pulse repetition frequency fr and 4.10) The error both the azimuth and the elevation angle-error detectors.Sec.

2 0. 0.22 0.1 0.2-nmf r \t • -^ + nfs ) cos [27r(mfr K„fr 7Tm = ln = l \mfr + sin Tn(mfr + nfs )t .06 0.9 1. (5.36 7. The Fourier series for a flat-top pulse train with a pulse repetition frequency/. 5. The angle <j>„ is assumed fixed. This combination removes the d-c component and passes only the fundamental of the conical-scan frequency.io 0.7. is 6 .5 1.18 >'• °B / /a 6 '§ 0.0 eT /eB target angle (6 T /6B) Fig. (5. Plot of the relative error-signal from the conical-scan radar [Eq.6 0.ttxxj t s nfs nfs ) cos \2ir{mfr noS ] fmf r Jr . Although this procedure was used in some early tracking radars and is illustrated in Fig. (5.2 0.-nnf r +1 Ttn) n= s s t s n^o) \ s S + .04 0. it is usually more common to employ a boxcar generator in the video to accomplish essentially the same function.9) in which the pulses were assumed to be amplitude-modulated by the conical-scan modulation rather than flat-topped. QJ 1 6 /a 8 *Q2 _ _ ~ O) 0.irnf r + s w ^ ]l) )/ (5..11)] as a function of and squint angle (6JB B).14) nf.95 4.12 V{t) = K'K"(t Tfr \ \ fr K n -^t sin nnf T cos (2nnf . 5.8 0.12 i > II // _ |o. - - nf.4 0. UJ - 0.3 should be noted that one of the properties of the phase-sensitive detector is that no cross talk exists between the two error channels even though the same error signal [Eq. modulated by the antenna conical-scan modulation of a Gaussian antenna pattern [Eq. The error-signal voltage from the boxcar .6)]. sin Mm/. generator will be slightly different from that given by Eq. .174 It Introduction to Radar Systems [Sec. Note that this differs from the waveform represented by Eq. 8 0.3 0. 1 1 1 1 1 i 1 1 0. but with more gain.)t .4 0.6 0. (5.5 0. 5 3.10)] is fed to both angle-error detectors. 7T 1 1 m=l 00 QO z ^ 2 sin TrmrL m cos.8 I I 0.10). db - 0. (5. In the above analysis the error signal was extracted by passing the video voltage through a detector followed by a low-pass filter. To derive the error signal with the boxcar generator we start with the Fourier-series development of a train of flat-top pulses whose amplitude is proportional to the scan modulation.08 S L.7 I i - 0.20 o> _.14 0.7 0.02 r 1 1 I 1 Antenna Bb crossover.

for proper s operation of the boxcar generator.. by a fluctuating target cross section. The error signal is s. There are two effects to be noted with the boxcar generator as compared with the conventional detector [Eq. There remains a d-c component. latter Ve (t) = 2K' exp {-a\d\ + d^I^a^O T ) "}.4. a filter may be used to r >/. unwanted modulation products will appear in the error signal. Conical scanning usually requires more than four pulses to derive the error signal. (5.Sec. the minimum number of pulses in sequential lobing is usually four one per antenna position.17) similar expression may be found for the elevation error. it is necessary that/. . the measurement of angular error in two orthogonal coordinates (azimuth and elevation) requires that a In practice. The names simultaneous lobing and monopulse are used to describe those tracking techniques which derive angle-error information on the basis of a single pulse. if 2Vr K' exp {-a\6l T is + 8 r)]'i(2a 2 W^ C2 sin ^ <j> cos fe + <£ ) (5.16) can cause an erroneous indication if not compensated.16) becomes Azimuth-error signal «a T cos (5. eliminate all but the scan frequency. 5. .4] Tracking Radar 175 The output of the boxcar generator (5. In the time interval during which a measurement is made with either sequential lobing or conical minimum of three pulses be processed. The effect of the fluctuating echo can be sufficiently serious in some applications to severely limit the accuracy of those tracking radars which require many pulses to be processed in extracting the error signal.14). 4 ^ sin fr is cos (lirfjt ^ -^St A ' (5. caused. . however. Simultaneous Lobing or Monopulse 13-17 In both the sequential-lobing and conical-scan tracking techniques.12)]. If this relationship does not hold. is found by setting the pulse width t r in Eq \\f All terms containing the pulse repetition frequency and its harmonics nf r disappear. and the phase shift irfjf r in the last term of Eq. Pulse-to-pulse amplitude fluctuations of the echo signal have no effect on tracking accuracy if the angular measurement is made on the basis of one pulse rather than many. ventional detector for the same angular error by the ratio C2 /Q fjfs Thus the boxcar generator is capable of greater gain. the tracking accuracy might be degraded. A = . (5. The angle of arrival of the echo signal may be determined in a single-pulse system by measuring the relative phase or the relative amplitude of the echo pulse received in each beam. >/ 5. the echo pulses the modulation produced by scanning. which utilizes one antenna beam on a time-shared basis. in contrast to the conical-scan or lobe-switching tracker. for example. More than one antenna beam is used simultaneously in these methods. especially if the frequency components of the scan. Also.15) The output of the azimuth Azimuth-error signal angle-error detector = Again. There are several methods by which angle-error information might be obtained with only a single pulse. The voltage output of the boxcar generator is greater than that of the con(5.16) the angle error small and/r >/„ Eq. a component at the conical-scan frequency = f components at the harmonics of/s and components at frequencies mf The r ± nf s components are negligible if f In any event. fluctuations were at or near the conical-scan frequency or the sequential-lobing rate. — must contain no amplitude-modulation components other than If the echo pulse train did contain additional modulation components.

8c. 5.8Z>.8. measurement and is also used as a reference to extract the sign of the error signal. Monopulse antenna patterns and error signal.Sa) to obtain the angular error in one coordinate. On reception. (c) difference pattern.9. The two overlapping antenna beams may be generated with a single reflector or with a lens antenna illuminated by two adjacent feeds.176 Introduction to Radar Systems [Sec. 13 The amplitude-comparison monopulse employs two overlapping antenna patterns (Fig. hence the name monopulse is quite appropriate. The sum pattern is used for transmission. 5. may be used if both elevation.) The sum of the two antenna patterns of Fig." or a short-slot coupler. (a) Overlapping antenna patterns. All the information necessary to determine the angular error is obtained on the basis of a single pulse. 5. feeds Signals received from the sum and the difference patterns are amplified separately and combined in a phase-sensitive detector to produce the error-signal characteristic shown in Fig.and azimuth-error signals are wanted. The two adjacent antenna feeds are connected to the two arms of a hybrid junction such as a "magic T. the outputs of the sum arm and the difference arm are each heterodyned The to an intermediate frequency and amplified as in any superheterodyne receiver. while both the sum pattern and the difference pattern are used on reception. The signal received with the difference The sum signal provides the range pattern provides the magnitude of the angle error. Left-hand diagrams in (a-c) are in polar coordinates. A block diagram of the amplitude-comparison-monopulse tracking radar for a single angular coordinate is shown in Fig. The sum and difference signals are multiplied in a phasesensitive detector to obtain both the magnitude and the direction of the error signal. 5. monopulse.4 An example of a simultaneous-lobing technique is amplitude-comparison monopulse. (b) sum pattern. or more simply. 5. (A cluster of four (a) (c) Angle (/>) (d) Fig. 5." a "rat race.8d. (d) product (error) signal. Amplitude-comparison Monopulse. 5. 5. right-hand diagrams are in rectangular coordinates. In this technique the RF signals received from two offset antenna beams are combined so that both the sum and the difference signals are obtained simultaneously. . 18 The sum and difference signals appear at the two other arms of the hybrid. and the difference in Fig.8a is shown in Fig.

duplexer is included in the sum arm for the protection of the receiver. A d > 0). The purpose of the phase-sensitive detector is to conveniently furnish the sign of the error signal. generating a pip. and the range output from the sum channel may be fed into an automatic-rangetracking unit. The sign of the difference signal (and the direction of the angular error) is determined by comparing the phase of the difference signal with the phase of the sum signal. The output of the phasesensitive detector. If the sum signal in the IF portion of the receiver were A s cos a> IF t. the sign of the difference signal may be measured by determining whether the difference signal is in phase with the sum signal or 180° out of phase. Range information is also extracted from the sum channel. The sum-channel signal operates the A-scope just as in a normal radar. — = + a part of the amplitude-comparison-monopulse by comparing the echo amplitudes from simultaneous offset beams. It gives an indication of the target range by deflecting the beam upward. the difference signal would be either A d cos a> 1F t or — A d cos a> lF t (A s > 0. The cluster of four feeds generates four partially overlapping antenna beams. depending on which side of center is the target. however.4] Tracking Radar is 177 connected to the sum arm. All four feeds generate the sum pattern. This presentation has been called the "Pisa" indicator. The angular-error signal may actuate a servo-control system to position the antenna. after the famous leaning tower. Block diagram of amplitude-comparison-monopulse radar (one angular coordinate).9 the output of the phase-sensitive detector (angle information) and the sum channel (range information) are shown presented on the A-scope. is Although a phase comparison is radar. 5. the angular-error signal basically derived A . 13 transmitter A Tronsmitter Sum TR chonnel Range IF Amplitude detector signal omplifier ^Hybrid junction Phose- LO sensitive detector Angleerror signal A-scope Antenno feeds Mixer IF amplifier Sum circuit Difference chonnel Sweep generator Fig. 5. The difference pattern in one plane is formed by taking the sum of two adjacent feeds and subtracting this from the sum of the other two adjacent feeds. The feeds might be used with either a parabolic reflector or a lens. The output of the phase-sensitive detector is an error signal whose magnitude is proportional to the angular error and whose sign is proportional to the direction. The amount of leaning is a measure of the magnitude of the angular error. 5.9. modifies the scope sweep to deflect the target pip either to the right or to the left. 5. depending upon the sign of the angular error.10. The difference pattern in theorthogonal plane is obtained by adding the differences of the orthogonal adjacent pairs. A block diagram of a monopulse radar with provision for extracting error signals in both elevation and azimuth is shown in Fig. The output of the monopulse radar may also be used to perform automatic tracking. In Fig.Sec. Since A d cos coIF t A d cos co IF (t tt). The phase relationship between the signals in the offset beams is not used.

with this technique. and bring the sum signal and the two sequence the time unscramble introduced to Phase detection occurs as in the conventional difference signals in time coincidence. 2 (5. Gaussian function. 13 the phase difference between channels must be maintained to within 25° or better for reasonably proper performance. the azimuth difference channel.) cosh a 2 6 Q 6 T (5.18) The one-way (voltage) difference pattern is 2G\ exp (e. Two phasesensitive detectors extract the angle-error information. 5. the other for elevation.. delayed signals difference two the followed by amplifier Most of the gain and gain control take place in the single IF amplifier. . in time by a suitable amount. one for each channel. Block diagram of two-coordinate (azimuth and elevation) amplitude-comparison-monopulse tracking radar. compensating delays are affect all three signals simultaneously. The one-way (voltage) pattern from one monopulse beam is % is the Gl exp (— aWjl). All three mixers operate from a single local oscillator in order to maintain the phase relationships between the three channels. where G is the maximum antenna gain.5° and ±| Monopulse Error Signal. + 0§o sinh a d a d. respectively.10. Range information is extracted from the output of the sum channel after amplitude detection. a = 2. Three separate mixers and IF amplifiers are shown. Since a phase comparison the difference channels. made between the output of the sum channel and each of important that the phase shifts introduced by each of the channels be almost identical. It is claimed that the phase and gain between channels have been mainsignal is tained to within ±2.776/0|. and 6% antenna beamwidth as measured between the half-power points.178 total Introduction to Radar Systems [Sec. and the elevation difference channel.4 of four hybrid junctions generate the sum channel.19) . db. An alternative approach to using three identical amplifiers in the monopulse receiver is to use but one IF channel which amplifies the sum 17 19 - signal and the two difference passed through the single IF The sum signals on a time-shared basis. is it is Amplitude detector Video amplifier - Range M^rj— o^litier} ^ Phase-sensitive detector Elevation- angle error IF amplifier Phase -sensitive detector Azimuthangle error LO Fig. The angular separation between the two antenna beams is 20a and the angle between the target and the axis Assuming no mutual coupling between the two feed horns. Assume that the antenna pattern is represented by the 19 one-way (voltage) sum pattern 2Gq exp is -| -a - (0* + (?*. the of symmetry is T . Any variations After amplification. 5. According to Page. monopulse. one for azimuth. The gains of the channels also must not differ by more than specified amounts.

the error signal is V„ = 4 exp [-a\d 2 q + d T )~] sinh a 6 q 6 r 2 2 2 (5. that is. The output of the phase-sensitive detector is a d-c voltage whose amplitude is proportional to the product of the sum and difference amplitudes. 1 8). especially for large signal-to-noise ratios. The slopes of the error signal at crossover in the monopulse. This is usually the case if accurate measurements are to be obtained.18) and (5. V = c 2 exp \_-a\6 2 a + 0|. The one-way (voltage) antenna pattern is represented analytically by the Gaussian function exp (—a 2 6 2 /2). (5.25) The error signal in the lobe-switching or the conical-scan radar (in one coordinate) is proportional to the difference of the two-way (voltage) antenna pattern exp (—aW). The receiver is assumed to be linear. [Eq. and the antenna gain is normalized. Thus the error signal in the monopulse radar is a linear function of the angular displacement of the target from the axis. The one-way (voltage) sum pattern in the monopulse radar is [from Eq.19).)] sinh 2a 6 Q d T 2 (5. (5.26) In a radar which transmits and receives on the difference pattern only.19)] is 2 The one-way 2 (voltage) difference pattern 2 exp L -f(^ + +6 r) sinh a 2 6 q 6 T (5.20)] are multiplied in the phase-sensitive detector to give the error signal.24) The monopulse error signal is [Eq. (5. the difference signal will be proportional to the product of IF difference signal (voltage) = 2Kexp [—a 2 (d\ + 0|)] X where sinh 2a d q 6 T cos 2 2-rrfIF t (5. (5.Sec.4] Tracking Radar 179 The IF voltage produced by Eqs. an assumption which should have little effect on the conclusions.18)] a2 = 2 exp "-{61 + 6%) cosh a 2 6 q 6 T (5. and a tracker which operates on the difference patterns only will be compared and used as a basis for relative accuracy. (5.21) with d%)~] q= 4] Vm = 4 exp [-2a 2 (6 2 cosh 2 a 2 6 Q Q T sinh 2a\6 T (5. (5.)] cosh 2 a 2 6 q 6 T sinh 2a 2 6 q 6 T (5. where the constant 2. G ~\.20) A' is a constant determined by the parameters of the radar equation. the more accurate is the measurement of angle. assuming small angular displacements.21) For small angular errors this reduces to Error signal = c2 6 T (5. The sum and difference signals [Eq.23) where 6q and d T are defined as before. 5.176/6%. or Error signal = cx 2 exp \_-2a 2 (6 q + #§. The two-way (voltage) antenna pattern for the sum channel is the square of Eq. The greater the signal-tonoise ratio and the steeper the slope of the error signal in the vicinity of zero angular error. the conical-scan tracker. where Comparison of Monopulse and Conical-scan Error Signals.27) .22) c x and c 2 are constants.

5. Since there is a physical limit to the minimum spacing between the feed horns.5 1 h 1 r. It should also be noted that more difficult to .7 L. Slope of the angular-error signal at crossover for monopulse and conical-scan tracking squint angle). the conical-scan radar occurs at a crossover of 2. however. 1 db. The maximum slope of the monopulse radar is seen to occur at a beam squint angle or the crossover point of the antenna patterns. it is 1 . The error-signal slope of the monopulse radar is slightly greater than the A = 5. Fig. achieve a low crossover with practical monopulse antennas than with conical-scan antennas. The (Vm V e. The maximum slope of The crossover values of The maximum of the curve is practical conical-scan radars are usually in this vicinity. radars (fi B half-power beamwidth. with some leeway allowed in the selection of the optimum. signals in practical conical-scan'and monopulse tracking radars will be comparable if the antenna pattern is the only factor of consequence. there will be a correspondingly lower limit to the separation between the two monopulse beams.11.3 0.30) T 'e T =o radar which utilizes the difference pattern for both transmitting and receiving does not have as suitable an error signal as the other two radars since its slope is zero for 6T 0. 5.10. 5.11. crossover of approximately In general. 0„ = slope of the conical-scan radar error signal over the range of squint-angle values of This is illustrated in Fig. However.29) Difference pattern (&) \dd =0 (5.2 db.1 5 0.6 0^ = 0. The monopulse radar usually generates its two (or four) overlapping beams from two (or four) adjacent feed horns.4 .4 \ 0.180 Introduction to Radar Systems slopes of these three error signals [Sec. The ordinate is the product of the error-signal slope times the antenna beamwidth. db 6 Fig. 2 Beam 4 3 crossover. the less will be the signal-to-noise ratio when the target is directly on axis. 5. the crossover level.0 Squint angle Bq /6s 0. and Vd) evaluated at 6T = are Monopulse te\ \au T le T =o =8a 2 e 9 exp(-2« 2 ^) =4a 2 e a exp(-a^) (5. The greater rather broad.28) Conical scan te) \aO T h T =o (5. and the abscissa is shown as either the practical interest.1 1 indicates that the jc/pssover can be much greater than the optimum and still Therefore the slope of the error result in a slope as high as %e?56onical-scan slope.2 0.

Wavefront phase relationships in phase-comparison-monopulse radar. The difference in amplitudes in the several antenna positions was proportional to the angular error. while the monopulse technique used two or more simultaneous beams.31) .11 Tracking Radar 181 is the product of the slope times the beamwidth. The angle of arrival (in one coordinate) may also be deterTarget mined by comparing the phase difference between the signals from two separate antennas. The distance to the target is R and is assumed large compared with the antenna separation d. The measurement of angle of arrival by comparison of the phase relationships in the signals Antenna from the separated antennas of a radio interNo. In actual practice the sum (or difference) pattern may not be the sum (or difference) of the two overlapping offset patterns. Unlike the antennas of amplitude-comparison trackers.12 two antennas are shown separated by a distance d. 5. but the phases are different. 5.Sec. The latter term is the one which will be used here. The tracking techniques discussed thus far in this chapter were based on a comparison of the amplitudes of echo signals received from two or more antenna positions. the larger will be the slope and the better the tracking accuracy. A tracking radar which operates with phase information is similar to an active interferometer and might be called an interferometer radar. those used in phase-comparison systems are not offset from the axis. 20 21 ' Phase-comparison Monopulse. the smaller the beamwidth of the antenna.1 ferometer has been widely used by the radio astronomers for precise measurements of the positions of radio stars. or phase-comparison monopulse. word of caution should be given concerning the nature of A the Gaussian approximation to the antenna beam shape assumed above. 5. time-shared antenna beam. The individual boresight axes of the antennas are parallel. The interferometer as Fig. The amplitudes of the target echo signals are essentially the same from each antenna beam. In Fig. 5. It has also been called simultaneous-phase-comparison radar.12. The distance from antenna 1 to the target is volume Rx =R+ target - sin i and the distance from antenna 2 to the is sin 6 The phase difference between the echo signals in the two antennas is approximately A<£ = — dsind A (5. the source of energy being radiated by the target itself. There will usually be some interaction between the two feed horns which can alter the two patterns. The sequential-lobing and conical-scan techniques used a single.4] since the ordinate in Fig. used by the radio astronomer is a passive instrument. causing the (far-field) radiation to illuminate the same in space. The line of sight to the target makes an angle 6 to the perpendicular bisector of the line joining the two antennas.

in front of the other receiver. Many of the grating lobes will be suppressed if directive antenna elements are used rather than the omnidirectional = = ^ . another to the azimuth receiver.13). First. The RF echo in the two channels. the phase comparison radar does not usually make efficient use of These two points are elaborated upon below. one of which is connected to the transmitter and receiver as in a conventional The transmitter is shown conradar. grating lobes of the pattern will be of the same amplitude.182 Introduction to Radar Systems [Sec. An additional antenna and receiving channel is necessary in order to track in two orthogonal coordinates.31) (where n is an integer) and lobes may be found by setting Aci The grating lobes appear when 0. (5. the total available antenna aperture. a multilobed pattern will be formed. 22 four antennas were arranged in a square to obtain tracking in both elevation and azimuth. The lobes of the pattern are called grating lobes.31). When two omnidirectional antennas are separated by distances of many wavelengths. while the other three were receivers. There are two reasons why this might be so. These are directive antennas. been have Although tracking radars based upon the phase-comparison principle widely as has not been technique this satisfactorily. The main lobe corresponds to n n and result in an ambiguous angle measurement. the sidelobe levels which result can sometimes be higher than those from a single reflector. and one to a common receiver which supplied the reference for both the Instead of obtaining the error signal from a phase elevation and azimuth receiver. Radiation patterns Mixer IF amplifier Local oscillator Phase detector Angle-error information Duplexer Mixer Transmitter IF amplifier Envelope detector ^_ Range information Fig.13. One antenna was connected to the elevation receiver. One of these antennas was a transmitter only. This voltage is used as the error-signal input to a servo-control loop which positions the antenna to make the error signal zero. In practice. as are the separations in the phase-comparison monopulse.4 For small angles where sin 6 «* 0. Block diagram of phase-comparison-monopulse radar (one angular coordinate).13. One of the receiving channels is envelope-detected. but to balance the phase shifts The two receiving channels should be identical. (5. a second duplexer might be inserted nected to the antenna via a duplexer. built and found to track aircraft used as some of the others discussed. as in the normal radar receiver. the sum and difference signals may be derived (as in the in Fig. A block diagram of a phase-comparison monopulse22 in one angular coordinate is shown in Fig. comparison (as 5. the phase difference is a linear function of the angular error and may be used to position the antenna via a servo-control loop. 5. by analogy with the They have also been called principal maxima. signals are heterodyned to an intermediate frequency with a common local oscillator. 16 amplitude-comparison monopulse) and compared in a phase-sensitive detector. not so much for protection. Each of the optical diffraction grating. to extract the range information. and second. while the other antenna feeds a receiver only. The outputs of the two IF amplifiers are compared in a phase detector whose output is a voltage proportional to A</> of Eq. Two antennas are shown side by side. solving for d. The positions of the grating 2nn in Eq. 5. 5. In one implementation of the phase-comparison monopulse radar.

on the other hand. that is. A factor of 4 in the effective antenna area can result in a factor-of-2 change = = =± = in range. = = = = 65A/D. or ±3.13.9°. Both the amplitude-comparison-monopulse and the phase-comparison-monopulse employ two antenna beams (for one coordinate tracking). with one antenna for transmission and the other three for reception.2 B and the first sidelobe approximately half a beamwidth farther. and the second grating lobe at d occur at 6 multiplies the .4] Tracking Radar The element pattern 183 elements assumed in the above illustration. With such close spacing the phase difference between the signals received in the two feeds is negligibly small. the first the half-power points correspond to 6 ±2. null for parabolic-reflector antennas occurs at approximately ±1. Therefore we see that the position of the first grating lobe lies between the half-power point and Depending upon its exact location. In the amplitude-comparison monopulse the two beams are offset. 33 The servo loop adjusts the phase shifter until the difference in phase between trackers .09°.31)] will The first grating lobe [n 1 or A</> just touching. The phase-comparison monopulse. Assuming that the ±1. The phase-comparison-monopulse tracking radar described above is but one method of employing phase information. measures phase differences only and is not concerned with amplitude difference. suppose that four parabolic reflectors were used to achieve tracking in two coordinates. but are directed to illuminate a common volume in space. One limitation of the phase-comparison-monopulse tracker described in Fig. For example. 5. assume that the antenna-reflector size is 30 wavelengths in diameter and that the separation between the antennas is also 30 wavelengths that is. 5. Since the feeds may be placed side by side. they could be as close as one-half wavelength. interference pattern of the two separated omnidirectional elements. The second grating lobe occurs create a shoulder lobe. The directivity of the element pattern will tend to reduce the magnitude of the grating lobes. its effective aperture might be as much as four times greater than if it were used with the type of phase-comparison monopulse shown in Fig.8°.Sec. if an amplitude-comparison-monopulse antenna or a conical-scan antenna occupied the same area as the four antennas. In other words. Any difference in the amplitudes between the two antenna outputs in the amplitude-comparison system is a result of differences in amplitude and not phase. As an example of the positions of the grating lobes relative to the element pattern. Separate antennas are needed since it is difficult to illuminate a single reflector with more than one feed and produce independent antenna patterns which illuminate the same volume in space. This type of pattern may be generated by using one reflector dish with two feed horns side by side (four feed horns for two coordinate data). The measurements made by the two systems are not the same. it may either widen the main lobe or the first null. Therefore the antenna beams are not offset. point in slightly different directions. 1 . The effective antenna area (or the gain) that is substituted into the radar equation is that of one of the antennas. where DjX half-power beamwidth of the element pattern is given by dB 30. consequently. or even a pronounced sidelobe. the two antennas are 2n radians in Eq. with the result that those lobes outside the coverage of the element pattern will be reduced. not that of all four together.13 is that it does not use its available antenna aperture as efficiently as other types of tracking radars. the resultant pattern is the product of the element pattern times the array pattern. in the vicinity of the first sidelobe of the element pattern. the characteristics of the antenna beams will also be different. the reduction will not be as complete as might be desired. In one embodiment of the phase-comparison principle as applied to missile guidance the phase difference between the signals in two fixed antennas is measured with a servo-controlled phase shifter located in one of the arms. but in general. (5. ±3.65°. 5. Therefore. with the likelihood that the first sidelobe level will be raised.74°. As a rough rule of thumb.

184
the

Introduction to Radar Systems

[Sec. 5.5

two channels is a null. The amount of phase shift which has to be introduced to make a null signal is a measure of the angular error. The phase- and amplitude-comparison principles can be combined in a single radar to produce two-dimensional angle tracking with only two, rather than four, antenna beams. 17 The angle information in one plane (the azimuth) is obtained by two separate antennas placed side by side as in a phase-comparison monopulse. One of the beams is tilted slightly upward, while the other is tilted slightly downward, to achieve the squint needed for amplitude-comparison monopulse in elevation. Therefore the horizontal projection of the antenna patterns is that of a phase-comparison system, while the vertical projection is that of an amplitude-comparison system.
5.5. Target-reflection Characteristics

and Angular Accuracy

will be influenced by such factors as the mechanical properties of the radar antenna and pedestal, the method by which the angular position of the antenna is measured, the quality of the servo system, the stability of the electronic circuits, the noise level of the receiver, the antenna beam width, atmospheric fluctuations, and the reflection characteristics of the target. These factors can degrade the tracking accuracy by causing the antenna beam to fluctuate in a random manner about the true target path. These noiselike fluctuations are sometimes called tracking noise, or jitter. In many cases the two factors which ultimately limit the angular accuracy of practical tracking radars are the mechanical errors and the targetreflectivity characteristics. The mechanical errors associated with tracking radars will not be discussed here. (An example of the mechanical errors experienced in a precise monopulse tracking radar, the AN/FPS-16, has been described by Barton. 24 25) A simple radar target such as a smooth sphere will not cause degradation of the angular-tracking accuracy. The radar cross section of a sphere is independent of the aspect at which it is viewed ; consequently, its echo will not fluctuate with time. The same is true, in general, of a radar beacon if its antenna pattern is omnidirectional. However, most radar targets are of a more complex nature than the sphere. The amplitude of the echo signal from a complex target may vary over wide limits as the In addition, the effective center of radar aspect changes with respect to the radar. Both of these effects amplitude fluctuations and wanderreflection may also change. ing of the radar center of reflection as well as the limitation imposed by receiver noise can limit the tracking accuracy. These effects are discussed below. Amplitude Fluctuations. A complex target such as an aircraft or a ship may be considered as a number of independent scattering elements. The echo signal can be represented as the vector addition of the contributions from the individual scatterers. as might occur because of motion If the target aspect changes with respect to the radar of the target, or turbulence in the case of aircraft targets the relative phase and amplitude relationships of the contributions from the individual scatterers also change. Consequently, the vector sum, and therefore the amplitude, change with changing target
-

The angular accuracy of tracking radar

aspect.

Amplitude fluctuations of the echo signal are important in the design of the lobeswitching radar and the conical-scan radar but are of little consequence to the monopulse tracker. Both the conical-scan tracker and the lobe-switching tracker require a This time corresponds in the finite time to obtain a measurement of the angle error. With lobe conical-scan tracker to at least one revolution of the antenna beam. switching, the minimum time is that necessary to obtain echoes at the four successive angular positions. In either case a minimum of four pulse-repetition periods are If required to make a measurement; in practice, many more than four are often used. the target cross section were to vary during this observation time, the change might be erroneously interpreted as an angular-error signal. The monopulse radar, on the

r

r

r

Sec. 5.5]

Tracking Radar

185

other hand, determines the angular error on the basis of a single pulse. Its accuracy will therefore not be affected by changes in amplitude with time. The echo signal from complex targets is best described in statistical terms. Some of the more useful statistical descriptions that have been applied to cross sections are the cumulative probability distribution, the autocorrelation function, and the power
spectral density.
effect

The power-spectral-density function

is

useful for describing the

of amplitude fluctuations on the performance of a conical-scan or lobe-switching

tracker.

typical power spectrum of the target amplitude fluctuations (fading) with a conical-scan tracking radar might appear as in Fig. 5.14. This curve is an analytical approximation to the experimental spectrum derived from 30 sec of azimuth data

A

T

— —— —
I
I

I

i

n

———
i i

1
12

I-

18

20

Frequency, cps

Fig. 5.14. Power spectral density of amplitude fluctuations for a C-47 aircraft on a crossover course. (Courtesy J. E. Ward and the MIT Servomechanism Laboratory. 42 )

obtained from a radar tracking a C-47 aircraft flying a crossover course. 42 The minimum range was 300 yd, and the maximum range was 5,000 yd. The autocorrelation function (which is the Fourier cosine transform of the power spectrum) corresponding to the spectrum of Fig. 5.14 is <f>(r) 1,410 exp (— 16.6t), where <f(r) is in For a radial trajectory over the same range limits the autounits of square mils. correlation function is </>(t) 40 exp (— 13t). To reduce the effect of amplitude noise on tracking, the conical-scan frequency should be chosen to correspond to a low value of amplitude noise. If considerable amplitude fluctuation noise were to appear at the conical-scan or lobe-switching frequencies, it could not be readily eliminated with filters or AGC. A typical scan frequency might be of the order of 30 cps. Higher frequencies might also be used since target amplitude However, this may not always be noise generally decreases with increasing frequency. Propeller-driven aircraft produce modulation components at the blade frequency true. and harmonics thereof and can cause a substantial increase in the spectral energy density at certain frequencies. 43 Also, the scan frequency cannot be made higher than one-quarter the pulse repetition frequency if a minimum of one hit per quadrant is to be obtained. It has been found experimentally that the tracking accuracy of radars operating with pulse repetition frequencies from 1,000 to 4,000 cps and a lobing or scan rate one-quarter of the prf are not limited by echo amplitude fluctuations. 26

=

=

186

Introduction to Radar Systems
signal

[Sec. 5.5

The percentage modulation of the echo
tions
is

due to cross-section amplitude

fluctua-

independent of range if AGC is used. Consequently, the angular error as a result of amplitude fluctuations will also be independent of range. Angle Fluctuations. 2 32 Changes in the target aspect with respect to the radar can cause the apparent center of radar reflections to wander from one point to another. (The apparent center of radar reflection is the direction of the antenna when the error signal is zero.) In general, the apparent center of reflection might not correspond to the target center. In fact, it need not be confined to the physical extent of the target and may be off the target a significant fraction of the time. The random wandering of the apparent radar reflecting center gives rise to noisy or jittered angle tracking. This form of tracking noise is called angle noise, angle scintillations, angle fluctuations, or target The angular fluctuations produced by small targets at long range may be of glint. little consequence in most instances. However, at short range or with relatively large targets (as might be seen by a radar seeker on a homing missile), angular fluctuations may be the chief factor limiting tracking accuracy. Angle fluctuations affect all tracking radars whether conical-scan, sequential-lobing, or monopulse. Consider a rather simplified model of a complex radar target consisting of two independent isotropic scatterers separated by an angular distance d D as measured from the radar. Although such a target may be fictitious and used for reasons of mathematical simplicity, it might approximate a target such as a small fighter aircraft with wing-tip tanks or two aircraft targets flying in formation and located within the same radar resolution cell. It is also a close approximation to the low-angle tracking problem in which the radar sees the target plus its image reflected from the surface. The qualitative effects of target glint may be assessed from this model. The relative, amplitude between the cross sections of the two scatterers is assumed to be a, and the relative phase difference is a. Differences in phase might be due to differences in range or to reflecting properties. The cross-section ratio a is defined as a number less than unity. The angular error A0 as measured from the larger of the two targets is 27

^

,

——
This
to
is

plotted in Fig. 5.15.

1 D The position of the larger of the two

+ a cos + a + 2a cos a
Q
<x

2

2

scatterers corresponds
1
.

while the smaller-scatterer position is at A0/0 7J Positive values of AS correspond to an apparent radar center which lies between the two scatterers; negative values lie outside the target. When the echo signals from both scatterers are in phase (a 0), the error reduces to a/(a 1), which corresponds to the so-called "center of gravity" of the two scatterers (not to be confused with the mechanical center
0,

A0/0Q

=

=+

=

+

of gravity).

Angle fluctuations are due to random changes in the relative distance from radar to is, varying values of a. These changes may result from turbulence in the aircraft flight path or from the changing aspect caused by target motion. In essence, angle fluctuations are a distortion of the phase front of the echo signal reflected from a complex target and may be visualized as the apparent tilt of this phase front as it
the scatterers, that
arrives at the tracking system.

Equation (5.32) indicates that the tracking error A6 due to
target
is

glint for the two-scatterer
.

This is probably a reasonable approximation to the behavior of real targets, provided the angular extent of the target is not too large compared with the antenna beamwidth. Since Q D varies inversely with distance for a fixed target size, the tracking error due to glint also varies inversely with distance.

directly proportional to the angular extent of the target Q D

A slightly more complex model than the two-scatterer target considered above is one
consisting of

many

individual scatterers, each of the

same cross

section,

arranged

Sec. 5.5]

Tracking Radar

187

uniformly along a line of length L perpendicular to the line of sight from the radar. The resultant cross section from such a target is assumed to behave according to the Rayleigh probability distribution. The probability of the apparent radar center lying outside
is the radians (in one tracking plane) is 0. 1 34, where 13.4 per cent of the time the radar will not be directed to a point on the target. Similar results for a two-dimensional model consisting of equal-cross-section scatterers uniformly spaced over a circular area indicate that the

the angular region of

LjR

R

distance to the target. 28

Thus

probability that the apparent radar center

lies

outside this target

is

0.20.

reduced by increasing the time constant of the AGC system (reducing the bandwidth). 26 33 34 However, this reduction in angle fluctuation is accompanied by a new component of noise caused by the amplitude fluctuations associated with the echo signal; that is, narrowing the AGC bandwidth

Angle fluctuations

in a tracking radar are

-

'

20

40

60

100 120 80 Phase difference ex

140

160

180

Fig. 5.15. Plot of Eq. (5.32).

amplitude a and relative phase

shift a,

Apparent radar center A0 of two isotropic separated by an angular extent 6 D
.

scatterers of relative

generates additional noise in the vicinity of zero frequency, and poorer tracking results. Amplitude noise modulates the tracking-error signals and produces a new noise component, proportional to true tracking errors, that is enhanced with a slow AGC. Under practical tracking conditions it seems that a wide-bandwidth (short-time constant)

should be used to minimize the over-all tracking noise. However, the servo bandwidth should be kept to a minimum consistent with tactical requirements in order to minimize the noise. Receiver and Servo Noise. Another limitation on tracking accuracy is the receiver The accuracy of the angle measurement is inversely proportional to the noise power. 35 Since the signal-to-noise ratio is square root of the signal-to-noise power ratio. 4 the angular error due to receiver noise proportional to l/# (from the radar equation),
proportional to the square of the target distance. Servo noise is the hunting action of the tracking servomechanism which results from backlash and compliance in the gears, shafts, and structures of the mount. The magnitude of servo noise is essentially independent of the target echo and will therefore be independent of range. 26 33 Summary of Errors. The contributions of the various factors affecting the tracking Angle-fluctuation noise varies inversely with range error are summarized in Fig. 5.16. receiver noise varies as the square of the range; and amplitude fluctuations and servo
is
'

AGC

noise are independent of range.

This

is

a qualitative plot showing the gross effects of

188

Introduction to Radar Systems

[Sec. 5.5

Curve A is the sum of and sequential-lobing tracking radars. Curve B does not include the amplitude fluctuations and is therefore representative of monopulse radars. In Fig. 5. 1 6 the amplitude fluctuations are assumed to be larger than servo noise. If not, the improvement of monopulse tracking over conical scan will be negligible. In general, the tracking accuracy deteriorates at both short and long target ranges, with the best tracking occurring at some intermediate range.
each of the factors.
all effects

Two different resultant curves are shown.

and

is

representative of conical-scan

1

1

1

1

1

1

1

1

1

1

1

1

hi
1

i

i

1

1

1

1

1

l

/

-

-

/° h

© ^ //
E
$0.01

Amplitude
fluctuations

^y. /
'^

Servo noise

III
i

1

1

1

1

1

1

i

I

i

I

i

10

nl 100

I

i

1

1

1

II

1,000

Relative radar range

Fig. 5.16. Relative contributions to angular tracking error due to amplitude fluctuations, angle and servo noise as a function of range. (A) Composite error for a conicalscan or sequential-lobing radar; (B) composite error for monopulse.
fluctuations, receiver noise,

At sufficiently long ranges, the signal-to-noise ratio may be too low to permit satisfactory tracking and the radar "loses track." Swerling analyzed the effect of receiver noise on the tracking performance using the loss rate as a criterion of performance, defined as the expected number of times per second the tracking error (in either range or angle) exceeds the maximum allowable value. 36 The loss rate can serve as a criterion
an optimum choice of servo parameters, transmitter power, maximum range, and other similar tracking-radar parameters. Swerling's analysis applies to either monopulse angle tracking or split-range-gate tracking (described in the next section). Three of the formulas derived by Swerling are presented below (in his notation). The output signal-to-noise (power) ratio Y versus loss rate A is
to find (

Kt Y( \2dJ ll

1

-

5/6

n Vln° J A

5/c

(5.33)

where

K=
t

=

a correction factor of order of unity and accounts for type of trackingerror circuit employed, distortion of pulse in IF, and particular type of linear approximation used instead of actual error vs. voltage curve

width between two-way half-power points of each or width of each range gate (range tracking)

beam

(angle tracking)

Sm 3

= maximum allowable tracking error = average tracking error f = equivalent square-band cutoff frequency of the servo regarded as an audio
e

filter

Sec

56]
relationship between the output signal-to-noise ratio

Tracking Radar

189

The

Y and

the IF signal-to-noise

ratio

X for square-law detection is

y=J^ +2X
1
is

(5.34)

where Nis the effective number of pulses integrated by the servo and The variance of the tracking error is bandwidth divided by 2fc
.

equal to the IF

o\

«

^
87

(5.35)

This is consistent with the form of the theoretical errors derived in Chap. 10 for other radar measurements. The greater the beamwidth (or the pulse width), the poorer will be the angle (or the range) accuracy. The rms tracking error (square root of the variance) is inversely proportional to the square root of the signal-to-noise
ratio.

5.6.

Tracking in Range

In most tracking-radar applications the target is continuously tracked in range as well Range tracking might be accomplished by a human operator who watches as in angle. an A-scope or J-scope presentation and manually positions a handwheel in order to maintain a marker over the desired target pip. The setting of the handwheel is a measure of the target range and may be converted to a voltage that is supplied to a data processor. The data processor in a fire-control radar predicts the future position of
the target for the purpose of aiming the weapon. The human operator tracking a target by positioning a handwheel can be considered 1 In pure displacement tracking, the turns of the handwheel are as part of a servo loop. made proportional to the displacement of the target. If the target's range changes at a constant rate, the operator must turn his handwheel at a constant rate. If he is lagging

behind the target, he will turn faster until the error is corrected if he is leading the In pure rate tracking, the position of the handwheel target, he will turn more slowly. determines the speed at which the movable marker on the CRT follows the target pip. When tracking a target moving with constant velocity the handwheel need not be turned once the proper adjustment has been made. Displacement and rate tracking may be combined so that the handwheel position automatically corrects for speed at the same time that the displacement error is corrected. This is called aided tracking. Aided tracking may also be used for manual tracking in
;

angle as well as range. As target speeds increase,

it is increasingly difficult for an operator to perform at the of efficiency over a sustained period of time, and automatic tracking becomes a necessity. Indeed, there are many tracking applications where an operator has no place, as in a homing missile or in a small space vehicle. The technique for automatically tracking in range is based on the split range gate.f Two range gates are generated as shown in Fig. 5.17. One is the early gate, and the other is the late gate. The echo pulse is shown in Fig. 5.17a, the relative position of

necessary levels

The the gates at a particular instant in Fig. 5.176, and the error signal in Fig. 5.17c. portion of the signal energy contained in the early gate is less than that in the late gate. If the outputs of the two gates are subtracted, an error signal (Fig. 5.17c) will result 37 The magnitude of the error which may be used to reposition the center of the gates. signal is a measure of the difference between the center of the pulse and the center of the
portions of a wave which exist during one or more selected t Gating is the process of selecting those time intervals (IRE definition").

>

190

Introduction to Radar Systems

[Sec. 5.7

The sign of the error signal determines the direction in which the gates must be gates. repositioned by a feedback-control system. When the error signal is zero, the range gates are centered on the pulse. The range gating necessary to perform automatic tracking offers several advantages as by-products. It isolates one target, excluding targets at other ranges. This permits
the boxcar generator to be employed. Also, range gating improves the signal-to-noise ratio since it eliminates the noise from the other range intervals. Hence the width of the gate should be sufficiently narrow to minimize extraneous noise. On the other hand, it must not be so narrow that an appreciable fraction of the signal energy is excluded. reasonable compromise is to make the gate width of the order of the

A

pulse width.
Echo pulse
j

U)

/
Early gate

n
Late

Time

(i>)

gate

Time

*•

Early gate I

(c)

signal

/

Late gate

Time

—*

/
Fig. 5.17. Split-range-gate tracking,
(a)

signal

Echo

pulse; (b) early-late range gates; (c) difference signal

between early and

late

range gates.

of finite length can cause noise in range-tracking circuits in an analogous to angle-fluctuation noise (glint) in the angle-tracking circuits. Rangetracking noise depends on the length of the target and its shape. It has been reported 26

A target

manner

that the

rms value of the range noise
is

tracking
5.7.

is approximately 0.8 of the target length when accomplished with a video split-range-gate error detector.

Tracking in Doppler

Tracking radars designed to extract doppler information, such as the or the pulse-doppler tracking radars, can also track the doppler frequency shift. This may be accomplished with a frequency discriminator and a tunable oscillator. Other techniques are, of course, possible. 44 Tracking the doppler frequency shift with a narrowband doppler filter (one which is wide enough to encompass the frequency spectrum occupied by the signal energy) offers two advantages: (1) the signal-to-noise ratio is improved, especially if the doppler frequency shift is large compared with the information bandwidth of the received signal; and (2) it may be used to resolve a desired target from a group of targets, especially in or pulse-doppler tracking radars.

CW

CW

5.8. Acquisition

tracking radar must first find and acquire its target before it can operate as a tracker. Therefore it is usually necessary for the radar to scan an angular sector in which the presence of the target is suspected. Most tracking radars employ a narrow pencilbeam antenna. Searching a volume in space for an aircraft target with a narrow pencil beam would be somewhat analogous to searching for a fly in a darkened auditorium with a flashlight. It must be done with some care if the entire volume is to be covered

A

Sec. 5.8]

Tracking Radar

191

uniformly and efficiently. Examples of the common types of scanning patterns employed with pencil-beam antennas are illustrated in Fig. 5.18. In the helical scan, the antenna is continuously rotated in azimuth while it is simultaneously raised or lowered in elevation. It traces a helix in space. Helical scanning was employed for the search mode of the SCR-584 fire-control radar, developed during World War II for the aiming of antiaircraft-gun batteries. 38 The SCR-584 antenna was rotated at the rate of 6 rpm and covered a 20° elevation angle in 1 min. The Palmer scan derives its name from the familiar penmanship exercises of grammar school days. It consists of a rapid circular scan (conical scan) about the axis of the antenna, combined with a linear movement of the axis of rotation. When the axis of rotation is held Because of this property, the stationary, the Palmer scan reduces to the conical scan. Palmer scan is sometimes used with conical-scan tracking radars which must operate with a search as well as a track mode since the same mechanisms used to produce conical
scanning can also be used for Palmer scanning. 39 Some conical-scan tracking radars increase the squint angle during search in order to reduce the time required to scan a given volume. The conical scan of the SCR-584 was operated during the search mode and was actually a Palmer scan in a helix. In general, conical scan is performed during the search mode of most tracking radars. The Palmer scan is suited to a search area which is larger in one dimension than another. The spiral scan covers an angular search volume with circular symmetry. Both the spiral scan and the Palmer scan suffer from the disadvantage that all parts of the scan volume do not receive the same energy unless the scanning speed is varied during the scan cycle. As a consequence, the number of hits returned from a target when searching with a constant scanning rate depends upon the position of the target within
the search area.

(/>)

U)

=P
(d)
(e)
(a) Trace of helical scanning Fig. 5.18. Examples of acquisition search patterns, scan; (c) spiral scan; (d) raster, or TV, scan; (e) nodding scan.

beam;

(b)

Palmer

raster, or TV, scan, unlike the Palmer or the spiral scan, paints the search area uniform manner. The raster scan is a simple and convenient means for searching Similar to the raster scan is the nodding scan a limited sector, rectangular in shape. rapidly in elevation and slowly in azimuth. beam produced by oscillating the antenna Although it may be employed to cover a limited sector as does the raster scan nodding scan may also be used to obtain hemispherical coverage, that is, elevation angle extending to 90° and the azimuth scan angle to 360°. The helical scan and the nodding scan can both be used to obtain hemispheric coverage with a pencil beam. The nodding scan is also used with height-finding radars.

The

in a

The Palmer,

spiral,

and

assist in the acquisition

raster scans are employed in fire-control tracking radars to of the target when the search sector is of limited extent.

192
5.9.

Introduction to Radar Systems
Examples of Tracking Radars

[Sec. 5.9

The major characteristics of three tracking radars will be presented for the purpose of illustration. The three trackers are (1) the SCR-584, (2) the MIT Lincoln Laboratory Millstone Hill radar, and (3) the AN/FPS-16. The SCR-58438 was the first successful operational tracking radar at microwave frequencies (Fig. 5.19). It was developed by the MIT Radiation Laboratory and became available in operational quantities during the latter half of World War 1 1. Its function was to provide the fire-control information necessary for operating a battery of four 90-mm antiaircraft guns. The SCR-584 was the first to use the conical-scan

Fig. 5.19.

SCR-584 tracking

radar.

(Courtesy McGraw-Hill Book Company, Inc.)

tracking technique. Its basic principle of operation was not too unlike that of modern conical-scan trackers. The radar was also designed to operate in a search mode to provide its own acquisition information. When searching, the beam scanned a helical pattern with 360° azimuth coverage and a reasonable amount of elevation coverage.

The

When

target information obtained during the search phase was displayed on a PPL a suitable target was found, the search pattern was stopped and the antenna was

positioned to acquire the target. The target-tracking data supplied by the radar were processed in an analog computer which smoothed the data, predicted the target's future position, and computed the lead angle for the guns. The output information actuated a servo system that positioned the guns according to orders from the computer. The SCR-584 was designed originally to operate at 5 band. An A"-band version was also produced. list of the parameters of the 5-band version is presented in Table 5.1. The SCR-584 considerably improved the capabilities of antiaircraft artillery when it was introduced during World War II. Although it was not the first fire-control radar used by the military for aiming antiaircraft guns, its accuracy and especially its angular resolution were superior to the VH F and U HF radars then in use. Its introduction was particularly important in World War II since the Germans had devised electronic

A

countermeasures against the existing SCR-268 tracking radars but did not have the

Sec. 5.9]

Tracking Radar

193

means for jamming the microwave frequencies. After the war, ready availability of tracking the SCR-584 made it popular as an instrumentation radar for drone or missile modified radar. A flexible a requiring and for research and development programs It employs a version of the SCR-584 is the AN/MPQ-12, also listed in Table 5.1. The Army M-33 fire-control radar that larger antenna and has more range capability. replaced the SCR-584 can also be considered of comparable performance, although of
different design.

radar is the MIT Lincoln Laboratory radar (Fig. 5.20) large conical scanning 40 This radar is similar in principle to the located on Millstone Hill in Westford, Mass. because its application is different. different SCR-584 tracker, but in detail it is quite

A

UHF

Table
Characteristic

5.1.

Comparison of Trackers

SCR-584
Conical scan 6

AN/MPQ-12
Conical scan 10
12

AN/FPS-16
Monopulse

Millstone Hill

Type of tracking Antenna size, ft
Frequency Beamwidth, deg Antenna gain, db Power: Peak Average Pulse width
Prf, cps

Sband
4
33

5 band
2.4

Cband
1.2

Conical scan 84 (440 Mc)

UHF
2.1

37

44.5

37.5
2.5

250 kw 340 watts
0.8 /isec

250

kw

1

Mw

Mw

lkw
0.25 ,usec
0.25,0.5, l.OjUsec

Receiver noise figure, db Receiver bandwidth, Mc

1,707 15
1.7

364-1,707 12

160-1,707
11

150 kw 2 msec 30 2

8.0 or 1.6

Accuracy

Range
Angle Range on 1-m 2
target,

20 yd 2 mils

10 yd

5

yd
mil

8

km

lmil

0.1

0.2°

nautical miles: For detection For accurate track

30
.

70
35

.

15

180 120

2,000

The Millstone Hill radar was designed to track objects such as satellites, missiles, the moon, and similar objects beyond or within the earth's atmosphere. Its parabolic
antenna is 84 ft in diameter and is supported 90 ft above the ground. Heavy-duty drive motors permit the antenna to track in azimuth and elevation at the rate of 4°/sec. A turnstile junction located just behind the circular feed horn allows the transmitted The turnstile also signal to have any polarization, from linear to circular to elliptical. necessary when feature is (This polarizations. permits receiving on two orthogonal rotation; see Faraday because of ionosphere the through propagated is energy UHF The size of waveguide used at this frequency is 10i by 21 in. Sec. 14.2.) The transmitter consists of two high-power klystron amplifiers operating in parallel,
similar to the

X626

described in Sec. 6.3.

Two identical receivers process the returned

echo

one for each orthogonal polarization. The signal in each receiver is divided into two channels. One channel contains a matched filter bank from which The other channel contains a digital-range and doppler information is extracted. coherent (phase) detector of wide dynamic range for extracting the phase and amplitude The radar output data are in digital form for characteristics of the returned signal. computer. digital transistorized the CG-24 in processing
signal,

example of a monopulse tracking radar is the AN/FPS-16 (Fig. 5.21), the charac15 25 The FPS-16 is an instrumentation teristics of which are tabulated in Table 5.1.

An

-

The trailerradar, designed especially for precision tracking of guided missiles. mounted mobile version is the AN/MPS-25. Angular accuracy of the FPS-16 after

194

Introduction to Radar Systems

[Sec. 5.9

Fig. 5.20.

MIT

Lincoln Laboratory Millstone Hill radar.

(Courtesy

MIT

Lincoln Laboratory.)

„.j-

"i"

A

f o :«*i..* I&f&ter
.

WrSfceaTWAfc..

Fig. 5.21.

AN/FPS-16

tracking radar.

(Courtesy Radio Corporation of America.)

Sec. 5.10]

Tracking Radar
1°.)

195

(A mil is one-thousandth of a radian, or is 0. 1 mil. This is an order of magnitude better than that of the SCR-584. For the achievement of an accuracy as good as 0.1 mil, careful designing was required. The four-horn monopulse feed is supported in front of the reflector by four invar rods. The antenna is 12 ft in diameter, and the entire azimuth turntable rotates on a ballbearing race 60 in. in diameter. The mechanical resonance on the entire structure is above 15 cps, permitting a closed-loop servo response of 5 cps. The radar is painted with a white, heat-reflecting paint to minimize mechanical errors caused by temperature gradients induced by solar radiation. Tracking in azimuth can be accomplished at a rate of 40°/sec. The elevation tracking rate is 30°/sec. The AN/FPQ-6, an improved version of the AN/FPS-16, has a 29-ft-diameter Cassegrain antenna and radiates 3
correction for propagation effects
17.4 mils

=

Mw

peak power
5.10.

at

C band. 45

Comparison of Trackers
four continuous-tracking-radar techniques that have been discussed (sequen-

Of the
tial

and phase-comparison monopulse), conical-scan and amplitude-comparison monopulse probably have seen more application than the other two. The phase-comparison monopulse has not proved too popular. It does not make efficient use of the available antenna aperture, and the sidelobe level might sometimes be higher than if a single antenna were used. Sequential lobing is similar to conical-scan tracking. Conical-scan tracking seems to be preferred in most applications because it is usually more practical to implement than Therefore, in this section, only the conical-scan radar and the sequential lobing. amplitude-comparison monopulse will be compared. (The latter will be referred to simply as monopulse.) The comparison of the monopulse and conical-scan trackers is made on the bases of detectability, tracking accuracy, and complexity. There is little significant difference between the detection capability of conical-scan and monopulse trackers when the major parameters of the radar (those which appear Both radar techniques result in some slight loss of in the radar equation) are the same. antenna gain over a nontracking antenna of the same size because of the offset antenna
lobing, conical scan, amplitude-comparison monopulse,

beams.

The tracking accuracy of the conical-scan radar is degraded

if

the target cross section

Amplitude fluctuates in amplitude at frequencies at or near the conical-scan frequency. However, wandering of fluctuations have essentially no effect on monopulse radar.
the apparent radar center (glint) increases the tracking error of both types of radars. For equal signal-to-noise ratios and antenna beamwidths, the tracking accuracy of the two systems should be comparable in the absence of amplitude fluctuations. Generally,

amplitude fluctuations are always present with complex targets so that the monopulse radar is preferred when tracking accuracy is important. The monopulse radar is the more complex of the two. Three separate receivers are necessary to derive the error signal in two orthogonal angular coordinates; only one Since the monopulse radar compares the receiver is needed in the conical-scan radar. amplitudes of the signals received in two or more channels, it is important that the gain of and phase shift through these channels be identical. Any differences in the gain or (A change in relative the phase might be interpreted as an erroneous angular error. phase can result in an amplitude difference.) It is usually difficult to maintain amplitude or phase stability in the RF portions of the receiver where the path length is many
wavelengths long. For this reason, the RF circuitry in the monopulse radar is usually placed as close to the antenna feed as possible. Many monopulse trackers employ lens antennas (Sec. 7.6) or Cassegrain reflectors (Sec. 7.5) which permit the RF circuitry to be placed directly at the feed without blocking the aperture. With the monopulse tracker it is possible, in principle, to obtain a measure of the

196

Introduction to Radar Systems

angular error in two coordinates on the basis of a single pulse ; a minimum of four pulses are necessary with the conical-scan radar. Thus the monopulse tracker is theoretically capable of obtaining an angle measurement in microseconds as compared with milliseconds for the conical-scan radar. If the signal-to-noise ratio per pulse were
sufficiently large, the monopulse tracker would be capable of responding to faster angular rates than the conical-scan tracker and would be limited only by the response of the mechanical structure of the radar antenna and the servos. If the two radars are to be evaluated on the same basis, however, the total echo energy ought to be the same. Therefore they should both integrate the same number of pulses. In essence, the monopulse radar first makes its measurement and then integrates a number of pulses to obtain the required output signal-to-noise ratio; the conical-scan radar does the
It integrates a number of pulses first and then extracts the angular error. The faster response of the monopulse tracker may be obtained only when the signal-tonoise ratio per pulse is of sufficient magnitude (20 to 30 db) to allow a good measurement to be made on a single pulse without the necessity for integration. Because the accuracy of monopulse is not degraded by amplitude fluctuations, it is

opposite.

countermeasures than is conical scan. In summary, it may be said that the performances of the conical-scan radar and the monopulse radar are quite comparable except when the amplitude of the target cross section fluctuates at a rate comparable with the conical-scan frequency. When target amplitude fluctuations are troublesome, they may be eliminated with the slightly more

less susceptible to electronic

complex monopulse radar.

REFERENCES
1.

2.

James, H. M., N. B. Nichols, and R. S. Phillips: "Theory of Servomechanisms," MIT Radiation Laboratory Series, vol. 25, McGraw-Hill Book Company, Inc., New York, 1947. Schafer, C. R.: Phase-selective Detectors, Electronics, vol. 27, no. 2, pp. 188-190, February. rr }
1954.

3.

Greenwood,

I.

A.,

Jr., J.

V. Holdam,

Jr.,

and D. Macrae,

Jr. (eds.):

MIT Radiation Laboratory New York, 1948.
4. Palma-Vittorelli,

"Electronic Instruments,"

Series, vol. 21, pp. 383-386,

McGraw-Hill Book Company, v } Inc

5.

M. B., M. U. Palma, and D. Palumbo: The Behavior of Phase-sensitive Deteccimento, vol. 6, pp. 1211-1220, Nov. 1, 1957. Krishnan, S.: Diode Phase Detectors, Electronic and Radio Engr., vol. 36, pp. 45-50, February,
tors,

Nuovo

6.

7.

and G. E. Uhlenbeck (eds.): "Threshold Signals," MIT Radiation Laboratory McGraw-Hill Book Company, Inc., New York, 1950. Oliver, B. M.: Automatic Volume Control as a Feedback Problem, Proc. IRE, vol. 36 yv pp 466-473, April,
J.

Lawson,

L.,

Series, vol. 24,

'

1948.

8.

J. C. G.: The Design of Automatic-gain-control Systems for Auto-tracking Receivers, Proc. IEE, pt. C, vol. 105, pp. 93-108, March, 1958.

Field,

Radar

9.

Locke, A.

S.:

"Guidance," pp. 402^108, D. Van Nostrand Company,

Inc., Princeton, N.J.,

10.

11.

12.

13.

14.

15.

16.

An Analysis of Conical Scan Antennas for Tracking, IRE pp. 39-47, 1956. Whittaker, E. T., and G. N. Watson: "Modern Analysis," 4th ed., p. 357, ex. 3, and p. 372 Cambridge University Press, New York, 1950. Kleene, S. C: Analysis of Lengthening of Modulated Repetitive Pulses, Proc. IRE, vol 35 pp. 1049-1053, October, 1947. Page, R. M.: Monopulse Radar, IRE Natl. Conv. Record, vol. 3, pt. 8, pp. 132-134, 1955. Dunn, J. H., and D. D. Howard: Precision Tracking with Monopulse Radar, Electronics, vol. 33 no. 17, pp. 51-56, Apr. 22, 1960. Barton, D. K., and S. M. Sherman: Pulse Radar for Trajectory Instrumentation, paper presented at Sixth National Flight Test Instrumentation Symposium, Instrument Society of America San Diego, Calif., May 3, 1960. Cohen, W., and C. M. Steinmetz: Amplitude and Phase-sensing Monopulse System Parameters,
J.

Damonte,

B.,

and D.

J.

Stoddard:
1,

Natl. Conv. Record, vol. 4, pt.

pts. I

and

II,

also discussion

Microwave J., vol. 2, pp. 27-33, October, 1959, and pp. 33-38, November, 1959by F. J. Gardiner, vol. 3, pp. 18, 20, January, 1960.

Natl. of Radio Engineers. pp. pt. 1946. Tracking Accuracy of Monopulse Radar Systems. Schenectady. A : W Made with the Millstone Hill 40 Petteneill G H. .. R. . New York. Karelitz. C. D. „ „ Aircraft Radar Targets at 5-band. Time Duplexed Monopulse Receiver. 66. in Tracking a Doppler Navigation Radar 44.„• ™ • • : : Howard D D. June 30. Control and Data Trans24 Barton D K Application of Precision Tracking . M. Rept. Contract Ward. vol.9-1. Natl. M. : & Sons'. Echo Signal Phase Front Distortion.: 45. J. November. June.: Aircraft Scintillation Spectra.. December.. and L. J. Pfeffer: The Effect of D AGC 36. and F. pt. O.. and J. and A. Proc.. Proc. IRE. December. 1954. New 18." sec.. vol. Effects of Automatic Gain Control Performance on the 1959 Muchmore. Jr. Electronics. Weimer. Delano. E. R. 3.„. Naval 43 Gardner R E Doppler Spectral Characteristics of Research Lab. 41. John Wiley & Sons. Proc. Inc. IRE Trans. 1961. R. pp. 1956. 39. H. Kirkpatrick: I 944 from the General Electric Co. 1778-1784. pt. 1959. IRE. : ' 1947 A ' . a Radar Noise in D Dunn. by J.. vol. H. 41. 5. AP-9. Company. pp. 197 York. E. pp. 1951. Research Laboratory. 855-863. Muchmore. report 22 Blewett J P S. L. 15. Troell.„. 2. J. March. Radar to Location. February. IRE. H. Sommer. 27. Siegel. vol. D Tracking Radar.. IRE Trans. Howard: The Dunn J H and 430^135. N. Sept. . and G. paper in "Avionics Research. pp. Natl. Mass. 1958. 840-849." Advisory Group for ment (AGARD). 1960. IRE.: "Principles of Noise. Electronics Conf. Peters. vol. (IRE). K. pp. W. Swerling. Barton D. Electronics.. June. pp.. ' D. Locke. • ' mission for an Unmanned Observation Platform. Princeton. 1956. R. Pergamon Press. MIT M. Dec. (IRE). and Radomes. W-33-038ac-13969. 110-113. IRE.. pp." chap. pp. vol. p. on Aeronaut." McGraw-Hill Book 1959 Inc. 5656. 19. Calif. 1945. Systems Based on 15.. vol AP-9. and R.: "Principles and Applications of Random Noise P . MIT Servomechanisms Laboratory. 1294-1306. B. 51-55. H. Cambridge. 1961. The SCR-584 Radar. 3. Tracking. F. (Chicago). B. Jr. „„. Third Natl. 1945.Y. 179-186. ' - 23. A. C. Rhodes. Inc. Tyrrell. Stryker. pp. Princeton. . R. Peters. King: Phenomena of Scintillation Tracking Systems. "Guidance. Improved Feed Design for Amplitude 1959. January. S. 405. New York. New York..: The Use of a Tuned Discriminator 228-231. AP-8. 227-229. . 1961. vol.J. N. 10. Proc. Downey. Ranges." pp.: Accuracy of a Monopulse Radar. Proc. 32 ' 33 ' 201-212. Patterns. M.: "Design of Fire-control Systems. 440-442. pp. IRE. J. pp. Mason. 1953. 1961 (ASTIA No.: Angular Jitter in Conventional Conical Scanning _ _ .: An Improved Simultaneous Phase Comparison Guidance vol. and L.. IRE Trans. vol." pp. 25 ' 26 ' 27 28 ' 29 30. M. ' Radiation Laboratory Series. March. vol. Functions for Monopulse Antenna Difference 21 Price. 1. Rept. N. Delano. Weimer. Conf. J. 35. 47. and R. 19.. S "Guidance. . Inc. pp. D. 3 Military Electronics Conv. R. 5.. Proc.. 1959. vol. P. AP-8. R. Electronics. .. vol. McGraw-Hill Book Company.. New York 1958 Freeman. Van Nostrand Company. 26. 1960. vol.: Hybrid Circuits for Microwaves. ANE-3. 110-117. 801-810. A. pp. H. AD 263478). under Noise in a Fire-Control System. F. 34. Theory. "IRE Dictionary of Electronics Terms and Symbols.. pp. 104-109. Inc. A. Howard. 1959. Jan. 1958. 1955. D. D. 1955. The Multilobe Tracking System.' pp. also discussions by Jr. Proc. Conf. ' Radar. P.. Munishian: Proc. on Military Electronics Conv. Automatic-tracking Radar Brockner. Van Nostrand Company. Inc.: Earth Satellite Observations Aeronautical Research and DevelopRadar. vol. Jr. Proc.: "Introduction to Monopulse. Turner (eds. 67-70. Hardin. Conf. Spectrum. : ' 26-28. 50. Conv. 37 38' 39 ' pp.) generally (Not 1950. 18. 1. May. pp. pp. Systems. E. New York. . 18. NATO. no. S. Hansen." Part II. on Radar Tracking Noise. and Analysis of . 47. E. Monopulse Radar Antennas. on Aeronaut. L. A Theory of Target vol. and I. Santa Monica. C. John Wiley Bendat. vol.. Hyneman: Distribution 1960. H. 1959.: Some Factors Affecting the Performance of a RM-989-1. Record. and by R. March. Radar Target Angular Scintillation in Tracking and Guidance vol. 20 Shelton. G. 104-109. Kraft. Electronics. 34. Proc. Radar in Scintillation Angular or Glint Delano. IRE Natl. Space for Readied Trackers New Pulse F. 35. K. W. J." Institute Measurement 42 Eisengrein R H. to the author Communicated available.J. January. B. 7.. 408-413. Rand Corp. J. November. 44.): "Radar Scanners Cady 1948. Proc.1 Tracking Radar 17. Locke pt... Aug.. p. 93-102. November. D.. 567-576.

Introduction Generation of adequate RF power is an important part of any radar system. An example is the klystron amplifier fed by a crystal-controlled. It is based upon the principle of crossed electric and magnetic fields just as is the magnetron oscillator. Transmitters that utilize selfmaster-oscillator excited power oscillators are usually smaller than transmitters with master-oscillator power amplifiers (MOPA).000 Mc. stable oscillator. klystron amplifier. 2 shows that the transmitter power varies as the fourth root of the range if all other factors are constant. frequency-multiplier chain. the potentialities of the klystron amplifier as a radar transmitter were realized. in the years following the war. Not only does a transmitter represent a large part of the initial cost of a radar system. The oscillator version is called the Stabilitron. Buying range at the expense of power is costly. from a single "bottle. an abbreviation for power amplifier. Conventional triodes and tetrodes were used since no other power tubes were available. Both of these transmitter configurations were encountered in the discussion of the MTI radar in Sec. The other utilizes a low-power. the power has to be increased 16-fold. will In this chapter various tube types used for high-power radar transmitter application be discussed. The radar equation of Chap. The Amplitron is characterized by high power. The invention of the cavity magnetron by Randall and Boot in the late thirties made possible the development of microwave radar in time for World War II. and it began to be applied early in the 1950s in radars where large power and good stability were required. especially where large transmitter bandwidth is required. The latter are more stable than self-excited oscillators and are usually capable of greater average power. Those considered are the magnetron oscillator. The high-power traveling-wave-tube amplifier may also be used for radar application.1." average power outputs greater than those obtained from any other tube type operating in this frequency range. sometimes referred to as MOPA. There are two basic transmitter configurations used in radar. and broad bandwidth. To double the range. further development of grid-controlled tubes for applications took place. Another high-power amplifier tube found in radar is the Amplitron. One is the self-excited oscillator. which is in turn amplified to the required power level by one or more poweramplifier tubes. The choice between the two is governed mainly by the radar application. Also. it requires a continual operating cost because of the need for prime power or fuel. 4. it is therefore important that the best transmitter be selected for any particular application. The magnetron oscillator has seen wide application in radar. These tubes were capable of delivering. but unlike many other parts of the radar. The earliest radar transmitters operated below the microwave region in the UHF and VHF bands.6 RADAR TRANSMITTERS 6. exemplified by the magnetron. In the years following World War II. high efficiency.1. Self-excited power oscillators therefore are likely to be found in applications where small size and portability are more important than the stability and high power of the MOPA. 198 . high-power television. primarily because of the needs of commercial UHF High-power triodes and tetrodes thus became available for radar application at frequencies up to approximately 1.

especially the application of cavity resonators. listed not necessarily in the order of their importance. and with a magnetic field parallel to the axis (perpendicular to the electric field). so they were not influenced by They were familiar. the pessimistic outlook this work seemed to dictate. the understanding of microwave circuitry was also being actively pursued at about this same time. The original magnetron device was a diode switch invented in 1921 by A. There are many texts available on microwave tubes. which turns the transmitter transmitted waveform. Randall and Boot applied the cavity resonators to the magnetron structure and produced a magnetron at 10 cm wavelength capable of better than 100 kw . Oscillations were due to an interaction between the electrons and the tangential component of a traveling. a more efficient and reliable form of magnetron is obtained based on negative-resistance oscillations.2. Postumus. The modern radar magnetron is based upon this principle. especially the role played by the cavity resonator. Cyclotron oscillations were observed at microwave frequencies by Zacek. they have not been too widely used since they suffer from low power.2] Radar Transmitters 199 traveling-wave tube. The application of the magnetic field deflected the electrons from their journey to the plate and cut off the conduction of the tube. The frequency of oscillation is determined by the resonant circuit. reported a third form of magnetron known as the travelingwave magnetron (not to be confused with the traveling-wave tube invented later). with the early experimental work with klystron amplifiers. Fortunately. but with application to the klystron amplifier rather than the magnetron. in 1935. 3 They were not too familiar with the earlier investigations of magnetrons. Various investigators reported microwave oscillations with this magnetron in the late 1930s. and low efficiencies at microwave frequencies (1 per cent for the cylindrical diode magnetron. 6. Also discussed is the modulator. at a wavelength of 29 cm as early as 1924. K. Emphasis will be on the properties of importance in radar applications. 1 2. however. in Prague. W.Sec. The name magnetron has been applied several different types of electron devices. The magnetron of Postumus consisted of an anode split into four segments. usually with cylindrical geometry. The basic operating principles of these tubes will be only briefly mentioned. who theoretically predicted and experimentally demonstrated their existence. and the grid-controlled tube. about 10 per cent for the split-anode magnetron at moderately long wavelengths). after Eric Habann. as well as a where the interested reader can find the details concerning the theory of each type of tube. This is sometimes called Habann-type oscillations. Hull also observed oscil- Historical Development. on and off to form the The Magnetron Oscillator ' in the past to All of these are diodes. The magnetic field is not as critical with the negative-resistance magnetron as it is with the cyclotron magnelations A tron. but further development for high-power applications had to await a better understanding of the principles of microwave circuits. Hull.wave RF field whose velocity was substantially equal to the average velocity of the electrons. in Japan. Amplitron. Stabilitron. The oscillations were due to the cyclotron resonance frequency that is characteristic of electrons in crossed electric and magnetic fields. Yagi. Although extremely high frequencies have been generated with cyclotron magnetrons. 6. erratic behavior. also obtained microwave cyclotron oscillations at about the same time. The first successful magnetron suitable for radar application was invented by Randall and Boot at the University of Birmingham in 1939. from his magnetron at a frequency of 30 kc. large published literature. With a magnetic field large compared with that used for the cyclotron magnetron. power of 8 kw was obtained with an efficiency of 69 per cent.

Another type of resonator structure found in some magnetrons is the interdigital The basic structure of one form of magnetron is shown in Fig. Description of Operation. formed entry into World War II. and the effective inductance is LjN. 6. the frequency of the magnetron is essentially that of an individual resonator. 1 The anode (1) is a which are cut holes (2) and slots (3). or a donutron. The resonant circuits to the inductance. Since the angular frequency is equal to (LC)~-. The development of the magnetron has probably relied less on theoretical results and more on the empirical approach to design than other tube types. Cross-sectional sketch of typical cavity magnetron illustrating component . Alternate segments are connected together at one end of the cavity. The shape of the The slot configuration has cavities determines the impedance (which is equal to L/C). are not readily susceptible to mathemati. . since cross-field devices. In the desired mode of operation The effective (the 77 mode) the individual C's and Us are connected in parallel. while the slots correspond to the capacity. where N is the number of resonators. The magnetron is a crossed-field device that is.2 of pulse power. The holes correspond. capacitance for the whole magnetron oscillator is NC. Fig. 6.2a or the vanes in 6. 1 electric fields cal analysis. the (both RF and d-c) are perpendicular to a static magnetic field. The resonator of the interdigital magnetron is a short cylindrical cavity. all lie within the vacuum envelope in the magnetron. 6. The anode segments extend as fingers from the two flat sides of the cavity. a lower impedance than the vane configuration. Other forms of resonators which might be used are the slots shown in Fig.1. including the magnetron and the Amplitron. 6. 4 This structure has been widely used in voltage-tuned magnetrons. The holes and slots function as the resonant circuits and serve the same purpose as the lumped-constant LC resonant circuits used at lower frequencies.2b. parts. and the remaining segments are connected together at the opposite end. in rough fashion. This type of anode has also been called a squirrel cage.200 Introduction to Radar Systems [Sec. the basis for the development of microwave radar in this country. large block of copper into anode. The British disclosed the principle of the cavity-resonator magnetron to the United States in 1940 during an exchange of technical information just prior to our The magnetron. a power considerably greater than had previously been achieved at these frequencies. more than any other single device.

coming bunched. Fig. results in JV/2 possible modes of operation.1) Not shown in Fig. An assembled The preferred mode of magnetron operation corresponds to an RF Stability}'* field configuration in which the RF phase alternates 180° between adjacent cavities. the electrons move along in a travelingwave field. (N 4)/(iV+ 4). Its frequency is approximately the resonant frequency of one of the cavity resonators. Figure 6.3 is an exploded view of the QIC 358 tunable L-band magnetron showing the 358 is shown in Fig. For this reason the heater power may be reduced or even turned off once oscillations have started. 6. 1) is usually a fat cylinder of oxide-coated material. cathode temperature during operation and causes secondary electrons to be emitted. The RF power is extracted by placing Slot type. 6.-type Since a refractory metal serves as the cathode cathode). 6.1) the electrons interact with the d-c electric field and the magnetic field in such a manner that the electrons give up their energy to the RF field. In the interaction space (5.Sec. or slightly exceeds. This is called the n mode.4.2. 5 under d-c conditions an oxide cathode is capable of an emission-current density of the order amp/cm 2 but under pulse conditions the emission current can be as great as The cathode must be rugged to withstand the heating and disintegration 00 amp/cm 2 caused by the back bombardment of electrons. 6 Another cathode that has been used is the dispenser cathode. 6. causing RF power to be de. a coupling loop in one of the cavities as shown (6. For example. shield disks located at each end of the cathode for the purpose of confining the If electrons are lost from the ends of the cathode. Fig. electrons to the interaction space. their power is not delivered to the RF field and the efficiency of the tube will decrease. 6. where The various modes are a result of mutual coupling between cavities. Fig. This traveling-wave field moves at almost the Fig. Fig. The presence of the crossed electric and magnetic fields causes the electrons to be completely bunched almost After beas soon as they are emitted from the cathode. the frequency stability will be poorer. The relatively fat cylindrical cathode can dissipate more heat than a thin cathode. 6. 6. porous surface for the impregnation of a tenacious layer of the oxide-emitting surface. The advantage of the oxide cathode is that higher emission currents can be obtained under pulse conditions than with other emitting materials.2] Radar Transmitters 201 The magnetron cathode (4.2 1 . to one-half for a 12-cavity magnetron. Each of the N/2 — N QK N . 1 are endor else by coupling one cavity directly to a waveguide. BacK bombardment increases the of 0. In addition. it can withstand high temperatures and severe arcing conditions better than can the matrix oxide cathode. Most magnetrons in the past used a cathode consisting of a matrix of nickel powder sintered on the nickel-alloy base metal providing a rough. (a) livered to the wave. The straps (7. Examples of cavitysame speed as the electrons. where This ratio is equal is the number of resonators. (b) vane type. 6. A fat cathode is also required for theoretical reasons. They improve the stability and efficiency of the tube. . base. The presence of more than one cavity in the magnetron is the total number of cavities. component parts.1) are metal rings connected to alternate segments of the anode block. The optimum ratio of the cathode diameter to anode diameter equals. 7 in which the oxide is impregnated in a tungsten (a) cylinder (called the impregnated type) or else the oxide material is made to diffuse through tungsten (called Z.magnetron resonators.

capable of of peak power at a duty factor of 0. modes corresponds to a different RF field configuration made up of a standing wave of All the modes except the n mode are degenerate.) 1 Mw one possible mode of operation means that the magnetron can oscillate at any one of and can do so in an unpredictable manner. TUNER POLE BELLOWS HOUSING AIR DUCT HOUSING output transformer AIR DUCT COUPLING FLANGE JBP\xeramic ^. these frequencies . A different mode means a different frequency and a different field configuration. output WINDOW CATHODE- CATHODE END SHIELDS- HEATER TERMINAL CUP ASSEMBLY Fig. they can oscillate at two different frequencies corresponding to a rotation of the standing-wave pattern.2 charge. An output circuit designed for one particular mode configuration may produce a weak or a zero output when the magnetron operates in a different mode.3. 6.^" »* 1 r. Therefore it is important that a magnetron be designed with but one mode dominant. that is. 6. Thus there are N — 1 possible frequencies in which the magnetron can oscillate.202 Introduction to Radar Systems [Sec. (Courtesy Raytheon Company. The presence of more than EXHAUST SEAL TUBULATION TUNING GEAR ASSEMBLY- GETTER TERMINAL SETTER MATERIAL MAGNET (ONE OF 4 SECTIONS). Exploded view of a Raytheon RK6517/QK358 tunable L-band magnetron.f.001 3. This is the essence of the stability problem. where the positions of the nodes and antinodes are interchanged.

Sayre. than 10 per cent.4. About a year after the invention of the cavity magnetron. Strapping is not the only method of obtaining mode stability. (Courtesy Raytheon Company. but it also ing to the 77 mode. but there is a closed transmission line with segments (digits) attached alternately to opposite conductors of The modes are readily separated in this construction. 6. 6. Four. 6. Strapping improved the stability and increased A the efficiency to 50 per cent. In the interdigital magnetron there are no individual resonators as such.2] Radar Transmitters it is 203 The it mode is usually preferred because separated from the others. increases the efficiency. The cross section of the straps may be either circular or rectangular.Sec. found that the stability and efficiency of the tube could be considerably improved by coupling together every other segment with a circular ring called a strap. .or six-segment anode blocks can be made to function in the -n mode without straps because there are few modes to separate. the frequency of the -rr mode was separated by less than 2 per cent from its 1 single ring strap increased the separation to greater next nearest degenerate mode. not degenerate and can be more readily in- The early magnetron invented by Randall and Boot suffered from frequency inefficiency as a result of stability and moding troubles. unstrapped magnetron operating at 1 cm wavelength (3. Photograph of RK 6517/QK 358 Z-band magnetron. as was shown in Fig. 1 . the early unstrapped British magnetrons were unstable and changing.1. at the University of Birmingham. 4. had efficiencies of 30 to 35 per cent. since higher powers can be obtained without fear of mode For example. Fig. Strapping not only improves the stability of operation. the lowest frequency is that correspondused. Chap.) In a particular eight-resonator. but the tube is limited the line. The straps connect all those segments which have the same potential in the n mode.000 Mc). Various forms of strapping are discussed by Walker in Ref. Even greater mode separation is possible if larger or more straps are In a strapped symmetrical magnetron. Another technique is the interdigital anode block. to low power.

! (3) load conductance. the large cavities. The rising-sun anode may be considered as two resonant systems. The current determines the voltage. more cavities in the rising-sun magnetron than in the conventional magnetron. Each of these systems by itself would have the mode spectrum of an unstrapped magnetron block with N/2 resonators. 1 11 Four parameters determine the operation of the magnetron. there are The rising-sun anode block is characterized by alternately large and small resonators. 6. voltage as a function of the four parameters mentioned above is greatly simplified since the input and output parameters operate nearly independently of each other. The coupling between the 7T modes of the two systems is strong. lengths. 6. In the rising-sun system the n mode Fig. while the last two are related to the output side. where it is difficult to manufacture strapped magnetrons because of their small size. It is quite suitable for high power because of the relatively large cavity resonators. and not vice versa. between the groups of other undesired at longer and at shorter wavelengths. except that a magnetron behaves generator.5). and they combine to produce the operating mode. It is configuration 9 (Fig. and (4) load susceptance.6. and the anode voltage.2 Another method of separating the modes in the magnetron is with the rising-sun anode This structure separates the modes without the need for of particular advantage at the very high radar frequencies (X band or above). the wavelength. The plot of the observed magnetron quantities as a function of the input circuit parameters. The observed quantities are usually the output power. large cathode and anode diameters.204 Introduction to Radar Systems [Sec. . These are (1) the magnetic field. and the other. is called the performance chart. (2) the anode current. the variation of the observed quantities can be studied as a function of the load presented to the magnetron. In many magnetrons the magnetic field is fixed by the tube designer and may not be a variable the radar designer has under his control. is called a Rieke diagram. The rising-sun magnetron has been used 10 to lies modes generate RF power at wavelengths as short as 3 mm. the straps have relatively high copper losses at these frequencies. 6. The problem of presenting the varia' tion of the three quantities power. with the input parameters magnetic field and current likewise chosen for convenience. The rising-sun structure may also be used for high-power applications at the longer wavestraps. Thus it is possible to study the effect of the magnetic field and the anode current at some value of load susceptance and conductance chosen for convenience. whereas in the strapped magnetron the tt mode corresponds to the longest wavelength. 6.5. for some fixed load. The plot of the observed quantities as a function of the load conductance and susceptance. or a load diagram. and long anodes. one comprising the small cavities. This is representative of — — — — t The anode voltage might be substituted more like a current generator than a voltage for the anode current. Rising-sun magnetron resonator. An example of a performance chart is shown in Fig. The first two parameters are related to the input side of the tube. In general. wavelength. for a fixed magnetic field and anode current. Also. Performance Chart andRieke Diagram. Weak coupling exists between the various modes of the two sets of resonators except for the 77 mode. Similarly. The results will not be greatly dependent upon the particular values of susceptance and conductance chosen.

For a constant output power the performance chart shows the compromises which can be made between the voltage. If the current gauss. usually a matched load.) {Courtesy Raytheon power output (heavy dashed lines) suggest the form of hyperbolas. The contours of constant (peak) 30 30 40 amp — — —-— — 6. the peak anode voltage is 14. is kept the same but the magnetic field reduced to slightly more than 1. Performance chart for the type 4J36-4J41 pulsed magnetron Fig. efficiency. The change in the oscillator frequency produced by a indicated on the chart as .5 kv. lines) approximate straight lines. of relatively small positive slope. is reduced to 35 per cent. This particular performance chart indicates that the magnetron will produce a power output of 500 kw with a peak anode voltage = 26 kv. and the ordinate is the peak anode voltage. peak anode current = 38 amp. There are four The curves of constant magnetic field (solid families of curves shown on this chart. 6. by the tube performance chart would be considerably simplified. Company. The abscissa is the peak anode current. Magnetic field Peak power output Magnetron efficiency Deviation in frequency from OMc oscillator.6. The dotted lines represent the deviation in frequency from a reference frequency Mc.2] the Radar Transmitters 205 Raytheon 4J36-4J41 pulse-type S-band magnetrons.Sec.500 gauss. only one curve of constant magnetic field would apply and the designer.500 The magnetron efficiency is 50 per cent under these conditions. and magnetic field. and magnetic field = 2. except for small If the magnetron were supplied with a fixed magnetic field chosen values of current. the peak power output drops to 200 kw. The data plotted in the performance chart are for a fixed load. Peak current. and the efficiency The lightly dashed lines are contours of constant efficiency. current.

The center of the Smith chart (Rieke diagram) coordinate. Although a large magnetic field requires a relatively heavy magnet. Therefore the greater the power output. a set of circles. The angle in a clockwise direction from this reference axis is proportional to the distance (in wavelengths) of the standing-wave-pattern minimum from the reference point. the following general features seem to be shown by most magnetrons (1 ) except for very low currents. The frequency contours are farther apart in that region where the output power is low. Although a cartesian set of load coordinates could be used. but the poorer the . the greater the efficiency. The shaded areas of the chart correspond to regions of poor magnetron performance. it may sometimes be the cheapest method of obtaining efficiency. Thus a given change in the phase of the VSWR will cause a greater change in the magnetron frequency if the operating point of the magnetron is in the region of high-power output rather than in the region of low-power output. . (2) decreasing the magnetic field at constant current results in a decrease in efficiency. or zero reflection coefficient.7. The region on the chart where the contours of constant frequency crowd together is also the region where the output power is greatest. The radial coordinate can also be specified by the reflection coefficient Y of the load since the VSWR p and reflection coefficient are related by the equation |T| = (p — l)/(/> + 1). and (3) a drop in efficiency occurs at very small currents. The contours of constant power approximate. The circumference of the chart corresponds to infinite VSWR. In general. In measuring the pushing figure the current must be changed rapidly in order to avoid frequency shifts due to temperature changes. while the contours of constant frequency approximate arcs of circles which are almost perpendicular to the contours of constant power. 12 The Smith chart is a form of circle diagram widely used as an aid in transmission-line calculations. An example of the Rieke diagram for a magnetron operating in the n mode is shown in Fig. Thus the region of low standing-wave ratio is toward the center of the chart. increasing the anode current while maintaining the magnetic field constant results in a decrease in efficiency. it is usually more convenient to plot the power and frequency on a set of load coordinates known as the Smith chart. 6. Although there is considerable variation in the performance charts for different magnetrons. in a rough manner. it is The latter is the easier for the corresponds to unity VSWR. The performance chart permits the radar designer to select the tube parameters which best satisfy the diverse requirements of a particular application. 6. An advantage of the Smith chart for plotting the effects of the load on the magnetron parameters is that the shapes of the curves are practically independent of the position of the reference point used for measuring the phase of the VSWR. The reference axis in the Rieke diagram usually corresponds to the output terminals of the magnetron or the output flange of the waveguide. The other plot of magnetron characteristics of interest in radar design is the Rieke diagram. 360° in the diagram is taken as a half wavelength. Plotted on the Rieke diagram are contours of constant power and constant frequency. or unity reflection coefficient. Thus the Rieke diagram gives the power output and the frequency of oscillation for any specified load condition. the performance chart indicates that large magnetic fields result in good efficiency. The standing.wave pattern along a transmission line repeats itself every half wavelength therefore.206 Introduction to Radar Systems [Sec. The coordinates of the Rieke diagram are the load conductance and susceptance (or resistance and reactance). A point on the Smith chart may be expressed in conductance-susceptance coordinates or by a set of polar coordinates in which the voltage-standing-wave ratio (VSWR) is plotted as the radius and the phase of the VSWR is plotted as the angular : more usual of the two possible coordinate systems since microwave engineer to measure the VSWR and the position of the voltage-standing-wave minimum (or phase) than it is for him to measure the conductance and susceptance directly.2 change in the anode current for a fixed load is called the pushing figure.

250 Mc. prf=391. duty cycle to the load). the operation may be unstable which moves the because of mode changes. However. ideal. start of the pulse. the starting characteristics of the magnetron will be enhanced. tube to operate ATR is unfired and presents an open circuit in the transIf the location of the ATR is such that the phase of the VSWR causes the in a favorable portion of the Rieke diagram (magnetron lightly coupled 270' 180° Example of a Rieke diagram for the RK 6517/QK 358 /.Sec. A change in the phase of the operating point of the magnetron into the low-power region (antisink) results in a VSWR lightly is The build-up of oscillations in a lightly loaded magnetron and the magnetron pulses start more uniformly. Pulse width = = 0. and frequency = 1 . sink region is not always desirable since the RF spectrum of the magnetron output will tend to broaden. Solid curves Dashed curves are contours of constant frequency. Fig. Also.7. more .) current = 50 amp. The phase and/or magnitude of the might vary because the antenna sees a different load impedance. peak anode {Courtesy Raytheon Company. average anode current = 62. 3. the starting characteristics will be poor. 6. the mission line.2 /«ec. reflection places the operating point in other hand. than if the load were matched. The region of highest power on the Rieke diagram is called the sink and represents Operation in the the greatest coupling to the magnetron and the highest efficiency. indicating a poor pulse shape. 6.2] Radar Transmitters 207 in a scanning radar frequency stability. if the phase of the On the ATR an un- favorable portion of the Rieke diagram.00125. The At the location of the ATR tube can also affect the performance of the magnetron.-band magnetron.5 ma d-c. are contours of constant power. a lightly loaded magnetron may perform poorly by showing signs of instability which take the form of arcing and an increase in the number of missing loaded magnetron. depending upon the VSWR objects it views.

6.5 Mc. however. The ability to tune a magnetron is a desirable operational feature. For the magnetron whose Rieke diagram is shown in Fig. when inserted into the hole of the resonator. the pulling figure is approximately 1. surface-to-volume ratio in a high-current region. The various methods for tuning a magnetron may be classified as (1) mechani- inductance tuning (L) and capacitance tuning (C). In the mechanically tuned magnetron. by electronic beams or space charge located within the magnetron cavity or some external cavity. to incorporate tuning into a magnetron than in other power tubes. Example of an inductance-type tance of the resonant circuit by altering the {Courtesy Raytheon Company. making RF discharges more likely. 6. Single hole-andslot resonator illustrating efficient VSWR is fixed at 1. tunable magnetron permits the radar to be operated anywhere within a band of frequencies and to be set to a precise frequency. An important characteristic of the magnetron is the pulling figure.8 shows a single hole-and-slot resonator. electronic tuning. Tuning. One of the limitations of . The pulling figure of a magnetron may be readily obtained from an inspection of the Rieke diagram. 6.8.9.5 (reflection co- = 0. In some applications. the frequency of oscillation is changed by the motion of some element in the resonant circuit associated with the magnetron. and (3) voltage-tuned. Figure 6.) tuner.5 is the pulling figure. 6.4.7. All rods are attached to a frame which is positioned by means of a flexible-bellows arrangement as illustrated in Fig. The latter property is not often available with fixed-tuned magnetrons whose frequencies might lie anywhere within a narrow band and are not under the control of the radar systems engineer. The difference between the minimum. (2) electronic. The insertion of the rods into each anode hole decreases the inductance of the cavity and therefore increases the resonant frequency. Mechanical tuning is accomplished by the movement of a tuning device. It is a measure of the effect a change in the output load has on the frequency of oscillation. or a crown-of-thorns tuner. if desired.208 Introduction to Radar Systems pulses. A tuner that consists of a series of rods inserted into each cavity resonator so as to alter the inductance is called a sprocket tuner. The pulling figure is defined as the difference between the maximum and the minimum frequencies of a magnetron oscillator when the phase angle of the load impedance varies through 360° and the magnitude of the Fig. It is usually more difficult. 6. changes the inducFig. [Sec. and voltage tuning is accomplished by designing the magnetron to operate in a region where a change in anode voltage results in a change in frequency.2 This poor operation results because the RF voltages inside the lightly loaded magnetron are large in the antisink region.20). The Cavity inductive tuning element L.9. frequency contours intercepted by the circle corresponding to a 1. such as the radar altimeter. The change in inductance results in a change in frequency. 6. it is absolutely essential that the transmitter be tunable over a wide range. An inductance tuner can also be seen in the exploded magnetron of Fig.and maximum- VSWR = A FM cal.

tunable magnetron. 6. thereby increasing the strap Because of the mechanical. A tuning mechanism in only one resonator hole does Fig. (Courtesy Raytheon and is a form of unsymmetrical tuning. and the angular symmetry of the n mode is preserved. in The two tuning methods may be used com- bination to cover a larger tuning range than Tuning is possible with either one alone. 6. limited tuning range. of high region The electron beam may be inserted directly into the magnetron cavity in a 1 A A - . Because the gap is narrowed capacitance slot in width. tuner is called the the two rings of a double-ring-strapped magnetron. this tuner is more suited for use at the longer wavelengths.000 to 9. and capacity-tuned magnetrons must A common form of capacity usually operate with low voltages and hence low powers. can be obtained rather simply by means A of a screw inserted in the side of one of the resonator holes.ridged waveguide as an auxiliary tuning cavity whose resonant frequency was varied by adjusting a short-circuiting plunger located at one end of the cavity.10. as a rule. Each cavity is affected in the same manner.Sec.5 to 1 are not uncommon with this arrangement. could be tuned over the frequency range from 13 9. It cutter (Fig. preserve angular symmetry tuner. This type of adjustment is useful Anode segment "Heavy dark line when it is desired to fix the magnetron represents tuner- frequency to a specified value within the normal ring position between the magnetron straps scatter band of untuned magnetrons. An electron beam injected into one or more of the cavities of a magnetron will change 14 15 the effective dielectric constant of the cavity and thereby change the frequency. Tuning rates of 100 Gc/sec have been achieved with servo-controlled. although there is some indication that the cookie cutter is more restricted in tuning range A than the crown-of-thorns tuner. the breakdown voltage will be lowered. A fixed-frequency magnetron also may be tuned over a limited frequency range of A about single-stub per cent by varying the load into which the magnetron operates. 10 per cent frequency change can be obtained with either of the two tuning methods described above. of the order of 1 per cent. in the change this purpose. Another form of unsymmetrical tuning is an auxiliary resonant cavity coupled to one of the magnetron cavities. Example of capacitance-type Company. The frequency of the auxiliary cavity may be changed by making one wall of the cavity flexible so that it can be moved in or out. for may be used the tube external to tuner located impedance of the single-stub tuner changes the operating point on the Rieke diagram and therefore changes the frequency.600 Mc with a peak power output of 140 kw. 6. using a section of a double. metal ring inserted between of a consists 6.8 increases the and decreases the resonant frequency.) not. with the cookie cutter. cookie 10).2] Radar Transmitters is 209 of the inductive tuning tube.and voltage-breakdown problems associated capacitance. The Both the capacitance and the inductance tuners described above are symmetrical. ranges of 1. that it lowers the unloaded Q of the cavity and the efficiency insertion of an element C into the cavity slot as shown in Fig. The auxiliary cavity is tightly coupled and determines the operating frequency of the magnetron. mechanically tuned magnetrons.

However. 15 18 19 The necessary conditions for voltage tuning are that the magnetron be heavily loaded and that the anode current be limited and not increase with an increase in anode voltage. The voltage-tuned magnetron therefore acts as a * . but it does determine. It was mentioned previously that the magnetron frequency will vary if the anode current or voltage is varied. . This phenomenon can be used to tune a magnetron. rapid-tuning applications than as a high-power radar transmitter. When the anode circuit is heavily loaded and the number of electrons in the interaction space is restricted. The frequency is usually a linear function of the anode voltage. but the power output will not be constant over the tuning range. under certain conditions of operation the magnetron frequency may be made quite sensitive to voltage changes.2 RF electric field. or an external cavity may be used with an electronic beam or a controlformed by a magnetron diode. Most of these tubes have an interdigital' CW resonant-cavity or split-anode structure. The tuning range of these tubes covers a two-to-one frequency band. although the first voltage-tuned magnetron generated about 100 watts. 6. This technique has been used to frequency-modulate a 4. The voltage-tuned magnetron seems more suitable to low-power. Ferrite materials might be used to tune an interdigital magnetron. The tube was also mechanically tuned over the frequency range of 720 to 900 Mc with a cylindrical element which varied the interstrap capaci- CW 1 tance. A ferrite cylinder can be placed near the shorted end of a coaxial line which is coupled to the interdigital resonator. This latter condition is usually met by operating the cathode with the electron emission temperature-limited rather than space-charge-limited. the anode circuit does not determine the frequency of oscillation. The ferrite material is kept out of the main magnetic field of the magnetron so that its permeability may be controlled by means of its own biasing magnetic field. The voltage-tuned magnetrons which have been reported in the literature 18 20 21 have been low-power tubes with power outputs of the order of a few watts or tens of watts. The power output from this tube was 25 watts.1 to 2 Mc/volt. as required for voltage tuning.000-Mc CW magnetron 16 with a frequency deviation of 2. 17 Electron beams were injected into 9 of the 12 cavities of a vane-type magnetron. Voltage tuning results in a considerably greater frequency change than does frequency pushing for the same change in anode voltage. It might be applicable to a low-power radar such as the altimeter or as a local oscillator in a wide-tuning- FM-CW range receiver.5 Mc (total frequency swing of 5 Mc). It has been found possible in some cases to tune magnetrons over a frequency range of 4: 1 by means of voltage tuning.210 Introduction to Radar Systems [Sec. and in most magnetron applications it is not a desirable characteristic. constant-current generator. in part. This is known asfrequency pushing.5 Mc was obtained with a power output of kw and an efficiency of 55 per cent. Typical voltage-tuning characteristics show a change of frequency of 0. The frequency is varied by electrically varying the density of the electronic beam. Electronic tuning has also been applied to a relatively high power magnetron. but their efficiencies are lower than conventional magnetrons. Electronic tuning is probably more appropriate to magnetrons such as might be used for altimeters or for microlable space charge CW FM wave communications. The number of electrons in the interaction space may also be limited by providing a region of cathode surface which is nonemitting. the RF power output. with an efficiency of 50 per cent. Larger frequency deviations are possible if some amplitude modulation can be tolerated. only a few per cent change in frequency can be obtained by changing the current. At 900 Mc a frequency deviation of 3. but with space-charge-limited current.

6. if the magnitude of the the load into which the magnetron operates changes or if the frequency of oscillation magnitude is large. The longer the transmission line. As the tuner setting is increased. in part. Frequencies in between. There may even be periodically spaced holes in the tuning curve where it is not possible for the magnetron to oscillate. Thus For this reason. Thus there will be holes in The frequency difference the frequency coverage when the mismatch is severe. 6. where c is the velocity of light. claimed that theoretical calculations show that a tuning range of 5 to 10 per cent be expected. The change in phase A<£ for a particular change in frequency A/ is equal to 4nL Af/c. 22 ' 25 The Rieke diagram of Fig. the phelong-line effect is absent if the line is sufficiently long. between alternate points where the tuning curve with mismatch crosses the tuning curve with no mismatch (dashed curve) is |(r/L)(A//l 9 ). or pass through. has been reflected there.11a to c.2] Radar Transmitters 211 A variation of permeability in the coupled coaxial line results in a change of frequency. The modulation characteristics of frequency-modulated tubes might The long-line effect is not a characteristic of the magnetron alone. 6. on the length of the pulse and is different with a pulsed tube. 6. VSWR is called the long-line The long-line effect causes the tuning curve (plot of frequency vs. the frequency of oscillation will increase uniformly The frequency then jumps discontinuously to point 3. The long-line effect may also result in a poor spectrum (poor pulse shape). 6. especially can the phase of the occur if Changes in the phase of the is large. It is not present in a transmission line whose two-way transit time exceeds the length of the pulse since the tube will not be on when the reflection returns. are not obtained. Examples of the manner in which the long-line effect alters the tuning curve of a pulsed magnetron are shown in Fig. The corresponds to a transmission line pereffect of a small mismatch is shown in With a small mismatch as in Fig.11*. The ordinate is the magnetron frequency.1 la fectly terminated in a reflectionless load. 6. 1 11 23 seriously affect the operation of a magnetron. also be distorted. The phase change is proportional to the length of the line as well as to the change in frequency. It is a property of any self-excited oscillator whose frequency is affected by the output impedance. although the term been sometimes has nomenon CW long-line effect seems to be more generally accepted.7 indicates that a change in Long-line Effect. and the abscissa is the position of the tuning mechanism of the magnetron.1 lc) has regions where more than one frequency is indicated for a given tuner Operation at more than one frequency for a given tuner setting is not possible. If the It is may ' ' VSWR VSWR VSWR VSWR phase places the magnetron operatoperation at certain frequencies where the ing point in an undesirable portion of the Rieke diagram. 6. The poor magnetron performance caused phase when the magnetron is connected via a long length by a change in VSWR = VSWR of transmission line to a mismatched load with large effect.116. with the two-way propagation along a transmission line of length L is (f> where / is the frequency of oscillation and c is the velocity of propagation. a tunable magnetron might encounter poor varies. the greater will be the phase change and the more likely it will be that the magnetron operating point will be in.11c. The long-line effect results from the influence of the wave that has traveled to the end of the transmission line. the spectrum of the magnetron might be abnormally broad over those parts of the tuning curve where the slope is nearly vertical. until point 1 is reached. Fig. in Fig. setting. tube than with a It depends. The tuning curve for large mismatch (Fig. and frequency jumping. The straight-line curve illustrated in Fig. a region of poor magnetron operation. 6. The phase angle associated 4-rrLflc. tuner position) of a tunable magnetron to be altered. such as point 2. the short-line to as referred effect. missing pulses. and has returned to the magnetron. L .Sec. and a large mismatch.

{klX )j{LJX a g) sinh 2 [(a^)L e /AJ . (d) large mismatch. X is the wavelength in free space. is given by .11. the pulling figure measured in megacycles. and X is the wavelength in the g guide or the transmission line. thereby reducing the VSWR (at the input to the line) and the long-line effect.212 is Introduction to Radar Systems [Sec. 6. mismatch. (b) small Magnetron tuning (c) large curves. A reduction of the VSWR seen by the tube increases the allowable skip length. CW If the skip length is measured in feet.2) 1) The above expressions for the critical line length assume a lossless transmission line. mismatch. regardless of the phase angle of the load. where L = PF = c skip length. terminated transmission line (no mismatch). (6. Pritchard shows that the is maximum VSWR p L which can be tolerated at the end of the transmission line _ {(kjX )l(L IX + if + (klk )l(2L IK) sinh [(2aA„)L g c given by g) g c c /Vl 1 . L e for which completely stable operation of the tube occurs. Pritchard 24 has shown that the longest length of transmission line. Eq.2 the length of the line. The effect of loss is to reduce the amplitude of the reflected wave. or critical length pulling figure of magnetron p = voltage-standing-wave ratio Tuner position (c) X (a) Perfectly Tuner position id) X Fig. 6.1) may be written L c (ft) = -^P F(P (6. magnetron.

since the loss may be prohibitively high.000 100.000 Fig. The behavior of the magnetron with a long transmission line differs from that of the tuning curve of a An example of the long-line effect pulse magnetron because of frequency hysteresis. parameters with several waveguide and representative values of p L as a function of line length for (From Pritchard. 6. in shown magnetron is curve of a on the tuning there are two different frequencies possible for a given frequency setting. 10. 213 where k\X Q = = <fiAnin)(flPF transmission-line attenuation per unit length XW /= frequency This equation losses. u IRE Trans. Plot of the maximum VSWR sizes.12. However.2 1 — I Mill H 1 — I I I I I I 2. 1 1 d. however. the fixed-tuned magnetron may also be adversely affected by the and a long transmission line. the other are stable. It is easier to mount a given transmitter at the feed of a lens antenna . but the percentage of inaccessible frequencies is less.1 la to c apply to a pulsed magnetron.12 for representative values of the parameters with 0. An obvious method of eliminating the long-line effect is to avoid the use of long transmission lines by locating the transmitter directly at the antenna terminals. The location of the breaks in presence of a large Therefore a variation the tuning curve varies uniformly with the phase of the reflection.2] Radar Transmitters a. 6. magnetron In the Fig. for a = Although a tunable pulse magnetron has been assumed in the above discussion of the long-line effect. always be some length of line for which large This length can be calculated by setting It is of little practical the denominator of Eq. may shift the entire tuning of the load which causes a change in the phase of the VSWR VSWR curve and might cause an unstable region to be shifted to the frequency of the fixedtuned magnetron. 6. For any lossy line. (6.Sec. not all radar transmitters are small enough to be located at the feed of a reflector- type antenna. no matter how small the and for a = VSWRs (essentially infinite) can be tolerated.0 J 100 Normalized ' 10 1.) 0.6 |l.3) equal to zero and solving for L c significance. 6.3) shows that there will .2 1. is several waveguide sizes. 2. 6. plotted in Fig. when the tuner setting is decreased.000 line length. The dashed portions of the curve correspond to case there are inaccessible frequencies just as with the frequency jumping. both of which One frequency is obtained when the tuner setting is increased. In the CW CW CW CW pulse magnetron. (6. Eq.4 1. The curves in Fig.0 1.

PL. U. On °.s to J ^ T3 v OJO 8S <N O Is J A °. si -J *8 © ^. >S T3 . ^^ 2 ^ crt aj -i— ++0OO 214 .^^ S>.> o>c^«On\OrtMOOHMnrt\onh)rt < ^oo^oo^rtin 1 •< vo <u o <_ o e w Z o < z JSP W c 3 J. '3 J" > ^* " ^ > a. o ~! go r -.»H •< > »i o C C ^ . 8-8 °° 00 p © . 3 3 V w i sp o O (U 4- 0/) G 6/> g - J g C u 3 g 3 P.sp « E ra (D PL. .0 r- n e2 0> 3 "^ <N 2 .0 o H — j" tJ S u r« _3 Oh Qw Pi 3 3 S Plh -c a "2 p. O/Jrr- -O Soox <*S 13 -O 2 a. ^ O M . V5 H 3 **u TO u 0. G S a Ph W <J " *o ~p > o ^p u o o"!° g^K^SJi© 9 s < DO ^ s 00 •*o w X 20.

but they permit the energy to flow unimpeded in a forward direction from the magnetron to the load. the In those applications where it is not possible to locate the transmitter directly at reflected length. Fig. Ltd. The selection of a magnetron for a specific applicawill little the tion is a job that requires careful examination of the various types available from All the tubes in Table 6. 7. the pulse one-half the than line longer transmission use a antenna or to wave from the load can be eliminated at the magnetron with a unilateral device such as an isolator or a circulator. The tubes were selected since they illustrate the parameters and characteristics of magnetrons. a decrease in efficiency.<5 EC . Examples of Pulse Magnetrons. several manufacturers in this country as well as abroad. tuning is not too high. 6. An oscillator followed by a power amplifier such as a klystron will not experience On the other hand. a phase shifter might be used to shift the entire If the curve to permit quasi-stable operation in between the points of frequency jump. where long life is a prime requirement.13. {Courtesy English Electric Valve Co.) Another technique for reducing the long-line effect is to decouple the magnetron to reduce its pulling figure. would be of little and efficient not is attenuation large with transmission line use of a since aperture blocking value for high-power radar transmitters. without electromagnet.. are relatively high power tubes.1) shows that a reduction of the pulling figure Decoupling the oscillator results in results in an increase of the skip length of the line. long-line effect because of the isolation provided by the amplifier. 1. 7. Photograph of type 7182 magnetron. There will usually be less over-all loss in efficiency for a given increase in skip length by decoupling the magnetron than by the deliberate insertion of loss. nor are they claimed to be typical of the many magnetrons which may be available. The tubes are all fixed-tuned. Equation (6. with the exception of the C-band tube.2] Radar Transmitters 215 Another antenna technique which is not a problem (Sec. as does the insertion of loss in the transmission line. The long-line effect may also be eliminated if the transmission line is longer than oneThe deliberate half the pulse length or if the attenuation in the line is sufficiently large.5). This will not be practical if the frequency separation between "skips" is small. The C-band tube is included since it was designed specifically for a commercial application (airborne weather-avoidance radar). 6. The tubes included in this table are not necessarily the best. Table 6. 1 lists the characteristics of five magnetrons covering the frequency range from L band to Xband.6). These devices prevent the energy reflected from the mismatched load from returning to the magnetron. . VSWR an Amplitron following an oscillator traveling in the reverse direction suffers not eliminate the long-line effect since energy or no attenuation. avoids long lengths of transmission line is the Cassegrain reflector geometry (Sec.

and a cross section illustrating the features of a 5-Mw version is shown in Fig.3 Most of the terms listed in the table have been defined previously except for the stability and the thermal factor. The velocity of the electrons constituting the electron beam in the klystron amplifier is modulated by the input signal. The length of the anode is two to three times that of ordinary designs. This arrangement results in more stable operation than is possible with a self-excited power oscillator. missing pulse is usually defined as one whose energy is 30 per cent less than normal. Because it is basically a power amplifier. A typical Jf-band marine radar might employ a fixedtuned magnetron generating 5 kw of peak power with an average power of 5 watts. Instead of coupling the output power through the side of one of the cavities. However.13. 6.14. it can be driven by a stable crystalcontrolled oscillator followed by a frequency-multiplier chain. The English Electric Valve Co. The stability is a measure of the percentage of missing pulses.3. and amplifier has The klystron . type 7182 Sband magnetron is of a radically different design from the usual magnetron. magnetron.) than conventional magnetrons of equivalent power. operating at a voltage of 5 kv and weighing 4 lb complete. one of the largest fields of application for magnetrons is in marine radars.14. stable output power with good efficiency and high gain. (Courtesy English Electric Valve Co.. resonant cavity extracts the RF power in the A density-modulated beam and delivers it to a useful load. The chief advantage of the klystron amplifier which makes it attractive as a radar transmitter is that it is capable of large. 1 are relatively high power magnetrons. Cross section illustrating the major features of a 5-Mw version of the outstanding characteristic of this design is its long life. 6. Klystron Amplifier Aerial plate The klystron amplifier overcomes the high-frequency limitations of conventional grid-controlled tubes by using to good advantage the transit-time phenomena in the electron beam. permitting cathodes large A Cothode and heater connection ^Locating ond / retaining flonge A Heoter Cothode Anode enough to dissipate the back-bombardment power generated by electrons returning to the cathode. The examples given in Table 6. it is coupled out symmetrically from one end. The reduced energy may be due to a lower-than-normal amplitude. Ltd.E. 26 photograph of the type 7182 without magnet is shown in Fig.216 Introduction to Radar Systems Heater connection [Sec.V. The Output window long anode plus the symmetrical output makes it natural to use a solenoid electromagnet. The resulting velocity modulation is converted into density modulation. had important application as a radar transmitter and fulfills a need which cannot be supplied by self-excited oscillators like the magnetron. shorter pulse width. which is an order of magnitude better E. 32 6. An Fig. 6. The thermal factor is a measure of the frequency change produced by a given change in anode temperature. A choke symmetrical output allows stable operation in the 77 mode without the need for strapping. 6. where small power tubes are widely used. or incorrect frequency.

6. always preferred in modern MTI radars whenever operational conditions gains In a klystron the RF input is well isolated from the RF output. instead. sketch of the principal parts of the klystron is shown in Description of Operation. power-handling capability. stable threein a are usual db to 40 from 30 of Gains "bottle." single are possible from a number of cavity tube. so that techniques. and cathode The klystron. Electrons are or voltage. conservative value of maximum emission density for short-pulse. the cavities is called the beam-accelerating A LC electrons are not intentionally collected by the anode as in other tubes. Heater Modulating anode three-cavity klystron.15. 6. anode voltage. the after the terminated by the collector electrode (shown on the right side of the diagram) cavity. RF cavities Interaction gaps Collector-^. The absence of a grid does not seriously poor because of its arrive at impair the coupling between the gap and the beam. the gap in high-power tubes does not usually contain a grid to the beam. beam has given up its RF energy to the output The RF voltage of the input signal is applied across the interaction gap of the first Low-power tubes might contain a grid structure at the gap to provide coupling cavity. necessitating special high-voltage High voltage produces X-ray radiation in the vicinity of the tube.3] Radar Transmitters 217 almost hence better MTI performance. for included as part of the electron-gun structure to provide a convenient means to the correspond which cavities RF The pulsing or modulating the electron beam. and high. and gains exceeding 80 db are possible. modulating anode is usually the electrons into a beam is called the electron gun. As a result. At the left-hand portion of the figure is the cathode.Sec. The lead shielding must be provided to protect operating personnel. However. 1 5. in the separate are tube the d-c and the RF portions of functions in an collector regions may therefore be designed to perform their respective optimum manner without concern for their effect on the RF fields. resonant circuits of lower-frequency amplifiers alsoserveas the anode since they are The positive potential applied to at a positive potential with respect to the cathode. Stable RF generators such as the klystron are permit. The voltages voltages. 2 The portion of the tube which focuses operation of klystrons is about 5 amp/cm A A life . handling required can be greater than 100 kv. which emits a stream of Fig. Fig. as the used can be Large gains mean that a low-power oscillator cavities. longelectrons. Those electrons which a experience wave) the sine of (peak maximum the gap when the input voltage is at a than those higher velocity a accelerated to be and will average than the voltage greater which arrive at the gap when the RF input is at a minimum (trough of the sine electrons . Diagrammatic representation of the principal parts of a chief limitations of klystrons are their relatively large size and high operating Large size is better suited to ground-based installations. the life of other types of microlife of a klystron may be made as good as or better than the wave power generators. depending upon the The input. 6.

Most high-power klystrons for radar application have one or more cavities between the input and the output cavities to provide additional bunching." or densitymodulated. The optimum magnetic CW Not all the power available in the klystron beam is delivered to the load. and beam collection are separate and independent in the klystron.watt amplifier. Each region can be designed to best perform its own particular function independently of the others.) the 41 per cent converted to RF power. power could be extracted from the density-modulated beam. and power to generate the magnetic focusing field if focusing is accomplished magnetically. losses in the output cavity account for 4 per cent. after traveling through the drift space. The intermediate cavities are not fed with energy The process whereby a time from the outside.15) is generally employed. RF interaction. 6.3 variation in velocity is impressed upon the beam of electrons is called velocity modulation. is dissipated at the collector while 18 per Of RF lost in a klystron without grids.450 Mc. the stray electrons would impinge upon the metal structure of the tube and cause it to overheat or possibly be destroyed. or slightly below. the electrons are collected by the collector electrode which is at. while the production of secondary electrons with virtually zero velocity accounts for an additional 6 per cent. only 25 per cent is delivered to the load. The over-all efficiency RF would have to include the heater power. If the interaction gap of the output cavity were placed at the point of maximum bunching. by means of an iris. since none are present. The magnetic field confines the electrons to a beam and prevents them from dispersing. an field strength is fairly critical and is not necessarily uniform along the length of the tube. operating at a frequency of 2. In low-power tubes the beam might even be focused by positive-ion space charge. 33 Of the total d-c power in the electron beam.218 wave). For example. The regions of beam formation. In some klystrons the electron beam may be confined by electrostatic fields designed into the tube structure and external magnets are not required. transit-time loading accounts for (Less power would be 6 per cent. Introduction to Radar Systems [Sec. utilizing grids at the RF interaction gaps and employing positive-ion focusing. or tunnel. CW Of the cent is 59 per cent not converted. the cathode is outside the RF field and need not be restricted to sizes small compared with a wavelength. or if a waveguide output is used. 6. The advantage of the klystron over other microwave tubes in producing high power due to its geometry. and hence higher gain. power required for cooling. In order to counteract the mutual repulsion of the electrons which constitute the axial magnetic field (not shown in Fig. If the beam were not properly confined in a high-power klystron. Large cathode area and large interelectrode spacings may be used to keep the emission current densities and voltage gradients to reasonable values. potential. RF conversion efficiencies of practical high-power klystrons seem to lie within the range of 35 to 45 per cent. In the drift space. In the three-cavity klystron the second cavity may be tuned to the frequency of oscillation. . only 41 per cent is converted to RF power. After the bunched electron beam delivers the RF power to the output cavity. Thus the conversion efficiency of this particular tube is 25 per cent (which is low for klystrons). the cathode Power is extracted from the output cavity to the load by a coupling loop. It can be of a shape and size most suited for satisfying the average or peak power requirements without regard for conducting RF currents. those electrons which are speeded up during the peak of one cycle catch up with those electrons slowed down during the previous cycle. An example of the division of power in the Sperry SAS-37 klystron was reported by Learned and Veronda for a 200. 41 per cent lost in the drift tube. The result is that the electrons of the velocity-modulated beam are "bunched. beam. The is only function of the collector electrode in the klystron is to dissipate heat. or even a slightly higher frequency for greater efficiency.

the phase shift between the RF pulse modulator. When the output window must be placed at the interaction gap. good RF performance and good heat dissipation. External and Internal Cavities. and it requires a modulator capable of handling the full power of the beam. between RF interval in the collector Of the three methods. Cavities external might be twice cavities external of range tuning practice the tune and maintain. the of efficiency the and pulses. When the cavities are wholly within the vacuum envelope. in klystrons. In most other tubes the functions of electron emission. like anything else. Good RF performance usually requires the tube electrodes to be small compared with a wavelength. only an input and an output window are required for the tube and they need not be placed directly at the By contrast. In one design. External cavities are often preferred because of their wider tuning range and more convenient method of tuning whenever the output window does not limit the powerhandling capability of the tube. (2) the RF input signal. change to any significant degree as not input and the RF output of the klystron will voltage is applied to the modulating anode. 34 in the be used may tubes vacuum Hard power. 6. peak input less than \ per cent of the Also. Unfortunately. the beam current should also be modulated when the purpose in the useful no to dissipated be will power modulated. these requirements cannot always be satisfied simultaneously. the entire beam current must be pulsed as well. RF interaction. windows of both external and internal cavities. One of the major factors which has restricted the power available from klystrons has been windows capable of coupling the output power from the the problem of obtaining vacuum envelope to the load. the modulating anode requires the least modulating power. while good heat dissipation requires large structures. The last mentioned is controlled by an electrode in the gun called the modulating anode. as in the case of externally tuned cavities.3] Radar Transmitters 219 The design flexibility available with the klystron is not present in other tube types considered in this chapter. The only control power necessary is the very small amount required to charge and this is discharge the capacitance of the klystron gun and its associated circuitry. . or (3) the Dielectric heating occurs in the electron klystron-beam current. cavities may be placed external to the vacuum by sealing the tuning The resonant This is called a interaction gap with a suitable low-loss vacuum-tight insulator. Modulating Anode.Sec. RF input signal is In method 2. otherwise beam tube will be low. and collection of electrons usually occur The design of such tubes must therefore be a compromise between in the same region. This heating is in addition to the RF heating due to the dielectric losses in the ceramic when transmitting RF power. except for the traveling-wave tube. of dissipating heat in the collector. system are easier to vacuum to the window and is usually ceramic. internal cavities are probably more suitable at the higher frequencies. minimizing any time jitter. Three possible methods of pulsing a klystron are by turning on and off (1) the klystron-beam accelerating voltage. is not unlimited. each cavity. When the beam is pulsed by turning the accelerating voltage on and off. the inside of the ceramic window may be bombarded by secondary electron emission emanating from the downstream tip of the output gap. the tube with external cavities requires a window at interaction gap. and is power modulating pulsed peak the independent of the pulse length. However. Two types of resonant cavities have been employed They differ in being within or outside the vacuum envelope of the tube. Other factors limiting large powers are the difficulty of RF operating with high voltages. The high-power capability of the klystron. This is similar to plate modulation of a triode or magnetron. The output window can become overheated and fail. and of obtaining sufficient cathode emission current. In that of similar cavities tuned from within the vacuum.

3 cutoff characteristics of the modulating anode permit only a few electrons to escape from the electron gun during the interpulse period when the beam is turned off. when the plans for the British magnetron were made available to the United States in 1940. The important in radar application since the receiver sensitivity will be degraded if during the interpulse period to cause the stray electron current noise to exceed receiver noise. where they were more directly involved at that thwarting aggression. where efficiency is not too important. They have also been operated with 30 of peak power and at an efficiency of 43 per cent. Mw Mw . and was developed for use in a linear accelerator. 1949. 6. relatively low power reflex klystron for use as local oscillators in microwave superheterodyne receivers.220 Introduction to Radar Systems [Sec. In such applications a Gaussian-shaped pulse is preferred since its spectrum falls off rapidly. but the incentive in the United States for a high-power microwave tube suitable for radar application was not as great as in Great Britain. Hansen concerning the interaction of electron beams and resonant microwave cavities. became the basic RF power generator for radar transmitters. in the United States. The largest pulsed klystron power before the development of these tubes had been about 30 kw and was achieved by the British during the war. These tubes developed a peak power of 20 at an efficiency of 35 per cent. Although the principle of the klystron was demonstrated by several investigators during the late 1930s. the potential worth of the klystron as a high-power microwave generator was first demonstrated by the university scientist as a by-product of his pursuit of fundamental knowledge. The failure to press the development of the high-power klystron during the war was due in large part to the wartime necessity of concentrating the relatively scarce technical efforts in a limited number of fields. W. resulting in reduced sideband energy. The modulating characteristics (output RF voltage vs. An important application of pulse shaping is in air navigation systems such as Tacan. a considerable effort was concentrated.857 Mc in-the 5 band and could be tuned over a frequency range of 100 Mc with a' flexible diaphragm which constituted one wall of a cavity. The Gaussianshaped pulse may be readily approximated with the modulating-anode technique. The tuning mechanism and the cavity resonators were located within the vacuum envelope. but at a sacrifice in tube life. Examples of High-power Klystrons. where it is necessary to transmit a pulse with as little sideband energy as possible in order to avoid interference between adjacent or nearby channels. 37 The klystron was based on the pioneering ideas of W. 38 It was designed to operate at a frequency of 2. time in Consequently. The reflex klystron may be used. 35 36 A Gaussian-shaped pulse may also be desired in radar applications in which interference to nearby radar receivers at slightly different frequencies may be more important than the loss in range accuracy and resolution obtained when a Gaussian pulse is used instead of a rectangular pulse. in low-power short-range radars. rather than the klystron. but the potentialities of the klystron for high power were not exploited until after the war. As is true of many devices which prove to be of practical value. modulating anode voltage) of the modulating anode can be made linear over a portion of the operating range so that the output waveform may be shaped as desired. on the Many experimental klystrons were built during the late thirties. The first high-power klystron tube capable of megawatts of peak power was operated at Stanford University in March. the magnetron. rather than by the systems engineer. however. The Varian klystron actually preceded the magnetron invented by Randall and Boot. The efficiency and power output of the reflex klystron is too low to be of consequence for high-power radar transmitters. is This sufficient electrons are present . However. credit for its invention is usually given to the Varian brothers. The peak power was greater than that delivered by any other tube at any frequency.

6. 60 cps.6 db. 2. The development of the Stanford 30-Mw klystron represented a considerable achievement and opened new possibilities for the radar systems engineer. The pulse-to-pulse phase assembly. - Mw A klystron.16.) is a three-cavity amplifier. £° within than less VA-87 is the of jitter gain is 57. available tubes. watts. It weighs The Stanford tube scientists also developed a family of sealed-off klystrons suitable 41 for radar application. S. 235-lb magnet requires a and 65 lb any 4-msec period. corresponding to an average power of 2. The pioneering work of the Stanford scientists was followed by the engineering and packaging of sealed-off klystrons by the tube industry.SEC . 6.500 hr. The tube could operate peak current of with a maximum beam accelerating voltage of 400 kv and a maximum Pumping had amp. The superior average power capabilities of the klystron permitted radars to be designed with considerably more power than possible previous to its introduction. The 39 40 operated at L band with 2 to Sperry SAL-36 of pulsed power with an efficiency exceed4 ing 40 per cent. and 185 were 325 kv values operating typical 250 amp. The first two klystrons commercially available for radar application were the Sperry Gyroscope Company SAL-36 and the Varian VA-87. not only bebut voltages. while the VA-87 has four cavities. These were based on the principles learned from the 30-Mw They operate in the L. It is a four-cavity tube with a largesynchronously tuned saturation gain of about 61 db and a synchronously tuned for tuned tube is the operation In normal signal bandwidth of about 20 Mc. 6. of approximately life average tubes have had an 1. large difficulties in working with cause of the onset of relativistic effects. only of an RF input 2.16. because of the represents a practical upper limit for accelerating voltages. Twenty-two of these klystrons were used Stanford University in the 220-ft linear accelerator to produce electrons with energies in the In operation. although A voltage of 400 kv to be employed to maintain the vacuum during operation.2. The basic design has apparently been incorporated in commercially respectively. it permitted the development of MTI radar systems far better than was possible with the magnetron. . these vicinity of 1 billion volts. Fig.5 and kv of 90 voltage beam maximum power output with a and the The half-power bandwidth under these conditions of operation is about 27 Mc. and the pulse repetition frequency. and 1 . results in an increase in electron mass rather than at velocity. Varian VA-87 pulsed-amplifier (Courtesy Varian Associates. The VA-87 delivered nominally The SAL-36 1 Mw of peak power at S band. At higher accelerating voltages the electron velocity is close Further increase in voltage to that of light.2 Mw.3] Radar Transmitters 221 The pulse width of the Stanford tube was 2 fisec. In particular.0. tube.4 kw when the peak power was 20 Mw. and X bands at pulse powers of 3. and the power gain was 35 db. A three-cavity design was used. photograph of the VA-87 is shown in Fig.

Eimac X626 pulsed-amplifier klystron. Because of these radiation. Inc. It is also possible in some applications to use electrostatic space-charge forces to focus the beam. was also developed by the Varian Associates and is known as the VA-842.17. 6. The tube utilizes three cavities external to the vacuum system.) surrounds the tube in order to protect operating personnel from X-ray A tube with similar characteristics. From 1. 6. p. The peak power is . The electron-gun portion of the tube is operated immersed in about 800 gal of oil. 42.17. lb of lead .222 Introduction to Radar Systems largest klystrons.3 and in terms of average power is the Eitel-McCullough X626 pictured in Fig. The drift spaces in a klystron with space-charge focusing are short and of large diameter. The gain is 30 db.06. The average power from this tube is 75 kw « It stands 10 ft 6 in. Even though the power output of the X626 klystron is large by any standards. both physically [Sec. The microwave cavities must be especially designed to operate with this type of beam. With electrostatic focusing the electron beam is first made to converge and then is allowed to diverge. It delivers a relatively long pulse of 2 msec at a 30-cps pulse repetition frequency.000 to 4 000 1 One of the Fig. {Courtesy Eitel-McCullough.25 Mw." The klystron tubes described above all require external electromagnets to confine the electrons to a beam. The conversion efficiency is 43 per cent. while the other parts of the tube are water-cooled. it has been claimed (Ref. 6. The interaction gaps are usually gridded to minimize stray coupling between the RF fields of the cavities and to give efficient interaction between the RF fields and the beam. 3) that "the design of a tube to handle 10 to 15 times the'power of the X626 would be a relatively straightforward (but not small) task. high and weighs 800 lb. and the tuning range is from 400 to 450 Mc. but with internal rather than external cavities. and the duty factor is 0. exclusive of auxiliary equipment The tube is designed for very long range radars.

Sec. 100 Fig. A Thirty watts of drive power is required. control grid generates a Gaussian-shaped pulse to minimize interference between adjacent navigation channels. 6.125 to 8.5 Gc.2 per cent of the beam voltage to get full rated power of 25 kw. which is only a modest power for most radar applications. The peak power available transmitters used for air-navigation aids such as Tacan. . has average power outputs falling on the curve.3] Radar Transmitters 223 restrictions the power output of space-charge-focused tubes has not been as large as with electromagnetic focusing. kw are permissible. low noise. The duty cycle is 0. and the efficiency is better than 30 per cent. the SAL-219. and comparatively good efficiency.21 5 Mc. The klystron amplifier may be used for klystron amplifier delivers 20 kw of electromagnetically focused VA-849 JV-band power with low incidental noise in the frequency range from 7. It pure C and is tunable over 60 Mc. It employs a single electron beam and contains two The first cavity encountered by the electron beam cavities separated by a drift space. A companion tube. from the tube is 25 kw.18. 6 8 10 Frequency. and paralleling of tubes to achieve greater The as well as pulse applications. that indicated by Fig. The Monofier is a form of klystron with good frequency stability. {Personal communication from C. An example of an electrostatic space-charge-focused klystron amplifier is the SAL-89 developed by the Sperry Gyroscope Company.18 since it is limited in power by the output connector and not by the focusing. 6. M.000 cps. Veronda of the Sperry Gyroscope Company?) A plot 500 of the maximum average power output from space-charge-focused klystron is Peak powers up to about power is comparaThe power output of the SAL-89 klystron described above falls short of tively simple.025.18. amplifiers as a function of frequency shown in Fig. 43 This is a three-cavity tube covering It was designed primarily for ground-based the frequency band from 960 to 1 . 6. 6. The grid requires a total voltage swing of only 3. has a gain of 50 db with a bandwidth of 30 applications signal with little extraneous noise is especially important in long-range C power at JVband) and The noise in the VA-823 series of klystrons (5 kw of CW CW CW Mc AM FM CW W A is 100 db below the carrier in any 1-kc channel more than 1 kc removed from the carrier. Average power of pulsed-klystron amplifiers with space-charge focusing as a function of frequency. and the pulse repetition frequency used in the air-navigation-aid application is 7.

3 a self-excited oscillator which velocity-modulates the beam. The Monofier may also be designed to operate pulsed.5 Dain. The broadbanding of a multicavity klystron may be accomplished in a manner somewhat analogous to the methods used for broadbanding multistage IF amplifiers. The increase in bandwidth is accompanied by a decrease in gain from about 57 db with synchronous tuning to about 44 db with stagger tuning. It has been reported that the half-power bandwidth of the S-band VA-87 klystron amplifier can be increased from its synchronously tuned bandwidth of 27 Mc to a value of 77 Mc by stagger tuning. 45 Consequently.8 per cent bandwidth. 47 This represents a 2. although there has been some cavity stages. with 30. One of the more common techniques is to stagger-tune the frequencies of the various Stagger tuning of a klystron is not strictly analogous to stagger tuning a conventional IF amplifier. The gain of a four-cavity klystron can be 60 db or greater. but the gain is increased and wider bandwidths may be achieved. Stagger tuning Mc Stagger tuning Number of cavities Synchronous tuning Synchronous tuning 2 3 20 45 70 4 30 40 2 1 6. The adjustment of the cavity frequencies for broad bandwidth in a multicavity stagger-tuned klystron is a complex procedure and is probably better performed at the factory than in the field. the RF conversion efficiency of a two-cavity klystron is of the order of 20 to 30 per cent. i5 &1 Almost all high-power klystron amplifiers employ more than two cavities. and the capability of obtaining wider bandwidths. In the UHF band.5 4 9 Two t two-cavity tubes in cascade 64 40 2. As the beam traverses the drift space. db Bandwidth. cavity. The entries in the table show that the gain and the bandwidth of a single four-cavity klystron are considerably better than those of a pair of two-cavity klystrons in cascade.2. using numerical parameters considered typical of a klystron operating at 700 Mc. 46 for klystrons with two.224 is Introduction to Radar Systems [Sec. 6. three. ~~ Comparison of Computed Gain and Bandwidth of Multicavity Klystrons at 700 Met Gain. Bandwidth of Multicavity Klystrons. The RF on passing through the second. does not increase significantly as more cavities are added. and four cavities.to 40-db gain. Thus the Monofier serves the function of CW systems. or two tubes. Electrostatically focused CW Monofiers at X band are capable of 1 or 2 kw of power with efficiencies of 20 to 25 per cent. Some of the results. greater efficiency. The efficiency of a three-cavity amplifier might vary from 35 to 45 per cent. because interactions between nonadjacent cavities cause the tuning of one cavity to affect the tuning of the others. The efficiency Table 6. the gain and bandwidth have been calculated by Kreuchen et al. work to serve as a guide. From the theory of electron bunching in a multicavity klystron. Possible areas of application include compact and pulse-doppler radar in the density-modulated is power beam extracted catcher. A special modification of the theoretical . as reported by Dain.2. The advantages of the additional cavities are an increase in gain. the velocity modulation is converted to density modulation. the broadbanding of klystrons is somewhat empirical in practice. The gain-bandwidth product increases significantly with additional cavities. 54 are presented in Table 6. with gains of the order of 20 db.

or 1 44 secondary cavity of the usual output cavity is coupled not only to the load but to a 52 of 40 db. Electrons emitted by the cathode of the traveling-wave tube are focused into a beam and pass through the RF interaction region. tubes.53 Bandwidths of multicavity klystrons may be as large as 1 to 1 2 per cent or greater. klystron amplifier In practice.wave tube. The velocity modulation is down transformed into density modulation (bunches) after traveling a short distance .54-59 4 The wide bandwidth of the traveling.500 volts. stagger tuning enables the bandwidth of the multicavity to £ per cent to values the to be increased from a synchronously tuned bandwidth of \ order of 5 per cent or more. 51. the to those of the klystron. The velocity of propagation of electromagnetic energy is For slowed down by the helix and is nearly equal to the velocity of the electron beam n the I line. 10. The adjustable loss and frequency. 6. 6. a popular form for low-power.wave tube in the form of traveling waves. .n/ .wave the radar engineer. velocitymodulates the electron beam just as in the klystron. .19. klystron in the form of standing waves. fers energy from the d-c beam to the RF wave.4] Radar Transmitters 225 VA-87 using a six-cavity driver section followed by an output section consisting of a halfdouble-tuned circuit with an inner cavity identical with that of the VA-87 results in a arrangement cavity double-tuned the (In Mc.4. causing the RF wave The RF signal. diagrammatic representation of a traveling-wave tube is shown in The electron optics of the traveling. wave this reason. After delivering their d-c energy to be the RF field. amplified enters via the input coupler and propagates along the periodic The periodic structure is shown as a helix in Fig. ) The gain of the tube was in excess and 10 kw peak of capable 5 commercial version of this tube is the VA-839. ii I H. ir I I 1*1 I (interaction region) I beam ' RF input RF output Fig. This is comparable with that available with high-power traveling-wave 6. electromagnetic wave and the electrons results in a cumulative interaction which transto be amplified. periodic delay or a structure. the electrons are removed by the collector electrode. Both employ the principle of velocity modulation. Description. power bandwidth of 4. and the electrons will be in synchronism with The synchronism between the the wave if they are accelerated by about 1. the radar within a is required (Sec. For example. if the beam the along will travel wave the helix. is 13 times as long as the axial length of the with one-thirteenth the speed of light. with a reduction in the gain. A large-bandwidth tubes. Electron S . 6. it is sometimes called a slowvelocity the but light speed of the about with wire the along travels wave the helical line wire of propagation in the direction of the beam is somewhat less. the traveling.Sec. 8 per cent.8) or where it is desired to rapidly tune wide frequency band to avoid deliberate jamming or mutual interference with nearby Bandwidths of the order of 10 to 20 per cent are possible with the travelingradars.wave tube are similar in many respects Fig. 6. The RF signal to structure.19. when applied to the input coupler of the traveling. resolution wave amplifier at the power levels required for long-range radar applications. Traveling-wave-tube Amplifier -47. It is Mw average power at an efficiency of 40 per cent. amplifier is its chief attribute of interest to A wide bandwidth is necessary in applications A ttenuatio n where good range CothodeHeater-* // anode J Gun 4#mm CollectorElector- >. Diagrammatic representation of the traveling-wave tube.19.

The cathode. is used in the traveling-wave amplifier. Other types of slow-wave structures must be used for high-power levels. The high-power traveling-wave tube is very similar to the klystron.wave tube from the reflection of a portion of the forward wave at the output coupler.4 When the electrons are bunched. The is well suited to low-power. One of the major differences between the usual klystron and the traveling-wave tube is that feedback along the periodic structure is possible in the traveling-wave tube whereas the back coupling of RF energy in the klystron is negligible. The feedback energy must be eliminated if the traveling. Another major limitation is the periodic structure. The attenuator must be designed so that the reflected wave from the output coupler is attenuated much more than the input wave is amplified. however. 6. the velocity-modulated electron beam is again converted to density modulation. In some traveling. 6. but it. If sufficient energy were fed back to the input. The faster space-charge wave is used in electron accelerators. enabling the structure to dissipate considerable heat. the traveling-wave tube might be considered as a limiting case of the multicavity klystron. Energy traveling in the backward direction may be reduced to an insignificant level in most tubes by the insertion of attenuation in the periodic structure. the concentration of space charge produces a repelling effect and the beam becomes debunched that is. if the traveling-wave tube is to achieve power is . Power and Bandwidth. Because of feedback the traveling-wave tube is inherently a less stable device than the klystron. or some substitute. The necessity for an attenuator capable of handling large average powers is one of the major restrictions on the output power of a traveling-wave tube not found in the klystron. These standing waves may be described as the beating of two space-charge waves traveling along the beam with different phase velocities. broadband applications but cannot be used at highpower levels (greater than approximately 10 kw) since it does not dissipate heat effectively. is necessary for proper operation. or it may be lumped. In addition. but this technique is probably better suited to low-power than to high-power tubes.226 Introduction to Radar Systems [Sec. The attenuation may be distributed. the traveling. the over-all size of the traveling-wave tube fortunately. One of these limitations is the problem of obtaining a feedback attenuator which can dissipate the necessary power in a small space and at low voltage-standing-wave ratio. Feedback energy might arise in the traveling. As the electrons travel farther along the tube. Thus standing waves of space charge exist along the beam. The slower space-charge wave. while the other wave has a phase velocity larger than that of the beam.wave tubes a considerable reduction in weight is possible by using permanent magnets periodically spaced along the tube. and the tube. It must be carefully matched to the periodic circuit. Unhelix powers does not have as wide a bandwidth as some of the lower-power structures. An axial magnetic field (not shown in Fig. the traveling-wave tube should be capable of as large a power output as the klystron. density modulation is converted back to velocity modulation. One of the waves has a phase velocity smaller than the beam velocity. The periodic structure usually associated with traveling-wave tubes is the helix. .wave amplifier is to function satisfactorily. Thus. The introduction of attenuation reduces the efficiency and power output. In some respects. In principle. when coupled to the electromagnetic wave. but it is usually found within the middle third of the tube. RF interaction region. In practice. the process is repeated. and the collector are all separate and can be designed to perform their required functions independently of the others. The electromagnets required for focusing might be quite large. the type of periodic structures suitable for high usually not small. It seems that those periodic circuits best suited for broad bandwidth have the lowest power-transfer and heattransfer dissipation capabilities. it is found that there are limitations to very high power output.wave tube would oscillate.19) confines the beam and prevents it from dispersing.

a sacrifice in bandwidth must be made. and the efficiency was about 14 per cent. per 10 more than of bandwidth over a The physical structure of the magnetron oscillator and the Amplitron are similar. The saturation gain of the S-band tube referred to above was about 20 db. exceptionally high efficiency. there is little advantage to be gained with a travelingwave tube as compared with multicavity klystrons. The peak power output is 2 Mw.001.or medium-power traveling-wave tubes. but low Space-charge hub Conducting vanes ties Space chargi spoke Phase velocity Conducting straps Catho roup velocity Anode Input pla Fig. broad bandwidth.Sec. with db.5] levels Radar Transmitters 227 comparable with other tube types. of of the order 10 are Gains cent. Amplitron and Stabilitron 60 " 6 Amplitron. bandwidth. not be as spectacular as the octave bandwidths possible with low. liquid-cooled tube intended to cover the major portion of the S-band radar frequency range. however. just as with the klystron. 6. it is nevertheless a significant bandwidth for most radar applicaThe gain and efficiency of a high-power.) tube. for example. {Courtesy Raytheon Company. A drawing of the workings of an Amplitron is shown in Fig. Basic structure of the Amplitron. The VA-125 is similar in many respects to the VA-87 klystron amplifier. the duty cycle is 0. Traveling-wave tubes have been built which operate at the megawatt level with the 57 Although a 10 per cent bandwidth may order of 10 per cent bandwidth at 5 band. A reduction in gain accompanies a large usually not as good as those of the klystron.20. 6. are capable of 5 to 10 efficiency conversion per cent than 85 better duty factors of approximately 0. They deliver about the same peak power and can be used interchangeably except that the VA-125 traveling. 6. The Amplitron is a crossed-field amplifier characterized by high peak and average power output.wave tube requires additional input power because of its lower gain.20. Example of Traveling-wave Tube for Radar. but The chief physical difference between the two their characteristics are quite different.002. The operation. A magnetic field is applied parallel to the axis of the of peak power at Amplitrons at L band. couplings (an input and an output) and the external two Amplitron uses the is that gain.000 Mc.5. heater power is usually required for starting the Amplitron or during . and The power gain is 33 db. Its bandwidth is 300 Mc at a frequency of 3. If the bandwidth is too small. Mw magnetron has but one. broadband traveling-wave tube are tions. The Varian traveling-wave-tube pulsed amplifier known as the VA-125 is a commercially available broadband. with a 2-^sec pulse width. The electrons No external originate from a continuously coated cathode coaxial to the RF circuit. 6.

but dependent on the d-c input. just as in the magnetron. traveling-wave tube in which the space-charge wave is coupled to a backward wave is known as a backward-wave amplifier. . . or waves whose phase velocity is opposite to the power flow. (Courtesy Raytheon Company?) RF periodic structure to support backward waves. 54 63 66 . In the travelingwave-tube amplifier the space-charge wave interacts with a forward wave. The Amplitron is a crossed-field device in that the electron beam is perpendicular to both the electric and magnetic fields. 6. The RF output and the RF input are decoupled. The nonreentrant circuit of the Amplitron permits a broader bandwidth oscillator. The Amplitron behaves as a saturated amplifier rather than as a linear amplifier. the RF circuit of the Amplitron is related more to that of the traveling-wave tube than to the magnetron. a wave whose phase velocity is in the same direction as the power flow. The space-charge waves in the Amplitron are formed by the interaction in between the electron beam and the crossed electric and magnetic fields. It acts [Sec.20 shows that the electron beam reenters the interaction space just as in the However.21. gains of 15 db. the RF circuit is not reentrant. unlike that of the magnetron. In this sense. that is. Although a saturated .wave RF circuit with which the electrons interact and as an electrode for the collection of electrons. Its operation is similar some respects to the traveling-wave-tube amplifier since amplification occurs because of an interaction between a traveling electromagnetic wave and a rotating space-charge wave. Figure 6. bandwidths of 1 5 per cent. magnetron than the reentrant circuit of the conventional magnetron oscillator. and output power of several megawatts peak. The crossed-field device called the magnetron amplifier uses the the A forward wave and can attain efficiencies of about 50 per cent. It is also possible for Fig. 6. The characteristic of a saturated amplifier is that the magnitude of the RF output is independent of the RF input.5 anode as both a slow. Space-charge waves in a crossed-field-magnetron-type device can couple with either a forward or a backward wave.69 The crossedfield device which couples the backward wave to the space-charge wave (Amplitron) has a higher rate of gain than the forward-wave device (magnetron amplifier) and for the same length will be more efficient. QK 622 5-band Amplitron.228 Introduction to Radar Systems consists of a series of vanes.

6. When used with a line-type modulator. Contours show constant modulator input. a McGraw-Hill publication.: Sec. 64 It produces a 622 pulse Amplitron is shown in Fig. for example. 1960. to its applied when the RF drive power peak power of 3 The gain is 7. 6. If. may QK Mw 10. The conversion efficiency of an Amplitron is defined as follows RF Efficiency power output — RF power input (6. and at some other frequency than the input signal. The Raytheon type input is no less than 550 kw. there are but few restrictions on its use as a pulsed amplifier for most radar AM A saturated amplifier is compatible with frequency modulation. 1960.22.000 Fig.22 for the Raytheon QK 520 L-band Amplitron. and it be used with radars designed with pulse compression. is shown in A plot of the RF power output as a function of the RF and d-c modulator input power 66 For constant d-c Fig. This departure from saturated amplifier behavior results from a slight increase in efficiency with large RF input power and because the input power reappears unattenuated at the output and adds to the RF power generated by the Amplitron itself. the amplifier.2 1. except when the RF input becomes comparable with the RF output. between the area in which the input does not control RF output and the area in which performance is satisfactory is well defined and of negligible width. 6. by permission. 6. 6. At reduced approach or even exceed 80 per cent at some frequencies.5] Radar Transmitters 229 voice communiamplifier cannot be used in some applications. poorly The transition region defined. kw 10. from the Apr. the RF output power is relatively independent oftheRF input power. such as. copyright.000 100 10 100 1. Plot of the RF power output as a function of the RF power input for the Raytheon QK 520 L-band Amplitron. it will cover the frequency band from 2. 6.100 Mc without mechanical or electrical Efficiencies greater than 70 per cent are observed over the entire frequency adj ustment.5 db. at a given level of d-c power.21 is is 700 kw. the device ceases to act as an In this region (shown shaded in Fig. issue of Electronics. 29.™ reprinted. (From Brown. cations.4) modulator power input to Amplitron . the RF output is noisy. RF input is reduced below a certain level.000 Peak RF power input. They applications.005. band.) power output. 1251b.22).000 1. input power. the gain is increased. Its duty cycle is 0. Eleven decibels is obtained when the peak power The weight of the completely packaged QK 622 as shown in Fig.900 to 3.

The high efficiency permits operation at considerably greater power levels than other tube types with similar heat-dissipation capabilities but of lower efficiencies. The effective over-all efficiency of a chain of Amplitrons therefore can remain high. An increase in efficiency by a factor of 4 from 20 to 80 per cent results in a 1 6-fold increase in the amount of output power delivered to a load. which describes a similar as the current is phenomenon in oscillators where the frequency is changed or pushed changed. An unusual feature of the Amplitron is its ability to operate without a cathode heater. typically 0. for example. (The advantages to be gained from high efficiency apply to any type of device and are not proprietary to the Amplitron alone. The low insertion loss makes it possible to pass the received echo signal back through the Amplitron before entering the duplexer.) The Amplitron has the advantage that it is one of the most efficient of the high-power microwave amplifiers.05 per cent. if the efficiency were 80 per cent. The phase shift varies less than 0. The Amplitron operates with low RF voltages and possesses good stability.5° for a 1 per cent variation in anode current with Amplitrons like the 622. Permanent magnets are usually used. being of the order of a fraction of a degree per ampere. 6.230 Introduction to Radar Systems [Sec. The tube starts without a cathode warmup period whenever RF drive power .0 db. This differs from other amplifiers in which the reversed signal is highly attenuated. magnetic field is required just as with the magnetron. It is reported that the output spectrum of a particular Z. the tube could deliver 40 kw while dissipating 10 kw. The Amplitron acts as a passive transmission line when the high voltage is removed. compared with other microwave QK QK low. The high efficiency of the Amplitron is one of the major reasons for its ability to generate large powers with structures of reasonable size. that a particular tube structure can safely dissipate 10 kw of heat and that this is the only limitation on the total power the tube can generate.5 This is a conservative definition since the RF input power is not lost but appears as part of the output. Amplitron voltages are lower than those of the klystron or the traveling-wave tube and are comparable with those of the magnetron. In a low-gain amplifier the input power which appears at the output may be a sizable fraction of the total. the useful power output would be 2. Such would be the case when several power tubes are operated in parallel or when individual transmitters feed individual elements of a phased array antenna. However. If the efficiency of the tube were 20 per cent. The magnitude of the A magnetic field represents a compromise between magnet weight and the higher efficiency which can be obtained with large magnets. by analogy with the term frequency pushing.5. Therefore an RF signal traveling in the from the output to the input suffers little attenuation. On the other hand. The percentage of missing pulses in the 622 is less than 0. Phase pushing in an Amplitron is usually quite small amplifiers.-band tube remains unperturbed regardless of phase position of output mismatch and VSWR up to a value of 2.5 kw and the power dissipated 10 kw. The phase shift through the Amplitron caused by a change in the d-c current applied to the device is called phase pushing. The quality of the output spectrum from the Amplitron is but little affected by changes 61 in load conditions.2 to . Assume. the low attenuation in the backward direction requires that a high-power circulator or some other isolation device be used between Amplitron and driver to prevent the reflected power from interfering with the driver portion of the transmitter or from building up into oscillation. Duplexing may therefore be accomplished at a lower power level than if it had to be placed at the output of the tube. Low phase pushing is important in radar applications where zero or negligible phase shift must be maintained between input and output. Isolation is also needed between Amplitrons when they Its insertion loss is 1 reverse direction operate in cascade.

The gain of an Amplitron can be increased at the expense of the bandwidth by the use transof positive feedback produced by inserting mismatches in the input and output mission lines.Q cavity absorbs that energy not at the resonant frequency of the cavity.5] is Radar Transmitters 231 present prior to application of the modulating pulse. Amplitron (Plotinotron) Partiol reflection (mismatch) Useful load Fio. The frequency of oscillation is also determined by the resonant frequency of the cavity. QK l. should be capable of delivering a useful RF average output power of more than 100 kw 71 at S band. These are inserted so that the RF energy reflected from the mismatch n reflected the output line will be returned to the mismatch in the input line and be again 70 can be db the order of of 30 Gains energy. The phase shift vs. to employ a on the input side to prevent unwanted oscillations from building up because of the reflections from mismatches at the output and input. one application of the Amplitron has been as a booster tube to increase the power output of existing radar equipments. 6. a slight change in frequency permits the cavity to correct for substantial phase shift which might be introduced by such factors . the output reflection to the cavity reflection and return is an integral multiple of 2tt radians and if the gain around this loop is greater than unity. No tuning of the tube is necessary because of its broad bandwidth. times the cavity reflection coefficient. A tube the size of QK 622 (average power of 15 kw). Energy which is at the resonant frequency is re-reflected and passes through the device in the forward Steady oscillations will occur if the total phase shift from direction with amplification. narrowband tunable cavity is connected to the tube input.23. 6. frequency characteristic of the stabilizing cavity has a larger slope than any other part of the circuit. and the duplexer can often be used without change on ferrite the input side of the Amplitron.Sec. of extremely large power. The absence of a heater 622 is claimed to be in excess of The life of the results in longer tube life. consequently. times the attenuation in the backward direction. The Amplitron the is capable. in principle. however. 6. 64 Because of its relatively low gain but high power and high efficiency. It is simply added to the output of the existing radar to give an order-of-magnitude increase in radiated power. A high-g. by the addition of RF feedback and the application of a stabilizing cavity (Fig. A mismatch is connected between the output of the tube and the load. input with the in such phase as to add obtained with bandwidths of the order of per cent. The high Q of the cavity resonator acts to stabilize the frequency of oscillation.23).000hr. A portion of the power output from the Amplitron is reflected by the mismatch and travels back through the tube in the direction of the input with little or no attenuation. The Amplitron can be made to operate as a highly stabilized oscillator Stabilitron. with an anode cooled with high-velocity liquid. Block diagram of Stabilitron oscillator consisting of an Amplitron with a high-Q cavity attached to the input and a broadband mismatch reflection on the output. The high. It is usually necessary.23 is to adjust the phase of the feedback loop to be compatible with the resonant frequency of the cavity. The purpose of the phase shifter shown in the diagram of Fig. 6. Mechanical tuning over a 1 per isolation device i \ cent range is possible. It is not necessary in fixed tuned devices or where the tuning range is small. The latter requires that the product of the output reflection coefficient. times the gain in the forward direction be greater than 1.

and. the stabilizing cavity must be inserted in the output rather Therefore. which is a measure of the change in frequency produced by a change in the external load. Platinotron. for example. consequence of the lower pulling figure is that the Stabilitron is less subject to long-line effect [Eq.5 impedance or frequency pushing.260-1. to further improve its frequency stability. in the magnetron oscillator. The Amplitron. which has the connotation of broad. The efficiency of the Stabilitron is quite good.232 Introduction to Radar Systems [Sec.150 Operating potential. for a given degree of frequency stabilization. It might vary from 45 to 60 per cent across the tuning range. because of the long-line effect. 42 550 550 1.560 1. .350 1. a much higher can be obtained with the Stabilitron than with the magnetron since the stabilizing cavity placed at the input to the Stabilitron absorbs less power than a cavity circuit efficiency as a change in antenna than the input. is about 5 to 20 times less than normally Table 6. 6. less frequently. A Phase pushing of the Amplitron is manifested as frequency pushing—change in frequency with change in anode current— in the Stabilitron. depending upon the type of frequency stability considered. whether the frequency pulling figure or the frequency drift due to temperature change is being compared. at the output. to amplify. % Peak output power. causes trouble. depending upon the particular type of magnetron used for comparison. flat. kv Typical efficiency. Its effect is minimized.350 1. Therefore the Stabilitron can be operated into transmission lines several times longer than is possible with the magnetron before frequency jumping. is called a Stabilitron. An improvement in the dynamic pushing figure of from 10 to 50 is possible. (6. when used in the manner described above to generate oscillations. Its to the basic crossed-field structure used for both name is derived from the Greek word platys. Physically.220-1. Peak operating amp 50-100 46 28 40 36 52 650 1.5 QK 630-629 4-0 6 1-4 Pulling figure. 1)] than is the magnetron.3. in the Stabilitron. however.3. kw Average power output. Thus Platinotron is meant to apply to an amplifier with broadband properties. The broadband properties of the Amplitron are also reflected The latter may be tuned over a 5 to 10 per cent frequency band by changing the resonant frequency of the stabilizing cavity and the phase shift in the line connecting cavity and tube. A comparison of typical operating values of a radar magnetron oscillator and a Stabilitron is shown in Table 6. the Amplitron and the Platinotron cannot be distinguished from one another. A high-g cavity can also be used in conjunction with a magnetron oscillator as it is in the Stabilitron. This is the name given the Amplitron and the Stabilitron. The pulling figure of the Stabilitron. Comparison of Typical Operating Values of Radar Oscillators Magnetron Stabilitron Characteristic 5J26 2-2. The frequency stability of the Stabilitron is from 5 to 100 times as good as that of the magnetron. Mc kc/amp current. Mc Operating magnetic field. However. Pushing figure. watts Tuning range. by the presence of the stabilizing cavity.400 gauss associated with the magnetron.

It is the direct descendant of the DeForest Audion. The inductance is caused by the connections made to the t Much Merle V. successful development of microwave radar early in World War II. VHF UHF employed in many applications at the lower radar of the cavity magnetron led to the frequencies (VHF and UHF). and the plate electrodes. equally significant developments were being made in the postwar development of new tube types such as the klystron. Both tubes were of relatively low power compared with postwar grid-controlled tubes. (4) reduced average power-handling capability due to smaller-size structures. power at the lower communication frequencies. 6. The process by which the is modulated in a grid-controlled tube is called density modulation. The potential applied to the control grid of the tube acts as a gate.000 Mc. the newer electronic generators have already demonstrated their microwave performance capabilities. and the lower This limited the development of the early radars to the The Navy's first prototype radar. Meanwhile. Grid-controlled Radar Transmitters Tubest 233 early radars developed in this country and abroad during the 1930s used conpower. and the Amplitron. the traveling-wave During the late 1950s. The Army's long-range search radar. However. the VT-122. The variation of potential applied to the grid imparted to the current traveling to the plate. be kept in mind that they are pertinent in the higher-frequency performance of all CW classes of tubes. These tubes were also used in the Army's first fire-control radar.Sec. The capacitance in the grid-controlled tubes is primarily that of the grids. used a 100T Eitel-McCullough triode tube 72 Six tubes were operated in a ring circuit to operating at a frequency of 200 Mc. RCA . During the postwar years. and radar. Considerable improvement was made after the war in the development of grid-controlled tubes for operation at The grid-controlled tube has been When the discovery The upward frequency scaling of grid-controlled increasingly higher frequencies. but it should capability.6. Some of the factors limiting the highfrequency performance of power tubes as the frequency is increased are (1) increased circuit reactances. A detailed description of the operating principles of these tubes may be found 73 in any classical text on vacuum tubes. stream electron Grid-controlled tubes are capable of megawatts of Limitations at High Frequency. applications spurred by tubes was UHF-TV.6] 6. ventional grid-controlled tubes since there existed no other source of large The RF bands. such as that by Spangenberg. interest in lowerfrequency radars waned. Hoover of of the material in this section was made possible from information kindly supplied by Tube Division. the cathode. the grid-controlled tube was capable of more average power per "bottle" than any other tube type at frequencies below 1 . downward in frequency and were highly competitive with grid-controlled tubes in the 400 to 1. the SCR-270.000-Mc region of the spectrum. Each tube had a plate dissipation of 100 watts. Although the grid-controlled tube can theoretically be scaled upward into the microwave region of the spectrum. the SCR-268. the needs for higher averagepower radar equipments and better MTI performance were two of the factors which renewed interest in the lower UHF and VHF radar bands. to control the number of electrons is traveling to the plate. Respectable power outputs have been obtained at frequencies as high as S band. used a Westinghouse tube. The type of grid-controlled tube considered here is the conventional triode or tetrode configuration operated in a vacuum. for which the author wishes to express his appreciation. scatter communications. (3) transit-time effects. in particle accelerators. the XAF. (2) RF losses in dielectrics. achieve greater power. or valve. and (5) reduced peak power All these factors will be discussed for the grid-controlled tube. these newer devices were scaled tube. and operated at 1 10 Mc. In any tube there will always be unavoidable capacitance and inductance.

Maximum operating temperatures can usually be increased 50 to 75°C over an equivalent glass insulated tube. instead of glass.234 Introduction to Radar Systems [Sec. When dielectrics must be used in high-power-tube construction. the minimum size of electrodes is determined in large part by the power dissipation required of the structure. 75 Ceramic tubes are more reliable than those of glass and are also easier to adapt to automatic production methods.5) . amp/cm 2 Q (6. Radiation loss is proportional to the square of the frequency. the electrode spacing cannot be made too large without encountering increased transit-time effects.78 T=6. Another factor contributing to the degradation of output power as the frequency is increased is the RF loss.10 where d = J= electrode spacing. 76 77 At low frequencies the time taken by an electron in traveling from the cathode to the plate can be considered to be instantaneous since the transit time is short compared with the period of RF oscillation. The reactance of the input capacity (grid-cathode capacity in a grounded-cathode tube) also shunts the input and is inversely proportional to the frequency. Some are designed with the resonant cavities entirely within the vacuum envelope. if the frequency is sufficiently high. 6. and hence the power loss. Reactances may be minimized. An important source of RF loss is the heating of the dielectric materials used in the construction of the tube for insulating supports or for the envelope which encloses the vacuum.500 Mc. For this reason most modern tubes use low-loss ceramic frequency. This causes a decrease in the power. is proportional to the square root of the frequency. relatively It is claimed 74 that replacing the glass envelope of the 2C39A (a low power triode) with a high-alumina (A1 2 3 ) ceramic envelope results in an increase in power of 10 per cent at a frequency of 2. If the dielectric is in the RF field. However. it is important to use as low a loss dielectric as possible. a practice almost always employed in modern high-power PR tubes. Also. Losses may also occur by radiation of electromagnetic energy from the tube elements or lead-ins. Spangenberg 73 shows that the cathode lead inductance reflects back to the input as a shunt resistance whose value is inversely proportional to the square of the frequency.6 electrodes. However. the field can excite molecular movements which result in heating. cm current density. Small-size electrodes spaced far apart result in small interelectrode capacity. both in operation and during bake-out. However. The inductance of the leads can be minimized by using coaxial transmission lines or waveguide and by designing the resonant circuits the same as microwave cavities. it is not always possible to do so. The reactive components act to shunt the input to the tube and short-circuit the tube as the frequency is raised. For example. Tubes with ceramic envelopes are mechanically stronger than tubes with glass envelopes and can withstand higher temperatures.7xlO. The finite time required for an electron to transit from cathode to plate places a limit on the upper usable frequency. The skin-effect resistance. Dielectric-heating losses are directly proportional to the They may be minimized by placing insulators outside the RF field or at least in regions of weak fields. the transit time of an electron traveling from the cathode to the plate in a planar diode under d-c conditions with a space-charge-limited - current is 54 . The resistance of the conducting parts of the tube increases with increasing frequency because of the skin effect. Most high-power tubes use microwave-circuit techniques. Both the skin-effect loss and the radiation loss can be minimized by operating the tube inside resonant-cavity structures. the time taken by an electron to transit the interelectrode distance will be comparable with the RF period and the transit time can no longer be considered zero. but never entirely eliminated.

54 Thus the choice of the grid-plate plate in a triode. Therefore. it is often desirable to increase the current density to avoid making the electrode spacing too small. the greater the acceleration of the electron and the less the time taken in traversing the space. the temperature of the cathode will Cathode back heating can be partially compensated by adjusting the heating increase.05 cm and a cathode current -10 2 This may seem rather short. In the trade between cathode current density and electrode spacing as given by Eq. When the transit time becomes an appreciable fraction of the RF period. the greater will be the current the cathode and the grid is usually determined by mechanical design considerations and by the amount of heat from the close spacing requires good fabrication cathode that the grid can safely dissipate. The minimum spacing between A technique and careful mechanical design if a grid structure is to be maintained only fractions of a millimeter away from a cathode surface operating at high temperature. as long as the back heating is small. efficiency. (2) heat generated by the interception of energetic electrons by the grids. power output are reduced. on the one hand. It has been mentioned that the transit time can be reduced by the use of high voltage. High voltage leads to increased current and power because it is. effective . The higher the voltage.Sec. associated with the RF voltages impressed across the interelectrode capacitances.5). density emitted. not possible to increase the shunt impedance of the resonant circuit to any great extent without reducing the circuit efficiency and the bandwidth. If enough electrons are turned back. and (3) ohmic losses due to the displacement currents . the less will be the bandwidth. on the other. and in the phase between the plate current and the grid voltage. (6. 6. . the smaller the spacing between the grid and the dissipation capabilities. and and gain high between compromise represents a spacing life wide bandwidth.5 X 10 of 900 Mc. power applied to the filament. in general. density of 1 amp/cm the transit time is 2. In the above example the transit angle would be 7r/2 radians. When the transit time is relatively large. Some of the electrons will fail to pass the grid and will be turned back to the cathode. The transit frequency a cycle at one-quarter of a about represents but it time is sometimes measured by the transit angle. spacing and the maximum voltage which can be used will be limited by the electrode In addition. The thermionically emitted primary electrons can cause damage by being accelerated with Electrons can also be emitted from sufficient energy to bombard other electrodes. the larger the cathode current density. the densitymodulated electrons are debunched because the transit time of electrons that leave the cathode at one moment of the cycle will be different from those departing at another moment. which is the product of angular frequency and the time taken by an electron to traverse the interelectrode space. The high-power performance might also be limited by the emission of electrons from the overheated grids. from this point of view.6] Radar Transmitters 235 For a triode or a tetrode the distance d is the spacing between the cathode and the plane of the control grid. The transit time in the grid-cathode region may be minimized by making the gridcathode spacing as small as possible and operating with a high grid voltage. For a spacing of 0. the less will be the of the cathode. 54 However. the tube must be capable of dissipating the The heating of the control grid and the screen grid is primarily caused heat generated by (1) ambient heat radiated from the hot cathode. a shift occurs 79 The gain. Also. sec. low-power it is concluded that it should be easier to build high-power tubes rather than tubes at the high frequencies if the heat generated can be safely dissipated and if the 81 cathode-emission limits are not reached. efficiency. To obtain large average power output. Excessive heat might cause the electrodes to sag or melt. the higher the voltage. 80 The spacing between grid and plate need not be as small as the grid-cathode spacing The minimum since the electrons do not start from rest as when they leave the cathode.

which is supposed to be capable of delivering 1 peak UHF UHF CW Mw . Tubes which use the beam-power electron optical system provide high power. The large negative current sheets of the beam power tube have the same effect as the pentode's suppressor grid in reducing the secondaries. An example of the design of a beam power tube capable of 1 kw of power at a frequency of 1. those cathodes with good emission properties such as oxide-coated cathodes are more susceptible to damage by positive-ion bombardment than the more rugged.or air-cooled. finite time is required after the application of the voltage for a breakdown to occur. 807. 73 This effect is accomplished by aligning the windings of the control grid and the screen grid. However. high gain. and if heat dissipation is proportional to area. across the vacuum envelope insulation. the more likely there will be an arc-over. The 6L6. the average power varies inversely as the square of the frequency. Very high power tubes use water-cooled grids or other means of conduction cooling to dissipate the heat. The high concentration of negative charge caused by the dense current sheets between the screen grid and the plate suppresses the flow of secondary electrons from the plate to the grid which occurs in a The effect of the secondary electron flow in the normal tetrode to distort the plate-current characteristic curves.01. One of the biggest single-unit beam power tubes is the RCA development type A-2581. The size of the tube is proportional to the wavelength. secondary-emission electrons are not necessarily harmful to tube performance. 829-B. and the 4X250 are all examples of beam power tubes. is normal construction The purpose of adding a suppressor grid to a tetrode. assuming that the voltage gradient required for breakdown is independent of frequency. Voltage breakdown may occur between the electrodes. Therefore both the averagepower and the peak-power capabilities decrease with increasing frequency. resistance. and good over-all performance. making it a pentode. The beam power tube has proved to be quite popular for conventional applications. 82 It is capable of delivering 100 kw of peak power at a duty factor of 0. 80 The peak power that a tube can withstand before breaking down is usually greater in pulse operation than in operation. 80 The peak power of a tube under pulsed conditions is often limited by the finite electron emission available and/or by voltage breakdown. It should be borne in mind that the above is only approximate. Beam Power Tubes. Therefore the longer the pulse duration. The beam power tube is a tetrode designed so that the electrons move from cathode to anode in dense sheets. cathodes such as pure tungsten or thoriated tungsten. Because of its success at the lower frequencies the beam-power-tube configuration has also been applied in high-power applications.6 low temperature electrodes by the process of secondary emission. CW A drawn from a particular cathode depends upon the material and the amount of life desired from the cathode. In general. but less efficient. plate dissipation is seldom the chief limitation on power output. The variation of average and peak power may be a complicated function of frequency in a specific tube design. or in the external cially in the tetrode. 6. especially in short-pulse application.236 Introduction to Radar Systems [Sec.000 Mc is the RCA-7214. If the dimensions scale directly as the wavelength. low-feedback effects. The amount of current that can be circuitry. The surface area will also vary as the square of the frequency. Although it must be properly taken into account in the design of a tube. Secondary emission can distort the characteristics to the point where the tube has the effect of a negative tetrode. esperelatively The heating of the control-grid and the screen-grid electrodes can be reduced with lower-temperature oxide cathodes and with electron-optical systems which minimize the interception of electrons by the grids. The plates of high-power tubes must be specifically designed to dissipate the heat generated and are frequently water. is to suppress the secondaries. the peak power will vary as the square of the wavelength.

19. length.Control grid Plate ^^Screen grid two UHF beam-power tetrode units.020 as is the spacing between Fig. (b) longitudinal section of unit tetrode element for beam power tube. all within the same vacuum envelope. is the power At a fixed frequency the unit cannot be made larger than some maximum size without encountering difficulties due to the generation of higher-order modes in the RF circuitry or to increased transit-time effects.25) is capable of providing a useful peak power output of 2 with a power gain of at least 20 db at a frequency of 425 Mc.Sec.f The single-unit beam power tubes are characterized by small size and rugged construction. The tube construction illustrated in this figure is "inverted. the Screen block smaller the tube and the smaller that the tube can dissipate. the 6952 (Fig.246. single cathode. The larger the tube. 6.' 3 IRE Trans. while the cathodes are located on the periphery. The multiplicity of units operating in parallel permits the attainment of high power from a single "bottle" since the heat to be dissipated is spread over a relatively large area. This is a ceramic tube with an over-all length of 6. Nov. a commercially available tube designed to give a power output of about RCA 10 kw in UHF-TV service. a 10-^asec pulse width." since the plate is a cylindrical structure located in the center of the tube.01 duty factor. Mw t Data are based on RCA Exhibit No. The 6952 is designed primarily for short-pulse operation the 2041 for long-pulse operation.or small-size transmitter. Figure 6. and a diameter of 5. The tubes described above are of medium power. RCA-6806 . In short pulse service with a pulse duration of 1 3 /^sec and a duty factor of 0. in. They differ from the tubes used in UHF-TV in that the plate-cathode ceramic insulating bushing is larger in the pulse version in order to permit the application of higher pulsed plate voltage. the more heat it can dissipate and the greater the power output.5 in.2 in. In long-pulse operation. pulse versions of the aforementioned . 80 83 84 A ' - ib) longitudinal view showing the aligned grid wires characteristic of the figuration is beam-power conThe spacing between the control-grid wires and the fila- shown in Fig.24a is a cross-sectional sketch of (RF by-passed to filaments and cathode shell) Cathode ^X . and a plate is limited in power capability by the amount of heat which a single unit can dissipate. and a 0. 6. 1957. They might be used in moderatecapability radars or in high-power array radars where the antenna is made up of a large number of individual radiating elements each fed by its own moderate. 6. One technique for increasing the power output is to employ in parallel a number of unit electron-optical structures arranged in a coaxial. . cylindrical configuration.004. (a) Cross-sectional sketch of two unit tetrode elements for beam power tube. PTO 920-4. 6.. Forty of these unit tetrodes are used in a cylindrical arrangement in the 6806.24. It has already been mentioned that the size of the tube structure is proportional to the waveMultiple-unit Tube Construction. The type of tube construction consisting of a one or two grids. tetrode.6] Radar Transmitters 237 power at a frequency of 500 Mc when operating as a pulse amplifier with a gain of 10 db. ments is 0.) UHF UHF the control-grid and the screen-grid wires. so that the higher the frequency. Similar construction is employed in the RCA-2041 and RCA-6952 tetrodes. {From Bennett.

. The output circuit of a double-ended tube is shown in Fig. The maximum voltage in the double-ended configuration appears in the active portion of the tube.06 duty factor). It is claimed 80 that this tube should perform creditably at frequencies up to at least 900 Mc since its continuous-power progenitor has been tested at these frequencies.238 Introduction to Radar Systems [Sec. 6.25. Similar arrangements must be provided for the grid-cathode input circuit. 6. Only the output circuits are shown in Fig. .6 with a pulse duration of 2 msec and a prf of 300 cps (0. it is possible to design the double-ended tube with a larger diameter before moding problems enter because of the elimination of the socalled "deadhead" space which exists in the single-ended tube between the upper portion of the active region and the position of the voltage maximum. It can be seen that the active length of C-D in the double-ended tube can be twice that of the single-ended configuration therefore at least twice the power output can be obtained as compared with the single-ended tube. The RCA 6952 mentioned above is an example of a single-ended tube. In essence. The single-ended tube can be considered as a coaxial transmission line.26a. RCA type 6952 tetrode. 6. the RCA-2041 is capable of delivering a peak power of 180 kw.26b. It is desirable to operate with the maximum voltage in the active portion in order to achieve maximum effectiveness. an average power of 1 1 kw.26a and b. envelope. The loci of maximum current lie in a horizontal of maximum voltage and. the single-ended circuit in a coaxial configuration is a figure of revolution whose electrical length is a quarter wave. 6. radial cavity external to the vacuum revolution between the points plane /. consisting of a cylindrical tube with a transition to a short. In addition. EV. with a power gain of 20 db and an efficiency in excess of 50 per cent at 450 Mc. Since the points a quarter wavelength. The maximum voltage exists at the center of £and V. whereas it appears outside the active region in the single-ended configuration. maximum current are separated by Fig. 80 Single-ended and Double-ended Configuration. 6. 60 '** A longitudinal cross-sectional view of the output circuit of a triode operating in a classical single-ended circuit arrangement is shown in Fig. The unit-tube principle described above for the beam-power tetrode has been applied to other electron-optical geometries such as the tetrode and triode. a double-ended tube is two single-ended tubes butted together at their high voltage ends EV.

6-in. In long-pulse operation of peak (2. it is well suited for radars whose targets are extraterrestrial.usec pulses at a duty factor of 0.26.275 in. the A2346 can develop a peak power of 10 Mw.6] Radar Transmitters 239 Plate cylinder- Vacuum envelope Plate cylinder Grid cylinder Vacuum envelope Fig.27. some of the highest powers obtainable with electronic tubes at power at a frequency of 500 Mc The A2346 is capable of generating 500 kw of when operated as a cathode-driven (grounded-grid) amplifier. sketch of the unit triodes employed in this tube is shown in Fig. so that less are common to both. Both operate by grid control of space current. 6.Sec. -diameter cylinder and has an active electronic length of 4 in.015 in. One of the largest developmental type A2346 triode shown in the photograph radar application is the of Fig.) power generators suitable for Superpower UHF Triode. Tetrode. yet it is capable of producing frequencies. Triode vs. Although a 2-msec pulse might be too long for many radar applications. This double-ended power tube employs 96 unit triodes arranged on a cross-sectional 6. (a) Longitudinal cross-sectional sketch of the output circuit for a triode tube arranged in "single-ended" circuitry. The grid-plate spacing is about 0. M 7S S5 Both the triode and the tetrode have been used as the The choice between the two types is a basis for high-power-tube design at UHF.0-msec pulses at a prf of 30 cps).01). as described in Chap.28. 6. and many design features difficult one. UHF RCA A UHF UHF CW Mw ' > . (From Hoover* 1 Proc.06 duty factor at frequencies below 500 Mc. The power gain is about 1 3 to 1 5 db with a conversion efficiency of 50 per cent. Direct-current plate-voltage blocking capacitors and power-output coupling circuitry not shown. 6. The tetrode has slightly higher gain than the triode. 6. and the grid wire to filament-strand spacing is 0. The grid wires are wound at a pitch of 72 turns per inch. 14. (b) Longitudinal cross-sectional sketch of the output circuit for a triode tube arranged in "double-ended" circuitry. the tube is capable of developing 5 power with a 0. In short-pulse application (10 . The electronics of this unit triode is similar to that of the original DeForest Audion. IEE.

the tetrode is the preferable configuration for moderate-to-large powers while the triode is to be preferred at the highest power levels. Although it may be dangerous to generalize. is The additional grid of the tetrode gives effects. 6.240 driving Introduction to Radar Systems [Sec.6 power is required for a given output. . The bandwidth is also greater because of the greater spacing. 6. it seems that if a choice must be made between the two.27. greater isolation between input and output and reduces internal feedback The lower because of the increased screen-to-plate spacing. (From Hoover. 6. IEE."* Proc. On the other hand. the additional grid of the tetrode requires a more complex construction than the triode and output capacity of the tetrode Fig. Cross-sectional sketch of unit triodes for UHF power tubes. There is also the possibility that the tetrode performance will deteriorate because of unforeseen parasitic oscillations there which might be generated in the cavity between the two grids. Liquid cooling applied here Copper plate Grid wires Wire-support fin i Thoriated-tungsten filamentary cathode Grid liquid cooling applied here Fig.) is some loss of space current to the additional grid.28. RCA developmental type A2346 superpower UHF triode.

The tube delivered more than 50 kw of output power with a 60 to 70 per cent plate As an amplifier efficiency and could be tuned over a frequency range of 350 to 650 Mc. cavity of the Harvard tube was placed between the control grid and the filament structure.6] Radar Transmitters 1 241 The resnatron is a particular form of tetrode characterized by ( ) the vacuum envelope. nor is it desirable. (2) the control grid and the screen grid operated at RF ground (grounded-grid tetrode). the grounded-cathode amplifier requires that the screen grid be bypassed to the cathode by some means. particularly of When used as an amplifier instead of an oscillator. and (3) the screen grid operated at the same d-c potential as the anode. the d-c voltage-supply leads do not become a Therefore the inductance of the supply leads has a part of the resonant circuits. The resnatron has been operated in the past as a grounded-grid (cathode-driven) amplifier. The . relatively poor. The tubes could not be sealed off since they were too big to be made to hold a vacuum with the techniques available at that time. although a grounded-cathode (grid-driven) tube is capable of higher gain. or Another benefit of placing the RF circuitry within the vacuum is that dielectrics such as glass or ceramics are external to the fields which could cause dielectric heating losses. for the transit time to play the same role that it played in the Sloan and application. it is not necessary. 6. while the output cavity was between the screen grid and the anode. In the laboratory it was also operated as a power amplifier. During World War II. RF circuitry located completely within the seen but limited application in the past. that is. and the RF isolation chokes all within the vacuum system of the tube. the power gain was 10 db. Salisbury and associates at the Harvard Radio Research Laboratory developed a high-power CW resnatron for jamming 87 88 The resnatron was operated as a self-excited oscillator. However. It was used operationally by troops in the field as a jammer with a noise-modulated bandwidth of 4 Mc. this tube represented a considerable increase in power capability over any other tube which operated at frequencies as high as UHF. they were continuously evacuated in order to maintain the vacuum. the bypass capacitors. Even by present standards the 50 kw delivered by this resnatron is quite a lot of power. The resnatron was originated about 1938 by Sloan and Marshall86 at the University A of California. screen grid at high d-c potential is that the effects of electron transit time are reduced. Operating the resnatron with the screen grid and plate at the same d-c potential eliminates the need for a d-c blocking capacitor between the screen grid and plate. high frequencies easier than with the conventional grid-controlled structure. The resnatron tubes were operated "on the pumps". limiting the life Even with continuous pumping the vacuum of the pure tungsten cathodes. it is of interest not only for historical reasons. the class C type. 81 ' ' Marshall oscillator. There is no reason to believe that sealed-off resnatron tubes could not be built if suffioscillator. circuit are separated from one another just as they are in the magnetron. but also because it is a potentially useful device for obtaining high power. Their original objective was to develop a high-power oscillator in which the phase delay caused by the transit time was compensated by introducing the proper phase shift in the feedback circuit. as further advantage of operating the required in the classical operation of tetrodes. The amplifier efficiency was the same as the self-excited The input Plate voltages of the order of 15 kv were employed for maximum power In its day. cient inside the tube development was effort were applied. These characteristics make the generation of high power at traveling. output. The RF and the d-c portions of the negligible effect on the operation of the tube. and in the high-power resnatron this presents tedious design problems. The resnatron was the first high-power tube capable of delivering tens of kilowatts of average power at UHF. klystron.Sec.wave tube. Although it has The Resnatron. They achieved an average power of 8 kw at a frequency of 860 Mc. By placing the resonant circuits.

All RF cavities. The pulse width is 300 /^sec at a prf of 60 cps (duty cycle of 0. where they are collected. number of experimental microwave resnatrons have been built at a frequency of 3.242 Introduction to Radar Systems [Sec. Being negative. As with other high-power resnatrons. 89-91 The tube described by Sheppard et al.5 tons and stood 15 ft. tuners. An interesting application of the resnatron and one analogous to the type of operation which might be encountered in radar is its use as a high-power pulse amplifier for the University of Minnesota linear proton accelerator. be made to deliver 1 of power. and plate current was 8 1 amp. each operating into high-g accelerator tank circuits. The performance obtained with the axial-flow resnatron described by McCreary et 89 al. The plate efficiency varied from 45 to 75 per cent. D. Sloan and his associates were able to achieve a peak power of I at S band with a 100-^sec pulse width. The plate voltage required for this power output was 70 kv.6 kw was obtained at a frequency of 560 Mc. Sloan claimed that the basic structure could. Four such resnatrons were used in the linear proton accelerator. In UHF A Mw Mw . depending upon the operating point of the tube. The advantage claimed for the reflex resnatron is that wideband modulation may be obtained with low modulation power by swinging the repeller voltage. and had to be periodically dismantled to have their cathodes replaced. 80 who quotes a personal communication from Dr. stream is directed parallel to was designed with a radial electron to the axis of symmetry of the Tubes operating with axial flow in which the electron the axis of symmetry have also been built. Three were power amplifiers. and the tubes had to operate "on the pumps. an efficiency of 38 per cent and a bandwidth of 8 Mc. A power output of 29 kw was obtained at a frequency of 420 Mc with power gains in excess of 10 db and an estimated bandwidth of 4 Mc." Exclusive of the 1.5 and an average power of 63 kw. The efficiency was better than 50 per cent. 6.000 Mc. that they had to be continuously pumped. A power gain of 10 db was achieved with an efficiency of 62 per cent. water-cooling coils. of emitting length per strand. The cathode consisted of 36 strands of pure tungsten with approximately 5 in. RRL resnatron' described above the electron beam was perpendicular cylindrical tube structure. they were operated successfully in truck-borne units during wartime conditions by army troops with no special educational background. This particular resnatron CW Mw utilized a radial-flow geometry. The tungsten filaments in the tube had to be replaced on the average of every 900 hr. with a power gain of 5. The reflex resnatron is similar to the conventional resnatron except that the anode is replaced by an electrode of sufficiently negative potential which repels the electrons and bends them back toward the screen. that is. the repeller collects little or no current. In one experimental device a power output of 2. In spite of the fact that these tubes were not small. Sloan of the University of California.400-cfm oil-diffusion vacuum pump. that they The Harvard flow. The tubes could be readily dismantled for filament replacement or other repairs. The resnatrons which have been described all operated in the portions of the frequency band.6 tubes were designed to be readily taken apart for replacing the cathodes. The advantage claimed for the axial-flow resnatron is its simpler construction as compared with that of a radial-flow tube. was similar to the performance achieved with the World War II radial-flow resnatron of Salisbury. Thus the tube actually consisted of 36 unit tetrodes operating in parallel. with minor modifications. and a 100-kv isolating choke were located within the vacuum enclosure. Hence the modulation power may be small. while the fourth acted as a driver for the other three. the tube weighed 2. These are mentioned by Hoover. 90 used a reflex principle of operation in combination with the axial geometry of the electron flow.018). the tube construction precluded any real bake-out and outgassing. 92 These tubes operate at a frequency of 202 Mc and deliver a peak power of 3. H.

When a fault occurs in the protected power tube. but it is believed that it can be triggered by sources ranging from cosmic rays to line-voltage transients. Simplified diagram of electronic-crowbar fault-protection circuit. Tubes may be protected from the damaging effects of arc-discharge currents by One such protection device is called an diverting the damaging current from the tube.asec The high-speed protection of this device permits safe. issue of Electronics. 80 - 93 little Tube Protection. sion. The low impedance of the thyratron when conducting results in the damaging current being shunted away from the power tube and through the thyratron crowbar tube. unprotected tube.29. It another design. except that the particular cathode used in this tube warped. 1956. here shown as a thyratron. .6] Radar Transmitters 243 was obtained with a 2-fj. parasitic oscillations. a peak power of 1. after the detection of the fault. from the January. full-power operation of the power tube almost immediately after the arc is quenched. Long Island. 6. The electronic crowbar is capable of providing fault protection within 1 to 5 . material whiskers. voltage across the crowbar tube to ensure its conduction despite severe low-impedance In a typical large power-tube installation the value of flash arcs in the protected tube. It was claimed that power gains of 10 to 30 db are power can reach hundreds of obtainable and that the average pulse or the maximum C Mw W kilowatts. The principle of operation of the electronic crowbar is illustrated by the diagram that placed of Fig. which in turn opens the circuit breaker and small series resistor R provides adequate deenergizes the primary source of power. It places a virtual short circuit across the rectifier output similar to on the rectifier by the flash arc. This impulse causes the thyratron to conduct.29. 9 * It is possible for power tubes to develop internal flash arcs with warning even though they are apparently of good design and operated in a conThis type of unexpected arc discharge is known as the Rocky Point servative manner.) does not seem to be well understood. AC Circuit -^-nm^ Rectifier breaker Electronic crowbar gas tube- Overlood-^ML^ relay _J_ (Reprinted. the sudden increase in current through the cathode resistor R k produces a positive voltage pulse which is coupled by C e to the grid of the electronic-crowbar gas tube. and photoelectrons. copyright. 6.sec pulse width. 1956. effect. a McGraw-Hill publication. electronic crowbar. The short-circuit current is transferred to a gas-discharge tube such as a hydrogen thyratron or an ignitron which is not damaged by the momentary short-circuit conditions. Its name is derived from experiences with power tubes for communications When a flash arc occurs in an transmitters at Rocky Point. New York. spurious renegade primary and secondary electrons.5 was thought that this tube should have delivered 5 Mw.Sec. A the series dropping resistor is only about 5 ohms. 6. by permisFig. the rectifier and filter-capacitor bank discharge large currents through The mechanism of the Rocky Point effect the arc and the tube can be easily damaged. The surge of current through the crow- bar tube actuates the overload relay.

and gridThe physical size of a transmitter controlled tubes are examples of power amplifiers. poor subclutter visibility The good stability of the amplifier plus the fact that the transmitted waveform results. The magnetron oscillator frequency can also be stabilized with an external cavity. The conclusions presented should therefore be considered as subject to revision as new developcharacteristics of the various radar will power tubes ments are reported. as with any component technology which is continuing to expand and grow. relatively modest. This is of Oscillators vs. 6. Since amplifier MTI radars are coherent from pulse to pulse. In addition to the power tube itself. Good frequency stability is of importance for MTI radar. Amplifiers. The frequency stability of a high-power self-excited oscillator is not as good as that of an amplifier driven by a stable. values of 8 to 10 db being typical. using a power oscillator is usually smaller than that of a power amplifier. is only valid as of the time of writing. but the stability is usually less than that obtained with the master-oscillator power amplifier. Tubes for radar transmitters are continually being improved. This is not true with oscillator MTI radars since they are coherent only over the duration of one pulse-repetition period. As mentioned previously. crystal-controlled low-power oscillator. Other advantages of the amplifier over the oscillator are that the amplifier is less .244 6. The magnetron and the Stabilitron are self-excited power oscillators. wth-time-around echoes from fixed targets are eliminated. one of the applications of a low-gain. It is not meant to convey the impression that some tubes should always be used in radar to the exclusion of others. The characteristics of each tube are sufficiently different so that each has its own area of application for which it is preferred. If the frequency of oscillation wanders excessively during the interpulse period. is a relatively low gain tube. The Amplitron. and driver amplifier stages are needed to amplify the power to the level necessary to drive the final The driver stages of the high-gain klystron or traveling-wave tube may be tube. is generated at low power level means that it is easier to achieve the sophisticated modulations required for pulse compression with amplifiers than with oscillators. A high-g cavity can be used to improve the self-excited oscillator frequency stability. Introduction to Radar Systems Comparison of Tubes one single tube is best suited for all radar applications. a frequency multiplier chain. but it would have to be placed at the output with a corresponding reduction in over-all efficiency. a stable. Before proceeding it might be worthwhile to inject a word of caution concerning the type of comparison presented here. may be classified as either self-excited advantage for radar applications in which mobility is required. high-efficiency tube like the Amplitron is as a booster to increase the range of lower-power radar sets. it also means that the maximum power output available from a small-size tube is less than from one of larger size. especially in those tubes with gains of the order of 50 to 60 db.7 No In this section the be compared and those factors which influence the selection of one tube instead of another will be discussed. on the other hand. and new principles of RF power generation will no doubt be discovered in the future. high-power. In the Stabilitron the stabilizing cavity is placed at the input and permits more efficient operation than if it were in the output. The various tubes considered for RF power generation power oscillators or as power amplifiers driven by stable low-power oscillators. The discussion of tube technology presented here. [Sec. but in general. traveling-wave tube. while the klystron. crystal-controlled oscillator. Amplitron. The low gain means that two or three high-power Amplitrons might have to be operated in cascade to achieve a reasonable over-all gain. The superior frequency stability and higher power output of the power amplifier are accompanied by a larger and heavier transmitter.7.

if size. the traveling-wave tube does not seem to be able to achieve as large a power output in practice as some of the other more narrowband tubes. as they are in many applications.wave tube should be significantly less than that of a klystron. all the amplifier tubes which were considered are capable of generating relatively large average power. The power variation with frequency is not always a simple one. 2. weight. The grid-controlled tube.7] affected Radar Transmitters 245 Power by imperfections in the modulator and it is not subject to long-line effect. high efficiency is one of the important attributes of a good power tube. more complete measure of transmitter efficiency from the operational point of view would be the ratio of the RF power output to the total power input. 3. the number of tubes combined is given by 2" (or 2. etc). The electronic efficiency is equal to the RF power delivered by the electron beam to the circuit. the characteristics of various power amplifiers will be compared. The efficiency usually quoted in this chapter is the RF conversion efficiency. Comparison of Power Amplifiers. it seems to 1 to be vary inversely as the square of the frequency. divided by the average power supplied to the electron beam. the klystron. for example. the power amplifier is probably to be preferred over the power oscillator for most radar applications in which high power and/or good MTI performance is amplifiers desired. but in general. were compared in the previous section. The power output of any particular tube type^vvill be less at the higher frequencies than at the lower frequencies. defined as the ratio of the RF power output available from the tube to the d-c power input of the electron stream. the remainder is wasted as heat because of copper losses. 16. For comparison. They are not often used above L band. The power input would include all A A . All tubes seem to suffer the same peak-power limitations imposed by voltage breakdown in waveguides or cavities. that is." The hybrid junction can be used to combine tubes in pairs. The greater the efficiency. This conclusion could very well be changed in the future since there is no fundamental reason why the power output of a traveling. and complexity are important consideraa magnetron oscillator would be preferred even at the sacrifice of radar performance. the magnetron and the Stabilitron. may also be combined to deliver more power than is possible with a 95-97 single "bottle. The high efficiency of the Amplitron is one reason why it is capable of much larger power than. The more efficient tube requires less prime power for a specified power output. In the remainder of this section. A frequency range. The peak-power limitation due to voltage breakdown also varies inversely with the square of the frequency. However. tions. The grid-controlled tube is capable of exceptionally high power output at UHF or lower. The tubes which will be considered include the grid-controlled amplifier. efficiency. a good average power for a magnetron oscillator is a few kilowatts. comparison of power oscillators will not be given here since the two oscillators most suited for radar application. although resnatrons have been operated experimentally at Sband with respectable power output. The circuit efficiency is the fraction of RF power going into the resonant system which appears as output power. In general. In principle.. the greater the power output for a structure of a given size and the easier it will be to dissipate the heat generated by the losses. the klystron. 8. and the Amplitron. The conversion efficiency is the product of the electronic efficiency times the circuit efficiency. and the operating costs are less. there seems no reason why the amplifier tubes discussed in this chapter could not be designed to operate anywhere within the normal radar frequency range from UHF to K band. 6. 4. the travelingwave tube. This is over-all efficiency. the relatively low efficiency traveling-wave tube. power. without a corresponding sacrifice in the bandwidth. With the exception of the grid-controlled tube. and the Amplitron are all capable of generating tens or even hundreds of kilowatts of average power. On the other hand. Sec.

with cent or more at high power levels. and the particular operating voltages and currents. 4. The klystron has a lower efficiency. but less than that of the multicavity synchronously tuned klystron. converted into kinetic energy of motion before conversion to RF energy can occur. The size and weight of a particular tube are probably not as important to the radar engineer as is the total weight of the transmitter.246 Introduction to Radar Systems [Sec. there is direct conversion of potential energy to RF energy in the crossed-field device. the smaller the power required of the source which drives the amplifier. the cooling requirements.wave tubes is of the order of 10 per This is followed closely by the Amplitron. 6. efficiency for a typical efficiency In the collinear beam device no additional kinetic energy is supplied to the beam after it The d-c energy from the power supply must all be enters the interaction space. The driver of low-gain tubes often has to be another power tube of the same capabilities as the output power tube. gain. tube is better than that of the grid-controlled tube. high-power klystron might be 35 to 45 per cent. Bandwidths from 3 to 5 per cent are possible with the klystron and might be as high as 12 per cent or more. the type of modulator. but is heavier than the grid-controlled tube. synchronously tuned klystrons. On the other hand. The multicavity klystron can also be made to have a respectable bandwidth by stagger tuning the various cavities and trading gain for bandwidth. The bandwidth of the grid-controlled tube is the Although smallest of all. The tube with the highest gain is the klystron. Only a modest driver is required for the klystron. The total weight depends on the gain and efficiency of the tube. the biggest of present radar power tubes. Typical magnetron oscillator efficiencies vary from 35 to 60 per cent. with typical values of from 8 to 10 db at high power levels. large bandwidths may be possible with broadband cavities within the vacuum envelope. and the power required for electromagnets if used. The electrons perform work on the RF field as they sacrifice their energy of position and drift to the collecting electrode (anode). but the over-all might be 25 per cent. bandwidths of about 7 to 10 per cent. Extremely high anode voltages as in the klystron and traveling-wave tube require lead shielding to attenuate harmful X-ray . The bandwidth of practical traveling. It is determined by the loaded Q of the output cavity. Gains of the order of 60 db or more are not uncommon with four-cavity. The grid-controlled tube with values from 10 The gain of the traveling-wave to 25 db is slightly better than that of the Amplitron. but the driver of the Amplitron. which is the lightest of all. bandwidth. crossed-field devices such as the magnetron or the Amplitron have higher efficiencies than collinear beam devices such as the klystron or the traveling-wave tube. bandwidth.7 power needed for the operation of the tube such as heater power. which is in between the efficiencies of klystrons and grid-controlled tubes. bandwidths of the order of 1 to 2 per cent are common. In general. The higher the gain of the power tube. The traveling-wave amplifier is theoretically capable of large 5. power for cooling The RF conversion devices. but the traveling-wave tube has the lowest of all. The Amplitron has the lowest gain of the amplifier tubes discussed here. might represent a substantial fraction of the total transmitter. for example. The klystron and the traveling-wave tube are the heaviest and 6. size and weight. The Amplitron with permanent magnet is lighter than the klystron or the traveling-wave tube. The efficiency of the grid-controlled tube is slightly less. The Amplitron seems to be capable of higher efficiencies (70 to 90 per cent) than any of the other amplifiers discussed. A broad-bandwidth transmitter is important in radar applications in which accurate range measurement or good range resolution is necessary or where it is required to change frequency rapidly over a wide frequency band. The need for an electromagnet and lead shielding greatly adds to the weight.

C o « "a c i? x H.Si.-* V"l *o O o ""> ""1 ""> >n >2 ' < < < 3 p.30 8l.2 . Si H B O J< 00 3 X E C T3 O c S £>< 5> 5 rt 1 00 z o -t vo o m t-i in Tt- o\ H r^ u-> ©o (^i 3 < o a! Iflffl ca < S< IS o PS £ 5 a> o o 00 . ^M oo oN « ~Hi/->~Hinr-»oooc> I! a! 5" §.-* ". p • * 00 •5 rt 5 £ I "|JSS 5 "•*ynS.3" C 3 2P -o .ro h C O CM <n o C c 3 1° C3 J? »/~l E*0! ^ © O .E ." „ .to &0 a •3 .8> j.3 g > £ u -a w — ca U S I 1 -H « OB. T3 13 o <a c e c U C § B 6." 3c S ^ . B. -~ <2' .o -o o a a £ * <u u •- OjO E O -c rt: JJ <u <o U-4B5 247 . o * nOTf « o t- -i 2 x 3 t- ra in \> .

especially Magnetrons such as the E. all operate at S band. Low-power grid modulators can be used with grid-controlled tubes such as the triode and the tetrode. must modulator be designed to the that so plate-modulated.8.000 to 10. is of the order of 1.4 are listed typical operating characteristics of 8.E.000 hr seems to be typical. high-power magnetron oscillators at the higher frequencies. have lives of the order of 10.760 hr in a year. life. 6. Modulators 98 " 101 The modulator is the device which turns the transmitting tube on and off in such a manner as to generate the desired waveform. Although plate modulation requires considerably higher modulation power than grid modulation. whereas the other klystron entry. Otherwise plate modulation is necessary.000 hr.V. No general statement can be made regarding the over-all size and complexity of radar power 'amplifiers because of the many factors involved. On the other hand. The number of electrons which manage to leak through the modulating anode of klystron amplifiers when it is cut off is claimed to be negligible for most radar applicaIn one type of klystron amplifier the interpulse noise due to electrons leaking past the modulating anode is stated to be at least 170 db below the pulse power level as measured over a 25-kc bandwidth.000 hr. full beam power of the klystron and the traveling. equipment and cables. large low-loss conductors or cable must be used for transmitting these high currents.) The life of a grid-controlled tube is equally good. if the tubes are designed with modulating anodes and provided the interpulse noise due to electrons leaking through the modulating anode is small compared with receiver noise. 102 . (There are 8. The magnetron particular characteristics which determine the type of modulator. This is a medium-power tube. the modulator is sometimes called apulser. oscillator. In addition. table of comparison. If. In Table 6. plate modulation must be employed to ensure that interpulse noise does not degrade the sensitivity of the receiver. is focused with an external electromagnetic focusing coil. With the exception of the grid-controlled power-amplifier representative The power output of these amplifiers is roughly comtetrode. as in grid-controlled tubes. which may be used for It applications or as a driver tube for higher-power amplifiers. included in this table since it is an example of a space-charge-focused klystron. is handle the full pulse power.248 Introduction to Radar Systems [Sec. Long life is desirable in order to minimize equipment downtime for repair. In those applications where the number of electrons which escape the cutoff action of the grid are large enough to induce a significant shot noise in the plate circuit. tubes. The Amplitron life is claimed to be greater than 1. particular tube will also depend upon how close to rated maximum power output it is Company) type operated. Screen-grid pulsing may also be employed in a tetrode. however. while the life of transmitters employing 7. large heater currents are necessary. it completely eliminates interpulse noise and improves the operational stability as regards missing pulses. For this reason grid-controlled tubes are sometimes plate-modulated. tions. for instance. the excessive replacement of tubes adds to the operating cost. When the transmitted waveform is a Each RF power tube has its own pulse. (English Electric Valve The life of any 7182. medium-power radar is 6. The Amplitron and the Stabilitron are plate-modulated similar to the magnetron. parable except for the SAS-61.000 hr. or perhaps less. the VA-820. The life of the klystron and the traveling-wave tube can be relatively high when compared with other high-power tubes.8 High voltages also require good insulation of the radiation generated by the tubes.wave tube can be switched by a modulator handling a small fraction of the total beam power. A life of 5.

an energy-storage element. Only a small well as initiating it. simple capacitor has the The simplest electrostatic storage element is a capacitor. (2) the pulsactor. 6. These are ( ) the line-type modulator with a gas-tube switch and a delay-line storage element. and (3) the hard-tube modulator incorporating a vacuum-tube switch and capacitor storage - I . Basic elements of one type of radar pulse modulator. Once the thyratron is triggered by its grid. The modulator containing a gas-tube switch and a delay line as the energy-storage element is called a The It is commonly employed with high-power magnetrons.Sec. 6. generate a pulse of width t. it cannot be shut off until the For this reason a capacitor is not a storage element completely discharges itself. upon discharge. in a relatively short time to form the pulse.30. and a switch. The modulator consists of a charging impedance. It is accumulated in the The charging energy-storage element at a slow rate during the interpulse period. the charging impedance prevents energy from the storage impedance device from being dissipated in the source.30. During the discharge cycle. 98 using an electrostatic storage element. portion of the stored energy in the capacitor is expended during the switching. which might be a magnetron. 6. the time constant of the discharge circuit must be large compared with the desired pulse duration and the switch must be capable of interrupting the discharge as vacuum tube can be used for this purpose. At limits the rate at which energy can be delivered to the storage element. Energy storage Charging impedance element I i Energy source I Switch\ I Load ! I ^. To obtain a relatively flat pulse from the discharge of a capacitor. The former is more often used in practice since it is The configuration of Fig.30 is that of a voltage-fed modulator easier to implement. which uses saturable reactances for both switching and storage purposes. which can be turned off or on with essentially equal facility. that the energy discharges exponentially with time and produces a poor pulse shape. 6. Charging path zT x ^-Discharge path Fig.8] Radar Transmitters 249 The basic elements of one type of radar modulator are shown in Fig. line-type modulator. gas tube such as the thyratron or the ignitron is capable of handling highpowerand However. the proper time. satisfactory storage element to use with a gas-tube switch since the discharge pulse A A A cannot be made rectangular. is unlike the operation of the vacuum tube. The energy-storage element might be either electrostatic (basically a capacitance) or electromagnetic (inductance). however. the switch is closed and the stored energy is quickly discharged through the load. The energy for the pulse is supplied from an external source. An open-circuited delay line of length t/2 will. However. 98 103 In the remainder of this section the three basic types of radar modulators which may be used to pulse the magnetron oscillator will be discussed. a gas tube cannot be turned presents a low impedance when conducting. This off once it has been turned on unless the plate current is reduced to a small value. a delay-line storage element can produce a rectangular pulse and is satisfactory for use with a gas-tube switch. disadvantage. delay line of the line-type modulator is called the pulse-forming network and is abbreviated PFN.

which might be of the order of 500 Fig. the switching tube. isolation element during the discharge cycle and prevents the pulse-forming network 6. Line-type Modulator. is called d-c resonant charging. Equivalent circuit of the A perfect match is not possible in to 1. i — ?. If a (Leh/C)value / suddenly applied to the input. 6. since the inductance of the line is The load is represented by a negligible compared with the charging inductance.32. .32. which is usually 50 ohms impedance. 6. The charging inductance L C h and the delay-line capacitance C form a resonant circuit. linear. The transformer would then be used to match the cable impedance to the impedance of I. 104 The charging impedance. since half the charging energy would be dissipated in the resistance. It is usually a lumped-constant its consists of an air-core inductance with taps along length to which are the pulse width fit attached capacitance to ground. Diagram of a line-type pulse modulator. oscillations will occur provided For small values of R. 6. the frequency of oscillation will approach the 1 The peak voltage across the delay-line capacitance C (L C hC)~ i Thus the pulse will be twice the supply voltage after the first half cycle of oscillation. The equivalent modulator circuit during the charging cycle is shown in Fig. .F pulse will also be discussed..8 The effect of modulator pulse shape on the magnetron R. A pure inductance absorbs no energy and is preferred in high-power applications. except that the maximum efficiency would then be 50 per cent. The delay line is represented by its capacitance only.31. diagram of a line-type pulse modulator is shown in Fig.000 ohms. 6. and the power supply. The number of taps depends upon level and the teristics fidelity required. These functions of the charging impedance could just as well have been obtained with a resistance. Introduction to Radar Systems [Sec. resistance R. ignoring the effect of the charging diode. nftowip- Ch-tfn. shown as an inductance. d-c voltage is > R/2. Some degree of flexibility is permissible in the selection of the delay-line impedance since a transformer may be used to match the delay line to that of the load. This method of operation. modulator of Fig. TB 77(L C hC) i repetition frequency/. 6. limits the rate at which It also acts as an current is drawn from the energy source during the charging cycle. It is sometimes convenient to design the delay line for an impedance of 50 ohms so as to make it unnecessary to match the delay line to *\^WiTOWJo^ the transmission cable. or pulse-forming network. diode Pulse-forming network -M. the magnetron. A from discharging into the energy source instead of into the useful load.31. Bypass diode c TTTT T Magnetron Energy source „ Trigqer_^_ Hydrogen thyratron Damping network Fig.250 element.»£. The effect of the charging diode may be ignored for the moment. delay line.31 during the magnetron impedance is nonall since the cases charging cycle. The impedance of the line is chosen to the charac- of the load. will be twice the resonant frequency/ or l// r where Tv is the pulse-repetition period. = {2ir)~ . The energy-storage element. = = .

remains on the delay line. a charge might be placed on the delay-line capacitance with polarity opposite to that normally placed on the capacitance during the charge cycle. Most of the switching devices for Gas-discharge devices line-type modulators are based on gas-discharge phenomena. fixed spark gap. range of operating frequency. If the charge were allowed to remain.31) permits the modulator to be readily operated at any pulse repetition frequency. semitheoretical parameter which has been used to evaluate the 101 It is defined as twice the product of the peak capability of thyratrons is the P b factor. The charging. the peak voltage on the network would increase with each cycle and build up to an abnormally high value. pulse The function of the damping network is to help reduce the trailing edge of the voltage and prevent postpulse oscillations which could introduce noise or false targets. hydrogen-filled thyratron over an inert-gas. The inverse charge may be dissipated by connecting a bypass diode and a series inductance LB in parallel with the thyratron as was shown in The diode conducts whenever an inverse voltage appears on the capacitance. output power times the pulse repetition rate in a typical line-type modulator. with the possibility of damaging the thyratron by exceeding its permissible operating voltage. The load no one switching device which is always better than the others. different characteristics as concerns life. although other gas fillings are sometimes used. 99 Switching Devices for Line-type Modulators. The polarity of the voltage reverts to the normal direction. 6. The series inductance LB the inductance of the transformer primary. do not seem to be used as frequently in modern radar as the hydrogen thyratron or the The saturable reactor also may be used as a switch. The resistance is chosen equal to the impedance of the pulse-forming network. and the capacitance . 6. A semiempirical. diode inserted in series with the charging inductance (Fig. have the advantage of relatively low impedance during the conduction state and can handle considerable power. and excessive build-up all is prevented. and impedance in the closed position. or hold-off. 6. precision of firing. is in magnetron radar 107 The advantage of a modulators. The /^-factor describes . the diode If is to keep the delay line from discharging until the thyratron fires. The hydrogen thyratron has been widely employed as the switch impedance is assumed to be equal to the impedance of the line. pulse. 106 The magnetron is a nonlinear impedance and will not be matched to the line under The mismatch can cause a spike to appear at the leading edge of the conditions. 6. maximum pulse repetition frequency. better capacity for high-peak currents and can be designed to be relatively insensitive to temperature.or mercury-filled thyratron is the rapid The hydrogen thyratron also has ionization and deionization time of hydrogen gas.8] Radar Transmitters 251 A disadvantage of d-c resonant charging is that the pulse repetition frequency is fixed once the values of the charging inductance and the delay-line capacitance are fixed.31. This is called the despiking circuit. The pulse modulator described above was assumed to operate from a d-c power Alternating current could also be used. Fig. The function of which is less than the prf as determined by the resonant frequency f .Sec. This voltage is in series with the d-c voltage of the power supply at the start of the next charging period. C form a resonant circuit that gives rise to an oscillation that reverses the voltage on the capacitance. This charge cannot be dissipated by the thyratron since its A small reversed voltage polarity is opposite to that needed to cause conduction. Two gas-discharge switches used in early radar moduThese two devices lators were the rotary spark gap and the enclosed. and the capacitance is chosen small enough so as to be almost completely charged after the oscillator draws 99 full-load current. Each of these has ignitron.31. 99 105 ' a mismatch occurs during the discharge cycle. The spike can be minimized by introducing an RC circuit in parallel with the primary as shown in Fig. There supply.

~ Saturable-reactor Modulator. since and C2 are in a series and Lls L C h. the voltage at B. The optimum number of stages will depend upon the desired width of the output pulse. 110 This limitation may be avoided by operating as indicated in Fig. Since the resonant circuit consisting of L C h and C\ is excited at its resonant frequency.000 or higher when using high-permeability nickel-iron alloys. current.34. As the current builds up in the resonant circuit.33. in Fig. Two-stage saturable-reactor modulator. the inductance L x reactance is lowered to a considerably smaller value L ls The switching action of L x allows the charge on C1 to transfer to C2 This is indicated by the current i C2 in Fig. across the capacitor n times the input voltage at A. C h and the capacitance Cl form a resonant circuit. Assume that the energy source feeding the modulator is a sine wave whose frequency is the same as the desired pulse repetition frequency. Since the reactance becomes a part of the discharge circuit. The reactance cannot be made arbitrarily small without permanently saturating the core. C h and Q. being 90° out of phase with the current. tion is its The advantage of the life. C x builds up to a maximum equal to reaches a maximum. the saturated condition can be as great as 2. Two stages are illustrated. the two inductances other than the charging inductance are biased to operate in the unsaturated condition and have high reactance. so that the resonant frequency of succeeding stages is higher. . decreases the voltage across the capacitor its When Q saturates and . 6. The voltage across C2 reaches a . 6. Initially. it limits the minimum width of the pulse. the voltage across the capacitor.252 Introduction to Radar Systems [Sec. The charging inductance Z. The circuit may be considered as a series of saturable reactors arranged in resonant circuits in which the networks are charged stage by stage.33. 6. the pulse-to-pulse jitter is less than with the thyratron. This change in inductance (impedance) may be used as the basis for switching action. and C2 forms a resonant circuit with a higher resonant frequency than that of Z. The ratio of inductance in the unsaturated condition to the inductance in current. peak repetition frequency for a particular thyratron. or pulsactor. relatively long It is saturable reactor in radar modulator applicaa passive device and uses neither electronic tubes nor Pola rizing /wine inas / \ Output transformer Energy source _J ^r J ^ JLZ2Z L Z | ° To load Fig. < Q and the inductance Lj returns to its unsaturated state. and the pulse These three parameters may be juggled so long as the P„ factor of the tube is not exceeded. but the inductance is low for large currents when the core is saturated. During the formation of the pulse the reactor becomes saturated and the pulseforming network must discharge through the saturated reactance of the inductor. . mechanical moving parts. Also. The combination of L ls C1. The inductance of each succeeding stage is made lower than that of the preceding stage. The chief disadvantage of the saturable reactor modulator is its poor and uncontrolled pulse shape.33. 6. 6. For this reason they have no effect on the initial charging action. 10S 112 The saturable reactor is an iron-core inductance so designed that its magnetic core is driven into saturation for normal values of coil The incremental inductance is high when the current through the coil is small and the core unsaturated.8 the trade-offs which can be made between the peak voltage.The current iC2 rises rapidly.

It derives its name from the fact that the switching is accomplished with "hard" vacuum tubes as opposed to "soft" gas tubes. This potential reaches a value which blocks or stops the conduction A A . increasing interval. The output pulse width may be changed to some extent by adjusting the polarizing current. rectangular The nonrectangular pulse shape is not a disadvantage in all cases. Build-up of the pulse in the two-stage saturable-reactor modulator of Fig. When operating in the manner described above. saturable-reactor winding and adding capacitance to ground so as to C The output waveform is simulate a lumped-constant delay line. The waveform generated by the driver is amplified by the power amplifier to the level required to pulse the transThere are any number of pulse-forming circuits which could be used for the mitter. During modulator. The energy stored in 2 is then transIn essence. 6. A more rectangular pulse may be had by tapping limitations of this modulator. 113 the conduction period. current flow.8] Radar Transmitters 253 maximum at the same time that L 2 saturates. 111 The capacitors from the taps to ground Time Fig.35. is more like that of a resonant sine wave than a rectangular not always well suited to radar application and is one of the This shape pulse. the last-stage. The frequency spectrum of a rounded pulse will fall off more rapidly on either side of the carrier frequency and tends to cause less interference. ' ' A single high-power tube operating as a blocking oscillator may be used as a pulse blocking oscillator is a self-excited.33. 6. energy-transfer compress the saturable reactors acts to of the chain the peak power and decreasing the pulse duration from stage to stage. 6. Block diagram of one type of hard-tube modulator. driver. overdriven oscillator.Sec. have negligible effect on the unsaturated operation. 6. used in this trimode capacity it is sometimes called apulsactor. causing grid current to biasing potential is developed across a capacitor in the grid circuit by the flow. Trigger from prf generator Driver (pulse shaping) (Video) power amplifier _To "transmitter Fig. 6. The driver generates the desired pulse waveshape. frequency interference with adjacent which might cause spectrum a wide produces pulse A bands. Hard-tube Modulator." The design of the power amplifier is similar to the design of conventional video amplifiers except that high-power tubes must be used. and a part of a resonant circuit. the grid is at a high positive potential. a switch. ferred to the load via the current i C3 . but at and during the switching operation. which is the output pulse waveform. they produce an improved output pulse waveform. The trigger pulses from the pulse-repetition-frequency generator initiate the driver at the proper instants of time. the saturable reactor performs the When functions of an inductive component. A block diagram of a particular hard-tube modulator is shown in Fig.34.^ 98 100 113 The hard-tube modulator is essentially a highpower video pulse generator.35.

15034 and the Machlett ML-7002 shielded-grid triode has made possible the use of hard-tube modulators at very high power levels. It can switch 3. These tubes have been employed as plate modulators for triodes and klystrons.8 The blocking oscillator when used for radar modulator application is not really an oscillator in the usual sense. should remain as nearly constant as possible over the required pulse duration (7 3 t 4 ). the voltage should fall as rapidly as possible (? 4 t ) to about 80 per cent of rated voltage. the voltage should be increased to the rated operating value very rapidly (f 2 In order to achieve a rectangular RF envelope the voltage / 3 ).05 when operated as a hard-tube modulator. The initial rise (t /j) of the waveform from zero volts to about 60 per cent of the rated magnetron operating voltage may be carried out at any convenient rate. If the top of the pulse were not flat but contained ripples.1 5030 and A.5 Mw Fig. The magnetron might start oscillating in an unwanted mode.36. usually the n mode. but not so rapidly that the Mw — — — — — . 11 * The transmitter pulse shape is not always the same as the modulator pulse shape. Introduction to Radar Systems [Sec. 6. or else the tube might not oscillate at all and an arc might form. The rate of rise during the next interval of time (f x t 2 ) must be long enough to permit oscillations to start in the desired mode of oscillation. 6. but is more a regenerative pulse generator. it would cause frequency pushing and broadening of the spectrum.) of 6-^sec pulse power. 80 In short-pulse application (6 ^sec ataprfof 500 cps) the A-l 5034 has been operated at plate voltages up to 55 kv.36 shows the ideal voltage pulse shape for a modulator required to pulse a magnetron with a rectangular shape.254 cycle. At the end of this interval. The 5 voltage may then be reduced as rapidly as convenient. 98 One of the limitations of the hard-tube modulator in the past has been the lack of vacuum tubes capable of handling the large power required for pulsing big radar transmitters. The development of tubes like the RCA types A. If the voltage were to rise too rapidly. Figure 6. The ML-7002 operates at 65 kv with a plate dissipation of 2 kw and is cooled by immersion in circulating oil. Ideal voltage pulse shape for a magnetron modulator. Modulator Pulse Shape. (After Gillette and Oshima. Once the oscillations are started in the desired mode at a low level. 114 IRE Trans. The differences between the regenerative pulse generator and the conventional blocking oscillator are discussed in Glasoe and Lebacqz. there would be too little time for the establishment of the desired oscillation mode. The voltage rises from 60 per cent to about 80 per cent of its rated value during this time. 75 The A-l 5030 is capable of switching 22 at a duty factor of 0.

as in airborne radar. 3. 1948. /.. (ed. 25. 17.: Operating Characteristics of Continuous-wave Magnetrons. Randals: High-power Interdigital Magnetrons. 255. McGraw-Hill Book Company. B. ED-6. 6. Eng. 1959. Boot. 1357-1363. M. Hok. 1950. IRE Trans. 12. 6. N.. a tube like the klystron presents no instead of the desired mode. Electronics. 2. T. F. 1953. 9. I. March. W. G. Hartman: The Magnetron as a Generator of Centimeter Waves.. 928-938. chap. pp. Elec. 29. September. C. L... 10.: Genesis of a Generator: The Early History of the Magnetron. Esperson. 115-118. IIIA. 271-288. System Tech. 647-654. Princeton. Proc. J. For different Other RF generators may impose example. A. It is light in weight and small in size and is particularly attractive in those applications where large and heavy packages are undesirable. Special circuits have been developed for approximating this shape. 156165. 1946. M. Nordsieck The Rising Sun Magnetron. vol. Comparison of Magnetron Modulators. Company. J. pp. Appl.: Pulsed Properties of Oxide Cathodes. The time jitter from pulse to pulse is usually worse with this modulator than with other types. 1948. easy to service. Hagstrum. G. vol. P. August. S. H. pp. and A. 182-183. pp. IRE. November.: Coupled Cavity Tunes X-band Magnetron. vol.. The ideal pulse shape can be only approximated in practice. G. J. and C. 14." D. Collins. 93. 4.): "Microwave Magnetrons.. Appl. Beltz./. 37. difficulty. Millman.. Van Nostrand C: Magnetron Mode : 13. on the modulator. Franklin Inst. 1954. most modulator problems encountered in practice can be 114 solved with properly designed conventional circuits. It pulse shape. F." MIT Radiation Laboratory Series. Kroll: Magnetron Research at Columbia Radiation Laboratory. Southworth. 1947. else operation may occur in the tt mode On the other hand. Fisk. July. 8. Okress.: Dispenser Cathode Magnetrons. "Principles and Applications of Waveguide Transmission. 1948. vol. Phys. February. requirements on the modulator. McGraw-Hill Book Company. pp. 644-657. IRE. 11. H. L. Advances in Electronics and Electron Physics. April. vol. It is less efficient than the others but systems engineer the greatest flexibility in operation. 5. Also contains an excellent bibliography on the magnetron. and A. J. Transitions. pp. 35. 1954. saturable reactor has the advantage of no active elements. The rate of rise of voltage must not be too slow. until the start of the next pulse.. pp. and P. vol. as for beacon interrogation. Wathen. pp. W. Inc. 1956. 1958. else the tube might oscillate and increase zero axis the 115 the noise level at the receiver or present false targets. E. Inc. vol. The hard-tube modulator can change pulse duration.J. vol. 167-348. and changing the pulse duration requires switching in another similar rise-time restrictions ' ' pulse-forming network.. T. A. MTT-2. January... pp. 7. Inst. but except in systems employing extremely short pulses. "Very High Frequency Techniques. New York... Schulman: Frequency Modulation and Control by Electron Beams. Inc. G. REFERENCES 1. J. 98 100 101 The line-type modulator is simple. and efficient. L. B. pt. vol. vol. offers the Pulse jitter is usually not bothersome. 1946. New York. 21 in Radio Research Laboratory Staff. Proc. The should be long. R. it should not recross is excessive. 33-37. A. 19.. J. 1947. . the Amplitron does not operate in the n mode as does the magnetron. Smith. or prf with it little may also be used to generate groups of pulses. for example. F.. H. vol. vol. Phys." vol. and its pulse shape is more difficult to control.Radar Transmitters backswing 255 Once the pulse has dropped below zero. Bell.. pp. Bernstein. Hull. and N. Coombes. See. E. 8. IRETrans. The Smith chart and its use are described in a number of microwave texts. Randall: The Cavity Magnetron. I. and J. consequently its life It is not as flexible as the other modulators.

vol. 33. Kilgore. March. IRE Natl. Hegbar: 1-kilowatt Frequency-modulated Magnetron for 900 Megacycles. 3. 1631-1653. H. Cuccia. Novick.. 23. Electonics Con/. 36. J. vol. Sperry Eng. and D. Learned. 20-29. 42. pp.256 15. 10. Ginzton. W. Varian: High Frequency Oscillator and Amplifier. IRE. pp.. Self: New A Design of High-power S-band Magnetron. 1947. G. and Basic Characteristics. pp. J. and R. MTT-4. Bush. C. 47. 10. March.. R„ C. pp. 97-1 10.) Dix. and J. IRE WESCON Conv. August.. and D. Technical Publication 1958. 1957. Electronics Conf. Hayter: Design High Power Pulsed Magnetron. pp. 105. vol. Kissinger: A Long-life C-band Weather Radar Applica- tions. 25. W. J. May. and H. San Bruno. Natl. and S. 168-173. H. C. Electronics. Polytechnic Institute : of Brooklyn. J. Proc. New York." chap. 1959. R. J. vol. L. Telecommun. A. 1958... M. 1955. Kurshan: Frequency-modulated Magnetron for Superhigh Frequencies.. vol. 4. Donal. J. May. vol. Natl. White. Frequency Pushing and Voltage Tuning.: Harnessing the Electron. and R. 162-163. C. pp. 2-18. B. Proc. H. pp. and P. 105. 1953.. 34. G. 1958. Developments in Broadband and High-power Klystrons. Inc. 1953. pp. New York. J. no. 10. 26. IEE. H. vol.. Cordray: How Long-line Effect Impairs Tunable Radar. Proc. R. H. R. 29-33. IRE Trans. 7. Conv. 41. Phys. November. 2. pt. A. 18. C.. 41. November. 332-338. Jr.: Multi-cavity Klystrons.. Varian. 368-378. F. R. 22. H. L. Introduction to Radar Systems Donal. pp. List ES/V/3. Welch. New York. Swanson. vol. vol. Eitel-McCullough. I. 35. 664-669. pp. June. 1. Shepherd. Proc. R. Appl. on Magnetron 7182(M543). P. "Vacuum-tube Oscillators. vol. Ginzton. Norris. Veronda: Recent Developments in High-power Klystron Amplifiers. Electronics. 1957. Lebacqz. Advances in Electronics. Jasberg. Wilbur: Magnetron Voltage Tuning in the S-band. 32. pp. 1939. pp.) facturing tion. The Generation of Shaped Pulses using Microwave Klystrons. suppl. Raytheon ManuCompany. IRE 28. Symposium on Modern Advances in Microwave Techniques.. 30. 37. vol. 3.. Okress. L. /. G. 35. pp. I. 123-132. 11. Voltage-tuned Magnetron for Application.: Modulation of Continuous-wave Magnetrons. H. 1954. F. vol. P. H. vol. March-April. W.. E. 115-121. . A. 657-664. vol. 833-838. 361-367. Bristol. 16. Chodorow. Willshaw: Microwave Valves: Survey of Evolution. vol. Record. (Chicago). Electronic Eng. 1952. 11. D.. 17. A. 161-171. IRE. 45. 43. Dalman. Jr. Preist. W. 18. Long-line Effect and Pulsed Magnetrons. Rev. vol. 19. pp. Engrs. M. 321-327. 24. C. Jr. Waltham. and Performance of a ED-4. Brochure on the VF10 Magnetron. Ferranti Electric. Speaks.) Type RK6410/QK338 Magnetron. 3496-9-55. Brit. S.: Prediction of Traveling-wave Magnetron Frequency Characteristics. 5. A. Shulman. (Additional data obtained via private communica- 29. J. Grid Controlled Pulsed Klystron Amplifier. Proc. 1956... IRE. 1958 (Paper 2659 R). E. 1957. A.. Proc. Calif. Proc. 1947. 31. IRE. Swearingen. January. E. UHF A Fla. J. Boyd. December. Veronda of the Sperry Gyroscope Company. S. 1955. November. Preist: Super-power Klystrons for Pulse Applications (brochure). Proc. April. IRE.: The Mitron: An Interdigital Voltage-tunable Magnetron. vol. English Electric Valve Co. H. Beck. Proc. July. vol.. 38. Foster. Mass. A. 12. pp. Sonkin: Design and Performance of a High-power Pulsed Klystron. Technical Information for the Trans. February. vol. and S. pt. and C. Griffin. J. Singh. Boot. 30. E. 41. E. Peters. Proc.... V. R. 1959. pt. IEE. and W. (Chicago). pp. pp. pp. 20. B. and G. Gainesville. W. (Additional data obtained via private communication. Neilson. Beltz. A. Veronda: The SAL-89. pp. pp. D.. 44. J. 30. 106-113. 35.. Inc. Hull. pp. F. May. A A : FM A : & : April. 20. 27. 577-609. pp. (Additional data obtained via private communication. 465-469. 40. July. A. 321-323. April. Proc. A. Personal communication from C.. Chodorow. vol. M. R. Jr. 72-76. 1957. 27. Inst.: Modulator Techniques for Gridded Klystrons and Traveling Wave Tubes. E. Principles of Operation. vol. 43. 1960. V. Deering: Three-cavity L-band Pulsed Klystron Amplifier.. Gleason. /. Ltd. IRE. and S. India. 1958 (Paper 2637 R). 1952. 39.. pp. Inc. IRE. Record. M. H. R. Pritchard. Rao: Proposed Ferrite-tuned Magnetron. H. A A : A 40. 419-425. Proc. suppl. T.. and R. Microwave J. July. 1584-1602.. and C. vol. F. 1954. John Wiley Sons. Shaw: Development of High-power Pulsed Klystrons for Practical Applications. Magnetron for W. 21 . IRE. V. 1953. and H. pt. J. 1955. Jr. Edson. pp. /. vol. and W. L.

1." Apr. 52. Traveling-wave 56^ Chodorow. G. vol. vol. 1957 Klystron Amplifiers. Proc.: "Traveling Wave Tubes. 1956. ED-6. IRE. Characteristics of a E. McBride.. (Paper 2668 R).. vol. Dohler. Proc. pp.. 2. Nalos.. 1958. 54. R. Beaver. 42." D. 5% Bandwidth 2. 486^95. A. S. Electronics. D. Maillart: An M-type Pulsed Amplifier. 1957. vol. Brown. April. 5. IRE. vol. IRE WESCON 48. pp. 106.'pt. pp. and W. W. pt. October. 3. Apr. 1958. vol. n 1950. 1957. W. 649-659. Contract 1959. 1958 (Paper 2624 R). C. J. vol. 1-19. Electronics. Patent 2. O. 33. Record. 2.. 65. A. An Investigation of the Magnetron 373 360. 45. Proc. Doehler. pt. Pantell: The Design and 57 Chodorow Megawatt Space-harmonic Traveling Wave Tube. 577-582. and D. Tubes. 1959. October. W. 8. and F. C: Crossed-field Microwave Tubes. New Microwave Platinotron 60. pp. 1950. J. Dixon: A May.: Some Recent Advances in Microwave „ . N. W. 1958 77. December. 1958. R. Applying the Amplitron and Stabilitron to 6. Transit Time Effects in Ultra-high-frequency Class-C Operation. May. J. New York. IRE. J. J. Dain. 46. 61. J. Groendijk. R.wave Tubes. 454^157. 18. pp. 63. vol. E. J. IRE Trans. (Project 207). V. A. Princeton. Stanford Electronics Lab. pp. 64. Electronics. pp. Feinstein. Moreno of Varian Associates. pp. Ultra-high-frequency Power Amplifiers. L.: Effect of Beam Coupling Coefficient on 1958.. IRE. D. pp. . available from Technical Services. pt. : : 76. rnr IRE . Natl. Dow. pt. J. 513-522. B. W.S. R. P. E. Caryotakis. /. E. 3. 31-38. 1954. 52-55. Brown. 78. 1960. . 1959 42-45. pp. and H.270 entitled "Regenerative Amplifier. 70 Brown X-band Amplitron. pp. xr Record. 38. Brown.: Present State of Art in High Power Traveling-wave Tubes.: Factors in the Design of Power Amplifiers for Ultra vol. 35^12. Weil. Smith. 1957. pt..: High-vacuum Power Tubes. C: Platinotron Increases Search Radar Range.: suppl. Nalos: The Design of High-power vol. Dubois. vol. vol. IRE WESCON Conv. Jr. J. Trans. IRE Intern. 72. 13. Inc.: Microwave Triodes. C. v . A. 1958. vol.: Traveling. G.5 50 King. Palo Alto. Nonr 225(24). 1. Kino: The Large Signal Behavior of Crossed-field Traveling-wave Devices. E. Electronic 58.. 68 Wiehtman. 66. Increasing Bandwidth of 47 Dodds. 51. 1. Proc. Symposium on York. P. 3. 48-53. Huber: wave Amplifier Tube. 45. H. 1959. Yadavalli. pp. 1955. 12. Apr. IEE.S. Waveguides. W. VanNostrand Company. vol. . M. 105. „ M . T. Electronics Conf. January. IRE. S. Moreno. 120-130. Doehler. pp. 30. Perloff: High Power : ONR NR W 71. pp. 2. /. J. Symons: IRE WESCON Conv. pt. and N. January.. 101-110. 813-820. A. Tubes. no. Auld. 1957. S-band Klystron. . Amplifier. 1947. 67 Proc. 3. Meacham. 7. October. pp. July. pt. no. IEE. 62.881. V. and G. Tech Rept 52 The Magnetron-type Traveling69. pp. 529-567. "Vacuum Tubes. Feb. Pierce. Brc>ad-band Operation of Multicavity 49. 29. pp.: Methods of Study of the Broad-band Frequency Response High Power Microwave Amplifiers. 1209-1222. 9. R. A. vol.J. Conv. December. Inc. Klystrons'. vol. {Chicago). Record. suppl. Proc.. vol. 33. pp. IRE. Record. vol. September." vol. 60-64. Polytechnic Institute of Brooklyn. H. Warnecke.: Calit. U.. T.. 111-114. Microwave : no. Proc. "Radar.„„ 73 Spangenberg. New 59. pp. C: U. 10. B. 105. 1735-1747. Proc. 15-kilowatt S-band Amplitron. Dain. vol. Kleen. : A 164-168. MTI Radar Systems. Lerbs. B. W. Wide Band High Power Klystrons. B. vol 35. G. Electronics. 55 ^T T . 75. J. A.. Personal communication from T. Division 14. 999-1008. B. May. pp. and R: L. Dombrowski. Otsuka. Walter: Wide Band 1957. Nalos. 75-79. Proc. IEE. K. Conv. Conv." McGraw-Hill Book Company. pp. R. Description and Operating Characteristics of the Tube Device. pp. vol. 10. and R. IRE. pp. Staprans. "OSRD Long History. Proc.: Theory of the Amplitron. W. Record. Brown. IRE. ED-6. vol. r _. 74. vol. December.. H. Natl. 1957. G. Electronics. and R. 12. W. vol. no. 29. Department of Commerce. B. and : CW ' Office ot H. 35-42. G. : 1. Proc. 105. Proc.. suppl. 1948. Beaver. IEE. 1959. Zawada: A 3-megawatt. K. 103-111. 1959. 18. O. 419-428. pp. S. and E. Jolly. November. pt.Radar Transmitters 46. L. R. 1957-1958. 257 Kreuchen. Aug. pp. Proc. Pierce. 3. of the Multicavity Klystron Amplifier. pp. . J. • A MW 53. 8. Jepsen. Guerlac. 1959.: . 44. wf. vol. vol. Advantages of Ceramics in Electron Tubes. High Frequency. pp. 1364-1373. A. O. Microwave J.

and H. H. 99. vol. vol. 18. M. J. H. 1957. no. 95. J. under USAF Contracts W-28-099 ac-216 W-33-038 ac-16649.: 109. 25. W. Glasoe. 1958. W. and J. 1953. 1953. I. W. W. J.. no 24 pp 70-72 ' rr June 12. 41. 2. M. Whinnery (ed. pp. H. 80-83.. 80. 92-97. pp. pp. 103. Sloan. Rev. Day. W. Garbuny. 38 vv pp 515-520 May.: Advances in the Techniques and Applications of Very-high-power Gridcontrolled Tubes. Tucker. 12. pp. 42^16.258 Introduction to Radar Systems A. IRE Trans.. Bennett. vol...: Large Signal Theory of Power Triodes. 185-207. pp. T. ' pp.. vol. pp. W. W. 1955. August. 107. vol. 10. Paralleled Amplifiers Increase R-F Power. September. vol 5 ' 5 ' ' M. Apr. 1959. 1946. Morrison. January 1959.Chicago). Marshall.): "Pulse Generators. Preist. Welch: The Generation of Ultra-high-frequency Power at the Fiftykilowatt Level. 101. June. H. Proc. and L. vol. J. 3.. Zinn. on Electron Tubes report. 51. Brown. C. Bennett. 36-41.. 98. California Microwave Lab. 98. F. Electronics Conf. vol. for a Compact Kilowatt UHF Beam Power Tube ' Record. D. G." advertising material of Eitel-McCullough. New York r j > Group 1947. New York. E. 52. vol 33 ' no. A. vol. N. RCA March. Dow.. May 15. 1950. RCA Rev. R. pp. The Use of Saturable Reactors as Discharge Devices for Pulse Generators Proc. McGraw-Hill Book Company Inc. and S..: The Application of Pulse Forming Networks. G. 10. and W-19-122 ac-38. Sloan. Nasley. Kazanowski: One-kilowatt Tetrode for Transmitters. 1940.. 943-999. A. Electronics. suppl. L. • > . vol. 1958. R.: The Resnatron. vol. Peterson. pp. UHF UHF UHF UHF 90. Hoover: Gas Tubes Protect High-power Transmitters. D. and E." MIT Radiation Laboratory Series. 58. 1956... r J 100. March. vol. J. Contract DA 36-039-SC-74981. Proc. M. Electronic Design vol s Electronic Design. E. pp. 86. chap. Electronics. pp. ED-6. G. Electronics. 62. Dec. Natl. 85. p. 41. and H. Review of Modulators and Their Requirements. W. 5. 93. pp. 41. January. Hoover. pp 116-133 ' 1 10. CP-3. D. Proc. M. vol. and J. B. and E. G. vol. IEE. pt. Phys." vol. B. pp. and H. 1949. : UHF 81. A. Busch. 16. W. pp 46^19 rr Graydon. 1957. 193A. Dow. : ' (. Inc. 1957. pp. 1071-1086. January. McBride. Rept. C. IRE vol. Parker. W. V. vol. Jr. 97. pt." Advisory 79. Zinn. 13-19. 1953. H. no. M. Grid-controlled Power Tubes for Radar Purposes. E. V in J. 1952. W. Salisbury. Williams: Economic Design of Saturating Reactor Magnetic Pulsers Trans. {Chicago). Hansen: Reflex Resnatron Shows Promise for UHFTV. 94. 169-171. Bell System Tech J / vol. and Electronics. 46. 92. Trinkous. V. 104. Sutherland. 603-614. vol. Sheppard.. also IRE Trans. H. Electronics. 1958. and G. 32.. W. 83. Gray. 102. February. Dec. Lampe: The Resnatron as a 200-MC Power Amplifier. Watrous.: Performance and Packaging of Modulators. Rev. W. vol. 10. 116-119.. G. F. D. 1960. G. 89. IRE. Coate: "Principles of Radar. C. pp. May. W. Behrend. September. 33. and M. McNees: An Axial-flow Resnatron for Proc. 34. 1956. I. V. McGraw-Hill Book Company. and J.. vol. 1956. Reiskind: Method of Multiple Operation of Transmitter Tubes Particularly Adapted for Television Transmission in the Ultrahigh Frequency Band. 161-172.. 3. Inc. Reintjes. 1960. W.. AIEE. 37-42. 23. 1948. N. Univ. May. no. W. 1955. 105. 1952. March. 105. H. Proc. Neitzert: Magnetic Pulse Modulators. IRE WESCON Conv. 87. Proc. J. 19. McArtney: Gas Clipper Tubes for Radar Service. Garbuny. CP-4. 7-13. Natl." chap. IRE. 1483-1492. Dec..). L. 96. Armstrong. "State-of-the-art Report on High Power Pulsed Tubes for Radar Purposes. and W.. 88. pp.: Pulse Forming Networks.. W. S. 1950. pp. Inc. 1958 (Paper 2752 R). McGraw-Hill Book Company. Electronics. Melville.. and C. 1951. L. pp. vol ED-3 pp rr 57-61 January. Hoover. W. P. and H. D. V. M. Beaver: Resnatron Tubes. Electronics Conf. IRE Trans. 1946. IEE. K. 64. IRE. for Service.. Marshall: Power. 144-147. vol. "Klystron Facts— Case Five. chap. 63-66. 19 in Radio Research Laboratory Staff. Welch: The Resnatron. Proc.: A New Design Approach 2. 26-29. P. 9. Commun. E.: New Beam Power Tubes IRE Trans. Schulte. "Very High Frequency Techniques. 3. January. A. pp. L. 35-47. 15. pt.. IRE. vol. 84. Electronics. Wittenberg. McCreary. January. 106. Proc. vol. 550-558. Using Silicon Diodes in Radar Modulators.. Lebacqz (eds.: Annular Circuits for VHF and UHF Generators. vol. 82. R. New York. M.: Thyratrons in ' Radar Modulator Service. H. 108.: Theory of the Reflex Resnatron. E. F. Mathias. 1949. 15. 29. pp.: A pp. H. 91.

pp. 34. Record. May 1958. A. Proc.. and K. Gillette. pp.. 1954. vol. Electronics. vol. Sets. Bell Labs. no. Radar Magnetrons. CP-3. pp. H. Lee. July 3.Radar Transmitters 111.: False Echoes in Line-type Radar Pulsers. pp. R. 113. Thomas. 27. 26-31. March. 31. A. vol. 1288-1295. 153-156.: "Hard Tube" Pulsers for Radar. P. no. 1956. Oshima: Pulser Component Design for Proper Magnetron IRE Trans. 259 19. 1959. April. Kunitz. 112. IRE. . 42. Operation.: Using Magnetic Circuits to Pulse Radar 42-43. Electronics. August. 32. 72-75. R. 115. E. 1956. vol.: Saturable Reactors Fire 9. pp. Reise. 114. vol. H.

one of the prime characteristics of radar. The type of antenna normally used for radar applications differs. but large in terms of wavelengths. High-gain antennas with narrow beamwidths are quite practical at microwave frequencies. At the microwave frequencies. Thus the radar antenna is called upon to fulfill reciprocal but related roles. Narrow beamwidths are important if accurate angular measurements are to be made or if targets close to one another are to be resolved. in general. depending on which is more convenient for the particular discussion. say. Microwave lenses have also found application in radar. On reception the antenna collects the energy contained in the echo signal and delivers it to the receiver. Antenna Parameters 1. The large apertures required for long-range detection result in narrow beamwidths. The directive gain is descriptive of the antenna pattern.8)] these two roles were expressed by the transmitting gain and the effective receiving aperture. Both definitions are of interest to the radar systems engineer. measure of the ability of an antenna to concentrate energy in a particular direction is called the gain. The advantage of microwave frequencies for radar application is that apertures of relatively small physical size. 1 Directive Gain. The earliest radars that operated in the or the bands used array antennas. surface-wave antennas are sometimes useful when the antenna must not protrude beyond the skin of the air frame. is extensively employed. the parabolic reflector. In the radar equation derived in Chap. whereas they would be difficult to achieve at.7 ANTENNAS 7. the radar antenna will be considered either as a transmitting or a receiving device. which is well known in optics. but the power gain is more appropriate for use in the radar equation. 1 [Eq. Two different. gd „ = maximum radiation intensity 2 jr— raverage radiation intensity : (7. The directive gain of a transmitting antenna may be defined as directivity. but related definitions of antenna VHF UHF A gain are the directive gain and the power gain. The vast majority of radar antennas have used the parabolic reflector in one form or another. are usually designed for omnidirectional coverage or for fixed point-topoint transmission.1) 260 . Results obtained for one may be readily applied to the other because of the reciprocity theorem of antenna theory. The former is sometimes called the while the latter is often simply called the gain. Radar antennas must generate beams with shaped directive patterns which can be scanned. from antennas used for communications. An antenna with a large effective receiving aperture implies a large transmitting gain. can be obtained conveniently. Most communication antennas. short-wave communication frequencies (HF). on the other hand. (1. In airborne-radar applications.1. In this chapter.2 The purpose of the radar antenna is to act as a transducer between free-space propagation and guided-wave (transmission-line) propagation. The two parameters are proportional to one another. The function of the antenna during transmission is to concentrate the radiated energy into a shaped beam which points in the desired direction in space.

spillover radiation by the reflector. 7.1 if the which decrease in intensity with lobes minor of series are a sidelobe first the Following most antennas the first vicinity of broadside (in this increasing angular distance from the main lobe. Proc.1. the classic spherical coordinate system shown often used. radiation antenna in appear always not does vestigial lobe beam. in Fig." The radiation pattern also has a pronounced lobe in leakage direct and to reflector the of effects backward direction (180°) due to diffraction through the mesh reflector surface. (0.1 is plotted as a function of one function as a intensity radiation P(6. distribution phase vestigial lobe as in Fig. In the the sidelobe level to rise. by P(0.) The first sidelobe is smeared into a sidelobe appears instead.1] Antennas 261 the direction where the radiation intensity is the power per unit solid angle radiated in radiation intensity as a function of the angular the plot of A denoted is and P(6. intercepted not is which feed the from radiated energy This is due to the Some of it "spills over." In patterns. angular radiation pattern shown in Fig. across the aperture is not constant. plotted as a function plotted on a relative basis.<f>) the coordinate.61 is In theoretical work. The 1 1 1 I F t — -Main lobe -10 -10 80 Degrees 100 off axis reflector antenna illustrating the main-lobe Fig. 7.<f>).<£) = P(6. an antenna with a symmetrical pencil-beam pattern can be for rectangular pattern radiation-intensity The coordinate. 7. power called a is angle of unit area. or "shoulder. Radiation pattern for a particular paraboloid 3 and the sidelobe radiation. When when both are called the antenna radiation pattern. An example of an antenna radiation pattern for a paraboloid antenna is shown 3 The main lobe is at zero degrees. on the side of the main particular radiation pattern is the vestigial lobe. 7. {After Cutler et al. 7. 7. causes feed the from example 100 to 115°). The first irregularity in this plotted in Fig. but the actual pattern is a plot of commonly employed with of the two angles 6 and <f>.Sec.Q)P(0& . when identical are pattern radiation-intensity and the plotted on a relative basis the maximum is normalized to a value of unity. IRE. .1. that is. or power per coordinates is called a radiation-intensity pattern. but any other convenient set of The ground-based angles can be used. angular a plot in one patterns in the apertures can often be written as the product of the radiation-intensity two coordinate planes for instance. The two-angle coordinates antennas are azimuth and elevation.<£) density. A complete three-dimensional plot of the radiation pattern is not always necessary. represented For example.. The power The power pattern pattern.

Equation (7. They would be the same if there were no losses The power gain and the directive gain may be related by the radiation efficiency factor p r as follows G= PrG D (7.<f>) iWW dd (7. or r if 41.5a) the half-power beamwidths are measured in radians. It does not take account of losses due to ohmic heating. If 6B and B are the half-power beamwidths in the two orthogonal planes. and the total power radiated is found by integrating the volume contained under the radiation pattern. The directive gain. The power gain. is defined as q = maximum radiation intensity from subject ante nna radiation intensity from (lossless) isotropic source with same power input "* This definition is the one which should be used in the radar equation since it includes the losses introduced by the antenna. (7.4) The beam area is the solid angle through which all the radiated power would pass if the power per unit solid angle were equal to P(6». One of the fundamental theorems of antenna theory concerns were described above in terms of a reciprocity. which will be denoted by G.<? % ax _4n B d<j> SSP(d. 1 Thus the gain definitions apply equally well whether the antenna is used for transmission or for .: 262 Introduction to Radar Systems [Sec. The maximum power per unit solid angle is obtained simply by inspection. It states that under certain conditions (usually satisfied in radar practice) the transmitting and receiving patterns of an antenna are the same. or resolution considerations and is more closely related to the antenna beamwidth.<f. the directive gain as defined by Eq. includes the effect of the antenna losses and any other loss which lowers the RF The power gain antenna efficiency. an equivalent antenna pattern. <f> Since the average radiation intensity over a solid angle of 4tt radians is equal to the total power radiated divided by 4n.253 the half-power beamwidths are measured in degrees. heating or a mismatched antenna.2) can be written as GD where 47TP(6>. Power Gain. is of importance for coverage.3) gives gd if = r-j- (7. 7. The definition of directive gain is based primarily on the shape of the radiation pattern.)d6d4> (73) B is defined as the beam area : B __ SSP(d. The difference between the two antenna gains is usually small.1 and the complete radiation pattern can be specified from the two single-coordinate radiation patterns in the 6 plane and the plane. It defines in effect. the beam area B is approximately equal to 0U„ Sub<f> stituting into Eq. which is always greater than the power gain.<£)max over the beam area.7) The definitions of power gain and directive gain transmitting antenna. (7 1) can be written as G = 47r(maximum power radiated/unit solid angle) total power radiated ' This equation indicates the procedure whereby the directive gain may be found from the radiation pattern. accuracy.

7.Sec. If too large a portion of the radiated energy were contained in the sidelobes. although the radar systems engineer might sometimes want to specify one or the other. maximum upon gain. Sidelobes of the order of 20 to 30 db below the main beam can be readily achieved with practical antennas. 7. Antennas The only practical distinction is 263 and receiving antennas greater power.2) . Effective Aperture.1] reception. the The polarization direction of the electric field vector is either vertical or horizontal. with a consequent lowering of the . No general rule can be given for specifying the optimum sidelobe level. This depends the application and how difficult it is for the antenna designer to achieve low If the sidelobes are too high. the effective the effective Another useful antenna parameter related to the gain is the It may be regarded as a measure of area presented by the antenna to the incident wave. the radar is more subject to interference from nearby friendly transmitters. Circular polarization and linear polarization are special it is cases of elliptical polarization. The relative amplitudes of the two waves and the phase relationship between them cart assume any values. A Also. With extreme care it might be possible to obtain sidelobes as low as 35 or 40 db. that the transmitting antenna which must be made between transmitting must be capable of withstanding effective receiving aperture. The gain G and area A e of a lossless antenna are related by A = = 4ir2 e 4ir Pa A (1 g) A A 2 where A K = PaA = wavelength A = physical area of antenna p a = antenna aperture efficiency (defined in Sec. the polarization is circular. appear as false targets. Circular polarization is often desirable in radars which must "see" through weather disturbances. depending upon the importance of ground reflections. If the amplitudes of the two waves are equal. Outline of Chapter.) and the current distribution across . but the specifications to the antenna designer must often be dictated by the practical limitations imposed by nature and not by the unattainable specifications of theory divorced from practice. the radiation pattern (beamwidth. An example of sidelobe radiation from a typical Sidelobes and Spurious Radiation. Most radar antennas are linearly polarized that is. which are perpendicular to each other in space. sidelobes. (7. high sidelobe level makes jamming of the radar easier. or effective area. In many applications the radar systems engineer might desire sidelobes of extremely low level. traveling in the same direction. and if they are 90° out of (time) phase. the The choice between horizontal and vertical linear polarization is often left to the discretion of the antenna designer. However. Low sidelobes are generally desired for radar applicaantenna was shown in Fig. etc. of the electric field vector. although there is no theoretical reason why they should not be possible. may also be elliptical or circular. Linear polarization is most often used in conventional radar antennas since easiest to achieve. considerably lower sidelobes seem difficult to achieve.9) The direction of polarization of an antenna is defined as the direction Polarization. there would be a reduction in the main-beam energy. Elliptical polarization may be considered as the combination of two linearly polarized waves of the same frequency. 7.1 tions. the relationship between practice of radar systems engineering. The purpose of this chapter is to summarize the results of antenna theory and technology which might be of particular interest in the study and In the next section. strong echo signals can enter the receiver and sidelobes.

D and I being measured in the same units. 10) J -a/2 I where A{z) = aperture distribution.(f>). It extends several antenna diameters from the aperture and. is usually of little importance to the radar engineer.2. Antenna Radiation Pattern and Aperture Distribution electric field intensity £(<£) produced by the radiation emitted from the antenna and the phase of the current distribution across the amplitude is a function of the aperture. The mathematical summation of all the contributions from the current elements contained within the aperture gives the This integral cannot be readily evaluated in the field intensity in terms of an integral. 1 4 E((f>) may be found by adding vectorially the contribution from the various current elements constituting the aperture. In the Fraunhofer region.11) where \A(z)\ T(z) = amplitude distribution = phase distribution . The demarcations among these three regions are not sharp and blend one into the other. and arrays. including both the amplitude and phase distributions. where D is the size of the aperture and /A or the distance R F either R F X is the wavelength. At a distance of 2D 2 /X. However.</>)\ is called the field-intensity pattern of the 2 antenna. the Fresnel region. In the Fraunhofer region. or A(z) = \A(z)\ exp/F(z) (7.94 that of the Fraunhofer gain at =D = D infinity. The effect on Several methods of pattern synthesis are discussed. approximations to the solution may be had by dividing the general case. The width of the rectangular aperture and coordinate system shown in Fig. the integral for electric field intensity in terms of current Consider the distribution across the aperture is given by a Fourier transform relation. 7. assumed to be flowing in x direction. ) (7. and the angle in the yz plane as measured from the y axis is <j>. for this reason. including the parabolic reflector. region. area about the antenna aperture into three regions as determined by the mathematical The - approximations that must be made. the gain of a uniformly illuminated antenna is 0. the gain is 0. may be written as a complex quantity. aperture in the z dimension is a. dz is £(«£) = ( " A(z) exp ()2n . 7. 7. defined in the previous section. assuming a </> > I. In the Fresnel region. The "boundary" R F between Fresnel and Fraunhofer regions is usually taken to be 2 ID^jX. The plot of the electric field intensity \E(0. the current at distance z.2. At a distance given by 2 jX. the radiation pattern of broadband signals and of errors in the aperture distribution is the antenna aperture considered. the radiating source and the observation point are at a sufficiently large distance from each other so that the rays originating from the aperture may be The near field is followed by considered parallel to one another at the target (observation point). The far-field electric field intensity. or far-field. A(z). Fresnel region in radar.264 Introduction to Radar Systems [Sec.2 This is followed by descriptions of the various is discussed. types of antennas which have been applied to radar.sin \ 2.99 that at infinity.<j>)\ is the power radiation pattern P(d. lenses. The farthest region from the aperture is the Fraunhofer. The plot of the square of the field intensity \E(Q. The region in the immediate neighborhood of the aperture is the near field. The chapter closes with brief discussions of radomes and focused antennas. rays from the are not parallel and the antenna target) point (or radiating aperture to the observation Little application is made of the radiation pattern is not constant with distance. The vast majority of radar antennas are operated in the Fraunhofer region.

14) This may be used as a basis for synthesizing an antenna pattern. = since A(z) =- f" E(<f>) exp t-j2n - sin <f>) d(sin <j>) (7.10). 7.10) can be extended over the from -co to +oo ±a/2. (7. since the aperture distribution is zero beyond z The Fourier transform permits the aperture distribution A{z) to be found for a given field-intensity pattern E(</>). It may also be The aperture distribution has been defined in terms of the current / in the x direction. It will be assumed only the effects of the that the phase distribution across the aperture is constant and amplitude distribution need be considered. In the remainder of this section. or integration.Sec. contributions across the face of the aperture. be advanced or retarded in phase by 2ir(zjX) sin <j> radians. that is. .13) The limits of Eq. Rectangular aperture and coordinate system for illustrating the aperture distribution and the far-field electric-field-intensity pattern. finding the aperture distribution A{z) which yields a desired antenna pattern E(<f>). . the antenna radiation pattern will be computed for various one-dimensional aperture distributions using Eq. polarization for defined in terms of the magnetic field component z or in terms of the electric field component Ez for polarization in the z direction. the contribution from a particular point on the aperture will principle.10) represents the summation. 7. The expression for the electric field intensity [Eq. (7. *"(/) = is J — 00 P /(0 ex P (-JW0 dt (7. applied to the calculation of the radiation or field-intensity patterns if the aperture The Fourier transform of a function /(f) is defined as distribution is known. of the individual contributions from the current distribution across the aperture according to Huygens' At an angle <f>. provided 5 these field components are confined to the aperture. Each of these contributions The field intensity is the integral of these individual is weighed by the factor A(z).10)] is mathematically similar to Therefore the theory of Fourier transforms can be the inverse Fourier transform.2.12) and the inverse Fourier transform /(0 -\: F(f)exp(j2irfi)df infinite interval (7. (7. H £W */ relationship between the Fig.2] Antennas 265 Equation (7.

7. 7.266 Introduction to Radar Systems [Sec.15) sin (j> Normalizing to make E(Q) = 1 results in A = I /a.88/l/a radians. One-dimensional Aperture Distribution. which is of the form (sin x)jx. and the beamwidth as measured between the half-power points is 0.sin A z \ <j> I dz [sin ir(ajX) sin Tr{ajX) sin <f>. It does not matter whether the distribution is produced by a reflector antenna.2 The inverse Fourier transform gives the electric field intensity when the phase and amplitude of the distribution across the aperture are known. The wider the aperture.3. therefore £(</>) = sin [Tr(ajX) sin <f\ (7. the antenna pattern as computed from Eq. or S\Xja deg. The aperture is defined as the projection of the antenna on a plane perpendicular to the direction of propagation. (7. distribution. (7. If the constant value of the aperture distribution is equal to A and if the phase distribution across the aperture is constant. .16) n(a/X) sin This pattern. line This might represent the distribution across a aperture extends in one dimension only. Perhaps the simplest aperture distribution to conceive (but not necessarily the easiest to obtain) is the uniform.10) is E(<f>) = A A f a/2 exp / 1 277 . — -477- -377- -27777- la A) sin <l> Fig. J -a/2 V / _ sin [7r(a/A) sin (77-/A) </>]_.16) is positive over the entire main lobe. The intensity of the first sidelobe is 13. a lens. </>] (7. 7. returning to a positive value in passing through the second zero. The solid curve is the antenna radiation pattern produced by a uniform aperture distribution the dashed curve represents the antenna radiation pattern of an aperture distribution proportional to the cosine function. The angular distance between the nulls adjacent to the peak is Xja radians.3.2 db below that of the peak. but changes sign in passing through the first zero. The $ shown by . or rectangular. source or the distribution in one plane of a rectangular aperture. The uniform distribution is constant over the aperture extending from For present purposes it will be assumed that the a/2 to +a/2 and zero outside. the narrower the beamwidth. and so on. is the solid curve in Fig. or an array. The voltage pattern of Eq.

aperture distributions.1 are (1) the relative gain produced by the particular antenna aperture distribution compared with the gain produced by the uniform aperture distribution. intensity db below maximum 13. A{z) = = = = 1 0.5 Triangular.575 0.W2) — 7r/2 J "| (7.1. A{z) Circular. 1 = 1— = Vl - |zl 0. A(z) t Silver.515 69A/a 83A/a 95A/a lllA/a 51 A/a 23 32 40 48 13.865 53A/a 56A/a 66A/a 73A/a 58. 1 0) can be conveniently carried out. (7. deg 51 A/a 51 A/a Intensity of first sidelobe.5A/a 20. and (3) the intensity of the first sidelobe as compared with the peak intensity.17) Tr{ajX) sin </>. 7.1 lists some of the properties of the radiation patterns produced by various 7. Type of distribution. they serve to illustrate how the aperture More complicated distributions which distribution affects the antenna pattern. (The pattern of a one-dimensional produced by the two-dimensional circular circular distribution is equivalent to that may not be the distributions these Although illumination. (2) the beamwidth in degrees as measured between the half-power points of the antenna pattern. Such phase reversals are characteristic of antenna 1 Also shown in Fig.833 0.75 0.4 17.8 17.) uniform with aperture employed with practical radar antennas. 7. machine computation. and circular distributions are included.667 0. Fourier transforms or which cannot be of tables from available cannot be readily found expressed in analytical form may be determined by numerical computation methods or triangular. An examination of the information presented in this table reveals that the gain of It is shown the uniform distribution is greater than the gain of any other distribution.2 Uniform. by Silver 1 that the uniform distribution is indeed the most efficient aperture distribution.2 13. .810 0.2] Antennas 267 odd-numbered sidelobes are therefore out of phase with the main lobe. (w (V ip — . The rectangular. Radiation-pattern Characteristics Produced by Various Aperture Distributions f A = wavelength a = aperture width .2 15. A(z) = 1 Cosine. 3 is the radiation patterns in which the minima are equal to zero.6 26. The properties of the antenna radiation patterns listed in Table 7.970 0. and the evennumbered ones are in phase. cosine to the nth power.Sec.6 z2 The aperture distributions are those which can be readily expressed in analytic form and for which the solution of the inverse Fourier transform of Eq. \z\ < 1 Relative gain Half-power beamwidth.0 1 0.1 = 1- (1 - A)z 2 : A A A A = = = = 1. pattern for the cosine aperture distribution 77Z A(z) = is cos — a \z\< The normalized radiation pattern E(<f>) = 77 sin {f y) where y Table Table = + tt/2) + + 77/2 sin .994 0.8 0. A(z) = cos" Oz/2): n = n 1 1 2 n 3 n 4 n Parabolic. 7.

In practical antennas. taking account of the proper phase relationships due to the difference in path lengths. Antenna Efficiency. losses are present and the over-all efficiency is the product of three factors :(1) the aperture efficiency. The relative gain is also may called the aperture efficiency [Eq. Thus low sidelobes and good efficiency run counter to one another. The aperture efficiency times the physical area of the aperture is the effective aperture. 40 db. albeit small. that is.18) where r is the radius of the aperture. The aperture efficiency is a measure of the gain of an antenna relative to the gain of a similar antenna with uniform aperture distribution. . For example. or most efficient. 1 is that the antennas with the lowest sidelobes (adjacent to the main beam) are those with aperture distributions in which the amplitude tapers to a small value at the edges.9)].2 the one which maximizes the antenna gain. There is a practical limit beyond which it becomes increasingly difficult to achieve low sidelobes even if a considerable amplitude taper is used. In a practical antenna this will not necessarily be true since there will always be some unavoidable phase variations caused by the inability to fabricate the antenna as desired. The over-all antenna efficiency would be the same as the aperture efficiency if the antenna were perfect.6). if all the energy from the feed were collected without loss by the reflector and if there were no losses in the antenna due to mismatch or to other causes. aperture distribution. where r is the radial distance from the center of the circular aperture.2 db of the uniform illumination. 7. For a circular aperture with uniform distribution. an aperture distribution which follows a cosine-squared law has a relatively large illumination taper. The greater the amplitude taper. The radiation efficiency defined by Eq. Therefore the relative-gain column be considered as the efficiency of a particular aperture distribution as compared with the uniform. word of caution should be given concerning the ability to achieve in practice low sidelobe levels with extremely tapered illuminations.268 that is. shall consider here the We antenna pattern produced by a two-dimensional distribution across a circular aperture. Introduction to Radar Systems [Sec.d) are used to describe the aperture distribution A(r.7) is the product of the last two factors. the tude. The field intensity at a distance R is thus proportional to 2 r E(R) =j "dd( °A(r. such as those of cos 3 and cos 4 It was assumed in the computation of these radiation patterns that the distribution of the phase across the aperture was constant. amplitudes of all the waves are added at the point. 6 The examples of aperture distribution presented previously in this section applied to distributions in one dimension. the lower the sidelobe level but the less the relative gain and the broader the beamwidth. Its sidelobe level is 32 db as compared with the 13. Another property of the radiation pattern illustrated by Table 7. The polar coordinates (r. Any practical device is never perfect. it will always be constructed with some error. all of the same phase but of different ampliTo find the field intensity at a point a distance 7? from the antenna. (7. Circular Aperture. The economic limit to the sidelobe level of conventional antennas seems to be of the order of 35 to A . Huygens' principle may be applied in the far field by dividing the plane wave across the circular aperture into a great many spherical wavelets.d) exp (- — )r dr (7. and 6 is the angle measured in the plane of the aperture with respect to a reference. The phase variations due to the unavoidable errors can cause the sidelobe level to be raised and the gain to be lowered. (7. (2) the spillover efficiency (if a reflector or lens) and (3) the efficiency of the feed.

radiation pattern produced by the obstacle from the radiation pattern of the undisturbed 1 . placed at the focus of One . One of the chief examples of aperture blocking is the feed in reflector-type antennas. and the shown in Fig.6 db is 30.3.6A/A and the first sidelobe is 24. 7. 1. = 0. where/? = 0. . but the gain aperture. relationship. Parabolic-reflector An Antennas 1 7S of the most widely used microwave antennas is the parabolic reflector (Fig. The sidelobe level Additional properties of this relative to a uniform distribution is 50 per cent. beamwidth is 58. The parabola is illuminated by a source of energy called the feed.5 the amplitude distribution of a circular aperture is similar to The sidelobes may be reduced. 7. Aperture blocking degrades the performance of an antenna by lowering the gain and raising the The effect of aperture blocking can be approximated by subtracting the sidelobes. table 6. 7.19) where f = 277(> /A) sin = first-order Bessel function j^) <j> A plot of the normalized radiation pattern is 17.. . This procedure is possible because of the linearity of the Fourier-transform in a example of the effect of aperture blocking caused by the feed 7 paraboloid-reflector antenna is shown in Fig. For p = the gain is reduced 72.4. The -10 -2 (.. Aperture Blocking. An obstacle in front of the aperture can cause an unavoidable blocking or shadowing and alter the effective aperture distribution.5A/Z). The first sidelobe is which has been considered in the past 1 is [1 — (rjr f] v . 2.6). 7. the distribution 1 . 2 = 2n{r /\) sin Fig. but at tapering the distribution of a linear aperture. 7. distribution can be found in Ref. Radiation pattern for a uniformly illuminated circular aperture. antenna the expense of broader beamwidth and less db below the main The effect of tapering lobe. p F1 When/? radiation pattern is of the form /„ hl (f )/£ and the radiation pattern reduces to that given above. is uniform 75 per cent.4.6 db down for/? = 2. 7. One aperture distribution gain.3] Antennas is 269 the field intensity £(<£) proportional to P'dfl = pexp (-27r-sin^cosfl)rdr = irrg2J 1 (f)/f (7. the half-power beamwidth is below the maximum.Sec.2. .5.

" Proc. is -20 -5 5 grees off axis Fig. is widely used beams are desired. An asymmetrical beam shape can be ob- tained by using only a part of the paraboloid. 7.3 the parabola and directed toward the reflector surface. 1. 1. or a paraboloid. The beamwidth in the plane containing the linear feed is determined by the illumination of the line source. Therefore a point source of energy located at the focus converted into a plane wavefront of uniform phase. When properly illuminated by a point pattern.11) is generated to by moving the parabolic contour parallel itself. although it is practical to use the parabolic cylinder for aspect ratios of this magnitude or larger. . It is not practical to use a paraboloidal reflector with a single horn feed for aspect ratios greater than about 8:1. Cutler. 7.19 to 5. 1 5 9 The para> > bolic cylinder (Fig. while the beamwidth in the perpendicular plane is determined by the illumination across the parabolic profile. Rotating the parabolic curve shown in Fig.270 Introduction to Radar Systems [Sec. rather than a point source.5. Fig. Parabolic-reflector antenna. the paraboloid generates a nearly symmetrical pencil-beam-antenna Its chief application has been for tracking-radar antennas. IRE. must be used to feed the parabolic cylinder. Examples of the paraboloid are shown in Figs. 7. Effect of aperture blocking caused by the feed in a parabolic-reflector antenna. 5. One of the advantages of the parabolic cylinder is that it can readily generate an asymmetrical fan beam with a much larger aspect ratio (length to width) than can a section of a paraboloid. source at the focus.21. A line source such as a linear array.6.6 about its axis produces a parabola of revolution called a circular parabola. This type of antenna. The reflector is made longer than the linear feed to avoid spillover and diffraction effects.) (From C. an example of which when Vertex or apex shown in Fig.6. The parabola is well suited for microwave antennas because (1) any ray from the focus is reflected in a direction parallel to the axis of the parabola and (2) the distance traveled by any ray from the focus to the parabola and by reflection to a plane perpendicular to the parabola axis is independent of its path. Another means of producing either a symis fan metrical or an asymmetrical antenna pattern is with the parabolic cylinder. 7. The basic parabolic contour has been used in a variety of configurations.

Scanning is accomplished in the parabolic torus by moving the feed. half of the energy radiated by the dipole would be radiated into space without striking With small paraboloids (apertures of a few square wavelengths). H however. antenna pattern is difficult to achieve in practice. There are other variations of parabolic reflectors such as cheeses. possible to phase the rearward primary energy to reinforce the secondary energy from the paraboloid reflector in order that the rearward energy contribute to the gain of the antenna and not be the reflector is lost. a perfectly symmetrical different. It is minimized with a shallow reflector. were generally fed by the simple half-wave-dipole First.3] Still Antennas 271 another variation of the parabola is the parabolic torus shown in Fig. 7. as well as the current distributions in these two planes.4. The which efficiency of the simple dipole feed can be increased with a more elaborate feed most of the energy radiated by the feed in the direction of the reflector. a half cylinder. for large antennas most of the energy not striking wasted. 7. for feeding an asymmetrical section of a paraboloid that generates a fan beam plane than in the E plane. - called the primary pattern. Most of the energy is directed in the forward direction. the dipole radiates feed for a paraboloid. In an ideal feed all the energy reflected polarized in the same direction. is its poor polarization from the paraboloid surface is the energy polarized at some other direc- depends upon the shape of the dish. the radiation pattern of the aperture when illuminated by the feed is called the secondary pattern. descriptions of which may be found in the literature. that is. However. or a hemisphere. open-ended waveguide operating in the TE U mode. The waveguide horn is probably the most popular method of feeding a paraboloid for radar application. wider in the When more directivity is required than can be obtained with a simple open-ended waveguide. The rectangular guide may be used. The radiation pattern produced by the feed is feed be independent of the angle. suffers from two major limitations. a plane sheet.17 and discussed in Sec. consists of a point source paraboloid Feeds for Paraboloids }^ The ideal feed for a aperture distridesired the achieve shape to proper of pattern of illumination with a emitted by the It is important in a paraboloid that the phase of the radiation bution. if radiating in the proper mode. It is generated by moving the parabolic contour over an arc of a It is useful where a scan angle less circle whose center is on the axis of the parabola. A rectangular guide operating in the TE 10 mode does not give a circularly symmetric radiation pattern since the dimenplanes. If the paraboloid reflector subtends a solid angle of 180° at the focus (a rather large angle). as a The element. This is accomplished by a parasitically excited reflector element placed behind the dipole to reflect energy toward the paraboloid. are sions in the E and As this is generally true of most waveguide feeds. 7. A circular paraboloid might be fed by a circular. primary of the to that reduction of the antenna gain and results in the generation of sidelobes with polarizaThe extent of the cross-polarized energy tion orthogonal to the primary polarization. than 1 20° is required and where it is not convenient to scan the reflector itself.wave dipole. tion is wasted because the antenna might not be designed to respond to a different The dipole feed causes some of the reflected energy to be perpendicular polarization. and the phase characteristic is usually good. pillboxes. If it is not. H . one with a large ratio of focal length to diameter. A better feed than the half-wave dipole is the open-ended waveguide. some form of waveguide horn may be used. The parasitic reflector can be another directs dipole. The early paraboloids Practical feeds for paraboloids only approximate the ideal.Sec. and 1 5 hoghorns. The second shortcoming of the dipole as a paraboloid feed characteristic. half. uniformly in a plane perpendicular to its length and radiates no energy in the direction of its length. This cross-polarized radiation causes an effective radiation. The resulting radiation pattern is therefore doughnut-shaped. it is the reflector.

7. Therefore. However. Also. Cutler. The less the spillover. (From C. more of the radiation from the feed will be intercepted by the reflector.F-nI i i i -5 -10 - / / / 1 \ i i - -15 - \ i\ 80 - ?o -100 i \ 1 1 1 -80 -60 -40 -20 6 degrees 20 off axis 40 60 100 Fig. the higher the efficiency.84A-diameter circular waveguide is shown in Fig. the illumination of the aperture can be determined and the resulting secondary beam pattern can be found by evaluating a Fourier integral or performing a numerical calculation. 7. On the other hand. causing a reduction in the aperture efficiency. a significant portion of the energy radiated by the feed would not intercept the paraboloid used. If one wished to obtain 1 relatively uniform illumination across a paraboloid 1 1 1^. 7. if the angle subtended by the paraboloid at the focus is large. IRE.'' Proc. {From C. 7.272 Introduction to Radar Systems [Sec. IRE. The radiation pattern of a 0. The aperture distribution at the edges will be even less than one-tenth the maximum because of the longer path from the feed to the edge of the reflector than from the feed to the center of the dish. When the primary feed pattern is 10 db down at the edges. the illumination is more tapered. Efficiency of a paraboloid as a is not critical.) aperture with a feed of this type.8 for the circular-waveguide feed whose pattern is efficiency shown in Fig. only a small angular portion of the pattern should be An antenna with a large ratio of focal distance to antenna diameter would be necessary to achieve a relatively uniform illumination across the aperture. so that the optimum angle subtended by the antenna at the focus Fig. there will be some angle at which these two counteracting effects result in maximum efficiency. 20 40 60 80 100 Half angle subtended by paraboloid at focus . 7. The maximum of the curve is relatively broad. Radiation pattern of 0. and would be 100 lost.84A-diameter circular-waveguide aperture.) from the feed in the direction of the edges is between 8 and 12 db below that at the center. 7.3 Optimum Feed Illumination Angle.7. This is illustrated in Fig.7. 7. Cutler. the intensity of the energy radiated toward the edge of the reflector should usually be about one-tenth the maximum intensity. If the radiation pattern of the feed is known. As a rough rule of thumb. 7 Proc. The lost "spillover" energy results in a lowering of the over-all and defeats the purpose of the uniform illumination (maximum aperture efficiency).8. The greatest efficiency is obfunction of the half angle subtended by the tained with a reflector in which the radiation paraboloid at the focus. the first minor lobe in the secondary pattern is in the vicinity of 22 to 25 db.

the in mismatch impedance blocking and and alter the effective porting structure intercept a portion of the radiated energy enters the feed and paraboloid the by reflected Some of the energy antenna pattern. so that there apex-matching raised portion of the reflector is called an matching than bandwidth broader the apex-matching plate has a transmission line. transmission of length minimum of compactness and utilizes a an example The antenna may also be fed in the manner shown in Fig. dipole.9 Fig. 10.9. frequency narrow of the reflecting surface at radiation intercepted by the feed is to raise a portion is made of such a size and surface raised The paraboloid. an impedance-matching the effect of the reflected reducing for technique Another band. and antenna mismatch reduce the efficiency to the order of 55 to 65 per cent for ordinary paraboloidal-reflector antennas.3] Antennas 273 aperture distribution set up by the Calculations of the antenna efficiency based on the efficiencies of about 80 per primary pattern as well as the spillover indicate theoretical uniformly illuminated ideal. in portion shown by the solid curve < The two reflected signals cancel at the feed. and 7. of the focus the at placed is removed. This is aperture. an with compared cent for paraboloidal antennas when polarization characpoor aperture. 1. The feed. Cutler is the 7. of a front feed.9b produces an in shown feed are examples of rear feeds. (a) Rear feed using half-wave Fig. Although devices inside the creases the minor-lobe level of the radiation pattern. direction in the transmission line. plate reflector. the across variations In practice. 7. Examples of the placement of the feeds in (c) front feed using horn. center the in not is line asymmetrical pattern since the transmission 7 rear feed in which the dual-aperture a feed. The rear feed has the line.9a Fig. and suptransmission feed. leaving that The major portion of the lower half of the parabola is axis the feed is out purposes practical all For 7. It is well suited for supporting horn aperture are above mentioned configurations feed the Two basic limitations to any of line. in shown as paraboloid the arranged to feed Fig. {b) rear feed using horn .9c. 7. the obstructs it but feeds. 7. 7.10. nated with the offset-feed parabolic antenna shown in is tipped with respect to the parabola s horn the but parabola. feed not shown in bend 1 80 at the end ot waveguide is in the center of the dish and the energy is made to A ^=^ [a] (b) (c) parabolic reflectors. phase aperture teristics. it causes a slight reduction in the gain and in- The . waveguide horn can be Feed Support. Fig. 1 1 Both the aperture blocking and the mismatch at the feed are elimiOffset Feed The center of the feed Fig. over a relatively only effective is remedy this but device. The resonant half-wave dipole and the These two arrangements b. the center (apex) of the at the focus a reflected produce to as contour reflector original the from distance reflected from the signal the to signal equal in amplitude but opposite in phase remainder of the is no mismatch. The waveguide rear rear of the dish.Sec. acts as any other wave traveling in the reverse impedance mismatch and Standing waves are produced along the line. causing an can be corrected by mismatch The a degradation of the transmitter performance. the waveguide advantage by a properly designed reflecting plate.

V8" V.0 3.0 1. it introduces problems of its own. f/D Ratio. ratio is based on both mechanical and electrical considerations. 1 Also. {After Ricardi and Devane.0 5." V2 V. the narrower must be the primary-pattern beamwidth and the larger _ 1 1 1 1 1 1 I 1—+- - 100 "i i i i 80 60 1 80 —— i "i — 60 B / / ^ 40 - 40 J / d/ / c/ / - 20 10 — r <^£. A small f/D ratio requires a deep-dish reflector.0 11. dimensions given in inches.0 7. 1 1 % open 92. so that there is no pattern deterioration due to aperture blocking nor is there any significant amount of energy intercepted by the feed to It produce an impedance mismatch. 7.8 - B C V. which may seriously deteriorate the radar system performance. Per cent transmission through aluminum mesh." courtesy Electronic Industries. The farther from the reflector the feed is placed." £mesh <0.11. or f/D ratio. An important design parameter for reflector antennas is the ratio of the focal length / to the antenna Fig. Vs. L 9. while a large f/D ratio requires a shallow reflector.10.3 1 8 6 - \) 4 '-I Somple 1 a 2 V? IV.1% 0.0 Frequency. 1 1 1 J 3. Gc (a) Frequency Kb) Gc Fig.0 9.0 1 I l i J .0 7. (6) polarization parallel to long dimension. but the feed must be supported farther from the reflector. 7.4 83. Table in («) applies to both figures. 1 b IV.8 1 1 Z L / A V« 90.0 11. parabola eliminates two of the major limitations of rear or front feeds.r n [Sec. Parabolic reflector diameter D. '/« Mesh IVz" T Vs.) .0 5. The shallow reflector is easier to support and move mechanically since its center of gravity is closer to the vertex. Vs. it is usually more difficult to properly support and to scan an offset-feed antenna than a circular paraboloid with rear axis is and The offset feed. 7. (a) Polarization perpendicular to long dimension.3 274 Introduction to Radar Systems of the path of the reflected energy. However.0 trons ' D I 1 1 V. should be noted that the antenna aperture of an its offset reflector) is the area projected parabola (or any parabolic on a plane perpendicular to Parabola not the surface area. Cross-polarization lobes are produced by the offset geometry. The selection of the proper///) with offset feed.1 1.5 87.

if the ice were to close the holes of a mesh antenna so that a solid rather than an open surface is presented to the wind. (After Para12 On the monov. light resistance. Ice adds to the weight of the antenna and makes it more difficult to rotate.1 0. age transmission ceases to increase and starts to decrease . 12 Radiotechnikq.1 la and b apply are given by table in Fig. spacing cm. wind nonsolid surface such as a mesh offers low reflector shaped variously to conform to ability the and of fabrication and assembly. to be stronger. 11 The transmission of linear-polarized plane waves at normal incidence to a plane sheet of expanded aluminum with a diamond mesh is o C ?60 E (/i shown in Fig.12. 7.0 or have///) beams monopulse-tracking Antennas used to generate beams with ordinary multiple the of level crossover proper the obtain more in order to waveguide it form. which the for screens sample the dimensions of the results of Fig. metal grating. On the other hand.1 la for polarization perpendicufor lar to the long dimension and in Fig. other face to increase. Most parabolic-reflector antennas seem to have//Z) ratios ratios of 0.116 §50 _ 3S The polarization parallel to the long dimension.5 0. The spacing the on incident wires is equivalent to a shortening of the wavelength of the surcoefficient transmission the causing between wires appears wider.11a. 40 30 20 10 - The presence of ice on the reflector surface is an important consideration for both the electrical and the mechanical design of the antenna. one hand. the result that both the backlobe of the antenna 100 and the relative intensity of the sidelobes adjacent I The reflecting surface may be made of a solid sheet material. it is difficult to reflector with small fjD a illuminate properly to necessary angle over the wide ranging from 0. mesh. With a properties of ice. spacing between wires 0. I I 1 to the main beam will increase and the antenna 90 - gain will decrease. 7. 7. . but Reflector Surfaces. cm 0.3] Antennas 275 obtain a feed with uniform phase must be the feed. In unfavorable mission. surfaces. (a) Wire diameter band (A 0. a grid of parallel through Two examples of transmission properties almost completely. In addition. (6) wire diameter X = = = = A (not shown in the figure).2 cm). 7. .) of a mesh reflecting surface is twofold. screen.4 0. with 10 11 may surface nonsolid a However.5.Sec. The leakage through several types of mesh screens has been measured by Ricardi and 80 70 (a) (b) Devane. or expanded is often preferable to use a wire feeds.3 0.12. 7. wires coated with ice are shown in Fig. - 0. is increased by the pressurface reflecting total the hand. ease low weight. Per cent transmission through a grid of parallel wires coated with ice at 3.2 Radius oi ice.02 cm. The relative importance of ence of ice reducing the transmission through the or a net decrease in transincrease net these two effects determines whether there is a can lose their reflecting meshes reflecting strongly even cases. dielectric the of result as a significantly mission can increase properties dominate and the percentfurther increase in the amount of ice. . A The expanded metal mesh made from aluminum is a popular metal mesh cost. On the mesh.1 cm. electrically. bigger motors would be needed to operThe structure also would have ate the antenna.6 The effect of ice on the electrical characteristics between wires = 1 cm. the reflecting Fig. 7. transpercentage the that indicate These 7. perforated metal.5 to 1.5 0.3 to 0. permit energy to leak through. ice which fills a part of the space bedielectric around the tween the mesh conductors may be considered a dielectric.

7. Reflector— —Stacked -beam feedhorn Turntable Fig. inside a radome.3 A r-inflated tube (lOpsi) —Fabric paraboloid Metolized fabric paraboloid \ \ // 1 1 // / / / / %" water internal -Feed horn 1 \ I Feed / 1 [ pressure i Feed-horn support-^ Inner tower Inner tower Fig.) AN/TPS-27.276 Introduction to Radar Systems [Sec. (Courtesy Westinghouse Electric Corporation.13.14. for the radar. Outline of the Paraballoon Electronics Division. Electronics Division. 3-D tactical . Cutaway view of a Paraballoon antenna. 7. 7.) antenna. (Courtesy Westinghouse Electric Corporation.

4] Antennas to icing conditions can be protected 277 Antennas which might be exposed by enclosure in radomes. 7. as one might positioning the often be used other than the brute-force technique of mechanically Phased array antennas and lens antennas offer the possibility of entire structure.4. 488). limited The beam produced by a simple paraboloid reflector can be scanned over a 1 17 18 too far scanned be cannot beam the However. If the paraboloid reflector is replaced by .000 about 4 hr.13).25 with paraboloid of a gain fjD The astigmatism.14). The plastic ding metallic silver particles inside the plastic coating A A RF energy. material such as vinyl-coated material are joined at the paraboloid-shaped halves. beam 30-ft-diameter oblate Paraballoon is used with the AN/TPS-27 stacked the in used is Paraballoon 50-ft-diameter a (3-D) tactical radar (Fig.50 can be scanned fjD ' - = D= A = changes with a 80 per cent of maximum (Ref. It is claimed the on Repeated inflation and deflation cycles have little effect as conventional metal tolerance a good as to that the contour can be maintained reflectors. 1316 parabolic reflector ParabaUoons somewhat different from that described inflatable antenna made from a plastic an previously is the Paraballoon (Fig. scanning the beam over a limited angle with a fixed reflector and a movable entire antenna position the is to it than feed the position mechanically much easier to manuIn addition. of increasing coma and is reduced to 80 per cent of its maximum diameter) antenna distance. forward for designed radar search long-range AN/TPS-22. feed. material of the Paraballoon has little effect on the inflated plastic The whole Paraballoon structure is enclosed within a protective air pressure surrounding the above kept are Both the radome and the antenna radome inflate the Pararequired to pressure small relatively the by blowers. as described in Sec. the positioning angle by pattern because without encountering serious deterioration of the antenna radiation 0. but is antenna pattern after scanning but a few beamwidths off axis. and total The tactical areas. without the necessity for moving large mechanical masses. The complete radar can be erected and in operation in 7. The advantage claimed for this antenna small size when deflated. A installed weight is approximately 9. during initial fabrication into two contoured fiberglass. on the fabric. . Because of 30-ftor puncture. and weight light of its because erected easily and portable contour.02 psi is sufficient rims and inflated. 30-ft-diameter factory operation of a attached to the inside aluminum vapor-deposited with Mylar of sheet the reflector a reflecting surface may also be made by imbedsurface of one of the paraboloids. 7. focal (/= paraboloid with value when the beam is scanned ±3 beamwidths off axis. balloon its operation is not affected by moderate leakage size the holes punctured by 50 if even can operate satisfactorily A diameter Paraballoon is that it is readily transof 20-mm shells. 7.12. 1 p. Hence scanning a simple in the deterioration of the because angle in limited generally feed is possible. gain is reduced to the before axis off beamwidths ±6. facture than antennas which must be moved about. The antenna impedance also scanning the antenna by paraboloid change in feed position. are discussed It is feed. These beam scanning the The present section considers the possibility of later in this chapter. 7.Sec. a radome for the complete antenna system of the AN/TPS-27 including sometimes better. Scanning-feed Reflector Antennas flexibility Large antennas are sometimes difficult to scan mechanically with as much must antenna large beam of a the scanning for technique Some like. The two paraboloids of plastic A for satispressure differential of as little as 0. lb. a spherical-reflector Spherical Reflectors. large fixed reflectors are usually cheaper and easier to structure.5 0. used as ParabaUoons early The Paraballoon.

and the pattern is generally poor. The beamwidth required of the primary feed pattern is determined by the illuminated portion of the aperture. A total useful scan angle of 140° was demonstrated. Stepped parabolic reflector.4 db gain) with a relative sidelobe level of 20 db. If the phase error from the sphere is to differ from that of a paraboloid by no more than A/ 16. a simple spherical reflector does not produce an equiphase radiation pattern (plane wave). 7 15 8. Li used a square-aperture horn with diagonal polarization in order to obtain the required primary beamwidth and low-primary-pattern sidelobes (better than 25 db).8° (39.16. 1 2 Gc The was 29. making a family of confocal parabopossible to scan the stepped reflector to slightly wider angles than a Generoting parabola Center of sphere Fig. much wider scan angles are possible than if the entire aperture were illuminated. 7.15. The resulting secondary beamwidth from the sphere was about 1. BMEWS . it is 278 Introduction to Radar Systems to is to step a parabolic reflector as shown in Fig. Principle of the parabolic-torus antenna. the maximum permissible diameter of the illuminated surface should be 3. If only a portion of the spherical reflector is illuminated at any one time.56 ft. Li 21 has described experiments using a 10-ft-diameter spherical reflector at a frequency of 1 focal length respects to the torus antenna described below. 7.23 The reduced in half-wavelength steps. It is is The second approach compensate for the spherical aberration with special feeds or correcting lenses These techniques yield only slightly larger scan angles than the single paraboloid reflector with movable feed. Wide scan angles can be obtained with a parabolic-torus conJiguration. However.5 in. A third technique to approximate the spherical surface and minimize the effects of is 7 4 [ Sec possible to achieve a wide scanning angle because of the symmetry of the sphere. The term spherical aberration is used to describe the fact that the phase front of the wave radiated by a spherical reflector is not plane as it is with a wave radiated by an ideal parabolic reflector There are at least three techniques which might be used to minimize the effect of spherical aberration. The principle of the parabolic-torus antenna is shown in Fig 7 16 and a photograph of an actual torus antenna used in (Ballistic Missile Early similar in many Parabolic Torus.spherical aberration focal length loids. One is to employ a reflector of sufficiently large radius so that the portion ot the sphere is a reasonable approximation " to a paraboloid. This type of antenna is simple paraboloid.22. Fig. but not as wide as with some other scanning techniques Disadvantages of this reflector are the scattered radiation from the stepped portions and the narrow bandwidth. 19 21 surface.

is not perfectly plane.) the of part is figure of left (Crane to Warning System (BMEWS). {Courtesy General Ekctnc This antenna is 165 ft high and 400 ft wide and uses 1 Heavy Military Electronics Department) torus The wave reflected from the surface of the parabolic of the ratio^of focal lengthy choice proper wave by plane it can be made to approach a between 0.c Miss le^ Fig. The parabolic torus is in the pattern. section of a parabolic arc about an axis in Fig.17. while the The cross section in one plane (the vertical plane beam angle may be scanned by The circular. of the torus itself. 7.45 lies The optimum ratio of///? to the radius of the torus R.500 tons of steel..A Antennas Sec. is cross section in the orthogonal plane half the radius of the torus approximately is radius whose moving the feed along a circle portion of the circular The radius of the torus is made large enough so that the circle from the surface of a appreciably differ not cross 'section illuminated by the feed will surface in the horireflector the of symmetry circular Because of the true parabola.43^and. Early surveillance radar of the Balhst . 7. Parabolic-torus antenna used in the erect ion eqinp ment.16) is parabolic.17. 7. . not the an enna. 7. Co. scanned in this plane without any deterioration zontal plane.0. but principal planes with sidelobes only radiation patterns are possible in the paraboloid. The sidelobes produced by the parabolic torus do surface due to its deviation from a true inherent phase errors of the parabolic-torus Good D .) The highest diameter of the illuminated area rather than the not lie within the principal planes. The larger the ratio of// slightly worse than those of a conventional in the parabolic torus is the diamete (The diameter the better the radiation pattern.4] 279 Warning System) generated by rotating a parallel to the latus rectum of the parabola. the beam can be readily is shown in Fig.

18. In principle the parabolic torus can be scanned 180°. the cost of the fixed reflector of the parabolic torus is relatively cheap compared with antennas which must be mechanically scanned. the maximum scan angle is usually limited to the vicinity of 120°. Principle of the organ-pipe scanner. Nonutilization of the entire aperture is probably not too important a consideration when over-all cost and feasibility are taken into account. This is the spherical reflector described above. reflector is Only a portion of the parabolic-torus particular time. consequently.4 parabola can cause sidelobes on the order of 15 db in intermediate planes. Hence scanning is possible in both planes. a torus with an elliptical cross section should result in less phase error and lower sidelobes than a torus with a parabolic cross section. 21 The sphere as an antenna is usually less effective (higher sidelobes. 25 Experimental measurements. 25 These sidelobes usually lie in the 45° plane and are called eyes. because of their characteristic appearance on a contour plot of the radiation pattern. However. Theoretically. had by moving the A limited amount of beam scanning in the plane of the parabolic cross section can be feed.280 Introduction to Radar Systems [Sec. 7. this plane can be obtained by substituting a circular cross section for the parabolic cross section so that the resulting reflector is a portion of a sphere. there is little basis for choosing between a torus with a parabolic or an elliptical cross section. illuminated by the feed at any This may appear to result in low aperture utilization or poor efficiency since the total physical area is not related in a simple manner to the gain as it is in a fully illuminated antenna. . just as in the conventional paraboloid. 7. however. but because of beam spillover near the end of the scan and self-blocking by the opposite edge of the reflector. do not show a great difference in performance between the two. Equal lengths of Wider scan angles in • ^wavequide Input horn Fig. lower gain) than either the parabolic or the elliptical torus.

or group of feeds in turn. Fig. but little until the beam reaches one end of the scanning secondary the found in are beams two and aperture energy appears at both ends of the The antenna cannot be used during this period of ambiguity.19. 7. For example.19. Naval Research Laboratory.Antennas Sec. practical in small-size of the rotating arranging a series of feeds on the locus of the transmitter power from one feed to the next the switching focal points of the torus and 27 " 29 of the organ-pipe scanner is shown in principle The scanner. 7. Thirty-six-horn organ-pipe scanner. The rotary horn may be flared to feed each to the transmitter All the transmission illuminate more than one elementary feed of the row of feeds. moving a single feed or by switching the transmitter the torus on an arm of length of center the about rotated single moving feed may be 120° torus antenna approximately one-half the radius of the torus. a A Fig. 7.4] 281 economical method for the parabolic torus is that it provides an large antenna aperture over a relatively wide rapidly scanning the beam of a physically Its disadvantages this angle of scan.S. called the pattern. In a model of the organ-pipe scanner shown in Scanning arm is large. 7.18. equal of be lines in the organ-pipe scanner must changes The radiation pattern from a torus with a well-designed organ-pipe scanner the point this At aperture. intermediate in obtained the large sidelobes parabolic torus is accomplished by Organ-pipe Scanner. scan angle with no deterioration of the pattern over for scanning and means other with compared when are its relatively large physical size planes. of a wheel so that one feed is always illuminating the if the radius problem mechanical difficult a becomes it antennas. Scanning the beam in the The advantage of between many fixed feeds. {Courtesy U.36 elements were deadtime. transferring power from periphery of a circle. feed horn is rotated within this circle..) spokes might be scanned by continuously rotating three feeds spaced 120° apart on the be may this Although reflector. may also be accomplished by A . organ-pipe with an The transmission lines from the feeds are arranged to terminate on the Fig. length.

7.^k beam 360 in one plane (Fig.™ ° IRE WESCON "^"t-own it faces. of 45° with the polarization of the feed is also at 45°. 90% reflection for A component 45° parallel wires Fig.20. This arrangement of the grid wires produces a barber-pole effect. is shown in Fig. a contemporary ot Jsaac Newton.) A (Courtesy Barab. Conv. The antenna consists of a parabolic torus extending f - .20). Its application to microwave reflector antennas permits a reduction in the axial dimension of the antenna. The Cassegrain principle is widely used in telescope design to obtain high magnification with a physically short telescope and allow a convenient The Cassegrain antenna rear The principle of the Cassegrain antenna at the vertex of the parabolic reflector. 7. 360° parabolic-torus antenna. Maraneoni. . the surface appears transparent and the energy passes through relatively unimpeded. Aperture blocking can be minimized by designing the parabolic portion of the torus as an offset parabola ' p just as in the case of a paraboloid. just as in optics. but in the parabolic torus they would he on the arc of a circle.19 the feeds are shown on a straight line. The feed is located and a subreflector is located in front of the parabola between the vertex and the focus.5. 30 " 32 36 The P rinci P'e of the parabolic torus may be applied to scanning a °°. 360° in azimuth and made up of parallel wire elements forming an angle The reflected. 7.282 f th Introduction to Radar Systems r SEC 7 5 a time 2? The deadtime for ^is model is equivalent to rotation past two f:u [felements. It also permits greater flexibility in the design of the feed system and eliminates the need for lone 6 transmission lines. 7. Parallel rays coming from a target (at is an adaptation to the microwave region of an optical technique invented in the seventeenth century by William Cassegrain.21a. consequently ot the 36 it was inoperative about 6 per cent of the time in Figs.18 and 7. Only linear polarization is possible with r this antenna. The many feed horns plus all the transmission lines of the organ-pipe scanner result in a relatively large structure with significant aperture blocking. Record. causing it to be Since the polarization of the reflected wave makes an angle of 90° with the grid wires on the opposite side of the structure. Cassegrain Antenna location for the observer. and Scott. The 45° tilt of the grid wires causes the wires on one side of the antenna to be perpendicular to those on the opposite side Radiation from the 45° feed is parallel to the grid wires which vertical meridians. 7.I°o'".

7. main distance from the since it removes a portion compromise. a. \M(e is (e from the subdistance the by divided focus real distance from the subreflector to the greater than always The eccentricity of a hyperbola is reflector to the virtual focus.5] infinity) are reflected 283 e the positio n of the feed. the larger used as the subreflector. The feed also contributes to aperture blocking antenna that was Cassegrain simple of a example Figure 7. (a) Cassegrain antenna showing the hyperbolic antenna.. The However. paIaboirat7andthefeedis atr. a secon set of incident on ^hyperbola will be at waves with foci at F and * hyperbola Any at F'. the locus of the hyperbola are the straight lhe chosen. the k sometimes placed forward of the reflector nearer to The point** is.) is point which matter no difference between focal radii is a constant. foci divided conjugate two the between unity and is defined as the ratio of the distance radii of a point on focal (The radii. but subreflector size must represent a of choice the Thus reflector. assembly. converging at the of toe point ot focal the at image it appears as a virtual reflector images the feed so that beam and are by the parabola as a convergent "«£«?£dbyfl£ ^ reflector ^geometry of the Cassegrain 8 is shown in Fig. Cassegrain of geometry (b) reflector. the main parabolic satisfies this surfaces which could be property. the of subreflector at the focus is also equal to the Magnification 1) where e is the eccentricity. A + - . Th hyperbolic subreflector. a large subreflector results in large aperture greater it has to be supported at a small subreflector reduces aperture blocking. be may which blocking. 7.21ft.22 is an of the energy Telephone Laboratories.hown at the ^Ws^nt 1 he points t The focus of the subreflector. reflector ' /\ Parabola [a) ib) at the vertex of subreflector and the feed Fig 7 21. designed by Wheeler Laboratories for the Bell be explained by considering the action may antenna Cassegrain of the principle The images the feed to a point behind the which mirror of the subreflector as a hyperbolic of the hyperbolic mirror magnification The parabola. reflect. convergent spherical waves centered cteS Parabolic Xd x . of the antenna reflector and the shorter will be the axial dimension undesirable.Antennas Sec. focal two the by the constant difference between The lines which join the point to the foci. LdTZtCconilte Convergent spherical foci of the hyperbolic subreflector. and there exists a family of hyperbolic it will be to the main nearer the subreflector.

(m where /= focal length diameter of antenna wavelength ^ (7. for simplicity. and the°u beam-deviation proportionality factor is taken to be 9 The fatter factor accounts for the deviation from Snell's law of reflection when the reflLtor a ° f a flat P. e (Ref In '' P" 488 >^is is a functln of h ^ } ratio. t^fn^u ? ' E f ' ^y . but butfofthe for the purpose of this . F' times application of the Cassegrain antenna Assume. Wheeler Labora- generated by each of the feeds are to overlap in space at their half-power ? points the spacing 5 between the effective phase centers of the two feeds must be i . example it is assumed to be constant. a monopulse radar with a conventionalfparabohc reflector fed by a two-feed-horn assembly tracking in a If the as a monopulse-trackine-radar wo beams r^'Sc EXamP e ° f 3 SimP e ' ' CaSSegraln antEnna - .284 Introduction to Radar Systems of the Cassegrain is effective focal length liS^S^iSL: One antenna^ ^ M ^" equal to the distance between ifi — is [Sec 7 5 Fand ********** JfJfS single plane.20) D= A = tWs rmU a k is assumed that the half-power beamwidth is given by 6SX/D.) (C°""^ °f P - W - »~.

conventional parabola because of the magnifying action can be reduced by a antenna paraboloid length of a It has been claimed that the over-all 33 configuration. Dielectricfrequency. to be suspended out at the problem. the at directly placed can be receivers low-noise the In the Cassegrain antenna require the receivers.20) gives a limitation on the minimum Eq. a Cassegrain antenna configuration using a at feeds The 1 of . would antenna paraboloid conventional the with same the To do focus in front of the or at least the front ends of the receivers. . effective///) an in results a hyperboloid with a magnification of 3 at the focus of a those than larger be must antenna Cassegrain the real focus F' of the of the hyperbolic reflector. _ handling capability. Open-ended waveguides may be « . The loss in the transmission line can significantly degrade the sensitivity of feed horn. For the first sidelobe is increased by (2DJD) -20-db sidelobe would be and a db 0. = the monopulse angle measurement. 34 36 - It can also be scanned by moving one of the antenna subreflector in front of the main reflector in the Cassegrain removed. use of a parabola ot The magnifying property of the Cassegrain antenna permits the For parabola with a larger//!) conventional///) ratio to obtain the same effect as a and with///) 35 paraboloid example. a approxigain by (power) the reduce will obstacle in the center of the aperture mately - 2(DJD) 2 ] 2 where . The conventional in a needed focus the long runs of transmission line to will be differences in there that chance the greater is the line. within The dielectric reduces the one-half wavelength if they are filled with dielectric. surfaces. each for feed paraboloid a separate vertex or by switching single beam by mechanically moving a single feed at the of a among many feeds as with an organ-pipe scanner. is relatively far out on structure feed the 1 type antenna. Cassegrain the with factor of two when redesigned An important advantage of the Cassegrain configuration for monopulse-radar the reflector.5] Antennas 285 value of///) ratio. wavelength one were horns centers of the two feed value off/D for a reflectorlarge This is a flD ratio would have to approximate unity. The closest The///) ratio in side) by side placed one-half wavelength (open-ended waveguides reflectors but is parabolic conventional most than larger 5.20) indicates that the apart. 7._ . This not only increases increased aperture blocking.. resulting in a configuration causes aperture blocking. If the phase Eauation (7. may be reduced by decreasing the size of the subreflector. important when low-noise receivers such as masers or parametric the receiver. the gain would reflector) and D is the diameter of D is the diameter of the main aperture. antenna is The elimination of long runs of transmission lines with the Cassegrain amplifiers are used. transmission longer the hence.122. which is still this case is closer than spaced manageable proportions. errors in the phases between the lengths of the transmission lines and. avoiding the application is that the RF plumbing can be placed behind paraboloid. In an antenna with f\D the feeds becomes supporting of problem mechanical the the axis of the reflector and might be spaced is feeds the more difficult than if the///) ratio were small.3 about by lower be to 0. 34 h the obstacle (hyperbolic sub- 2 . overThe Cassegrain antenna configuration can be used to generate a multitude ot the of vertex the of vicinity the in placing by lapping beams from a single reflector scanning permits geometry Cassegrain The beam. increased to about Aperture blocking — 1 8 db. By . but it also results in design mechanical the antenna.05. Part of the energy is the parabolic reflector If sidelobes. particular a used at be can minimum size of waveguide which usually has less powerand guide air-filled than loss higher of is filled guide. The relative (voltage) level of if the factor DJD were equal example. however. (7. the in increase and an beam of the main The presence of the reduction is circular [1 circular small and assumed to have a completely tapered parabolic illumination. Sec.

However the feed cannot be made too large since it partially shadows the energy reflected by moving it closer to the subreflector. The twistreflector is equivalent to a quarter-wave plate which produces a 90° rotation of the plane of polarization (Ref. 1. 7. just as in the optical case. and Teflon are suitable for small microwave lenses. Some aperture blocking does occur. which passes waves with negligible atten- ^""poi. The microwave paraboloid reflector is analogous to an automobile headlight or to a searchlight mirror. 34 If operation with a single polarization is permissible. A One of the large weight. (2) metal-plate lenses. | Subreflector with polarization - wave radiated by the feed. Polarization-twisting Cassegrain antenna. 7. Three types of microwave lenses applicable to radar are (I) dielectric lenses. Lens Antennas dish is vertically polarized and passes through the subreflector with negligible effect.23 can considerably reduce aperture blocking. and (3) lenses with nonuniform index refraction. 12. or vice versa. a converging lens is thicker in the middle than at the outer edges. dielectric-lens antenna of Fig 7 24a is similar to the conventional optical lens. point at the focus of the lens produces a plane wave on the opposite side of the lens. Minimum (twist reflector) total aperture blocking occurs when from the main the feed size and Main reflector with polarization twister distance are such that the shadows produced by the subreflector and the feed are of equal area. the technique diagramed in Fig. 7. Aperture blocking by the subreflector is reduced with this design. however. The analogy of an optical lens is also found in radar (an example is the radar of Fig. because of the feed. 7. It is equal to the square root of the dielectric constant. solid. ( — —_* - uation but reflects the horizontally polarized ^"SJ Hor. Lens and reflector antennas are often interchangeable in microwave systems since they both convert a spherical wave to a plane wave.8). Dielectric lenses may be designed using the principles of classical geometric optics. The subreflector is transparent to vertically polarized waves and does not block the aperture. Focusing action is a result of the difference in the velocity of propagation inside the dielectric as compared with the velocity of propagation in air. the size of the parabolic reflector.6.286 Introduction to Radar Systems directive. poi.5 making the feed more subreflector may be reduced without incurring a spillover loss.24*). 2 The homogeneous.23. polystyrene Plexiglas. The index of refraction n of a dielectric is defined as the speed of light in free space to the speed of light in the dielectric medium. The horizontally polarized wave reflected by the subreflector is rotated by the twistreflector at the surface of the dish. Zoning is considerably bv based on the fact that a 360° change . but this blocking can be made small and comparable with that of an ordinary parabolic-reflector design. limitations of the solid homogeneous dielectric lens is its thick size Both the thickness and the weight may be reduced and stepping or zoning the lens (Fig. main dependent surface Sec. The subreflector consists of a horizontal grating of wires. They have low loss and may be easily shaped to the desired contour. Materials such as polyethylene. Since the velocity of propagation is greater in air than in the dielectric medium. 1. called a vertically polarized transreflector. 10). 34 The most common type of radar antenna is the parabolic reflector in one of its various forms. or [Sec. 7. of Dielectric Lenses. The wave reflected from the main Fig.

index of refraction) of a solid dielectric lens. on the order of 1 or more. the greater will be the larger the However. the the dielectric constant.24. The the lens design of the lens is again increased in the direction of the axis according to another step may until the path length in the dielectric is once more 360°. .5 quency. it electric consists of molecular particles of microscopic size. Even with these limitations. Direct micro- wave analogy of optical lens. where fx is the index of refraction. t. than the next outer zone. 7. in the dielectric path the wavelength. it is not without disadvantages. If the thickness of the path length removed from the lens is t. when the increased toward the lens axis as in the design of a normal lens. a equal to is dielectric path length introduced by the thickness can be reduced to zero without altering the phase across the aperture. the net change in the optical path length is /ut or to some integral multiple wavelength one equal to This change in path length must be of a wavelength. mismatch between the lens and free space and the greater the loss in energy due to Compromise values of the index of refraction lie reflections at the surface of the lens. the thickness of the dielectric However. be may such as polystyrene foam.6] Antennas . The larger the dielectric constant (or between in optics. thinner it will be. The effect of the steps may be minimized by using a design with large// D. The Focus v/////a Focus (*> Fig. but the artificial dielectric The particles consists of discrete metallic or dielectric particles of macroscopic size. The ordinary diconstruct to possible is antennas. Lens reflections may also be reduced with transition surfaces as These surfaces should be a quarter wave thick and have a dielectric constant which is the square root of the dielectric constant of the lens material. 2 Instead of using ordinary dielectric materials for lens Artificial Dielectrics. however. caused by the shadowing produced by the steps. at which time wavelength less one zones is the of each through length path optical be made. zoning results in a frequencyAnother limitation is the loss in energy and increase in sidelobe level sensitive device. spacing between particles should be small compared well as the as field electric to the with a wavelength. If these conditions are met. Although zoning reduces the size and weight of a lens. and 1.Sec. The particles are arranged in some particular configuration The dimension of the particles in the direction parallel in a three-dimensional lattice. 287 Starting with distribution of phase at the aperture has no effect on the aperture phase is progressively zero thickness at the edge of the lens. (a) Converging-lens antenna constructed of homogeneous solid dielectric.™-* them of artificial dielectrics. 7.6. constant dielectric low material of in a imbedded rods strips. the lens will be insensitive to fre1 . a stepped lens is usually to be preferred because of the significant reduction — in weight. (b) Zoned dielectric lens. Dielectric lenses are normally wideband. spheres. or disks.

7. and the direction of the rays through the lens is governed by the usual optical laws involving the index of refraction. A dielectric thinner at the edges. The spacing s between the plates of the metal-plate lens must lie between A/2 and X if only the dominant mode is to be propagated. 5 For example.**. Snell's law is obeyed in an £-plane lens. For of the antenna this reason. 7.21) assuming no interaction between the spheres. the surface closest to the feed is an ellipsoid of revolution if the surface at the opposite face of the lens is plane. and end views of a converging lens antenna constructed from parallel-plate v waveguide. The phase velocity in parallel-plate waveguide is greater than that in free space hence the index of refraction is less than unity.25. as opposed to a converging dielectric lenses . The index of refraction for this type of metal-plate lens is is lens which V 1 -<3\ (7. Lenses size made from artificial dielectrics are generally of less weight than those from solid dielectrics.25. 1 when the Artificial- may be designed in the same manner as other dielectric lenses. in general. The metal-plate lens shown in Fig. converging metal-plate lens is therefore thinner at the center than at the edges. elevation. 41 An artificial dielectric may also be constructed by using a solid dielectric material with a controlled pattern of voids. but the latter are easier to machine. at the lower radar frequencies. This is a form of Babinet inverse of the more usual ©- £ E £ . (£-plane metal-plate lens. for example.25 is an £-plane lens since the electric-field vector is parallel to the plates.288 Introduction to Radar Systems of the artificial [Sec. 43 composed of particles imbedded in a low-dielectric-constant maThe voids may be either spheres or cylinders.) artificial dielectric terial. Plan." An artificial dielectric may be constructed with parallelplate waveguides as shown in Fig.6 When the particles are metallic spheres dielectric constant of radius a and spacing s between centers the dielectric is approximately *=!+'477a 3 (7.1 L » Direction of propagation Fig. as. The surface contour of a metal-plate lens is. Metal-plate Lens. This is opposite to the usual optical refracting medium. 7. not parabolic as in the case of the reflector. 7. artificial dielectrics are often preferred is large.22) .

47 guided or The rays are Still another type of constrained lens is shown in Fig. 7.25 per cent.28.26) just as with a dielectric lens. The constrained 37 Other techniques for lens of Fig.27 is usually unsuited for radar applications.6] Antennas 289 where A is the wavelength in air. respectively (Ref. Focusing action is obtained by constraining the waves to pass between the plates in such a manner that the path length can be increased above that in free space. The closer the Howspacing. a In be completely reflected. with high sidelobes on one side of the axis. the electric field is perpendicular to the plates (H field parallel). but the lens of Fig. When this lens has a flat front surface as the direction of the plates and the lens axis. Equation (7.6 is compromise practice.wavelength lens at a wavelength of 1 . the index of reThe wave incident on the lens will fraction is zero and the waveguide is beyond value of /a between 0. shown in the figure. the thickness of the metalbecomes large unless inconveniently long focal lengths are used. 47 . and the E field is parallel The latter characteristic might cause this lens rather than perpendicular to the plates. The direction of the rays is'not affected by the refractive index. where d is the angle between in Fig. 7. -*-£" Fig. cutoff. the curved side toward the feed is a hyperboloid of revolution. reduce the gain. Another class of metal-plate lens is the constrained lens. 7. and increase the sidelobe level. Zoned metal-plate lens. For a value of s = A/2. often selected. Even with an index of refraction in the vicinity of 0. of index the ever. arbitrarily be made cannot refraction. to be classified with the £-plane lens of Fig. constrained lens.26. The bandwidth of a zoned metal-plate lens is larger than that of an unzoned lens.Sec. A disadvantage of this constrained lens is that the £-plane radiation pattern has a low gain and is very distorted. or path-length lens. the spacing. corresponding to plate spacings of 0.25. thus the velocity of the wave which propagates through the plates is relatively unaffected provided the plate spacing is greater than 1\2. and therefore small since the reflection from the interface between the lens and air will increase just as in the case of the solid-dielectric lenses. This lens differs from obeyed. 1.557A and 0. Snell's law is not and the metal plates. The thickness may be reduced by zoning (Fig. the index of refraction is equal to 0. 7.25 cm plates (normal to the electric vector). but plate lens the steps in the lens contour scatter the incident energy in undesired directions. s = X.E-plane lens employs Snell's law to achieve focusing action. constrained by the constrained lens described above since it is cylindrical.28 focuses by constraining the wave while the other .27. An example of a particular constrained lens with plates slanted at an angle d is shown The index of refraction is simply n = sec 6. At the upper limit of spacing. 7. in which In the //-plane metal-plate the rays are guided or constrained by the metal plates. 37 41 obtaining //-plane metal-plate constrained lenses are described in the literature.6.5 to 0. 7.625A and to power reflections at normal incidence of 1 1 and 6. 7. 1. 7.22) is always less than unity.866. The construction of the £-plane constrained lens is simple since it Focusing is obtained normal to the constrained consists only of rectangular plates.5 and 0. p.5 produced a 1° beam which could be scanned over a 100° sector by with an///) - A = positioning the feed. 72. the less will be the index of refraction and the thinner will be the lens. 410). and Snell's law does not apply.

(After Kock. (b) top view of lens. Mechanical errors in the lens contour contribute but once to the phase-front error.6 Focus Fig. 7.27.) Index of re- Lens Tolerances. 7. are the main advantage of a mechanical tolerances involved in the metal lens over a reflector. sec 8. .." Although there may (a) --©-- E Focus < (b) Fig. IRE.) £ field parallel to metal plates. Dashed lines represent ray paths. (a) Cylindrical constrained lens with (After Ruze.28. A Fry and Goward 5 state that "the easy plate lens . In general." Proc..290 Introduction to Radar Systems [Sec. IRE. given error in the contour of a mechanical reflector contributes twice to the error in the wavefront because of the two-way path on reflection. the mechanical tolerances for a lens antenna are less severe than for a reflector. 7. fraction = Example of a constrained metal-plate lens constructed of slanted plates. 46 Proc.

in which the index of refraction varied in some prescribed manner within the Although such lenses had interesting properties. from sample to sample or even within the same sample. Stabilization of the .) lens types.23 ) P p) 1 <- 7 24 ) - Tolerance on the index of refraction for a fully zoned lens: Dielectric lens dp = ±—(p-l) =± p P 2 > < 1 ( 7 .Sec. In many attached to a loss of gain and to a deterioration of the The ' maximum of A/8) is applications a phase variation across the wavefront of ±A/16(a tolerances and antenna acceptable engineering practice. (The relationship between patterns is described in Sec.25 ) Metal-plate lens dp — 3 (I . the mechanical tolerance of a reflector antenna must be ±A/32 if the phase-front error is to be ±A/16.p) 1 ( 7 . because of the two-way path due to reflection. source of error in lenses not found in reflector antennas is the variation in the Both real and artificial dielectrics are not perfectly uniform properties of the material. the property that a plane wave incident on the sphere is focused to a point on the surface Likewise. a transmitting point source on the surface at the diametrically opposite side. A with refraction of materials (p is the square root of the dielectric constant e). and lenses a nonuniform index of refraction are practical. Workers in the field of optics have from time to time devised lenses lens. a plane wavetront derivation of the mechanical tolerances necessary to achieve The antennas. or ±2/16 (p is the index aperture the across variation phase total The of refraction). 7. 7.29).11. The tolerances required for lens antennas are given below for several is taken to be A/8. at practical. 1)/ «a A. 7. It might also have application rapidly scanned antenna over a wide angle is required. Because of the spherical symmetry of the lens. A where the antenna is mounted on an unstable base such as a ship.27) + p) By comparison. might advantage the "main" and what consideration. of the sphere is converted to a plane wave on passing through the lens (Fig. the focusing property does not depend Luneburg lens might be used where a upon the direction of the incident wave. we get dp Since in a zoned lens (p lens metal-plate zoned s in a spacing on plate Tolerance 6s — = =± ^ 16(1 (7.26 ) 16 ±A /16r. Luneburg Lens. they were only of academic interest were not since optical materials with the required variation of index of refraction index of control the possible to it is frequencies microwave However.6] Antennas 291 lens over a reflector be room for discussion concerning the advantages of a is nevertheless an important tolerances of question the be. Tolerance on lens thickness Dielectric lens dt t: =± =± — 160 16(1 •\ Metal-plate lens dt —16 - p 1) > < 1 ( 7 . 1 texts on standard in found be may value specified accurate to a importance of degree the maximum permissible error in the phase front depends upon antenna pattern. One of the most important of the variable-index-of-refraction lenses in the field of 48 The Luneburg lens is spherically symmetric and has radar is that due to Luneburg.

28) The index of Fig.1 to 2. The diameter of this stepped-index lens is 18 in. and decreases to a value of 1 on the periphery. a maximum at the Luneburg-lens geometry showing rays from a point source radi7. 49 50 The dielectric constant of the individual shells varies from 1. ated where it equals Vl. Hemispherical-half-shell construction of stepped-index Luneburg lens.292 Introduction to Radar Systems [Sec. Inc.) .refraction is (7. The Luneburg lens can also generate a number of fixed beams and is competitive in many applications with arrayantenna beam forming.1.29. 7.0 in increments of 0. as continuous variation of dielectric constant such needed for the Luneburg-lens antenna was one of the limitations which had to be overcome by early experimenters.30) 10 concentric spherical shells are arranged one within the other. In one example of a Luneburg lens (Fig. 7. relationship j" = e* = 2. each of constant index of refraction. The index of refraction (x or the dielectric constant e varies with the radial distance in a Luneburg lens of radius r according to the . {Courtesy Emerson and Cuming. and the frequency of operation is X band.30. as a plane wave after passage through the lens. Discrete changes in index of refraction approximate a continuous variation. The beam may be scanned by positioning a single feed anywhere on the surface of the lens or by locating many feeds along the surface of the sphere and switching the radar transmitter or receiver from one horn to another as with an organ-pipe scanner.. 7. Practical threedimensional Luneburg lenses have been constructed of a large number of spherical shells. As many as 50 steps might be used in this type of design. - Fig.6 beam may be obtained by adjusting the feed to compensate for the ship's motion. The development of materials which exhibit a center.

the sidelobe level of a Luneburg lens to achieve extremely low sidelobes. The 1 8-in. 7. A proper with the full spherical lens is that the feed causes aperture blocking. or the reflecting reflect these rays in the hemisphere to plane can be extended beyond the base of the when compared hemispherical lens of a disadvantage possible direction. 7. opposite direction. the lens is included between two plane reflectors which pass through the lens center. spherical lens. Since the two-dimensional version is simpler than a threedimensional one. the size of called expandable beads. A plane reflecting is needed (2tt solid radians) the lens shown in Fig. They should They should be easily temperature. yet they must be strong enough to have low dielectric loss and support their own weight without collapsing. Jf-band lens mentioned above was constructed from a polystyrene material 49 These are discrete spheroids of polystyrene. is not as fig. a dry mixture capable of being molded is obtained. but the sidelobe level is 52 " 56 This is due to the fact that the greater. Another promising technique for the construction of large Luneburg lenses. reflector. assuming the same type of with plane reflecting surface on the base.6] Antennas 293 must not be too heavy. surface is placed at the base of the hemisphere to image the feed at S into a virtual source Movement of the source 5 causes a corresponding movement of the beam in the at S'. ing the lens A baglike radome cover- provides weather protection. that is. corner than the angle greater solid much over a reflector is effective The Luneburg principle may also be applied to a two-dimensional lens which scans a fan beam in one plane. int source/ P° tend to concentrate energy toward the edge of JX« tinq surface Thus the aperture distribution the aperture. it was the first type to be constructed. is the use of artificial dielectrics. S^'A"'' " 5 ~\by the rays in a Luneburg lens followed radiation jm„gi! f paths . 56 When the full 4tt radians of solid coverage is not required. in not be affected by the weather or by changes and isotropic if the homogeneous must be and properties manufactured with uniform The dielectric materials performance characteristics are to be independent of position. lens Luneburg of the of a paraboloid. The Luneburg-lens principle can also be applied as a passive reflector in a manner 51 If a reflecting cap is placed over a portion of the analogous to a corner reflector. The cap may be made The Luneburg to cover a sector as large as a hemisphere.3 1 can be used. j ^S 20 to 22 db.Sec.31 since a portion of the energy emitted by the source misses the The sidelobes resulting from the missed radiation may be reflector entirely and is lost. that Luneburg lens tapered as Hemispherical 7. The antenna pattern of a Luneburg lens has a slightly narrower beamwidth than that of a \ \ paraboloidal reflector of the same circular \\ A cross section. tendency for illuminated energy to concentrate at the edges of the lens makes it difficult In practice. during controlled which is a higher-dielectric-constant material in proper proportions. By mixing the partially expanded beads with manufacture.31. A geodesic analog of a . a smaller portion of the 51 57 If only hemispherical coverage lens can be used. There is some deterioration of the feed pattern in the hemispherical lens shown in Fig. seems to be in the vicinity of - 51 reduced with absorbent material to absorb the nonreflected energy. If the scanning sector is less than 277 radians. with a saving in size and weight. especially 51 at UHF. The natural feed illuminates both antennas. a smaller lens can be had by constructing only a spherical wedge of the Luneburg lens. an incident wave emerges in the same direction from which it entered. 7.

The lens acted as the feed for the reflector. Array Antennas array antenna consists of a number of individual radiating elements suitably spaced with respect to one another. The "tin-hat" geodesic analog of a two-dimensional Luneburg lens. reflections from the lens radians).7. Reflections from the lens surface which cause a significant mismatch at the feed can also be eliminated or reduced without significantly degrading the pattern by tilting the feed slightly off axis to avoid the back reflection. Although it is dangerous to from these sources in a stepped lens might be 1 or 2 db. The relative amplitude and phase of the signals applied to each of the elements are controlled to obtain the desired radiation pattern from the combined action of all the elements.294 Introduction to Radar Systems [Sec. An array consists of no less than two An . Other types of lenses based on the principle of nonuniform index of refraction have been described by Kelleher. or scattering generalize. 60 and others. such as the Luneburg. Another advantage of the lens is that mechanical and electrical tolerances are more relaxed than in the reflector antenna. 40 The lack of suitable solid or artificial dielectric materials has limited the development of lenses. Considerable equipment can be placed at the focus of the lens without interfering with the resultant antenna pattern. Conventional lenses are usually large and heavy. the ratio of the focal length /to the antenna diameter D must be made large.32. - Two^ vertical beams were generated in this radar. Theoretically. 56 Huynen. 61 A homogeneous dielectric sphere may be scanned through 4tt solid radians if the index of refraction is not too high and if the diameter is not greater than about 30A. In the mortar radar application a fan beam generated by the two-dimensional Luneburg was converted to a pencil beam by a cylindrical reflector. The problem of dissipating heat from large dielectric lenses. The lens is capable of scanning the radiated beam over a wide angle by positioning the feed. Each beam had a width of 0. 7.06° in the vertical and horizontal planes. the Luneburg lens can scan the complete sphere (4n solid Fig. surfaces. The radiating elements might be dipoles. waveguide horns. 51 59 The result is a dome-shaped parallel-plate region as shown in Fig. Solid dielectric lenses can also achieve reasonably wide scan angles by properly designing the contour of both surfaces of the lens. A large///) requires a greater mechanical structure because the feeds are bigger and must be supported farther from the lens. or any other type of antenna. Lenses which must scan by positioning of the feed should also have large fjD ratio. 62 One of the disadvantages of the lens is that it is usually less efficient than comparable reflector antennas because of dielectric losses in the materials.32. The//Z> ratio of zoned lenses might be of the order of unity or more. The antenna scanned a 40° azimuth sector at a rate of 17 scans per beam per second. 184 Evaluation of Lenses as Antennas. One of the major advantages of a lens over a reflector antenna is the absence of aperture blocking. and was separated in the vertical by an angle of 1. 7. compared with the from the steps in a zoned lens.85°.7 two-dimensional Luneburg has been applied in a ^-band mortar-location radar. additional losses 7. can sometimes restrict their use to moderate-power or receiving applications. respectively. 58 In the geodesic analog the variation in dielectric constant is obtained by the increased path length for the RF energy traveling in the TEM mode between parallel plates. unless zoned. Constrained metal-plate lenses are capable of very wide scan angles as limited scanning possible by moving the feed in a paraboloid reflector.76 and 1. To reduce the loss caused by scattering from the steps in a zoned lens. 7.

. The relative mum radiation is perpendicular. to the line (or plane) of the . array when the beam is The endfire array is a special case of the linear or the planar widely used in radar been not have arrays linear Endfire directed along the array. . Just as with any other radiating from Fourier-transdetermined be may pattern radiation far-field achieve the desired efficiency but maximum in results The uniform amplitude distribution form theory expense ol the at sidelobes lower in results tapered distribution large sidelobes the other radar antennas. the or lens the of action the by wave plane It is converted to a reflector spherical are applied to the signal before it is the array antenna. gains since an endfire linear medium or low to limited usually are They applications. or more) could not be built and the reflector. The two-dimensional planar array is . main beam. the proper phase relationships individual elements. (thousands. „ ^ „ probably the one of most interest in radar A antennas. phases between the elements determine the position of the of the Scanning fixed. beam when radiation parallel to the array. array and the planar A planar array is a two-dimensional configuration of elements line in one dimension. same with the beams tracking search and/or The antennas Many of the early radars developed in the late 1930s used array with compared UHF) lower or (VHF low relatively frequencies of these radars were . are plane in the orthogonal The combination of the linear-array feed and the parabolic bolic-cylinder antenna. . than broadside. tive phase shift between the elements of the linear Two common geometrical forms of array antennas of interest in radar are straight in a arranged elements of consists array linear A array. ten of thousands. array-antenna the array. structure. the phase relationships are such that the When the radiation is at some angle other The broadside linear- plane and narrow beamwidth array antenna may be used where broad coverage in one also act as a feed for a paracan array linear The desired. that is. versatile of all radar applications since it is fundamentally the most A square or a circular aperture rectangular aperture can produce a fan-shaped beam. The lens lens the both from concept in differs The array antenna limited only and the it is apply the proper phase relationships to the wavefront after is feed the leaving on wavefront the of radiated by the point-source feed. . antennas. generate many simultaneously to made be can produces a pencil beam. also is pattern radiation If the phases are fixed. if directive elements are array broadside of a sometimes used as the radiating elements as low-silhouette employed also are elements endfire of arrays Linear required. possible with either a simple cylinder generates a more controlled fan beam than is of a linear array and combination The parabola. . reflector. The shape In reflector. the radiation pattern is a conical-shaped beam. of as a linear array of linear arranged to lie in a plane. array An endfire array has its maximum The linear array generates a fan radiation is perpendicular to the array. The planar array may be thought the direction of maxiwhich in array planar or a linear is a A broadside array arrays or almost perpendicular.s 7 7 -i Antennas The maximum number is 295 elements by practical considerations. in the transmission lines feeding needed across the array to distribution the aperture. The array aperture. Small endfire arrays are antenna of high gain requires an excessively long array. the antenna entire the moving mechanically by accomplished be beam formed by an array can also be steered by varying the relacan beam the However. There antennas containing large numbers of elements is no fundamental reason why array if it were necessary to do so. . the radiated. of a linear array or with a section parabolic cylinder can also generate a pencil beam. the phase distribution across A reduced Unlike which can be readily controlled and is one of the characteristics array an of aperture or a lens a uses which one from array an distinguish the design of a radar system using efficiency.

Lower sidelobes can be achieved. Radiation Pattern?. 7.7 although they were high for that period. the receivers. the results obtained apply equally well to a transmitting antenna.™. large peak powers can be radiated.* Consider a linear array made up of elements equally spaced a distance d apart (Fig. This is an important advantage if the antenna is large.296 Introduction to Radar Systems [Sec. where 6 is the direction of the incoming radiation It is further assumed that the amplitudes and phases of the signals at . It is the flexibility ottered by the many individually controlled elements in an array antenna that makes it attractive for radar applications. Other types of array antennas are possible than the linear or the planar arrangements For example. long-range radars. since more elements are required to fill the same physical aperture at the higher frequencies. Because large antenna apertures are necessary for high-performance.33). the array is more practical at lower frequencies than at higher frequencies. the array antenna received renewed attention. The array antenna was also used extensively in communications work prior to World War II and its performance and design were probably better understood by the preradar 'antenna engineer than were reflector antennas or lenses. but such a radar could not be converted efficiently to an electronically scanned array by simple replacement of the antenna alone because of the interdependence of the antenna and the other portions of the radar. in principle. by rnore sophisticated techniques. The elements are assumed to be isotropic point sources radiating uniformly in all directions with equal amplitude and phase. Although isotropic elements are not realizable in practice. and the aperture illumination can be more readily controlled than in a singlefeed antenna. but because of the reciprocity principle. An array in which the relative phase shift between elements is controlled by electronic devices is called an electronically scanned array. In an electronically scanned array the In the 1950s. The outputs of all the elements are summed via lines of equal length to give a sum output voltage Ea Element 1 will be taken as the reference signal with zero phase. The array is shown as a receiving antenna for convenience. Large-aperture antennas can be designed at VHF and UHF with relatively few radiating elements. and the data-processing portions of the radar are often designed as a unit. antenna elements. = each . the ability to control each individual element results in a complex and expensive radar. 7. primarily because of its inherent ability to electronically steer a beam without the necessity of moving large mechanical structures. provided the phase at each element is that needed to give a plane wave when the radiation from all the elements is summed in space. especially for the computation of radiation patterns. The difference in the phase of the 1 compared with more conventional antennas are its high cost and the complexity resulting from the many additional components.d/X) sin 6. as interest in extremely long range radars increased. A given radar might work equally well with a mechanically positioned array. Its major disadvantages when N . or indeed on an object of any shape. Interest in array antennas for radar applications waned with the development of microwave radar and the application of optical techniques to microwaves The reflector proved to be a simpler antenna than the array in the sizes required for microwave frequencies. or a reflector antenna if they each had the same radiation pattern. Other advantages of the array are that more than one beam can be generated with the same aperture. the transmitters. The effect of practical elements with nonisotropic patterns will be considered later. It was more convenient to design and manufacture and was reliable in operation The reflector is a popular form of antenna and one difficult to displace later radars. signals in adjacent elements is y> 2M. However. the elements might be arranged on the surface of a cylinder to obtain 360° coverage (360 coverage may also be obtained with a number of planar arrays) The radiating elements might also be mounted on the surface of a sphere. they are a useful concept in array theory. a lens.

7.7] Antennas 297 element are weighted uniformly. The sum of all the voltages from the individual elements. Equation 7.29) co is the angular frequency of the signal.5 db below the main beam. and the The pattern of a uniformly illumi268). (—90° < 6 < 90°) is considered. radiation pattern is equal to the normalized square of the amplitude. p.Sec. when the phase difference between adjacent elements is y. or The G a(6) If the spacing \Ef 2 sin 2 2 [NirjdlX) sin 0] 2 N ~N (7. TV-element linear array. the phase shift would be zero). the last sidelobe is jN 2 (Ref. second term represents an amplitude factor of the form sin (iVy/2)/sin (y/2). for N sufficiently 1 large. but sidelobe. displaced by 180°. height of nated array with elements spaced A/2 apart is similar to the pattern produced by a continuously illuminated uniform aperture [Eq. When the spacing between . the half-power beamwidth is approximately equal = The first 10L8 N is (7 32) 13. To avoid ambiguities and confusion. 1 .31 predicts a second beam equal in magnitude to the main beam.33. can be written Ea = where sin cot + sin (cot + y) + sin (cot + 2y) + y h sin [cot + (N 66 l)y] (7. for convenience. As long as the spacing between elements of the array is K\2 or less. (7.3 1) to is between antenna elements is A/2 and if the sine in the denominator of replaced by its argument. The sum can be sin (JVy/2) sin (y/2) 1) written sin cot + (N - (7. Therefore the amplitudes of the voltages in each element are the same and.16)].31) sin \jr(djX) sin 0] Eq. while the Incoming signal — Fig. (7. the sidelobe radiation will be small compared with the main beam.30) The first l)y/2 (if the phase factor is a sine wave of frequency to with a phase shift (N reference were taken at the center of the array. 7. will be taken to be unity. this backward radiation is usually eliminated by placing a reflecting or an absorbing screen behind the antenna For this reason only the radiation over the forward half of the antenna array.

and the signals at each element are assumed where M = number TV = J number of radiating elements in d e . additional lobes can appear in the antenna radiation pattern with amplitude equal to that of the main beam. Therefore the element spacing should be no greater than a half wavelength if tt(<//A) sin = 0. If the radiation patterns in the two principal the two-dimensional antenna pattern is G(6M = Note that the angles 6 e . The beam of an array antenna may be steered rapidly in space without moving large mechanical masses by properly varying the phase of the signals applied to each element. or secondary principal maxima.298 Introduction to Radar Systems is [Sec. 70 183 The normalized radiation pattern of a uniformly illuminated rectangular array is G(0 d ) " ° = si " 2 l>K<*M) sin sin 2 fl„l sin 2 iMnjd/X) sin flj ( N 2 |>(d/A) sin 0J M 2 sin 2 [^d/A) } sin B in d a dimension with spacing d dimension Beam Steering. etc. 2v. 7. The resultant radiation the product of the element factor Ge {6) and the array factor G (0).7 greater than half a wavelength. n. 68 These effects cause the element radiation pattern to be different when located within the array in the presence of the other elements than when isolated in free space. the latter a being the pattern of an array composed of isotropic elements. they might lead to confusion. When directive elements are used. These are called grating lobes. It should be used with caution.33) many problems of array tain the principal axes of the antenna. Grating lobes caused by a widely spaced array may therefore be eliminated with directive elements which radiate little or no energy in the directions of the undesired lobes. The positions of the grating lobes can be found from Eq. GtfJGJBJ (7.34) and 6a are not necessarily the elevation and azimuth angles normally associated with radar. since targets viewed by the grating lobes cannot be distinguished from targets viewed by the main beam. The spacing between adjacent elements is d. the numerator and the denominator are zero. (7. If less than full coverage is satisfactory. planes are G^dJ and G 2 (da ). which may be seriously inadequate for design. grating lobes will occur at 6 = ±30° and 6 = ±90°. only an approximation. or when elements is They occur whenever both 6 For example. the grating lobes produced in a widely separated array can be reduced or eliminated over a smaller scanning range by elements with directive rather than isotropic radiation patterns. when the spacing d between two wavelengths. The array pattern may be found by summing the contributions of each element. The array factor has e (Q) G is also been called the space factor. 67 and it does not take account of the scattering or diffraction of radiation by the adjacent array elements or of the outward-traveling-wave coupling. taking into account the proper amplitude and phase. the resultant array antenna radiation pattern is G(6) = GM N N^H^l = <***&) sin"* (7-33) lTT(d/A) sin v\ where pattern is the radiation pattern of an individual element. In a two-dimensional. full ±90° coverage is desired. Consider an array of equally spaced elements.31). the radiation pattern may sometimes be written as the product of the radiation patterns in the two planes which conis Equation (7. the pattern of each element must be measured in the presence of all the others. In order to obtain an exact computation of the array radiation pattern. They are due to the radiation from the elements elements adding in phase in those directions for which the relative path lengths are integral multiples of 2tt radians. rectangular planar array. It ignores mutual coupling. If the grating lobes are allowed to exist.

If the same phase is applied position of the main beam will be the difference between adjacent elements is zero and beam will point in a direction main The 0. 1. ' I (7.sin fl )] O )] L . than rather If variable.44377.36) pIG 7. u Q_ to normal to the antenna. where m- </> normalized radiation pattern of the array when the relative phase difference beis given by tween adjacent elements is <f> Zm | sin 2 [JV77(rf/A)(sin 2 () N 2 sin |XdM)(sin .37) can be neglected when cos small - sin 6 s» sin (0 - O) O (7. the two angles corresponding to the 3-db antenna pattern are Q. . shifters mechanically positioning the entire antenna structure. 7.7] Antennas 299 to all elements. . sin Nd cos O Nd cos 6 The half-power beamwidth 6B = + - 0.sin . that when point + ing to the half-power O 69 expression for u can be written the in 6 term sin 6 The sin —0. Change ofBeamwidth with Steering Angle. angle broadside to the array at an between elements is other than other than broadside if the relative phase difference difference .443A ?& a = is . = is therefore </> c + m0.36) states that the main lobe of the antenna pattern may be positioned to the array. The phase-shifting device might be either phase electronic or mechanical with Steering of the beam or electronically controlled by obtained can be than greater magnitude of orders speeds scanning results in . This may be proved by assuming that the sine in the denomipattern is nator of Eq.36) can be replaced by its argument. the Therefore. 2. the relative phase of equal amplitude. (7. ~ -^Nd O (7-39) cos 6 n off broadside. an angle O by the insertion of the proper phase shift <£ at each element of steered as the be beam may the used. = — < . = sin 6 - sin O = sin (0 - O) cos O - [1 - cos (0 - O)] sin O O is (7.34).443A p 443A 0+ " 0o = Sm ^ _. beamwidth in the .Sec. when the beam is positioned an angle _1 plane of scan increases as (cos O) . so that the radiation 2 antenna 2 2 2 (sin u)/u The sin X)(sin Nw(d/ ).38) points of the Using the above approximation. where u u)/u O of the form (sin Denote by + the ±0. all to applied phase stant 2t7(J/A) sin . V The elements. so that sin side of Eq. are shifters phase fixed.443A -0.37) The second term on the right-hand (beam is near broadside). O angle corresponding to the half-power point when and +0.34. O. direction O .443tt. Steering of an antenna beam with The maximum occurs when sin of the radiation pattern variable-phase shifters (parallel-fed array). = sin Equation (7. 7. and c 0. = - = > . (7. mechanically phase is changed (Fig.is of the main beam is at an angle O when the phase = zero </> The .443tt to u corresponds is. pattern is reduced to half its maximum value when u correspondangle the and 0_. The phase at each element is any con(N 1). The half-power beamwidth in the plane beamwidth is of scan increases as the beam is scanned off the broadside direction. The from the measured angle the is where cos proportional to inversely O approximately O . . . N</cos0o ~N<icos0 o -0.

It certainly does not apply when the energy is Eq.35). the two beams overlap.38)] radiated in the endfire direction. The antenna radiation pattern is plotted in spherical coordinates as a function of the two direction cosines. (7. of the radius vector specifying the point of observation. as O approaches 90° (the endfire condition). Radiation pattern of an array in free space = 1 0. The shape of the beam at the other angular positions is the projection of the circular beam shape on the surface of the unit sphere. both beams exist and merge to form the endfire beam at O = 90°. and x is An interesting technique axis perpendicular to the cos a x and cos av 0=0 ' .443 —. It can be seen that as the axes. If the antenna elements are in free space.36. The antenna actually produces two beams. The angle 0' defining the boundary between the endfire and the broadside regions is that value of O which makes the argument of the first sin. showing beams at 6 „ _ = d and „ . (7. sin O ) (7 40) sin Care must be taken in the interpretation of Eq.443 —+ sin 6U + sin" 1 (o. 7.36) for the linear array. 7. assumed large compared with the wavelength 1) In a practical array. The gain of an antenna will also vary with scan angle in a manner determined by the pattern of the element within the arrav varies approximately inversely as cos Although the effect [Eq.35) unity. The above analysis applies to the linear array. or a similar expression for the planar array. more exact expression for the beamwidth [obtained without the approximation of A is eB = sin 1 ^0. 7. The broadening of the beam is a direct consequence of the fact that the maximum gain remains constant. 70 an example of which is shown in Fig.443(A/AW). provided certain assumptions are fulfilled. At (beam broadside to the array) a symmetrical pencil beam of half-power width B is assumed.1 term of Eq.7 The change in beamwidth with angle as derived above is not valid when O the antenna beam is too far removed from broadside. 7. it cannot be concluded that the gain of the antenna always decreases in a similar fashion! Equation (7. Therefore. The preceding statement is not readily obvious from an examination of the pattern in two dimensions since the antenna radiation pattern is three-dimensional. cos a„ and cos a„. where a is the antenna length. (Schelkunoff and Friis 64 prove that the directive gains of continuous linear antennas in the limiting cases of broadside and endfire radiation are equal to 4a/ A. y In Fig. the beamwidth in the plane of the scan . since a value of the sine greater than unity has no meaning. O .40) when the argument of the first 1 term is greater than unity. shows that the maximum value of the gain or the field intensity is independent of the scan angle.33)]. sin = sin (tt — 0).35. 6o=0 one of which is at an angle 6 the other at the angle 77 — This follows from the fact that O (Fig.36. „„. <f> is taken to be a constant value of 90° and the beam is scanned in the coordinate.300 Introduction to Radar Systems [Sec. The angle <f> is measured from the cos ol axis. (7. If the array is above a reflecting or an absorbing ground plane. 7. or sin 0' Fig. however. of scanning is to broaden the beamwidth in the plane of the scan. the gain will change with scan angle because of changes in the mutual coupling between elements. measured from the for graphically portraying the variation of the beam shape with scan angle has been described by Von Aulock. the antenna cannot radiate at angles greater than 90° and the antenna pattern must be modified accordingly. Bickmore has shown that a similar result applies to a planar aperture 69 that is. (7.

beam shape is not symmetrical about the center of Thus the beam direction is slightly different from that the computed by standard formulas. The circuitry (not shown) which follows . the beam. The output of each amplifier is subdivided into a number of independent signals which are individually processed as if they were from separate receivers. Amplifiers may be placed between the individual antenna elements and the beam-forming (phase-shifting) networks to amplify the incoming signal and compensate for any losses in the beam-forming networks. Another set of three phase shifters generates a beam in the d = +0 O The third set of phase shifters generates a beam in the direction = — 6 direction. 7. The inherent flexibility of the array antenna permits a number of beams to be generated simultaneously from the same aperture. Low-noise amplifiers should be used if the signal-to-noise ratio is to be maximized.7] Antennas scanned in the d direction. but is eccentric.) antenna can be made to look in all directions at once. 7. but is constant in direction. within the The ability to limitation imposed by the radiation pattern of the antenna elements. For 0^0. {From Von Aulock. In addition to the changes in the shape of the main beam. Thus a Fig. 7.'"' courtesy Proc. The simple array in this figure is shown with but three elements. When beams are formed in networks placed after the RF amplifiers the antenna is The beamcalled a postamplification beam-forming array. phase difference inserted between adjacent elements. the sidelobes also change in appearance and position.Sec. it 301 beam the cf> is broadens in that direction. than reception on easier usually is beams form many necessarily a disadvantage since it is a useful method of operating an array in many systems applications. One set of phase shifters produces a beam-directed broadside to the array (6 = 0). each with three sets of phase shifters. This is not transmission.37. abbreviated PABFA. Therefore the beam-forming array will be discussed primarily single receiving as a receiving antenna. where A<£ is the . The simple linear array which generates a single beam can be converted to a multiplebeam antenna by attaching additional phase shifters to the output of each element. as shown in Fig. Each beam to be formed requires one additional phase shifter. Beamwidth and eccentricity of the scanned beam. forming networks may be at either RF or IF. 1 The angle 6 is determined by the relationship O = sin" (^Xjl-nd). IRE.36. Beam-forming Array.

1 Fig.37. Beam No. . Mixer Mixer Mixer 1 > (^ \Anno/ no/ f^ > ' \Amp/ . 7.t Toppe i lines 1 c o o _». <h = |277(rf/A)sin 0„|.7 Beam No. constant pha r 7 1 7 ^7 1 1 — L0 I r .38. 3 Til Beam No.2 o o o o . 3 F[G. Beam-forming network using tapped delay lines at IF. o Beam No. Beam No.2 Beam No. 7.302 Introduction to Radar Systems [Sec. Simultaneous postamplifier |0i — 9o\ = |A0| = beam formation using array antenna. 7.

height finder used 30 miles of 5-band waveguide to produce 333 beams. beams with an ^-element array.Sec. The Butler. The output of a being PABFA radar can be applied directly to a data-processing device without displayed to an operator. of course. to generate multiple beams directly. The waveguide transmission Energy is tapped from each lines act as the delay lines. 10. which uses a parallel-plate lens. and the Bootlace 190 antenna. 71 In this capacity it acts as an in conjunction with a circular or a spherical array. phase relationships for the correct gives the analog computer which automatically spherical array. 7.3) built by the Maxson Corporation for the Federal Aviation Agency. The RF. Maxson. Another RF beam-forming device is the parallel-fed network attributed to " 188 By properly utilizing 3-db directional couplers or hybrid junctions. to form beams at various couplers directional by the appropriate points waveguide at Considerable waveguide is used in arrays of this type.7] Antennas in Fig. The signal-to-noise ratio at the output of the summing .39. beam-forming principle shown in Fig. preserved during a frequency translation (except for the constant phase shift introduced by the local oscillator). The Luneberg lens can be used as a beam-forming network to form multiple beams it is possible to form n overlapping The Mubis 189 antenna. Mubis. The indicator first display for a PABF A is slightly different from that of the usual radar. with Butler. RF beam-forming network using tapped transmission lines. 185 fixed phase shifters. 7. Signal-to-noise Ratio. 7. 7.38. as described in Sec. which is a form of parasitic array lens. and the Bootlace beam-forming devices are passive and therefore can be used for both transmission and reception. The Maxson elevation angles. The tapped delay The phase of the IF is the same as that of the RF since phase is lines are shown at IF. are also capable of RF beam forming. convenient method of obtaining a receiving beam-forming network at IF is with the use of a series of tapped delay lines as illustrated in Fig.6. 7.39 is used in the AHSR-1 height finder (Fig. Crossed-line directional coupler A >^ Waveguide Fig. 7. The Luneberg lens can also be used.37 is 303 the summing networks conventional radar receivers.

but the gain of the transmitting antenna used with PABFA is Therefore less than that of the scanning array since it has a considerably broader beam. and the composite effect is that of single-channel radar. the total signal-to-noise ratio with postdetection integration is less than n times the signal-to-noise ratio of a single pulse and The smaller the signal-to-noise ratio per pulse. beam radar are of larger signal-to-noise ratio than those obtained with the PABFA radar. In the above example. pulses of small signal-to-noise ratio obtained in the PABFA must be integrated before detection (coherent integration) if the total signal-to-noise ratio is to be equal to n times the signal-to-noise ratio of a single pulse. Postdetection. There need be no loss in the signal-tonoise ratio due to the forming of the beams in an array antenna. In practice. There are fewer of them. this is compensated by the fact that the fixed receiving beams of the PABFA receive many more pulses per unit time from a target than does a scanning time-shared beam. ponents are added coherently in the summing networks of the array. A many fixed radar beams operating in parallel. the signal-to-noise ratio of each received pulse will be less with the PABFA radar than with the radar which uses a single scanning beam. in principle. a single 1 by 1° transmitting and receiving pencil beam would cover the 90 by 2° volume by making an observation in each of the angular resolution cells in sequence. if the energy available in the received signals is processed properly in both cases. In principle.304 Introduction to Radar Systems [Sec. many narrow receiving beams are fixed in space. the detection capability will be the same and the performance of the two radars will be equivalent. let it be assumed that the PABFA radar consists of a receiving array generating a number of overlapping narrow beams fixed in space. a proof. there is an integration loss (Sec. The receiving antennas of the PABFA and the scanning radar are assumed to be of the same effective area. 7. A scanning radar with a single narrow beam must cover the volume by time sharing.7 network is theoretically the same as that of a conventional radar using a single large antenna to produce the same antenna beam. The separate transmitting array is assumed to generate a single broad beam illuminating the same volume of space as covered by the multitude of receiving beams. the nonlinear effects of the second detector. provided the array and its circuitry are properly designed. integration is more often used. The amplifiers must have sufficient gain to overcome any losses in the beam-forming networks. while the noise components are added incoherently. The individual pulses obtained with the scanningthe greater the integration loss. It can be readily shown that the total energy contained in the many small pulses from the PABFA radar is the same as the total energy contained in the few large pulses received from a scanning single-beam radar. Coherent integration is not always Because of practical. a radar with a postamplification beam-forming array is equivalent in over-all performance to a radar with a single scanning beam. Therefore. all other factors being equal. The transmitting pattern The single broad transmitting beam and the is therefore a single fan beam 90 by 2°. or noncoherent. 72 Comparison of PABFA with Scanning-beam Radar. no loss as compared with a The above explanation is qualitative and does not constitute mathematical proof of the equivalence of the signal-to-noise ratio of a PABFA and a conventional radar was given by Rush. the receiving antenna might generate one hundred and eighty 1° pencil beams arranged to cover an angular sector 90° in azimuth and 2° in elevation. For example. the two radars may not be exactly equivalent because it may not always be convenient or possible to process the signals in an optimum manner in both cases. there is. However. Noise components from parts of Since the signal comthe receiver other than the RF amplifier should be kept small.6). 2. provided the comparison is made on a similar basis and For that the received signals are processed in the optimum manner in each case. making the total integration loss less with the The n . purposes of comparison.

resolution cell. proper phase change is introduced by the phase shifters in each of the lines feeding the When the phase of the first element is taken as the reference. which is less than 2-rr radians. 3<f>.406). (c) parallel-fed array with power-dividing network. Only one while the PABFA radar with its fixed beams views the target continuously. Arrangements for applying phase relationships in an array. 7.(N required in the succeeding elements are <f>. the energy may be transmitted from one end of the line (Fig. 7. Therefore the PABFA radar may be slightly less efficient than the scanning-beam radar when the integration is performed noncoherently. All the phase shifters are identical and introduce the same amount of phase shift. or it may be fed from the center out to each end (Fig. the phase shifts element.Sec. PABFA . fed from one end. detection decision radar in this time since it must integrate all is made by the the received by the scanning radar on a order to equal energy energy in available the single observation.40. .7] Antennas 305 scanning-beam radar. In the parallel-fed array of Fig. The relative phase shift between adjacent elements of the 2ir(d/X) sin 6 in order to position the main beam of the radiation array must be ef> The necessary phase relationships between the elements may pattern at an angle d be obtained with either a series-fed or a parallel-fed arrangement. . 7. (a) Series array. The maximum phase change required of each phase shifter in the parallel-fed array Because phase shift is periodic with period 2n. radar is the time taken by the beam to cover the entire volume and return to the same The scanning radar views the target but once during the scan time. ^-/V \ > I > / V\ * 1 > /f\ » 1 > />V (a) V i i —*-/j\ * * * -9\ / ib) Fig. Parallel Feeds. = . (6) series array. 2<f>. In the series-fed arrangement. . 7.40c. Series vs. The data rate of the two radars can be shown to be the same for equivalent detection The data rate of the scanning-beam capability. it is possible is many times 277 radians. The adjacent elements are connected by a phase shifter with phase shift <f>. the energy to be radiated is divided between the elements by a power splitter. — .40a). \)<f>. assuming ideal coherent integration. . 7. center-fed. Equal lengths of line transmit the energy to each element The so that no unwanted phase differences are introduced by the lines themselves.

1 Hence the aperture is uniformly illuminated. A similar phenomenon occurs in the series-fed array when the energy is radiated or If a short pulse is applied at one end of a received at or near the broadside direction. only a single control signal is needed to steer the beam. Thus the series-fed array introduces more loss than a parallel-fed array. The usable bandwidth of a resonant array of elements is N . the system response may be degraded. is to smear or distort the echo pulse. The chief limitation of the resonant array is its very narrow bandwidth. The combined outputs from the parallel-fed elements will fail to coincide or overlap. It for each phase shifter. The A^-element parallel-fed linear array similar to that of Fig. but by adjusting a short-circuiting plunger at the end of the The short dissipates no power. If the operating frequency is changed from the design frequency. The impedance match is obtained not only by choosing the impedances of the elements properly. any energy is reflected by the short circuit is radiated as a array. Since the elements are spaced a half wave array. amplifiers can be inserted in each element to compensate for the signal attenuation. It is possible to compensate for the delay in the series-fed array and avoid distortion of the main beam when the signal spectrum is wide by the insertion of individual delay lines of the proper length in series with the radiating elements. radiation of energy by the first element might be completed On reception. 7. For example. the signal suffers the insertion loss of a single phase shifter n times. Since each phase shifter in the series-fed linear array of Fig.7 in some maximum of but 2tt radians. and the received pulse will be smeared. Resonant and Nonresonant Series-fed Arrays. This situation may be relieved by replacing the 277 modulo phase shifters with delay lines. and its impedance is well matched at the design frequency. apart. In a parallel-fed array the insertion loss of the phase Hence the phase shifter in a series-fed array must be shifter is introduced but once. signals.40a has the same value shift. if the pulse width is short compared with the antenna response time (if the signal bandwidth is large compared with the antenna bandwidth). just as all the energy propagated in the forward direction. of lower loss compared with that in a parallel-fed array. 7. and the array is not properly matched. but it is easier to program the necessary phase shifts. Series-fed arrays radiating in the broadside direction may be classed as either resonant or nonresonant. however.306 Introduction to Radar Systems applications to use a phase shifter with a [Sec. The final choice between the two will usually depend upon the system application. the entire array would not be excited simultaneously. the spacing between the elements is no longer a half wavelength. the array is no longer uniformly illuminated and the radiated beam is not perfectly normal to the array. 7. seems unequivocally to excel the other in all situations. In addition. Neither feeding arrangement. the impedance contributions of the elements do not all add at the input. A resonant array is one in which the elements are spaced exactly one-half wavelength apart. beam normal to the It can be shown that the elements of the resonant array couple equal power from the waveguide or transmission line and that there is no attenuation in a line loaded with pure series or pure shunt elements spaced at half-wavelength intervals. the effect before the remainder of the energy reaches the last element. If the series phase shifters are too lossy. dimensional parallel-fed array of elements requires 2 separate control The two-dimensional series-fed array requires but two control signals. 73 In a series-fed array containing n phase shifters. However. The radiation pattern and the impedance of the array deteriorate with a change in frequency. if the energy were to arrive in a direction other than broadside. or N— MN M+ N— radiates a beam normal to the array. series-fed transmitting array.40c requires a separate control signal of phase A two1 total (one phase shifter is always zero).

Another phase shifter which has been used in array radar is the rotating-arm mechaniphase shifter (Fig. In some cases might prove to be a limitation. The limited-bandwidth restriction of a resonant array can be removed by making the spacing between the elements differ from a half wavelength. 7. the postamplification beam-forming array described previously. The radiated beam ± N is this not normal to the array. any There are Devices. The electrical length may be changed by physically shortening or lengthening the line. especially when the narrow bandwidth of the resonant array Fig. Hence one line can supply the necessary phase variation to two elements. means. it introduces other problems. Even with these limitations. The telescoping section may be in the shape of a U. 5 Therefore practical resonant arrays cannot be approximately 50/ made too long. 7. if the original +6. match and improves the bandwidth. cannot be tolerated. and (3) variable phase shifters controlled by electronic means. for example. The power end of the array must therefore be Ordinarily dissipated in a matched load. Fixed phase shifts are utilized in array antennas which generate fixed as. (2) variable phase shifters array. Small variations in the element spacings have a similar effect on a resonant array as a change in frequency. Any reflected energy will radiate from the antenna as an undesired lobe at an angle —6. Principle of rotating-arm mechanical phase shifter. - cal 65 74 It consists of a number of concentric transmission trough with an insulated conductor passing square three-sided Each line is a lines. but it is the type of problem which can be tolerated in To arroy elements Open transmission nes». actuated by mechanical means. and the total length of line is changed in a manner similar to a slide trombone. 7. Phase-shifting number of devices which can provide the proper phase shifts at the elements of an They may be classed as (1) fixed phase shifters.) with each circular assembly. the phase at the other end is decreasing. the nonreso^Contacts to transmission line nant array is useful. Variable-phase-shift devices are based on changing the electrical length of a transmission line. One of the simplest methods of obtaining a fixed phase shift is with a length of transmission line. (Details of the line are not shown. most applications.41).7] Antennas 307 per cent. Electronic phase shifters operate by changing the (electrical) length of line by electronic beams. When the phase at one A end of the concentric line is increasing.Sec. An array of this type is Although the nonresonant array eliminates the poor impedance called nonresonant. This is called a line stretcher.41. one on either side of . The arms are rotated to produce a continuous and uniform variation of phase across the elements of the array. One of the simplest forms of mechanical phase shifters is a transmission line designed with a telescopic section whose length can be varied. In the nonresonant array there must be no power reflected from the end of the array after the energy in the transmission line has passed the last element. Most mechanical phase shifters are based on this principle. about 5 per cent of the total power gets beyond the last element and is dissipated as wave radiates at an angle at the heat. moving arm makes contact down the middle.

2* Concentric coupled helix sections 69 Fig. 79 80 A schematic representation is shown in Fig. 74 A change in phase in a waveguide transmission line may be obtained by changing the The wavelength of the radiation propagated in the guide is dimensions of the guide.308 Introduction to Radar Systems [Sec. The outer rings.43. 79 . 7.42. and 8 are not excited by the signal traveling from terminals 1 to 5. 7. For this reason a given mechanical motion produces more phase shifter. through the device is varied by mechanically positionin the process. while the inner rings feed the inner elements. Each of the short coupled helices behaves as a directional coupler which transfers all the power principle. The helical-line phase shifter acts as a trombone line stretcher for a signal propagated from terminal 1 to terminal 5. to terminal 7 ing the coupled helices. Likewise. a complementary phase shifter may be had by bridging terminals 2 and 6 and passing the second signal from terminal 4 to terminal 8. elements. feed the outer elements of the array. phase change than would a line stretcher in conventional transmission line. 2. Principle of the rotary-waveguide The signal crosses the bridge and is completely transferred to terminal 5. Since terminals 4. The reduction in length is essentially equal to the wind-up factor of the helix. The rotating-arm phase shifter has been used in a VHF height-finder radar 75 and in the Air Force Camarray center. 76 A - from or minal 3. The path length. which is the ratio of the circumference to the pitch. Schematic representation of helical-line trombone phase shifter. 5 ' dependent on the guide width. terminals 6 and 8 are not excited.7 A total of N/2 concentric rings are required for a linear array of N + I The progressively greater phase variation required at the outer elements of the array as compared with the phase variation at the inner elements is readily obtained with the concentric-ring configuration. Thus a shorter phase shifter can be had which is especially advantageous at VHF or UHF frequencies. bridge Research Center VHF experimental scanning radar called Billboard. The phase velocity on the helical transmission line is considerably less than the velocity of light. Several phase-shifting devices have been based on this This technique has been applied to ground-controlled-approach (GCA) scanning radar 77 and to the AN/APQ-7 (Eagle) scanner. A signal incident at terminal completely transferred to terTerminals 2 and 4 are not excited is Fixed section Fig. Wind-up factors may range from 10 to 20 in practical designs. 78 mechanical device which gives more phase shift for a given amount of motion than a conventional line stretcher is the helical-line phase shifter due to Stark. Two helical lines 1-4 and 5-8 are coupled electromagnetically to one another by the helices 2-3 and 6-7. and hence the phase shift.42. to the 1 main helix. being larger. 7. 6. 7.

An interesting property of the 180° section is that it converts circular The phase-shifted circularly polarized polarization to the opposite sense of rotation.1 sec or the type of phase shifter and its design. mechanical with better are readily achieved antenna beams to be scanned considerably faster than is possible with a large antenna which must be positioned mechanically. Although these speeds permit devices. Transition sections (not shown) might be employed and to to convert from rectangular waveguide to the round waveguide of section I Sections I and III are rectangular. An array of spiral elements makes a simple scanning antenna. even shorter switching times can be had with Switching times on the order of milliseconds electronically controlled phase shifters. electronic most with commonplace are second switching times or better. The obtained by the rotation of a round waveguide and can be made quite phase The mechanically rotating section II may be replaced by a ferrite Faraday rapid. 85 86 The linearly polarized beam radiated by a flat. 81 phase shifter has been described in detail by Fox and was applied in the Bell Telephone scanning radar. The signal enters section I as a linearly polarized wave and is converted to a circularly Section II acts as a half-wave plate. shifter is a two-port RF transmission line in which the phase of the output signal is 89 Phase varied by changing the d-c magnetic field in which the ferrite is immersed. can be shifter - A hybrid junction such as a magic T or its equivalent may be operated as a microwave 84 manufactured with printed circuit techniques. 7. The switching actuated. two-dimensional array of spirals may be scanned by rotating the individual spiral antenna elements. but possibly excluding the rotary joint. The entire assembly.7] Antennas shifter is the 309 Another mechanical rotary-waveguide phase shifter based on the The rotary-waveguide properties of circularly polarized waves in round waveguide. 82 83 ' by placing mechanically adjustable short circuits in the collinear arms. or traveling-wave tubes. and vice versa. back to III section of waveguide round convert the equivalent to quarter-wave plates. 7. No additional phase-shifting devices are required. One degree of mechanical rotation corresponds to a phase change of one electrical degree. rotation by an angle 6 results in a 20-radian change in the time phase of the section. They convert linearly polarized waves into Fox calls these 90° differential-phase-shift circularly polarized waves. Sections I and III are fixed. 76 The rotary-waveguide phase shifter Laboratories' waveguide (Fig. A different form of mechanical beam steering is used in an array with spiral antenna elements. .43). gaseous discharges. may be obtained An electronically controlled phase shifter at microwave frequencies 88 A ferrite phase with ferrite materials. output signal. rotator to produce an electronically controlled phase shifter with no mechanical moving phase parts. mechanically were above described All the phase shifters time required to position the phase shifter through a phase change of 360° depends on Switching times on the order of 0. including the spiral radiators and feed networks. Some devices are capable of microshifters. Peak powers of several kilowatts are possible. by the action of the quarter-wave polarized wave linearly back to a converted is wave A plate of section III. Helical radiating elements have also 87 been used in arrays to obtain phase shifts by rotation of the elements. A phase shifter of this type is simple and compact and has shift is little attenuation. or a 180° differential-phase-shift polarized wave. round of sections three of consists FH MUSA while section II is free to rotate. component in a particular plane is polarization of the velocity phase sections since the speeded up by 90° with respect to the polarization component in the orthogonal plane. shifts of 360° can be obtained in a structure of relatively small size with magnetic fields of 100 oersteds or less and with insertion loss less than 1 db. It is primarily useful in those applications where a broadband element is required and the power is not too high.Sec. and switching times can be made as short as tens of microseconds.

The output of the delay line is a signal of frequency/ with a phase delay equal to 2-nfc T. The traveling-wave tube is not a bilateral device as are most of the devices mentioned. Gaseous-discharge phase shifters are based upon the variation of the dielectric constant of the gaseous medium with the number of free electrons. The spreading resistance of the variable-capacitance diodes causes an RF > > A . The result is a signal with the same frequency as the input signal f but with the phase advanced by an amount <f>. Furthermore.. If. W. A X UHF . -The phase shift through the device is varied by changing the reactance of the termination. Separate phase-shifting circuitry would therefore be required for transmitting and for receiving. H. Relatively little voltage variation is required to obtain the necessary phase shifts. The phase-shifted control signal and the output of the first mixer are heterodyned in the second mixer. (Coaxial and strip-line equivalents can also be used. It is difficult to obtain stable operating characteristics with long life in sealed-off tubes. 7. gaseous-discharge phase shifters are often noisy. Another method of obtaining an electronically controlled phase shift (due to Prof. 7. 360° of phase change was obtained for a change of 18 volts on the helix. Ferrite phase shifters are to millimeter wave frequencies. 88 An advantage of the traveling-wave tube as phase shifter is that the same device can give amplification over a wide bandwidth and can provide a low-noise figure. the difference frequency is selected from the second mixer. A </> . is a function of the current through the discharge. and large phase variation per wavelength and can be adapted to a wide range of frequencies. Huggins of The Johns Hopkins University) is shown in Fig. or the device might be operated in a temperature-controlled environment. 96 An example is shown in Fig. In a particular tube. portion of the control frequency is passed through a delay line of length r. Phase shifters based on this principle have been constructed at frequencies ranging from to band. and hence the phase shift. 7. and hence the phase. 93 Gaseous-discharge phase shifters can handle about 1 kw of power and have fast switching time.310 Introduction to Radar Systems [Sec. signal of frequency/. the difference frequency were taken from the first mixer and the sum frequency from the second mixer. the output would be delayed in phase by the amount <f>. as are ferrite phase shifters. is controlled by the bias voltage applied to the diodes. in the first mixer.) The signal input at arm 1 is divided equally between arms 2 and 3. electronically controlled phase shift by variation of the helix voltage. however. The number of free electrons. The development of better materials might relieve this problem. 71 94 95 signal of frequency f whose phase is to be shifted an amount is mixed with a control <f>. ease of control. The traveling-wave tube may be made to provide a fast. 88 They are limited. to relatively low peak powers. on the other hand. The capacitance.45 in which two variable-capacitance diodes are coupled by a waveguide short-slot hybrid junction.7 Ferrite devices are sensitive to temperature changes and hysteresis effects. Ferroelectric phase shifters are may be eliminated phase - also possible.44. (The phase of the energy transferred from one line to the other is advanced 90° in the shortslot coupler. The reactive termination may be operated in a circuit such as a circulator or a properly phased balanced circuit to separate the incident and the reflected waves. phase shift may also be obtained by terminating a transmission line with a pure reactance.) The two signals are reflected by the diodes with a change in phase depending on the value of capacitance and recombine in arm 4. The energy incident upon the reactive termination is reflected with a phase change that is a function of the magnitude of the reactance. 90 The inconvenience of a tempera- ture-controlled environment ferrite phase shifter to precisely control the available from 10 Mc with a feedback control loop about the 91 92 shift. If the sum frequency is selected from the first mixer. One method of obtaining an electrically variable reactance is with the varactor (variable-capacitance) diode.

/* Fig. change in relative phase between adjacent elements may be Frequency Scanning. than the shifter phase transmission-line switched loss with the also be used as switches other RF phase shifters discussed. A lines might be of binary lengths (1. When the frequency is exactly /„. etc.7] Antennas 311 These phase shifters are compact. such as stubs Irises. The lines connecting adjacent elements of the series-fed frequency-scanned array are of equal length and chosen so that the phase at each element is the same when the frequency is the center frequency /„. the slot radiation can be controlled by means of discontinuities. Fig. and insertion loss of approximately 1 db. these spurious lobes are increment 22. 4 8. 1st fo + fc . For example. coupling Still another approach to the design of a steerable array is by controlling the 98 99 Normally. 7. Ferrites or gas tubes can are discrete rather than shifts phase available the with this type of shifter. might be represented by the series-fed array shown in Fig.5°) the antenna gain the phase shift is quantized into four bits (smallest lobe is about 24 db below the spurious largest the and db approximately 0. the beam points straight ahead. phase of increments needed the mate with digital-computer logic. An antenna consisting of elements spaced one-quarter wavelength apart with alternate elements in phase quadrature can be made to steer its beam by varying the amplitude at each element. transmission Variable-capacitance diodes may also be used to switch fixed lengths of 191 192 is switched line that of length The lines as a discretely variable line stretcher. 97 No phase shifters in the conventional sense are required. and have limited phase shift-bandwidth levels power efficient. if negligible. This principle can be used to scan a beam from an frequency-scanned antenna array if the phase shifters are frequency-dependent. A variable-capacitance diode phase shifter using a short-slot hybrid junction. within the guide can be made to electronically control the amplitude and phase of the Variations in coupling are obtained by varying the radiation coupled from the slot. but the insertion microseconds of order the Diode switching times can be of is greater. 7. The total phase through a fixed length / of transmission line is lirfl/c. but are limited to low product. units) so that the phase shift can be controlled or less. Schematic representation of the Huggins electronic phase shifter. for example. ' d-c magnetic field applied to the ferrite with an external electromagnet.. zation level is sufficiently small. obtained by a change in frequency.40a with fixed lengths of A A transmission line connecting the elements. fast. waveguide.2. be designed to be mechanically positioned by means external to the guide.45.44. the amplitude and phase of the energy coupled from the slots are fixed.Sec. can or irises. in an array made by cutting slots in the side of a of slot radiators. generally. judiciously placed within the waveguide near the slot. 7. The shift. However. and thus is a function of the frequency/.1 . order to approxiavailable be must lengths of line number determines the phase. Changes in Ferrite discontinuities iris position change the amount of coupling and the phase. Because If the quantipattern. 7. radiation antenna the in appear can lobes spurious continuous. 2d mixer f Q & Variable-capacitance diode termination t mixer Output© Delay Input © Short-slot hybrid junction Variable-capacitance diode termination © A T f . phase shift may also be obtained with amplitude adjustments only. m is reduced main beam. As .

The proper phase is determined needed to scan an angle 6 in the azimuth plane and 0o m the elevation plane. provided the switching time is long compared with the time for the wave to transit the length of the array. by superimposing the phase shifts It is possible to operate the array in such a manner that each phase shifter need not be controlled separately. They are reciprocal devices and may be used on transmission as well as on reception. At frequencies below f the beam moves in the opposite direction.™ The beam generated by a two-dimensional planar array may be scanned in space by applying to each element the necessary phase shift required to position the beam in the desired direction (Fig. if the antenna beamwidth were 1° the array would consist of approximately 10. the number of elements will be large and a large number of individual control signals are required to adjust all phase shifters to the correct value. Transmission lines can handle large power with low loss.46). Even when the wide-frequency band is available.Systems [Sec.000 control signals would be necessary if the phase of each element were controlled independently. 7 7 the frequency is increased above/ the phase through each length of transmission line increases and the beam rotates to one side. the use of the spectrum to accomplish frequency scanning may preclude the use of frequency for other purposes such as for electronic counter-countermeasures (ECCM). . especially at the The implementation of a frequency-scanned-array radar is relatively straightforward The phase shifters are simple lengths of transmission line. An independently controlled phase shifter is attached to each element. considerable saving in the number of control signals can be A . and systems which require wide bandwidths must justify the need. Frequency scan is more appropriate for one-dimensional rather than two-dimensional in principle.000 elements.V^ '—A^r< —/\-< -VX: \-^ P anar ' —Aa~ «3 Srray W th ' SSmilthf P hase " shift voIume tnc scan in two angular coordinates (elevation and lower radar frequencies. If the antenna gain is high. The simplicity of frequency scanning is complicated by the relatively large frequency spectrum which must be available in order to scan. the beam over a reasonable angular sector. The beam can be steered as rapidly as the frequency can be changed. The electromagnetic frequency spectrum is quite crowded. 7. For example. arranged in a square with 100 elements on a side. accurate range measurement' or resolution. A total of 10. Some other phase-shifting technique may be used in conjunction with frequency scan to steer in the other dimension. . Two-dimensional Scanning. —$kr< ~^p£^< 7" v* 3 V-^ '4v. 312 Introduction to Radar scanning.

47. might There are other combinations of series. To steer the beam in the in one plane. uses an array in A which steering in elevation is accomplished with frequency scan and steering in azimuth acts by mechanical rotation.7] Antennas elevation (Fig. (7. Volumetric scanning of a planar array with separate (parallel-parallel structure). example of the previous paragraph. The antenna is diagramed in Fig. the elements in the same column may of control from the independence follows Applying the phase shifts by rows and by columns unit beam may be Eq. ^^ Input I i A Azimuth controls azimuth and elevation control signals Fig.48. 7.47). as do all the series phase shifters in the Therefore. 313 had by steering the beam independently in azimuth and An array both scan since the phasing networks of this type is called a parallel-parallel structure receive the column same the in lie which elements the All planes are parallel-fed. phase shifting were carried out by rows and columns.11) developed by Hughes Aircraft Co. parabolic-cylinder for a feed line-source as a of less be scanned an angle of more than 100° in elevation with a frequency excursion . plane take the same value. 1.34). may be considered as a number of frequency-scanned arrays placed side by side. by given as planes of radiation patterns in the principal proper horizontal the selecting by volume coverage the within steered to any position If. (azimuth) displacement and the proper vertical m A /^ /j^. 7. in the (elevation) displacement.000 necessary dependent controls were used. 7. An end-fed frequency-scanned 40-element linear array is claimed that the beam can It reflector.and parallel-fed planar arrays which elevation be employed. 67 In the series-series planar array all series phase shifters in the azimuth plane. The Frescan radar (Fig. 7. only two control signals are required. elevation planar array using frequency scan in azimuth and phase shifters to scan in This is an example of a parallel-series array. all the elements that lie along the one unit lor purposes as considered be may row The elements in the same phase shift beconsidered one control Likewise. when inonly 198 control signals would be needed instead of the 10. identical phase shift in order to steer the beam also receive the same row same orthogonal plane._ _i „ Sec.

When the required not too large. The last-mentioned element (as well wen as ine the spiral radiator) is capable of large bandwidths. Array Elements.. A slot array is easier to construct at the higher microwave frequencies than a dipole array. Frescan incorporates an electronic pitch-and-roll stabilization system that modifies the elevation and scanning pattern to compensate for r ship s motion.10 per cent. where large angular coverage with a single array broad radiation pattern and are used s1 " ^^^J > 8 den? ents ° f array antennaS must not onl have the proper radiation y P S bC tWeen elementS muSt be smaII 6? AIso the element pattern Tu ade n Wh u n plaC ed the P resence of the °thers because of scattering and 68 rl ff diffraction effects. some of the radiating elements commonly found in arrays will be briefly J J mentioned. m . slot Chang6S phaSC by 18 °° in half a uide wavelength. The dipole is a simple radiating element which has been widely employed with both mechanically scanned and electronically scanned arrays. and the dipole produce a the dipole relatively is desired. Alternating the S elements causes the phase of the signal radiated from each element to be P a radlatCd by a Sl0t Can be Cha "ged 1 80 ° b tiltin ^ in the opposite y J? g T In a dipole array the phase can be reversed by reversing 1 . such as a 5E nSt naSrn wll" ' T - ." Those elements for which the mutual coupling is low. The power coupled out of the guide by the slot is a function of the angle at which the slot is cut. Polyrods helices spira or logarithmically periodic-. When slots or dipoles spaced half wavelengths along the walls of a waveguide are fed in a series lashion. Almost any type of radiating element can be used as the building block of an array antenna. Detailed descriptions of the various radiators used for antennas may be found in the standard texts on antennas and will not be discussed here However. more directive elements can be used. the phase of the elements must be alternated along the array since the field V Delay line Delay line V Delay line w T Variable frequency signal - V V T&r v V Delay line - V -A fn — • VOlUmetriC SCa " ning f a lanar arra usin ° P y g eievadon fre q uency «=an in azimuth and phase-shift scan !!£? phase of the ' ^f is TJT°' The coverage direction. .radiators have been directive elements are desired. than.— 314 Introduction to Radar Systems [ Sec 7 7 In shipboard use. Another simple element related to the dipole is the slot cut into the side of a waveguide.

The element spacing in the length of the array under these conditions is 1 00A.49. 7. it may not always be pattern. in antennas with modest requirements on of the beam is desired. the unequally spaced array can be analyzed in terms of an equivalent whose pattern is uniformly spaced array. Therefore beamwidth "thinning out" by removing elements decreases the gain even though the unchanged as eleremains beamwidth If the might remain essentially unchanged. the at is 7. The advantages of considered. to prescribe are based on trial and of optimum configurations is a difficult task. In array.65°. is scan angle arrays may be discrete-element of precise aperture control which is characteristic 107 negated by mutual coupling. with a nonuniform amplitude distribution. However. ±30.49 Fig. The axis. array consequently the radiation pattern is also symmetric about its show the relative phase to axis the Z side of either on plotted is pattern radiation relations between the various sidelobes. The beamwidth of the 25-element thinned-out array of Fig. thinned-out in a ments are removed compensate for the decrease in gain. be must effects mutual-coupling large. One method of obtaining an array with unequal spacing is to remove elements pseudorandom from an equally spaced array. obtained that with beamwidth would be comparable and ±90°. unequal the of smallest the is sin 6 ). = ±48. a function of the scan angle. Most designs at . at appear would lobes but grating . or if the positioning precise if or but if extremely low sidelobes.106 Mutual coupling can be accounted for in the design of the array. of A 100A array with 200 elements spaced A/2 apart has a theoretical beamwidth 51°. B ±14. mismatched and a level. are usually physically large enough broadening due to diffraction by adjacent elements.5. More controlled techniques of a synthesis The spacings. tends to unequally rather than equally (with a spacing greater than a half wavelength) produced. a raised radiation poor in a result can elements mutual coupling between coupling is not important mutual general. The abscissa Z is a universal pattern A is the spacings. 7. between elements. space because practical to do so. 109 array. be which the main beam is steered. be otherwise would that lobes grating "smear" the undesired An example of the radiation pattern of an unequally spaced linear array of 25 factor equal to elements is shown in Fig. it is Large 108 Two advantages claimed for that might be of advantage in certain applications. Arrays with unequally spaced elements.5. sidelobe sidelobe-level or beam-position accuracy. Spacing the elements of a thinned-out array error. beam-steering of or frequency analyzed either as a function of - an abscissa scale in degrees which applies to the case of 6 = and dmin = 2A. where dmin M„inM)(sin and d is the angle to array. to increase must level sidelobe average the array. spaced unequally the the best mean-square representation for The gain of an array is proportional to the number of elements it contains.q 7 71 Antennas 315 to cause significant pattern directive polyrod. If aperture with 4A spacing. The closest center of the array.49 is 0. have equal spacing Unequally Spaced Arrays. shown in antenna whose computed pattern is from the array outward manner controlled in a monotonically The spacings increase The element spacings are symmetrically placed with respect to center of the center. the the same 25 elements were equally spaced over a 100A with a full aperture of 200 elements. input impedance of a Mutual coupling between the elements of an array causes the it would have if isolated in free impedance the from different be to radiating element but 103 . the normal to from the measured angle 6 is the wavelength. element the random nature may be used. The vast majority of array antennas have properties however. Plotting the pattern in terms of Z permits it to is shown Also angle. used as compared with an can be elements fewer that are arrays unequally spaced broadband operation is equally spaced array of comparable beamwidth and that possible. however.

where is the number of elements. each radiating the same power. 7.0 1. ^(sintf-sintfo) {Courtesy Electronic Computed radiation pattern of a 25-element unequally spaced array.7 It has also been shown that the sidelobe level of a nonuniformly spaced array with uniform excitation of the elements theoretically can be reduced in height to approximately 2/N times the main-lobe level.6 J 0. Communications. 174 A linear array of 266 uneq ually spaced elements replaced a 388 equally spaced element array.6 0. without N increasing the beamwidth of the main lobe. 182 The principle of unequally spaced thinned-out arrays has been applied to radio- astronomy telescopes to effect an over-all increase in economy without a significant decrease in beamwidth.316 Introduction to Radar Systems [Sec. This is an example of how unequal spacings can effect a "space taper" 20 S.2 ' i i 0.2 1. 7. Radar Applications of the Array.8 1.4 0. whereas the reflector antenna must have some depth in order to support a feed at a distance from the main part of the antenna.8 2. One important . application of the array in radar has been as a fixed-beam antenna scanned by the mechanical rotation of the entire antenna structure. The element spacing was chosen to approximate a cosine-squared aperture illumination.3° beam was obtained which could be scanned ±30°. The mechanically scanned array has the advantage of a compact structure as compared with an equivalent reflector or lens antenna. The array can be made relatively flat.0 /= Fig. Inc. In transmitting applications space tapering permits a form of tapered aperture illumination with vicinity of the identical transmitting elements.2X 1.) across the array aperture with the same approximate pattern properties (at least in the main lobe) as if an amplitude taper had been used. At the lower frequencies only a relatively small number of antenna elements are needed to obtain an array with large receiving cross section. A 0. deg.49. Large mechanically rotated array antennas are more competitive with other antenna types at the lower radar frequencies than at the higher microwave frequencies. for 30 40 50 60 70 90 d m \„ .

X A The beamwidth was 0. antenna for shipboard radar The compact size of an array makes it an attractive 111 rectangular-slotteddescribe a 12-ft-long Byers and Katchky applications slots operating at inclined alternately 128 of consisting waveguide linear-array antenna because of its antenna desirable especially an is waveguide The slotted band the slots between spacing the To avoid resonant effects. . . The beam vertical plane hence the antenna was basically . of angle restricted is a reducing grating lobes in this manner The radiation gain of 16. - arranged vertical vertical plane column was fed in phase so as to provide more directivity in the steering was applied in the beam No alone.7] „_. for radar application was Another example of a rotating linear-array antenna 112 a linear array of 80 waveguide of consisted antenna The al. . 30° flared horn was used in the vertical plane to obtain more directivity One of .Sec. rhombic antennas extending f which stood for multiple-unit steerable antenna. rotated at 1 rpm. was scanned omnidirectional elements were used and 2 68 wavelengths in the vertical dimension. The frequency of operation was 5-band.4 db antenna! the directive polyrod elements each had a sidelobes with 12 db down. grating radiate little or no energy in the direction of the the In scan.7 by 16°. in the horizontal direction phase shifters scanned the by beam with The rotary-waveguide 6. application particular sidered a limitation for the in elevation. Six of 16 m and could be wavelength a at 2.5° beamwidth endfire mile in length generated a The array was later shifters. phase the 23° angle by elevation and 12 scanned between than less beamwidth 1 increased to 16 rhombics extending 2 miles with a War II in a microWorld during applied was The MUSA beam-steering technique polyrod elements 115 116 of array an 42 was antenna The MUSA. grating lobes produce grating lobes at wavelengths 2 spaced elements measurements. 7. coaxial-line power reported. of ±9 pattern of the polyrods permitted a scanning range db. and the entire structure 28° in azimuth and 1 each. element with one than would be obtained a 14-element linear array. For example.5° in the vertical direction.. simplicity and compactness. elements that These grating lobes may be removed with directive 30 and ±90° in disadvantage The lobes. By making the center of the array. in the horizontal plane with rotary-waveguide in the horizontal apart wavelengths 2 spaced were elements dimension + FH MUbA MUSA aboard large Navy ships. wave radar called FH Each of the three elements in a in three horizontal rows of 14 elements each. If appear and give ambiguous angle would in an array with this spacing. was made slightly different from a half transmission line waveguide in a occurs which 180° reversal phase to accommodate the array with nonresonant slot spacing. The beamwidth was . pattern antenna The dividers. The total gain of the antenna was 29 fire-control purposes radar operated at S band and was used for The FH a fan beam 2 generated and length in ft The antenna was 1 The polyrod phase shifters. a uniform motion at the rate of 10 scans per second. were of 25 db or better over a 35 per cent bandwidth which steered the beam with The first application of a stationary array antenna Telephone Laboratories for Bell the by developed was mechanical phase shifters 113 The array was given the name MUSA short-wave radio reception in the early 1930s.. radar surveillance SCR-270 a ground-based of dipoles arranged in four columns of eight elements Its antenna was a planar array 0° in elevation. inclined alternately were slots The wavelength. Antennas 317 mechanically rotated array antenna was the the early radars which used a operating at a frequency of 106 Mc. and the length of the horns and rigid waveguide of network complex by a The waveguide horn elements were fed Sidelobes a was 1 i by 30 fan beam. the beam every half wavelength. but it was not conmechanical pointed slightly to one side of the for which the antenna was designed. et McCoy described by array was 20 ft.

A change in beam is accomplished by mechanically display to indicate beams.5° in azimuth and is shaped in elevation to give an approximately cosecant-squared coverage down to 70° angle of depression.-band log-periodic antenna elements The array has provision for 8. while electronically controlled phase shifters can scan a beam at rates many orders of magnitude greater.: applied to ground-controlled-approach (GCA) radar. - acoustic arrays. waveguide dimensions changes the phase velocity of the radiation along the guide and hence changes the phase between the dipoles. There is close analogy between radar array techniques and elevation ESAR (Fig. A particular aperture distribution may be more readily obtained with the array than with the lens or the reflector since the amplitude and phase of each element of the . 1. Mechanically actuated phase shifters can scan the beam through its coverage as fast as 0. 2. a high-resolution Z-band. scanning beams or both at the same time. Even so. The motion of the scanning technique is reciprocating rather than rotaryconsequently perfectly uniform beam scanning can only be approximated ' moving one wall of the waveguide. The beam can be scanned in less than 20 //sec. At resonance the becomes quite large.*". Scanning of the t Sec 77 Another electromechanically steerable antenna developed during World War II was the Eagle scanner used in the AN/APQ-7. 4. five rows in can be generated by the ESAR system A separate transmitter feeds each of the /. 1 sec or better. while the next is from +1 to -30° As the beam scans through the perpendicular to the antenna." The Eagle scanner is a 16-ft linear array of 250 dipole radiators mounted with half-wavelength spacing on a waveguide feed line. Large peak and/or large average powers may be obtained with separate transmitters at each of the elements of the array. The array might generate fixed beams. The beam can be made to scan a 60° azimuth sector in f sec.12) is an example of an electronically steerable array radar using a frequency-conversion phasing scheme."* The same scanning principle has also beef 318 Introduction to Radar Systems The beamwidth is 0. Simultaneous-lobing (monopulse) tracking beams can be formed. The beam is scanned through an angle of ±30 from the perpendicular to the array by alternately feeding the array from opposite ends. Advantages and Limitations. it has been successfully manufactured for use in operational radars. The array antenna has the following characteristics not generally enjoyed by other antenna types 1. The steerable feature of an array means that the beam from a shipboard or airborne radar may be stabilized electronically rather than by mechanically moving large 5. or a single beam might be programmed to generate conical-scan tracking 3. One scan is from -1 to +30°. cluster of 25 scanning One of the major limitations of the Eagle scanner is the mechanical precision required in its construction. Electronic scanning techniques have been applied to acoustic arrays for sonar 1 9 detection. Since the phase reverses along the waveguide every half wavelength adjacent dipoles are reversed. ground-mapping radar for navigation and bombing. The beam may be scanned continuously or moved discretely from one point in space to any other point in space. The beam from an array can be rapidly scanned over the coverage of the antenna without the necessity of moving the entire antenna structure.768 elements. the array becomes resonant because of the half-wave spacing between elements. both airborne and ground-based. The large VSWR is used to place a mark on the cathode-ray-tube the center of the scan and calibrate the position of zero degrees VS WR A and five columns in azimuth.4 to 0. The antenna is 50 ft in diameter. The array has the ability to generate simultaneously many independent beams from the same antenna aperture. The basic Eagle mechanism is capable of scanning rates as high as 20 times per second.

r.7] 319 The ability to control the aperture distribution array can be individually controlled. this is limited by mutual angle scan the importantly. for example.Antennas Sec. a single plane single plane aperture. by the radiation pattern of the much Practical arrays might scan ±30° without to avoid grating lobes. provided other gain any negate not do networks. it may surface of a sphere or a cylinder in The elements could conceivably be arranged on the not all the elements of the sphere However. are gains high when not usually competitive in cost bands or the lower economical for radars operating in the Ihe probably more than at the higher microwave frequencies. If wider coverage is necessary.000 hr. per one of failure at the average rate the failures degraded. cheaper be even might antenna lower-frequency value means that the array is reasonable to a The need to keep the number of elements The array antenna is required.. amplifiers is necessary that the phase of In order to achieve small. The cost probably will antenna (gain) Hence the same beamwidth the number of elements. beam-forming losses in the phase shifters or the obtained from the absence of spillover. order to obtain more complete direction because of the self-shielding ot can be used to generate a beam in a particular the spherical array itself. the lower-frequency antenna is larger than a similar one at higher frequency. significantly accumulated before the performance of the radar is must eventually be found and replaced. ™<>ntc elements. elements that make up an array and by the coupling. 7. amplifiers must not conditions and the voltages applied to the loop to servo-controlled of sort some use to maintaining stable phase conditions is to Markow array.. operating need to maintain phase stability even under adverse phase only the that assumed was it preceding analysis and discussion of array antennas It designer. radar the by introduced knowingly changes were those deliberately and other and mixers. phase-sensitive output with that of the input in a i VHF UHF . the transmission lines. wavelength half spaced a are that they large number of elements tail. more but shape changes with scan angle. However. almost 1 million while an array with a 0. limited coverage available from a One of the disadvantages of the array antenna is the should be able to scan a aperture Theoretically.000 assuming elements. such that of other antennas. makes it theoretically possible to achieve an or lens antennas. . Larger scan angles difficulty. a failures may be such many Although hour. coverage. possible. of are. antenna with low sidelobes. the environment in which approach One vary. of an array is the An important factor which contributes to the cost and complexity In all the conditions. the life of the components is single beam. the antenna beam since practical seldom is complete hemisphere. A r*u a limitations to the widespread use of the Cost and complexity are perhaps the biggest o proportional of an array is roughly array antenna in radar applications. be obtained with more than one aperture. rw. the finite. as at the higher frequencies even though cost about the same at the lower frequencies Se . course. relatively a even array will be degraded but little when might a conventional radar with a as once. the of networks major the through maintain constant the phase shift maintains the phase between the output describes a servo phase-control system which 2° by comparing the phase of the amplifier and the input of a UHF amplifier to within detector. components used in the array be constant or negligibly at constant maintained must be operates radar the this ideal. but is absent in the 6 Spillover loss is common in the reflector antenna can be slightly higher than array an of For this reason the efficiency arrav as losses characteristic only of the array. would experience a 1° array beamwidth average life of each element were 10.1° beamwidth requires performance of the many-element The apart." at all "die not does Thus the array antenna If. and has more effective antenna area. An array which generates a 1° beamwidth requires approximately 10. However.

and produced and assembled automatically Perhaps the most promising area for future component development is that of solid-state devices. reliable. The discussion in part. Pattern- when the number of elements of the discrete antenna is most this section applies. where A(x. If the array is to compete economically its components must be cheap. for the must be a further restriction in that aperture distributions which give rise to large reactive-power components are to be avoided. d e The Fourier-integral Synthesis. to linear one-dimensional apertures or to rec- = A(x)A(y). Large reactive power is characteristic of supergain antennas and results in excessive losses and narrow of finite dimension. Although it cannot steer as rapidly as an 320 Introduction to Radar Systems of radar applications. or for that matter the radar system of the future. The radar antenna of the future. The distribution A(z) across a continuous aperture was given by Eq. methods may be divided into two classes. The element spacing is generally taken to be a half wavelength. If the elements were not isotropic but had a pattern E£6). The current distributions derived for continuous apertures may sometimes be used to approximate the discrete-aperture distributions and vice synthesis versa. the more or less conventional mechanically scanned antenna has been able to meet most of the requirements of radar. 64 . 7. multiple beams & of the array. tangular apertures where the distribution is separable.electronically scanned array. 4(z) = J_ E(4>) exp f -j2n * sin <f>) d(sin c/>) (7. and if the desired over-all pattern were denoted E a {6). the resulting antenna pattern would be £„(<£) = J ^(z)exp (pTr^sin <f>\ dz (7. practical antennas There is bandwidth (high 0. the pattern to be found by synthesis using techniques derived for isotropic elements would be given E (6)/E (d).41) . 121 synthesis techniques which apply to array antennas usually assume uniformly spaced isotropic elements. 7.8.y) All pattern-synthesis methods are approximations since large. depending upon whether the aperture is continuous or discrete. The radar systems engineer must await the efforts of the research scientist and the component-development engineer before the lull theoretical potentialities of large-array antennas can be economically exploited for general radar application. (7. might well be an all-solid-state device.14) where z E(<f>) = distance along aperture = radiation pattern If only that portion of the aperture distribution which extends over the finite-aperture dimension d were used. The array will be used when it can perform a function better and/or more cheaply than other competitive antennas.14). Pattern Synthesis it has been able to scan as fast as required for the majority Similar statements can be made for the property of forming from the same aperture as well as for most of the other stated advantages The problem of pattern synthesis in antenna design is to find the proper distribution of current across a finite-width aperture so as to produce a radiation pattern which approximates the desired pattern under some condition of optimization. r SEC 7 8 Even though the steerable array is often more costly and more complex than other antennas.2. it would undoubtedly be used where its high-speed beam-steering or multiplebeam-forming capability is needed and cannot be obtained by any other means However. The Fourier-integral relationship between the radiation pattern and the aperture distribution was discussed in Sec. that is.

consists in reconstructing the antenna pattern from in The principle is analogous to the sampling theorem of circuit theory number of samples. The least-square criterion of the Fourier-integral The Fourier series may be used to synthesize m . Woodward-Levinson Another method is but one technique upon which antenna synthesis can be based.) g 7 g-i Antennas 321 the variable of integration from Substituting Eq.(«A/<0 with a Hrf/A) sin about each of the sampled values. distribution aperture finite with a method of approximating the desired antenna pattern a finite number of sampled values. the approximate is EM-IT W WAXsMaW _ J.14) into the above and changing becomes pattern radiation antenna to £ to avoid confusion. radians Ijd spaced points sampled the and E(</>) Fie 7 50a is shown the antenna pattern pattern. theory as given by Shannon is: It a information The classical sampling theorem of it is completely determined by Jfcps. 122 Fourier integral may be used to synthesize the pattern of a practice to The Fourier-series method is restricted Similar conclusions apply. Method. than higher frequencies function f(t) contains no seconds apart.[ \ °° £(!) exp j2n . where y> f. is the Fourier method elements in the array). giving W radiation pattern £„(<£) from sampling process applied to an antenna pattern is that the by a series of values spaced determined completely d\% aperture an antenna with a finite In n is an integer. that is." The analogous spaced 1/2 its ordinates at a series of points ' which a time ' 125 . wavelength produce undesired grating lobes.EQ) si" sing)] d| (7 43) . the desired pattern changes decrease in magnitude as overshoot (Gibbs's phenomenon). m The antenna pattern pattern of the form The (sin y>)/v> function Ea is (sin y>)/y can be constructed from the sample values £. waveform of limited bandwidth may be reconstructed from a finite by Levinson introduced was values sampled on The antenna-synthesis technique based apparently developed was and forties early the in Laboratory Radiation at the MIT 5 124 independently by Woodward in England. that used in same as the and is called the composing function (<f>) = . the pattern of a discrete array. and the approximate patterns is a greater the number of the (or aperture the larger The optimum. the better will be the approximation. are shown The sampled values Es {nXjd). just as the continuous aperture.42) antenna pattern Interchanging the order of integration. Closer spacing wavelength. by the sample values E^d). but approaches a value £ (<£) is discontinuity. value ot Whenever the desired antenna pattern has discontinuities or whenever the oscillatory an in results method Fourier the rapidly.(sin <f> - sin |) d| dz (7. which determine the antenna apart. the £o($ = m . half of a vicinity the in spacing arrays with element 126 127 Spacings larger than a results in supergain arrays which are not practical. aperture finite radiation pattern £(<£) when A(z) is restricted to a derived on the pattern antenna 121 the to approximation has shown that the Ruze method for Fourier-series the (or antennas continuous for integral basis of the Fourier the desired between deviation discrete arrays) has the property that the mean-square that square) mean (least sense this in It is minimum. The overshoot does not about of 9 per cent of the total the aperture is increased. (7. . where Kid radians apart. A oo ^(d/AKsin <£ sin f where approximates the desired the Fourier-integral radiation pattern which of dimension d.

(A) sampled values EjinX/d). which specify the antenna pattern (cTrecon"C ° mP ° Sing fUnC "° n the desiS Sdiadon '° a PP roxima ^ "^ ^ ^ * K) - The (sin v»)/y composing function is well suited for reconstructing the pattern Its value at a particular sample point is unity. the antenna pattern from a finite aperture is reconstructed from a sum of (sin wMw l\ i 1 i i 1 1 >.14). (7. In n UnCt n Can b£ readil y generated with a uniform aperture J . distribution The Woodward-Levinson synthesis technique consists in determining the amplitude and phase of the uniform aperture distribution corresponding to each of CS and Perf° rming a summation t0 obtain the required over-all aperture totribulion dSS. S ^I ™y be fou "d by substituting the antenna pattern l n -T'T (/.-.322 Introduction to Radar Systems is [Sec ? g antenna pattern given by oo Ea (4>)= 2 (nl\ E. Eq. but it is zero at all other sample points. The aperture . ! l i 1 *—^*l_J 1 i 1 ! 3X 2X d 3 d- X 2X 3X 4X 5X sin d ( d d d # i) E s Kn\/d) 1 | 3X _2 X 4 c l 12 2 X 3 X 4X 5X d d sin ^ (*) Stlern^T" Fig 7.44) that is. sin |>(rf jd/X)(sin -nidi A idI X)(sin <f> 4> - nXld)~\ Ad/ - nl/d) (7.50 («) Radiation pattern E(</>) with sampled values spaced X/d radians apart where d aperture dimension.44) into the Fourier-transform relationship given by istribution T £ of Eq.

If \N\ d\K the antenna will have 1 \Nl\d\ 12 *. . the least-mean-square criterion is not 121 preference on theoretical no commands it Ruze. or -1 < sin < 1. shape and cannot be specified by simple analytical before the aperture even pattern synthesized the of latter permits one to "see" the nature to obtain the desired If necessary. to According is ^The^Fourier-integral method useful when the antenna pattern can be specified The be readily performed analytically and when the integrations can pattern to be approximated is of a complicated the when useful more is method Levinson The flexibility oi the expressions. aperture of width d is 2d\l approximate the radiation pattern from a finite .„ nl Sec. the across radians of nix the desired antenna pattern £(g is The number of samples needed to approximate Therefore W2. N < f < > Woodwardbetween Fourier-integral synthesis and the mean-square whose pattern radiation a Levinson method is that the former gives and the Woodward-Levinson method deviation from the desired pattern is a minimum. antenna synthesizing "optimum" method for aperture distribution An important synthesis problem in antenna design is to find the for a specified sidelobe beamwidth narrowest that produces a radiation pattern with the broadside symmetric for Dolph given by was problem this level The solution to 128 The optimum aperture in phase. Fourier ward-Levinson synthesis technique is not based on the error than the mean-square greater therefore possesses a The essential difference in the least-square sense Fourier synthesis. reproduced and the level of the sidelobe 133 The Fourier and Woodward-Levinson tech- Woodward- Dolph-Chebyshev Arrays. . 323 ^"U-Mvj^l-V) ^ i nirnzX 7 45) 4V (7 / ' > generates the „th (sin vO/v OTm P 0S Therefore the aperture distribution which jJf sampled value Es {nlld) is proportional to the pattern has uniform amplitude and are patterns composing the individual The phase across the aperture is such that is here beamwidth the (where beamwidth Solaced f'om one another by a half a The nulls which surround the mam beam) defined as the distance between the two change phase linear a and represents phase is given by the exponential term of Eq. sidelobe level when the beamwidth is specified. but sidelobe level is specified. which exactly fits the desired pattern at a finite gives an antenna radiation pattern pattern between the sampled number of points. There are any number of patterns.8] distribution Antennas becomes «. determined by the condition that -n\l < <£ undesirable an is which supergain. The briefly sketched here The principle of the Dolph-Chebyshev method will be in the literafound be may distribution aperture required details of calculating the 129 . and necessarily the However. 7. .132 ture. be cannot method points of the Woodward-Levinson it is not optimum integral.. in the pattern can be changes balance between the faithfulness with which abrupt ripples. .45) aperture.™'™radiation pattern in an optimum niquesare but two methods of synthesizing ar arbitrary might serve to specify an which criteria other manner. arrays of equally spaced point sources energized between the two distance the as (defined distribution that minimizes the beamwidth is described in terms of the level sidelobe given a for nulls enclosing the main beam) half-wavelength spacing. 1 Therefore the number of samples required to condition and is to be avoided. best in all cases. adjustments can be made distribution is computed. Not Chebyshev polynomials for the discrete linear array of beamwidth when the minimum the yield only does the Dolph-Chebyshev distribution to produce the lowest shown be also can it conversely. (7. . The behavior of the synthesized Since the Woodcontrolled.

46a) in symmetrica" E2x(<t>) = 2~ZA k i-=i cos (2k l)7r — A sin 6 \ where Ak is symbol^have been the amplitude of the *th element. ViW = )t0 ( 48 d Substitutin S the g l n" given by (7. or vice versa.48) the ratio of the main beam V to the sidelobe level. The aperture distribution.324 Introduction to Radar Systems that the array consists of [Sec 7 g an even number of elements 2N. ^bes are !? determined by the oscillating portion. hJtl^uT r! T ?™ ^ . defined previously. in order to restrict tq. and the other x = CO s {<dJX) In the^radiation & N E2 y(<f>) = 2gA t cos [(2fe - 1) cos" 1 x] = 2%A k Tik *=i N 1 (x) (7.49) eZ«t nM N 2^fcCOS sides E l definiti °n ^ the Chebyshev polynomial J 2 (2k- l)7r^sin<£ = cos I (2N- l)cos" x cos { 77 -^ sin <f> (7. *). 7.51 is shown Ux 128* -256* + - + ^ 'T? ^ > * ^ ^th/So^S < < mum amplitude of the sidelobes. (7. the characteristics of the optimum Dolph-Chebyshev antenna pattern can be seen from an examination of the Chebyshev polynomial. the pattern is Dolph™ is equated to a Chebyshev polynomial the a minimum for a radiation pattern is that the beamwidth The Dolph-Chebyshev E2n(</>) where x is = T2N _ 1 (x x) = T2N-1 2 x cos it d. number of elements of this equation are polynomials of degree 27V - 1. times the maxij= It was assumed in the above that there was an even number of elements in the array Similar expressions can be derived for the case where the number of elements is odd Qualitatively. or one Ak less than the specify the aperture and may be found from 160* 32* 1. and the main beam is .50) A Both constituting the array. while the side iltTl . is completely specified from either the beamwidth or the sidelobe level once the number of elements is given. The coefficients distribution necessary to obtain the optimum radiation pattern trie above equation. — sin . (7. and hence the antenna pattern.47) Tn(x) = cos (« cos" 1 x) for |x| < Equation (7.46a) may therefore be expressed as a polynomial of degree » and the properties of known polynomials may be used to synthesize has shown that when the radiation pattern resultmg pattern is optimum in the sense specified sidelobe level. pattern may be found by summing the patterns from elements taken pairs about the center of the array. It oscillates between the values ±i a " d nc eases monotonically for x The argument x x is used in 1. In Fig. a thC Dol P h -Chebyshev distribution is g characterized by having all its sidelobes of equal magnitude.47) gives the following: by (7.46b) where T^_ x (x) is the Chebyshev polynomial of degree 2k polynomial of degree « is defined as — 1 1 133 The Chebyshev (7. The radiation pattern is therefore Assume The radiation p^Sdto (7. ! Fa 7 4^ instead of* .48) x to the range -1 x 1 required bv its aCC ° rdi " 8 * detrmnTed bTth" 1? ^creasing portion ™° not ° mca y «y of the polynomial. X related to r/. If de is the element spacing.

or edge. . b in the expansion (a For a six-element array the relative amplitudes = + N m applied to the elements would be proportional This is called the binomial to 1. The pattern is interferometer with spacing d similar to that produced by a two-element discussions sidelobes which previous (In beam. 70 A /d 65 \/d - distribution proposed by John 134 It is not often used in practice Stone Stone.51. main the to All sidelobes are equal lobes) grating were equal in magnitude to the main lobe were called sidelobes do not exist.Antennas Sec. of equal magnitude. distribution. 7. The uniform distribution. is Side-lobe level. These values are function given in Fig. r. 1 . 7. when co. where of a. If directive pattern and he element pattern. The Dolph-Chebyshev dis- x 50 \/d 25 30 35 db tribution includes all distributions between the binomial and the interferometer. the resultant element the by multiplied increasing sidelobes so that.52. 10. 7. Dolph-Chebyshev arrays... because of its relatively wide beamwidth and and was first H ! 60A/2 55A/2 the large current ratios required across the aperture.52 as a dimension ^ 12°). howspecial case of the Dolpha not Approximate beamwidth Fig.(The half-power beamwidth of an array vaHd for small beamwidtli { 6 B < aperture tolxaminey When . elements is the number of A)*" . 5. for Chebvshev distribution because its sidelobes are not Dolph-Chebyshev array pattern assumes The uniform sidelobe level produced by the elements were used. the over-all an antenna with isotropic elements.8] 325 Dolph-Chebyshev array of The broadside half-power beamwidth of a of sidelobe level. degree Fig. = 1 . especially when the number of ele- ! ments is large. that is. ever. 7. array the of pattern would be equal to the product A small reduction in beamsidelobes. 10. Chebyshev polynomial of 8.U) is concentrated at the edges. the pattern array pattern consisting of an with width can be obtained by designing the antenna factor. the energy across the aperture 7-. . decreasing have would In general. when 135 pattern has uniform sidelobes. The other limiting case occurs when to the coefficients proportional are elements array the on This occurs when the currents 1 the array. 5.

Taylor Distribution. The antenna pattern produced by a Taylor distribution has uniform sidelobes. Dolph-Chebyshev array approaches the radiation pattern of a -Af hW ~ [cosh t^A'-u^ costt(u E(d)) -. the sidelobe level decreases with increasing Hence ±n divides <f>. where n is a finite integer. 7. Van der Maas 130 has shown that as the number of elements of a Dolph-Chebyshev array is increased. One example is ^ M TZ? + = T7 = n J - + - V V I W V shown in Fig. the currents in the end elements of the array become large compared with the currents in the rest of the elements and the radiation pattern becomes sensitive to changes in the excitation of the end element. y In the region where lid/A) sin <f>\ n. 7. The region in which the sidelobe level is ls defined h sin 0| < n. For a given beamwidth. . Using as a basis the relationship T2m 2{Tm f 1 2(Tm V2/2)(Tm V2/2).51) Hn m in the region u A An infinite number of equal sidelobes appears in the region « 2 > A 2 This ideal radiation pattern is physically unrealizable because of the behavior of the radiation pattern in the remote sidelobe region. 7.The two-way pattern for P a radar antenna ? of of 1 elements designed with a Dolph-Chebyshev distribution is (T )* Although this is a polynomial of degree 2m. (7. Unlike the Dolph-Chebyshev pattern. > MV . the equivalent one-way minor-lobe improvement is distribution.™ Proc. the sidelobes of the Taylor pattern decrease outside a specified angular region.53.326 Introduction to Radar Systems h lph ' Ch [Sec. In radar however it that ° ™P™ tanc *. IRE. the transmitter On reception. but only in the region of the main beam. 2 2 . it is not equal to the Chebyshev polynomial of the same degree (T2 and the two-way pattern does not represent the optimum relationship between beamwidth and sjdelobe level. The power divider apportions power in accordance with Tm + a/2/2. the isolator introduces sufficient attenuation in the reverse direction to correspond to the According to Mattingly. Taylor137 has shown that design procedures may be obtained for approximating the ideal radiation pattern of Eq.53. as the number of elements approaches infinity. I 2 2 = (djX) sin d = aperture dimension = angle measured from normal to array cosh -n-A = sidelobe ratio where u <j> <f> A2 A2 <u >u 2 2 (7. (After Mattingly.51) with a physically realizable aperture distribution. Mattingly^ indicates mum two-way pattern can be achieved with antennas in which slightly that an optidifferent transmit and receive patterns are obtained using nonreciprocal devices. just as does the Dolph-Chebyshev pattern. Nonreciprocal array with approximately 4 to 5 db.8 ySheV distribution ives an optimum one-way § pattern in the sense thl that X?u the beamwidth is a minimum for a specified sidelobe level. two-way Chebyshev pattern. The main lobe appears < . and the corresponding aperture distribution contains infinite peaks at the edges of the antenna However. This sets a practical upper limit to the number of elements which can be used in a Dolph-Chebyshev array and therefore sets a lower limit to the width of the main beam which can be achieved in practice. Chebyshev design will improve the beamwidth by about 10 per cent over the conventional Chebyshev design. Tm — V2/2 Power divider the extended Duplexer I | JReceiv |Tronismitter Fig.) In the limit.

to rapidly. Dolph-Chebyshev distribution. which defines the boundary In selecting the integer «. 138-139 Modified (sin -nu)\-nu Patterns. —40 db. the factor a = V^ + 2 n (n (7. sacrifice some beamwidth and low near-in sidelobes for sidelobes which decay A one-parameter family of line-source distributions suitable for radar applications was suggested by Taylor for achieving radiation patterns with a main lobe of adjustable 140 amplitude and a sidelobe structure similar to that of a uniformly illuminated aperture. synthesizing the pattern of circular. 7. Taylor states that n must be at least 3 it is essential to avoid values that are too small.8] Antennas 327 the main beam and a the radiation pattern into a uniform sidelobe region surrounding decaying sidelobe region. the sharper will be the beam. in certain instances. interfering signals would be more likely to cluster They would be easier to recognize as false targets in the vicinity of the main beam. .52) - \) The value of n does not have to be very large in order to make a only a few per cent For example. if 40 db antenna gain is required. the average sidelobe level must be less than sidelobes it is since is preferred over an equal-sidelobe pattern is in low-noise applications important that the portion of the radiation pattern which illuminates the relatively "hot" ground be kept to a minimum. the same The distributions difficulties as arise in a Dolph-Chebyshev distribution will occur. The Taylor distribution does not produce an optimum pattern as does the DolphDolph-Chebyshev array by Its beamwidth will be broader than that of a Chebyshev. Hence it may be better. The larger n is. that is. A value of n = 8 gives a difference in beamwidth of 5.5 per cent. However. that within the main beam.7 per cent greater than the optimum by the ideal. if n is too large. the region of uniform sidelobes and decreasing sidelobes. applied to also been has distribution The Taylor two-dimensional apertures. The radiation-pattern synthesis technique in which intensity all the sidelobes are of equal intensity (Dolph-Chebyshev) or of almost equal In certain (Taylor) may not always be desirable from an operational point of view. produced bution with n = 5 gives a beamwidth 7.Sec. for high values of h are peaked at the center and at the edge of the aperture. Narrow-beamwidth antennas with Taylor distributions can be realized without significant reduction in gain by 164 properly choosing the value of n as described by Hansen. Very large antennas with narrow beamwidths may exhibit a compared with severe degradation in gain because of the large energy contained within the sidelobes as This may be avoided by requiring the. but unobtainable. average sidelobe level to be less than the gain. because of the symmetry of the antenna pattern than if they appeared far removed from Another example of where an antenna pattern with rapidly decaying the main beam. radar applications it may be of advantage to have the sidelobe level decay rapidly on For example. Care must be exercised in the selection of the sidelobe level of a Taylor or a DolphChebyshev distribution. while low values of n produce distributions which taper from a maximum value at the aperture center to a minimum at the edge. for a design-sidelobe ratio of 25 db and at least 6 for a design-sidelobe ratio of 40 db. a Taylor distrigreater than unity. interfering or spurious signals which enter either side of the main beam. increasing angle from the main beam. the radar receiver via the sidelobes might appear from any angle when the antenna If the antenn