RADIO SYSTEMS – ETIN15

Lecture no:
5
Digital modulation

Anders J Johansson, Department of Electrical and Information Technology

anders.j.johansson@eit.lth.se

3 April 2017 1
 Contents

• Brief overview of a wireless communication link

• Radio signals and complex notation (again)
• Modulation basics
• Important modulation formats

3 April 2017 2
 STRUCTURE OF A WIRELESS
COMMUNICATION LINK

3 April 2017 3
 A simple structure

Speech Speech Chann.

A/D Encrypt. Modulation
encoder encoding

Data

Key

Speech Speech Chann.

D/A Decrypt. Demod.
decoder decoding

Data

(Read Chapter 10 for more details)

3 April 2017 4
 RADIO SIGNALS AND
COMPLEX NOTATION
(from Lecture 3)

3 April 2017 5
 Simple model of a radio signal
FR R
O M EP
E
• A transmitted radio signal can be written LE ATE
CT D
UR
E
3

Amplitude Frequency Phase

• By letting the transmitted information change the
amplitude, the frequency, or the phase, we get the tree
basic types of digital modulation techniques

– ASK (Amplitude Shift Keying)

– FSK (Frequency Shift Keying)
Constant envelope
– PSK (Phase Shift Keying)

3 April 2017 6
 Example: Amplitude, phase and
frequency modulation
FR R
O M EP
s  t   A  t  cos  2 f ct    t   E
LE ATE
CT D
A t    t Comment: UR
E
3
00 01 11 00 10

- Amplitude carries information

4ASK - Phase constant (arbitrary)

00 01 11 00 10

4PSK - Amplitude constant (arbitrary)

- Phase carries information

00 01 11 00 10
- Amplitude constant (arbitrary)
4FSK - Phase slope (frequency)
carries information

3 April 2017 7
 The IQ modulator
FR R
O M EP
E
LE ATE
CT D
I-channel UR
E
(in-phase) Transmited radio signal 3

fc

-90o

Q-channel
(quadrature)
Take a step into the complex domain:
Complex envelope

Carrier factor

3 April 2017 8
 Interpreting the complex notation
FR R
O M EP
E
Complex envelope (phasor) Transmitted radio signal LE ATE
CT D
UR
E
3

Polar coordinates:
By manipulating the amplitude A(t)
and the phase Φ(t) of the complex
envelope (phasor), we can create any
type of modulation/radio signal.

3 April 2017 9
 MODULATION
BASICS

3 April 2017 10
 Pulse amplitude modulation (PAM)
The modulation process
Bits Complex domain
Radio
bm cm sLP  t  signal
Mapping PAM Re{ }

Complex
exp  j 2 f c t 
numbers

Symbol
∞
Many possible pulses
PAM: s LP t = ∑ c m g  t−m T s  time g  t
m=−∞

∞
“Standard” basis pulse criteria t
∫ ∣g t ∣2 dt =1 or =T s (energy norm.) g  t
−∞
∞

∫ g t g *  t−mT s  dt=0 for m≠0 (orthogonality) t

−∞ Ts
3 April 2017 11
 Pulse amplitude modulation (PAM)
Basis pulses and spectrum
Assuming that the complex numbers cm representing the data
are independent, then the power spectral density of the
base band PAM signal becomes:

∞ 2

∣
S LP  f ~ ∫ g t e
−∞
− j 2 f t
dt
∣
which translates into a radio signal (band pass) with

1
S BP  f    S LP  f  f c   S LP   f  f c  
2

3 April 2017 12
 Pulse amplitude modulation (PAM)
Basis pulses and spectrum
Illustration of power spectral density of the (complex) base-band
signal, SLP(f), and the (real) radio signal, SBP(f).

S LP  f  S BP  f 

f  fc fc f
Can be asymmetric,
since it is a complex Symmetry (real radio signal)
signal.

What we need are basis pulses g(t) with nice properties like:

- Narrow spectrum (low side-lobes)

- Relatively short in time (low delay)
3 April 2017 13
 Pulse amplitude modulation (PAM)
Basis pulses
TIME DOMAIN FREQ. DOMAIN
Rectangular [in time]

Normalized time t / Ts Normalized freq. f ×T s

(Root-) Raised-cosine [in freq.]

Normalized time t / Ts Normalized freq. f ×T s

3 April 2017 14
 Pulse amplitude modulation (PAM)
Interpretation as IQ-modulator
For real valued basis functions g(t) we can view PAM as:

sI  t   Re  sLP  t  
Re  cm 
g  t
cos  2 f ct 
Radio
bm cm Pulse fc signal
Mapping shaping
filters
-90o
 sin  2 f ct 
g  t
Im  cm 
sQ  t   Im  sLP  t  

(Both the rectangular and the (root-) raised-cosine pulses are real valued.)

3 April 2017 15
 Multi-PAM
Modulation with multiple pulses
Bits Complex domain
Radio
bm cm sLP  t  signal
Mapping multi-PAM Re{ }

exp  j 2 f c t 

∞
multi-PAM: s LP t = ∑ g c  t −mT s 
m
m−∞

“Standard” basis pulse criteria Several

different
2
∫∣g c t∣ dt =1 or =T s
m
(energy norm.)
∫ g c t g *c  t−kT s  dt =0 for k ≠0 (orthogonality) pulses
m m

∫ g c t g *c  t  dt =0 for c m≠c n

m n
(orthogonality)

3 April 2017 16
 Multi-PAM
Modulation with multiple pulses
Frequency-shift keying (FSK) with M (even) different transmission
frequencies can be interpreted as multi-PAM if the basis functions
are chosen as:
− j k  f t
g k t =e for 0≤t≤T s
and for k = +/- 1, +/- 3, ... , +/- M/2

S LP  f  S BP  f 

 fc fc
f
Bits: 00 01 10 11

3 April 2017 17
 Continuous-phase FSK (CPFSK)
The modulation process
Bits Complex domain
Radio
bm cm sLP  t  signal
Mapping CPFSK Re{ }

exp  j 2 f c t 

CPFSK: s LP t = A exp  j  CPFSK t  

where the amplitude A is constant and the phase is
∞ t

 CPFSK t =2 h mod ∑ c m ∫ g  u−mT  du

m=−∞ −∞
Phase basis
where hmod is the modulation index.
pulse

3 April 2017 18
 Continuous-phase FSK (CPFSK)
The Gaussian phase basis pulse
In addition to the rectangular phase basis pulse, the Gaussian is the
most common.

BTs=0.5

Normalized time t / Ts

3 April 2017 19
 IMPORTANT MODULATION
FORMATS

3 April 2017 20
 Binary phase-shift keying (BPSK)
Rectangular pulses

Base-band

Radio
signal

3 April 2017 21
 Binary phase-shift keying (BPSK)
Rectangular pulses
Complex representation Signal constellation diagram

3 April 2017 22
 Binary phase-shift keying (BPSK)
Rectangular pulses

Power spectral
density for BPSK

Normalized freq. f ×T b

3 April 2017 23
 Binary phase-shift keying (BPSK)
Raised-cosine pulses (roll-off 0.5)

Base-band

Radio
signal

3 April 2017 24
 Binary phase-shift keying (BPSK)
Raised-cosine pulses (roll-off 0.5)
Complex representation Signal constellation diagram

3 April 2017 25
 Binary phase-shift keying (BPSK)
Raised-cosine pulses (roll-off 0.5)

Power spectral
density for BAM

Normalized freq. f ×T b
Much higher spectral
efficiency than BPSK
(with
rectangular
pulses).

3 April 2017 26
 Quaternary PSK (QPSK or 4-PSK)
Rectangular pulses

Complex representation

Radio
signal

3 April 2017 27
 Quaternary PSK (QPSK or 4-PSK)
Rectangular pulses

Power spectral
density for QPSK

Twice the spectrum

efficiency of BPSK
(with rect. pulses).
TWO bits/pulse
instead of one.

3 April 2017 28
 Quadrature ampl.-modulation (QAM)
Root raised-cos pulses (roll-off 0.5)

Complex representation

Much higher spectral

efficiency than QPSK
(with
rectangular
pulses).

3 April 2017 29
 Amplitude variations
The problem
Signals with high amplitude variations leads to less efficient amplifiers.

Complex representation of QPSK

It is a problem that the signal

passes through the origin, where
the amplitude is ZERO.
(Infinite amplitude variation.)

Can we solve this problem in a simple

way?

3 April 2017 30
 Amplitude variations
A solution
Let’s rotate the signal constellation diagram for each
transmitted symbol!

 /4 2× /4
etc.

3 April 2017 31
 Amplitude variations
A solution
Looking at the complex representation ...

QPSK without rotation QPSK with rotation

A “hole” is created in
the center. No close to
zero amplitudes.

3 April 2017 32
  / 4- Differential QPSK (DQPSK)

Complex representation

Still uses the same rectangular pulses as QPSK - the power

spectral density and the spectral efficiency are the same.
This modulation type is used in several standards for mobile
communications (due to it’s low amplitude variations).

3 April 2017 33
 Offset QPSK (OQPSK)
Rectangular pulses

In-phase
signal

Quadrature
signal

There is one bit-time offset between the in-pase and the quadrature
part of the signal (a delay on the Q channel). This makes the transitions
between pulses take place at different times!

3 April 2017 34
 Offset QPSK
Rectangular pulses
Complex representation

This method
also creates a
hole in the
center, giving
less amplitude
variations.

3 April 2017 35
 Offset QAM (OQAM)
Raised-cosine pulses
Complex representation

This method
also creates a
hole in the
center, but
has larger
amplitude
variations
than OQPSK.

3 April 2017 36
 Continuous-phase modulation

Phase
MSK/FFSK
Basic idea: 2
- Keep amplitude constant
3
- Change phase continuously 
2

1 0 1 0
In this particular example 1

we change the phase in 2 1 0 1 0
a piecewise linear fashion
by +/- /2, depending on Tb 0 1 t
the data transmitted. 1
− 
2
−
3
This type of modulation − 
2
can be interpreted both as
phase and frequency −2 
modulation. It is called
MSK (minimum shift keying) or
FFSK (fast frequency shift keying).
3 April 2017 37
 Minimum shift keying (MSK)

Simple MSK implementation

0 1 0 0 1

Rectangular Voltage MSK signal

01001 controlled
pulse
filter oscillator
(VCO)

3 April 2017 38
 Minimum shift keying (MSK)

Power spectral
density of MSK

3 April 2017 39
 Gaussian filtered MSK (GMSK)

Further improvement of the phase: Remove ’corners’

(Simplified figure)
Phase Phase
2 2
3 3
 
2 2
1 1
 
1 1 0 1 1 1 0 1
 
2 1 0 1 0 1 2 1 0 1 0 1

Tb t Tb t
1 1
−  − 
2 2
− −
3 3
−  − 
2 2
−2  −2 

MSK Gaussian filtered MSK - GMSK

(Rectangular pulse filter) (Gaussian pulse filter)

3 April 2017 40
 Gaussian filtered MSK (GMSK)

Simple GMSK implementation

0 1 0 0 1

Gaussian Voltage GMSK signal

01001 controlled
pulse
filter oscillator
(VCO)

When implemented this “simple” way, it is usually called Gaussian

filtered frequency shift keying (GFSK).

GSFK is used in e.g. Bluetooth.

3 April 2017 41
 Gaussian filtered MSK (GMSK)

Digital GMSK implementation

D/A
cos  2 f c t 
Digital
Data baseband
fc
GMSK
modulator
-90o
 sin  2 f ct 
D/A

Digital Analog

This is a more precise implementation of GMSK, which is used in

e.g. GSM.
3 April 2017 42
 Gaussian filtered MSK (GMSK)

BT = 0.5 here
(0.3 in GSM)
Power spectral
density of GMSK.

3 April 2017 43
 How do we use all these spectral
efficiencies?
Example: Assume that we want to use MSK to transmit 50 kbit/sec,
and want to know the required transmission bandwidth.

Take a look at the spectral efficiency table:

The 90% and 99% bandwidths become:

B90%  50000 /1.29  38.8 kHz
B99%  50000 / 0.85  58.8 kHz

3 April 2017 44
 Summary

BPSK with
root-raised
cosine
pulses

TABLE 11.1 in textbook.

3 April 2017 45