IA ,

D T
IN EN
.
Analog Communication

A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
by
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

Dr. Amitava Chatterjee
AV I A

Electrical Measurement and Instrumentation Laboratory,

Electrical Measurement and Instrumentation Laboratory,
Electrical Engineering Department,
D TR

Jadavpur University, Kolkata, India.
JA EC
EL
Analog Signal Transmission and Reception

IA ,
D T
IN EN
.
 The purpose of a communication system is to transmit

A , TM
information-bearing signals through a communication

O E E
AT R
, K G D JE
channel
h l separating
ti g the
th transmitter
t itt from
f the
th receiver.
i

LK PA
TY IN ER
SI R TT
ER E A
IV IN H
 A large number of such information sources are analog
N G C
E
p
sources. Speech, U N A
, image
g and video are examples
p of analog
g
R L E AV
sources of information.
PU CA MIT
AV I A
D TR

 Each analog source is characterized by its bandwidth,

d
dynamic
i range, and
d th
the nature
t off th
the signal.
i l
JA EC
EL
Analog Signal Transmission and Reception

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
Elements of a Communication System.
PU CA MIT

There are two basic modes of communication:

AV I A

Broadcasting: Involves the use of a single powerful

transmitter and numerous receivers, that are relatively
D TR

inexpensive to build.
JA EC

Point-to-point Communication: It takes place over a link

EL

between a single transmitter and a receiver.

Analog Signal Transmission and Reception

IA ,
D T
Communication channels

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
Based on guided propagation Based on free propagation

SI R TT
ER E A
IV IN H
N G C
E
U N A
 Telephone channels  Wireless broadcast
R L E AV
PU CA MIT

using twisted pairs. channels.

AV I A

 Coaxial cables.  Mobile radio channels.

 Optical fibers.  Satellite channels.
D TR
JA EC
EL
Analog Signal Transmission and Reception

IA ,
D T
Analog signal transmission

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
Baseband communication Carrier communication

SI R TT
ER E A
IV IN H
N G C
E
It does not use modulation.
U N A It makes use of modulation.
R L E AV
PU CA MIT
AV I A

Baseband refers to the band of frequencies representing the

original signal as delivered by a source of information. The
D TR

baseband in Telephony is the audio band with the range: 0 –

JA EC

3.5 kHz and in Television is the video band with the range: 0
EL

– 4.3 MHz.
Analog Signal Transmission and Reception
Features

IA ,
D T
IN EN
.
 In baseband communication, signals are transmitted

A , TM
without any shift in the range of frequencies of the signal.

O E E
AT R
However baseband signals produced by various information

, K G D JE
LK PA
sources are not always suitable for direct transmission.

TY IN ER
SI R TT
ER E A
 These signals are further modified to facilitate

IV IN H
transmission, using a process called modulation.
N G C
E
U N A
R L E AV
PU CA MIT

 Modulation causes a shift in the range of frequencies. This

is called carrier communication.
AV I A
D TR

 A carrier c(t) is a sinusoid of high frequency, and one of its

JA EC

parameters i.e. amplitude, frequency, or phase, is varied in

proportion to the baseband signal m(t).
EL
Analog Signal Transmission and Reception

IA ,
D T
Analog modulation techniques

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
Amplitude Modulation (AM) Angle Modulation

SI R TT
ER E A
IV IN H
N G C
E
Amplitude of c(t) is varied with
U N A Instantaneous phase or frequency
R L E AV
m(t). of c(t) is varied with m(t).
PU CA MIT
AV I A

When carrier frequency is modulated, it is called Frequency

D TR

Modulation (FM).
(FM) When carrier phase is modulated
modulated, it is
JA EC

called Phase Modulation (PM).

EL
 Analog Signal Transmission and Reception

IA ,
Purpose
p off Modulation …

D T
IN EN
.
A , TM
To translate the low-pass signal in frequency to the passband of
the channel so that the spectrum of the transmitted bandpass

O E E
AT R
, K G D JE
LK PA
signal
i l match
h the
h passband
b d characteristics
h i i off the
h channel.
h l

TY IN ER
To accommodate for simultaneous transmission of signals from

SI R TT
several message sources using frequency-division
frequency division multiplexing

ER E A
IV IN H
(FDM).

N G C
E
U N A
R L E AV
To expand the bandwidth of the transmitted signal in order to
increase its noise-immunity in transmission over a noisy channel.
PU CA MIT
AV I A

What is Demodulation ??
D TR

At the receiving end of the system, the original bandpass signal is

restored
t d byb performing
JA EC

f i a process called
ll d demodulation.
d d l ti
Demodulation can be viewed as the reverse of the modulation
EL

process.
 Analog Signal Transmission and Reception

IA ,
Modulation Process

D T
IN EN
.
A , TM
For all modulation processes analog signal m(t) is considered as a
low-pass signal of bandwidth W i.e. M(f)  0, for |f|> W. The signal

O E E
AT R
, K G D JE
LK PA
i assumed
is d as a power signal
i l with
i h power Pm:

TY IN ER
SI R TT
T

T mt  dt
1 2 2
Pm  lim

ER E A
IV IN H
T 
T
N G C
E
2
U N A
R L E AV
PU CA MIT

This message signal is carried through the communication

channel by
y impressing
p g it on a carrier signal
g c(t):
( )
AV I A

ct   Ac cos2f c t   c 
D TR
JA EC

Ac: carrier amplitude, fc: carrier frequency, c: carrier phase.

EL
 Review of Fourier Transform

IA ,
D T
IN EN
 The
Th Fourier
F i iintegral
t l transform
t f pair
i can be
b given
i as:

.
A , TM


G  f    g t e  j 2 ft dt

O E E
AT R
, K G D JE
LK PA


TY IN ER


g t    G  f e j 2 ft df

SI R TT


ER E A
IV IN H
N G C
where f = frequency measured in Hz. This pair can be used to

E
U N A
describe the time-frequency
q y relationship
p for nonperiodic
p signals.
g
R L E AV
The relationship between the time and frequency domains is
PU CA MIT

indicated by the double arrow, given as:

g t   G  f 
AV I A

 G(f) is specified by a magnitude characteristic and phase

D TR

characteristic:
JA EC

G  f   G  f  e j  f 
EL
 p
Properties
EL
Property

JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
f

ER E A
E
SI R TT
off the Fourier transform

TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
Review of Fourier Transform

Mathematical Description

IA ,
.
 p
Properties
EL
Property

JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
f

ER E A
E
SI R TT
off the Fourier transform

TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
Review of Fourier Transform

IA ,
.
Mathematical Description
 EL
f

JA EC
D TR
Fourier transform

AV I A
p

PU CA MIT
pairs
Time Function

R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
Fourier Transform

IN EN
D T
Review of Fourier Transform

IA ,
.
 EL
f

JA EC
D TR
Fourier transform

AV I A
p
Time Function

PU CA MIT
pairs

R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
Fourier Transform

D T
Review of Fourier Transform

IA ,
.
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
RANDOM PROCESSES

D T
IA ,
.
 Random Processes

IA ,
D T
IN EN
These are those
Th th processes iin nature
t that
th t are best
b t

.
A , TM
characterized in statistical terms. For example, air
temperature and air pressure fluctuating randomly as

O E E
AT R
, K G D JE
LK PA
functions of time, thermal noise voltages generated in

TY IN ER
resistors of electronic devices etc.

SI R TT
ER E A
Random signals are usually modeled as infinite-duration,

IV IN H
N G C
infinite-energy signals. The set of all possible waveforms is

E
U N A
called an ensemble of time functions or a random process.
process
R L E AV
PU CA MIT

The set ((ensemble)) of all p

possible waveforms
f off a random
AV I A

process is denoted as X(t,S), where t is the time index and S

represents the set or sample space of all possible sample
D TR

functions. A single waveform in this set is called x(t,s).

JA EC

Usually s or S is dropped from the notation.

EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
Random Processes

An ensemble of sample functions.

AT R
A , TM
IN EN
D T
IA ,
.
 Random Processes

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A

For a fixed sample point sj, the graph of the function X(t, sj)
versus time t is called a realization or sample function of the
random process. These sample functions are denoted as:
D TR

x j t   X t , s j , j  1,2 , , n
JA EC
EL
 Random Processes

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

For a fixed time tk inside the observation interval, the set of

numbers
AV I A

x t , x t , , x t   X t
k k n k k
, s1 , X t k , s2 , , X t k , sn 
D TR

1 2
JA EC

constitutes a random variable. This gives an indexed

ensemble (family) of random variables {X(t,s)}, called a
EL

random process.
 Random Processes

IA ,
D T
IN EN
Let us consider a random process X(t) at different time

.
A , TM
instants t1 > t2 >…> tn, n is a positive integer. In general, the

O E E
Xti  x(t

AT R
samples
p ( i)), i = 1,2,…,n are n random variables

, K G D JE
LK PA
statistically characterized by their joint probability density

TY IN ER
function (PDF) denoted as p(xt1, xt2,…, xtn) for any n.

SI R TT
ER E A
IV IN H
What is the difference between a random variable and a
N G C
E
random process
p ???U N A
R L E AV
PU CA MIT

For a random variable, the outcome of a random

AV I A

experiment is mapped into a number.

For a random process, the outcome of a random
D TR

experiment is mapped into a waveform that is a function of

JA EC

time.
EL
 Stationary Random Processes

IA ,
D T
IN EN
Let us consider we have n samples of the random process

.
A , TM
X(t) at t = ti, i = 1,2,…,n, and another set of n samples

O E E
y , as, Xti+   X(t
( i+ ),

AT R
displaced
p in time from the first set by ) i=

, K G D JE
LK PA
1,2,…,n.

TY IN ER
SI R TT
If the
th joint
j i t PDFs
PDF off ththe ttwo sets
t off random
d variables
i bl are

ER E A
IV IN H
identical i.e. p(xt1, xt2,…, xtn) = p(xt1+ , xt2 + ,…, xtn + ) for all 
N G C
E
and n, the random process is called stationary in the strict
U N A
R L E AV
sense.
PU CA MIT
AV I A
D TR

The statistical properties of a stationary random process

JA EC

are invariant to a translation of the time axis.

EL
 Autocorrelation Function

IA ,
D T
IN EN
Let a random process X(t) be sampled at t = ti. Then X(ti) is

.
A , TM
a random variable with PDF p(xti). For a stationary process,

O E E
p(xti) for all . Hence the PDF is independent

AT R
p(xti+ ) = p(
p( p of

, K G D JE
LK PA
time.

TY IN ER
SI R TT
ER E A
Let us consider two random variables Xti = X(ti), i = 1,2,

IV IN H
N G C
corresponding to samples of X(t) taken at t = t1 and t = t2.

E
U N A
The autocorrelation function of the random process X(t) is
R L E AV

measured by the expectation of the product of the two

PU CA MIT

random variables Xt1and Xt2. Mathematically speaking,

AV I A
D TR

RXX t1 ,t 2   RX t1 ,t 2   E  X t X t     xt xt p xt , xt dxt dxt

 
JA EC

1 2 1 2 1 2 1 2
 
EL
 Autocorrelation Function

IA ,
D T
IN EN
For a stationary process X(t), the joint PDF of the pair

.
A , TM
(Xt1,Xt2) is identical to the joint PDF of the pair (Xt1+,Xt2 +),

O E E
y . Hence the autocorrelation ffunction of

AT R
for anyy arbitrary

, K G D JE
LK PA
X(t) depends on the time difference, t1- t2 = .

TY IN ER
SI R TT
ER E A
IV IN H
N G C
For a stationary, real-valued random process, the

E
U N A
autocorrelation ffunction is:
R L E AV

RXX t1 ,t 2   RXX t1  t 2   RXX   E  X t   X t 

PU CA MIT

1 1
AV I A

On the other hand:

 
RXX    E  X t  X t   E X t X t    RXX  
D TR

/ /
1 1
JA EC

1 1

RXX 0   E  X t2   average power of the random process

EL

1
 Autocorrelation Function

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
The physical significance of the autocorrelation function is
PU CA MIT

that it provides a means of describing the interdependence

AV I A

of two random variables obtained by observing a random

process X(t) at times  seconds apart.
D TR

The
e more
o e rapidly
ap d y tthe
e random
a do p process
ocess X(t) changes
(t) c a ges with
t
JA EC

time, the more rapidly will the autocorrelation function RX()

EL

decrease from its maximum RX(0) as  increases.

 Crosscorrelation Function

IA ,
D T
The crosscorrelation function of X(t) and Y(t) is defined by

IN EN
.
the joint moment as:

A , TM
O E E
AT R
RXY t1 ,t 2   E  X t Yt     xt yt p xt , yt ddxt dy
 

, K G D JE
LK PA
d t

TY IN ER
1 2 1 2 1 2 1 2
 

SI R TT
The random processes X(t) and Y(t) are said to be

ER E A
IV IN H
statistically independent iff:

N G C
E
   
U N A
p xt , xt , , xt , yt , yt , , yt  p xt , xt , , xt  p yt , yt , , yt
R L E AV
/ / / / / /
1 2 n 1 2 n
PU CA MIT

1 2 m 1 2 m

for all choices of ti, ti/ and for all positive integers n and m.
AV I A

The random processes X(t) and Y(t) are said to be

D TR

uncorrelated if:
JA EC

RXY t1 ,t 2   E  X t E Yt 

EL

1 2
 The Wiener-Khintchine Theorem

IA ,
D T
Let xn be a real signal
signal. Then:

IN EN
.
A , TM
RXX l   S XX 
F

O E E
AT R
, K G D JE
LK PA
TY IN ER
This means the energy spectral density of an energy signal

SI R TT
q
is the Fourier transform of its autocorrelation sequence.

ER E A
IV IN H
N G C
E
U N A
R L E AV
This result is very important because it implies that the
PU CA MIT

autocorrelation sequence of a signal and its energy spectral

AV I A

density contain the same information.

The power spectral density or power spectrum of a
D TR

stationary process X(t) is defined as:

JA EC

S XX  f   S X  f    RXX e  j 2 f d


EL
 Power Spectral Density

IA ,
D T
IN EN
.
The power spectral density SXX(f) and the autocorrelation

A , TM
function RXX() of a stationary process X(t) form a Fourier

O E E
AT R
p
transform pair:

, K G D JE
LK PA
TY IN ER
S XX  f    RXX  e  j 2 f d


SI R TT
ER E A
RXX    S XX  f e j 2 f df


IV IN H
N G C
E
U N A
R L E AV

The cross power density spectrum or cross spectral

PU CA MIT

density of two jointly stationary random processes X(t) and

AV I A

Y(t) is defined as:

S XY  f   F RXY    RXY e  j 2 f d


D TR
JA EC
EL
 Power Spectral Density (PSD)

IA ,
Properties
p …

D T
IN EN
.
The zero-frequency value of the PSD of a stationary

A , TM
process equals the total area under the graph of the

O E E
AT R
, K G D JE
LK PA
autocorrelation
t l ti ffunction.
ti

TY IN ER
S XX 0    RXX  d


SI R TT
ER E A
IV IN H
N G C
The average power of a stationary process equals the total

E
area under the g
U N A
graph
p of the PSD.
R L E AV

E  X t   RXX 0    S XX  f df

PU CA MIT

2 
AV I A

The PSD of a stationary process is always nonnegative and

the PSD of a real-valued random process is an even function
D TR

of frequency.
frequency
JA EC

S XX  f   S XX  f 
EL
 Noise in Communication Systems

IA ,
D T
There are many potential sources of noise (external or

IN EN
.
internal) in a communication system.

A , TM
O E E
AT R
, K G D JE
LK PA
The external
Th t l sources off noise
i include
i l d atmospheric
t h i noise,
i

TY IN ER
galactic noise, man-made noise etc.

SI R TT
ER E A
The internal sources of noise include spontaneous

IV IN H
N G C
fluctuations of currents or voltage in electrical circuits. Two

E
U N A
most common examples p are shot noise and thermal noise.
R L E AV
PU CA MIT

Shot noise arises in electronic devices e.g. diodes and

AV I A

transistors because of the discrete nature of current flow in

transistors,
these devices.
D TR
JA EC

Thermal noise arises from the random motion of electrons

EL

in a conductor.
 White Noise

IA ,
D T
The noise analysis of communication systems is

IN EN
.
customarily based on an idealized form of noise called white

A , TM
noise. The PSD of white noise is independent of operating

O E E
AT R
frequency.
frequency

, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR

Characteristics of white noise.

noise (a) Power spectral density and
JA EC

(b) Autocorrelation function.

EL
 White Noise

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV

The PSD of N0 is in watts

PU CA MIT

white noise: per Hertz.

AV I A

The parameter N0 is usually referred to the input stage of

D TR

the receiver of a communication system. It is given as N0 =

JA EC

kTe where k = Boltzmann’s constant and Te = the equivalent

EL

noise temperature of the receiver.

 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
White Noise

O E E
LK PA
AT R
transform of the PSD, for white noise, we have:

A , TM
IN EN
D T
IA ,
.
Since the autocorrelation function is the inverse Fourier
 White Noise

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
Strictly speaking, white noise has infinite average power

TY IN ER
and, as such, it is not physically realizable.

SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV

As long as the bandwidth of a noise process at the input of a

PU CA MIT

system is appreciably larger than that of the system itself,

AV I A

then the noise process may be modeled as a white noise.

noise
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
(AM)
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
IA ,
.
AMPLITUDE MODULATION
 Amplitude Modulation

IA ,
D T
IN EN
In Amplitude Modulation (AM),
(AM) the amplitude of the

.
carrier signal Ac is varied in proportion to the

A , TM
basedband/message signal m(t) and fc and c are kept

O E E
AT R
, K G D JE
LK PA
constant.

TY IN ER
SI R TT
Different types of amplitude modulation are:

ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

Double-sideband suppressed carrier (DSB-SC) AM.

AV I A

Conventional double-sideband (DSB) AM.

D TR

Single-sideband
g ((SSB)) AM.
JA EC

Vestigial-sideband (VSB) AM.

EL
Double-sideband Suppressed Carrier

IA ,
D T
(DSB SC) AM
(DSB-SC)

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
u(t)

IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

Ac cos2f c t  c 
AV I A

u t   mt ct   Ac mt  cos2f c t  c 

D TR
JA EC

m(t): message signal, c(t): carrier signal, u(t): modulated signal.

EL
 EL
JA EC

()
u(t)
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
(DSB-SC)

SI R TT
TY IN ER
, K G D JE
O E E
(DSB SC) AM

LK PA
AT R
A , TM
IN EN
D T
IA ,
.
Double-sideband Suppressed Carrier
 Double-sideband Suppressed Carrier
(DSB-SC)
(DSB SC) AM

IA ,
D T
IN EN
.
A , TM
Using Fourier transform, the spectrum of u(t):

O E E
AT R
U  f   F mt  F  Ac cos2f c t  c 

, K G D JE
LK PA
TY IN ER
Ac j
 M  f   e  f  f c   e  j  f  f c 

SI R TT
c c

ER E A
2

IV IN H
Ac
N G C
 M  f  f c e j  M  f  f c e  j 

E
U N A c c
R L E AV
2
PU CA MIT
AV I A

Let us consider m(t) is limited to the interval –W  f  W. The

D TR

process of DSB-SC modulation simply translates the

JA EC

magnitude of the spectrum of m(t) by an amount fc. The

phase of m(t) gets translated in frequency and offset by the
EL

carrier phase c.

 Double-sideband Suppressed Carrier
(DSB-SC)
(DSB SC) AM

IA ,
D T
IN EN
.
A , TM
U  f   F mt  F  Ac cos2f c t  c 

O E E
AT R
, K G D JE
LK PA
Ac j
 M  f   e  f  f c   e  f  f c 

TY IN ER
c  j c

SI R TT
2

ER E A
Ac
 M  f  f c e  M  f  f c e 
IV IN H
j  j

N G C
c c

E
2
U N A
R L E AV
PU CA MIT
AV I A

Bandwidth of the message signal = W

D TR

Bandwidth occupancy of the modulated signal = 2W

JA EC

Channel bandwidth required to transmit u(t) is Bc = 2W

EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
Magnitude and phase spectra of m(t) and u(t).

D T
IA ,
.
Double-sideband Suppressed Carrier (DSB-SC) AM
Double-sideband Suppressed Carrier (DSB-SC) AM

IA ,
Conclusions …
C

D T
IN EN
.
 The modulated signal spectrum centered at fc is composed

A , TM
of two parts:

O E E
AT R
, K G D JE
LK PA
(i) the portion in the frequency band f> fc is called the

TY IN ER
upper sideband (USB) of U(f) and

SI R TT
(ii) the
th portion
ti iin th
the ffrequency b d f< fc is
band i called
ll d the
th

ER E A
IV IN H
lower sideband (LSB) of U(f).
N G C
E
U N A
R L E AV
PU CA MIT

This is why y it is called double sideband ((DSB)) AM signal.

g The
AV I A

modulated signal does not contain a discrete component of

the carrier frequency fc. Hence it is called a suppressed
D TR

g
carrier signal i.e. u(t)
( ) is a DSB-SC AM signal.
g
JA EC
EL

Note: One must always maintain fc  W.

 Power Content of DSB-SC Signals

IA ,
D T
IN EN
.
Without loss of generality,

A , TM
we assume c = 0.

O E E
AT R
, K G D JE
LK PA
TY IN ER
T

Ru   lim T u t u t  dt

1 2

SI R TT
T 
T2

ER E A
IV IN H
N G C
T

T Ac mt mt  cos2f c t cos2f c t  dt

E
1 2 2
 lim U N A
R L E AV
T 
T2
PU CA MIT

T
Ac2
T mt mt  cos4f c t  2f c   cos2f c dt
1 2

AV I A

lim
2 T 
T2
D TR
JA EC
EL
 Power Content of DSB-SC Signals

IA ,
Now,

D T
IN EN
.


 mt mt  cos4f t  2f dt

A , TM
c c

O E E


AT R
, K G D JE
LK PA


  F mt   F mt cos4f c t  2f c   df



TY IN ER


SI R TT
 using Parseval's relation : xt  y  t dt  X  f Y   f df 
 

ER E A
   

IV IN H
 
N G C
- -

E
U N A
 M  f  2 f c e M  f  2 f c e


 j 2 f c  j 2 f c 
R L E AV


 e  j 2 f
M  f    df
PU CA MIT

  2 2 
AV I A

0
D TR
JA EC

This is because M(f) is limited in [-W, W] and W << fc. So

EL

there is no frequency overlap between M(f) and M(f2fc).

 Power Content of DSB-SC Signals

IA ,
D T
IN EN
.
A , TM
Then,

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT


 mt mt   cos4f t  2f dt  0

ER E A
c c

IV IN H


N G C
E
T
Ac2
1 2
Ru  
U N A
lim T mt mt   cos2f c dt  0
R L E AV
2 T 
T2
PU CA MIT

Ac2
Rm  cos2f c  
AV I A


2
D TR
JA EC
EL
 Power Content of DSB-SC Signals

IA ,
D T
A 2

R    R   cos2f 

IN EN
c

.
u m c
2

A , TM
O E E
Fourier transform
f

AT R
, K G D JE
LK PA
TY IN ER
PSD of the modulated signal:

SI R TT
 Ac 
2

S u  f   F Ru    F  Rm  cos2f c  

ER E A
IV IN H
2 
N G C
E
U N A
Ac2
 S m  f  f c   S m  f  f c 
R L E AV
PU CA MIT

4
 From modulation p 
AV I A

property
p y:
 
 F xt  cos2f t   1 X  f  f   1 X  f  f 
 0 
D TR

 
0 0
2 2
JA EC

Total power in the modulated signal is obtained by putting 

EL

= 0 in Ru().
 Power Content of DSB-SC Signals

IA ,
D T
IN EN
.
Therefore,

A , TM
O E E
AT R
Ac2
Rm  cos2f c  0

, K G D JE
LK PA
Pu 

TY IN ER
2

SI R TT
Ac2
 Rm 0 

ER E A
IV IN H
2
N G C
E
U N AAc2
 Pm
R L E AV

2
PU CA MIT
AV I A

where Pm = Rm(0) = power in the message signal.

D TR
JA EC
EL
 Demodulation of DSB-SC AM Signals

IA ,
D T
IN EN
To recover the original signal m(t) from the modulated

.
signal, it is necessary to retranslate the spectrum to its

A , TM
original position.

O E E
AT R
, K G D JE
LK PA
TY IN ER
For unique detection of m(t), we assume an ideal

SI R TT
channel and assume there is no noise.

ER E A
IV IN H
N G C
E
U N A
r t   u t   Ac mt cos2f c t  c 
R L E AV
PU CA MIT
AV I A

This r(t) is demodulated by first multiplying r(t) by a

D TR

locally generated sinusoid Ac/cos(2fct+) and then

JA EC

passing the product signal through an ideal low-pass

EL

filter of bandwidth W.
 Demodulation of DSB-SC AM Signals
Then,

IA ,
D T
IN EN
r t Ac/ cos2f c t    vt   Ac Ac/ mt cos2f c t  c cos2f c t  

.
A , TM
 Ac Ac mt cosc    Ac Ac/ mt cos4f c t    c 
1 1

O E E
AT R
/

, K G D JE
LK PA
2 2

TY IN ER
SI R TT
ER E A
IV IN H
The low-pass filter rejects the double frequency
N G C
E
components and passes only the low-pass components.
U N A
R L E AV

output
PU CA MIT
AV I A

yl t   Ac Ac/ mt cosc  

1
2
D TR

signal, we must have c = .

For recovering m(t) from the message signal 
JA EC

yl t   Ac Ac/ mt 
1
EL

Then,
2
 Demodulation of DSB-SC AM Signals

IA ,
D T
It is required that the phase  of the locally generated

IN EN
.
sinusoid should ideally be equal to the phase c of the

A , TM
received carrier signal.

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
This method of demodulation is called coherent

ER E A
IV IN H
detection or synchronous demodulation.
N G C
E
U N A
R L E AV
PU CA MIT

u(t) yl(t)
AV I A
D TR
JA EC
EL

Coherent detector for demodulating DSB-SC modulated wave.

 Demodulation of DSB-SC AM Signals

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
The cut off frequency
q y of the low-pass
p filter is g
greater
R L E AV
than W but less than (2fc – W). This requirement is
PU CA MIT

satisfied by choosing fc > W.

AV I A

If c -  = 90°, the demodulated signal becomes zero.

This is called quadrature null effect.
D TR
JA EC

Observation: In coherent detection based system, the system

complexity is higher for suppressing the carrier wave to save
EL

transmitted power.
 Conventional DSB AM

IA ,
D T
IN EN
IIn conventional
ti l AM
AM, the
th amplitude
lit d off the
th carrier
i signal
i l

.
A , TM
u(t) is varied about a mean value, linearly with the
baseband signal m(t).
( )

O E E
AT R
, K G D JE
LK PA
TY IN ER
transmitted signal

SI R TT
ER E A
IV IN H
u t   Ac 1  mt cos2f c t  c 
N G C
E
U N A
R L E AV
PU CA MIT

Acm(t)cos(2fct + c): a DSB AM signal and

AV I A

Accos(2fct + c): the carrier component.

D TR
JA EC

Note: One must always have m(t)1.

EL
 Conventional DSB AM

IA ,
D T
mt   amn t 

IN EN
We can express:

.
A , TM
O E E
AT R
, K G D JE
LK PA
mn(t) is normalized such that its minimum value is –1.

TY IN ER
SI R TT
ER E A
mt 
IV IN H
N G C
mn t  

E
U N A a: modulation
max mt  index
R L E AV
PU CA MIT
AV I A

The modulated signal: u t   Ac 1  amn t cos2f c t  c 

D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
()
m(t)

PU CA MIT
R L E AV
U N A
N G C
mp

IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
-mp

LK PA
AT R
Conventional DSB AM

A , TM
Conventional AM signal and its envelope.
IN EN
D T
IA ,
.
 Conventional DSB AM

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
Conclusion: The envelope of conventional AM

SI R TT
ER E A
has the information about m(t) only if the AM

IV IN H
N G C
signal u(t) = Ac[1+m(t)]cos(2fct + c) satisfies the

E
U N A
R L E AV
condition
diti Ac[1+m(t)]
[1+ (t)] > 0 for
f all
ll t.
t
PU CA MIT
AV I A
D TR
JA EC
EL
 Conventional DSB AM
Spectrum of the modulated signal:

IA ,
D T
U  f   F u t   F  Ac 1  mt  cos2f c t  c 

IN EN
.
A , TM
 F  Ac 1  amn t  cos2f c t  c 

O E E
AT R
, K G D JE
LK PA
 F amn t  F  Ac cos2f c t  c   F  Ac cos2f c t  c 

TY IN ER
SI R TT
Ac j
 aM n  f   e  f  f c   e  j  f  f c 
c c

ER E A
IV IN H
2
N G C
E
Ac j
 e  f  f c   e  j  f  f c 
U N A
R L E AV
c c

2
PU CA MIT

Ac e aM n  f  f c   e  f  f c  
j c j c
AV I A

   j 
2  e aM n  f  f c   e  j  f  f c 
c c
D TR

Conclusion: The spectrum of a conventional AM signal

JA EC

occupies a bandwidth twice the bandwidth of the message

EL

signal.
 S t
Spectrum off b
EL
JA EC
D TR
AV I A

baseband i
PU CA MIT
R L E AV

b d signal
U N A
N G C
IV IN H

l m(t)
ER E A
E

(t) and
SI R TT
TY IN ER
, K G D JE
O E E ti
LK PA
AT R
d conventional
Conventional DSB AM

A , TM
IN EN
D T
i

IA ,
.
l AM signal (t)
l u(t).
Power Content of Conventional AM Signal

IA ,
D T
IN EN
Conventional AM signal is similar to DSB when m(t) is

.
substituted with (1+amn(t)). In DSB-SC signal, power content

A , TM
in the modulated signal:

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
Ac 2

ER E A
Pu  Pm Pm: power in the message signal m(t)

IV IN H
N G C
2

E
U N A
R L E AV
For conventional AM:
PU CA MIT

T 2 T

liT   1  a 2 m 2 t   2amn t dt

liT   1  amn t  dt  lim
1 1 2
AV I A

2
Pm  lim dt
T T 2 T T 2 n
D TR

T 1  a m t dt (assumption : av. of mn t   0 )

1 2
JA EC

 lim 2 2

T 
T2
EL

n
Power Content of Conventional AM Signal

IA ,
For conventional AM:

D T
IN EN
.
T

 

A , TM
1 2
Pm  lim T 1  a m t  dt  1  a Pm
2 2 2

O E E
AT R
T 
T2

, K G D JE
n n

LK PA
TY IN ER
SI R TT
ER E A
IV IN H
Ac2 Ac2 Ac2 2
N G C
Pu  Pm   a Pm

E
U N A
R L E AV
n
2 2 2
PU CA MIT

(Ac2/2): exists due to carrier (does not carry any information)

AV I A

(Ac2/2)a2Pmn: the information carrying component

D TR

Conclusion:
l i
JA EC

Conventional
i l AM signals
i l are far
f less
l power
efficient compared to DSB-SC systems. But conventional AM
EL

can be easily demodulated.

 Drawbacks of Conventional AM

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
It is wasteful of power. Here only a fraction of the total

TY IN ER
transmitted power is actually affected by m(t).

SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

It is wasteful of bandwidth. It requires a transmission

AV I A

bandwidth equal to twice the message bandwidth.

bandwidth
D TR
JA EC
EL
 Demodulation of Conventional DSB AM Signals

IA ,
D T
IN EN
.
Here demodulation can be easily carried out. There is

A , TM
no need for a synchronous demodulator.

O E E
AT R
, K G D JE
LK PA
TY IN ER
The received signal is rectified. The rectified signal is

SI R TT
u(t) when u(t) > 0 and is zero when u(t) < 0.
0

ER E A
IV IN H
N G C
E
U N A
The message signal is recovered by passing the rectified
R L E AV

signal through a low-pass filter whose bandwidth

PU CA MIT

matches that of the message signal.

AV I A

The combination of the rectifier and the low-pass filter

D TR

is called an envelope detector.

detector
JA EC
EL
 Single-Sideband (SSB) AM

IA ,
D T
IN EN
.
A , TM
 In SSB modulation, only the upper sideband (USB) or the

O E E
AT R
lower sideband (LSB) is transmitted
transmitted. This is sufficient to

, K G D JE
LK PA
reconstruct the message signal m(t) at the receiver.

TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

 This reduces the bandwidth of the transmitted signal to W

AV I A

Hz that of the baseband signal.

Hz, signal
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER

SSB AM spectra.
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
Single-Sideband (SSB) AM

IA ,
.
 Generation of SSB AM Signals

IA ,
Frequency-discrimination
q y method

D T
IN EN
.
A , TM
Stage 1: Generate a DSB-SC AM signal, using a product

O E E
AT R
modulator.

, K G D JE
LK PA
TY IN ER
Stage 2: Employ a band-pass filter, designed to pass USB
or LSB of the modulated wave.

SI R TT
ER E A
IV IN H
N G C
E
U N A Bandpass
p
mt  u t 
R L E AV
filter
PU CA MIT
AV I A

Ac cos 2f c t
D TR
JA EC
EL

Generation of SSB AM signal by frequency-discrimination method.

 Generation of SSB AM Signals

IA ,
D T
Let the DSB-SC AM signal generated be:

IN EN
.
u DSB t   2 Ac mt cos 2f c t

A , TM
O E E
AT R
, K G D JE
LK PA
The upper single-sideband AM (USSB AM) is generated

TY IN ER
by employing a high pass filter with T.F.:

SI R TT
ER E A
1, f  fc
IV IN H
Hf  
N G C
E
U N A
0,
0, otherwise
R L E AV
PU CA MIT
AV I A

H  f   u 1  f  f c   u 1  f  f c 
D TR

u1  : unit

JA EC

where
h it step
t function
f ti
EL
 Generation of SSB AM Signals

IA ,
D T
Spectrum of the USSB AM signal is:

IN EN
.
U u  f   Ac M  f  f c u1  f  f c   Ac M  f  f c u1  f  f c 

A , TM
O E E
AT R
 Ac M  f u1  f  f  f  f  Ac M  f u1  f  f  f  f

, K G D JE
LK PA
TY IN ER
c c

SI R TT
… ((1))

ER E A
From Fourier transform:

IV IN H
N G C
E
U N A
R L E AV
1 j 
F  t    u1  f 
PU CA MIT

2 
2t 
AV I A

1 j 
F  t     u1  f 
2t 
D TR

2
JA EC
EL
 Generation of SSB AM Signals

IA ,
D T
Taking inverse Fourier transform of

IN EN
.
(1) and employing modulation
xt e  X  f  f 

A , TM
j 2 f 0t
property of Fourier transform: F

O E E
AT R
0

, K G D JE
LK PA
TY IN ER
SI R TT
uu t   Ac mt   F 1 u 1  f e j 2 f t  Ac mt   F 1 u 1  f e  j 2 f t

ER E A
c c

IV IN H
N G C
1 j  j 2 f t 1 j   j 2 f t
 Ac mt    t    Ac mt    t  

E
U N A
 e c

 e c
R L E AV
2 2t  2 2t 
PU CA MIT

Ac Ac
 mt   jmˆ t e j 2 f t
 mt   jmˆ t e  j 2 f t
c c
AV I A

2 2
…(2)
D TR

This relation mt   t   mt ; mt  

1
 mˆ t 
JA EC

used the t
identities: mˆ t  : Hilbert transform of mt 
EL
 Generation of SSB AM Signals

IA ,
D T
Hilbert transform can be viewed as a linear filter with

IN EN
.
impulse response h(t) = (1/t) and frequency response:

A , TM
 jj,, f 0

O E E
AT R
, K G D JE
LK PA

H  f    j,

TY IN ER
f 0

SI R TT
 0, f 0


ER E A
IV IN H
N G C
From eq. (2):

E
U N A
R L E AV
e
j 2 f t
 e  j 2 f t  e
j 2 f t
 e  j 2 f t 
uu t   Ac mt   jAc mˆ t 
c c c c
PU CA MIT

 
 2   2 
AV I A

 Ac mt cos 2f c t  Ac mˆ t sin 2f c t

D TR

uu((t)) is an USSB AM signal

g in time domain. The LSSB AM
  
JA EC

signal can be derived from: u t  u t  u t

u l DSB
EL
 Generation of SSB AM Signals

IA ,
D T
IN EN
.
Now,

A , TM
uu t   ul t   u DSB t 

O E E
AT R
, K G D JE
LK PA
or, Ac mt cos 2f c t  Ac mˆ t sin 2f c t  ul t   2 Ac mt cos 2f c t

TY IN ER
SI R TT
or, ul t   Ac mt cos 2f c t  Ac mˆ t sin
i 2f c t

ER E A
IV IN H
N G C
E
U N A
General time domain representation
p of SSB AM signal:
g
R L E AV
PU CA MIT

u SSB t   Ac mt cos 2f c t  Ac mˆ t sin 2f c t

AV I A

(- sign: USSB AM signal; …(3)

+ sign: LSSB AM signal)
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
signal
Generation of SSB AM signal.
AT R
A , TM
IN EN
D T
IA ,
.
Generation of SSB AM Signals
 Demodulation of SSB AM Signals

IA ,
D T
From the USSB signal in eq
eq. (3):

IN EN
.
r t cos2f c t    u t cos2f c t  

A , TM
O E E
AT R
, K G D JE
LK PA
 Ac mt cos   Ac mˆ t sin   double frequency terms
1 1

TY IN ER
2 2

SI R TT
ER E A
This signal is passed through an ideal low-pass filter

IV IN H
N G C
whose output is:

E
U N A
R L E AV

yl t   Ac mt cos   Ac mˆ t sin 

1 1
PU CA MIT

2 2
AV I A
D TR

 The
Th second
d component contributes
ib to the
h distortion
di i off
JA EC

the demodulated SSB signal. This was not present in a DSB-

EL

SC signal.
 Demodulation of SSB AM Signals

IA ,
D T
IN EN
.
For phase coherent demodulation,  = 0 and yl(t) =

A , TM
(1/2)Acm(t).

O E E
AT R
, K G D JE
LK PA
This can be done by multiplying u(t) by a locally generated

TY IN ER
carrier and then low-pass filtering the product.

SI R TT
ER E A
IV IN H
How to achieve perfect synchronism ??
N G C
E
U N A
R L E AV
A low power pilot carrier or pilot tone is transmitted at the
PU CA MIT

carrier frequency, in addition to the selected sideband.

AV I A
D TR

 Note:
N SSB AM is
SSB-AM i very popular
l iin voice
i communication
i i over
JA EC

telephone channels (wires and cables).

EL
 Demodulation of SSB AM Signals

IA ,
D T
IN EN
A ti l constraint
Any practical t i t ??

.
A , TM
O E E
AT R
YES. It is difficult to eliminate one sideband if the

, K G D JE
LK PA
message signal has large power concentrated in the vicinity

TY IN ER
of f = 0.

SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
Wh
Whatt is
i an ideal
id l situation
it ti ??
PU CA MIT
AV I A

For useful generation of SSB-AM

SSB AM signal,
signal the message
spectrum must have an energy gap centered at the origin.
D TR
JA EC
EL
 Demodulation of SSB AM Signals

IA ,
D T
IN EN
|M(f)|

.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
|U(f)|
|U(f)|

N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR
JA EC

(a) Spectrum of a message signal m(t) with an energy gap of width 2fa
centered on the origin. (b) Spectrum of corresponding SSB signal
EL

containing the upper sideband.

Implementation of AM Modulators and Demodulators
Conventional DSB AM Modulators

IA ,
D T
IN EN
.
Power-law Modulators

A , TM
O E E
AT R
 A nonlinear device like a P-N diode, with a voltage-current

, K G D JE
LK PA
characteristic shown below, can be used for modulation.

TY IN ER
SI R TT
i The input-output characteristic

ER E A
of the nonlinear device may be

IV IN H
governed by a square law:
N G C
E
U N A
R L E AV
PU CA MIT
AV I A

vo t   a1vi t   a2 vi2 t 
v
vi t  : input;
p vo t  : output;
D TR

p
Voltage-current
Voltage current
JA EC

characteristic of P-N diode. a1 , a2 : constant parameters

EL
 Conventional DSB AM Modulators

IA ,
D T
Power law Modulators
Power-law

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
mt  + Nonlinear Bandpass filter
u t 
tuned to fc

SI R TT
device

ER E A
IV IN H
N G C
Ac cos 2f c t

E
U N A
R L E AV
PU CA MIT
AV I A

Block-diagram of power-law AM modulator.

D TR
JA EC
EL
 Conventional DSB AM Modulators

IA ,
Power-law Modulators

D T
IN EN
.
The input to the nonlinear device: vi t   mt   Ac cos 2f c t

A , TM
O E E
AT R
, K G D JE
LK PA
vo t   a1 mt   Ac cos 2f c t   a2 mt   Ac cos 2f c t 

TY IN ER
2

SI R TT
 2a2 

ER E A
 a1mt   a2 m t   a2 Ac cos 2f c t  Ac a1 1 
2 2 2
mt  cos 2f c t

IV IN H
N G C
 a1 

E
U N A
R L E AV

The output of the bandpass

PU CA MIT

filter with bandwidth 2W and By design, one maintains:

AV I A

centered at f = fc is:
2a2
 2a2  mt   1
u t   Ac a1 1  mt  cos 2f c t
D TR

a1
 a1 
JA EC
EL
 v2
EL
JA EC
D TR
AV I A
PU CA MIT
Switchingg Modulators

v1
R L E AV
U N A
N G C
IV IN H
ER E A
E

-Tp
SI R TT
TY IN ER
, K G D JE
s(t)

O E E
LK PA
AT R
A , TM
IN EN
D T
IA ,
Tp

.
Conventional DSB AM Modulators

t
 Conventional DSB AM Modulators

IA ,
Switchingg Modulators

D T
IN EN
vi t   mt   Ac cos 2f c t

.
The input to the diode:

A , TM
O E E
AT R
, K G D JE
LK PA
The input output characteristic of the diode is shown by the

TY IN ER
v2-v1 curve, with Ac >> m(t). Output across the load resistor:

SI R TT
vi t , ct   0

ER E A
vo t   

IV IN H
 0, ct   0
N G C
E
U N A
R L E AV
This can be viewed as:
PU CA MIT

vo t   vi t st   mt   Ac cos 2f c t st 

AV I A

st  : switching function shown in the figure.

For a periodic function s(t), its Fourier series representation:
D TR
JA EC

1 2 
st     cos c t  cos 3c t  cos 5c t  
1 1
EL

2  3 5 
 Conventional DSB AM Modulators

IA ,
Switchingg Modulators

D T
IN EN
.
Then,

A , TM
vo t   mt   Ac cos 22
f c t st 

O E E
AT R
, K G D JE
LK PA
 mt   Ac cos 2f c t 

TY IN ER
SI R TT
1 2  1 1 

ER E A
 2    cos c t  3 cos 3c t  5 cos 5c t  

IV IN H
  
N G C
E
U N A
 
R L E AV
Ac
1  A mt  cos 2f c t  other terms
4

PU CA MIT

2  c 
AV I A

The bandpass filter tuned at fc (i.e. c) suppresses other

terms and the output is the desired conventional DSB AM
D TR

signal:
g
Ac  
JA EC

u t   1  A mt  cos 2f c t

4
EL

2  c 
 DSB-SC AM Modulators

IA ,
Balanced Modulators

D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
u t   4 Ac a2 mt  cos 2f c t

SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR

Bl k di
Block-diagram off a balanced
b l d modulator.
d l t
JA EC
EL
 DSB-SC AM Modulators

IA ,
Ringg Modulator

D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A

A ring modulator
modulator.
D TR

Mathematically speaking,
speaking vo(t) = m(t)c(t) where c(t) is a square
JA EC

wave of frequency fc.

EL
 DSB-SC AM Modulators

IA ,
Ringg Modulator

D T
IN EN
.
As c(t) is a periodic function, its Fourier series

A , TM
representation is:

O E E
AT R
, K G D JE
4 

LK PA
ct    cos c t  cos 3c t  cos 5c t  
1 1

TY IN ER
 3 5 

SI R TT
 vo t   mt ct 

ER E A
IV IN H
N G C
4 
  mt cos c t  mt cos 3c t  mt cos 5c t  
1 1

E
U N A
R L E AV
 3 5 
PU CA MIT

When this vo(t) is passed through a bandpass filter with

AV I A

center frequency fc and bandwidth 2W, the output:

D TR

u t   mt cos 2f c t

4
JA EC


EL
 SSB AM Modulators

IA ,
D T
IN EN
.
A , TM
O E E
AT R
The Hilbert transform based method uses two mixers i.e.

, K G D JE
LK PA
two balanced modulators in addition to a Hilbert

TY IN ER
transformer.

SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
The Frequency
Th F di
discrimination
i i ti b
basedd method
th d uses a single
i l
PU CA MIT

balanced modulator and a sideband filter.

AV I A
D TR
JA EC
EL
 DSB-SC AM Demodulators

IA ,
D T
Synchronous or Coherent or Homodyne Demodulator

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
r t  Balanced Low-pass filter mt 

SI R TT
modulator

ER E A
IV IN H
N G C
E
U N A
Phase-locked
R L E AV
loop
PU CA MIT
AV I A
D TR

The low pass filter is of bandwidth W Hz. It passes the

d i d signal
desired i l and
d rejects
j all
ll signal
i l and
d noise
i components
JA EC

above W Hz.
EL
 Conventional DSB AM Demodulators

IA ,
Rectifier
f Detector

D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A

Without loss of generality, we assume c = 0. Here,

D TR
JA EC

r t   Ac 1  mt cos 2f c t

EL
 Conventional DSB AM Demodulators

IA ,
Rectifier
f Detector

D T
IN EN
r t   Ac 1  mt cos 2f c t

.
A , TM
O E E
AT R
The output
p across resistor in the input
p side can be viewed

, K G D JE
LK PA
as the AM signal multiplied by the periodic function s(t):

TY IN ER
vR t   Ac 1  mt cos c tst 

SI R TT
ER E A
1 2  
 Ac 1  mt cos c t   cos c t  cos 3c t  cos 5c t  
IV IN H
1 1
N G C
E
U N A2   3 5 
R L E AV

Ac 1  mt 
PU CA MIT

  other terms of higher frequencies


AV I A

The output of the

low-pass filter Ac 1  mt  The final Ac mt 
D TR

(with cut-off
cut off output:

JA EC

frequency W Hz):

EL
 Conventional DSB AM Demodulators

IA ,
Envelope
p Detector

D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR

The time constant must be so selected that

(1/fc)<<RC<<(1/W),
) RC (1/W) where
h W = highest
hi h ffrequency iin m(t).
( ) The
Th
JA EC

envelope detector output = Ac[1+m(t)].

EL
 Conventional DSB AM Demodulators

IA ,
D T
Envelope Detector

IN EN
.
A , TM
O E E
AT R
The simplicity of this demodulator has made conventional

, K G D JE
LK PA
DSB AM a practical choice for AM radio broadcasting.

TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
There are literally billions of radio receivers and an
R L E AV

inexpensive implementation of the demodulator is very

PU CA MIT

important.
AV I A
D TR
JA EC
EL
 SSB AM Demodulator

IA ,
D T
IN EN
.
Like DSB-SC scheme, this also requires a phase coherent

A , TM
reference. For signals like speech, which have relatively

O E E
AT R
little or no power content at d.c., one can use a demodulator

, K G D JE
LK PA
scheme like this.

TY IN ER
SI R TT
ER E A
IV IN H
r t  Balanced
mt 
N G C
Low-pass filter

E
modulator
U N A
R L E AV
PU CA MIT

Estimate carrier
AV I A

component
D TR

Here the
h balanced
b l
JA EC

d modulator
d l i used
is d for
f frequency
f
conversion of the bandpass signal to low-pass or baseband
EL

signal.
 Frequency Division Multiplexing (FDM)

IA ,
D T
IN EN
.
In multiplexing, a number of independent signals can be

A , TM
combined into a composite signal suitable for transmission

O E E
AT R
over a common channel.

, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
TDM is used in digital transmission only. But FDM is used
N G C
E
both in analog and digital transmission.
U N A
R L E AV
PU CA MIT
AV I A

In FDM, each message signal modulates a carrier of a

different frequency, where the minimum separation between
D TR

two adjacent carriers is either 2W (DSB AM) or W (SSB AM)

JA EC

(W = bandwidth of each message signal).

EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
Block diagram of FDM System.

AT R
A , TM
IN EN
D T
IA ,
.
Frequency Division Multiplexing (FDM)
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
Practical implementation of an FDM System.

IN EN
D T
IA ,
.
Frequency Division Multiplexing (FDM)
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
ANGLE MODULATION

D T
IA ,
.
 Angle Modulation

IA ,
D T
IN EN
.
In Angle Modulation, the angle of the modulating sinusoidal carrier

A , TM
wave is varied according to the baseband signal.

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
Angle
g Modulation Techniques
q

ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

Frequency Modulation Phase Modulation

AV I A

(FM) (PM)
D TR
JA EC
EL
 Angle Modulation

IA ,
D T
Merits and Demerits …

IN EN
.
A , TM
Angle Modulation methods are more complex to implement and

O E E
AT R
more difficult to analyze.
y In manyy cases,, only
y an approximate
pp

, K G D JE
LK PA
analysis can be done.

TY IN ER
SI R TT
Angle Modulation possesses a bandwidth expansion property.
property FM

ER E A
IV IN H
and PM generally expand the bandwidth such that the effective
N G C
E
bandwidth of the modulated signal is usually many times the
U N A
R L E AV
b d idth off th
bandwidth the message signal.
i l
PU CA MIT
AV I A

So increased transmission bandwidth and increased complexity

p y are
the principle demerits of angle modulation methods.
D TR

However, the major benefit is that these systems provide high degree
However
JA EC

of noise immunity i.e. better discrimination against noise and

EL

interference than AM signals.

 Angle Modulation

IA ,
D T
IN EN
Let i(t) be the ‘angle’ of a modulated sinusoidal carrier and

.
A , TM
i(t) be a function of the message signal m(t).

O E E
AT R
, K G D JE
LK PA
Implication
l

TY IN ER
SI R TT
The angle modulated wave is: u(t) = Accos[i(t)].
(t)]

ER E A
IV IN H
Ac = amplitude of the carrier wave.

N G C
E
U N A
R L E AV
A complete oscillation occurs whenever i(t) changes by 2
PU CA MIT

radians. If i(t) increase monotonically with time, the

average frequency in Hertz over an interval from t to t+t is:
AV I A

i t  t   i t 
f t t  
D TR

2t
JA EC
EL
 Angle Modulation

IA ,
D T
IN EN
Then one can define ‘instantaneous frequency’ of the angle

.
A , TM
modulated wave:

O E E
AT R
 i t  t   i t  1 di t 

, K G D JE
LK PA
f i t   lim f t t   lim
t 0   

TY IN ER
t 0
 2t  2 dt

SI R TT
Implication
li i

ER E A
IV IN H
N G C
E
U N A
Hence we can say u(t) is a rotating phasor of length Ac and
R L E AV

angle i(t). Angular velocity of such a phasor = (di(t)/dt) (in

PU CA MIT

rad/sec.).
AV I A

In the simple case of an unmodulated carrier, i(t) = 2fct +

c. This phasor rotates with a constant angular velocity of
D TR

2 fc. c = value
2f l off i(t) when
h t = 0.
0
JA EC
EL
 Angle Modulation

IA ,
D T
IN EN
There are two common methods of varying i(t).

.
A , TM
O E E
AT R
) Here i((t)) is varied linearly
1. Phase Modulation ((PM): y with

, K G D JE
LK PA
m(t).

TY IN ER
SI R TT
ER E A
IV IN H
i t   2f c t  k p mt 
N G C
E
U N A
R L E AV
2f c t  angle of the unmodulated carrier
PU CA MIT

k p  phase sensitivity of the modulator (rad/volt)

AV I A

Assumptions: m(t) is a voltage waveform. Also, angle of the

D TR

unmodulated carrier is chosen zero at t = 0 ((for convenience).

)
JA EC

 
Therefore, u t   Ac cos 2f c t  k p mt 
EL
 Angle Modulation

IA ,
D T
2. Frequency Modulation (FM): Here instantaneous frequency
2

IN EN
.
fi(t) is varied linearly with m(t).

A , TM
O E E
AT R
, K G D JE
LK PA
f i t   f c  k f mt 

TY IN ER
SI R TT
f c  frequency of the unmodulated carrier

ER E A
IV IN H
k f  frequency sensitivity of the modulator (Hz/volt)
N G C
E
U N A
R L E AV
PU CA MIT

i t   2f c t  2k f  md

t
AV I A

0
D TR

Here we have made the same assumptions as before.

JA EC

Therefore, u t   Ac cos 2f c t  2k f  m d

t
EL

 0 
 Angle Modulation

IA ,
D T
An FM signal may be regarded as a PM signal whose

IN EN
.
modulating wave is  md in place of m(t).
t

A , TM
0

O E E
AT R
, K G D JE
LK PA
TY IN ER
mt  FM
u t  (FM signal)

SI R TT
modulator

ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

mt  Integrator PM
u t  (FM signal)
AV I A

modulator
D TR

Ac cos2f c t 
JA EC
EL

Generation of FM signal using a Phase Modulator.

 Angle Modulation

IA ,
Similarly, a PM signal can be generated by first

D T
IN EN
differentiating m(t) and then using the result as an input to a

.
A , TM
frequency modulator.

O E E
AT R
, K G D JE
LK PA
TY IN ER
mt  PM
u t  (PM signal)

SI R TT
modulator

ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

mt  Differentiator FM
u t  (PM signal)
AV I A

modulator
D TR

Ac cos2f c t 
JA EC
EL

Generation of PM signal using a Frequency Modulator.

 Frequency Modulation

IA ,
D T
IN EN
i l u(t)
FM signal (t) is
i a nonlinear
li f ti
function off m(t).
(t) Hence
H

.
A , TM
frequency modulation is a nonlinear modulation process.
Hence spectrum of FM signal is not related in a simple

O E E
AT R
, K G D JE
LK PA
manner to m(t) as in the case of AM.

TY IN ER
SI R TT
One can consider two situations.

ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

Situation I: In the simplest case, we consider a single-tone

modulation that p produces a Narrowband FM signal.
g
AV I A

Situation II: In the general case, we consider a single-tone

D TR

modulation where the FM signal is wideband.

JA EC
EL
 Frequency Modulation

IA ,
D T
IN EN
mt   Am cos2f m t 

.
Let the sinusoidal

A , TM
modulating signal be:

O E E
AT R
, K G D JE
LK PA
The instantaneous f i t   f c  k f Am cos2f m t 

TY IN ER
frequency of the
 f c  f cos2f m t 

SI R TT
resulting
g FM signal:
g

ER E A
IV IN H
N G C
E
f = kfAm = frequency deviation = maximum departure of fi from the
U N A
R L E AV
carrier frequency fc.
PU CA MIT
AV I A

i t   2 f i  d 2  f c  f cos2f m d

t t

The angle of 0 0
D TR

the FM signal: f
 2f c t  sin2f m t 
JA EC

fm
EL
 Frequency Modulation

IA ,
D T
IN EN
i t   2 f i  d 2   f c  f cos2f m d

.
t t

A , TM
The angle of 0 0

O E E
AT R
the FM signal: f

, K G D JE
LK PA
 2f c t  sin2f m t 

TY IN ER
fm

SI R TT
ER E A
IV IN H
f
N G C
  modulation index of FM signal.

E
We denote: U N Afm
R L E AV

i t   2f c t   sin2f m t 
PU CA MIT
AV I A

 FM signal: u t   Ac cos2f c t   sin2f m t 

D TR

For narrowband FM,  is small compared to 1 radian.

JA EC
EL

For wideband FM,  is large compared to 1 radian.

 Narrowband Frequency Modulation

IA ,
D T
Let us consider FM signals generated using a

IN EN
.
sinusoidal modulating signal:

A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
u t   Ac cos2f c t   sin2f m t 

SI R TT
ER E A
IV IN H
N G C
E
u t   Ac cos2f c t cos sin2f m t   Ac sin2f c t sin sin2f mt 
U N A
R L E AV
PU CA MIT

for  small compared

AV I A

to 1 radian
di
u t   Ac cos2f c t   Ac sin2f c t sin2f m t 
D TR
JA EC
EL
 Narrowband Frequency Modulation

IA ,
D T
IN EN
.
Narrowband

A , TM
FM signal
-

O E E
mt 

AT R
Product
Integrator 

, K G D JE
LK PA
modulator

TY IN ER
modulating +

SI R TT
wave

ER E A
IV IN H
carrier

N G C
E
U N A -90 phase wave
R L E AV
shifter
Ac cos 2f c t
PU CA MIT
AV I A

Narrowband phase modulator

D TR
JA EC

Block diagram for generation of narrowband FM signal.

EL
 Narrowband Frequency Modulation

IA ,
D T
IN EN
Id ll an FM signal
Ideally, i l has
h a constant
t t envelope
l and,d ffor the
th

.
case of a sinusoidal modulating frequency fm, i(t) is also

A , TM
sinusoidal with same frequency.

O E E
AT R
, K G D JE
LK PA
TY IN ER
But the modulating signal produced by the narrowband

SI R TT
modulator differs from the ideal condition in two

ER E A
fundamental aspects:

IV IN H
N G C
E
U N A
R L E AV
PU CA MIT

I. The envelope contains a residual amplitude modulation

AV I A

and, therefore, varies with time.

D TR

II. For a sinusoidal modulating wave, the angle i(t) contains

JA EC

harmonic distortion in the form of third- and higher-order

EL

harmonics of the modulation frequency fm.

 Narrowband Frequency Modulation

IA ,
D T
The signal u(t) can be expanded as:

IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
u t   Ac cos2f c t   Ac cos2 f c  f m t   cos2 f c  f m t 
1

TY IN ER
2

SI R TT
This is similar to an AM signal:

ER E A
IV IN H
N G C
E
U N A
R L E AV

u AM t   Ac cos2f c t   Ac cos2 f c  f m t   cos2 f c  f m t 

1
PU CA MIT

2
AV I A

  modulation
d l ti factor
f t off an AM signal.
i l
D TR

The only difference is the sign of the lower side frequency in

JA EC

the narrowband FM which is reversed. Hence a narrowband

FM requires essentially the same bandwidth of transmission
EL

(2fm) as the AM signal.

 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
IA ,
.
Narrowband Frequency Modulation

Phasor comparison of (a) narrowband FM signal and (b) AM signal.

 Wideband Frequency Modulation

IA ,
D T
Let us consider the single-tone
single tone FM signal once more:

IN EN
.
A , TM
O E E
AT R
for an arbitrary
u t   Ac cos2f c t   sin2f m t 

, K G D JE
LK PA
value of 

TY IN ER
SI R TT
Let us assume that fc is large
g enough
g compared
p to the

ER E A
bandwidth of the FM signal:

IV IN H
N G C
E
U N A
R L E AV

u t   Re Ac exp j 2f c t  j sin2f m t 

PU CA MIT

 Reu~ t exp j 2f t 

AV I A

u~ t  : complex envelope of the AM signal u t 

D TR
JA EC

u~ t   Ac exp j sin2f m t 
EL
 Wideband Frequency Modulation

IA ,
D T
u~ t   A exp j sin2f t 

IN EN
.
c m

A , TM
O E E
AT R
, K G D JE
LK PA
Unlike the original FM signal u(t), the complex envelope ũ(t)

TY IN ER
is a periodic function of time with a fundamental frequency

SI R TT
of fm.

ER E A
IV IN H
Expanding ũ(t) in form of a complex Fourier series:
N G C
E
U N A
R L E AV
PU CA MIT

u t    cn exp j 2nf m t 

~
AV I A

n  
fm fm
D TR

cn  f m  u~ t  exp j 2nf m t dt  f m Ac  exp j sin2f m t   j 2nf m t dt

2 2
JA EC

f m f m
 
EL

2 2
 Wideband Frequency Modulation

IA ,
D T
IN EN
u t    cn exp j 2nf m t 

~

.
A , TM
n  

O E E
AT R
fm fm

, K G D JE
LK PA
cn  f m  u~ t  exp j 2nf m t dt  f m Ac  exp j sin2f m t   j 2nf m t dt
2 2

TY IN ER
f m f m

SI R TT
 
2 2

ER E A
IV IN H
Let x = 2fmt. Then cn can be rewritten as:
N G C
E
U N A
R L E AV
PU CA MIT

Ac 
cn  exp j  sin x  nx dx
AV I A

2
D TR

This expression contains nth order Bessel function of the

JA EC

first kind and argument . This is commonly denoted by the

symbol Jn().
EL
 Wideband Frequency Modulation

IA ,
D T
IN EN
.
1 
J n    exp j  sin x  nx dx

A , TM
2

O E E
AT R
Then cn  Ac J n . Hence,

, K G D JE
LK PA
TY IN ER
u t   Ac  J n  exp j 2nf m t  and

~

SI R TT
 
ER E A
n  

IV IN H
u t   Ac . Re  J n  exp j 2 f c  nf m t


N G C
E
U N A n  
R L E AV
PU CA MIT

Interchanging the order of summation and evaluation

of the real part in the right
right-hand
hand side of this equation:
AV I A
D TR

u t   Ac  J n  cos2 f c  nf m t 
JA EC


EL

n  
 Wideband Frequency Modulation

IA ,
D T
IN EN
.
u t   Ac  J n  cos2 f c  nf m t 

A , TM


O E E
AT R
n  

, K G D JE
LK PA
TY IN ER
This is the desired form for the Fourier series representation

SI R TT
off the
h single-tone
i l FM signal
i l u(t)
( ) for
f an arbitrary
bi  The
. Th

ER E A
IV IN H
discrete spectrum of u(t) is obtained by taking the Fourier
N G C
E
transforms of both sides of the last equation.
U N A
R L E AV
PU CA MIT
AV I A

Ac 
Uf   J n  f  f c  nf m    f  f c  nf m 
D TR

2 n
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
IA ,
Plot of Bessel functions of first kind for varying order.

.
Wideband Frequency Modulation
 Wideband Frequency Modulation

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A

 The properties of Bessel function Jn():

D TR

Jn() = (-1)n J-n() for all n both +ve and –ve.

JA EC

 For small , J0()  1, J1()  (/2), Jn()  0 (n > 2)

EL

 J   1

2
n
n  
 Wideband Frequency Modulation

IA ,
Observations and Conclusions…
C

D T
IN EN
.
A , TM
The spectrum of an FM signal contains a carrier

O E E
AT R
, K G D JE
LK PA
component and an infinite set of side frequencies localted

TY IN ER
symmetrically on either side of the carrier at frequency

SI R TT
separations of fm, 2fm, 3fm … This is unlike AM where a

ER E A
sinusoidal modulating signal gives rise to only one pair of

IV IN H
side frequencies.
N G C
E
U N A
R L E AV
PU CA MIT

For the special case of  small compared to unity, only

AV I A

Bessel coefficients J0() and J1() have significant values,

values so
that the FM signal is effectively composed of a carrier and
single pair of side frequencies at fc  fm. This situation
D TR

corresponds to the special case of narro

narrowband
band FM
FM.
JA EC
EL
 Wideband Frequency Modulation

IA ,
C I: Spectra
Case p off the FM signal,
g , when frequency
f q y off the

D T
IN EN
.
modulating signal is fixed, but its amplitude is varied.

A , TM
O E E
AT R
, K G D JE
LK PA
( )
(a) (b)

TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV

(c)
PU CA MIT
AV I A
D TR
JA EC

Discrete amplitude spectra of an FM signal, normalized w.r.t. the carrier

EL

amplitude (only positive frequencies are shown).

 Wideband Frequency Modulation
Case II: Spectra of the FM signal, when amplitude of the

IA ,
D T
IN EN
.
modulating signal is fixed, but its frequency fm is varied.

A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR
JA EC

Discrete amplitude spectra of an FM signal, normalized w.r.t. the carrier

EL

amplitude (only positive frequencies are shown).

 Transmission Bandwidth of FM Signals

IA ,
D T
IN EN
An effective
effecti e band
bandwidth
idth can be specified for the transmission of an

.
A , TM
FM signal.

O E E
AT R
, K G D JE
LK PA
For large , the bandwidth approaches, and is only slightly greater

TY IN ER
than, the total frequency excursion 2f.

SI R TT
ER E A
For small , the spectrum of the FM signal is effectively limited to fc

IV IN H
N G C
and one pair of side frequencies at fc  fm, so that the bandwidth

E
U N A
approaches 2fm.
R L E AV
PU CA MIT

Carson’s Rule:
AV I A

The transmission bandwidth of an FM signal generated by a

single-tone modulating signal of frequency fm:
D TR
JA EC

 1
BC  2f  2 f m  2f 1    2  1 f m
EL

 
 Transmission Bandwidth of FM Signals

IA ,
D T
IN EN
Generally
G ll speaking,
ki the
th effective
ff ti bandwidth
b d idth off the
th

.
A , TM
modulated signal, is given as, by Carson’s rule, as:
BC  2  1W

O E E
AT R
, K G D JE
LK PA
TY IN ER
W = bandwidth of the message signal.

SI R TT
ER E A
Note: In wideband FM, usually   5.

IV IN H
N G C
E
U N A
conclusion
R L E AV
PU CA MIT
AV I A

The b
Th bandwidth
d idth off an angle-modulated
l d l t d signal
i l iis much
h
greater than the bandwidth of various AM schemes,
D TR

which is either W ((in SSB)) or 2W ((in DSB or

JA EC

conventional AM).
EL
 Implementation of Frequency Modulators
and Demodulators

IA ,
D T
IN EN
d l
Frequency Modulators

.
A , TM
i) Direct FM

O E E
AT R
, K G D JE
LK PA
In the direct method, the carrier frequency is directly

TY IN ER
varied in accordance with the input baseband signal. This is

SI R TT
p
accomplished using
g a VCO and a VCO can be ppopularly
p y

ER E A
designed using a varactor diode.

IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR
JA EC
EL

Varactor diode implementation of Direct FM modulator.

 Implementation of Frequency Modulators
and Demodulators

IA ,
D T
IN EN
Frequency Modulators
d l

.
A , TM
i) Direct FM

O E E
AT R
, K G D JE
LK PA
Let the inductance of the

TY IN ER
inductor in the tuned condition be

SI R TT
L0 and the capacitance of the
varactor diode is: C t   C0  k 0 mt 

ER E A
IV IN H
N G C
E
U N A
1
R L E AV
When m(t) = 0,
0 the frequency
fc 
of the tuned circuit is: 2 L0C0
PU CA MIT
AV I A

Generally, when m(t)  0:

D TR

f i t  
1 1 1 1
  fc
2 L0 C0  k 0 mt  2 L0C0
JA EC

k k
1  0 mt  1  0 mt 
EL

C0 C0
 Implementation of Frequency Modulators
and Demodulators

IA ,
D T
IN EN
Frequency Modulators
d l

.
A , TM
i) Direct FM

O E E
AT R
, K G D JE
LK PA
Assuming that:

TY IN ER
SI R TT
k0
  mt   1

ER E A
IV IN H
C0
N G C
E
U N A

R L E AV
Then, using
Then
approximations:
1     1  and 1
 1 
PU CA MIT

2 1 
AV I A

 k0 
Then, we obtain: f i t   f c 1  mt 
 2C0 
D TR
JA EC

This is the relation for the frequency modulated signal.

EL
 Implementation of Frequency Modulators
and Demodulators

IA ,
D T
IN EN
Frequency Modulators
d l

.
A , TM
ii) Indirect FM

O E E
AT R
, K G D JE
LK PA
Narrowband

TY IN ER
FM signal
m t  Product -


SI R TT
Integrator
g
modulator
d l t

ER E A
+
IV IN H
modulating
N G C
E
wave U N A
R L E AV
carrier
i
wave
PU CA MIT

-90 phase
shifter
AV I A

Ac cos 2  f c t

Narrowband phase modulator

D TR
JA EC
EL

Generation of narrowband FM signal.

 Implementation of Frequency Modulators
and Demodulators

IA ,
D T
IN EN
Frequency Modulators
d l

.
A , TM
ii) Indirect FM

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV

Indirect generation of FM signals.

PU CA MIT

Let the narrowband u n t   Ac cos2f c t  t 

AV I A

signal be:

Then the output of the y t   Ac cos2nf c t  nt 

D TR

frequency multiplier:
JA EC

The final wideband u t   Ac cos2nf c  f L 0 t  nt 

EL

FM signal is:
 FM Demodulators

IA ,
D T
FM Demodulator –Scheme
Scheme I:

IN EN
.
A , TM
FM demodulators are implemented by generating an AM

O E E
signal whose amplitude is proportional to the instantaneous

AT R
, K G D JE
LK PA
frequency of the FM signal.

TY IN ER
SI R TT
To achieve the first step to transform theFM signal into an

ER E A
AM signal, the FM signal is passed through an LTI system

IV IN H
whose frequency response is approximately a straight line in
N G C
E
U N A
the frequency band of the FM signal.
R L E AV

BC
PU CA MIT

Let the frequency response

H  f   V0  k  f  f c  for f  f c 
of such a system be: 2
AV I A

u t   Ac cos 2f c t  2k f  m d 

and if the input t

to the system is:

D TR

 0 
JA EC

vo t   Ac V0  kk f mt cos 2f c t  2k f  m d 

Then the t
EL

output will be:  0 

 FM Demodulators

IA ,
D T
FM Demodulator –Scheme
Scheme I:

IN EN
.
A , TM
The next step is to demodulate this signal so that the

O E E
message signal m(t) can be recovered.

AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
FM signal AM signal

IV IN H
FM to AM AM output signal

N G C
E
converter
U N A demodulator
R L E AV
PU CA MIT
AV I A

A general FM demodulator
demodulator.
D TR

One candidate for implementation of FM to AM conversion

JA EC

is a simple differentiator with H(f)= 2f.

EL
 FM Demodulators

IA ,
D T
FM Demodulator –Scheme
Scheme I:

IN EN
.
A , TM
Another candidate for implementation of FM to AM

O E E
conversion is the rising half of the frequency characteristic

AT R
, K G D JE
LK PA
of a tuned circuit.

TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
PU CA MIT
AV I A
D TR

To obtain a linear characteristic over a wide range of

f
frequencies,
i usually
ll two circuits
i i tuned d at two frequencies,
f i f1
JA EC

and f2, are connected in a configuration called a balanced

EL

discriminator.
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
FM Demodulator –Scheme

IV IN H
Scheme I:

ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
FM Demodulators

LK PA
AT R
A , TM
IN EN
D T
IA ,
.
The frequency response of a balanced discriminator circuit.
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
FM Demodulator –Scheme

IV IN H
Scheme I:

ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
FM Demodulators

LK PA
A balanced discriminator circuit.
AT R
A , TM
IN EN
D T
IA ,
.
 FM Demodulators
The Frequency Discriminator:

IA ,
D T
IN EN
.
A , TM
Slope circuit Envelope
H1(f) detector

O E E
AT R
+

, K G D JE
FM

LK PA
Baseband
B b d

TY IN ER
signal
wave -

SI R TT
Slope circuit Envelope

ER E A
IV IN H
H2(f) detector
N G C
E
U N A
R L E AV

u t   Ac cos 2f c t  2k v  m d 

t
The incoming
PU CA MIT

FM signal:  0 
AV I A

u~ t   Ac exp  j 2k f  m d

t
The complex
envelope of u(t):  
D TR

0
JA EC

Let ũ1(t) denote the complex envelope of the slope circuit

H1(f) response, due to ũ (t). The desired response of the slope
EL

circuit H1(f) i.e. u1(t) is a hybrid modulated signal.

 FM Demodulators
The Frequency Discriminator:

IA ,
D T
IN EN
.
A , TM
Slope circuit Envelope
H1(f) detector

O E E
AT R
+

, K G D JE
FM

LK PA
Baseband
B b d

TY IN ER
signal
wave -

SI R TT
Slope circuit Envelope

ER E A
IV IN H
H2(f) detector
N G C
E
U N A
R L E AV

u1 t   Ac V0  kk f mt cos 2f c t  2k f  md 

t
The slope circuit
PU CA MIT

response:  0 
AV I A

kk f
Let us choose, for all t, mt   1
V0
D TR
JA EC

 kk f 
The resulting envelope u1 t   AcV0 1 
~ mt 
EL

detector output:  V0 
 FM Demodulators
The Frequency Discriminator:

IA ,
D T
IN EN
.
A , TM
Slope circuit Envelope
H1(f) detector

O E E
AT R
+

, K G D JE
FM

LK PA
Baseband
B b d

TY IN ER
signal
wave -

SI R TT
Slope circuit Envelope

ER E A
IV IN H
H2(f) detector
N G C
E
U N A
R L E AV
The complementary slope circuit H2(f) is so designed that
PU CA MIT

Ĥ2(f) = Ĥ1(-f) where Ĥ1(f) is an equivalent low pass filter with

q
a frequency y response
p that can be used to replace
p the band
AV I A

pass filter with frequency response H1(f).

D TR

Let u2(t) denote the response

off the
th complementary
l t slope
l  kk f 
u 2 t   AcV0 1 
~ mt 
JA EC

circuit produced by u(t). Then  V0 

EL

it’s envelope detector output:

 FM Demodulators
The Frequency Discriminator:

IA ,
D T
IN EN
.
A , TM
Slope circuit Envelope
H1(f) detector

O E E
AT R
+

, K G D JE
FM

LK PA
Baseband
B b d

TY IN ER
signal
wave -

SI R TT
Slope circuit Envelope

ER E A
IV IN H
H2(f) detector
N G C
E
U N A
R L E AV

The final output: u~1 t   u~2 t   2 Ac kk f mt 

PU CA MIT
AV I A
D TR
JA EC

This output is a scaled version of the original message

signal m(t) and free from bias.
EL
 FM Demodulators

IA ,
D T
FM Demodulator –Scheme
Scheme II:

IN EN
.
A , TM
Another FM signal demodulator uses feedback in the
demodulator to narrow the bandwidth of the FM detector.

O E E
AT R
, K G D JE
LK PA
This reduces the noise power at the output of the

TY IN ER
demodulator. An FM demodulator with feedback (FMFB)

SI R TT
employs a VCO in the feedback branch.
branch

ER E A
IV IN H
N G C
E
U N A

Received Output
R L E AV
BP LP
Discriminator
PU CA MIT

signal Filter Filter signal

AV I A

VCO
D TR
JA EC
EL

FMFB demodulator.
 FM Demodulators

IA ,
D T
FM Demodulator –Scheme
Scheme II:

IN EN
.
A , TM
An alternative to FMFB demodulator is to use PLL.

O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
Modulated Output

ER E A
Phase Loop filter

IV IN H
N G C
signal comparator G(f)

E
U N A signal
R L E AV
PU CA MIT

VCO
AV I A
D TR

PLL-FM demodulator.
JA EC
EL
 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
Modulated Output
Phase Loop filter

O E E
AT R
, K G D JE
LK PA
signal
i l comparator
t G(f) signal

TY IN ER
SI R TT
ER E A
VCO

IV IN H
N G C
E
U N A
R L E AV

The input to the PLL is the angle-modulated wave:

PU CA MIT

t   2k f  m d

AV I A

u t   Ac cos2f c t  t  For FM:


D TR

Instantaneous frequency of the VCO: f v t   f c  kv vt 

JA EC
EL

v(t) = control voltage of the VCO; kv = a deviation constant

 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
Modulated Output
Phase Loop filter

O E E
AT R
, K G D JE
LK PA
signal
i l comparator
t G(f) signal

TY IN ER
SI R TT
ER E A
VCO

IV IN H
N G C
E
U N A
R L E AV

Therefore, the VCO output:

PU CA MIT
AV I A

yv t   Av sin
i 2f c t  v t 
t

v t   2k v  vd
0
D TR
JA EC

The phase comparator is basically a multiplier and a filter

that rejects the signal component centered at 2fc.
EL
 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
Modulated Output
Phase Loop filter

O E E
AT R
, K G D JE
LK PA
signal
i l comparator
t G(f) signal

TY IN ER
SI R TT
ER E A
VCO

IV IN H
N G C
E
U N A
R L E AV

The phase comparator output:

PU CA MIT

et   Av Ac sint   v t 
1 t   v t   phase error
AV I A

2
D TR

The signal
g e(t)
( ) is the input
p to the loop
p filter.
JA EC

If the PLL is in lock, the

sint   v t   t   v t   e t 
EL

phase error is small:

 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
(t) + e(t)
Loop filter
v(t)

O E E
AT R
, K G D JE
LK PA
g(t)
(t)
-

TY IN ER
SI R TT
v(t)

ER E A
t

2kv  v d

IV IN H
N G C
E
0
U N A
R L E AV

Linearized PLL.
PU CA MIT
AV I A

The phase error is: e t   t   2kv  v d

0
D TR

d d d 
d
JA EC

 e t   2k v vt   t  or e t   2k v  e  g t  d  t 

dt dt dt dt
EL

0
 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
(t) + e(t)
Loop filter
v(t)

O E E
AT R
, K G D JE
LK PA
g(t)
(t)
-

TY IN ER
SI R TT
v(t)

ER E A
t

2kv  v d

IV IN H
N G C
E
0
U N A
R L E AV

The Fourier transform of this integro-differential equation:

PU CA MIT

 j 2f   f   2k   f G  f    j 2f  f 

AV I A

e v e

1
Hence,  e  f    f 
D TR

 kv 
1   G  f 
JA EC

 jf 
EL
 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
(t) + e(t)
Loop filter
v(t)

O E E
AT R
, K G D JE
LK PA
g(t)
(t)
-

TY IN ER
SI R TT
v(t)

ER E A
t

2kv  v d

IV IN H
N G C
E
0
U N A
R L E AV

G f 
V  f    e  f G  f    f 
PU CA MIT

Then control
voltage of the  kv 
1   G  f 
AV I A

VCO is:
 jf 
D TR

Now G(f) is G f 
JA EC

designed such kv  1 in the frequency band |f|< W

EL

jf of the message signal.

that:
 FM Demodulators
PLL-FM Demodulator:

IA ,
D T
IN EN
.
A , TM
(t) + e(t)
Loop filter
v(t)

O E E
AT R
, K G D JE
LK PA
g(t)
(t)
-

TY IN ER
SI R TT
v(t)

ER E A
t

2kv  v d

IV IN H
N G C
E
0
U N A
R L E AV

j 2πf
Vf   f 
PU CA MIT

j 2k v
AV I A

H
Hence:
1 d kf
Equivalently, vt   t   mt 
D TR

2k v dt kv
JA EC

Conclusion: As the control voltage of the VCO is proportional

EL

to the message signal, v(t) is the demodulated signal.

 AM Radio Broadcasting

IA ,
D T
IN EN
In a broadcasting system, the receiver performs some

.
A , TM
system functions in addition to demodulating the incoming

O E E
modulated signal.

AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
 Carrier-frequency tuning: to select the desired signal.

N G C
E
U N A
R L E AV

 Filtering: required to separate the desired signal from

PU CA MIT

other modulating signals that may be picked up along the

AV I A

way.
way
D TR

 Amplification: intended to compensate for the loss of

JA EC

signal power incurred in the course of transmission.

EL
 AM Radio Broadcasting

IA ,
D T
IN EN
AM radio
AM radio FM radio
FM radio

.
A , TM
O E E
AT R
RF carrier range
g 535‐1605 kHz 88‐108 MHz

, K G D JE
LK PA
TY IN ER
SI R TT
Midband  455 kHz 10.7 MHz

ER E A
frequency of IF 

IV IN H
N G C
section

E
U N A
R L E AV
IF bandwidth
IF bandwidth 10 kHz
10 kHz 200 kHz
200 kHz
PU CA MIT
AV I A

Typical frequency parameters of commercial

AM and FM radio receivers.
D TR

The
h use off conventional
JA EC

i l AM for
f broadcast
b d i justified
is j ifi d from
f an
economic standpoint. The major objective is to reduce the
EL

cost of implementing the receiver.

 Superheterodyne AM Receiver

IA ,
The receiver most commonly used in AM radio broadcast is

D T
IN EN
called superheterodyne receiver.

.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV

Superheterodyne AM receiver.
PU CA MIT

The common IF frequency chosen is: f IF  455 kHz

AV I A

The bandwidth of the IF amplifier is designed as 10 kHz.

The frequency of the local oscillator is:
D TR

f LO  f c  f IF ;
JA EC
EL

f c : carrier frequency of the desired AM radio signal.

 Superheterodyne AM Receiver

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
The tuning range of the local oscillator: 995 - 2055 kHz.
U N A
R L E AV
PU CA MIT

The RF amplifier is tuned to fc and its output is mixed

with the local oscillator frequency
q y fL0.
AV I A

The bandwidth of the RF amplifier

BC  BRF  2 f IF
D TR

is limited to the range:

JA EC

(Bc = Bandwidth of the AM radio signal (10 kHz)).

EL
 Superheterodyne AM Receiver

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
The local
Th l l oscillator
ill t output
t t cos2f
2 fL0t is
i mixed
i d with
ith th
the received
i d

ER E A
IV IN H
signals:

N G C
r1 t   Ac 1  m1 t cos 2f c t; r2 t   Ac 1  m2 t cos 2f c / t

E
U N A
R L E AV

f c  f L 0  f IF and f c /  f L 0  f IF
PU CA MIT

The mixer output consists of two signals:

AV I A

y1 t   Ac 1  m1 t cos 2f IF t  double

d bl frequency
f term
y2 t   Ac 1  m2 t cos 2f IF t  double frequency term
D TR

m1(t) = the
th desired
d i d signal
ig l and
d
JA EC
EL

m2(t) = the signal transmitted by the radio station transmitting

at the carrier frequency fc/ = fL0 + fIF.
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
Superheterodyne AM Receiver

IA ,
.
Frequency response characteristics of IF and RF amplifiers.
 Superheterodyne AM Receiver

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
The output of the IF amplifier is passed through an
R L E AV
envelope detector which produces the desired audio-band
PU CA MIT

message signal m(t).

AV I A
D TR

Finally, the output of the envelope detector is amplified

JA EC

and the amplified signal drives a loudspeaker.

EL
 FM Radio Broadcasting

IA ,
D T
IN EN
.
Commercial FM radio broadcasting utilizes the frequency

A , TM
band 88 – 108 MHz for transmission of voice and music

O E E
AT R
, K G D JE
LK PA
channels.
h l

TY IN ER
SI R TT
ER E A
IV IN H
The carrier frequencies are separated by 200 kHz and the
N G C
E
peak frequency deviation is fixed at 75 kHz.
U N A
R L E AV
PU CA MIT
AV I A

 The receiver most commonly used in FM radio broadcast

is a superheterodyne type.
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
receiver
Superheterodyne FM receiver.
AT R
FM Radio Broadcasting

A , TM
IN EN
D T
IA ,
.
 FM Radio Broadcasting

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
The mixer brings U N A
g all FM radio signals
g to a common IF
R L E AV
bandwidth of 200 kHz, centered at fIF = 10.7 MHz.
PU CA MIT

The amplitude limiter removes any amplitude variations in

AV I A

the received signal at the output of the IF amplifier by band-

limiting the signal.
D TR
JA EC

A band-pass filter centered at fIF = 10.7 MHz with a

bandwidth of 200 kHz is included in the limiter to remove
EL

higher order frequency components.

 FM Radio Broadcasting

IA ,
D T
IN EN
.
A , TM
O E E
AT R
, K G D JE
LK PA
TY IN ER
SI R TT
ER E A
IV IN H
N G C
E
U N A
R L E AV
A balanced frequency discriminator is used for frequency
PU CA MIT

demodulation.
AV I A

The output of the audio amplifier is filtered by a low pass

filter to remove out-of-band noise.
D TR
JA EC

The output of the low pass filter is used to drive a

EL

loudspeaker.
 References

IA ,
D T
IN EN
.
A , TM
 John G. Proakis and Masoud Salehi, Communication Systems

O E E
AT R
, K G D JE
LK PA
Engineering 2nd Edition,
Engineering. Edition Pearson Education,
Education 2008.
2008

TY IN ER
SI R TT
ER E A
IV IN H
 Simon Haykin, Communication Systems. 4th Edition, Wiley India

N G C
E
Edition, 2008. U N A
R L E AV
PU CA MIT
AV I A

 B
B. P.
P Lathi,
L thi Modern
M d Di
Digital
i l anddA
Analog
l C Communication
i i S Systems. 3rd
Edition, Oxford University Press, 2000.
D TR
JA EC
EL
 EL
JA EC
D TR
AV I A
PU CA MIT
R L E AV
U N A
N G C
IV IN H
ER E A
E
SI R TT
TY IN ER
, K G D JE
O E E
LK PA
AT R
A , TM
IN EN
D T
IA ,
.