SVEC-10

ANALOG
COMMUNICATION
By
T. Ravi Kumar Naidu,
Assistant professor,
Dept of ECE,
SVEC.

III-I SEM
ECE

ALL THE BEST

ANALOG COMMUNICATION
What Does Communication (or Telecommunication) Mean?
The term communication (or telecommunication) means the transfer of some
form of information from one place (known as the source of information) to
another place (known as the destination of information) using some system to
do this function (known as a communication system).

So What Will we Study in This Course?

In this course, we will study the basic methods that are used for
communication in todays world and the different systems that implement
these communication methods. Upon the successful completion of this course,
you should be able to identify the different communication techniques, know
the advantages and disadvantages of each technique, and show the basic
construction of the systems that implement these communication techniques.

Old Methods of Communication

Pigeons
Horseback
Smoke
Fire
Post Office
Drums
etc

Problems with Old Communication Methods

Slow
Difficult and relatively expensive
Limited amount of information can be sent
Some methods can be used at specific times of the day
Information is not secure.

Examples of Todays Communication Methods

All of the following are electric (or electromagnetic) communication systems
Satellite (Telephone, TV, Radio, Internet, )
Microwave (Telephone, TV, Data, )
Optical Fibers (TV, Internet, Telephone, )
Copper Cables (telephone lines, coaxial cables, twisted pairs, etc)

Advantages of Todays Communication Systems

Fast
Easy to use and very cheap
Huge amounts of information can be transmitted

T.R.K.NAIDU

Page 2

Secure transmission of information can easily be achieved

Can be used 24 hours a day.

Basic Construction of Electrical Communication System

A communication system may transmit information in one direction such as TV and

radio (simplex), two directions but at different times such as the CB (half-duplex), or
two directions simultaneously such as the telephone (full-duplex).

Basic Terminology Used in this Communications Course

A Signal:

A System:
Analog Signals:

Digital Signals:
T.R.K.NAIDU

is a function that specifies how a specific variable changes versus

an independent variable such as time, location, height (examples:
the age of people versus their coordinates on Earth, the amount of
money in your bank account versus time).
operates on an input signal in a predefined way to generate an
output signal.
are signals with amplitudes that may take any real value out of an
infinite number of values in a specific range (examples: the height
of mercury in a 10cmlong thermometer over a period of time is a
function of time that may take any value between 0 and 10cm, the
weight of people setting in a class room is a function of space (x
and y coordinates) that may take any real value between 30 kg to
200 kg (typically)).
are signals with amplitudes that may take only a specific number
of values (number of possible values is less than infinite)
Page 3

Noise:

Signal to Noise
Ratio (SNR):

(examples: the number of days in a year versus the year is a

function that takes one of two values of 365 or 366 days, number
of people sitting on a one-person chair at any instant of time is
either 0 or 1, the number of students registered in different classes
at KFUPM is an integer number between 1 and 100).
is an undesired signal that gets added to (or sometimes multiplied
with) a desired transmitted signal at the receiver. The source of
noise may be external to the communication system (noise
resulting from electric machines, other communication systems,
and noise from outer space) or internal to the communication
system (noise resulting from the collision of electrons with atoms
in wires and ICs).
is the ratio of the power of the desired signal to the power of the
noise signal.

Signal
Bandwidth (BW): is the width of the frequency range that the signal occupies. For
example the bandwidth of a radio channel in the AM is around 10
kHz and the bandwidth of a radio channel in the FM band is 150
kHz.
Rate of
Communication: is the speed at which DIGITAL information is transmitted. The
maximum rate at which most of todays modems receive digital
information is around 56 k bits/second and transmit digital
information is around 33 k bits/second. A Local Area Network
(LAN) can theoretically receive/transmit information at a rate of
100 M bits/s. Gigabit networks would be able to receive/transmit
information at least 10 times that rate.
Modulation:
is changing one or more of the characteristics of a signal (known
as the carrier signal) based on the value of another signal (known
as the information or modulating signal) to produce a modulated
signal.

Analog and Digital Communications

Since the introduction of digital communication few decades ago, it has been gaining
a steady increase in use. Today, you can find a digital form of almost all types of
analog communication systems. For example, TV channels are now broadcasted in
digital form (most if not all Kuband satellite TV transmission is digital). Also, radio
now is being broadcasted in digital form (see sirus.com and xm.com). Home phone
systems are starting to go digital (a digital phone system is available at KFUPM).
Almost all cellular phones are now digital, and so on. So, what makes digital
communication more attractive compared to analog communication?
Advantages of Digital Communication over Analog Communication
Immunity to Noise (possibility of regenerating the original digital signal if
signal power to noise power ratio (SNR) is relatively high by using of
devices called repeaters along the path of transmission).
Efficient use of communication bandwidth (through use of techniques like
compression).
Digital communication provides higher security (data encryption).
The ability to detect errors and correct them if necessary.
T.R.K.NAIDU

Page 4

Design and manufacturing of electronics for digital communication

systems is much easier and much cheaper than the design and
manufacturing of electronics for analog communication systems.

Modulation
Famous Types
Amplitude Modulation (AM): varying the amplitude of the carrier based on
the information signal as done for radio
channels that are transmitted in the AM radio
band.
Phase Modulation
(PM): varying the phase of the carrier based on the
information signal.
Frequency Modulation (FM): varying the frequency of the carrier based on
the information signal as done for channels
transmitted in the FM radio band.

Purpose of Modulation

For a signal (like the electric signals coming out of a microphone) to be

transmitted by an antenna, signal wavelength has to be comparable to the
length of the antenna (signal wavelength is equal to 0.1 of the antenna
length or more). If the wavelength is extremely long, modulation must be
used to reduce the wavelength of the signal to make the length of the
required antenna practical.
To receive transmitted signals from multiple sources without interference
between them, they must be transmitted at different frequencies (frequency
multiplexing) by modulating carriers that have different frequencies with
the different information signals.

Classification of Signals
Some important classifications of signals
Analog vs. Digital signals: as stated in the previous lecture, a signal with a
magnitude that may take any real value in a specific range is called an analog
signal while a signal with amplitude that takes only a finite number of values
is called a digital signal.
Continuous-time vs. discrete-time signals: continuous-time signals may be
analog or digital signals such that their magnitudes are defined for all values
of t, while discrete-time signal are analog or digital signals with magnitudes
that are defined at specific instants of time only and are undefined for other
time instants.
Periodic vs. aperiodic signals: periodic signals are those that are constructed
from a specific shape that repeats regularly after a specific amount of time T0,
[i.e., a periodic signal f(t) with period T0 satisfies f(t) = f(t+nT0) for all integer
values of n], while aperiodic signals do not repeat regularly.
Deterministic vs. probabilistic signals: deterministic signals are those that can
be computed beforehand at any instant of time while a probabilistic signal is
one that is random and cannot be determined beforehand.
Energy vs. Power signals: as described below.

T.R.K.NAIDU

Page 5

Energy and Power Signals

The total energy contained in and average power provided by a signal f(t)
(which is a function of time) are defined as
Ef

| f (t ) |

dt ,

and
1
Pf lim
T T

T /2

| f (t ) |2 dt ,

T / 2

respectively.
For periodic signals, the power P can be computed using a simpler form based
on the periodicity of the signal as

PPeriodic f

T t0

| f (t ) |

dt ,

t0

where T here is the period of the signal and t0 is an arbitrary time instant that is
chosen to simply the computation of the integration (to reduce the functions you have
to integrate over one period).

Classification of Signals into Power and Energy Signals

Most signals can be classified into Energy signals or Power signals. A signal is
classified into an energy or a power signal according to the following criteria
a)
b)

Energy Signals: an energy signal is a signal with finite energy and

zero average power (0 E < , P = 0),
Power Signals: a power signal is a signal with infinite energy but
finite average power (0 < P < , E ).

Comments:
1.
2.
3.
T.R.K.NAIDU

The square root of the average power P of a power signal is what is

usually defined as the RMS value of that signal.
Your book says that if a signal approaches zero as t approaches then
the signal is an energy signal. This is in most cases true but not always
as you can verify in part (d) in the following example.
All periodic signals are power signals (but not all nonperiodic signals
are energy signals).
Page 6

4.

Any signal f that has limited amplitude (| f | < ) and is time limited
(f = 0 for | t | > t0 for some t0 > 0) is an energy signal as in part (g)
in the following example.

Exercise 11: determine if the following signals are Energy signals, Power signals,
or neither, and evaluate E and P for each signal (see examples 2.1
and 2.2 on pages 17 and 18 of your textbook for help).
a)

a (t ) 3sin(2 t ), t ,
This is a periodic signal, so it must be a power signal. Let us prove it.
Ea

| a (t ) |

dt

| 3sin(2 t ) |

dt

1
1 cos(4 t ) dt
2

1
9 dt 9 cos(4 t )dt
2

J
Notice that the evaluation of the last line in the above equation is
infinite because of the first term. The second term has a value between
2 to 2 so it has no effect in the overall value of the energy.
Since a(t) is periodic with period T = 2/2 = 1 second, we get
1

Pa

1
| a (t ) |2 dt | 3sin(2 t ) |2 dt

10
0
1

1
1 cos(4 t ) dt
2
0

9
0

1
9 dt 9 cos(4 t )dt
2
0
0

9 9

sin(4 t )
2 4

9
W
2
So, the energy of that signal is infinite and its average power is finite
(9/2). This means that it is a power signal as expected. Notice that the
average power of this signal is as expected (square of the amplitude
divided by 2)

b)

b (t ) 5e 2|t | , t ,
Let us first find the total energy of the signal.

T.R.K.NAIDU

Page 7

Eb

2
| b (t ) | dt

5e 2|t |

dt

25 e 4t dt 25 e 4t dt

25 4t 0
25
e
e 4t

0
4
4
25 25 50

J
4
4
4
The average power of the signal is

1
Pb lim
T T

T /2

1
| b (t ) | dt lim

T T
T / 2

1
T T

25 lim

c)

d)

1
T T

e 4t dt 25 lim

T / 2

T /2

5e 2|t |

dt

T / 2
T /2

4t

dt

0
T /2
25
1
25
1

lim e 4t

lim e 4t
T / 2
0
4 T T
4 T T
25
1
25
1

lim 1 e 2T
lim e 2T 1
4 T T
4 T T
00 0
So, the signal b(t) is definitely an energy signal.
So, the energy of that signal is infinite and its average power is finite
(9/2). This means that it is a power signal as expected. Notice that the
average power of this signal is as expected (the square of the amplitude
divided by 2)
4e 3t , | t | 5
c (t )
,
| t | 5
0,

1
, t 1

d (t ) t
,
0, t 1

Let us first find the total energy of the signal.

Ed

1
dt
t
1

2
| d (t ) | dt

ln t 1

0 J
So, this signal is NOT an energy signal. However, it is also NOT a
power signal since its average power as shown below is zero.

T.R.K.NAIDU

Page 8

The average power of the signal is

1
Pd lim
T T

T /2

1
| d (t ) | dt lim

T T
T / 2

T /2

1
dt
t

1 T

1
T / 2
1
ln t 1 lim ln ln 1
T T
T T

2 T

T
ln

1 T
2
lim ln lim
T T
T

T
2

Using Lehopitals rule, we see that the power of the signal is zero.
That is
lim

e)

T
2
ln 2
lim T 0
Pd lim

T
T T 1

So, not all signals that approach zero as time approaches positive and
negative infinite is an energy signal. They may not be power signals
either.
e (t ) 7t 2 , t ,

f)

f (t ) 2 cos 2 (2 t ),

g)

12 cos 2 (2 t ), 8 t 31
g (t )
0,
elsewhere

t .
.

Basic Signal Operations

A)

Time Shifting: given the signal f(t), the signal f(tt0) is a time-shifted version
of f(t) that is shifted to the LEFT if t0 is positive and to the RIGHT if t0 is
negative.

B)

Magnitude Shifting: given the signal f(t), the signal c + f(t) is a magnitudeshifted version of f(t) that is shifted UP if c is positive and shifted DOWN if
c is negative.

C)

Time Scaling and Time Inversion: given f(t), the signal f(at) is a time-scaled
version of f(t), where a is a constant, such that f(at) is an EXPANDED
version of f(t) if 0 < |a| < 1, and f(at) is a COMPRESSED version of f(t) if |
a| > 1. If a is negative (i.e. a < 0), the signal f(at) is also a time-inverted
version of f(t).

D)

Magnitude Scaling and Magnitude Inversion: given f(t), the signal bf(t) is a
magnitude-scaled version of f(t), where b is a constant, such that bf(t) is an
ATENUATED version of f(t) if 0 < |b| < 1, and bf(t) is an AMPLIFIED
version of f(t) if |b| > 1. If b is negative (i.e. b < 0), the signal bf(t) is also
a magnitude-flipped version of f(t).

T.R.K.NAIDU

Page 9

Unit Impulse Function

The unit impulse function is one of the most important functions that we will be using
extensively in this course. Yet, it is one of the most difficult functions to understand
and use.
A couple of definitions exist for the unit impulse function (t), which is sometimes
also called the Dirac delta function. The following are two definitions:
Graphical Definition:
The rectangular pulse shape shown below approaches the unit impulse function as
approaches 0 (notice that the area under the curve is always equal to 1).

Mathematical Definition:
The unit impulse function (t) satisfies the following conditions:
1.
2.

(t) = 0

(t ) dt

if

t 0,

1.

(therefore it is non-zero only at t = 0).

(so, all of the area under it is concentrated at t =

0)
Properties of the Unit Impulse Function:
a)

Multiplication of a function by the unit impulse response:

f (t ) (t t 0 ) f (t 0 ) (t t 0 )

b)

Sampling of a function using the unit impulse response:

f (t) (t T )dt f (T ) (t T )dt f (T ) (t T )dt f (T )

c)

Obtaining the unit step function from the unit impulse function

T.R.K.NAIDU

Page 10

 ( )d

t0

0,
1,

t0

u (t )

du (t )
(t )
dt

Relation between Fourier Series and Transform

The Fourier Transform (FT) is derived from the definition of the Fourier Series (FS).
Consider, for example, the periodic complex signal gTo(t) with period T0 = 2/0. The
exponential FS of that signal allows the representation of gTo(t) as
gTo (t )

D e

jn 0 t

Where the complex coefficients Dn are evaluated as

Dn

1
T0

T0

gTo (t )e jn 0 t dt

for any period T0 . For convenience, we will use the period

which gives
1
Dn
T0

T0 / 2

To

T0 /2 t T0 /2,

(t )e jn 0 t dt ,

T0 / 2

Note that the magnitudes and phases of the coefficients Dn are, respectively, known
as the amplitude and phase spectrums of gTo(t), or
| Dn | is amplitude spectrum of gTo(t) , and
Dn is phase spectrum of gTo(t).
Now, let us separate the different periods of gTo(t) with zeros such that the different
periods of gTo(t) move away from each other (by holding one period of gTo(t) and
increasing the period duration T0 until it becomes infinite). When T0 , the
signal gTo(t) can be renamed g(t) since it is no longer periodic but only contains one
period of gTo(t). In this case, 0 = 2/T0 0, and the discrete-time signal represented
by the coefficients Dn vs. n becomes (in the limit) a continuous-time signal in terms
of a new variable = n0.
Therefore,
Dn

1
T0

g (t )e

jn 0 t

dt

can be written as
G ( )

g (t )e

jt

dt

where = n0 , and Dn = G(n0)/T0 .

T.R.K.NAIDU

Page 11

Fourier Transform (FT) and Inverse Fourier Transform (IFT)

The FT of signal g(t) is denoted F [g(t)] and is defined as

g (t )e

G ( ) F [ g (t )]

j t

and the IFT of G() is denoted F

g (t ) F

1
2

[G ( )]

dt

(1)

[G()] is defined as

G ( )e

jt

(2)

We say that g(t) and G() for a FT pair, or g(t) G().

Notice that the exponent term in (1) has a negative sign but no negative sign exists in
the exponent in (2). Also, notice that the integration in (1) is in terms of t and it is in
terms of in (2).
Also notice that G() in general is a complex signal that has both a magnitude |G()|
and a phase G() , or
G ( ) G ( ) e j G ( ) .

General Properties of the FT

1)

Symmetry
For REAL g(t)

G() = G*()

|G()| = |G*()| = |G()|

G() = G()
For REAL g(t) with
G() is PURLEY REAL
g(t) = g(t) (even functions)
For REAL g(t) with
G() is PURELY IMAGINARY
g(t) = g(t) (odd functions)
2)

Existence of the FT

For a signal g(t), if | g (t ) | dt , FT F [g(t)] exists.

The opposite is not necessarily true.

This above existence condition comes from the fact that

g (t )e

jt

dt

T.R.K.NAIDU

| g (t )e

jt

| dt

| g (t ) | . | e

jt

| dt

| g (t ) | dt

Page 12

Therefore, if the integration of the magnitude of a function is finite, then the FT

integral is also finite and therefore, the FT exists.
3)

Linearity
If

then

g(t) G()

and f(t) F(),

a g(t) + b f(t) a G() + b F().

Meaning of Negative Frequency

We know that the frequency of any signal is always given in terms of a
positive number. You, for example, would say that the frequency of a radio channel is
650 kHz and never say that it is 650 kHz. So, why does the FT have two parts, a
part that falls to the right of the y-axis with positive frequency and a part falls to the
left of the yaxis with negative frequency (i.e., the part with < 0)?
It is a fact that you can describe the frequency spectrum of any real signal
(such as all the signals that you can generate in the lab) using only one half of the
frequency range (either positive or negative, so no need for both). So, what happens
on the other side? It is known that the magnitude of the spectrum of any real function
is an even function of frequency (it is symmetric about the y-axis), and the phase of
the spectrum is an odd function (it is anti-symmetric about the y-axis, or symmetric
about the origin), so if we know either the negative or the positive halves of the
spectrum of any real function, we can get the other half using these symmetry
properties. For a purely imaginary signal, a similar thing happens and therefore, only
one half of the frequency spectrum is needed. Now, a complexvalued signal is a
combination of a real signal and an imaginary signal and therefore, it carries twice the
amount of information as a realvalued or an imaginary-valued signal. So, a
complexvalued signal would require twice the frequency range to describe its
contents as a real or imaginary signal. This double the amount of information is
basically described on the two halves of the spectrum. In fact, you can use only the
positive half, but you would need two frequency spectrums to describe the frequency
contents of that complex signal.

FT of Important Functions
1)

FT of the Unit Impulse Function (t)

F [ (t )]

jt
(t )e dt

j ( 0 )
dt
(t )e

(t )dt

(t) 1

2)

FT of the Gate Function

T.R.K.NAIDU

0 | t | / 2

t
rect 1 / 2 | t | / 2

1 | t | / 2

Page 13

 /2

1
t
t
F rect rect e jt dt e jt dt
e j t
j

/ 2

/ 2

e
2j

/2

/ 2

j
1 j 2
e
e 2

sin( / 2)
/ 2

The function sin(x)/x is called sinc(x)

rect(t/)
3)

4)

IFT of

sinc(/2)

( 0 )

( 0 )

1
2

)e jt d

1
2

)e j 0 t dt

1
e j 0 t
2

ejot / 2 (0)
FT of cos(0t )
F cos(0t )

jt
cos(0t )e dt

1
e j 0 t e j 0 t e jt dt
2

1
1
e j 0 t e jt dt e j 0 t e jt dt

2
2

1
1
1
F e j 0 t F e j 0 t ( 0 ) ( 0 )
2
2
2

cos(0t) [( 0) + ( + 0)]
Similarly,

sin(0t) (/j) [( 0) + ( + 0)]

What does Spectrum of a Signal Mean?

Now we know how to get the FT G() from a signal g(t) and how to get g(t) back
from G() using the IFT formula. But, what does the FT of a signal physically
mean? The FT of a signal represents the frequency contents of that signal. That is,
what sines or cosines add up together to form the signal. So, it appears that the FT
does the same thing as the FS. In fact, that is true. The difference between the FS and
the FT is that the FS shows what sines and cosines with frequencies that are multiples
of some fundamental frequency 0 combine to produce the PERIODIC signal. The
periodicity of any signal that the FS simulates (even if the signal we are applying the
FS to is not periodic, the FS automatically assumes the periodicity of the signal that is
given in the period that we are integrating over) causes the spectrum of the signal to
be a discretefrequency signal that is defined only at multiples of 0. For general
signals that are not periodic, the FT (which is a form of the FS) becomes a
continuousfrequency signal that is defined for all values of . So, how much
energy (or power) does a sine function, for example, with a specific frequency
contribute to a signal that is not periodic? The answer is generally zero. Only if we
T.R.K.NAIDU

Page 14

take a range of frequencies, such as the range of 150 to 160 Hz, we can say that the
contribution of the sine waves with frequencies in this range is 10 J (this comes from
the fact that the integration of a signal with finite magnitude between the points t =
5 (just before 5) to t = 5+ (just after 5) is always zero.

Properties of the Fourier Transform

If

g(t) G()
G ( )

(1)

g (t )e

jt

dt

G ( )e

jt

(2)

g (t )

a)

1
2

(3)

Symmetry between the FT and IFT

Let s be a time variable and be a frequency variable.
F [G (t )] H ( )

G (t )e

jt

dt

1
2

G (t ) e

jt ( )

dt

By comparing the term between the brackets with Equation (3) above, we get
F [G (t )] H ( ) 2g ( )

G(t) 2g()
b)

Time Scaling
F [ g ( at )]

g (at )e

j t

dt

Let = at

if a > 0:

1
( at )
a

j
1
F [ g (at )] g (at )e
a

1
( at )
a

j
1
F [ g (at )] g ( at )e
a

g ( at )

d = a dt
1
( )
a

j
1
d (at ) g ( )e
a

Let = at

if a < 0:

c)

1
d ( at )
a

d = a dt

g ( )e

1
( )
a

1 1
G
a a

1 1
d G
a a

1
G

a a

Time and Frequency Reversals

F [ g ( t )]

g ( t ) e

jt

dt

Let = t d = dt =

T.R.K.NAIDU

Page 15

g ( )e j ( ) d

g ( )e

j ( )

d G ( )

g(t) G()

d)

(This can also be obtained using (b) above with a = 1)

Time Differentiation
dg (t )
d

F g (t )

dt
dt

d 1

G ( )e jt d

dt 2

to

Since the differentiation is with respect to t and the integration is with respect
, we can bring derivative inside the integral as
1
dg (t )
F

dt
2

d
G ( )e jt d
dt

1
2

F F

jG ( )e

j t

jG ( ) jG ( )

dg(t)/dt jG()
e)

-1

and

dng(t)/dtn (j)nG()

Time Integration

g (t )dt F

g (t ) dt

1

G ( )e jt d dt

2

t

Also here, the two integrals are in terms of different variables, so we can
switch the order of integration as

g ( )d F

g ( ) d

G ( )

d d

Since the period of integration in the inner integral is t , we can

insert a unit step function u(t) that is equal to 1 in this region and 0 outside
and change the limits of integration to be from to , which gives
1

g ( )d F

T.R.K.NAIDU

G ( )

u (t )e

d d .

Page 16

By setting s = , we get
1

g ( ) d F

1
2

G
(

u (t s )e js ds d

G ( )

u (t s)e

js

ds d

Notice that the inner integral is nothing but the FT of u(t+s),

1 jt

e d
G
(

g ( )d F

1
j t
G
(

)
e
d

F F -1 G (0) ( ) F F -1
G ( )
j

1
G (0) ( )
G ( )
j
F

g ( )d

G (0) ( )

f)

1
G ( )
j

Time and Frequency Shifting

g(t t0) G()ejt

g)

1
G ( ) e jt d
j

g(t) ej t G( 0)

and

Multiplying by a Sinusoid
Using the frequency shifting property in (e) above,
g(t)cos(0t) (1/2)[G( 0) + G( +0)]
g(t)sin(0t) (1/j2)[G( 0) G( +0)]

h)

Convolution of Two Signals

The convolution of two signals g(t) and f(t) is defined as
g (t ) * f (t )

g ( ) * f (t )d

f ( ) * g (t )d

Similarly,
G ( ) * F ( )

G ( s) * F ( s)ds

F ( s ) * G ( s)ds

The FT of the convolution of two signals is the product of the two FTs, and the
IFT of the convolution of two signals is the product of the two IFTs
g (t ) * f (t ) G ( ) F ( )

and

g (t ) f (t )

1
G ( ) * F ( )
2

Signal Transmission Through a Linear System

A communication system is usually described by its impulse response h(t). The
impulse response of a system is basically the output of that system when the input
signal to that system is a unit impulse function (t). The impulse response of the
T.R.K.NAIDU

Page 17

system is the timedomain representation of that system. The FT of the impulse

response denoted H() is known as the frequency response of the system.
A signal g(t) that is transmitted through the system with the impulse response h(t)
produces an output signal y(t) that is given by the convolution equation
y (t ) g (t ) * h(t ) .

In frequency domain, this can be represented as

Y ( ) G ( ) H ( ) .

Decomposing this into a magnitude and a phase component gives

| Y ( ) | e j Y ( ) | G ( ) | | H ( ) | e j G ( ) H ( ) .
Distortionless Transmission
When transmitting a signal g(t) through a communication system, the system may or
may not distort the transmitted signal. A system that does not distort the transmitted
signal is allowed to possibly change its magnitude and possibly delay it. If the output
signal a specific communication system is an amplified/attenuated and delayed form
of the input signal, than that system is called and distortionless communication
system.
Therefore, the output of a distortionl ess communication system is
y (t ) kg (t td ) ,

where k is a constant, and td is a time delay that is greater than zero. In frequency
domain, this gives

Y ( ) kG ( )e jt d

H ( ) ke jt d

| H ( ) | k

&

H ( ) t d

A system that is described by the above frequency response is known as a

distortionless system or a linear phase system (the phase of the frequency response
changes linearly with the frequency).
Notice that the impulse response of the system described by the frequency response
H() given above is
h(t ) k (t td ) .

Therefore, inputting an impulse function into this system produces a scaled and
delayed impulse function at the output. The important thing here is that the input
signal is not distorted but only delayed and scaled.

Electric Filters

T.R.K.NAIDU

Page 18

Filters are electric devices that allow part of their input signals to pass and block part
of their input signals. The distinction between the parts that are blocked and the parts
that are allowed to pass is based on frequency. The range of frequencies that are
allowed to pass is called the PASSBAND and the range of frequencies that are
blocked is called STOPBAND. A LowPass Filter (LPF) is a filter that allows low
frequencies up to a specified frequency to pass and block the rest of the frequencies. A
HighPass Filter, on the other hand, allows all frequency components that are above a
specific frequency to pass and block the rest. A BandPass Filter is a filter that allows
frequencies in a specific range that is greater than zero and less than infinity to pass
and blocks frequencies above or below that range.
a)

LowPass Filters (LPF): a major characteristic of LPFs is the bandwidth of

the filter.
The bandwidth of a LPF is half the width of pulse of its
frequency response (i.e., the width of the part of the pulse that is in the
positive range of the frequency which is W1 ). The frequency W1 is also
known as the CUTOFF frequency of the filter.

b)

HighPass Filters (HPF): no bandwidth is defined for a HPF since the

frequency response of that filter extends up to infinite. However, this filter is
characterized by its CUTOFF frequency, which is W1 as shown below.

c)

BandPass Filters (BPF): the BPF is characterized by two frequencies, W1

known as the LOWER CUTOFF frequency, and W2 known as the UPPER
CUTOFF frequency. The bandwidth of that filter is also the width of pulse that
is in the positive frequency region, or BW = W2 W1.

Ideal vs. Real Filters:

The frequency responses for the three types of filters shown above are those of ideal
filters. The reason is that there is an extremely sharp transition between the passbands
and stopbands of these filters. The sharpness of the transition between passband and
T.R.K.NAIDU

Page 19

stopband is determined by something called the ORDER of the filter. The order of the
filter is generally determined by the number of reactive components (capacitors and
inductors) that are used in that filter. A zeroorder filter (no capacitors or inductors) is
basically a flat filter that allows all signals to pass. A firstorder filter (one capacitor
or inductor) is a filter that has very smooth transition between the passband and
stopband. A secondorder filter (number of reactive elements = number of capacitors
+ number of inductors = 2) has a sharper transition. The ideal filters shown above
have in fact an infinite order (require an infinite number of inductors or capacitors,
which makes them unrealizable (cannot be built in practice). Also an ideal filter would
result in an infinite amount of delay between the input and output signals, which
would make it useless even if you were able to build it.

Baseband vs. Passband Communication Systems

Communication systems can be classified into two groups depending on the
range of frequencies they use to transmit information. These communication systems
are classified into BASEBAND or PASSBAND system. Baseband transmission sends
the information signal as it is without modulation (without frequency shifting) while
passband transmission shifts the signal to be transmitted in frequency to a higher
frequency and then transmits it, where at the receiver the signal is shifted back to its
original frequency.
Almost all sources of information generate baseband signals. Baseband signals
are those that have frequencies relatively close to zero such as the human voice (20
Hz 5 kHz) and the video signal from a TV camera (0 Hz 5.5 MHz). A plot of an
audio signal and its frequency spectrum are shown below, where it is seen that the
most of the power of the audio signal is concentrated in the frequency range from (0
4 kHz). The telephone system used for homes and offices, for example, may transmit
the baseband audio signal as it is when the call is local (from your home to your
neighbors home). However, when the telephone call is a longdistance call that is
T.R.K.NAIDU

Page 20

transmitted via microwave or satellite links, the baseband audio signal becomes
unsuitable for transmission and the communication system becomes a passband
system. Similarly, transmitting the video signal from your camera to your TV using a
wire represents a baseband communication while transmitting that video signal via
satellites passband transmission. Therefore, baseband transmission, which is easier
than passband transmission, is usually used when communicating over wires, while
overtheair transmission requires passband transmission. Notice that even over
wires, the transmission may be passband transmission in specific applications.

(a)
(b)
An audio signal in (a) timedomain, and (b) in frequencydomain.
The process of shifting the baseband signal to passband range for transmission is
known as MODULATION and the process of shifting the passband signal to baseband
frequency range at the receiver is known as DEMODULATION. In modulation, one
characteristic or more of a signal (generally a sinusoidal wave) known as the carrier is
changed based on the information signal that we wish to transmit. The characteristics
of the carrier signal that can be changed are the amplitude, phase, or frequency, which
result in Amplitude modulation, Phase modulation, or Frequency modulation.

Types of Amplitude Modulation (AM)

AM is itself divided into different types:
1.

Double Sideband with carrier (we will call it AM): This is the most widely
used type of AM modulation. In fact, all radio channels in the AM band use
this type of modulation.

2.

Double Sideband Suppressed Carrier (DSBSC): This is the same as the AM

modulation above but without the carrier.

T.R.K.NAIDU

Page 21

3.

Single Sideband (SSB): In this modulation, only half of the signal of the
DSBSC is used.

4.

Vestigial Sideband (VSB): This is a modification of the SSB to ease the

generation and reception of the signal.

Double Sideband Suppressed Carrier (DSBSC)

Assume that we have a message signal m(t) with bandwidth (BW) 2B rad/s (or B
Hz) that has a FT
m(t) M().
Let the signal c(t) be a carrier signal (itself carrying no information at all) that is
given by
c(t) = cos(ct),
such that the frequency of the carrier c is much larger than the highest frequency in
the information signal (we set the amplitude of the carrier to be 1, but it can be any
value).
DSBSC Modulation
The DSBSC signal is simply obtained by multiplying the information signal with the
carrier signal as shown in the modulator (or transmitter) block diagram shown below
gDSBSC(t) = m(t)cos(ct) (1/2) [M( c) + M( + c)].

This signal gDSBSC(t) is a modulated signal that has its spectrum centered around c
and c . Therefore, this signal becomes a passband signal with frequency that is
much larger than the maximum frequency in m(t) and can be transmitted using a
relatively short antenna. Also, other similar information signals can be modulated
using cosine functions with different frequencies from c and therefore, will not
overlap or interfere with this modulated signal when transmitted over the same
channel like a air or a coaxial cable.
DSBSC Demodulation
The demodulation process of a DSBSC signal involves obtaining the original
information signal or scaled version of it from the modulated signal. This can be done
by multiplying the modulated signal with another carrier signal that has EXACTLY
T.R.K.NAIDU

Page 22

the same frequency and phase as the carrier signal in the modulator block as seen in
the demodulator block diagram shown below. The amplitude of the two carrier signals
in the modulator and demodulator are not important since they just affect the
magnitude of the different intermediate signals and final output signal of the
demodulator.

The signal labeled e(t) in the demodulator becomes

e (t) = gDSBSC(t)cos(ct) = m(t)cos2(ct) = (1/2) m(t) [1 + cos(2ct)]
= (1/2) m(t) + (1/2) m(t) cos(2ct)
(1/2) M() + (1/2) [M( 2c) + M( + 2c)].
However, as seen in the FT of e(t), the original message signal (scaled by 1/2) is
present but also other components with frequencies centered around 2c and 2c.
These components are undesired and must be removed fop us to get the message
signal. This can be done using a LPF (a filter centered around zero frequency that
permits low frequencies to pass and rejects high frequencies). The BW of the filter
must be 2B rad/s (or B Hz) or possibly slightly higher (but not much higher that it
will allow the high-frequency components around 2c and 2c to partially or
completely pass).
Therefore, the output signal f(t) of the LPF will be
e (t) = (1/2) m(t)

(1/2) M().

This is simply a scaled version of the original transmitted signal that can be easily
amplified to obtain the original signal exactly.

T.R.K.NAIDU

Page 23

Timedomain representation of the different signals obtained in the DSBSC

modulationdemodulation process.

T.R.K.NAIDU

Page 24

Frequencydomain representation of the different signals obtained in the DSBSC

modulationdemodulation process.

T.R.K.NAIDU

Page 25

Double Sideband Suppressed Carrier (DSBSC) (Continuation)

The modulation and demodulation technique discussed last lecture require the
existence of high quality multipliers (usually called mixers in communication
applications). The use of multipliers is generally undesirable for reasons that are
beyond the scope of this course. So, we need to find DSBSC modulation techniques
that do not depend on multipliers.
To avoid the use of multipliers, several multiplier-less methods exist.
NonLinear Modulators
In the following block diagram for DSBSC modulation, the message signal m(t) with
a BW of 2B rad/s and the carrier signal c(t) = cos(Ct) are not multiplied, but are
added the upper path and subtracted in the lower path.

The signals x1(t) and x2(t) therefore are

x1 (t ) c (t ) m(t ) cos( C t ) m(t )
x1 (t ) c (t ) m(t ) cos( C t ) m(t )

These signals are passed through two exactly similar nonlinear devices that have
scale there input signals and add it to a scaled version of the square of their input
signals.
y1 (t ) a cos(C t ) m(t ) b cos(C t ) m(t )

a cos(C t ) am(t ) bm 2 (t ) 2bm(t ) cos(C t ) b cos 2 (C t )

am(t ) bm 2 (t ) 2bm(t ) cos(C t ) a cos(C t )

Undesired

Undesired

Desired

y2 (t ) a cos(C t ) m(t ) b cos(C t ) m(t )

Undesired

b
2

Undesired

b
cos( 2C t )
2
Undesired

a cos(C t ) am(t ) bm 2 (t ) 2bm(t ) cos(C t ) b cos 2 (C t )

am(t ) bm 2 (t ) 2bm(t ) cos(C t ) a cos(C t )




Undesired

So,

Undesired

Desired

Undesired

b
2

Undesired

b
cos(2C t )
2
Undesired

z (t ) y1 (t ) y2 (t )
2am(t ) 4bm(t ) cos(C t )


Undesired

T.R.K.NAIDU

Desired

Page 26

The sum (or actually the different) of the outputs of the two nonlinear devices
contains two terms that can be described as follows:
2am(t )

is the original message signal. This is an UNDESIRED

BASEBAND signal with bandwidth BW = 2B rad/s.

4bm(t ) cos(C t )

is the message signal multiplied by the carrier. This is

the DESIRED signal with frequency centered around C.

It is obvious that since the desired signal 2bm(t ) cos(C t ) occurs around C ,
we can use a BPF with a passband region centered around C and BW = 4B rad/s
(or 2B Hz) to allow this signal and reject the first component 2am(t).
Notes:
Many nonlinear devices exist such as transistors and diodes. These devices
operate nonlinearly around their biasing regions. The nonlinearity of these
devices may be in the form of an exponential relationship that can be
approximated as a square relation for signals with low amplitudes in specific
operation regions of these devices.
The modulation system shown above can be used for demodulation too. Just
replace the BPF with a LPF of BW = 4B rad/s and feed the carrier signal to
one input and the DSBSC modulated signal to the other input. (Exercise: show
that the output of that system is a scaled version of the message signal)
The following block diagram is a simpler DSBSC modulator, where the non
linear device has a = 0 (Exercise: verify that this system is able to do DSBSC
modulation). However, this system can be used for demodulation only if the
magnitude of the message signal is significantly small such that the square of
that signal is much lower (and therefore can be ignored) than the magnitude of
the message signal.

Switching Modulators
Another type of DSBSC modulator/demodulators is switching modulation. The idea
of switching modulation is the other carriers such as square waves can be used
instead of sinusoidal waves to modulate the message signal. Since a square wave can
be represented in terms of a sum of sinusoids with fundamental frequency o equal
to the frequency of the square wave. So, if a message signal is modulated using a
square wave with frequency equal to the desired carrier frequency C and then this
modulated signal is filtered using a BPF centered at C with bandwidth twice the
bandwidth of the message signal, the resulting signal is a DSBSC signal.
The square wave modulation can be performed using one of many configurations:
A) Diodebridge Modulator:
T.R.K.NAIDU

Page 27

A typical configuration of diodebridge modulator is shown below, where c(t)

= cos(Ct).

When c(t) < 0, all diodes are turned off and therefore, the circuit simplifies
to the following

Therefore, the current of the message source m(t) passes through the 1
resistor and creates a voltage across the resistor that is equal to m(t) Volts.
However, when c(t) > 0, all diodes become forwardbiased (they become like
conductors), and therefore the circuit simplifies to

So, all current of the message source passes through the short circuit and no
current passes through the resistor. This leaves the voltage across the resistor
to be zero.
T.R.K.NAIDU

Page 28

Hence, the signal at the input of the BPF is equal to the message signal when
the carrier is negative and equal to zero when the carrier is positive. This is
simply like multiplying the message signal with a square wave that has a
frequency equal to the carrier frequency. The BPF removes the DC term and
all higher harmonics of this signal resulting in a DSBSC signal at its output.
This circuit can also be used for demodulating the DSBSC signal by feeding
this signal in place of the message signal and replacing the BPF with a LPF.
The ring modulator works in a similar way except that it results in having a
bipolar square wave multiplied by the message signal (see page 159 of your
textbook for details).
b) Ring modulator
Consider the scheme shown in Fig.

We assume that the carrier signal c (t ) is much larger than m(t ) . Thus c (t ) controls
the behavior of diodes which would be acting as ON-OFF devices. Consider the
carrier cycle where the iterminal 1 is positive and terminal 2 is negative. T1 is an
audio frequency
transformer which is essentially an open circuit at the frequencies near about the
carrier. With the polarities assumed for c (t ) , D1, D4 are forward biased, where as
D2, D3 are reverse biased. As a consequence, the voltage at point a gets
switched to a' and voltage at point b to b' . During the other half cycle of c (t ) ,
D2 and D3 are forward biased where as D1 and D4 are reverse biased. As a
result, the voltage at a gets transferred to b' and that at point b to a' . This
implies, during, say the positive half cycle of c (t ) , m(t ) is switched to the output
where as, during the negative half cycle, m(t ) is switched. In other words,
v (t ) can be taken as
v(t) = X P t m(t),
where X P t is square wave
X P t

n 1

n 1

/ 2n 1 cos 2f c t 2n 1

V t 4 m(t ) cos( wc t )

Amplitude Modulation (AM)

The DSBSC modulation is one type of modulation in which the information (or
message) is carried on the amplitude of a sinusoidal signal. Another type of this
T.R.K.NAIDU

Page 29

modulation is what we can call Double Side Band Including Carrier or Full AM (or
simply AM).
Problem 1: The problem with DSBSC modulation is that its demodulation ALWAYS
requires the availability of the carrier signal in the demodulator. This was assumed in
our previous discussion without giving any method for obtaining this carrier from the
received signal at the demodulator. Notice that the carrier at the demodulator must
have the same frequency and phase of the carrier at the transmitter or some parts of
the message signal will be lost (try the Matlab program named DSBSC on WebCT).
In fact, the generation of the carrier signal at EXACTLY the same frequency and
phase of the carrier at the modulation is relatively expensive and may drive the cost of
the demodulator to be higher (you will study later in the course methods for obtaining
the carrier frequency from a received signal).
Conclusion: For applications where ONLY ONE modulator but MANY
demodulators are required, as it is the case for radio broadcasting, use any method for
modulation even if it relatively expensive if it will reduce the cost of the demodulator.
This will save a lot of money for many people at the expense of increasing the cost for
the broadcasting entity.
Solution 1: The cost of the demodulator can be significantly reduced by using a
modulator that DOES NOT require the generation of the carrier at the demodulator
but uses what is called an ENVELOPE DETECTOR method to demodulate the
amplitude modulated signal.
Problem 2: An envelope detector is a device that tracks the envelope (the upper of
lower cover) of a modulated signal. Since most message signals are bipolar in nature
(their amplitude ranges from a negative value A to a positive value +A), therefore,
when the modulated signal is a DSBSC, where the envelope of that signal sometimes
touches zero, the envelope detector does not follow the message signal m(t), but
follows either |m(t)| or |m(t)|. So, the DSBSC modulation method is not suitable
when we want to use envelope detectors to do the demodulation.
Solution 2: To avoid the crossing or touching of the upper and lower envelopes of the
modulated signal, the signal that multiplies the carrier must always be positive or
negative but not positive sometimes and negative sometimes. This can be achieved by
adding a constant to the original message signal to lift it up so that the sum of the
constant and the message signal is always positive (or always negative).
Assume we have a message signal m(t) such that A < m(t) < A, where A is a
positive constant. Therefore,
q (t ) A m(t ) 0

The AM signal is obtained by using the same modulation process of the DSBSC
where a carrier signal c(t) = cos(Ct) is multiplied by the signal q(t) shown above to
give
g AM (t ) q (t ).c (t ) [ A m(t )] cos(C t ) A cos(C t ) m(t ) cos(C t .

Notice that gAM(t) contains a DSBSC signal (m(t)cos(Ct)) and a scaled carrier term
(Acos(Ct)). The carrier term CARRIER NO INFORMATION at all. It is there to
make sure that the upper and lower envelopes do not touch each other. This is the
T.R.K.NAIDU

Page 30

reasoning for naming this type of modulation FULL AM, while naming the DSBSC as
such (because it is similar to Full AM but after suppressing (removing) the carrier
term).
The FT of the AM signal becomes
g AM (t ) A ( C ) ( C )

1
M ( C ) M ( C ) .
2

Notice that the two delta functions centered at C represent the added carrier.
So, this signal can be demodulated using an envelope detector, which is extremely
simple and is very cheap. As it is almost always the case, you cannot get something
for free. So, where is the cost for this convenience? The cost for this convenience is
that part of the power of the modulated signal carries no information (in fact most of
the power of the modulated signal carries no information). A transmitted signal that
carries no information to some degree may be considered as wasted.
To find the efficiency of the AM transmission, we need to find the ratio of the power
of the signal that carries the information to the total power in the AM modulated
signal. Assume that the message signal is cos(mt) (what is called a singletone
signal) with amplitude A where 0 1 (i.e., a fraction of the amplitude of the
carrier component in the AM ( is called the modulation index)), or
z (t ) A cos(mt ) cos(C t )
A
cos (C m )t cos (C m )t
2
A
A

cos (C m )t
cos (C m )t
2
2

The power of this signal is the sum of the two powers of the two sinusoids (because
they have different frequencies (refer to page 39 of your textbook Parsevals
Theorem)
A

2
Pz
2

2
2

A .

The power of the carrier term in the modulated signal is

w(t ) A cos(C t ) ,

Pw

A2
2

Therefore, the efficiency of the AM transmission becomes

2

Pz

Pz Pw

T.R.K.NAIDU

2
2

.
2
2
A
A2 2

2
2

Page 31

Since 0 1 to avoid the touching of the upper and lower envelopes of the
modulated signal, the MAXIMUM efficiency of the AM signal is
max

1
0.333 33.3% .
1 2

So, no matter what we do, we cannot bring the efficiency of AM modulation to more
than one third, or stated in other words, at least 2/3 of the power of the AM signal is
wasted.
Generation of AM Signals
Since, AM signals are simply similar to DSBSC modulation but the information
signal is shifted by a constant first and then modulated by the carrier, the generation
of AM signals can be performed using ANY DSBSC modulation technique. Also, the
demodulation of AM signals can be performed using ANY DSBSC demodulation
technique. However, the opposite is not always true. So, not all AM modulation and
demodulation techniques will work for DSBSC modulation and demodulation.
The advantage of AM over DSBSC is the ability of easily modulating and
demodulating it. A simple AM modulator is shown below.

The signal generated by the combined sources is the sum of the message and carrier.
The signal at the other side (the right hand side) of the diode is the halfwave rectified
signal of the sum of the message and carrier. The spectrum of this halfwave rectified
signal contains many components at frequencies around 0, c, 2c, 3c, etc.,
including the terms cos(ct) and m(t)cos(ct). In fact, these two signals are the
only ones around c and therefore, will only be the signals that pass through the BPF
to create the AM signal.
Demodulation of AM Signals
1.

Rectifier Detector

The circuit below is similar in nature to the circuit above where the AM signal is fed
to the diode which rectifies the signal. The signal to the right of the diode is one that
contains many components at frequencies around 0, c, 2c, 3c, etc., including
the message unmodulated signal m(t) and a DC component A. All components other
than these two are filtered by the LPF and the capacitor at the end has the function of
blocking that DC component so that the output of the signal is a scaled version of m(t)
without any DC.
T.R.K.NAIDU

Page 32

2.

Envelope Detector

The envelope (cover of a sinusoidal wave) detector shown below is a modification of

the rectifier detector. The diode is either forwardbiased (when the AM signal is
higher in value than the voltage across the capacitor), or reversebiased (when the
AM signal is lower in value than the voltage across the capacitor). When the diode is
forward biased, it acts like a short circuit and the voltage across the capacitor follows
the voltage of the source. When the diode is reversebiased, it is acting like an open
circuit and the capacitor simply discharges through the resistor. If the value of the
timeconstant of the capacitor and resistor = RC is suitable (not too large or too
small), the charging and discharging of the capacitor results in a signal that follows
the message signal with some small ripples. If the value of = RC is too large, the
discharge may be too slow that some parts of the envelope of the AM signal are not
followed. If the value of = RC is too small, the discharge may be too fast that the
output signal contains extremely large ripples and it may be hard for any added
lowpass filter to remove these large ripples.

The following figure illustrates the modulation and demodulation of AM signals in

timedomain.

T.R.K.NAIDU

Page 33

The following figure illustrates modulation and demodulation of AM in frequency

domain (Note: the demodulation techniques used above and below are slightly
different (above an envelope detector is used and below, a rectifier detector is used)).

T.R.K.NAIDU

Page 34

Quadrature Amplitude Modulation (QAM)

If you noticed, you would see that so far we have always used a carrier that is c(t) =
cos(Ct) to modulate the message signals both in AM and DSBSC. There is nothing
special about using a cosine instead of a sine carrier. In fact, we can transmit two
T.R.K.NAIDU

Page 35

signal using the same carrier frequency but using cos(Ct) for one of the message
signals and sin(Ct) for the other signal. These transmitted signals if transmitted over
the same channel would not interfere with each other and can be demodulated.
Consider the following block diagram of a Quadrature Amplitude Modulation (QAM)
and Demodulation system:

The modulator/demodulator system shown above clearly is able to modulate and

demodulate two different signals without any interference. However, if the generation
of the carrier at the demodulator had even small phase or frequency errors, the
demodulated signals will interfere at the outputs. The following figure illustrate what
happens when the carrier at the demodulator has a small frequency error (must
be a small value much less than c) and/or a small phase error .

T.R.K.NAIDU

Page 36

If the carrier at the receiver has a small frequency error (but a phase error =0),
we see that the two output signal become
1
m1 (t ) cos(t ) m2 (t ) sin(t )
2
.
1
r2 (t ) m1 (t ) sin( t ) m2 (t ) cos( t )
2
r1 (t )

Clearly, in this case, the output signals are not purely either of the two message
signals but a combination. The ratio of message 1 to message 2 at the different outputs
changes as a sinusoid with a frequency equal to the frequency error .
If the carrier at the receiver has a phase error (but a frequency error = 0), we see
that the two output signal become
1
m1 (t ) cos( ) m2 (t ) sin( )
2
.
1
r2 (t ) m1 (t ) sin( ) m2 (t ) cos( )
2
r1 (t )

In this case, the output signals are not also mixed signals of the two messages.
However, the ratio of the two messages in each output is a constant.

Single Side Band Suppressed Carrier (SSBSC or SSB for short)

T.R.K.NAIDU

Page 37

In QAM, we have seen that you can modulate two signals at the same frequency using
sine and cosine carriers and demodulate them without having interference. However,
still each signal would require twice the bandwidth of the original message signal. In
fact, for real timedomain signals, the parts of the spectrum of that signal that fall in
the positive region and negative region frequency carry exactly the same information
and therefore, need not both to be transmitted. This type of amplitude modulation is
called Single Side Band (SSB) modulation. Since the carrier is usually not transmitted
in this type of modulation, it is called SSBsuppressed carrier (SSBSC). For short,
we will call it SSB modulation.
To understand the process of transmitting only one of the side bands of a signal, we
will start discussing the process in frequencydomain and then try to find the time
domain representation of the SSB signal. So, our aim in the next few pages is to write
an expression of gUSB(t) and gLSB(t), which are the timedomain signal
representations of the upper and lowerside band signals, in terms of the original
message m(t).
Consider the baseband message signal m(t) with the frequency spectrum M()
shown in part (A) of the figure in the next page. Assuming that the signal m(t) is a
real signal, the magnitude of its spectrum is an even function and the phase of its
spectrum is an odd function (so, the information contained in the part of the spectrum
with positive frequency is exactly the same as the information contained in the part
with negative frequency). The spectrum M() can be split into two parts called M+
() and M() as shown in parts (B) and (C). M+() is the part of M() with
positive frequency and M() is the part with negative frequency. It is clear that
M ( ) M ( ) M ( ) .

Both M+() and M() can be represented in terms of M() by noticing that each
is a step function that multiplies M(). That is
M ( ) M ( )u ( )
M ( ) M ( )u ( )

(1)

Since the step function u() and the sign function sgn() are related by a scaling
and a shifting up/down relationship given by
sgn( ) 2u ( ) 1
1 1
u ( ) sgn( )
,
2 2
1 1
u ( ) sgn( )
2 2

T.R.K.NAIDU

(2)

Page 38

We can substitute for the step functions in the two equation labeled (q) with the
sgn() function as shown above in (2). This gives
T.R.K.NAIDU

Page 39

 1 1

sgn( )
2 2

M ( ) M ( )u ( ) M ( )

1
1
M ( ) M ( ) sgn( )
2
2 jM ()
h

1 1

M ( ) M ( )u ( ) M ( ) sgn( )
2
2

1
1
M ( ) M ( ) sgn( )
2
2

(3)

jM h ( )

Let us define the term M()sgn() to be jMh(), where the subscript h stands for
Hilbert. Therefore,
M h ( ) jM ( ) sgn( ) .

(4)

So, we can write M+() and M() as

1
M ( )
2
1
M ( ) M ( )
2

M ( )

j
M h ( )
2
.
j
M h ( )
2

(5)

If we define m+(t) and m(t) to be the inverse Fourier transforms of M+() and M
(), respectively, and mh(t) to be the inverse Fourier transform of Mh(), we get
1
m(t )
2
1
m (t ) m(t )
2
m (t )

1
jmh (t )
2
.
1
jmh (t )
2

(6)

Now, let us look carefully at the Mh(), which is defined in (4). This signal is a
transformed version of M(). The transformation is known as the Hilbert transform.
Based on the shape of sgn(), we see that this transform simply flips the positive part
of the spectrum of M() and multiplies the whole thing by j. If we define the
transfer function of this transform to be H(), where
H ( ) j sgn( ) ,

(7)

we see that H() can be represented graphically using any of the two equivalent
forms shown below.

T.R.K.NAIDU

Page 40

So, this transform can be though simply as a transform that produces input signal
shifted by an angle of /2 to the right (shifted by /2).
***
As an exercise, apply this transform to the Fourier transform of a cosine
function
cos(ct). You should get a sin(ct). If you apply it to sin(ct), you will
get cos(ct).
To get the representation of the Hilbert transform in timedomain (i.e., to find mh(t)
in terms of m(t)), from the table of Fourier transforms we notice the FT pair
sgn(t )

2
.
j

Using the symmetry between time and frequency property given in the table of FT
properties, we get
2
2 sgn( ) ,
jt

or
1
j sgn( )
t
Now, since
M h ( ) j sgn( ) M ( ) ,

we see that mh(t) is the convolution in timedomain of the inverse Fourier transforms
of two functions jsgn() and M(), which gives

T.R.K.NAIDU

Page 41

1
* m(t )
t
.

1 m( )

d
t

mh (t )

(8)

This is the Hilbert transform of a signal in timedomain.

Now we are ready to represent the SSB signals in terms of the message signals. First,
we notice that the USB and LSB signals in frequencydomain shown in parts (E) and
(F) of the above figure can be represented in terms of shifted versions of M+() and
M() as
GUSB ( ) M ( C ) M ( C )
GLSB ( ) M ( C ) M ( C )

(9)

So,

gUSB (t ) m (t )e jC t m (t )e j C t
g LSB (t ) m (t )e jC t m (t )e j C t

(10)

Since,
1
m(t )
2
1
m (t ) m(t )
2
m (t )

1
jmh (t )
2
1
jmh (t )
2

(11)

Substituting this (11) in (10), we get

1
1
1
1
m(t )e j C t jmh (t )e j C t m(t )e j C t jmh (t )e j C t
2
2
2
2
m(t ) cos(C t ) mh (t ) sin(C t )

gUSB (t )

1
1
1
1
m(t )e j C t jmh (t )e j C t m(t )e j C t jmh (t )e j C t
2
2
2
2
m(t ) cos(C t ) mh (t ) sin(C t )

g LSB (t )

(12)
where mh(t) is given by (8).
These are the representations of the upper and lowerside band signals in terms of
message signal.

T.R.K.NAIDU

Page 42

Vestigial Side Band Modulation

As mentioned last lecture, the two methods for generating SSB modulated signals
suffer some problems. The selectivefiltering method requires that the two side bands
of the DSBSC modulated signal which will be filtered are separated by a guard band
that allows the bandpass filters that are used to have nonzero transition band (so it
allows for real filters). An ideal Hilbert transform for the phaseshifting method is
impossible to build, so only an approximation of that can be used. Therefore, the SSB
modulation method is hard, if not impossible, build. A compromise between the
DSBSC modulation and the SSB modulation is known as Vestigial Side Band (VSB)
modulation. This type of modulation is generated using a similar system as that of the
selectivefiltering system for SSB modulation. The following block diagram shows
the VSB modulation and demodulation.

T.R.K.NAIDU

Page 43

The above example for generating VSB modulated signals assumes that the VSB filter
(HVSB()) that the transition band of the VSB filter is symmetric in a way that adding
the part that remains in the filtered signal from the undesired side band to the missing
part of the desired side band during the process of demodulation produces an
undusted signal at baseband. In fact, this condition is not necessary if the LPF in the
demodulator can take care of any distortion that happens when adding the different
components of the bandpass components at baseband.
To illustrate this, consider a baseband message signal m(t) that has the FT shown in
the following figure.
The DSBSC modulated signal from that assuming that the carrier is 2cos(Ct) (the 2
in the carrier is placed there for convenience) is
g DSBSC (t ) m(t ) cos( C t )

In frequencydomain, this gives

G DSBSC ( ) M ( C ) M ( C )

T.R.K.NAIDU

Page 44

Passing this signal into the VSB filter shown in the modulator block diagram above
gives
GVSB ( ) H VSB ( ) M ( C ) M ( C ) .

Note that the VSB filter is not an ideal filter with flat transfer function, so it has to
appear in the equation defining the VSB signal.
Now, let us demodulate this VSB signal using the demodulator shown above but use a
nonideal filter HLPF() (the carrier here is also multiplied by 2 just for convenience)

X ( ) H VSB ( C ) M ( 2 C ) M ( )

at 2
Baseband
C

H VSB ( C ) M ( ) M ( 2 C )


baseband
at 2C

Passing this through the nonideal LPF in the demodulator gives an output signal that
we will call Z(). This signal is given by

Z ( ) H LPF ( ) H VSB ( C ) M ( ) H VSB ( C ) M ( )

H LPF ( ) H VSB ( C ) H VSB ( C ) M ( )

For this communication system to not distort the transmitted signal, the output signal
Z() must be equal to the input signal (or a scaled and shifted version of it).
Z ( ) M ( ) H LPF ( ) H VSB ( C ) H VSB ( C ) M ( ) .

This gives us the following relationship between the LPF at the demodulator and the
VSB fitler at the modulator
H LPF ( )

1
.
H VSB ( C ) H VSB ( C )

So, this filter must be a LPF that has a transfer function around 0 frequency that is
related to the VSB filter as given above. To illustrate this relationship, consider the
following VSB BPF example.

T.R.K.NAIDU

Page 45

Another example follows.

Carrier Acquisition
T.R.K.NAIDU

Page 46

In the different AM modulation methods where a carrier is not transmitted (DSBSC,

SSB, etc), we assumed that the carrier that was used in the transmitter is also
available at the receiver. This is a big assumption since it may be easy to generate a
sine wave, however, it is difficult to get a sine wave with the SAME FREQUENCY
and PHASE as the sine wave that was used in the transmitter. Therefore, the concept
of carrier acquisition is important for these types of modulation.

Carrier Acquisition in DSBSC

Since at the input of a DSBSC demodulator system, only the transmitted signal (and
random noise) is available, the demodulator must have some method to extract the
carrier from the received signal. It is known that the carrier frequency is located at the
middle of the DSBSC signal as shown below.

In theory, a filter with an extremely narrowband (few Hz) that is centered at the
carrier frequency c would be able to extract a signal that represents the carrier
frequency. However, this assumes that the message signal has a DC component that
when modulated this DC component moves to the carrier frequency. Having a DC
component in the message signal that will be modulated using DSBSC is not
necessarily the case. In fact, it is generally not the case. Take for example an audio
signal. Audio signals do not contain DC since and audio signal travels through the air
molecules in the form of variations in the air pressure around the atmospheric
pressure. A microphone receives these variations in the air pressure and generates an
electric signal with no DC. Also, digital data in the form of binary numbers is
generally represented by 1s and 0s that have voltages of +5 V and 5 V, respectively.
Such a signal generally has no DC. So, in practice, the message signals that are
modulated using DSBSC (or other modulation techniques) do not have a DC
component as shown in the following figure.

When this signal is DSBSC modulated, we get the following signal

T.R.K.NAIDU

Page 47

Clearly, in this case, the use of a narrowband BPF to extract the carrier frequency
from the modulated signal is not possible because no components in the modulated
signal fall at the carrier frequency. Therefore, another method must be used in general.
Signal Squaring Method
The block diagram of this system for extracting a carrier from the received DSBSC
signal is shown bellow.

As mentioned earlier, most message signals contain no DC components and therefore

their DSBSC modulated signals contain no components at the carrier frequency.
However, passing such a DSBSC modulated signal through a squaring device as
shown in the block diagram above will give a signal x(t) that is given by
x (t ) m(t ) cos( c t ) m 2 (t ) cos 2 ( c t )
2

m 2 (t )
m 2 (t ) m 2 (t )
1 cos(2 c t )

cos( 2 c t )
2
2
2

This signal contains a form of the square of the message signal (this is a baseband
signal) and the modulation of this signal around twice the carrier frequency. Although
m(t) may have no DC component, m2(t) will ALWAYS contain DC since it is always
positive.

Notice that F{m2(t)} is NOT EQUAL TO M 2() but is equal to the M()*M(). So, it
has twice the bandwidth of M() and contains a DC component. So, the spectrum of
x(t) will be

T.R.K.NAIDU

Page 48

Passing this signal through a narrowband BPF with center frequency of 2c will pass
the sinusoid at that frequency a small part of m2(t).cos(2ct) around that frequency.
So, the output of the BPF will look like the following.

To purify the sinusoid with twice the carrier frequency from the remaining part of
m2(t).cos(2ct), we feed this signal to a phase locked loop (PLL), which produces a
sinusoid with the same frequency and phase as that of the sinusoid at the input but
rejects the additional signal that represents a distortion (discussion about PLLs will be
next). So, the output of the PLL will have the following spectrum.

This is a sinusoid with twice the carrier frequency. Passing this signal through a
device that divides the frequency of its input signal by two will give a sinusoid with
the carrier frequency (the desired signal).

Phase Locked Loop (PLL)

In the block diagram shown above, one of the blocks was a phase locked loop (PLL).
The PLL is constructed as shown below.
T.R.K.NAIDU

Page 49

VoltageControlled Oscillator (VCO): is a device that produces a sine wave with a

frequency that is linearly proportional to the value of the input signal (when the input
increases, the frequency of the output increases and vice versa). The frequency of the
output of the VCO when the input is zero is called the Free Running Frequency
(FRF).
Assuming that the input signal to the PLL is Asin(ct+i), this signal is multiplied by
the output of the PLL. Let us assume that the output of the PLL was at some instant
equal to Bcos(ct+o). Therefore, both the input and output have the same frequency
but different phases. The signal x(t) becomes
x(t ) AB sin( c t i ) cos( c t o )

AB
sin( i o ) sin(2 c t i o ) .
2

The first term in x(t) is a baseband signal since the sine does not contain the carrier
frequency. The second component is a high frequency component at twice the carrier
frequency. Therefore, the LPF passes the first component and blocks the second
component. So, y(t) is
y (t )

AB
sin( i o ) .
2

Now looking at this signal carefully, we see that it is zero when the two angles i and
o are equal. So, assuming that i and o were equal at some time (i = o), which
means that the signal z(t) is Bcos(ct+i), the signal input to the VCO will be zero.
With this, the output of the VCO will be the same signal as what it has been
outputting before (i.e., Bsin(ct+i)). Now assume that the phase i of the input to
the PLL started to increase (or started to decrease) slowly such that i is slightly
greater than o (or i is slightly less than o). The signal z(t) will be a small positive
(or negative) value. This signal will force the VCO to speed up (or slow down) and
therefore produce a sine wave with an increasing (or decreasing) phase until the phase
of the output of the VCO becomes equal to the phase of the input signal. At this
moment, the signal z(t) will be zero and this will inform the VCO that it has the same
phase as the phase of the input signal so it to stop speeding up (or slowing down) and
continue at that frequency and phase.

Characteristics of a PLL
PLL have several characteristics that are summarized as follows:

T.R.K.NAIDU

Page 50

1.

2.

3.

CAPTURE Range of a PLL: every PLL is built to operate in a specific range

of frequencies. If the input signal has a frequency in this range, the PLL will
be able to capture the phase and frequency of the input signal and start to
produce a sine with that frequency and phase.
LOCK Range of a PLL: once the PLL has been able to capture a signal and
attach its frequency and phase to it, the frequency and phase of the input
signal can change and the PLL will still be able to track these changes. The
LOCK range of a PLL is always LARGER THAN the CAPTURE range.
Noise filtering: if the input to the PLL contains some lowpower noise or
some lowpower distortion that makes the input signal a nonpure sinusoid,
PLLs can in general purify the input signal from the noise or distorting
signal and produce a very clean sinusoid (So, in the signal squaring method
discussed above, inputting the sinusoid with the distortion signal to the PLL
produces a pure sinusoid with the same frequency and phase as the input
sinusoid).

Costas Loop
This is another method for extracting the carrier from the received signal in a DSBSC
receiver. The block diagram of the Costas loop is given below.

The signals x1(t) and x1(t) in the block diagram above are simply obtained by
multiplying the received DSBSC modulated signal respectively by cos(ct+o) and
sin(ct+o), which is obtained by shifting cos(ct+o) by /2. So,
x1 (t ) m(t ) cos( c t i ) cos( c t o )

m(t )
cos(2 c t i o ) cos( i o )

2
around 2c
around 0

m(t )
x 2 (t ) m(t ) cos( c t i ) sin( c t o )
sin( 2 c t i o ) sin( i o )

2
around 2 c
around 0

Since both of these signals has a component around 2c and a component around 0
frequency, the highfrequency components of these filters will be rejected by the LPF
and only the lowfrequency components will pass. So,
m(t )
cos( i o )
2
.
m(t )
y 2 (t )
sin( i o )
2
y1 (t )

These two signals are multiplied to produce the signal z(t), which is given by
T.R.K.NAIDU

Page 51

m 2 (t )
m(t )
m(t )

cos( i o )
sin( i o )
cos( i o ) sin( i o )
4
2
2

m 2 (t )
.

sin(0) sin 2( i o )
8
m 2 (t )

sin 2( i o )
8

z (t )

The signal z(t) is a baseband signal. It is zero only if i = o. When this is the case,
the output signal w(t) of the following Narrowband LPF will be zero, and therefore,
the input to the VCO is zero. This zero signal tells the VCO to just stay at its current
frequency and phase because they are the correct values. Assuming that we started
with i = o but eventually the carrier of the received signal started to speed up (or
slow down) a little so (i o) became a small positive (or negative) value, the signal
z(t) will become the positive (or negative) square of the message signal. This signal
contains a positive (or negative) DC value since it is the square of the message signal.
Therefore, using a narrowband LPF that is designed to extract only the DC value of
this signal will produce a positive (or negative) number that is proportional to (i
o). So, the signal w(t) will be
w(t )

D
sin 2( i o ) ,
8

Where D is the DC value of m2(t). Assuming the difference (i o) is small, the value
of w(t) will be a small positive number when (i > o) and a small negative number
when (i < o). This signal will either force the VCO to speed up to catch up with the
advancing phase of the received signal or slow down to allow the received signal to
catch up with it.
Note that the output of this system according to your book is the signal y1(t). This is in
fact the demodulated output signal. However, if you only needed to extract the carrier
signal from the received modulated signal, then the signal you are interested in is r(t).

Carrier Acquisition in SSB

The above techniques do not work for SSB (try using the signal squaring method for a
SSB signal). Therefore, the process of obtaining the carrier of the transmitted signal at
the receiver when using SSB requires some help from the transmitter by transmitting
a carrier with a very small amount of power compared to the message signal. The
purpose of this carrier is different from the carrier in Full AM. The amplitude of the
carrier in Full AM is to simplify the demodulator by using an envelope detector. This
carrier in fact contains most of the power of the transmitted signal. Transmitting the
lowpower carrier in SSB allows the receiver to know the frequency and phase of the
carrier used in the modulation. This is done by using a narrowband BPF at the
frequency of the carrier to extract that lowpower signal and then amplifying it. The
power of that carrier signal is just a small fraction of the power of the transmitted
signal (1/20 the power of the transmitted signal for example).
Superheterodyne AM Radio Receiver
Since the inception of the AM radio, it spread widely due to its ease of use and more
importantly, it low cost. The low cost of most AM radios sold in the market is due to
the use of the full amplitude modulation, which is extremely inefficient in terms of
power as we have seen previously. The use of full AM permits the use of the simple
and cheap envelope detector in the AM radio demodulator. In fact, the AM
T.R.K.NAIDU

Page 52

demodulator available in the market is slightly more complicated than a simple

envelope detector. The block diagram below shows the construction of a typical AM
receiver and the plots below show the signals in frequencydomain at the different
parts of the radio.

Description of the AM Superheterodyne Radio Receiver

Signal a(t) at the output of the Antenna: The antenna of the AM radio receiver
receives the whole band of interest. So it receives signals ranging in frequency from
around 530 kHz to 1650 kHz as shown by a(t) in the figure. Each channel in this band
occupies around 10 kHz of bandwidth and the different channels have center
frequencies of 540, 550, 560, . , 1640 kHz.
Signal b(t) at the output of the RF (Radio Frequency) Stage: The signal at the
output of the antenna is extremely week in terms of amplitude. The radio cannot
process this signal as it is, so it must be amplified. The amplification does not amplify
the whole spectrum of the AM band and it does not amplify a single channel, but a
range of channels is amplified around the desired channel that we would like to
receive. The reason for using a BPF in this stage although the desired channel is not
completely separated from adjacent channels is to avoid possible interference of some
channels later in the demodulation process if the whole band was allowed to pass
(assume the absence of this BPF and try demodulating the two channels at the two
edges of the AM band, you will see that one of these cannot be demodulated). Also,
the reason for not extracting the desired channel alone is that extracting only that
channel represents a big challenge since the filter that would have to extract it must
have a constant bandwidth of 10 kHz and a center frequency in the range of 530 kHz
to 1650 kHz. Such a filter is extremely difficult to design since it has a high Qfactor
(center frequency/bandwidth) let alone the fact that its center frequency is variable.
Therefore, the process of extracting only one channel is left for the following stages
where a filter with constant center frequency may be used. Note in the block diagram
above that the center frequency of the BPF in the RF stage is controlled by a variable
capacitor with a value that is modified using a knob in the radio (the tuning knob).
Signal c(t) at the output of the Local Oscillator: This is simply a sinusoid with a
variable frequency that is a function of the carrier frequency of the desired channel.
The purpose of multiplying the signal b(t) by this sinusoid is to shift the center
frequency of b(t) to a constant frequency that is called IF (intermediate frequency).
Therefore, assuming that the desired channel (the channel you would like to listen to)
has a frequency of fRF and the IF frequency that we would like to move that channel to
T.R.K.NAIDU

Page 53

is fIF, one choice for the frequency of the local oscillator is to be fRF + fIF. The
frequency of the local oscillator is modified in the radio using a variable capacitor that
is also controlled using the same tuning knob as the variable capacitor that controls
the center frequency of the BPF filter in the RF stage. The process of controlling the
values of two elements such as two variable capacitors using the same knob by
placing them on the same shaft is known as GANGING.
Signal d(t) at the output of the Multiplier (Usually called frequency converter or
mixer): The signal here should contain the desired channel at the constant frequency
fIF regardless of the original frequency of the desired channel. Remember that this
signal does not only contain the desired channel but it contains also several adjacent
channels and also contain images of these channels at the much higher frequency 2fRF
+ fIF (since multiplying by a cosine shifts the frequency of the signal to the left and to
the right). When this type of radios was first invented, a standard was set for the value
for the IF frequency to be 455 kHz. There is nothing special about this value. A range
of other values can be used.
Signal e(t) at the output of the IF Stage: Now that the desired channel is located at
the IF frequency, a relatively simple to create BP filter with BW of 10 kHz and center
frequency of fIF can be used to extract only the desired channel and reject all adjacent
channels. This filter has a constant Q factor of about 455/10 = 45.5 (which is not that
difficult to create), but more importantly has a constant center frequency. Therefore
the output of this stage is the desired channel alone located at the IF frequency. This
stage also contains a filter that amplifies the signal to a level that is sufficient for an
envelope detector to operate on.
Signal f(t) at the output of the Envelope Detector: The signal above is input to an
envelope detector that extracts the original unmodulated signal from the modulated
signal and also rejects any DC that is present in that signal. The output of that stage
becomes the original signal with relatively low power.
Signal g(t) at the output of the Audio Stage (Power Amplifier): Since the output of
the envelope detector is generally weak and is not sufficient to drive a large speaker,
the use of an amplifier that increases the power in the signal is necessary. Therefore,
the output of that stage is the original audio signal with relatively high power that can
directly be input to a speaker.

T.R.K.NAIDU

Page 54

T.R.K.NAIDU

Page 55

T.R.K.NAIDU

Page 56

In the previous chapter, we studied the different AM technique in which the amplitude
of some carrier signal is modified according to the message signal. The frequency and
phase of the carrier of the carrier signal in all AM modulation techniques were
constant. In this chapter, we will study a different method for transmitting information
by changing the phase or frequency (changing the angle) of the carrier signal and
keeping its amplitude constant.

Instantaneous Frequency
The frequency of a cosine function x(t) that is given by
x(t ) cos c t 0
is equal to c since it is a constant with respect to t, and the phase of the cosine is the
constant 0. The angle of the cosine (t) = ct +0 is a linear relationship with
respect to t (a straight line with slope of c and yintercept of 0). However, for other
sinusoidal functions, the frequency may itself be a function of time, and therefore, we
should not think in terms of the constant frequency of the sinusoid but in terms of the
INSTANTANEOUS frequency of the sinusoid since it is not constant for all t.
Consider for example the following sinusoid
y (t ) cos (t ),

where (t) is a function of time. The frequency of y(t) in this case depends on the
function of (t) and may itself be a function of time. The instantaneous frequency of
y(t) given above is defined as
d (t )
i (t )
.
dt
As a checkup for this definition, we know that the instantaneous frequency of x(t) is
equal to its frequency at all times (since the instantaneous frequency for that function
is constant) and is equal to c. Clearly this satisfies the definition of the instantaneous
frequency since (t) = ct +0 and therefore i(t) = c.
If we know the instantaneous frequency of some sinusoid from to some time t, we
can find the angle of that sinusoid at time t using
(t )

( )d .
i

Changing the angle (t) of some sinusoid is the bases for the two types of angle
modulation: Phase and Frequency modulation techniques.

Phase Modulation (PM)

In this type of modulation, the phase of the carrier signal is directly changed by the
message signal. The phase modulated signal will have the form
g PM (t ) A cos c t k p m (t ) ,

where A is a constant, c is the carrier frequency, m(t) is the message signal, and kp
is a parameter that specifies how much change in the angle occurs for every unit of
change of m(t). The phase and instantaneous frequency of this signal are

T.R.K.NAIDU

Page 57

 PM (t ) c t k p m (t ),
dm (t )
c k p m&(t ).
dt
So, the frequency of a PM signal is proportional to the derivative of the message
signal.

i (t ) c k p

Frequency Modulation (FM)

This type of modulation changes the frequency of the carrier (not the phase as in PM)
directly with the message signal. The FM modulated signal is
t

g FM (t ) A cos c t k f m ( )d ,

where kf is a parameter that specifies how much change in the frequency occurs for
every unit change of m(t). The phase and instantaneous frequency of this FM are

FM (t ) c t k f

m ( )d ,

i (t ) c k f

dt

m ( )d

c k f m (t ).

Relation between PM and FM

PM and FM are tightly related to each other. We see from the phase and frequency
relations for PM and FM given above that replacing m(t) in the PM signal with
t
dm (t )
m ( )d gives an FM signal and replacing m(t) in the FM signal with dt
gives a PM signal. This is illustrated in the following block diagrams.

T.R.K.NAIDU

Page 58

Continuity of the Phase of PM and FM signals

Message signals that are phase or frequency modulated can be classified into
a)
continuous signals WITH NO delta functions,
b)
discontinuous signals WITH NO delta functions,
c)
continuous signals WITH delta functions.
c)
discontinuous signals WITH delta functions.
For the above signals, we can summarize the continuity or discontinuity of the phase
of PM and FM signals in the following table.
Signal

PM (t ) c t k p m (t )

gPM(t)

FM (t ) c t k f

m ( )d

gFM(t)

a)
b)

Continuous
Discontinuous

c)

Cont. with delta

functions

d)

Discont. with delta

functions

Cont.
Discont.
Cont. with
random phase
at deltas
Discont. with
random phase
at deltas

Continuous
Continuous

Cont.
Cont.

Discontinuous WITH NO
delta functions

Discont.

Discontinuous WITH NO
delta functions

Discont.

Example 1: Sketch the FM signal that results when modulating the message signal
m(t) shown below with kf = 2(2) and c = 2 (10) rad/s.

T.R.K.NAIDU

Page 59

Example 2: Sketch the PM signal that results when modulating the message signal
m(t) shown below with kp = 2 and c = 2 (14) rad/s.
To sketch the PM signal, we can compute dm(t)/dt and sketch the frequency
modulated signal when dm(t)/dt is input to an FM block similar to Example 1.

Bandwidth of FM and PM Signals

T.R.K.NAIDU

Page 60

The bandwidth of the different AM modulation techniques ranges from the bandwidth
of the message signal (for SSB) to twice the bandwidth of the message signal (for
DSBSC and Full AM). When FM signals were first proposed, it was thought that their
bandwidth can be reduced to an arbitrarily small value. Compared to the bandwidth of
different AM modulation techniques, this would in theory be a big advantage. It was
assumed that a signal with an instantaneous frequency that changes over of range of
f Hz would have a bandwidth of f Hz. When experiments were done, it was
discovered that this was not the case. It was discovered that the bandwidth of FM
signals for a specific message signal was at least equal to the bandwidth of the
corresponding AM signal. In fact, FM signals can be classified into two types:
Narrowband and Wideband FM signals depending on the bandwidth of each of these
signals

Narrowband FM and PM
The general form of an FM signal that results when modulating a signals m(t) is

g FM (t ) A cos c t k f

m
(

)
d

A narrow band FM or PM signal satisfies the condition

k f a (t ) = 1
For FM and
k p m (t ) = 1
For PM, where
t

a (t )

m ( )d ,

such that a change in the message signal does not results in a lot of change in the
instantaneous frequency of the FM signal.
Now, we can write the above as
g FM (t ) A cos c t k f a (t ) .
Starting with FM, to evaluate the bandwidth of this signal, we need to expand it using
a power series expansion. So, we will define a slightly different signal
j t k a (t )
g FM (t ) A e c f A e j c t ejk f a (t ) .

Remember that
j t k a (t )
g FM (t ) A e c f A cos c t k f a (t ) jA sin
c t

T.R.K.NAIDU

kf a (t ) ,
Page 61

so
g FM (t ) Re g FM (t ) .
Now we can expand the term e jk f a (t ) in g FM (t ) , which gives

g FM (t ) A e

j c t

j 2 k f2a 2 (t ) j 3k f3a 3 (t ) j 4 k f4a 4 (t )

1 jk f a (t )

K
2!
3!
4!

j t

k f2a 2 (t ) j ct jk f3a 3 (t ) j ct k f4a 4 (t ) j ct

j c t
c
A e
jk f a (t )e

e
K
2!
3!
4!

Since kf and a(t) are real (a(t) is real because it is the integral of a real function m(t)),
and since
Re{ej t} = cos(ct)
and
Re{ jej t} = sin(ct),
then
c

g FM (t ) Re g FM (t )

k f2a 2 (t )
k f3a 3 (t )
k f4a 4 (t )
A cos(c t ) k f a (t )sin(c t )
cos(c t )
sin(c t )
cos(c t ) K
2!
3!
4!

The assumption we made for narrowband FM is ( k f a (t ) = 1 ). This assumption will

result in making all the terms with powers of k f a (t ) greater than 1 to be small
compared to the first two terms. So, the following is a reasonable approximation for
g FM (t )
g FM ( Narrowband ) (t ) A cos(c t ) k f a (t ) sin(c t )

when k f a (t ) = 1

.
It must be stressed that the above approximation is only valid for narrowband FM
signals that satisfy the condition ( k f a (t ) = 1 ). The above signal is simply the
addition (or actually the subtraction) of a cosine (the carrier) with a DSBSC signal
(but using a sine as the carrier). The message signal that modulates the DSBSC signal
is not m(t) but its integration a(t). One of the properties of the Fourier transform
informs us that the bandwidth of a signal m(t) and its integration a(t) (and its
derivative too) are the same (verify this). Therefore, the bandwidth of the narrowband
FM signal is
BW FM ( Narrowband ) BW DSBSC 2 BW m (t ) .

We will see later that when the condition (kf << 1) is not satisfied, the bandwidth of
the FM signal becomes higher that twice the bandwidth of the message signal. Similar
relationships hold for PM signals. That is
g PM ( Narrowband ) (t ) A cos(c t ) k p m (t ) sin(c t )

when

k p m (t ) = 1 ,
T.R.K.NAIDU

Page 62

and
BW PM ( Narrowband ) BW DSBSC 2 BW m (t ) .

Construction of Narrowband Frequency and Phase Modulators

The above approximations for narrowband FM and PM can be easily used to
construct modulators for both types of signals.

Narrowband FM Modulator

Narrowband PM Modulator

Wideband FM and PM
For the FM signal shown below, the value of kf may not satisfy the condition
k f a (t ) = 1 , and therefore the approximation used in narrowband FM may not be
applicable. For FM signals

g FM (t ) A cos c t k f

T.R.K.NAIDU

m
(

)
d

Page 63

that do not satisfy k f a (t ) = 1 , computing the bandwidth is more difficult than for
narrowband signals. To compute the bandwidth of wideband FM, we will approximate
the signal g FM (t ) by another signal that results from modulating a sampled version
of the message signal. That is, instead of using the signal m(t) as the message signal
for modulating the carrier, we will use an approximation m (t ) that is obtained by
sampling m(t) as shown in the following figure.

The value of t that we will use is the maximum allowable value that will insure that
the m(t) can be reconstructed from m (t ) without loss of information. This maximum
value for t is obtained using the Nyquist sampling theorem, which states that for a
signal m(t) with a bandwidth of Bm (Hz), the minimum sampling frequency is 2Bm.
Therefore, the maximum for t is given by t = 1/2Bm. Of course a smaller value for
t will be better since it gives a better approximation for m(t) but is unnecessary.
So, to find the approximate bandwidth of FM signal, let us assume that the original
message signal m(t) is bounded in amplitude by the two values mp and mp. Therefore,
m p m (t ) m p .
Then, we can straight forward say that
m p m (t ) m p .
Now, the approximate signal g FM (t ) to g FM (t ) is obtained by FM modulating m (t )
as

g FM (t ) A cos c t k f

m ( )d

Since m (t ) is constant over periods of t = 1/2Bm, the instantaneous frequency of

g FM (t ) will be constant over periods of t = 1/2Bm. The signal g FM (t ) will look like
as follows.

T.R.K.NAIDU

Page 64

Since m p m (t ) m p , the instantaneous frequency of g FM (t ) (and also g FM (t ) )

will be in the range

c k f m p i (t ) c k f m p .
This means that the instantaneous frequency changes over a range of = kf mp (this
can also be written as f = kf mp/2) on each side around the carrier frequency c.
So, to approximate the bandwidth of the of original FM signal g FM (t ) , we will
compute the approximate bandwidth of the approximation signal g FM (t ) by finding
its frequency spectrum. Since g FM (t ) is composed of blocks of sinusoids with
different frequencies that are in the range of frequencies of
c k f m p i (t ) c k f m p , we can find the spectrum of each of these blocks
independently and then add these spectrums to get the overall spectrum of g FM (t ) .
Consider the part of g FM (t ) that shown below

T.R.K.NAIDU

Page 65

The signal z(t) is given by

z (t ) A rect 2B m t t 0 cos i t ,
where c k f m p i (t ) c k f m p . The Fourier transform of z(t) is
Z ( )

i

i
A
j i t 0
j i t 0
sinc
sinc

e
e
4B m
4B m
4B m

Remembering that the sinc function looks like the following

Sketching the magnitude of Z() (i.e., magnitude spectrum of z(t)) will give the
following (since the complex exponentials have a magnitude of one and the two
components shifted to the left and right almost do not interfere with each other)

Adding the spectrum of the different signals like z(t) given above will give us the
spectrum of the approximation signal g FM (t )

If we assume that the sidebands (the small humps at the two edges of a sinc function)
have negligible power, and knowing that = kf mp, we see that the bandwidth of an
FM signal is approximately equal to
T.R.K.NAIDU

Page 66

BW FM 2k f m p 8 B m
2 8 B m

(rad/s)
(rad/s)

Using the fact that f = kf mp/2 , the bandwidth in Hz becomes

BW FM

2k f m p

4B m (Hz)
2
2 f 4B m (Hz)
2 f 2B m

(Hz)

In practice, this bandwidth is higher than the actual bandwidth of FM signals.

Consider for example narrowband FM. Using this formula for the bandwidth, we see
that the bandwidth is twice the actual bandwidth. In fact, a more accurate relationship
is known as CARSONs Rule, which is given by
(Hz) 2 f B m

BW FM 2 f 2B m
2 2 B m

(Hz)

(rad/s)

where
Bm = Bandwidth of the Message Signal m(t) in Hz,
and
= kf mp

f = kf mp/2.

Verification of Carsons Rule (for computing the FM Bandwidth)

The Casrons rule says that the BW of any FM signal is given by
BW FM 2 f 2B m
2 2 B m

(Hz) 2 f B m

(Hz)

(rad/s)

where
Bm = Bandwidth of the Message Signal m(t) in Hz,
and

f = /2 = kf mp/2 .

We can verify the Carsons rule using a simple message signal that is a sinusoid with
frequency m. That is
m (t ) cos(m t ) ,
where the magnitude of the message signal is a constant. The bandwidth of this
signal Bm is clearly m. The signal a(t), which is the integration of m(t), becomes

T.R.K.NAIDU

Page 67

a (t )

m ( )d

sin(m t ) .

So, the FM signal becomes

k
g FM (t ) A cos c t f sin(m t ) .
m

Let us define a function g FM (t ) that is the complexexponential form of g FM (t ) , or

g FM (t ) A e

k
j c t f sin m t
m

A e

j c t

kf
sin m t
m

(1)

The second exponential above is a periodic signal with period Tm=2/m. So, it can be
expanded using the exponential Fourier series as
e

kf
sin m t
m

De
n

jn m t

(2)

where according the exponential Fourier series expansion (page 53 in your text book)
1
Dn
Tm

f (t ) e

jnm t

Tm

dt m
2

kf
sin m t
m

e jnm t dt .

This integration can be simplified if we use the substitution x m t

x m t

dx m dt

t coveres the range
,
m m
Let us define to be

1
dx
m

dt

x coveres the range ,

k f

. Substituting for and for x m t in the

m
m

integration of Dn, we get

Dn

1
2

j sin x

e jnx dx .

Verification of Carsons Rule (Continued)

T.R.K.NAIDU

Page 68

kf

We have seen that the complex Fourier series coefficients of the term e j m
be written as
Dn

1
2

j sin x

sin m t

can

e jnx dx .

Unfortunately, the above integration does not have a closed form (cannot be written in
terms of basic functions). However, this function can be integrated numerically. The
integration above is called the Bessel function of the first kind with order n that is
evaluated at . This is abbreviated by Jn(). So,

1
J n ( ) D n
2

j sin x

e jnx dx .

One important property of this Bessel function is that

J n ( ) 1 J n ( ) .
n

So, J n ( ) J n ( ) .
Substituting Jn() for Dn in Equation (2) above and using the result in Equation (1)
gives
g FM (t ) A e j c t
A

( ) e

( ) e jnm t
j c nm t

The original function g FM (t ) is related to g FM (t ) by

g FM (t ) Re g FM (t ) ,
So,
g FM (t ) A

( ) cos c n m t ,

where

k f

m
m

is know as the MODULATION INDEX of the FM signal.

Conclusion:
The spectrum of an FM signal that results from modulating a sinusoidal
message is an infinite sum of sinusoids with frequencies c, cm, c2m,
T.R.K.NAIDU

Page 69

c3m, c4m . The amplitude of these sinusoids is the values of a

Bessel function of the first kind that is evaluated at .
Plotting the Bessel function of the first kind Jn() for different orders n and different
values of is shown below.

Looking at the above plots, we can fill in the following table

Jn()
n=0
n=1
n=2
n=3
n=4
n=5
n=6
n=7
n=8
n=9
n = 10

=1
0.7652
0.4401
0.1149
0.0196
0.0025
0.0002
0.0000
0.0000
0.0000
0.0000
0.0000

=2
0.2239
0.5767
0.3528
0.1289
0.0340
0.0070
0.0012
0.0002
0.0000
0.0000
0.0000

=3
0.2601
0.3391
0.4861
0.3091
0.1320
0.0430
0.0114
0.0025
0.0005
0.0001
0.0000

=4
0.3971
0.0660
0.3641
0.4302
0.2811
0.1321
0.0491
0.0152
0.0040
0.0009
0.0002

=5
0.1776
0.3276
0.0466
0.3648
0.3912
0.2611
0.1310
0.0534
0.0184
0.0055
0.0015

=6
0.1506
0.2767
0.2429
0.1148
0.3576
0.3621
0.2458
0.1296
0.0565
0.0212
0.0070

The power of the FM signal is A2 since it has a magnitude of A. Looking carefully at

the table, we notice that the values of the elements below the thick line are all below
(0.1). Therefore, the magnitude of each component that corresponds to one of these
T.R.K.NAIDU

Page 70

values is less than (0.1)A. So, the power of these components is [(0.1)A]2 = (0.01)A2
W. Since these components are on both sides around the carrier frequency, we see that
the total power contained in all of these components is less than 0.02 W. This means
that the power contained in the components above the thick line is at least 98% of the
power of the FM signal. Therefore, we can safely ignore the components below that
line.
It is seen from the table that for each value of , only the components with |n| +1
are significant and the rest are small. Eliminating all the components in Equation (3)
with negligible amplitudes, the bandwidth of the FM signal becomes
BW FM 2m 1 2 m m
2 2 B m

(rad/s)

This verifies that the BW of an FM signal is given by the Carsons rule given above.

Effect of NonLinearity on AM and FM signals

Sometimes, the modulated signal after transmission gets distorted due to non
linearities in the channel, for example. In general, when the transmitted modulated
signal is affected by nonlinearity in the channel, the demodulated signal becomes a
distorted version of the message signal. We can easily show that the effect non
linearities on different types of amplitudemodulated signals is devastating, while
frequencymodulated signals are immune to nonlinearities. In fact, the effect of non
linearities on FM signals can be used for generating wideband FM signals from
narrowband FM signals, which is an important feature.
NonLinearity in AM: Consider the channel shown below with a DSBSC input
signal. The channel is a nonlinear channel in which the output signal of the channel
is the sum of the input signal and other powers of the input signal.

Let gDSBSC(t) be
g DSBSC (t ) m (t ) cos c t .
The output signal of the channel f(t) becomes

T.R.K.NAIDU

Page 71

f (t ) a1m (t ) cos ct a2 m (t ) cos c t

a 3 m (t ) cos
c t
a1m (t ) cos ct a2 m 2 (t ) cos 2 c t a3m 3 (t ) cos
3 c t
2

a3 m 3 (t )
a2 m 2 (t )
1 cos 2c t
a1m (t ) cos ct
cos c t 1 cos 2c t
2
2
a m 3 (t )
a3 m 3 (t )
a m 2 (t )
a2 m 2 (t )
2
a1m (t ) 3

c
os

cos
2

cos c t cos 3c t
c
c

2
2
2
4

3a m 3 (t )
a3m 3 (t )
a2 m 2 (t )
a2 m 2 (t )
a1m (t ) 3

cos

cos
2

cos 3c t

c
c
4
2 44 2 4 4 43 1 44 44 2 4 4 43
14 22 43 1 4 4 4 4 4 2
1
4
4 4 4 4 43
Around 0
Around 2
Around 3
Around c

So, unless a3 is zero, it is clear that the original modulated signal gDSBSC(t) given
above cannot be extracted from the received signal f(t) because the terms with
frequency around c does not contain only m(t) but also m3(t). So, DSBSC (and AM
modulation in general) is vulnerable to nonlinearities.
NonLinearity in FM: Again, consider the same channel shown given above with an
FM input signal.

g FM (t ) A cos c t k f

Let gFM(t) be

m
(

)
d

The output signal of the channel q(t) becomes

q (t ) a1A cos c t k f m ( )d a2 A cos c t k f

a3 A cos c t k f m ( )d

m
(

)
d

a2 A
a1A cos c t k f m ( )d
1 cos 2c t 2k f m ( )d
2

aA
3 1 cos 2c t 2k f m ( )d cos c t k f m ( )d
2

3a A
aA
aA
2 a1A 3 cos c t k f m ( )d 2 cos 2c t 2k f m ( )d
4
2

{2
1 4 4 4 4 4 44 2 4 4 44 4 4 43
1 4 4 4 4 44 2 4 4 4 4 4 43
DC

Around c with k f k f

Around 2c with k f 2 k f

a3 A
cos 3c t 3k f m ( )d
4

1 4 4 4 4 44 2 4 4
4 4 4 43
Around 3c with k f 3 k f

T.R.K.NAIDU

Page 72

In this case, it is clear that all the coefficients of the different terms are simply
constants. The terms are in fact different FM signals with different frequencies and
different values of the parameter kf. But, the important conclusion is that the original
FM signal can easily be extracted from the output signal of the nonlinear channel
using a filter centered at the carrier frequency and bandwidth equal to the bandwidth
of the FM signal. So, FM signals are IMMUNE to (do not get damaged by) non
linearities in the channel or in the components.

Use of Nonlinearity to Manipulate FM Signals

The fact that passing an FM signal through a nonlinear device results in a set of FM
signals with different carrier frequencies and different parameter kf (and therefore
different frequency variation parameter ), we can use this process for manipulating
the carrier frequency c and/or the frequency variation . In general, passing the
FM signal through a nonlinear device with a maximum nonlinearity power of P
will give P different FM signals as shown in the block diagram below.

Passing the output of the Ppower nonlinear device through a BPF with center
frequency Pc and bandwidth equal to the bandwidth of the last FM signal will
extract that signal and reject the other FM signal. This process can be used to obtain
wideband FM signals from narrowband FM signals as will be described through a set
of examples next. We will assume that the nonlinear devices that we use in the
following examples have built-in BPFs to eliminate the undesired FM components.

Generation of Wideband FM Signals

Indirect Method for Wideband FM Generation:
Consider the following block diagram

T.R.K.NAIDU

Page 73

A narrowband FM signal can be generated easily using the block diagram of the
narrowband FM modulator that was described in a previous lecture. The narrowband
FM modulator generates a narrowband FM signal using simple components such as
an integrator (an OpAmp), oscillators, multipliers, and adders. The generated
narrowband FM signal can be converted to a wideband FM signal by simply passing it
through a nonlinear device with power P. Both the carrier frequency and the
frequency deviation f of the narrowband signal are increased by a factor P.
Sometimes, the desired increase in the carrier frequency and the desired increase in f
are different. In this case, we increase f to the desired value and use a frequency
shifter (multiplication by a sinusoid followed by a BPF) to change the carrier
frequency to the desired value.
Example 1: A narrowband FM modulator is modulating a message signal m(t) with
bandwidth 5 kHz and is producing an FM signal with the following specifications
fc1 = 300 kHz,
f1 = 35 Hz.
We would like to use this signal to generate a wideband FM signal with the following
specifications
fc2 = 135 MHz,
f2 = 77 kHz.
Show the block diagram of several systems that will perform this function and specify
the characteristics of each system
Solution: We see that the ratio of the carrier frequencies is
f c 2 135*106

450 ,
f c 1 300*103
and the ratio of the frequency variations is
f 2 77 *103

2200 .
f 1
35
Therefore, we should feed the narrowband FM signal into a single (or multiple) non
linear device with a nonlinearity order of f2/f1 = 2200. If we do this, the carrier
frequency of narrowband FM signal will also increase by a factor of 2200, which is
higher than what is required. This can easily be corrected by frequency shifting. If we
feed the narrowband FM signal into a nonlinear device of order fc2/fc1, we will get
the correct carrier frequency but the wrong value for f. There is not a way of
correcting the value of f for this signal without affecting the carrier frequency.

T.R.K.NAIDU

Page 74

System 1:

In this system, we are using a single nonlinear device with an order of 2200
or multiple devices with a combined order of 2200. It is clear that the output of
the nonlinear device has the correct f but an incorrect carrier frequency
which is corrected using a the frequency shifter with an oscillator that has a
frequency equal to the difference between the frequency of its input signal and
the desired carrier frequency. We could also have used an oscillator with a
frequency that is the sum of the frequencies of the input signal and the desired
carrier frequency. This system is characterized by having a frequency shifter
with an oscillator frequency that is relatively large.
System 2:

In this system, we are using two nonlinear devices (or two sets of nonlinear
devices) with orders 44 and 50 (44*50 = 2200). There are other possibilities
for the factorizing 2200 such as 2*1100, 4*550, 8*275, 10*220, .
Depending on the available components, one of these factorizations may be
better than the others. In fact, in this case, we could have used the same
factorization but put 50 first followed by 44. We want the output signal of the
overall system to be as shown in the block diagram above, so we have to
insure that the input to the nonlinear device with order 50 has the correct
carrier frequency such that its output has a carrier frequency of 135 MHz. This
is done by dividing the desired output carrier frequency by the nonlinearity
order of 50, which gives 2.7 Mhz. This allows us to figure out the frequency of
the required oscillator which will be in this case either 13.22.7 = 10.5 MHz
or 13.2+2.7 = 15.9 MHz. We are generally free to choose which ever we like
unless the available components dictate the use of one of them and not the
other. Comparing this system with System 1 shows that the frequency of the
oscillator that is required here is significantly lower (10.5 MHz compared to
525 MHz), which is generally an advantage.

T.R.K.NAIDU

Page 75

System 3:

We also can bring the frequency shifter before all the nonlinear devices and
therefore reduce the frequency of the required oscillator to the minimum value
by finding the required carrier frequency at the input of each nonlinear device
to insure that the carrier frequency of the final output of nonlinear devices is
the desired final carrier frequency.
Direct Method of Generating WB FM Signals
This method is simple in the sense that it uses a single component: the voltagecontrolled oscillator (VCO). As described in the section of Carrier Acquisition for
DSBSC systems, VCOs are devices that produce a sinusoid with a frequency that is
proportional to the input signal. So, if the input signal to a VCO is the message signal,
the output of the VCO will be an FM modulated signal of the message signal since the
frequency of this FM signal changes according to the input message signal.

Demodulation of FM (and PM) Signals

There exist many methods for demodulation of FM signals. One of these methods is
phaselocked loops (PLL) that were studied in a previous lecture. PLLs when fed
with an FM signal directly produce an output signal that is proportional to the
message signal. Here we will discuss two other methods that are directly related to
each other for FM and PM demodulation.
Signal Differentiation Method
An FM signal has the following form

g FM (t ) A cos c t k f

m ( )d

So, its magnitude is constant with value of A. The information is contained in the
frequency (or angle) of the FM signal. To extract the message signal contained in an
FM signal, we can transfer the information from the angle to the magnitude by simply
differentiating the FM signal. Since the derivate of a sinusoid results in multiplying
T.R.K.NAIDU

Page 76

the magnitude of the sinusoid by the derivate of its angle, the derivative of the above
FM signal becomes

dg FM (t )
A c k f m (t ) sin c t k f
dt

m ( )d .

So, the message signal of the above derivative is contained in the frequency of the
sinusoid and also in its magnitude. Passing the derivative of the FM signal through an
envelope detector will give the desired message signal at the output. Therefore, the
following block diagram is an FM demodulator.

The same idea can be used for PM demodulation. A PM signal has the form
g PM (t ) A cos c t k p m (t ) .
So, if we differentiate it, we get
dg PM (t )
dm (t )

A c k f
sin c t k p m (t ) .
dt
dt

If this signal is passed through an envelope detector, the output will be proportional to
the derivative of the message signal. Passing this signal through an integrator will
give us what we want. Therefore, the block diagram of a PM demodulator will be as
follows.

Frequency Discrimination Method

T.R.K.NAIDU

Page 77

The same concept as described above in the signal differentiation method for
demodulation FM and PM signals can be used but after replacing the derivative block
at the beginning with a filter. Assume that we would like to demodulate the following
FM signal which has a carrier frequency of c and a frequency band from 1 to 2.

We can use the following BPF which is centered not at c but at a higher frequency
such that the range of frequency [1 , 2] falls in the transition band of the filter (in
the region where the filter changes from not passing to passing the input signal). If the
transition band of the filter has a linear response (a line with some nonzero slope),
the different parts of the FM signal, which is input to the filter, will be amplified (or
attenuated) by different factors depending on the frequency of the these parts. The
higher the frequency, the higher the amplitude of the output signal of the filter, and the
lower the frequency, the lower the amplitude of the output signal. This process is very
similar to what the differentiator in the signal differentiation FM demodulation
method does.

T.R.K.NAIDU

Page 78

Therefore, passing the signal s(t) that is outputted by the BPF into an envelope
detector gives a signal that is proportional to the message signal.
Therefore the following block diagram is the frequency discrimination FM
demodulator.

T.R.K.NAIDU

Page 79