Angle M od u lation

In Chapter 3, we considered amplitude modulation (AM) of the carrier as a means for

transmitting the message signal. Amplitude-modulation methods are also called linear­
modulation methods, although conventional AM is not linear in the strict sense.
Another class of modulation methods includes frequency modulation (FM) and phase
modulation (PM), which are described in this chapter. In FM
systems, the frequency of the
carrier fc
is changed by the message signal; in PM systems, the phase of the carrier is
changed according to the variations in the message signal. Both FM and PM are nonlin­
ear, and often they are jointly called angle-modulation methods.
In the following sections,
we will show that, angle modulation, due to its inherent nonlinearity, is more complex
to implement and much more difficult to analyze. In mariy cases, only an approximate
analysis can be done. Another property of angle modulation is its bandwidth-expansion
property. The FM and PM systems generally expand the bandwidth such that the effective
bandwidth of the modulated signal is 1,1sually many times the bandwidth of the message
1
signal. With a higher implementation complexity and a higher bandwidth occupancy, we
would naturally question the usefulness of these systems. As our analysis in Chapter 6
will show, the major benefit of these systems is their high degree of noise immunity. In
fact, these systems sacrifice bandwidth for high-noise immunity. That is the reason that
FM systems are widely used in high-fidelity music broadcasting and point-to-point com­
munication systems, where the transmitter power is quite limited. Another advantage of
angle-modulated signals is their constant envelope, which is beneficial when the signal is
amplified by nonlinear amplifiers.

4.1 REPRESENTATION OF FM AND PM SIGNALS

An angle-modulated signal generally can be written as

u(t) = Ac cos(2n fct + ¢(t)), (4.1.1)

1 Strictly speaking, the bandwidth of the modulated signal is infinite. That is why we talk about the effective
bandwidth.

1 61
162 Angle Modulation Chapter 4

le
where denotes the carrier frequency and ¢(t) denotes a time-varying phase. The instan­
taneous frequency of this signal is given by

f; (t) = le + 2n1 dt
d
</>(t). (4.1.2)
Ifm(t) is the message signal, then in a PM system, the phase is proportional to the mes­
sage, i.e.,

¢(t) = kpm(t), (4.1.3)

and in an FM system, the instantaneous frequency deviation from the carrier frequency is
proportional to the message signal, i.e.,

l;(t) - le = k1m(t) = 2Jt1 dtd </>(t), (4.1.4)

where kp and kt are phase and frequency deviation constants. From the preceding rela­

{
tionships, we have

t
kpm(t), PM
2nkt J_00 m(r)dr,
¢(t) = (4.1.5)
FM

The foregoing expression shows a close and interesting relationship between FM and PM
systems. This close relationship allows us to analyze these systems in parallel and only
emphasize their main differences. First, note that if we phase modulate the carrier with the
integral of a message, it is equivalent to the frequency modulation of the carrier with the

{
original message. On the other hand, this relation can be expressed as

d kp !:_
dt m(t), PM
-¢(t)
dt = (4.1.6)
2nk1m(t), FM

which shows that if we frequency modulate the carrier with the derivative of a message, the
result is equivalent to the phase modulation of the carrier with the message itself. Figure 4.1
shows the above relation between FM and PM. Figure 4.2 illustrates a square-wave signal
and its integral, a sawtooth signal, and their corresponding FM and PM signals.
The demodulation of an FM signal involves finding the instantaneous frequency of
the modulated signal and then subtracting the carrier frequency from it. In the demodula­
tion of PM, the demodulation process is done by finding the phase of the signal and then
recovering m(t). The maximum phase deviation in a PM system is given by

!:!.. ¢max = kp max[ l m(t)I], (4.1.7)

and the maximum frequency deviation in an FM system is given by

Illmax = kt max[ l m(t)I]. (4.1.8)

Section 4.1 Representation of FM and PM Signals 1 63

FM modulator

�.�
�
Ill
Integrator PM
modulator

PM modulator

�IPMI____._
�
Ill
FM
Differentiator Figure 4.1 A comparison of frequency and
modulator
phase modulators.

m(t)
1 r-------.
0 2 4
-1 -------

Figure 4.2 Frequency and phase modulations of square and sawtooth waves.
1 64 Angle Modulation Chapter 4

Example 4.1.1
The message signal

Ac c
m (t) = a cos(2rr:fmt)
is used to either frequency modulate or phase modulate the carrier cos(2nf t ). Find the
modulated signal in each case.

Solution In PM, we have

</> (t) = kµm(t) = kpa cos(2rtfmt), (4. 1 .9)

and in FM we have
,

11
k a
-oo
</> (t) = 2rr:k1 m (r) dr = _J_ sin(2rtfmt). (4. 1 . 10)
fm

{AcAc cc
Therefore, the modulated signals will be

(n
cos (2nf t + kµa cos(2rr:fmt) ) ,
) FM
PM
u(t) = k1a •
(4. 1 . 1 1)
cos 2 f t + Im sm(2rr:J,nt) ,

By defining

(4. 1 . 12)

and
(4. 1.13)

we have

u (t) =
{ AcAc cos (2 f t + {3p cos(2rtfmt) ) ,
nc
cos (2nfc t + f31 sin(2rtfmt)) ,
PM
FM · (4. 1 . 14)

The parameters {3p and f31 are called the modulation indices of the PM and FM systems,
respectively. •

We can extend the definition of the modulation index for a general nonsinusoidal
signal m (t) as

/3p = kp max[l m (t) I]; (4.1.15)

/31 = kt max[lm(t) I] (4.1.16)
w . '

where W denotes the bandwidth of the message signal m (t). In terms of the maximum
phase and frequency deviation !l</Jmax and D.fmax we have
'
(4.1.17)
(4.1.18)
�ection 4. 1 Representation of FM and PM Signals 1 65

Narrowband Angle Modulation2• Consider an angle-modulation system in

which the deviation constants and kp kt
and the message signal m (t) are such that for
all t, we have </> (t) « 1 . Then we can use a simple approximation to expand u (t) in

=
Equation (4. 1 . 1 ) as

u (t) Ac cos 2nfct cos </> (t) - Ac sin 2nfct sin </> (t)
(4. 1 . 19)

where we have used the approximations cos </> (t) � 1 and sin </> (t) � </> (t) for </> (t) « 1 .
Equation (4. 1 . 1 9) shows that in this case, the modulated signal is very similar to a con­
ventional AM signal given in Equation (3.2.5). The only difference is that the message
signal m (t) is modulated on a sine carrier rather than a cosine carrier. The bandwidth of
this signal is similar to the bandwidth of a conventional AM signal, which is twice the
bandwidth of the message signal. Of course, this bandwidth is only an approximation of
the real bandwidth of the FM signal. Phasor diagrams for this signal and the compara­
ble conventional AM signal are given in Figure 4.3. Compared to conventional AM, the
narrowband angle-modulation scheme has far less amplitude variations. Of course, the
angle-modulation system has constant amplitude and, hence, there should be no amplitude
variations in the phasor-diagram representation of the system. These slight variations are
due to the first-order approximation that we have used for the expansions of sin(</> (t)) and
cos(</> (t)). As we will see later, the narrowband angle-modulation method does not pro­
vide better noise immunity than a conventional AM system. Therefore, narrowband angle
modulation is seldom used in practice for communication purposes. However, these sys­
tems can be used as an intermediate stage for the generation of wideband angle-modulated
signals, as we will discuss in Section 4.3.

-sin ) Vu(t)

·IAC am(t) . /'

1' 1' ­

I·
cos

(a)

Figure 4.3 Phasor diagrams for the

conventional AM and narrowband angle
(b) modulation.

2Also known as low-index angle modulation.

 1 66 Angle Modulation Chapter 4

4.2 SPECTRAL CHARACTERISTICS OF ANGLE-MODULATED SIGNALS

Due to the inherent nonlinearity of angle-modulation systems, the precise characterization

of their spectral properties, even for simple message signals, is mathematically intractable.
Therefore, the derivation of the spectral characteristics of these signals usually involves
the study of simple modulating signals and certain approximations. Then the results are
generalized to the more complicated messages. We will study the spectral characteristics
of an angle-modulated signal when the modulating signal is a sinusoidal signal.

4.2.1 Angle Modulation by a Sinusoidal Signal

Consider the case where the message signal is a sinusoidal signal (to be more precise, sine

=
in PM and cosine in FM). As we have seen in Example 4.1.1,
in this case for both FM and
PM we have
u(t)
A cos c (2nf t
c + f3 sin ), 21tfmt (4.2.1)
where f3 is the modulation index that can be either /3p or [31, and in PM sin 2nfm t is
substituted by cos 2nfmt. Using Euler's relation, the modulated signal can be written as

n =
(4.2.2)
Since sin 2 fm t is periodic with period Tm Jm , the same is true for the complex expo­
nential signal

Therefore, it can be expanded in a Fourier-series representation. The Fourier-series coeffi­

cients are obtained from the integral

1 Tm ej/3 sin2rcfmte-jn2rcfmt dt
I

Cn fm
u=2:1:.fmt _!_ f 2TI ej(/3 sinu-nu) du. (4.2.3)
21t Jo

This latter expression is a well-known integral called the Bessel function of the first kind
of order n and is denoted by ln (/3) . Therefore, we have the Fourier series for the complex

L
exponential as
00
ejf3 sin2rcfm t =
ln ({3) ej2rcnfmt . (4.2.4)
-
n= oo

)
(4.2.4) into Equation (4.2.2), we obtain

(
By substituting Equation

U (t) = Re Ac
'f;,
n oo
ln ([3 ) ej2rcnfmt ej2rcfct

L Acln (/3)
00
= COS (21t(fc + nfm)t) . (4.2.s:
n=-oo
Section 4.2 Spectral Characteristics of Angle-Modulated Signals· 1 67

The preceding relation shows that, even in this very simple case where the modulating sig­
nal is a sinusoid of frequency fm,
the angle-modulated signal contains all frequencies of
the form fc + nfm n
for = 0, ± 1 , ±2, . . . . Therefore, the actual bandwidth of the modu­
lated signal is infinite. However, the amplitude of the sinusoidal components of frequencies
fc nfm
± n
for large is very small. Hence, we can define a finite for the effective bandwidth
modulated signal. For small f3, we can use the approximation

(4.2.6)

Thus, for a small modulation index {3, only the first sideband corresponding to = n 1
is important. Also, we can easily verify the following symmetry properties of the Bessel
function:
even
(4.2.7)
n
odd n
n
Plots of In ({3) for various values of are given in Figure 4.4. Table 4. 1 shows the
number of harmonics required to include 80%, 90%, and 98% of the total power of the FM
signals, respectively.

Plots of Bessel functions ln(f3)

1
0.8

0.6

0.4

0.2

- 0.2

-0.4

-0.6
0 1 2 3 4 5 6 7 8 9 10
(3

Figure 4.4 Bessel functions for various values of n .

TABLE 4.1 REQUIRED NUMBER OF HARMONICS IN FM

Power (%) fJ = 0.5 fJ = I fJ = 2 fJ = 5 fJ = 8 fJ = 10 fJ = 15

80 2 4 7 9 14
90 2 5 8 10 15
98 2 3 6 9 11 16
1 68 Angle Modulation Chapter 4

Example 4.2.1
Let the carrier be given by c(t) = 10 cos(2rcfct), and let the message signal be cos(20rct).
Further assume that the message is used to frequency modulate the carrier with k1 = 50.
Find the expression for the modulated signal and determine how many harmonics should be
selected to contain 99% of the modulated signal power.
Solution The power content of the carrier signal is given by
A� 100
Pc = l = l = 50. (4.2.8)

The modulated signal is represented by

(
u (t) = l 0 cos 2rcfct + 2rck1 f_100 cos(20rcr) dr )
(
= 10 cos 2rcfct + �� sin(20rct) )
= 10 cos(2rcfct + 5 sin(20rct)). (4.2.9)

The modulation index is given by Equation (4. 1.16) as

f3 = kt max[lm(t)
fm
I]
= 5; (4.2.10)

00
therefore, the FM-modulated signal is

L Acfn (/3)
u(t) =
n=-oo
00
COS (2rc(fc + nfm)t)

L
=
n=-oo 10ln (5) cos (2rc(fc + lOn)t) . (4.2. 11)

The frequency content of the modulated signal is concentrated at frequencies of the form
fc + lOn for various n . To make sure that at least 99% of the total power is within the effective

n=k
bandwidth, we must choose a k large enough such that

n=-k
10012 (5)
z= ; :::: o.99 x 5o. (4.2. 12)

This is a nonlinear equation and its solution (for k) can be found by trial-and-error and by using
tables of the Bessel functions. In finding the solution to this equation, we must employ the
symmetry properties of the Bessel function given in Equation (4.2.7). Using these properties,
we have

(4.2. 13)

Starting with small values of k and increasing it, we see that the smallest value of k for which
the left-hand side exceeds the right-hand side is k = 6. Therefore, taking frequencies fc ± l Ok
for 0 ::S k ::S 6 guarantees that 99% of the power of the modulated signal has been included
and only 1 % has been left out. This means that if the modulated signal is passed through au
ideal bandpass filter centered at fc with a bandwidth of at least 120 Hz, only 1 % of the signal
Section 4.2 Spectral Characteristics of Angle-Modulated SignalS 1 69

Be = 120 Hz

-Jc
f

Figure 4.5 The harmonics present inside the effective bandwidth of Example 4.2. 1 .

power will be eliminated. This gives us a practical way to define the effective bandwidth of the
angle-modulated signal as 120 Hz. Figure 4.5 shows the frequencies present in the effective
bandwidth of the modulated signal. •

In general, the effective bandwidth of an angle-modulated signal, which contains at

98% of the signal power, is given by the relation
least

Be = 2({3 + l)fm, (4.2.14)

where fJ is the modulation index and fm is the frequency of the sinusoidal message signal. It
is instructive to study the effect of the amplitude and frequency of the sinusoidal message
signal on the bandwidth and the number of harmonics in the modulated signal. Let the
message signal be given by
m(t) = a cos (2rrfmt). (4.2.15)
Using Equations (4.2.14), (4.1.12), and (4.1.13), the bandwidth3 of the modulated
signal is given by

(4.2.16)

or
PM
(4.2.17)
FM
The preceding relation shows that increasing
in PM and FM
a,
the amplitude of the modulating signal,
has almost the same effect on increasing the bandwidth On the other Be.
hand, increasing fm ,
the frequency of the message signal, has a more profound effect in
increasing the bandwidth of a PM signal as compared to an FM
signal. In both PM and FM ,

the bandwidth Be
increases by increasing fm;
but in PM, this increase is a proportional
increase, and in FMthis is only an additive increase which usually (for large {J) is not
,

3For the remainder of the text, "bandwidth" refers to the effective bandwidth, unless otherwise stated.
1 70 Angle Modulation Chapter 4

PM FM

fc + 2fm fc + 2fm

f',,, > fm f',,, > fm

fc + 2f',,, fc + 2f',,,
f
Figure 4.6 The effect of doubling the bandwidth (frequency) of the message in FM and PM.

{l
substantial. Now if we look at the number of harmonics in the bandwidth (including the
carrier) and denote it by Mc,
we have

2 Lkµaj + 3, PM
Mc = 2 ( L.BJ + 1) + 1 = 2 L.BJ + 3 = ka
2 t + 3, FM
j (4.2.18)
fm
In both cases, increasing the amplitude a increases the number of harmonics in the band­
width of the modulated signal. However, increasing fm has no effect on the number of har­
monics in the bandwidth of the PM signal, and it almost linearly decreases the number of
harmonics in the FM signal. This explains the relative insensitivity of the FM-signal band­
width to the message frequency. First, increasing fm
decreases the number of harmonics in
the bandwidth, and at the same time, it increases the spacing between the harmonics. The
net effect is a slight increase in the bandwidth. In PM, however, the number of harmonics
remains constant and only the spacing between them increases. Therefore, the net effect is
a linear increase in bandwidth. Figure 4.6
shows the effect of increasing the frequency of
the message in both FM and PM.

4.2.2 Angle Modulation by an Arbitrary Message Signal

The spectral characteristics of an angle-modulated signal for a general message signal m(t)
is quite involved due to the nonlinear nature of the modulation process. However, there
exists an approximate relation for the effective bandwidth of the modulated signal. This is
known as Carson's ruleand is given by

Be = 2(,8 + l ) W, (4.2.19)
 Section 4.3 Angle Modulators and Demodulators 1 71

f3 =
{kkp
where f3 is the modulation index defined as

max[Jm (t) J],

1 max[Jm( t)J ]
PM
(4.2.20)
, FM
w
and W t
is the bandwidth of the message signal m ( ) . Since wideband FM has a f3 with a
value that is usually around 5 or more, the bandwidth of an angle-modulated signal is much
greater than the bandwidth of various amplitude-modulation schemes. This bandwidth is
either W (in single sideband) or 2W
(in double-sideband (DSB) or conventional AM).

Example 4.2.2
Assuming that m(t) = 10 sinc(l04t), determine the transmission bandwidth of an FM­
modulated signal with k1 = 4000.

Solution For FM, we have Be = 2(f3 + l ) W. To find W, we have to find the spectrum of
m (t). We have M(f) = 10-3 rr c10-4 f), which shows that m(t) has a bandwidth of 5000 Hz.
Since the maximum amplitude of m(t) is 10, we have
k1 max[lm(t) I ] 4000 x IO
f3 = = =8
w 5000
and

Be = 2(8 + 1) x 5000 = 90, 000 Hz = 90 kHz. •

4.3 IMPLEMENTATION OF ANGLE MODULATORS AND DEMODULATORS

Any modulation and demodulation process involves the generation of new frequencies that
were not present in the input signal. This is true for both amplitude- and angle-modulation
systems. Thus, consider a modulator system with the message signal m(t)as the input
and with the modulated signal u(t)
as the output; this system has frequencies in its output
that were not present in the input. Therefore, a modulator (and demodulator) cannot be
modeled as a linear time-invariant (LTI) system, because an LTI system cannot produce
any frequency components in the output that are not present in the input signal.

Angle Modulators. Angle modulators are generally time-varying and nonlinear

systems. One method for directly generating an FM signal is to design an oscillator whose
frequency changes with the input voltage. When the input voltage is zero, the oscillator
generates a sinusoid with frequency fc ;
when the input voltage changes, this frequency
changes accordingly. There are two approaches to designing such an oscillator, usually
called a VCO or voltage-controlled oscillator. One approach is to use a varactor diode. A
varactor diode is a capacitor whose capacitance changes with the applied voltage. There­
fore, if this capacitor is used in the tuned circuit of the oscillator and the message signal is
applied to it, the frequency of the tuned circuit and the oscillator will change in accordance
1 72 Angle Modulation Chapter 4

Lo To oscillator circuit

Figure 4.7 Varactor-diode implementation

of an angle modulator.

with the message signal. Let us assume that the inductance of the inductor in the tuned
circuit of Figure is4.7 Lo
and the capacitance of the varactor diode is given by

C(t) = Co + kom(t). (4.3.1)

When m (t) = 0, the frequency of the tuned circuit is given by fc = �.
Lo Co In general,
2TC
for nonzero m(t) , we have

f; (t) -n-,Jr;L=; o (:;:Co:=:; o1=+::::k=::; om==;=(t;:::;=))

=

2n,JLoCo J1 + Ckoo m(t)

1 1

J
= Jc 1 + 1��m(t) . (4.3.2)
Assuming that
ko
E =
Co « 1
-m(t)
and using the approximations

(4.3.3)
--
and
1 � 1 - E' (4.3.4)
1 +E
we obtain
1 � 1 - -E ·
---,,
,JI+E 2
f;(t) � fc (i - 2�0 m(t)) ,
Hence,
(4.3.5)
which is the relation for a frequency-modulated signal.
A second approach for generating an FM signal is to use a In
reactance-tube implementation, an inductor whose inductance varies with the applied volt­
reactance tube.
age is employed; the analysis is very similar to the analysis presented for the varactor
Section 4.3 Angle Modulators and Demodulators 1 73

Integrator
FM
m(t) </>(t)
+
PM

Carrier 90 ° phase
oscillator shift
� _ � Ac cos
- _
2nfct
Figure 4.8 Generation of a narrowband angle-modulated signal.

diode. Although we described these methods for the generation of FM

signals, basically
the same methods can be applied for the generation of PM signals (see Figure 4.1), due to
the close relation between FM and PM signals.
Another approach for generating an angle-modulated signal is to generate a nar­
rowband angle-modulated signal and then change it to a wideband signal. This method is
usually known as the indirect method
for the generation of FM and PM signals. Due to the
similarity of conventional AM signals, the generation of narrowband angle-modulated sig­
nals is straightforward. In fact, any modulator for conventional AM generation can be eas­
ily modified to generate a narrowband angle-modulated signal. Figure 4.8 shows the block
diagram of a narrowband angle modulator. Next, we use the narrowband angle-modulated
signal to generate a wideband angle-modulated signal. Figure 4.9 shows the block diagram
of such a system. The first stage of this system is to create a narrowband angle modula­
tor, such as the one shown in Figure 4.8. The narrowband angle-modulated signal enters
a frequency multiplier which multiplies the instantaneous frequency of the input by some
constant n.This is usually done by applying the input signal to a nonlinear element and
then passing its output through a bandpass filter tuned to the desired central frequency. If
the narrowband modulated signal is represented by

U n (t) = Ac cos(2rtfct + </J(t)), (4.3.6)

Frequency
Narrowband
r-----�
Frequency
nfc BP Output
angle modulator Xn filter

Local
oscillator
�---� Frequency fw
Figure 4.9 Indirect generation of angle-modulated signals.
1 74 Angle Modulation Chapter 4

the output of the frequency multiplier (which is the output of the bandpass filter) is
given by
y(t) = Ac cos(2nnfct + n<f>(t)). (4.3.7)

In general, this is a wideband angle-modulated signal. However, there is no guarantee that

the carrier frequency of this signal, nfc.
will be the desired carrier frequency. In the last
stage, the modulator performs an up/down conversion to shift the modulated signal to the
desired center frequency. This stage consists of a mixer and a bandpass filter. If the fre­
quency of the local oscillator of the mixer is Ao and we are using a down converter, the
final wideband angle-modulated signal is given by

u(t) = Ac cos (2n(nfc - fw)t + n<f>(t)) . (4.3.8)

Since we can freely choose n and Ao. we can generate any modulation index at any desired
carrier frequency using this method.

Angle Demodulators. FM demodulators are implemented by generating an AM

signal, whose amplitude is proportional to the instantaneous frequency of the FM signal,
and then using an AM demodulator to recover the message signal. To implement the first
step, i.e., to transform the FM signal into an AM signal, it is enough to pass the FM signal
through an LTI system, whose frequency response is approximately a straight line in the
frequency band of the FM signal. If the frequency response of such a system is given by

IH(f)I = Vo + k(f - fc) for

Be
If - !cl < 2 (4.3.9)

( /_100 )
and if the input to the system is

u(t) = Ac cos 2nfct + 2nkf m(r) dr , (4.3. l OJ

( )
then the output will be the signal

v0(t) = Ac C Vo + kk1m(t)) cos 2nfct + 2nkf

The next step i s t o demodulate this signal to obtain
:00
f m(r) dr .
AcCVo + kk1m(t)),
from which th1
(4.3. n

message m (t)
can be recovered. Figure 4. 10 shows a block diagram of these two steps.
Many circuits can be used to implement the first stage of an FM demodulator, i.e.
FM to AM conversion. One such candidate is a simple differentiator with

IH(f)I = 2n f. (4.3. 12

Output
FM signal AM signal .------, signal
FM to AM AM
Figure 4.10 A general FM
convertor demodulator
demodulator.
Section 4.3 Angle Modulators and Demodulators 1 75

)\
H (f)

�
- - - - - - - - - - - - - - -

--------------- I
I
I
I
Figure 4.11 A tuned circuit used in an
Jc f FM demodulator.

Another candidate is the rising half of the frequency characteristics of a tuned circuit,
as shown in Figure 4. 1 1 . Such a circuit can be easily implemented, but usually the linear
region of the frequency characteristic may not be wide enough. To obtain linear charac­
teristics over a wide range of frequencies, usually two circuits tuned at two frequencies
Ji and h are connected in a configuration, which is known as a balanced discrimina­
tor. A balanced discriminator with the corresponding frequency characteristics is shown in
Figure 4. 12.
These FM-demodulation methods, which transform the FM signal into an AM sig­
nal, have a bandwidth equal to the channel bandwidth Be
occupied by the FM signal.
Consequently, the noise that is passed by the demodulator is the noise contained within Be.
A different approach to FM-signal demodulation is to use feedback in the FM demod­
ulator to narrow the bandwidth of the FM detector and, as will be seen in Chapter to 6,
reduce the noise power at the output of the demodulator. Figure 4. 13 illustrates such a sys­
tem. In this figure, the FM discriminator is placed in the feedback branch of a feedback
system that employs a VCO path.
The bandwidth of the discriminator and the subsequent lowpass filter are designed
to match the bandwidth of the message signal m(t).
The output of the lowpass filter is the
desired message signal. This type of FM demodulator is called an FM demodulator with
feedback (FMFB). An alternative to the FMFB demodulator is the use of a phase-locked
loop (PLL), as shown in Figure 4.14 (PLLs are studied in detail in Section 8.8.1).
The input to the PLL is the angle-modulated signal (we will neglect the presence of
noise in this discussion)
u(t) = A e cos [2nfet + ¢ (t)], (4.3 . 13)

where, for FM,

¢ (t) = 2nkt /_100 m(r) dr . (4.3. 14)

The VCO generates a sinusoid of a fixed frequency; in this case, it generates the carrier
frequency fe ,
in the absence of an input control voltage.
Now, suppose that the control voltage to the VCO is the loop filter's output, denoted
as v(t).
Then, the instantaneous frequency of the VCO is

fv (t) = fe + kv v(t), (4.3. 1 5)

1 76 Angle Modulation Chapter 4

..____,,_...,

l��-L--'-�--'-���[H-1(-f-)[
ll���[H-2(-f�
Linear region
(a)
Q)

!2
(b)
f !1

'------y--'
Linear region
(c)

Bandpass filters Envelope detectors

,-R.- - -- - - - --,: :,-_o - - - - - - - - - - �I
I

c.

I m(t)
I
I
I
c.
I:
I
I
I R II : Figure 4.12 A balanced
D
discriminator and the corresponding
� - - - - - - - - - - �(d
�-
) -----------� frequency response.
Section 4.3 Angle Modulators and Demodulators 1 77

Output
f---.---l�
signal
Bandpass Lowpass
Discriminator
filter filter

vco

Figure 4.13 Block diagram of an FMFB demodulator.

Input Output
signal Phase gnal
Loop filter r- si_
comparator
G (f) -.- _-;�

vco Figure4.14 Block diagram of

a PLL-FM demodulator.

kv
where is a deviation constant with units of Hz/Y. Consequently, the VCO output may be
expressed as
Yv (t) = A v sin[2nfct + </>v(t)], (4.3. 16)

where
</>v (t) = 2nkv la t v(r) dr. (4.3. 17)

The phase comparator is basically a multiplier and a filter that rejects the signal
component centered at 2fc ·
Hence, its output may be expressed as
1
e(t) = l A v Ac sm. [</>(t) - <f>v (t)], (4.3. 1 8)

where the difference </>(t) - <f>v (t)

= ¢;(t)
constitutes the phase error. The signal e(t) is
the input to the loop filter.
Let us assume that the PLL is in lock position, so the phase error is small. Then,

sin[</>(t) - </>v(t)] � </>(t) - </>v (t) = </>e (t) (4.3 . 1 9)

under this condition, so we may deal with the linearized model of the PLL, shown in
Figure 4. 15.

<f> (t) +
___<P
_
(
e_t)____ Loop filter
v (t)
g (t)

<f>v(t)
...._
..__ _____-! 2nkvfo� (r) dr �---'

Figure 4.15 Linearized PLL.

1 78 Angle Modulation Chapter 4

We may express the phase error as

<Pe (t) = ¢(t) - 2nkv la t v(r) dr, (4.3.20)

or equivalently, either as
d d
dt <Pe (t) + 2nkv v(t) = -
- dt ¢(t) (4.3.21)
or as
d (t) 2rckv r oo <P (r)g(t - r) dr = d ¢(t). (4.3.22)
dt <Pe + o eJ dt
The Fourier transform of the integro-differential equation in Equation (4.3.22) is
(j2rcf) c'Pe (f) + 2nkv c'Pe (f)G(f) = (j2rcf) c'P (f); (4.3.23)

( ) c'P (f).
hence,
1 (4.3.24)
c'Pe (f) =
1 + �f G(f)
The corresponding equation for the control voltage to the VCO is

V(f) = <P e (f)G(f)

G(f)
O!) c'P (f) . (4.3.25)
1+ G(f)

\ I»
Now, suppose that we design G(f) such that
kv G(f) 1 (4.3.26)
jf
in the frequency band \f l < W of the message signal. Then, from Equation (4.3.25), we

�:
have
·2
V(f) = fv c'P (f), (4.3.27)
TC
or equivalently,

1 d ¢(t)
v(t) = 2nk
v dt
=
kt m(t). (4.3.28)
kv
Since the control voltage of the VCO is proportional to the message signal, v(t)
is the
demodulated signal.
We observe that the output of the loop filter with the frequency response G(f)
i�
the desired message signal. Hence, the bandwidth of G (f)
should be the same as the
 Section 4.4 FM Radio Broadcasting 1 79

bandwidth Wof the message signal. Consequently, the noise at the output of the loop
filter is also limited to the bandwidth W.
On the other hand, the output from the VCO
is a wideband FM signal with an instantaneous frequency that follows the instantaneous
frequency of the received FM signal.
The major benefit of using feedback in FM-signal demodulation is to reduce the
threshold effect that occurs when the input signal-to-noise-ratio to the FM demodulator
drops below a critical value. The threshold effect is treated in Chapter 6.

4.4 FM RADIO BROADCASTING

Commercial FM radio broadcasting utilizes the frequency band 88-108 MHz for the trans­
mission of voice and music signals. The carrier frequencies are separated by 200 kHz
and the peak frequency deviation is fixed at 75 kHz. Preemphasis is generally used, as
described in Chapter 6, to improve the demodulator performance in the presence of noise
in the received signal.
The receiver most commonly used in FM radio broadcasting is a superheterodyne
type. The block diagram of such a receiver is shown in Figure 4. 16.
As in AM radio reception, common tuning between the RF amplifier and the local
oscillator allows the mixer to bring all FM radio signals to a common IF bandwidth of
200 kHz, centered at fw = 10.7 MHz. Since the message signal m (t) is embedded in
the frequency of the carrier, any amplitude variations in the received signal are a result of

RF ,
amplifier

I
I
I
I
I fw = 10.7 MHz
I
Audio
IF
I Mixer Limiter Discriminator amplifier and
I amplifier
I deemphasis
Common '
tuning :
I
I
I
I
I
I Automatic
I
I frequency
I �
control (AFC)
!_./ Local
Lowpass
oscillator
filter

Figure 4.16 Block diagram of a superheterodyne FM radio receiver.

 1 80 Angle Modulation Chapter 4

additive noise and interference. The amplitude limiter removes any amplitude variations in
the received signal at the output of the IF amplifier by hardlimiting the signal amplitude.
A bandpass filter, which is centered at /IF = 10.7 MHz with a bandwidth of 200 kHz,
is included in the limiter to remove higher-order frequency components introduced by the
nonlinearity inherent in the hard limiter.
A balanced frequency discriminator is used for frequency demodulation. The result­
ing message signal is then passed to the audio-frequency amplifier, which performs the
functions of deemphasis and amplification. The output of the audio amplifier is further fil­
tered by a lowpass filter to remove out-of-band noise, and this output is used to drive a
loudspeaker.

FM Stereo Broadcasting. Many FM radio stations transmit music programs in

stereo by using the outputs of two microphones placed on two different parts of the stage.
Figure 4.17 shows a block diagram of an FM stereo transmitter. The signals from the left
and right microphones, m1 (t)
and mr(t),
are added and subtracted as shown. The sum signal
m1(t) + mr(t) is left unchanged and occupies the frequency band 0-1 5 kHz. The difference
signalm1 (t) - mr(t) is used to AM modulate (DSB-SC) a 38-kHz carrier that is generated
from a 19-kHz oscillator. A pilot tone at the frequency of 1 9 kHz is added to the signal for
the purpose of demodulating the DSB-SC AM signal. We place the pilot tone at 1 9 kHz
instead of 38 kHz because the pilot is more easily separated from the composite signal at
the receiver. The combined signal is used to frequency modulate a carrier.
By configuring the baseband signal as an FDM signal, a monophonic FM receiver
can recover the sum signal m1 (t) + mr(t)
by using a conventional FM demodulator. Hence,

Frequency FM
doubler modulator

DSB-SC
Preemphasis 1----.i AM
modulator
FM stereo transmitter

0 15 19 23 38 53 /(kHz)
Baseband signal spectrum

Figure 4.17 FM stereo transmitter and signal spacing.

,\I i
 Section 4.5 Summary and Further Reading 181

Lowpass
filter
0--15 kHz

Signal
from IF �---�

amplifier Narrowband To audio­

FM Frequency To stereo band
tuned filter
discriminator doubler indicator amplifier
19 kHz

38 kHz

Bandpass
Synchronous
filter
demodulator
23-53 kHz

Fignre 4.18 FM-stereo receiver.

FM stereo broadcasting is compatible with conventional FM. In addition, the resulting FM

signal does not exceed the allocated 200-kHz bandwidth.
The FM demodulator for FM stereo is basically the same as a conventional FM
demodulator down to the limiter/discriminator. Thus, the received signal is converted to
baseband. Following the discriminator, the baseband message signal is separated into the
+
two signals, m1(t) mr(t) and m1(t) - mr(t), and passed through de-emphasis filters, as
shown in Figure 4.18. The difference signal is obtained from the DSB-SC signal via a syn­
chronous demodulator using the pilot tone. By taking the sum and difference of the two
composite signals, we recover the two signals, m1(t) and mr(t) . These audio signals are
amplified by audio-band amplifiers, and the two outputs drive dual loudspeakers. As indi­
cated, an FM receiver that is not configured to receive the FM stereo sees only the baseband
signal m1(t) +
mr(t) in the frequency range 0-15 kHz. Thus, it produces a monophonic
output signal that consists of the sum of the signals at the two microphones.

4.5 SUMMARY AND FURTHER READING

This chapter covered frequency modulation (FM) and phase modulation (PM) for analog
signal transmission. The spectral characteristics of FM and PM signals were described, as
well as their bandwidth occupancy. The implementation of FM and PM modulators and
demodulators was also covered. In the final topic of this chapter, we described the use of
FM in radio broadcasting.
As in the case of amplitude modulation, angle modulation (FM and PM) is treated in
numerous books on communication systems, including those that were previously cited in
Section 3.6.
 182 Angle Modulation Chapter 4

PROBLEMS

4.1 The message signal m (t) = 10 sinc(400t) frequency modulates the carrier
c(t) = 100 cos 2nfct. The modulation index is 6.
I
'

1. Write an expression for the modulated signal u (t).

I 2. What is the maximum frequency deviation of the modulated signal?
3. What is the power content of the modulated signal?
4. Find the bandwidth of the modulated signal.

Q
m(t)
; signal m (t) is shown in Figure P-4.2; this signal is used once to frequency modulate
a carrier and once to phase modulate the same carrier.

0 1 2 3

-1 - - - - - - - - - - - - ....�
... ��
Figure P-4.2

1. Find a relation between kp and kf such that the maximum phase of the modu­
lated signals in both cases are equal.
2. If kp = fd = 1 , what is the maximum instantaneous frequency in each case?
" iL t-
4.3 Determine the in-phase and quadrature components as well as the envelope and the
phase of FM and PM-modulated signals.
-

4.4 An angle-modulated signal has the form

u (t) = lOO cos [2nfct + 4 sin 2000nt] ,

where fc = 10 MHz.

1. Determine the average transmitted power.

2. Determine the peak-phase deviation.
3. Determine the peak-frequency deviation.
4. Is this an FM or a PM signal? Explain.

4.5 Find the smallest value of the modulation index in an FM system that guarantees
that all the modulated-signal power is contained in the sidebands and no power is
transmitted at the carrier frequency when the modulating signal is a sinusoid.
Problems 183

4.6 To generate wideband FM we can first generate a narrowband FM signal, and then
,

use frequency multiplication to spread the signal bandwidth. Figure P-4.6 illustrates
such a scheme, which is called an Armstrong-type FM modulator. The narrowband
FM signal has a maximum angular deviation of 0.10
radians to keep distortion under
control.

1. If the message signal has a bandwidth of 15 kHz and the output frequency from
the oscillator is 100
kHz, determine the frequency multiplication that is neces­
sary to generate an FM signal at a carrier frequency of = fc
MHz and a 104
frequency deviation of f = 75
kHz.
2. If the carrier frequency for the wideband FM signal is to be within ±2 Hz, deter­
mine the maximum allowable drift of the I 00 kHz oscillator.

Narrowband Frequency
FM modulator multipliers

A (2nf0t)
cos

Frequency
Jo = lOO kHz multipliers

Figure P·4.6 Armstrong-type FM modulator.

4.7 Determine the amplitude and phase of the various frequency components of a PM

{
signal with kp = 1 and with m(t) a periodic signal given by
1, 0 <- t < T2m
m(t) =
-1, T,2n < t < T.
- - m

in one period.

[ /_100 m (r) dr J
4.8 An FM signal is given as

u(t) = 100 cos 2nfct + 100 ,

where m(t) is shown in Figure P-4.8.

1. Sketch the instantaneous frequency as a function of time.
2. Determine the peak-frequency deviation.
 1 84 Angle Modulation Chapter 4

m(t)

-1 1 2

-5 Figure P-4.8

4.9 The carrier c(t) = 100 cos 2nlet is frequency modulated by the signal m (t) =
5 cos 20,000nt, where le = 108 Hz. The peak-frequency deviation is 20 kHz.

1. Determine the amplitude and frequency of all signal components that have a
power level of at least 10% of the power of the unmodulated carrier component.
2. From Carson's rule, determine the approximate bandwidth of the FM signal.

I,
fl
4.10 The carrier c(t) = A cos 2n l 06 t is angle modulated (PM or FM) by the sinu­
soid signal m (t) = 2 cos 2000nt. The deviation constants are kp = 1 .5 rad/V and
kf = 3000 HzN.

1. Determine f3t and f3p·

2. Determine the bandwidth in each case using Carson's rule.
3. Plot the spectrum of the modulated signal in each case. (Plot only those fre­
quency components that lie within the bandwidth derived in Part 2.)
4. If the amplitude of m (t) is decreased by a factor of 2, how would your answers
to Parts 1-3 change?
5. If the frequency of m (t) is increased by a factor of 2, how would your answers
to Parts 1-3 change?

4.11 The carrier c(t) = 100 cos 2nlet is phase modulated by the signal m (t) =
5 cos 2000nt. The PM signal has a peak-phase deviation of n/2. The carrier fre­
quency is le = 108 Hz.

1. Determine the magnitude spectrum of the sinusoidal components and sketch the
results.
2. Using Carson's rule, determine the approximate bandwidth of the PM signal and
compare the result with the analytical result in Part 1 .
Problems 185

4.12 An angle-modulated signal has the form

u(t) = lOO cos [2nfct + 4 sin 2nfmt] ,

where fc = 10 MHz and fm = 1000 Hz.

1. Assuming that this is an FM signal, determine the modulation index and the
transmitted-signal bandwidth.
2. Repeat Part I if fm is doubled.
3. Assuming that this is a PM signal, determine the modulation index and the
transmitted-signal bandwidth.
4. Repeat Part 3 if fm is doubled.

4.13 It is easy to demonstrate that amplitude modulation satisfies the superposition princi­
ple, whereas angle modulation does not. To be specific, let m1 (t)
and m2 (t) represent
two message signals and let u1 (t)
and u2 (t)
represent the corresponding modulated
versions.

1. Show that when the combined message signal m 1 (t) m t

+ 2 ( ) DSB modulates a
carrier Ac 2n Jct,
cos the result is the sum of the two DSB amplitude-modulated
signals u1 (t) + u2 (t).
2. Show that if m1 (t) + m 2 (t) frequency modulates a carrier, the modulated signal
is not equal to u1 (t) + u 2 (t).

4.14 An FM discriminator is shown in Figure P-4. 14. The envelope detector is assumed
C if the
to be ideal and has an infinite input impedance. Select the values for L and
discriminator is to be used to demodulate an FM signal with a carrier fc 80 MHz
=
and a peak frequency deviation of 6 MHz.

L C
�lO kf!
u (t)
Envelope
m (t)
detector

Figure P-4.14

4.15 An angle-modulated signal is given as

u(t) = 100 cos [2000nt + </>(t)] ,

where (a) <f>(t) = 5 sin 20nt and (b) </>(t) = 5 cos 20nt. Determine and sketch the
amplitude and phase spectra for (a) and (b), and compare the results.
1 86 Angle Modulation Chapter 4

4.16 The message signal m(t) into an FM modulator with a peak frequency deviation
fd = 25 HzN is shown in Figure P-4. 16. Plot the frequency deviation in Hz and the
phase deviation in radians.

m (t)

2 ------------------

1 ---------

-1

-2 -----------------
Figure P-4.16

4.17 A message signal m(t) has a bandwidth of 10 kHz and a peak magnitude jm(t)J
of 1 volt. Estimate the bandwidth of the signal u(t) obtained when m(t) frequency
modulates a carrier with a peak-frequency deviation of (a) fd = 10 HzN, (b) 100
HzN, and (c) 1000 HzN.
4.18 The modulating signal that is the input into an FM modulator is
m(t) = 10 cos 16nt.

[ i� m (r) dr] ,
The output of the FM modulator is

u(t) = l O cos 4000nt + 2nkf

where kf = 10. (See Figure P-4. 18.) If the output of the FM modulator is passed
through an ideal BPF centered at Jc = 2000 with a bandwidth of 62 Hz, determine
the power of the frequency components at the output of the filter. What percentage
of the transmitter power appears at the output of the BPF?

BPF
FM
m (t) demodulator u (t) Output
-1 62 Hz I-
Jc = 2000 Hz 1
k1 = 10
2000 f
Figure P-4.18

4.19 The message signal m1 (t) is shown in Figure P-4.19 and the message signal m 2 (t) =

sinc(2 x 104 t) volts.

 Computer Problems 187

0 1 3 4 t(sec) Figure P-4.19

1. If m1 (t)is frequency modulated on a carrier with frequency 106 Hz and a
frequency-deviation constant (k 1) equal to 5 HzN, what is the maximum instan­
taneous frequency of the modulated signal?
2. If m1 (t) is phase modulated with a phase-deviation constant kp = 3 radN, what
is the maximum instantaneous frequency of the modulated signal? What is the
minimum instantaneous frequency of the modulated signal?
3. If m2 (t) is frequency modulated with k1 = 103 HzN, what is the maximum
instantaneous frequency of the modulated signal? What is the bandwidth of the
modulated signal?

4.20 A superheterodyne FM receiver operates in the frequency range of 88-108 MHz.

The IF and local-oscillator frequencies are chosen such that /IF < Ao. We require
that the image frequency f; fall outside of the 88-108 MHz region. Determine the
minimum required fw and the range of variation in Ao .

COMPUTER PROBLEMS

{
4.1 Frequency Modulation
The message signal
1 0 .:::: t t0/3
<

m(t) = - 2 to /3 .:::: t 2to /3

<

0 otherwise
frequency modulates the carrier c(t) = cos 2nfct, where fc = 200 Hz and to =
0.15 sec. The frequency-deviation constant is kI = 50. Therefore, the frequency­
modulated signal is

1. Plot the message signal m(t) and its integral on two separate graphs. The sam­
pling interval is ts = 0.0001.
1 88 Angle Modulation Chapter 4

2. Plot the FM signal u(t).

3. Use MATLAB's Fourier-transform routine to compute and plot the spectra of
m (t) and u (t) on separate plots.
4. Suppose we define the bandwidth W of m(t) as the width of the main lobe
of its spectrum. Determine the modulation index fJ and the modulation-signal
bandwidth Be using Carson' s rule.
4.2 Frequency Modulation

The message signal

m (t) = {� nc( lOOt) ltl ::::; to
otherwise
frequency modulates the carrier c(t) = cos(2rclet), when le = 250 Hz and to = 0. 1 .
The frequency-deviation constant i s k1 = 100. Therefore, the frequency-modulated

( 1: m(r) dr) .
signal is
u (t) = cos 2rclet + 2rck1

00 separate graphs. The sampling

1. Plot the message signal and its integral on two
interval is ts = 0.0001.
2. Plot the FM signal u(t).
3. Use MATLAB ' s Fourier-transform routine to compute and plot the spectra of
m (t) and u (t) on separate graphs.
4. Demodulate the FM signal u (t) to obtain the message signal and compare the
result with the original message signal. The FM signal can be demodulated by
first finding the phase of u (t), i.e., the integral of m(t), which can be differenti­
ated and divided by 2rck1 to yield m(t). Use MATLAB function unwrap . m to
undo the effect of 2rc-phase foldings. Comment on how well the demodulated
message signal matches the original message signal m (t).

{
4.3 Frequency Modulation

The message signal

t O<t < l
m(t) = -t + 2 1 :::: t < 2
0 otherwise
frequency modulates the carrier c(t) = cos(2rclet), when le 1000 Hz. The
frequency-deviation constant is kf = 25.
1. Plot the message signal and its integral on two separate graphs.
2. Plot the FM signal
Computer Problems 1 89

3. Use MATLAB's Fourier-transform routine to compute and plot the spectra of

m(t) and u(t) on separate graphs.
4. Determine the modulation index, the bandwidth, and the range of the instanta­
neous frequency of u(t).
5. Demodulate the FM signal u(t) to obtain the message signal and compare the
result with the original message signal. The FM signal can be demodulated by
first finding the phase of u(t), i.e., the integral of m(t), which can be differenti­
ated and divided by 2rtk/ to yield m(t). Use the MATLAB function unwrap . m
to undo the effect of 2rc-phase foldings. Comment on how well the demodulated
message signal matches the original message signal m(t).