r-:l-

,:ii,$
:i4E

$ .,,,.,.
it:tr!;:;t&+rid,;Ir!,':
f
.€t
;;i.

ii+ .:

11., iit;ii
;li1,:,,:r :i,ii

iS..
,:ur 'I

::t:L, I l

titi l

rr$ ''

**#
ritil

,*ffi iiwtr,
Angle Modulation
4.1 INTRODUCTION
As we mentioned in Sec. 3.1, angle modulation encompasses phase modulation (PM) andfrequencv
modulation (FM) and refers to the process by which the phase angle of a sinusoidal carrier wave is var-
ied according to the message signal. As we studied in Chap. 3, in amplitude modulation the spectrum
of the modulated signal is essentially the translated message spectrum, and the transmission bandwidth
never exceeds twice the message bandwidth. In angle modulation, the spectral components of the
modulated waveform are not related in any simple fashion to the message spectrum. Furthermore,
superposition does not apply (see Solved Problem 4.5), and the bandwidth of the angle-modulated
signal is usually much greater than twice the message bandwidth. The increase in bandwidth and system
complexity is compensated for by the improved performance in the face of noise and interference
(see Chap. 8).

4.2 ANGLE MODULATION AND INSTANTANEOUS FREQUENCY

For angle modulation, the modulated carrier is represented by [see Eq. (3.1)]

x,(t)-lcosla,t+OQ)1 (4.1)

where A and uc are constants and the phase angle @(r) is a function of the message signal m(t).lf we
rewrite Eq. (4.1) as
x,(t): A cos 0(t) (4.2)

where 0(t) - a,t * qQ) (4.3)

then we can define the instantaneous radian frequency of x.(r), denoted by w,, as

do(') do$)
o,- 'dt'dt -,t-* (4.4)

Note that when QQ): constant, then u.,, : e,. r

 Analog and Digital bnmunications

The function /(r) and d$(t)ldt are known as the instantaneous phose deviotion and instantaneous
frequency deviation of x"(t). The quantity A,a defined by
L,u la,_ r"l^^* : (4.s)
is called the maximum (or peak) radianfrequency deviation of the angle-modulated signal.

4.3 PHASE AND FREQUENCY MODULATION

The two basic tlpes of angle modulation are phase modulation and frequency modulation. In PM, the
instantaneous phase deviation of the carrier is proportional to the message signal; that is,

kom(t) (4.6)
Q(t) -
where to is the phase deviation constant, expressed in radians per unit of m(t).
In FM, the instantaneous frequency deviation of the carrier is proportional to the message signal;
that is,

(4.7)
ry:krm(t)
or : o, mQ)il, + 6(to) (4.8)
dQ)
[),
where k, is the frequency deviation constant, m(t)
expressed in radians per second per unit of
m(t), and $(t) is the initial phase angle at /
- to. It is usually assumed that /o : -oo and
d(-oo) - 0.
carrier
Thus, we can express the angle-modulated
signal as

xpr,a(/)
-I cos fw"t * kpm(t)) (4.9)

xrr,a(/)
- ,n cos * o,
I_* m(^)a.l] t+.rol
lr,,
From definition (4.4), we have

ei: a, n o,
ry for PM (4.r l)

ai: a, * ktm(t) for FM (4.12)

Thus, in PM, the instantaneous frequency r,,,,

varies linearly with the derivative of the modu-
lating signal, and in FM, u,r, varies linearly with
the modulating signal.
Figure 4.1 illustrates AM, FM, and PM
waveforms produced by a sinusoidal message
wave-form. Fig. 4.1 AM, FM, and PM waveforms
 Angle Modulotion

4.4 FOURIER SPECTRA OF ANGLE.MODULATED SIGNALS

An angle-modulated carrier can be represented in exponential form by writing Eq. (4.1) as

x,(t) - Re (Asi@"t+ftt))) - Re (Aei',t eioo)) (4.l3)

where Re means the "real part of." Because of this representation, the angle modulation is also referred
to as the exponentiol modulation.
Expandin g slQQ) in a power series yields

x,(t) -x"fe"'*' io.a-ry-... + i' o'!!-.

[,* ]]

- ,a a"t -@(r) sin a,t - ry cosact . ry sin w"t. (4.14)

[.o, ]

Thus, the angle-modulated signal consists of an unmodulated carrier plus various amplitude-modulated
terms, such as @(r) sin w,t, $2(t) cos cr.r"/, d3(r) sin u"t, ..., and so on. Hence, its Fourier spectrum consists
of an unmodulated carrier plus spectra of $(t), O'Q), dt Q),. . ., and so on, centered at w,.
It is clear that the Fourier spectrum of an angle-modulated signal is not related to the message signal
spectrum in any simple way, as was the case in AM.

4.5 NARROWBAND ANGLE MODULATION

ld(r)l*.. << I (4.15)

then Eq. (4.1+) can be approximated by [neglecting all higher-power terms of d(r]
x"(t) x, A cos @,t - AO(t) sin u"t (4.16)

The signal represented by Eq. (4. 16) is called the narrowband (NB) angle-modulated signal. Thus,

xNsprra(/)
=,4 cos a,t - Akrm(t) sin w"t (4.17)

xNerv(/) = I cos a"t - ulo, [_* *txldA]sin,,.,"r (4.18)

Equation (4.16) indicates that a narrowband angle-modulated signal contains an unmodulated carrier
plus a term in which d(t) la function of m(t)l multiplies arl2 (rad) phase-shifted carrier. This multiplica-
tion generates a pair of sidebands, and it Q(t) has a bandwidth WB, the bandwidth of an NB angle-modu-
lated signal is2Wr. This is reminiscent ofAM.

4.6 SINUS0IDAL (0R ToNE) M0DULATToN

A. Modulation Index
If the message signal m(t) is a pure sinusoi{ that is,
l' 4.4 | Anolog and Digitol Communications

sin,rl*t for PM
m(t) (4.te)
- I:: cos a*t for FM

then Eqs. (4.6) and (4.8) both give

dG) - t3 sin a^t (4.20)

lk oo, for PM
I
(4.2r)
where ,, - 1krr, for FM
t. *,,
|
The parameter p is known as the modulation index for angle modulation and is the maximum value of
phase deviation for both PM and FM. Note that B is defined only for sinusoidal modulation and it can
be expressed as

t4
P- _L,A (4.22)
-
u)m

where Ai.,,is the maximum frequency deviation, defined in Eq. (4.5).

B. Fourier Spectrum
Substituting Eq. (4.20) into Eq. (4.1), we obtain

x"(t) : A cos (utrt * p sin u*t) (4.23)

which is the angle-modulated signal with sinusoidal modulation. It can be shown by the use of the Fourier
series that this signal can also be written as (see Solved Problem 4.6)

x"(t) - u ,\ J,(P) cos (r^u" t nu^)t (4.24)

where J"(P) is the Bessel function of the first kind of order n and argument p.Table B-1 (inApp. B) lists
some selected values of J,(0). From Eq. @.2a) and Table B-1, we observe that

I . The spectrum consists of a carrier-frequency component plus an infinite number of sideband com-
ponents at frequencies o,,. *. nu*(n : 1,2,3, ......).
2. The relative amplitudes of the spectral lines depend on the value of J"(p) and the value of J,(p)
becomes very small for large of n.
3. The number of significant spectral lines (that is, having appreciable relative amplitude) is a
function of the modulation index p. With P << 1, only Jo and J, arc significant, so the spec-
trum will consist of carrier and two sideband lines. But if P >> l, there will be many sideband
lines.

Figure 4.2 shows the amplitude spectra of angle-modulated signals for several values of O.
 t-n-l

ec- am iic ac+ (rm

,d: 10

uca
Fig. 4.2 Annpt'itude spectra of sinusoidatLy modulated FM signals (w, fixed)

4.7 BANDWIDTH OF ANGLE-MODULATED SIGNALS

A. Sinusoidal Modutation
From Fig. 4.2 and Table B.1 we see that the bandwidth of the angle-modulated signal with sinusoidal
modulation depends on p and a..In fact, it can be shown that 98 percent of the normalized total signal
power is contained in the bandwidth
Wnx2(0+l)a* (4.2s)
When P << 1, the signal is an NB angle-modulated signal and its bandwidth is approximately equal to
2w*. Ustally a value of P < 0.2 is taken to be sufficient to satisfy this condition. A1l the bandwidths can
be expressed in hertz (Hz) simply by replacing Ac,,,with Lf and wn wrthf,.

B. Arbitrary Modutation
For an angle-modulated signal with an arbitrary modulating signal m(t)band-limited to arradians per
second (radls), we define the deviation ratio D as

^
I):
maximum frequency deviation L,a
(4.26)
bandwidth of m(t) aM
The deviation ratio D plays the same role for arbitrary modulation as the modulation index p plays for
sinusoidal modulation. Replacing /by D and arby ,, in Eq. (4.25), we have

Wa=2(D+l)u, (4.27)

This expression forbandwidth is generally referred to as Carson's rule.If D << l, the bandwidth is
approximately 2wp1, and the signal is known as a narrowband (NB).angle-modulated signal (see Sec. 4.4).
 Analog and Digital Communicotions

If D >) 1, the bandwidth is approximately 2Dw, - 2La, which is twice the peak frequency deviation.
Such a signal is called awideband (WB) angle-modulated signal.

4.8 GENERATION OF ANGLE-MODULATED SIGNALS

A. Narrowband Angle-modutated Signals
The generation of narrowband angle-modu-
lated signals is easily accomplished in view of
Eq. (4.16) or Eqs. (4.17) and (4.18). This is illus-
trated in Fig. 4.3.

B. Wideband Angte-modutated Signals

There are two methods of generating wideband
(a) NBPM
(WB) angle-modulated signals ; the indirect method
and the direct method.

7. Indirect Method In this metho4 an

NB angle-modulated signal is produced first (see
Fig. 4.3) and then converted to a WB angle-
modulated signal by using frequency multipliers
Gig. a.a). The frequency multiplier multiplies the
argument of the input sinusoid by n. Thus, if the (b) NBFM
input of a frequency multiplier is Fig. 4.3 Generation of narrowband
angte-modulated signa Ls
xQ):lcos[a.t+O(t)7 (4.28)

then the output of the frequency multiplier is

y(t) -l cos fna"t + n$(t)) (4.2e)

Use of frequency multiplication normally increases

the carrier frequency to an impractically high xn
value. To avoid this, a frequency conversion (using
a mixer or DSB modulator) is necessary (Fig. a.5)
Fig, 4.4 Frequency muttiptier
to shift the spectrum.

2. DireCt MethOd In the direct method of generating an FM signal, the modulating signal directly
controls the carrier frequency. A common method used for generating FM directly is to vary the
inductance or capacitance of a tuned electric oscillator. Any oscillator whose frequency is controlled
by the modulating signal voltage is called a voltage controlled oscillator (VCO). The main advantage
of direct FM is that large frequency deviations are possible, and thus less frequency multiplication is
required. The major disadvantage is that the carrier frequency tends to drift, and so additional circuitry is
required for frequency stabilization.
 Angle Modulation

NBFM signal WBFM signal

r", f t1|,".,
I "r:
Lf2: n61.
I ar,
I 1

Fig. 4.5 NB-to-WB conversion

4.9 DEMODULATION OF ANGLE.MODULATED SIGNALS
Demodulation of an FM signal requires a system that produces an output proportional to the instanta-
neous frequency deviation of the input signal. Such a system is called afrequency discriminator. If the
input to an ideal discriminator is an angle-modulated signal
x,(t)-Acosfw^t+d@l
then the output of the discriminator is

. do1)
ra(t)-rtoi (4.30)

where fr, is the discriminator sensitivity.

For FM, d(4 is given by [Eq. (4.8)]

6$) - o, I_* m(\)d\

so that eq. 1+.:O) becornes
ya(t): kakrm(t) (4.31)

The characteristics of an ideal frequency discriminator are shown in Fig. 4.6.

The frequency discriminator also can be used to demodulated PM signals. For PM, $(t) is given by
[Eq. (4.6)]
kom(t)
d(t) -
Then yaQ), given by Eq. (4.30), becomes

ya4)
- trokrT (4.32)

Integration of the discriminator output yields a signal that is proportional to m(t). A demodulator for PM
can therefore be implemented as an FM demodulator followed by an integrator.
A simple approximation to the ideal discriminator is an ideal differentiator followed by an envelope
detector (Fig..7). If the input to the differentiator is

x,(t) --.4 cos la"t + SQ))

i aR I Anolog and Digital Communications

lnput frequency l1
,Jr) i [7l xLo l-E""t,ro" I i, vo?)

l_____

Fig. 4.6 Characteristics of ideal frequency discriminator Fig. 4.7 Frequency discriminator

then the output of the differentiator is

x;(r): -olr, + !99)sin[,",.r

'( + o(,,)] (4.33)
[' dt ]
The signal x!(t,) is both amplitude- and angle-modulated. The envel ope of x/,(r) is

r
ol', *;)
_
doT))

The output of the envelope detector is, by Eq. (4.4),

Y,t(t): oi (4.34)

which is the instantaneous frequency of the x,(t).

There are many other techniques that can be used to implement a frequency discriminator. (See Solved
Problems. 4.25,4.26, and 4.27.)

Instantaneous Frequency

4.1 Determine the instantaReous frequency in hertz of each of the following signals:

(a) I o cos
(ron*,. ;J
(b) l0 cos(207rt + ri)
(c) cos Z}Amcos (5 sn}nt) * sin 200ntsin (5 sin 2zrr)

(a) 0(t)
- 2A0rt +:
J

,,: 4dt :2oon- 2z-(loo)

The instantaneous frequency of the signal is 100 Hz, which is constant.
 ergb Uodutotirn f 4.g 1
(b) 0(t)
- 20rt + rt2
o,: * =. 2(h -r 2irt : 2r(10 +0
dt
The instantaneous frequency of the signal is 10 Hz at t : 0 and increases linearly at a rate of
I Hzls.
(c) cos 200rt cos (5 sin2trt) * sin200urrsin(5 sin 2nt)
- cos (200trt - 5 sin2rt)
0(t):200rt - 5 sin2trt

,r, - (,Jt * 200r - I0r cos 2trt - 2r(100 - 5cos 2zrr)

The instantaneous frequency of the signal is95Hz at t 0 and oscillates sinusoidally between
95 and 105 Hz.
-

4.2 Consider an angle-modulated signal

, x,{t):10 cos [(108)ar * 5 sin ur(t}3)tJ
Find the maximrmr phase deviation and the maximum frequency deviation.

Comparing the given x"(r) with Eq. (4.1), we have

0(t)
- a,t * 0(t) - (108)zrr * 5 sin 2r(103)t

and dU)-5sin2tr(103)t
Now a/(t) 5(2n)(103) cos 2tr(103)t
-
Thus, the maximum phase deviation is

ld(t)l,"",. - 5 rad
and the maximum frequency deviation is
La : lQ, Q)l^ * - 5(2r)(t03) radls
Lf : 5 kllz

4J 4.8: rr,(0
For the message sigual rr(r) showu in the Fig.
(a) If m(t) is frequency modulated on a carrier with -^,i
10"
frequency 10dHz, *itt u frequency deviation con-
stant (frr) eqrral to 5 HzN what is the maximum
ihstantaneous frequency of the modulated signal?
(b) If m(t) is phase modulated with phase-deviation
constant kr: 3 radiansfl{ what is the maximum 13
instantaneoue frequency of the modulated signal? Fig. 4.8

(a) The instantaneous frequency is

f,(t) : f, * k, m(t)
-
maxlf,(t)) - maxlf, + krm(t))- (106 * 5 x 105) - 1.5 MHz.
 Analog and Digitol Communications

(b) The phase of the phase modulated signal is @(l)

- kp m(t) and the instantaneous frequency
: f, * Oal (see Eq. 4.1 1)
f,(t) * fi
k.d
:f,+ -i*rr,
The maximum of f,(t)is achieved for r in [0, 1] where *Ur: 105.
*
Hence, maxlf,(t)J: 106 +: x 105
- 1.048 MHz
2r
Phase Modutation and Frequency Modutation
4.4 Anangle-modulated signal is described by
, r. ' ,x"{t)
=',10 cos [2n'(106;r + 0.1 sin (I03]n"r-J
:,:, fl
r".'
*
,, :

(a) Conside*ngx"(fl as a PM signal with *; 10, f,ind t1,4,t:).

(b) Consided-ng {,(D as an FM signal with'fu+ I0r,1ffi,ilr'({}.,- '.i

(a) rpr,a(/)
- I cos fc.',.t * kom(t)l
- 10 cos [2zr(106;r + 10m(t)]
- 10 cos [2zr(t06;l + 0.1 sin (103)zrr]
Thus, m(t1 :0.01 sin (lO3)tt
lnrl
(b) x.y(I) -- A cos + kt J _* m(\)d),)
l*'.r
: 10 cos [2a(10!r + 0.1 sin 1tO3)z'r]

Assuming
m(t) : am cos (103)nt
we get

fin f'* m(^)d).: lora* ['_* cos (10312') d)

sin (103)nt: 0.1 sin (1 03)nt

ft
Thus, a^: 10, and

m(t) 10 cos (103)z'r

-
Letmr(t) andmr(t) be twe mp-$sage signals, and let x*(r)and,xo,(f),, ,,!he msdplxed sig+
co.responding to m(t) and mr(t), respectively.

(a) Show that if the modulation is DSB (AM), then n,(r) * mr(t)will produce a modulated signal
equal to :ua (r) +:r- .(f),. (!his is why sometilnes referred,to as a line*r modulation.)
 _ Angle Modulotion fal
, .---

(b) Show that if the modulation is PM, then the modulated signal produced by mr(t) * mr(t)
will not be x", (r) + x",(t); that is, superposition does not apply to angle-modulaied signals.
,fl,.his is why anEle modulation is sometimes referred to as a nonlineqr modulation.)

(a) For DSB (AM), form Eq. (3.3) we have

mlt) - (t) : mlt) cos act
r.,
mz(t) --- .rc: (r)
- mz(t) cos c,,r"/
m{t) * mr(t) n x,,(r) - l*r(t) + mr(t)l cos r,r"/
: mlt) cos urt * mr(t) cos r.r,r"/

: xr,(/) + xr,(t)
Hence, DSB (AM) modulation is a linear modulation.
(b) For PM, by Eq. (a.9) we have
,(/) -' x",(t) : I cos fw"t * komr(t)l
mz(t)
- *",(r) : I cos lal + krmr(t)l
mlt) * m,r(t) -- x,(t): I cos {a,t * kolmlt) + mr(t))}
* *r,(r) + xr,(t)
Hence, PM is a nonlinear modulation.

4,6 Derive Eq. @.2$.

In a sinusoidal angle modulation, the modulated signal [Eq. (4.23)]
- x"(t) A cos (a,t * p sin a*t)
-
can be expressed as

x"(t) -I Re (ei'J
"io'i"'"'t
(4.3s)
The function ei?"in'-' is clearly a periodic function ' with peri od T* - 2r/a*.It therefore has a
Fourier series representation

e' "' -- S
oiSsinu;,,,r ^ ^-ina,,,t
"
,4cn€
By Eq. (1.3), the Fourier coefficients cn can be found to be

"'"'
,, : ? Jf _ri*,,, ,lt)sin-""r e- rne""t dt
2V

Setting e^t: x, we have

,,: y'tjsinx*nx)d*: J,(p)

* [:_
where J,(Ais the Bessel function of the first kind of order n andargument p (see App. B).
f 4i.j1 Anolog ord Digitol Co**unirotion,

Thus, ,jfsinn,,t : D Jr(o)ej",' (4.36)

Substituting Eq. (4.36) into Eq. (4.35), we obtain

tool Ioo I
x,(t) -l Re lr,'l n:-*
D J n(P)ei"'.' | - A
I
I
nel z-t J,(g)"j('"*"^"
I I Ix=-oo )

Taking the real part yields

x"(t) - n J'(acos (.'''" * na*)t

,\
4.7 Find the normalixed average power in an argle*rnodrllafed signal wjth sigusoidal modulation.
From Eq. $.24), an angle-modulated signal with a single-tone modulation can be expressed as

, x",, AJ,(Acos (c.,,r. I nu*)t

,i:*
The normalized average power in x,(t) is given by

'- ,t--- )n"',(a: lo',4 r',(a: )u' (4.37)

slnce D
n:_@
J',(a -| (4.38)

4.8 A certain sinusoid at a frequency f* Hz is used as the modulating signal in both a corventional
AM system and a FM system. When modulate4 the peak frequency deviation of the FM system
is set to three times the bandwidth of theAM system. The magnitudes of those sidebands spaced
at *f*Hz from carrier in both systems are equal, and the total average pov/ers are equal in both
systems.
Determine the (a) modulation index of the FM systern (b) modulation index of the AM systern.

(a) Given that Lf : 3 x 2f* : 6f*

mr: Lf 6f'
t f* -- -f^ - 6

(b) Let Arbe the peak amplitude of the carrier in the AM system and Arbe peak amplitude of the
carrier in the FM system.
Total power in AM System: 42 * I4',t-,',q J
where pt is the modulation index.

Total power in FM system- 4

2
Magnitude of the first sideband in FM systemis Arll,(6)l
 Magnitude of the first sideband in AM system is F ,q,
2r
(i) '2
/2 ,4? p, Al
urvvrr\r,,i'
Given
2 -- T
f
4
(1i) A2U'(6)l : tU'
From Bessel function Tables, 46) - -0.28
(i) and (ii) can be solved simultaneously to yield
2146)l :0.61

{.s,.,ffii$ffiffiiui,Iffiffii E#ffi.ii$.fiequene5r:mo ated by the signal m{t) -* 5 cos (20,000 zrr),

'urh$r;4 ; i flz fho eeak frcquency deviation is 20 kHz.
:::
D€tnrmine the amplitqde and frequency of all signal components that have a power level of at least
l0% of the porrer of the unmodulated carrier component.

The modulation index, a Lf

L/-
20x103 .
f_ 10,000
The modulated signal has the form

u(t) - tor,i=* J,(2) cos (2tr1to8 +n. t04)t)

1002
The power of the unmodulated carrier signal is P
- ; : 5000; l0% of 5000 : 500

rc02 4 Q)
^nffih^66n + t-: f,
,ran^r, componentf
Power in the frequency t *t kym i^ ID
t, -^ is
p + kfm --
2
Using Bessel function tables the amplitude and power of the frequency components are calculated as
The signal components that have a power level greater than 500 are those with frequencies 108 + 104
and 108 *2x 104. Since J,'(P)-J-|(B),thereforethefrequencies 108- l04and 108 -2 x 104
will also satisff the minimum power level condition.
Hence these are the four signal components. Their amplitudes and power are already shown in
Table 4.1.

Table 4.1
k JoQ) Frequency Hz Amplitude 100 JkQ) Power Pf, * of*

0 0.224 108 22.4 2s0.6

I 0.577 + lo4
108 57.5 1663.1
2 0.3s3 108+2xl}a 3s.3 622.s
3 0.129 108+3x104 12.9 83. I
4 0.034 108+4xl}a 3.4 5.7
 -2-1 0123
, Fig. 4.9(a)
(a) It is an FM signal, because the message signal is integrated to produce a phase variation of
the carrier ,
(b) Message bandwidth for the periodic signal is approximately W ! Hz
- )
Using Eq. 4.10 and 4.12
kfr , 5
k"-10-AL-..:
,.J m(t\ I
2f,,"r", I_u* -:
^"r-:/max
7f

a Lf^u* 5 10
W rU2 7T

o- -
Modulation index {J:
1o

Transmission Bandwidth is 2(W + A.f)

012
: z lL+:l - 1''o + rl H, -1
l.z n) lz' ) Fig. 4.9(b)
(c) The instantaneous frequency is given using F,q.4.12.

/;(t):loo+
'2n l romQ)

'(d) Peak freq deviation (A/)-u*

5
-

Fourier Spectra of Angte-modu,.*O Signats

4.11 A carrier is angle-moduta,ld blthe t;rm."tt:"
:**o*.^prsinurt)
x,(t);:r{ cos (urt * 0,
K wrt * (4.39)
where c,;, and uzarenot harmonically related. Find"the spectrum of xr(r).
 I-4.15 I
ln a manner similar to Solved Problem 4.6, x,(t) can 6e expressed as

x"(t)::::""0:,';:'""':":":)'u'""."^':''l
(4.40)

Using Eq. (a.36), we have

ujglsin',t : D
,l j,(gr)sin',t

_ J,,(gr)si*r,t
"iozsrnazt I

Substituting these expression into Yq f+ iO>-iO,uO,rg the real part, we obtain

I
@ oo

x"(t): A D J, (|rV*(pr) cos (r" * ma1 * mar)t (4.41)

m:-@
From Eq. (a.a1) we see that the spectrum of x"(t) consists of four categories: (l) the carrier line;
(2) sideband lines at a, * na, due to one tone alone; (3) sideband lines at a" * me2 due to the
other tone alone; and (4) sideband lines at w" * na1 * ma2 due to the nonlinear property of angle
modulation.

4.1I':, ,.a t**e=mod. iffi[*' 6&.fts niltdulated signal xr(t) is [see Bq, (4.23)]
x"(t): I cos (wJ + B sin w*t)
When P << l, we have NB angle modulation.
(a; Find the 'oftfris $B angle-modulated signal.
(b) .Compare the rqsrqlt,with that of a tone-modulated AM signal.
,,r,.:::,,,:r,(s},:.Biqcuss &$i i#i{ffilaftd diftfe*ees by drawing their phasor representations.

(a) x"(t) -,4 cos (u"t * p sin a*t)

- A cos w"t cos (p sin a^t) -,,4 sin w"t sin (p sin a*t)
when p << l, we can write
cos (p sin rr,rrt) = I

sin (p sin a*t) x p sin u*t

Then the NB signal can be approximated by

xNs"(/) =I cos a"t - PA sin u*t sin urt

BA OA-
*
= .,4 cos e"t - 7 cos (a, - a^)t + cos (o, u*)t (4.42)
1,
Note that Eq. (4.42) also can be easily obtained from'Eq. (4.16) by letting 6(t) : p sin a*t.
 lrrtrg ,rd Dfgitrl Ct*
From Eq. @.a2) we see that the spectrum of x*u"(l) consists of a carrier line and a pair of
side lines at w" a em.
(b) The preceding result is almost identical to the situation for a tone-modulated AM signal given
by (see Solved Problem 3.7)

xerr.l(/) : I cos urt * 1tA cos a*t cos t,trt

:I uAcos (o..
* r--- +
u,A
* ,,,;-)t (4.43)
cos ,,t,.t - ar,)t
il-
cos (,r;,.

where p is the modulation index forAM.

Comparison of Eqs. (4.42) and (4.43) shows that the main difference between NB angle
modulation and AM is the phase reversal of the lower sideband component.
(c) By using Eq. (4.35) and
eiasin''l xl + j P sin it,,t for P <.1
Eq. (a.a2) can be written in phasor form as
' x,va"(/) : Re[l ej'J (l + j 0 sin cu,r)]

: Relurt*'' f, * !"i*',,t C
22 -
(4.44)
Ilt "-
Similarly, Eq. (a.a3) can be written in phasor form as
xav(r) - Re lAei'J 0+ pcos c,,,,/)]

ur,.,, h * L"i*., - L,
- RelL122 (4.4s)

By taking thq term Aei'J as the ref-

erence, the phasor representations of L
Eqs. (4,44) and (4.45) are shown in 2
Fig. 4.10. From Fig. 4.10, the differ-
1
ence befween Eqs. (4.44) and (4.45) \a
is obvious. In NB angle modulation, 2
It
D
1
the modulation is added in quadra- (a) NBFM wave (b) AM wave
ture with the carrier, which results Fig. 4.10 Phase representation
in phase fluctuation with little ampli-
tude change. In the AM case, the modulation is added in phase with the carrier, producing
amplitude fluctuation with no phase deviation.

Bandwidth of Angle-Modulated Signals

4.13 Given the angle-modulated signal
x,(tl'* I0,cos(221108 r + 200 cos 2r103 r) : '

what is its bandwidth?

 O*gtt_Or*nr, ;* 1

The instantaneous frequency is

ai:2r(lot) - 4r(los) sin 2zr(lo3)r
So Lu - 4r(l0s), e*: 2r(103), and
L' +zrtlo])
:2oo
IJ: uJm- 2zr(10')
By Eq. (4.25),
Ws: 2(3 + l)a*: 8.04r(105) radls

Since p >> | , W, = 2 L,,t - 8z'( 105) radls or fs :400 kHz

4.14 A Z0-megahertz (MHz) carrier is frequency-modulated by sinusoidal signal such that the maxi-
mum frequency deviation is 100 kHz. Determine the modulation index and the approximate
bandwidth of the FM signal if the frequency of the modulating signal is (a)l kHz, (b) 100 kl7z,
and (c) 500 kHz.
' Af: 100 kLlz,f,:20 MHz >> f,,
For sinusoidal modulation, 0 : LJlk.
(a) Withl,
- I kHz, 0 : 100. This is a WBFM signal, andfu x 2L1- 200 kH,
(b) With, - 100 kHz, 0: l. Thus, by Eq. (4.25),

fs = 2(0 + l)f^: 400 kHz

(c) Withf* - 500 kLlz, 0:0.2. This is an NBFM signal, andf, x 2f,,: 1000 kHz:1 MHz.

4.15 Consider an angle-modulated sigual

, xc(r)
- l0 cos (uU + 3 sin i,.,,r)
Assrung PM and$ == I &tfz. Calculate the modulation index and find the bandwidth when (a) "f*
is doubled and (b)f*is decreased by bne-half.

rpr,a(/)
- A coslu.,..t * kom(r)] : l0 cos (a,t * 3 sin r.,.,,,/)

Thus, m(t)
- a- sin a*t, and
xprra(/): l0 cos (c,.r./ * kra^sinu*t)
From Eq. (a.2\ or Eq. (4.23),
B - koa^: 3

We see that the value of B is independent of f*. By Eq. (4.25),whenf*

- I l<112,

.ft: 2(0 + lV*: 8 kHz

(a) When f^ it doubled, 0 : 3,.f* : 2 kHz, and
fn : 2(3 + l)2 - 16 kHz
(b) When k it decreased by one-hatf, 0 : 3, f,,:0.5 kHz, and
fr:2(3 + 1) (0.5) :41<I1z
fZ.181l Arolog ond Digitol Co**rrirotion,

4.16 Repeat Solved Problem 4.15 when FM is assumed.

xrrur(/) : I cos * O, [-,- m(^)r^] : l0 cos (,.,t,t t 3 sin u^t)

[r",
Thus, m(t) : am cos u*t and

x.1a(/): 10.o, .*rin r,r]

[r.,
From Eq. @.21) or Eq. (4.23),
o^k-f arkf
o
u-- ',nk-f 2

% --if.- znlo\-'-
We see that the value of p is inversely proportional to f*. Thus, by Eq. (4.25), whenf,, I l.Jlz,
-
(a) When h is double4 0 :312,f* - 2kHz, and

.fa:2(g t- r)f,,:
"12),11 *rl , : l0 kHz
(b) Whenf, is decreased by one-half, 13 - 6,f^:0.5 kHz, and
.fa:2(0 + l)f*:2(6 + l) (0.5) :7 kLlz

4.17 Acarrier is frequency-modulated with a sinusoidal signal of 2 kHz, resulting in a maximum

frequency deviation of 5 kHz.
(a) Find the bandwidth of the modulated signal.
(b) The amplitude of the modulating sinusoid is increased by a factor of 3, and its frequency
is lowered to 1 kHz. Find the maximum frequency devjation and the bflq+,ryt4$,of,,tho,new
modulated signal.

(a) From Eq. (4.21),

o
k-fo* Lf_ s(t03)
,- - - 2.5
% .f_ z1to3;
By Eq. (4.25), the bandwidth is

fa: 2(0 + l)f*: 2(2.5 + l)2 : l4kuz

(b) Let /rb" the new modulation index. Then

gr: ry, : 6kf" :6p:6(2.5)- 15

iu* u)m

Thus, Lf : 0rf^, _ (15)(1) _ 15 kHz

.fa : 2(gr + 1).f.,, : 2(15 + I ) (l) 32k1lz

 Angle Modulation

4.18 Iu addition to Carsonb rule (4.27),the f\ltowing formula is often used to estimate the bandwidth
of an FM signal:

WBx2(D*2)w, for Daz

where aM: 2trfu and-fuis the highest frequency of the signal in hertz. Compute the bandwidth,
using this formula, and compare it to the bandwidth, using Carson's rule for the FM signal with
Lf : 7 5 ktlz andfu 15 kHz.
-
Note that commercial FM broadcast stations in the United States are limited to a maximum
frequency deviation of 7 5 kHz, and modulating frequencies typically cover 50 Hz to I 5 kHz.
:
Using Eq. @.26) with a M 2rfu, where f, :
I 5 kHz, we have

Lf 75(103) _ J.
"lt-^ ----:-
f, - ts1to31 -
and by using the given formula, the bandwidth is

J's : 2(D + 2)fu : 210 kHz

Using Carson's rule, Eq. (4.27), we see that the bandwidth is

-fn:2(D + l)fu: 180 kHz

Note: High-quality FM radios require a bandwidth of at least 200 kHz. Thus, it seems that Car-
son's rule underestimates the bandu,idth.

Generation of Angle-modulated Signats

4.19 Consider the frequency multiplier of Fig. 4.4 and an NBFM signal
. xy"na(/) : I cos (wrt * p sin u^t)
with rg < 0"5 wdf"* 200 kHz. Letf*range from 50 Hz to 15 khlz,and let the maximum fre-
quency deviation A/at the output be 75 kHz. Find the required frequency multiplication n and
the maximum allowed frequency deviation at the input.

From Eq. @.22), 0 : Lf/f.. Thus,

n
rmrn
75(103) _<
_J D
tJ^u*: jfrj
75(103)
: l5oo
l5(l 03)
lf Pt - 0.5, where p, is the input lj, then the required frequency multiplication is

n* fu - l5oo :3ooo
o, 0.5
The maximum allowed frequency deviation at the input, denoted A1, is

Lf
Lfr:
' n - qs
3000
:25H2.
f 4ro_l Analog and Digital Communications

4.20 A block diagram of an indirect (Armstrong) FM transmitter is shown in Fig. 4.11. Compute ttre
maximum frequency deviation A/of the output of the FM transmitter and the carrier frequency
f,if ft 200 kllz,fro - 10.8 MHz, Lft:25 Hz, n1 - 64, andn, - 48.
-
Frequency
multiplier

Fig. 4.1I tstock diagram of an indirect FM transmitter

Lf -- ( f,) (n,,) (nr) : (2s) (64) (48) IIz :76.8 kHz

.fz: nrft -- $4) (200) (10') : 12.8(1 061Hz: 12.8 MHz
. fzt.e MHz
, Jz:fz+fro: (12.9 + 10.9) (10.)' Hz: 1_
12.0 MHz
Thus, whenf : 23.6 MHz, then
f" : nrfr: (48) (23.6) : I I 32.8 }dHz
Whenf, -- 2MHz then
.f" : ,r.f, - (48) (2) : 96 MHz

4.21 Design an Armstrong indirect FM modulator to generate an FM carrier with a carrier frequency
of 96 MHz and A/- 20 VJtz. A narrowband FM generator withf : 200 kHz and adjustable A/
in the range of 9 to 10 Hz is available. We also have an oscillator with adjustable frequency,in
the range of 9 to 10 MHz there is a bandpass filter with any center frequency, aud only frequency
doublers are available.

From Fig. 4.11, we see that

Lf : nrn2 Lf1, where Lf : 20 kHz

since 9 < Lft< lo + 9 < 2o'ooo < lo

fttflz
Since only doublers are available, therefore nrnrhas to be power of 2. By trial and error, ftrflz:
2048, nr: 64, nz:32
The output of the first multiplier is 64 x 200 kHz: 12.8 MHz

lnput to the second multiplier has to be 5:q : 3 MHz, for which the local oscillator
32
frequencyf,o must be-fro
- 12.8 - 3- 9.8 MHz which is in the given range.

4.22 Inan Armstrong-type FM generator of Fig. 4.11 (Solved Problem 4.20), the crystal oscillator
frequency is 200 kHz. The maximum phase deviation is limited to 0.2 to avoid distortion.Letf^
range from 50 Hz to 15 kHz. The carrier frequency at the output is 108 i:NdH:z, and the maxirnum
frequency deviation is 75 kHz. Select multiplier and mixer oscillator frequencies.
 Referring to Fig. 4.11, we have
Lf, : 0f* : (0.2) (50) _ t0 Hz

!{Lft -ry:7500 10
:n1ft2

fz: ntft: nr(2) (105; Uz

Assuming down conversion, we have

r r -f,
Jz-Jto n2

f" 7s00(2)(10') - r 0811 06; 1392 L

Thus, -fLo: nJr - nz - n2 n2
( 10") Hz

Letting frz: 150, we obtain

nr : 50 and fro : 9.28 MHz

4,23 Agiven 4ngle-modtdated signat has a maximum frequency deviation of 50 Hz for an input sinu-
,. soid o,f : iffil*,: ency of l20 Hz. Determine the required frequency multiplica-
:

tion factor n to produce a maximum frequency deviation of 20 kHz when the input sinusoid has
unit amptitude and a frequency of 240 Hz andthe angle modulation used is (a) PM and (b) FM.

(a) From Eqs. (4.21) and (4.22) we see that in sinusoidal PM, the maximum frequency deviation
A/ is proportio nal to f^. Thus,

\f,J' : {'19)(50) : loo Hz

l.120]' ',

Lf., 20(103) __
Hence, 2OO
Lfr 100
(b) Again from Eqs. (4.21) and (4.22) we see that in sinusoidal FM, the maximum frequency
deviation A/is independent of f*. Thus,

n __
Lf, _ 20(103)
:400
Lfl 50

4.24 At low carrier frequencies it may be possible to generate

an FM signal by varying the capacitance of a parallel reso-
nant circuit. Show that the output x.(t) of the tuned circuit
shown in Fig. 4.12 is an FM signal if the capacitance has a
time dependence of the form

C(t) * go km(t)
Fig. 4.!2
-
and <<1
l*^,1
| 4Jf1 Anrlog orrt Digitol Corrunirotion,

If we assume km(t) is small and slowly varying, then the output frequency u, of the oscillator is
given by

Since l(klC)m(t)l << l, we can use the approximatton

(r- z)-tt2-r*),

and obtain u)i (l * 12co

I x u)c[, 9*r,l) : a, + krm(t)
l

.i,:: and k,:l''k

where
' Jtco ' 2 co

Thus, by Eq. (4.12),x.(r) is an FM signal.

4,25 ArFM signal

,rna(r)
-,4 cos * o,
[-*,In(A)d.\]
[r",
is applied to the system shourn in Fig. 4.13 consisting of a high-pas$.RC tjlter and an envelope
detector. Assume that wRC << I in the frequency band occupied by xr*(r). Determine the
output signaly(r), assuming that ktlm(t)l I u" for all r.

The frequency response H(r) of the RChigh-pass c

filter is

H(a\- R jwRC
\ / R+U(juC)- l+iaRC
If aRC < < l, then Fig. 4.13

H(a) x juRC
Since multiplication by jw rn the frequency domain is equivalent to differentiation in the time
domain [see Eq. (1.23)], the output v(r) of the RC filter is

v(t) x
^, *[xrr(r)]
- -ARClu, t k m(t)lsin * k,
[-* m(\d^)
[r"r
The corresponding envelope detector ou@ut is

y(t)-ARClu"-krm(t)l
which shows that, except for a dc term ARCI", the output is proportional to m(t).
 Gi;1
4,25 Delay lines might be used to approximate the derivative of the signal by reatizing that
x(t) * x,{t - 1)
x\t) x (4.46)
T
Draw the qystem, and suggo$t ho'w srnall r must be in order for the right side to be a good
approximation of the derivative.

A system to realizeEq. @.a6) is shown in Fig. 4.14.

y(t): ltrtO -x(t-r)l

T
Taking the Fourier transform of both sides yields
ll
:lX(r)
Y(w) - s-)ar X@)l: :-XGo)(l - e't*''',
- TT
lf ar << l, then I - e-)" = jwr and
' Y(u) - iwX(a)
which indicates that y(t) is approximately
equal to the derivative of x(r) and r must
satisff the following condition:

r, . - -
I I
a w,*La Fig. 4.74

,t;27, Congidffffi E$$$1ffi*i,r .:..;,,.

. t - 7t m(X)dX)
..1
rrpr(t): A cos * kr J _*
t:
[u,'ct

Let t, and t, (t, > /r) denote the times associated with two adjacent zero crossings of xor(r)
Gig. a.ls). If
nt.
I " m(^) dA x m(t) (tr- /r) tt < t < tz
JI,

tt tz

Fig. 4.15 Zero crossings of an FM signaI

| 4,24 | Analog and Digital Communications

then show that k,m(t)=

* -u)c
where L,t-tz-tt.
Let xrv(/):Acos0(t)
where 0(t) : a"t * tt, m(\) il,
I' _
Let t, and t, (tz > r, ) be the time associated with two adjacent zero crossings; that is,
xrrra(/r):xpy(r/:0
Then 0(tr) - 0(tr1: r - w,(t, - t) + b I,,' m(\) d)
The bandwidth of the message m(t) is assumed much less than the bandwidth of the modulated
signal. T'hen m(t) is essentially constant over the interval ltr, trl and we have

lr, * ktm(t)l (t, - tr): n

Thus, by Eq. (4.12),

ai: e, + k m(t):
;i
krm(t):
* -
u)c

where A,t tz - tr.

-
4.28 The result of Solve Pfo,bleru 4.27 indicates that rc(r) can be rEcovered by eounting tlrc zero
crossings xr*(r). Let N denote the number of zero crossings in time 1'. Show that if I satisfies:
in
the con<dition

l'r<<+
t*r(t) * *N *f"J'
2T
Let tp t2, t3,...... denote the times of zero crossings and 7, - tz - tp T2 - t3 - t2, .... Assume
that there are N zero crossings in
T--Tr*Tz+... *T*
From the result of Solved Problem 4.27, we have
TI
ksmlt):
' - irc
=
ll

Tr:
a, * krm(t)
. Angle Modulation f 4r5l
,

This is true for 72,73,...; that is,

Thus

Hence, we obtain

or !*(tl-km(t)L
21 2T -t
The condition I lf" < 7 ensures that within T there will be some zero crossings, and the condition
T << llf,
offers no excessive averaging (or smoothing) of m(t).

4.29 A supe*-eterodyne receiver has been designed to receive radio stations,,in the frequency band
108 MHz to 1.57 MHz.
(a) If/o is chosen to be 12lfrlz, then show that the image frequency band overlaps the RF

(b) Determine the minimum required/* such that the image frequencies fall outside the I08 to
157 MHz region. Assume that fro <f"
(a) fLu: 108 - 12 : 96MHz fin : 96 - 12 : 84 MHz
{ 157 - 12 : 145 MHz .fi*z: 145 - 12 : 133 MHz
J -
Loz

Cleariy it is seen that the image band and RF band overlaps.

(b) f,*t: 108 - 2fw
For fu ) 0,'fi*r is always less than 108 MHz, so it never overlaps the RF band.
fi,n2: 157 - zfw
To avoid overlap,f,*2 < 108
+ ts7 - 2J; < 108
+ 2fu > 49 + fw > 24.5 MHz

4.1 An angle-modulated signal is given by nonlinear device -is an ideal square-law

device with input-output characteristics
x,(t) - 5 cos l2n(106;r + 0.2 cos 200nt)
eo(t): oez,1t1
Can you identify whether x.(r) is a PM or an
FM signal? I----- ----------l
t'
fAns. No. It can be either a PM or an FM
I

,,(t) i -l Nontin""rl ,,(0 I :- I iNtl -

signal.l i loeviceffi
i*-
4.2 The frequency multiplier is a nonlinear r!
device followed by a bandpass filter,
l-
- - - - - - - - -l

as shown in Fig. 4.16. Suppose that the , Fi9.4.16

IEq.II Analog ond Digital Communications

Find the output y(t) if the input is an FM deviation of 20 kHz when the modulating
signal given by sinusoid has unit amplitude and a fiequency
of 200 Hz.
ei(t)
-,4 cos (w,t * p sin w*t)
lAns. n - 8001
lAns. y(t) -A/ cos (2u,t + 20 sin w*t), 4.5 A block diagram of a typical FM receiver,
: I covering the broadcast range of 88 to
where A' o*.This result indicates
108 MHz, is shown in Fig. 4.18. The IF
that a ,Orrurr-r1* device can be used as a amplifier frequency is 10.7 MHz. The limi-
frequency doubler.] ter is used to remove the amplitude fluctua-
4.3 Assume that the 10.8 MHz signal in tions caused by channel imperfection. The
Fig. 4.17 is derived from the 200-kHz oscil- FM receiver is tuned to a carrier frequency
lator (multiplication by 5\
and that the of 100 MHz,
200-kElz oscillator drift is 0.1 Hz. (a) A 10 Hz audio signal frequency modu-
(a) Find the drift in the 10.8 MHz signal. lates a 100 MHz carrier, producing p
(b) Find the drift in the carrier of the result- 0.2. Find the bandwidths required for
-
ing FM signal. the RF and IF amplifiers and for the
'Wnt.(a) + 5.4 Hz, @) a8 Hzl audio amplifier.
4.4 A given FM signal has a maximum frequency (b) Repeat (a) if p 5.
deviation of 25 Hz for a modulating sinusoid -
lAns. (a) RF and IF amplifiers: 24kHz;
ofunit amplitude and a frequency of 100 kIIz. audio amplifier: l0 kHz (b) RF and IF
Find the required value of frequency multi- amplifiers: 120 k}{z, audio amplifier:
plication n to produce a maximum frequency l0 kHz.l

Fig. 4.17

Loudspeaker

Fig. 4.18 FM receiver