Analog Communications Unit 4

Frequency Modulation
Frequency modulation is that form of angle modulation in which the instantaneous
frequency fi(t) is varied linearly with the message signal m(t) as shown by
fi(t) = fc + kf m(t) (1)
the term fc represents the frequency of the unmodulated carrier , and the constant kf
represents the frequency sensitivity of the modulator expressed in hertz per volt .This
assumes that m(t) is a voltage waveform.
Integrating eq.1 with respect to time and multiplying the result by 2π ,we
get

θ (t)=  2πf c t + 2πk f ∫ m(t )dt 

t
(2)
 0

The frequency modulated wave is therefore described in time domain as

s1 (t ) = Ac cos 2πf c t + 2πk f ∫ m(t )dt 

t
(3)
 0

Where

k f = nk1 (2)

FM can be generated by first integrating m(t) and then using the result as the input to the
phase modulator as shown in fig.1

Fig 1 Generation of FM Wave using Phase Modulator

Analog Communications Unit 4

Fig 1 a)Modulating wave b)Frequency modulated wave c)Derivative of m(t)

d)Phase modulated wave

Fig.1a shows a sinusoidal modulating wave m(t) with 2 full cycles and FM wave
produced by this modulating wave is shown in fig.1b.fig.1c shows the derivative
dm(t)/dt,which consists of original sinusoidal modulating wave shifted in phase by 900
and the desired PM plot is shown in fig.1d.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Single tone frequency modulation

Consider then a sinusoidal modulation wave defined by

m(t ) = Am cos(2πf m t ) (1)

The instantaneous frequency of the resulting FM wave equals

f i (t ) = f c + k f Am cos(2πf m t )
(2)
= f c + ∆f cos(2πf m t )

Where

∆f = k f Am (3)

The quantity is called the frequency deviation, representing the maximum departure of
the instantaneous frequency of the FM wave from the carrier frequency. A fundamental
characteristic of an FM wave is that the frequency deviation is proportional to the
amplitude of the modulating wave and is independent of the modulation frequency.
Using Eq 2, the regular argument of the FM wave is obtained as

t
θ (t ) = 2π ∫ f i (t )dt
0
(4)
∆f
= 2πf c t + sin (2πf m t )
fm

The ratio of the frequency deviation ∆f to the modulation frequency f m is commonly

called the modulation index of the FM wave. We denote it by β, so that it can be written

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

∆f
β= (5)
fm

And

θ (t ) = 2πf c t + β sin (2πf m t ) (6)

From Eq 6 we see that, in a physical sense, the parameter represents the phase deviation
of the FM wave: that is, the maximum departure of the angular argument θ (t ) from the
angle 2πf c t of the unmodulated carrier

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Spectrum Analysis of sinusoidal FM wave

The FM wave for sinusoidal modulation is given by

s (t ) = Ac cos[2πf c t + β sin 2πf m t ] (1)

Using a well-known trigonometric identity, we may expand this relation as

s (t ) = Ac cos(2πf c t ) cos[β sin( 2πf m t )] − Ac sin( 2πf c t ) sin [β sin( 2πf m t )] (2)

From this expanded from, we see that the in-phase and quadrature components of the FM
waves s(t) for the case of sinusoidal modulation are as follows.

s1 (t ) = Ac cos[β sin 2πf m t ] (3)

s Q (t ) = Ac sin [β sin 2πf m t ] (4)

Hence, the complex envelope of the FM wave equals

s1 (t ) = s1 (t ) + js Q (t )
(5)
= Ac exp[Jβ sin (2πf m t )]

The complex envelope s(t) retains complete information about the modulation process.
Indeed, we may readily express the FM wave s(t) in terms of the complex envelope s(t)
by writing

s (t ) = Re[ Ac exp( j 2πf c t ) + jβ sin( 2πf m t )]

= Re[s (t ) exp( j 2πf c t )] (6)

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

From Eq 5, the complex envelope is a periodic function of time, with a fundamental

frequency equal to the modulation frequency f m .Therefore it can be expanded as the
form of a complex Fourier series as follows

∞
s (t ) = ∑c
n = −∞
n exp( j 2πnf m t ) (7)

Where the complex Fourier coefficient Cn equals

− 12 f
cn = f m ∫
−1
2 fm
m
s (t ) exp(− j 2πnf m t ) dt
(8)
− 12 f
= fm A c∫−1
2 fm
m
exp[ jβ sin( 2πnf m t ) − − j 2 π n f m t ] dt

For convenience, we define the variable

x = 2πf m t (9)

In terms of which we may rewrite Eq 8 as

Ac π
cn =
2π ∫ π exp[ j(β sin x − nx )] dx
−
(10)

The integral on the right side of Eq 10 is recognized as the nth order Bessel function of
the first kind and argument β . This function is commonly denoted by the symbol Jn(β)
that is

1 π
cn =
2π ∫ π exp[ j(β sin x − nx )] dx
−
(11)

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Hence, we may rewrite Eq 10 as

c n = Ac j n (β ) (12)

Substituting Eq 12 in 7 , we get, in terms of the Bessel function Jn(β), the following

expansion for the complex envelope of the FM wave

∞
s (t ) = Ac ∑ j (β ) exp( j 2πnf t )
n = −∞
n m (13)

Next, substituting Eq 13 in eq.6, we get

 ∞ 
s (t ) = Ac Re  ∑ j n (β ) exp[ j 2π ( f c + nf m )t ] (14)
n =−∞ 

Interchanging the order of summation and evaluating the real part of the right side of Eq
14, we get

 ∞ 
s (t ) = Ac  ∑ j n (β ) cos[2π ( f c + nf m )t ] (15)
 n= −∞ 

This is the desired from for the Fourier series representation of the single-stone FM wave
s(t) for an arbitrary value of . The discrete spectrum of s(t) is obtained by taking the
Fourier transforms of both sides of Eq (15)thus
∞
S(f)=Ac ∑ j (β ) [∆( ( f − f
n = −∞
n c − nf m ) +∆( ( f + f c + nf m ) ] (16)

Fig.1 shows the plot between the Bessel function versus the modulation index for n 0, 1,
2, 3, 4. These plots show that for fixed n, alternates between positive and negative values

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

for increasing and that approaches zero as approaches infinity. Note also that for fixed
we have

J-n(β) = Jn(β), n even

J-n(β) = - Jn(β) n odd
Accordingly, we need only plot or tabulate for positive values of order n.

Fig 1 Plots of Bessels function of first kind

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Narrow-Band FM

For small values of the modulation index compared to one radian, the FM wave assumes
a narrow-band form consisting essentially of a carrier, an upper side-frequency
component, and a lower side-frequency component.

Fig 1 plot of Bessel function

This property follows from the face that for small values of β , we have

j0 (β ) ≅ 1
β
j0 (β ) ≅ (1)
2
j0 (β ) ≅ 0, n >1

The approximations indicated in Eqs 1 are closely justified for values of the modulation
index defined byβ ≤ 0.3 rad. Thus, substituting Eqs 1 in Eq.2
 ∞ 
s (t ) = Ac  ∑ j n (β ) cos[2π ( f c + nf m )t ] (2)
 n= −∞ 

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

We get
βAc β Ac
s (t ) ≅ Ac cos(2πf c t ) + cos[2π ( f c + f m )t ] − cos[2π ( f c − f m )t ] (3)
2 2
This equation shows that for small β, the FM wave s(t) may be closely approximated by
the sum of a carrier of amplitude Ac, an upper side-band frequency component of
amplitudeβAc/2 and a lower side-frequency component of amplitude βAc/2and phase-
shift equal to 180o. An FM wave so characterized is said to be narrow-band.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Transmission bandwidth of FM waves

An FM wave contains an infinite number of side-frequencies so that the bandwidth

required to transmit such a signal is similarly infinite in extent. In practice, however, we
find that the FM wave is effectively limited to a finite number of significant side-
frequencies compatible with a specified amount of distortion. We may therefore specify
an effective bandwidth required for the transmission of an FM wave. Consider first the
case of an FM wave generated by a single-tone modulating wave of frequency fm. In such
an Fm wave, the side-frequencies that are separated from the carrier frequency fc by an
amount greater than the frequency deviation ∆f decrease rapidly toward zero, as that the
bandwidth always exceeds the total frequency excursion, but nevertheless is limited.
Specifically, for large values of the modulation index β, the bandwidth approaches, and is
only slightly greater than the total frequency excursion 2∆f . On the other hand, for small
values of the modulation index , the spectrum of the FM wave is effectively limited to the
carrier frequency and one pair of side-frequencies at f c ± f m , so that the bandwidth

approaches 2 f m . We may thus define an approximate rule for the transmission bandwidth

of an FM wave generated by a single-tone modulating wave of frequency f m as

 1
B ≅ 2∆f + 2 f m = 2∆f 1 +  (1)
 β

This relation is known as Carson’s rule.

Table shows the total number of significant side-frequencies (including both the upper
and lower side-frequencies) for different values of β, calculated on the 1% basis just
explained. The transmission bandwidth B calculated using this procedure can be
presented in the form of a universal curve by normalizing it with respect to the frequency
deviation ∆f , and then plotting it versus β. This curve is shown in Fig.1 , which is drawn
as a best fit through the set of points obtained by using Table 1 in Fig.1 we note that as

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

the modulation index β is increased, the bandwidth occupied by the significant side-
frequencies drops toward that over which the carrier frequency actually deviates. This
means that small values of the modulation index β are relatively more extravagant in
transmission bandwidth than are the larger values of β.

Table-1

G. Kiran Kumar, Assistant Professor, ECE@NITs

www.jntuworld.com www.jwjobs.net

Analog Communications Unit 4

Fig.1Universal curves for evalvting the 99% bandwidth of an FM wave

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

FM Generation

There are essentially two basic methods of generating frequency-modulated

waves, namely, indirect FM and direct FM. In the indirect method of producing
frequency modulation, the modulation wave is first used to produce a narrow-band FM
wave, and frequency multiplication is next used to increase the frequency deviation to the
desired level. On the other hand in the direct method of producing frequency modulation
the carrier frequency is directly varied in accordance with the incoming message signal.
In this subsection, we describe the important features of both methods.

Fig 1 a)Narrow band frequency modulator

Fig.1b)frequency multiplier

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Fig.1c)wide band frequency modulator

Indirect FM :
Consider first the generation of a narrow-band FM wave. To do this, we begin
with the expression for an FM wave s1(t) for the general case of a modulating wave m(t),
which is written in the form

s1 (t ) = A1 cos[2πf1t + φ1t ] (1)

Where f1 is the carrier frequency and A1 is the carrier amplitude. The angular argument
of s1(t) is related to m(t) by

t
φ1 (t ) = 2πk1 ∫ m(t ) dt (2)
0

Where k1 is the frequency sensitivity of the modulator. Provided that the angle is small
compared to one radian for all t, we may use the following approximations:

cos[φ (t )] ≅ 1 (3)
sin [φ (t )] ≅ φ (t ) (4)

Correspondingly, we may approximate Eq 1 as follows

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

s1 (t ) = A1 cos(2πf1t ) − A1 sin (2πf 1t )φ1 (t )

t (5)
= A1 cos(2πf1t ) − 2πk1 A1 sin (2πf1t ) ∫ m(t )dt
0

Eq 5 defines a narrow-band FM wave. Indeed, we may use this equation to set up the
scheme shown in Fig 1a for the generation of a narrow-band FM wave; the scaling factor
2πk1 is taken care of by the product modulator. Moreover, bearing in mind the
relationship that exists between frequency modulation and phase modulation , we see that
the part of the frequency modulator that lies inside the dashed rectangle in Fig 1a

represents a narrow-band modulator of Fig. 1a which differs from an ideals FM wave in

two respects.

1. The envelope contains a residual amplitude modulation and, therefore, varies with
time.
2. For a sinusoidal modulating wave, the phase of the FM wave contains harmonic
distortion in the form of third and higher-order harmonics of the modulation
frequency fm

However, by restricting the modulation index to β ≤ 0.3rad, the effects of residual AM

and harmonic PM are limited to negligible levels.
The next step in the indirect FM method is that of frequency multiplication.
Basically, a frequency multiplier consists of a nonlinear device (e.g. diode or transistor)
followed by a band-pass filter, as in Fig 1b. The nonlinear device is assumed to be
memory less, which means that there is no energy storage. In general, a memory less
nonlinear device is represented by the input-output relation.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

s 2 (t ) = a1 s 2 (t ) + a1 s12 (t ) + − − − − − − + a n s1n (t ) (6)

Where a1,a2,----an are constant coefficients. Substituting Eq 5 in 6 expanding and then

collecting terms, we find that the output s2(t) has a dc component and n frequency-
modulated waves with carrier frequency f1,2f1-------nf1 and frequency deviations
∆f1, 2∆f 1, − − − − − n∆f 1, respectively. The value of is determined by the frequency

sensitivity k1 of the narrow-band frequency modulator and the maximum amplitude of the
modulating wave m(t). The motivation for using the band-pass filter in Fig 1b.
Specifically, the filter is designed with two aims in mind.

1. To pass the FM wave centered at the carrier frequency nf1 and with frequency
deviation n ∆f1,

2. To suppress all other FM spectra

Thus, connecting the narrow-band frequency modulator and the frequency multiplier
as depicted in Fig 1c, we may generate a wideband FM wave s(t) with carrier
frequency f c , = nf 1 and frequency deviation, as desired. Specifically, we may write

s1 (t ) = Ac cos 2πf c t + 2πk f ∫ m(t )dt 

t
(7)
 0

Where

k f = nk1 (8)

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

In other words, the wideband frequency modulator of Fig 1c has a frequency sensitivity n
times that of the narrow-band frequency modulator of Fig 1a, where n is the frequency
multiplication ratio. In Fig 1c shows a crystal-controlled oscillator as the source of
carrier; this is done for frequency stability.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Direct FM

Direct method of FM generation:

In the direct method of FM generation, the instantaneous frequency of the
carrier wave is varied directly in accordance with the message signal by means of a
device known as voltage-controlled oscillator. One way of implementing such a device is
to use a sinusoidal oscillator having a relatively high-Q frequency-determining network
and to control the oscillator by incremental variation of the reactive components. An
example of this scheme is shown in Fig 1, showing a Hartley oscillator. We assume that
the capacitive component of the frequency-determining network consists of a fixed
capacitor shunted by a voltage or varicap, is one whose capacitance depends on the
voltage applied across its electrodes. The variable –voltage capacitance may be obtained,
for example, by using a p-n junction diode that is biased in the reverse direction; the
larger the reverse voltage applied to such a diode, the smaller the transition capacitance
of the diode.

Fig 1 Hartely oscillator for the generation Direct FM

The frequency of oscillation of the Hartley oscillator of Fig 1 is given by

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Lecture Notes 34

1
f1 (t ) = (1)
2π ( L1 + L2 )C (t )

Where C(t) is the total capacitance of the fixed capacitor and the variable-voltage
capacitor, and L1 and L2 are the two inductances in the frequency determining network.
Assume that for a modulating wave m(t) the capacitance C(t) is expressed as follows

C (t ) = C 0 + k c m(t ) (2)

Where C0 is the total capacitance in the absence of modulation, and Kc is the variable
capacitor’s sensitivity to voltage change. Substituting Eq 2 in 1, we get

−1 2
 k 
f1 (t ) = f 0 1 − c m(t ) (3)
 C0 

Where fo is the unmodulated frequency of oscillation.

1
f0 = (4)
2π C 0 (L1 + L2 )

Provide that the maximum change in capacitance produced by the modulating wave is
small compared with the unmodulated capacitance C0, we may approximate Eq 3 as
follows

−1 2
 k 
f1 (t ) = f 0 1 + c m(t ) (5)
 2C 0 

Define

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

f 0kc
kf = (6)
2C 0

We then obtain the following relation for the instantaneous frequency of the oscillator

f1 (t ) = f 0 + k f m(t ) (7)

Where Kf is the resultant frequency sensitivity of the modulator, defined by Eq 6

An FM transmitter using the direct method as described herein, however,
has the disadvantage that the carrier frequency is to obtained from a highly stable
oscillator. It is
therefore necessary, in practice, to provide some auxiliary means by which a very stable
frequency generated by a crystal will be able to control the carrier frequency. One
method of effecting this control is illustrated in fig.2. The output of the FM generator is
applied to a mixer together with the output of a crystal-controlled oscillator, and the
difference frequency term is extracted. The mixer output is next applied to a frequency
discriminator and then low-pass filtered. A frequency discriminator is a device whose
output voltage has an instantaneous amplitude that is proportional to the instantaneous
frequency frequency of the FM wave applied to its input; this device is described later in
the section. When the FM transmitter has exactly the correct carrier frequency, the low-
pass filter output is zero. However, deviations of the transmitter carrier frequency from
its assigned value will cause the frequency discriminator-filter combination to develop a
dc output voltage with a polarity determined by the sense of the transmitter frequency
drift. This dc voltage, after suitable amplification, is applied to the voltage-controlled
oscillator of the FM transmitter in such a way as to modify frequency of the oscillator in
a direction that tends to restore the carrier frequency to its required value.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Fig 2 A Feed back scheme for the frequency stabilization of a frequency modulator

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Balanced Frequency Discriminator

Considering an idealized form of the circuit shown in fig.1 we introduce the notion of an
ideal slope circuit that is characterized by a purely imaginary transfer function, varying
linearly with frequency inside a prescribed interval. Such a circuit includes the
differentiator as a special case.

Fig 1 Ideal balanced frequency descriminator

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Fig 2 a)balanced frequency descriminator b)Frequency response

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Fig 3 a)Frequency response of ideal slope circuit

b)Frequency response of filter
c)ideal slope circuit

To be specific, consider the transfer function depicted in Fig 3a, which is defined by

  B B B
 j 2πa f − f c + , fc − ≤ f ≤ fc +
  2 2 2
  B B B
H 1 ( f ) =  j 2πa f + f c + , − fc − ≤ f ≤ − fc + (1)
  2 2 2
0 elsewhere



Where a is a constant. We wish to evaluate the response of this slope circuit, denoted by
s1(t), for an input FM signal s(t) of carrier frequency fc and transmission bandwidth B. It
is assumed that the spectrum of s(t) is essentially zero outside the frequency band . For
evaluation of the response s1(t), which involves replacing the slope circuit with an

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

equivalent low-pass filter and driving this filter with the complex envelope of the input
FM wave s(t)
Let H1(f) denote the complex transfer function of the slope circuit defined by Fig
3 a. This complex transfer function is related to H1(f) by

H 1 ( f − f c ) = H 1 ( f ), f >0 (2)

Hence, using Eqs 1 and 2, we get

  B B B
 j 2πa f c + , − ≤ f ≤
  2 2 2
H 1 ( f ) = 0, elsewhere (3)




Which is shown in Fig 3 b

The incoming FM wave s(t) is defined by Eq, which is reproduced here for
convenience:

s (t ) = Ac cos 2πf c t + 2πk f ∫ m(t ) dt 

t
(4)
 0

The complex envelope of this FM wave is

s (t ) = Ac exp i 2πk f ∫ m(t ) dt 

t
(5)
 0

Let s1(t) denote the complex envelope of the response of the slope circuit defined
by Fig 3a. Then we may express the Fourier transform of s1(t) as

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

s (t ) = H 1 ( f )S ( f )
  B B B
 j 2πa  f +  S ( f ), ≤ f ≤ (6)
=  2 2 2
0
 elsewhere

Where S(f) is the Fourier transform of s(t). The multiplication of the Fourier transform of
a signal by the factor is equivalent to differentiating the signal in the time domain. We
thus deduce from Eq 6 that

 ds (t ) 
s1 (t ) = a  + jπB s (t ) (7)
 dt 

Substituting Eq 5 in 7, we get

The response of the slope circuit is therefore

s1 (t ) = Re[s1 (t ) exp ( j 2πf c t )]

 2k f   t π (8)
= πBa Ac 1 +
 B
m(t ) cos  2πf c t + 2πk f
 
∫0
m(t ) dt + 
2

The signal s2(t) is a hybrid-modulated wave in which both the amplitude and frequency of
the carrier wave vary with the message signal m(t). However, provided that we choose

2k f
m(t ) < 1
B

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

for all t, then we may use an envelope detector to recover the amplitude variations and
thus, except for a bias term, obtain the original message signal. The resulting envelope
detector output is therefore.

 2k f 
s1 (t ) = πBaAc  m(t ) (9)
 B 

The bias term πBaAc in the right side of Eq.9 is proportional the slope a of the

transfer function of the slope circuit. This suggests that the bias may be removed by
subtracting from the envelope detector output s1 (t ) the output of a second envelope

detector preceded by the complementary slope circuit with a transfer function H2(f) as
described in Fig 3 c. That is, the respective complex transfer functions of the two slope
circuits are related by

H 2 ( f ) = H 1 (− f ) (10)

Let s2(t) denote the response of the complementary slope circuit produced by the
incoming FM wave s(t). Then, following a procedure similar to that described herein, we
find that the envelope of s2(t) is

 2k f 
s1 (t ) = πBaAc 1 − m(t ) (11)
 B 

Where s2(t) is the complex envelope of the signal s2(t). The difference between the two
envelope in Eqs 9 and 11 is

s 0 (t ) = s1 (t ) − s 2 (t )
(12)
= 4πk1 aAc m(t )

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Which is free from bias

We may thus model the ideal frequency discriminator as a pair of slope circuits
with their complex transfer functions related by Eq10, followed by envelope detectors
and a summer, as in Fig 1. This scheme is called a balanced frequency discriminator of
back-to-back frequency detector.The idealized scheme of Fig1 can be closely realized
using the circuit shown in Fig 2a. The upper and lower resonant filter sections of this
circuit are tuned to frequencies above and below the unmodulated carrier frequency fc,
respectively. In Fig 2b we have plotted the amplitude response of the these two tuned
filters, together with their total response, assuming that both filters have a high-Q filter.
The linearity of the useful portion of this total response, centered at fc, is determined by
the separation of the two resonant frequencies. As illustrated in Fig 2b, a frequency
separation of 3B0 gives satisfactory results, where 2B0 is the 3-dB bandwidth of either
filter. However, there will be distortion in the output of this frequency discriminator due
to the following factors:

1. The spectrum of the input FM wave s(t) is not exactly zero for frequencies outside
the range
2. The tuned filter outputs are not strictly band-limited, and so some distortion is
introduced by the low-pass RC filters following the diodes in the envelope
detectors.
3. The tuned filter characteristics are not linear over the whole frequency band of the
input FM wave s(t).

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Zero-crossing Detector

The detector exploits the property that the instantaneous frequency of an FM wave is
approximately given by

1
fi ≅ (1)
2∆t

Where ∆t is the difference between adjacent zero crossings of the FM wave, as

illustrated in Fig 2. Considering an interval T chosen in accordance with the following
two conditions:

1. The interval T is small compared to the reciprocal of the message band-width W

2. The interval T is large compared to the reciprocal of the carrier frequency fc of
the FM wave

Condition 1 means that the message signal m(t) is essentially constant inside the interval
T. Condition 2 ensures that a reasonable number of zero crossing of the FM wave occurs
inside the interval T. The FM waveform shown in Fig 2 illustrates these two conditions.
Let no denote the number of zero crossings inside the interval T. We may then express the
time ∆t between adjacent zero crossing as

T
∆t = (2)
n0

Fig 1 block diagram Zero crossing detector

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Fig 2 FM waveform

Hence, we may rewrite Eq.1

n0
fi ≅ (3)
2t
Since, by definition, the instantaneous frequency is linearly related to the message signal
m(t), we see from Eq 3 that m(t) can be recovered from a knowledge of no. Fig 1 is the
block diagram of a simplified from of the zero-crossing detector based on this principle.
The limiter produces a square-wave version of the input FM wave, the limiting of FM
waves. The pulse generator produces short pulses at the positive-going as well as
negative-going edges of the limiter output. Finally, the integrator performs the averaging
over the interval T as indicated in Eq 3, thereby reproducing the original message signal
m(t) at its output.

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

PHASE LOCKED LOOP

The phased-locked loop (PLL) is a negative feedback system that

consists of three major components: a multiplier, a loop filter, and a voltage controlled
oscillator (VCO) connected together in the form of a feedback loop, as in Fig 1.
The VCO is a sine-wave generator whose frequency is determined by a
voltage applied to it from an external source. In effect any frequency modulator may
serve as a VCO
Initially adjust the VCO so that when the control voltage is zero, two conditions
are satisfied:
(1) the frequency of the VCO is precisely set at the unmodulated carrier frequency fc,
(2) the VCO output has a 90o phase-shift with respect to the unmodulated carrier
wave. Suppose that the input signal applied to the phase-locked loop is an FM
wave defined by
s(t)=Ac sin (2∏fct + φ1 (t ) ) (1)

Fig 1 Phase locked loop

Where Ac is the carrier amplitude. With a modulating wave m(t), we have

t
φ1 (t ) = 2πk f ∫ m(t ) dt (2)
0

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

Where kf is the frequency sensitivity of the frequency modulator. Let the VCO output be
defined by

r (t ) = Av cos[2πf c t + φ 2 (t )] (3)

Where Av is the amplitude. With a control voltage v(t) applied to the VCO input, we have

t
φ 2 (t ) = 2πk v ∫ v(t ) dt (4)
0

Where kv is the frequency sensitivity of the VCO, measured in hertz per volt. The
incoming FM wave s(t) and the VCO output r(t) are applied to the multiplier, producing
two components:

1. A high-frequency component represented by

k m Ac Av sin [4πf c t + φ1 (t ) + φ 2 (t )] (5)

2. A low-frequency component represented by k m Ac Av sin [φ1 (t ) − φ 2 (t ) ] where km is

the multiplier gain, measured in volt-1

The high-frequency component is eliminated by the low-pass action of the filter and the
VCO. Therefore, discarding the high-frequency component, the input to the loop filter is
given by

e(t ) = k m Ac Av sin [φ e (t ) ] (6)

Where φe (t ) is the phase error defined by

G. Kiran Kumar, Assistant Professor, ECE@NITs

Analog Communications Unit 4

φ e (t ) = φ1 (t ) − φ 2 (t )
t (7)
= φ1 (t ) − 2πk v ∫ v(t ) dt
0

The loop filter operates on its input e(t) to produce the output

∞
v(t ) = ∫ e(τ )h(t − τ ) dτ (8)
−∞

Where h(t) is the impulse response of the filter,

Using Eqs 6 through 8 to relate φe (t ) and φ1 (t ) differentiating with respect to time,
we obtain
∞
d/dt [ φe (t ) ]=d/dt[ φ1 (t )] − 2πk 0 ∫ −∞
sin φ e (τ )h(t − τ ) dτ (9)

Where K0 is a loop parameter defined by

K 0 = k m k v Ac Av (10)
Eq.9 suggests the representation or model of Fig. In this model we have also included the
relationship between v(t) and e(t) as represented by Eqs and . We see that the block
diagram of the model resembles Fig. The multiplier is replaced by a subtractor and a
sinusoidal nonlinearity, and the VCO by an integrator
The loop parameter K0 plays an important role in the operation of a phase-locked
loop. It has the dimensions of frequency; this follows from Eq, where we observe that the
amplitudes Ac and Am are both measured in volts and the multiplier gain km is measured
in volt-1

G. Kiran Kumar, Assistant Professor, ECE@NITs