The Hong Kong Polytechnic University EIE331: Communication Fundamentals

File name: eie331 04fm.pdf

ANGLE MODULATION:
FREQUENCY (FM) and PHASE (PM) MODULATIONS
• Basic definitions
• Narrow-band and wide-band frequency modulations
• Transmission bandwidth of angle modulated signals
• Phase-locked loop (PLL)
• Generation and demodulation of angle modulated signals
• FM stereo multiplexing

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 1

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

References for angle modulation:

Haykin: Section 3.10, pp. 154–193
Lecture notes
Tutorial notes

Angle modulation: Carrier angle is varied according to the slowly-varying

message signal
An important feature of angle modulation:
• It can provide a better discrimination (robustness) against noise and
interference than AM
• This improvement is achieved at the expense of increased
transmission bandwidth
• In case of angle modulation, channel bandwidth may be exchanged
for improved noise performance
• Such trade-off is not possible with AM

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 2

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

BASIC DEFINITIONS
Relationship between the angle and frequency of a sinusoidal signal
Sinusoidal carrier
c(t) = Ac cos[θi(t)]
Angle of carrier
θi(t) [rad]
Instantaneous frequency of carrier

1 1 dθi(t) 1
fi(t) = ωi(t) = = θ̇i(t) [Hz]
2π 2π dt 2π
In the case of an unmodulated carrier, the angle becomes

θi(t) = 2πfct + φc

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 3

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Phase modulation (PM)

Sinusoidal carrier: c(t) = a(t) cos[θi(t)]

PM: The angle θi(t) of carrier is varied linearly with the message signal m(t)

θi(t) = 2πfct + kpm(t)

Amplitude of carrier is constant: a(t) = Ac

Phase-modulated waveform

s(t) = Ac cos[2πfct + kpm(t)]

where • fc denotes the carrier frequency (i.e., frequency of unmodulated signal)

• kp is the phase sensitivity of the PM modulator expressed in radians
per volt
• It is assumed that the angle of unmodulated carrier is zero at t = 0

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 4

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Frequency modulation (FM)

Sinusoidal carrier: c(t) = a(t) cos[θi(t)]

The instantaneous frequency fi(t) of carrier is varied linearly with the message
signal m(t)
1 dθi(t) 1
= θ̇i(t) = fc + kf m(t)
2π dt 2π
Angle of carrier
Z t Z t
θi(t) = 2π fi(τ )dτ = 2πfct + 2πkf m(τ )dτ
0 0

Amplitude of carrier is constant:

a(t) = Ac

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 5

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Frequency-modulated waveform
· Z t ¸
s(t) = Ac cos 2πfct + 2πkf m(τ )dτ
0

where • fc denotes the carrier frequency (i.e., frequency of unmodulated signal)

• kf is the frequency sensitivity of the FM modulator expressed in Hertz
per volt
• It is assumed that the angle of unmodulated carrier is zero at t = 0

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 6

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

PM and FM signal in the time domain produced by a single tone message signal

Note: • The similarity between the angle modulated signals

• Amplitude of angle modulated signals is constant

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 7

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

MOST IMPORTANT PROPERTIES OF PM and FM SIGNALS – Part I

PM signal in the time domain:

sP M (t) = Ac cos[2πfct + kpm(t)]

FM signal in the time domain:

· Z t ¸
sF M (t) = Ac cos 2πfct + 2πkf m(τ )dτ
0

1. Amplitude of PM and FM signals is constant

2. Because the information is carried by the angle of carrier, a nonlinear
operation (also nonlinear distortion) that preserves the angle has no influence
on the angle modulation systems (i.e., it does not cause distortion).
Consequently, even a hard limiter may be used to fix the amplitude of a
PM or FM signal

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 8

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

MOST IMPORTANT PROPERTIES OF PM & FM SIGNALS – Part II

h Rt i
sP M (t) = Ac cos[2πfct + kpm(t)] sF M (t) = Ac cos 2πfct + 2πkf 0
m(τ )dτ

3. A close relationship exists between the PM and FM signals:


 

  
Phase
modulator FM wave    Frequency
modulator PM wave

 

   

 

kp
kp = 2πkf kf = 2π

Conclusions:
• A PM/FM modulator may be used to generate an FM/PM waveform
• FM is much more frequently used than PM
• All the properties of a PM signal may be deduced from that of an FM signal
• Henceforth, in the remaining part of our studies we deal only with FM
signals

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 9

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

FREQUENCY MODULATION

Goal: Determine the spectrum and transmission bandwidth of an FM signal

h Rt i
FM signal sF M (t) = Ac cos 2πfct + 2πkf 0
m(τ )dτ is a nonlinear function of
message signal m(t)

Angle modulations (including FM and PM) are nonlinear modulation processes

Consequently, spectrum of FM signal may not be determined in the frequency

domain using Fourier transform

Empirical approach is required to determine the spectrum and transmission

bandwidth of FM signal where the following single-tone sinusoidal message
signal is considered
m(t) = Am cos(2πfmt)

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 10

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Definition of frequency deviation and modulation index

Let m(t) = Am cos(2πfmt) denote the single-tone message (modulating) signal
Then the instantaneous frequency of FM signal becomes

fi(t) = fc + kf Am cos(2πfmt) = fc + ∆f cos(2πfmt) (1)

In (1), ∆f = kf Am is the frequency deviation, representing the maximum

departure of instantaneous frequency of FM signal from the carrier frequency fc
Angle of FM signal is
Z t
∆f
θi(t) = 2π fi(τ )dτ = 2πfct + sin(2πfmt) = 2πfct + β sin(2πfmt) (2)
0 fm

In (2), β = ∆f /fm is the modulation index, representing the maximum

departure of angle of FM signal from angle 2πfct of unmodulated carrier

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 11

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Units of frequency deviation and modulation index

∆f
∆f = kf Am [Hz] β= [rad]
fm

Recall the relationship between FM and PM signals

FM: s = (t) = sF M (t) = Ac cos[2πfct + β sin(2πfmt)]
PM: sP M (t) = Ac cos[2πfct+kpAm cos(2πfmt)] =⇒ !! β ⇔ kpAm

Two cases are distinguished:

kf A m
• Narrow-band FM, for which β = fm << 1 rad
kf Am
• Wide-band FM, for which β = fm >> 1 rad

The spectrum of narrow- and wide-band FM signals are completely different

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 12

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Spectrum of narrow-band FM signal

Narrow-band FM signal in the time domain is

s(t) = Ac cos[2πfct + β sin(2πfmt)] where β << 1 rad

Using a trigonometric identity cos(α + β) = cos α cos β − sin α sin β we get

s(t) = Ac cos(2πfct) cos[β sin(2πfmt)] − Ac sin(2πfct) sin[β sin(2πfmt)] (3)

If the modulation index is much smaller than 1 radian then

Since lim {cos α} = 1, =⇒ cos[β sin(2πfmt)] ≈ 1

α→0

Since lim {sin α} = α, =⇒ sin[β sin(2πfmt)] ≈ β sin(2πfmt)

α→0

Hence, (3) simplifies to

s(t) ≈ Ac cos(2πfct) − βAc sin(2πfct) sin(2πfmt)

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 13

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

From the equation developed on the previous transparency

s(t) ≈ Ac cos(2πfct) − βAc sin(2πfct) sin(2πfmt)

we get using the trigonometric identity sin α sin β = 21 [cos(α − β) − cos(α + β)]

1
s(t) = sF M (t) ≈ Ac cos(2πfct)+ βAc {cos[2π(fc + fm)t] − cos[2π(fc − fm)t]}
2
Recall, the single-tone modulated AM signal in the time domain is
1
sAM (t) ≈ Ac cos(2πfct) + µAc {cos[2π(fc + fm)t] + cos[2π(fc − fm)t]}
2

Comparing sF M (t) and sAM (t) we conclude:

• If β = µ then the only difference between AM and narrow-band
FM signals is that in the latter the sign of the lower side frequency
component is reversed

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 14

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Spectrum and transmission bandwidth of narrow-band FM signal

Spectrum of narrow-band FM and AM signals
Narrow-band FM: AM:
SF M (f ) = F {sF M (t)} SAM (f ) = F {sAM (t)}
   

 

 





       






Transmission bandwidth of narrow-band FM is equal to that of AM:

BT = 2W

where W is the bandwidth of the low-pass message signal

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 15

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Spectrum of wide-band FM signal

Equation giving the FM signal in the time domain was developed on page 13,
see (3) on page 13

For convenience, that equation is repeated here:

s(t) = Ac cos(2πfct) cos[β sin(2πfmt)] − Ac sin(2πfct) sin[β sin(2πfmt)] (4)

Recall, only periodic signals may be expanded as a trigonometric Fourier series.

Fourier series and, consequently, periodic signals have a discrete spectrum

Even though s(t) is not necessarily periodic, the terms cos[β sin(2πfmt)] and
sin[β sin(2πfmt)] are periodic and each can be expanded as a Fourier series

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 16

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

It has been shown in applied mathematics that

∞
X
cos[β sin(2πfmt)] = J0(β) + 2Jn(β) cos(2πnfmt) (5)
n even
∞
X
sin[β sin(2πfmt)] = 2Jn(β) sin(2πnfmt) (6)
n odd

where n is positive, and the Bessel functions of the first kind, of order n and
argument β are defined as
Z π
1
Jn(β) = exp[j(β sin λ − nλ)]dλ
2π −π

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 17

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Substituting (5) and (6) (see page 17) into (4) (given on page 16) and expanding
products of sine and cosine functions using trigonometric identities finally yields
the FM signal for arbitrary modulation index β in the time domain
s(t) =AcJ0(β) cos(2πfct)
X
∞
+ AcJn(β)[cos(2πfc + n2πfm)t − cos(2πfc − n2πfm)t]
n odd

X
∞
+ AcJn(β)[cos(2πfc + n2πfm)t + cos(2πfc − n2πfm)t]
n even

This equation is valid for both the narrow-band and wide-band FM signals
Recall the Fourier transform of a cosine signal with infinite duration is
1
F {cos(2πfct)} = [δ(f − fc) + δ(f + fc)]
2

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 18

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

To get the spectrum of an FM signal, the Fourier transform of

s(t) =AcJ0(β) cos(2πfct)
X
∞
+ AcJn(β)[cos(2πfc + n2πfm)t − cos(2πfc − n2πfm)t]
n odd

X
∞
+ AcJn(β)[cos(2πfc + n2πfm)t + cos(2πfc − n2πfm)t]
n even

has to be calculated
Observe
• Single-tone modulated FM signal has a discrete spectrum
• Spectrum contains a carrier-frequency line plus an infinite number of sideband
lines =⇒ Distortion-less transmission requires infinite bandwidth!
• All lines are equally spaced by the modulating frequency
• Odd-order lower sideband lines are reversed in phase

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 19

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Properties of Bessel functions of the first kind

1. For small values of modulation index β the following approximations may be
used (Recall, this is the case of narrow-band FM)
J0(β) = 1 and J1(β) = β2 and Jn(β) = 0, n > 2
2. Bessel functions Jn(β) as a function of modulation index β

Note: At certain values of Jn(β), the carrier disappears

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 20

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Spectrum of a single-tone FM signal

Conditions: • Single-tone sinusoidal modulation m(t)
• Spectra are normalized with respect to the carrier amplitude
• Magnitude of spectra is shown only for positive frequencies

m(t): Frequency fixed, amplitude increased m(t): Amplitude fixed, frequency decreased

β = 1.0 β = 1.0

β = 2.0 β = 2.0

β = 5.0 β = 5.0

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 21

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

TRANSMISSION BANDWIDTH OF FM SIGNAL

In theory, an FM signal contains an infinite number of side frequencies =⇒
Bandwidth required for distortion-free transmission is infinite in extent whether
or not the message is band-limited
But implemented FM systems using finite bandwidth do exist and perform well
Explanation: Amplitude of side frequencies decays if we move away from
the carrier frequency and sufficiently far away from the carrier the spectral
components becomes negligible
Experiments showed that if the amplitude of side frequency components is
1 % then a distortion may not be noticed, if the amplitude of side frequency
components is 10 % then a small but noticeable distortion exists
By definition: The transmission bandwidth of an FM signal is the separation
between the two frequencies beyond which none of the side frequencies is
greater than 1 % of carrier amplitude obtained when the modulation is removed

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 22

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Determination of transmission bandwidth of a single-tone FM signal

In practice, the frequency deviation ∆f is fixed
Carson’s rule µ ¶
1
BT ≈ 2(∆f + fm) = 2∆f 1+
β
Easy to use, but Carson’s rule somewhat underestimates the bandwidth
requirement of an FM system

Exact 1 % bandwidth of
an FM signal
BT
(β)
∆f

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 23

The Hong Kong Polytechnic University EIE331: Communication Fundamentals

Transmission bandwidth of an arbitrary modulating signal

Let W denote the highest frequency component of the spectrum of message

signal m(t). In case of a low-pass modulating signal, W is equal to the
bandwidth of m(t)

Let D denote the deviation ratio that is defined as the ratio of maximum
possible frequency deviation to W . Recall, in built FM systems the frequency
deviation ∆f is fixed

Then the bandwidth of FM signal may be estimated by the Carson’s rule

changed according to the parameters of the arbitrary message signal
µ ¶
1
BT = 2∆f 1 +
D

Géza KOLUMBÁN — Dept. of Electronic and Information Engineering 24