Chapter 2 Continuous-Wave

Modulation

“Analog Modulation” is the subject concerned

in this chapter.
 2.1 Introduction
 Analog communication system
 The most common carrier is the sinusoidal wave.

(Analog)

Carrier wave
2.1 Introduction
 Modulation
 A process by which some characteristic of a carrier
is varied in accordance with a modulating wave
(baseband signal).
 Sinusoidal Continuous-Wave (CW) modulation
 Amplitude modulation
 Angle modulation
 Sinusoidal carrier

2.1 Introduction
Baseband signal

Amplitude Modulation

Frequency Modulation
2.2 Double-Sideband with Carrier or simply
Amplitude Modulation
Carrier c(t )  Ac cos(2f c )
Baseband m(t )
Modulated Signal s(t )  Ac [1  ka m(t )] cos(2f ct ),
where ka is amplitude sensitivity or modulation index
 Two required conditions on amplitude sensitivity
 1 + ka m(t)  0, which is ensured by |ka m(t)| ≦ 1.
 The case of |ka m(t)| > 1 is called overmodulation.
 The value of |ka m(t)| is sometimes represented by “percentage”
(because it is limited by 1), and is named (|ka m(t)|100)%
modulation.
 fc >> W, where W is the message bandwidth.
 Violation of this condition will cause nonvisualized envelope.
2.2 Overmodulation

overmodulaion
| ka m(t ) | 1
2.2 Example of Non-Visualized Envelope
2.2 Example of Visualized Envelope
 2.2 Transmission Bandwidth
s(t )  Ac [1  ka m(t )] cos(2f ct )

 S ( f )   ( f  f c )   ( f  f c )  M ( f  f c )  M ( f  f c )
Ac ka Ac
2 2
Transmission bandwidth BT = 2W.

2W
2.2 Transmission Bandwidth
 Transmission bandwidth of an AM wave
 For positive frequencies, the highest frequency
component of the AM wave equals fc + W, and the
lowest frequency component equals fc – W.
 The difference between these two frequencies
defines the transmission bandwidth BT for an AM
wave.
2.2 Transmission Bandwidth
 The condition of fc > W ensures that the sidebands
do not overlap.
2.2 Negative Frequency
 Operational meaning of “negative frequency”
in spectrum
 If time-domain signal is real-valued, the negative
frequency spectrum is simply a mirror of the
positive frequency spectrum.
 We may then define a one-sided spectrum as
Sone-sided( f )  2S ( f ) for f  0.
 Hence, if only real-valued signal is considered, it is
unnecessary to introduce “negative frequency.”
2.2 Negative Frequency
 So the introduction of negative frequency part is
due to the need of imaginary signal part.
 Signal phase information is embedded in imaginary
signal part of the signal.

mQ(t)
phase
mI(t)
2.2 Negative Frequency
 As a result, the following two spectrums contain the
same frequency components but different phases
(90 degree shift in complex plane).
 ( f  fc )  ( f  fc )
2 cos(2f ct )
 fc fc

 ( f  fc )
2 j sin( 2f ct )   ( f  fc )
 fc fc
2.2 Negative Frequency
 Summary
 Complex-valued baseband signal consists of
information of amplitude and phase; while real-
valued baseband signal only contains amplitude
information.
 One-sided spectrum only bears amplitude
information, while two-sided spectrum (with
negative frequency part) carries also phase
information.
2.2 Virtues of Amplitude Modulation
AM receiver can be implemented in terms of simple
circuit with inexpensive electrical components.
 E.g., AM receiver

v1 (t )  s 2 (t )  Ac2 [1  k a m(t )]2 cos2 ( 2f c t )

Ac2
 [1  k a m(t )]2 [1  cos(4f c t )]
2
2.2 Virtues of Amplitude Modulation
 The bandwidth of m2(t) is twice of m(t). (So to
speak, the bandwidth of m(f)*m(f) is twice of m(f).)

Lowpass filter
2.2 Virtues of Amplitude Modulation
 So if 2fc > 4W,
Ac2
 v2 ( t )  [1  ka m(t )]2
2
Ac
 v3 (t )  [1  ka m(t )]
2
Ac ka
block DC
if m(t ) is zero mean  m(t )
2
By means of a squarer, the receiver can recover the information-
bearing signal without the need of a local carrier.
2.2 Limitations of Amplitude Modulation
(DSB-C)
 Wasteful of power and bandwidth

s(t )  Ac [1  ka m(t )] cos(2f c t )

 Ac cos(2f c t )  ka m(t ) cos(2f c t )
    
withcarrier

Waste of power in the information-less “with-carrier” part.

2.2 Limitations of Amplitude Modulation
 Wasteful of power and bandwidth

Only requires half of bandwidth after modulation

2.3 Linear Modulation
 Definition
 Both sI(t) and sQ(t) in

s(t) = sI(t)cos(2fct) – sQ(t) sin(2fct)

are linear function of m(t).

2.3 Linear Modulation
 For a single real-valued m(t), three types of
modulations can be identified according to
how sQ(t) are linearly related to m(t), at the
case that sI(t) is exactly m(t):
 (Some modulation may have mI(t) and mQ(t) that
respectively bear independent information.)
1. Double SideBand-Suppressed Carrier modulation
(DSB-SC)
2. Single SideBand (SSB) modulation
3. Vestigial SideBand (VSB) modulation
 2.3 DSB-SC and SSB
Type of modulation sI(t) sQ(t)

DSB-SC m(t) 0

SSB m(t) mˆ (t ) Upper side band transmission

SSB m(t)  mˆ (t ) Lower side band transmission

ˆ (t )  Hilbert tr ansform of m(t ) , which is used to completely “suppress” the other sideband
*m

DSB-SC

SSB usb

SSB lsb
2.3 DSB-SC
 Different from DSB-C, DSB-SC s(t) undergoes
a phase reversal whenever m(t) crosses zero.
s(t )  m(t ) cos(2f ct )

Require a receiver that can

recognize the phase reversal!
2.3 Coherent Detection for DSB-SC
 For DSB-SC, we can no longer use the “envelope
detector” (as used for DSB-C), in which no local
carrier is required at the receiver.
 The coherent
detection or
synchronous
demodulation
becomes necessary.
2.3 Coherent Detection for DSB-SC
v(t )  Ac' cos(2f c t   ) s(t )
 Ac Ac' cos(2f c t ) cos(2f c t   )m(t )
1 1
 Ac Ac' cos(4f c t   )m(t )  Ac Ac' cos( )m(t )
2 2
LowPass1
 Ac Ac' cos( )m(t ),
2
provided f c  W .
2.3 Coherent Detection for DSB-SC
 Quadrature null effect of the coherent detector.
 If  = /2 or –/2, the output of coherent detector for DSB-
SC is nullified.
 If  is not equal to either /2 or –/2, the output of coherent
detector for DSB-SC is simply attenuated by a factor of
cos(), if  is a constant, independent of time.
 However, in practice,  often varies with time; therefore, it is
necessary to have an additional mechanism to maintain the
local carrier in the receiver in perfect synchronization with
the local carrier in the transmitter.
 Such an additional mechanism adds the system complexity of
the receiver.
2.3 Costas Receiver for DSB-SC
In-phase coherent detector
 An exemplified
design of
synchronization
mechanism is the
Costas receiver,
where two
coherent
detectors are
used.

Quadrature-phase coherent detector

2.3 Costas Receiver for DSB-SC
 Conceptually, the Costas receiver adjusts the
phase  so that it is close to 0.
 When  drifts away from 0, the Q-channel output
will have the same polarity as the I-channel output
for one direction of phase drift, and opposite
polarity for another direction of phase drift.
 The phase discriminator then adjusts  through the
voltage controlled oscillator.

sin( )
-pi/2 0 pi/2
2.3 Single-Sideband Modulation
 How to generate SSB signal?
 1. Product modulator to generate DSB-SC signal
 2. Band-pass filter to pass only one of the sideband
and suppress the other.
 The above technique may not be applicable to a
DSB-SC signal like below. Why?
filter filter

SSB
2.3 Single-Sideband Modulation
 For the generation of a SSB modulated signal
to be possible, the message spectrum must have
an energy gap centered at the origin.
2.3 Single-Sideband Modulation
 Example of signal with 300 Hz ~ 300 Hz
energy gap
 Voice : A band of 300 Hz to 3100 Hz gives good
articulation
 Also required for SSB modulation is a highly
selective filter
2.3 Single-Sideband Modulation
 Phase synchronization is also an important
issue for SSB demodulation. This can be
achieved by:
 Either a separate low-power pilot carrier
 Or a highly stable local oscillator (for voice
transmission)
 Phase distortion that gives rise to a Donald Duck voice
effect is relatively insensitive to human ear.
2.3 Vestigial Sideband Modulation
 Instead of transmitting only one sideband as
SSB, VSB modulation transmits a partially
suppressed sideband and a vestige of the other
sideband.
SSB

VSB

Still, no information loss in principle.

2.3 Requirements for VSB filter
1. The sum of values of the magnitude response |H(f)| at any
two frequencies equally displaced above and below fc is
unity. I.e., |H(fc  f)| + |H(fc  f)| = 1 for fv < f < fv.
2. H(f  fc) + H(f + fc) = 1 for W < f < W.

ffc f f+fc

fc fc
So the transmission band of VSB filter is BT = W + fv.
2.3 Generation of VSB Signal
 Analysis of VSB
 Give a real baseband signal m(t) of bandwidth W.
 Then, M (  f )  M * ( f ) and M ( f )  0 for | f | W .
 Let MVSB( f )  M ( f )[1  H Q ( f ) / j ] / 2, where

1, f   fv
1 
H Q ( f )   (0,1),  f v < f < 0, and H Q (  f )  H Q* ( f ).
j 0,
 f 0

The filter is denoted by HQ because it is used to generate sQ(t) (cf. Slide 2-23)
 2.3 Generation of VSB Signal
*
1 1 1  1 
LQ ( f )  H Q ( f ) is real.  LQ (  f )  H Q (  f )  H Q* ( f )   H Q ( f )   L*Q ( f )
j j j  j 
  LQ ( f ).
1  1 
LQ ( f )  HQ ( f ) [1  LQ ( f )] / 2  1  H Q ( f )  / 2
j
 j 
1.0

W
2.3 How to recover from VSB signal?
M VSB( f )  M VSB
*
( f )

 M ( f )[1  LQ ( f )]  M  (  f )[1  LQ (  f )]* 

1
2
 M ( f )[1  LQ ( f )]  M ( f )[1  LQ (  f )],
1
2
since [1  LQ (  f )] is real, and M  (  f )  M ( f )

 M ( f )[2  LQ ( f )  LQ (  f )]
1
2
 M ( f ), because LQ (  f )   LQ ( f ).
2.3 VSB upper sideband transmission

sVSB(t )
mVSB(t )
 M L ( f ) FL ( f )  M R ( f ) FR ( f )  [ M L ( f )  M R ( f )][ FL ( f )  FR ( f )]
if M L ( f ) FR ( f )  M R ( f ) FL ( f )  0.

 M ( f  f c )1{ f c  W  f   f c  W }  M * (  f  f c )1{ f c  W  f  f c  W }
1
cont.

2
 [1  LQ ( f  f c )] [1  LQ (  f  f c )] 
  1{ f   f c  W }  1{ f  f c  W }
 2 2 

1
2
M ( f  f c )  M * ( f  f c )

 [1  LQ ( f  f c )] [1  LQ (  f  f c )] 
  1{ f   f c  W }  1{ f  f c  W }
 2 2 
 [1  LQ ( f  f c )] [1  LQ (  f  f c )] 
 sDSB ( f )   1{ f   f c  W }  1{ f  f c  W }
 2 2 
 sDSB ( f ) 
1
[1  LQ ( f  f c )]1{ f  f c  W }  [1  LQ ( f  f c )]1{ f  f c  W }
2
 sDSB ( f )  2  LQ ( f  f c )  LQ (  f  f c )  (See next slide.)
1
2
2
[1  LQ ( f  f c )]1{ f  f c  W }

2
[1  LQ (  f  f c )]1{ f  f c  W }

2 [1  LQ ( f  f c )]1{ f  f c  W }
 [1  LQ (  f  f c )]1{ f  f c  W }

1
LQ ( f  f c )

LQ (  f  f c )

2
LQ ( f  f c )  LQ (  f  f c )  2
Consequently,

sVSB( f )  sDSB( f ) LQ ( f  f c )  LQ (  f  f c )  2 
1
 2
sVSB( f )  sDSB( f ) H ( f )

 H( f ) 
1
LQ ( f  f c )  LQ ( f  f c )  2
2
1
H( f )

1
H ( f  fc )

 H ( f  fc )

1

1
H ( f  fc )  H ( f  fc )
 1
H ( f  fc )  H ( f  fc )

 LQ ( f )  H ( f  f c )  H ( f  f c ) for | f | W
 H Q ( f )  j[ H ( f  f c )  H ( f  f c )] for | f | W
2.3 Mathematical Representation of VSB
signal
M VSB ( f )  M ( f )[1  jH Q ( f )] / 2
1 1
 M ( f )  jM ( f ) H Q ( f )
2 2
1 1
 M ( f )  jM ' ( f )
2 2
where M ' ( f )   M ( f ) H Q ( f )   jM ( f ) LQ ( f ).
Notably, m' (t ) is real. This is an extension of Hilbert Transform.

M ' (  f )   jM (  f ) LQ (  f )  jM * ( f ) LQ ( f )
 (  j )* M * ( f ) L*Q ( f )  [ jM ( f ) LQ ( f )]*  [ M ' ( f )]* .
2.3 Application of VSB Modulation
 Television Signals
1. The video signal exhibits a large bandwidth and
significant low-frequency content.
 Hence, no energy gap exists (SSB becomes
impractical).
 VSB modulation is adopted to save bandwidth.
 Notably, since a rigid control of the transmission VSB
filter at the very high-power transmitter is expensive, a
“not-quite” VSB modulation is used instead (a little
waste of bandwidth to save cost).
2.3 Application of VSB Modulation
VSB Filter for Television Signal Transmissions

55.25 MHz 59.75 MHz

2.3 Application of VSB Modulation
 As the transmission signal is not quite VSB modulated,
the receiver needs to “re-shape” the received signal
before feeding it to a VSB demodulator.
2.3 Application of VSB Modulation
2. In order to save the cost of the receiver (i.e., in
order to use envelope detector at the receiver), an
additional carrier is added.
 Notably, additional carrier does not increase
bandwidth, but just add transmission power.

s(t )  Ac cos(2f c t )  Ac ka m(t ) cos(2f c t )  m' (t ) sin( 2f c t ) 

1
2
 1  1
 Ac 1  ka m(t ) cos(2f c t )  ka Ac m' (t ) sin( 2f c t )
 2  2
2.3 Application of VSB Modulation
 Distortion of envelope detector
2
 1  1
s 2 (t )  Ac2 1  ka m(t ) cos2 ( 2f c t )  ka2 Ac2 ( m' (t )) 2 sin 2 (2f c t )
 2  4
1  1 
 ka Ac2 m' (t ) 1  ka m(t ) sin( 2f c t ) cos(2f c t )
2  2 
 

1 2  1
LowP ass
2
 1 2 
 Ac  1  ka m(t )  ka ( m' (t )) 
2

2  2  4     
 Distortion 
The distortion can be compensated by reducing the amplitude
sensitivity ka or increasing the width of the vestigial sideband. Both
methods are used in the design of Television broadcasting system.
2.3 Extension Usage of DSB-SC
 Quadrature-Carrier Multiplexing or Quadrature
Amplitude Modulation (QAM)

Synchronization is critical in QAM modulation, which is often achieved by a

separate low-power pilot tone outside the passband of the modulated signal.
2.4 Frequency Translation
 The basic operation of SSB modulation is simply a
special case of frequency translation.
 So SSB modulation is sometimes referred to as frequency
changing, mixing, or heterodyning.
 The mixer is a device that consists of a product modulator
followed by a band-pass filter, which is exactly what SSB
modulation does.
2.4 Frequency Translation
 The process is named
upconversion, if f1 + f
is the wanted signal,
and f1 – f is the
unwanted image
signal.
 The process is named
downconversion, if f1
 f is the wanted
signal, and f1 + f is
the unwanted image
signal.
2.5 Frequency-Division Multiplexing
 Multiplexing is a technique to combine a
number of independent signals into a composite
signal suitable for transmission.
 Two conventional multiplexing techniques
 Frequency-Division Multiplexing (FDM)
 Time-Division Multiplexing (TDM)
 Will be discussed in Chapter 3.
2.5 Frequency-Division Multiplexing
Example 2.1
2.6 Angle Modulation
 Angle modulation
 The angle of the carrier is varied in accordance with
the baseband signal.
 Angle modulation provides us with a practical
means of exchanging channel bandwidth for
improved noise performance.
 So to speak, angle modulation can provide better
discrimination against noise and interference than
the amplitude modulation, at the expense of
increased transmission bandwidth.
2.6 Angle Modulation
 Commonly used angle modulation
 Phase modulation (PM)
s(t )  Ac cos[2f ct  k p m(t )], where k p is phase sensitivity.
 Frequency modulation (FM)

 
s(t )  Ac cos 2 0( f c  k f m( ))d
t

 A cos2f t  2k  m( )d ,

t
c c f 0

whe re k f is frequency sensitivity.

2.6 Angle Modulation
 Main differences between Amplitude
Modulation and Angle Modulation
1. Zero crossing spacing of angle modulation no
longer has a perfect regularity as amplitude
modulation does.
2. Angle modulated signal has constant envelope; yet,
the envelope of amplitude modulated signal is
dependent on the message signal.
2.6 Angle Modulation
 Similarity between PM and FM
 PM is simply an FM with 0m( )d in place of m(t).
t

sPM (t )  Ac cos[2f ct  k p m(t )]

sFM (t )  Ac cos 2f ct  2k f  m( )d 

 t

 0 
 Hence, the text only discusses FM in this chapter.
2.7 Frequency Modulation
 s(t) of FM modulation is a non-linear function of m(t).
s(t )  Ac cos 2  f i ( )d  Ac cos 2  ( f c  k f m( )) d 
  
t t

 0   0 

 Ac cos 2f c t  2k f  m( )d 

 t

 0 
 So its general analysis is hard.
 To simplify the analysis, we may assume a single-tone
transmission, where
m(t )  Am cos(2f mt )
 From the formula in the previous slide,
f i ( t )  f c  k f m( t )
 f c  k f Am cos(2f mt )
 f c  f  cos(2f mt )
where f  k f Am is the frequency deviation.
 s(t )  Ac cos2 0 f i ( )d 
 t

 
 Ac cos 2  [ f c  f  cos(2f m )]d 
 t

 0 
 f 
 Ac cos2f c t  sin( 2f m t )
 fm 
where   f / f m is often called the modulation index
of FM signal.
 Modulation index  is the largest deviation from 2fct in
FM system.
s(t )  Ac cos2f ct   sin( 2f mt )
 As a result,

f c  f m  f c  f  fi (t )  f c  f  cos(2f mt )  f c  f  f c  f m

1. A small  corresponds to a narrowband FM.

2. A large  corresponds to a wideband FM.
 2.7 Narrowband Frequency Modulation

s(t )  Ac cos2f c t   sin( 2f m t )

 Ac cos(2f c t ) cos[ sin( 2f m t )]  Ac sin( 2f c t ) sin[  sin( 2f m t )]
 Ac cos(2f c t )  Ac sin( 2f c t )  sin( 2f m t ) (Often,  < 0.3.)
 2.7 Narrowband Frequency Modulation
 Comparison between approximate narrowband
FM modulation and AM (DSB-C) modulation
sFM (t )  Ac cos(2f c t )  Ac sin( 2f ct )  sin( 2f mt )
Ac Ac
 Ac cos(2f c t )  cos(2 ( f c  f m )t )  cos(2 ( f c  f m )t )
2 2
s AM (t )  Ac [1  ka m(t )] cos(2f c t )
 Ac [1  ka Am cos(2f mt )] cos(2f c t )
ka Am k A
 Ac cos(2f c t )  cos(2 ( f  f m )t )  a m cos(2 ( f  f m )t )
2 2
2.7 Narrowband Frequency Modulation
 Represent them in terms of their low-pass
isomorphism.
~ Ac
sFM (t )  ( Ac  j 0)  [cos(2f mt )  j sin( 2f mt )]
2
A
 c [cos(2f mt )  j sin( 2f mt )]
2
~ ka Am
s AM (t )  ( Ac  j 0)  [cos(2f mt )  j sin( 2f mt )]
2
ka Am
 [cos(2f mt )  j sin( 2f mt )]
2
2.7 Narrowband Frequency Modulation
 Phaser diagram

Ac
 [cos(2f mt )  j sin( 2f m t )]
~ 2
sFM (t )

Ac
(cos(2f mt ) Ac
2 [cos( 2f m t )  j sin( 2f m t )]
 j sin( 2f mt )) 2
( Ac  j0)
~
sAM (t )
Let Ac  ka Am .
2.7 Spectrum of Single-Tone FM
Modulation

s(t )  Ac cos2f c t   sin( 2f m t )

 ReAc exp  j2f c t   sin( 2f m t )
 Re~
s (t ) exp( j 2f t )
c

s (t )  Ac exp  j sin( 2f mt )

~ (See Slide 1-247)
 
~
s (t )  Ac  n
J (  ) e j 2f m nt
 n
J (
n  
x ) e jn
 e jx sin( )

n  

where J n () is the nth order Bessel function of the first kind.
2.7 Spectrum of Single-Tone FM
Modulation

 S~( f )  ~

 j 2ft
s ( t ) e dt


  
j 2f m nt   j 2ft
   Ac  J n (  )e e dt

 n   
 
 Ac 
n  
J n (  )  e  j 2 ( f nf m ) t dt



 Ac J
n  
n (  ) ( f  nfm )
2.7 Spectrum of Single-Tone FM
Modulation

Consequently,


1 ~ ~*
S ( f )  S ( f  fc )  S ( f  fc )
2

Ac 
  J n (  ) ( f  f c  nfm )   (  f  f c  nfm )
2 n  

(  ) ( f  f c  nfm )   ( f  f c  nfm )
Ac

2
J
n  
n
 2.7 Spectrum of Single-Tone FM
Modulation
 The power of s(t)
 By definition, the time-average autocorrelation function
is given by:
1 1
T E[s(t )s(t   )]dt  lim 
T T
Rs ( )  lim
T  T  T
s(t ) s(t   )dt
2T 2T
 Hence, the power of s(t) is equal to:
1
T
1 T 2
  
T
Rs (0)  lim s2
( t ) dt  lim Ac
cos 2
2  f c
t   sin( 2f m
t ) dt
T 
2T T 
2T T
1 1  cos  4 f t  2  sin( 2 f t )  A 2

T c
T
 lim
T 
A 2 c m
dt  c

2T 2 2
 2.7 Spectrum of Single-Tone FM
Modulation
 The time-average power spectral density of a
deterministic signal s(t) is given by (cf. Slide 1-117)
1
PSD ( f )  lim
T 
S ( f ) S *
2T
(f)
2T
where S2T ( f ) is the Fourier transform of s(t ) 1| t | T .
 From 
s2T (t )  Ac  J n (  ) cos(2 ( f c  nfm )t )  1| t | T 
n  
we obtain:

S2T ( f )  AcT  J n (  )sinc(2T ( f  f c  nfm ))  sinc(2T ( f  f c  nfm ))
n  
For simplicity, assume that 2T increases along the multiple of
1/fm., i.e., 2T = p/fm, where p is an integer. Also assume that fc is
a multiple of fm, i.e., fc = qfm., where q is an integer. Then
PSD( f )
1
 lim
T 
S ( f ) S *
2 T ( f )
2T
Ac2 
 lim
p 
 J 
k (  )  ( f  f c  kf m )   ( f  f c  kf m )

4 k  

  J n (  )sinc( p( f  f c  nfm ) / f m )  sinc( p( f  f c  nfm ) / f m )
n  

f c  (n  1) f m f c  nfm f c  (n  1) f m
fm
f c  nfm 
p
 Ac2   
PSD( f )  lim   J n (  )sinc( p( f  f c  nfm ) / f m )  J k (  ) ( f  f c  kf m )
4 p 
n   k  
 
 J
n  
n (  )sinc( p( f  f c  nfm ) / f m )  J k (  ) ( f  f c  kf m )
k  
 
 J
n  
n
(  )sinc( p( f  f c  nfm ) / f m )  J k (  ) ( f  f c  kf m )
k  
 

 
n  
J n
(  )sinc( p ( f  f c
 nf m
) / f m  k
)
k  
J (  ) ( f  f c
 kf m 
)

Ac2   2 
   J n (  ) ( f  f c  nfm )   J n (  )J n 2 q (  ) ( f  f c  nfm )
4 n   n  

 

  J n (  )J n2 q (  ) ( f  f c  nfm ) 
n  
 n
n  
J 2
(  ) ( f  f c
 nf m 
)

Ac2   2 
  n J (  )[ ( f  f  nf )   ( f  f  nf )]
4 n   
c m c m
2.7 Average Power of Single-Tone FM
Signal
 Hence, the power of a single-tone FM signal is
given by:
 Ac2   2 

 PSD ( f ) df   n J (  )   J (  ) J n 2 q (  ) 
2  n   
n
n  

Ac2  

  1 
2  n  
 (  1 ) n
J n (  ) J n2 q (  ) 

Ac2

2
 Question: Can we use 2f to be the bandwidth of a
single-tone FM signal?
Example 2.2
 Fix fm and kf,
but vary  = f/fm =
kf Am/fm.
Example 2.2
 Fix Am and kf,
but vary  = f/fm
= kf Am/fm.
2.7 Spectrum of Narrowband Single-Tone
FM Modulation

J 0 ( )  1
 

When  is small,  J 1 (  )  .
 2
 J n (  )  0 for n  2
2.7 Spectrum of Narrowband Single-Tone
FM Modulation
 This results in an approximate spectrum for
narrowband single-tone FM signal spectrum as
Ac2  2
PSD( f )   J n (  ) ( f  f c  nfm )   ( f  f c  nfm )
4 n  
Ac2 2
 J 1 (  ) ( f  f c  f m )   ( f  f c  f m )
4
Ac2 2
 J 0 (  ) ( f  f c )   ( f  f c )
4
Ac2 2
 J 1 (  ) ( f  f c  f m )   ( f  f c  f m )
4
 J n (  )  (1)n J n (  )  J n2 (  )  J 2n (  )
Ac2 2
PSD ( f )  J 1 (  ) ( f  f c  f m )   ( f  f c  f m )
4
Ac2 2
 J 0 (  ) ( f  f c )   ( f  f c )
4
Ac2 2
 J 1 (  ) ( f  f c  f m )   ( f  f c  f m )
4
 2 Ac2
  ( f  f c  f m )   ( f  f c  f m )
16
Ac2
  ( f  f c )   ( f  f c )
4
 2 Ac2
  ( f  f c  f m )   ( f  f c  f m )
16
 Ac2 Ac2
4 4

 2 Ac2  2 Ac2  2 Ac2  2 Ac2

16 16 16 16
fm

 fc fc
2.7 Transmission Bandwidth of FM signals
 Carson’s rule – An empirical bandwidth
 An empirical rule for Transmission Bandwidth of
FM signals
 For large , the bandwidth is essentially 2f.
 For small , the bandwidth is effectively 2fm.
 So Carson proposes that:
 1
BT  2f  2 f m  2f 1  
 
2.7 Transmission Bandwidth of FM signals
 “Universal-curve” transmission bandwidth
 The transmission bandwidth of an FM wave is the
minimum separation between two frequencies
beyond which none of the side frequencies is
greater than 1% of the carrier amplitude obtained
when the modulation is removed.

(  ) ( f  f c  nf m )   ( f  f c  nf m )
A
S( f )  c
2
J
n  
n

Ac cos(2f c t ) 
Ac
 ( f  f c )   ( f  f c )
2
  2nmax BT / f
BT 2nmax f m 2nmax 0.1 2 20.0
 
f f m  0.3 4 13.3
0.5 4 8.0
 For fixed f , a smaller  causes a larger BT . 1.0 6 6.0
2.0 8 4.0
5.0 16 3.2
10.0 28 2.8
20.0 50 2.5
30.0 70 2.3
 2.7 Bandwidth of a General FM wave
 Now suppose m(t) is no longer a single tone but a
general message signal of bandwidth W.
 Hence, the “worst-case” tone is fm = W.
 For nonsinusoidal modulation, the deviation ratio D = f / W is
used instead of the modulation index .
 The derivation ratio D plays the same role for nonsinusoidal
modulation as the modulation index  for the case of
sinusoidal modulation.
 We can then use Carson’s rule or “universal curve” to
determine the transmission bandwidth BT.
2.7 Bandwidth of a General FM wave
 Final notes
 Carson’s rule usually underestimates the
transmission bandwidth.
 Universal curve is too conservative in bandwidth
estimation.
 So, a choice of a transmission bandwidth in-
between is acceptable for most practical purposes.
Example 2.3
 FM radio in North America requires the
maximum frequency derivation f = 75 kHz.
 If some message signal has bandwidth W = 15 kHz,
then the deviation ratio D = f / W = 75/15 = 5.
 Then
 1  1
BT ,Carson  2f 1    2  751    180 kHz
 D  D
2nmax 16
BT ,UniversalCurve  f  75  240 kHz
D 5
Example 2.3
 In practice, a bandwidth of 200 kHz is allocated to
each FM transmitter.
 So Carson’s rule underestimates BT, while
“Universal Curve” overestimates BT.
2.7 Generation of FM Signals
 Direct FM
 Carrier frequency is directly varied in accordance with the
message signal as accomplished using a voltage-controlled
oscillator.
 Indirect FM
 The message is first integrated and sent to a phase modulator.
 So, the carrier frequency is not directly varied in accordance to
the message signal.
2.7 Generation of FM Signals

 m( )d (See next slide.)

2.7 Generation of FM Signals
 Frequency multiplier
 2 1
cos ( x )  2 (cos(2 x )  1)

cos3 ( x )  1 (cos(3x )  3 cos( x ))
 4
 1
cos ( x )  (cos(4 x )  4 cos(2 x )  3 cos( x ))
4

 8
 5 1
cos ( x )  16 (cos(5 x )  5 cos(3x )  10 cos( x ))



2.7 Demodulation of FM Signals
 Indirect Demodulation – Phase-locked loop
 Will be introduced in Section 2.14
 Direct Demodulation
 Balanced frequency discriminator
s1 (t ) |~
s1 (t ) |
~
so (t )
s(t )
differentiation filters

s2 (t ) |~
s2 (t ) |
   BT  BT
 j 2a  f  fc  , | f  f c |
  2  2
  BT  BT
H 1 ( f )   j 2a  f  f c  , | f  f c |
  2  2
 0, elsewhere


  BT  BT
 j 2a  f  f c  2 , | f  f c | 2
  
  B  B
H 2 ( f )    j 2a  f  f c  T , | f  f c | T
  2  2
 0, elsewhere


2.7 Analysis of Direct Demodulation in
terms of Low-Pass Equivalences
 B
~ 2 H 1 ( f  f c ), | f | T
H1 ( f )   2
 0, otherwise
  B  B
 j 4a  f  T , | f | T
  2  2
 0, otherwise
  BT ~ BT
~ 1 ~ ~  j 2a  f   S ( f ), | f |
 S1 ( f )  H1 ( f ) S ( f )    2  2
2  0, elsewhere
 d~
s (t ) 
~
 s1 (t )  a   jBT ~
s ( t )
 dt 
s(t )  Ac cos 2f ct  2k f  m( )d 
 t

 0 

 s (t )  Ac exp j 2k f  m( )d 

 t
~
 0 

 d~s (t ) 
~
s1 (t )  a   jBT ~
s (t )
 dt 

 a  jAc 2k f m(t ) exp j 2k f  m( )d  
 t

  0  

 jBT  Ac exp j 2k f  m( )d  
  t

  0  
 2k f 
m(t ) exp j 2k f  m( )d 
 t
 jBT aAc 1 
 BT   0 
 s1 (t )  Re~
s1 (t ) exp( j 2f c t )
  2k f    
m(t ) exp j 2k f  m( )d exp( j 2f c t ) 
t
 Re  jBT aAc 1 
 B   
 
0
T

 2k f  
m(t ) sin  2f c t  2k f  m( )d 
t
 BT aAc 1 
 BT   0 
 2k f   
m(t ) cos 2f c t  2k f  m( )d  
t
 BT aAc 1 
 BT   0 2
2k f
 If m(t ) < 1 and f c  W , then envelope detector can be used
BT
to obtain the amplitude of the lowpass equivalent message.
~  2k f 
| s1 (t ) | BT aAc 1  m( t ) 
 BT 
Similarly, 2

  BT  BT
~  j 4a  f  , | f |
H2 ( f )    2  2
 0, otherwise

  BT ~ BT
~ 1 ~ ~  j 2a  f   S ( f ), | f |
 S2 ( f )  H 2 ( f ) S ( f )    2  2
2  0, elsewhere

~  d~
s (t ) 
 s2 (t )  a   jBT ~s (t )
 dt 
 2k f 
m(t ) exp j 2k f  m( )d 
 t
 jBT aAc 1 
 BT   0 
 s2 (t )  Re~
s2 (t ) exp( j 2f c t )
 2k f   
m(t ) cos 2f c t  2k f  m( )d  
t
 BT aAc 1 
 BT   0 2

~  2k f 
| s2 (t ) | BT aAc 1  m( t ) 
 BT 
~ s (t ) | ~
o s (t ) |  | ~
1 2 s (t ) | 4k aA m(t )
f c

Final Note: a is a parameter of the filters, which can be used to

adjust the amplitude of the resultant output.
2.7 FM Stereo Multiplexing
 How to do Stereo Transmission in FM radio?
 Two requirements:
 Backward compatible with monophonic radio receivers
 Operate within the allocated FM broadcast channels
 To fulfill these requirements, the baseband message
signal has to be re-made.
 m

fm = 19 kHz

m(t )  [ml (t )  mr (t )]  [ml (t )  mr (t )] cos(4f m t )  K cos(2f m t )

          
For monophonic reception For coherent detection
Demultiplexer in receiver of FM stereo.

m
2.8 Nonlinear Effects in FM Systems
 The channel (including background noise, interference and
circuit imperfection) may introduce non-linear effects on
the transmission signals.
 For example, non-linearity due to amplifiers.
0.8
0.6
 1 3
 i v ( t )  vi (t ), | vi (t ) | 1 0.4
3
 2
0.2

vo ( t )   , vi ( t )  1 0
 3 -0.2
 2
 , vi (t ) < 1 -0.4
 3 -0.6
-0.8
-1.5 -1 -0.5 0 0.5 1 1.5
2.8 Nonlinear Effects in FM Systems
 Suppose

 ov ( t )  a v
1 i ( t )  a v
2 i
2
( t )  a 3
3 i (t )
v

vi (t )  Ac cos2f c t   (t )

 (t )  2k f 0m( )d
t

Then ov ( t )  a v
1 i ( t )  a v
2 i
2
( t )  a 3
3 i (t )
v
 a1 Ac cos[2f c t   (t )]  a1 Ac2 cos2 [2f c t   (t )]
 a1 Ac3 cos3 [2f c t   (t )]
vo (t )  a1 Ac cos[2f c t   (t )]  a2 Ac2 1  cos[4f c t  2 (t )]
1
2
 a3 Ac3 3 cos[2f c t   (t )]  cos[6f c t  3 (t )]
1
4
1  3 
 a1 Ac   a1 Ac  a3 Ac3  cos[2f c t   (t )]
2

2    4         
BT ,Carson  2 f  2W

1 1
 a2 Ac cos[4f c t  2 (t )]  a1 Ac3 cos[6f c t  3 (t )]
2

2          4         
BT ,Carson  4 f  2W BT ,Carson  6 f  2W

Thus, in order to recover s(t) from vo(t) using band-pass filter

(i.e., to remove 2fc and 3fc terms from vo(t)), it requires:
2 f c  (4f  2W ) / 2  f c  (2f  2W ) / 2
 or equivalent ly, f c  3f  2W .
The filtered output is therefore:

 3 
vo ,filtered(t )   a1 Ac  a3 Ac3  cos[2f ct   (t )]
 4 
 Observations
 Unlike AM modulation, FM modulation is not affected
by distortion produced by transmission through a channel
with amplitude nonlinearities.
 So the FM modulation allows the usage of highly non-
linear amplifiers and power transmitters.
2.8 AM-to-PM Conversion
 Although FM modulation is insensitive to amplitude
nonlinearity, it is indeed very sensitive to phase
nonlinearity.
 A common type of phase nonlinearity encountered in
microwave radio transmission is the AM-to-PM conversion.
 The AM-to-PM conversion is owing to that the phase
characteristic of amplifiers (or repeaters) also depends on the
instantaneous amplitude of the input signal.
 Notably, the nonlinear amplifiers discussed previously will
leave the phase of the input unchanged.
 Often, it requires that the peak phase change for a 1-dB
change in input envelope is less than 2%.
2.9 Superheterodyne Receiver
 A commercial radio communication system
contains not only the “transmission” but also
some other functions, such as:
 Carrier-frequency tuning, to select the desired
signals
 Filtering, to separate the desired signal from other
unwanted signals
 Amplifying, to compensate for the loss of signal
power incurred in the course of transmission
2.9 Superheterodyne Receiver
 A superheterodyne receiver or superhet is designed to
facilitate the fulfillment of these functions, especially
the first two.
 It overcomes the difficulty of having to build a tunable highly
selective and variable filter (rather a fixed filter is applied on
IF section).

heterodyning function
2.9 Superheterodyne Receiver
Example AM Radio FM Radio
RF carrier range 0.535-1.605 MHz 88-108 MHz
Midband frequency of IF section 0.455 MHz 10.7 MHz
IF bandwidth 10 kHz 200 kHz

First detector Second detector

2.9 Image Interference
 Fix fIF and fLO at the receiver end. What is the
fRF that will survive at the IF section output?
 Answer: f RF | f LO  f IF |
 Example. Suppose the receiver use 1.105MHz local
oscillator, and receives two RF signals respectively
centered at 0.65MHz and 1.56MHz.
f IF f IF
0.455 MHz 0.455 MHz
f RF f Interference
0.65 MHz
1.56 MHz
f LO
1.105 MHz
2.9 Image Interference
Then at the output of the IF section, we respectively obtain:

f IF  f LO  f RF f Image  f LO  f RF  1.755 MHz

0.455 MHz

f Image Interference  f Interference  f LO  0.455 MHz f Interference  f LO  2.665 MHz

2.9 Image Interference
 A cure of image interference is to employ a highly
selective stages in the RF session in order to favor the
desired signal (at fRF) and discriminate the undesired
signal (at fRF + 2fIF or fRF – 2fIF).
2.9 Advantage of Constant Envelope for FM
modulation
 Observations
 For FM modulation, any variation in amplitude is
caused by noise or interference.
 For FM modulation, the information is resided on
the variations of the instantaneous frequency.
 So we can use an amplitude limiter to remove the
amplitude variation, but to retain the frequency
variation after the IF section.
2.9 Advantage of Constant Envelope for FM
modulation
 Amplitude limiter
 Clipping the modulated wave at the IF section
output (almost to the zero axis) to result in a near-
rectangular wave.
 Pass the rectangular wave through a bandpass filter
centered at fIF to suppress harmonics (due to
clipping).
 Then the filter output retains the frequency
variation with constant amplitude.
2.10 Noise in CW Modulation Systems

 To simplify the system analysis, we assume:

 ideal band-pass filter (that is just wide enough to pass the
modulated signal s(t) without distortion),
 ideal demodulator,
 Gaussian distributed white noise process.
 So the only source of imperfection is from the noise.
2.10 Noise in CW Modulation Systems

BT

So after passing through the ideal bandpass filter, s(t) is

unchanged but w(t) becomes a narrowband noise n(t).
Hence,
x (t )  s(t )  n(t ),
where n(t )  nI (t ) cos(2f ct )  nQ (t ) sin( 2f ct ).
2.10 Noise in CW Modulation Systems
 Input signal-to-noise ratio (SNRI)
 The ratio of the average power of the modulated
signal s(t) to the average power of the filtered noise
n(t).
 Output signal-to-noise ratio (SNRO)
 The ratio of the average power of the demodulated
message signal to the average power of the noise,
measured at the receiver output.
2.10 Noise in CW Modulation Systems
 It is sometimes advantageous to look at the lowpass
equivalent model.
 Channel signal-to-noise ratio (SNRC)
 The ratio of the average power of the modulated
signal s(t) to the average power of the channel noise
in the message bandwidth, measured at the receiver
input (as illustrated below).
2.10 Noise in CW Modulation Systems
 Notes
 SNRC is nothing to do with the receiver structure, but
depends on the channel characteristic and modulation
approach.
 SNRO is however receiver-structure dependent.
 Finally, define the figure of merit for the receiver as:

SNRO
figure of merit 
SNRC
2.11 Noise in Linear Receivers Using
Coherent Detection
 Recall that for AM demodulation
 when the carrier is suppressed, linear coherent
detection is used. (Section 2.11)
 when the carrier is additionally transmitted,
nonlinear envelope detection is used. (Section 2.12)
 The noise analysis of the above two cases are
respectively addressed in Sections 2.11 and
2.12.
2.11 Noise in Linear Receivers Using
Coherent Detection

cos(2f ct )

m(t ) : stationary with zero meam and PSD SM ( f ) bandlimite d to W .

s(t )  CAcm(t ) cos(2f ct )
2.11 Noise in Linear Receivers Using
Coherent Detection
 Average signal power

  
1 T 1 T
 E[ s (t )]dt  lim E C 2 Ac2 cos2 (2f ct )m 2 (t ) dt
2
lim
T  2T T T  2T T

 lim
1
T  2T 
T
T
 
C 2 Ac2 cos2 (2f ct ) E m 2 (t ) dt

1 T
 C A P lim  (2f ct )dt
2 2 2
c cos
T  2T T

C 2 Ac2 P

W
2
where P  E[m (t )]   SM ( f )df is the message power.
2
W
2.11 Noise in Linear Receivers Using
Coherent Detection
 Noise power in the message bandwidth
W W N0
W
S w ( f )df  
W 2
df  WN 0
2.11 Noise in Linear Receivers Using
Coherent Detection

cos(2f ct )

 Channel SNR for DSB-SC and coherent detection (observed at x(t))

C 2 Ac2 P 2 C 2 Ac2 P
SNRC ,DSB-SC  
WN o 2WN o
 Next, we calculate output SNR (observed at y(t)) under the condition
that the transmitter and the receiver are perfectly synchronized.
x (t )  s(t )  w(t )
 CAc m(t ) cos(2f ct )  wI (t ) cos(2f ct )  wQ (t ) sin(2f ct )

 v (t )  x (t ) cos(2f ct )
 [CAc m(t ) cos(2f ct )  wI (t ) cos(2f ct )  wQ (t ) sin(2f ct )] cos(2f c t )
 CAc m(t ) cos2 ( 2f c t )  wI (t ) cos2 (2f ct )  wQ (t ) sin(2f ct ) cos(2f ct )
1
LowPass 1
 CAc m(t )  wI (t )
2 2 (See Slide 1-219.)
1 1
 y (t )  CAc m(t )  wI (t ) (Recall E[w2 (t )]  E[wI2 (t )]  E[wQ2 (t )].)
2 2

 SNRO ,DSB-SC 

E C 2 Ac2m 2 (t ) / 4


C 2 Ac2 P

C 2 Ac2 P
2

E wI (t ) / 4  2
E[ w (t )] 2WN 0
 Figure of merit for DSB - SC and coherent detection  1.

Similar derivation on SSB and coherent detection yields the

same figure of merit.
 Conclusions
 Coherent detection for SSB performs the same as
coherent detection for DSB-SC.
 There is no SNR degradation for SSB and DSB-SC
coherent receivers. The only effect of these modulation
and demodulation processes is to translate the message
signal to a different frequency band to facilitate its
transmission over a band-pass channel.
 No trade-off between noise performance and bandwidth.
This may become a problem when high quality
transceiving is required.
2.12 Noise in AM Receivers Using Envelope
Detection

s(t )  Ac [1  ka m(t )] cos(2f ct )

lim
T 
1
T
T
E [ 2
s ( t )]dt  A2
c
E (1  k a
m ( t 2
)) lim
T 
1 T
T
cos2
( 2f ct )dt
2T 2T
Ac2
 (1  ka2 P ) (Assume m(t ) zero mean.)
2
 N0
Also, W Sw ( f )df  W
W W
df  WN 0
2
 Hence, channel SNR for DSB-C is equal to:
Ac2 (1  ka2 P )
 SNRC ,AM 
2WN o
 Next, we calculate output SNR (observed at y(t)) under the condition that
the transmitter and the receiver are perfectly synchronized.

x (t )  s (t )  n (t )
 Ac [1  ka m(t )] cos(2f c t )  nI (t ) cos(2f c t )  nQ (t ) sin(2f c t )
 [ Ac (1  ka m(t ))  nI (t )] cos(2f c t )  nQ (t ) sin(2f c t )
 envelop detector

x
x

y (t )  x ( t ) 
2
LowP ass

1
 [ Ac (1  ka m(t ))  nI (t )]2  nQ2 (t )
2
1
 [ Ac (1  ka m(t ))  nI (t )] if Ac [1  ka m(t )]  n~(t )
2
(Refer to Slide 2-136 and 2-138.)
block DC
1
 [ Ac ka m(t )  nI (t )]
2
 E Ac2 ka2 m 2 (t ) / 2 Ac2 ka2 P Ac2 ka2 P
 SNRO ,AM   
E nI (t ) / 2
2 2
E[n (t )] 2WN 0

SNRO ,AM Ac2 ka2 P ( 2WN 0 ) ka2 P

  2  <1
SNRC ,AM Ac (1  ka P ) ( 2WN o ) 1  ka P
2 2

Conclusion
 Even if the noise power is small compared to the
average carrier power at the envelope detector output,
the noise performance of a full AM receiver is
inferior to that of a DSB-SC receiver due to the
wastage of transmitter power.
 Example 2.4 Single-Tone Modulation
 Assume m(t )  Am cos(2f mt )
 s(t )  Ac [1  ka Am cos(2f mt )] cos(2f c t )
Hence,
lim
T 
1 T
T
E [ s 2
( t )]dt  A2
c lim
T 
1 T
T
1  k A
a m cos( 2 f m t ) 2
cos2
(2f ct )dt
2T 2T
 Ac lim
2
T 
1 T
T
cos 2
( 2f c
t )  2 k A
a m
cos( 2 f m
t ) cos2
( 2f ct )
2T
 ka2 Am2 cos2 ( 2f mt ) cos2 ( 2f ct ) dt
 1 k 2 2
A  A2

 Ac2   0  a m   c (1  ka2 P )
2 4  2
 2
1 T 2 A
where P  lim   
2 m
T 
Am cos ( 2 f m t ) dt .
T 0
2
 Following similar procedure as previous discussion,
SNRO ,AM ka2 P ka2 Am2 / 2
  .
SNRC ,AM 1  ka P 1  ka Am / 2
2 2 2

So even if for 100% percent modulation (kaAm = 1), the figure of merit
= 1/3. This means that an AM system with envelope detection must
transmit three times as much average power as DSB-SC with coherent
detector to achieve the same quality of noise performance.
 2.12 Threshold Effect
 What if Ac [1  ka m(t )]  n~(t ) is violated (in AM modulation
with envelope detection)?
x (t )  s (t )  n (t )
 Ac [1  ka m(t )] cos(2f c t )  nI (t ) cos(2f c t )  nQ (t ) sin(2f c t )
 [ Ac (1  ka m(t ))  nI (t )] cos(2f c t )  nQ (t ) sin(2f c t )

B | n~ | ( B  nI )2  nQ2  B 2  2nI B  nI2  nQ2  B 2  2nI B  nI2  B  nI

B <<| n~ | ( B  nI )2  nQ2  B 2  2nI B  | n~ |2  B 2  2 | n~ | B  | n~ |2  B  | n~ |
y (t )  x ( t ) 
2
LowP ass Assume Ac [1  ka m(t )] << n~(t ) .
1
 [ Ac (1  ka m(t ))  nI (t )]2  nQ2 (t )
2
1
 Ac2 (1  ka m(t ))2  2nI (t ) Ac (1  ka m(t )) | n~(t ) |2
2
1
 Ac2 (1  ka m(t ))2  2 | n~(t ) | Ac (1  ka m(t )) | n~(t ) |2
2
1
 ( Ac (1  ka m(t )) | n~(t ) |) 2
2
1
 ( Ac (1  ka m(t )) | n~(t ) |)
2
block DC
1
 [ Ac ka m(t ) | n~(t ) |]
2
  Ac ka m(t )  nI (t ), Ac [1  ka m(t )] | n~(t ) |
2 y (t )   ~(t ) |,
 c a
A k m ( t )  | n Ac [1  ka m(t )] <<| n~(t ) |

E Ac2 ka2 m 2 (t ) Ac2 ka2 P Ac2 ka2 P

 SNRO ,AM   
E | n (t ) | 
~ 2
E[nI (t )]  E[nQ (t )] 4WN 0
2 2

SNRO ,AM Ac2 ka2 P ( 4WN 0 ) ka2 P 1

  2  <
SNRC ,AM Ac (1  ka P ) ( 2WN o ) 2(1  ka P ) 2
2 2
2.12 Threshold Effect Ac2 / 2 Ac2
r 
 Threshold effect 2WN 0 4WN 0
 For AM with envelope detection, there exists a
carrier-to-noise ratio r (namely, the power ratio
between unmodulated carrier Ac cos(2fct) and the
passband noise n(t) ) below which the noise
performance of a detector deteriorates rapidly.
2.12 General Formula for SNRO in Envelope
Detection
 For envelope detector, the noise is no longer additive; so the
original definition of SNRO (which is based on additive
noise) may not be well-applied.
 A new definition should be given:
 Definition. The (general) output signal-to-noise ratio
for an output y(t) due to a carrier input is defined as
so2
SNRO 
Var[ y(t )]
where so  E[ y (t )]  E[ yo (t )], and yo (t ) is equal to y(t )
in the presense of noise alone. y(t )  so  yo (t )
2.12 General Formula for SNRO in Envelope
Detection
 so is named the mean output signal.
 Var[y(t)] is named the mean output noise power.
 Example. y(t) = A + nI(t), where nI(t) is zero mean..

 so  E [ A  n I (t )]  E [n I (t )]  A

 Var [ y ( t )]  Var [ n I ( t )]  E [ n 2
I ( t )]

A2
 SNRO 
E[nI2 (t )]
This somehow shows the backward compatibility of the new
definition.
2.12 General Formula for SNRO in Envelope
Detection
 Now for an envelope detector, the output due to a carrier
input and additive Gaussian noise channel is given by:

y (t )  ( A  n I (t )) 2  nQ2 (t )
I0( ) = modified Bessel function
of the first kind of zero order.
 y (t ) is Rician distribute d with pdf

 yo (t )  nI2 (t )  nQ2 (t ) is Rayleigh distribute d with pdf

 Appendix 4 Confluent Hypergeometric
Functions
 The Kummer confluent hypergeometric function is a
solution of Kummer’s equation
d2y dy
x 2  (b  x )  ay  0 for a, b complex
dx dx
with boundary conditions y (0)  1 and y ' (0)  a / b.
 For b  0, 1, 2, …., the Kummer confluent
hypergeometric function is equal to 1F1(a;b;x).
Generalize d hypergeometric function
rr 
(a1 )k (a2 )k  (a p )k x k (a )k  a (a  1)  (a  k  1)
p Fq ( a ; b ; x )    , where  .
k 0 (b1 ) k (b2 ) k  (bq ) k k! (a )0  1
A4.2 Properties of the Confluent
Hypergeometric Function
a
1. 1 F1 ( a; b; x )  1  x as x  0.
b
2. 1 F1 ( 1;1; x )  1  x.
 x  x  x 
3. 1 F1 ( 1 / 2;1; x )  exp      (1  x ) I 0    xI2   
 2  2  2 
x
2 as x  .

 ( m / 2 )   m 1 
 exp( b u ) I 0 (u )du 
m 1
1 F1  ;1; 2 
2 2
4. u 
  2 4b  
m
0 2b
5. exp( u ) 1 F1 ( ;  ; u ) 1 F1 (    ;  ;u )
2.12 General Formula for SNRO in Envelope
Detection
 Hence,
N   u 2

E [ y (t )]  exp(  r )  u exp  
2
 I 0 (u )du
(2r ) 3/ 2 0
 4r 
N ( 3 / 2 )   3 
 exp(  r ) 3 / 2  1 1 
F ;1; r  By Property 4
(2r ) 3/ 2
2( 4 r )   2 
  3 
  N exp(  r )  1 F1  ;1; r 
2  2 
   1 
  N  1 F1   ;1; r  By Property 5
2   2 
2.12 General Formula for SNRO in Envelope
Detection
 As a result,
   1 A2  
so  E[ y (t )]  E[ yo (t )]   N  1 F1   ;1; 2   1.
2   2 2 N  
 Similarly, we can obtain:
  A2    1 A 2

2

Var[ y (t )]  2 N  1 F1   1;1; 2
2
   1 F1   ;1; 2  
  2 N  4  2 2 N  
 
 A2
   1 A2

2

 2 N2 1    1 F1   ;1; 2   By Property 2
 2 N 4   2 2 N 
2
  
2.12 General Formula for SNRO in Envelope
Detection
 This concludes to:
(Continue from the previous slide.)
2.12 General Formula for SNRO in Envelope
Detection
Curve of SNRO 
1 F1  1 / 2;1; r   12
(1  r )  1 F1  1 / 2;1; r 
4 2


and the two limiting approximat es.

1000

100

SNRO,AM  2ka2 Pr
10

1 SNRO,AM  ka2 Pr

0.1
0.1 1 10 100
2.12 General Formula for SNRO in Envelope
Detection
 Remarks
 For large carrier-to-noise ratio r, the envelope detector
behaves like a coherent detector in the sense that the output
SNR is proportional to r.
 For small carrier-to-noise ratio r, the (newly defined) output
signal-to-noise ratio of the envelope detector degrades faster
than a linear function of r (decrease at a rate of r2).
 From “threshold effect” and “general formula for SNRO,” we
can see that the envelope detector favors a strong signal. This
is sometimes called “weak signal suppression.”
2.13 Noise in FM Receivers

 To simplify the system analysis, we assume:

 ideal band-pass filter (that is just wide enough to pass the
modulated signal s(t) without distortion),
 ideal demodulator,
 Gaussian distributed white Noise process.
 So the only source of imperfection is from the noise.
 nI (t )  r (t ) cos( (t ))

nQ (t )  r (t ) sin( (t ))
s(t )  Ac cos[2f ct   (t )], where  (t )  2k f  m( )d .
t

x (t )  Ac cos[2f ct   (t )]  r (t ) cos[2f ct   (t )]
 Ac cos[2f ct   (t )]  r (t ) cos[2f ct   (t )   (t )   (t )]
  Ac  r (t ) cos[ (t )   (t )] cos[2f ct   (t )]
 r (t ) sin[ (t )   (t )] sin[2f ct   (t )]
  Ac  r(t ) cos[ (t )   (t )]2  r 2 (t ) sin2 [ (t )   (t )] cos[2f ct   (t )]
 r(t ) sin[ (t )   (t )] 
where  (t )   (t )  tan 1

 c
A  r ( t ) cos[ ( t )   ( t )] 
x (t )  A c  r ( t ) 
cos[ ( t )   ( t )]
2
 r 2
( t ) sin 2
[ (t )   (t )] cos[2f ct   (t )]
Limiter
 A  cos[2f ct   (t )]

Next, the signal will be passed through a Discriminator.

Recall on Slides 2-94 ~ 2-101, we have talked about the

Balanced Frequency Discriminator, whose input and output
satisfy:
Input s(t )  Acos(2f ct   (t )) Output ~
so (t )  2aA ' (t )
Specifically, with ~ s (t )  A exp( j (t )), we have :
 ~ 
s (t )  aAj ' (t )  BT exp j (t )
~ ds (t )
s1 (t )  a   jBT ~
 dt 
 ~ 
s (t )  aAj ' (t )  BT exp j (t )
~ ds (t )
s2 (t )  a   jBT ~
 dt 
~
so (t ) | ~
s1 (t ) |  | ~
s2 (t ) | 2aA ' (t )

Thus, after passing through the discrimina tor

 1  r ( t ) sin[  ( t )   ( t )] 
d   (t )  tan   
  Ac  r (t ) cos[ (t )   (t )]  
v(t )  2aA ' (t )  2aA
dt
The above assumption is true for any deterministic (constant) (t)!
It has been shown that Assumption 1 is justifiable for high carrier-
to-noise ratio (or equivalently, under Assumption 2).
 We then obtain the desired “additive” form.

d
Table 6.2 : 8. g (t )  j fG( f ) nQ (t ) j 2f nd (t )
dt H( f ) 
2Ac
 jf j (  f ) f2
 SN ( f )  H ( f )H ( f )SN ( f )  SN ( f )  2 SN ( f )
d Q
Ac Ac Q
Ac Q

f2 S N ( f  f c )  S N ( f  f c ),
B 
 2 N 0 , | f | T SN ( f )   for | f |< BT / 2
 S N ( f )   Ac 2
0, otherwise
Q
d

 0, otherwise 
vo (t )  k f m(t )  no (t )

Bandwidth W < BT/2

that is just enough
to pass m(t).
f2
 N , | f | W
 S N ( f )   Ac2 0
o

 0, otherwise
 N0 2 N 0W 3

W
 E[n (t )]  2
2
o f df 
2

Ac W 3 Ac2

Observation from the above formula:

In an FM system, increasing carrier power Ac2 / 2
 Decreasing noise power. This is named the noise quieting effect.
As vo (t )  k f m(t )  no (t ),

k 2f E [m 2 (t )] 3 Ac2 k 2f P
 SNRO ,FM  3
 3
, provided Ac  r (t ).
2 N 0W 2 N 0W
3 Ac2

We next turn to SNRC,FM.

s(t )  Ac cos[2f ct   (t )], where  (t )  2k f  m( )d .

t

 average power in the modulated signal s(t ) is Ac2 / 2.

N0
Average noise power in the message bandwidth is 
W
df  WN 0 .
W 2
Ac2 / 2 Ac2
 SNRC ,FM   .
N 0W 2 N 0W
 3 Ac2 k 2f P
2
SNRO ,FM 2 N 0W 3 3k f P
  2
 2
.
SNRC ,FM Ac W
2WN 0

SNRO ,FM
Remarks : For fixed W , increasing BT  increasing .
SNRC ,FM
f k f P1 / 2
1. Deviation ratio D   . f  k f max | m(t ) |
W W
SNRO ,FM
Hence,  D2.
SNRC ,FM
 1  1
2. BT ,Carson  2f 1    2 DW 1    2W ( D  1)
 D  D
2.13 Summary
 Specifically,
 for high carrier-to-noise ratio r (equivalently to the
assumption made in Assumption 1), an increase in
transmission bandwidth BT provides a corresponding
quadratic increase in figure of merit of a FM system.
 So, there is a tradeoff between BT and figure of merit.
 Notably, figure of merit for an AM system is
nothing to do with BT.
Example 2.5 Single-Tone Modulation
 m(t) = Am cos(2fmt)
 Then we can represent the figure of merit in terms of
modulation index (or deviation ratio)  as (cf. Slide 2-63):
SNRO ,FM 3k 2f P 3k 2f ( Am2 / 2) 3 f 2 3 2
  2
 2
 2
  .
SNRC ,FM W W 2W 2
 In order to make the figure of metric for an FM system to
be superior to that for an AM system with 100%
modulation, it requires:
3 2 1 2
     0.471
2 3 3
2.13 Capture Effect
 Recall that in Assumption 1, we assume Ac >> r(t).
 This somehow hints that the noise suppression of an FM
modulation works well when the noise (or other unwanted
modulated signal that cannot be filtered out by the bandpass or
lowpass filters) is weaker than the desired FM signal.
 What if the unwanted FM signal is stronger than the desired FM
signal.
 The FM receiver will capture the unwanted FM signal!
 What if the unwanted FM signal has nearly equal strength as the
desired FM signal.
 The FM receiver will fluctuate back and forth between them!
2.13 FM Threshold Effect
 Recall that in Assumption 1, we assume Ac >> r(t)
(equivalently, a high carrier-to-noise ratio) to simplify (t)
so that the next formula holds.
3 Ac2 k 2f P
SNRO ,FM  3
.
2 N 0W
 However, a further decrease of carrier-to-noise ratio will
break the FM receiver (from a clicking sound down to a
crackling sound).
 As the same as the AM modulation, this is also named the
threshold effect.
2.13 FM Threshold Effect
 Consider a simplified case with m(t) = 0 (no message
signal).
 1  r (t ) sin[ (t )]  
d  tan   
  Ac  r (t ) cos[ (t )]  
2v (t )   ' (t ) 
dt
Ac r (t ) ' (t ) cos[ (t )]  Ac r ' (t ) sin[ (t )]  r 2 (t ) ' (t )

Ac2  2 Ac r (t ) cos[ (t )]  r 2 (t )

To facilitate the understanding of “clicking” sound effect,

we let r(t) = l Ac, a constant ratio of Ac.
2.13 FM Threshold Effect
cos[ (t )]  l
 2v(t )   ' (t )  l ' (t )
1  2l cos[ (t )]  l2

Then at the time, say,  (t )   , and l  1 but l  1

cos[ (t )]  l l
 2v(t )   ' (t )  l ' (t )   ' (t )
1  2l cos[ (t )]  l2
l 1
Thus a sign change in ’(t) will cause a spike!

Notably, when l = 0 (no noise), the output equals m(t)

= 0 as desired.
2.13 FM Threshold Effect
cos[ sin( t )]  l
 (t )   sin( t )  2v(t )   ' (t )  l cos(t )
1  2l cos[ sin( t )]  l2

l  1.05
l 5
l  0.05
2.13 Numerical Experiment
 Take BT/(2W) = 5.
 The average output signal power is calculated in the
absence of noise.
 The average output noise power is calculated when
there is no signal present.
3 Ac2 k 2f P
SNRO ,FM  3
, provided Ac  r (t )
2 N 0W
3 Ac2 f 2 Am2
 3
, P and f  k f Am
4 N 0W 2
3
 BT  BT Ac2 / 2
 3r   , f  and r  is the carrier to noise ratio.
 2W  2 BT N 0
2.13 Numerical of the
formula in Slide 2-156
 Curve I: The average noise
power is calculated assuming an
unmodulated carrier.
 Curve II: The average noise
power is calculated assuming a
sinusoidally modulated carrier. The deviation of
 As text mentioned, for r < curve II from
11 dB, the output signal-to- curve I shows that
noise ratio deviates the average noise
appreciably from the linear power is indeed
curve. dependent on the
 A true experiment found that modulating signal.
occasional clicks are heard at
r around 13 dB, only
slightly larger than what
theory indicates.
2.13 How to avoid “clicking” sound?
For given modulation index (or deviation ratio)  and message
signal bandwidth W,
1. Determine BT by Carson’s rule or Universal curve.
2. For a specified noise level N0, select Ac to satisfy:
 Ac2  Ac2
10 log 10    13 dB or equvalentl y,  20.
 2 BT N 0  2 BT N 0
2.13 Threshold Reduction
 After our learning
that FM
modulation has
threshold effect,
the next question
is naturally on
“how to reduce
the threshold?”
2.13 Threshold Reduction
 Threshold reduction in FM receivers may be
achieved by
1. negative feedback (commonly referred to as an
FMFB demodulator), or
2. phase-locked loop demodulator.
 Why PLL can reduce threshold effect is not covered in
this course. Notably, the PLL system analysis in
Section 2.14 assumes noise-free transmission.
 x (t ) v (t )

s(t )  Ac cos[2f ct   (t )], where  (t )  2k f  m( )d .

t

svco (t )  2 cos[2f vcot  vco (t )], where vco (t )  2k f  m( )d .
t

s(t ) svco (t )  2 Ac cos[2f ct   (t )] cos[2f vcot  vco (t )]

Bandpass
 Ac cos[2 ( f c  f vco )t  (1   ) (t )]
The new frequency deviation f new  (1   )f original.
Thus, the bandpass filter can conceptually have a smaller passband
as wide as (1   ) BT , centered at ( f c  f vco ).
w(t ) svco (t )  2 w(t ) cos[2f vcot  vco (t )]
 The noise at the Mixer output can be treated white with the same
noise level as the input whit e noise.

Since E [n I2 (t )]  E [nQ2 (t )] is smaller, and Ac remains the same,

the condition of Ac  r (t )  nI2 (t )  nQ2 (t ) holds with higer probabilit y.

Experiments show that an FMFB receiver is capable of realizing a

threshold extension on the order of 5~7 dB.
2.13 Threshold Reduction of an FMFB
receiver
 To sum up:
 An FMFB demodulator is essentially a tracking
filter that can track only the slowly varying
frequency of a wideband FM signal (i.e., the
message signal part), and consequently it responds
only to a narrowband of noise centered about the
instantaneous carrier frequency.
2.13 Pre-Emphasis and De-emphasis in FM
 Recall that the noise PSD at the output shapes like a bowel.
 If we can “equalize” the signal-to-noise power ratios over the
entire message band, a better noise performance should result.
2.13 Pre-Emphasis and De-emphasis in FM

 In order to produce an undistorted version of the original

message at the receiver output, we must have:

H pe ( f ) H de ( f )  1 for  W  f  W .

 This relation guarantees the intactness of the message power.

 Next, we need to find Hde(f) such that the noise power is
optimally suppressed.
2.13 Pre-Emphasis and De-emphasis in FM

 Under the assumption of high carrier-to-noise ratio, the

noise PSD at the de-emphasis filter output is given by:
 N0 f 2
 2 | H de ( f ) |2 , | f | W
| H de ( f ) | S N ( f )   Ac
2
o

 0, otherwise
2
W N f
 Average noise power   0
2
| H de ( f ) |2
df
W Ac
 2.13 Pre-Emphasis and De-emphasis in FM
 Since the message power remains the same, the
improvement factor of the output signal-to-noise ratio after
and before pre/de-emphasis is:
N0 f 2

W


W
df f 2 df
W Ac2
2W 3
I  W

 3
2 W W
N0 f

W 2 2
| H ( f ) |2
df f | H de ( f ) | df f 2 | H de ( f ) |2 df
W 2 de W W
Ac
 Example 2.6
Assume R << r and 2fCr << 1 for | f | W .

Pre-emphasis filter De-emphasis filter

1
H pe ( f )  1  j( f / f 0 ) and H de ( f )  , where f 0  1 /( 2Cr).
1  j( f / f 0 )
 Example 2.6
2W 3 2W 3 (W / f 0 )3
I   
3
W
f 2
3[(W / f )  tan 1
(W / f 0 )]
W 1  ( f / f 0 )2
2 2 W
f | H de ( f ) | df 3 df 0
W

With f0 = 2.1 KHz and W = 15 KHz, we obtain I = 22 = 13 dB.

A significant noise performance improvement is therefore obtained.
2.13 Pre-Emphasis and De-emphasis in FM
 Final remarks:
 The previous example uses a linear pre-emphasis
and de-emphasis filter to improve the noise.
 A non-linear pre-emphasis and de-emphasis filters
have been applied successfully to applications like
tape recording. These techniques, known as Dolby-
A, Dolby-B, and DBX systems, use a combination
of filtering and dynamic range compression to
reduce the effects of noise.
2.14 Computer Experiments: Phase-Locked
Loop
 Phase-locked loop

s(t )  Ac sin[ 2f ct  1 (t )], where 1 (t )  2k f  m( )d .

t

r(t )  Av cos[2f ct  2 (t )], where 2 (t )  2kv  v( )d .

t

km

Loop filter = low pass filter + filter h().
e( t )  k m s ( t ) r ( t )
 km Ac sin[ 2f c t  1 (t )]  Av cos[2f c t  2 (t )]


km Ac Av
sin[ 4f ct  1 (t )  2 (t )]  sin[ 1 (t )  2 (t )]
2
Low P assk A A
 m c v sin[ e (t )], where e (t )  1 (t )  2 (t ).
2

Also, v (t )   {e( )}LowP assh(t   )d .
km

e (t )  1 (t )  2 (t )  1 (t )  2kv  v( )d
t

 de (t ) d1 (t )
  2kv v (t )
dt dt
d1 (t ) 
  2kv  {e( )}LowP assh(t   )d
dt
d1 (t ) 
  2k0  sin[e ( )]h(t   )d
dt
where k0  kmkv Ac Av / 2.
 de (u )
e (t )  0
t
du
du
 0
t d ( u )
1

du
t 
 
du  2k0 0  sin[e ( s )]h(u  s )ds du

 1 (t )  02k0 sin[e (u )] * h(u ) du

t
The previous formula suggests an equivalent analytical model for PLL.

When e(t) = 0, the system is said to be in phase-lock.

In this case, 1(t) = 2(t) or equivalently, kvv(t) = kfm(t).

When e(t) is small (< 0.5 radians), the system is said to be in near
phase-locked.
In this case, we can approximate sin[e(t)] by e(t); hence, a linear
approximate model is resulted.
 Linearization approximation model for PLL.

We can transform the above time-domain system to its

equivalent frequency domain to facilitate its analysis.
 Linearization approximation model for PLL.

e ( f )
1 ( f ) H( f )

2 ( f )

G( f )
1
G( f ) 
j 2f
e ( f ) e ( f )

1 ( f ) [1 ( f )   2 ( f )]   2 ( f )
e ( f )

 e ( f )  2k0 e ( f ) H ( f )G ( f )
1 jf
 
1  2k0 H ( f )G ( f ) jf  k0 H ( f )
2.14 Experiment 0: First-Order PLL
H ( f )  1.

e ( f ) j ( f / k0 )

1 ( f ) 1  j( f / k0 )

A parameter k0 controls both the loop gain and bandwidth of the

filter. In other words, it is impossible to adjust the loop gain
without changing the filter bandwidth.
2.14 Experiment 1: Second-Order
Acquisition PLL

e ( f ) jf jf ( jf ) 2
  
1 ( f ) jf  k0 H ( f ) jf  k0 (1  a /( jf )) ( jf ) 2  k0 ( jf )  k0a
( jf / f n ) 2

1  2 ( jf / f n )  ( jf / f n ) 2

where natural frequency f n  ak0 and damping factor   k0 /( 4a ) .

2.14 Experiment 1: Second-Order
Acquisition PLL
Fast response but require
longer time to stablelize

Slow response but quick

stablelizatoin.
2.14 Experiment 2: Phase-Plain Portrait
H ( f )  1  a /( jf ) and using nonlinear PLL model.

 50

 n 2 2 Herz
f
Take 
  1 (critical damping)
 2
50
Let m(t )  Am cos(2f m t ), where f m  Hz.
2 2
1 de (t ) 1 d1 (t ) 
   sin[ e ( )]h(t   )d
2k0 dt 2k0 dt 

k f Am 
 cos(2f mt )   sin[ e ( )]h(t   )d
k0 
 Phase-Plane Portrait

1 d
e (t )
2k0 dt

Initial value e (t )
Remarks on the previous figure
 Each curve corresponds to different initial frequency error
(i.e., boundary condition for the differential equation).
 The phase-plane portrait is periodic in e(t) with period 2,
but the phase-plane portrait is not periodic in de(t)/dt.
 There exists an initial frequency error (such as 2 in the
previous figure) at which a saddle line appears (i.e., the
solid line in the previous figure).
 In certain cases, the PLL will ultimately reach an
equilibrium (stable) points at (0,0) or (2,0).
 A slightest perturbation to the “saddle line” causes it to shift
to the equilibrium points.
2.15 Summary and Discussion
 Four types of AM modulations are introduced
 (expensive) DSB-C transmitter + (inexpensive) envelope
detector, which is good for applications like radio
broadcasting.
 (less expensive) DSB-SC transmitter + (more complex)
coherent detector, which is good for applications like limited-
transmitter-power point-to-point communication.
 (less bandwidth) VSB transmitter + coherent detector, which
is good for applications like television signals and high speed
data.
 (minimum transmission power/bandwidth) SSB transmitter +
coherent detector, which is perhaps only good for
applications whose message signals have an energy gap on
zero frequency.
2.15 Summary and Discussion
 FM modulation, a representative of Angle
Modulation
 A nonlinear modulation process
 Carson’s rule and universal curve on transmission
bandwidth
2.15 Summary and Threshold effect

Discussion
 Noise performance
I. Full AM (DSB-C) with
100 percent modulation
(See Slide 2-5)
II. Coherent DSB-SC &
SSB
III. FM with  = 2 and 13
dB pre/de-emphasis
improvement
IV. FM with  = 5 and 13
dB pre/de-emphasis
improvement
2.15 Summary and Discussion
 Normalized transmission bandwidth in the previous figure

DSB-C, DSB-SC SSB FM(=2) FM(=5)

Bn 2 1 8 16
 24  5 3.2
BT BT / f BT
Bn ,UniversalCurve    (Refer to Slide 2-85.)
W W / f f
2.15 Summary and Discussion
 Observations from the figure
1. SSB modulation is optimum in noise performance
and bandwidth conservation in AM family.
2. FM modulation improves the noise performance of
AM family at the expense of an excessive
transmission bandwidth.
3. Curves III an IV indicate that FM modulation
offers the tradeoff between transmission bandwidth
and noise performance.