Communications II

Lecture 4: Effects of Noise on AM

Professor Kin K. Leung

EEE and Computing Departments
Imperial College London
© Copyright reserved
 Noise in Analog Communication Systems

• How do various analog modulation schemes perform in the

presence of noise?

• Which scheme performs best?

• How can we measure its performance?

© KKL 2
 We must find a way to quantify (=to measure) the performance of a
modulation scheme.

We use the signal-to-noise ratio (SNR) at the output of the receiver:

average power of message signal at the receiver output PS

SNR0  
average power of noise at the receiver output PN

© KKL 3
 Model of an analog communication system

© KKL 4
 PT : The transmitted power

Limited by: equipment capability, cost, government restrictions,

interference with other channels, etc

The higher it is, the more the received power (PS ), the higher the SNR

For a fair comparison between different modulation schemes:

PT should be the same for all

We use the baseband signal to noise ration SNRbaseband to calibrate the

SNR values we obtain

© KKL 5
 A Baseband Communication System

• It does not use modulation

• It is suitable for transmission over wires

• The power it transmits is identical to the message

power: PT = P

• The results carry over to band-pass systems

© KKL 6
© KKL 7
 Average signal(=message) power:
P =the area under the triangular curve

Assume:
Additive, white noise with power spectral density PSD=N0/2

Average noise power at the receiver:

PN = area under the straight line = 2WN0/2 = WN0

© KKL 8
 SNR at the receiver output:

PT
SNRbaseband 
N 0W

Note: Assume no propagation loss PT  PS

Improve the SNR by:

(a) increasing the transmitted power (PT ↑),
(b) restricting the message bandwidth (W ↓),
(c) making the receiver less noisy (N0 ↓).

© KKL 9
 REVISION: Amplitude Modulation

General form of an AM signal:

s(t ) AM  [ A  m(t )] cos(2f ct )

A: the amplitude of the carrier

fc: the carrier frequency

m(t): the message signal

© KKL 10
 Modulation index:

mp

A
mp: the peak amplitude of m(t), i.e., mp = max |m(t)|

© KKL 11
 Signal recovery

1)   1  A  m p : use an envelope detector

2) Otherwise: use
synchronous detection=product demodulation=coherent
detection

© KKL 12
 Synchronous detection
• Multiply the waveform at the receiver with a local carrier of the
same frequency (and phase) as the carrier used at the transmitter:

• Use a LPF to recover (m(t) + A)/2 and finally m(t)

• Problem: At the receiver you need a signal perfectly

synchronised with the transmitted carrier
© KKL 13
 REVISION Amplitude Modulation: Double-sideband suppressed
carrier (DSB-SC)

s(t ) DSBS  Am(t ) cos(2f ct )

Signal recovery: With synchronous detection only

© KKL 14
 Noise in DSB-SC

The true signal received is:

x (t )  s (t )  n(t )
 s(t )  nc (t ) cos(2f c t )  ns (t ) sin( 2f c t )
 Am(t ) cos(2f c t )  nc (t ) cos(2f c t )  ns (t ) sin( 2f c t )
 [ Am(t )  nc (t )] cos(2f c t )  ns (t ) sin( 2f c t )

© KKL 15
 For synchronous detection:

• multiply with 2cos(2πfct):

y (t )  2 cos(2f ct ) x (t )
 Am(t ) cos2 ( 2f ct )  nc (t )2 cos2 ( 2f ct )  ns (t ) sin(4f ct )
 Am(t )[1  cos(4f ct )]  nc (t )[1  cos(4f ct )]
 ns (t ) sin(4f ct )

• Use a LPF to keep

~
y  Am(t )  nc (t )

© KKL 16
 Signal power at the receiver output:

PS  E{ A2m 2 (t )}  A2 E{m 2 (t )}  A2 P

Power of the noise signal nc(t):

W
PN   N 0df  2 N 0W
W

© KKL 17
 SNR at the receiver output:

A2 P
SNR0 
2 N 0W

To which transmitted power does this correspond?

PT  E{ A2m(t ) 2 cos2 (2f c t )}

A2 P

2

© KKL 18
 So
PT
SNR0   SNRDSB  SC
N 0W

PT
Comparison with SNRbaseband 
N 0W

SNRDSB  SC  SNRbaseband

Conclusion: a DSB-SC system provides no SNR performance gain over a

baseband system.

© KKL 19
 Noise in standard AM, Synchronous Detection

Pre-detection signal:

x (t )  [ A  m(t )] cos(2f ct )  n(t )

 [ A  m(t )] cos(2f ct )  nc (t ) cos(2f ct )  ns (t ) sin(2f c t )
 [ A  m(t )  nc (t )] cos(2f c t )  ns (t ) sin(2f ct )

© KKL 20
 Signal Recovery:

• Multiply with 2 cos(2πfct):

y (t )  A[1  cos(4f c t )]  m(t )[1  cos(4f ct )]

 nc (t )[1  cos(4f c t )]  ns (t ) sin( 4f ct )

• LPF
~
y  A  m(t )  nc (t )

© KKL 21
 Signal power at the receiver output:

PS  E{m 2 (t )}  P

Noise power:

PN  2 N 0 P

© KKL 22
 SNR at the receiver output:

P
SNR0   SNRAM
2 N 0W
Transmitted power:

A2 P A2  P
PT   
2 2 2
SNR of a baseband signal with the same transmitted power:
A2  P
SNRbaseband 
2 N 0W

© KKL 23
 Thus:

P
SNRAM  2 SNRbaseband
A P

Note:
P
1
A P
2

Conclusion: the performance of standard AM with synchronous recovery

is worse than that of a baseband system.

© KKL 24
 Noise in standard AM, Envelope Detection

Phasor diagram of the signals present at an AM receiver

Ei(t): receiver output= y(t)

© KKL 25
 y (t )  envelope of x (t )
(95)
 [ A  m(t )  nc (t )]  ns (t )
2 2

Equation too complicated

Must use limiting cases to put it in a form where noise and message are
added

© KKL 26
 1st Approximation: (a) Small Noise Case

n(t )  [ A  m(t )]

Then
ns (t )  [ A  m(t )  nc (t )]

Then

y (t )  [ A  m(t )  nc (t )]

© KKL 27
 Thus
P
SNR0   SNRenv
2 N 0W

And in terms of baseband SNR:

P
SNRenv  SNRbaseband
A P
2

Valid for small noise only!

© KKL 28
 2nd Approximation: (b) Large Noise Case

n(t )  [ A  m(t )]

Isolate the small quantity in (95):

y 2 (t )  [ A  m(t )  nc (t )]2  ns2 (t )

 ( A  m(t )) 2  nc2 (t )  2( A  m(t ))nc (t )  ns2 (t )
 ( A  m(t )) 2 2( A  m(t ))nc (t ) 
 [n (t )  n (t )]1  2
2 2
 
 
c s 2 2 2 2
 nc ( t ) n s ( t ) n c ( t ) n s ( t ) 

© KKL 29
  2[ A  m(t )]nc (t ) 
y (t )  [n (t )  n (t )]1 
2 2 2

nc (t )  ns (t ) 
c s 2 2

 2[ A  m(t )]nc (t ) 
 E (t )1 
2
n 2

 En ( t ) 

where

En(t )  nc2 (t )  ns2 (t ) : the envelope of the noise

© KKL 30
 From the phasor diagram: nc(t) = En(t) cosθn(t)

Then:
2[ A  m(t )] cos  n (t )
y ( t )  En ( t ) 1 
En ( t )
x
Use 1  x  1  for x  1 :
2
 [ A  m(t )] cos  n (t ) 
y (t )  En (t )1  
 En ( t ) 
 En (t )  [ A  m(t )] cos  n (t )

© KKL 31
 Noise is multiplicative here!

No term proportional to the message!

Result: a threshold effect, as below some carrier power level (very low
A), the performance of the detector deteriorates very rapidly.

© KKL 32
 SSB modulation

Single (lower) sideband AM:

A A
s(t ) SSB  m(t ) cos 2f c t  mˆ (t ) sin 2f ct
2 2
where m ˆ (t ) is the Hilbert transform of m(t).
mˆ (t ) is obtained by passing m(t) through a linear filter with transfer
function −jsgn(f ).
mˆ (t ) and m(t) have the same power P .
The average power is A2P/4.

© KKL 33
 Noise in SSB

Receiver signal x(t) = s(t) + n(t).

Apply a band-pass filter on the lower sideband.

Using coherent detection:
y (t )  x (t )  2 cos(2f c t )
A  A 
  m(t )  nc (t )    m(t )  nc (t )  cos(4f c t )
2  2 
A 
  mˆ (t )  ns (t )  sin(4f ct )
2 
After low-pass filtering,
A 
y (t )   m(t )  nc (t ) 
2 

© KKL 34
 Signal power A2P/4

Noise power for nc(t) = that for band-pass noise = N0W

SNR at output
A2 P
SNRSSB 
4 N 0W
For a baseband system with the same transmitted power A2P/4

A2 P
SNRbaseband 
4 N 0W
Conclusion: SSB achieves the same SNR performance as DSB-SC (and
the baseband model) but only requires half the band-width.
© KKL 35
 Summary

© KKL 36