Analog Communication Systems

Lecture 5
Effect of Noise on Analog Communication Systems

Dr. Adnan Ismail Al-Sulaifanie

Department of Electrical and Computer Engineering

College of Engineering
University of Duhok

2020 - 2021
Outline

I Noise signal: definition, sources, and classification.

I White noise: definition and mathematical representation.
I Noise power in bandlimited systems.
I Effect of noise on amplitude modulation systems.
I Effect of noise on angle modulation systems.

2 / 26
References

1. Proakis, Fundamentals of communication systems, chapter 6

2. Haykin, Digital Communication systems, chapter 6

3 / 26
Noise

I Noise is defined as unwanted signal that randomly added to the desired signal.
I Noise signals are small at point of origin usually in mV or in µV.
I Problem occurs at the receiver where the message signal is also small. Therefore,
amplifying the information signal also amplifies the noise.

I Noise level depends on bandwidth, temperature, amount of current, and gain.

I Noise degrades system performance and reduces its efficiency.
4 / 26
Noise Categories
A. External noise: this category is uncontrollable and has different sources: interference
from nearby channels, man-made, and natural sources.

1. Manmade Noise often produced by spark-producing mechanisms such as engine

ignition systems and fluorescent lights.
2. Atmospheric Noise is caused by naturally occurring disturbances in the earths
atmosphere.
3. Space Noise arrives from outer space. Solar noise is generated by our sun and cosmic
noise comes from all other stars.
B. Internal noise: is generated by the communication devices themselves:
1. Shot noise: the electrons are discrete and are not moving in a continuous steady flow,
so the current is randomly fluctuating.
2. Thermal noise: caused by the rapid and random motion of electrons within a conductor
due to thermal agitation.
Thermal agitation =⇒ increases free electrons =⇒ increases current flows through the
conductor.
5 / 26
White Noise

I White noise is a random signal having equal intensity at different frequencies, giving it
a constant power spectral density.
I It is analogous to the term white light in the sense that all frequency components are
present in equal amounts.
I The power spectral density (PSD)of white noise is constant over all frequencies
N0
SN (f ) =
2
I Factor 1/2 is included to indicate that half the power is associated with positive
frequencies and half with negative.
I Systems with limited bandwidth reduces amount of noise power that appeared at the
output (it is called narrowband noise).
I Narrowband noise can be represented mathematically using in-phase and quadrature
components. For the narrowband noise of bandwidth 2B and centered at frequency fc ,
we may represent N(t) in the form:
n(t) = nc (t) cos(2πfc t) − ns (t) sin(2πfc t)
6 / 26
nc (t) and ns (t) are low-pass random signal and they have the same PSD. Also they have
the same variance as the band-pass noise n(t)
7 / 26
Noise Power
For ideally filtered narrowband noise, the PSD of nc (t) and ns (t) is therefore given by
Sc (f ) = Ss (f ) = N0 |f | < B
The average power in each of the baseband noise signal of nc (t) and ns (t) is identical and
equal to the average power of the pandpass noise.
P(nc (t)) = P(ns (t)) = P(n(t))
For ideal LPF, the average power of the noise is determined by:
∫ B
N0
Pn = df = No B
−B 2
For non ideal filter, the same result can be achieved:

8 / 26
 Pr Pr
The SNR of the baseband signal at receiver side is =
Pn No B

9 / 26
Noise in Analog Communication Systems

I How do various analog modulation schemes perform in the presence of noise?

I Which scheme performs best?
I How can we measure its performance?

Pt transmitted power of s(t)

Pm power of m(t)
Pr received power. It is assumed no loss during transmission, hence Pr = Pt
n(t) noise signal n(t) = nc (t) cos(2πfc t) − ns (t) sin(2πfc t)
Pn noise power at the input of the receiver
Pno noise power at the output of the receiver
Pr
SNR = SNR of the output of the receiver
Pno

10 / 26
Effect of Noise on Baseband transmission
The transmitted power Pt = Pm
∫ B
N0
At the receiver side, the average noise power is given by: Pno = df = N0 B
−B 2
Assume Pt is the transmitted power and there is no loss during the transmission, i.e.
Pr = Pt , the SNR at receiver output is
Pt
SNR of baseband transmission is SNRb =
N0 B
SNR can be improved by
1. Increasing transmitted power.
2. Restricting message bandwidth.
3. making the receiver less noisy.

11 / 26
Example
Find the SNR in a baseband system with a bandwidth of 5 kHz and with N0 = 10−14 W/Hz.
The transmitter power is 1 kilowatt and the channel attenuation is 10−12 Pt
Pr 10−9
SNR = = −14 = 20
N0 B 10 ∗ 5000
SNRdB = 10 Log10 20 = 13 dB

12 / 26
Effect of Noise on DSB-SC-AM I
s(t) = m(t) ∗ Ac cos(ωc t)
r (t) = s(t) + n(t) = Ac m(t) cos(ωc t) + nc (t) cos(ωc t) − ns (t) sin(ωc t)
[ ]
r (t) = Ac .m(t) + nc (t) cos(ωc t) − ns (t) sin(ωc t)
Synchronous detection
[ ]
y (t) = r (t) cos(ωc t) = Ac .m(t) + nc (t) cos2 (ωc t) − ns (t) sin(ωc t) cos(ωc t)

1[ ][ ] 1
y (t) = Ac .m(t) + nc (t) 1 + cos(2ωc t) − ns (t) sin(2ωc t)
2 2
1[ ] 1[ ] 1
y (t) = Ac .m(t) + nc (t) + Ac .m(t) + nc (t) cos(2ωc t) − ns (t) sin(2ωc t)
2 2 2
Ac 1
The output of LPF is yo (t) = .m(t) + nc (t)
2 2
A2c Pt
The signal power at the receiver output is Pmo = Pm =
4 4

13 / 26
Effect of Noise on DSB-SC-AM II
Pn
The noise power at the receiver output is Pno =
4
Pn = 2 No B for bandpass signal
Pt
Pmo A2 Pm Pt
SNRDSB−SC = = 4 = c =
Pno Pn 2N0 B N0 B
4
SNRDSB−SC = SNRb
The SNR in DSB-SC is same as the SNR for a baseband system, i.e. we lose nothing in
performance by using a band-pass modulation scheme compared to the baseband
transmission, even though the bandwidth of the former is twice as wide.

14 / 26
Homework

1. Derive SNRo if the local carrier at receiver side is shifted by ϕ

2. A DSC-SC modulated signal is transmitted over a noisy channel, having a noise
spectral density N20 = 2 ∗ 10−17 of watts per hertz. The message bandwidth is 4 kHz
and the carrier frequency is 200 kHz. Assume that the average received power of the
signal is -80 dB. Determine the post-detection signal-to-noise ratio of the receiver.

15 / 26
Effect of Noise on DSB-C AM
[ ]
s(t) = Ac 1 + µ mn (t) cos(ωc t)

A2c
Pt = (1 + µ2 Pm )
2
[ ]
r (t) = s(t) + n(t) = Ac + µAc mn (t) + nc (t) cos(ωc t) − ns (t) sin(ωc t)

Under high SNR and with the aid of phasor diagram

Ac + µAc mn (t) + nc (t) ≫ ns (t)
the output of the rectifier is y (t) ≈ Ac + µ Ac mn (t) + nc (t)
This new expression for the demodulated signal has three components: dc component,
signal component, and noise component. The dc term can be removed with a capacitor.
Then the output of the filter is
yo ≈ µ Ac mn (t) + nc (t)
The signal power at the receiver output is Pmo = A2c µ2 Pm
The noise power at the receiver output is Pno = Pn = 2 N0 B
16 / 26
 A2c
Pmo A2cPm µ2 A2c
Pm µ2 (1 + µ2 Pm )
SNRDSB−C = = = 2
Pno 2 N0 B 2 N0 B A2c
(1 + µ2 Pm )
2
A2c
µ2
Pm (1 + µ2 Pm )
SNRDSB−C = 2
(1 + µ2 Pm ) N0 B

µ 2 Pm
SNRDSB−C = SNRb
1 + µ2 Pm
The figure of merit for this system is always less than 0.5. Hence, the noise performance of
an envelope-detector receiver is always inferior to a DSB-SC receiver, the reason is that at
least half of the power is wasted transmitting the carrier as a component of the modulated
(transmitted) signal.

17 / 26
Effect of Noise on SSB-AM I
Ac Ac
s(t) = m(t) cos(ωc t) ∓ m̂(t) sin(ωc t)
2 2
A2c A2 A2
The transmitted power is Pt = Pm + c Pm = c Pm
8 8 4
The input of the modulator is
Ac Ac
r (t) = m(t) cos(ωc t) ∓ m̂(t) sin(ωc t) + n(t)
2 2
[ Ac ] [ Ac ]
r (t) = .m(t) + nc (t) cos(ωc t) + ∓ m̂(t) − ns (t) sin(ωc t)
2 2
Using synchronous detection, the output will be
1 [ Ac ] 1[A
c
] 1[ ]
y (t) = .m(t) + nc (t) + .m(t) + nc (t) cos(2ωc t) ∓ Ac .m̂(t) − ns (t) sin(2ωc t)
2 2 2 2 2
Ac 1
The output of LPF is yo = .m(t) + nc (t)
4 2

18 / 26
Effect of Noise on SSB-AM II
A2c Pt
The signal power at the receiver output is Pmo = Pm =
16 4
Pn
The noise power at the receiver output is Pno =
4
Pt
Pmo Pt Pt
SNRSSB = = 4 = =
Pno Pn Pn N0 B
4
Pt
SNRSSB = = SNRDSB−SC = SNRb
N0 B
SSB achieves the same SNR performance as DSB-SC (and the baseband model) but only
requires half the band-width.

19 / 26
Effect of Noise on Frequency Modulation
[ ] ∫t
s(t) = Ac cos 2πfc t + ϕ(t) ϕ(t) = 2πkf −∞ m(τ ) dτ
n(t) = nc (t) cos(ωc t) − ns (t) sin(ωc t)
n(t) = An (t) cos(ωc t + ϕn )
√ [ n (t) ]
s
An (t) = nc2 (t) + ns2 (t) ϕn = tan−1
nc (t)
√ [ [ n (t) ]]
s
n(t) = nc2 (t) + ns2 (t) cos ωc t + tan−1
nc (t)
[ ] [ ]
r (t) = s(t) + n(t) = Ac cos ωc t + ϕ(t) + An (t) cos ωc t + ϕn (t)

The amplitude of noise is An (t) and the phase difference is Ψn (t) = ϕn (t) − ϕ(t)

20 / 26
 [ An (t) sin[Ψn (t)] ]
θ(t) = ϕ(t) + tan−1
Ac + An (t) cos[Ψn (t)]
We don’t consider the envelope of r (t) because the envelope variations at the bandpass
filter output are removed by the limiter.
Ac ≫ An (t) cos[Ψn (t)]
and
tan−1 (x) ≈ x if x ≪ 1
The previous equation can be approximated to
An (t) sin[Ψn (t)]
θ(t) = ϕ(t) +
Ac
This expression can be further simplified by and replacing Ψn (t) = ϕn (t) − ϕ(t) with ϕn (t).
This is justified because the phase ϕn (t) is uniformly distributed between 0 and 2π radians
and, since ϕ(t) is independent of ϕn (t), it is reasonable to assume that the phase difference
ϕn (t) − ϕ(t) is also uniformly distributed over radians. Theoretical considerations show that
this assumption is justified provided that the carrier-to-noise ratio is high.

21 / 26
ns (t) = An (t) sin[ϕn (t)]
ns (t)
θ(t) = ϕ(t) +
Ac

22 / 26
 ∫t ns (t)
θ(t) = 2πkf −∞ m(τ ) dτ +
Ac
dθ(t)
With an ideal discriminator, its output is proportional to the derivative
dt
The output of the discriminator scaled by 2π is
1 dθ(t)
v (t) = = kf m(t) + nd (t)
2π dt
The noise term nd (t) is defined by
1 dns (t)
nd (t) =
2πAc dt
The carrier-to-noise ratio is high and the discriminator output v (t) consists of the original
message signal m(t) multiplied by the constant factor kf plus an additive noise component
nd (t). The additive noise at the discriminator output is determined essentially by the
quadrature component ns (t) of the narrowband noise n(t).
The signal power at the receiver output is Pmo = kf2 Pm
2N0 B 3
The noise power at the receiver output is Pno =
3A2c
23 / 26
 3 A2c kf2 Pm
SNRFM =
2N0 B 3
3 A2c kf2 Pm
SNRFM 2N0 B 3 kf2 Pm
Figure of merit = = = 3 = 3 β2 = 3 D2
SNRb A2c B2
2N0 B
SNRFM 3 ( BT )2
Figure of merit ≈
SNRb 4 B

24 / 26
Summary
Pn = Pnc = Pns = No ∗ Bt Bt : transmission-bandwidth

SNR
Technique Pt Pn Pmo Pno SNR
SNRb
Pt
Baseband Pm No .B Pm Pn
No .B
A2c A2c Pn Pt
DSB-SC Pm 2No .B Pm 1
2 4 4 No .B
A2c µ2 Pm Pt µ2 Pm
DSB-C (1 + µ2 Pm ) 2No .B A2c µ2 Pm Pn 1+µ2 Pm No .B
2 1 + µ2 Pm
A2c A2c Pn Pt
SSB Pm No .B Pm 1
4 16 4 No .B
A2c 2No B 3 3 A2c Kf2 Pm
FM No .2B(1 + β) Kf2 Pm 3A2c 2No B 3
3 β2
2

25 / 26
Example
Derive expression of SNR at the output of the AM demodulator when the phase difference
between two local carriers is ϕc in the following two cases:
1. SSB
2. DSB-SC

26 / 26