Communications Chapter Two PDF

CHAPTER TWO
Communication System Elements

(Signal Analysis):
Signal classification of periodic and non-periodic
signals (Fourier series and Fourier transform), classification
of systems, power spectral density and correlation.
(CH 2) Signal Analysis By Dr. Hikmat Al-Shamary & Dr. Tariq M. Salman
Chapter 2
Signal Analysis
The signal is that serves to start some action. In communication systems, the
signal is an electrical voltage v(t) or current i(t).
Signal classification
 A signal is said to be deterministic if it could be expressed mathematically as

a function of time (e.g. f(t)=cosωt), otherwise it is said to be random (e.g.
noise).
 A signal is said to be periodic if it repeat itself after a fixed length of time
(i.e. f(t)=f(t+T)) when -∞<t<∞), otherwise it is said to be non-periodic.
 A signal is said to be power, if it has a finite value of average power while it
is said to be energy signal if it has a finite value of energy.
𝑇
1
Signal power 𝑃𝑎𝑣 = ∫|𝑓(𝑡)|2 𝑑𝑡 watt (2-1)
𝑇
0
Signal energy 𝐸 = ∫ |𝑓(𝑡)|2 𝑑𝑡 joul/c (2-2)

−∞
At R =1 Ω.
Notes:
1. For a given value of resistance R.

𝑇 ∞
1 1
𝑃𝑎𝑣 = ∫|𝑓(𝑡)|2 𝑑𝑡 , 𝐸 = ∫ |𝑓(𝑡)|2 𝑑𝑡 … f(t) voltage
𝑇𝑅 𝑅
0 −∞
𝑇 ∞
𝑅
𝑃𝑎𝑣 = ∫|𝑓(𝑡)|2 𝑑𝑡 , 𝐸 = 𝑅 ∫ |𝑓(𝑡)|2 𝑑𝑡 … f(t) current
𝑇
0 −∞
7
2. Periodic signal is almost power signal.

3. Nonperiodic signal is almost energy signal.
4. Some of signals are neither power nor energy,( increasing or divergent
signals like tan t, 1/t, eat…etc.).
5. The power of energy signal is zero while the energy of power signal is
infinity.
Ex 2-1:
Classify the following signals:
1-𝑓(𝑡) = cos 3 𝑡
2-𝑓(𝑡) = 𝑒 −2|𝑡|
1----
3-𝑓(𝑡) = 𝑒 +2𝑡
Solution:
1-𝑓(𝑡) = cos 3 𝑡
 Deterministic T=2π/3
 Periodic
 Energy or power?
∞
Test energy; 𝐸 = 𝑅 ∫−∞|𝑓(𝑡)|2 𝑑𝑡 = ∞, 𝑛𝑜𝑡 𝑒𝑛𝑒𝑟𝑔𝑦 𝑠𝑖𝑔𝑛𝑎𝑙
1 T
Test Power; 𝑃𝑎𝑣 = ∫0 |𝑓(𝑡)|2 𝑑𝑡
𝑇
2𝜋 2𝜋
3 2 3 1 1
= ∫ |cos 3𝑡| 𝑑𝑡 =2𝜋 ∫0 (2 + 2 cos 6 𝑡) 𝑑𝑡
2𝜋 0
3 3
=1/2 watt ‫ ؞‬power signal

f(t)
−2|𝑡|
2-𝑓(𝑡) = 𝑒
e+2t e-2t
 Deterministic
 Non periodic t
1 T
Text power: 𝑃𝑎𝑣 = ∫0 |𝑓(𝑡)|2 𝑑𝑡 = 0, 𝑠𝑖𝑛𝑐𝑒, 𝑇 = ∞
𝑇
∞ ∞ 𝑒 −4𝑡 ∞
Test energy; 𝐸 = 𝑅 ∫−∞|𝑓(𝑡)|2 𝑑𝑡=2 ∫0 𝑒 −4𝑡 𝑑𝑡=2 | =1/2 joule
−4 0
8
3-𝑓(𝑡) = 𝑒 +2𝑡
 Deterministic
 Non periodic
 Neither power nor energy (since its divergent increasing signal)
f(t)
e+2t
1----
Some important signals

1- Impulse (Dirac)
δ(t)
∞, 𝑡=0
𝛿(𝑡) = {
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒 t
Properties:
∞
 ∫−∞ 𝛿(𝑡)𝑑𝑡 = 1
∞
∫−∞ 𝑓(𝑡)𝛿(𝑡)𝑑𝑡 = 𝑓(0)
 ∞ } sampling property
∫−∞ 𝑓(𝑡)𝛿(𝑡 − 𝑡0 )𝑑𝑡 = 𝑓(𝑡0 )
f(t)
δ(t-t0) f(t0)
t t
Ex 2-2:
∞
Solve ∫−∞ 𝑡 2 𝑒 −𝑠𝑖𝑛𝑡 𝑐𝑜𝑠2𝑡 𝛿(𝑡 − 𝜋)𝑑𝑡
Solution: = 𝜋 2 𝑒 −𝑠𝑖𝑛𝜋 cos 2𝜋 = 𝜋 2 . 𝑒 0 . 1 = 𝜋 2
9
H.W:
1
 Show that (a) 𝛿(𝑎𝑡) = |𝑎| 𝛿(𝑡), (b) 𝛿(𝑡) = 𝛿(−𝑡),
∞ 1 1
 Solve ∫−∞ 𝛿(𝑡 − 𝜋𝑡)cos( ) 𝑑𝑡… Ans: ( )
𝑇 𝜋
2- Unit step:
1, 𝑡 ≥ 0
𝑢(𝑡) = { … (2-5)
0, 𝑡 < 0
𝑑 𝑡
𝛿(𝑡) = 𝑢(𝑡) ; 𝑢(𝑡) = ∫−∞ 𝛿(𝑡)′ 𝑑𝑡 ′ … (2-6)
𝑑𝑡
3- Signum:
−1, 𝑡 < 0
𝑠𝑔𝑛(𝑡) = { 0, 𝑡 = 0 … (2-7)
+1, 𝑡 > 0
𝑠𝑔𝑛 (𝑡) = 2𝑢(𝑡) − 1 …(2-8)
4- Sinc:
𝑠𝑖𝑛𝜋𝑡
𝑠𝑖𝑛𝑐(𝑡) = …(2-9)
𝜋𝑡
5- Sa:
𝑠𝑖𝑛𝑡
𝑆𝑎(𝑡) = …(2-10)
𝑡
H.W: Show that, 𝑠𝑖𝑛𝑐(0) = 𝑆𝑎(0) = 1
10
Signal spectrum:
The spectrum is the frequency representation of a signal:
1- Spectrum of the sinusoidal signals:
𝑓(𝑡) = 𝐴 cos(2𝜋𝑓0 𝑡 + 𝜃)
a) Single sideband spectrum:
Phase
A θ
f f
f0 f0
b) Double sideband spectrum:
Amplitude
Phase
A/2
θ
f f
-f0 f0 -f0 f0
-θ
Rules:
1. The amplitude is always positive.
𝐴 cos 2𝜋𝑓0 𝑡 ⟹ 𝐴 cos(2𝜋𝑓0 𝑡 − 1800 )

2. The phase ids always measured from the real axis.
𝐴 sin 2𝜋𝑓0 𝑡 ⟹ 𝐴 cos(2𝜋𝑓0 𝑡 − 900 )
3. If the angular frequency is used ( ω ), the amplitude is multiplied
by 2π.
Ex 2-3:
Plot the amplitude and phase spectrum (single and double sided) for the following
signal:
𝑓(𝑡) = 4 cos(2𝜋𝑓𝑜 𝑡 + 30𝑜 ) + 2sin(4𝜋𝑓𝑜 𝑡)
11
Solution:
𝑓(𝑡) = 4 cos(2𝜋𝑓𝑜 𝑡 + 30𝑜 ) + 2cos(4𝜋𝑓𝑜 𝑡 − 90)
Amplitude Phase
2 30o
f f
f0 2f0 f0 2f0
-90o
Amplitude Phase
90o
2
1 30o
f f
-2f0 -f0 f0 2f0 -2f0 -f0 f0 2f0
-30o
-90o
Amplitude Phase
90o
4π 30o
2π
ωo ω
-2ωo -ωo ωo 2ωo -2ωo -ωo ωo 2ωo
-30o
-90o
If the x-axis is ω ( not f), the amplitude is multiplied by 2π.

H.W:
Plot the amplitude and phase spectrum (single and double sided) for the following
signal:
1. 𝑓(𝑡) = 5 + 3 sin(2𝜋 ∗ 103 𝑡 + 30𝑜 ) − 4 cos(4𝜋 ∗ 103 𝑡 − 20𝑜 )
2. 𝑓(𝑡) = −2 + 10 sin(2𝜋 ∗ 105 𝑡 − 20𝑜 ) − 14 cos(6𝜋 ∗ 105 𝑡 + 30𝑜 ) +
7 sin(4𝜋 ∗ 105 𝑡 − 50𝑜 )
3. 𝑓(𝑡) = 1 + 4 cos(10000𝑡 + 30𝑜 )
12
2- Spectrum of periodic signals

The spectrum of any periodic signal double sided) can be obtained by
plotting |𝐶𝑛 |versus 𝑛𝑓𝑜 and 𝜃𝑛 versus 𝑛𝑓𝑜 , where :
T
1 𝑎𝑛 𝑏𝑛
|𝐶𝑛 | = ∫ 𝑓(𝑡)𝑒 −𝑗𝑛𝜔𝑜 𝑡 𝑑𝑡 = −𝑗 … (2-11)
𝑇 2 2
0
𝑏𝑛 … (2-12)
|𝜃𝑛 | = 𝑡𝑎𝑛 −1
𝑎𝑛
Where T is the period of the periodic signal
𝑎 2 𝑏 2
|𝐶𝑛 | = √( 𝑛 ) + ( 𝑛) … (2-13)
2 2
Ex 2-4:
Plot the double-sided amplitude and phase spectrum of the signal shown below:
f(t)
Solution:
T
1 e-t
|𝐶𝑛 | = ∫ 𝑓(𝑡)𝑒 −𝑗𝑛𝜔𝑜 𝑡 𝑑𝑡
𝑇
0
t
0.5 1
1 0.5 −𝑡 −𝑗𝑛𝜔 𝑡
= ∫ 𝑒 𝑒 𝑜 𝑑𝑡
0.5 0 T
0.5
= 2 ∫0 𝑒 −(1+𝑗4𝜋𝑛)𝑡 𝑑𝑡 T=0.5, 𝑤𝑜 = 4π, fo= 2 Hz
0.79
=
1+𝑗4𝜋𝑛
0.79
|𝐶𝑛 | = 𝜃𝑛 = −𝑡𝑎𝑛−1 (4𝜋𝑛)
√1+16𝜋2 𝑛
13
n │Cn│ n θn
0 0.79 0 0
±1 0.063 ±1 .
±2 . ±2 .
±3 . ±3 .
. . . .
Cn
0.79
0.063
nf0 Hz
-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12
θn
π/2
2 4 6 8 10 12 nf0 Hz
-12 -10 -8 -6 -4 -2 0
-π/2
We conclude that the spectrum of periodic signals is discrete.
14
Ex 2-5:
Plot the double sided amplitude and phase spectrum of the periodic signal shown
below (rectangular pulse) when
a) τ=1, To=5, 10 and 20 sec.
b) To=20, τ=4, 2 and 1 sec.
f(t)
Solution: Pulse duration
τ: pulse duration A
To : period
-τ/2 τ/2
t
1 2𝜋
fo : ; 𝜔 = 2𝜋𝑓𝑜 = To
𝑇𝑜 𝑇𝑜
2𝜋
≤ 1 “Duty cycle”
𝑇𝑜
𝜏
𝐴, 𝑡 ≤
2
𝑓(𝑡) = { 𝑇𝑜 𝑇𝑜
0, − < |𝑇| <
2 2
𝜏
1 T 1
|𝐶𝑛 | = ∫ 𝑓(𝑡)𝑒 −𝑗𝑛𝜔𝑜 𝑡 𝑑𝑡 = ∫ 𝜏 𝐴𝑒 −𝑗𝑛𝜔𝑜 𝑡 𝑑𝑡
2
𝑇 0 𝑇𝑜 −
2
𝑗𝑛𝜔𝑜 𝜏 𝑗𝑛𝜔𝑜 𝜏
𝐴
= (𝑒 − 2 − 𝑒+ 2 ) n≠0
−𝑗𝑛𝜔𝑜 𝑇𝑜
𝑗𝑛𝜔𝑜 𝜏 𝑗𝑛𝜔𝑜 𝜏
−
2𝐴 𝑒 2 −𝑒 2
= ( ) n≠0
𝑛𝜔𝑜 𝑇𝑜 2𝑗
2𝐴 𝑛𝜔𝑜 𝜏
= 𝑠𝑖𝑛( ) *{τ/τ}
𝑛𝜔𝑜 𝑇𝑜 2
𝑛𝜔 𝜏
𝑜
𝐴𝜏 sin( 2 )
= 𝑛𝜔 𝜏
𝑇𝑜 ( 𝑜 )
2
𝐴𝜏 𝑛𝜔𝑜 𝜏
𝐶𝑛 = 𝑆𝑎( )
𝑇𝑜 2 …(2-14)
General for the rectangular pulses of amplitude A,
duration τ, and period To.
𝐴𝜏 𝑛𝜋𝜏
𝐶𝑛 = 𝑆𝑎( )
𝑇𝑜 𝑇𝑜
15
𝐴𝜏 𝑛𝜋𝜏
|𝐶𝑛 | = |𝑆𝑎( )|
𝑇𝑜 𝑇𝑜
θn=0 or ±180o , (since there is no imaginary part)
𝜏 1
a) τ=1, To=5, ⟹ =
𝑇𝑜 5
𝐴 𝑛𝜋
|𝐶𝑛 | = |𝑆𝑎( )|
5 5
2π Cn
A/5(2π)
Main lobe Finite zero crossing

point
Envelope Sa
function
nω0
-2ωo
ωo
2ωo
-ωo
-20ωo
-10ωo
-15ωo
-5ωo
5ωo
10ωo
20ωo
15ωo 6π/τ
2π/τ
4π/τ
Side lobes
θn
nω0
-2ωo
15ωo
-ωo
5ωo
-10ωo
2ωo
-15ωo
-5ωo
ωo
10ωo
-π
16
Cn
Cn
τ /T 0 =0.2; 5 comp
nf0 nf0
τ /T 0 =0.1; 10 comp
nf0 nf0
τ /T 0 =0.05; 20 comp
nf0 nf0
τ: fixed To: fixed

To:variable τ:variable
When τ is fixed:
As To increases (1) The amplitude decreases as 1/ To.
(2) Spacing between lines decreases as 2π/ To.
When To is fixed:
As To increases (1) The amplitude increases proportional to τ.
(2) The frequency content of the signal is compressed in narrower
range.
Parseval Power theorem:
The power of periodic signals can be computed in frequency domain rather than
time domain using the spectrum function:
∞
𝑃𝑎𝑣 = ∑ |𝐶𝑛 |2 Watt (2-15)
−∞
1
Note: if |𝐶𝑛 | in volt ⟹ 𝑃𝑎𝑣 = ∑∞
−∞|𝐶𝑛 |
2
𝑅
If |𝐶𝑛 | in ampere⟹ 𝑃𝑎𝑣 = 𝑅 ∑∞ 2

−∞|𝐶𝑛 | ,
For given R, if R=1, we use the form of equation (2-15).

17
Ex 2-6:
𝜏
For the rectangular pulse below if = 0.25 find;
𝑇𝑜
1- The total average power.
2- The ratio of average power in the first three harmonics to the total
average power.
Solution: f(t)
A
-T o To
t
-τ/2 τ/2 T o /2
To
1- Using time domain
𝜏
1 𝑇 1 𝜏
𝑃𝑎𝑣 = ∫ |𝑓(𝑡)|2 𝑑𝑡 = ∫ 𝐴2 𝑑𝑡 = 𝐴2 = 0.25 𝐴2 𝑤𝑎𝑡𝑡
2
𝜏
𝑇 0 𝑇 − 𝑇 𝑜
2
Using frequency domain

𝜏 𝜏 𝑛𝜋𝜏
𝐴 ∏(
𝑇𝑜
) ⟹ 𝐶𝑛 = 𝐴 𝑇 𝑆𝑎( 𝑇 ).. General formula for rectangular pulse
𝑜 𝑜
𝐶𝑛 = 0.25 𝐴 𝑆𝑎(0.25 𝑛𝜋)

|𝐶𝑛 | = 0.25 𝐴|𝑆𝑎(0.25 𝑛𝜋)| n |𝐶𝑛 |
∞ 0 0.25 A
𝑃𝑎𝑣 = ∑ |𝐶𝑛 |2 ±1 0.225 A
−∞ ±2 0.159 A
±3 0.075 A
2- |𝐶𝑛 | = 0.25 𝐴|𝑆𝑎(0.25 𝑛𝜋)|
𝑃3 =2|𝐶1 |2 + 2|𝐶2 |2 + 2|𝐶3 |2
Cn
=0.163 A2
𝑃3 0.163𝐴2 Co
= = 65.2 % C1
𝑃𝑇 0.25𝐴2 C2
C3
nfo
-3fo
-2fo
3fo
2fo
-fo
fo
18
H.W:
For the signal shown, find:
1- The total average power.
2- The average power in the fundamental frequency.
3- The average power in the first five harmonics.
4- The dc power.
5- The ratio of average power in the frequency range (0→ 3kHz) to the total
average power. f(t)
t
-3 -2 -1 0 1 2 3 m sec
3- Spectrum of non-periodic signals

The spectrum of any non-periodic signals (double sided) can be
obtained by plotting │F(ω)│verses ω and θ(ω) versus ω, where:
∞
𝐹(𝜔) = ∫ 𝑓(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 =Re+j Im (2-16)a

−∞
𝐼𝑚 Phase spectrum (2-16)b

𝜃(𝜔) = 𝑡𝑎𝑛−1
𝑅𝑒
|𝐹(𝜔)| = √𝑅𝑒 2 + 𝐼𝑚2 Amplitude spectrum (2-16)c
F(ω) is called Fourier Transform (F.T.). F(ω) can be transformed

back to time domain (f(t)) using Inverse F.T. (I.F.T) given by:
∞
1
𝑓(𝑡) = ∫ 𝐹(𝜔)𝑒 𝑗𝜔𝑡 𝑑𝜔
2𝜋 (2-17)
−∞
19
Ex 2-7:
Plot the double-sided amplitude and phase spectrum of the signal shown below:
Solution: f(t)
∞
𝐹(𝜔) = ∫−∞ 𝑓(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 1----
e-atu(t)
∞
= ∫−∞ 𝑒 −𝑎𝑡 𝑢(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
∞ 1
𝐹(𝜔) = ∫0 𝑒 −(𝑎+𝑗𝜔)𝑡 𝑑𝑡 = ;𝑎 > 0 t
𝑎+𝑗𝜔
1 𝜔
= −𝑗
𝑎2 +𝜔2 𝑎2 +𝜔2
1
𝐹(𝜔) =
√𝑎2 +𝜔2
𝜔
𝜃(𝜔) = −𝑡𝑎𝑛−1 ( )
𝑎
F(ω)
1/a ----
θ(ω)
π /2 ----
-π /2 ----
The spectrum of non-periodic signals is continuous.
20
Some Fourier Transform Properties:

Property 𝒇(𝒕) 𝑭(𝝎)
Linearity 𝑎1 𝑓1 (𝑡) + 𝑎2 𝑓2 (𝑡) 𝑎1 𝐹1 (𝜔) + 𝑎2 𝐹2 (𝜔)
1 𝜔
Scaling 𝑓(𝑎𝑡) 𝐹( )
|𝑎| 𝑎
Delay 𝑓(𝑡 − 𝑡𝑜 ) 𝑒 −𝑗𝜔𝑡𝑜 𝐹(𝜔)
Frequency translation 𝑒 𝑗𝜔𝑜 𝑡 𝑓(𝑡) 𝐹(𝜔 − 𝜔𝑜 )
1 1
Amplitude translation 𝑓(𝑡) cos 𝜔𝑜 𝑡 𝐹(𝜔 + 𝜔𝑜 ) + 𝐹(𝜔 − 𝜔𝑜 )
2 2
∞
Time convolution ∫ 𝑓1 (𝜏)𝑓2 (𝑡 − 𝜏) 𝑑𝜏 𝐹1 (𝜔)𝐹2 (𝜔)
−∞
1 ∞
Frequency convolution 𝑓1 (𝑡)𝑓2 (𝑡) ∫ 𝐹 (𝑢)𝐹1 (𝜔 − 𝑢)𝑑𝑢
2𝜋 −∞ 1
Duality 𝐹(𝑡) 2𝜋𝑓(−𝜔)
𝑑𝑛
Time differentiation 𝑓(𝑡) (𝑗𝜔)𝑛 𝐹(𝜔)
(𝑑𝑡)𝑛
𝑡 1
Time integration ∫ 𝑓(𝑡)𝑑𝜏 + 𝜋𝐹(0)𝛿(𝜔)
−∞ 𝑗𝜔
21
Selected F.T Pairs:
𝒇(𝒕) 𝑭(𝝎)
1
𝑒 −𝑎𝑡 𝑢(𝑡)
𝑎 + 𝑗𝜔
1
𝑡𝑒 −𝑎𝑡 𝑢(𝑡)
(𝑎 + 𝑗𝜔)2
2𝑎
𝑒 −𝑎|𝑡|
𝑎2 + 𝜔 2
2
𝑠𝑔𝑛(𝑡)
𝑗𝜔
𝑗
𝑠𝑔𝑛(𝜔)
𝜋𝑡
1
𝑢(𝑡) 𝜋𝛿(𝜔) +
𝑗𝜔
𝛿(𝑡) 1
1 2𝜋𝛿(𝜔)
𝑒 ±𝑗𝜔𝑜 𝑡 2𝜋𝛿(𝜔 ∓ 𝜔𝑜 )
𝑐𝑜𝑠𝜔𝑜 𝑡 𝜋[𝛿(𝜔 − 𝜔𝑜 ) + 𝛿(𝜔 + 𝜔𝑜 )]
𝑠𝑖𝑛𝜔𝑜 𝑡 −𝑗𝜋[𝛿(𝜔 − 𝜔𝑜 ) − 𝛿(𝜔 + 𝜔𝑜 )]

f(t)
𝜔𝜏 Rectangular
𝐴∏(𝑡⁄𝜏)
A
𝐴𝜏𝑆𝑎 ( )
2 pulse -τ τ
t
0
2 2
𝜔𝜏 2 Triangular f(t)
𝐴⋀(𝑡⁄𝜏) 𝐴𝜏 [𝑆𝑎 ( )] A
2 pulse
t
-τ 0 τ
22
Ex 2-8:
Find the Fourier Transform of the trapezoidal signal shown below:
Solution:
By differentiating the signal twice, we got:
𝐴𝑡 −𝐴𝑡
The line equations are 𝑔(𝑡) = ,and 𝑔2 (𝑡) =
𝑏−𝑎 𝑏−𝑎
f(t) F(ω )
A
-b -a a b
t
𝑑 𝐴 𝑢(𝑡 + 𝑏) − 𝑢(𝑡 + 𝑎)
𝑓(𝑡) = [ ]
𝑑𝑡 𝑏 − 𝑎 −𝑢(𝑡 − 𝑎) + 𝑢(𝑡 − 𝑏)
-a a
t
-b b
𝑑2 𝐴 𝛿(𝑡 + 𝑏) − 𝛿(𝑡 + 𝑎)
𝑓(𝑡) = [ ]
𝑑𝑡 2 𝑏−𝑎 −𝛿(𝑡 − 𝑎) + 𝛿(𝑡 − 𝑏)
Taking F.T
--
t
-b -a a b
--
(𝑗𝜔)2 𝐹(𝜔) = [ 𝑒 𝑗𝜔𝑏 − 𝑒 𝑗𝜔𝑎 ]
𝐴
𝑏−𝑎 −𝑒 −𝑗𝜔𝑎 + 𝑒 −𝑗𝜔𝑏
2𝐴 cos(𝑎𝜔)−cos(𝑏𝜔)
∴ 𝐹(𝜔) = [ ]
𝑏−𝑎 𝜔2
23
Rayleight’s Energy Theorem.

It corresponds to Parseval’s power theorem but concern with energy of non-
periodic signals.
∞ ∞
1
𝐸 = ∫ |𝑓(𝑡)|2 𝑑𝑡 = ∫ |𝐹(𝜔)|2 𝑑𝜔 (2-18)
2𝜋
−∞ −∞
Ex 2-9:
Find the ratio of Energy up to the first zero crossing point to the total energy
for the rectangular pulse of amplitude A and width τ. f(t)
Solution:
∞ 𝜏/2 A
𝐸𝑡 = ∫−∞|𝑓(𝑡)|2 𝑑𝑡 = ∫−𝜏/2 𝐴2 𝑑𝑡 = 𝐴2 𝜏 joule t
-τ/2 τ/2
2𝜋
The first zero crossing point for 𝐴Π(τ/T) occurs at 𝜔 = rad/sec
𝜏
2𝜋
𝜏 F(ω) =A τ Sa (ωτ/2)
1
𝐸2𝜋 = ∫ |𝐹(𝜔)|2 𝑑𝜔
𝜏 2𝜋 From Table
2𝜋
𝜏
2𝜋 2
1 𝜔𝜏
=
2𝜋
∫ (𝐴𝜏)2 |𝑆𝑎 ( 2 )| 𝑑𝜔
2𝜋
𝜏
= 0.92 A2τ
𝐸2𝜋
𝜏 0.92𝐴2 𝜏
= ω
𝐸𝑡 𝐴2 𝜏 -6π/τ -4π/τ -2π/τ 2π/τ 4π/τ 6π/τ
= 92 %
H.W:
Calculate the ratio of energy of the frequency range (10→20) Hz up to total energy
|𝑡|, −0.5 < 0 < 0.5
for the signal 𝑓(𝑡) = {
0, 𝑒𝑙𝑠𝑒 𝑤ℎ𝑒𝑟𝑒
24
System:
A group of objects that can interacts harmoniously and combined in a manner
intended to achieve a desired objective.
Liner system:
The system that satisfies the following:
Signal Response
If g1(t) is the output when f1(t) is the input and f(t) g(t)
System
g2(t) is the output when f2(t) is the input, then,
when the input is f1(t)+ f2(t),
the output would be a1 g1(t)+ a2 g2(t).
System
g(t)=f(t)*h(t)
System Impulse Response: f(t)
h(t)
When an input signal is applied to a system,
δ(t) h(t)
then the output would be; System
𝑔(𝑡) = 𝑓(𝑡) ∗ ℎ(𝑡) = ∫ 𝑓(𝜏)ℎ(𝑡 − 𝜏) 𝑑𝜏 (2-19)

−∞
Where h(t) is called the impulse response of the system, h(t) is found by
applying δ(t) at the system input, where the output would be h(t),
𝑔(𝑡) = ℎ(𝑡) when the input is δ(t).
Ex 2-10:
f(t) + y(t) g(t)
Determine the impulse response + ʃ
-
of the system shown below. Delay
Solution: to
Let 𝑓(𝑡) = 𝛿(𝑡)
25
𝑦(𝑡) = 𝛿(𝑡) − 𝛿(𝑡 − 𝑡𝑜 )

h(t)
𝑡 𝑡
1
𝑔(𝑡) = ∫ 𝑦(𝑡)𝑑𝑡 = ∫ [𝛿(𝑡) − 𝛿(𝑡 − 𝑡𝑜 )]𝑑𝑡
−∞ −∞
t
ℎ(𝑡) = 𝑢(𝑡) − 𝑢(𝑡 − 𝑡𝑜 ) t0
Ex 2-11:
Find the output of the system shown below:
Solution: g(t)
f(t)=Asinπtu(t) h(t)
𝑓(𝑡) = 𝐴 𝑠𝑖𝑛𝜋𝑡 𝑢(𝑡)
ℎ(𝑡) = 𝛿(𝑡) − 𝛿(𝑡 − 2) A h(t)
𝑔(𝑡) = 𝑓(𝑡) ∗ ℎ(𝑡) t
1 2 3 4 +1
∞
= ∫−∞ 𝑓(𝜏)ℎ(𝑡 − 𝜏)𝑑𝜏 2
t
-1
∞
= ∫−∞[𝐴 𝑠𝑖𝑛𝜋𝑡 𝑢(𝑡)][𝛿(𝑡 − 𝜏) − 𝛿(𝑡 − 2 − 𝜏)] 𝑑𝜏
∞ ∞
= ∫ 𝐴 𝑠𝑖𝑛𝜋𝜏 𝑢(𝑡)𝛿(− (𝜏 − 𝑡))𝑑𝜏 − ∫ 𝐴 𝑠𝑖𝑛𝜋𝜏 𝑢(𝑡)𝛿(− (𝜏 − (𝑡 − 2))𝑑𝜏
−∞ −∞
=𝐴 𝑠𝑖𝑛𝜋 𝑡 𝑢(𝑡) − 𝐴 𝑠𝑖𝑛𝜋 𝑡(𝑡 − 2) 𝑢(𝑡 − 2)
0, 𝑡<0
∴ 𝑔(𝑡) = {𝐴𝑠𝑖𝑛𝜋𝑡, 0 < 𝑡 < 2 g(t)
0, 𝑡>2
1 2 t
26
H.W:
Find the output of the following system (plot the waveforms)
1)
System
g(t)
f(t) h(t)
1
h(t)
t 3 --
-1 +1
t
0 3
2)
System
f(t) g(t)
h(t)
4
h(t)
2
t 1 1
+2
t
0 1 2
Step response
The output of the system when the input is unit step u(t).
𝑑
ℎ(𝑡) = [𝑔𝑢 (𝑡)]… (2-20)
𝑑𝑡
System
u(t) h(t) gu(t)
27
Solution of H.W1:
f(t) h(τ )
∞
𝑔(𝑡) = 𝑓(𝑡) ∗ ℎ(𝑡) = ∫−∞ 𝑓(𝜏)ℎ(𝑡 − 𝜏) 𝑑𝜏1 3 --
τ
τ τ
-1 +1 0 3
h(-τ )
τ
-3 0
h(t-τ )
1) When 𝑡 ≤ −1 τ
t-3 t
Replica
𝑔(𝑡) = 0
2) When −1 < 𝑡 ≤ 1
𝑡
𝑔(𝑡) = ∫−1(𝑡 − 𝜏)(1)𝑑𝜏
1
= (𝑡 + 1)2
2
3) When 1 < 𝑡 ≤ 2
1
𝑔(𝑡) = ∫−1(𝑡 − 𝜏)(1)𝑑𝜏
= 2𝑡
4) When 1 < 𝑡 < 2
1
𝑔(𝑡) = ∫𝑡−3(𝑡 − 𝜏)(1)𝑑𝜏
1
= 4 + 𝑡 − 𝑡2
2
5) When 𝑡 ≤ −1
𝑔(𝑡) = 0
g(t)
-3 -2 -1 0 1 2 3 4 t
28
System Transfer Function:

It is the F.T of the system impulse response
𝐻(𝜔) = 𝐹[𝑓(𝑡)]
… (2-21)
Using the time convolution f(t) h(t) 𝑔(𝑡) = 𝑓(𝑡)⨂ℎ(𝑡)

property of Fourier transform, T.D
we have: F(ω) H(ω) 𝐺(𝜔) = 𝐹(𝜔)𝐻(𝜔)
F.D
𝐺(𝜔) = 𝐹(𝜔). 𝐻(𝜔)
… (2-22)
The amplitude and phase spectrum for g(t) would be:
|𝐺(𝜔)| = |𝐹(𝜔)|. |𝐻(𝜔)|

… (2-23)
𝜃𝑔 (𝜔) = 𝜃𝑓 (𝜔) + 𝜃ℎ (𝜔)

… (2-24)
Equivalent Frequency Transfer Function:
Cascaded Series: H(ω)
F(ω) H1(ω) H2(ω) G(ω)
𝐻𝑒𝑞 (𝜔) = 𝐻1 (𝜔)𝐻2 (𝜔) ...(2-25)
Parallel:
H1(ω)
F(ω) + G(ω)
H2(ω)
𝐻𝑒𝑞 (𝜔) = 𝐻1 (𝜔) + 𝐻2 (𝜔) ...(2-26)
29
Feedback:
+
F(ω) + H1(ω) G(ω)
±
H2(ω)
𝐻1 (𝜔)
𝐻𝑒𝑞 (𝜔) = ...(2-26)
1±𝐻1 (𝜔)𝐻2 (𝜔)
Ex: 2-12:
Calculate the energy of input and output signal for the system shown below:
Solution:
Idial
For the input signal f(t) g(t)
LPF
=e-5t u(t)
Gain=1
∞ BW=5 rad/sec
𝐸𝑓 = ∫−∞|𝑓(𝑡)|2 𝑑𝑡
∞ H(ω)
= ∫0 𝑒 −10𝑡 𝑑𝑡 = 0.1 𝑗𝑜𝑢𝑙𝑒
1
ω
-5 +5
For the output signal

1 ∞ 1 ∞
𝐸𝑔 = ∫ |𝐺(𝜔)|2 𝑑𝜔 = 2𝜋 ∫−∞|𝐹(𝜔)|2 |𝐻(𝜔)|2 𝑑𝜔
2𝜋 −∞
1 1
𝑒 −𝑎𝑡 𝑢(𝑡) ⟺ ⇒ 𝐹[𝑒 −5𝑡 𝑢(𝑡)] =
𝑎 + 𝑗𝜔 5 + 𝑗𝜔
1 5 1
𝐸𝑔 = ∫ 𝑑𝜔 = 0.05 𝑗𝑜𝑢𝑙𝑒
2𝜋 −5 25 + 𝜔 2
30
H.W: A signal 𝑓(𝑡) = 2𝑒 −𝑎𝑡 𝑢(𝑡) is applied to the input of system have frequency
𝑏
transfer function |𝐻(𝜔)| = determine the required relations between
√𝜔2 +𝑎2
the constants a and b such as exactly 50% of the input energy on a 1-ohm
basis, is transferred to the output.
Note:
Is some systems, H(ω) can be found using:
𝑍𝑜𝑢𝑡
𝐻(𝜔) = … (2-28)
𝑍𝑖𝑛
Ex 2-13:
Find H(ω) for the system shown below:
1 R
𝑍𝑖𝑛 = 𝑅 +
𝑗𝜔𝑐
1 f(t) C g(t)
𝑍𝑜𝑢𝑡 =
𝑗𝜔𝑐
1
𝑍𝑜𝑢𝑡 𝑗𝜔𝑐 1 H(ω)
𝐻(𝜔) = = 1 =
𝑍𝑖𝑛 𝑅+ 1+𝑗𝜔𝑅𝑐
𝑗𝜔𝑐
31
Spectral Density and Correlation:

Energy Spectral Density (ESD):
It shows the distribution of energy at each frequency component of
nonperiodic signal.
𝜓𝑓 (𝜔) = |𝐹(𝜔)|2
joule/Hz … (2-29)
To find the total energy from the spectrum, we use:
1 ∞
𝐸= ∫ 𝜓 (𝜔) 𝑑𝜔 joule … (2-30)
2𝜋 −∞ 𝑓
Power Spectral Density (PSD):

It shows the distribution of power at each frequency component of periodic
signal.
𝑆𝑓 (𝜔) = 2𝜋 ∑ |𝐶𝑛 |2 𝛿(𝜔 − 𝑛𝜔𝑜 ) Watt/Hz … (2-31)

𝑛=−∞
To find the total power from the spectrum, we use:
1 ∞
𝑃= ∫ 𝑆 (𝜔) 𝑑𝜔 Watt … (2-32)
2𝜋 −∞ 𝑓
Note :
Power spectral density exists for deterministic and random signals, such as noise.
32
Ex 2-14:
A given voltage signal 𝑓(𝑡) = cos 20 𝜋𝑡 + 2 cos 30 𝜋𝑡 across 2 Ω resistor.

a) Determine PSD of f(t).
b) Sketch Sf(ω).
c) Calculate the average power [(i) using time domain, (ii) using spectral
density]
Cn
(a) 𝑆𝑓 (𝜔) = 2𝜋 ∑∞ 2
𝑛=−∞|𝐶𝑛 | 𝛿(𝜔 − 𝑛𝜔𝑜 )
= 2𝜋(1)2 𝛿(𝜔 + 30𝜋) 2 2

1 1
+2𝜋(2)2 𝛿(𝜔 + 20𝜋)
nf0
+2𝜋(2)2 𝛿(𝜔 − 20𝜋) -15 -10 0 10 15
Sf (ω)
+2𝜋(2)2 𝛿(𝜔 − 30𝜋)
8π 8π
2π 2π
(b)
nω0
-30π -20π 0 20π 30π
1 𝑇 42 22
(c) −(𝑖): 𝑃𝑎𝑣 = ∫0 |𝑓(𝑡)|2 𝑑𝑡 = + = 8 + 2 = 10 𝑣𝑜𝑙𝑡 2 /𝑅
𝑇 2 2
10
𝑃𝑎𝑣 = =5 watt
2
1 ∞ 2 ∞
(𝑖𝑖): 𝑃𝑎𝑣 = ∫ 𝑆 (𝜔)𝑑𝜔 = 2𝜋 ∫0 𝑆𝑓 (𝜔)𝑑𝜔
2𝜋 −∞ 𝑓
1 ∞
=
𝜋
∫0 [8𝜋𝛿(𝜔 − 20𝜋) + 2𝜋𝛿(𝜔 − 30𝜋)] 𝑑𝜔
1
= (8𝜋 + 2𝜋) = 10 𝑣𝑜𝑙𝑡 2 /𝑅
𝜋
10𝑣 2
𝑃𝑎𝑣 = =5 watt (the same result)
2Ω
33
Correlation:
It is the inverse Fourier Transform of the power spectral density. It is a
measure of similarity between two signals or a signal and its replica shifted by 𝜏
seconds.
𝑅𝑓 (𝜏) = 𝐹 −1 {𝑆𝑓 (𝜔)}

Watt … (2-33)
Cross Correlation:
∞
𝑅𝑥𝑦 (𝜏) = ∫−∞ 𝑥(𝑡)𝑦(𝑡 + 𝜏)𝑑𝑡 𝑁𝑜𝑛𝑝𝑒𝑟𝑖𝑜𝑑𝑖𝑐 𝑠𝑖𝑔𝑛𝑎𝑙𝑠 …(2-34)a
1 𝑇
𝑅𝑥𝑦 (𝜏) = ∫0 𝑥(𝑡)𝑦(𝑡 + 𝜏)𝑑𝑡 𝑃𝑒𝑟𝑖𝑜𝑑𝑖𝑐 𝑠𝑖𝑔𝑛𝑎𝑙𝑠 …(2-34)b
𝑇
Auto Correlation:
∞
𝑅𝑥 (𝜏) = ∫−∞ 𝑥(𝑡)𝑥(𝑡 + 𝜏)𝑑𝑡 𝑁𝑜𝑛𝑝𝑒𝑟𝑖𝑜𝑑𝑖𝑐 𝑠𝑖𝑔𝑛𝑎𝑙𝑠 …(2-35)a
1 𝑇
𝑅𝑥 (𝜏) = ∫0 𝑥(𝑡)𝑥(𝑡 + 𝜏)𝑑𝑡 𝑃𝑒𝑟𝑖𝑜𝑑𝑖𝑐 𝑠𝑖𝑔𝑛𝑎𝑙𝑠 …(2-35)b
𝑇
Properties of Correlation:
(1) When 𝜏 = 𝑜
𝑅𝑓 (0) = 𝐸 𝑓𝑜𝑟 𝑒𝑛𝑒𝑟𝑔𝑦 𝑠𝑖𝑔𝑛𝑎𝑙𝑠
𝑅𝑓 (0) = 𝑃𝑎𝑣 𝑓𝑜𝑟 𝑝𝑜𝑤𝑒𝑟 𝑠𝑖𝑔𝑛𝑎𝑙𝑠
(2) 𝑅𝑓 (𝜏) ≤ 𝑅𝑓 (0)
(3) If 𝑧(𝑡) = 𝑥(𝑡) + 𝑦(𝑡) 𝑡ℎ𝑒𝑛,
𝑅𝑧 (𝜏) = 𝑅𝑥 (𝜏) + 𝑅𝑥𝑦 (𝜏) + 𝑅𝑦𝑥 (𝜏) + 𝑅𝑦 (𝜏)
34
Ex 2-15:
Determine and sketch the autocorrelation function of periodic square wave shown
below:
Solution:
f(t)
A
Replica
Edges t
-3T -T -T 0 T T 3T
-T T
τ- T
4
T
4 +τ 4 2 4 4 2 4
1 𝑇 τ- T T
𝑅𝑓 (𝜏) = ∫0 𝑓(𝑡)𝑓(𝑡 + 𝜏)𝑑𝑡 4 4 +τ
𝑇 f(t)
–𝑇
(1) When <𝜏<0 A
2 A
𝑇 t
1 +𝜏 -T 0 T
4 2 -T T
𝑅𝑓 (𝜏) = ∫ 𝑇 𝐴 𝑑𝑡 4 4
𝑇 −4
1 𝜏
= 𝐴2 ( + ) τ- T T
+τ
2 𝑇 4 4
f(t)
–𝑇
(2) When 0<𝜏< A
2 A
𝑇 t
1 4 2 -T 0 T
𝑅𝑓 (𝜏) = ∫ 𝑇 𝐴 𝑑𝑡
-T
4 4
T
𝑇 𝜏− 4
1 𝜏 Rf(τ)
= 𝐴2 ( − )
2 𝑇 1 2
A
2
0
t
-3T -T -T T T 3T
2 2 2 2
The autocorrelation is useful for the detection of signals, in which masked by

additive noise, see the following figures.
35
H.W:
A sinusoidal waveform, 3√2𝑐𝑜𝑠𝜔1 𝑡, is added to a second 4√2𝑐𝑜𝑠𝜔2 𝑡,
determine the rms value of the sum, if (a) 𝜔1 = 𝜔2 , (b) 𝜔1 ≠ 𝜔2
Ans: a=7, b=5
H.W:
For the system shown below, find:
a) 𝑔(𝑡)
b) Average power at the system 𝑖/𝑝 & 𝑜/𝑝.
c) PSD of 𝑓(𝑡) and 𝑔(𝑡).
d) Average power at system 𝑖/𝑝 using 𝑃𝑎𝑣 = 𝑅𝑓 (0)
H(ω)
f(t)
1
t ω
-2π -3π -π -π 0 π π 3π 2π -7 0 7
2 2 2 2
θh(ω)
π
2
-π
2
36
Problem Sheet of Signal Analysis

Q1: Sketch the signal and double amplitude and phase spectrum of the following
signals:
(a) 𝑓(𝑡) = −7 sin(3𝜋𝑡) − 5cos(6𝜋𝑡 + 90𝑜 )

(b) 𝑓(𝑡) = −4 sin(106 𝜋𝑡) + 8cos(107 𝜋𝑡 + 170𝑜 )
(c) 𝑓(𝑡) = ∑3𝑛=0(−0.5)𝑛 𝑐𝑜𝑠[𝑛(𝜔𝑜 𝑡 + 10𝑜 )]
Q2: If 𝑓(𝑡) is a periodic signal in the period − 𝜏⁄2 < 𝑡 < 𝜏⁄2 and` is given by:
𝑓(𝑡) = 2𝑡;
Final the double-sided spectrum and the ratio of the power in first three
harmonics to the total average power of the signal.
Q3: Sketch the two – sided amplitude and phase spectrum of the signals shown
below.
f(t)
Vm
Half wave rectified

sinsoidal signal
a t
-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5
f(t)
b t
-τ 0 τ
f(t)
c t
-τ/2 0 τ/2
𝑡
𝑉𝑐𝑜𝑠𝜋 |𝑡| < 𝜏⁄2
𝑓(𝑡) = { 𝜏
0 𝑒𝑙𝑒𝑠𝑒 𝑤ℎ𝑒𝑟𝑒
37
f(t)
d t
-τ -τ/2 0 τ/2 τ
𝑉 𝜋𝑡
(1 − 𝑐𝑜𝑠𝜋 𝜏 ) |𝑡| < 𝜏
𝑓(𝑡) = { 2
0 𝑒𝑙𝑒𝑠𝑒 𝑤ℎ𝑒𝑟𝑒
Q4: The complex exponential Fourier series of a given signal 𝑓(𝑡) over the
interval (0, T) is:
∞
3
𝑓(𝑡) = ∑ 𝑒 𝑗𝜋𝑛𝑡
4 + (𝜋𝑛)2
𝑛=−∞
a) Determine the numerical value of 𝑇.

b) What is the mean (average) value of 𝑓(𝑡) over the given interval.
c) The component of 𝑓(𝑡) at a certain frequency can be expressed as 𝑎𝑐𝑜𝑠3𝜋𝑡.
Determine the numeric value of a.
3 6
Ans: 2 𝑠𝑒𝑐, ,
4 4+9𝜋2
Q5: Write a representation of 𝑓(𝑡) shown below as a complex exponential fourier

series. f(t)
0 1 2 Over interval
a t
(0,2)
f(t)
b Over interval
t
-4 2 0 2 4 (-4, 4)
38
2
Ans: (a) 𝑓(𝑡) = ∑∞
𝑛=−∞ 𝑒 𝑗𝑛𝜔𝑜 𝑡
𝑗𝑛𝜋
𝜋 2𝜋 𝜋 2𝜋
1 𝜋 1 1
(b) 𝑓(𝑡) = ( + 𝑒 𝑗 4 𝑡 − 𝑒 𝑗 4 𝑡 + ⋯ + 𝑒 −𝑗 4 𝑡 − 𝑒 −𝑗 4 𝑡 + ⋯ )
𝜋 2 3 3
Q6: Represent 𝑓(𝑡) = 𝑡 2 in a trigonometric Fourier series over the interval (0.2).
4 4 4 1
Ans: 𝑓(𝑡) = + ∑∞ ∞
𝑛=1 𝑐𝑜𝑠 𝑛𝜋𝑡 − ∑𝑛=1 𝑠𝑖𝑛 𝑛𝜋𝑡
3 𝜋2 𝜋 𝑛
𝜔2 +2
Q7: For a signal 𝑓(𝑡), 𝐹(𝜔) = (1 − 𝑗𝜔) + 𝜋𝛿(𝜔)
𝜔4 +3𝜔3 +2
Determine:
(a) The average (dc) value of the signal.
(b) The signal 𝑓(𝑡).

1
Ans: (a) 1/2 (b) + 𝑒 −𝑡 𝑢(𝑡)
2
Q8: Final the Fourier transform of the pulse waveform shown in fig. below using
only the transform of an impulse together with appropriate transform properties.
f(t)
t
0 τ 2τ
Q9: A certain function of time, 𝑓(𝑡), has a Fourier transform shown in Fig. below.
Sketch the Fourier transform of
F(ω)
(a) 𝑓(2𝑡) (b) [𝑓(𝑡)]2
4
ω
-10 0 10
39
Q10: A given periodic signal (in volts),
𝑓(𝑡) = 4 sin 8𝜋𝑡 + 2 cos 12𝜋𝑡 Volts developed across Ωresistor.
(a) Determine the highest fundamental frequency possible for the signal.
(b) Sketch two-sided amplitude and phase spectrum of 𝑓(𝑡).
(c) Find the total mean (average) power in 𝑓(𝑡).
(d) Calculate the percentage of the total mean power contained in each
harmonic up to the fifth harmonic using the fundamental of the part (a).
Ans: (a) 4π (c) 10 watt (d) n %

0 0
1 0
2 80
3 20
4 0
Q11: Determine the required numerical value of the positive real constant a if it is
given that one – half the energy in 𝑓(𝑡) = 𝑒𝑥𝑝 (−𝑎𝑡) 𝑢 (𝑡) lies in the spectrum
range from zero to one 𝐻𝑧.
Ans: 2𝜋
Q12: Find the ratio of average power in the spectral range 0 → 2 𝐻𝑧to the total
average power of the signal shown below.
f(t) Ans: 86.55%.
-2 -1 1 2 3 4
t
40
Q13: The two – sided exponential voltage 𝑓(𝑡) = 10𝑒 −|𝑡| volt is developed across
a 50 Ω resistor.
(a) Calculate the total energy dissipated in the resistor.

(b) What fraction of this energy is in the frequency range of 0 – 1 rad/ sec?
Ans: (a) 2 joule (b) 40%.
Q14: Evaluate 𝑓1 (𝑡)⨂𝑓1 (𝑡) for the functions shown

f1(t) f2(t)
Ans:
2 2
4(t+1), -1<t 1
t t
0 2 -1 0 1
4(3-t), 1<t 3
f1(t) f2(t)
2 2 t(4-t), 0 <t 2
(t-4)2 , -1<t 1
t t
0 2 0 2
f1(t) f1(t)
2 2 -4t, 0<t 1
12t-16, 1<t 2
t t
0 1 2 0 1 2 2<t 3
-2 -2
4(t-4), 3<t 5
Q15: The spectral density of the input to a given linear time invariant system is:
𝐹(𝜔) = [𝑒𝑥𝑝(−𝑗𝜋𝜔)]/(1 + 𝑗𝜔)
The corresponding output spectral density is:
𝐺(𝜔) = [𝑒𝑥𝑝(−𝑗2𝜋𝜔)]/(1 − 𝜔2 + 2𝑗𝜔).
(a) Determine the transfer function of the system.

(b) Determine the impulse response of the system.
41
Q16: Determine the magnitude of the – frequency transfer function of the transversal
filter, shown in Fig. below:
f(t) T T
1 2 1
g(t)
𝜔𝑇
Ans: |𝐻(𝜔)| = 4𝑐𝑜𝑠 2 ( )
2
Q17: Determine the output, g (t), of a linear time – invariant system whose input and
frequency transfer function are shown in Fig. below:
f(t)
H(ω)
2 1
t ω
-1.0 -0.75 -0.5 -0.25 0 0.25 0.5 0.75 1.0 0.5 -3π 0 3π
4
Ans:𝑔(𝑡) = 1 + cos 2𝜋𝑡, 𝑃𝑓 = 2𝑤, 𝑃𝑔 = 1.811 𝑤.
𝜋
Q18: In frequency, transfer function of a given system is shown in Fig. below. If the
input signal to this system is: 𝑓(𝑡) = 𝑆𝑎(𝑡 − 𝑡𝑜 ), sketch the spectral density of
the corresponding output signal.
H(ω) θh(ω)
1
π/2
ω ω
-2 2
-π/2
42
Q19: The bandwidthe of a given ideal LPE is 4 rad/ sec and the low frequency gain
is one. Calculate the output energy, on one ohm basis, if the input f(t):
(a) 𝑓(𝑡) = 𝛿(𝑡) Ans: (a) 4⁄𝜋 joule
(b) 𝑓(𝑡) = 𝑒 −4𝑡 𝑢(𝑡) (b) 1⁄16 joule
Q20: In the system shown below. If 𝐻1 (𝜔) = 10𝑒 −𝑗2𝜔 𝑎𝑛𝑑 𝐻2 (𝜔) = 𝑒 −𝑗2𝜔 .
(a) Find the step response of the system.

(b) If 𝑓(𝑡) = cos 10𝜋𝑡 + 3 cos 20𝜋𝑡, find 𝑔(𝑡).
f(t) + H1(ω) g(t)
H2(ω)
10
Ans: (a) 𝑔𝑢 (𝑡) = 𝑢(𝑡 − 2)
11
10 30
(b) 𝑔(𝑡) = [ 𝑐𝑜𝑠10𝜋(𝑡 − 2) + 𝑐𝑜𝑠20𝜋(𝑡 − 2)] 𝑢(𝑡 − 2)
11 11
Q21: A certain signal 𝑓(𝑡) has the following PSD (assume 1Ω load):
𝑆𝑓 (𝜔) = 𝜋[𝑒 −|𝜔| + 𝛿(𝜔 − 2) + +𝛿(𝜔 + 2)]
(a) What is the mean power in the bandwidth 𝜔 ≤ 1 rad/ sec?
(b) What is the mean power in the bandwidth 0.99 𝑡𝑜 1.01 rad/sec?
(c) What is the mean power in the bandwidth 1.99 𝑡𝑜 2.01 rad/ sec?
43
(d) What is the total mean power in (𝑡) ?
Ans: (a) 0.632 w (b) 7.36 mw.
(c) 1.0027 w (d) 2 w.
Q22: A source with a resistance of 6 k  resistor. Compute the mean power across
the 6-kΩ resistor if the power spectral density of the source (in watt per Hz) is
(a) 10−2 , (b) 𝑆 (𝜔 + 10) + 𝑆 (𝜔– 10)
Ans: (a) 3.7 μw . (b) 11.79 μw.
Q23: A given voltage signal 𝑓(𝑡) = 4𝑐𝑜𝑠 2 20𝜋𝑡 + 2𝑐𝑜𝑠30𝜋𝑡 across 1Ω.
(a) Determine PSD of 𝑓(𝑡).
(b) Sketch 𝑆𝑓 (𝜔).
(c) Calculate the mean (average) power, both in the time domain and in frequency
domain, that is dissipated by 𝑓(𝑡) across the 1Ω resistor.
Ans: (c) 8 watt.
Q24: Sketch the output of the filter whose frequency transfer function is shown
in Fig. below, if the input signals are:
H(ω)
(a) 𝑓(𝑡) = 2𝑢(𝑡)
(b) 𝑓(𝑡) = 10 𝛿(𝑡) 2
ω
-10 0 10
θn(ω)
-ω
44
Q25: In the system shown below, if the PSD of g(t)
𝜂
is : ⁄𝐴 watt/Hz when −2𝜋𝐵 ≤ 𝜔 ≤ 2𝜋𝐵
(a) Sketch PSD of 𝑦(𝑡).

f(t) d/dt g(t)
(b) Find the average power at system output

Ans: (a) 𝑆𝑦 (𝜔) = 𝜔2 𝑆𝑔 (𝜔).
8𝜋2𝜂𝐵3
Q26: For the cascaded systems in Fig. below: (b) watt.
3𝐴2
𝜋
𝐻1 (𝜔) = |3𝜔|∠ − , when 0 < 𝜔 ≤ 10 rad/sec
2
𝜋
𝐻2 (𝜔) = |𝜔 + 1|∠ − 𝜔 − , when 0 < 𝜔 ≤ 20 rad/sec
2
If the power spectral density of f (t) is 10(1 + 𝜔⁄10) watt/Hz in the range
0 < 𝜔 ≤ 30 rad/ sec.
(a) Find and sketch magnitude and phase 𝐻𝑒𝑞 (𝜔).
(b) Compute input and output power of the system.
f(t) H1(ω) H2(ω) g(t)
27: A system is governed by the differential equation:

𝒅𝒈(𝒕) 𝒅𝒇(𝒕)
+ 𝒂𝒈(𝒕) = 𝒃 + 𝒄𝒇(𝒕)
𝒅𝒕 𝒅𝒕
(a) Find and plot |𝐻(𝑓)| and ∠|𝐻(𝑓)| for c = 0.

(b) Find and plot |𝐻(𝑓)| and ∠|𝐻(𝑓)| for b = 0.
Q28: Determine the autocorrelation function of the pulse waveform
𝑒 −𝑎𝑡 [𝑢(𝑡) − 𝑢(𝑡 − 𝜏)] when it is repeated periodically every 2𝑇 seconds.
1
Ans: 𝑅𝑓 (𝜏) = [𝑒 3𝑎𝜏 − 𝑒 𝑎𝜏 𝑒 −2𝑎𝑇 ] 𝑤ℎ𝑒𝑛 − 𝑇 < 𝜏 < 0
4𝑎𝑇
𝑒 −𝑎𝜏
𝑅𝑓 (𝜏) = [𝑒 −2𝑎𝑇 − 𝑒 −2𝑎𝜏 ] 𝑤ℎ𝑒𝑛 0<𝜏<𝑇
−4𝑎𝑇
45
Students Notes:
46

Communications Chapter Two PDF

Uploaded by

Copyright:

Available Formats

Communications Chapter Two PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Communications Chapter Two PDF

Uploaded by

Copyright:

Available Formats

CHAPTER TWO

Communication System Elements

 A signal is said to be deterministic if it could be expressed mathematically as

Signal energy 𝐸 = ∫ |𝑓(𝑡)|2 𝑑𝑡 joul/c (2-2)

1. For a given value of resistance R.

2. Periodic signal is almost power signal.

=1/2 watt ‫ ؞‬power signal

Some important signals

Solution: = 𝜋 2 𝑒 −𝑠𝑖𝑛𝜋 cos 2𝜋 = 𝜋 2 . 𝑒 0 . 1 = 𝜋 2

H.W: Show that, 𝑠𝑖𝑛𝑐(0) = 𝑆𝑎(0) = 1

b) Double sideband spectrum:

𝐴 cos 2𝜋𝑓0 𝑡 ⟹ 𝐴 cos(2𝜋𝑓0 𝑡 − 1800 )

If the x-axis is ω ( not f), the amplitude is multiplied by 2π.

2- Spectrum of periodic signals

We conclude that the spectrum of periodic signals is discrete.

Main lobe Finite zero crossing

τ: fixed To: fixed

Parseval Power theorem:

If |𝐶𝑛 | in ampere⟹ 𝑃𝑎𝑣 = 𝑅 ∑∞ 2

For given R, if R=1, we use the form of equation (2-15).

Using frequency domain

𝐶𝑛 = 0.25 𝐴 𝑆𝑎(0.25 𝑛𝜋)

3- Spectrum of non-periodic signals

𝐹(𝜔) = ∫ 𝑓(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡 =Re+j Im (2-16)a

𝐼𝑚 Phase spectrum (2-16)b

|𝐹(𝜔)| = √𝑅𝑒 2 + 𝐼𝑚2 Amplitude spectrum (2-16)c

F(ω) is called Fourier Transform (F.T.). F(ω) can be transformed

The spectrum of non-periodic signals is continuous.

Some Fourier Transform Properties:

Linearity 𝑎1 𝑓1 (𝑡) + 𝑎2 𝑓2 (𝑡) 𝑎1 𝐹1 (𝜔) + 𝑎2 𝐹2 (𝜔)

Delay 𝑓(𝑡 − 𝑡𝑜 ) 𝑒 −𝑗𝜔𝑡𝑜 𝐹(𝜔)

Frequency translation 𝑒 𝑗𝜔𝑜 𝑡 𝑓(𝑡) 𝐹(𝜔 − 𝜔𝑜 )

Duality 𝐹(𝑡) 2𝜋𝑓(−𝜔)

Selected F.T Pairs:

𝑐𝑜𝑠𝜔𝑜 𝑡 𝜋[𝛿(𝜔 − 𝜔𝑜 ) + 𝛿(𝜔 + 𝜔𝑜 )]

𝑠𝑖𝑛𝜔𝑜 𝑡 −𝑗𝜋[𝛿(𝜔 − 𝜔𝑜 ) − 𝛿(𝜔 + 𝜔𝑜 )]

Rayleight’s Energy Theorem.

𝑔(𝑡) = 𝑓(𝑡) ∗ ℎ(𝑡) = ∫ 𝑓(𝜏)ℎ(𝑡 − 𝜏) 𝑑𝜏 (2-19)

Let 𝑓(𝑡) = 𝛿(𝑡)

𝑦(𝑡) = 𝛿(𝑡) − 𝛿(𝑡 − 𝑡𝑜 )

=𝐴 𝑠𝑖𝑛𝜋 𝑡 𝑢(𝑡) − 𝐴 𝑠𝑖𝑛𝜋 𝑡(𝑡 − 2) 𝑢(𝑡 − 2)

u(t) h(t) gu(t)

System Transfer Function:

Using the time convolution f(t) h(t) 𝑔(𝑡) = 𝑓(𝑡)⨂ℎ(𝑡)

|𝐺(𝜔)| = |𝐹(𝜔)|. |𝐻(𝜔)|

𝜃𝑔 (𝜔) = 𝜃𝑓 (𝜔) + 𝜃ℎ (𝜔)

F(ω) H1(ω) H2(ω) G(ω)

𝐻𝑒𝑞 (𝜔) = 𝐻1 (𝜔)𝐻2 (𝜔) ...(2-25)

𝐻𝑒𝑞 (𝜔) = 𝐻1 (𝜔) + 𝐻2 (𝜔) ...(2-26)

For the output signal

Is some systems, H(ω) can be found using:

Spectral Density and Correlation:

Power Spectral Density (PSD):

𝑆𝑓 (𝜔) = 2𝜋 ∑ |𝐶𝑛 |2 𝛿(𝜔 − 𝑛𝜔𝑜 ) Watt/Hz … (2-31)

To find the total power from the spectrum, we use:

A given voltage signal 𝑓(𝑡) = cos 20 𝜋𝑡 + 2 cos 30 𝜋𝑡 across 2 Ω resistor.

= 2𝜋(1)2 𝛿(𝜔 + 30𝜋) 2 2

𝑅𝑓 (𝜏) = 𝐹 −1 {𝑆𝑓 (𝜔)}

The autocorrelation is useful for the detection of signals, in which masked by