Professional Documents
Culture Documents
1.1. INTRODUCTION
1.1.1. BASIC DEFINATIONS
COMMUNICATION SYSTEM:
The system which is used to transmit information bearing signal from one
point (source) to another (destination).
SOURCE:
The system which generates information is called source.
The information which is coming out of source may be in the form of
analog or digital in nature.
Practically almost all sources generate analog signals.
Examples for information are: voice, music, television picture,
teletype data, atmosphere temperature and pressure & soon.
The above signals can be converted into electrical signals for
transmission through channel.
The signals can be of three forms:
i. Analog: these are continuous in amplitude and time.
ii. Discrete: These are continuous in amplitude and discrete in
time.
iii. Digital: These are discrete in both amplitude and time.
SYMBOL RATE:
The rate at which the information source generates source alphabets.
It is expressed in symbols/sec.
INFORMATION RATE:
The information rate represents the minimum average data rate required to
transmit information from source to destination.
Therefore Information rate = (Symbol rate ) X ( Source entropy).
This is expressed in bits/sec.
Where source entropy is the average information content per symbol
[bits/symbol]
DEMERITS:
NOTE:
1) Sampling.
2) Quantization and
3) Encoding.
Let us consider a digital communication system as shown fig1.1. the basic signal
operations performed here are source coding channel coding and modulation.
Page 2
Source:
Assume the signal generated by the source is a digital or an analog signal is converted
into digital.
Channel encoder / decoder is used to map the source encoder output with the channel
input and channel output to source decoder input respectively.
In this process some controlled redundancy is added at the encoder, which helps to
detect the errors at the receiver.
This is necessary because error may occur due to addition of noise in the channel.
Therefore in source encoding redundancy is removed & in channel coding controlled
redundancy is added.
Modulator/ Detector:
1.3. Sampling:
To convert an analog signal into digital data, sampling, quantization and encoding has
to be performed.
An analog signal is first sampled, which can be defined as “the process of converting
an analog signal into discrete time signal by measuring the signals at periodic instant
of time”.
The sampling rate must be chosen properly, so that the discrete time signals obtained,
uniquely defines the continuous-time signals.
Statement:
“If a finite energy signal g(t) contains no frequencies higher than ‘ W ‘ hertz, it
can be completely recovered from its ordinates at a sequence points spaced ‘1/2W’
seconds apart”.
Page 3
Proof:
The sampling theorem for band limited signal can be proved in two parts:
1) Representing g (t) in terms of its samples.
2) Reconstructing it from its samples.
To prove this, let us consider an analog signal g (t), of finite energy and infinite
duration as shown in fig1.2 (a).
g (t ) g (nTs) (t nTs)
n
--------------- >(1)
δ (t-nTs) indicates the samples placed at ±Ts, ±2Ts, ±3Ts, ±4Ts………….& so on.
Eq (1) can be graphically represented as shown in fig1.2 (b)
Since g(nTs) are samples of g(t)
g (t ) (t nTs ) g (nTs ) (t nTs ) ……………..(2)
Therefore, Eq (1) can be written as :
g (t ) g (t ) (t nTs )
n
Page 4
If we consider g (t) is message signal & Ts (t) carrier signal then gδ (t) is the
modulated wave as shown below
We know that multiplication of two time function is equivalent to the convolution of their
respective Fourier transforms.
By definition g(t) ------------------> G(f)
From the properties of the delta function convolution of G(f) and δ(t − nTs) is G(f-nfs).
Therefore eq(4) can be rewritten as
G ( f ) f s G (f - nf
n
s ) ---------------------- > (5)
Therefore from the above equation it is understood that G (f) is placed at ±fs, ±2fs, ±3fs,
±4fs………. & hence G (f) is periodic in fs.
Page 5
Reconstruction of signal from its samples.
g (t )
G( f ) exp( j 2ft)df
1 n jnf
g (t )
2 w
n
g (
2 w
) exp(
w
) exp( j 2ft)df
n 1 e j 2f (t nTs ) w
g (t ) g ( ) [ ] w
n 2w 2w j 2 (t nTs)
Page 6
n 1 e j 2w(t nTs ) e j 2w(t nTs )
g (t ) g ( ) [ ]
n 2w 2w j 2 (t nTs)
n 1 sin 2w(t nTs)
g (t ) g ( 2w ) 2w [
n (t nTs)
]
n sin 2w(t nTs)
g (t ) g ( 2w )[ (2wt 2wnTs) ]
n
n sin (2wt n)
g (t ) g ( 2w )[
n (2wt n)
]
n
g (t ) g ( 2w ) sin c(2wt n)
n
---------->(10)
Equation 10 is called an interpolation formula for reconstructing the original signal g (t)
Fig.1.3(c) shows the response of the reconstruction filter.
W.K.T the complex envelope g(t) of band pass signal ğ(t) are related as
g(t)=Re[ğ(t)exp(j2πfct)] ----------->(1)
Since ğ(t) is complex, ğ(t) = gI(t)+jgQ(t) ------------->(2)
And exp (j2πfct) = cos (2πfct) +jsin (2πfct) -----------> (3)
Substituting equ (2) and equ (3) in equ (1)
g(t) = gI(t)cos(2πfct) – gQ(t)sin(2πfct) ------------>(4)
gI(t) and gQ(t) are obtained by multiplying cos and sin functions to band pass g(t) as shown
in fig.1.5(a) . To reconstruct g(t) from these components gI(t) and gQ(t) are multiplied by
cos and sin functions as shown in fig.1.5(b).
Page 7
Fig1.5(a): Generation of gI(t) and gQ(t)
Let g(t), c(t) and s(t) be the analog i/p, sampling function and o/p signals respectively as
shown in fig.1.6.
Page 8
If fs is the sampling frequency= (1/Ts), mathematically the sampled signal s(t) can be
written as
s (t) = c(t)g(t)………….(1)
TA
Where, Cn sin c( f nT )...................(3)
Ts
Where, T = pulse duration
A = amplitude of pulses
fn = nfs
Page 9
Fig1.7: Spectrum of s (f)
Therefore the shifted version of h(t) is to be multiplied with basedband signal to get the
samples
s(t ) g (nT )h(t nT ) -------------> (1)
n
s s
g (t ) g (nT ) (t nT ) ------ > (4)
n
s s
Page 10
g (nT ) (t nT )h(t )d --------- > (5)
n
s s
g (t ) h(t ) g (nT )h(t nT ) ---------------- > (6)
n
s s
We know that, G ( f ) f s G (f - nf
n
s )
S( f ) fs G(f - nf
n
s )H(f)
Page 11
Fig.1.8: (a) one pulse of rectangular pulse train.
(b) spectrum of the pulse train of (a)
Practical method of obtaining flat top samples from a continuous time signal g (t) is
known as sample & hold circuit. A sample and hold circuit is depicted in fig .1.10(a) & it
make use of 2 FET switches & a capacitor connected across second FET transistor. A
pulse applied at G1 closes the sampling switch & capacitor holds the sampled voltage
until discharged by a pulse applied at G2.
{0 otherwise
Flat top sampling is obtained by convoluting the output of ideal samples by a rectangular
pulse h(t) as shown in fig1.10 (b)
i.e . v(t ) g (t ) h(t )
Taking FT on both side we get
Page 12
V ( f ) G ( f ).H ( f )
V ( f ) fs G(f - nf
m
s )H(f)
------------ > (2)
Where
H ( f ) Sinc ( f ) exp( jf )
H (f) acts like a LPF & attenuates the spectral components located at integer multiples of
fs
However, it also attenuates the HF components of the massage spectrum centered at f=0,
known as aperture effect. Fig.1.10(c) shows the reconstruction process. The
reconstruction filter [LPF] removes unwanted frequencies component present at integer
multiples of f s . An equalizer is cascaded with the reconstruction filter to compensate
for aperture effect. It may be notated that the amplitude frequency response of the
1
equalizer is H ( f ) .
Fig1.10 (a)
Fig.1.10 (b)
Fig.1.10(c)
Page 13