UNIT - 2

BASE BAND PULSE TRANSMISSION

SYLLABUS:
Introduction, matched filter, properties of matched filter, matched filter for rectangular
pulse, Error due to noise, Inter symbol Interference(ISI), Nyquist’s criterion for distortion less base band
binary transmission, ideal Nyquist channel, Raised cosine Filter & its spectrum, correlative coding duo
binary & modified duo binary signalling schemes, partial response signalling, base band M-array PAM
Transmission, Eye diagrams.

INTRODUCTION:
Generally the digital data transmission takes place in two ways.

i Base Band Pulse Transmission.

ii Pass Band Pulse Transmission.

 In base band pulse transmission, the digital data transmitted directly over the channel without
using any modulation technique. This one can be used for shorter distance communication.
 In pass band pulse transmission, the digital data can be transmitted over the channel by using
modulation techniques. This one can be used for the longer distance communication.
 In case of base band data transmission, there is a requirement to use low frequency channel
whose bandwidth is large enough to pass the input data stream.
 In case the channel is dispersive then the each received pulse is affected somewhat by adjacent
pulses. This type of interference is called Inter Symbol Interference (ISI). This is the major
source of bit errors. This one can be controlled by maintaining certain pulse shape in the
overall system.
 Another source of bit errors is channel noise. Now introducing matched filter for detection of
pulse signal effected by the channel noise. Matched filter uses linear time invariant filter.

MATCHED FILTER:
The device uses linear time invariant filter for the detection of pulse transmitted over a channel that
is corrupted by channel noise is called marched filter, which is so called because its impulse response is
matched to the pulse signal. It is used to improve the S/N ratio.

Let us consider a receiver model shown in fig, with LTI filter of impulse response h(t).
1
Page

FIG: Linear receiver

The filter input x(t) consist of pulse signal g(t) & noise w(t), it is expressed as

-------------------- (1)

- Where T is observation interval, g(t) is a binary symbol 1 or 0, w(t) is a sample function of white
noise, zero mean, psd N0/2

Since the filter is linear, the filter output y(t) is expressed as

-------------------- (2)

Where go(t) & n(t) are signal & noise components produced by signal x(t).

Here, the filter has to make at time t=T, the instantaneous power of the output signal as maximum as
possible than the average power of the output noise n(t). This is equivalent to maximizing the peak signal
to noise ratio, defined as

-------------------- (3)
Where the instantaneous power in the output is signal and is a measure of the average
output noise power.

 The requirement is to specify the impulse response of the matched filter in order to maximize the
output signal to noise ratio in equation 3.

OUTPUT SIGNAL POWER:

Let G(f) denotes the Fourier transform of the known signal g(t) & H(f) denote the frequency response of
the filter. Then the F.T of the output signal go(t) is equal to ( ). ( ). then

-------------------- (4)

When the filter output is sampled at time t=T, then take squared magnitude in eq.3.

-------------------- (5)

AVERAGE OUTPUT NOISE POWER:

The power spectral density SN(f) of the output noise ( ) is equal to the power spectral density of the
input noise ( ) times the squared magnitude response , i.e.,

-------------------- (6)
2
Page
Thus the average power of the output noise ( ) is:

-------------------- (7)

Substituting eq.s 5,7 into eq.3, the expression for peak signal to noise ratio is rewritten as,

-------------------- (8)

Apply Schwarz’s inequality theorem to the numerator of eq. 8.

 The Schwarz’s inequality theorem stated as, if we have two complex functions in
the real variable x, satisfying the conditions

 | 1 ( x) |2 dx  


Then we may write,

-------------------- (9)

The equality in eq. 9 holds if, we have -------------------- (10)

Where K = arbitrary constant and * = complex conjugation.

From the Schwarz’s inequality in eq.9, setting

then the numerator of eq. 8 written as

-------------------- (11)

Substitute eq.11 in eq.8 -------------------- (12)

The signal to noise ratio does not depends on the frequency response ( ) of the filter but only
on the signal energy and noise power spectral density. The max signal to noise ratio of o/p signal is,
3
Page

-------------------- (13)
 Assume that ( ) is the optimum value of ( ) to maximize the signal to noise ratio. From eq.10

-------------------- (14)

Then the optimum value of the filter ℎ ( ) is

∗
For real signal g(t), ( ) = (− )

-------------------- (15)

The optimum value of the filter, is except factor , it is time inversed & delayed version of the input
signal g(t). i.e. it is matched to the i/p signal.

PROPERTIES OF MATCHED FILTER:

Matched filter is a optimum device for the detection of received pulse signal effected by noise. By using
LTI filter, to improve the output signal to noise ratio.

i The impulse response of the matched filter is, except the scaling factor k, is the time inversed &
delayed inversion of the input pulse signal g(t).
i.e. hopt(t) = k . g(T-t)

Proof: From the Schwartz’s inequality theorem, from eq.10

Then the optimum value of the filter ℎ ( ) is

∗
For real signal g(t), ( ) = (− )
4
Page
ii The max signal to noise ratio of the output signal does not depends on the response of the filter
but only depends on signal energy & noise power spectral density.
i.e.

Proof: We know that the signal to noise ratio of output signal

-------------------- (1)

From the Schwarz’s inequality theorem,

The Fourier transform of filter o/p go(t) may written as

using the inverse FT

-------------------- (2)

According to Rayleigh’s Energy theorem, the integral of the squared magnitude spectrum
of a pulse signal with respect to frequency is equal to the signal Energy E.

-------------------- (3)

The average output noise power

-------------------- (4)

The peak signal to noise ratio of pulse signal is

5
Page

-------------------- (5)
 The peak signal to noise ratio of the o/p signal does not depends on the response of the filter but
only depends on signal energy and noise power spectral density.

MATCHED FILTER FOR RECTANGULAR PULSE:

Consider a signal g(t) is a rectangular pulse with an amplitude A and duration T is shown in fig.a.

 When it is passed through the matched filter whose impulse response h(t) is exactly matches with
the input pulse signal g(t), the output response of matched filter go(t) is a triangular wave form is
shown in fig.b.
 It has the maximum energy at time t = T i.e. max energy at the end of the pulse signal g(t)
interval. It is equal to KA2T.

For a special case of rectangular pulse, the matched filter is implemented using a
circuit known as integrate and dump circuit in fig d.

6
Page
 The integrator computes the area under the rectangular pulse, & the resulting o/p then sampled at
time t = T. Immediately after t = T the integrator restored to initial conditions.

. ., ( ) = =

It is shown in fig.1. i.e., during the interval 0≤ ≤ ,the o/p response of matched filter exceeds to o/p
response of integrate & dump circuit.

ERROR RATE DUE TO NOISE [PROBABILITY ERROR OF MATCHED FILTER]:

Consider a binary PCM system based on polar non return to zero signalling. In this form of
signalling, symbol 1 & 0 are represented by positive & negative rectangular pulses with amplitude A for
equal duration. The channel noise is modelled as additive white Gaussian noise w(t) of zero mean &
power spectral density N0 /2

In the signalling interval 0≤ ≤ b, the received signal is written as,

-------------------- (1)

Where Tb = bit duration, A = transmitted pulse amplitude

Given the noisy signal x(t), the receiver is required to make a decision in each signalling interval as to
whether the transmitted symbol is a ‘1’ or a ‘0’.

The structure of the receiver used to perform this decision make process is shown below.

Fig: Receiver for Baseband Transmission

The noisy signal x(t) is passed through the matched filter. Then the o/p of matched filter sampled at
time t = T. If this o/p sample value is more than threshold ‘ ’, the receiver make a decision as symbol 0.

If the o/p sample value is exactly equals to threshold value, the receiver may choose it as 1
(or) 0. Here there two possible kinds of errors to be considered

i. Symbol 1 is chosen where a 0 was actually transmitted.

7

ii. Symbol 0 is chosen where a 1 was actually transmitted.

Page
To determine the average probability of error, we consider these two situations separately.

Symbol ‘0’ was sent

The received signal is expressed as

-------------------- (2)

The output of matched filter is expressed as

-------------------- (3)

The above equation 3 represents the sample value of random variable. It is characterised as,

 The random variable y is Gaussian distributed with a mean of –A.

 The variance of the random variable y is

-------------------- (4)

Where Rw(t,u) is the auto correlation function of the white noise w(t).

-------------------- (5)

Where is the time shifted delta function

-------------------- (6)
8
Page
Then the conditional probability density function of random variable y where binary ‘0’ is transmitted is

-------------------- (7)

This function representation is shown in below fig.(a)

Fig: (a) Pdf of random variable y at matched filter output when ‘0’ is transmitted
Fig: (b) Pdf of random variable y at matched filter output when ‘1’ is transmitted

Let P10 denote the conditional probability of error, when symbol 0 was sent. In figure a, the shaded area
under the curve fy(y/0) over the limits λ to ∞ gives the P10

-------------------- (8)

Introduce the complementary error function for simplification of equation 8

-------------------- (9)

Simplify the equation 8, let us take

-------------------- (10)
9
Page
Symbol 1 was sent

The conditional probability density function of random variable ‘y’ where binary ‘1’ is transmitted is

-------------------- (11)

This function representation is shown in figure b

Let P01 denote the conditional probability error, when symbol 1 was transmitted. In figure b, the shaded
area under the curve fy(y/1) over the limits. -∞ to λ gives P01.

-------------------- (12)

For the simplification of eq. 12, we take

Interchanging the limits

-------------------- (13)

From equation 9 & 13 is written as

-------------------- (14)

Therefore the average probability of symbol error, Pe is

If symbol 0 & 1 are equiprobable, then

10
Page
Take the optimum value of threshold is equal to zero

i.e.

-------------------- (15)

The energy of transmitted signal, Eb is

------------------- (16)

Substitute eq. 16 in 15

------------------- (17)

The above expression gives the average probability of symbol error. (or) This equation gives the
probability error of matched filter.

INTER SYMBOL INTERFERENCE:

This is another source of bit errors in base band data transmission. Suppose the channel is in dispersive
nature, the present pulse is affected by the adjacent pulses, this type of interference is called as Inter
Symbol Interference.

Let us consider a base band binary data transmission system in order to describe ISI
mathematically.

FIG: Base band binary data transmission system

The incoming binary service {bk} consists of symbols 1 & 0, each with Tb duration. It is given to
Pulse Amplitude Modulator, it modifies binary sequence into a new sequence of short pulses, whose
amplitude ak is represented in polar form

-------------------- (1)
11
Page
These are given to transmit filter, whose impulse response is g(t), producing the transmitted signal i.e.

-------------------- (2)

This signal s(t) is modified and transmission over the channel whose impulse response h(t). The
channel adds the white noise ( ). Then the noisy signal x(t) is then passed through the receiving filter
output y(t) sampled & given to decision device. Sampled value is more than the threshold we will say it
as binary ‘1’. If the sampled value is lower than the threshold we will say it as binary ‘0’.

The receiver filter output is written as

-------------------- (3)

Where is the scaling factor & the pulse p(t) is to be defined. For precise value, we have to consider
transmission delay to, but in simplification we consider t0=0.

The scaled pulse .p(t) is obtained by double convolution of impulse response of transmitter g(t), the
impulse response of channel h(t) and impulse response of receiving filter c(t).

-------------------- (4)

p(t) is normalized by setting p(0) = 1

The above eq. gives the use of scaling factor to obtain amplitude changes in the signal transmission.

Convolution in time domain transformed into multiplication in frequency domain.

-------------------- (5)

The received filter o/p y(t) is sampled at time ti=iTb

-------------------- (6)

In the above eq. first term is contribution ith transmitted bit. The second term indicates residual effect
of all other transmitted bits on the decoding of the ith bit, this is because presence of pulses at starting &
ending of the sampling instants. This is called ISI. The third term indicates noise at the sample time t=ti.

In the absence of noise & interference eq. 6 written as

y(ti)=µai -------------------- (7)

12

Under these ideal conditions, the ith transmitted bit is decoded correctly. In order to minimize the ISI
Page

we have to specify the frequency response of the filters and transmitted signal pulse shape.
NYQUIST’S CRITERION FOR DISTORTION LESS BASE BAND BINARY TRANSMISSION:

In order to eliminate ISI, we need to specify frequency response & pulse shape. First, we need to
determine the frequency response of the transmit and receive filter for the better reconstruction of binary
sequence. For this, receiver does extracting & decoding. The extraction involves sampling the o/p y(t) at
time t=iTb. The decoding done from pulse at k=i, i.e. p(t) is shown as

-------------------- (1)

Where p(0)=1

If P(t) satisfies the condition in eq. 1 the receiver o/p

y(ti)=µai for all i

The condition in eq. 1 gives the perfect reception in the absence of noise.

From design point of view, it is used to transform eq. 1 in to the frequency domain.

From the Fourier transform

-------------------- (2)

Where fb is bit rate in bits per second (b/s)

( ) Fourier transform of infinite periodic sequence of delta functions of period Tb.

The time domain representation of pulse signal ( ) is

-------------------- (3)

13
Page
Let the integer n = i-k

-------------------- (5)

The frequency function p(f) eliminates inter symbol interference for samples taken at intervals Tb
provided that it satisfy eq. 5.
14
Page
IDEAL NYQUIST CHANNEL:

The simplest way of satisfying p(f − nfb)=Tb, this eq. is to specify the frequency function p(f)
to be in the form of a rectangular function, it is shown as

-------------------- (1)

Where rect(f)→rectangular function of unit amplitude.

The overall system bandwidth ‘w’ is defined by

-------------------- (2)

By applying inverse Fourier transform to eq. 1, we will get

-------------------- (3)

The special value of bit rate fb=2w is called the Nyquist rate & w itself called the Nyquist bandwidth. The
ideal base band data transmission system described by eq. 1 in frequency domain and eq. 3 in time
domain is called the Ideal Nyquist channel.

The plots for p(f) & p(t) shown in below fig a & b respectively.

15
Page
  The normalised form of frequency function p(f) is plotted for positive & negative frequencies
shown in fig a.
 The function p(t) can be regarded as the impulse response of ideal low pass filter with pass band
magnitude response 1/2w & BW w.
 The function p(t) has its peak value at the origin & goes through zero at integer multiplies of the
bit duration Tb. It is shown in fig b.
 The received pulse sampled at Tb, we does not have ISI.
 This ideal Nyquist channel have two practical difficulties.
i The p(f) is plot from –w to w & zero elsewhere. This is physically unrealizable because of the
abrupt transitions at the edges ±w.
ii The function p(t) decreases as 1/t for large t, resulting in a slow rate of decay.

To evaluate the effect of this timing error, consider y(t) sample at t=∆ . In the absence of noise

In the above eq., first term defines the desired symbol, where as remaining term represents the inter
symbol interference caused by timing error ∆ in sampling the o/p y(t).

RAISED COSINE SPECTRUM:

The practical difficulties of Ideal Nyquist channel is eliminated by extending the bandwidth from the
minimum value w=fb/2 to an adjustable value between w & 2w. Now we specify the frequency response
p(f) to satisfy conditions more elaborate than ideal Nyquist channel i.e. shown by

-------------------- (1)

Here we are considering raised cosine spectrum to satisfy eq. 1. This function frequency response
consists of flat portion and roll off portion that has a sinusoidal form, as follow
16
Page
 -------------------- (2)

The frequency parameter f1 & bandwidth w are related by

f1 -------------------- (3)
  1
W
The parameter is called the roll off factor, it indicates the excess Bandwidth over the ideal solution,
w.
The transmission Bandwidth BT, is defined by

-------------------- (4)

The frequency response p(f) is normalized by multiplying it by 2w, shown in fig(a). For three values of
namely 0,0.5,1

Fig (a) : frequency response of different roll off factors

17

For = 0.5 & 1, the p(f) cut off gradually as compared with the ideal Nyquist channel.
Page
 The time response p(t) is the I.F.T of p(f) i.e.

-------------------- (5)

The plot of p(f) shown in fig(b) for = 0,0.5,1

Fig (b) : Time response for different roll of factors

P(t) consists of two factors sinc(2wt) characterizing the ideal Nyquist channel & the second factor that
decreases as 1/t2 for large |t| .

For = 1, we have the most gradual roll off in that amplitude of oscillators tails of p(t) are smallest.

 The ISI reduces by increasing from zero to unity.

 In case = 1, number of zero instant increases i.e. we have additional zero instants at ±3/2Tb,
±5/2Tb etc.... it reduces the ISI.

The special case = 1 is known as the full cosine roll off characteristics [full raised cosine spectrum],
then frequency response

-------------------- (6)

The time response p(t) is

18
Page

-------------------- (7)
CORRELATIVE LEVEL CODING:
By adding inter symbol interference to the transmitted signal in a controlled manner, it is possible
to achieve a signalling rate equal to the Nyquist rate of 2w symbols per second in a channel of bandwidth
W Hertz. Such schemes are called correlative level coding (or) partial response signalling schemes.

The design of these schemes based on the following assumptions, i.e. ISI
introduced into transmitted signal is known value, its effect can be taken at the receiver in a deterministic
way.

This is the practical method for receiving theoretical maximum signalling rate of
2w sym/sec in a bandwidth of w Hz.

DUOBINARY SIGNALLING:
 The basic idea of correlative level coding can be implemented by using this signalling scheme.
Here Duo implies that doubling of the transmission capacity of a straight binary system.
 This type of correlative coding also called as class 1 practical response.

Consider a binary i/p sequence {bk} consisting of uncorrelated binary symbols 1 and 0, each having
duration Tb. First this sequence is applied to a pulse amplitude modulator, it produces two level sequence
of short pulses, whose amplitude ak is defined as,

-------------------- (1)

This sequence is next applied to the duo binary encoder, it is converted into a three level o/p, i.e. -2, 0 &
+2. This is produced by the following block diagram.

Fig: Duobinary signalling scheme

The two level sequence {ak} is first passed through a simple filter involving a single delay
element and summer. For every unit impulse applied to the i/p of this filter, we get two unit impulses
19

spaced Tb seconds apart at the filter o/p.

Page
Therefore the duo binary coder o/p Ck is the sum of present i/p pulse ak and its previous value ak-1, i.e.

Ck = ak + ak+1 -------------------- (2)

The above eq. describes that, uncorrelated two level sequence converted into three level correlated
sequence. This correlation between adjacent pulses is viewed as introducing ISI into a transmitted signal
in an artificial manner.

An ideal delay element, producing a delay of Tb seconds, has the frequency response

The frequency response of simple delay line filter is 1 + . The overall frequency response
this filter connected in cascade with ideal Nyquist channel is

-------------------- (3)

Where H1(f)→class 1 partial response.

For ideal Nyquist channel of bandwidth w=1/2Tb, we have

-------------------- (4)

Thus the overall frequency response of the duo binary scheme is half cycle cosine function, i.e.

-------------------- (5)

For the above eq. magnitude & phase response spectrums shown below
20

Fig: Frequency response of the duo-binary conversion filter.

(a) Magnitude response. (b) Phase response.
Page
From the frequency response the impulse response of the duo binary signalling scheme is

From eq. 3

-------------------- (6)

The impulse response spectrum of h1(t) is shown below

Figure: Impulse response of the duo-binary conversion filter.

In the figure, the response of the input pulse is spread over more than one signalling interval. It is
also stated as response in any signalling interval is partial.

The original two level sequence {ak} is detected from the duo binary sequence {ck} by using eq. 2
i.e.

-------------------- (7)

Where →Estimate of original pulse at t=kTb

→Previous Estimate.

The receiver detect correctly if there is no errors in ck & ak-1.

The technique of using a stored Estimate of the previous symbol, is called “Decision Feedback”.
It is a just reverse process of delay line filter at transmitter. The major problem in this detection is, either
ck (or) ak-1 having error i.e. propagate through the o/p. This error propagation can be eliminated by using
proceeding before the duo binary coder this is shown in below fig.
21
Page
 Fig: Precoded Duobinary scheme

From the above diagram, precoder converts i/p binary sequence {bk} into another sequence {dk}, i.e.

-------------------- (8)

-------------------- (9)

This precoded sequence is applied to PAM, it generates two level sequence {ak}, i.e. ak=±1. After
that, this two level sequence given to duo binary coder, it generates three level sequence {ck}, i.e. -2, 0,
+2

-------------------- (10)

This sequence is applied to the decision device, it is shown in fig. 22

Page

Fig: Detector for recovering original binary sequence from the pre-coded duo-binary coder output.
 This device taken the decisions on the duo binary sequence based on following conditions

When |Ck| = 1, the receiver simply makes a random guess in favour of symbol 1 or 0.

MODIFIED DUO BINARY SIGNALLING:

In the duo binary signalling technique the frequency response H(f) & the power spectral density
of the transmitted pulse is non zero at the origin. This is considered to be an undesirable feature in some
applications. This is overcome by using the class IV partial response (or) modified duo binary technique,
which involves correlation span of two binary digits (Delay 2Tb). This is shown in below figure.

Modified duo binary encoder involves subtractor & delay 2Tb. The o/p of the modified duo binary
conversion filter is expressed as

-------------------- (1)

Here, the two level sequence (ak=±1) converted into three level sequence (-2, 0, +2).

The overall frequency of the delay line filter cascade with the ideal Nyquist channel is 23

-------------------- (2)
Page
The corresponding magnitude & phase response is shown in fig.

From the above spectrum we don’t have any frequency response & PSD at zero frequency. It
means there is no DC component at o/p.

From the first line of eq. 2, the impulse response of the modified duo binary coder is expressed as
below

-------------------- (3)

The impulse response is plotted below

24
Page
 To eliminate the possibility of error propagation in the detection process of modified duo binary
signalling, we use a precoding procedure similar to duo binary scheme. From the fig of modified duo
binary, the precoded sequence is expressed as

-------------------- (4)

It is given to PAM, it generate two level sequence {ak}. This two level sequence ak is given to modified
duo binary coder, it will generate three level sequence based on the following expression

-------------------- (5)

For detection, this sequence {ck} given to decision device, it will gives the binary 0 (or) 1 based on the
following conditions.

If |ck|>1, say symbol bk is 1

If |ck|<1, say symbol bk is 0 -------------------- (6)

If |ck|=1, the receiver makes a random guess in favour of symbol 1 (or) 0.

GENERALISED FORM OF CORRELATIVE LEVEL CODING:- (PARTIAL RESPONSE OF

SIGNALLING)

This Generalised form of correlative level coding (or) partial signal response gives the
straight forward techniques for achieving the different classes of partial response. This Generalization is
shown below

25
Page
  It involves a tapped delay line filter with tap weights w0, w1, ..........., wn-1.
 Different classes of partial response signalling schemes are achieved by using weighted linear
combination of N ideal Nyquist pulses, it is

-------------------- (1)

 The below table gives the specific details of five different classes of partial response signalling
schemes.

Type of class N w0 w1 w2 w3 w4 Comment

Duo binary coding
I 2 1 1
II 3 1 2 1
III 3 2 1 -1
Modified duo binary coding
IV 3 1 0 -1
V 5 -1 0 2 0 1

Different classes of partial response signalling schemes.

For example, in the duo binary coding (or) class I partial response signalling,

W0=+1

W1=+1

In case of modified duo binary signalling, W0=+1

W1=0

W2=-1

BASE BAND M-ARY PAM TRANSMISSION:-

In the base band binary PAM transmission, the pulse amplitude modulator produces binary pulses
with one of two possible amplitude levels. But in case of base band M-ary PAM transmission, the pulse
amplitude modulator produces one of M possible amplitude levels with M>2. This one can be explained
with the following example, inarternary system with binary data sequence 0 0 1 0 1 1 0 1 1 1 .

 The above fig a shows the different amplitude levels of gray coded sequence. Fig b shows the
electrical representation of given binary data using inarternary system.
 Pulse duration of binary PAM system is denoted by Tb, but in case M-ary PAM transmission the
pulse duration T=2Tb. We refer 1/T as the signalling rate, which is expressed as sym/sec (or)
bauds. In case of M-ary PAM system, 1 baud =log2M bits/sec.
 The symbol duration of the M-ary PAM system is related to the symbol duration of binary PAM
system is related as
26

T  Tb log 2 M
Page
 Fig: Output of a quaternary system. (a) Waveform. (b) Representation of the 4 possible dibits,
based on Gray encoding.

Therefore, in case of M-ary PAM transmission in a given channel bandwidth, it is possible to transmit
log times faster than the binary PAM system. The transmitter power must be increased by the
factor compared to a binary PAM system.

 In a base band M-ary PAM transmission system, the input binary sequence converted into a
M-level PAM pulse train. The detection process is same as the binary PAM at the receiver
section. That is, first this M-level PAM pulses are transmitted over the channel, which is
affected by the noise & distortion. The received signal is passed through the receive filter and
then sampled at appropriate time intervals. Each sample is compared with the predefined
threshold value, and decision is made as to which symbol was transmitted. In order to reduce
the bit errors which are introduced due to ISI & noise, we have to design the transmit &
receive filter and pulse shape as same as the base band binary PAM system.

27
Page
EYE PATTERNS:

It is an experimental tool for observing the combined effect of inter symbol interference and channel
noise on the performance of the base band pulse transmission system.

 The eye pattern names comes from human eye because the appearance of binary waves closely
equal to human eye.
 The interior portion of the eye pattern is called the eye opening.

 The width of the eye opening define the time interval over which the received signal can be
sampled without error from ISI.
 The best sampling time is at which the opening eye is widest.
 The height of the eye opening at a specified sampling time, defines the noise margin of the
system.
 The sensitivity of the system to timing error is determined by the rate of closure of the eye as the
sampling time is varied.

28
Page