Chapter 4

Digital Transmission

4.1 Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display.
4-1 DIGITAL-TO-DIGITAL CONVERSION

In this section, we see how we can represent digital

data by using digital signals. The conversion involves
three techniques: line coding, block coding, and
scrambling. Line coding is always needed; block
coding and scrambling may or may not be needed.

Topics discussed in this section:

 Line Coding
 Line Coding Schemes
 Block Coding
 Scrambling
4.2
Line Coding
 Converting a string of 1’s and
0’s (digital data) into a
sequence of signals that
denote the 1’s and 0’s.
 For example a high voltage
level (+V) could represent a
“1” and a low voltage level (0
or -V) could represent a “0”.

4.3
Figure 4.1 Line coding and decoding

4.4
Relationship between
data rate and signal
rate
The data rate defines the number of


bits sent per sec - bps. It is often

referred to the bit rate.
 The signal rate is the number of
signal elements sent in a second and
is measured in bauds. It is also
referred to as the modulation rate.
 Goal is to increase the data rate
whilst reducing the signal rate(baud
rate).

4.5
Figure 4.2 Signal element versus data element

4.6
Data rate and Baud
 rate
Data Rate:

The number of bits(data elements) sent in 1s

The unit is bps
 Signal Rate:

The number of bits(data elements) sent in 1s

The unit is baud
 The baud or signal rate can be expressed as:
S = c x N x 1/r bauds
where N is data rate, c is the case factor (worst,
best & avg.)
r is the ratio between data element & signal element

4.7
Example 4.1

A signal is carrying data in which one data element is

encoded as one signal element ( r = 1). If the bit rate is
100 kbps, what is the average value of the baud rate if c is
between 0 and 1?

Solution
We assume that the average value of c
is 1/2 . The baud rate is then

4.8
Note

Although the actual bandwidth of a

digital signal is infinite, the effective
bandwidth is finite.

4.9
Example 4.2

The maximum data rate of a channel (see Chapter 3) is

Nmax = 2 × B × log2 L (defined by the Nyquist formula).
Does this agree with the previous formula for Nmax?

Solution
A signal with L levels actually can
carry log2L bits per level. If each
level corresponds to one signal element
and we assume the average case (c =
1/2), then we have

4.10
Considerations for choosing
a good signal element
referred to as line
encoding
 Baseline wandering - a receiver will

evaluate the average power of the

received signal (called the baseline)
and use that to determine the value
of the incoming data elements. If the
incoming signal does not vary over a
long period of time, the baseline
will drift and thus cause errors in
detection of incoming data elements.
 A good line encoding scheme will
prevent long runs of fixed amplitude.
4.11
Line encoding C/Cs
 DC components - when the voltage
level remains constant for long
periods of time, there is an
increase in the low frequencies of
the signal. Most channels are
bandpass and may not support the
low frequencies.
 This will require the removal of
the dc component of a transmitted
signal.

4.12
Line encoding C/Cs
 Self synchronization - the
clocks at the sender and the
receiver must have the same
bit interval.
 If the receiver clock is
faster or slower it will
misinterpret the incoming bit
stream.
4.13
Figure 4.3 Effect of lack of synchronization

4.14
Example 4.3

In a digital transmission, the receiver clock is 0.1 percent

faster than the sender clock. How many extra bits per
second does the receiver receive if the data rate is
1 kbps? How many if the data rate is 1 Mbps?
Solution
At 1 kbps, the receiver receives 1001
bps instead of 1000 bps.

At 1 Mbps, the receiver receives 1,001,000 bps instead of

1,000,000 bps.

4.15
Line encoding C/Cs
 Error detection - errors occur
during transmission due to line
impairments.
 Some codes are constructed such
that when an error occurs it can be
detected. For example: a particular
signal transition is not part of
the code. When it occurs, the
receiver will know that a symbol
error has occurred.
4.16
Line encoding C/Cs
 Noise and interference - there
are line encoding techniques
that make the transmitted
signal “immune” to noise and
interference.
 This means that the signal
cannot be corrupted, it is
stronger than error detection.

4.17
Line encoding C/Cs
 Complexity - the more robust
and resilient the code, the
more complex it is to
implement and the price is
often paid in baud rate or
required bandwidth.

4.18
Figure 4.4 Line coding schemes

4.19
Unipolar
 All signal levels are on one side of
the time axis - either above or below
 NRZ - Non Return to Zero scheme is an
example of this code. The signal
level does not return to zero during
a symbol transmission.
 Scheme is prone to baseline wandering
and DC components. It has no
synchronization or any error
detection. It is simple but costly in
power consumption.

4.20
Figure 4.5 Unipolar NRZ scheme

4.21
Polar - NRZ
 The voltages are on both sides of the
time axis.
 Polar NRZ scheme can be implemented
with two voltages. E.g. +V for 1 and
-V for 0.
 There are two versions:
 NZR - Level (NRZ-L) - positive voltage
for one symbol and negative for the other
 NRZ - Inversion (NRZ-I) - the change or
lack of change in polarity determines the
value of a symbol. E.g. a “1” symbol
inverts the polarity a “0” does not.
4.22
Figure 4.6 Polar NRZ-L and NRZ-I schemes

4.23
Note

In NRZ-L the level of the voltage

determines the value of the bit.
In NRZ-I the inversion
or the lack of inversion
determines the value of the bit.

4.24
Note

NRZ-L and NRZ-I both have an average

signal rate of N/2 Bd.

4.25
Note

NRZ-L and NRZ-I both have a DC

component problem and baseline
wandering, it is worse for NRZ-L. Both
have no self synchronization &no error
detection. Both are relatively simple to
implement.

4.26
Example 4.4

A system is using NRZ-I to transfer 1-Mbps data. What

are the average signal rate and minimum bandwidth?

Solution
The average signal rate is S= c x N x R
= 1/2 x N x 1 = 500 kbaud. The minimum
bandwidth for this average baud rate is
Bmin = S = 500 kHz.

Note c = 1/2 for the avg. case as worst

case is 1 and best case is 0
4.27
Polar - RZ
 The Return to Zero (RZ) scheme uses
three voltage values. +, 0, -.
 Each symbol has a transition in the
middle. Either from high to zero or
from low to zero.
 This scheme has more signal transitions
(two per symbol) and therefore requires
a wider bandwidth.
 No DC components or baseline wandering.
 Self synchronization - transition
indicates symbol value.
 More complex as it uses three voltage
level. It has no error detection
capability.
4.28
Figure 4.7 Polar RZ scheme

4.29
Polar - Biphase:
Manchester and
Differential

Manchester
Manchester coding consists
the NRZ-L and RZ schemes.
of combining

 Every symbol has a level transition in the

middle: from high to low or low to high.
Uses only two voltage levels.
 Differential Manchester coding consists
of combining the NRZ-I and RZ schemes.
 Every symbol has a level transition in the
middle. But the level at the beginning of
the symbol is determined by the symbol
value. One symbol causes a level change the
other does not.

4.30
Figure 4.8 Polar biphase: Manchester and differential Manchester schemes

4.31
Note

In Manchester and differential

Manchester encoding, the transition
at the middle of the bit is used for
synchronization.

4.32
Note

The minimum bandwidth of Manchester

and differential Manchester is 2 times
that of NRZ. The is no DC component
and no baseline wandering. None of
these codes has error detection.

4.33
Bipolar - AMI and
Pseudoternary
Code uses 3 voltage levels: - +, 0,
-, to represent the symbols (note not
transitions to zero as in RZ).
 Voltage level for one symbol is at
“0” and the other alternates between
+ & -.
 Bipolar Alternate Mark Inversion
(AMI) - the “0” symbol is represented
by zero voltage and the “1” symbol
alternates between +V and -V.
 Pseudoternary is the reverse of AMI.

4.34
Figure 4.9 Bipolar schemes: AMI and pseudoternary

4.35
Bipolar C/Cs
 It is a better alternative to
NRZ.
 Has no DC component or baseline
wandering.
 Has no self synchronization
because long runs of “0”s results
in no signal transitions.
 No error detection.

4.36
Multilevel Schemes
 In these schemes we increase the
number of data bits per symbol
thereby increasing the bit rate.
 Since we are dealing with binary data
we only have 2 types of data element
a 1 or a 0.
 We can combine the 2 data elements
into a pattern of “m” elements to
create “2m” symbols.
 If we have L signal levels, we can
use “n” signal elements to create Ln
signal elements.
4.37
Code C/Cs
 Now we have 2m symbols and Ln signals.
 If 2m > Ln then we cannot represent the
data elements, we don’t have enough
signals.
 If 2m = Ln then we have an exact mapping
of one symbol on one signal.
 If 2m < Ln then we have more signals than
symbols and we can choose the signals
that are more distinct to represent the
symbols and therefore have better noise
immunity and error detection as some
signals are not valid.

4.38
Note

In mBnL schemes, a pattern of m data

elements is encoded as a pattern of n
signal elements in which 2m ≤ Ln.

4.39
Representing
Multilevel Codes
 We use the notation mBnL, where
m is the length of the binary
pattern, B represents binary
data, n represents the length of
the signal pattern and L the
number of levels.
 L = B binary, L = T for 3
ternary, L = Q for 4 quaternary.

4.40
Figure 4.10 Multilevel: 2B1Q scheme

4.41
Redundancy
 In the 2B1Q scheme we have no redundancy
and we see that a DC component is
present.
 If we use a code with redundancy we can
decide to use only “0” or “+” weighted
codes (more +’s than -’s in the signal
element) and invert any code that would
create a DC component. E.g. ‘+00++-’ ->
‘-00--+’
 Receiver will know when it receives a “-”
weighted code that it should invert it as
it doesn’t represent any valid symbol.

4.42
Multitransition Coding
 Because of synchronization requirements we
force transitions. This can result in very
high bandwidth requirements -> more
transitions than are bits (e.g. mid bit
transition with inversion).
 Codes can be created that are differential
at the bit level forcing transitions at bit
boundaries. This results in a bandwidth
requirement that is equivalent to the bit
rate.
 In some instances, the bandwidth requirement
may even be lower, due to repetitive
patterns resulting in a periodic signal.

4.43
Figure 4.13 Multitransition: MLT-3 scheme

4.44
MLT-3
 Signal rate is same as NRZ-I
 But because of the resulting
bit pattern, we have a
periodic signal for worst
case bit pattern: 1111
 This can be approximated as
an analog signal a frequency
1/4 the bit rate!
4.45
Scrambling
 The best code is one that does not
increase the bandwidth for
synchronization and has no DC components.
 Scrambling is a technique used to create
a sequence of bits that has the required
c/c’s for transmission - self clocking,
no low frequencies, no wide bandwidth.
 It is implemented at the same time as
encoding, the bit stream is created on
the fly.
 It replaces ‘unfriendly’ runs of bits
with a violation code that is easy to
recognize and removes the unfriendly c/c.

4.46
Figure 4.18 AMI used with scrambling

4.47
For example: B8ZS substitutes eight
consecutive zeros with 000VB0VB.
The V stands for violation, it violates the
line encoding rule
B stands for bipolar, it implements the
bipolar line encoding rule

4.48
Figure 4.19 Two cases of B8ZS scrambling technique

4.49
HDB3 substitutes four consecutive
zeros with 000V or B00V depending
on the number of nonzero pulses after
the last substitution.
If # of non zero pulses is even the
substitution is B00V to make total # of
non zero pulse even.
If # of non zero pulses is odd the
substitution is 000V to make total # of
non zero pulses even.
4.50
Figure 4.20 Different situations in HDB3 scrambling technique

4.51