Digital Transmission

Most modern communication systems are digital, meaning that the transmitted information signal
carries bits and symbols rather than an analog signal. The effect of C/N ratio increase or decrease on
digital transmissions appears in the form of errors in the transmitted digital information.

Difference between a Symbol and a Bit

A symbol is a pulse with duration TS seconds (this is a period with units of “seconds” not
temperature as was used before) that may contain a bit or more of data. For:

• a symbol that takes one of 2 values, each symbol contains 1 bit

Ö TS = T B & fS = fB
• a symbol that takes one of 4 values, each symbol contains 2 bits
Ö TS = 2TB & fS = fB/2
• …..
• A symbol that takes one of M = 2N values, each symbol contains N bits.
Ö TS = N TB & fS = fB/N

There is an inherent advantage to transmitting data using symbols with large values of M. The
advantage is that more data is transmitted by using the same BW. That is, the element that determines
the bandwidth in digital transmission is number of transitions between symbols not the number of
levels over which we change. So, two signals with 10 changes per second but one of them changes
between 2 levels while the other changes among 10 levels will require the same bandwidth.

Types of Signaling (Line codes)

The reality is that information sources that provide digital data actually provide numbers that are not
suitable for transmission on any channel as they are. An analog to digital converter (ADC) that
converts an analog audio signal to a digital format provides sample values every 50 to 100
microseconds. The sample values must be formatted in a proper way to make them suitable for
transmission through the communication channel. When each digital value is represented using a
pulse for each bit (in binary communication) or a pulse for multiple bits simultaneously (for M-ary
communication), the signal generated by stacking different pulses is called a line code. To put it in a
simple way, the process of line coding is using the digital data obtained from an information source to
generate a voltage signal that represents the information. There are different forms of line codes that
can be used to represent the information. The terms Return to Zero (RZ) and Non–Return to Zero
(NRZ) will be used in describing these signal. Some line codes are shown next along with their power
spectral densities, advantages and disadvantages. We will consider only binary line codes although
many lines codes can work with binary and M-ary transmissions:

1. Unipolar (On–Off) Non-Return to Zero (NRZ): In this form of line codes, a bit of 1 is
represented by some positive voltage (+5 volts for example) and a bit of 0 by 0 volts
(justifying calling this signal On–Off). The pulses corresponding to binary 1 remain at the
positive voltage for the whole duration of the bit period (it does not return to zero at any time
during the bit period justifying calling this line code NRZ).

TS
RS 1
ω
0 = RS
2 2TS

Advantages:
1. Very simple to generate (has only two levels which can be easily generated
using simple digital electronics)
2. Spectrally efficient (requires the minimum amount of bandwidth for a
specific bit rate)

Disadvantages:
1. Does not provide any form of bit or frame synchronization (for long
sequences of ones or zeros, the transmitter and receiver may get
unsynchronized)
2. Does not provide any form of error detection.
3. Generally has non-zero DC (even if voltages of -5 V and +5 V are used,
unless the number of zeros and number of ones are equal). This may be a
problem for some communication systems that cannot transmit DC values.

2. Bipolar Non-Return to Zero (NRZ): In this line codes, a bit of 1 is represented in an

alternating form by some positive voltage (+5 volts for example) once and the next time a bit
 of 1 appears it will be represented by the same voltage but with a negative value (–5 volts). A
bit of 0 is represented by zero volts. The pulses corresponding to binary 1 remain at the
positive and negative voltages for the whole duration of the bit period (they do not return to
zero). The advantage of this line code over the Unipolar (On–Off) NRZ is that it has zero–DC
value because bits of 1 alternate in using the positive and negative voltages. A line code with
zero–DC is desired in some applications that require that the transmitted signal to have no
DC.

TS
RS 1
ω
0 = RS
2 2TS

Advantages:
1. Spectrally efficient (requires the minimum amount of bandwidth for a
specific bit rate)
2. Provide synchronization between the transmitter and receiver for long
sequences of Logic 1.
3. Allows some form of error detection (since two consecutive logic 1’s are
represented by a positive and negative pulses, if the receiver detects two
consecutive pulses that have the same polarity (both are positive or both are
negative), it can easily detect that there must have been an error in the
transmission).
4. Has zero DC value regardless of the number of Logic 1s and Logic 0s in the
information to be transmitted.

Disadvantages:
1. Does not provide synchronization information for long sequences of zeros.
2. Requires more sophisticated electronics to be generated because it uses
signals with 3 levels (for example, +5, 0, and -5 V)

3. Manchester (Bi-Phase): In this line code, a bit of 0 is represented by some positive voltage
for the first half of the bit period and some negative voltage for the second half of the bit
period. A bit of 1 is simply the negative of the zero bit so it is represented by the negative
voltage for the first half of the bit period and the positive voltage for the second half of the
bit. Unlike previously discussed line codes, which carry the information in the levels of
the pulses, since each of the two binary values (0 and 1) in this line code are transmitted
using pulses that have half of their duration being a high voltage (or positive voltage)
and the other half being a low voltage (or negative voltage), the information is not
carried in the levels but in the transition from high to low voltage or vice versa in the
middle of the pulse representing each digital bit. A transition from high to low may
represent a zero while a transition from low to high would then represent a one. This
line code is very good for insuring synchronization between the transmitter and receiver. For
consecutive bits that are equal, a transition may occur at the border of bits. These transitions
are simply ignored and do not carry information.

RS 1
= RS 1.5 RS
2 2TS

Advantages:
1. Provides full synchronization information for long sequences of zeros and
long sequences of ones. Also, this line code can easily be used to also
provide frame synchronization information by simply transmitting bits with
both of their two parts being high voltage or both are low voltage, where the
transmitter can use this method to signal to the receiver that a byte, for
example, has ended and a new byte it starting.
2. Allows some form of error detection (since a bit has two parts, so if the two
received parts of a bit have the same value, an error may have happened (this
assumes that frame synchronization as discussed in point 1 above is not used)
3. Has zero DC value regardless of the number of Logic 1s and Logic 0s in the
information to be transmitted.

Disadvantages:
1. Requires more bandwidth for transmission (approximately 1.6 times the
bandwidth of the previous line codes) since there are more transitions in the
signal of this line code compared to the previous line codes for the same bit
rate.
2. Although this line code has only two levels, it is slightly more complicated to
generate compared to Unipolar NRZ line code.
Digital Modulation Methods

Several types of digital modulations exist. The three basic types of digital modulations are named:

1. Amplitude Shift Keying (ASK)

2. Phase Shift Keying (PSK)
3. Frequency Shift Keying (FSK) (Will not be discussed here)

Next we describe several of these modulations and shown their time domain pulses and their
constellation.

Amplitude Shift Keying (ASK)

Binary Amplitude Shift Keying (BASK)

In ASK, the different pulses all have the same phase but different amplitudes. In the binary amplitude
shift keying (BASK) modulation technique we transmit one of two pulses for each bit:

1) for logic “0”, we transmit nothing ( s0 (t ) = 0 for 0 ≤ t ≤ TS )

2) for logic “1”, we transmit a modulated pulse with magnitude A
( s1 (t ) = A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS )

Since all the pulses have the same phase (you can think of the first signal as
s 0 (t ) = 0 ⋅ cos ( 2π f c t ) for 0 ≤ t ≤ T S ), the constellation of this modulation technique (the
representation of the pulses of the modulation algorithm) becomes

M-ary Amplitude Shift Keying (M-ary ASK)

In this digital modulation technique, we transmit one of M pulses, where M is a power of two
number such that
 M = 2n .

The quantity n here is equal to the number of bits that are carried by each transmitted pulse. To
determine which pulse to transmit in each case, we will have to divide the sequence of information
bits into groups of n consecutive bits. The combination of bits will be one of M = 2 n possible
combinations that will determine the pulse to be transmitted for these n bits. So, in this modulation
we will transmit one of the following pulses:

1) for bit sequence “00..00”, transmit s00…00 (t ) = 0 for 0 ≤ t ≤ TS

2) for bit sequence “00..01, transmit s 00…01 (t ) = A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS
3) for bit sequence “00..10, transmit s 00…10 (t ) = 2A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS
4) for bit sequence “00..11, transmit s 00…11 (t ) = 3A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS

: :

M) for bit sequence “11..11, transmit s11…11 (t ) = ( M − 1) A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤ T S

Since all pulses of this modulation technique have the same phase, the representation of all pulses in
the constellation fall on a straight line. The constellation of this modulation algorithm is shown
below.

Clearly in this case, the power required for transmitting different pulses is different and the average
power of transmission can be obtained easily by averaging all powers assuming that different bit
sequences have equal probabilities.

Phase Shift Keying (PSK)

Binary Phase Shift Keying (BPSK)

In this digital modulation technique, we transmit one of two pulses for each bit that are similar in
magnitude but have a 180° phase shift between them (one of them is the negative of the other):

1) for logic “1”, we transmit the pulse s1 (t ) = A ⋅ cos ( 2π f ct ) for 0 ≤ t ≤ TS

2) for logic “0”, we transmit the pulse s0 (t ) = − A ⋅ cos ( 2π f ct ) for 0 ≤ t ≤ TS

The constellation of the modulation algorithm in this case becomes.

It is worth mentioning that although this algorithm is very similar to the BASK, this is better because
with the same probability of error as a BASK, this modulation technique would require less power
(verify this).

Quadrature Phase Shift Keying (QPSK)

In this digital modulation technique, we transmit one of four pulses that correspond to a group of 2
bits. All pulses have equal magnitudes but are separated in phase by 90° (each two are negative of
each other). Because of the phase shift of 90° between the different pulses, two of the pulses will be
represented by dots on the x-axis and two will be represented by dots on the y-axis. This means that
the 4 pulses are

1) for logic “00”, we transmit the pulse s 00 (t ) = A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS

2) for logic “01”, we transmit the pulse s 01 (t ) = A ⋅ sin ( 2π f c t ) for 0 ≤ t ≤TS
3) for logic “10”, we transmit the pulse s10 (t ) = −A ⋅ cos ( 2π f c t ) for 0 ≤ t ≤TS
4) for logic “11”, we transmit the pulse s11 (t ) = −A ⋅ sin ( 2π f c t ) for 0 ≤ t ≤TS

The constellation now is a 2-D constellation that looks like the following

−A A

−A
8-ary Phase Shift Keying (8-ary PSK)

The QPSK modulation discussed above can be extended further where more symbols are added so
instead of having 4 symbols in the QPSK, we now have 8 symbols. We transmit one of 8 pulses such
that each pulse now carries 3 bits. All pulses will have equal magnitudes but are separated in phase by
45°. Therefore, the transmitted pulses are

⎛ 0π ⎞
1) s 000 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 2 ⎠
⎛ π⎞
2) s 001 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4⎠
⎛ 2π ⎞
3) s 010 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠
⎛ 3π ⎞
4) s 011 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠
⎛ 4π ⎞
5) s100 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠
⎛ 5π ⎞
6) s101 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠
⎛ 6π ⎞
7) s110 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠
⎛ 7π ⎞
8) s111 (t ) = A ⋅ cos ⎜ 2π f c t − ⎟ for 0 ≤ t ≤TS
⎝ 4 ⎠

So the constellation of this modulation can easily be shown to be the following (remember that
cos (α ± β ) = cos (α ) cos ( β ) ∓ sin (α ) sin ( β ) ):
 A

−A A

−A

PSK Modulations with higher power of two values of M are possible but as the value of M
increases, the modulation algorithm becomes more inefficient as different points in its constellation
become very close to each other resulting in high probability of error compared to the required power
of transmission. We can easily modify the pulses that are required to transmit the information such
that the average power of transmission is reduced and at the same time, the probability of error is
either lowered or at least maintained at its level. In this case, other modulation techniques in which
the constellation pints are distributed more evenly are used. These modulation techniques are called
Quadrature Amplitude Modulation (QAM), and from their name, they are a mix of both ASK and
PSK.

Quadrature Amplitude Modulation (QAM)

Often, we need to transmit more bits per symbol than what simple digital modulation techniques
described in above and in the previous lecture can practically do. We have seen that M-ary PSK can
be used for transmitting 3 bits per symbol or higher bits per symbol but once the number of bits per
symbol increases, the efficiency (probability of error) of the modulation increases significantly
because the points of the constellation are confined to a circle of a specific radius. Redistributing the
constellation points in a different type of modulation gives better performance in terms of probability
of error at the expense of a more complicated system and a no longer constant amplitude for the
different pulses as it is the case with all PSK modulations.

We can generate M signals that have different amplitudes and/or phases such that each carries n
bits ( n = log 2 M ) by having the different pulses in the form
 s i (t ) = ai ⋅ cos ( 2π f c t ) + bi ⋅ sin ( 2π f c t ) for 0 ≤ t ≤Ts
and i = 1, 2, … , M

where {ai , bi } is a pair of coefficients that are given by a matrix of pairs depending on the desired
M-ary QAM modulation scheme (or sometimes called M-QAM for short). For example, 16-QAM
will have a matrix for {ai , bi } given by
⎡ (−3,3) (−1,3) (1,3) (3,3) ⎤
⎢ (−3,1) (−1,1) (1,1) (3,1) ⎥⎥
{ i i} ⎢
a , b = ⎢
(−3, −1) (−1, −1) (1, −1) (3, −1) ⎥
⎢ ⎥
⎣(−3, −3) (−1, −3) (1, −3) (3, −3) ⎦

So in 16-QAM, each 4 bits are combined and based on the pattern of these 4 bits (giving 16
possibilities), the corresponding pair {ai , bi } is selected from the above matrix and the corresponding
pulse for these 4 bits is generated and transmitted.

The constellation of the above 16-QAM is shown below:

More complicated constellations are obtained for higher levels of QAM. Once the pulses have been
transmitted, the received pulses can be plotted on the same constellation. The pulse assumed to have
been transmitted is the one with the point in the constellation that is closest to that of the received
pulse. QAM is widely used in digital communication systems including the V.92 56k bps modem.