ELECTRICAL AND ELECTRONICS

ENGINEERING

EEE 401: COMMUNICATION

PRINCIPLES
 COMMUNICATION SYSTEMS

Communication Systems refer to the transmission, reception and processing of information or messages by
electrical means. The commonest forms are telegraphy, telephony, radio and TV broadcasting.
The transmission system is usually through metallic wires and cables, optical fibers or through space when
radio signals are employed.

The block diagram of a basic communication system is as shown Figure 1.

Information
Transmitter Channel Receiver Destination
Source

Noise Source

Figure 1: Basic Communication System.

(a) Information Source: The information or message to be communicated may originate as speech
(or music), picture (moving or stationary), printed material or data generated by a computer or
obtained from some physical phenomena, whatever the source, the message must first be
converted into electrical signals before transmission takes place. Then signal is received by the
receiver and reconverted to a desired form at the destination. Suitable transducers are required
to perform the conversion processes at the transmitter and receiver ends. The table below
illustrates various types of transducers that may be employed.
(b) Transmitter: Electrical signals generated by the transducer are processed by the transmitter.
Such signal processing techniques involve amplification to boost the signal current or voltage
amplitude or power level; others are modulation, filtering, and coding e.t.c.
(c) Receiver: The reverse process of converting the electrical signal back into the form of message
desired at the destination takes place in the receiver. Filtering, demodulation and decoding are
signal processing techniques involve in the receiver.
(d) Disturbing Influences: Electrical noise is the most important disturbing influence that limits the
performance of a communication system. There are various types of noise depending on the
type of transmission medium.
(i) Thermal Noise: exists on conducting wires, resulting from the statistical fluctuations of the
electrons in the wire. It increases with temperature and is prominent when the signal level is
low. Noise also occurs at the transmitter and receiver ends, but this can be reasonably controlled
by appropriate circuit design.

In radio communication, noise introduced in the transmission channel can be grouped into two
broad categories; Man-made noise and Natural noise.

2
 Man-made Noise: This noise is from electric motors, switches, automobile ignition, electric
welding equipment or high-voltage power lines e.t.c.
Natural Noise: This is from lighting discharges during thunderstorms, usually referred to as static.
Other sources of natural noise are Solar noise arising from the radiation from the sun and
disturbances occurring in the sun, and cosmic noise coming from distant stars in the galaxy.

Table 1: Types of the Transducer that may be employed

Message Transmitter Receiver Message

(Input)) (Transducer) (Transducer) (Output)

Speech

Sound

Music Microphone Loudspeaker Sound

TV

Picture

Fax (Printed Material) TV Camera TV Camera Picture or text

Photodetector Picture or text

Physical Temperature Bimetallic Strip Electrical Signal

Phenomenon Thermocouple

Pressure Piezoelectric Crystal Recording

or strain gauge Instruments

NOTE: The form of the message output at the destination needs not be the same as the form
transmitted since an electrical signal can be converted into any form of message using
transducers.

Classification of Communication Systems

This can be done in a number of ways:

3
(a) By types of Systems
(i) Telegraphy: Transmission of written texts. Familiar form of this is the telegram which is a
written material intended to be transmitted by telegraphy for delivering to an address. Telex is
another form, which is the message through a network of telegraphy exchanges.
(ii) Telephony: A form of telecommunication set up for the transmission of speech or other sounds.
(iii) Broadcasting: A radio communication service in which the transmission is intended for direct
reception by the general public. The most familiar forms are sound Broadcasting and Television
Broadcasting. We have terrestrial broadcasting when the transmission from the transmitter to the
receiver is close to the earth surface.
(iv) Satellite-Broadcasting: occurs when satellites are used as repeater stations between the
transmitter and receiver.

(b) Classification by Signal types

There are two types: Analog and Digital

Analog Message: A physical quantity that varies with time, usually in a smooth and continuous version. e.g.
the output signal from a microphone when speech (or an acoustic signal) impinges on it. The light intensity
variation at some points in a TV signal is also analogue in form.

Digital Message: This is an ordered sequence of symbols selected from a finite set of discrete elements e.g.
the keys pressed on a computer.

This signal is periodic with period 𝑇 = 2𝜋

A
T
A

Figure 2: Periodic signal

Digital and Analog Signals

A digital information source produces a finite set of possible messages. A typewriter is a good example of a
digital source. There are a finite number of characters (messages) that can be emitted by this source. The
output voltage describes the information in the sound, and it is distributed over a continuous range of values.

A digital communication system transfers information from a digital source to the intended receiver (also
called the sink).

An analog communication system transfers information from an analog source to the sink.

4
Strictly speaking, a digital waveform is defined as a function of time that can have only a discrete set of
amplitude values.

An analog waveform is a function of time that has a continuous range of values.

Advantages of Digital Communication

1. Relatively inexpensive digital circuits may be used.
2. Privacy is preserved by using data encryption.
3. Greater dynamic range (the difference between the largest and smallest values) is possible.
4. Data from voice, video, and data sources may be merged and transmitted over a common
digital transmission system.
5. In long-distance systems, noise does not accumulate from repeater to repeater.
6. Errors in detected data may be small, even when there is a large amount of noise on the
received signal.
7. Errors may be often be corrected by the use of coding.

Disadvantages of Digital Communication

1. Generally, more bandwidth is required than that for analog system.
2. Synchronization is required.

(c) Classification by types of Services

There are two types: Fixed service and Mobile service.
Fixed Service: This occurs when the transmitter and receiver are fixed.
Mobile Service: This occurs when either the transmitter or receiver or both are moving at
different speeds. Mobile service can further be divided into
(i) Land-mobile service e.g. cellular mobile like GSM.
(ii) Aero-nautical mobile, used in communicating with aircraft.
(iii) Maritime mobile, used for communicating with ships at sea.
Any of these services can involve the use of the satellite and we then have fixed satellites, mobile-
satellite or satellite broadcasting services.

(d) Classification by Frequency of Operation

The radio frequency spectrum used for communication purposes is divided into frequency
bands. Different frequency bands are better suited for different types of services and
transmission media.

See Table 1.

5
 MODULATION PROCESSES IN A COMMUNICATION SYSTEM
Modulation is one of the important signals processing technique undertaking at the transmitter stage. To be
able to transmit the signal efficiently, the information bearing signal (baseband signal) must be processed in
some manner before sending the signal into the transmission medium.

Most commonly, the baseband signals must be shifted to higher frequencies for efficient transmission. It is
a well-known theory of electromagnetic radiation that an efficient radiator of an electric energy (that is the
antenna or aerial) must have a dimension of the order of a wavelength in size.

Take, for instance, an audio signal of 𝑓 = 1 𝑘𝐻𝑧 (𝑜𝑟 𝜆 = 300 𝑘𝑚). This would require a radiator of about
100 – 200 km in size if it were to radiate this signal efficiently. However, by translating the baseband to a
higher frequency of say, 1 𝑀𝐻𝑧, the size of an efficient radiator is reduced by a factor of 1000, down to about
100 - 200 m, which is more practicable. The use of higher frequencies also provides wider bandwidth which
can accommodate other baseband signals for increased information transfer.

6
Modulation is the process of superimposing a modulating signal into a carrier wave. Modulation process
involves varying the amplitude, frequency or phase (or combination of these) of the high frequency sine-
wave carrier, in accordance with the information to be transmitted. This is referred to as sine wave or
continuous wave (c-w) modulation.

Digital modulation occurs when the carrier is a pulse train of high bit-rates instead of a high frequency
continuous carrier wave. The pulse parameters suitable for modulation include amplitude, duration or width
and pulse-position referred to as PAM, PDM (or PWM) and PPM systems, respectively.

Modulation is required for the following reasons:

i. For large range transmission: 𝐻𝑚 = 𝜆⁄4, the frequency of carrier should be high for 𝐻𝑚 to be small.
ii. For the quality of transmission: 𝑃 ∝ 1⁄ 2 = 𝑙⁄ 2. Where 𝑙 is the length of antenna.
𝜆 𝜆
iii. To avoid overlapping of signals: Use different carrier frequency to transmit different signal.

AMPLITUDE MODULATION
Amplitude Modulation is defined as a system of modulation in which the amplitude of the carrier is made
proportional to the instantaneous amplitude of the modulating (baseband) signal.

Figure 3: Amplitude modulation process.

Let 𝑒𝑐 = 𝐸𝑐 𝑠𝑖𝑛 𝑤𝑐 𝑡 be the carrier signal 1

𝑒𝑚 = 𝐸𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡 be the modulating signal 2

𝑒 = 𝐴 𝑠𝑖𝑛 𝑤𝑐 𝑡 be the modulated signal 3

Where 𝑤𝑚 = 2𝜋𝑓𝑚 and 𝑤𝑐 = 2𝜋𝑓𝑐 . Then, by the definition above,

𝐴 = 𝐸𝑐 + 𝑘𝑒𝑚 4

Substitute equation (2) into (4)

𝐸
𝐴 = 𝐸𝑐 (1 + 𝑘 𝐸𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡) 5
𝐶

7
We may put 𝑘 = 1, for simplicity, then

𝐸
𝐴 = 𝐸𝑐 (1 + 𝐸𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡) 6
𝐶

𝐴 = 𝐸𝑐 (1 + 𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡) 7
𝐸𝑚
where 𝑚 = 𝐸𝑐
is the modulation index 8
if 𝑚 < 1, this implies under modulation,

if 𝑚 = 1, this implies perfect or moderate modulation and;

if 𝑚 > 1, over modulation occurs which leads to interference and the carrier experiences a 180° phase
reversal which causes additional sidebands and hence, the wave gets distorted. Such overmodulated wave
causes interference, which cannot be eliminated.

Substitute equation (7) into equation (3)

𝑒 = 𝐸𝑐 (1 + 𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡) 𝑠𝑖𝑛 𝑤𝑐 𝑡 9
1
Using the trigonometric relation 𝑠𝑖𝑛 𝑥 𝑠𝑖𝑛 𝑦 = 2 [𝑐𝑜𝑠(𝑥 − 𝑦) 𝑐𝑜𝑠(𝑥 + 𝑦)], equation (9) becomes
𝑚𝐸𝑐 𝑚𝐸𝑐
𝑒 = 𝐸𝑐 𝑠𝑖𝑛 𝑤𝑐 𝑡 + 2
𝑐𝑜𝑠(𝑤𝑐 − 𝑤𝑚 ) 𝑡 − 2
𝑐𝑜𝑠(𝑤𝑐 + 𝑤𝑚 ) 𝑡 10

Em
E m sin  m t
Ec
Emax
Ec A

Emin

Figure 4: AM wave

8
Equation (10) shows that the equation of an AM wave contains three terms: the first is identical with the
unmodulated carrier signal, the other two represent two side-frequencies with the lower frequency (𝑓𝑐 −
𝑓𝑚 ) and the upper frequency (𝑓𝑐 + 𝑓𝑚 ) as illustrated in Figure 5.
𝑤
Note that 𝑓 = 2𝜋, 𝑓𝑚𝑖𝑛 = 𝑓𝑐 − 𝑓𝑚 and 𝑓𝑚𝑎𝑥 = 𝑓𝑐 + 𝑓𝑚 .

carrier

mEc mEc
2 2

fc − fm fc + fm

2 fm
Figure 5: Frequency representation of AM wave

𝐵𝑊 = 𝑓𝑚𝑎𝑥 − 𝑓𝑚𝑖𝑛 = 𝑓𝑐 + 𝑓𝑚 − (𝑓𝑐 − 𝑓𝑚 ) = 2𝑓𝑚 11

It can be seen that the AM signal occupies a bandwidth of 2𝑓𝑚 , which is twice the frequency of the
modulating (baseband) signal. From Figure 4, we have
𝐸𝑚𝑎𝑥 −𝐸𝑚𝑖𝑛
𝐸𝑚 = 12
2

𝐸𝑚𝑎𝑥 +𝐸𝑚𝑖𝑛
and 𝐸𝑐 = 2
13

𝐸𝑚 𝐸 −𝐸
𝑚= 𝐸𝑐
= 𝐸𝑚𝑎𝑥 + 𝐸𝑚𝑖𝑛 14
𝑚𝑎𝑥 𝑚𝑖𝑛

Where 𝐸𝑚𝑎𝑥 = 𝐸𝑐 + 𝐸𝑚 is the maximum amplitude of the modulated wave at 𝑠𝑖𝑛 𝑤𝑚 𝑡 = 1 and 𝐸𝑚𝑖𝑛 =
𝐸𝑐 − 𝐸𝑚 is the minimum amplitude of the modulated wave at 𝑠𝑖𝑛 𝑤𝑚 𝑡 = −1.
This gives a practical method of calculating the modulation index from a waveform displayed on an
oscilloscope.
If the modulating signal is not a single sinusoid but contains a range of frequencies 𝑓1 to 𝑓2, the modulated
signal will in addition to the carrier signal, contain two sidebands as shown in Figure 6.

f1 f2 f
(a)

9
 carrier

fc − f2 fc − f1 fc fc + f1 fc + f2

2 f2

(b)
Figure 6: Side band frequency

Note that the message to be transmitted is contained only in the sidebands and not in the carrier signal. The
same message contained in the lower sideband is duplicated in the upper side band; and the bandwidth of
the modulated wave is 2𝑓2, which is twice the highest frequency contained in the modulating signal.

Power Relations in AM Waves

If the AM wave is developed across a load of resistance (for example an antenna resistance), the total power
in the modulated wave is expressed as
2 2
𝐸̅𝑐2 𝐸̅𝐿𝑆𝐵 𝐸̅𝑈𝑆𝐵
𝑃𝑡 = 𝑅
+ 𝑅
+ 𝑅
15

where the voltages 𝐸̅𝑐 , 𝐸̅𝐿𝑆𝐵 , and 𝐸̅𝑈𝑆𝐵 are the rms values of the voltage amplitude.
The power in the unmodulated carrier is

𝐸𝑐 2
𝐸̅𝑐2 ⁄ 𝐸𝑐 2
√2
𝑃𝑐 = 𝑅
= ( 𝑅 ) = 2𝑅
16

From equation (16), 𝐸𝑐 2 = 2𝑃𝑐 𝑅

𝑚𝐸𝑐 2
⁄
2√2
Similarly, 𝑃𝐿𝑆𝐵 = 𝑃𝑈𝑆𝐵 = ( 𝑅 ) 17

𝑚2 𝐸𝑐 2
𝑃𝐿𝑆𝐵 = 𝑃𝑈𝑆𝐵 = ( 8𝑅
) 18

Substitute 𝐸𝑐 2 into equation (18)

𝑚2 𝑃𝑐
𝑃𝐿𝑆𝐵 = 𝑃𝑈𝑆𝐵 = ( 4
) 19

𝑚2 𝑃𝑐 𝑚2 𝑃𝑐
𝑃𝑡 = (𝑃𝑐 + 4
+ 4
) 20

10
 𝑚2
𝑃𝑡 = (1 + ) 𝑃𝑐 21
2

The maximum power in the AM wave occurs when 𝑚 = 1 and the total power becomes

𝑃𝑡𝑚𝑎𝑥 = 1.5𝑃𝑐 22

Example 1: A broadcast radio transmitter radiates 10 kW of power when the modulation index is 60%.
(i) How much of this is the carrier power?
(ii) What is the maximum power the transmitter can deliver without signal distortion?
(iii) How much power is contained in one of the side band?

Solution
𝑚2
(i) 𝑃𝑡 = (1 + 2
) 𝑃𝑐

𝑃𝑡 10
𝑃𝑐 = 𝑚2
= 0.62
= 8.47 𝑘𝑊
1+ 1+
2 2

(ii) 𝑃𝑡 𝑚𝑎𝑥 = 1.5𝑃𝑐 = 1.5 × 8.47 = 12.71 𝑘𝑊

𝑚2 𝑃𝑐 0.62
(iii) 𝑃𝐿𝑆𝐵 = 𝑃𝑈𝑆𝐵 = ( 4
) = 4
× 8.47 𝑘𝑊 = 0.76 𝑘𝑊

Examples 2: When the modulation index is 0.75, an AM transmitter produces 10 kW of power,

(i) How much of this is carrier power?
(ii) What are the percentage power saving if the carrier signal and one sideband were suppressed and only
one sideband is transmitted.

Solution
𝑃𝑡 10
(i) 𝑃𝑐 = 𝑚2
= 0.752
= 7.8 𝑘𝑊
1+ 1+
2 2
(ii) Power contained in one sideband is
𝑚2 𝑃𝑐 0.752
𝑃𝐿𝑆𝐵 = ( 4
) = 4
× 7.8 𝑘𝑊 = 1.1 𝑘𝑊
(7.8 + 1.1)
Percentage power saved = × 100% = 89%
10

Modulation by Several Sine Waves

Let 𝐸1, , E2, 𝐸3, etc be the amplitudes of the simultaneous modulating voltages, the total modulating voltage
is given by

𝐸𝑡 = √(𝐸1 2 + 𝐸2 2 + 𝐸3 2 +. . . . . . . . . . ) 23

11
 𝐸 𝐸 2 𝐸 2 𝐸 2
𝑚𝑡 = 𝐸 𝑡 = √(𝐸1 2 + 𝐸2 2 + 𝐸3 2 +. . . . . . . . . ) 24
𝐶 𝐶 𝐶 𝐶

or 𝑚𝑡 = √(𝑚1 2 + 𝑚2 2 + 𝑚3 2 +. . . . . . . . . . ) 25

A number of communication systems derived from AM process are:

(i) Double Side Band Total Carrier (DSBTC): This occurs when the two sidebands and the carrier
signal are transmitted that is, the full AM signals are transmitted.

Figure 7: DSBTC

(ii) Double Side Band Suppressed Carrier (DSBTC): Here, the unmodulated carrier frequency is
suppressed before transmission and only the two sidebands are transmitted. There is no loss of
information by doing this, since the information is contained only in the sidebands. The process
2
leads to a swing in power since 𝑃𝑐 = 3 𝑃𝑡(𝑚𝑎𝑥) . That is two-third (or 67%) of the total power
resides in the suppressed carrier when m = 1. Commercial AM radio transmission makes use of
this system.

Figure 8: DSBSC

(iii) Single Side Band (SSB): occurs when only one of the sidebands is transmitted; the other
sideband and the carrier signal being suppressed before transmission. This technique has the
following advantages:
(a) No loss of information, since the information is contained in each sideband.
(b) There is much power saving.
(c) There is saving in frequency bandwidth: The information is transmitted with half of the
bandwidth of the full AM signal.
This is important when several other services demand the use of radio frequency spectrum e.g.
for mobile systems or for TV systems.
The major disadvantage is the more complex circuit of the transmitter than the DSBSC leading
to more expensive radio equipment. It is mainly used for professional communication
equipment. However, the bandwidth saving advantage is becoming of greater importance; with
the result that future commercial AM transmitter will be required to adopt the system.

12
 Figure 9: SSB

(iv) Vestigial Side Band (VSB): This technique is mostly employed for video signals used in TV
systems. The bandwidth of the TV video signal is about 4 MHz. When this is used to amplitude
modulate a carrier signal, the bandwidth of the full AM signal is 8 MHz. It is therefore important
to use SSB in order to conserve the frequency spectrum. However, the characteristics of the
band pass filter required to separate the required sideband demand must be transmitted so as
to avoid some harmful effects on the TV signal.

Figure 10: VSB

Advantages of AM
1. It is easier to implement.
2. Simple circuit can be designed for demodulation.
3. It is less expensive.

Disadvantages of AM
1. It consumes much power.
2. It requires high bandwidth.
3. It has a noticeable high noise interference.

Areas of application of AM
1. Portable two-way radio.
2. Citizen band radio.
3. VHF aircraft radio.
4. Modems for computer.

13
Types of AM Wave Modulator
Square Law Modulator
This is a non-linear device like a diode which is closely related to square law. The block diagram and the circuit
are as shown in Figure 11.

(a)

(b)
Figure 11: Block and circuit diagram of square law modulator

The derivation of the modulator is as shown:

𝑉1 (𝑡) = 𝑚(𝑡) + 𝐴𝑐 cos 𝜔𝑐 𝑡 26
The mixed/ superimposed signal is passed through the non-linear device to give
𝑉2 (𝑡) = 𝑘1 𝑉1 (𝑡) + 𝑘2 𝑉12 (𝑡) 27
Where 𝑘1 and 𝑘2 are constant. Then substitute equation (26) into (27)
𝑉2 (𝑡) = 𝑘1 (𝑚(𝑡) + 𝐴𝑐 cos 𝜔𝑐 𝑡) + 𝑘2 (𝑚(𝑡) + 𝐴𝑐 cos 𝜔𝑐 𝑡)2 28
𝑉2 (𝑡) = 𝑘1 𝑚(𝑡) + 𝑘1 𝐴𝑐 cos 𝜔𝑐 𝑡 + 𝑘2 𝑚2 (𝑡) + 2𝑘2 𝐴𝑐 𝑚(𝑡) cos 𝜔𝑐 𝑡 + 𝑘2 𝐴2𝑐 cos 2 𝜔𝑐 𝑡 29
Rearranging equation (29) gives
2𝑘2
𝑉2 (𝑡) = 𝑘1 𝑚(𝑡) + 𝑘2 𝑚2 (𝑡) + 𝑘2 𝐴2𝑐 cos2 𝜔𝑐 𝑡 + 𝑘1 𝐴𝑐 (1 + 𝑚(𝑡)) cos 𝜔𝑐 𝑡 30
𝑘1

The last term in equation (30) represents the desired AM wave with the first three terms unwanted. The
bandpass filter helps to eliminate the unwanted part and the desired signal 𝑠(𝑡) is obtained as
𝑠(𝑡) = 𝑘1 𝐴𝑐 (1 + 𝑘𝑜 𝑚(𝑡)) cos 𝜔𝑐 𝑡 31
2𝑘2
where 𝑘𝑜 = 𝑘1
.

14
Switching Modulator
The major difference between the square law and switching modulator is that instead of the operating the
diode as a non-linear device. It is gated or switched ON and OFF.

(a)

(b)
Figure 12: Block and circuit diagram of switching modulator

The derivation of the modulator is as given:

𝑉1 (𝑡) = 𝑚(𝑡) + 𝑐(𝑡) 32
Where 𝑐(𝑡) = 𝐴𝑐 cos 𝜔𝑐 𝑡.
Assume that the modulating signal magnitude is very small as compared to 𝐴𝑐 (carrier signal amplitude), the
diode will be forward biased with the carrier signal controlling the switch 𝑐(𝑡) > 0 i.e., ON and reversed
biased 𝑐(𝑡) < 0 i.e., OFF. Therefore, the output of the diode is
𝑉 (𝑡), 𝑖𝑓 𝑐(𝑡) > 0
𝑉2 (𝑡) = { 1 33
0, 𝑖𝑓 𝑐(𝑡) < 0
Equation (33) is approximated to be
𝑉2 (𝑡) = 𝑉1 (𝑡)𝑥(𝑡) 34
1
Where 𝑥(𝑡) is the periodic pulse train with time period 𝑇 = 𝑓 and Fourier series is expressed as
𝑐

Figure 13: Periodic pulse train

15
 1 2 (−1)𝑛 −1
𝑥(𝑡) = + ∑∞
𝑛=1 cos(2𝜋(2𝑛 − 1)𝑓𝑐 𝑡) 35
2 𝜋 2𝑛−1
1 2 2
𝑥(𝑡) = + 2 𝜋
cos(2𝜋𝑓𝑐 𝑡) − 3𝜋 cos(6𝜋𝑓𝑐 𝑡) + ⋯ 36
Substitute equation (36) into (34)
1 2 2
𝑉2 (𝑡) = [𝑚(𝑡) + 𝐴𝑐 cos 𝜔𝑐 𝑡] [2 + 𝜋 cos(2𝜋𝑓𝑐 𝑡) − 3𝜋 cos(6𝜋𝑓𝑐 𝑡) + ⋯ ] 37
Simplifying equation (37) gives
𝐴𝑐 4 𝑚(𝑡) 2𝐴
𝑉2 (𝑡) =
2
[1 + (𝜋𝐴 ) 𝑚(𝑡)] cos(2𝜋𝑓𝑐 𝑡) + 2 + 𝜋 𝑐 cos2(2𝜋𝑓𝑐 𝑡) −
𝑐
2𝑚(𝑡) 2𝐴
3𝜋
cos(6𝜋𝑓𝑐 𝑡) − 3𝜋𝑐 cos(2𝜋𝑓𝑐 𝑡) cos(6𝜋𝑓𝑐 𝑡) + ⋯ 38
The first term is the desired AM wave while others are unwanted signals which are filter out by the
bandpass filter. So, 𝑠(𝑡) is obtained as
𝐴𝑐 4
𝑠(𝑡) = [1 + (𝜋𝐴 ) 𝑚(𝑡)] cos(2𝜋𝑓𝑐 𝑡) 39
2 𝑐
Exercise
1. The equation of AM wave is given as 𝑠(𝑡) = 20(1 + 0.8 cos(2𝜋 × 102 𝑡)) cos(4𝜋 × 105 𝑡). Find:
a. Carrier power.
b. Total sideband power.
c. Bandwidth of the AM wave.
2. A modulating signal 𝑚(𝑡) = 10 cos(2𝜋 × 103 𝑡) is amplitude modulated with a carrier signal 𝑐(𝑡) =
50 cos(2𝜋 × 105 𝑡). Find the:
a. Modulation index.
b. Carrier power
c. Power required for transmitting the AM wave.
3. A carrier wave frequency 𝑓𝑐 = 1 𝑀𝐻𝑧 with a peak voltage of 20𝑉 is used to modulate a signal at
frequency 1 𝑘𝐻𝑧 with a peak voltage of 10𝑉. Find the:
a. Modulation index.
b. Frequency of the modulated wave.
c. Bandwidth.
4. 𝑦(𝑡) = 10 cos(1800𝜋𝑡) + 20 cos(2000𝜋𝑡) + 10 cos(2200𝜋𝑡). Find the modulation index of the
given wave.

ANGLE MODULATION
This is the process of modulation in which the frequency or phase of a carrier signal varies according to the
message. The angle modulation can be classified into: frequency and phase modulations.

FREQUENCY MODULATION (FM)

Frequency modulation is a system in which the amplitude of the unmodulated carrier signal is made constant
while its frequency is varied in accordance with the variations of the modulating signal. The amplitude and
phase of the carrier signal remains constant.

16
 Figure 14: Frequency modulation process.

The instantaneous frequency ƒ, of the FM signal is given by

𝑓 = 𝑓𝑐 (1 + 𝑘 𝑚(𝑡))
40
𝑓 = 𝑓𝑐 (1 + 𝑘𝐸𝑚 𝑐𝑜𝑠 𝑤𝑚 𝑡)

where 𝑘 is a constant of proportionality, 𝑚(𝑡) = 𝐸𝑚 𝑐𝑜𝑠 𝑤𝑚 𝑡 is the modulating signal, 𝑓𝑐 is the carrier
frequency.
Let the instantaneous amplitude of the FM signal be written as
𝑒 = 𝐴 𝑠𝑖𝑛 𝜃 41

Where 𝐴 is the amplitude of the modulated signal and also amplitude of the carrier signal, 𝜃 is the angle of
the modulated signal and is given as
𝜃 = ∫ 𝑤𝑑𝑡 = 2𝜋 ∫ 𝑓𝑑𝑡 42

Substitute for 𝑓 in equation (40) into equation (42), we have

𝜃 = 2𝜋 ∫ 𝑓𝑐 (1 + 𝑘𝐸𝑚 𝑐𝑜𝑠 𝑤𝑚 𝑡)𝑑𝑡 43

𝐸𝑚
𝜃 = 2𝜋𝑓𝑐 (𝑡 + 𝑘 𝑠𝑖𝑛 𝑤𝑚 𝑡) 44
𝑤𝑚

𝐸𝑚 2𝜋𝑓𝑐
𝜃 = 2𝜋𝑓𝑐 𝑡 + 𝑘 2𝜋𝑓𝑚
𝑠𝑖𝑛 𝑤𝑚 𝑡 45

𝐸𝑚 𝑓𝑐
𝜃 = 2𝜋𝑓𝑐 𝑡 + 𝑘 𝑓𝑚
𝑠𝑖𝑛 𝑤𝑚 𝑡 46

Substitute equation (46) into equation (41) gives

𝐸𝑚 𝑓𝑐
𝑒 = 𝐴 𝑠𝑖𝑛( 2𝜋𝑓𝑐 𝑡 + 𝑘 𝑓𝑚
𝑠𝑖𝑛 𝑤𝑚 𝑡) 47

From equation (40), the instantaneous frequency of the FM signal deviates from the carrier frequency 𝑓𝑐 by
an amount 𝑘𝐸𝑚 𝑐𝑜𝑠 𝑤𝑚 𝑡, the maximum value is
𝛿 = 𝑘𝐸𝑚 𝑓𝑐 48

That is, when 𝑐𝑜𝑠 𝑤𝑚 𝑡 = 1 is the maximum frequency deviation. Equation (48) in (47) becomes
𝛿
𝑒 = 𝐴 𝑠𝑖𝑛( 2𝜋𝑓𝑐 𝑡 + 𝑓 𝑠𝑖𝑛 𝑤𝑚 𝑡) 49
𝑚

or 𝑒 = 𝐴 𝑠𝑖𝑛( 2𝜋𝑓𝑐 𝑡 + 𝑚𝑓 𝑠𝑖𝑛 2𝜋𝑓𝑚 𝑡) 50

17
 𝛿 𝐸𝑚 𝑓𝑐
where 𝑚𝑓 = =𝑘 51
𝑓𝑚 𝑓𝑚

Equation (51) is the modulation index of the FM signal. The value of the modulation index in equation (51)
can be used in classifying whether the FM is narrowband or wideband.
Note: The following points:
(i) The amplitude of the FM signals is constant only the frequency varies within the range 𝑓𝑐 ± 𝛿;
(ii) The maximum deviation of the frequency from the carrier frequency 𝑓𝑐 is proportional to the
amplitude 𝐸𝑚 of the modulating signal. That is 𝛿 ∝ 𝐸𝑚 ;
(iii) The FM modulation index 𝑚𝑓 is inversely proportional to the frequency 𝑓𝑚 of the modulating signal
1 𝛿
when 𝛿 (or 𝐸𝑚 ) is constant i.e., 𝑚𝑓 ∝ 𝑓 = 𝑓 .
𝑚 𝑚

Frequency Spectrum of FM Signal

From equation (49)
𝑒 = 𝐴 𝑠𝑖𝑛( 2𝜋𝑓𝑐 𝑡 + 𝑚𝑓 𝑠𝑖𝑛 𝑤𝑚 𝑡) 52

This is a complete expression involving the sine of the sine of an expression. The solution involves the use of
the Bessel function as follows:
𝑒 = 𝐴{𝐽𝑜 (𝑚𝑓 ) 𝑠𝑖𝑛 𝑤𝑐 𝑡 + 𝐽1 (𝑚𝑓 )(𝑠𝑖𝑛(𝑤𝑐 + 𝑤𝑚 ) 𝑡 − 𝑠𝑖𝑛(𝑤𝑐 − 𝑤𝑚 ) 𝑡) + 𝐽2 (𝑚𝑓 )(𝑠𝑖𝑛(𝑤𝑐 +
2𝑤𝑚 ) 𝑡 + 𝑠𝑖𝑛(𝑤𝑐 − 2𝑤𝑚 ) 𝑡) + 𝐽3 (𝑚𝑓 )(𝑠𝑖𝑛(𝑤𝑐 + 3𝑤𝑚 ) 𝑡 − 𝑠𝑖𝑛(𝑤𝑐 −
3𝑤𝑚 ) 𝑡)+. . . . . . . . . } 53

Each pair of sidebands is multiplied by 𝐽 coefficients which are Bessel functions of the first kind and order
denoted by the subscript with argument 𝑚𝑓 . 𝐽1 (𝑚𝑓 ) may be shown to be a solution of an equation of the
form.
2 𝑑2 𝑦 𝑑𝑦
(𝑚𝑓 ) 𝑑𝑚𝑓 2
+ (𝑚𝑓 ) 𝑑𝑥 + (𝑚𝑓 2 − 𝑛2 )𝑦 = 0 54

The solution of this 2nd order differential equation is

2 4 6 8
(𝑚𝑓 ) (𝑚𝑓 ) (𝑚𝑓 ) (𝑚𝑓 )
𝑚𝑓 1 𝑛
2 2 2 2
𝐽𝑛 (𝑚𝑓 ) = ( ) [ − + − + ......]
2 𝑛! 1! (𝑛 + 1)! 2! (𝑛 + 2)! 3! (𝑛 + 3)! 4! (𝑛 + 4)!

In practice, it is not necessary to calculate 𝐽𝑛 (𝑚𝑓 ) from the above equation. The values are readily available
in tabular form or graphical form in standard mathematical texts.

Features of Narrowband FM
1. It has a small bandwidth.
2. Modulation index 𝑚𝑓 < 1.
3. Its spectrum consists of a carrier, the USB and LSB.
4. It is used in mobile communication such as in ambulances, taxis cab and police wireless.

Features of Wideband FM
1. It has an infinite bandwidth.

18
 2. Modulation index 𝑚𝑓 > 1.
3. Its spectrum consists of a carrier, infinite sidebands which are located around it.
4. It is used in FM radio and TV.

Exercise
1. A sinusoidal modulating waveform of amplitude 5𝑉 and a frequency of 2𝑘𝐻𝑧 is applied to FM
generator which has a frequency sensitivity of 40𝐻𝑧/𝑉. Calculate:
a. Frequency deviation.
b. Modulation index.
c. Bandwidth.

2. A FM wave is given by 𝑠(𝑡) = 20 cos(8𝜋 × 106 𝑡 + 9 sin(2𝜋 × 103 𝑡)). Calculate the:
a. Frequency deviation.
b. Bandwidth of the signal.
c. Power of the FM wave.

PHASE MODULATION (PM)

Phase modulation is similar in some respects to frequency modulation. The two are usually regarded together
as Angle Modulation. Phase modulation is used in mobile communication systems.

Figure 15: Phase modulation process.

Expressing the unmodulated carrier signal as

𝑒 = 𝐴 𝑠𝑖𝑛(𝑤𝑐 𝑡 + 𝜑) 55

If the phase 𝜑 is varied such that its magnitude is proportional to the instantaneous amplitude of the
modulating signal, the resulting signal is phase-modulated. We have for PM
𝑒 = 𝐴 𝑠𝑖𝑛(𝑤𝑐 𝑡 + 𝜑𝑚 𝑠𝑖𝑛 𝑤𝑚 𝑡) 56

19
where 𝜑𝑚 is the maximum value of the phase change introduced by the modulating signal, and is
proportional to the maximum amplitude of the modulating signal. For similarity with the expression for FM,
in equation (49), We may write
𝑒 = 𝐴 𝑠𝑖𝑛(𝑤𝑐 𝑡 + 𝑚𝑝 𝑠𝑖𝑛 𝑤𝑚 𝑡) 57

where 𝑚𝑝 is the modulation index for PM. Although the forms of equations (49) and (57) are similar
suggesting similarities between FM and PM, there is a significant difference between them, arising from their
definitions.
𝛿 𝑓𝑐
In FM, 𝑚𝑓 = = 𝑘𝐸𝑚
𝑓𝑚 𝑓𝑚

𝐸𝑚
That is 𝑚𝑓 ∝ 𝑓𝑚
the modulation index is proportional to the amplitude of the modulating voltage and
inversely proportional to the modulating frequency.

In PM, 𝑚𝑝 and 𝐸𝑚 are independent of 𝑓𝑚 .

Under identical conditions, FM are PM are not distinguishable for a single modulating frequency, 𝑓𝑚 but
when 𝑓𝑚 is changed, however, 𝑚𝑝 will remain constant whereas 𝑚𝑓 will change in inverse proportion to 𝑓𝑚 .

DEMODULATOR (OR DETECTOR OR RECEPTION) OF MODULATED SIGNALS

The Superheterodyne Radio Receiver
The superheterodyne radio receiver is the commonest form of receiver used in practice to receive AM and
FM transmissions.

Audio &
antenna power
amplifiers
loudspeaker
f0 − fs
fs IF
RF range mixer detector
amplifier

f0

Local
oscillator

Gang tuning

Figure 5.1: Block diagram of a superheterodyne receiver.

The modulated signal coming in through the antenna is first passed into the Radio Frequency (RF) stage,
which is a tuned and tunable circuit.
The wanted signal is selected and unwanted frequency rejected through filtering. The signal is then combined
with another RF signal (local oscillator) to generate a signal at low, fixed frequency (intermediate frequency,
IF). The IF signal contains the same modulation as the original modulated signal. A constant frequency

20
difference (𝑓0 − 𝑓𝑠 ) is maintained between the local oscillator and RF circuits, normally ganged together and
operated in union by one control knob.
The IF of a standard AM broadcast receiver at medium wave and IF range is usually fixed at 455 kHz. FM
receivers using standard 88 – 108 MHz band have an IF of 10.7 MHz. The characteristics of an IF amplifier are
independent of the frequency to which the receiver is tuned. Therefore, the sensitivity (ability to receive low
signal strength) and selectivity of the receivers are usually fairly uniform throughout its tuning range.
The RF circuits are mainly used to select the wanted frequency, to reject any interfering signal and to reduce
the noise figure of the receiver. The most important of mixer are the bipolar transistor, FET and IC. All these
are generally self-excited so that the device could act as both oscillator and mixer. At UHF and above, crystal
diode (e.g., silicon) are used because of their low noise figures.
The Hartley oscillator can be used as local oscillator in receivers operating up to the limit of the short-wave
broadcasting frequencies that is 36 MHz. The Colpitts oscillator is better for VHF and above.

AM Demodulator (or Detector): The Diode Detection or The Envelope Detector

The diode detector is the commonest device used for demodulating AM signal. Consider the following circuit:

Figure 16: The Diode Detector circuit and waveform obtained.

RF amplitude-modulated signal is feed at the input, C is small capacitor and R is a large resistance. The Parallel
combination of R and C is the local impedance across which the rectified output voltage 𝐸𝑜 is developed at
the output.
At each positive peak of the RF voltage, C charges up to a voltage level almost equal to 𝐸𝑠 . Between peaks,
the capacitor discharges a little through R, but it is replenished at the next positive peak. Hence, 𝐸𝑜 is
produced as a near replica of the modulating voltage with a little amount of RF ripple.
The time constant 𝜏 of the RC combination must be slow enough to keep the RF ripple as small as possible
but also must be sufficiently fast for the detector circuit to follow the fastest modulation variation. Hence,
choosing the correct values of R and C is important for an AM detector circuit.

Image Frequency and Its Rejection

In a standard broadcast receiver (and in fact, in the vast majority of all receivers made), the local oscillator
frequency is made higher than the incoming signal frequency for reasons that will become apparent. It is
made equal at all times to the signal frequency plus the intermediate frequency.
Thus, 𝑓𝑜 = 𝑓𝑠 + 𝑓𝑖 or 𝑓𝑠 = 𝑓0 − 𝑓𝑖 , no matter what the signal frequency may be. When 𝑓𝑠 and 𝑓𝑜 are mixed,
the difference frequency, which is one of the by-products is equal to 𝑓𝑖 . As such, it is the only one passed

21
and amplified by the IF stage. If a frequency 𝑓𝑠𝑖 manages to reach the mixer, such that 𝑓𝑠𝑖 = 𝑓0 + 𝑓𝑖 , that is,
𝑓𝑠𝑖 = 𝑓𝑠 + 2𝑓𝑖 , then this frequency will also produce 𝑓𝑖 when mixed with 𝑓𝑜 .
Unfortunately, this spurious intermediate frequency signal will also be amplified by the I.F stage and will
therefore produce interference. This has the effect of two stations being received simultaneously and is
naturally undesirable. The term 𝑓𝑠𝑖 is called the image frequency and is defined as the signal frequency plus
twice the intermediate frequency.
𝑓𝑠𝑖 = 𝑓𝑠 + 2𝑓𝑖 , 58

The rejection of an image frequency by a single-tuned circuit that is, the ratio of the gain at the signal
frequency to the gain at the image frequency is given as
𝛼 = √1 + 𝜌2 𝑄 2 59

where
𝑓𝑠𝑖 𝑓
𝜌= 𝑓𝑠
− 𝑓𝑠 60
𝑠𝑖

𝑄 = loaded 𝑄 of tuned circuit.

Image rejection depends on the front-end selectivity of the receiver and must be achieved before the IF
stage. Once the spurious frequency enters the first IF amplifier, it becomes impossible to remove it from the
wanted signal.

Example: In a broadcast superheterodyne receiver having no RF amplifier, the loaded Q of the antenna
coupling circuit (at the input to the mixer) is 100. If the intermediate frequency (IF) is 455 kHz,
Calculate (a) the image frequency and its rejection ratio at 1000 kHz
(b) the image frequency and its rejection ratio at 25 MHz

Solution
(a) 𝑓𝑠𝑖 = 1000 + 2 × 455 = 1910 𝑘𝐻𝑧
1910 1000
𝜌= − = 1.910 − .524 = 1.386
1000 1910

𝛼 = √1 + 1002 × 1.3862 = 138.6

𝛼 [𝑑𝐵] = 20 log10 𝛼 = 20 log10 138.6 = 42 𝑑𝐵

This is 42 dB and is considered adequate for domestic receivers in the MF band.

(b) 𝑓𝑠𝑖 = 25 + 2 × 0.455 = 25.91𝑀𝐻𝑧

25.91 25
𝜌= − = 1.0364 − 0.9649 = 0.0715
25 25.91

𝛼 = √1 + 1002 × 0.07152 = 7.22

𝛼 [𝑑𝐵] = 20 log10 𝛼 = 20 log10 7.22 = 17.17 𝑑𝐵

22
FM Demodulation: The Slope Detector
The function of a FM demodulator is to change the frequency deviation of the incoming frequency modulated
signal into an audio-frequency amplitude variation (identical to the one that originally caused the frequency
variation). The conversion should be done linearly and efficiently.
In addition, the detector circuit should as much as possible, be insensitive to amplitude variation. The
resulting amplitude modulated wave is then applied to the detector arrangement to produce a voltage whose
amplitude depends on the frequency variation of the input voltage.
Consider an FM signal fed to a tuned circuit whose resonant frequency is to one side of the center frequency
of the FM signal as shown in the Figure 17.

Figure 17: FM detector waveform.

The output of this tuned circuit will have amplitude that depends on the frequency deviation of the input
signal, as shown in Figure 17. The frequency variation produces an output voltage proportional to the
frequency deviation of the carrier signal.
This output voltage is then applied to a diode detector with an RC load of suitable time constant, as in case
of AM detection.
The characteristic of the single slope detector described above does not give a linear conversion, except along
a very limited frequency range.
To overcome this, two slope detectors are used in a balance configuration as shown in Figure 18.

Figure 18: Slope Detector

23
The two slope detectors are connected back-to-back to the opposite ends of a centre tapped transformer.
One point is tuned above the IF by 𝑓𝑐 + 𝛿𝑓, and the other part is tuned below the IF 𝑓𝑐 − 𝛿𝑓, each tuned
circuit is connected to a diode detector with an RC load, the characteristics of the balanced detector is as
shown in Figure 19.

Figure 19: Characteristics response of the balanced slope detector.

Figure 19 indicates that the characteristic of the balanced slope detector is linear over a wide frequency
variation than for the single slope detector.

BINARY DIGITAL COMMUNICATIONS

Instead of using a sinusoidal signal as baseband to modulate a sinusoidal carrier signal, which we have so far
considered, we shall now consider modulation processes involving binary pulse signals as baseband.
This is important because digital communication is the current and future technology adopted in all forms of
communication systems and services and analogue of whatever form can be easily converted into digital
forms by sampling, quantizing and coding processes.
Therefore, digital modulation provides more information capacity, high data security, quicker system
reliability with great quality communication. As before, we shall consider the three types of modulations
namely; Amplitude, Frequency and Phase.

AMPLITUDE SHIFT KEYING (ASK)

The simplest form of bandpass data modulation is Amplitude Shift Keying (ASK). Here, the symbols are
represented as various discrete amplitudes of a fixed frequency carrier oscillator. This involves switching the
amplitude of the carrier signal ON or OFF in accordance with the binary baseband signal (i.e., between two
symbol state). It is referred to as ON-OFF keying (OOK). If more than two symbol states are used, then an M-
ary ASK process is adopted, an example being the 8-ASK format shown here. Consider a sequence of binary
pulses, and a sinusoidal carrier signal as shown Figure 20.

24
 Figure 20: Amplitude Shift Keying Modulation

The 1’s in the binary signal turn ON the carrier signal, and the ‘0’s turn it OFF
𝑒 = 𝐴𝑒𝑚 𝑐𝑜𝑠 𝑤𝑐 𝑡 61

Where 𝑒𝑚 = 1 𝑜𝑟 0. The process of multiplying 𝑒𝑚 by 𝑐𝑜𝑠 𝑤𝑐 𝑡has the effect of shifting the frequency
spectrum of 𝑒𝑚 by 𝑤𝑐 = 2𝜋𝑓𝑐 while preserving the shape of the spectrum of 𝑒𝑚 as shown.

Generation ASK Modulator

The ASK modulator comprises of the carrier signal generator, the binary sequence from the message signal
and the band-limited filter as shown in Figure 21.

Figure 21: Binary ASK Modulator

The carrier generator, sends a continuous high-frequency carrier. The binary sequence from the message
signal makes the unipolar input to be either High or Low. The high signal closes the switch, allowing a carrier
wave. Hence, the output will be the carrier signal at high input. When there is low input, the switch opens,
allowing no voltage to appear. Hence, the output will be low. The band-limiting filter, shapes the pulse
depending upon the amplitude and phase characteristics of the band-limiting filter or the pulse-shaping filter.
For more than two symbol states, this approach becomes quite complicated with the requirement to gate
on carriers with differing amplitudes to represent the required number of symbol states.

25
 Figure 22: M-ary ASK Modulator

ASK Demodulator
The ASK demodulator or detector are of two types namely;

Asynchronous ASK detector

The Asynchronous ASK detector consists of a half-wave rectifier, a low pass filter, and a comparator.
Following is the block diagram for the demodulator in Figure 22. The modulated ASK signal is given to the
half-wave rectifier, which delivers a positive half output. The low pass filter suppresses the higher frequencies
and gives an envelope detected output from which the comparator delivers a digital output. Here, clock
frequency at the transmitter does not matches that at the receiver.

Figure 23: Binary Asynchronous ASK Demodulator

Synchronous ASK detector

The ASK modulated input signal is passed to the square law detector whose output voltage is proportional
to the square of the amplitude modulated input voltage. The low pass filter minimizes the higher frequencies.
The comparator and voltage limiter help to get a clean digital output. Here, clock frequency at the transmitter
must matches that at the receiver.

Figure 24: Binary Synchronous ASK Demodulator

26
FREQUENCY SHIFT KEYING (FSK)
This is a digital modulation in which the frequency of the carrier signal varies according to the digital signal
changes. In FSK, the carrier is switched between two predetermined frequencies, either by modulating one-
sine-wave oscillator or by switching between two oscillators of difference frequencies, which are locked in
phase.
𝑒 = 𝐴 𝑐𝑜𝑠 𝑤1 𝑡
Consider 62
𝑒 = 𝐴 𝑐𝑜𝑠 𝑤2 𝑡

where 𝑓1 corresponds to ‘1’ and 𝑓2 corresponds to ‘0’

Figure 25: Frequency Shift Keying Modulation

Sometimes, it is arranged such that

𝑓1 = 𝑓𝑐 − 𝛥𝑓
63
𝑓2 = 𝑓𝑐 + 𝛥𝑓
i.e., the frequency deviates ±𝛥𝑓 from 𝑓𝑐
−𝑇 𝑇
or 𝑒 = 𝐴 𝑐𝑜𝑠( 𝑤𝑐 + 𝛥𝑤)𝑡, ≤𝑡≤ 64
2 2

𝛥𝑓 is commonly called the frequency deviation. The frequency spectrum of the FSK is somewhat complex.
The bandwidth is shown to be
𝐵𝑊(𝐹𝑆𝐾) = 2𝛥𝑓 + 2𝑓𝑚 65
where 𝑓𝑚 is the bandwidth of the baseband binary signal.
(i) If 𝛥𝑓 ≥ 𝑓𝑚 (wideband FM), the bandwidth approaches 2𝛥𝑓; i.e if one uses a wide separation of tones
in an FSK system, the bandwidth is essentially the frequency separation of the tones, independent of
the bandwidth of the baseband signal.
(ii) If 𝛥𝑓 ≤ 𝑓𝑚 (Narrowband FM), the bandwidth approach 2𝑓𝑚 , similar to the bandwidth of to the OOK
modulation.
𝛥𝑓
Equation (65) may be expressed in terms of the modulation index, (𝑚𝑓 = 𝑓 ) i.e.,
𝑚

𝐵𝑊𝐹𝑆𝐾 = 2𝑓𝑚 (1 + 𝑚𝑓 ) 66

Wideband FM corresponds to 𝑚𝑓 ≥ 1
Narrow band FM corresponds to 𝑚𝑓 ≤ 1

27
Generation FSK Modulator
The system consists of 2 oscillators that generate high and low-frequency signal separately. A binary message
signal is provided to the transmitter circuitry. The carrier wave from the two oscillators and binary modulating
signal operates the switch. In the case when the modulating signal bit is high i.e., 1, the switch gets closed
forming a path for a high-frequency wave to get transmitted that is generated by oscillator 1. Thus, a high-
frequency signal is achieved in case of bit 1 of the message signal. As against when the input bit is level low
i.e., 0, the switch now gets closed in a manner to form the path for the low-frequency carrier to get
transmitted. This low-frequency carrier is generated by the oscillator 2. Hence, it is clear that a low-frequency
signal is achieved in case of a low data bit of the modulating signal. However, in order to eliminate phase
discontinuities of the signal at the output, an internal clock is provided to the oscillator. Thus, the high or
low-frequency signal is selected according to the digital modulating signal. Thus, an FSK modulated wave is
transmitted and achieved at the output.

Figure 26: FSK Modulator

FSK Demodulator
The FSK demodulator or detector are of two types namely;

Asynchronous FSK detector

It consists of two separate Bandpass filter that is tuned to two different frequencies. The output of this BPF
is then fed to an envelope detector which generates two separate output according to the signal achieved
from the BPF. A decision-making device then compares the output of the two-envelope detector. When the
output of detector 1 exceeds detector 2 then logic 1 is obtained and when the output of detector 1 is decided
by detector 2 then logic 0 is obtained as the output. Thus, the original bitstream is generated by the detector.

Figure 26: Asynchronous FSK Modulator

28
Synchronous FSK detector
This demodulator is more complex than the asynchronous type. Synchronous detection of FSK is very similar
to that for ASK but in this case there are two detectors tuned to the two carrier frequencies. As for ASK,
synchronous detection and matched filtering minimize the effects of noise in the receiver. Recovery of the
carrier references in the synchronous receiver is made simple if the frequency spacing between the symbols
is made equal to the symbol rate (Sunde's FSK), as the modulated spectrum contains discrete spectral lines
at the carrier frequencies. The drawback of using Sunde's FSK is that the bandwidth of the FSK signal is
approximately 1.5 to 2 times that of an optimally filtered ASK or PSK binary signal.

Figure 27: Synchronous FSK Modulator

Advantages of FSK
1. FSK is a constant envelope modulation and hence insensitive to amplitude (gain) variations in the
channel and compatible with non-linear transmitter and receiver systems.
2. The detection of FSK can be based on relative frequency changes between symbol states and thus
does not require absolute frequency accuracy in the channel. FSK is thus relatively tolerant to local
oscillator drift and doppler shift.

Disadvantages of FSK
1. FSK is slightly less bandwidth efficient than ASK or PSK (excluding MSK implementation).
2. The bit/symbol error rate performance of FSK is worse than for PSK.

Phase Shift Keying (PSK)

In this case, we have
−𝑇 𝑇
𝑒 = ± 𝑐𝑜𝑠 𝑤𝑐 𝑡; 2
≤𝑡≤2 67
Here ‘1’ in the baseband binary stream corresponds to positive polarity of the signal, and ‘0’ corresponds to
negative polarity (or a change of polarity by 𝜋 radians or 180∘ ) as shown.

Figure 28: Phase Shift Keying Modulation

29
Whenever there is a change from 1 to 0 and from 0 to 1, there is a discontinuous phase transition in the
signal. The frequency spectrum is similar to the OOK modulation.

Synchronous PSK Modulator

Figure 29: PSK Modulator

Synchronous PSK Detector

There is no asynchronous equivalent detection process for PSK, and various forms of synchronous detection
must be employed. The ideal detector thus requires perfect knowledge of the unmodulated carrier phase at
the receiver. As with ASK, any phase error of the locally generated carrier reference reduces the signal
voltage at the output of the detector by a factor cos(𝜃). This in turn (see example) degrades
the Es/N0 performance of the detector by a factor 𝑐𝑜𝑠 2 (𝜃). Thus, the need for zero phase error for optimum
detection and must re-visit the whole area of carrier recovery.
Note that if the phase error reaches 90°, the output falls to zero.

Figure 30: PSK Demodulator

Analogue Pulse Modulation
This is another form of modulation where the carrier signal is a periodic pulse train but the modulating signal
is analogue. A number of parameters of the carrier pulse train may be changed in accordance with the
variations of the analogue baseband and the following types as shown in the Figure 29.
Pulse Amplitude Modulation (PAM)
Pulse Width or Pulse Duration Modulation (PWM or PDM) and
Pulse Position Modulation (PPM).

30
 Figure 29: Analogue Pulse Modulation

The modulated parameter of the pulse namely: Amplitude, Duration or Position varies in direct proportion
to the sample value of 𝑥(𝑡). This form of modulation is important in application such as Time-Division
Multiplexing; data telemetry and instrumentation systems.

Exercise
1 (a) A broadcast transmitter radiates 100 𝑘𝑊 of carrier power. What power will be radiated in a
SSBSC System at 65% modulation?
(b) A 1 𝑀𝐻𝑧 carrier signal is simultaneously modulated with 800 𝐻𝑧 and 2 𝑘𝐻𝑧 audio sine signals.
Determine the frequencies present in the output.

2. An AM wave is defined by the equation 𝐴(𝑡) = (15 + 3 𝑐𝑜𝑠(2𝜋 𝑥 103 ) 𝑡) 𝑐𝑜𝑠(2𝜋 𝑥 106 𝑡).
(a) Identify the amplitude and frequency of the unmodulated carrier signal and the modulating signal.
(b) Determine the modulation index, the power in the two sidebands and the power in the carrier.

3. (a) When the modulating frequency in an FM system is 400 𝐻𝑧 and the modulating voltage is 2.8 𝑉, the
modulation index is 70. Calculate the maximum deviation.
(b) What is the modulation index when the modulating frequency is reduced to 260 𝐻𝑧 and the
modulating voltage simultaneously raised to 3.2 𝑉?

31
 SOURCES AND EXAMPLES OF CHANNEL DEGRADATION
This section is aimed at injecting a bit of reality into the design process by considering the characteristics of
a range of typical channels.

Figure 30:

All communications links are ultimately limited by background noise in the system, and so it is essential to
have a working knowledge of the statistical properties of noise as they affect data communications
performance. Also, many channels are subject to interference, usually man-made, which can be equally
detrimental to communications integrity. Thirdly, no communications link is truly distortion free, whether
this be caused by imperfections in the processing hardware or defects within the channel, and the nature
and impact of distortion on a digital communications system must be understood if good design choices are
to be made. In order to 'root' the discussion of noise, interference and distortion in reality, this chapter
concludes with sections outlining the characteristics of two channel types in everyday use: the telephone
channel and the radio channel.

Doppler Shift
Whenever a signal source moves towards or away from a receiver, the frequency of the signal as observed
at the receiver increases or decreases respectively. This is known as the Doppler effect. The degree of
frequency shift is a linear function of the speed of motion and the carrier signal frequency. For example, a
source moving at 70 mph using a carrier frequency of 900 MHz will experience a Doppler shift of up to +/–
100 Hz at the receiver.

Figure 31: Effect of doppler shift

Correcting for Doppler shift can be very difficult, particularly in a multipath environment, where signals
arriving from different angles experience different Doppler shifts.

Interference
This arise owing to contamination of the channel by extraneous signals, for example, from power line,
machinery, ignition systems, other channel users and so on. If the characteristics are known, then
interference can often be suppressed by filtering or subtraction, for example car suppresser. Interference is

32
often impulse-like in nature and we know from our knowledge of the Fourier transform and Fourier series
expansion that an impulse contains energy over a very wide bandwidth. In the case of ignition noise, the
ignition system may be firing at only 4000 Hz (1000 rpm), yet significant high frequency energy will exist at
frequencies of several MHz.

Sources of interference
Most interference encountered in digital communications systems (except for deep space missions!) arises
from either other communications systems or machinery. For example, crosstalk in telecommunications
lines is classed as interference, as is the ignition noise generated by a car engine.

Figure 32: Crosstalk effect

In radio systems in particular, a major source of interference is from other users of the radio spectrum. For
example, equipment radiating on frequencies close to the wanted channel can pass through the receiver
selection filtering, causing what is termed adjacent channel interference. In cellular applications, mobile
users in different geographical locations are assigned the same frequency for their calls, and if they are not
separated by sufficient distance, co-channel interference occurs.

Figure 33: Co-channel interference

In both radio and television, multipath interference is common, manifest as ghosting on the television
screen, caused by signals travelling by many different paths between transmitter and receiver each with
slightly different time delay.

33
Dealing with interference
Because most interference (excluding noise) in communications systems is generated by other items of
equipment, it is often possible with good design to minimize the effects of interference. This can be achieved
both by judicious selection of modulation and coding format to be the least sensitive to a given type of
interference, and also by tackling the causes of the interference.

Crosstalk in telephone lines, for example, can be reduced by careful layout of cables, or by replacing cables
with optical fibre, which has no external radiation to cause crosstalk.
Ghosting caused by multipath can often be cured by using directive antennas to avoid picking up reflections.
Some modern cellular base-stations and even some mobile handsets are using adaptive antennas which in
real time change the direction of the antenna beam to null out interferers and focus on the wanted signal.

Co-channel and adjacent channel interference are again controllable by good system planning and good
selective filtering within the receiver modem.

Figure 34:
Noise
Noise is characterised as random, unpredictable electrical signals from natural sources for example, thermal
noise, shot noise, atmospheric noise etc.
Because of the multiplicity of noise sources in communication link, it is difficult to define the properties
(frequency range, level and instantaneous phase) of noise and hence equally difficult to reduce its effect on
the communications link performance. For convenience, most textbooks and indeed practising engineers
assume noise to fall predominantly into the class of Additive White Gaussian Noise (AWGN) which does
indeed form an adequate classification for most cases. However, we should not forget that this is a general
simplification of the whole noise issue.
Unlike interference, noise originates predominantly from within the communications link itself and is usually
totally random in nature, making it very difficult to deal with. There is a variety of mechanisms by which noise
is generated, the most commonly referenced forms being thermal noise, shot noise, flicker noise and
atmospheric noise.
Characteristics of noise
Noise is usually classified as white or coloured depending on the spectral density of the noise power with
frequency.

34
White Noise is defined as having a flat power spectral density over all frequencies of interest, with a value
usually denoted as N0 Watts/Hz.

Figure 35: White noise spectral

Coloured noise has a non-uniform spectral distribution; however, over a finite bandwidth corresponding
perhaps to a single communications channel, the power spectral density might appear flat, hence the
term bandlimited white Gaussian noise.

Figure 36: Coloured noise spectral

Not only is it necessary for the spectrum to be flat, but the statistics of the noise must be such that the
envelope distribution of the bandlimited noise must be Gaussian in nature to fully satisfy the Shannon
condition. Fortunately, this holds approximately true for the majority of practical narrowband
communications systems.

Sources of noise
Thermal noise often dominates in communications systems and originates from the free movement of
electrons within a conductor. The name arises because the energy and degree of movement of electrons
increases proportionally with the temperature of the conductor.

Figure 37: Thermal noise generation

The current and voltage generated by this movement has a waveform that is entirely random in nature and
which will, over time, have an average power spectrum that is flat over all frequencies. This property of
thermal noise to contain all frequencies has resulted in it being called white noise to mirror the property of
white light to contain all colours.

35
Thermal noise can clearly be reduced by cooling the noise source and this very principle is being applied in
some radio receivers using cryogenic coolers, to improve the receiver sensitivity.
The noise power of a thermal noise is directly proportional to the product of the temperature and bandwidth
of interest i.e.,
𝑷𝒏 ∝ 𝑻𝜹𝒇 = 𝒌𝑻𝜹𝒇 68

Where 𝑻 is absolute temperature

𝒌 = 𝟏. 𝟑𝟖 × 𝟏𝟎−𝟐𝟑 𝑱/𝑲 is Boltzmann’s constant
𝛿𝑓 is the bandwidth of interest

Figure 38:

𝑉2 𝑉𝑛
The noise voltage can be derived from 𝑃𝑛 = , where 𝑉 = , so
𝑅 2

𝑉𝒏 = √4𝑘𝑇𝛿𝑓𝑅 69

Shot noise: is generated within semiconductor junctions when electrons cross a potential barrier. Whereas
thermal noise power is proportional to temperature, shot noise power is proportional to the bias current in
the semiconductor. The nature of shot noise is also purely random and has a flat power spectrum with
frequency. The shot noise current is given as
𝒊𝒏 = √𝟐𝝆𝒊𝒑 𝜹𝒇 70
−𝟗
Where 𝝆 = 𝟏. 𝟔 × 𝟏𝟎 𝑪 is the charge of an electron
𝒊𝒑 is the direct diode current

Flicker noise: is also generated in semiconductors and is proportional to the dc bias current, but differs in
that the noise power decreases with frequency. Because this power variation is almost directly proportional
to 1/f, it is sometimes called 1/f noise.
𝒊𝒃𝒊𝒂𝒔 𝒊𝒃𝒊𝒂𝒔
𝑷𝒏 ∝ 𝒇
=𝒌 𝒇
71
Atmospheric noise: is a general term given to noise arising from electromagnetic radiation from solar and
galactic sources. Certain stars, for example, emit definite and regular amounts of noise which are best
avoided by pointing the antenna away from the noise source. The compound effect of this noise is usually
expressed as an equivalent sky noise temperature and is generally much less than thermal noise. The level
of noise varies considerably with frequency, with the higher levels of noise occurring in the microwave region
of the spectrum (300 MHz to 300 GHz).

36
 Figure 39:
Noise Calculation
Addition of noise due to several sources: Let’s assume there are 𝒎 several sources of thermal agitation
noise generated in series, the total noise is
𝑽𝒏𝑻 = 𝑽𝒏𝟏 + 𝑽𝒏𝟐 + ⋯ + 𝑽𝒏𝒎 72

𝑽𝒏𝑻 = √𝟒𝒌𝑻𝜹𝒇(𝑹𝟏 + 𝑹𝟐 + ⋯ + 𝑹𝒎 ) 73

Addition of noise due to several amplifiers in cascade: This occurs in receivers with a number of amplifying
stages in cascade.

Figure 40:

It is important to find the equivalent input noise voltage. In doing so, it is better to go one step further and
find an equivalent resistance from such output voltage.

𝑉𝑛3 = √4𝑘𝑇𝛿𝑓𝑅3 74

This is the rms voltage at the output due to the temperature.

′ 𝑉𝑛3 √4𝑘𝑇𝛿𝑓𝑅3
𝑉𝑛3 = 𝐴2
= 𝐴2
= √4𝑘𝑇𝛿𝑓𝑅3′ 75

If 𝑅3 is referred to the input of the second stage,

𝑅
𝑅3′ = 𝐴32 76
2
Therefore, the equivalent resistance is
𝑅
′
𝑅𝑒𝑞 = 𝑅2 + 𝑅3′ = 𝑅2 + 𝐴32 77
2
Referring 𝑅2 to the input,

37
 𝑅
′ 𝑅2 + 32
𝑅𝑒𝑞 𝐴2 𝑅2 𝑅3
𝑅2′ = = = + 78
𝐴22 𝐴21 𝐴21 𝐴21 𝐴22

The noise resistance at the input of the first stage 𝑅1 is

𝑅 𝑅
𝑅𝑒𝑞 = 𝑅1 + 𝑅2′ = 𝑅1 + 𝐴22 + 𝐴2 𝐴3 2 79
1 1 2

Noise in reactive circuits

Figure 41:

𝑉𝑛 𝑉𝑛
𝑖𝑛 = =𝑅 80
𝑍 𝑠 +𝑗(𝑋𝑙 −𝑋𝑐 )

𝑉𝑛
At resonance, 𝑖𝑛 = and the voltage across the capacitor is
𝑅𝑠
𝑉𝑛 𝑉𝑛
𝑉 = 𝑖𝑛 𝑋𝑐 = 𝑋 = 𝑄𝑅𝑠 = 𝑄𝑉𝑛 81
𝑅𝑠 𝑐 𝑅𝑠
Where 𝑋𝑐 = 𝑄𝑅𝑠 at resonance,
𝑄 is the magnification factor
𝑉 2 = 𝑄 2 𝑉𝑛2 = 𝑄 2 4𝑘𝑇𝛿𝑓𝑅𝑠 82

𝑉 2 = 4𝑘𝑇𝛿𝑓(𝑄 2 𝑅𝑠 ) = 4𝑘𝑇𝛿𝑓𝑅𝑝 83
Where 𝑅𝑝 = 𝑄 2 𝑅𝑠 is the equivalent parallel impedance of tuned circuit at resonance

𝑉 = √4𝑘𝑇𝛿𝑓𝑅𝑝 84

Noise Figure: This is a figure of merit used to evaluate the performance of an amplifier or radio receiver
with lower value indicating better performance. This is the ratio of signal power to noise power at the same
point.
𝑆 𝑉 2
𝑁
= (𝑉𝑠 ) 85
𝑛

It is also the ratio of the signal-to-noise power supplied to the input terminal of a receiver or amplifier to
the signal-to-noise power supplied to the output or load resistance.

𝑖𝑛𝑝𝑢𝑡 𝑆⁄𝑁
𝐹= 86
𝑜𝑢𝑡𝑝𝑢𝑡 𝑆⁄𝑁

For an ideal receiver, 𝐹 = 1. In practical receiver, the 𝑜𝑢𝑡𝑝𝑢𝑡 𝑆⁄𝑁 will be lower than the 𝑖𝑛𝑝𝑢𝑡 𝑆⁄𝑁, so
𝐹 > 1.
4𝑘𝑇𝑒𝑞 𝛿𝑓𝑅𝑒𝑞
𝐹 =1+ 87
4𝑘𝑇𝑛 𝛿𝑓𝑅𝑛

38
If 𝑅𝑒𝑞 = 𝑅𝑛 , then
𝑇𝑒𝑞
𝐹 =1+ 𝑇𝑛
88
𝑇𝑒𝑞 = 𝑇𝑛 (𝐹 − 1) 89

Continuation of Noise Figure

Figure 42:

i. How to calculate 𝑷𝒔𝒊 : use voltage divider rule to obtain 𝑉𝑠𝑖

𝑅𝑡
𝑉𝑠𝑖 = 𝑉 90
𝑅𝑎 +𝑅𝑡 𝑠

𝑉𝑠𝑖 2
Obtain the power at 𝑅𝑡 by substituting equation (90) into 𝑃𝑠𝑖 = 𝑅𝑡
gives

2
𝑅𝑡
𝑉𝑠𝑖 2 ( 𝑉) 𝑉 2𝑅
𝑅𝑎 +𝑅𝑡 𝑠
𝑃𝑠𝑖 = 𝑅𝑡
= 𝑅𝑡
= (𝑅 𝑠+𝑅𝑡)2 91
𝑎 𝑡

ii. How to calculate 𝑽𝒏𝒊 : use equation (69) and substitute 𝑅 = 𝑅𝑎 //𝑅𝑡

𝑅 𝑅
𝑉𝑛𝑖 2 = 4𝑘𝑇𝛿𝑓 (𝑅 𝑎+𝑅𝑡 ) 92
𝑎 𝑡
2
𝑉𝑛𝑖
Obtain the power at 𝑅𝑡 by substituting equation (92) into 𝑃𝑛𝑖 = gives
𝑅𝑡

𝑅 𝑅
𝑉𝑛𝑖 2 4𝑘𝑇𝛿𝑓( 𝑎 𝑡 )
𝑅𝑎 +𝑅𝑡
𝑃𝑛𝑖 = 𝑅𝑡
= 𝑅𝑡
93

4𝑘𝑇𝛿𝑓𝑅𝑎
𝑃𝑛𝑖 = 94
𝑅𝑎 +𝑅𝑡
𝑺
iii. How to calculate input signal-to-noise ratio ( ): find the ratio of equations (91) and (94)
𝑵𝒊

𝑆 𝑃 𝑉 2𝑅 4𝑘𝑇𝛿𝑓𝑅𝑎
𝑁𝑖
= 𝑃 𝑠𝑖 = (𝑅 𝑠+𝑅𝑡)2 ÷ 𝑅𝑎 +𝑅𝑡
95
𝑛𝑖 𝑎 𝑡

𝑆 𝑉 2 𝑅𝑡
= 4𝑘𝑇𝛿𝑓𝑅𝑠 96
𝑁𝑖 𝑎 (𝑅𝑎 +𝑅𝑡 )

39
 𝑺
iv. How to calculate output signal-to-noise ratio ( ): obtain the power at 𝑅𝐿 and substitute
𝑵𝒐
equation (90)
2
𝑅𝑡
𝑉𝑜 2 (𝐴𝑉𝑠𝑖 )2 (𝐴𝑉𝑠 )
𝑅𝑎+𝑅𝑡
𝑃𝑠𝑜 = = = 97
𝑅𝐿 𝑅𝐿 𝑅𝐿

𝐴2 𝑉𝑠2 𝑅𝑡2
𝑃𝑠𝑜 = 𝑅 2 98
𝐿 (𝑅𝑎 +𝑅𝑡 )

𝑆 𝑃𝑠𝑜 𝐴2 𝑉𝑠2 𝑅𝑡2

= = 99
𝑁𝑜 𝑃𝑛𝑜 𝑃𝑛𝑜 𝑅𝐿 (𝑅𝑎 +𝑅𝑡 )2

v. How to calculate noise figure: find the ratio of equations (96) and (99)
𝑆
𝑁𝑖 𝑉 2 𝑅𝑡 𝐴2 𝑉𝑠2 𝑅𝑡2
𝐹= 𝑆 = 4𝑘𝑇𝛿𝑓𝑅𝑠 ÷ 𝑃𝑛𝑜 𝑅𝐿 (𝑅𝑎 +𝑅𝑡 )2
100
𝑎 (𝑅𝑎 +𝑅𝑡 )
𝑁𝑜

𝐿 𝑎 𝑅 (𝑅 +𝑅 )
𝑡
𝐹 = 4𝑘𝑇𝛿𝑓𝐴2 𝑅 𝑅 𝑃𝑛𝑜 101
𝑎 𝑡

Obtain 𝑃𝑛𝑜 using

𝑉𝑜 2 (𝐴𝑉𝑛 )2 𝐴2 4𝑘𝑇𝛿𝑓𝑅
𝑃𝑛𝑜 = 𝑅𝐿
= 𝑅𝐿
= 𝑅𝐿
102

Substitute equation (102) into (101)

𝑅𝐿 (𝑅𝑎 +𝑅𝑡 ) 𝐴2 4𝑘𝑇𝛿𝑓𝑅

𝐹= × 103
4𝑘𝑇𝛿𝑓𝐴2 𝑅𝑎 𝑅𝑡 𝑅𝐿

𝑅(𝑅𝑎 +𝑅𝑡 )
𝐹= 𝑅𝑎 𝑅𝑡
104
′
it is convenient to define 𝑅𝑒𝑞 , which is a noise resistance that does not incorporate 𝑅𝑡 and which
is given by
′
𝑅𝑒𝑞 = 𝑅𝑒𝑞 − 𝑅𝑡 105
The total equivalent noise resistance for this receiver will now be
′ 𝑅 𝑅
𝑅 = 𝑅𝑒𝑞 + (𝑅 𝑎+𝑅𝑡 ) 106
𝑎 𝑡
Substitute equation (106) into (104) gives
′ 𝑅 𝑅𝑡 (𝑅 +𝑅 )
𝐹 = (𝑅𝑒𝑞 + (𝑅 𝑎 )
)( 𝑎 𝑡 ) 107
𝑎 +𝑅𝑡 𝑅𝑎 𝑅𝑡

(𝑅𝑎 +𝑅𝑡 ) ′
𝐹 =1+ 𝑅𝑎 𝑅𝑡
𝑅𝑒𝑞 108
𝑅𝑎 +𝑅𝑡
Under mismatch condition, 𝑅𝑡
→1
′
𝑅𝑒𝑞
𝐹 =1+ 𝑅𝑎
109

40
Noise Figure for Measurement
The preceding section showed how the noise figure may be computed if the equivalent noise resistance is
easy to calculate. When this is not practicable, as under transit-time conditions, it is possible to make
measurements that lead to the determination of the Noise Figure in Communication System. A simple
method, using the diode noise generator, is often employed.

Figure 43:

The noise voltage supplied to the input of the receiver by the diode will be given by
𝑅 𝑅𝑡
𝑉𝑛 = 𝑖𝑛 𝑍𝑛 = 𝑖𝑛 (𝑅 𝑎 110
𝑎 +𝑅𝑡 )

Substitute equation (70) into (110) gives

𝑅 𝑅
𝑉𝑛 = √2𝜌𝑖𝑝 𝛿𝑓 × (𝑅 𝑎+𝑅𝑡 ) 111
𝑎 𝑡

Obtain 𝑃𝑛𝑜
𝑉𝑛𝑜 2 (𝐴𝑉𝑛 )2 𝐴2 2𝜌𝑖𝑝 𝛿𝑓𝑅𝑎2 𝑅𝑡2
𝑃𝑛𝑜 = = =𝑅 2 112
𝑅𝐿 𝑅𝐿 𝐿 (𝑅𝑎 +𝑅𝑡 )

Obtain the noise figure using equation (101)

𝐿 𝑎 𝑅 (𝑅 +𝑅 )
𝑡
𝐹 = 4𝑘𝑇𝛿𝑓𝐴2 𝑅 𝑅 𝑃𝑛𝑜
𝑎 𝑡

𝑅 (𝑅 +𝑅 )
𝐿 𝑎 𝑡 𝐴2 2𝜌𝑖𝑝 𝛿𝑓𝑅𝑎2 𝑅𝑡2
𝐹 = 4𝑘𝑇𝛿𝑓𝐴2𝑅 𝑅 × 𝑅 2 113
𝑎 𝑡 𝐿 (𝑅𝑎 +𝑅𝑡 )

𝜌𝑖𝑝 𝑅𝑎 𝑅𝑡
𝐹= 114
2𝑘𝑇(𝑅𝑎 +𝑅𝑡 )

Under mismatch condition, 𝑅𝑡 >> 𝑅𝑎 , then

𝜌𝑖𝑝 𝑅𝑎
𝐹= 2𝑘𝑇
115

41
As a final simplification, we substitute into Equation (115) the values of the various constants it contains.
These include the standard temperature at which such measurements are made, which is 17°C or 290 K. This
gives a formula which is very often quoted:
1.6×10−19 𝑖 𝑅
𝐹 = 2×1.38×10−23𝑝×290
𝑎
= 20𝑖𝑝 𝑅𝑎 116

DIGITAL COMMUNICATION TECHNIQUES

Many signals used in communication systems are analogue in nature e.g., Speech, TV, fax telemetry signals.
Others are, however, digital e.g., teletype data, computer output, pulsed radar signal. For analogue signals
to be transmitted by digital techniques the signal must first be sampled, i.e., narrow samples are taken at
regular time interval as shown in Figure 44, where 𝑇 is the period of sampling and 𝜏 is the width of sample
pulse.
Analouge
signal
f (t )


t
T
Figure 44: Analogue sampling technique

The samples are usually further processed before transmission. Usually, sampling is done electronically by
gating the signal ON and OFF at the desired gate as shown in Figure 45.

f (t )
o/p samples

clock

Figure 45: Sampling circuit

If the samples are to faithfully represent the shape of the analogue signal, the rate of sampling should be as
high as possible. To determine how high the sampling rate should be, would require determining the highest
frequency components contained in the signal to be sampled. If 𝐵 𝐻𝑧 is the highest frequency of the sample.
Theoretically, analysis shows that the sampling rate, 𝑓𝑐 , should be greater than twice this highest frequency

42
 1
i.e., 𝑓𝑐 ≥ 2𝐵 in order to give faithfully reproduction of the signal or the period of the sample 𝑇 ≤ sec. This
2𝐵
1
is called the Nyquist Criterion. The minimum sampling rate of 2𝐵 samples per second or 2𝐵
is the Nyquist
Criterion interval.

Fourier Analysis of a Periodic Signal

The theory credited to Fourier states that if 𝑓(𝑡) is a periodic function of time with period 𝑇, it may be
expressed by the following trigonometric expression.
𝑎𝑜 2
𝑓(𝑡) = 𝑇
+ 𝑇 ∑∞
𝑛=1(𝑎𝑛 𝑐𝑜𝑠 𝑤𝑛 𝑡 + 𝑏𝑛 𝑠𝑖𝑛 𝑤𝑛 𝑡) 117

2𝜋𝑛
where 𝑤𝑛 = .
𝑇
To determine the coefficient 𝑎𝑛 , we multiply both sides of equation (117) by 𝑐𝑜𝑠 𝑤𝑛 𝑡 and integrate over the
period 𝑇. All the terms on the RHS are vanished except the terms containing 𝑎𝑛 . Since
𝑇
2
∫ 𝑐𝑜𝑠 𝑤𝑗 𝑡 𝑐𝑜𝑠 𝑤𝑛 𝑡𝑑𝑡 = 0,
−𝑇 𝑗≠𝑛 118
2

𝑇
2
∫−𝑇 𝑠𝑖𝑛 𝑤𝑗 𝑡 𝑐𝑜𝑠 𝑤𝑛 𝑡𝑑𝑡 = 0, 119
2

𝑇 𝑇
2𝑎𝑛 2 2𝑎𝑛 𝑇
2
∫−𝑇 𝑓(𝑡) 𝑐𝑜𝑠 𝑤𝑛 𝑡 𝑑𝑡 = ∫ 𝑐𝑜𝑠 2 𝑤𝑛 𝑡
𝑇 -T
𝑑𝑡 = .
𝑇 2
= 𝑎𝑛 120
2 2

1
Note that: 𝑐𝑜𝑠 2 𝑤𝑛 𝑡 = 2 (1 + 𝑐𝑜𝑠 2 𝑤𝑛 𝑡) 121

𝑇
2
𝑎𝑛 = ∫−𝑇 𝑓(𝑡) 𝑐𝑜𝑠 𝑤𝑛 𝑡𝑑𝑡, 𝑛 = 0, 1, 2, . . . . . 122
2

𝑇
2
Similarly, 𝑏𝑛 = ∫−𝑇 𝑓(𝑡) 𝑠𝑖𝑛 𝑤𝑛 𝑡𝑑𝑡, 𝑛 = 1, 2, . . . . . 123
2

Equation (117) may be written as

1
𝑎𝑜 2 2 2
𝑓(𝑡) = 𝑇
+ 𝑇 ∑∞ 2
𝑛=1[𝑎𝑛 + 𝑏𝑛 ] 𝑐𝑜𝑠[𝑤𝑛 𝑡 + 𝜃𝑛 ] 124

−𝑏
where 𝜃𝑛 = 𝑡𝑎𝑛−1 ( 𝑎 𝑛 ) 125
𝑛
A third form of expression for 𝑓(𝑡) is the complex exponential form
1
𝑓(𝑡) = 𝑇 ∑𝛼𝑁=0 𝐶𝑛 𝑒 𝑗𝑤𝑡 126

where 𝐶𝑛 is a complex number defined by

1
𝐶𝑛 = 𝑎𝑛 − 𝑗𝑏𝑛 = (𝑎𝑛 2 + 𝑏𝑛 2 )2 𝑒 𝑗𝜃𝑛 127

43
 𝑇
2
Therefore, 𝐶𝑛 = ∫−𝑇 𝑓(𝑡)𝑒 −𝑗𝑤𝑡 𝑑𝑡 128
2
1
−𝑏
|𝐶𝑛 | = (𝑎𝑛 2 + 𝑏𝑛 2 )2 is the amplitude spectrum and 𝜃𝑛 = 𝑡𝑎𝑛−1 ( 𝑛 ) is the phase characteristic.
𝑎 𝑛
Note that the relationship between equations (117), (126), and (128) is obtained by expressing
1
𝑐𝑜𝑠 𝑤𝑛 𝑡 = 2 (𝑒 𝑗𝑤𝑛 𝑡 + 𝑒 −𝑗𝑤𝑛 𝑡 ) 129
1
𝑠𝑖𝑛 𝑤𝑛 𝑡 = 2
(𝑒 𝑗𝑤𝑛 𝑡 − 𝑒 −𝑗𝑤𝑛 𝑡 ) 130

Example: Determine the Fourier analysis of the following periodic rectangular pulses.

f (t )
Am


t
−T −T 2 0 T 2 T

𝜏
Using 𝐶𝑛 = ∫ 𝑓(𝑡)𝑒 −𝑗𝑤𝑡 𝑑𝑡
2
−𝜏
2

𝜏
2
𝐶𝑛 = ∫−𝜏 𝐴𝑚 𝑒 −𝑗𝑤𝑡 𝑑𝑡
2

𝜏 𝜏
𝑗𝑤𝑛 −𝑗𝑤𝑛
𝑒 2 −𝑒 2
= 𝐴𝑚 [ 𝑗𝑤𝑛
]

2𝐴𝑚 𝜏
= 𝑠𝑖𝑛 (𝑤𝑛 )
𝑤𝑛 2

𝜏
𝑠𝑖𝑛 𝑤𝑛 𝑠𝑖𝑛 𝑥
2
= 𝜏𝐴𝑚 𝜏 = 𝜏 𝐴𝑚
𝑤𝑛 𝑥
2

𝑤𝑛 𝜏 𝑠𝑖𝑛 𝑥
where 𝑥 = 2
and 𝑥
= 𝑠𝑖𝑛 𝑐 𝑥 is a common function occurring in many signal analysis problems, a plot
of which is as shown in the figure below.
𝑠𝑖𝑛 𝑥
The 𝑥
, the envelope of the amplitude spectrum of the signal since n has discrete values n = 0, 1, 2, 3…….
2𝜋𝑛 2𝜋 −𝑤 𝜏 2𝜋 𝜏 𝑛𝜋𝜏
𝑤𝑛 takes on discrete values 𝑤𝑛 = 𝑇
, also 𝑤1 = 𝑇
and 𝑇𝑖 = 2𝑛 = 2𝑇𝑛 = 𝑇 .
2𝜋 2𝜋𝑛 2𝜋
The spacing between successive lines is 𝛥𝑤𝑛 = (𝑛 + 1) − = 𝑇 which is the fundamental angular
𝑇 𝑇
frequency of the periodic pulses. Also, lines of the frequency spectrum occur at multiples of these
fundamental frequencies

44
Note that: as 𝑇 decreases (or as there are more pulses per second of the periodic signal), the fuel lines more
out further. Conversely, as 𝑇 increases the frequency lines crowd together and ultimately approaches an
almost smooth spectrum.
As the pulse width, 𝑇 decreases, the frequency content of the signal extends out over a larger frequency
2𝜋
range. The first zero crossing at 𝑤𝑛 = 𝜏
moves out in frequency. There is, therefore, an inverse relation
between the pulse width (or pulse duration) and the frequency spread (or bandwidth) of the periodic wave,
that is, a very narrow pulse gives rise to a wide bandwidth and vice-versa.

Sampling frequency spectrum

The process of sampling can be considered mathematically by writing.
𝑓𝑠 (𝑡) = 𝑓(𝑡)𝑠(𝑡)
1
where 𝑠(𝑡) is a periodic series of pulses of unit amplitude, width, 𝜏, and period 𝑇 = 𝑓 as
𝑐

f s (t ) = f (t ) s (t )
f (t ) x 

s (t ) t
−T −T 2 0 T 2 T
Figure 46: Sampling frequency spectrum

applying the Fourier series relation to 𝑠(𝑡) gives the sampled signal 𝑓𝑠 (𝑡) to be of the form
𝜏 𝑠𝑖𝑛 𝑛𝜋𝜏
𝑓𝑠 (𝑡) = 𝑓(𝑡) [1 + 2 ∑𝛼𝑛=1 𝜏 𝑐𝑜𝑠 2𝜋𝑛𝑓𝑐 𝑡] 131
𝑇 𝑛𝜋
𝑇
The Fourier transform, 𝐹𝑠 (𝑤), of 𝑓𝑠 (𝑡) or the amplitude spectrum of the signal samples is by
𝑛𝜋𝜏
𝜏 𝜏 𝑠𝑖𝑛
𝐹𝑠 (𝑤) = 𝑇 𝐹(𝑤) + 𝑇 ∑𝛼−𝛼 𝑛𝜋𝜏
𝑇
𝐹(𝑤 − 𝑛𝑤𝑐 ) 132
𝑇
where 𝐹(𝑤) is the Fourier transform of 𝑓(𝑡) i.e.,
𝛼
𝐹(𝑤) = ∫−𝛼 𝑓(𝑡)𝑒 −𝑗𝑤𝑡 𝑑𝑡 133

45
Equation (113) implies that the frequency spectrum of the sampled signal consists of a sum of the frequency
spectrum of the original signal translated to and centered on the multiple of the sampling frequency as shown
below.
F ( )

f c  2B

0 B

Fs ( ) Low-pass filter
response

fc − B fc fc + B

Figure 47: Signal obtained for 𝒇𝒄 > 𝟐𝑩

𝑛𝜋𝜏
𝑠𝑖𝑛 𝑠𝑖𝑛 𝑥
𝑇
The amplitude of each successive component decreases as 𝑛𝜋𝜏 = 𝑥
𝑇
As long as the different spectra are separated, the signal 𝑓(𝑡) can be filtered out from 𝑓𝑠 (𝑡) using a low-pass
filter whose response is shown in broken lines in the Figure 47. The minimum sampling rate is obtained when
the spectra are just touching each other.

Fs ( )
f c = 2B

0 B = fc − B fc
Figure 48: Signal obtained for 𝒇𝒄 = 𝟐𝑩

This occurs when 𝐵 = 𝑓𝑐 − 𝐵 or 𝑓𝑐 = 2𝐵 (Nyquist Sampling Rate). If the spectra components are closer than
this, they will overlap and it will be impossible to filter out 𝐹(𝑤) without distortion.
Fs ( )
f c  2B

0 fc
Figure 49: Signal obtained for 𝒇𝒄 < 𝟐𝑩

From the three scenarios presented above, it is clear that the baseband signal can be recovered from the
samples without distortion when the Nyquist criterion is satisfied i.e., 𝑓𝑐 ≥ 2𝐵.

46
The phenomenon of overlapping spectra due to samples that are too widely spaced (i.e., sampling rate too
low) and the distortion that result from this is called ALIASING. The filter response required under the
condition 𝑓𝑐 = 2𝐵 is one with a very sharp out- off (rectangular response) which is difficult to achieve.
In practice, therefore, the sampling rate is usually made slightly larger than the Nyquist rate to ensure easy
separation of the frequency spectra and to simplify the problem of low pass filtering in order to retrieve𝑓(𝑡).
For example, speech transmitted by telephone line is generally filtered to 𝐵 = 3.4 𝑘𝐻𝑧, the Nyquist rate is
6.8 𝑘𝐻𝑧, but for digital transmission, the standard practice is to adopt a sampling rate of 8 𝑘𝐻𝑧 (or 8000
samples per seconds).

Coding
The purpose of source coding is to transform the information type in the source to a form best suited to the
transmission process. Often this involves converting an analogue signal such as voice or light intensity in an
image to a digital binary representation for transmission using a modem.

These days, the source coding process will usually implement an algorithm to realize bit or symbol content
compression in addition to the standard process of quantization and A/D conversion. Standard image
compression algorithms are the MPEG and JPEG formats for moving and static images respectively. At the
present time, the source coding for music and speech is not so well standardized, with many different formats
in use throughout the world. In most, but not all cases, a complementary decoder is implemented to restore
the signal to near or exactly its original format.

Waveform coding – A/D converters

The term waveform coding is applied to source encoding methods that seek to digitize the incoming analogue
waveform, without making use of any information about its frequency content or source parameters, and
thus achieve more efficient coding. This method is the most flexible source coding technique, able to
accommodate any type of input waveform regardless of the generating source, and is typified by
the analogue to digital converter.

47
The A/D conversion process, often more grandly called pulse code modulation in many textbooks, usually
involves regular sampling of the input signal level and then a conversion of this sampled value into a number
representing the level. The accuracy of the representation is governed primarily by the resolution of the A/D
converter – that is, how many data bits it uses to represent each measured value. Typical A/D converters use
eight bits for telephony voice digitization, and 16 or 18 bits for HiFi music digitization. Some professional
mixing desks use 24-bit converters! Most communications links will make use of the complementary digital
to analogue (D/A) converter in order to restore the analogue waveform samples from the received data
words.

Nyquist sampling
One of the key goals in waveform sampling is to digitize only the minimum number of samples necessary to
represent the waveform accurately and hence allow accurate reconstruction on reception. The minimum
rate at which an arbitrary waveform can be sampled without loss of information is in fact twice the
bandwidth of the input waveform.

This is known as the Nyquist sampling criterion. Sampling at less than twice the bandwidth of the input signal
(equivalent to twice the maximum modulation frequency for baseband signals) results in what is
termed aliasing.

The Nyquist sampling requirement can be derived intuitively from our knowledge of Fourier
Series summarized earlier. We can view the sampling process as the mixing of the input signal with a train of
very narrow data sampling pulses which will result in sum and difference components appearing at the mixer
output for each harmonic of the pulse waveform mixing with the signal waveform as shown. This is the
spectrum that would effectively appear at the output of a D/A converter. In order to reconstruct the input
waveform correctly, the D/A output needs to be filtered so that only the spectral components present within
the source signal remain.

Nyquist sampling theorem – aliasing

If the Nyquist sampling criterion is not met, the effect is for the sum and difference components associated
with each harmonic of the sampling signal to overlap with those of adjacent harmonics in the sampling

48
spectrum. Clearly it is now not possible to filter out the wanted from the unwanted signals and thus perfect
reconstruction of the original signal is not achieved.

It is immediately apparent from this analysis that sampling at twice the maximum input signal frequency is
thus the minimum sampling rate needed for A/D conversion. It is also evident that removal of the possible
alias components when sampling at this minimum rate requires a 'brick wall' filter for signal recovery after
D/A conversion. In practice, a sampling rate of at least three times the maximum sampling frequency is used
in order to reduce the specification of these 'anti-aliasing' filters.

Some modern A/D or D/A converters, called sigma-delta converters, use many-fold oversampling (x64 or
x128 is typical) with in-built decimation or interpolation filters.

Dynamic range
The ability of an A/D converter to cope with both large and small signals is an important factor in waveform
encoding, and the ratio of 𝑉𝑚𝑎𝑥 to 𝑉𝑚𝑖𝑛 over which a converter will operate is called the dynamic range. This
parameter depends heavily on the resolution of the A/D, that is, the number of bits available to represent
any given sample. The more bits in the converter, the more quantization levels the converter is using to
match to any given waveform sample.
It is not difficult to see that an n-bit converter can differentiate between 2𝑛 = 𝑀 discrete signal levels and
𝑉
that the minimum signal variation that it can detect and represent is 𝑚𝑎𝑥⁄𝑀 Volts. This is often called
the quantization step size.

The dynamic range for an A/D converter is thus given as

49
 𝑉𝑚𝑎𝑥 𝑉𝑚𝑎𝑥
𝐷𝑦𝑛𝑎𝑚𝑖𝑐 𝑟𝑎𝑛𝑔𝑒 = = = 𝑀 𝑜𝑟 2𝑛
𝑉𝑚𝑖𝑛 𝑉𝑚𝑎𝑥⁄
𝑀

Expressed in 𝑑𝐵, this gives us the well known formular; the dynamic range of a linear A/D is 𝑛 × 6.02 𝑑𝐵.
An 8-bit converter thus has a dynamic range of approximately 48 dB.

Quantization noise
Another very important parameter in any source encoding scheme is the level of noise and distortion
introduced by the coding process. For waveform encoding, the main noise source is quantization error, that
is, the amplitude errors which the A/D and D/A conversion process introduces into the signal by not having
infinite precision. The level of quantization noise is dependent on how close any particular sample is to one
of the M levels in the converter. For a speech input, this quantization error will be manifest as a noise-like
disturbance at the output of a D/A converter.

The sampled analogue signal is further digitized and encoded before transmission. Binary coding is the most
frequently used form of coding. The first step is the quantization of the sampled signal, where by each sample
amplitude is broken up into a prescribed number of discrete amplitude levels as shown.

The signal to quantization noise ratio for an A/D converter is thus clearly a function of the number of bits
used and is given by

3𝑀2
𝑃𝑒𝑎𝑘 𝑆𝑁𝑅 =
2

An 8-bit converter thus has a signal to quantization noise ratio of approximately 50 dB for a full-scale input
signal.

Companding
A method of reducing the number of bits required in a converter while achieving an equivalent dynamic range
or signal to quantization noise ratio is to use a technique known as companding. The term companding comes
from a combination of the words COMPressing and expANDING, which adequately describe the process
involved.

50
Essentially, in order to improve the resolution of weak signals within a converter, and hence enhance the
signal to quantization noise ratio, the weak signals need to be enlarged, or the quantization step size
decreased, but only for the weak signals. Strong signals, on the other hand, can potentially be reduced
without significantly degrading the signal to quantization noise ratio, or alternatively the quantization step
size is increased. This compression process must be matched with an equivalent expansion process in the
D/A converter if the waveform integrity is to be maintained.

Because this technique is so effective at reducing the number of A/D and D/A bits needed to provide
adequate signal to quantization noise ratio for speech signals in particular, international standards have been
set defining the compression and expansion ratios to be used for telephone interconnect throughout the
world. There are in fact two standards, one predominantly a US standard called 𝜇-law companding, and the
other a European and ITU (International Telecommunications Union) called A-law companding.

5
4

3
2

0 1 2 3 4 5 6 7 8 9 10 11 t
(3) (4) (4) (6) (5) (5) (5) (5) (5) (4) Sampling instants

0011 0100 0100 0110 0101 0101 0101 0101 0101 0100

4 digit binary representation

Figure 50: Quantization of sampled signal

Figure 50 shows an analogue signal, sampled every second with each sample quantized such that the total
amplitude swing (6 V) is divided into equally spaced amplitude levels of 1 volt apart. The nearest discrete
level to the actual sample amplitude is taken as the value of the sample of the sampling instant. The
quantization process obviously leads to some error in the nearest discrete level to represent it.

51
The overall effect is as if additional noise had been introduced into the system. It is in fact referred to as
quantization noise. The noise is reduced as the quantization interval is made smaller. In practice, 8 to 16
quantization levels are found to be good enough for sound to be intelligible, but in practice, 128 or 256 levels
are used as standard in digital communication system.

Pulse Code Modulation (PCM)

If the quantized signal is encoded into binary pulses of equal amplitude, we have the PCM system. Figure 51
illustrates the 4-digit binary pulses resulting from the quantized signal. The number of binary digits used
depends on the number of quantization levels used. An eight level PCM (8 = 23 ) requires 3 binary digits, 16
level PCM requires 4 digits (16 = 24 ) and a 128-level system requires 7 digits (128 = 27 ) and so on. In
general, M levels PCM requires n digits, the bandwidth of the PCM signal is increased by the factor.
The information carried by binary signals is represented by either 1 or 0 (or sometimes 1 or -1). All that
the receiver needs to recognize is the presence (or absence) of a pulse (or the polarity of a pulse) to be able
to decode the signal. The shape of the pulse is not important for the correct interpretation of the message,
provided the pulse is large enough to be detected. Hence, the undesirable effect of noise introduced into an
analogue signal during transmission is much reduced in digital systems.
Consider the 6-digit code word shown 010101

(a) message sent

(b) message received

(corrupted pulses)

1 0 1 0 1

Figure 51: 4-digit binary pulses

Figure 51b is the corrupted received version of the rectangular pulses transmitted. Though corrupted, the
signal can be correctly detected through a regenerating circuit, for example, a Schmitt trigger circuit
provided the amplitude of the corrupted pulse is above the threshold of the regenerator. This signal
reshaping is usually done at repeater stations in a digital transmission.

Delta Modulation
The sampling rate of a signal should be higher than the Nyquist rate, to achieve better sampling. If this
sampling interval in Differential PCM is reduced considerably, the sample-to-sample amplitude difference is
very small, as if the difference is 1-bit quantization, then the step-size will be very small i.e., ∆ delta.
The type of modulation, where the sampling rate is much higher and in which the step-size after quantization
is of a smaller value ∆, is termed as delta modulation.

Features of Delta Modulation

i. An over-sampled input is taken to make full use of the signal correlation.

52
 ii. The quantization design is simple.
iii. The input sequence is much higher than the Nyquist rate.
iv. The quality is moderate.
v. The design of the modulator and the demodulator is simple.
vi. The stair-case approximation of output waveform.
vii. The step-size is very small, i.e., ∆ delta.
viii. The bit rate can be decided by the user.
ix. This involves simpler implementation.

Advantages of DM Over DPCM

i. Transmits 1-bit for a sample. The sampling rate and transmission bandwidth is quite small.
ii. Very easy to design the modulator and the demodulator.
However, there exists some noise in DM.
i. Slope overload distortion i.e., start-up (when ∆ is small with high dynamic range)
ii. Granular noise i.e., hunting (when ∆ is large)

Figure 52: Delta modulation

Slope Overload Distortion (Start-up error)

This distortion arises because of the large dynamic range of the input signal. As can be seen from the figure
above, the rate of input signal 𝑥(𝑡) is so high that the staircase signal 𝑢(𝑡) cannot approximate it, the step
size 𝛿 becomes too small for staircase signal to follow the steep segment of 𝑥(𝑡). Thus, there is a large error
between the staircase approximated signal and the original signal 𝑥(𝑡). This error is called slope overload
distortion. To reduce this error, the step size should be increased when the slope of the signal of 𝑥(𝑡) is high.
Since the step size of delta modulation remains fixed, its maximum or minimum slopes occur along straight
lines. Therefore, the modulator is called linear delta modulation.

Granular Noise (Hunting)

This noise occurs when the step size is too large compared to small variation in the input signal. This is for
small variation in the input signal, the staircase signal is changed by large amount of 𝛿 because of large step
size. From the figure, when the input signal is almost flat, the staircase 𝑢(𝑡) keeps on oscillating by ±𝛿 around
the signal. The error between the input and approximated signal is called granular noise. The solution to this
problem is to make step size small.
Thus, large step size is required to accommodate wide dynamic range of the input signal (to reduce slope
overload distortion) and small steps are required to reduce granular noise. Adaptive delta modulation is the
modification to overcome these errors.

53
 MULTIPLEXING AND DEMULTIPLEXING
Multiplexing is the process of transmitting a number of separate signals (or channels) simultaneously over
the same communication medium. Multiplexing is a technique of sending different signals over a common
channel. The device which performs this task is known as a multiplexer. In multiplexing, there is a conversion
from parallel (many inputs) to serial (single output). So, we can say that multiplexing works on the principle
of many-to-one. Multiplexing can be done in time, frequency, code, or wavelength domain.

Demultiplexing separates original signals from a common channel. The device which performs this task is a
de-multiplexer. There is a conversion from serial (single input) to parallel (many outputs). The principle
followed by demultiplexing is one-to-many. Just like multiplexing, demultiplexing also can be done in time,
frequency, code, or wavelength domain.

Frequency division multiplexing (FDM)

This is the most common multiplexing technique used. In FDM, user signals are modulated using different
carrier frequencies. The carrier frequencies are separated by a band of frequencies known as guard bands to
prevent them from interfering. The bandwidth of the frequency channel is divided among all users and guard
bands. FDM is possible only if the combined bandwidth of transmitting signals is less than the bandwidth of
the frequency channel. In FDM every user won't get the full bandwidth but only a part of the bandwidth for
the entire duration. The higher the number of users, the less bandwidth is allotted. If a user is not transmitting
data, then the frequency band allotted to that user goes unutilized.

One example of FDM is a television in which multiple channels are received over different frequencies
without overlap and through the common wireless media.

Figure 53: Block diagram of Frequency division multiplexing

Time division multiplexing (TDM)

Time division multiplexing is a technique in which multiple user signals are placed on a single output media
by dividing them into fixed-length time slots. Every user will transmit only when its slot appears. These slots
are in micro or nanoseconds and are chosen with the help of a switch that performs switching action in micro-
or nanoseconds. Unlike FDM, in TDM every user will get the full bandwidth but only for the time slot allotted
to that user. In TDM the time slot may go unutilized if a user is not transmitting.

54
 Figure 54: Block diagram of Time division multiplexing

Wavelength division multiplexing (WDM) - WDM works on the same principle as FDM. The only differences
are that the media is an optical fibre cable and the frequencies to be transmitted are very high. WDM
combines many light beams into a single light beam and uses an optical fibre cable to transmit. WDM is
becoming more popular because of the very high bandwidth available in fibre optics. WDM works on the
principle that when any light beam falls on the prism it bends the beam based on the angle of incidence and
its frequency. Just like in FDM, the WDM multiplexer also combines various lights into a single light and the
demultiplexer then separates these light signals using again a prism or grating.

Figure 55: Block diagram of Wavelength division multiplexing

Code division multiplexing (CDM)

CDM is a multiplexing technique used in the domain of wireless communications. It uses the spread spectrum
technique. The user data is encoded in a digital format such that its frequency spectrum spreads over the
entire available frequency spectrum. One code is used to spread the data of one user and the same is used
to de-spread it later. This scheme is more secure because of the coding and spreading of the spectrum
involved which makes it more immune to jamming and interference. The user data is demultiplexed by the
receiver as it is aware of the spreading code.

Each user is allotted the full bandwidth and time. The user data is multiplied by the code to spread the signal
and then transmitted. There is less waste of bandwidth and time slots in this technique. If a user's code is
multiplied by another user's code, the received value will be 0.

55
 Figure 55: Block diagram of code division multiplexing

Figure 56: Structure of frequency, time and code division multiplexing

Difference between FDM and TDM

Tutorial Problem
The bandwidth allowed for voice transmission over a telephone line is 3kHz. A number of telephone
conversations are to be multiplexed over a single communication channel by TDM using PCM. It is found in
a particular system that for reliable identification of bits at the receiving end, an interval of at least 1μsec
must be allowed for each bit. If the signals are to be quantised into 16levels, estimate the numbers of voice
signals to the nearest multiple of five, that can be multiplexed.
Solution

56
 Nyquist sampling rate, fc = 6 kHz or Nyquist sampling interval, T = 1/fc ≤ 1/6 x 10-3sec. This is the
period between any two samples of a telephone signal.
Suppose there are n signals to be multiplexed, each quantised to 16 levels(i.e. 24 levels), then each
signal is represented by 4 bits.
There are therefore 4n bits between two samples of the same signal. If the interval between bits is
1μsec, then the sampling interval, T = (4n x 1)10-6sec.
4n x 10-6 ≤ 1/6 x 10-3
or n ≤ 1/24 x 103signals = 41.67signals
The number of signals to be multiplexed, to the nearest multiple of fine is n = 40 signals.

57