15.

8
Communication Systems
The angle ,, between the two vectors s, and s, is given as

cos ;k (12)
The two vectors s, and s, are orthogonal or perpendicular to each other ifthei.
inner product s s =0 and , = 90°.

15.3 BINARY FREQUENCY SHIFT KEYING (BFSK)

15.3.1 INTRODUCTION
FSK is a form of constant-amplitude low-performance type of digital
modulation similar to the standard Frequency Modulation (FM). It is also
called as Binary Frequency Slhift Keying (BFSK).
a Definition:
The frequency of asinusoidal carrier is shifted between two discrete values
according to the input binary signal (0s or Is).
15.3.2 GENERATION OF COHERENT BINARY FSK
Here the incoming binary data sequence is first applied to an on-off level
encoder. At the encoder. output, symbol 1 is represented by constant
amplitude of + / E, volts, where Eb is the transmitted signal energy per bit
and symbol '0" is represented by - E, volts.
When we have binary bit 1 at the input, the oscillator with
frequency f in the
upper channel is Switched on, while the oscillator with frequency f in
the
lower channel is switched off, (inverter converts binary 1into
binary 0) with
the result that frequency f is transmitted.

fH S+ Af for symbol '1 (la)

S=f-Af for symbol 0° ...
(1b)
where, Afis the frequency deviation.
 Modulation Scheme l 15.9

PH(t)
Binary FSK
Binarydata signal
sequencebt) On-off level d(t) VBrSK(t)
encoder 1(t) = Cos (2r fH t)

,PL(t)
Inverter

COs (2T f t)

Fig 15.4 FSK Generation

When we have binary bit 0 at the input, the oscillator with frequency fH. in the
upper channel is switched off and the oscillator with frequency f in the lower
channel is switched on (inverter converts binary 0into binary 1), with the result
that frequency f is transmitted.
carrier
4 We assume that the two oscillators are synchronized and generate two
signal o, (t) and ¢, (t) ,which are orthogonal to each other.
Data bit to be value of Value of
Value of d() PH() PL()
transmitted b(t)
+1V +1V OV
Binary 1
-1V 0V +1V
Binary 0
Fig 15.5 Conversion table for BFSK representation
* The general expression for BFSK is given as
s(1)=VEsk(t) = V cos [2 G +d) Af]; 0stsT, . (2)
where, V, - Analog carrier amplitude (Volts).
S.- Analog carrier frequency (Hertz).
and
Af - Frequency shift in the analog carrier frequency (Hz),
(Volts).
d()) - Binary input signal
|15.10
Communicati
From equation (2), it can be seen that the frequency shift carrier
in
on Systems
(4) is proportionalto the amplitude of binary input signald(). frequency
() For a binary 1 input, d(t)= +1, then equation (2) becomes
VESK () = Vcos [2n G. + Af) ]
*..(3)
(ii) For a binary Oinput, d() =-1, then the equation (2) becomes,
VESK () = V, cos(2r fe- Af) ]
"..(4)
15.3.3 FSK WAVEFORMS

In BFSK the center frequency G) is shifted up and down in the

frequency
domain based on the binary input signal. When binary input signal changes
from a logic 0 (binary 0) to a logic 1(binary 1) and Vice versa, the outout
frequency shifts between two frequencies,
() Binary 1- Mark frequency Gm), and
(i) Binary 0 - Space frequency G).
Binary input Frequency output
0 Space f,)
1 Mark m)
Fig 15.6 FSK tine domain - truth table

-Af + Af

fm
Logic 0 Logic 1

Fig 15.7 FSK frequency domain

Digia!Modulation Scheme l
|15.11

Tb
1
Binary 1 0
input

Ts
FSK

Output
fm fm f Tm fm ts fs

fm- mark frequency

fs - space frequency

Fig 15.8 FSK Waveforms

A Frequency Deviation (Af)
In FSK, frequency deviation is defined as he difference between either the
mark or space frequency and center frequency or half the difference beween
mark and space frequencies" and expressed mathematically as

Af = 2
... (5)

where, Af - Frequency deviation (Hz).

Lf-fI - Absolute difference between the mark and space
frequencies.
15.3.4 BAUD RATE AND BANDWIDTH
i) Baud Rate
equals the time of
3 Here a symbol contains only one bit. Therefore, bit time (Tb)
baud.
an FSK symbol time (T.), i.e., bit rate equals the
J (6)
Baud rate N
becomes
Number of bit per symbol in FSK (N=1), then equation (6)
Baud rate of FSK =f ... (7)
where,Sh - Input bit rate (bps)
 15.12|
(iü) BandWidth (BW) Communication Systems
o Theminimum bandwidth for FSK is given as
B = |G,-f)-m-f)l
(8)
=|G-f) |+2f,
From (5), 2Af = |G-f)| -(9)
o By substituting equation (9) in equation (8), we get
= 24f +2f
B = 2(Af +f) .. (10)
15.3.5 FSK NON-COHERENT RECEIVER

Analog mark
Rectified
frequency signal
dc
BPF Envelope
FSK detector Comparator
input Data
Power output
Splitter

BPF Envelope
detector
dc
Analog space
frequency
Fig 15.9 Non-coherent FSK demodulator
This type of FSK detection is referred to as non-coherent detection because
there n0 frequency is. involvd in the demodulation process that 5
synchronized either in phase, frequency, or both with the incoming FSK signal.
The FSK input signal is simultaneously applied to the inputs of both Band Pass
Filters(BPFs) through the power splitter.
The upper BPF passes only the mark frequency on to its respective envelope
detector. Similarly the lower BPF passes only the spac frequency on to its
respective envelope detector.
DigitalModulation. Scheme l 15.13|

The envelope detectors indicate the total power in each pass band, and the
comparator responds to the largest of two powers as an output.
15.3.6 COHERENT FSK RECEIVER

X dt

Choose 1 ify>0
x()
Decision
device
Choose 0 ify<0

dt
Threshold = 0

Fig 15.10 Coherent FSK demodulator

* The incoming noisy FSK signal is multiplied by a locally generated carrier
Signal that has the exact samne frequency and phase as the transmitter carrier. So
it is called coherent (or) synchronous FSK detection.
to
The received noisy FSK signal is applied to the coherent FSK receiver,
detect the original input binary sequence.
which are
The receiver consists of two correlators with common input,
supplied with locally generated coherent carier signals 1(t) and ¢2(t).
the resulting
The correlator outputs are then subtracted, one from the other, and
of the
difference Y is compared with a threshold of zero volts. The output
decision device will be the digital information signal.
the receiver decides
() Ifhe correlator output is exceeded the threshold,
in favour ofsymbol "I"
15.14|
(1) f he correlator output is not exceeded the Communi cati
threshold, the
on Systems
decides in favour of symbol "0" receiver
Y<0 a 0
(iii) If Yis exactly zero, the receiver makes a random
Or 1.
guess in favour of 0
Y=0 s 0(or) I
15.3.7 GEOMETRICAL (OR) SIGNAL SPACE REPRESENTATION OF
BFSK
For geometrical
ORTHOGONAL
representation of BFSK, two carriers () and ,() of tuwe
different frequencies fu andf are used for modulation.
To make d,(t) and ,() orthogonal, the
frequencies and f should be some
fH
integer multiple of base band frequency f':
"..(1)
1
...(2)
Here, f,-T. then the carriers will be

cos (2T mf, t) ... (3a)

2
cos (2 nf, )
=

(3b)
The carriers o, (t) and ¢,(1) are orthogonal over the
period T;. Then, we can
write expressions for BFSK as

COS (2rnf t) ...(4a)

and s() =yPs T, 2

T cos (2f ) (46)

Here
2 and f-f. 2
f ’ Carrier frequency
Af ’ Frequency deviation. This is increase (or) decrease frequency fromJe*
 Modulation Scheme l
Digia,
15.15

Using the equations (3a) and (3b), we can write the above equations as,
... (5a)
SH) = P; T, ,()
SL() = Ps T, -(1) ... (5b)

Based on the above two equations we can draw the signal space diagram as
shown in Fig.15.11.

d2(t)

SH(t)

d=2P;Tb=2E,
1(t)
SL(t)

Decision
boundary

Fig. 15.I Signal space diagram of BESK

(1) Distance between Signal Points:
the distance between these
There are two signal points in the signal space and
two points can be obtained as,
d² = (vP, T, ) + VPsT, ' - 2 P, T%
d =V2 P, T, ...(6)

Energy per bit, E, = Ps T,

|d = 2E, ".. (7)
 15.16|
15.3.8 POWER SPECTRAL DENSITY (PSD) OF BFSK SIGNALS Communication Systemy
We may express the binary FSK signal as follows.

2Eb
s(1) = cos 27t f, t t T,
0stsTb
Using trigonometric identity, we get
2E, 2 Ep
s(1) = cos cos (2nf.1)-1 sin
T sin (2rf.i)
2Eb
T;
cOS
cos(27 f, t)+ 2Eb sin
sin(2rÍ,!) .. 2)
In above equation the plus sign coesponds to transmitting symbol 0 and
minus signcorresponds to transmitiny symbol 1.
() The in-phase component is completely independent of the input binary
wave. It equals 2bb
T,
cos ( t/T)) for all values of time t. The power

spectral density of this component consists of two delta functions,

1
weighted by the factor and occurring at =*,T.
2Tb
(ii) The quadrature component is directly related to the input binary wave,
when we have symbol '1' itequals - g() and for 0° its equals +g(1).
The symbol shapingfunction g(t) isdefined by
2 E
g) = Vin)0s/sT,
T, ...(3)

0 elsewhere

The energy spectral density of this symbol shaping function equals

8 E, T,cos? (r T,f ... (4)

r² (4 T f2-1²
 | Modulation, Scheme 1
Digital
15.17

The in-phase and quadrature components of the binary FSK wave are
independent of each other. The baseband power spectral density of the binary
FSK Wave equals the sum of the power spectral densities of these wo
components as,

1 1 8 E, cos² (r T,f) ...(5)

Sp) = 2 T, +
(4T f2- 1
We can find the power spectrum of the binary FSK signal contains two discrete
frequency Components located at,
1 ... (6a)

1
(6b)

ERROR RATE
15.3.9 PROBABILITY OF ERROR (OR)ERROR PROBABILITY (OR) BIT
(BER)OF BFSK
defined as,
4 The observation vector 'x' has two elements x, and x, and are

x() ¢,) dt ... (1a)

0

x() ¢,0) dt (1b)

transmitted symbol.
where, x(t) is the received signal, which depends upon
is equal to the
The Gaussian random variable L whose sample value y
difference between xË and x; that is
... (2)
the random variable L
Ifsymbol 1 was transmitted then conditional mnean of
is given by
E[L|1] = E[X,|1]-E(X,|1]
... (3)
15.18
Communication Systems
For symbol '0 the conditional mean of the random variable L is
E[L|0] = E[X, |0]-E [X, |0]
=
-y E *.. (4)
The variance of the random variable L is independent of which binary symbol
-was transmitted.

Var (L] = Var [X]+ Var [X,]

No +No - N
2 2

If symbol 0 was transmitted but receiver making a decision in favour of

symbol 1, then the conditional probability of error is

P,(0) = P(Q >0|symbol Owas sent) = Q |0) dl

0

1 (+yE,
exp
2No

= Conditional Probability Density Function

1 (l+yE, )
P,(0) = exp| 2 No
dl ...
(5)
V 2r No
By changing the variable of integration from to z, in equation (5) put
l+ E = Z, then we get
V2N
P(0) = exp (-z) dz = E; .. (6a)
2 No

Similarly, for symbol '1 was transmitted. Then the probability of error aS,
1
P. (1) = 2 erfc E ... (6b)
2 No
Digial.ModulationScheme 1 15.19|

4 By
averaging P.(0) and P(1), we get the average probability of bit error (or)
BitError Rate (BER) for coherent BFSK with equiprobable symbols as,

Pe =

, erfe 2 No
s 3.10 ADVANTAGES, DISADVANTAGES AND APPLICATION
Advantages of FSK
Easy to implement,
(i) It has better noise immunity than ASK. So probability of error free
reception of data is high.
Disadvantages of FSK
) High bandwidth requirement.
(ii) It is used only for low bit rate (1200 bps) applications.
(i) It cannot be used in high speed modems, because with increase in speed
the bit rate increases.
(iv) It has a poorer error performance than PSK or QAM.
A,Application
It is used in low-performance, low-cast, asynchronous data mnodems that are
used for data communications over analog voice-bandtelephone lines.

15.4 SOLVED PROBLEM

Determine:
(a) Peakfrequency,
(b) Minimum bandwidth,
() Baud rate for a BFSK signal with a mark frequency of 49 kH, a space
Jrequency of 51 kH, and an input bit rate of 3kbps.
Given: f, = 49 kHz; f-51 kHz; , = 3 kbps
9 Solution:
|49 x 103-51 x 103 |
() Peak frequency: Af = 2 2 =1kHz