Wireless Communication and Networks Upena Dalal

Chapter 7

Problem 7.1 What is the bandwidth efficiency for FSK, ASK, BPSK, and QPSK for a BER of 10−5 on
a channel with an SNR of 15 dB?

Solution: Bandwidth efficiency = R/W bps/Hz where R= data rate in bps and W=bandwidth in Hz.

Using given SNR value maximum bandwidth efficiency can be derived using Shanon’s equation and
decided by the channel capacity. For 15 dB SNR it comes out to be 5 bps/Hz

For solving this example for the given BER, experimental results are necessary as shown below

Now Eb/No = 15 dB – (R/W)dB

For FSK and ASK for given BER Eb/No = 12.5 dB

 (R/W)dB = 2.5 dB =1.778 bps/Hz

For BPSK, for given BER Eb/No = 9.5 dB

 (R/W)dB = 5.5 dB = 3.548 bps/Hz

For QPSK baud rate is R/2 => R/W =7.096 bps/Hz which is greater than max capacity which means
more Eb/No is required compared to BPSK for faithful transmission.

Note: If the BER is to be less than 10-5 i.e. 10-6 or 10-7, Eb/No improves, it means spectral efficiency
reduces (for the given SNR) but the reliability in transmission improves.

Problem 7.2 In a QPSK transmitter, the information bit rate is 100 kbps. If the available bandwidth is
200 kHz, find the spectral efficiency of the scheme. Compare it with the case of 16PSK.

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

Solution: Ideally, for QPSK the spectral efficiency is 2 bps/Hz and for 16PSK it is 4 bps/Hz as we
transmit the symbols rather than bits.

Now For this particular case the spectral efficiency is to be calculated w.r.t. available BW. So, the
spectral efficiency = 100 kbps/ 200 KHz = 0.5 bps/Hz.

For 16PSK scheme the spectral efficiency is 2 times the QPSK scheme, but if we consider same bit
rate as in QPSK (100 kbps) the spectral efficiency reduces by two compared to QPSK.

Problem 7.4 Calculate the minimum bandwidth requirement for a non-coherent BFSK system if the
symbol duration is 0.2 ms. What is the minimum bandwidth required for an MSK system having the
same symbol duration?

Solution: For M ary FSK frequency deviation can be found as follows

2 x 2f x Ts = 2P

Now for BFSK P = 1

 2f x Ts =1
 f = 1/2Ts ( for minimum bandwidth)
 Peak deviation = 2f
As per Carson’s rule the transmission BW ≈ 2(f + R)

= 2[(1/ 2 x 0.2 x 10-3)+(1/0.2 x 10-3)

= 3/(0.2 x 10-3)

= 15 kHz

There is another formula for MFSK found in the literature considered for rectangular wave

Transmission bandwidth BT = R x M/ log2 M here M =2

BT = 1 x 2/[(0.2 x 10-3) x log2 2] = 10 kHz

Now due to high roll off, the MSK signal there is no significant spectral power for frequencies below
fc-1/4Ts and fc + 1/4Ts, so the transmission BW taken as 1/2Ts ≈ 2.5 kHz

Problem 7.5 If it is required to send 1.024 × 106 binary digits (bits) per second with Pe ≤ 10−6 with
BPSK and 16PSK, the channel noise power spectral density is No = 10−8 .
Determine the transmission bandwidth and the signal power required at the receiver input in both the
cases.

Solution: Assume that the channel is not bandlimited so the transmission bandwidth to be found is
sufficient to support the required bit rate 1.024 Mbps.

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

If we assume Eb/No ≈ 11.5 dB from the available graphs of BPSK (Prob.1) for Pe  10-6 then

Eb/No(dB) = S/No(dB.Hz) – R(dB)

Problem 7.6 In the case of PSK, prove that bit error probability is

Solution: If there are total of M possible signals, the modulation signal set S can be represented as

S = { s1(t), s2(t), ……….sM(t) } ……………(1)

For binary modulation schemes, bit Is mapped directly to a signal and S will contain only two signals.
For M-ary keying signal set will contain more than two signals and each signal (or symbol) will
represent more than a single bit of information. Elements of S can be considered as points in a vector
space.

Any finite set of physically realizable waveforms in a vector space can be expressed as a linear
combination of N orthonormal waveforms which form the basis of that vector space. Once the basis is
determined, any point in that vector space can be represented as a linear combination of the basis
signals { Φk(t)| k=1,2,…,N} such that

N
sj(t) =  s jk  k (t ) ……………(2)
k 1

The basis signals are orthogonal to one another in time such that


  j (t )  k (t )dt  0 j≠k ……….. (3)


Each of the basis signals is normalized to have unit energy


  j (t )dt  1
2
So, E= ……………. (4)


Now, for BPSK scheme Eb = bit energy, Tb = Bit period

2 Eb
s1 (t )  cos( 2 f c t ) 0 ≤ t ≤ Tb ……..(5a)
Tb

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

2 Eb
s2 (t )   cos( 2 f c t ) 0 ≤ t ≤ Tb ……..(5b)
Tb

Rectangular pulse shape rect[(t-Tb/2)/Tb] is assumed.

 j (t ) for this signal set simply consist of single waveform 1 (t )

2
=> 1 (t )  cos( 2 f c t ) 0 ≤ t ≤ Tb …….. (6)
Tb

Using this basis signal the BPSK signal set can be represented as

SBPSK = { E b 1 (t ),  E b 1 (t ) }……………..(7)
This signal set can be shown on the constellation diagram. Now the probability of bit error is
proportional to the distance between the closest points in the constellation, hence the modulation
scheme with a dense constellation is less energy efficient compared to a modulation scheme that has a
sparse constellation. A simple upper bound for the probability of symbol error in an AWGN channel
with a noise spectral density No for an arbitrary constellation can be obtained using the union bound,
which provides some estimate for average probability of error for a particular modulation signal.
Conditional probability

 d jk 
Pe ( / s j )   Q  ……………..(8)
k 1,k  j  2 N o 
 

Here, djk is the Euclidian distance between jth and kth signal point in the constellation and

 1
Q( x)   exp(  x 2 / 2)dx …………..(9)
x 2

PSK schemes exhibit symmetric constellations, the distance between all constellation points are

equivalent. The conditional error probability Pe ( / s j ) is the same as for all j , hence equation
(8) gives the average probability of symbol error.

Now in BPSK the Euclidian distance is 2 E b between adjacent points in the constellation.
Substituting this into equation (8) the probability of bit error is obtained as

2 Eb
Pe  Q
No

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

In case of QPSK Euclidian distance is 2 E s but Es = 2Eb => effectively the distance between

neighboring points = 2 E b , so

2 Eb
Pe  Q
No

Problem 7.7 A QPSK scheme is used to modulate the bit stream 001010011100. Sketch the
transmitted waveform. Assume fs = fb /2. Also, get the waveform for 8PSK for the same stream but
assume fs = fb /3.

Solution:

PHASE CHANGE (Degrees) Dibit

0 01

90 00

180 10

270 11

QPSK WAVEFORM

or as per phases shown below

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

For 8PSK should be done on MATLAB

Problem 7.8 For the sequence in Problem 7.7, draw the MSK signal waveform.

Solution:

MSK waveform

001010011100.

Better waveforms can be observed using MATLAB.

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

There are two important differences between MSK and QPSK

1. In MSK the baseband waveform, that multiplies the quadrature carrier, is much smoother than
the abrupt rectangular waveform of QPSK. While the spectrum of MSK has a main center
lobe which is 1.5 times as wide as the main lobe of QPSK, the side lobes in MSK are
relatively much smaller in comparison to the main lobe, making filtering much easier.
2. The waveform of MSK exhibits phase continuity , that is , there are no abrupt changes as in
QPSK. As a result we avoid the ISI by nonlinear amplifiers

Problem 7.10 Prove that the phase in MSK is continuous.

Solution:

The resulting signal for CPFSK is represented by the formula

where aI(t) and aQ(t) encode the even and odd information respectively with a sequence of square
pulses of duration 2T. Using the trigonometric identity, this can be rewritten in a form where the
phase and frequency modulation are more obvious,

where bk(t) is +1 when aI(t) = aQ(t) and -1 if they are of opposite signs, and φk is 0 if aI(t) is 1, and π
otherwise. Therefore, the signal is modulated in frequency and phase, and the phase continuously and
linearly changes.

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

MSK is encoded with bits alternating between quarternary components, with the Q component
delayed by half the symbol period. However, instead of square pulses as OQPSK uses, MSK encodes
each bit as a half sinusoid. This results in a constant-modulus signal, which reduces problems caused
by non-linear distortion.

Problem 7.11 For a band-pass channel using BPSK signaling, the data rate is 9.6 kbps. Find the
transmission bandwidth between (a) the first nulls and (b) the second nulls. Also, find the peak value
of the spectrum.

Solution: for the BPSK signal 1 bit per symbol is used so Rs = 1/Ts.

Bandwidth between first nulls = 2/2Ts = 9.6 kHz

Bandwidth between second nulls = 3/Ts = 28.8 kHz

Problem 7.13 Show that the periodic PN signal normalized autocorrelation function corresponding to
the binary case with equal probability of 1’s and 0’s (+1 and −1) and with N = 7 is of the form shown
in Fig. 7.62 .

Solution: Normalized autocorrelation for N=7 is 1/7

Taking the sequence with equal probability of 0’s and 1’s cyclic shift is necessary

+1 +1 +1 -1 -1 +1 -1

-1 +1 +1 +1 -1 -1 +1

-1 +1 +1 -1 +1 -1 -1 = -1

+1 +1 +1 -1 -1 +1 -1

+1 -1 +1 +1 +1 -1 -1

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

+1 -1 +1 -1 -1 -1 +1 = -1

Problem 7.14 An aperiodic PN waveform is defined for the interval N c = 7 with a sequence +1, −1,
−1, +1, −1, +1, +1. Find the autocorrelation function of this process. Is this an ideal aperiodic
sequence? (Hint : Different shifts will give different values of the correlation functions.)

Solution:
Without shift +1 -1 -1 +1 -1 +1 +1
+1 -1 -1 +1 -1 +1 +1
____________________

+1 +1 +1 +1 +1 +1 +1 =7

With 1 shift +1 -1 -1 +1 -1 +1 +1
+1 -1 -1 +1 -1 +1 +1
____________________

0 -1 +1 -1 -1 -1 +1 0 =2

Similarly with 2 shifts autocorrelation is -1 and so on….

Thus the sequence does not satisfy the basic property of being two valued autocorrelation => not an
ideal aperiodic sequence.

Problem 7.15 Evaluate the 15-bit spreading code 100110101111000 for balance, run property, and
correlation.

Solution:

(1) Balance -> 8 1’s and 7 0’s => The number of binary 1’s differ from binary 0’s by at most one
digit so, property satisfied.
(2) Run-> Total runs =8

1 00 11 0 1 0 1111 000

1 2 2 1 1 1 4 3

Thus half of the runs are of length 1, ¼ are of length 2, 1/8th are of length 3…. Property satisfied.

(3) Correlation property-> substituting +1 for logic 1 and -1 for logic 0 and applying one shift
autocorrelation comes out to be -1

Problem 7.16 A DSSS system is operating under the conditions of two independent jammers having
very close centre frequency. The chip rate is 125 Mbps and the message bit rate is 2500 bps. If the
combined bandwidth of the two jammers is 60 kHz and their respective received signal powers are 0.2
× 10−7 W and 10−6 W, what is the approximate interference power at the output of the correlator?

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

Solution: Bs = 2/tchip = 2 x 125 = 250 MHz

Bm = 1/tm = 2500 Hz = 2.5 kHz

PG = Bs/Bm = (250 x 106)/ (2.5 x 103) = 1,00,000

The combined bandwidth of two jammers is 60 kHz which is less than transmission bandwidth so the
equation we can use is

[(0.2 x 10-7)2 + (10-6)2 ]/2= J/PG

=> [0.04 x 10-14 + 10-12]/2 = J/PG

=> J = 5.002 x 10-8 W

=> J = -73 dB

Problem 7.17 In a DSSS system, the message bit rate is 4800 bps and the PN clock rate is 220 Mbps.
(a) Find the processing gain.
(b) Find the output SNR if the received signal power is 6 × 10−10 W and one-sided noise spectral
density is 10−20 W/Hz.
(c) If the coherent reference signal is off by 10° and the synchronization error is 30%, will the code
noise term be a determining factor in the output SNR? How?

Solution: (a) clock rate = chip rate

Processing gain for biphase modulation = 2tm/ tchip = 2 x 220 x 106/ 4800 = 91666.66

Processing gain for quadriphase modulation = tm/ tchip = 45833.33

(b) SNR out = PG x Pr / No x Bs (neglecting interfering signals)

[91666.66 x 6 x 10-10]/ 10-20 x 220 x 106 ≈ 2500 x 104 = 2.5 x 107

(c) Whatever degree it is, if there is a loss of synchronization, the code noise term will be a
determining factor in output SNR because, synchronization loss results in increased probability of bit
errors, so on the BER-> SNR plot, this indirectly represents the reduction in SNR, means to maintain
the same BER we need to increase output SNR. 30% synchronization error code noise is a
considerable factor in the output SNR

Problem 7.18 A speech signal band-limited to 4 kHz and 256 quantization levels is to be transmitted
by a DSSS transmitter. Find the required chip rate to obtain a processing gain of 6400.

Solution: Assuming biphase modulation.

4 kHz signal should be sampled with 8 kHz sampling rate,

256 quantization levels, so 8 bits per sample => 8 x 8 = 64 kbps bit rate= Bm

Now PG = Bs/Bm

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

=> Bs = PG/Bm = 6400 x 64 x 103 = 409.6 Mbps

Problem 7.19 An FHSS system utilizes a fast hop system that contains 10 hops/message bit and 1024
hopping frequencies. The message bit rate is 2400 bps and the final RF multiplication factor is 10.
Find the (a) RF signal bandwidth, (b) processing gain in decibels, (c) PN code generator clock rate,
and (d) frequency separation in kHz. For the found value of processing gain, what must be the
equivalent PN code rate when t m = 1/42,000. [Hint : PN code generator clock rate = k ( c − 1) ×
message bit rate]

Solution: K=10, k=10, M=1024

(a) RF signal bandwidth = Bs= KMf1 = KM/t1= KM/(tm/k) = 10 x10x 1024/(1/2400) = 245.76
MHz
(b) Processing gain = kKM = 10 x 10 x 1024=102400
PG in dB = 50.1 dB
(c) PN code generator clock rate, i.e. chip rate,
PN code generator output c-1 bits
=>1024 = 1010
=> c=10
=> c-1=9
Now PN code generator clock rate = k(c-1) x message bit rate
= 10 x 9 x 2400
= 2,16,000 Hz = 216 kHz
(d) There are total 1024 frequencies in 245.76 MHz. If we remove multiplication factor of 10 =>
the bandwidth is 24.576 MHz due to total number of frequencies
 Frequency separation= 24.576 x 106/ 1024 = 24 kHz

For the same value of processing gain, PN code rate depends on k, c and tm
Now as tm changes to 1/42,000
=> code rate=10 x 9 x 42000 = 3.78 MHz

Problem 7.20 In an FH system, the switching speed of the synthesizer is 5 μs and the message bit rate
is 5 kbps after error-correction coding. There are 5 hops/message bit and the final frequency
multiplication is 8. What is the maximum processing gain obtainable?

Solution: M frequencies produced by frequency synthesizer are separated by duration t chip = 5 μs 1/

tm = 5 kbps, k=5, K=8

Clock rate = k(c-1) x (1/tm)

 1/(5 x 10-6) = 5(c-1) x 5000

 c-1 = 1/(5 x 25000 x 10-6)
 =8
 c=9
 M = 2c =512

Now, PG = k K M = 5 x 8 x 512=20,480

=> PG = 43.11 dB

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

Problem 7.21 Consider a hybrid system PN/FHSS. The PN code rate is 250 kchips/s. For the FH,
8192 frequencies are used with a spacing of 250 kHz. For error correction, 1/3 rate convolution
coding is used. Find the processing gain if the final message rate is (a) 75 bps and (b) 2400 bps.

Solution: PN rate of 250,000 chips per second

8192 frequencies separated by 250 kHz

=> FH spreading = 2.048 GHz

Rate 1/3 convolution coding is ignored and need not to touch as the final message rate is given.

(a) Final message rate = 75 bps

 PG for PN/ FH stage = Bs/Bm = (2.048 x 109)/75 ≈ 27,306,666

 PG = 74.32 dB

(b) Final message rate = 2400 bps

 PG for PN/ FH stage = Bs/Bm = (2.048 x 109)/2400 ≈ 8,53,333

 PG = 59.31 dB

Problem 7.22 For the polynomials h ( x ) = x4 + x3 + x2 + 1 and h ( x ) = x4 + x2 + x + 1, assuming

initial condition (1111), (a) calculate and plot the autocorrelation functions for each, (b) calculate and
plot the cross-correlation functions, and (c) draw the block diagrams using both the Fibonacci and
Galois methods.

Solution: (1) h(x) = x4+x3+x2+1

Fibonacci Implémentation Galois implémentation

Séquence : 1 1 1 1 1 1 1 1

1 1 1 1 1 1 0 0

1 1 1 1…… 0 1 1 0….

If we check the balance property, autocorrelation property of these sequences we can conclude that It
is not a valid m-sequence. Valid is x4+x3+1

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

(2) h(x) = x4+x2+x+1

Fabonacci Implementation Galois implementation

Same as the above case

Problem 7.23 For the polynomials in Problem 7.22, find the Gold sequence.

Solution: Polynomials in Prob. 11 are not valid sequences but in general the method of finding the
Gold-sequence is as follows.

Gold sequences are generated by modulo-2 addition of two m-sequences, clocked by the same chip
clock. The important part is that only special pairs of m-sequences exhibit the desired correlation
properties. Both the m-sequences have equal length and use the same clock, hence the created Gold
sequence is of the same length as the former ones.

It can be shown that for any number of shift in the initial conditions between the two m-sequences , a
new Gold-sequence is generated (Gold, 1967). If each m-sequence is of length N, the same number of
different shifts between the two m-sequences is available. Thus a Gold-sequence generator can create
a number of different N Gold sequences.

Problem 7.24 In a PN sequence with the length 28 − 1, how many runs of 1111 would be expected?

Solution: 28 -1 = 256-1 = 255 bits

Total runs = x (including bits 1,2,3,4,5,6,7

half of the runs of length 1 => x/2 runs of 1

one fourth are of length 2 => x/4 runs of length 2

x/8 runs of length 3 and so on…….

x x x x x x x x
(1)  (2)  (3)  (4)  (5)  (6)  (7)  (8)  255
2 4 8 16 32 64 128 128

 64x + 64x + 48x +32x +20x +12x + 7x + 8x = 255 x128

 255x =255 x 128
 Total 128 runs
 64 runs of bit 1 (64 bits)
 32 runs of bits 2(64 bits)
 16 runs of bits 3(48 bits)

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

 8 runs of bits 4 (32 bits) => 4 runs of 1111 are expected as equal probability of 0000
and 1111
 4 runs of bits 5 (20 bits)
 2 runs of bits 6 (12 bits)
 1 run of bits 7 (7 bits)
 1 run of bits 8 (8 bits)

Problem 7.25 In an FHSS system, the total bandwidth Bs = 400 MHz and the individual channel
bandwidth Bm = 100 Hz. What is the minimum number of PN bits required for each frequency hop?

Solution: Assuming K=1

Bs = Mf1 = 2c x100

=> 400 x 106 /100 = 2c

=> 2c = 4 x 106

=> c = log2(4 x 106) = log10(4 x 106)/ log102 =21.93

For minimum number of PN bits c = 21 is required

Problem 7.26 For an OFDM system, if 52 subcarriers, spaced at 312.5 kHz, are defined, find the total
occupied bandwidth excluding the secondary lobes.

Solution:

Total occupied bandwidth = (N + 1) x f

= (52 + 1) x 312.5 kHz

= 16,562.5 kHz

 16.56 MHz

Problem 7.27 If the OFDM bandwidth is 19.2 MHz and 48 subcarriers are modulated with
convolution-coded 2/3 rate data using (a) BPSK and (b) QPSK, in each case what must be the input
data rate?

Solution:

Bandwidth = (N + 1) x f

19.2 x 106 = (48 + 1 ) x f

 f = 391.84 kHz
Now f = 1/Ts => Ts = 2.55 s

Now considering 1 bit/symbol for BPSK and 2 bits/symbol for QPSK

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

Tb = 2.55 s for BPSK => bit rate for BPSK after convolution encoder stage = 391.84 kbps

Tb = 1.276 s for QPSK => bit rate for QPSK after convolution encoder stage = 783.7 kbps

So, the input rates will be

391.84 x 2/3 = 261.22 kbps (for BPSK)

783.7 x 2/3 = 522.47 kbps (for QPSK)

Problem 7.28 In an OFDM system, 52 subcarriers are defined with 300 kHz spacing. If, for the
channel estimation purpose, 8 pilot carriers are added at equal distance maintaining the same spacing,
what will be the percentage rise in the occupied spectrum?

Solution:

Occupied spectrum without pilots W = (N + 1) x f

= (52+1) x 300 x 103

= 15.9 MHz

Now adding the pilots at equal spacing, total subcarriers will be now 52 + 8= 60

Now spectrum will be W = (N+ P + 1) x f

= (52 + 8 + 1) x 300 x 103

= 18.3 MHz

Now, 18.3 MHz – 15.9 MHz = 2.4 MHz => approx. 15 % rise in the spectrum

Problem 7.29 In an OFDM system, the frequency domain setting is such that the spacing between
two subcarriers is 312.5 kHz. If the frequency spacing includes a cyclic prefix of 800 ns added to the
OFDM symbol, answer the following:
(a) What will be the final symbol duration with the cyclic prefix?
(b) What will be the OFDM symbol duration without the cyclic prefix?
(c) Will there be any loss of spectrum efficiency due to addition of the cyclic prefix?

Solution:

(a) Assuming cyclic prefix is added before IFFT stage as per assumption for spectrum setting,

Final symbol duration with cyclic prefix Ts = 1/f = 1/(312.5 x 103)

= 3.2 μs

(b) Now CP is 800 ns = 0.8 μs

So the OFDM symbol duration is 3.2 – 0.8 = 2.4 μs without cyclic prefix

© Oxford University Press 2014

Wireless Communication and Networks Upena Dalal

(c) Without CP obviously frequency spacing is more as per above case leading to more spectral
consumption. Hence, there will not be any loss of spectral efficiency due to CP addition once
the spectral setting is not altered.

Problem 7.30 Find the additions and multiplications required for a 256-point FFT and a 1024-point
FFT. By selecting the 1024-point IFFT instead of the 256-point FFT, will you be able to get good
performance? If yes, at what cost?

Solution: For 256 point FFT

No. of complex multiplications = (N log10N)/(2 log102)

= (256 x log10(256))/ (2 log102)
= 616.5/0.6
= 1024
No. of complex additions = (N log10N)/(log102)
= (256 x log10(256))/ 0.3
= 2048
 The total operations are 1024 + 2048 = 3072
For 1024 point FFT same way we can calculate No. of multiplications = 5120 and No. of additions =
10240, so total operations are 15360.

Taking 1024 point IFFT better resolution of the OFDM baseband s achieved but at the cost of
(1) more processing time
(2) More number of subcarriers or more zero padding
(3) Faster processor as far as the hardware is concerned.

Problem 7.31 A data rate of 5 Mbps is targeted in a multipath radio environment by using BPSK
modulation. The maximum delay spread is 25 μs, and the number of subcarriers is 128 for multicarrier
transmission. Compare the ISI effect if the system is (a) single carrier and (b) multicarrier.

Solution: For single carrier scheme

ISI = max / Tsc = max x Rsc = 25 x 10-6 x 5 x 106

= 125

For multicarrier scheme

ISI = max / Tmc = (max x Rsc) = (25 x 10-6 x 5 x 106)/128

= 0.976

Which is very less compared to first case

© Oxford University Press 2014