Scribd
Upload
120 views

Midterm Examination-II Solutions

The document is a handout for the solutions to a midterm exam in CS477 taken on November 13, 2005. It contains the solutions to 5 problems, each worth a different number of points, for a total of 150 points. The handout provides the questions, steps to the solutions, and final answers for each of the 5 exam problems related to topics like analog audio transmission, sampling, time division multiplexing, frequency modulation, and noise/noise equivalent bandwidth.

Uploaded by

saadqaisar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
120 views

Midterm Examination-II Solutions

The document is a handout for the solutions to a midterm exam in CS477 taken on November 13, 2005. It contains the solutions to 5 problems, each worth a different number of points, for a total of 150 points. The handout provides the questions, steps to the solutions, and final answers for each of the 5 exam problems related to topics like analog audio transmission, sampling, time division multiplexing, frequency modulation, and noise/noise equivalent bandwidth.

Uploaded by

saadqaisar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

CS477: Autumn 2005-2006

Zartash Uzmi

Handout # 25

Midterm Examination-II Solutions


(Open Books, Open Notes)
Time Allowed: 100 minutes (Nov. 13, 2005; 1:05pm-2:45pm)
I certify that I have neither received nor given unpermitted aid on this examination and
that I have reported all such incidents observed by me in which unpermitted aid is given.

Signature
Name

Student ID

Problem 1

[30]

Problem 2

[40]

Problem 3

[20]

Problem 4

[30]

Problem 5

[30]

TOTAL

[150]

SOLUTIONS

Problem 1: Analog Baseband Audio Transmission [30 points] Consider an analog


audio transmission cable used for music transmission at the baseband, where the bandwidth
for music can be considered to be 20 kHz. The cable introduces a signal power loss of L = 10
from source to destination. We want to evaluate the eect of using m equal-length repeater
sections (including the receiver), where the power loss in each segment is L1 .
Assume that the noise added to the signal is modeled as additive white gaussian noise
with power spectral density

N0
,
2

where N0 = 1 108 W/Hz. At the destination, a typical

baseband receiver is employed whose front end is a unity passband gain lowpass lter with
passband characteristics given by:


H(f ) =
5

|f | 20 kHz
20 kHz |f | 25 kHz

f
5000

else

No additional gain is provided at the receiver. When there are no repeater sections, the
destination signal-to-noise ratio ( NS )D is 40 dB.
(a) [5 points] What is the transmitted signal power ST ?
Solution: From Chapter-9 of text, ( NS )D =

ST
,
LN0 W

where W is the one-sided noise equivalent

bandwidth. Since the lter fall-o |H(f )| is linear from 20 kHz to 25 kHz, we have that
2

W =22.5 kHz. This yields, ST =22.5 Watts.


(b) [7 points] Now consider that the transmission uses m repeater sections (including the
receiver), but we intend to keep the same transmit power ST and same destination signalto-noise ratio ( NS )D as in the case of transmission without repeaters. What relation must
exist between m and L such that the destination signal-to-noise ratio ( NS )D remains 40 dB
for the same transmit power you found in part (a)?
Solution: For ST and ( NS )D to remain the same as before, we must have mL1 = L which is
1

obvious from the expressions given in text. But L1 = L m . Thus, we get, m = L

m1
m

(c) [6 points] Under the condition of part (b), what is the transmitted signal power ST ?
Solution: Same as computed before, i.e., 22.5 Watts.
(d) [6 points] If ( NS )D is xed at 40 dB, nd ST for each value of m ranging from 2 to 10.
1
Solution: ST = ST (old) mL
, which gives, ST = 22.5
L

10

m1
m

(e) [6 points] If ST is xed at the value computed in part (a), nd ( NS )D for each value of m
ranging from 2 to 10.
Solution: ( NS )D =

L
(S)
mL1 N D(old)

40

m1

= 10 10 10 mm

Problem 2: Sampling [40 points] Consider a signal x(t) whose Fourier transform X(f )
is shown in the gure below (the frequency axis is in Hz).
X(f)
1

-2W

-W

2W

The signal x(t) is sampled using an ideal sampling function at a sampling rate fs , and
the resulting sampled signal is passed through a low pass lter whose frequency response is
f
given by H(f ) = ( 2W
).

(a) [10 points] What is the Nyquist sampling rate for x(t)? If we sample x(t) at the Nyquist
rate, can it be recovered by passing the sampled signal through the low pass lter H(f )?
If not, why? And if not, how can we recover x(t) from the sampled signal? Explain in
legible writing. Legible and brief answers will get lot more credit than long and unreadable
explanations.
Solution: The Nyquist sampling rate for x(t) is 4W which is twice the maximum frequency
component in x(t). Since H(f ) has a bandwidth W , we can not recover x(t) after the
sampled signal is passed through H(f ). In order to recover the signal x(t) after passing the
samples through a low pass lter, the lter bandwidth must be increased to 2W . That is,
f
use ( 4W
) as the low pass lter.

For each of the following parts, what is the signal x(t) at the output of the low pass lter,
when the sampling rate is:
(b) [10 points] fs = W samples per second.
) = W  ( f ). Therefore, x(t) = 2W 2 sinc 2W . Note that sampling in time
Solution: X(f
2W
domain leads to repetition of spectrum in the frequency domain scaled by a factor equal to
the sampling frequency.
(c) [10 points] fs = 2W samples per second.
) = 2W ( f ). Therefore, x(t) = 2W 2 sinc2 (W )
Solution: X(f
W

(d) [10 points] fs = 3W samples per second.


) = 0. Therefore, x(t) = 0
Solution: X(f

Problem 3: Time Division Multiplexing [20 points] Eighteen voice signals, each of
bandwidth 3400 Hz, are to be multiplexed together using a TDM scheme. Each signal is
sampled using at-top sampling, where the duration of each sampling pulse is 2 s. The
multiplexing operation includes provision for synchronization by adding two extra pulses,
each of duration 2 s.
(a) [10 points] If the sampling rate is 10 kHz, what is the spacing between successive pulses
of the multiplexed signal?
Solution: Inter-sample interval is

1
10000

= 100 s. If G denotes the spacing between two

channels, then (18 + 2)(G + 2) = 100 which gives G = 3 s.


(b) [10 points] Repeat part (a) if the sampling is carried out at the Nyquist rate?
Solution: The inter-sample interval is now

1
6800

147.06 which yields, G = 5.353 s.

= 147.06 s. Once again, (18 + 2)(G + 2) =

Problem 4: Frequency Modulation [30 points] A tone-modulated FM signal with


modulation index = 2.0 is transmitted through an ideal bandpass lter with unity passband
amplitude, midband frequency fc and bandwidth kfm , where fc is the carrier frequency and
fm is the frequency of the modulating tone, and k is some constant number. You may use
the Bessel function plot from your book, replicated here for convenience.

(a) [6 points] If k = 3, plot the spectrum of the output of the ideal bandpass lter.
Solution: For positive frequencies, we observe three spectral lines of amplitude 0.22Ac at
fc , and 0.58Ac , at fc fm .
(b) [7 points] For k = 3, what are the minimum and maximum values of the envelope?
Solution: First note that it is NOT narrowband FM, it is wideband FM passed through a
narrow lter which in no way makes it a narrowband FM signal. Thus, while the NBFM has
an approximately constant envelope, this signal at the output of the bandpass lter may not
have an approximately constant envelope. The envelope is obtained either from the phasor
diagram or from,

A(t) = Ac J02 () + 4J12 () sin2 2fc t
(c) [8 points] If k = 5, plot the spectrum of the output of the ideal bandpass lter.
Solution: As in part (a) except that we get a total of 5 spectral lines for positive frequencies.
(d) [9 points] For k = 5, what are the minimum and maximum values of the envelope?
Solution: Similar to part (c).

Problem 5: Noise and Noise Equivalent Bandwidth [30 points]


(a) [10 points] White noise with two-sided power spectral density

N0
2

= 106 W atts/Hz is

input to a unity-gain low-pass lter of bandwidth W . i.e., H(f ) = 1 for |f | < W , and is
zero, otherwise. What should be the bandwidth of the low pass lter if the output noise
power is set to be 30 dBm. Be careful about the denition of dBm.
W
Solution: Noise power is N0 W , thus, 10 log N1003
= 30, which implies W =500 kHz.

(b) [10 points] Find the noise equivalent bandwidth of a lter whose time domain response
is rect(t).

Solution: NB =

sinc2 (f )df
sinc(f)|max


0

sinc2 (f )df =

1
2

sinc2 (f )df =

1
2

2 (t)dt =

1
2

(c) [10 points] Find the noise equivalent bandwidth of a lter whose frequency response is
given in the following gure.
H(f)

45 o
-4

Solution: Noting

4
0

x2 dx =

64
.
3

Thus, NB = 4 +

4
3

16
3

You might also like