Exercises

Module Basics Sampling and Reconstruction

Notes:
• Only the answers are available.
• The symbol [P] in the margin of an exercise denotes there is a pencast available.

1
 Exercise 1
[P1] Let x(t) = 3 sin(20πt − π/3) be the input signal of the ideal C/D convertor as depicted in Fig. (1).

Ideal
x(t) x[n]
C/D
convertor

fs = 1/Ts

Figure 1: Ideal C/D converter.

In each of the following cases the discrete-time signal samples x[n] are obtained by sampling
x(t) at a sample rate fs [samples/sec] and the resulting signal samples x[n] can be written as

x[n] = A sin (θn + ϕ)

with A > 0, 0 < θ < π and |ϕ| < π. For each case below, determine the values of A, ϕ and
θ. Make a spectral plot of x[n] as a function of the relative frequency θ [-] in the Fundamental
Interval (|θ| ≤ π) and compare this plot with the spectral plot of x(t) as a function of absolute
frequency f in [Hz].
a. Let the sampling frequency be fsa = 24 [samples/sec].
b. Let the sampling frequency be fsb = 16 [samples/sec].
c. Let the sampling frequency be fsc = 8 [samples/sec].

Exercise 2
The continuous-time signals x1 (t) and x2 (t) are given as

x1 (t) = 2 cos (4πt) , x2 (t) = 2 cos (6πt) .

a. Carefully plot the signals x1 (t) and x2 (t) in one figure such that at least four periods of each
signal are shown.
b. In your plot, identify the points for which x1 (t) and x2 (t) are equal. At what frequency do
these points occur (i.e., how many times per second do we have x1 (t) = x2 (t))?
c. Suppose that the signals x1 (t) and x2 (t) are now sampled with exactly this frequency. We
thus obtain the sampled signals x1 [n] and x2 [n]. Obtain a formula for x1 [n] and x2 [n] to
show that these samples are equal.

Exercise 3
A signal x(t) has the two-sided spectrum representation shown in Fig.(2).

Spectrum x(t)
10
8ej 3 8e−j 3
π π

4 4

f [Hz]
-300 -100 0 100 300

Figure 2: Frequency spectrum of x(t).

2
 The signal x(t) is sampled with sampling frequency fs = 1/Ts [samples/sec] to obtain the
discrete-time signal samples x[n] = x(t)|t=n·Ts .
a. Give an equation for x(t) as sum of sinusoids.
b. Give an equation for x[n] and plot the spectrum of x[n] as a function of the relative frequency
θ within the Fundamental Interval −π ≤ θ ≤ π when fs = 900 [samples/sec].
c. Do the same as in the previous question, but now for fs = 300 [samples/sec].

Exercise 4
A signal x(t) has the two-sided spectrum representation shown in Fig.(3).

Spectrum x(t)
10ej 4 10e−j 4
π π

6e−j 3 6ej 3
π π

f [Hz]
-50 -20 0 20 50

Figure 3: Spectrum of x(t).

a. Write an equation for the signal x(t) and a sum of sinusoids.

b. Suppose that the signal is sampled to produce the sequence x[n] = x(t)|t=nTs , where 1/Ts =
fs = 200 [Hz]. Sketch the spectrum of the sampled signal as a function of the relative
frequency θ in the range: −2π < θ < 2π (i.e. show the alias frequencies). Thus make
a spectral plot of all the frequency components of x[n] not only within the Fundamental
Interval, but also outside this interval, i.e. show also alias frequencies.
Carefully label the complex weights of the frequency components and verify which frequency
components are the same.

Exercise 5
A signal x(t) has the two-sided spectrum representation shown in Fig. (4).

Spectrum x(t)

4ej 3 4e−j 3
π π

2ej 2 2e−j 2
π π

f [Hz]
-40 -12 0 12 40

Figure 4: Spectrum of x(t).

a. Is x(t) a real valued signal? If so, explain why?

b. Write an equation for the signal x(t) and a sum of sinusoids.
c. Is x(t) a periodic signal? If so, what is its period?

d. Determine the minimum sampling rate fs [samples/sec] that can be used to sample x(t)
without any aliasing.

3
 Exercise 6
Signal x(t) is represented by the following equation:
2
(αk ) ej2400πkt
X
x(t) =
k=−2

with α0 = √1 ,
3
αk = √ 1√
3+j( 3)k
for k = 1, 2 and α−k = αk∗ .

a. Is this signal periodic? If so, what is the Fundamental period T0 ?

b. Write an equation for the signal x(t) and a sum of sinusoids.
c. Determine the minimum sampling rate fs [samples/sec] that can be used to sample x(t)
without any aliasing.

Exercise 7
[P2] Consider the sampling system shown in Fig. (1) with sampling rate 1/Ts = fs = 2500 [sam-
ples/sec]. Suppose that the discrete-time signal x[n] is given by the formula
 
2π π
x[n] = 10 cos n− .
5 3

Many different continuous-time signals could have been the input to the above system. Determine
two such input signals with frequency less than 2500 [Hz]; i.e., find a continuous time x1 (t) =
A1 cos(2πf1 t + φ1 ) and another continuous time signal x2 (t) = A2 cos(2πf2 t + φ2 ) such that
x[n] = x1 (t)|t=nTs = x2 (t)|t=nTs .
Exercise 8
Consider the D/C convertor as depicted in Fig. 5.

Ideal
y[n] y(t)
D/C
convertor

fs = 1/Ts

Figure 5: Ideal D/C converter.

The discrete-time signal samples can in general be described with the equation: y[n] = sin(θ0 n+
π
3 ).The D/C convertor runs at a sample rate of fs [samples/sec]. The continuous-time signal can
in general be described as y(t) = A sin(2πf0 t + φ), with A > 0, f0 > 0 and φ > 0.
a. Assume θ0 = π2 and fs = 20 [samples/sec]. Derive a mathematical expression for y(t) and
give a plot of the spectra from y[n] and y(t).
b. Assume θ0 = π2 and fs = 32 [samples/sec]. Derive a mathematical expression for y(t) and
give a plot of the spectra from y[n] and y(t).
c. Assume θ0 = 3π 2 and fs = 20 [samples/sec]. Derive a mathematical expression for y(t) and
give a plot of the spectra from y[n] and y(t).
d. Assume θ0 = 3π 2 and fs = 32 [samples/sec]. Derive a mathematical expression for y(t) and
give a plot of the spectra from y[n] and y(t).

4
Exercise 9
Consider the sampling and reconstruction system shown in Fig. (6).

Ideal Ideal
x(t) x[n] y(t)
C/D D/C
convertor convertor

fs = 1/Ts fs = 1/Ts

Figure 6: Sampling and reconstruction system.

In this figure, x[n] = x(t)|t=nTs and x(t) is given by the formula

 π
x(t) = 10 cos(10πt) + 2 cos 40πt + .
3
a. What condition must be satisfied by the sampling rate, fs = 1/Ts , such that y(t) = x(t)?
b. What is the output y(t) if the sampling rate is fs = 20 [samples/sec]? Sketch the spectrum
of the sampled signal x[n] (as a function of the relative frequency θ for relative frequencies
−2π < θ < 2π, i.e. show also aliased frequencies and label the axis carefully.

Exercise 10
Consider again the sampling and reconstruction system of Fig. (6). The ideal C-to-D converter
samples x(t) with a sampling period Ts = 1/fs [sec] to produce the discrete-time signal samples
x[n]. The ideal D-to-C converter then forms a continuous-time signal y(t) from the samples x[n].
Suppose that x(t) is given by

x(t) = (10 + 10 cos(500πt − π/2)) · cos(2000πt).

a. Use Eulers formulas to expand x(t) in terms of complex exponentials and sketch the two-sided
spectrum of this signal. Be sure to label important features of the plot. Is this waveform
periodic? If so, what is the period?
b. What is the minimum sampling rate fs that can be used in the above system such that
y(t) = x(t)?

c. Plot the spectrum of the discrete-time signal samples x[n] for the case when fs = 5000 [Hz]
as a function of the relative frequency θ with |θ| ≤ π.

Exercise 11
Consider again the sampling and reconstruction system of Fig. (6). In all parts below, the sampling
rates of the C/D and D/C converters are equal, and the input to the Ideal C/D converter is
 π
x(t) = 2 cos 100πt + + cos (300πt) .
2
a. If the output of the ideal D/C Converter is
 π
y(t) = x(t) = 2 cos 100πt + + cos (300πt) ,
2
what general statement can you make about the sampling frequency fs in this case?

5
 b. If the sampling rate is fs = 250 [samples/sec], determine the discrete-time signal samples
x[n], and give an expression for x[n] as a sum of cosines. Make sure that all frequencies in
your answer are positive and less than π [radians]. Plot the spectrum of this signal over the
range of relative frequencies −π < θ < π. Label the frequency, amplitude and phase of each
spectral component.

c. If the output of the ideal D/C Converter is

 π
y(t) = 2 cos 100πt + + 1,
2
determine the value of the sampling frequency fs .

Exercise 12
[P3] Consider the sampling and reconstruction system shown in Fig. (7).

Ideal Ideal
x(t) x[n] y(t)
C/D D/C
convertor convertor

fsi = 1/Tsi fso = 1/Tso

Figure 7: Sampling and reconstruction system with different sampling rates fsi and fso .

The input signal x(t) is given as

 π
x(t) = cos 400πt + .
4
Both the C-to-D converter and the D-to-C converter can be operated at two different sampling
frequencies, which are 400 [Hz] and 800 [Hz]. Explain which sampling frequency you would choose
for each converter to make sure that the output signal is given as
 π
y(t) = cos 200πt + .
4

Exercise 13
Consider the sampling and reconstruction system of Fig. (7).
a. Suppose that the input x(t) is given by
3π
x(t) = 10 + 10 cos(4000πt − π) + 8 cos(14000πt − ).
4
Determine the spectrum for x[n] when fsi = 10000 [samples/sec]. Make a plot for your
answer, making sure to label the frequency, amplitude and phase of each spectral component.
b. Using the discrete-time spectrum from part (a), determine the analog frequency components
in the output y(t) when the sampling rate of the D-to-C converter is fso = fsi = 10000 [Hz].
c. Again using the discrete-time spectrum from part (a), determine the analog frequency
components in the output y(t) when the sampling rate of the D-to-C converter is fso =
20000 [Hz]. In other words, the sampling rates of the two converters are different.

6
Exercise 14
Consider again the sampling and reconstruction system of Fig. (6). We can do some interesting
things with sampling. One of them is that we can change the period of a periodic waveform. This
problem illustrates how this can be done for the specific periodic input signal

x(t) = 2 cos(66πt) + cos(198πt).

In all the following parts, assume that the sampling frequency is fs = 30 [Hz]. Note that this
sampling rate does not satisfy the conditions of the Shannon sampling theorem, so aliasing will
occur.
a. Plot the spectrum of the periodic continuous-time signal x(t). What is the fundamental
frequency of x(t)?

b. Determine an expression for the discrete-time signal x[n] as a sum of discrete-time cosine
signals. Be sure that all of the relative frequencies are positive and less than π radians. Plot
the spectrum of x[n] over the range of relative frequencies −π ≤ θ ≤ π.
c. Now the continuous-time output signal y(t) that is created by the ideal D-to-C converter
operating with sampling rate fs = 30 [Hz] will also be a sum of cosine signals. Write an
expression for y(t) and plot its spectrum. What is the fundamental frequency of y(t)?
d. How are the fundamental frequencies of x(t) and y(t) related? Do you think that it would
be possible to change the fundamental frequency by a different factor by using a different
sampling rate?