Digital Signal Processing for Problems Page 1

Engineering Applications

Problem 1

A sine wave with the frequency f0 = 50 kHz and the amplitude A = 2.5 V is sampled and
quantized with a 14 bit Analog-to-Digital converter (ADC). The ADC has an input range of
±2.5 V and the sampling rate is fs = 200 kHz.

a) Find the signal-to-noise ratio (related to the quantization noise) of the sampled signal
(expressed in dB and as a ratio).

b) Find the bit rate of the serial digital signal.

c) Sketch the Fourier spectrum (magnitude) of the sampled signal in the range ±300 kHz.

Problem 2

Consider again the ADC of Problem 1.

a) Find the signal-to-noise ratio (related to the quantization noise) of the sampled signal
when the amplitude of the sine wave is A = 0.5 V.

b) If the sine wave has a DC offset of 1 V, find its maximum acceptable amplitude to avoid
clipping.

Problem 3

A digital oscilloscope samples a sine wave of frequency f0 with a sampling rate of

fs = 100 kHz. The FFT function of the oscilloscope displays a spectral line at
frequency fmeas = 33 kHz. Find five possible frequencies f0 which match the
measurement.

Problem 4

Consider an ADC having a sampling rate of 5 MHz. The figure below shows the amplitude
response of the anti-aliasing filter. At the input we have a dual tone signal with frequencies
f1 = 100 kHz and f2 = 3 MHz. Both components have amplitudes of 0 dBV.

Find the frequencies and the amplitudes of the sampled signal in the frequency range
from 0 to fs /2.

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 2
Engineering Applications

Problem 5

Consider the discrete-time signal x(n) = [u(n) − u(n − 4)] (10 − n).

a) Sketch the signal x(n).

b) Sketch the signal y(n) = x(3 − n).

Problem 6

For the discrete-time rectangular pulse

1 0 ≤ 𝑛𝑛 ≤ 3
𝑥𝑥(𝑛𝑛) = �
0 else
determine the signal y(n) = x(n) ∗ x(n) by convolution (make a table of y(n)).

Hint: Use also Python to calculate and plot y(n) scipy.signal.convolve()

Problem 7

A discrete-time LTI system has the impulse response

a n n≥0
h( n) = 
0 else
with 0 < a < 1. Find the step response y(n) = h(n) ∗ u(n) of the system by discrete-time
convolution:

a) Sketch h(k) and below u(n − k) over k for n = 3.

b) Make a table of y(n) for −2 ≤ n ≤ 5.

c) For a = 0.7, calculate the numerical value for y(3).

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 3
Engineering Applications

Problem 8

Consider the discrete-time signal x(n) = (2, −1, 5, 4, −3, −2, 0, −1).

a) Calculate the mean mx of x(n).

b) Calculate the power P of x(n).

c) Calculate the variance 𝜎𝜎𝑥𝑥2 of x(n).

d) Calculate the rms value xRMS of x(n).

e) Calculate the crest factor CF of x(n).

f) Calculate the skewness sk of x(n).

g) Calculate the kurtosis ku of x(n).

Hint: Use also Python to calculate these values.

Problem 9

Calculate the crest factor and the peak-to-average power ratio (in dB) for the following
signals:

a) A sine wave with the amplitude A.

b) A dual tone signal with different frequencies and amplitudes A1 and A2 = A1 /2. Assume a
peak amplitude of Amax = A1 + A2, and that the two sine waves are uncorrelated and that
the total power is the sum of the individual power of the two sine waves.

Problem 10

Consider an ideal lowpass filter with the transfer function Hid( f ) and the bandwidth 1/8 fs.

a) Find the impulse response h(n) of the corresponding discrete-time system.

b) Find the filter coefficients of a corresponding FIR filter of order N = 6.

c) Find the filter coefficients if a Hamming window is applied.

Hint: Use also Python to calculate the filter coefficients scipy.signal.firwin()

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 4
Engineering Applications

Problem 11

Consider an ideal highpass filter with the transfer function Hid( f ) and the cutoff frequency
1/4 fs. Find the impulse response h(n) of the corresponding discrete-time system.

Problem 12

The Figure below shows the amplitude response of the highpass filter with a cutoff frequency
1/8 fs. Assume a sampling rate of fs = 48 kHz.

a) In the Figure, mark the passband and the stopband.

b) From the Figure, determine the minimum attenuation in the stopband (in dB) and the
corresponding frequency (in Hz).

c) For the frequency found in (b), determine the amplitude at the filter output for a sine
wave with an amplitude 𝑢𝑢�𝑖𝑖𝑖𝑖 = 1 V at the input

Problem 13

Compare the following two number formats:

a) Express the number −96 (decimal) as an 8 bit number in two's complement notation.

b) What decimal number results, if the binary number of (a) is interpreted as a Q.7 number?

c) Express the number −96 (decimal) as a 16 bit number in two's complement notation.

d) What decimal number results, if the binary number of (c) is interpreted as a Q.15
number?

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 5
Engineering Applications

Problem 14

The block diagram below shows a typical analog front end of a sensor system. The voltage
source U0 with source impedance R0 represents the sensor. It is followed by a variable gain
amplifier and an antialiasing filter connected to the analog-to-digital converter.

R0 = 200 Ω Antialiasing Filter ADC

VGA
BN = 100 kHz 16 bit

U0 fs = 1 MHz
Ri = 200 Ω
nV
�𝑁𝑁0,𝑉𝑉𝑉𝑉𝑉𝑉 = 30 Uin = 1.4 Vpp
√Hz
SNR = 90 dB
U0 = 10 mVpp VGA: Variable Gain Amplifier
ADC: Analog-to-Digital Converter
SNR: Signal-to-Noise Ratio

In this problem, we will determine the signal-to-noise ratio (SNR) of the signal chain.

a) Find the voltage gain V of the VGA, which is required to obtain a full-scale sensor signal
at the ADC input (pay attention to the voltage divider R0 – Ri). Convert V to dB.

b) Find the signal power S at the ADC input (assume a sine signal).

c) The noise of the sensor is determined by the thermal noise of the source impedance R0.
White noise has the power spectral density N0/2; with N0 given by
N0 = 4 k T R0
k = 1,38⋅10−23 Ws/K is the Boltzmann constant. Assume an absolute temperature
T = 300 K. Find N0 and the rms value of the noise voltage �𝑁𝑁0 .

d) Find the rms value of the thermal noise voltage �𝑁𝑁0,𝑡𝑡ℎ at the ADC input (pay attention to
the voltage divider R0 – Ri).

e) Find the power Nth of the thermal noise at the ADC input. The power is obtained by
multiplication of N0,th with the noise bandwidth BN of the antialiasing filter.

f) Find the power NVGA of the amplifier noise at the ADC input. N0.VGA is the power spectral
density of the amplifier noise at the output, while �𝑁𝑁0,𝑉𝑉𝑉𝑉𝑉𝑉 is its rms value (output-
referred noise).

g) Find the power NADC of the ADC noise (essentially the quantization noise).

h) The total noise power N is the sum of thermal noise, amplifier noise and ADC noise. Find
the SNR (in dB) at the output of the system.

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 6
Engineering Applications

Solutions

P1 a) s = 86.04 dB, S/Nq = 401.79⋅106

b) rb = 2.8 Mbit/s

P2 a) s = 72.06 dB
b) A = 1.5 V

P3 f0 = 33 kHz, 67 kHz, 133 kHz, …

P4 1. f = 100 kHz, −1 dBV 2. f = 2 MHz, −40 dBV

P6 a) n y(n)
−2 0
−1 0
0 1
1 2
2 3
3 4
4 3
5 2
6 1
7 0
8 0

P7 a)
n y(n)
−2 0
−1 0
0 a0
1 a0 + a1
2 a0 + a1 + a2
3 a0 + a1 + a2 + a3
4 a0 + a1 + a2 + a3 + a4
5 a0 + a1 + a2 + a3 + a4 + a5
… …

b) y(3)= 2.533

P8 mx = 0.75 P = 7.5 𝜎𝜎𝑥𝑥2 = 6.94 xRMS = 2.74 CF = 1.83

sk = 0.23 ku = 1.83

Fakultät Elektrotechnik Prof. Dr. C. Roppel

Digital Signal Processing for Problems Page 7
Engineering Applications

P9 a) CF = 3.01 dB, PAPR = 3.01 dB

b) CF = 5.56 dB, PAPR = 5.56 dB

1 1
P10 a) ℎ(𝑛𝑛) = 4 si �4 𝜋𝜋 𝑛𝑛�

b) c)
i bi w(i) bi w(i)
0 0.0750 0.08 0.0060
1 0.1592 0.31 0.0493
2 0.2251 0.77 0.1733
3 0.2500 1 0.2500
4 0.2251 0.77 0.1733
5 0.1592 0.31 0.0493
6 0.0750 0.08 0.0060

1 1
P11 a) ℎ(𝑛𝑛) = si(𝜋𝜋 𝑛𝑛) − 2 si �2 𝜋𝜋 𝑛𝑛�

P12 b) Amin ≈ −22.5 dB at f ≈ = 3.84 kHz

c) 𝑢𝑢�𝑜𝑜𝑜𝑜𝑜𝑜 = 75 mV

P13 a) 1010 0000

b) −0.75
c) 1111 1111 1010 0000
d) −0.002929

P14 a) G = 48.94 dB
b) S = 0.245 V2
c) �𝑁𝑁0 = 1.82 ∙ 10−9 V/√Hz
d) �𝑁𝑁0,𝑡𝑡ℎ = 2.5478 ∙ 10−7 V/√Hz
e) 𝑁𝑁𝑡𝑡ℎ = 6.4915 ∙ 10−9 𝑉𝑉 2
f) 𝑁𝑁𝑉𝑉𝑉𝑉𝑉𝑉 = 9 ∙ 10−11 𝑉𝑉 2
g) 𝑁𝑁𝐴𝐴𝐴𝐴𝐴𝐴 = 2.45 ∙ 10−10 𝑉𝑉 2
h) SNR = 75.55 dB

Fakultät Elektrotechnik Prof. Dr. C. Roppel