EMONA ETT101-40

Lab Manual

Signals & Systems Experiments for

the ETT-101/C BiSKIT
 EMONA ETT101-40

Lab Manual

Signals & Systems Experiments for

the ETT-101/C BiSKIT
Signals & Systems Experiments with EMONA ETT101/C

Authors: Robert Radzyner PhD

Carlo Manfredini B.E., B.F.A.

Editor: Carlo Manfredini

Issue Number: 1.0

Published by:
Emona Instruments Pty Ltd,
78 Parramatta Road
Camperdown NSW 2050
AUSTRALIA.

web: [Link]
telephone: +61-2-9519-3933
fax: +61-2-9550-1378

Copyright © Emona TIMS Pty Ltd and its related entities. All rights
reserved. No part of this publication may be reproduced or distributed in
any form or by any means, including any network or Web distribution or
broadcast for distance learning, or stored in any database or in any network
retrieval system, without the prior written consent of Emona Instruments
Pty Ltd.

For licensing information, please contact Emona TIMS Pty Ltd.

NI LabVIEW are registered trademarks of National instruments Corp.

The TIMS logo is a registered trademark of

Emona TIMS Pty Ltd

Printed in Australia
 EMONA TIMS Signals & Systems Lab Manual
for the ETT-101/C BiSKIT Experimenter

Contents

Introduction (i)
How to handle, install the ETT-101-40 plug-in module
ESD Warning, Unpacking, Handling, Set-up and ESD Preventative Measures

Special signals – characteristics and applications Lab 1

Pulse sequence speed throttled by inertia
Isolated step response of a system
Isolated pulse response of a system
Sinewave input

Systems: Linear and non-linear Lab 2

Getting started with linearity
The VCO as a system
A feedback system
Frequency response

Unraveling convolution Lab 3

Unit pulse response
The superposition sum
A rectified sinewave at input
A sinewave input
Mystery application – FIR filter
Comparing impulse response with simulation

Integration, correlation & matched filters Lab 4

Auto-correlation function of PRBS sequences
Cross-correlation function of PRBS sequences
Auto-correlation & matched filtering
Matched filtering using “integrate & dump” circuitry
Exploring complex and exponential functions Lab 5
Complex numbers
Complex functions
Exponential functions

Fourier series analysis Lab 6

Constructing waveforms from sine & cosine
Computing Fourier coefficients
Build a manually swept spectrum analyzer

Spectrum analysis of various signal types Lab 7

Spectrum of the impulse train
Spectrum of the filtered impulse train
Duty cycle & sampling
‘’Sinc’’ pulse train
Spectrum of PN sequences
Analog noise generation (AWGN)
Non-linear processes

Time domain analysis of an RC circuit Lab 8

Step response of the RC network
Impulse response of the RC network
Exponential pulse response of the RC network
Synthesizing transfer functions

Poles and zeros in the Laplace domain Lab 9

System with feedback only – allpole
Feedback and feedforward – poles & zeros
Viewing frequency response using FFT

Sampling and Aliasing Lab 10

Through the time domain – PAM, Sample & Hold
Through the frequency domain
Aliasing and the Nyquist rate

Getting started with analog-digital conversion Lab 11

PCM encoding & quantization
PCM decoding & reconstruction
The recovered analog signal & reconstruction
TDM: Time division multiplexing
Discrete-time structures: Finite Impulse (FIR) filters Lab 12
Notch filter creation using two-delay FIR
Using notch filter to eliminate interference
REAL WORLD FILTER DESIGN EXERCISE: Implementing
a higher order FIR BPF using online software design tools

Poles and zeros in the z plane with IIR systems Lab 13

Relating roots to coefficients in the quadratic polynomial
IIR without feedforward – second order resonator
IIR with feedforward – second order filters
Viewing spectrum with broadband noise input
Dynamically varying the poles & zeros

Practical aspects of discrete time structures Lab 14

Dynamic range at internal nodes
Transposed Direct form 2 IIR
Implementations with high sampling rates

Digital filter design from specification Lab 15

Reviewing a LPF design
Designing a BPF to meet a specification

Exploring digital filter structures Lab 16

Execution time per order
Saturation of internal stages
Using simulation to assist design decisions
DF1 and DF2 Transpose structures

Appendix A: ETT101-40 Lab to Textbook chapter table

Lathi.B.P. , “Signal processing & Linear Systems”, Oxford University Press

Oppenheim.A.V.,Wilsky.A.S., “Signals & Systems”, Prentice Hall, 2nd edition

Haykin, Van Veen, “Signals and Systems”, Wiley, 2nd edition

Ziemer.R.E,Tranter.W.H, Fannin.D.R, “Signals & Systems: Continuous and Discrete”, Prentice

Hall, 4th edition

Boulet.B.: “Fundamentals of Signals & Systems”, Thomson/Delmar Learning

McClellan.J.H, Schafer.R.W, Yoder.M.A, “DSP First”, Prentice Hall

References
 Electrostatic Discharge Sensitive Product

CAUTION: The inputs/outputs of this product can be damaged if subjected to

Electrostatic Discharge (ESD).

To prevent damage, industry-standard ESD prevention measures must be

employed during installation, maintenance, and operation.

v
Introduction
This experiment manual is designed to provide a practical “hands-on”, experiential, lab-based
component to the theoretical work presented in lectures on the topics typically covered in
introductory “Signals and Systems” courses for engineering students.

Whilst it is predominantly focused on all electrical engineering students, this material is not
exclusively for electrical engineers. With an understanding of differential equations, algebra of
complex numbers and basic circuit theory, engineering students in general can reinforce their
understanding of these important foundational principles through practical laboratory course work
where they see the “math come alive” in real circuit based signals. This provides a foundation for
further study of communications, control, and systems engineering in general.

Students take responsibility for the construction of the experiments and in so doing learn from their
mistakes and consolidate their knowledge of the underpinning theory, which at times is particularly
abstract and hard to grasp for these early engineering students. They are not constrained by the
software and need to be systematic in debugging their own systems when results do not meet their
expectations.

The common reaction of early students when confronted with “complex analysis” is one of confusion
and regression to “rote-learning” in order to survive the examination process. This manual has as its
predominant aim to create real, “hands-on” implementation of the theory, in such a way that the
student can directly articulate and connect the mathematical abstractions with real world
implementations. It is a journey of personal discovery where the motto is “why is it so ?”

The use of “modeling” is the fundamental tool in this, and other EMONA TIMS equipment (such as the
TIMS-301/304C), and it has been shown that experimenting with scaled models of real world systems
allows students early-on to get a tangible “feel” for principles that they may later utilize in real world
commercial workplace environments. As well, students tend to “believe” results from “real hardware”
rather than from software simulations, and this supports their “learning by doing”.

The authors sincerely hope that students using this equipment and guided by this manual will complete
with a sense that complex numbers and systems analysis “makes sense” and is somehow more “real” and
applicable to real world problems. In this way they may successfully use these principles in solutions to
future problems they will encounter.

Carlo Manfredini
Sydney
Experiment 1 – Special signals: characteristics and
applications

Achievements in this experiment

Time domain responses are discovered: step and impulse responses as paradigms for the
characterization of system inertia; sinewaves were used as probe signals; clipping was applied to
the recovery of a digital signal.

Preliminary discussion
Bandwidth is a term that has been in the engineering vocabulary for many decades. Its
usage has extended over time, especially in the context of digital systems. It has become
commonplace now to mean information transfer rate, and all Internet users know that
broadband stands for fast, and better. There are highly competitive markets demanding
top performance – ever higher speed whilst maintaining a low probability of corruption.
However, as speed is increased, obstacles emerge in the form of noise, interference and
signal distortion. At the destination these limitations become digital errors, resulting in
pixellated images, and audio breaking up.

Engineers involved in the design of these systems must assess the suitability of numerous
components and sub-units e.g. adequate speed of response ?, too noisy, distorted ? They will
need to benchmark the behaviour of subsystem. The procedures that are used for modelling and
testing must be universally accepted.

The most important consideration affecting the speed of a digital signal is the switching
process to produce a change of state. The switching time can never be instantaneous in a
physical system because of energy storage in electronic circuitry, cabling and connecting
hardware. This energy lingers in stray capacitance and inductance that cannot be completely
eliminated in wiring and in electronic components. The effect is just like inertia in a mechanical
system.

A universal procedure is needed to characterize, measure and specify ‘inertia’. Various

paradigms have become established over many years of application. One of these is the step
response. For this reason, the step function has become one of the special signals in systems
engineering.
There are other signal types of importance. The sinusoid or sinewave heads the list of the
range of applications. There are many others, including the impulse function, ramps, pseudo
noise waveforms and pseudorandom sequences, chirp signals.
This Lab has its focus on signals that are most needed for basic operations. Other signals will
be introduced progressively in succeeding labs.

Figure 1: step, impulse and sinusoid signals

The experiment

In Part 1 we investigate how signals are distorted when a system's response is affected by
inertia, and discover signals that are useful for probing a system's behaviour.
In Part 2 we introduce the sinewave, and observe how the systems investigated in Part 1
respond to inputs of this kind.
Signals that have been subjected to amplitude limiting, also known as clipping, are commonly
encountered when excessive amplification is used, such as in audio systems, resulting in overload
distortion. In Part 3 we generate clipped signals and examine a useful application of clipping.

As this experiment is a process of discovery, we will name the blocks which represent the
channel “ System Under Investigation”, S.U.I, until we have familiarized ourselves with their
actual characteristics.

It should take you about 45 minutes to complete this experiment.

Pre-requisites

Familiarization with the ETT-101 conventions and general module usage. A brief review of the
operation of the SEQUENCE GENERATOR module. No theory required.
Equipment

PC with LabVIEW Runtime Engine software supplied by EMONA (LabVIEW RTE2016).

ETT-101 unit

Two BNC – 2mm leads & assorted patch leads

External oscilloscope or built-in digital scope (ETT-101C)

Figure 2: ETT-101-40 control panel screenshot with tab Lab 1 selected

Procedure

Part A – Setting up the ETT-101 and software control panel

-Plug in the ETT-101-40 Signals & Systems module into the Expansion slot with the power OFF
-Powerup the ETT-101/C
-Turn on the PC (if not on already) and wait for it to fully boot up.
-Launch the ETT-101-40 SigSys software control panel and confirm ‘USB OK’ message
-Select the required tab on the ETT-101-40 SigSys software control panel

Note: To stop the ETT-101 VI when you’ve finished the experiment, it’s preferable to use
the STOP button on the ETT-101 control panel itself rather than the LabVIEW window
STOP button at the top of the window. This will allow the program to conduct an orderly
shutdown and close any USB connections which have been opened.
Part 1a – Pulse sequence speed throttled by inertia

In this set of exercises we continue the digital theme introduced above and explore the
behaviour of signals in transit through a channel that has a limited speed of switching.

SEQUENCE
S.U.I.
SOURCE

Figure 3: block diagram of the setup for observing the effect of

a system (SUI) on a digital pulse sequence.

Figure 4: patching diagram for Fig 1a

11. Patch up the model in Figure 1b. The settings required are as follows:

VCO: FREQUENCY=1000 Hz (LO range);

SEQUENCE GENERATOR: DIP switch to UP:UP for a short sequence.
TUNEABLE LPF: FREQ to maximum; GAIN to midway (unity gain)
SCOPE: Timebase 500us; EXTERNAL TRIGGER: Rising edge trigger; Trigger level=1V

Set up the CHA scope lead to display the LINE CODE output of the SEQUENCE GENERATOR

2. Measure the smallest interval between consecutive transitions . Compare this with the
duration of one period of the clock by moving the scope lead to view the SEQUENCE
GENERATOR CLK input from the VCO.
Question 1
What is the minimum interval of the SEQUENCE GENERATOR data ?

We could think of these sequences as streams of logic levels in a digital machine, possibly
representing digitized speech or video. The information elements in this stream are the unit
pulses. They are sometimes called symbols. Verify that there is one symbol per clock period.
Since the clock frequency is 1000 Hz, the symbol rate is 1000 per second. The symbols in this
sequence have only two possible values, so they are called binary symbols, and the transmission
rate is commonly expressed as bits/sec.

3. Connect the CHA lead to the

output of the BASEBAND LPF module
(BBLPF) and connect the CHB lead to
the output of the TUNEABLE LPF
module (TLPF).

Set the TLPF FREQ so the output

appears similar to that from the
BASEBAND LPF, as shown in Figure 1c.

Aim to make the risetime similar. Figure 5: example signals

After this point, DO NOT change the TLPF controls for the rest of this experiment. As it
is tuneable it must be left at this current setting for consistent comparisons.

Note the presence of oscillations on both signals and the differences between them. Where
possible you should venture comments. You are not expected to have any prior knowledge of
these waveforms.

Question 2
Describe the signal transitions for both outputs:

4. With the clock remaining unchanged on 1000 Hz measure the time for each signal to
change state. Is it the same for low to high (amplitude) as for high to low ? Specify the
reference points you are using on the amplitude range, eg 1% to 99%, 10% to 90%. Note these
values in the table below. “Freeze” the signals using the “RUN/STOP” switch on the oscilloscope
in order to take your measurements, and use the TRIGGER SLOPE control to select between
rising and falling edge capture. Change the timebase to improve measurement accuracy.
NOTE: Disconnect the RC LPF when measuring the other systems as it loads the output LINE
CODE signal slightly and affects the measurements.

TIP: Calculate the levels you wish to measure and use the X & Y cursors as guidelines.

Table 1: transition times for sequence data

Range BBLPF@1kHz TLPF@1kHz BBLPF@1.5kHz TLPF@1.5kHz

(%) (us) (us) (us) (us)

10-90 rising

10-90 falling

1-99 rising

1-99 falling

5. Next, increase the clock frequency to around 1.5 kHz. Repeat the measurements in Task
14 above, and compare the two sets of results.

6. Progressively increase the clock frequency, and carefully observe the effect on the
output waveform. Note that something significant occurs above 2 kHz. Confirm that below
2 kHz the original transitions can be unambiguously discerned at the channel output, even
though they are not sharp. Describe your observations as the clock is taken to 3 kHz and above.
Are you able to correctly identify the symbols of the original sequence from the distorted
output waveform? Estimate the highest clock frequency for which this is possible. Venture an
explanation for the disappearance of transitions in this channel.

Question 3
Describe the signal transitions for both outputs:

In the next segment we will closely examine the shape of the transition corresponding to an
isolated step excitation.
Part 1b – isolated step excitation of a system

STEP S.U.I.
SOURCE

Figure 6: block diagram of step excitation arrangement

Figure 7: patching diagram for step response of SUI

7. Connect signals as shown in Figure 1d above. Connect CHA to the BBLPF output and CHB
to the TLPF output, and view both signal on the scope. Settings are as follows:

8. VCO: FREQUENCY=approx. 1kHz (LO

range);
DIVIDER: Dips set to :[DOWN:DOWN] (/8)
SCOPE: Timebase 1ms; Rising edge trigger on
EXT TRIG; Trigger level=1V

Confirm that the scope time base is set to

display not more than two transitions. Use
RUN/STOP to freeze scope display.
Figure 8: example signals

Adjust the TLPF GAIN and FREQ to give a similar risetime, as shown in Fig 8.
Observe the channel's response to a single transition (you can use scope trigger and other time
base controls to display a LO to HI transition or a HI to LO transition). Confirm that the shape
of the output transition is similar to the shapes you observed in Task 13 above.

When the response to a step excitation is isolated in this way, so that there is no overlap
with the responses of neighbouring transitions, it is known as the step response.

Note the presence of oscillations and the relatively long settling time to the final value
(sometimes known as ringing -- a term that goes back to the days of manual telegraphy and
Morse code). Compare with the waveform in Task 3 .

Note that some of the transitions observed in Task 3 occur before the previous transition
response has completely settled.

The risetime of the step response is an indicator of the time taken to traverse the transition
range. Various definitions can be found according to the application context. The frequently
used 90% criterion is suggested as a convenient choice for this lab.

9. Measure and compare the risetime of the three step responses. Use this to estimate
the maximum number of transitions per second that could be accommodated in each case (ignore
the effect of the oscillations). Compare this with the results in Task 3

Table 2: transition times for step input

Range BBLPF TLPF RCLPF

(%) (us) (us) (us)

10-90 rising

10-90 falling
Graph 1: step response waveforms
Part 1c – isolated pulse response of a system

An isolated pulse can also be used as an alternative to the use of an isolated step as the
excitation to “probe” the behaviour of the system. The SYNC pulse from the SEQUENCE
GENERATOR serves as source of this signal.

PULSE
S.U.I.
SOURCE

Figure 9: block diagram of pulse response investigation

Figure 10: patching diagram for pulse response investigation

10. Change patching to use the SYNC pulse from the SEQUENCE GENERATOR as shown the
Leave the patching as per the previous section, with the pulse output connected to both S.U.I.
With the frequency of the VCO set to approx.. 1000 Hz, progressively increase the frequency
while viewing the SYNC pulse, which becomes progressively narrower. Start with SYNC pulse
width of 1 ms. Reduce it gradually to 0.9, 0.8, 0.7, 0.6 ms.

When you reach 0.5 ms, move in more slowly and observe the effect on the pulse width. As you
continue to reduce the input impulse width, the flat top between transitions gets shorter, and
ultimately disappears. You are aiming to eliminate the flat top and have the signal reach a
consistent shape. Since the rising transition is not able to reach its final value (the flat top), it
is not surprising that the amplitude of the pulse gets smaller.
Question 4
Describe what happens when you reach 0.4 and 0.2ms ?

11. Are you able to determine the ‘demarcation’ pulse width -- i.e. after which the response
shape remains unchanging and any “flat top” is no longer visible ?
Record the pulse width for all SUI’s in the table below.

Table 3: pulse response readings

BBLPF TLPF RCLPF

“demarcation” pulse width

(us)

Previous step response

(from Table 2)

% of step response
(pulse/step)

Period of oscillations (us)

12. Using the oscilloscope cursor to measure the pulse width, which is stable as you have
triggered the scope on it, to get accurate measurements.

13. Express this as a percentage of the step response risetime, using the values from the
previous section on step response, and note these values in the table above.

Reflect on this for a moment, i.e. the response shape remaining apparently independent of
the input pulse width -- this is an interesting discovery.

14. Move the scope leads so as to view the input pulse as CHA and one of the SUI outputs
on CHB. Note that for the both there are oscillations. The presence of these oscillations
provides an opportunity for additional observations of shape changes as the width of the input
pulse is reduced. There are many ways of testing this, eg. the number of sidelobes, their
relative amplitudes, the intervals between zero crossings.

15. For each SUI, set the pulse width to the “demarcation” value and measure the period of
the oscillations following the pulse. Note these in the table above.
You have demonstrated that, provided the time span of the excitation signal is sufficiently
concentrated, the shape of the response pulse is entirely determined by the characteristics of
the system. We could think of this as the striking of a bell, or tuning fork, or of the steel
wheel of a train to detect a crack. The system is hit with a short sharp burst of energy.

INSIGHT: The response shape is not affected by the input signal.

The energy burst used as input is called an impulse. The resulting response is called the impulse
response. An impulse function is a mathematical construct derived from a physical pulse. The
idea is straightforward. The pulse width is reduced to an infinitesimal value while maintaining
the energy constant. Naturally this implies a very large amplitude. The impulse function plays a
central role as one of the fundamental signals in systems theory, with numerous ramifications.

In the above exploration we discovered practical conditions that make it possible to generate a
system's natural response or characteristic, i.e. a response that is not affected by the exact
shape of the input excitation. Concurrently we have discovered a path to the definition of the
impulse function and a vital bridge to link this mathematical abstraction to the world of physical
signals.

16. With the setup unchanged, measure the delay at the peak of the output pulse and
compare this with the delay of the step response measured earlier.

17. Return to your records of the step responses obtained previously. For each case, carry
out a graphical differentiation with respect to time (approximate sketches are sufficient,
however take care to achieve a good time alignment to identify key features). Compare these
results with the records obtained . As a useful adjunct exercise, consider a slightly modified
step function in which the transition is a ramp with a finite gradient, though still quite steep.
Carry out the differentiation with respect to time on this function, and compare with the above.
Record your conclusion.
 Graph 2: differentiations of step response waveforms

You have demonstrated that, provided the time span of the excitation signal is sufficiently
concentrated, the shape of the response pulse is entirely determined by the characteristics of
the system. We could think of this as the striking of a bell, or tuning fork, or of the steel
wheel of a train to detect a crack. The system is hit with a short sharp burst of energy. The
response shape is not affected by the input signal.

The energy burst used as input is called an

impulse. The resulting response is called the
impulse response. An impulse function is a
mathematical construct derived from a
physical pulse. The idea is straightforward.
The pulse width is reduced to an infinitesimally
narrow value while maintaining the energy
constant. Naturally this implies a very large
amplitude. The impulse function plays a
central role as one of the fundamental signals
in systems theory, with numerous
ramifications.
In the above exploration we discovered practical conditions that make it possible to generate a
system's natural response or characteristic, i.e. a response that is not affected by the exact
shape of the input excitation. Concurrently we have discovered a path to the definition of the
impulse function and a vital bridge to link this mathematical abstraction to the world of physical
signals.

In other words, imagine that you had an arbitrarily shaped input signal to your system. You could
decompose that signal into a series of impulses corresponding to each point in that signal. eg a
series of thin samples. If you passed each one of those samples through your S.U.I, one at a
time, and then added them all up (correctly spaced in time), you would have the response of the
SUI to that complete arbitrary signal.

Since you already know the response of the SUI to a single impulse (ie: its impulse response),
you could construct the SUI’s response to any more complex signal through this process of
addition. This process, known as ‘’superposition theorem’’ states that the response of the system
to the string of impulses is just the sum of the response to the individual impulses.

You will be exploring ‘’superposition theorem’’ further in Experiment 3 – Unravelling Convolution

but for now you are familiar with the insights provided by the humble impulse response.
Part 2 – Sinewave input

As mentioned in the introduction, sinewaves are encountered in a large number of applications.

The special role of the sinusoidal waveshape for system characterization is explored in
Experiment 2, and further developed in Experiment 6. In this segment we just get our toes
wet. We carry out some basic observations and compare the sinewave response of the various
S.U.I’s with the impulse response obtained above.

S.U.I.

Figure 11: block diagram of setup for sinewave investigation

Figure 12: patching model for sinewave investigation

18. Connect the VCO SINE output to the inputs of both S.U.I.

Set VCO GAIN to maximum and VCO FREQ to minimum. Then adjust VDC anticlockwise until
the VCO output frequency is just below 300. You have now setup the VCO range to start from
below 300 Hz and using the VCO FREQ control you can adjust to up until beyond 10kHz in a
single action.

SCOPE: Connect CHA of the scope to the SINE output of the VCO, and CHB to output of S.U.I.
Trigger the scope on the SINE output.
Progressively increase the frequency from 300 Hz to 10 kHz, by only adjusting the FREQ
control (leave the others as they are setup) and observe the effect on the amplitude of the
output signal. You can assume the input amplitude will remain constant. Make a record of your
findings in the form of a table of amplitude vs frequency. Once all readings are completed,
enter your results into the table on the TAB1 control panel, which will plot those results.
Consider the possible advantage of using log scales.
To enable a “log” Y axis, STOP the SnS control program, right click the plot graph, select Y
scale > Mapping > Log and then RUN again. To return to Linear, repeat this process and select
“Linear”.

Table 4: amplitude vs frequency readings

Frequency (Hz) BBLPF (Vpp) TLPF(Vpp) RCLPF(vpp)

300

600

1000

2000

2500

3000

4000

6000

10,000

19. Refer to the results you obtained and sketched of the step response in Question 19.
Notice the similarity of the step response shape to a half cycle of a sinewave. Estimate the
frequency of the matching sinewave. Examine the graph obtained in the above task and see
whether any feature worth noting appears near this frequency.

Question 5
What frequency would a matching sinewave have ?
Question 6
Describe what happens to the frequency response plotted on the control panel tab at this
frequency ?

20. Return to the observations you recorded in Task 19. A physical mechanism was proposed
there to explain the reduction in pulse response amplitude as the width of the input pulse was
progressively made smaller. Consider whether the reduction in output amplitude of the sinewave
with increasing frequency could be explained through a parallel argument.

Question 7
What was the mechanism described earlier ?
Tutorial questions (solutions not provided)

Q1 The impulse function was described in Part 1. Explain why the step function is a
better alternative in a practical context. Show how the impulse response can
be obtained from the step response. Is this indirect procedure for measuring
the impulse response theoretically equivalent, or does it involve an
approximation?

Q2 Consider a system with step response rise-time of 4 s. What information does this
provide about the impulse response?

Q3 a. Consider the waveform at the yellow X output of the SEQUENCE GENERATOR (as
in Part 1). Suppose the p-p voltage is 3.9 Volt and the clock is 2 kHz. What is
the average power into a 1 Ohm load?

b. Suppose the waveform is passed through BASEBAND LOW PASS FILTER and
the p-p output amplitude is also 3.9 Volt. Is the power greater or less than
at the channel input? State the reasoning (hint: consider the waveform
shape required to have the average power exceed that of the waveform at
the channel input).

c. Consider two different sequences as above. One has N transitions per period,
the other has N + 4. Explain why the number of transitions does not affect
the average power for the signal format at the channel input. Is the
answer the same at the output? If no, in which case will the average power
be greater? Indicate why. Hint: math not required, just consider how the
average is worked out.

Q4 A 60 kHz sinewave is applied at one input of a MULTIPLIER, and a 59 kHz sinewave

at the other input. The amplitudes are both 2 Volt p-p. Use a suitable
formula to show that the MULTIPLIER output is the sum of two sinewaves.
Calculate their respective frequencies. The MULTIPLIER output is fed to a
system similar to BASEBAND LOWPASS FILTERS, with step response rise-
time 300 s. Describe the signal at the output of this system, if any.
Lab 2 – Systems – Linear & Nonlinear

Achievements in this experiment

Understanding of the special role of sinusoids in linear systems. Knowledge and experience of
testing for linearity.

Preliminary discussion

Lab 1 was about special signals. The aspect that was special was their use for getting
characteristic information about systems. However, although the focus of that lab was on
signals, a system was nevertheless needed as a vehicle for the action. In this lab the focus will
be on the system, and as might be expected, it is signals that will be doing the talking.
Moreover, it will emerge that the story in Lab 1 was incomplete. We will discover a signal set
that is perfectly matched to a special class of systems.

We get started with explorations of various linear and nonlinear systems, and look into memory
effects.
Next, tests for linearity are implemented.
The lab concludes with a ‘lateral’ investigation of frequency response.

Pre-requisite work

Question 1
Write down a formula to express the square of a sinusoid in terms of a double angle argument.

Question 2
What is the meaning of differential linearity ?

Consider the two conditions for linearity for a system S defined by y = S(x)
Scaling or homogeneity: a.y = S(a.x)
Additivity: y1 + y2 = S(x1 + x2)

Question 3
How would you apply these formulas in testing systems for linearity in this Lab ? How many
replicas of the system are needed for the additivity test ?
Equipment

PC with LabVIEW Runtime Engine software supplied by EMONA (LabVIEW RTE2016).

ETT-101 unit with ETT-101-40 plug-in module

Two BNC – 2mm leads & assorted patch leads

External oscilloscope or built-in digital scope (ETT-101C)

Figure 1: ETT101-40 control panel SFP screenshot with tab Lab 2 selected

Procedure

Part A – Setting up the ETT-101 and software control panel

-Plug in the ETT-101-40 Signals & Systems module into the Expansion slot with the power OFF
-Powerup the ETT-101/C
-Turn on the PC (if not on already) and wait for it to fully boot up.
-Launch the ETT-101-40 SigSys software control panel and confirm ‘USB OK’ message
-Select the required tab on the ETT-101-40 SigSys software control panel

Note: To stop the ETT-101 VI when you’ve finished the experiment, it’s preferable to use the
STOP button on the ETT-101 control panel itself rather than the LabVIEW window STOP
button at the top of the window. This will allow the program to conduct an orderly shutdown and
close any USB connections which have been opened.
Experiment

Part 1: Getting started with linearity

Figure 2: patching diagram for a “rectifying” system

1. Connect the SINE terminal of the VCO via the BUFFER AMPLIFIER to the RECTIFIER
block input. Connect CHA to this input. Connect CHB to the RECTIFIER output, as per Figure 3.

The settings required are as follows:

VCO: FREQUENCY=1000; AMPLITUDE after BUFFER AMP: 1 V pp;
SCOPE: Timebase 500us; Rising edge trigger on CHA; Trigger level=0V

2. Adjust the amplitude of the SINE waveform from 1 Vpp to around 6 Vpp. Take various
readings throughout this range and document in the table 1 above. As the output signal appears
only to be positive going, measure its positive amplitude (which is also its total peak to peak
value). Show that this system satisfies the homogeneity test over part of the input range (i.e.
where the range excludes the zero input point).

After entering the readings into the Table, enter them into the Lab 2 tab table in the SnS
control panel to see an XY plot of the relationship.
 Table 1

Input amplitude RECTIFIER amplitude

(Vpp) (Vpp)

Question 4
Does this system (RECTIFIER) satisfy the scaling test for linearity? Show your reasoning.

In the above, consider the meaning and applications of the term ‘differential linearity’ (also
known as 'incremental linearity').

For comparison, you will now repeat the above with an analog MULTIPLIER. Observe the effect
of changing the AMPLITUDE. Record your findings concerning linearity.

After entering the readings into the Table, enter them into the Lab 2 tab table in the SnS
control panel to see an XY plot of the relationship.
3. Connect the SINE terminal of the VCO via the BUFFER AMPLIFIER to the
MULTIPLIER inputs. Connect both inputs in parallel. Connect CH0 to this input. Connect CH1 to
the MULTIPLIER output, as per Figure 4.
The settings required are as follows:

VCO: FREQUENCY=1000; AMPLITUDE: 1 V pp; SINE output selected

SCOPE: Timebase 1ms/div; Rising edge trigger on CHA; Trigger level=0V

Figure 3: patching diagram for a “multiplying” system

 Table 2

Input amplitude MULTIPLIER amplitude

(Vpp) (Vpp)

Question 5
Does this system (MULTIPLIER) satisfy the scaling test for linearity? Show your reasoning.

Note that the application of the additivity test is not needed when the homogeneity test has
failed. A complication with the additivity test is the need for replicas of the system under test,
hence the homogeneity test is normally applied first.

Unlike the BBLPF channels, investigated in Lab 1, the responses of the above systems appear as
effectively instantaneous, i.e. the systems are ‘memoryless’ over the time scales of interest
above. However, closer examination with the scope time base in the nanosecond ranges reveals
that the responses are not as ‘instantaneous’ as might seem. Thus the idea of a ‘memoryless’
system is only relative in a practical context.
Part 2 – the VCO as a system

In this part we explore the output-input characteristics of the VCO (voltage controlled
oscillator) and apply linearity tests.

Figure 4: patching diagram for investigation of VCO as a system

4. Connect as shown in Figure 5.

Settings are as follows:

VDC: set to -2.0V
VCO: adjust FREQ=2000 Hz when VDC set to -2.0V.
GAIN set to midpoint (although this is arbitrary it is adequate);
SCOPE: Timebase 1ms/div; Rising edge trigger on sinewave output; Trigger level=0V

5. Display both the input DC voltage and the output sinusoid. Ensure that the scope is
triggering on the sinusoid signal for a stable display.

6. Observe the effect of varying the DC input voltage. Start with the VDC set to minimum.
Document output frequency vs input DC voltage in the table below. (The scope can be used to
measure the DC voltage and frequency).
 Table 3

Input DC voltage VCO output frequency

(V) (Hz)

7. Consider whether the VCO is a linear system, i.e. in terms of output frequency vs input
DC voltage. For this purpose should you focus on incremental frequency rather than absolute
frequency ? (cf differential linearity). Plotting the readings will help.

After entering the readings into the Table, enter them into the Lab 2 tab table in the SnS
control panel to see an XY plot of the relationship.

Question 6
Is the VCO a linear system ? Explain your reasoning.
 Figure 5: patching diagram for investigation of VCO as a system
with rapidly “varying” input level

Settings are as follows:

VDC: set to minimum (fully anticlockwise)
VCO: set to HI range; adjust FREQ=100kHz when VCO INPUT is not connected
VCO: GAIN set to midpoint (although this is arbitrary it is adequate);
SCOPE: Timebase 100us/div; Rising edge trigger on sinewave output; Trigger level=0V

8. Rewire the experiment as per Fig.6. Leave the VCO INPUT of the VCO unconnected and
adjust the VCO FREQ to give 100kHz output.

Connect the MULTIPLIER’s output to the VCO INPUT of the VCO. This effectively gives you
amplitude control over the fixed frequency sinewave by using the DC voltage as the
multiplication factor. Observe the effect of varying the DC input voltage on the MULTIPLIER
output, as well as the VCO output, over the entire range of the VDC.

9. Observing the VCO output and noticing how it varies in relation to the input voltage.
Consider possible applications (hint: synonym for ‘varying’ is ‘modulating’). View the input and
output of the VCO with both scope leads. Trigger the scope on the input signal for a stable
display.

A note about scope triggering:

As a general rule, it is best to trigger the scope on the lowest frequency signal being displayed.
This causes the scope to display the slower signal with stability, and have the faster signal
moving relative to that. In this case, although the two signals displayed are not synchronised, it
is easy to see the influence of the lower frequency signal on the VCO output. Try both ways and
see if you agree.
Question 7
What applications could the VCO with varying output frequency be used for ?
Part 3 – a feedback system

We will apply both linearity tests to a simple feedback system implemented with the LAPLACE
module. First we carry out some basic exploration of the operation of the INTEGRATOR.

10. Patch up the model in Figure below.

Settings are as follows:

Select Lab 2 tab on the Signals & Systems software control panel
Load the AWG with the preset signals.
ARB 2: Frequency = 1000 Hz; Square wave amplitude: 4 V pp; Trim the ARB 2 offset to 0 V
SCOPE: Timebase 500us; Rising edge trigger on CHA; Trigger level=0V

Figure 6: patching model for sawtooth generation

After patching and viewing input and output signals from the INTEGRATOR, you most likely will
find a significant DC offset on the output. Any small DC offset at the input will cause an
integrator to drift in the dircetion of that offset.

To compensate for this, trim the DC offset of the ARB 2 output such that the INTEGRATOR’s
output is at 0V. In the example shown below, an offset of -5 was applied to ARB 2.
We are not concerned about DC offsets in this experment. Rather we are looking at rates of
change in the signals.

Remember to LOAD AWG each time you make any changes to the settings.
 Figure 7: “offset ARB2“control in Lab 2 tab

Figure 8: scope display for sawtooth generation, after offset trimming.

11. Measure the p-p amplitude of the output sawtooth waveform. Calculate the slope (slew
rate) of the sawtooth and use this to write a formula for the INTEGRATOR output in terms of
the input.

Question 8
What is the formula for the INTEGRATOR output ?

Next we apply negative feedback to the “open loop” integrator above.

12. Patch up the modules according to the figures below to add a feedback path to the
system.

Settings are as follows:

Select Lab 2 tab on the Signals & Systems software control panel
Load the AWG with the preset signals.
ARB 2: Frequency = 1000 Hz; Square wave amplitude: 4 V pp; Trim the ARB 2 offset to 0 V
ADDER GAINS: Set A & B to midway (gain = -1)
BUFFER : set GAIN to -1
SCOPE: Timebase 500us; Rising edge trigger on CHA; Trigger level=0V

Figure 9: block diagram for a feedback system

Figure 10: patching diagram for a feedback system

To start with, set the gain at the BUFFER to approx -1 by using the 2kHz SINE signal. Do not
adjust it after doing this.

NOTE: the ADDER has negative gain on each input as does the BUFFER.
For this experiment, you are after negative gain in the feedback path, but positive gain
everywhere else.

By placing the BUFFER before the ADDER input, you will have a net positive gain on that path,
while the B input path will remain a negative gain path.

View the output and input of the INTEGRATOR as CHA and CHB respectively.
Your display should look like Figure 11 though with different levels depending on your settings.

Figure 11: example signals

Observe the displayed waveforms of Fig 12 and satisfy yourself that they represent step
responses (briefly display the input waveform as a reminder). Note that these responses are
made up of exponential segments.

Examine the waveform at the ADDER output (input to the INTEGRATOR), and confirm that it
looks like the impulse response corresponding to the step response at the INTEGRATOR output.

13. On the scope, measure the time taken for the exponential curve to decay by 50% and
use this to calculate the time constant.

14. Refer to your preparation work on tutorial Q1c and compare the theoretical step
responses with those observed above.
Question 9
Use the value of the b2 gain, and INTEGRATOR constant you measured above to determine the
time constant of the exponential responses. Compare this with the value obtained from your
measurement.

15. Observe the effect of varying the b2 gain, and check that this is in general agreement
with the predictions of the theoretical formulas. In particular, take a close look at b2 values of
0, -0.1, -0.2, ... and confirm that you can see effect addition (or subtraction) of the output from
the input, as you subtly introduce feedback.

16. Explain why this feedback system is not memoryless in the time scale of our
observations. Determine the location of the system memory (see Tutorial Q1d).

17. Change the input amplitude and confirm that the system satisfies the homogeneity
condition for linearity. Switch the FUNCTION GENERATOR back to SINEWAVE output to
demonstrate that the result is the same with a sinusoidal input.

Question 10
Write a differential equation for this first-order feedback system. Assume initial conditions
are zero. Show that with a sinusoidal function of time as input, the output is also sinusoidal.
Show that this also happens when the input is a complex exponential. Which special property of
complex exponential functions provides the key?
Part 5 – frequency response

In Lab 1 (Part 2) we found that when a sinewave of constant amplitude was applied to the input
of a BASEBAND LPF (BBLPF) channel the output amplitude was dependent on the frequency. A
sinewave was chosen for that test because we were making a comparison with the shape of the
impulse response of the BBLPF, which was very similar to a sinusoidal segment.

The idea of frequency response using sinewaves is well known in the audio culture, for example,
as used in the testing of amplifiers, microphones, etc. . Here we ask: why have sinewaves been
singled out for this -- could another kind of waveshape be used to measure frequency response,
for example, a squarewave? There is a quick way to find out.

S.U.I. S.U.I.

Figure 12:block diagram for investigation

with a squarewave and a sinewave

The settings required are as follows:

VCO: LO range; FREQUENCY set to minimum; GAIN not used;
TUNEABLE LPF: Gain set to midway (approx. = 1). Fc set to maximum
SCOPE: Timebase 500us; Rising edge trigger on CHA; Trigger level=1V

18. Measure the frequency response of a different filter, the TUNEABLE LPF with a square
wave. Use the setup shown above. Use the square wave available at the VCO DIGITAL output.
Explain your criterion for the amplitude measurement at the output (specifically, to deal with
the shape changes and ringing oscillations). Enter your readings into the table below.
 Input frequency Output using Square Output using Sine
(Hz) wave input (Vpp) wave input (Vpp)

1000

2000

3000

4000

5000

7000

9000

11000

13000

15000

Question 11
Are you able to use the square wave for this test ?

19. Next, try with a sinewave. The measurement is very straightforward. Unlike the
squarewave, the sinewave emerges at the output with no apparent change of shape. The issue of
amplitude measurement criterion does not arise. Enter your readings into the same table.

Enter the readings into the software control panel for Lab 2 to quickly see a plot of the
relationships between input and output for both.
 Figure 13:example of use of onscreen plotter for two relationship pairs

Consider the possibility of other waveforms that may also have this property. Relate this to the
math in Tutorial Q 1.

And now, a role for sinusoids in quantifying nonlinearity.

In the world of the discerning audiophile, speed of response is of paramount importance, and we
have seen that this can be quantified in terms of a time domain response, and also through the
medium of the sinewave frequency response. An equally essential facet of sound quality is
linearity. The presence of nonlinearities in amplifiers, microphones, etc generates distortion and
diminishes the listening experience. Hence it is important to have paradigms for quantifying
deviation from linearity.
It is interesting that the sinusoidal waveform is also a basis for nonlinearity modelling. Think
about how this might be implemented. However, you will find it more rewarding to reconsider
this after discovering the fertile valleys of the frequency domain, in Labs 8 & 9.

The systems used so far have not included a discrete-time example. This is done in later
experiments.
Tutorial questions (solutions not provided)
Q1 a. Write a differential equation for the first-order feedback system in Part 4.
Assume initial conditions are zero. Show that with a sinusoidal function of
time as input, the output is also sinusoidal. Show that this also happens when
the input is a complex exponential. Which special property of complex
exponential functions provides the key?

b. For each of the two possible outputs, write expressions for the ratio of output to
input when the input is exp(jt). Show how this result can be used to obtain
a formula for the amplitude response when the input is the real signal
sin(t).

c. Modify the differential equation in (a) to include nonzero initial conditions. Make
the input zero and solve for the natural response (i.e., the homogeneous
solution). Use this to obtain the step responses at each output. Apply the
findings in Lab 1 to convert the step responses to impulse responses.

d. Find out how to make an integrator with an operational amplifier.

e. Suppose that in (b) j is replaced by the general complex variable s =  + j.

Suggest an interpretation for s in a practical context.

f. Consult your favourite reference to see how the Laplace transform is applied to
solve linear differential equations with constant coefficients.

Q2 What is the difference between a halfwave rectifier and a fullwave rectifier ? Show
how to make a full-wave rectifier with two half-wave rectifiers.

Q3 Consider a multiplier with 1.5V DC to input y. Is this system linear wrt input x?
Repeat this with a 2 Volt p-p sinewave at input y. Consider this test with
other waveforms at input y (for example a squarewave, a rectified sinewave).

Next, apply a 2 kHz sinewave to the two inputs connected in parallel. Write an
equation for the output in terms of the input. Is the scaling (homogeneity)
test for linearity met? Show that the output waveform is sinusoidal. What
is its frequency? Suggest an application for this system ?

Q4 From the Lab, should we conclude that nonlinear systems are generally memoryless?
Suggest examples to the contrary. Hint: a speed control governor in a steam
engine, analogous method for amplitude compression control in audio systems.

Q5 Consider a cascade of two of the first-order systems in Q 1 . The INTEGRATOR

constants are 1 and 2 (in seconds), the feedback gains are 1 and 2,
respectively. The input is sin(t). Write down the output/input amplitude
ratio of the overall system. Plot this vs frequency (use convenient numerical
values for the constants).

Q6 Consider again the two first-order systems in Q 5, however this time connect them
in parallel, with output 2 subtracted from output 1. Repeat Q 5 for this case.
Compare the two configurations.

Repeat with the outputs added and compare with the above.

Q7 Consider the sawtooth waveform generated with the integrator. Draw a block
diagram showing how this waveform could be used with a VCO and a scope in
x-y mode to produce a rudimentary frequency response display.
Experiment 3 – Unraveling Convolution

Achievements in this experiment

Carry out a step-by-step dissection of the convolution process in a discrete-time system. Use
this to discover the convolution formula. Demonstrate that convolution can be visualized as a
running average of successive values of the input signal. Observe a special property applying
to sinewaves. Demonstrate the operation of a filter in the time domain.

Preliminary discussion

For many students, the first encounter with convolution is an abstract mathematical formula in
a textbook. This lab offers a more evocative experience. By tracing the passage of some basic
signals through a simple linear system, you will be able to observe the underlying process in
action, and, with a little arithmetic, discover a formula as it emerges from the hardware.

Refresh your basic trigonometry: you will need sin(.t) + [Link](.t + ) expressed as a single
sinusoid.

If any of the modules is unfamiliar, spend a little time with the ETT-101 User Manual. This will
give you a head start in setting up the lab.

We patch up a delay line with two unit delays and three taps with independently adjustable gains
as in Figure 1. In the first exercise you will set these gains to given values, and observe the
output when the input is a single pulse (more exactly, a periodic sequence of single pulses). This
is an important preliminary as it introduces the unit pulse response.

+
OUTPUT

b0 b1 b2

INPUT
UNIT UNIT
DELAY DELAY

Figure 1: general block diagram of the system

Next, you will observe the output when a pair of adjacent pulses is used as input. This near
trivial example provides us with a springboard to the general case. It will demonstrate how
convolution operates as an overlapping and superposition of unit pulse responses. From the
dictionary, superposition : the action of placing one thing on or above another, especially so
that they coincide. A second more general input sequence is then used to reveal a deeper
insight and to provide a vehicle for setting up the key formulas.
In the remainder of the lab we use sinewaves as inputs. This takes us to the rediscovery of the
special role of sinusoids in linear time-invariant systems.

In the final exercise we carry out and unravel a disappearing act.

It should take you about 40 minutes to complete this experiment.

Pre-requisites:

Familiarization with the ETT-101-40 conventions and general module usage. No theory required.