Signal Processing - July 2016 - 85

the students to the signals they will be
working with during the activity.
This article describes our experiences with using these types of experiments
and the lessons learned that increase
their effectiveness.

Samples of application-oriented
activities
When dealing with application-oriented
activities, there is a tradeoff that has to be
made between realism and complexity. It
is difficult to teach the theory from realistic signals such as quadrature-amplitudemodulated signals or speech signals
because their spectra are complicated
and/or difficult to model mathematically.
However, if students never see such realistic signals, they will never understand
the purpose of the theory they are
learning. In contrast, square waves,
impulses, and other such simplistic signals are easy to model mathematically
but don't always capture the variation
that is seen in real-world signals. To fully
understand a system, such as a cell phone
or heart monitor, requires a high level of
understanding of concepts such as Fourier transforms and series, system linearity
and time invariance, wireless communications, electromagnetics, system frequency response, multipath, and many
other concepts. To introduce such systems all at once would be overwhelming
for a novice student.
When we speak of applicationoriented activities, we are referring to
using realistic signals operating on simplified systems that are placed into the
context of these more complex systems.
For example, students may explore the
modulation of a speech signal, which in
and of itself is not a cell phone. However, the simplified experiment provides students with the opportunity to
manipulate the signals and systems in
real time and observe the effects in both
the frequency and time domains. Then
context can be provided about how
modulation of speech signals relates to
how a cell phone works. It is up to the
instructor to control the students' exposure to appropriate levels of detail. We
have taken two different approaches to
designing application-oriented activities for the students. In one approach

a single general-purpose and configusensor signals such as load cells and therrable platform, the Signals Exploration
mocouples. All signals are relative to a
Board (SEB), was designed along with a
common ground. These signals can then
set of weekly lab activities. In the other
be modulated using v y (t) , have noise or
approach, several platforms (boards)
other signals added to them using v z (t) ,
were used to illustrate the concepts.
sampled, and filtered. The output of the
The SEB was designed to be easy to
system can be observed on the oscilset up, flexible enough to relate to many
loscope and/or spectrum analyzer and
types of applications, and robust enough
heard through a speaker or headphones.
to handle several years of student use.
The applications that can be explored
As shown in Figure 1, the signal path
include amplitude modulation (AM) and
of the SEB is composed of four stages:
frequency shifting, signal-to-noise ratio,
input, sampling, filtering, and output.
harmonic distortion, music and speech
The input stage produces the signal
signal processing, ECG signal processing, and many others.
v 1 (t) = v x (t) v y (t) + v z (t) , where v x (t)
The SEB has been used in a weekly
is one of several different input signal
three-hour lab during a ten-week quarchoices that are selected with a jumper;
ter in the introductory CTSS course
v y (t) can be chosen with a switch to be
for the last seven years. Lab activities
either 1 V, a dc voltage between −5 V
have been designed for each week of
and 5 V set by a potentiometer, or a timethe quarter to give students experience
varying signal; and v z (t) can be chosen
with many course concepts that are
with a switch to be either 0 V or a timemore difficult to understand. The activivarying signal. The signal v 1 (t) can be
ties have gone through a lot of evolution
sampled with various forms of samduring this period based on lessons that
pling such as pseudo-impulse, pulse,
were learned. In their
or zero-order-hold
most recent format,
(ZOH) sampling or
When dealing with
each activity follows
the sampling can
application-oriented
a similar sequence.
be bypassed. The
activities, there is a
Students start by
filtering stage has
tradeoff that has to be
working with a simthree paths that can be
made between realism
ple signal such as an
selected with another
impulse train, progjumper: a bypass with
and complexity.
ress to a slightly more
no filtering, a prewired
complicated signal such as a sum of four
operational amplifier for a first-order filter,
harmonically related sinusoids, and end
or three prewired operational amplifiers
with the application-oriented signal such
for multifeedback filters that can impleas a speech signal.
ment up to a sixth-order filter. Students
One of the weekly lab activities begins
create the filter characteristics by insertwith students exploring what happens to
ing the appropriate passive devices into
periodic signals as they go through a linsockets on the board. The output stages
ear-time-invariant (LTI) system and ends
contain a low-power driver that can drive
with students measuring their own ECG
the 50- X input impedance of a spectrum
signal. Students use a first-order active
analyzer and a high-powered driver for 8low-pass filter with a variety of signals.
X speakers and headphones.
They start with a pseudo-impulse train,
The SEB can be easily configured with
filter a square wave, and then measure
jumpers and switches to facilitate a wide
and filter their ECG signal. In each of the
range of activities. As shown in Figure 1,
stages, students are asked to complete a
the inputs header allows students to select
more open-ended activity such as to figbetween an applied time-varying voltage
ure out how to adjust the input signal so
signal such as from a function generator
that the filter changes the fifth harmonic
(signal), the output of an on board microby −3 dB. Students are asked to turn in
phone (MIC), their measured electrocara report with screen captures showing
diography (ECG) signal, or the output
signals in both the time and frequency
of an onboard instrumentation amplifier
domains and write a paragraph that
(inst amp) that can measure differential
IEEE Signal Processing Magazine

July 2016

Table of Contents for the Digital Edition of Signal Processing - July 2016