Margarita D. Takach and Robert G. Heeren
Department of Electrical Engineering
Seattle University
Our teaching of laboratories has traditionally consisted of unconnected experiments. Many were verification oriented of pre-set designs; in the past students often failed to connect empirical data back to analytical predictions and more recently to computer simulations. In addition we observed that students were not developing as strong an intuitive feeling for frequency domain as they were for time domain in these introductory laboratories.
It has been our experience that the lack of an over arching goal in our entry level laboratory program resulted in students not having the motivation and interest to seriously investigate on an individual basis the collection of experiments set before them. With the above in mind, a set of design oriented experiments is being developed for our introductory circuits and electronics laboratory program. These entry level junior courses constitute the laboratory studies segment of the EE Core Curriculum required of electrical engineering majors at Seattle University.
The new experiments are centered around the use and design of operational amplifiers. Our central theme-operations amplifiers: external behavior, internal design-is the connective thread that ties both laboratories together. The experiments also incorporate the study of spectrum analysis to build an intuitive understanding of frequency domain concepts. Modeling is also central to the students' thought process since SPICE and/or MATLAB simulations are compared closely to both observed circuit behavior and analytical predictions.
The diverse nature of EE students at Seattle University has forced us in somewhat different directions in laboratory instruction from our style of the recent past. Historically, our students had significant pre-college technical experience with electronic components via hobbies, military training, or job related skills. Recently, however, a major shift seems to have occurred. Many of our students come to the study of EE with a weakly developed ``technological world view''. This paper presents some of our experiences teaching technically naive students their first laboratory courses in electrical engineering. Our only assumption is they have encountered the ideal operational amplifier in their sophomore electrical circuits course along with standard passive circuit elements. We have found a design oriented sequence a significant and absorbing challenge for the kinds of students our department is attracting. These challenges seem to be of a different character than those our previous instructional practice faced.
The first course, Electrical Circuits Laboratory, uses standard passive circuit components as well as the traditional 741 operational amplifier IC chip. External behavior, not internal chip design, is the paramount experience here. Mastery of manual and computer read test instrumentation at an operation level is a significant achievement in this first laboratory course. The Tektronix Fourier Analyzer is the major test instrument that is used to develop their frequency domain intuition.
Our second course, Electronic Circuits Laboratory, explores internal design of the operational amplifier as a means of ``hooking the interest'' of our students in laboratory electronics. Their curiosity has already been peaked through their experiences with 741 op-amp circuits, so the promise of being able to design and test their own ``home made'' operational amplifier allows us to investigate multi-stage designs as well as traditional amplifier configurations which have historically proven useful in electronics. Particular emphasis is placed upon designing all amplifiers ``from scratch'' as opposed to just verifying circuit performance of pre-packaged ICs. Circuits with current source active loads, which have found repeated use in IC op-amp designs, are considered as well. The culminating experience here is the design of their own ``home made'' operational amplifier by quarter's end.
All laboratory experiences are assigned over a one week period during a nine or ten week quarter. Students begin the experiment during scheduled lab time and subsequently make extensive use of open access to the lab facilities to complete their work. Each of the six student test benches in our laboratory room is equipped with a digital/analog oscilloscope, a digital multimeter, a 3 output power supply, a function generator, and a 486 personal computer. We also have two Tektronix 2642 200 KHz Fourier Analyzers and a Tektronix plotting curve tracer each mounted in roll-a-way carts. One Philips digital RCL meter is available in the room to provide precise component values.
The PC serves several purposes: it allows the scope screen to be printed via a software program, it is the platform for SPICE and MATLAB simulation and EXCEL data base software, and is the external controller for the Fourier Analyzer. All computers are connected to a single HP laser printer in the room through a spooler. Each PC also has 3 video games installed which students are free to use as a means of venting their frustrations when they encounter difficulty in getting their circuit design to function properly, or to just take a short break from design/testing work.
Students are required to buy their own breadboards which are conveniently available through our IEEE Club, and they are also encouraged to purchase their own DMM. We require them to keep a bound design notebook in both laboratories. We expect them to glue scope printouts and simulation outputs on the notebooks' pages. Instead of formal written reports, students instead learn to keep a ``live'' lab journal.
The experiments in the circuits laboratory were chosen with the following goals in mind: using modern test instrumentation, studying both the time and frequency response of passive and op-amp circuits, examining spectral content and harmonic distortion of filtered signals, and making the connection between the passive circuits laboratory and the active electronics laboratory via our thematic approach with operational amplifiers.
Our underlying philosophy is simple: we believe that when accurate comparisons between analytical predictions (equivalent circuits), empirical results, and SPICE/MATLAB simulations are achieved the students can then truly claim they have mastered the circuit under study!
Merely for descriptive purposes, we have clustered our experiments together in similar groups.
Cluster A: Time Domain Test Instruments (2 sessions)-use of power supplies, function generators, DMMs and printing oscilloscopes. DC and AC measurements. Design, construction, testing, and simulation of a first order passive low/high pass RC filter.
Cluster B: Frequency Domain Test Instrumentation (2 sessions)-use of the Fourier Analyzer for spectral display of various signals. The low/high pass filters designed in the previous lab are used to explore the spectrum of a square wave before and after filtering. The Fourier analyzer is then used to measure the frequency response of the passive filters using both the Fourier technique and the sweep frequency method. A passive second order resonant filter is designed, built, tested, and simulated. Specific harmonics of a square wave are then filtered. SPICE and MATLAB simulations are compared to spectrum and network measurements.
Cluster C: Single Stage Op-Amp Circuits (3 sessions)-Students design, build, and test non-inverting, inverting and integrating op-amp circuits using the classic 741 op-amp. Saturation, unity-gain frequency, and slew rates are measured. Gain-BW tradeoff is observed with the Fourier Analyzer. Total harmonic distortion is measured using the Fourier Analyzer. SPICE simulations are compared with measurements. Next students learn, design, build and test active low/high pass filters employing the 741 op-amp.
Cluster D: Multiple Stage Op-Amp Circuits (3 sessions)-Students build, test, and simulate an active band pass filter made up of cascaded active low/high pass filters built in the previous experiments. The culminating experience of the quarter is achieved with a cascaded class project, the active equalization filter. Each octave of this active filter is designed by separate student teams and then cascaded into a multiple octave filter the last day of the quarter! Boost and cut are illustrated via a loud speaker. Time domain and frequency domain abstractions are thus tied back to the more intuitively accessible audio domain!
The theme in this lab, internal design of operational amplifiers, was chosen to provide a clear motivation for the study of traditional transistor amplifier configurations, and to experience multi-stage designs that have proven historically useful in operational amplifier design. Students will have designed and tested an elementary ``home made'' operational amplifier by the end of this second laboratory course. We believe this theme provides a natural way of connecting all experiments in the two quarter sequence into a cohesive whole giving the students a sense of fulfillment at the end of the total experience.
Students design and build different variations of the common emitter and emitter follower configurations in the first cluster of experiments. These experiments introduce the students to some of the fundamental concepts necessary for the design of operational amplifiers: cascading, choice of coupling between stages, use of current sources for biasing and active loads, and input/output impedance. By the seventh session students are designing the first stage of their operational amplifier, the emitter coupled pair. In the remaining sessions, students build two additional stages. This elementary three-stage operational amplifier is then tested in various simple configurations such as inverting, non-inverting, invertor-adder op-amp circuits, and a positive feedback oscillator to confirm that it is capable of ``doing operations''.
Most designs are accomplished by the use of load lines since this technique encourages visualization of DC bias considerations, thereby also insuring operational familiarity with the printing curve tracer. We generally use the 2N2222 npn BJT, both in discrete and quad IC formats, since this unit is readily available and thus reduces costs with our rather high burn out rate. Teams of two students design, construct, test, and simulate their own circuit in each experiment and document their results individually in their own design notebook.
Cluster A: Common Emitter Amplifier Designs (3 sessions)-CE amplifiers with passive and active loads, AC and DC coupling, and cascaded amplifiers. Exploration of non-linear distortion and clipping with over driven inputs and changeable bias point. Dependance of low frequency response upon the coupling is explored manually by scanning the signal frequency and automatically with the Fourier Analyzer. Gain and input/output impedance measurements are compared to small signal equivalent circuit calculations and SPICE simulations.
Cluster B: BJT Current Sources (1 or 2 sessions): Current mirrors. Four versions with different arrangements of positive/negative power supplies are designed and individually displayed on the curve tracer. Output impedance is estimated.
Cluster C: Emitter Follower Amplifier Designs (2 sessions)-AC coupled and current source biased EF amplifiers. Level setting. Gain and input/output impedance are measured and compared to predictions and simulations.
Cluster D: Operational Amplifiers (3 sessions)-Differential amplifiers (DA), common-mode-rejection-ratio measured and compared to predictions and DA simulations, emitter coupled pair, DA with active load and current source biasing, cascaded DAs, DC and AC coupling, double inputs and single output, emitter follower output stage. Simple three stage operational amplifier is built and tested. To verify their design can do ``operations'', students connect it in four different configurations: inverting, non-inverting, summer op-amp circuits, and a Wien-bridge oscillator. With these feedback configurations students quickly realize that stability is an issue requiring special attention when designing commercial op-amps. Students use very low amplitude input signals to keep the op-amp circuit in a stable state.
We have taught the circuits laboratory twice during the current year in this format and the electronics laboratory four times over the last two years. Having test instrumentation on the design bench right next to a computer has definitely increased the students' interest in comparing experimental results to paper designs and computer simulations. Challenging our students to do their own design of an operational amplifier circuit has significantly increased their motivation and interest in doing careful weekly design experiments leading up to the last project!
Without careful attention to modeling, SPICE outputs can easily differ from empirical results as demonstrated in Figures 1 &2. Accurate modeling of the wiring resistances in the RLC resonant filter dramatically improves the comparison of Fourier Analyzer measurements with a SPICE simulation. The schematic for the culminating experiment-the ``home made'', AC coupled, 3 stage operational amplifier is shown in Figure 3. Each team may produce variations on this schematic, DC versus AC coupled, and emitter resistors versus current mirrors.
In general, we believe these two laboratory oriented courses are giving students an intuitive understanding of passive and active circuit behavior in both time and frequency domains. Our students have had considerable success in making accurate comparisons between model predictions, computer simulations, and bench measurements. Current seniors, who were the first groups to pass through these new labs last year, appear to have stronger lab skills and to be more highly motivated in laboratory oriented investigations than we have experienced the last several years.
We plan to develop a laboratory manual based upon these experiences that should be ready for the 1995-96 academic year.
