Electronic Systems Thinking (and Circuits)

 

Electronic Systems Thinking (and Circuits)

IIT Madras decided to launch the Bachelor of Science in Electronic Systems(BS-ES), in 2023, following a fairly successful launch of BS in Data Science (DS) in 2020. This is a Web-Enabled programme where students access lecture material remotely but appear for exams and evaluations in person at exam centers across the country, which helps control the quality of the degree. I am excited to be part of such ambitious thoughts and giant leaps that an educational institute looks to make every now and then. The BS-DS programme had enough evidence to suggest that students can learn theory quite well. What about the labs? Can labs be conducted remotely but evaluated in person? Can electronics be taught in such a mode? While there are numerous such challenges to surmount for an ambitious idea like this to succeed, they also present opportunities to try things out differently. A chance to experiment. 

The Curriculum:

The first opportunity was the design of the curriculum itself. We were free to deviate from the traditional B.Tech in electronics curriculum.  Considering the nature of the proposed BS degree where students could drop off at various points to earn a foundational certificate, a diploma, or a full-fledged BS degree it was important that students were exposed to theory and practice in equal proportion all throughout. Conventional B.Tech degrees focus on theory initially and gradually expose students to practice. A couple of my colleagues and I came up with a sequence of courses to address this shortcoming. Did we just reorder the conventional B.Tech curriculum and pull some of the labs up front? No. We went back to the drawing board and asked a few fundamental questions. (a) Does a B.Tech student understand the need for theory courses in a stream, like EE, that is so application-oriented? (b) Does a student relate the theory being taught to the functioning of gadgets that are part of our daily lives? The answer to both questions happened to be a resounding NO and that was the first problem we wanted to address. My colleague Prof. Andrew Thangaraj was very clear that we needed a course in the very first semester that excited students about electronics and allowed them to appreciate the functioning of these gadgets to first order. The course was called Electronic Systems Thinking (EST) and I was entrusted with the job of designing the contents and learning objectives. 

Figure: A typical electronic system

 

Learning Objectives:

Preparing the material to teach a conventional course for the first time is quite a task in itself but at least we have an idea of what it should look like. We have either been taught the course earlier or we have some example course to follow. What do we do in the case of a course like EST which, to the best of our knowledge, doesn't exist anywhere? The only precedence to such a course was Computational Thinking which was introduced with a similar objective to serve the needs of the data science degree. The idea of motivating the theory using real-life examples was common and I must acknowledge it definitely helped to get a basic idea of what EST should look like. The course would revolve around the sense-compute-communicate block of a typical electronic system. That said, the EST course was very different in that we had to enable students to relate to various gadgets we use in daily life. Personally, I had already taken enough gadgets apart and put some of them back together as well and therefore had enough experience doing gadget teardowns. That was not the hard part. The question was deeper. What do I demonstrate for 12 weeks? Should the course be a gadget-teardown demonstration? What are the learning objectives? In other words, what do I expect the students to be able to do after completing this course? Disassemble and assemble their own gadgets? Definitely not! What then? 

Prerequisites:

The other important aspect to consider is what the students of the EST course are equipped to handle at their point of entry? They are class XII students who have passed mathematics and physics. The burning question was, therefore, how much electronics can one appreciate with limited knowledge of mathematics and physics? Concepts like the Fourier Transform and Laplace Transform are assumed to be the foundation that allows one to appreciate most electronics but that requires a course in itself to introduce to students for the first time. Teaching transforms in such a course also defeats the purpose of designing a course like EST as the students are forced to learn a whole lot of theory without appreciating the need for it. This one realization gave me much-needed clarity:- I wasn't allowed to foray into the frequency domain and had to stick to the time domain entirely. Coming back to the gadget-teardown and experimental demos, I had to be conscious about not trivializing the need for theory or handwaving complex ideas. The idea was to therefore demonstrate experiments that can be analyzed in the time domain, appreciated with first-order analysis, but motivate the need for theory in future courses that allow you to perform higher-order analysis. In some sense, first-order analysis is sufficient to reverse engineer an already-designed gadget but much more is needed to design one from scratch. Our students need to be able to do both!

How much of electronics can one understand with just time-domain analysis? Turns out it's quite a bit.  

Initial weeks:

The course begins with an iPhone-5s teardown where the objective is to allow students to appreciate the miniaturization of modern electronics and observe that the battery in its current form decides the size of the mobile. We move on to study the battery, charger, and even a kettle and discuss various aspects of voltage, current, resistance, power, and energy. This leads us to the idea of circuits, Ohm's law, and Kirchoff's current & voltage laws. The theory is motivated by experimental observations using demos on a breadboard. The sinusoidal waveform is the heart of electrical engineering and deserves a dedicated experiment! A computer fan, used in the heatsink, is operated as a generator by manually rotating the fan and observing the waveform on the leads of the motor! The sinusoid is evident on the oscilloscope and is motivated by Faaday's law of electromagnetic induction which students do in fair detail in class XII physics. Based on student feedback we decided to introduce the idea of "problem solving" every four weeks where we take a couple of problems and solve them with a detailed discussion about how and why we approached the problem in that particular method. The whole course is run as a discussion between two electrical engineers.

Analog Computation:

Week 5 begins with a video where the legendary (late) S. P. Balasubramaniam is performing in a live concert. A sound mixer is at the heart of any live performance and primarily allows volume balancing, frequency equalization, and sound mastering. The subsequent weeks focus on the first-order implementation of these features. The application motivates a simple resistive analog mixing circuit, whose solution is obtained using KCL, KVL, or superposition theorem. The simplicity is evident from the circuit shown here and this is exactly what governs the principle of analog addition in an actual mixer.

Figure: Analog resistive mixer

The next topic is frequency equalization which is typically analyzed in the frequency domain. The challenge was to stick to the time domain and this was achieved by solving the first-order differential equation of an RC low-pass filter. The key idea of filtering is demonstrated in practice on a breadboard and also matched to the solution of the differential equation. One of the key goals of this course is to tell students a circuit can be analyzed from first principles in the time domain which will exactly match the frequency domain analysis they will perform later in future courses. Simple low-pass, high-pass, and band-pass filters are studied in great detail. 

Digital Computation and Communication:

All computation until now was performed in the analog domain. However, any system today will inevitably switch to digital computation/ storage and the digital computer forms an essential part of any computing chain today. An experiment that highlights analog vs digital computation is one where two mics are used to capture the same voice signal (the lecture itself). One is fed directly to an oscilloscope, emulating an analog oscilloscope, and the other through a sound card to a computer. The wav data on the computer is captured by software like Audacity. The idea of digital storage is apparent from being able to save the file and the discussion that ensues is centered around the key ideas that enable storage:- amplitude and time quantization. Eventually, the exact size of a wave file is estimated using the sampling rate and resolution of each sample. A teardown of a laptop fits nicely into the discussion allowing us to discuss various components of the digital computer like memory, motherboard, processor, and even heat sink. The final part of the course focuses on communication, both wired and wireless. The crux of amplitude modulation is conveyed using elementary trigonometry. 

What next?

The course is available as a YouTube playlist. It was quite a revelation for me that so much of electronics can be understood with elementary mathematics and physics. From my experience in the VLSI industry and academia, the concepts of KCL, KVL, Ohm's law, and charge sharing between capacitors are fundamental and help to understand nearly any circuit. The learning objectives therefore revolve around solving various problems in these topics! The "(and Circuits)" was appended to the title as the course eventually helps gain expertise to solve problems in circuits. The intermediate feedback about the course was very good and encouraging but we await the end-of-course feedback. Assuming the feedback is as good, what next? Can this course be taught and brought into the regular B.Tech curriculum? That's exactly the advantage of the online mode of lecture delivery. It allows us to shoot and edit a video of all experiments, which hardly works as expected in the first attempt! Live demonstrations in a 50-minute lecture are not suited for such complex experiments. That said, in this era of artificial intelligence and data science, core disciplines like electrical engineering need to put in the extra yard to excite students about the stream, and the EST course may be one way to do it. What about the challenges of offering such a course in a live class? Well, they need to be converted into opportunities.

Acknowledgments: I sincerely thank my colleagues Prof. Aniruddhan and Prof. Boby George for their valuable inputs, and discussions, and for being part of the course. I would like to thank the lab in charge Mr. Anand and Mr. Udaykumar of the EE department at IITM for helping with all the demos. Finally, I would like to thank Prof. Andrew, Prof. GV, and Dr. Gowri Suryanarayana for their valuable inputs at various times.