01/15/03 Class (and a little beyond) - Comments and Outline:

• Comments on Syllabus. Books at Yale Book Store. Check that they contain a CD-ROM.

• Will email out a list of auxiliary reference books, which are in our Becton Library, some on reserve for the course.

• Syllabus is linked at <classes.yale.edu>, a simple course Web page will evolve at
http://www.eng.yale.edu/ee-labs/morse/courses/ee325/ . This is part of The Morse Labs Web site with which you should be familiar. Email will get used frequently, check it at least once a day.

• Although we will cover only roughly 7 chapters worth in Sedra & Smith, plus portions of others, it is an excellent book for future reference. Still, it's gotten too expensive. If you want to try to save on textbooks, try http://isbn.nu/ or http://www.bookpool.com/. Our book is ISBN 0-19-511663-1, and has an associated Web site http://www.sedrasmith.org/.

• Much of chapt.1 is an overview of fundamentals, ought to correspond roughly to previous exposure, though nomenclature may have differed. For Wednesday 1/22/03, read pp 9-28, familiarize yourself with what's to be found in other parts of chapt.1. As you probably know by now, a lot of reading of technical material is done in "layers." There will be no class this Friday 1/17, because of the Martin Luther King holiday re-shuffle. At the end of the semester we’ll have a Monday class to make up for it.

• Much of chapter 2 is about the ideal op amps that this semester will strive to assemble. We could strictly follow the book, doing 'ideal' op amps first. Or we could skip chapt.2 and come back to it, op amp finally in hand. The former is a bit too idealistic; the latter too much delayed gratification. We'll take a parallel track where we are aware of what we can do with op amps as per chapt.2 while we work on assembling them. Thus, start scanning over chapt.2 and acquaint yourself with its structure.

• Why study analog electronics?

• all digital circuitry is analog (this becomes important at high speed and at low power)
• circuit configurations teach a kind of visual thinking about relationships between components.
• This is the level of 'circuit shapes' (O'Dell)
• Getting good at the simultaneous awareness of simple relationships, 'rule daemons' (e.g. Ohm's Law) This is a good skill for navigating any multi-parameter design space.
• Learning the art of 'tradeoffs', an essential part of engineering, is very explicit in analog circuit design. Those tradeoffs include the important dimension of cost.

• Review the "semiconductor food chain", and what makes it possible:

• components are cheap, chip-to-chip interconnects are expensive
• functionality of each 'layer' is relatively independent of others - modularity
• reserve performance of the lower layers allows modularity. (Lower layer limits only become a factor in performance extremes.)
• Approximations work because of modularity. For the designer this implies more local optimization, less stress.

• What about design?

• very much my viewpoint, as you can see from my background at
  http://www.eng.yale.edu/ee-labs/morse/about/pjk.html
• synthesis of higher constructs from lower modules
• envisioning these possibilities by familiarity with approximate relationships, therefore being able to hold those relationships in your head, not just SPICE (the generic "core" program of what we will use--Electronic Workbench). SPICE doesn't generate ideas, i.e. the 'circuit shapes'.
• Examples of useful approximation methods are
• hierarchy of device parameters
• limiting cases
• Another part of design are certain structuring principles which we'll encounter throughout the semester, though again I'm more conditioned to think that way than Sedra & Smith. Among these are
• biasing (lets unipolar devices handle bipolar signals)
• superposition (a fundamental benefit of linear systems)
• symmetry (selectively eliminates device shortcomings, the cornerstone of analog integrated circuits)
• equilibrium (of processes, e.g. charge conservation)
• In an information-rich world you can get away without ideas of your own for prolonged periods. Engineering has always been a very cumulative process. WWW and the searchable electronic literature and bibliographic databases (Compendex+, INSPEC) push even further this opportunity to adopt and adapt. But be honest about it.
• Design is a 'linguistic' process, you learn a 'vocabulary' and then learn to 'write' with it. In that sense our component and circuit vocabularies are practice in a larger process of integrative thinking, especially if you stay aware of the design methods. Even laudable books like Sedra & Smith are sometimes more preoccupied with individual circuit attributes, and push method into the background.