With so much analog-circuit design being done by automated tools and circuit-performance analysis done with tools such as Spice, it may seem anachronistic to actually go through a basic circuit in detail, in order to understand what each component does, and what the tradeoffs associated with component selection may be. It's all too easy, though, for circuit-designers (especially aspiring ones) to be so far removed from the down-in-the-dirt circuit design that they lack the needed understanding of the basics. Yet it’s the grasp of these basics which translate into solid designs with fewer unpleasant surprises.
The dominant role which ICs have doesn’t help here, either, for two reasons. First, by their nature, ICs conceal their inner workings. Second, there are many valuable design tricks and techniques which are viable only with ICs, but which don’t make sense for discrete circuits. Examples are the use of matched components on the die to cancel out some errors and temperature-related drifts, or the use of a large number of “free” components such as extra transistors and resistors to achieve the desired performance, options which are not possible or attractive in a discrete circuit.
Nonetheless, there are alternatives for those who want to learn about the reality of basic analog-circuit design. A good place to start, not surprisingly, is with the venerable 555 timer—but not as an IC. Check out this all-discrete version of the 555 from Evil Mad Scientist, Figure 1 , which they cleverly call a “dis-integrated circuit. Students and aspiring analog designers can build it, then use it to try many of the thousands of 555-based circuits available on the web or in books, while observing the internal waveforms and effects of changing the external component values and arrangements. Since it is all relatively low speed, a low-cost scope or even DVM can be used. (The same company also makes an all-discrete version of the 741 op amp.)
Another “learning tool” is a basic, all-analog Lightning Detector circuit (not, you don't use it by setting it up at spots it will likely strike ground, so there's no danger). Nothing is critical about this four-transistor circuit: use of alternate parts, component tolerance, layout, packaging, or supply. Yet it allows the user to observe how a circuit reacts to a stimulus—and if natural lighting is not available, you can simulate it by placing one of those low-cost piezoelectric igniters near the antenna or using a sparking circuit such as unsnubbed relay contacts.
Of course, observing a circuit's waveforms and behavior is one thing, but at some point you have to do some mathematical analysis of currents, voltages, and topology to become a better engineer. That's why a recent “Power House” application note from Texas Instruments really caught my attention. In Power Tips: How to design a robust series linear regulator with discrete components, author Manjing Xie walks you step-by step through the design of this relatively simple analog circuit.
She shows you how to start with a basic one-transistor circuit, Figure 2 (yes, we still need our discrete devices) and then add what you need to make the regulator work properly. To do this, you have to anticipate shortcomings and then provide enhancements which overcome them. You also have to do basic real-world analysis, such as the power dissipation of the transistor if the output is shorted (here, it's 2.4 W, which is not good; so you have to add a current-limiting resistor).
This basic, discrete-component analog regulator circuit is a teaching tool, a learning tool, and even a potential testing tool: if you have a job candidate who claims some analog expertise, talking through this circuit will show how much he or she actually knows of basics, before getting to the more difficult higher-frequency issues such as parasitics.
There really are a lot of good options for learning basic, realistic analog circuits and analysis. Are there teaching tools, circuits, and projects do you use or recommend?