>
curriculum/Electronic Circuit Analysis
section 9 of 102 min read

9. Things to Try

  1. Build the simplest CE amplifier with a 2N3904, voltage-divider bias, RC=4.7R_C = 4.7 kΩ, VCC=12V_{CC} = 12 V. Measure midband gain. Sweep input frequency from 10 Hz to 1 MHz, plot the magnitude response. You should see the low cutoff (set by coupling caps and emitter-bypass cap) and the high cutoff (set by Miller multiplication of CμC_\mu).

  2. Wrap negative feedback around it. Add a 1:10 feedback divider from output back to base via a coupling cap. Re-measure the gain, which should be much lower (closer to 10, set by 1/feedback ratio) but with much wider bandwidth. Verify that the gain-bandwidth product is approximately conserved.

  3. Build a Wien bridge oscillator with an op-amp and an incandescent flashlight bulb (the bulb's positive temperature coefficient stabilizes the amplitude, the lovely trick from Bill Hewlett's HP 200A). The output is a beautiful, low-distortion sine wave. Measure the frequency. Change the R or C and re-measure.

  4. Build a class-AB push-pull with a 2N3904 (NPN) and a 2N3906 (PNP). Drive it with a signal generator, observe the output on a scope. Notice how it crosses zero smoothly (no crossover distortion) when biased correctly with a couple of small-signal diodes between the bases, but distorts ugly if you remove the bias diodes.

  5. Try LTspice simulation of a simple cascode and a simple CE, both with the same transistor. Compare bandwidths. The cascode should be at least 5 to 10 times wider for the same low-frequency gain. Also compare their phase responses; the cascode keeps cleaner phase out to higher frequency.

  6. Build a single-tuned RF amplifier at 10 MHz with a slug-tuned inductor and an air-variable cap. Measure the bandwidth, then deliberately reduce the load resistance (lower the Q) and see the bandwidth widen. Verify BW=f0/QBW = f_0/Q.

  7. Measure the Miller effect directly. Build a CE amp without a cascode; sweep the source impedance (use a variable resistor in series with the input) from 100 Ω to 100 kΩ. Plot the high-frequency cutoff versus source impedance. The cutoff should fall as source impedance rises, by roughly 1/R1/R, because the Miller-multiplied CμC_\mu forms an RC with the source. This is the most direct experimental confirmation of Miller's theorem.

When the above feel intuitive, you are through this chapter.