Key Takeaways
- Gain from 10 to 100 → from stable to violent oscillation
- Root cause: LC double pole creates 180° phase shift; high gain makes loop gain > 1
- Dilemma: high gain = good accuracy, but = instability — compensation resolves it
Lesson 1
Too Much Gain Causes Instability: Where Oscillation Comes From
Raising the error amplifier gain from 10 to 100 makes the output stop regulating — it oscillates violently.
Ch4 concluded: higher gain can further reduce error, but requires compensation for stability. Let's verify directly: change the error amp (VCVS) gain from -10 to -100, change nothing else — see what happens.
Result: the output no longer holds 6V — it oscillates violently between ~3V and ~9V. Why? The LC filter has a double pole at 504Hz, creating ~180° phase lag near crossover. If loop gain still exceeds 0dB at that frequency, the Barkhausen oscillation criterion is met — positive feedback replaces negative.
This is the core tradeoff in all feedback design: higher gain → smaller steady-state error (good), but → easier to oscillate (bad). Solution? Don't simply reduce gain — we need to 'fix' the phase while keeping high gain. That's what compensation networks do.
High-Gain Oscillation: Gain=100
Error amp gain=-100, Vref=0.495V. LC filter phase lag + high gain = oscillation.
After reading this section, run the simulation and observe the waveforms. To explore further, open the example in a standalone page.
Watch Items
- out waveform: no longer smooth DC, large oscillation (~3V-9V)
- pwm waveform: duty cycle jumping wildly between 0 and 100%
- Compare with Ch4's stable waveform: everything else identical, only gain changed