There’s a beautifully cheap way to control AC power: don’t convert it to DC at all, just chop a piece out of every half-cycle. A single-phase AC voltage regulator does exactly that with two thyristors wired inverse-parallel — one for the positive half-cycle, one for the negative — and the only control knob is when, within each half-cycle, you let them start conducting. I tested one on a 240 V RMS supply, sweeping the firing angle from 0° to 180°.
The idea is the firing angle, α. Fire the thyristor early (small α) and it conducts for almost the whole half-cycle — high output. Delay it (large α) and it conducts only the tail end — low output. You’re not scaling the voltage, you’re deleting the front of each half-wave and keeping the rest. With a purely resistive load that’s clean enough: voltage and current stay in phase, both just shrink as α grows, and output power falls with them.
But “delete the front of every half-cycle” is the same thing as “introduce sharp edges into a sine wave,” and sharp edges are harmonics. The more I delayed firing, the more chopped the waveform got and the higher its total harmonic distortion climbed. The power control is real, but it isn’t free — you’re trading a clean sinusoid for a controllable but dirty one, and that distortion is current your supply and transformer carry without it doing any useful work.
An inductive load — a motor, which is the realistic case — makes it stranger. The current lags the voltage, and because an inductor stores energy, current keeps flowing after the supply voltage has already crossed zero. So the thyristor doesn’t turn off when you’d expect; conduction stretches past the zero-crossing, the waveform shape changes again, and the active power delivered drops relative to what the voltage alone would suggest. The load isn’t a spectator to the regulator — it rewrites the regulator’s behavior.
So phase-angle control is a bargain with a fine-print clause. It’s cheap, direct, and needs no DC link — which is why it runs lamp dimmers, heaters, fan speeds, and soft-starters — but every degree of control you buy by delaying the firing is paid for in harmonic distortion, and an inductive load changes the exchange rate. “It controls the power” was never the whole question. “What does it do to the waveform, on this load” is the rest of it.