Close but not exactly. It is for regulation.
In an audio amp when you crack it the rails drop and there is very little v oltage across the outputs at the peaks of current. Now turn it down to say ten watts. That is a little over one amp going into eight ohms.
If you have 60 volts, low output voltage but at one amp then you conceivabl y might dissipate 60 watts - making a little over a watt output.
The M-400s have a three step commutating power supply in a way. It is not e xactly but it shares the load a different way than other amps. There is the low, intermediate and high supply. The low supply is about +/- 40. the nex t around 60 and the high at about 80.
The whole idea is to only give the main outputs the voltage they need. the headroom is what heats them up. In fact that is why testing on amps like by hersch-Houck was preceded by a warmup of an hour at one third power. It is about as hot as it can get then and some amps went into thermal shutdown b ecause they were just running a continuous sine wave through it. Way differ ent than music, and specifically designed to beat on the amp.
Originally I thought the commutators switched but they don't. They operate linear, tracking the audio and keeping the supply voltage ahead of it.
All the dissipations are calculated and they didn't want power supply fluct uations to screw that all up. The thing is right on the edge in the first p lace, but seems to hold up.
You are right in that being a glorified light dimmer.
Know what else ? I started brainstorming a circuit that could take a TRIAC like that and make a signal by changing the phase at which it fires asymmet rically. Since TRIACs are not the fastest thing on the planet it would be l imited to servo control or possibly subwoofer amps. But we are talking one output device fed by a switcher with some tuning to make it put out more of a sine wave at some frequency the TRIAC can handle.
Since I started this business I have run into a bunch of configurations abo ut which I had misconceptions. Forced to study in detail has revealed quite a bit. For example, before I had the print I was taking some quick ohm/dio de tests for shorts right away and saw how it reads and thought WTF ? It wa s a Crown grounded bridge. Well that splains it...
Anyway, the Carver scheme is something I have not seen in anything else. Ev en if it was patented that surely has run out by now. I think most engineer s and manufacturers don't like the component count, because there is no chi p that does this.