Part of the secret is that there is a bridge rectifier that I missed on my first look at the circuit. So the AC waveform coming out of the triac gets rectified to DC before going to the motor.
The next question is, how do we avoid burning out a 30 volt DC motor with rectified 115 volt line voltage? This must be done by being careful about what fraction of the AC power gets delivered by the chopping algorithm. In other words, the firmware can only trigger the triacs in limited ways.
Another fellow measured voltages at the motor as follows:
Fan setting 1 - 21 volts Fan setting 2 - 33 voltsThe timings relative to zero crossings that I measured would seem to send about 75 percent of each cycle to the motor, which seems virtually certain to burn out the motor. Is it possible that the fan only gets triggered on positive cycles (thus reducing 75 to 37 percent?). Maybe.
Tom's coffee pages / tom@mmto.org