Did you rule out the chuffing-in-idle scenario (the one solved with a tach resistor)?
I suppose you would have said if you put the scope on the tach signal but to your point in the other thread perhaps something is modulating/pulsing the 5V power supply which ripples into false tach pulses. If the PS board has a direction-relay I believe those are powered from 5V and you mentioned this started when you "cycled direction" which I assumed meant you put it in reverse. Likewise I believe the smoke fan motor PWM is sourced from 5V - so for example if you turn OFF the smoke did the audio still chuff with no motion?