Yes, the voltage across that power resistor looks very different when
scoped. No clean pulses at all; just constant NOISE and lots of it up to
about 20 or more volts so no wonder it was getting hot.
I also scoped the pwm output from the controller chip (TP.2 on the
diagram) and it looks like it's outputting a pulse train but is totally
overwhelmed by noise which I would imagine must be virtually saturating
V1812 resulting in a close to 100% duty cycle at no load.
Clearly major stability issues here.