Hi again,
Ok, i just wanted to get an idea what you are doing here.
I took a look at the schematic you linked to, and i believe tcmtech was on the
right track. The only problem with that exact mod is that the resistor overheats as
you said but even more serious is that the triac gate gets overpowered.
One of the goals in triac circuits is to minimize gate power, and that isnt
going to do it. That idea however can be implemented in a slightly different way
though as i will explain.
It looks like the problem right now is that although the gate appears to be
'triggered' by the PIC output (via the opto) it is not actually synced with that
pulse, but with that pulse AND the voltage at MT2. The voltage at MT2 is
of course dependent on the inductor voltage and that is not good.
The idea then is to develop a new trigger signal, but one that is dependent
solely on the PIC output pulse, and virtually nothing else.
The main way to do this is to create a DC power supply that is common to MT1,
and use that with say a normal opto (transistor output) to trigger the triac.
Of course to test this circuit first, you can use a battery as the dc supply.
Keep in mind however that when you use a dc supply you can not still use the
triac opto, it has to be a transistor type.
What i would do is use a battery, use a transistor opto, and test the circuit
again and see how it works out. I would bet it would be better.
One thing though, you may hear the transformer hum a bit more because it
will be being banged by higher than usual harmonics.
I hope you can try this soon, and let us know of the outcome.
I also have to ask:
Does the PIC circuit really have to be electrically isolated from the
power line? Many circuits can run with a common line from the
PIC ground (or +supply) to the line power neutral.