clamp smps
Gomez
Difficult to say what has been damaged.
Maybe optocoupler, start-up capacitor or TOP itself.
Have you tried to replace P1.5KE200 ? Often this is the problem.
Maybe the P1.5KE200, though you have used snubber resistors, got still very hot before you made short. Value of 10K//10K you have used sounds quite low. This could says that you have not solved the main problem; the primary leakage inductance.
3uH seems good over, let say, 250/350uH of primary inductance, however have you measured output trace inductance ?
Trace inductance will have big impact over primary leakage, as it increases the secondary inductance and secondary inductance will be transferred to primary multiplied by the square of tourns ratio.
For instance, if primary wounds are 40 tourns and secondary are 4 tourns (low voltage output) the ratio is N=10. Should the trace + secondary indunctance be equal to 50nH, (i.e. long tracks of your PCB), you will have 5uH of additional leakage on primary.
As MADURAN said, energy stored with this power is terrific bigger. So, though before you made the short your SMPS seemed working, it doesn't means it works nice.
When fixed the fault I suggest:
1) split primary. about 50% reduction of primary leakage.
2) get measured the trace inductance in your application.
3) use snubbers resistor and
4) get monitored the temperature over clamp's body. Temperatures greather then 100°C will not allow for any safety margin.
5) split P1.5KE200 in two P1.5KE100 in series. It works, and helps to reduce a lot the temperature over the clamp.
6) use multi-wounds technique to reduce loss in copper. It doesn't matter if you have used wire greather then 0.4mm. The skin effect at 100-130kHz will reduce useful sectional area.