You could try soldering in the new Fairchild MOSFET, for a couple bucks it might be worth a try. I do see that there is some minor discoloration on the through holes were the resistor legs are on the board. I do see that the Fairchild FET is rated at 12 amps and the original Asians were 6 amp. For this exercise to be accurate, you would want to find if the quiescent current through that line of resistors and driver transistors, adjacent, is within the ratings of the new FET's. At a minimum, I would be checking resistance of those resistors and continuity of the small driver transistors. In most cases, this means desoldering those from the board.
It's difficult to find a true fail mechanism with this since we have no idea how the circuits are designed with respect to circuit use and the currents involved. It may be that your friends at Cheapowatt were running everything right to the ragged edge with no headroom (typical) and just a minimum of hot-rodding pushed something over the edge. We can't really say that the MOSFET was at fault since we have no idea about the rest of the circuits that interact with them. Since power supplies are fairly crude, a good EE with circuit design experience could calculate the resistance and voltages present on the board and derive a calculation of whats involved. That person isn't me, I have a hard enough time identifying component values.
Oh well, at least you got to see some flames and hear some electronic sizzle. :B
Thanks for posting the pictures and continuing the discussion. BTW, mineral spirits works pretty good for thermal paste. And, yeah, they must tell the workers commuting from their far North to just slather the stuff on.
