k7000 (19") Blows Fuse: Shorted HOT but Not What You Think

[johnczerwinski]

This ones for the books. I've never run into it before, but maybe someone else has

Symptom:
k7000 19" (P447) kept blowing fuses

Cause:
Inside the screw hole was exposed to the collector and other leg of the transistor, so would short out if you fastened the screw to the heatsink (even with an insulator).
My theory is someone, who worked on it previously, had tried using a bigger screw or man-handled the HOT in removing it.

Troubleshooting process

1. HOT would show shorted with the multimeter
2. When you pulled the HOT and tested out of circuit, all the legs showed no short with normal readings
3. Solder it back in, it would show shorted again.
4. Unscrewed the screw from the heatsink, and the HOT read normal in circuit
5. Fastened it back, showed shorted
6. Removed it completely out of circuit and put my multimeter in ohms/continuity with the probe in the hole (only touching the plastic package) and would show shorted to the collector and other legs.
7. Replaced the HOT with a brand new one, works without shorting. But the initial short took out other components to cause a vertical collapse. That's for another log.

Conclusion:
Even though the transistor tested normally with standard test procedures (black probe on collector, red on emitter and base), there was damage to the plastic package, inside the screw hole, that exposed the collector to the screw, then to the heatsink causing it to short.

Hope this help someone else

 

[DuffCon]

This looks similar to when someone uses PC CPU thermal compound instead of thermal grease. The PC compound is conductive and fries the HOT. When I did this mine looked like yours.