GWInstek PSP-603 repair

[crispus]

I am the happy owner of this power supply. It arrived on October 6th.
Here is a short review and here some similar schematics.

I knew that had a short, so I started to look after it. Didn't take too long to found out that the main switching transistors (V9/V10) are in dead short. Changed them temporarily with other 2 from a PC PSU.
Connect the power supply from a transformer (~75Vac)

using the "debug cable" - up to 3 light bulbs in series with the load:
 
and seemed to work (above the multimeter you can barely see that the light bulb is starting to glow)

There were some issues with Farnell, made another order to TME and two weeks later I got the replacement transistors from TME. I replaced them with BUV48A.
While playing with the 'now working' power supply I considered that the problem is fixed and the debug cable is no longer needed.
It was so wrong.

I was testing the short protection when the hell broke loose and nearly fell off my chair. Something was terribly wrong.


Now, the painful part.
Changed the transistors again after one week when Farnell finally shipped the order, the control was working but no output.
I started with TL494 which had no Vref. Pull it out, solder a socket (good choice), put a good one in place. But still no output.

Followed the output of TL494 and seemed OK. After the control transformer (T2) R29/R32 are burned (open circuit). The transistors were pretty determined not go alone.
Changed them too, but no output.

Pull the TL494 out, measure with the multimeter R & C, all seems to be fine. Noticed though that the frequency is kind of odd:


I made on a breadboard a test circuit with TL494, similar values: ~37 Khz, but in the PSP is almost 10 times higher. Hmmm. The absolute max frequency in the datasheet is 300 Khz.
Accidentally I touch the capacitor on the breadboard and suddenly the frequency goes beyond 300Khz. Fix it, 37 KHz again. Take the again the capacitor out, it goes above 300 KHz.
It seems like a déjà vu.

Measure again the capacitor, this time with my LCR meter (thanks iloveelectronics). The capacity is right, but the other parameters not nearly what a capacitor should look like.
Go again to the donor PSU, and replace the culprit.

And now, I DO have the output.


Many thanks to ealex who guided me through this journey.