PC Power supply - 3.3V shorted to ground

[sasa]

Some 10 years old "Chieftec HPC-360-202 DF" PC power supply finally ended his life with my old PC, blowing away two main 680uF  bypass caps...

I do not consider it is worthed to do any repair for such old PSU, nor I would actually made any repair myself. However, it is interesting that 3.3V line is actually shorted to ground. I wonder any professional EE repaired PC PSUs had such experience?

Probably more than caps are fallen out or blown away.

[Kjelt]

Better to ground than directly put the input to the output 
Modern PC PSUs have only 12V powerrails and the other voltages are DC-DC converters from 12V to the other voltage, I do hope they also go to ground or open if they fail.

[Armadillo]

Older power supply likes pyrotechnic exit while the newer ones will do the more graceful cycle dance.   

[SaabFAN]

Some Power Supplies do crayz things on the primary side when something on the secondary goes wrong.

An amplifier I repaired a few weeks ago, for example, blew both switching-transistors (Push-Pull-Configuration) when the output-cable that normally goes into the speaker touched the case repeatedly.
I was unsure if something else was broken, so I ordered a few transistors more than needed, but it really was just the 2 N-FETs that were blown. The rest of the amp was fine and worked like nothing happened after replacing the FETs.

I'm guessing that the PSU tried to bring the 3,3V-Rail up and started oscillating because the 12V-Rail and 5V-Rail went overvoltage, which the old caps didn't like (excessive ripple current paired with high ESR due to age = Lots of heat).
Try fixing the short on the 3,3V-Line and replace the two filter caps. Old PC PSUs are good if you want to test stuff that needs a bit more power than the standard 31V / 5,1A Lab-PSU can deliver.
Just connect the green wire to a black one and the PSU switches on the moment you flick the power-switch.
Unless it is one of those abominations without a power-switch. Those need to be purged by the cleansing flames of the Emperor

[sasa]

Thank you all for comment.

This actually was quite nice PSU included with a big tower Chieftec case some 10 years ago and surprisingly for that time had built in quad protection common today (OVP, OPP, SCP, OTP). It worked fine exactly according estimated life, with only dust cleaning periodically and that is quite impressive.

http://www.highpowersupply.com/product-ps-hpc360202df.htm

If protection circuits worked fine, It should perhaps work a bit longer with changed critical components. Perhaps I will have a bit more free time for diagnostic during weekend as finding exact reason for short is more interesting than actually fixing such old device.

[sasa]

After more diagnostic and desoldering all 3.3V rail cables, it was obvious that two small resistors in vicinity are burned out, as well as one small cap was dried out and that causes the short with ground. As well, some diodes seems to leaks.

Two more 2.2uf 400V caps on external voltage doubler board are dried out as well. All these dried out caps had no visible signs they are faulty and again, thanks to Jay_Diddy_B 5 transistor ESR, they are easily located in-circuit.

As mentioned, it is not worthed to do any repair on such old PSU. Perhaps I will salvage large heat-sink and coils (which seems to be intact).

Thanks to all for comments.