Interesting PSU Component Failure

[steenss]

Hi all.

Anyone seen a metalized poly cap fail like this before?

531s-2.jpg (114.5 KiB) Viewed 20869 times

AcBel compact PSU from a Dell 531s. Possibly marginal with only 17A @ 12V & A64x2 5000+, 4MB DDR2 800 & Radeon HD3450.

531s-1.jpg (192.03 KiB) Viewed 20869 times

Torn down PSU with component spot between the coils. The unit had arcing sound on being plugged in & sometimes while running. The electrolytics are LTEC except for primaries which are Elite... Heat's a killer for these things. Replaced the stamped fan grill guard with a wire one. Seems to run fine now.

531s-3.jpg (188.19 KiB) Viewed 20869 times

[c_hegge]

Interesting failure indeed. That usually only happens when the PSU gets a power surge. I assume that the one attached to the AC receptacle is fine?

[steenss]

The other MPX cap is OK. The MOV should have mopped up the surge, but with X caps in parallel I guess they cop the spike.

[c_hegge]

That's what I would have thought. Maybe it was just a dud cap.

[LongRunner]

I know it's an old thread, but I found something to add.

Apparently it's not that rare for X2 capacitors to fail. It tends to go unnoticed much of the time though as they normally fail with reduced capacitance or open-circuit (as they are designed to!). It seems that there are some complaints about Carli MPX caps (as seen there). I made a new thread about these "eX-capacitors" (yeah, I know that pun is bad). It's when they are used as droppers that the failure shows (although "high stability" variants, with series instead of mono construction, should really be used there).

At least the ceramic Y capacitors seem to be quite bulletproof, even if they have miserable temperature coefficients of capacitance (Y5U is about as good as they get, with the exception of sub-nF models). Film Y capacitors aren't often seen (certainly not in PC PSUs) so I can't really say.