This is why you always replace those little ventless caps

[Pentium4]

These Fuhjyyu caps came out of my "working" Dynapower PSU I posted in the build quality pictoral ( https://www.badcaps.net/forum/showpo...postcount=1419) Look at how high the ESR is! And the capacitances are out of spec too. The little ones were the caps in between the primary heatsink and the transformer. The 220uF cap (no vent) was on the -12V rail. What kind of havoc could this much ESR wreak into a PSU?!

I recapped the PSU:
https://www.badcaps.net/forum/showpo...postcount=1431

[Th3_uN1Qu3]

Re: This is why you always replace those little ventless caps

The -12v rail isn't that critical. On the other hand, the caps between the primary heatsink and the transformer, are part of either the 5vsb or the speedup circuit for the base driver of the primary transistors - if they dry out completely, they will cause the primary transistors to blow.

[Pentium4]

Re: This is why you always replace those little ventless caps

Yeah I knew the -12V wasn't critical but still crazy how the cap got so bad on a minor rail. And thanks I always wondered what those little ones were for! Wouldn't this severely effect the overall operation of the PSU?

[Wester547]

Re: This is why you always replace those little ventless caps

It would have helped to have added a heatsink-mounted voltage regulator for the -12V rail on their part, though judging by the images in the power supply build thread it seems they at least added a PI filter for that rail, but I don't think that helps as much and even those PI filter coils do generate heat...

I bring up the voltage regulator because I've only ever seen 10uF (usually 50V) capacitors on the -12V or -5V rail PSUs from decent brands with a voltage regulator added... I know they still added a 220uF capacitor but even then. Fuhjyyu are bad enough to fail on their own, though, especially in a gutless PSU. Unique is right - the most important small capacitors in a PSU are the startup capacitors, whatever drives the PWM chip, or the +5VSB chip, or the UC384x variant, or the circuit dealing with the main switchers (like Unique said), or even the rather inefficient two-transistor +5VSB circuit (the "critical" capacitors though the ones without them are less worrying), since they are stressed. I don't understand why they almost always use general purpose and small capacitors for this, other than for costs and to save space - they are more likely to fail in that line of duty, especially from Taiwanese and Chinese brands...

The other small capacitors, excepting the ones on the negative rails (of course, if they fail, they can mess up the regulation of the whole PSU...), are usually just general purpose decoupling and coupling capacitors, that's why you sometimes see even 85*C capacitors used there. Since they are only decoupling and coupling capacitors, which only pass a signal, not much ripple goes through them at all, so even subpar brands can hold up there for a long time. Not sure if they have it as easy as the voltage doubler does or even easier but I can say that they certainly have it easier than the output capacitors or the startup capacitors.

[Pentium4]

Re: This is why you always replace those little ventless caps

Yeah it's a decent size inductor for the -12V but yeah I assume the Fuhjyyu just dried out, what's strange is the 3300uF 16V Fuhjyyu on the 12V rail read 0.01 ESR and 3500uF...

So high ESR on those little caps isn't that big of a deal even 30+ ? I still like to replace them all especially with how cheap those little caps are when you buy in quantity.

This PSU used a lot of GP caps, there was a Teapo SEK series cap on the 5V and lots of Fuhjyyu TN series

Thank you for the great post as usual