the gutless, bloated, and fried power supply hall of shame

[c_hegge]

All of the FSPs I've seen from the last few years have had DM311 based 5vsb circuits, even the low end ones with whimpy heat sinks and out of spec ripple.

[mockingbird]

Originally Posted by c_hegge

All of the FSPs I've seen from the last few years have had DM311 based 5vsb circuits, even the low end ones with whimpy heat sinks and out of spec ripple.

None of them have it... They all use the same pi filter for the 5vsb... I've got 20-pin 250 watt, and recent 300/350-watt 24-pin models (FSP300-60THN, FSP350-PN), Remarked FSPs from Aopen, etc... They all have an almost identical 5vsb circuit and there's always two 8mm Teapo SC caps there, and they're almost always what's stopping the thing from working. I replace them with 6.3v 1200uF KZEs because I can get them cheap and they have a better ripple rating.

[momaka]

The PI filter has nothing to do with the DM311 chip. The DM311 chip would be on the primary side of the PSU. Look for an 8 pin DIP chip. If those FSPs are single-transistor forward converter designs (and they probably are), you might see two 8 pin DIP chips (one being the PWM controller).

[c_hegge]

Omni 400W

Small primary caps, No PI filter coils, not enough caps on the secondary, and the caps which are there are all from unknown brands, and it blew up at 320W on the load tester

[mockingbird]

The PI filter has nothing to do with the DM311 chip. The DM311 chip would be on the primary side of the PSU. Look for an 8 pin DIP chip. If those FSPs are single-transistor forward converter designs (and they probably are), you might see two 8 pin DIP chips (one being the PWM controller).

Why are those two Teapo caps always bloating in the pi filter where the purple +5VSB goes in?

[Jooo]

Some nasty capacitors in a old Fortron PSU.
Surprisingly, this PSU was pulled from a working machine, and the voltages were stable... And the only capacitor that isn't bulging is Fuhjyyu, oh well.

[Wester547]

Originally Posted by mockingbird

Why are those two Teapo caps always bloating in the pi filter where the purple +5VSB goes in?

I don't know how right I am on this but maybe Teapos are more prone to fail by way of heat exposure than many other capacitors.

Originally Posted by Jooo

Some nasty capacitors in a old Fortron PSU.
Surprisingly, this PSU was pulled from a working machine, and the voltages were stable... And the only capacitor that isn't bulging is Fuhjyyu, oh well.

Ironic, considering Fuhjyyu is at the bottom of the chain pretty much when it comes to capacitor quality. At least that has Panasonic primary capacitors, though those heatsinks are not impressive but at least they're not finless. I'm not surprised that it was still stable. I think it's because FSP (at least in that timeframe of PSUs) PSUs are well designed. I had a FSP300-60BTV (it had similar heatsinks) that worked fine outside of some cold booting problems; (refusing to boot unless I let 5VSB sit for two minutes or so before turning on the computer) it was later discovered that a Fuhjyyu bloated on the 5VSB rail, and that PSU had a mingling of Jamicon, Teapo, and Fuhjyyu capacitors (Teapo LXK primaries, 680uF/200V). It still worked for many years. I also noticed that your PSU has less than 1A on the 5VSB rail... not sure how common that is, for usually I don't see 5VSB as rated below 1A in PSUs, even the older ones.

[Heihachi_73]

Same story with my Teapo-equipped FSP200-60ATV I scrapped a few weeks ago. It worked right up to the day when I had to power off the PC (it's on 24/7), and it never came back on afterwards. I'm now seeing how long a Fuhjyyu-equipped "420W" Thermal Master will hold up in the same system.

[Behemot]

Originally Posted by c_hegge

Omni 400W

Small primary caps, No PI filter coils, not enough caps on the secondary, and the caps which are there are all from unknown brands, and it blew up at 320W on the load tester

Looks like some LC to me. It even has input rectifier, thermistor and some stuff, it does not seem to be so bad, and 320 W is something you can work with, right…? Gonna repair it?

[c_hegge]

Nah, I've got lots of spare 400W+ PSUs, so a cheap 250 Watter isn't really much good to me.

[ben7]

Originally Posted by c_hegge

Nah, I've got lots of spare 400W+ PSUs, so a cheap 250 Watter isn't really much good to me.

Time to blow the rest of it up :P

[Pentium4]

Is that light blue cap a Lelon? The PSU looks honestly rated. Are those primary Panasonic caps still good?

[ben7]

Originally Posted by Pentium4

Is that light blue cap a Lelon? The PSU looks honestly rated. Are those primary Panasonic caps still good?

They all look to be 'YEC' brand caps.
YEC = YUK

[Pentium4]

Originally Posted by c_hegge

Omni 400W

Small primary caps, No PI filter coils, not enough caps on the secondary, and the caps which are there are all from unknown brands, and it blew up at 320W on the load tester

That thing looks pretty bad...It looks similar to one I threw away a while back. Wow, look at all those spots on the secondary for more caps...And to be honest, I'm surprised it could even do 320W! It probably wasn't in spec though. What died when it blew up, the secondary diodes? Those heatsinks don't look too good

Originally Posted by ben7

They all look to be 'YEC' brand caps.
YEC = YUK

Ohhh that's right, YEC are also light blue... YUK indeed

[momaka]

Originally Posted by mockingbird

Why are those two Teapo caps always bloating in the pi filter where the purple +5VSB goes in?

Good question, but I don't think I can answer it. Most likely they just get overexposed to heat, as Wester547 suggested. Perhaps there's a hot spot there. Or maybe the caps were not correctly specced for that circuit.

I just fixed the 2-transistor 5vsb circuit of an old 235-watter. The output caps were failed, and putting any load over 100 mA would make the voltage oscillate all over the place between 4V and 5V with lots of audible noise. When that happened, the second rail produced by the 5VSB transformer (normally 12V - used for powering the driving transformer for the BJTs on the main power supply) would jump to over 30V! That was all bad news for the 16V, 47uF cap that was filtering that rail. I then changed the output caps on the 5VSB with 2x very low ESR caps (one is actually motherboard grade), and the 5VSB is now solid as a rock and stable right at 5.01V, even with 800 mA of load. I have no doubts it will do fine at higher loads.

[Wester547]

Originally Posted by Pentium4

That thing looks pretty bad...It looks similar to one I threw away a while back. Wow, look at all those spots on the secondary for more caps...And to be honest, I'm surprised it could even do 320W! It probably wasn't in spec though. What died when it blew up, the secondary diodes? Those heatsinks don't look too good.

Switching transistors are far more commonplace in terms of failure than diodes or rectifying bridges (because they usually get decent airflow that's why they get away with using diodes in place of rectifying bridges in some very cheap PSUs, as has been noted before).

As for those Teapo capacitors, it could just be an inadequate batch as well. But Teapos may also fail around or on the 5VSB rail because the 5VSB rail (I might be wrong on this) is more often active than any other rail on a PSU (as it is still active when the system is 'soft-off' or 'soft-on"), so the capacitors on that rail are almost always working to some extent unless AC power is cut.

[c_hegge]

Originally Posted by Pentium4

...I'm surprised it could even do 320W! It probably wasn't in spec though. What died when it blew up, the secondary diodes? Those heatsinks don't look too good

No, it wasn't in spec. The switchers exploded at 320W. They were 13007s, and that's usually where they blow up.

[tleu8472]

I have one of those Omni 400W PSUs which I want to use to power an old P4 machine. Was the ripple bad? Can you suggest some modifications I can make to improve it? What caps should I add/replace on the output filtering? Many thanks.

[ben7]

Originally Posted by tleu8472

I have one of those Omni 400W PSUs which I want to use to power an old P4 machine. Was the ripple bad? Can you suggest some modifications I can make to improve it? What caps should I add/replace on the output filtering? Many thanks.

Seeing how very few filter caps there were, I would guess that the ripple was terrible. It's also missing PI filter coils.

[tleu8472]

What uF and voltage ratings should I use to add/replace the secondary filtering caps? It does have resistors instead of coils - will this suffice for what will probably be a 250W max load system?