FSP400-60GLN worth fixing?

[Elitist]

Re: FSP400-60GLN worth fixing?

For the record, that R44 on my chassis is a zero ohm item, i.e it is a circuit link. Got no issues over that side of the board, but would very much appreciate help at the R92 position.

[Behemot]

Re: FSP400-60GLN worth fixing?

In BlueStorm II 500 W R92 is "510", that would be 51 ohms.

[Elitist]

Re: FSP400-60GLN worth fixing?

Thanks a £$m, Behemot. That's bad news, though. Seems the resistor has survived the thermal storm but not the markings on it. So must look for the cause elsewhere.

[Elitist]

Re: FSP400-60GLN worth fixing?

Whilst I was dismantling this beast, I had a strong feeling of deja vu. Sure enough, whilst searching for spares, I found another FSP400-60GLN with burn marks around
R92 (vapourised), D90, and R93 with some colouration around D92, R9F & R9G.
However the failure symptoms were different! Voltages were good but unable to sustain even modest loads and the power good signal was completely absent.

Are we any nearer a definitive diagnosis? After all, both ran for about two years.

[Elitist]

Re: FSP400-60GLN worth fixing?

"Less jewellery, more gold into electrotech industry! Half of the computer problems is caused by bad contacts"
Phosphor-bronze may be marginally cheaper & better ?

[Elitist]

Re: FSP400-60GLN worth fixing?

Update: working on it! re. Everell, #27, my D41 had already been uprated to a UF1003, listed as "Diffused Junction Ultra-Fast Switching for High Efficiency Low Reverse Leakage Current Surge Overload Rating to 30A Peak" which would seem to meet your recommendations and, out-of-circuit, shows appropriate diode characteristics. C15 was a C(r)apXon 100/25 which I've replaced with an 100/35 Rubycon. Still searching to cannibilise more components, but might be forced into purchasing a few SMC/SMD/SME/W.H.Y. at the exhorbitant cost of a few pennies each + VAT+minimum order+postage - could work out cheaper to buy a new PSU.
Incidentally, why is C41 missing?

[Elitist]

Re: FSP400-60GLN worth fixing?

Excuse not having read through the entire thread, but curiosity won out and I removed C32 to find out what it was - C(r)apXon 2200/6.3 ! It's the one wrapped in a black sleeve to hide its parentage. Apologies if it's been reported already - let me know and I'll delete this. If not, would be helpful to know what function it performs.

[Behemot]

Re: FSP400-60GLN worth fixing?

Is it the one between the heatsink and something (a coil IIRC)? If so than it's for +3.3 V filtering. Or +5, but I think 3.3. You can beep your way from it to the output where you'll see which rail it is from the wire coloring…

[Elitist]

Re: FSP400-60GLN worth fixing?

Dobry den, Behemot! Yes, that's the one - behind discrete M30. I'll beep it - if it is 3.3v line, that explains the whole sorry saga. Dekuji !

[Elitist]

Re: FSP400-60GLN worth fixing?

Well, exasperation has set in. Enormously grateful for this thread and the helpful guys who've posted - thanks. Replaced caps, checked diodes/zeners and resistors, inserted missing components, all on two identical units. They both run now, but the 3.3v line still drops to 2.3v with a 50ohm ww load. Knowledge (especially), time available may bring this project to an untimely termination. Particularly annoying that these PSU were specifically recommended by several sources when purchased a couple of years ago. Their physical configuration makes them a little more difficult to run on the bench than average. Despite some opinions, not convinced that higher cost units are more reliable than el cheapos like these. At least with the latter, if something goes bang the fix can be simple and cheap. Just as many dead behemoths come this way, almost impossible to access, diagnose and fix, although, in their favour, stripping them yields more spares!

[everell]

Re: FSP400-60GLN worth fixing?

Just got one of the Behemot favorite power supplies, a Sparkle FSP400-60GLN. Looked clean inside, and the fan spins freely. Three capacitors visually bulging. Measured the APFC FETs and switching FETs for shorts. No problem.

First things first - get the 5vsb operational. D41 the feedback diode was a slowpoke 1N4007. Replaced it with a 1N4935. D42 snubber diode was also a slowpoke 1N4007 diode. Replaced it with a MUR2100E high speed diode. C15 the feedback filter capacitor was a Teapo 22 uF/50 volt capacitor measuring 19.3 uF and ESR 0.95. Replaced it with a Panasonic 22 uF/50 volt capacitor measuring 21.7 uF and ESR 0.27. The input pi filter capacitor C45 was a bulging Capxon measuring 27 uF and ESR 2.9. The output pi filter capacitor C46 was a Capxon measuring 1030 uF (should be 1000 uF) and esr 0.03. I replaced both with Nichicon PW 1000 uF both measuring esr 0.09. Next, I applied power and the 5vsb came up fine measuring 5.10 volts.

Next, I checked to see if the D35 diode was missing in the 3.3 volt circuit. Yes. So I installed one. Two of the capacitors in the output circuits were bulging, C32 and C35. Both were in the 3.3 volt pi filter circuit. Both were Capxon 2200 uF/6.3 volt capacitors. Strange thing both measured esr 3.6. Since the capacitors were not close together, I would not have expected the esr to be the same. Since all the other capacitors seem to be OK, this may have been a bad batch problem. So what do you think about this?

The power supply now seems to be working fine.

[Wester547]

Re: FSP400-60GLN worth fixing?

Well, those CapXons IIRC have a capacitance tolerance of -+20% (M). So 1030uF is definitely within spec for a 1000uF 16V? capacitor. 0.09Ω ESR sounds a bit high for a 1000uF 10V or 16V? 8mm or 10mm (whatever you replaced it with ) Nichicon PW unless the reading in question was taken with its full leads (0.09Ω ESR would be spot-on for a 10mm 1000uF 6.3V PW). In accordance to the datasheet, a 1000uF 10V or 16V PW should not be higher than 0.07Ω or 0.06Ω ESR respectively. "Bad batch"... seems to me that almost all CrapXons have the same manufacturing process flaws so they are bound to fail the same way.

[Behemot]

Re: FSP400-60GLN worth fixing?

It may depend on the load, or sun erruptions, don't know. I get these (and similar) units once with bad 3.3V caps, than 12V caps, than all of them, than just stand-by…it may be completely random. The single thing remains - they are garbage and go bad here.

[Elitist]

Re: FSP400-60GLN worth fixing?

Just received the last of the batch fitted ~3yrs ago. Was it working? Yes, but only for
~30s. No caps were bulging. Only sign of problems were the 'usual suspects' - burnt R92/R93/D90 and peripheral sme s described on previous page. Might attempt resurrection based on advice. Glad to see this thread is still open & thanks again, guys.

[Behemot]

Re: FSP400-60GLN worth fixing?

Check also the nearby caps, they do not always bulge as they are small, but ESR skyrockets due to the heat, while capacity goes to zero.

[Elitist]

Re: FSP400-60GLN worth fixing?

Thanks again, Bh, working through the caps replacement; would be cheaper to replace the whole PSU! Curiosity - why is C41 missing? Any mileage in filling it in? Value?

[Behemot]

Re: FSP400-60GLN worth fixing?

I believe they are both in parallel. I have seen it in some variants, most lacked it though.

[Elitist]

Re: FSP400-60GLN worth fixing?

Fascinating. When I read your response, Bh, I thought maybe it was for an hf bypass C. But no, there's an sme one underneath and screen print shows C41 polarised.
Changed a load of caps, Ds & Rs as recommended. The beast runs smoothly, fan/voltages/etc and continues to do so when connected to PC. Only problem - PC won't start up even though all voltages remain correct (no unexpected drops on any line) and stable. Guess this is diagnostic for a guru - sadly not for me...

[Behemot]

Re: FSP400-60GLN worth fixing?

Does it output PWR Good?

[Elitist]

Re: FSP400-60GLN worth fixing?

Excellent question - no! Only got 0.13V. Where did the other 0.37V go...