Epox 8k9a2+ bad caps

[Phaz]

all affected caps are the 6.3 2200 microfarad other caps arent affected
Machine stability is not affected, mbm5 reports my 3.3v rail @ 1.7v
all suspect caps bulged, some discharge on the top of them

[RJARRRPCGP]

Re: Epox 8k9a2+ bad caps

all affected caps are the 6.3 2200 microfarad other caps arent affected
Machine stability is not affected, mbm5 reports my 3.3v rail @ 1.7v
all suspect caps bulged, some discharge on the top of them

It's probably still unstable. If you run any heavy tasks, then it probably would crash. Also, I had the 3.3V rail be read as 3.1V and 2.8V while using my PC one day and amazingly, was still stable when running a 4-hour loop of 3D Mark 2001 SE! That was just because of the ATX power connector being loose!

FYI, if you ever come across a PC with the 3.3V (or another rail) reading being majorly low, such as 2.8V, then check the ATX power connector and just push it down and make sure it's clipped. Mine for some reason didn't clip after putting in another processor last August. I was surprised that the ATX power connectors didn't get charred!

[Rainbow]

Re: Epox 8k9a2+ bad caps

I've got this board for repair. There are GSC (1000uF, 2200uF) and TEAPO (3300uF) caps bulging. There are also GSC 1500uF that look OK but I'll replace them anyway.

[japlytic]

Re: Epox 8k9a2+ bad caps

I'd get those GSC capacitors out first (According to www.gsc-cap.com, their business has been established in Mainland China since 2003... maybe this explains something?).
Avoid the Nichicon extra low-ESR units... as many of us know, they have been bulging in Optiplex GX270s.

[Rainbow]

Re: Epox 8k9a2+ bad caps

Replaced the 6 3300uF Teapos (connected in parallel) with 4 Kome 4700uF/16V+1 Kome 1000uF/16V, 1000uF GSCs with Kome 1000uF/16V, 2200uF GSCs with Kome 2200uF/16V and finally the still-good-looking 1500uF GSCs with Kome 2200uF/16V too.
The board had some problems with powering up but then finally powered up and worked. But after a while, it suddenly turned off - but strangely without smell So powered it up again - and it turned off in a couple off seconds again. Next power up attempts were unsuccessful.
So I traced where the PS_ON# signal from the ATX connector goes - and found a small 8-pin chip - Attansic ATTP1 which had some white substance between its pins. Cleaned it using a toothbrush&alcohol - and the board then powered up and worked fine
The ATTP1 is over-temperature protection controller and that white substance was probably conductive enough for it to activate.

[japlytic]

Re: Epox 8k9a2+ bad caps

I have never heard about KOME capacitors before... see www.pro-an.com.hk and you will find they have been established in 1992.

[PeteS in CA]

Re: Epox 8k9a2+ bad caps

Contamination and small IC pin spacing sucks. Welcome to the wonders of SMT. Don't have shaky hands when probing a SMT IC pin with a DMM probe or scope probe .

[Rainbow]

Re: Epox 8k9a2+ bad caps

That white substance was probably something from a water-cooling system that the owner uses with that board. The ATTP1 chip connects directly to a thermistor or diode so it's very sensitive to resistance changes.

[Rainbow]

Re: Epox 8k9a2+ bad caps

I've measured the capacitance of the bad caps from this board (using a multimeter that can do up to 20uF):
GSC 2200uF/10V:
12.11uF
16.09uF
19.40uF
>20uF

GSC 1000uF/6.3V:
16.10uF (bottom of came out)
9.33uF
5.64uF
0.21uF
13.90uF
0.00uF (748pF )
>20uF
6.99uF

GSC 1500uF/16V - all >20uF (none of them bulged)
TEAPO 3300uF/6.3V - all >20uF (including the single bulged one)

Summary: GSC is REAL crap. No wonder that the board failed with values like these.

[sulbert]

Re: Epox 8k9a2+ bad caps

Need to recap one of these boards too. The board lying on my desk has Sanyo MV-WX caps instead of 3300uF Teapos so I have a bit less work
But I have trouble finding replacement for the 2200uF/10V GSC caps located between the FETs - 10mm is the largest diameter that will fit there but I have only found 12.5mm 10V caps so far (and I think I can't replace them with the 6.3V ones since they seem to be connected to the 5V rail).

[Rainbow]

Re: Epox 8k9a2+ bad caps

I used 2200uF/16V that have the same diameter as the original GSC and are even lower.

[AK0R]

Re: Epox 8k9a2+ bad caps

But I have trouble finding replacement for the 2200uF/10V GSC caps located between the FETs - 10mm is the largest diameter that will fit there but I have only found 12.5mm 10V caps so far (and I think I can't replace them with the 6.3V ones since they seem to be connected to the 5V rail).

Computek (which usually supplies Rubycon MBZ) and Digi-Key (Panasonic FC) both have parts available in 2200uF, 10v, 10mm size. It would be marginal, but acceptable, to use a 6.3v cap on the 5v line.

[AK0R]

Re: Epox 8k9a2+ bad caps

Sulbert,
How many do you need? Panasonic lists 0 for their inventory. I need to order some of these for a power supply I'm refurbishing, and will be placing an order with Computek for those and some other caps.

[sulbert]

Re: Epox 8k9a2+ bad caps

Thank you but I already bought 12.5mm Sanyos from a local electronics store (the owner of the board wanted it back ASAP). Looks ugly IMHO but it works.