Gigabyte GA-8IE533 with busted caps

[c_hegge]

A customert brought their computer in with the board in the title. It wouldn't POST. After Removing the side, I notices 2 leaking GSC 3300uF Caps. The other caps on the board are as follows:

- 3x GSC 1500uF 16V
- 4x Sanyo WG 1500uF 6.3v
- 7x Supacon 6.3v 1000uF

Will probably use Panny FJ for the 2 busted 3300uF 6.3v and the supacons, FC for the 1500uF 16v.

Will also replace that horrible little northbridge heatsink for something bigger

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Aren't those WG's 8mm? FC doesn't have 8mm in 1500, only 1200 and less.

From the sheet I have, the FJ's ripple is very close to WG's and it shows an 8mm in 1500.

You could also drop in one 4v, 470 or 560 poly on each end of that row (2 caps total). Right next to the cooler bracket clips. The polys are only 10-12mm tall and should not pose a clearance issue with those clips. That's Vcore power and super low ESR from the polys would be a plus there. The other 4 will give you the capacitance.

Good luck!
Toast

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

There are 3 GSC 1500uF caps which are 10mm. They're the ones I'll replace with Panny FC. The Sanyo WG will be left in place as they aren't bad caps.

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

Might need some help here. I've re-capped it and it is now posting reliably but it locks up (mouse doesn't move, num lock and caps lock don't work, etc) after ~10hrs on Prime95. I've replaced all of the caps on the board except the four sanyo WG next to the CPU as sanyo are normally reliable. Should I have replaced them? My options for replacements to them are Panny FJ with the same capacitance/voltage (1500, 6.3) or Chemi-Con PSA Polys which are 680uF and 2.5v. Also, there are four blank spaces. Should I fill them up?
Thanks

PS. I've tried a different PSU and memtest says that the RAM is fine

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

Seems as though nobody has any ideas. Will try replacing the Sanyo WG and adding some 560uF polys, as per Toast's suggestion. Will post with results.

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Can you take a voltage reading on those open spaces? You might be able to go to 4v or 2.5v polys, which I think would be less expensive.

If you're stuck using the 6.3v, I just ordered some UCC PSC's @ 1500uF for a Sapphire Radeon HD3870 I'm sprucing up. Ran $2.06 each.

Toast

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

I've already done the re-cap. I replaced the Sanyo WG with 1500uF, 6.3v Panny FJ and filled the remaining slots up with 2.5v 560uF Fujitsu polys (NOT the Yellow, K-vent type), although I didn't measure the voltage as I was sure they were on the vcore. It's been running for a few hours on Prime95 without incident. I'm sure that if they were going to explode from over-voltage, they would've done it by now. I'm going to leave it running for the next few days and see how it goes. I'll post back with the results.

PS. I'm also testing a re-capped Antec SmartPower 350 by using it to power the board during the test. The video card I'm using is also re-capped.

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Did you test the Sanyo's?

Good to hear it sounds like it's going to be okay.

Total recap on the SP350, or just the outputs?

Toast

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

I Don't have an ESR meter or capacitance meter, and don't re-cap enough boards to make it worthwile, so I didn't test the sanyos. The SP350 has had all of the caps replaced with either Panny FC, FM, Nichicon HE or Chemi-con KZE, except for the really small ones and the two big ones on the mains.

The pics are of the old caps. The first pic has text. First line is what series they are, second is the ratings, third is what they were replaced with. You can probably tell from my signature what I plan on doing with these old caps

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

Just had to re-start the testing as my little brother turned it off at the wall

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Well, it has run 24 hrs with no problems...

Seems as though the *ahem* *cough* "The Sanyo WG will be left in place as they aren't bad caps." *ahh-choo* had some issues. *ahem*



Toast

[PCBONEZ]

Re: Gigabyte GA-8IE533 with busted caps

Just had to re-start the testing as my little brother turned it off at the wall

When he gets older, interrupt him while he's making out.

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

OK Toast, this is the first time I've seen Sanyo caps fail (and the first ones I'll be throwing in the fire ). Do you think that 24 hours on prime95 is a good enough test? It was running stable for at least that before he turned it off (although I was hoping for 48 hours).

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Run it for another 24 if you care too. I'm certain the FJ's & Polys did the trick though.

Nice job!

Toast

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

OK, will post tomorrow with results

[PCBONEZ]

Re: Gigabyte GA-8IE533 with busted caps

OK Toast, this is the first time I've seen Sanyo caps fail (and the first ones I'll be throwing in the fire ). Do you think that 24 hours on prime95 is a good enough test? It was running stable for at least that before he turned it off (although I was hoping for 48 hours).

I think what he was saying is that the Sanyo were fine and didn't need to come off.

[Toasty]

Re: Gigabyte GA-8IE533 with busted caps

Ummm... ahhhhh....?



Read from the beginning. The only caps not replaced initially were the Sanyo's. It locked up around 10 hours in on Prime95. Since replacement (and poly enhancement), no problems.

( Until little bro entered the scene )

[gdement]

Re: Gigabyte GA-8IE533 with busted caps

If the board lets you, I'd overclock the bus slightly (3-5% overclock on everything) and test it that way. Any stable board should be able to pass stress tests with an extra margin like that. It helps to account for imperfection of stability tests, and the general degradation that will happen over time.

[PCBONEZ]

Re: Gigabyte GA-8IE533 with busted caps

YOU read from the beginning. [With think cap on.]

If you look at what happened there's nothing at all here that proves the Sanyo were even slightly bad. [I thought you realized that.]

1) If it lasted 10 hours in on Prime95 then it's really unlikely that Vcore caps were the problem. - It's MUCH more likely it locked up because of an OS [presumably Windows] problem, a random RAM error, a PSU spike, or an over heat issue with RAM, CPU, Chipset, or --> VRM MOSFETs <--.

2) The 16v VRM-high got replaced with FC which aren't even close to being low enough ESR. The VRM-low [the Sanyos in Vcore] were then having to compensate [work harder] to keep Vcore clean. That may well have overheated the Sanyo and killed them after 10 hours but they were good for 10 hours before that. The VRM MOSFETs are also over worked trying to compensate for SMPS noise that the VRM-high caps are no longer removing as much of. That will eventually over heat them and will absolutely cause a lock-up.

3) When he pulled the Sanyo, Polymars were added to Vcore. That effectively unloads the Lytics in Vcore [almost completely] and helps compensate for the under rated FC's in VRM-high thereby reducing the load on the MOSFETs - somewhat. The high side MOSFETs are still dealing with more SMPS noise than they should be and that may eventually fry them.

Sorry,
Sequence of events just doesn't prove bad Sanyo AND may have been the cause of their death if they are dead now. [Which I doubt.]
.

[c_hegge]

Re: Gigabyte GA-8IE533 with busted caps

Looks like it's fixed. It ran for two Prime95 tests longer than 24 hours. In fact, I'm using it for this post, as well as the Antec SP-350 and video card I re-capped. Here are the after pics.