• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

weird stability problems (thorough details inside)

K

knob

Limp Gawd
Joined
Feb 16, 2008
Messages
291
Friend of mine has been having some bad stability trouble lately.

First the system spec:
PD @ stock clock (3.0GHz)
3GB Corsair (2x1GB of 6400C4, 2x512MB of 5400C4, all running at 667MHz, 2.1v, 5-5-5)
Abit AW8
Generic (aspire?) 420W psu
XFX 7600 GT (stock clocks)

He's owned the box for 2 years, since initially built, with the stability trouble only arising recently, in the last 2 weeks. He can be playing WoW or surfing the web when without warning, the box will restart.

He and I have been going around trying to figure this out... we've tried upping CPU volts, DRAM volts and NB volts, all seemingly to no avail.

He ran memtest continuously for about 6 hours with no errors at all.

I had him download Orthos today and try it out. His CPU will do Small FFTs for up to around a half hour, before the pc will shut off and restart, without Orthos throwing an error... though sometimes Orthos will throw an error (the normal rounding error kind). More interestingly, if he takes the side panel off his case, he can go hours without a restart, and the volts on 12v stay about 0.1v higher on Orthos load (11.9v, that is)

I sat down with a volt meter and measured the volts on the 12v rail when he was running Orthos. At idle, there is about 12.05v at the free molex connector we were measuring at. At full Orthos load, his 12v rail volts will hover a bit above 11.8v, with the PSU putting out decidedly warm air, even though his case temp was only couple degrees above ambient.

Oddly, after it restarts, sometimes it will lockup on the POST screen, having detected only 1 of 2 CPU cores.

Personally, I want to suspect a PSU problem, though the idea of a failing CPU keeps nagging. Any ideas are appreciated.
 
update/bump: He ran Orthos overnight (8 hours) on Small FFTs and got 0 errors.

This points to a failing psu? Anybody?

EDIT: Apparently I didn't know the whole story about that Orthos run... apparently when Orthos went idle, after he manually stopped it, his box restarted
 
This type of thing has always been a PSU problem for me. Just about every machine I've had where I go low budget on the PSU has gone bad after a few years and it's always been fixed by a PSU replacement. If you're considering it to be a CPU problem, you should also consider that it could be a motherboard problem.
 
The POST lockup definately points to a PSU issue, as during drive spin-up the system is using alot of power on the 12v rail.............. I'd be willing to bet within a short timeframe the machine will start being difficult to power on at all; with the power button sometimes taking several minutes to have any effect.............. I'd also be willing to bet (having experience with Aspire PSUs) that if you were to open it, some or all of the caps would be bulging/leaking................
 
Ok, another update.

I had my friend pull apart his power supply and I took a look inside it. There is a cap with a bulging top over on the secondary side, kinda between two toroidal transformers. (In the same spot as the central blue capacitor in this wikipedia picture - so, coolie_d, looks like you called it)

I figure the best thing to do now would be to test a known working psu in for his possibly failing one and see if his stability issues go away. If so, I'll have him pick up a 450VX.
 
+1 for a probable bad PSU. No real way to test them unfortunately. Best way is just to go get a new one. Fortunately, that is the cheapest of your options at this point.

I have used el cheapo PSU's over the years with pretty good results. I only had 1 that outright died, but it did not take anything with it. Just lucky I guess. But I never stressed them anywhere near their rated output. But after I started hanging out in various hardware site forums, I have been a little more picky about my last couple of builds.

My old XP 2800 system is running an Ultra 400 watt, and my Q6600 is running a Coolermaster 500 watt supply. I only have one graphics card now, but if I ever decide to run Crossfire, I would prolly bump that one to at least a 600 watt supply. I think I paid 30 bux for the Ultra from Tiger Direct OUtlet, and 40 (-30 MIR) for the Coolermaster at Microcenter last month.

The Ultra has been great. That system runs for months at a time without requiring reboots.
 
You must log in or register to reply here.
Back
Top