Trying to diagnose faults - what can I try

Soldato
Joined
1 Feb 2006
Posts
8,188
Hi,

I've had numerous threads on my various issues since getting my new system on Feb 10th.

First I had a gfx card RMA'd, then still having issues so got the OC'd bundle RMA'd. RAM was replaced and apparently all tested with 10hours + on Prime95.

I get the bundle back and I'm still having issues with any of the overclocked profiles. It fails OCCT, IBT and Prime95 within minutes. I did a PSU test in OCCT and it lasted almost an hour before a BSOD. Think this suggests PSU is ok.

I have tried memtest for 10hours + with about 20 passes.

I have tried a different gfx card. Same issues.

I have tried different wall sockets. Same issues.

I have tried different hard drive with fresh install of Windows. Same issues.

I have tried different cables for the modular PSU. Still the same issues.

I really want this thing to work but I think the only thing faulty could be the CPU/mobo.

OCUK say it was working 100% but everything here including temperatures are grand so I'm so confused.

Any other suggestions on what to test? At stock settings I don't get any of the problems.
 
Just ran Prime95 with a different gfx card, fresh install of windows 7 at stock profiles. No issues at all after 15 minutes.

Switched back to an OC'd profile from OCUK and Prime95 has 2 threads fail after 2 minutes.

Surely surely surely this is a bundle issue.
 
I would run it for a few days on stock profiles, not 15mins - then switch to the overclocked profile - see what happens
 
Will try that and see what happens. Can't really afford the time for running diagnostics though. It will be 3 weeks on wednesday since delivered. Joke!
 
im thinking PSU at this point , what model is it how are the rails under load before the BSOD with the OCUK profile and your full system spec would be handy as well to know
 
It is a Corsair 750W and that should be well capable of running this setup. I had an old gfx card in, with one hard drive so not much extra loads from drives etc.

What readings is it you would need to see? I'm not too up in this OC'ing game.
 
just check on the voltages with something like HWMoniter and see mainly how the 12v line handles when under load. I only say the PSU as its the only real question mark unless your hard drive is screwing it up some how
 
PSU/motherboard incompatibility looks like the most probable explanation. There were loads of issues with UD5s and a seasonic psu iirc, though I don't remember any involving a corsair psu. It's not that it cant run the system, more some ghost in the machine which occasionally makes certain hardware take offense at each other.

Stress testing at stock seems a sensible plan. If it's fine at stock but not at their overclock, then they're going to struggle to argue the overclock is fine.

Software readings of psu voltages are somewhere between guesses and blatent lies, trust a multimeter but not a software reading.
 
Thanks for that. I am running OCCT and it gives me a reading of vcore, 3.3V, 5V and 12V. Anything I should look for there?

I did run the OCCT PSU test for an hour earlier and it was fine.
 
Not really any way of knowing if it's a psu vs motherboard problem except by trying another psu. The voltages will probably look fine either way. OCCT has no idea what voltages the rails are at, whatever it may claim. OcUK testing it as fine lends credence to this theory, but it's still fairly unlikely.

I think it's most likely to be a fault with the bundle. The processor is probably fine, motherboard may well be faulty, likewise ram. Depends how much faith you put in the testing done by ocuk I suppose.
 
Back
Top Bottom