Computer Build Mystery Theater

E

eloj

2[H]4U
Joined
Dec 31, 2000
Messages
3,612
Gigabyte H55M-USB3
Intel Core i5 650
4x4GB Corsair XMS DDR3
Corsair VX550

http://bayimg.com/IakeJaAdc

What's the first thing you'll suspect?

A bad stick! Memcheck+ 4.20 those suckers!

What if I told you that moving the memory sticks around didn't change the error address?

Hmmm...

What if I told you I'd tried the memory sticks one by one and they all passed?

Wait.. maybe it's not the memory, but the MB?

What if I told you I'd tried another, albeit same model, motherboard -- with the same result?

Uhh.. bad CPU?

What if I told you I tried with another CPU?

WTH.

I've ordered another 16GB of memory (Kingston this time). If this error persists I'm at a total loss. PSU? Some sort of problem with the chassi shorting the MB? (I've double-checked the raisers)
 
Did memtest pass with just 2 sticks? With 4 sticks, you'll want to bump the vdimm a tiny bit. This is a common issue with populating all DIMMs on a mobo. Hopefully, the slight voltage bump will solve your issues.
 
I may try that if I have to, but the Kingston ValueRAM I got today just made its first full pass through memtest+ without errors.

http://yfrog.com/z/mg60ambj

This has been one instance where knowing just enough to be dangerous about the integrated memory-controller and so forth really hasn't been to my benefit.
 
The happiness lasted about 40 minutes, which is how far I got into the OS install before I realized I wasn't happy with the partitioning, restarted, and since then the machine has been spontaneously rebooting quite a lot. memcheck+ passed another 7h hours after that started.

At this point it's got to be either the PSU or a fundamental incapability of these Gigabyte boards to actually work (or something stupid with the case)

I guess I should try with just two modules first (but I don't want to run it with just two so...) and try another PSU. It's just that my only other suitable PSU is in my gaming rig and starting to tear into something that actually is WORKING reliably isn't that attractive at this point.

Might just buy a spare PSU and see.

Jebus, this is the ...... build from hell.
 
I had a similar issue that just randomly went away... that'll trip your brain-circuits a bit too.
 
I have not gone back and retested the Corsair XMS since the Kingston RAM was been behaving under memcheck+. When the machine started its rebooting behaviour I did bump the voltage, which made no difference.

Today I tried with a new PSU, a CX500 V2 running outside the case, which has actually gotten me not only through the debian install, it also behaved completely stable for the one hour I played around with it post-install before shutting it down gracefully.

If this continues to be stable -- I'll probably run some cpuburn -- I'll go back and retest the original XMS memory again.

I don't want to claim success yet, been burned too many times on this build. If it reverts to rebooting on this PSU too, I'm officially out of ideas (and a whole lot of $$$)
 
I had some seriously confusing issues with my desktop in my sig that I eventually decided was likely the PSU. My issues only showed up every few months, making diagnosis difficult. The PSU also did well on my PSU tester. I replaced just about everything else and still have problems, so it's got to be my PSU...
 
You must log in or register to reply here.
Back
Top