New build crashes randomly, frequently - help?

Calica

Distinguished
Dec 20, 2011
15
0
18,510
Making a new thread for this because the old one is under graphics cards, and I don't think it belongs there. The old thread is here: http://www.tomshardware.com/forum/341317-33-radeon-6950-output

I'm getting frequent, random crashes on my new build. Sometimes it freezes at "Windows is starting," other times it crashes a few minutes into a session. Steps taken so far:

Updated the BIOS, because hey, why not.

At first I thought it might be the graphics card, so I removed that and tried booting up. Computer crashed at startup. Took the motherboard out of the case to rule out a short, and tried booting up with nothing attached except the RAM. This worked fine. Hooked up the hard drive (which was still mounted in the case) and tried booting again. This also worked fine - Windows started normally and everything. I installed 3DMark and PCMark from a USB drive, but didn't try to run either one because it was late in the day.

Next day: Installed the graphics card (motherboard is still sitting on the table at this point). Attempted to boot; crash on startup. Second attempt to boot got me into Windows, but the computer crashed a few minutes into PCMark. Removed the graphics card; successfully booted into Windows, but again, crashed a minute or so into PCMark. I had CoreTemp running, and it showed temperatures under 40 C the whole time. So yesterday, it appeared to work without the graphics card; today it doesn't. Must have been good/bad luck?

When it crashes, it crashes either right at startup (after POST), or less than five minutes after boot. Yesterday, I was puttering around on it for at least half an hour, and it was fine. Crashes don't just happen when I'm stressing the CPU; they also happen when I'm just using Firefox or something.

It's never crashed when I'm in the motherboard settings, but that could just be because it hasn't happened yet.

I've tried the sticks of memory separately and didn't see crashing, but I don't know if that's because the memory is fine, or just because it didn't crash THIS time.

Bad motherboard, bad PSU, bad CPU, bad memory, or some evil combination of the above? About all I can say right now is that it probably isn't the graphics card, since it crashes whether or not the card is installed, but of course I can't prove that the card isn't also damaged. Also, yesterday when it was miraculously crash-free, I ran disk check and it came up clean.

I've got no spare parts whatsoever, so I can't see if parts work in a different system or vice versa. I'm lost. Help?

Motherboard: ASRock Z68 Extreme3 Gen3
Graphics: AMD Radeon 6950 1 GB
Hard drive: Western Digital Caviar Black 500 GB
Memory: 2x GSkill 4GB Sniper Low Voltage Series
PSU: PC Power & Cooling Silencer Mk III 500W
Processor: Intel Core i5 2500K, not overclocked
OS: Windows 7 Professional
 

aqe040466

Distinguished
Nov 29, 2011
886
1
19,065



Do you have an Anti-Virus installed?
 

Calica

Distinguished
Dec 20, 2011
15
0
18,510


Oh man I would be so happy if that were the problem. Trying that now.

Edit: No antivirus, no. I haven't had enough time between crashes to do that.
 

aqe040466

Distinguished
Nov 29, 2011
886
1
19,065



Try Formatting your Hard drive and Reinstall windows and all the drivers that comes with your MOBO and an Anti-virus. And then reboot your system.
 
I'd say it's almost certainly a RAM issue. Probably the leading cause of "random" crashing with a new system. Especially with the hang in memtest86+.

Either the mobo isn't properly recognizing the RAM, the mobo has a bad memory slot, or you have a bad stick of RAM. Find out which one it is.
 

Calica

Distinguished
Dec 20, 2011
15
0
18,510

Sorry, but how do I do that? Memtest86+ completed with no errors on the second try; does that mean I can't trust it to identify bad RAM? If so, how do I figure out which it is?

(Yep, I'm frustrated, but not with anyone here. I'm very grateful to everyone who's taking the time to read and respond! Thank you!)
 
A stick of RAM can pass memtest if the settings are wrong but nothing is physically wrong with the RAM itself. So first thing, make sure the voltage and timings are correct in the BIOS; try setting them manually to the manufacturer specs instead of "auto" and see if that helps. Who knows for sure, but on default settings, the board could be doing one or the other wrong. Perhaps with this "low voltage" RAM, you have a set that takes 1.25V or 1.35V and the board won't go below 1.5V on auto, and the difference is too much ... so sooner or later it goes haywire. A lot of boards give an allowance upward on auto (e.g from 1.5V-1.65V), but the standard for DDR3 is 1.5V and requiring less than 1.5V is not a usual thing, so it's quite likely some boards don't account for it without manual intervention. It's impossible to know what your sticks require without posting the actual specs of that particular model of RAM, so I really hope you do that.

I would also recommend running a few more passes with memtest. One clean pass does not necessarily mean everything is a-ok. If there was an error message EVER, it indicates a definite defect. A hangup without an actual error message might just be the timings or voltage, or it might not. That's why when you suspect a RAM issue, they recommend running several passes of memtest until you have a definite idea of whether it's clean or dirty.