Motherboard causing crashes?

Colzy

Colzy

Colzy

Associate
Joined
31 Jan 2014
Posts
157
Location
Bristol, UK
Hi, I have been battling with an unstable PC for a while now and I was hoping the experts over here at OCUK might be able to help :)

I'll start by describing the problem, and go on to the specs and all the things I've tried which is directing me to a motherboard issue.

The PC will crash under load, usually in fairly graphics intensive games - could be Civ V, BF4, Saints Row, Endless space. It's a hard crash - screen goes black and requires a hard reboot. Can crash a few minutes in or an hour in. Seems stable in desktop conditions.

Specs:
Motherboard: GA-Z77-D3H (rev. 1.0); latest bios
CPU: Intel ivybridge i5 3570K
RAM: Corsair dual channel 2x4gb 1600MHz
PSU: Corsair CXM 750W
Graphics: Radeon HD 7850
Others: 250GB SSD, 1TB HDD, DVD drive
I haven't done any overclocking.

Things I've tried:
The most obvious thing was potential overheating. I've been running temp checks on both GPU and CPU, neither are problematic and will crash even when they are fairly low, when a game has only been running a few mins. Tried with side off case in the middle of winter etc. I just don't think thats it.

Software: Reinstall windows 7, and now I've installed Win 10. Try about 6 different ATI drivers (currently on the latest). Tried many different BIOS versions. Updated all hardware drivers. I think I've covered all bases here.

Memory: Tried memory in any combo of slots, with one or two sticks. Unless both sticks are iffy, I think its OK.

PSU: Upgraded to from an older 550W supply to the current 750W supply. No change.

Graphics:
This is the sticking point. the HD 7850 causes a hard crash, but I've tried two brand new 280Xs (gigabyte and MSI) and both showed a lot of artifacting, even after short periods. The machine did not hard crash with these cards. These cards are designed to run hot but I struggle to believe both would artifact in open case, free airflow conditions. I believe they also have a thermal throttle if they start overheating.

CPU: This (with the mobo) is the only thing I haven't been able to check. I've reseated it and re-pasted it with the heatsink. Temps are fine under 80C under load (stock)

BIOS:
I have tried SO many settings and versions as to be ridiculous now. Stuff from various threads, memory timings, CPU states, god knows what else. Even tried disabling onboard sound.

Others:
The SSD is new(ish). I've had windows on both HDD and the SSD, no difference. I've of course reseated everything, and there is minimal dust in the case.

So my question is: If I have ruled out all these possibilities (tell me if you don't think I have), is it possible to do any tests to see if there is an issue with the mobo? Could it indeed be causing the crash, and the artifacting I mention in the 280X cards?

Thank you for taking the time to read through this,

Col

NOTE: I no longer have access to the 280X cards.
 
tried setting the power limit to 20% in catalyst/overdrive for the 7850?

also download whocrashed and analyse to see what caused the crash
 
The overdrive is on default (I can't remember the exact setting). I have in the past underclocked the card to see if that had an effect - it still crashed. Still, I can give it a shot.

I'll grab whocrashed and see if that gives any further insight when I get home.
 
Sorry for the delay, had a trip to London get in the way.

the +20% power didn't work at all unfortunately, and whocrashed actually isn't reporting anything oddly.

However, I've just upped the vcore a touch (to 1.3) and haven't had a crash in over an hour, so seeing how that goes - ofcourse its pretty random so might not be fixed. I'll report back...
 
No nothing - I'll check the settings (log locations etc) tonight but I'd expect the defaults to be correct (unless they differ in Win10?)

Incidentally no crash in ~3hrs gaming last night with the vcore change. I think I got the idea for changing that from an old thread that you were helping a chap on :)

PS. Sorry this seems to have deviated away from motherboards somewhat. Mods - feel free to move it if appropriate.
 
Last edited:
Ive had same problem on my old 2500k shuttle xpc system for a long time random crashing during heavy gaming runs perfect other times will try out Whocrashed my windows error logging says something about power forget now need to try some fault finding been busy/lazy only tried different gpu card and got a psu upgrade reinstall a fresh OS changed SSD have a lot of other things still to check.

Not overclocking anything and its not heat related thinking its the rubbish low budget motherboard that came with my shuttle which I'm stuck with all drivers an bios are up to date.
 
Last edited:
If you are still having problems, I'd be tempted to put the memory through 8+ passes of memtest.

It sounds like you may have fixed it already though (fingers crossed :))
 
It seems not unfortunately :(

Just had another hard lock after being in BF4 for just a few minutes. Strange after being in it a few hours last night! Whocrashed reports nothing at all (I checked, it does work).

So that puts out the +20% gfx card voltage and upping the vcore to 1.3. I've already run memtest for several hours in the past with no errors, although I could always run it again.

This issue has me absolutely baffled - and I'm back to 'is the mobo faulty'. :(

Any way I can check anything else?

Thanks for the tips,

Col

Edit - bit more info on the RAM. its Corsair XMS 3, 2x 4gb dual channel 9-9-9-24. I've tried them singularly, in diff slots, with XMP profile, without.. they were one of the first things I tried to test. ideas welcome though I may well have missed something.
 
Last edited:
wazza300 said:
what is the cpu clocked at?

bf4 is cpu heavy so might show up unstable cpu/not enough cpu voltage
Click to expand...

It's clocked at default 3570K settings, so 3.4GHz boost to 3.8 GHz. I overvolted it to 1.3V earlier as I mentioned, but not sure how high I should go..

Worth noting the machine does also crash in more mundane games like Civ5, endless space, saints row 3, although possibly less often; hard to say due to the random nature of it.

Edit - looking at the interwebs 1.3 is pretty high and used for overclocks, should be stable much lower :/
 
Last edited:
It should be fine on stock voltage tbh

Did you try with CPU/vtt at 1.08v? And set dram voltage to 1.65v

Might be worth looking in aida64 or hwmonitor to see if your PSU 12v and 5v are in spec
 
OK, just upped the vtt. I reset the core to defaults though. RAM is already at 1.65V as that's the specification on it.

HW monitor shows everything running well and voltages correct.

I'll give this a test with the new vtt see if we get anywhere or if the voltages move.

Cheers!
 
Brief update. Still no luck on the fix, tried running RAM at lower specs (1333) but still at the rated 1.65v. no luck.

I'm really at a loss here, running out of ideas.
 
You must log in or register to reply here.
Back
Top Bottom