P8Z77-V LX has me baffled

[Kivorn]

I recently built the following computer:

P877-V LX
3450 ivy bridge
xfx 6870
8GB PC1333 Viper RAM
800 watt Thermaltake PSU
PNY 128GB SSD

Here is the problem. The computer runs fine at idle, browses the web. Idle temperatures are low on both the CPU, MB sensor and the GPU. When the computer is cold everythign runs fine. When it warms up I start getting black screens while playing 3d games or running benchmarks, etc.

For instance on a cold boot, I can start running a 3d game and all is fine. A few minutes later the screen will go black, but everything else stays fine, a few seconds later the screen will return to normal and this will keep happening. The longer the computer is running, the more frequent this happens. It will eventually get to where it does this immediately upon engaging anything 3d...and then eventually I will get the screen going black and then coming back EVEN while staring at the bios (hence no software issues).

Things I've done :

Pulled each stick of RAM, swapped each to the other slot etc.. doesn't help.
Tried another GPU...same problem.
I have used the onboard video and it seems stable.

Today I moved my computer from the 75 degree F space beneath my desk to my kitchen table where it stays around 68 degrees. With the case side on it took a bit of time to encounter the black screen. So I took the side off and put a fan on it and the problem went away.

It appears to me at this point that I think I have a weak motherboard that is overly heat sensitive.

GPU temps rarely get above 55C, CPU 40C max. Also, once the computer has been running for some time the temps will be cold and immediately upon starting 3d rendering..bam, black screen.

A few times the video is lost completely turning off the monitor and requiring a reboot.

Any help will be appreciated.

edit: wanted to add that voltages on all sensors in the Asus utility all stay strong, never dipping below 5v, 12v etc.

 

[jaquith]

Lets start with simple:

Connections: Have the monitor connected directly to the HD 6870 and not the iGPU ports.

Drivers: Update to the latest AMD drivers (HD 6000 series + Windows 64-bit) - http://support.amd.com/us/gpudownload/windows/Pages/radeonaiw_vista64.aspx
Catalyst Software Suite & AMD Catalyst Application Profiles

Virtu: If running Virtu either 'Stop' or uninstall for now.

Clock Speeds: Once in a while changing the Default clock speeds can help to diagnose an issue. The Reference is "900MHz Engine Clock and 1050MHz Memory Clock (GDDR5)" try increasing the values by +50MHz each using MSI Afterburner - http://event.msi.com/vga/afterburner/download.htm

AI Suite: I don't like it and I don't recommend it. Uninstall for now, reboot, then reboot once again into the BIOS:
press F5 (Load Optimized Defaults)
AI Overclock Tuner -> XMP and Profile 1 ; if kits are XMP
press F10 (Save & Exit = Yes)

--

Failure, Uninstall the HD 6870's driver manually in the Device Manager then reboot. Re-install the AMD Drivers downloaded as shown above.

Example to Uninstall a driver

 

[Kivorn]

Curious...after the computer gets "warmed up" and starts doing this very frequently, it will continue to do this even while looking at the bios page...since no drivers are used at this point doesn't that rule out windows environment software?

Nevertheless I've already unistalled the drivers, used older drivers, reformatted and reinstalled windows 7 a couple of times already. On this reformat I haven't installed the asus AI suite (proble started before I installed anything aside from windows and world of warcraft... and yes, I've tested other 3d applications).

Also since this happened with another older video card that works perfectly fine in its respective computer I think the clock speeds on this one are ok (but I've already bumped up the clocks a little and as far down as they'll go when i first started troubleshooting to no avail).

 

[jaquith]

IF the problem occurs with ONLY the HD 6870 then it's either the GPU itself or the PSU. Voltage doesn't mean the AMPS are available. IF the HD 6870 works fine in another PC then PSU by process of elimination.

The MOBO would have to get VERY hot for the MOBO to be the culprit. Try removing the side panel and monitor the temps.

RAM:
Use the 'Blue' DIMM slots

BIOS:
press F5 (Load Optimized Defaults)
AI Overclock Tuner -> XMP and Profile 1 ; if kits are XMP
press F10 (Save & Exit = Yes)

Power connections use (2) PCIe 2 x 6 Pin connectors and not Molex to PCIe 6 pin adapters.

 

[Kivorn]

The problem occurs with my current new 6870, and with my older video card which works fine in the other computer. The older video card requires no external power source and has the same proble when placed in this computer, and I have the two PCIe connectors connected properly.

The on board video doesn't have this problem.

I have removed the side panel and with a fan blowing into the case the problem gets less and or goes away. While doing this the temps (all of them) never get anywhere near hot.

Basically if I have the ambient room temp at 68 F or below, case fan open with fan it rarely does it. If I put the case side back on (highly ventilated case btw) in my 72 degree bedroom eventually it will start doing this.

 

[jaquith]

My home office starts at 72F then in no time flat it's 75F with the PC running. However, I know folks all over the World where their room temperatures are 80F~85F and their PC's run fine.

My thought is still the PSU. Try re-seating the 24 pin main. Could it be the MOBO sure anything is possible, but typically when a MOBO runs hot the entire system becomes unstable which doesn't seem to be your situation.

IF the PSU is running hot then that could explain the problem to a degree forgive the unintended pun.

 

[Kivorn]

I enjoyed the pun, and I concur with your reasoning, I'm getting no other errors. I have a motherboard and PSU on the way from a friend, we'll get to the bottom of this and i'll let you know.

 

[jaquith]

No doubt the best way to find a problem is to eliminate all other possibilities.

Good Luck! Please let me know your findings.

 

[Kivorn]

New PSU... no help.. new motherboard (same model).. no help.

 

[jaquith]

Duh the GPU?! Just for the extra burden of 'maybe's' how about different monitor and cable.

I must admit once I had a wacko issue that was dumb on my part -- I forgot to screw-in the DVI cable to my monitor and I had a similar experience...shhh.

 

[Kivorn]

I have already tried a different monitor and cable.. fresh out of the box actually... I took the computer and monitor out to my shop and it worked fine... so I think I have an electrical problem with the wiring in my house.... go figure. I must note that it was 88 degrees ambient temp out there and ran the full furmark benchmark so I would say it's stable enough.

 

[jaquith]

Well you clearly ruled-out every single thing all the way up to an exorcism.

Glad to hear you figured it all out.