Brand New Build - Display turns off, GPU fans to 100%

[apicomplexa]

Hey, everyone.

I just finished putting together the following components and have ran into the above problem:

Gigabyte UD5h Z77X mobo
3770k
G.skill 2x4gb ddr3 2400
Radeon 6950 (from my old build)
Creative Titanium HD soundcard
H100i cooler

I'm using my 1000w Seasonic platinum psu and SSD's/HDD from my old build as well. I can't seem to pinpoint whats causing this problem. It occured the first time I tried to power up the system - went to onboard graphics for a few and finally got my actual 6950 working. It seems to be cutting off randomly - I was able to run both the cpu/gpu at 100% all of last night... and left it on idle today with no problems.

Things I have tried:

initially thought I might have wired too many fans into my controller or too many cathodes on 1 molex... unplugged them all except for the ones on CPU and it went well (this was when I was able to test all night). However, even with them unplugged I just had the problem occur again.

I plugged all the fans in, turned all cathodes on, all fans to 100% via my fan controller, prime95 for cpu and kombuster for gpu and its all running now without a problem

One thing I have done is remove the sound card, as it was in the PCI slot right above the graphics card and only left about 1.5mm of space between the back of the graphics card and the black hard cover front of the Titanium HD. No problems as of yet, but I anticipate it to begin again anytime. usually turning off the computer and jossling the graphics card will fix it, however, I don't think its a loose connection because it has occured both when I am slightly moving the card in the slot as its on too see a response... and when very security screwed into place.

Could being so close to the sound card, and having the heat trapped cause the card to shut off like that? My display goes off but the system still runs - the PSU does not trigger an entire systems shutdown. If the rear of the graphics card touches the Titanium's plastic cover while under load interfere?

All driver's are up to date - I haven't flashed the bios on the mb yet, as I just got it and the revision should be quite recent - or so I figure.

Also I must note that I have an Arctic Cooling Twin Turbo II on the graphics card, so even at max it rarely breaches 60c... and also this initially started happening upon my first power up of the system - everything was ice cold.

Any ideas?

 

[sygyzy]

Are you starting up a game when this happens? In other words, something that stresses the GPU.

 

[apicomplexa]

Thanks for the quick response! No, I haven't been able to tie any game to the problem. Like I said I've been able to run MSI Kombuster that simulates 100% stress on the GPU all night and then again today, it still just shuts off randomly. It did shut off once as I was 7 minutes or so into Crysis 2, but it has also done it when I am sitting at my desktop after a fresh start.

Also, out of frustration and to test my heat-buildup-because-of-soundcard theory, I just literally laid the soundcard on the back of the GPU and all went well for about 20 minutes of combuster. Stupid, but I think I can eliminate that as being a potential problem.

Also - I'm pretty sure it has to be GPU related because its only the GPU that loses connection with my monitor and causes the blackout. That happens, then GPU fans immediately go to 100% (so I assume its still receiving power from the mb or psu?) and my entire system -fans, cathodes, etc. stay on. I've had a system power down before by accidently plugging/unplugging something while power was live, but this isn't it.

 

[badb0y]

Check the power connectors going to the GPU. Something similar happens to me where my monitor went to a black screen and my GPU started blowing air like crazy. Upon closer inspection it turned out that one the 6-pin power connectors was loose and causing issues.

 

[tviceman]

Do what badboy recommends, and also reseat the GPU, make sure it is firmly in the PCIe slot.

 

[apicomplexa]

Checked the connectors and reseated the GPU - all looks well at the moment. My fear is this is going to happen very intermittently until I run out of item return time on my new purchases. I wish i could pinpoint whats causing the problem. ;/

 

[apicomplexa]

My worst fear is that it is the motherboard. My PSU is modular, so if it was a cable, getting a replacement wouldn't be an issue. If I AM able to run graphical benchmarks overnight with no problems, could it still be caused by the mobo when it randomly shuts off? I don't receive an error that I know of (the UD5H has the little red LED that spits out error codes in the top right). Also, sound card has been back in for a few hours and no loss of connection has occured.

I've tried to switch to a different PCI-E slot, but it happens so rarely that I can't really tell. Would there be any settings that I should check on in the BIOS? There's usually not a lot for the gpu in there. ;/

The strangest thing to me is that this problem was a lot worse when I first booted up the system for the first time. I couldn't get the GPU to detect at all.... driver issue maybe - with the mobo? (I did install using the disk - I think shortly after that is when I got it semi-working for a bit) I don't know if a BIOS flash of the mobo would be appropriate just yet.

 

[BD231]

Probably just a connection issue like others have stated, time will tell. Report back if it happens again.

 

[apicomplexa]

I will. Thanks everyone.

 

[apicomplexa]

Problem just happened again tonight, roughly 10 minutes in a Civ 5 game. I just did a clean install last night with all new drivers for everything.

 

[tviceman]

Does your GPU normally power the fans up all the way for a second or two when you first turn the computer on? EDIT: It sounds like a motherboard problem to me.

 

[apicomplexa]

I actually just thought of that earlier tonight. When my comp first starts, the gpu fans do go to 100% for just a split second and then back down to normal. Since I'm losing connection to my screen, I can't plug it into onboard graphics or anything to see whats going on - I believe that instead of just my GPU going out, this is a full on crash that just leaves the fans/LEDS in my case still running, making me think that its just my display that went out.

So far, I've tried my old ram from my other system just in case that was the problem - problem still persists. Also, the PCI slot (for my Titanium HD soundcard) right above the PCI-E (for 6950 gpu) just went out it seems. I can get the sound card to detect in both of the other 2 pci slots, but not the one above the motherboard.

What do you guys think? - If I did decide to rma the motherboard... do you think I should go through newegg or Gigabyte? I'm sure newegg would make me send the board back first before having a new one sent out (which would put me without a computer for over a week - with finals around the corner). I wonder how gigabyte would handle the RMA? Any experiences? Or is it even the mobo? Thanks again

 

[Eureka]

Newegg will probably get it to you faster and with less complications.