GPU fan goes 100%, screen goes black

[Alex511]

Alright so im currently putting together a new build with a couple of parts i have lying around, and my recently installed GTX 970 seems to be causing some issues.

The card was quite loud on idle with fan speed of about 50-75% up until i installed the drivers, which fixed that completely, however now whenever i launch something even slightly gpu-intensive, it's fan immediately jumps to 100% and after about 3 seconds the screen goes black, and the only solution is to reboot. Some situations this happens in is when i try to launch any game, when i have too many internet tabs/programs open, or even sometimes when i logon to Windows after startup. This makes it very difficult to get much of anything done

At first I thought maybe it's overheating and crashing, however that wouldn't make any sense since my cpu and motherboard idle at around 35°C, so I don't see how my gpu could be much higher. I also considered that maybe the card has too much power draw for the psu, but that wouldn't make much sense either, given that the 970 is around a 145W device, and it should easily be getting that.

Here's my current specs:

Win7 64-bit
8GB RAM
240GB SSD
Intel i3-2100 (soon to be replaced with i7-3770k)
ASUS P8H61-M LE/CSM motherboard (soon to be upgraded to Intel DZ75ML45K)
Nvidia Geforce GTX 970 Reference edition
EVGA 500w PSU

Anyhow, if anyone has any knowledge or suggestions about this issue, I would be very grateful for any advice

 

[Scaramorey]

I would go ahead and download something like MSI Afterburner or HW Monitor, if you haven't already, to check the gpu temp. I've never had a reference 970 or any reference card for that matter, so I don't know but I would think with it being external exhaust it might be hot and not affect other parts too much. Maybe worth a shot.

EDIT: Also here is another post on this site where a guy had a similar problem with 970
http://

Hope at least that link will help you. I'm dealing with an overheating card right now as well lol. Good luck

 

[Alex511]

I would go ahead and download something like MSI Afterburner or HW Monitor, if you haven't already, to check the gpu temp. I've never had a reference 970 or any reference card for that matter, so I don't know but I would think with it being external exhaust it might be hot and not affect other parts too much. Maybe worth a shot.

EDIT: Also here is another post on this site where a guy had a similar problem with 970
http://

Hope at least that link will help you. I'm dealing with an overheating card right now as well lol. Good luck

Alright so I went ahead and got HW Monitor to take a look at the GPU temps. Under idle it seemed fairly decent averaging around about 55°C, and I took a screenshot of the data to show the overall temps and also the CPU temps as a comparison.

IDLE:

However, when I went into a game with the HW Monitor open, the GPU temps slowly started to climb up to finally 96°C in only about 10 seconds, while the CPU temps stayed almost entirely the same. I took a picture with my phone less than a second before the screen went black, when it was at 94°C.

LOAD:

So apparently it is overheating and shutting off, although im still baffled as to how this could be, seeing how the rest of the components are perfectly fine under load, and that I have ample amount of cooling in my case to ensure this couldn't happen.

 

[Scaramorey]

What's the history of the card? How old is it? Was it one of the parts you had lying around or is it new?

Maybe the thermal paste isn't getting good contact.

 

[Alex511]

What's the history of the card? How old is it? Was it one of the parts you had lying around or is it new?

Maybe the thermal paste isn't getting good contact.

Well I bought it used from a well-rated amazon vendor that described it as being purchased and tested to boot before being resold, therefore i don't know much about it's history, but visually the card appears to be in perfect condition on the outside, though i'd still just have to guess about the interior.

 

[Scaramorey]

Well I would take it out and check if the fan or exhaust is completely clogged with dust (probably won't be since you said it looks perfect on the outside). If it is already free of dust I would open the card up, blow out any dust on the inside or in the fan with compressed air. After this I would clean off the old thermal paste and apply some good-quality paste. Over time the paste can get really hard and start to crack, greatly reducing the contact between the gpu and the cooler, which leads to horrendous gpu temps.

Not sure if you have taken apart a graphics card before, but if you haven't, watch some Youtube videos (linustechtips and jayztwocents both have videos related to this), be careful, and make sure you are comfortable with the process before you try.

Same as preparing to build a pc, which you have obviously already done so you should be fine. Good Luck

 

[Alex511]

Well I would take it out and check if the fan or exhaust is completely clogged with dust (probably won't be since you said it looks perfect on the outside). If it is already free of dust I would open the card up, blow out any dust on the inside or in the fan with compressed air. After this I would clean off the old thermal paste and apply some good-quality paste. Over time the paste can get really hard and start to crack, greatly reducing the contact between the gpu and the cooler, which leads to horrendous gpu temps.

Not sure if you have taken apart a graphics card before, but if you haven't, watch some Youtube videos (linustechtips and jayztwocents both have videos related to this), be careful, and make sure you are comfortable with the process before you try.

Same as preparing to build a pc, which you have obviously already done so you should be fine. Good Luck

Ok so I went ahead and took apart the card, cleaned off the thermal paste, and then applied some more, and after putting it back together I'm still getting the same results. Interestingly enough, I took out the GTX 660 from my previous build and tried it in this one to see if it did the same thing, but it was perfectly fine. Even under load it stayed beneath 50°C and it never crashed.

 

[Karadjgne]

55ish is very hot for idle temps, most times idle is @7-10 above ambient, so gpu should be running within a few ° C of the cpu.

Check the fan and it's motor. Make sure the fan is actually spinning and pushing air. The 970's run cool enough not to actually need a fan for most applications, with many models having eco switches that'll not turn on the fan til it hits @60-65.
When reapplying the thermal paste, did you happen to check the thermal tape on the VRM's? All of that circuitry is supposed to attach thermally to the heatsink as well as the actual processor. Wondering if that card had been used with a waterblock or liquid cooler and the tape was not reapplied before being returned to stock condition before resale.

 

[Alex511]

55ish is very hot for idle temps, most times idle is @7-10 above ambient, so gpu should be running within a few ° C of the cpu.

Check the fan and it's motor. Make sure the fan is actually spinning and pushing air. The 970's run cool enough not to actually need a fan for most applications, with many models having eco switches that'll not turn on the fan til it hits @60-65.
When reapplying the thermal paste, did you happen to check the thermal tape on the VRM's? All of that circuitry is supposed to attach thermally to the heatsink as well as the actual processor. Wondering if that card had been used with a waterblock or liquid cooler and the tape was not reapplied before being returned to stock condition before resale.

Yeah all the thermal tape seemed well in place on the VRMs when i opened up the card. Overall I didn't see much evidence of anyone having opened it up before. The GPU fan usually idles at around 1400 rpm, and im sure it doubles or even triples that when it crashes. What I decided to do was take my large room fan I've been using to cool my apartment and put it right up next to the open case, hoping to get some better results. It was blowing quite a lot of air and I managed to get the idle GPU temp down to around 49°C, but nonetheless when I started up a game, it still climbed slowly all the way back up to +90°C and crashed.