Fan at 100% and card still overheats :(

H

Hooligan

Gawd
Joined
Aug 7, 2001
Messages
834
Hi all,

Recently I purchased Borderlands on Steam cuz it was on sale... Naturally I fired it up and set resolution to 1920x1200, Antialiasing at 16x and Shadows enabled.

The game looked pretty good and ran smoothly with 0 issues. About 5 minutes into the game, however, I notice on my G15 keyboard (using eVGA Precision) that my eVGA GeForce GTX 285 is running really hot (80 degrees celcius) and the fan is at 100%.

Artifacts start appearing on screen, card temperature goes up to 84 degrees celcius and then the PC locks up.... have to restart hitting the button...

Once restarted, I disabled Shadows and lowered Antialiasing down to 4x. Now temperatures stay between 76 and 79 degrees celcius in game while the fan is still at 100%.

This obviously makes me pretty bummed, since I saw that the card is fully capable of running the game with the GFX settings maxed out with no stutter or FPS loss, but can't because it overheats, artifacts on screen and eventual freeze up.

I am running the latest nVidia drivers (fresh Win7 64 bit install last weekend and downloaded drivers off nVidia website the same day)....

I have a Rocketfish full sized aluminum tower case... It's got some air flow (1 intake fan in front of the hard drive cages and 2 exhaust fans in the back

I do however, have 5 hard drives in the cages (including 2 WD Velociraptors) so the intake air might be a bit warm... but theres tons of open space on the case, since I have no other PCI cards other than the GTX 285.

Here's a pic of the case (not mine, but very similar) - the square on the side is an exhaust fan
IMG_0355.jpg


what can I do to be able to play this game w- everything enabled?
 
Maybe check the cooler on the card and see if it is dusty? Make sure the fan is really spinning up when it thinks it is - in case the fan is going bad or something. Try it with the side panel off and a fan blowing in, just to make sure it is/isn't a case cooling problem?
 
There might be something up with the card as a whole, artifacting and locking up/rebooting the entire PC at "only" 84C is uncharacteristically low; they're rated for 105C and should be able to run artifact-free up to around 100C (not saying you want it to be there of course, just that it is supposed to be able to handle it without issues).
 
I was able to play for 2 hours with Shadows disabled and lowered Antialiasing while temps stayed under 80 degrees Celsius.

As soon as I enabled shadows and bumped AA up, temperatures over 80 degrees caused artifacting and I preemptively shut down the game to avoid another lock up.
 
temps seem similar to mine, and i have to run mine at 100% fan while playing also, or it will artifact...but it won't crash either way. Ya got a clean it before rma'ing so you might as well try it
 
Just because the fan is at 100% doesn't mean that heat transfer is working properly. You may have a cooler full of dust (interferes with heat transfer from the sink to the air), or you may have a TIM problem (limiting heat transfer from the chip to the sink), or a combination of both.

I play Borderlands @1920x1200 with everything maxed out on my GTX260OC and I don't even hear the fan ramp up noticeably. If I think of it I'll run some instrumentation next time.
 
My cousin also has a 285 that artefacts. As others have said, clean out the dust, change the thermal paste and thermal pads, check your drivers, reseat the card, check the power cables and so on. There's a bucket load of things that could be causing issues.
 
I'm with trombe here. I had a GTX 280 overclocked like crazy and flashed with a custom bios and the card would heat up to over 85C and give no problems. The card is rated to run much hotter than that.

1. Check card's cooler for dust and clear out if any.

2. RMA card possibly faulty

3. Remove cooler, take apart and clean thoroughly, and also apply new TIM while your at it if card has no warranty
 
Perhaps I'm way off base here but if that big square blocking the GPU is an exhaust fan then maybe it's starving the card of fresh air, if it's possible I would try using the fan as an intake fan or if you can't reverse it just turn it off to see if that helps.
 
Ebernanut said:
Perhaps I'm way off base here but if that big square blocking the GPU is an exhaust fan then maybe it's starving the card of fresh air, if it's possible I would try using the fan as an intake fan or if you can't reverse it just turn it off to see if that helps.
Click to expand...

strange, I didn't notice that but good suggestion, the other thing is how the hell can that fan do either exhaust or intake. Now that I look at it, it's covered or is there something I'm missing. In the case that I'm on crack and there is a fan there and it's not covered and pointless in either configuration definitely change it to an intake. If that helps get a more powerful fan for the back of the case one that has high cfm's but isn't too loud like a cooler master excaliber etc... this way you can dump alot more hot air out of your case as it rises.

Edit, Okay so looking at it closer it looks like it's the rad for the water cooling enclosed loop you got going to the proc. In that case if it is blowing hot air on the video card then that's not going to cut it and could very well be causing the problem. The last place you want that is near your gpu.
 
Borderlands doesn't have AA, you mean AF (unless you are enabling AA via the nvidia control panel, which I wouldn't do as it doesn't really add much for the increased load).

But a GTX200 shouldn't lock up at 85c.

What PSU are you using out of intrest?
 
yeah, these are unusual circumstances. i also run a GTX260SC EVGA but i never get any hotter than 69c at the same settings. i also have an air compressor that i use monthly on my system to blowout all the dust. my fan also only runs at 70% manual setting. the 285 has had alot of reports of being a loosely screwed card when it comes to the heat sinks. Many have just taken it apart and checked over the screws to find them loose and not having the heat sinks well secured.
 
DeathPrincess said:
Borderlands doesn't have AA, you mean AF (unless you are enabling AA via the nvidia control panel, which I wouldn't do as it doesn't really add much for the increased load).

But a GTX200 shouldn't lock up at 85c.

What PSU are you using out of intrest?
Click to expand...

ok it might be AF.. its a setting in game that goes 2x, 4x, 8x, 16x... figured it was AA... might be AF... not sure Im at work right now.

I have an Antec 750 watt PSU
 
Also, I am not sure if I should remove the heatsink to replace the thermal compound... from what I've seen/heard online, is that some video cards use a thick thermal compound on some of the memory chips because there is some space between heatsink and chip... so if I remove the heatsink to put some artic silver then there might not be contact between heatsink and chip.
 
Hooligan said:
Also, I am not sure if I should remove the heatsink to replace the thermal compound... from what I've seen/heard online, is that some video cards use a thick thermal compound on some of the memory chips because there is some space between heatsink and chip... so if I remove the heatsink to put some artic silver then there might not be contact between heatsink and chip.
Click to expand...

From my post before, can you say if that metal square blocking your videocard is the radiator for your enclosed loop and if it is blowing hot air on your videocard? If so flip in the other way to blow the air out of the side of the case and not directly on your videocard. Does you use your case with the side panel on or off? Does the side pannel have a spot for a fan?

Edit: Nevermind, I just saw that the pic is not your actual setup. Sorry bout that
 
Hooligan said:
Also, I am not sure if I should remove the heatsink to replace the thermal compound... from what I've seen/heard online, is that some video cards use a thick thermal compound on some of the memory chips because there is some space between heatsink and chip... so if I remove the heatsink to put some artic silver then there might not be contact between heatsink and chip.
Click to expand...

I suggested this if you don't have a warranty and thus no other option. It's a better option than throwing the card away or replacing it if it's not absolutely necessary. That card is ample to run borderlands. If your card has no warranty options and you are at your wits end then I think it's worth a shot.
 
Lord_Exodia said:
strange, I didn't notice that but good suggestion, the other thing is how the hell can that fan do either exhaust or intake. Now that I look at it, it's covered or is there something I'm missing. In the case that I'm on crack and there is a fan there and it's not covered and pointless in either configuration definitely change it to an intake. If that helps get a more powerful fan for the back of the case one that has high cfm's but isn't too loud like a cooler master excaliber etc... this way you can dump alot more hot air out of your case as it rises.

Edit, Okay so looking at it closer it looks like it's the rad for the water cooling enclosed loop you got going to the proc. In that case if it is blowing hot air on the video card then that's not going to cut it and could very well be causing the problem. The last place you want that is near your gpu.
Click to expand...

The GTX285 does not draw air from the side... the card has no vents on the side (the ribs you see in the pic below are not vents), it draws air from outside the case - below the DVI connectors.

Oi7jQ.jpg
 
Hooligan said:
The GTX285 does not draw air from the side... the card has no vents on the side (the ribs you see in the pic below are not vents), it draws air from outside the case - below the DVI connectors.

Oi7jQ.jpg
Click to expand...

No.

It draws air from the bottom through the fan, and exhausts it out the back.
 
yeah. you're right. im dumb. anyways - my point was that the side exhaust fan does not starve it from air.
 
you are missing the point, the gpu fan needs fresh air, so nothing should be blocking it inside the case.
 
Also, I am not sure if I should remove the heatsink to replace the thermal compound... from what I've seen/heard online, is that some video cards use a thick thermal compound on some of the memory chips because there is some space between heatsink and chip... so if I remove the heatsink to put some artic silver then there might not be contact between heatsink and chip.
Click to expand...

Just wanted to share with you: I own a GTX285 (ASUS Matrix) and during gaming it would heat up until about 80C. Around 80C it would ALWAYS lockup my PC! So today I replaced the thermal compound (used Arctic MX-4) and removed all dust (there was some in the fan and on the PCB) and now it is at 73C max! And no lockups yet! So maybe you could give it a try also!
 
zerounleashednl said:
Just wanted to share with you: I own a GTX285 (ASUS Matrix) and during gaming it would heat up until about 80C. Around 80C it would ALWAYS lockup my PC! So today I replaced the thermal compound (used Arctic MX-4) and removed all dust (there was some in the fan and on the PCB) and now it is at 73C max! And no lockups yet! So maybe you could give it a try also!
Click to expand...

Ill pick up some canned air on the way home from work today and give it a shot. Thanks for your input
 
well god dammit.

I took that beast apart and lo and behold, there are thick ass thermal pads on all the chips... and when I removed the heatsink, some of them tore a bit.

So i cleaned the copper heatsink and gpu and applied some thermaltake paste i had (the chip is huge, had to use all of the paste).

I just booted up windows and temp is 53 at idle without the case closed... fan at 40%... i dont remember what idle temps used to be. will load up borderlands now... just wanted to vent my frustrations at my previous fears becoming a reality when noticing thermal pads :(

edit:

Update: OK much improvement in game.... 73 degrees tops instead of 80. Thats without the side exhaust fan or side case cover.

will update soon w/ fan and cover on.

edit #2:

case cover and exhaust fan back on and temps went up to 78 max.... well at least its still a slight improvement over 80
 
Last edited:
If you didn't replace the thermal pads (it was unclear from your post), make sure the memory chips are still touching the heatsink. They use those thick pads because there is often a larger gap between the memory chips and heatsink, so they need to fill the space.
 
Forceman said:
If you didn't replace the thermal pads (it was unclear from your post), make sure the memory chips are still touching the heatsink. They use those thick pads because there is often a larger gap between the memory chips and heatsink, so they need to fill the space.
Click to expand...

I didnt replace them because I have none to replace them with :(
 
Snapped this pic after cleaning off the paste off the GPU w/ nail polish remover and cotton balls - you can see how some of the thick pads tore a bit :(

bw8fc.jpg
 
Those thermal pads look really... uhhh... old and dried-out... :eek:

The thermal pads on my GTX285 are grey and look more fresh and alive... :(
 
Most of the heat come from the GPU anyway, I did the same with my 295GTX (cleaned the goo and applied Arctic Silver Ceramique and cleaned the solid carpet of dust between the fan and the fins of the heatsink) and the pads on the memory look the same and the temperature droped from 105c to 70-73 tops in Dawn of War 2 (witch load my 4 GPU core to 95% utilisation).

If you don't overclock the memory I doubt you'll have any problems.
 
That sounds exactly like when my 8800GTX died. It just kept overheating for no reason. Fan was clean, new paste, etc. didn't help. The core was just overheating all the time because the card was dying.
 
Once you do all the cleaning and reapplying of thermal paste and it still does not work. Try this: Lower your core/mem speeds a bit and see if that makes any difference. I would be liberal about it at first so you can see if it really does make any difference so maybe try to take 200-300mhz off both core and mem. I had an issue with a card where after 10-15mins games would crash, I tried everything, what helped me was lowering the clock speeds which stabilized the card. Sure you will lose some performance, but you have no warranty so if you can get a working 280 out of your 285 it maybe worth it. Let us know what happens and good luck
 
Spare-Flair said:
The core was just overheating all the time because the card was dying.
Click to expand...

Chip don't just get hotter because they are old. There was a problem getting heat to the sink, or from the sink to the air. Maybe the fan wasn't turning properly, or whatever, but it isn't just "old" so therefore hot.
 
You must log in or register to reply here.
Back
Top