Hi,

I bought an MSI gtx970 OC , (us.msi.com/Graphics-card/GTX-970-4GD5-OC.html#hero-overview) about 6 months ago when it was still winter from pcandparts.com

From the day I got it it ran at 80 C , so I tweaked the fan curves in afterburner to drop it to 69~73 C in the winter , but that is while the fans run at around 70% (linear : 60% for 60C , 70% for 70C ...). It seemed wrong to me that I had to run the fans that high all the time during winter.

Now during the summer, I'm struggling to keep it under 80 C ,the temp at which it throttles down.

Initially, I had 2 fans, one front intake and one exhaust at the back . I tried adding a 3rd fan as side intake, then I tried that same fan as bottom intake. Nothing worked, the only thing that kinda worked is keeping the case open, which dropped it to about 78C.

The card is still under warranty, but I want to make sure It's my setup not the card. It could always be a thermal paste issue even if the card is new, but I don't want to void the warranty by re-applying paste.

Any idea what I should do?

Thank you
Since opening the case helped with the temps my guess would be an airflow issue, what fans are you using in your case?
However it still shouldn't run that hot... Have you overclocked it or changed any parameters in softwares like afterburner/precision X? And what exactly do you mean by struggling to keep it under 80? What temperatures are we talking about?
TjMax for cards like the 970 is 95 degrees, anything below that will not throttle the card.
He told you, he bought a GTX970 OC - stands for OverClocked...?
EzikMezik wroteSince opening the case helped with the temps my guess would be an airflow issue, what fans are you using in your case?
However it still shouldn't run that hot... Have you overclocked it or changed any parameters in softwares like afterburner/precision X? And what exactly do you mean by struggling to keep it under 80? What temperatures are we talking about?
TjMax for cards like the 970 is 95 degrees, anything below that will not throttle the card.
The exhaust fan is a cooler master fan, and the intake is a Thermaltake.
The case is a mid tower cooler master
Temps I'm talking about are 77~80C , It doesn't go above 80 because that's the temp limit I set in afterburner, it throttles when it goes over that.

I know that anything below 95 is "safe" but still like you said it shouldn't go up to 80C without me OC'ing it.
Johnaudi wroteHe told you, he bought a GTX970 OC - stands for OverClocked...?
OC stands for factory overclocked, he can still push it further himself.

What I thought you meant by "keeping it under 80C" was that it reached temperatures above 80. But you confirmed that it just stays there which is perfectly acceptable.

What I know is that GPU boost 2.0 is known to boost the clocks as high as they can go until they reach a temperature threshold, which was 80C in the case of my 770, and then will throttle down from there.

I looked around for a bit and it turn out that's the 970 works very similar to my 770, both use GPU boost 2.0 and will behave in the same fashion.

https://www.reddit.com/r/buildapc/comments/3dmdck/build_help_gtx_970_working_at_80%C2%BAc_after_playing/
http://www.tomshardware.com/answers/id-2709218/gtx-970-overheating-game-70c-80c.html

However what I would do opt for a better airflow, so more/better fans in your case should make the temps somewhat better
Your card is overclocked and running above the standard clock rate set by nVidia. Whether that overclock was set by you or MSI does not change that fact. If your priority is to keep the card cool, lowering the clock to stock will allow you to lower the voltage, which will make a significant difference in the heat output of the card. I will admit that your temps are unusually high, however every single card is different (and by that, I mean every single unit, not every model) and it might be that you just got the short end of the straw.

Lower your voltages, clean out the fans on the card and on your chassis, and make sure your room's temperatures are fine. I do not think PCandParts will accept your RMA.
This is a blower design cooler, so it takes the air from the inside of the case and exhausts from the back, make sure the intake front fan is not blocked by cables or HDD, most of cases has fix HDD cage that will play a big role in blocking airflow.
At the end that the max what you can do, the blower cooler is known for its low performance, thats why models with such cooler are cheaper I guess
Anyway, check this video by linustechtip that test the temps with different case fans configuration
https://www.youtube.com/watch?v=8OmkmluAYAQ
tt400 wrote Lower your voltages, clean out the fans on the card and on your chassis, and make sure your room's temperatures are fine. I do not think PCandParts will accept your RMA.

I email-ed pcandparts and they are whining about my power supply. I know I don't have the best PSU out there, but I monintored the card and it's running fine except for the heat. I can feel air coming out the exhaust and I don't think PSU has anything to do with heat here.
They said they can take it in for testing, but as usual I'm sure it will result in me paying for shipping.

What do you think?

*I know my psu isn't great but I pay for uni and I can't afford a new one in the time being
I'm not sure if a bad PSU cause overheating, I suggest to get some readings about voltage, fan and max temperature on load, install Furmark and run it and monitor the value inside MSI Afterburner, a GTX970 would not kill itself, if it overheats to a dangerous level it will auto decrease its core frequency. The last solution is to down-clock it, reduce its core for 100mhz to 150mhz, it will still perform well for 1080p
I always monitor it, and the down-clocking it does that on it's own so I don't think I need to do it. But I just find it weird that they had to say it's the PSU. I will try taking it to a friend who has a thermaltake 700watts PSU . If that doesn't help, I'll try changing the thermal paste. I don't even care about their warranty since they'll charge 20% of it's price. I'd rather have a card working at 80 degrees that clocks 200+ mhz from it's stock clock than pay 20% (of 400$) for nothing.
I recall reading that items not receiving optimal power can overheat. This may explain why a low end PSU can cause overheating.

A PSU can also overheat itself and with case fans rotating air around, neighboring component temperatures will increase. Check if your PSU fan is working properly.
6 days later
I also have the exact same part of PCandParts. I've been using it for a couple months, it always gets to 80C in intensive games but never higher. I thought that was supposed to be normal?
What is your PSU ?

Inadequate power can lead to an overheating GPU and can also lead to permanent damage to other components as well !

If i were you i won't risk my computer, till you get a decent PSU drop your GPU's base clock to the stock 1050mhz and eliminate your CPU's OC "if any", doing so will reduce the stress on your PSU and lead to a significant drop in temps..