Overheating XFX R9 280x

[thorpj]

So my XFX R9 280X is overheating. In every demanding game that i've got, i have to lower the settings below normal, not because of frame rate, rather because temperatures begin to exceed 70C.

 

Strangely, this issue never occurred when i was running an FX-6300 with a Gigabyte GA-970A-D3P.

 

Thing is, i need some advice as to whether i have enough proof to get the GPU replaced at the store that i bought it from. They're normally quite good with warranties (and believe me, i have way too much experience with that...), but i'm worried that when i take the PC in and tell them that the GPU is overheating and leave it to them for testing, they might come back and tell me that nothing is wrong, and then i'll have to pay a fee, because they worked on it, and no fault was found.

 

I've read about a large number of overheating issues with this GPU, and many also report that XFX aren't a particularly good brand for GPUs. Something else i'm concerned with is that part of the cooler is loose (as shown in image)

 

I'm going to add a little info about temps below, but my plan is to provide any info that people ask for.

 

nb. At 1080p

Ambient Temp: 25-30 C

 

Battlefield 4: 68-70C @ Ultra, but have to limit fps to 70 and turn of Ambient Occlusion and Post AA, which is annoying.

 

Crysis 3: Manages a solid 60fps @ medium/high/very high (i tweaked the settings using tweaktown article). But was running at about 75C. Turning down to medium/high, still about 75C, locked to 60fps.

 

I've sat on this problem for a while, and it's really frustrating for me. The last PC i owned was a CrapDell with XP, and since then, a CrapBook, so the number of issues i've had with this pc have been... an experience...

 

thanks in advance!

[officiallyasimulator]

how is the airflow in your case?

[Guest]

These cards do run very hot due to their architecture. 

I'd suggest checking the case airflow.

[dragoon20005]

see if you are able to ramp up the fan speed to max in Afterburner and test if the temps run lower than what you experienced

 

actually 70 degrees is still ok for the GPU

 

you should worry if the temps goes up to 90-100 degrees on load

[thorpj]

see if you are able to ramp up the fan speed to max in Afterburner and test if the temps run lower than what you experienced

 

actually 70 degrees is still ok for the GPU

 

you should worry if the temps goes up to 90-100 degrees on load

hi, thanks for helping.

 

I'm happy with 70C, but when it goes above (ie around 75c) i become distinctly unhappy :0

[Hecknologist]

You have high temps and a loose cooler. Take off the cooler and replace TIM, put the cooler back on and check temps. If you're under 70°C = great, if under 85°C = still ok (old gpu + xfx cooler so don't be surprised), if over 85°C = check your case cooling for adequate airflow. If still not happy you could always try underclocking or contacting xfx/retailer. Also factor in your delta temps as it's the middle of summer and you're in Aus.

[thorpj]

how is the airflow in your case?

hi, thanks for helping

 

My case isn't particularly good. It's a Coolermaster Storm Enforcer. (specs are in sig). I've got a front 200mm intake, top 200mm exhaust, rear 120mm exhaust. I'd love to swap the top 200mm around to achieve positive air pressure, and see if it improves cooling, but for some reason this Bitfenix spectre 200mm only fits one way!

 

Recently, i also took these two small 80mm fans out of a laptop cooling pad, and cable tied them to fit somewhere near the front, pointing more or less at the GPUs. Totally ghetto, and totally useless - i haven't noticed a difference.

[thorpj]

You have high temps and a loose cooler. Take off the cooler and replace TIM, put the cooler back on and check temps. If you're under 70°C = great, if under 85°C = still ok (old gpu + xfx cooler so don't be surprised), if over 85°C = check your case cooling for adequate airflow. If still not happy you could always try underclocking or contacting xfx/retailer. Also factor in your delta temps as it's the middle of summer and you're in Aus.

Hi,

 

Doing that would void my warranty. I would prefer to keep the warranty, since this card has been so unreliable, i may well need it again.

 

My hope is that i can get the card replaced with a variant from a vendor like MSI or Gigabyte, or perhaps use the cash for a 970.

 

Edit: Yes, bloody Australian summers

[Hecknologist]

Hi,

 

Doing that would void my warranty. I would prefer to keep the warranty, since this card has been so unreliable, i may well need it again.

 

My hope is that i can get the card replaced with a variant from a vendor like MSI or Gigabyte, or perhaps use the cash for a 970.

 

Edit: Yes, bloody Australian summers

Yea, definitely a fair point, if you feel it would do more bad than good just leave it.

[Badger906]

70c is well below thermal cut off, start worrying when it's over 100c!

amd cards always run hot, the r9 290x runs at 95c with its stock cooler.

my 770 chugs a long all day at 80c without any issues, and even then the fans no higher than 65% on auto.

start worrying about gpu temps when it starts to crash and blue screen, not when you think it's hot!!

[Badger906]

plus it's all relative to ambient temps. if you're in a country where it's 40c and your gpu is running at 70c that's a delta of 30c, but if it's only 20c in the house and at 70c that's 50c.. but again 95c+ I'd start to worry.. not at 70

[thorpj]

70c is well below thermal cut off, start worrying when it's over 100c!

amd cards always run hot, the r9 290x runs at 95c with its stock cooler.

my 770 chugs a long all day at 80c without any issues, and even then the fans no higher than 65% on auto.

start worrying about gpu temps when it starts to crash and blue screen, not when you think it's hot!!

There's definitely some truth in what you're saying, but i'm more concerned with longevity. I dislike running the card that hot.

 

ps, AMD said the 290/x could run safely at 95. We all know how it really happened. They made the card, then found that it runs at 95. They didn't design it to run at 95c.

 

But if everyone here will agree that i should be content with the card running at 70-76c, then i'll change my views

[thorpj]

Also, should i be worried about the cooler being loose? Perhaps they'll refund because of that (getting a 970 sure would be nice!)

[Jb0nez]

Yes I know old thread, but it's number 1 on a Google search.
For my own XFX r9 280x, I've found it's not the gpu temp but the VRM temps that bring it down. Gpu can be humming along at 72C while the VRMs run minimum 10C hotter. Once either of them hit 85C, game over. Computer freezes and screen turns black. If I touch the PCB it's almost hot enough to burn my finger. It seems their "double dissipation" 3-fan cooler hits only the gpu. I've tried adding as many fans as I can, I've tried both positive and negative air pressure, nothing helps. Since the card is so massive it touches my hard drive cage (not removable in my case) so poor airflow over back. To play GTA V at 2560*1440, I have to use Speedfan and turn every fan in the computer (including GPU) to 100%, AND open the side and place a small fan blowing right into the gfx card. Even then I still crash with black screen on occasion. But for the most part the VRMs stay closer to the gpu temp. Very poor cooling design on this card. Or, poor choice for a moderate size atx case.

 

 

Edit: I should mention GPU-Z was never able to read the VRM temp correctly. Nor did numerous other free tools. I had to use Aida 64.

[don_svetlio]

It's a bit late so I skipped a bit. If it's been posted already, sorry for missing it

7970 (R9 280X) is rated at 77*C normal working temperature by AMD. In fact, IIRC, the reference designs ran even hotter. If it starts going over 80-85*C then you might have an issue. Check if fans are ramping up. Maybe something's wrong with a setting in bios that locks them to 1 speed? I'd suggest leaving them on auto and if that's not working, create your own fan curve.

EDIT:http://www.overclock.net/t/1399444/7970-manufacturers-specs-for-max-temp

[thorpj]

It's a bit late so I skipped a bit. If it's been posted already, sorry for missing it

7970 (R9 280X) is rated at 77*C normal working temperature by AMD. In fact, IIRC, the reference designs ran even hotter. If it starts going over 80-85*C then you might have an issue. Check if fans are ramping up. Maybe something's wrong with a setting in bios that locks them to 1 speed? I'd suggest leaving them on auto and if that's not working, create your own fan curve.

EDIT:http://www.overclock.net/t/1399444/7970-manufacturers-specs-for-max-temp

Hi, since i posted, i've made a fan curve in MSI afterburner, and temperaturess haven't been an issue

 

thanks

[thorpj]

Yes I know old thread, but it's number 1 on a Google search.

For my own XFX r9 280x, I've found it's not the gpu temp but the VRM temps that bring it down. Gpu can be humming along at 72C while the VRMs run minimum 10C hotter. Once either of them hit 85C, game over. Computer freezes and screen turns black. If I touch the PCB it's almost hot enough to burn my finger. It seems their "double dissipation" 3-fan cooler hits only the gpu. I've tried adding as many fans as I can, I've tried both positive and negative air pressure, nothing helps. Since the card is so massive it touches my hard drive cage (not removable in my case) so poor airflow over back. To play GTA V at 2560*1440, I have to use Speedfan and turn every fan in the computer (including GPU) to 100%, AND open the side and place a small fan blowing right into the gfx card. Even then I still crash with black screen on occasion. But for the most part the VRMs stay closer to the gpu temp. Very poor cooling design on this card. Or, poor choice for a moderate size atx case.

 

 

Edit: I should mention GPU-Z was never able to read the VRM temp correctly. Nor did numerous other free tools. I had to use Aida 64.

Try pointing a fan at the GPU

 

Below is an Antech Spotcool fan

[EcoBoost]

I've read about a large number of overheating issues with this GPU, and many also report that XFX aren't a particularly good brand for GPUs. 

Correct, but honestly as long as you're below 80-85c you're not doing any damage.

 

These cards do run very hot due to their architecture. 

No they don't. My HIS IceQ 280 on the same Tahiti architecture barely hits 65.

[don_svetlio]

Correct, but honestly as long as you're below 80-85c you're not doing any damage.

 

No they don't. My HIS IceQ 280 on the same Tahiti architecture barely hits 65.

Eh, my Asus DC2 also rarely reaches 70*C but that doesn't change the fact that the Tahiti chip produces a decent amount of heat. It all depends on how potent the cooler is and how much copper it has. From what I saw of the XFX cooler, it lacked copper

[EcoBoost]

Eh, my Asus DC2 also rarely reaches 70*C but that doesn't change the fact that the Tahiti chip produces a decent amount of heat. It all depends on how potent the cooler is and how much copper it has. From what I saw of the XFX cooler, it lacked copper

I never liked the XFX coolers, but I don't think mine has copper either, and it still works very well. Tahiti still produces a decent amount of heat (250W being what Nvidia's flagship Titan X puts out) but the engineers at HIS and Asus were able to cool these chips decently, why not XFX? Oh yeah, cause a while ago everyone was a suckup to these cards because they were cheap and "looked nice" (which IMO they don't).

[don_svetlio]

I never liked the XFX coolers, but I don't think mine has copper either, and it still works very well. Tahiti still produces a decent amount of heat (250W being what Nvidia's flagship Titan X puts out) but the engineers at HIS and Asus were able to cool these chips decently, why not XFX? Oh yeah, cause a while ago everyone was a suckup to these cards because they were cheap and "looked nice" (which IMO they don't).

Honestly, in terms of pure aesthetics, I love the reference AMD designs the most

EDIT: I should mention mine has more copper than all the electrical wires in our house lol, I can get a pretty penny if I recycle it

Btw, how are your VRM temps? Asking since mine tend to max out at about 70*C-72*C during Witcher 3 prolonged sessions (3-4 hours-ish). I know mine don't have contact with the main heatsink so was wondering whether you were fairing any better?

[Guest]

Correct, but honestly as long as you're below 80-85c you're not doing any damage.

 

No they don't. My HIS IceQ 280 on the same Tahiti architecture barely hits 65.

Silicon lottery, and it depends on so many factors.

Case airflow, aftermarket cooler, ambient temperature... it's impossible to compare one card to another unless you own both.

[winningsince1337]

I never liked the XFX coolers, but I don't think mine has copper either, and it still works very well. Tahiti still produces a decent amount of heat (250W being what Nvidia's flagship Titan X puts out) but the engineers at HIS and Asus were able to cool these chips decently, why not XFX? Oh yeah, cause a while ago everyone was a suckup to these cards because they were cheap and "looked nice" (which IMO they don't).

My 290 tops out at 70-72 and it's way hotter than Tahiti.... IMHO if you have a black and white build, like me, the white lighting on the XFX logo makes the card 10/10

[don_svetlio]

My 290 tops out at 70-72 and it's way hotter than Tahiti.... IMHO if you have a black and white build, like me, the white lighting on the XFX logo makes the card 10/10

Qeustion, do your VRM1 modules run hotter, cooler or identical to the GPU die? or is it on a game-to-game difference basis?

[winningsince1337]

Qeustion, do your VRM1 modules run hotter, cooler or identical to the GPU die? or is it on a game-to-game difference basis?

VRM 1 temps > GPU Core Temps > VRM 2 Temps.

 

That trend always seem to be correct (at least the VRM 1 > GPU Core if my memory serves me correct) but core temps depend on how demanding the game is/my room ambient temps. I'll go load up a game and report back with some temps for you.