hawkinsa21
New member
Hi guys,
I rarely post anywhere these days, grown old with a house full of kids etc. But, been getting back into PC gaming as I've learned to juggle time a bit better
So, I have an XFX 7950. Its 14 months old now and outside of warranty, and I've not really been hitting it that hard. But the stock temps weren't great. It was idling in the low 40's and the ambient temp in my living room isn't high. It was hitting over 80 degrees on stock speeds in Far Cry 3 and after an hour it once hit 87, which I know is a couple of degrees over the manufacturers rated 85. After reading posts elsewhere about stock TIM application sucking on these cards and people reapplying their own with good success, I took the plunge to do it myself.
So I took it apart and low and behold the manufacturers TIM application was a mess; all over the place, and in great big wads and there were a couple of small areas of no TIM contact between GPU and HS at all. Took a good amount of cleaning. I carefully reapplied my trusty AS5 (I know there is better but it is what I had), and carefully reassembled. Stock temps now high 40s
and Far Cry 3 hit 97 degrees!!!
I know AS5 has a cure time, but an immediate 10 degree increase into a dangerous zone looks bad. I carefully took apart and checked the contact. The AS5 contact was even and precise and covers the entire GPU die, as good as I've ever gotten it so I don't think there is an issue with HS to die contact / pressure. I cannot see any evidence of uneven spread or air bubbles. The vram thermal pads that came with the card were still attached OK as well. I'm hesitant to just lump it on like the manufacturer did, but theirs certainly seemed to be better.
Any suggestions welcome! Options I see are new thermal paste as the AS5 is probably 8 years old (although shouldn't have a shelf life). I'm also buying a friends H60 AIO which I was going to use for my CPU but now I might use on the GPU (red mod / GPUCool). The 7950 is a stellar card, I can't let it become a paper weight.
Thanks,
Alex
EDIT: Currently idling on Windows 7 desktop at 47 degrees (Afterburner)
I cannot read my VRAM temps via either GPU-Z or HWiNFO, must not be supported. One thing I noted is that the thermal pads came loose when I removed the PCB, but they sat back in place OK. I'm not sure whether 'sitting' back in place is sufficient for cooling, if they aren't making proper contact? Also, would the vram temp increase the temp that Afterburner is reporting for GPU temperature? Could it be that the core is actually OK but the vram temp has gotten worse and drastically screwed the temps?
I rarely post anywhere these days, grown old with a house full of kids etc. But, been getting back into PC gaming as I've learned to juggle time a bit better

So, I have an XFX 7950. Its 14 months old now and outside of warranty, and I've not really been hitting it that hard. But the stock temps weren't great. It was idling in the low 40's and the ambient temp in my living room isn't high. It was hitting over 80 degrees on stock speeds in Far Cry 3 and after an hour it once hit 87, which I know is a couple of degrees over the manufacturers rated 85. After reading posts elsewhere about stock TIM application sucking on these cards and people reapplying their own with good success, I took the plunge to do it myself.
So I took it apart and low and behold the manufacturers TIM application was a mess; all over the place, and in great big wads and there were a couple of small areas of no TIM contact between GPU and HS at all. Took a good amount of cleaning. I carefully reapplied my trusty AS5 (I know there is better but it is what I had), and carefully reassembled. Stock temps now high 40s

I know AS5 has a cure time, but an immediate 10 degree increase into a dangerous zone looks bad. I carefully took apart and checked the contact. The AS5 contact was even and precise and covers the entire GPU die, as good as I've ever gotten it so I don't think there is an issue with HS to die contact / pressure. I cannot see any evidence of uneven spread or air bubbles. The vram thermal pads that came with the card were still attached OK as well. I'm hesitant to just lump it on like the manufacturer did, but theirs certainly seemed to be better.
Any suggestions welcome! Options I see are new thermal paste as the AS5 is probably 8 years old (although shouldn't have a shelf life). I'm also buying a friends H60 AIO which I was going to use for my CPU but now I might use on the GPU (red mod / GPUCool). The 7950 is a stellar card, I can't let it become a paper weight.
Thanks,
Alex
EDIT: Currently idling on Windows 7 desktop at 47 degrees (Afterburner)
I cannot read my VRAM temps via either GPU-Z or HWiNFO, must not be supported. One thing I noted is that the thermal pads came loose when I removed the PCB, but they sat back in place OK. I'm not sure whether 'sitting' back in place is sufficient for cooling, if they aren't making proper contact? Also, would the vram temp increase the temp that Afterburner is reporting for GPU temperature? Could it be that the core is actually OK but the vram temp has gotten worse and drastically screwed the temps?
Last edited: