Thermal Paste for CPU and GPU

F

Forum Reader

Expert Member
Joined
Aug 25, 2019
Messages
1,968
Reaction score
1,446
My PC has random shutdowns, usually about 15mins after starting it up and in certain games. Pretty sure it's due to the CPU or GPU overheating since if I open up the case and blow out the dust it works fine for a week or two again with no issues, and then the shutdowns start again.

The CPU and GPU are quite old, purchased back in 2015. So I think the thermal paste has hardened and is not making a good contact with the heatsink. It's a FX8350 and a GTX970. What would be a decent thermal paste to use? Will be getting it from Takealot and they have so many different makes/brands.

Takealot.com: Online Shopping | SA's leading online store

South Africa's leading online store. Fast, reliable delivery to your door. Many ways to pay. Shop anything you can imagine: TVs, laptops, cellphones, kitchen appliances, toys, books, beauty & more. Shop the mobile app anytime, anywhere.
www.takealot.com www.takealot.com

I assume I will also need 90%+ rubbing alcohol to remove the old paste. This is the only one I can find (on Takealot):

Takealot.com: Online Shopping | SA's leading online store

South Africa's leading online store. Fast, reliable delivery to your door. Many ways to pay. Shop anything you can imagine: TVs, laptops, cellphones, kitchen appliances, toys, books, beauty & more. Shop the mobile app anytime, anywhere.
www.takealot.com www.takealot.com

I think it's most likely the CPU that is the problem but will redo the thermal paste on the GPU just in case.
 
You don't need to do both. Install a PC TEMP monitoring app such as 'Open Hardware Monitor' and find out exactly which one is overheating.

There is a reason we don't try and fix things that aren't broken in IT.
 
RunningFromFat said:
You don't need to do both. Install a PC TEMP monitoring app such as 'Open Hardware Monitor' and find out exactly which one is overheating.

There is a reason we don't try and fix things that aren't broken in IT.
Click to expand...

Tried that a while back and temps seemed fine. So it must have shut down before recording the temp spike. Will try again and see.
 
Forum Reader said:
Tried that a while back and temps seemed fine. So it must have shut down before recording the temp spike. Will try again and see.
Click to expand...
Turn it back on immediately and try see the temps, if it's in a rebooting loop, then you know for sure something is overheating. Can also take out the gfx card + use onboard and do some stress tests to get the temps going.

Reason being is that I wouldn't like fiddling with a gfx card heat sink and you risk breaking your card.
 
RunningFromFat said:
Turn it back on immediately and try see the temps, if it's in a rebooting loop, then you know for sure something is overheating. Can also take out the gfx card + use onboard and do some stress tests to get the temps going.

Reason being is that I wouldn't like fiddling with a gfx card heat sink and you risk breaking your card.
Click to expand...

Sometimes it doesn't reboot immediately. Have to restart it a few times then for it to boot to windows. I noticed this started at the beginning of the year when I installed Windows 10. Usually the crash happened around the time when WindowsUpdate started running in the background. Saw it on event viewer. Once I disabled WindowsUpdate, it stopped for a while. Now I don't bother. But now gaming crashes it too (usually in menus where the GPU isn't being used heavily). I guess stuff running in the background pushes the CPU too high at times.

What's weird is that it will only crash once for the day, then for the rest of the day I'll have zero issues (like for 12hrs+). I do notice that the fans spin up faster (more noise) after the crash. When I first start up the PC for the day the fans don't make a lot of noise. This is why I haven't bothered trying to fix it at all for all this time. Just have the one restart and then I'm fine until the next day.
 
Last edited:
Forum Reader said:
Sometimes it doesn't reboot immediately. Have to restart it a few times then for it to boot to windows. I noticed this started at the beginning of the year when I installed Windows 10. Usually the crash happened around the time when WindowsUpdate started running in the background. Saw it on event viewer. Once I disabled WindowsUpdate, it stopped for a while. Now I don't bother. But now gaming crashes it too (usually in menus where the GPU isn't being used heavily). I guess stuff running in the background pushes the CPU too high at times.

What's weird is that it will only crash once for the day, then for the rest of the day I'll have zero issues (like for 12hrs+). I do notice that the fans spin up faster (more noise) after the crash. When I first start up the PC for the day the fans don't make a lot of noise. This is why I haven't bothered trying to fix it at all for all this time. Just have the one restart and then I'm fine until the next day.
Click to expand...

...and thermal paste is the issue why?

Also no need for that expensive alcohol ... don't you have sanitizer left from Covid days ?
 
  • Wow
Reactions: Yuu
Dimpie (COMPUTEK) said:
...and thermal paste is the issue why?
Click to expand...

What else would it be? If I blow out the dust and the PC works fine for a few weeks then it has to be heating related correct?

I have sanitizer but it's the 70% alcohol which has other stuff in it.
 
Forum Reader said:
So you have no clue. Don't bother responding and let other's help instead.
Click to expand...

Jumping from “blowing dust out and it works” to “need thermal paste on the cpu and gpu” is a huge jump based on no data.

Runaway temperatures are easy to see in hwinfo
 
  • Like
Reactions: Yuu
_kabal_ said:
Jumping from “blowing dust out and it works” to “need thermal paste on the cpu and gpu” is a huge jump based on no data.

Runaway temperatures are easy to see in hwinfo
Click to expand...

Another one. So what is your guess?
 
Could be temps. Very easy to test, just use HWInfo or similar and monitor temps, set it to log them. If it crashes, open the log file and check the temps.

Also, check event viewer, could be clues there as to the cause.

If it isn’t temps, my guess would be PSU or RAM.

Either way, log temps and check them. And check event viewer.

And stop being rude to those in your support thread trying to help.
 
UrBaN963 said:
Could be temps. Very easy to test, just use HWInfo or similar and monitor temps, set it to log them. If it crashes, open the log file and check the temps.

Also, check event viewer, could be clues there as to the cause.

If it isn’t temps, my guess would be PSU or RAM.

Either way, log temps and check them. And check event viewer.

And stop being rude to those in your support thread trying to help.
Click to expand...

Thanks and I'm not being rude. Those guys are just coming here to troll and post useless stuff that doesn't help me in any way. I don't know why they bothered posting when they got nothing useful to say.
 
Forum Reader said:
Thanks and I'm not being rude. Those guys are just coming here to troll and post useless stuff that doesn't help me in any way. I don't know why they bothered posting when they got nothing useful to say.
Click to expand...
To be fair, both @Dimpie (COMPUTEK) and @_kabal_ made valid points. Thermal paste may help but it won't do anything fore the dust.

And the assumption that thermal paste is required for either the CPU, GPU or both is based on no data, at least as far as you've shared here. It's not a bad idea, but it's just an assumption because we don't know if there's an actual thermal issue.

First step would be to check temps and event viewer, that will narrow the options down and we can proceed with more useful answers and suggestions based on more data.
 
Last edited:
UrBaN963 said:
To be fair, both @Dimpie (COMPUTEK) and @_kabal_ made valid points. Thermal paste may help but it won't do anything fore the dust.

And the assumption that thermal paste is required for either the CPU, GPU or both is based on no data, at least as far as you've shared here. It's not a bad idea, but it's just an assumption because we don't know if there's an actual thermal issue.

First step would be to check temps and event viewer, that will narrow the options down and we can proceed with more useful answers and suggestions based on more data.
Click to expand...

Did a bit of light gaming for 20 mins.

- The CPU temp went up quite high to 63c. Apparently 61c is its max operating temp so that looks like a problem.
- The GPU temp seems ok but the GPU hot-spot temp went up to 87.5c. Though seems like the GPU fan speed only went up to 27% of max. I'd expect it to automatically speed up more if really overheating?

Will monitor again when it actually starts crashing in a week or two.

temps.png
 
Forum Reader said:
My PC has random shutdowns, usually about 15mins after starting it up and in certain games. Pretty sure it's due to the CPU or GPU overheating since if I open up the case and blow out the dust it works fine for a week or two again with no issues, and then the shutdowns start again.

The CPU and GPU are quite old, purchased back in 2015. So I think the thermal paste has hardened and is not making a good contact with the heatsink. It's a FX8350 and a GTX970. What would be a decent thermal paste to use? Will be getting it from Takealot and they have so many different makes/brands.

Takealot.com: Online Shopping | SA's leading online store

South Africa's leading online store. Fast, reliable delivery to your door. Many ways to pay. Shop anything you can imagine: TVs, laptops, cellphones, kitchen appliances, toys, books, beauty & more. Shop the mobile app anytime, anywhere.
www.takealot.com www.takealot.com

I assume I will also need 90%+ rubbing alcohol to remove the old paste. This is the only one I can find (on Takealot):

Takealot.com: Online Shopping | SA's leading online store

South Africa's leading online store. Fast, reliable delivery to your door. Many ways to pay. Shop anything you can imagine: TVs, laptops, cellphones, kitchen appliances, toys, books, beauty & more. Shop the mobile app anytime, anywhere.
www.takealot.com www.takealot.com

I think it's most likely the CPU that is the problem but will redo the thermal paste on the GPU just in case.
Click to expand...
I have the same problem i5/gtx960
 
Forum Reader said:
Did a bit of light gaming for 20 mins.

- The CPU temp went up quite high to 63c. Apparently 61c is its max operating temp so that looks like a problem.
- The GPU temp seems ok but the GPU hot-spot temp went up to 87.5c. Though seems like the GPU fan speed only went up to 27% of max. I'd expect it to automatically speed up more if really overheating?

Will monitor again when it actually starts crashing in a week or two.

View attachment 1417139
Click to expand...
I believe what you've circled is the CPU socket itself - the CPU seems to have maxed out at 60.1. Still very close to max but not quite there. Also, hitting MAX shouldn't cause an instant crash, but you are running very close to the edge. Fresh thermal paste may help with that.

As for the GPU, the hot spot temp is high but your average temps are okay. It is hitting it's thermal limit though, so definitely needs to be addressed. I'd use something like MSI afterburner to adjust the fan curve, make it a bit more aggressive to get some extra cooling sooner.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X