FX-8350 spontaneously "overheats" then returns to normal temperatures

Toggle sidebar Toggle sidebar
I

Itroitnyah

Distinguished
Apr 14, 2013
9
0
18,510
System:

FX-8350
Cooler Master Hyper 212 EVO
Asus M5A97 LE R2.0
8gb Kingston HyperX Grey
Powercolor Radeon HD 7850
Topower 650w (I know)

The issue is exactly as the title states. When I am playing videogames, the cpu fan will randomly start spinning as fast as it can. But only briefly. Not more than 3-8 seconds before returning to normal speed. This leads me to believe that it may be an overheating issue, but it feels different from every other overheating issue I've heard about because it is not a sustained overheating. When I monitor CPU temps they seem to indeed start to rise before the fans go berserk and the temps return to normal.

This is not an issue that I had when I first built the system, it is one that developed only a few months ago, and the system is now 3 years old. I tried replacing the thermal paste and that did not fix the issue. This leads me to believe that either A) the motherboard is starting to have issues or B) The power supply is.

I would not be surprised by the motherboard, as it's been giving me issues since day 1 with a broken LAN port and flakey rear USB ports that would magically stop detecting things that are already plugged in when I would plug in a new device. Perhaps it is starting to grow senile in its' old age? I would be surprised if it were the power supply since it seems illogical to me that it would cause this type of issue, but I don't believe there would be anything else that would be making my cpu cooler go berserk randomly.

Any ideas or solutions would be greatly appreciated.
 
Solution
Onus
Hmmm, I believe your mobo does not have heatsinks on its VRMs. They could also be what is overheating, even if the CPU stays cool. A test would be to take the side off your case and point a common household fan into it and see if the behavior changes.
HWMonitor can also keep track of the maximum temp reached, and should include a few mobo sensors in its readouts.
Sort by date Sort by votes
Onus

Onus

Titan
Jan 27, 2006
37,037
12
89,965
Did you clean all the dust out of the cooler when you replaced the paste?
You say it is three years old, so it is possible the cheap fan that Coolermaster uses has begun to bind, reducing its RPM. A new fan may fix that, or see if you can adjust your fan curve. If it was at 30% up to a certain temp, change it to 40%, or 50% and see if that prevents the overheating.
 
Upvote 0 Downvote
Shaun o

Shaun o

Illustrious
Oct 7, 2007
6,113
1
41,960
Well if the temps of the FX cpu seem fine, and you have not noticed a rise in the max temp it is reaching under gaming load ect.

It would suggest that the PWM chip found on the fan it`s self may be at fault.
If you look at a fan that has four wires on it, it`s usually a PWM fan.

It sends polling information to the motherboard, on the rpm the fan is spinning at on a regular pulse.
As you know the motherboard monitors the heat the cpu is giving out and then if the cpu starts to become hot based on a fan power curve setting it increases the voltage supplied to the Pwm enabled fan to make it rotate faster in rpm to push more air at a greater flow rate.

So what I would do, is at least test the cpu cooler fan with another Pwm based fan if you have a spare one.
If the same thing happens then it would confirm that it is, or may be board side related to power distribution.

Or just an aged fan as the problem.
 
Upvote 0 Downvote
I

Itroitnyah

Distinguished
Apr 14, 2013
9
0
18,510


I did indeed clean out the dust. I do have a spare fan lying around the case so I may have to put that on the hyper 212 evo and see if that helps. Thanks for the tip.

The only way to know that is to know the actual temps.
Click to expand...

Sure. In game, I seem to stay around 40C. That's just from WoW though, since I mostly play that. When the fan went nuts I was still at 40C.

So what I would do, is at least test the cpu cooler fan with another Pwm based fan if you have a spare one.
If the same thing happens then it would confirm that it is, or may be board side related to power distribution.
Click to expand...

Will do, thanks for the information. I was unaware that PWM fans had a chip inside of them that could malfunction. TIL

 
Upvote 0 Downvote
Belphegore

Belphegore

Respectable
Mar 27, 2016
366
0
2,160
FX-8350 chips are known to have awful, incorrect temp reads. For example, my current 8350 has gone as low as 4C and idles at 11C, and doesn't go past 30C under 100% load (water cooled). My first 8350 chip would randomly spike a false 99C temp and cause my machine to go into shut down mode, even though it only lasted one second and went from 30C - 99C in that one second.
 
Upvote 0 Downvote
I

Itroitnyah

Distinguished
Apr 14, 2013
9
0
18,510
Alright, OP again.

So I tried what Onus and Shaun suggested and replaced my Hyper 212 EVO fan. I had a spare Arctic Cooling F12 laying around, so it fit the bill. It worked, but only temporarily. I did a 6.5 hour long gaming session today. During the session my temps remained at a solid 40C (Assuming they were being read correct, Belphegore), and there were only a few incidents where the fan speed wildly increased. Much farther apart in timing than before. The fan I put in its' place could possibly be on its way out the door too? Unless another solution is suggested I will be swapping out the current fan for one of the case fans that I know is working fine and seeing if that fixes the solution.

try a load test and monitor the temp. see if the temp is just a up and down thing or a always up thing.
Click to expand...
It's a steady temp reading. Correct me if I'm wrong, but even though videogames aren't "stressful" on your CPU like an actually stress test, playing games can be an equivalent. I did manage to maintain a steady 40C throughout 6.5 hours.
 
Upvote 0 Downvote
Belphegore

Belphegore

Respectable
Mar 27, 2016
366
0
2,160


While games aren't exactly "stressful" on the CPU, more and more games are tending to pull it into action along with the GPU. It's intentional most of the time, but sometimes it's due to poor optimization. For instance, when I was playing FFV on Steam, my CPU was heating up to 40C... for a sprite based game... and I'm using water cooling (normally what my temp will read if I'm in an AAA title with top graphics) and my GPU was cool as a cucumber. That was poor optimization, and it was later fixed.

40C isn't bad at all though for the 8350. You'll never want it to exceed 70-75C, however. My first chip would heat up to that range when I was playing Wolfenstein: The New Order when it first released and had horrendous optimization, and it would trigger a shut down. That was the same chip that would spike 99C for no reason while idling and trigger an instant shut down. I'd come home from work and the PC would be off.
 
Upvote 0 Downvote
G

gonf

Distinguished
Jun 17, 2008
300
0
18,860
even with stressful cpu game. you wont get 100% on all core all the time. thats when Prime95 comes in. so just run it. if you don't feel like doing a long one. just do a 30min one and see if the temp go up and down. if not than it not the cpu problem.
the evo212 is a very good budget heat sink. i have one with my fx8350 too. am getting 32C idle and 47C with it with load (room temp is about 19).
 
Upvote 0 Downvote
Onus

Onus

Titan
Jan 27, 2006
37,037
12
89,965
Hmmm, I believe your mobo does not have heatsinks on its VRMs. They could also be what is overheating, even if the CPU stays cool. A test would be to take the side off your case and point a common household fan into it and see if the behavior changes.
HWMonitor can also keep track of the maximum temp reached, and should include a few mobo sensors in its readouts.
 
Upvote 0 Downvote
Solution
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom