• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

AMD FX-8350 idling at 60 degrees with stock cooler

BrownBrown

New Member
Joined
Jun 30, 2020
Messages
4 (0.00/day)
I've been using the same fx 8350 for over three years now, and only recently discovered it's overclocking capabilities. Just a couple weeks ago I got it running stable at 4.6ghz, good temps and all, usually idling at ~33 degrees and only reaching around 62 degrees under load.
I leave the computer to sit for a week and come back figuring I'd try overclocking memory and northbridge next, and so I run some stress tests before changing anything. But I notice it's running a bit hotter than before, not once dipping below 45c before and during testing, usually sitting around 50 on desktop, and would easily break 70 degrees under load. After a number of restarts, nothings changed.
I figured bad thermal paste was to blame, since the last application was over three years ago. Using cheap $5 compound from a local computer shop, I applied some fresh paste, and booted into windows. Checking temps the moment I log in, I find it sitting at ~62c, not dipping below 57 degrees, sometimes breaking 70 degrees just browsing chrome. Since writing this, it's stayed consistent at 60+ degrees, despite only running AIDA64 and a chrome tab

Reapplying thermal paste seems to have added a good 10 degrees onto my average temp. I've tried rereapplying with a smaller glob, nothing. Before writing this conclusion, I tried downclocking to stock frequency/voltage, which has brought average temp down to ~52 degrees on desktop, down from the 70 degrees displayed in bios.
50 degrees still seems abysmal, and I'm at a loss for solutions. Is the thermal paste to blame? I've put in an order for some thermal grizzly conductonaut compound, will that make a significant difference? The stock cooler has otherwise serviced me well, fan speeds aren't any different, reaches 6000+ rpm no problem. Not sure what could be the problem other than the paste, and it doesn't seem right that 3 year old preapplied compound would beat out a new application, even if cheap
 
Conductonaut is liquid metal FYI, not thermal paste. It is very effective but there are a lot of risks involved. Make sure you know what you're doing if you start using compounds that are electrically conductive. It's not just paste it on and put the cooler on, it requires precise usage so you don't kill components. It's not easy to clean either so if you spill any on your motherboard it's dead. LM also requires regular maintenance and will destroy your IHS and cooler cold plate if it's not nickel-plated. If your cooler uses aluminium it will fuse with the gallium (liquid metal) and be destroyed. Just as a fair warning.

Is "6000+" a typo or are you running server fans?

It seems strange to increase temps by almost 20 degrees at idle and 10 degrees at load by changing thermal paste, even if the new paste is abysmal. Are you sure the application was good and the mounting pressure was good too?
 
Conductonaut is liquid metal FYI, not thermal paste. It is very effective but there are a lot of risks involved. Make sure you know what you're doing if you start using compounds that are electrically conductive. It's not just paste it on and put the cooler on, it requires precise usage so you don't kill components. It's not easy to clean either so if you spill any on your motherboard it's dead. LM also requires regular maintenance and will destroy your IHS and cooler cold plate if it's not nickel-plated. If your cooler uses aluminium it will fuse with the gallium (liquid metal) and be destroyed. Just as a fair warning.

Is "6000+" a typo or are you running server fans?

It seems strange to increase temps by almost 20 degrees at idle and 10 degrees at load by changing thermal paste, even if the new paste is abysmal. Are you sure the application was good and the mounting pressure was good too?

oops, guess I could've done a little more research, saw Conductonaut under "best thermal paste for overclocking" and went straight to purchase. 6000 rpm is what the sensor reads as far as I can tell, I'm using the basic stock cooler
images

ie this one. AIDA64 reads 6000 rpm, bios read 5000 shortly after powering on, and it sure sounds like 6000 when it gets hot. I don't think there's much to get wrong when mounting the cooler. I screwed the mounting bracket in, and clamped it on, there was obvious contact with the thermal paste both times i went to put the cooler in. Not sure what else I could have done wrong, strange indeed
 

Attachments

  • Screenshot (198).png
    Screenshot (198).png
    122.1 KB · Views: 396
I'd first replace the $5 paste with something decent like MX-4 before going for something as extreme and potentially hazardous as Conductonaut.
 
I'd first replace the $5 paste with something decent like MX-4 before going for something as extreme and potentially hazardous as Conductonaut.

subtext is great advice for me two hours ago, will watch reviews and tutorials before purchasing this time around
 
Temp does seem high for the CPU.
While the system is running and you can monitor temps on screen, push the cooler down on the CPU with your fingers to see if increased pressure drops the temps.
Be careful of burning your fingers or getting them in the fan.
 
oops, guess I could've done a little more research, saw Conductonaut under "best thermal paste for overclocking" and went straight to purchase. 6000 rpm is what the sensor reads as far as I can tell, I'm using the basic stock cooler
images

ie this one. AIDA64 reads 6000 rpm, bios read 5000 shortly after powering on, and it sure sounds like 6000 when it gets hot. I don't think there's much to get wrong when mounting the cooler. I screwed the mounting bracket in, and clamped it on, there was obvious contact with the thermal paste both times i went to put the cooler in. Not sure what else I could have done wrong, strange indeed
Just googled it and yeah, 6000RPM is correct. That is a hell of a fan. Can't be nice on the ears, though.

I'd try using decent thermal paste like @Assimilator mentions. That seems to be the only variable that seems to have changed, although mounting pressure and thermal paste spread are still reasonable to consider as a second and third variable to look into and test if the good thermal paste makes no difference.
 
The cooler isn't properly attached. Anyway, you're wasting your time getting better thermal paste, you're problem is the cooler it self, it doesn't dissipate enough heat.
 
Looking at this a bit further, it doesn't make sense. I'd expect the CPU temperature to increase directly after the overclock, not a week after the overclock.

@BrownBrown I would suggest setting everything back to stock and recording the temperatures, then applying the overclock again. When overclocking, check what voltages you're applying and where - if you're setting volts to "auto" your board could be pushing more through the CPU than it needs, causing the increased temps.

Another possibility is that the CPU and/or board is degraded and/or failing (possibly bad caps on the board?).
 
i must agree that cooler is no good to a overclocked fx8350, you need a better one a tower air cooler or watercooled aio and some good paste.
 
A stock cooler is only good for stock useage, not for OC'ing.
When those are made it's to handle such a chip when ran at stock so it's no suprise to me it's getting warm.
You will need something better than a stock cooler, also know since it's an 8 cored chip it's going to run a bit warm anyway and be worse when OC'ed as you have yours. Either big air or water is what to go for here and I'd get one of the larger cooling capacity AIO's if you go that way with it.
For reliability and ease of maintenance big air in itself will do fine but you will still see temps getting warm, just not as warm as you're seeing them now. An AIO will let it run cooler but there is the maintenance side of it to consider if you do that.

In any case you really need to ditch the stock cooler or just run it at stock specs.
 
Looking at this a bit further, it doesn't make sense. I'd expect the CPU temperature to increase directly after the overclock, not a week after the overclock.

@BrownBrown I would suggest setting everything back to stock and recording the temperatures, then applying the overclock again. When overclocking, check what voltages you're applying and where - if you're setting volts to "auto" your board could be pushing more through the CPU than it needs, causing the increased temps.

Another possibility is that the CPU and/or board is degraded and/or failing (possibly bad caps on the board?).

Did just this, reset defaults and put core cpu and cpu nb (i think) voltage to +0. Booting with defaults didn't revert temperatures back to "normal," but the overclocking process seems normal more or less. Got to 4.4ghz before increasing voltage, resting/loaded temperature did seem to increase with frequency foremost, like normal. As far as I can tell, the cooler is able to maintain stable temps as it is. Seems I've managed to get 4.7ghz with 1.5 volts on the cpu, havent run cinebench yet though. Doing nothing, it rests at around 66 degrees, but under load it doesn't go above 70 strangely enough, cooler maintains 69 pretty "easily," at the expense of my ears. This was the case for just about every clock speed I tried, it'd rest at 60+ degrees, but not break 70 under load, and no throttling from what I can tell. Age wise, both cpu and motherboard have only been in use for 3 years now, as long as I've had em, as far as I know. Not sure how "old" that is, I feel like 3 years running stock speeds is a short life.

SO from what I can tell, the cooler is working as intended, but the cpu itself is running hotter for no reason, regardless of load, and I have no way of diagnosing? Still confident some decent thermal paste will help some, but I really don't know what could have changed in the span of two weeks.

Last thing of note would be when running stress test on stock speed, cpu temp would peak, and it'd take about 100 seconds before the cooler could bring it back down to something maintainable. Not sure if that says more about the cooler or the paste. Thinking the paste

Temp does seem high for the CPU.
While the system is running and you can monitor temps on screen, push the cooler down on the CPU with your fingers to see if increased pressure drops the temps.
Be careful of burning your fingers or getting them in the fan.

Attempted this, had strong pressure for about a minute, saw no obvious change in temperature, or shifting of the cooler so, it feels properly mounted.
 
I also have an 8350 in my media server. The stock cooler simply isn't sufficient, even at stock clocks. Pulled an old tower cooler out of the closet and zip tied it to the board. Idles at 35 C now lol. I have no idea how you were running decent temps before, as I never could on the stock cooler. Maybe ambient temp was way cooler at the time? Edit: nvm, just read again that you've been running it like that for three years. I dunno then.
 
its a 8350 with the stock cooler
replace the cooler or replace the board/cpu/ram
they run hot they are loud everything here is normal
 
@BrownBrown
My 2 Cents,
Upgrade your cooler at the very least. A Hyper 212 EVO would do wonders for temps on that CPU.

But ultimately, that CPU/Mobo could be going faulty and you'd need to make a replacement. Might want to plan for that soon. As the Ryzen 4000 series comes out, getting into a Ryzen 3000 based CPU & board will be less expensive and Ryzen 2000 is a great value right now.
 
Back
Top