[SOLVED] [SOLVED]FX 8350 overheating, lost as to what to do next...

[xxsm1ttyxx]

Hey guys,

I am at my wits end here... I have a FX 8350 which I have had for several years now, stock speeds no crazy OC. About a month ago I noticed thermals getting a bit warm and decided it was time to clear it of dust and debris. My office is in the basement of my house, and I have animals, which prompts regular cleaning. This is nothing new, and have been doing this process every few months for several years now. After full clean and put back together with new thermal paste (noctua pea sized) all is well. single digits to low teens C for idle temps. A few days go by of gaming, movies, streams, etc and nothing is out of the ordinary. Then out of nowhere I come home from work, boot the machine from off, launch Apex legends to update it and the CPU hits 80c within seconds, and is frozen. I reboot and same thing occurs also noticing chipsets on the mobo reaching very high temps as well. At this point I take the cooler back off, clean thermal paste and reapply. Same deal as prior, great thermals for a few days then boom back to heat of the sun.... I have reset my cmos to clear any possible OC settings I may have accidentally set, loaded optimized defaults, reseated CPU. Nothing changes still overheats.

After all of the above I was curious if it was possible that a bad motherboard would cause these strange things to happen. I have an older board that I swapped from a few years back, cleaned it up and placed it into the system. Updated bios ensured all was working properly and loaded optimized defaults and to my surprise I was again met with the very low temps I was used to seeing. This was about 4 days ago now, last night while simply watching netflix, the CPU was hitting 65+. Only Chrome(Netflix) and the OS was running. Finish the movie and go to launch apex, and the PC blows up to thermal limit and freezes again.

I have verified that all fans in the case spin up, are not faulty, and run at 1500 RPM+. The AIO is working, I can feel flow through the tubes, and distinctly can feel a temperature difference when touching the tubes. The AIO is mounted to the top of my case with a push/pull config. Fans controlled through the Extreme curve setting through Corsair iCue. Pump also set to Extreme curve pushing at 3000 RPM.

I am so confused as to what the problem could be as I am getting very similar issues happening after almost the same amount of time of 2 separate boards now. Looking for any sort of insight or direction on what to check next.

Build details are below.

FX-8350
Corsair H115i AIO cooler (push/pull config)
Kingston HyperX Fury 16gb (4x4) 1600
Gigabyte 990FX UD3/MSI 970 Gaming mobos (2 diff ones)
Corsair CX750M psu
MSI GTX 960 2g
Corsair case (dunno model, large wide open mid tower case - dual front intake/1 rear exhaust )
Ambient temps in my office at this time of year are around 65F

Thanks for reaching my wall of text.

-Smitty

 

[maxamillionfeettall]

The next time it spikes, try pushing down on the pump to see if it's a contact issue and watch the temps. Might be the backplate needs a reseat or something similar.

 

[xxsm1ttyxx]

The next time it spikes, try pushing down on the pump to see if it's a contact issue and watch the temps. Might be the backplate needs a reseat or something similar.

So this was something I was also curious about, as it was the same on both mobos i tried. The back plate has a bit of up and down play in it after putting in the standoffs. After screwing the pump down onto the socket it doesnt move. It has always been like this since I originally installed the cooler like 3 years ago.

Thinking more on this and doing some research regarding the Corsair AIO coolers and AMD motherboards. It seems that the backplate that must be used is the default one that comes with the board and stock cooler mounts. But when installing the standoffs it leaves a gap between the PCB and the stand offs. Now several folks claim that this is normal and you simply tighten down the AIO and all is good. On the other side of that coin, there are equally as many folks who have attempted to do this and had terrible thermals as a result. As such using rubber washers to fill the gap so the standoffs are flush with the PCB.

Is it possible that in my prior reinstallations of the CPU/AIO I have used "too much" thermal compound. Resulting in what could be considered a better seal between IHS and Heatsink simply because there is enough compound in there if the AIO backpate being loose was the issue? I have been paying for more close attention to amount of compound and such when reapplying it after cleaning. So I guess I am grasping at straws here. But I supposed my next test would be to get washers and get that backplate flush with the PCB and go again.

 

[maxamillionfeettall]

As long as its snug when tightened, it should be good....stress on "should" as after all this time, there may be some "settling". If the contact isnt the issue, perhaps you cpu/board has the voltages set too high or vrm's are heating up, though i doubt that tbh.

If you have the stock fan or any other cooler, I'd try that

 

[xxsm1ttyxx]

As long as its snug when tightened, it should be good....stress on "should" as after all this time, there may be some "settling". If the contact isnt the issue, perhaps you cpu/board has the voltages set too high or vrm's are heating up, though i doubt that tbh.

If you have the stock fan or any other cooler, I'd try that

Ive factory defaulted both mobos i have, the votages are normal from what i can see. Dropping to .9 while on idle hitting 1.42 while under load.
The temps for the chipsets outside of the cpu socket are pushing 60+ while under load and i did see a 90 during a crash.

I dont have any original AMD stock cooler mounting brackets anymore as i never run the stock coolers :/
May snag a Hyper212 i guess and see what it does...

 

[maxamillionfeettall]

Well shucks, vrm's are fine then, since the pump seems to operating nominally, I doubt theres a blockage in the line or something wrong with the cooler itself, other then the slight possibility of contact(dont worry about the thermal paste too much unless you've been caking it). The only other thing I can think of is ,manually setting the voltage, though they seem fine, to whatever is good for the cpu, 1.3 i think i recall, or the psu is underpowered and causing issues with the boards power delivery, or the boards power delivery in general is shagged...but thats a real stretch.

hmmmmmm

 

[xxsm1ttyxx]

Well shucks, vrm's are fine then, since the pump seems to operating nominally, I doubt theres a blockage in the line or something wrong with the cooler itself, other then the slight possibility of contact(dont worry about the thermal paste too much unless you've been caking it). The only other thing I can think of is ,manually setting the voltage, though they seem fine, to whatever is good for the cpu, 1.3 i think i recall, or the psu is underpowered and causing issues with the boards power delivery, or the boards power delivery in general is shagged...but thats a real stretch.

hmmmmmm

I also did research into this as my PSU is getting to be on the dated side as well. I have probably had it nearly the entire lifespan of the CPU. Threads on this forum seem to have conflicting opinions regarding if a PSU could be causing CPU overheating. I know my entire PC does not draw nearly the 750 Watts the PSU I have is rated for. But due to its age and run time, could it be possible that it has become unstable in the voltage it is able to supply? Can the power delivery from it become unstable to the point it would be causing what I am seeing? It seems like it might make sense as I have the same experience on 2 different boards, both boards with different chipsets.

 

[maxamillionfeettall]

Well, I'd personally get an heatsink to test with as that is the cheapest option and would immediately rule out the cooler if results repeated. after that, we'll see.

I've only ever had psu's fail out right, but I wont rule out the possibility.

 

[xxsm1ttyxx]

More food for thought....
I supposed its possible that the coolant in my AIO could have evaporated some over the years? Even though its a 280mm rad I suppose with it running 24/7 for the last few years it may be time.

All seems to point at getting a different cooler and testing as the next steps.

 

[mdd1963]

AOI's gradually deteriorate in funtion over the course of a few years, and, then just seem to stop working, all the while confusing their owners by providing a reassuring buzzing sound to mislead folks into thinking their pumps are still fine....

A radiator at room temp should imply no warm fluid making it to it...

As to the cause being fluid loss or contaminants/hard water deposits, unknown....

 

[xxsm1ttyxx]

A radiator at room temp should imply no warm fluid making it to it...

Can you elaborate just a bit on what you are trying to say here? I think im understanding you as meaning the rad should itself be warmer than the room temp, or else its a good possibility it has failed?

 

[maxamillionfeettall]

Considering the op said

I can feel flow through the tubes, and distinctly can feel a temperature difference when touching the tubes.

,I'd expect the pump to be working, unless there was some sort of blockage reducing flow.

 

[ClassicSin]

I am currently working on my buddy's rig right now experiencing similar issues. Granted his is a much older build (fx-4170, Asus m5a97 rev. 1.00) I have personally run into this issue with AMD back in the dual cores days. CPU randomly just starts overheating. As was my case, I literally watched my dual cores go out with a nice little puff of smoke. Prior to this, even chrome was beating the crap out of it. As cooling wasn't the issue, I was able to determine the chip itself just bit the dust.

Fast forward to my buddy's rig, same thing, chrome started beating the crap out of his CPU and games would crash after so long. Cleaned the heatsink, reapplied thermal paste, checked the Mobo and booted it. Ran a little warm for about another 6-7 months, but ran, but eventually bit the dust as well. I can currently boot it and windows will start, but put any load on the CPU at all and temps skyrocket and asus's on board overheat protection shuts down the system. I've done research into it and opinion is that the chipset on board wasn't optimized for the FX series or that the PSU is failing. But in almost every situation, when test results are posted, this is ruled out.

In my honest opinion, having experienced this personally myself once, my friend now dealing with it, and from Google searches (seems to be a more common problem than you think) your CPU is dying just outright and the instablitiy in the chip is causing abnormal power draw all over the board.

I don't trust AMD's quality control at this point. And is the main reason I switched to Intel and have never looked back. A CPU shouldnt just out right quit 5-6 years down the road, especially with adequate cooling.

Im going to continue looking into this myself. But would recommend opening chrome and a couple tabs and let em sit. Chrome is CPU intensive for a browser and if my assumptions are right, the more tabs you open the higher your temps will climb until overheating causes a system crash.

I will keep checking and see what you find out however. Good luck man.

 

[mdd1963]

Lots of folks with failed/failing AOIs feel one very warm fluid line, radiator at room temp, other line at room temp....

(Oddly enough, many seem to believe that the fluid miraculously dropping to room temp in a .5 sec trip thru the radiator sounds normal, but, it is not.)

There may be some fluid flowing, but, many suffer severely reduced flow from blockages, lack of fluid due to microleaks, etc... Given an AIO, it's failure is several times more likely than the CPU just 'suddenly heating up abnormally'...

Try an ordinary air cooler tie strapped down to the CPU surface, if necessary.....; you'll know in 3-4 minutes of CPU-Z /bench-stress CPU if the cooler or CPU/MB was /is bad...

 

[xxsm1ttyxx]

AOI's gradually deteriorate in funtion over the course of a few years, and, then just seem to stop working, all the while confusing their owners by providing a reassuring buzzing sound to mislead folks into thinking their pumps are still fine....

A radiator at room temp should imply no warm fluid making it to it...

As to the cause being fluid loss or contaminants/hard water deposits, unknown....

This got me questioning my AIO and prompted me to make the drive to Microcenter and grab a air cooler.

Considering the op said ,I'd expect the pump to be working, unless there was some sort of blockage reducing flow.

I also would have expected this to be the case as well as iCue from Corsair was showing the pump working at max RPM as well as the fans moving to match based on thermals of the AIO liquid.

Spoke with the dude at microcenter, whos first move was "bro ur AIO is dead".

Bought a Cooler Master MA620P air cooler, wanted something just a bit more than the Evo212.

Ran Heaven 4.0 for around 2 hours on ultra settings to induce a load and let it sit. Nothing major to report back on that but CPU load was also something less than 30%. Loaded up Origin/Apex, and prior to this as i mentioned i was hitting thermal margin and then shutdown simply updating. Temps stayed no higher than 55c under full gaming load for 4 hours with stock fan curves for the CPU fans. Which didnt reach above 35% at that temp.
Im going to have to say, I 100% didnt even care to test the AIO before posting here. The software was giving me information I felt meant it couldnt be the AIO. SADFACE.

This has been a great learning lesson of not being too trusting of the software tools as they can lead you down a path of incorrect troubleshooting and possible lost monies... I was willing to drop another 100 bucks on trying a completely diff mobo from ebay. 60$ air cooler that will probably last me for the next untold amount of hours on this old rig, 1 of the simplest troubleshooting steps outside of "did you put it on right?"

I appreciate the help and words of insight. I hope to be able to return the favor in the future.

--Smitty--