AMD FX 8120 Overheating Issue

[logykphlo]

Hi,

I'm hoping someone here will be able to help me. I'm having an issue at the moment with my CPU idling at high temps and causing system shutdowns due to overheating when under load.

To give an example of the types of temperatures I'm getting, I benched my tower for testing a couple of days ago, with both sides off and all fans set to 100% speed, with the PC sat in the BIOS I was idling at 49 degrees. Going into an OS bumped the idle temperature to 60+ degrees

Some relevant details:

CPU: AMD FX 8120 - NOT OC'd - Using BIOS defaults. Voltage is currently at 1.380.
Motherboard: ASUS Sabertooth 990FX - BIOS fully updated.
Heatsink/Fan: Arctic Freezer 13

My case is a Coolermaster Silencio 550, I have 2 120mm fans mounted on the front (also removed the front door for better cooling) that are positioned to both draw air into the case, a 120mm fan mounted on the rear to expel air and I've also machined an extra grill onto the top of the case, directly above the heatsink with another 120mm fan expelling air.

The CPU/Heatsink have been removed, old thermal paste removed from both and a fresh layer applied, the thermal paste used is Arctic Silver 5, with a small drop applied, and then spread in a thin layer to give 100% coverage of the contact area, there is no excess paste. The CPU/Heatsink have been reseated and have a solid, complete contact.

All fans/vents/heatsink/case internals etc... have been cleaned out and are completely dust free. Internal cabling has been routed/tied back to allow maximum possible air flow through the case. The case is positioned on the floor but is raised by a couple of pieces of wood to allow air flow under the case, it is not obstructed on any side and is not positioned near any external heat sources.

I have underclocked the CPU as well as reducing the voltage, this only made a couple of degrees of difference to the temperature, and not enough to prevent the overheating while under load. As mentioned above, the case has been tested with the sides off, and again whilst the temperature dropped a couple of degrees the difference was only marginal (we're talking 49 degrees idle in BIOS with sides off vs 52 degrees with sides on).

I'm utterly at a loss as to what could be the problem, any help/advice would be much appreciated and if any more information about my setup is required please feel free to ask.

Kind regards

Gerry

 

[Morfinel]

Welcome

First thing to check is those stupid plastick pins on the cooler if you tighten screws too much this pins come out.
Scond is the thermal paste, people in this situation put too much thinking it will be better
This cpu cooler is not ideal for AMD cpu as the plate is smaller therefore doesn't cover the cpu entirely.
And finally hot weather

 

[Hagar]

I too would be looking at the cooler as the first culprit. Probably putting in a higher spec thermalright, phanteks or noctua would be a better option.

Prior to watercooling, I used a phanteks 140mm dual fan on my 8350 at 4.7GHz which worked well through last summer.

 

[logykphlo]

Thank you for the responses

@Morfinel - as I stated, there is no excess to the thermal paste I'm well aware that too much can cause an insulating effect rather than a conductive one.

My initial thoughts were that it could be a heatsink issue as you both have stated, however I've been running this system for a while with this exact setup and it has consistently idled at 28-30 degrees in Windows up until recently (with nothing changing in the system) and all of a sudden it's shot up by 20 - 30 degrees with no apparent cause.

There is no damage to the heatsink/fan, the fan still spins up to full RPM without a problem, not sure which plastic pins on the cooler you're talking about Morfinel as the only plastic component is the fan. But in any case, nothing is loose/popped out/disconnected on the heatsink, and it is firmly seated.

I think if I do change the heatsink I'll probably get a new case with a top mounting for a liquid cooling reservoir and stick something like the corsair h100i in there, unless you think a single 120mm mounted liquid cooling solution would suffice? I'm just kinda hesitant to throw money at the problem when it may not necessarily solve it.

 

[Morfinel]

I realised after posting this plastic pins are only for Intel, did your gpu tems went up as well ? Try to run it when it's colder and if you have any software check what rpm this fan is running at, if you have any other pwn fan swap them even if it require to holding the other fan next to cooler

 

[logykphlo]

No the GPU temps and the ambient temperatures inside the case are just a few degrees higher than normal, which can be attributed to the current weather, it's just the CPU that's having issues.

I currently have the fan locked to run at 100% which equates to 2300 RPM ish.

 

[pandem0nium]

I would've said contact with the heatsink as well. What ive done with a similar cooler is ghetto mod two 120 mm either side of the hs. Attach them to each other at the corners with cable ties.

 

[deshepherd]

One thought - when running in BIOS does CnQ come into play or is the CPU pinned at full speed. I saw something similar when I upgraded to a 8350 last week. Temp in BIOS was in 40s (nb in hot top floor room which was at 28C) and reduced a bit when I disabled the fan speed control for CPU. In windows at idle when CPU reduced freq to 1.4GHz temps were much lower.

 

[Phixsator]

A friend of mine had the same issues with his Fx-8350, found out it that the fan on the heatsink was broken and didnt spin as it should, sometimes even stopping completely. Easy way to test this would be to remove the CPU fan from the heatsink and put a 120mm or 140 mm fan (whatever you have really) infront of the intake of the heatsink.. While temps wouldnt be amazing they should drop a fair bit and you will know if its the fan not performing as it should.