Are my VRMs kicking the bucket?

[Ace331FTW]

I have recently began experimenting with overclocking my system again, for what I think is the first time since I rebuilt in this new case. I have been running an AMD FX-8350 on a Gigabyte GA-990FXA-UD3 r4.0 motherboard for the past 4.5 years, with basically no problems. When I last tried overclocking, I was able to get my CPU up to almost 4.7 GHz, cooled under a Hyper 212 Evo in a Corsair 200R mid-tower. Now I am using a Corsair H100i GTX that I bought second hand, inside a Raijintek Thetis case. The case is a bit unorthodox in that it features an inverted layout with the PSU in front, and I am pulling fresh air into the case through the radiator to cool my CPU.

Under stock conditions, the system runs fine and only maxes out at around 45 degrees or so while gaming. When I tried to reapply my OC profile I had saved from a while back, the CPU initially idled in the mid 40s. When I started an Aida64 stress test (no GPU load) the CPU temp warning on my mobo wouldn't stop screaming and the reported CPU temp was around almost 80ish degrees, package temp around 65... The fans also kicked up to jet takeoff speeds. This was weird to me, considering I had set this to only around 4.4 GHz and I had previously gone higher. Right now, I am stable at 4.3 GHz with about 1.3v, at the stock 1.4v it gets a bit too hot for comfort... When I then tried stressing my GPU only, its temps were basically unchanged to what they used to be in the other case. This case only has one 120mm exhaust near the CPU and the PSU exhausting air, and the rad bringing in air at the top, which blocked a bit by the GPU.

Could it be that the VRMs are burnt out or something, or just too warm because of ambient case temps?

There has also been an issue with Windows locking up entirely, only letting me move my mouse, forcing me to hard reboot, but that happens at stock speeds as well so I wouldn't necessarily call that related (unless the motherboard is the problem).

 

[EarthDog]

Probably not giving up the ghost, but I would imagine since you reduced airflow around the socket and VRM area by moving to the CLC that could be what is causing some issues.

That said, sounds like a mounting issue or thermal paste application issue as well as that case not having a lot of airflow it appears. I would look at remoutning and see what the paste looks like. If you want to see if it is the case, take the side panel off and see if temps drop.

Remember with your airflow, extended play will easily raise temps inside that case since you are dumping all of that load in the case through the rad and its orientation. Maybe put that as exhaust and be sure to have some intake fans for airFLOW.

 

[wingman99]

Usually when the VRM overheat they will throttle the processor clock speed down first or shutdown the PC.

 

[Ace331FTW]

Thanks for the initial suggestions but I should clarify. I don't think this is a mounting issue as the CPU is cooled just fine when at stock or my 4.3 GHz overclock, but touching the voltage at all is what makes it tick. Right now it's undervolted technically, as the stock voltage (at 1.4-1.4125) causes throttling when overclocked. Adding another fan to give some fresh air to the bottom half of the case or get some more of the hot air out would be nice, but there are no more fan spots to occupy

 

[WhitehawkEQ]

Thanks for the initial suggestions but I should clarify. I don't think this is a mounting issue as the CPU is cooled just fine when at stock or my 4.3 GHz overclock, but touching the voltage at all is what makes it tick. Right now it's undervolted technically, as the stock voltage (at 1.4-1.4125) causes throttling when overclocked. Adding another fan to give some fresh air to the bottom half of the case or get some more of the hot air out would be nice, but there are no more fan spots to occupy

IIRC 1.4V is not stock for the FX-8350, I have 5 FX-8350 systems and they all run at 4GHZ @ 1.28V +/- .02V with max temps below 40c.
Where did you read that 1.4V is stock? I would like to see it.
Try setting the CPU to 1.25V and start your OC there and see how high you get, and bump the voltage as needed.

 

[Ace331FTW]

IIRC 1.4V is not stock for the FX-8350, I have 5 FX-8350 systems and they all run at 4GHZ @ 1.28V +/- .02V with max temps below 40c.
Where did you read that 1.4V is stock? I would like to see it.
Try setting the CPU to 1.25V and start your OC there and see how high you get, and bump the voltage as needed.

The voltage is set to 1.4V when I restore optimized defaults in the BIOS. I've heard people say before that default settings sometimes set voltages way higher than what they need to be and that's definitely true with this motherboard. When I started my overclock, I did my standard procedure of:
1.) Restore optimized defaults.
2.) Take all of the auto settings and set them statically to their given values.
3.) Bump up speeds slightly
4.) Reboot and test
5.) Rinse and repeat

Only, when I got to rebooting, the temperature warnings started going off, even when I just set all the voltages statically and left the speeds stock.

I'll try overclocking with the side panel off to see if it is an air exhaust issue and then if it is, maybe try to ghetto rig a vrm fan in there somehow.

 

[WhitehawkEQ]

Turn off all C1E & C6 states, Cool & Quite and Core Performance Boost and see if that helps. Oh and any power saving settings.

 

[Ace331FTW]

I will try that when I get home from work later today, but they seem like the only things keeping my temps down in my current setup at 4.3GHz. I'm going to restart my process but start the voltage at 1.25 instead.