This morning I updated to the latest version of Windows (likely not relevant) as well as my Adobe Creative Cloud suite (also likely not relevant) and turned my computer off. After a few hours I returned home and started my computer up. It ran for roughly 5 minutes before starting to slow right down and eventually turning off. At first I thought it was a power issue so continued turning it on although it continued to turn off.
After some inspection it appears that on startup the CPU begins at 60C but then makes it way up to 90C+ before the PC shuts off completely. There's only 1% utilisation on the CPU and hardly any on the RAM either. I bought some thermal paste and reapplied to the CPU although it didn't make a difference.
My only thought is that the Corsair H100i V2 is running fairly hot but I hadn't noticed any corrosion or damage where the thermal paste was. The PC has run for 5 years without any issues. Is there anything else that could be causing this or do I just have a dodgy Threadripper CPU?
PC Specs:
AMD Ryzen Threaripper 1920x
Corsair H100i V2
4 x 8GB Corsair DDR4 Vengeance 3000MHz
Gigabyte X399 AORUS Gaming 7 TR4 ATX Gaming Motherboard
Seasonic Prime 750W 80PLUS
Solved! Go to Solution.
Update 24.06
I RMA’d the cooler through Corsair as it still had a few months of the warranty left. Six weeks later I received a replacement H100x (I had to wait an additional few weeks for a TR4 bracket to be sent out separately!).
After fitting the cooler these are the new clock speeds and temps the CPU is getting:
Does this looks normal? It’s been running like this for over an hour now.
Thanks for your help in diagnosing the issue Milchspende.
Some screenshots before the CPU completely overheats and the computer shuts off:
Instead of saying you have a dodgy threadripper, go buy a new cpu cooler!
All your information indicates a faulty cpu cooler or incorrect installation of the cooler, yet you think its the cpu
Thanks for your reply. What makes you think that it's the cooler?
Is there something that I can do to test this?
@TobyH wrote:
begins at 60C but then makes it way up to 90C+
That is a clear indication, that the cpu cooler has no, or little to none contact to the cpu to transfer heat away from it
Easiest way to be sure would be to check with a new cooler, since you already changed the thermal compound
Did you check for leaks of the corsair cooler?
Yes I have checked for leaks but found nothing! The computer is cleaned regularly, roughly every 1-2 months so not a dust issue either. On startup the CPU is only running at 0.5MHz (usually sits at 3.5GHz+) which is why I'm leaning towards CPU failure.
As I mentioned previously, the entire build has been untouched and running for the past 5 years without failure - this is why there shouldn't be a contact issue between the cooler and CPU.
Your CPU is thermal throttling.
If you have been running a AIO Watercooler for 5 years, it definitely looks like a cooler problem.
AIOs tend to leak after some years.
My first AIO (Corsair H60) leaked after almost 2 years.
Anyway, if you want to lower your temperatures and prevent thermal throttling you either have to check for contact between cpu and cooler plate, thermal compound, leaks, or just get a new cooler, to save time and hassle
Thanks for your response.
Is it not odd that the CPU was running at 0.5GHz? It usually sits closer to the 3.5GHz mark so I assumed that this meant it was on the way out.
iCue was giving me readings for the cooler's temperature, pump speed and fan speed so I would have assumed this would be enough to say that it's working properly. Regardless of that I have submitted an RMA under warranty to Corsair as it's just under 5 years old.
I'll let you know how I go.
@TobyH wrote:Thanks for your response.
Is it not odd that the CPU was running at 0.5GHz? It usually sits closer to the 3.5GHz mar
The CPU Downclocks to 0.5Ghz to lower the heat output and protect itself
Lower Clockspeeds->lower temperature
Makes sense.. Thanks again for your help.
I'm hoping that I haven't damaged the CPU by running it to it's max temp several times.
This sounds more a like an issue with the CPU itself.
A 1920x requires only modest cooling, just the standard air cooler and a couple of venting fans should keep it idling under 40'C in the summer and 30'C in the winter.
Some suggestions:
Update 24.06
I RMA’d the cooler through Corsair as it still had a few months of the warranty left. Six weeks later I received a replacement H100x (I had to wait an additional few weeks for a TR4 bracket to be sent out separately!).
After fitting the cooler these are the new clock speeds and temps the CPU is getting:
Does this looks normal? It’s been running like this for over an hour now.
Thanks for your help in diagnosing the issue Milchspende.