CPU underclocks and thermal safeguard shuts it down within minutes after power outage

[Clorith]

I had a power outage today (caused by a kitchen stove triggering the breakers in my home). After this, my PC immediately themral throttles while looking at the UEFI BIOS, under-clocks the CPU, and also over-heating it at the same time, leading to thermal safeguards kicking in, causing it to power down immediately. Eventually landing at 550MHz clock speed and shutting down once it hits 90C.

 

Components:

- CPU: Threadripper 1950x

- Motherboard: ASUS ROG Strix X399-E Gaming, Socket-TR4

- CPU Cooler: Corsair Hydro Series H115i

- PSU: Corsair AZ1200

- Memory: Corsair Vengeance LPX DDR4 3600MHz 4x16GB

 

When I boot up the machine seems fine, but the CPU temperature is growing by about 1 degree Celsius per second. I am able to get into UEFI BIOS I've seen the temperature start at about 55-60C (which is still way too high for a system that hasn't even reached an OS), and can watch the temp go up, thermal throttling reduce my clock speed, and eventually hit the magic 86-90C mark where it shuts down for safety.

 

I can feel that the cooler pump is running, as there is noticeable movement in the tubes when you hold them, the motherboard is also letting the CPU cooler know that temperatures are getting high, as it ramps up the fans at around 70C.

 

I am unable to make it into Windows to actually check iCUE for the coolant temperatures and such, as the time to boot is just barely too long, and the throttling has slowed windows sing-in to a crawl by that time before it shuts down.

I am not sure what information is helpful, my PC knowledge these days is... very lacking, so I've snapped some UEFI images of any screen that may potential have values of interest. These are all values with the "optimized defaults", as I have not wanted to make any modifications to the system after I originally built it 5 years ago (I am also aware that the ram is under-performing, being an early Threadripper adopter meant it wasn't doing everything right, and the official response was that you had to overclock your ram and CPU to be aligned to get the full effect, no idea if that was true or not, but I did not wish to mess around as I had no knowledge in these areas).

[glenalz81]

If you're confident the CPU is not actually overheating you can find the thermal limits option in BIOS and disable it, this should at least let you get to windows for further diagnosis.

[Clorith]

I'm not feeling very confident about it, the CPU cooler is cool to the touch, but the brackets for the CPU are very warm, I've no way of saying just how warm though. I worry that if I disable the thermal limits I may inadvertently damage the CPU in the process, but am open to suggestions/recommendations in that regard (Finding a new motherboard for the socket seems impossible right now, same for PSUs around me for some reason, so either of those needing replacement will be a more time consuming affair).

[PDifolco]

Cooler aint working if at idle you gain +1C /sec

Either it has been disconnected or it's failing

[Kid.Lazer]

Feel the temperature of the radiator where the lines enter the tank. If it's still stone cold when your CPU is overheating, the pump has failed.

[Clorith]

Great tip @Kid.Lazer! I did so, and it remained stone cold and didn't shift in temperature at all... beginning to think the feeling of movement in the tubes might have been reverberations then

 

I'll have to go looking for a new cooler then as my first step of remediation, it being an older, slightly odd socket it's going to take a few days (thinking I will go with a normal cooler this time, at least then I can see when it isn't working).

[SpookyCitrus]

This is 100% a failing/failed cooler, your idle temps in bios are way to high. Either the power surge caused it to fail or it was already going out. Just feeling the tubing and or parts of the AIO and seeing if they are warm to the touch isn't the right way about checking to see if it's working. I'd say with confidence you should just replace the AIO. 

[Kid.Lazer]

1 minute ago, Clorith said:

Great tip @Kid.Lazer! I did so, and it remained stone cold and didn't shift in temperature at all... beginning to think the feeling of movement in the tubes might have been reverberations then

 

I'll have to go looking for a new cooler then as my first step of remediation, it being an older, slightly odd socket it's going to take a few days (thinking I will go with a normal cooler this time, at least then I can see when it isn't working).

It's possible the pump motor still runs, hence you feeling the vibrations. But the pump seals may have failed or some other issue preventing it from actually moving fluid.

[DreamCat04]

14 hours ago, Clorith said:

I had a power outage today (caused by a kitchen stove triggering the breakers in my home). After this, my PC immediately themral throttles while looking at the UEFI BIOS, under-clocks the CPU, and also over-heating it at the same time, leading to thermal safeguards kicking in, causing it to power down immediately. Eventually landing at 550MHz clock speed and shutting down once it hits 90C.

 

Components:

- CPU: Threadripper 1950x

- Motherboard: ASUS ROG Strix X399-E Gaming, Socket-TR4

- CPU Cooler: Corsair Hydro Series H115i

- PSU: Corsair AZ1200

- Memory: Corsair Vengeance LPX DDR4 3600MHz 4x16GB

 

When I boot up the machine seems fine, but the CPU temperature is growing by about 1 degree Celsius per second. I am able to get into UEFI BIOS I've seen the temperature start at about 55-60C (which is still way too high for a system that hasn't even reached an OS), and can watch the temp go up, thermal throttling reduce my clock speed, and eventually hit the magic 86-90C mark where it shuts down for safety.

 

I can feel that the cooler pump is running, as there is noticeable movement in the tubes when you hold them, the motherboard is also letting the CPU cooler know that temperatures are getting high, as it ramps up the fans at around 70C.

 

I am unable to make it into Windows to actually check iCUE for the coolant temperatures and such, as the time to boot is just barely too long, and the throttling has slowed windows sing-in to a crawl by that time before it shuts down.

I am not sure what information is helpful, my PC knowledge these days is... very lacking, so I've snapped some UEFI images of any screen that may potential have values of interest. These are all values with the "optimized defaults", as I have not wanted to make any modifications to the system after I originally built it 5 years ago (I am also aware that the ram is under-performing, being an early Threadripper adopter meant it wasn't doing everything right, and the official response was that you had to overclock your ram and CPU to be aligned to get the full effect, no idea if that was true or not, but I did not wish to mess around as I had no knowledge in these areas).

That really sounds like a dead pump. Do you have any cooler that you could test it with? Because if it doesn't occur with a different cooler, then your AIO is dead (most likely the pump)

[Clorith]

Thank you all for the great suggestions, I really appreciate it!

 

This happened at the most inconvenient time, the nearest I could find to a cooler I can get to replace it with is a 7 day special order due to the somewhat dated socket; Unfortunately my vacation starts tomorrow and I will be away for a full month, so I won't get to test this until I get home.

 

I'll keep you all updated in this thread as soon as I get home and can confirm the suspected resolution is in fact the correct one (I feel very confidently about it, given all of your feedback).

 

Have a great summer y'all!

[Clorith]

Following up here as promised, got my new cooler (no more water cooling!), a heatsink with some lovely fans, as well as redoing my case airflow while I was at it (without the AiO radiator, I now have room for that). Temps are now way lower than what the AiO pump managed even just on idle, so I would call this a great success