Jump to content
Corsair Community

H100i PRO AIO might be going bad


Daslim

Recommended Posts

So here's the thing. I posted on Reddit a couple of days ago about an issue with my Ryzen 2700x CPU. The temp would fluctuate from 40C all up to 54C while on idle. I had a few responses which one told me my pump might be bad. Now my question is how would I know if it's bad. Is there a way to check? While putting the comp under load, I reached temps of 100C. I did overclock it to 4.1 Ghz however I had to Underclock it to keep temperatures down to 85C underload. I used Prime 95 and AIDA 64 to stress test it and used Ryzen Master to check temps.
Link to comment
Share on other sites

Which temp would fluctuate up to 54C? CPU temp or H100i Pro Temp (coolant temperature)?

 

Obviously CPU temp can move around for a multitude of reasons, legitimate or not. The coolant temp shift would be more concerning. If so, check to see if the CPU temp floats up along with it. They won't be 1:1, but if the coolant temp rise is legitimate, then you might see a 27C coolant and 37C CPU temp on boot, then 10 minutes later your coolant is 40C and the CPU has a 50C minimum. It is also possible the sensor has gone bad. In that case the coolant temp may be going up, but the CPU temp remains unchanged. Both are problems that require replacement, but its better to have the cause narrowed down before submitting the Support Ticket.

Link to comment
Share on other sites

Which temp would fluctuate up to 54C? CPU temp or H100i Pro Temp (coolant temperature)?

 

Obviously CPU temp can move around for a multitude of reasons, legitimate or not. The coolant temp shift would be more concerning. If so, check to see if the CPU temp floats up along with it. They won't be 1:1, but if the coolant temp rise is legitimate, then you might see a 27C coolant and 37C CPU temp on boot, then 10 minutes later your coolant is 40C and the CPU has a 50C minimum. It is also possible the sensor has gone bad. In that case the coolant temp may be going up, but the CPU temp remains unchanged. Both are problems that require replacement, but its better to have the cause narrowed down before submitting the Support Ticket.

I just want to say thank you for taking the time to help me figure this out. I just booted my desktop and I am checking on ICUE. It reports the coolant at 32-33C at 1110 rpm while the CPU keeps jumping from 41C to 52C and then goes back down before jumping again.
Link to comment
Share on other sites

OK, that isn't all that unusual anymore since our CPUs aren't sedated with heavy C-states like from a few years back. Check to see if the corresponding temp jump matches a change in Vcore. It likely does.

 

Second part of the check-up is load temps, but be careful with casual synthetic testing. Most of these will respond very aggressively, especially if you are using standard Auto voltage. The BIOS and CPU are designed to keep you stable at all costs and that generally means pouring on voltage. I am not a 2700x owner and I can't give you exact voltages and temperature ranges, but that info is out there. What I would recommend is using a milder and even type of stress test, like the "Bench Tab" stress test in CPU-Z. AIDA and OCCT are good for checking stability. We want to check thermal behavior and that is easier with a constant load vs deliberate fluctuations. Also, if there is a contact problem between CPU and cooler, you'll flunk the easy stress test too. No need to drop an anvil on it like unmodified Prime 95. You should see the CPU jump to a specific temp the moment you start the test. It will then hold steady, except for the occasional +1C to coolant and CPU temp and coolant heats up. This should be slow, maybe +1C every 30 seconds. I would expect the 2700x and 240mm radiator to peak out around +6-8C after 10 minutes. No need to run any more than that or if it levels off after 5 minutes there is no need to continue either.

Link to comment
Share on other sites

OK, that isn't all that unusual anymore since our CPUs aren't sedated with heavy C-states like from a few years back. Check to see if the corresponding temp jump matches a change in Vcore. It likely does.

 

Second part of the check-up is load temps, but be careful with casual synthetic testing. Most of these will respond very aggressively, especially if you are using standard Auto voltage. The BIOS and CPU are designed to keep you stable at all costs and that generally means pouring on voltage. I am not a 2700x owner and I can't give you exact voltages and temperature ranges, but that info is out there. What I would recommend is using a milder and even type of stress test, like the "Bench Tab" stress test in CPU-Z. AIDA and OCCT are good for checking stability. We want to check thermal behavior and that is easier with a constant load vs deliberate fluctuations. Also, if there is a contact problem between CPU and cooler, you'll flunk the easy stress test too. No need to drop an anvil on it like unmodified Prime 95. You should see the CPU jump to a specific temp the moment you start the test. It will then hold steady, except for the occasional +1C to coolant and CPU temp and coolant heats up. This should be slow, maybe +1C every 30 seconds. I would expect the 2700x and 240mm radiator to peak out around +6-8C after 10 minutes. No need to run any more than that or if it levels off after 5 minutes there is no need to continue either.

 

So as I said before I did stress test before since I was trying to get this CPU at 4.2 - 4.1 at all cores. I was able to get 4.15 however once I run the stress test, it runs but just shuts down. I tried messing around with the voltage and the most stable I could get was 4.0ghz at 1.25v It stayed around 80- 85C and didn't go any higher than that. I didn't check the coolant temps during that time though.

 

EDIT: Also I'm not too familiar with my cooler since it was given to me by a friend. Is it one to be used with overclocking CPU's?

Link to comment
Share on other sites

All coolers can be used for overclocking. It's not the cooler that is the limitation. The CPU and voltage is the liming factor.

 

Be aware not all 100% loads are equal. Prime 95 may very well take your CPU to uncomfortable temperatures, but that does not mean you will hit that if you are heavily loaded doing normal things. That is why I suggested a more representative type of test that usually does come close to matching real world temp number.

Link to comment
Share on other sites

×
×
  • Create New...