Corsair iCue H115i RGB PRO XT Poor Cooling

[CDGames]

I've recently had issues with cooling with my water cooler. They've been manageable issues before, but now they're getting out of hand. At stock speeds (without any undervolting, although my board tends to overvolt my CPU so I would be undervolting anyways), my temperatures sit at around 80 degrees at idle. After applying a more appropriate voltage, temperatures sit closer to 60, sometimes dipping down to 60. This is in my new almost-bare computer room, which is converted from my garage. It gets fairly cold in here (when it's at its coldest, CPU temperatures average around 40 degrees). It's actually fine at sustaining that temperature through a gaming session (usually getting to 70, which is okay, but when you're paying any amount for a cooler rather than using the stock one, you want more than "okay"), but a run of cinebench throttles immediately and crashes my computer

I was using a Thermaltake cooler for a short time after getting this one because, for a short period of time, I was using a case that wouldn't actually fit this CPU cooler due to the square profile of the radiator, and temperatures were significantly better with that than I ever got with the H115i. I switched back, however, because iCue is a vastly superior software to Thermaltake's software, but looking online, people are getting much better performance than I have been

I was probably going to get a new cooler eventually anyways (maybe get a custom loop, which lots of people seem to be doing with this case anyways), so I didn't want to RMA it if I'm going to switch it out very soon, so I was wondering what everybody thought I should do, or if everybody has any suggestions on what I should try

I've tried adjusting the screws on the mount itself, but those are as tight as they can go, and loosening them any amount hasn't proved effective either. I've tried separate settings (I have my own "Balanced/Quiet" profile that I made and then abandoned because it wasn't performing well enough), I've tried re-orienting the tubes to keep them away from other warm components, and I've tried new memory, which I recently just upgraded. I have not tried using another motherboard, however. While I was using my Thermaltake cooler, I was using a Mini-ITX motherboard due to it being a Mini-ITX build, but now I'm in a full-sized ATX case, so I switched back to my old Micro-ATX motherboard. The board I had before also utilized an LGA-1151 mounting style, rather than the AM4 mounting style, so my mounting pressure may be more even between the two. With all the times I've removed the cooler, however, I've gotten to see the thermal paste application and it's been already (albeit a bit heavy a few times, but I've been applying less and less)

 

My case: Lian Li 011D Razer Edition

My CPU: Ryzen 9 3900x

My GPU: EVGA RTX 2080 Super Black Gaming

My cooler: H115i RGB PRO XT

My power supply: Corsair SF750

[c-attack]

Anything this extreme is either full blown cooler failure or a contact issue. For the cooler, a clear sign of problems is the H115i Temp starts relatively cool but keeps ticking upwards along with the CPU temp, about +1C every few seconds until it stops around 50-60C. That is a zero or low flow state. Since coolant temp is minimum possible CPU temp, it will keep it elevated at all times and makes load next to impossible. Once you heat up, it takes hours of powered off for it too cool down since the heat can't get to the radiator. Coolant temp is high, but radiator exhaust is cool.

 

On the other hand if the H115i Temp is sitting in the 20-30C range and your CPU temps are banging between 70-100C instantly, that is probably a contact issue. This is fiddly stuff and I don't use the AM4 bracket. Make sure you are using the backplate from this cooler and not another. Too much TIM will make a bit of a mess, but not cause this type of issue. In rare cases where someone puts gobs and gobs on, it can create little heat bubbles in the TIM. So your temps seem normal, but there are these weird short term blimps in temp and then the bubble pops and the CPU temp drops again. That does not seem like what is happening hear and this is rare in general. I wouldn't worry too much about a razor thin layer of TIM. The 3900X die size is not super large. A dot in the center should do it and the cold plate will smear it out. Even if you are painting the processor lid, this is not the normal result so I think the focus should be on brackets, standoffs, and the general socket area. DO NOT over tighten the screws. Nothing good can ever come from that. Normal hand tightening is always enough.

Edited February 25, 2021 by c-attack

[CDGames]

The AM4 mount doesn't use a backplate, rather it attaches with two clips around the CPU; one on top and one on bottom, which attaches with a screw and a weird standoff that wraps around the clips rather than screwing directly into the motherboard. I would love to be able to use a proper mounting mechanism, but unfortunately, this is the only thing that works with this cooler. If there's another method of mounting this, please let me know, 'cause I might be doing something wrong, although this is what it says to do in the manual. Please let me know!

[c-attack]

Can you clarify which of the two scenarios you find yourself in? Is the CPU temp slowly building from boot getting a few degrees warmer at steady intervals? Or is it instantly 70-90C from the moment you press power on?

[CDGames]

Can you clarify which of the two scenarios you find yourself in? Is the CPU temp slowly building from boot getting a few degrees warmer at steady intervals? Or is it instantly 70-90C from the moment you press power on?

Sorry for the delayed response

It doesn't go up to 70-90C when I power it on, it only goes to 60 (undervolted, which is how I prefer it regardless of whether or not it overheats. With the stock voltage, that being around 1.4 volts, it instantly rises to 80C). It does do so instantly, however

My main thing is that it rises to 100C instantly during a Cinebench load regardless of my voltage. I can keep it at 70C while gaming, and even around 75C while rendering videos, however temperatures do slowly rise during this situations, and I wouldn't want to leave them like that

[Ih78isme]

If it was me I'd remove the Corsair pump/CPU block, and run it uninstalled to check its performance, but that requires you having another cooler you can use in its place so that you can still boot into windows and run the ICUE software.

 

Run it for 30 minutes, and is the coldplate still cold after use, if it's getting warm to the touch then clearly something wrong with the pump or water distribution, can you apply different pump speeds via the ICUE software, are they within expected ranges? What is the temp of the pump block uninstalled via the ICUE software.

 

Any sloshing of water when you shake the radiator?

[c-attack]

It seems like you are saying the CPU is idling around 60C after boot. That would be the low voltage situation which then becomes unmanageable once full voltage is applied. However, it would simplify things considerably if you would look at the H115i XT Temp in CUE. It will start off and room temp when powered down and slowly warm up to about +4-7C over ambient after the PC is turned on in a normal, non-maximal load state. If the H100i XT Temp is sitting in the 40-60C range, that clearly indicates a flow problem of some type and the unit needs to be replaced.