iCUE random fan RPM spikes

Hi,

I have a H110iGTX and the latest iCUE drivers.

I have noticed that sometimes the fan RPM goes from ~ 30% to 100% for ~1-5 seconds and then back to normal without any reason. CPU temp and Sensors are all fine. I tried Quiet and Custom curves. But the problem is reoccuring.

It is quite disdurbing because even at load the fans are running at ~30-40% and are quiet.

Here is a log of today.

 

I also tried Link4 but had the problem too. I did not had any issues when I bought the AIO. I think I first noticed this issue ~1-3 months ago.

corsair_cue_20181201_09_10_30.zip

There were always a few issues of this with the GTX coolers.What kind of coolant temperature are you seeing (idle and load)? One of the quirks is the fans will spike when too close to the last graph point on the curve. One of the workaround is to put that last point safely out of reach at 50C or so. Then use the penultimate dot for your actual required final control point.

 

However, if it is doing this on the quiet curve as well that may not be it. Are you running any other monitoring software or motherboard control software that might interfere with iCUE/Link?

Temps of AIO are: idle ~ 28 and with some load ~33-38 (depending on the load). Actuall CPU temps are idle ~28 and load ~50-55.

With my custom curve the remps are a bit higher because it set to be even quieter.

 

I have now changed the curve a bit. My custom curve has a max value of 70% whereas the spikes are (according to the RPM) 100% speed.

 

I only have the AMD GPU drivers that are monitoring the GPU. Nothing else is monitoring any sensors.

 

PS: It seems that those spikes only occur durin the first 15-45 min after CPU got load. After that I have not noticed any spikes.

Hmm... cooler is working and your temps are good. Sometimes that model would spike to 100% at 40C regardless of settings, but that does not seem to be the case.

 

Things like AI Suite, MSI Commander Center, etc. sometimes can interfere with cooler operation. Something like HWiNFO might cause a glitch when both programs poll the cooler at the same time and that can create a weird value (216C, 0C, etc.) causing momentary odd behavior. This goes not seem to be the issue either.

 

Does this happen continuously during the first 45 minutes? Or just once and done? I am afraid the GTX were quirky coolers and I don't miss them. Perhaps a current owner has seen this as well and knows another workaround.

And make sure the temperature sensor for your curves are set to the CLC's coolant, not the CPU.

What is CLC?

I do not see any option to change to sensor source. When looking at the custom curve the headline is "Sensor H110iGTX Temp". So I think the sensor is not the CPU itself but the internal sensor of the corsair AIO.

CLC = Closed Loop Cooler

 

Yup, "Sensor H110iGTX Temp" is the coolant temperature.

Yes, on the GTX series coolers you could not change the fan control variable. It will always be coolant/liquid/“H110i GTX Temp”.

 

Is this happening constantly when you first start? Or once and then not again? If it is frequent, that would suggest a problem with the internal fan controller or possibly the temp sensor. If so, I think you should start a support ticket with Corsair enroute to replacing the cooler.

Yesterday after I booted up and started a game it started like after 1-2 min. Random fan RPM spikes. After about half an hour there were no more RPM spikes.

I closed the game and let the CPU to rest (water was at ~30°). Then tried prime95 and furmark. No spikes.

The system did some resting and then I did a restart and started a game (water was at ~30°). No spikes.

Today I booted up (water was at ~23°). Started a game and RPM spikes occurred.

Now the water temp is at ~40° (~30-50min) and there are no spikes at the moment.

Do you mind if I ask what game you were playing?

Your water temp from 23 to 40? Heavy GPU game with a lot of waste heat? Normal coolant delta should be +5-7C for cpu only loads. When the GPU warms the case, that complicates the analysis but that heat does add to cpu temp. If the coolant is jumping around a bit or shows unnatural highs, it’s possible air bubbles or some other temporary blockage is causing a temp spike that results in a fan spike. Is there any variance in the pump speed? It should be steady all the time.

I think that a temperature rise from 23 (idle) to 40° of the water (after ~30-45 min of gameplay) is not much when taking into account that my fans are normally spinning at 400-700 rpm.

The CPU itself is at ~60° almost constantly (during gameplay).

Pump speed is between 2800-2900rpm (Extreme mode). I also tried quiet mode which results in constant somwhat 1800-1900rpm.

 

If there are bubbles, should not the spikes re-occur? After sometime of gameplay they are gone and do not reoccur (at least I have not heard any) for hours.

Today I booted up and just did some surfing and office. CPU had a max of ~15% usage with a temp of ~25°. Water temp has a temp of 26-26.6°.

Fans are running at 360-420RPM. And running for about an hour. (No gaming during that time).

Now I got some spikes. Not to many and that much RPM but going up to ~1000RPM (where 1500 is max spikes which happen during gaming).

Those spikes lasted now for about 1-3min. Right now they are gone again.

Keep an eye on the coolant temp when these spikes occur. There was a trigger to increase fan speed to max at 40C if the final curve point was less than that value (or something thereabouts). That was the basis of the original suggestion to set the last point out at 50C to see if this would settle it down.

The collant did not have spikes at 40°. The spikes occur somewhere around 30°.

 

And as written before: today I had spikes where collant was ~26° and CPU was ~25-26°. Just browsing and writting documents.

Now I got some spikes. Not to many and that much RPM but going up to ~1000RPM (where 1500 is max spikes which happen during gaming).

Those spikes lasted now for about 1-3min. Right now they are gone again.

 

Sorry, not how I read the above. If this is a new issue and is persistent, I would contact Corsair Tech Support through the ticket system with an eye toward replacing the cooler. I don't have another explanation other than the fan controller is failing.

Update:

I have cleaned the radiator (there was only some dust) and tightened the screws a bit.

This is now about some months ago, but after that there were no spikes anymore (since ~3 months now).