iCue is not doing what it's supposed to do - some fans are not spinning correctly even though properly set up.

[GiNT0NiC]

 

Hello Corsair Community,

I have noticed the few months that my few AIO fans(only Fans #1,#2 and #3) do not always spin properly, even during gaming. No problems had with pump.

In the iCue Software "Gaming" cooling setting it is set that way. See picture. The AMD R7 5800x CPU package is selected as the sensor.
In the picture it looks like everything is running OK so far.  So i have ~2000RPM on around ~85% Fan curve it works properly.

German words in english:
Kühlungsvoreinstellungen = Cooling presets
Kühler = Cooler
Lüfter = Fans
Lüfterkurve = Fan curve

 I have created 2 Coolersettings:

1.Gaming you can already see in this picture



And 

2.Maximum = everything turned up to maximum

 

The Problem:

However, even though I am currently in the game now and the temperature is so high, the fans spin very low and constantly. Sometimes all fans are affected and sometimes it's fan#3 and the other fans#1 and #2 spin normally.

Recorded Screenvideo:

https://drive.google.com/file/d/1NGU-NNvwFMzqYbkGOAAZAVjJBtB4JOa6/view?usp=share_link

You can see the Sensor shows 66°C and the Fans shows you ~900 RPM something is wrong. You can see on Fan curve ~90% and all fans should turn up to ~90%. Around 1900 - 2100 RPM.

 

Here it comes. The funny thing is, if I change the fan settings of all or depending on which fan is affected the cooling settings e.g. of fan#1 from "Gaming" to "Maximum" or switching the "Sensor" as in the video, then the fan updates/synchronizes properly. 

And if i start the PC and surfing on the Internet or playing game and from the beginning if leave the cooler settings of all fans at "Maximum", i have no problems at all all fans rotate properly to maximum. However, with "Gaming" cooler settings there is the problem.  In my opinion this is a bug in the iCue software.

 Hwinfo, Aida oder Sandra are not installed on my PC.  I have TechPowerUP GPU-Z, MSI Afterburner, Armoury Crate, G.Skill, L-Connect3 and Roccat Swarm for Keyboard, Mouse and Headset RGB Controlling.

I have already reinstalled/uninstalled the Icue several times with Guide and installed newest version but nothing happens. 
I have now V4.32.129. Would RMA help ? The problem is that I don't have a spare AIO cooler... It seems Softwareissue (i think maybe I am wrong after all)

 

Greetz

GiNT0NiC

 

 

 

Edited February 1 by GiNT0NiC
Update

[c-attack]

The first issue is your have set your CPU temp as the control source for the radiator fans.  This is not the correct variable.  Your CPU is cooled conductively.  Fan speed does not affect that process.  The pump speed and fans are responsible for transporting that heat to the radiator where it is expelled -- or effectively it takes out the trash.  You don't need to do that every time you drop something in the bin.  With your fans you are managing the H150i Pro XT Temp or coolant temperature.  This is the baseline or lowest possible CPU temp with zero volts on the CPU.  +-1C to the coolant also is +-1C to the CPU.  So if you set up a stress test and the coolant has increased 0.7C in the 10 seconds you have been running, then the fans do not have an impact at this point.  If after 3 minutes the coolant has increased by +6C, then the most you can reduce the CPU temp by is 6C at maximum speed.  In reality you will not be able to get back to zero in a single radiator system.  

 

If you set the sensor choice back to H150i XT Temp and then select any of the shape tools in the lower right corner of the custom graph, that will give you a good starting point for the expected coolant range.  Every environment is unique and you should tweak the baseline and maximum speeds to your preference.  Radiator fan speed has a low to moderate effect on coolant temperature but a very small effect on the over all CPU temperature on larger radiators.  The entire point of the 360mm size is you don't need to blast the fans.  Typical coolant delta for 300W on a 360mm radiator might be +10C at 1300 rpm.  However, I don't think you can generate that kind of wattage so I would expect something more like +6-7C at maximum.  Be aware if you heat up the case with the GPU, this will add temperature to every component in the case, including the radiator system.  Most users in mixed use systems will see more coolant temp rise in gaming from the GPU than if they ran a full bore CPU stress test.  

 

A few possible reasons why the system may not be responding the way you expected in the curves above.

1) Your CPU temp is extremely dynamic when playing games.  The Commander XT and Core now have some type of hysteresis to prevent the inevitable up/down/up/down reaction when you link fans to a volatile source.  We don't know a lot about the parameters for the hysteresis and I don't use either controller and can't add much here.

2) Issue with the data.  There have been some CPUID temp value issues over the past few months. I believe AMD package temps were affected at one point.  Aside from the reasons stated above, this is another reason to use the coolant temp data not just for radiator fans but the remaining case fans as well.  That sensor is native to the controller and will work in all operational states, not just when CUE is running.  It should make more smoother, less noisy fan control with no loss of performance.  

[GiNT0NiC]

On 2/3/2023 at 11:16 PM, c-attack said:

The first issue is your have set your CPU temp as the control source for the radiator fans.  This is not the correct variable.  Your CPU is cooled conductively.  Fan speed does not affect that process.  The pump speed and fans are responsible for transporting that heat to the radiator where it is expelled -- or effectively it takes out the trash.  You don't need to do that every time you drop something in the bin.  With your fans you are managing the H150i Pro XT Temp or coolant temperature.  This is the baseline or lowest possible CPU temp with zero volts on the CPU.  +-1C to the coolant also is +-1C to the CPU.  So if you set up a stress test and the coolant has increased 0.7C in the 10 seconds you have been running, then the fans do not have an impact at this point.  If after 3 minutes the coolant has increased by +6C, then the most you can reduce the CPU temp by is 6C at maximum speed.  In reality you will not be able to get back to zero in a single radiator system.  

 

If you set the sensor choice back to H150i XT Temp and then select any of the shape tools in the lower right corner of the custom graph, that will give you a good starting point for the expected coolant range.  Every environment is unique and you should tweak the baseline and maximum speeds to your preference.  Radiator fan speed has a low to moderate effect on coolant temperature but a very small effect on the over all CPU temperature on larger radiators.  The entire point of the 360mm size is you don't need to blast the fans.  Typical coolant delta for 300W on a 360mm radiator might be +10C at 1300 rpm.  However, I don't think you can generate that kind of wattage so I would expect something more like +6-7C at maximum.  Be aware if you heat up the case with the GPU, this will add temperature to every component in the case, including the radiator system.  Most users in mixed use systems will see more coolant temp rise in gaming from the GPU than if they ran a full bore CPU stress test.  

 

A few possible reasons why the system may not be responding the way you expected in the curves above.

1) Your CPU temp is extremely dynamic when playing games.  The Commander XT and Core now have some type of hysteresis to prevent the inevitable up/down/up/down reaction when you link fans to a volatile source.  We don't know a lot about the parameters for the hysteresis and I don't use either controller and can't add much here.

2) Issue with the data.  There have been some CPUID temp value issues over the past few months. I believe AMD package temps were affected at one point.  Aside from the reasons stated above, this is another reason to use the coolant temp data not just for radiator fans but the remaining case fans as well.  That sensor is native to the controller and will work in all operational states, not just when CUE is running.  It should make more smoother, less noisy fan control with no loss of performance.  

I haven't Commander XT ^^ my  CPU AIO is directly connected via USB. I have normal 7 Fans and 3 AIO Fans runs already actually cool. So i have set all my fan settings to "Intensive".  I will monitor this for the next few days.
 

But the problem with the H150 XT Temp is that it takes the sensor from the pump temperature ~28 -30°C. Most of the time it is around 25-35°C.  So when I play a resource consuming game like CP2077 or BF2042, the CPU temperature increases massively. I mean the CPU 5800x Packet Temp is the main sensor for CPU heat or not ? And if the actual CPU temperature gets much higher than with the H150 XT sensor, yes it reacts slower to the increased temperature, which could be a problem if the 5800x quickly reaches >70°C and the h150 xt fan can't keep up ?  

 

[c-attack]

The CPU is cooled conductively. Fans don’t cool the cpu whether it is a water or an air cooler. They blow the heat conducted into the radiator elsewhere. Your cpu is hot because you are applying voltage to the underside and nothing you do with the fans will change that. If all that mattered was fan speed, then we could all run 7GHz by using 4000 rpm fans. It doesn’t work that way and difference between small and large cpu coolers is not a large reduction in temperature but a large reduction in noise. You are blasting your fans without purpose and loosing your advantage. 
 

The AIO was designed to use coolant temperature as its control variable. If you keep forcing it to use cpu temperature you are likely to continue to encounter problems with the controller.