iCUE showing a drastically different temperature to ryzen master (2700x)

Im trying to set up fan profiles using my 2 commander pro's, but the cpu temp is way off, by 50c This is making it so i cannot make a fan profile and all 11 of my ML RGB fans are roaring 24/7.

 

I think that the problem is with how the temperature is read with the 2700x

 

https://i.imgur.com/wiHlIgX.png

Pretty much only Ryzen Master and HWInfo read it correctly, best off to just set custom curves with the offset in mind.

Just setup a curve with a +50c offset, and its keeping my fans under control, hopefully this will be fixed in iCUE or LINK as the problem persists with both applications

Or use a different control variable than CPU temp. There is no direct correlation between that and any fan functionality.

Or use a different control variable than CPU temp. There is no direct correlation between that and any fan functionality.

 

I understand that, but there isn't really a temp sensor that I would like to use other than the CPU. I'm using a fractal design s36 so there's no temp sensor I can use other than the CPU one.

 

It just needs to read the actual temperature instead of having a +50c offset, baring that, it needs to have an advanced options menu in iCUE to set your own offset for the temperature.

 

Up to yet I haven't seen the temps go higher than 65c in Ryzen master, but they go all the way up to 115c in iCUE and Link

Question for you: What is it that you are directly trying to cool with your case fans?

 

Answer for you: That'd be the internal case temperature. NOT the CPU. If you have an air cooled system, the CPU temp is related to the internal case temp ... but not exclusively. So is the GPU. By basing case fans on the CPU temperature, you may not get enough airflow when it's your GPU that's doing all the works and heating things up. Like when you're gaming.

 

Everyone defaults to the CPU temp because, for a long time, that was the only choice that we had. But that doesn't mean that it's the best choice. You have a Commander Pro. It has 4 temperature probes. These make for excellent source values as you can place on in the warmest part of the case and use that to control your case fans.

Easy choice would be coolant temperature if you have a Corsair i-series cooler. The coolant temp is affected by both case temp and actual waste heat temps, so it winds up being a good indicator of need. If not, utilize any of the temp probes for the Commander Pro, as suggested above. You can run one up to your CPU radiator exhaust (if applicable) to mimic coolant temperature. Another spot is above the GPU, toward the rear exhaust. Both locations tend to give you enough data range to make the control gradual and probably better than straight case air temp, which can have a very tiny range in a well ventilated case.

Question for you: What is it that you are directly trying to cool with your case fans?

 

Answer for you: That'd be the internal case temperature. NOT the CPU. If you have an air cooled system, the CPU temp is related to the internal case temp ... but not exclusively. So is the GPU. By basing case fans on the CPU temperature, you may not get enough airflow when it's your GPU that's doing all the works and heating things up. Like when you're gaming.

 

Everyone defaults to the CPU temp because, for a long time, that was the only choice that we had. But that doesn't mean that it's the best choice. You have a Commander Pro. It has 4 temperature probes. These make for excellent source values as you can place on in the warmest part of the case and use that to control your case fans.

 

I have 11 fans in my system, the bottom intake, bottom front intake and rear exhaust are based on my GPU temp, and the top 2 intake and 6 fans on my radiator in push-pull on my CPU temp, I will see if I can figure something out using the temp probes

 

-Edit-

So I found that my VRM temperatures scale 1:1 with my CPU temp with a -10c offset, so I figured I could just use the VRM probe as my temp sensor. This temp sensor doesn't exceed 100c (actually 50c) so I can keep control of my fans even when it goes above 50c

I have 11 fans in my system, the bottom intake, bottom front intake and rear exaust are based on my gpu temp, and the top 2 intake and 6 fans on my radiator in push-pull on my cpu temp, i will see if i can figure something out using the temp probes

 

For a water cooled system, your radiator fans shouldn't be based on the CPU temperature either.

 

What do the fans on a radiator actually cool? Is it the CPU? No! It's the coolant. So the fan speed should be based off of the coolant temperature. Most of Corsair's AIOs provide the coolant temperature as a control value and you can use this. If you don't have this available, a thermistor in the exhaust of the radiator will give you a good approximation of the coolant temperature.

 

You see a similar design in automobiles. Your "engine temp" in your car isn't the actual temperature of the engine, but the temperature of the coolant. Your car's thermostat controls the radiator fan and the coolant flow based on the coolant temperature, not the temperature of the actual engine where fuel is being burned. The same thing applies to liquid cooling for your system.

For a water cooled system, your radiator fans shouldn't be based on the CPU temperature either.

 

What do the fans on a radiator actually cool? Is it the CPU? No! It's the coolant. So the fan speed should be based off of the coolant temperature. Most of Corsair's AIOs provide the coolant temperature as a control value and you can use this. If you don't have this available, a thermistor in the exhaust of the radiator will give you a good approximation of the coolant temperature.

 

You see a similar design in automobiles. Your "engine temp" in your car isn't the actual temperature of the engine, but the temperature of the coolant. Your car's thermostat controls the radiator fan and the coolant flow based on the coolant temperature, not the temperature of the actual engine where fuel is being burned. The same thing applies to liquid cooling for your system.

 

My cooler is a fractal design s36, so I don't have a thermister inside my loop or one that I can monitor at the least.

So, unfortunately, I cannot base it on any loop temp or even case temp as the air is being blown directly outside of the chassis.

I know using CPU temp, or even VRM temp is not ideal, but it seems to work. All the air coming out of my system when stress testing is cool to lukewarm meaning it is now doing its job in my eyes.

My cooler is a fractal design s36, so I don't have a thermister inside my loop or one that I can monitor at the least.

So, unfortunately, I cannot base it on any loop temp or even case temp as the air is being blown directly outside of the chassis.

I know using CPU temp, or even VRM temp is not ideal, but it seems to work. All the air coming out of my system when stress testing is cool to lukewarm meaning it is now doing its job in my eyes.

 

And that's why I suggested the thermistor in the exhaust flow. But if it's working for you ... excellent.

I like VRM temp as the control variables for some boards. It was very reliable with a predictable range on Intel X99. Not often an option for many other chipsets. As suggested previously, you can run the end of one of the thermistor wires from the Commander Pro kit to the exhaust side of the S36. Within 1C, exhaust air temp is coolant temp. You have 8 probes slots with 2 C-Pros. Shame not to use at least one them. One of the benefits of using the probe temps is C-Pro fan controllers will respect the variable and fan curve, even when Link/iCUE is not running. That includes the boot cycle. No more power blast on start-up/wake. Runs even if the software glitches or you need to quit the program to use a different monitoring tool. I find that part to be very valuable.