Monitor CPU and GPU fan RPM on Clevo laptops

[Martin]

The difference is because of EC, which runs proprietary firmware and thus fan speed monitoring and control doesn't follow any standards or known methods and is not disclosed to public.
Desktop systems are mostly using dedicated circuits, which are easier to understand and control, in many cases also compatible with other well-known circuits.
v5.73-3350 is the actual Beta version.

 

[mtuan293]

Just updated to version 5.74. So exited to see my fans RPM for the first time yay!!!
[attachment=2716]
Edit:
Why does it sometimes grey out like this?
[attachment=2717]

 

[Martin]

Grey values mean HWiNFO was not able to read the current value in the actual refresh cycle.

 

[mtuan293]

Grey values mean HWiNFO was not able to read the current value in the actual refresh cycle.

Oh thanks for letting me know. Btw, I'm just curious to know if these values are accurate? Because I see some differences between GPU1 and GPU2 fan. Oh, and when I set the fan to "Custom" in Clevo Control Center, I only hear the CPU fan running, the two GPU fans seem to shut off. However the GPU2 is still reporting at 2k RPM in HWinfo. And when the CPU fan actually shuts off (rarely, since Clevo sets a pretty aggressive profile makes the CPU fan running quite often), I see GPU2 is zero as well.
[attachment=2718]

 

[Martin]

It's possible that the order of fans reported by HWiNFO is not correct. In that case I'd need you to provide additional RW dumps when only one of the fans is spinning.
If that's not possible, then at least if one of them is stopped, but I will need all possible combinations. If you can provide such dumps, please name the files accordingly, so I can analyze them and determine the correct registers.

 

[mtuan293]

It's possible that the order of fans reported by HWiNFO is not correct. In that case I'd need you to provide additional RW dumps when only one of the fans is spinning.
If that's not possible, then at least if one of them is stopped, but I will need all possible combinations. If you can provide such dumps, please name the files accordingly, so I can analyze them and determine the correct registers.

So I have 3 different combinations as follow:
1. All 3 fans are ON at maximum speed 
(This one is guaranteed since it can be toggled via Fn + 1 shortcut).

2. All 3 fans are OFF 
(I took 3 attempts to make sure since the CPU fan is kicked off very often).

3. Only the CPU fan is ON, the two GPU fans are OFF 
(3 attempts, same reason as above).

If there's any inconsistencies in the EC dumps, please tell me so I can redo the process.

 

[Martin]

Thanks for the data. It seems that GPU2 fan speed reported was just a mirror of the CPU fan speed.
I will fix this in the next build.

 

[mtuan293]

Thanks for the data. It seems that GPU2 fan speed reported was just a mirror of the CPU fan speed.
I will fix this in the next build.

Thank you!
Also I want to ask if it's possible to control these fans? Since you can read values from the EC, is it possible to change them and make the fans behave as desired?

 

[Martin]

I'm sorry, I don't know how to control those fans.

 

[MegaCookie]

Hi there,
Can you help me getting the RPM readings for my Clevo P950HR? It got three fans (for reference, here is also the service manual: https://drive.google.com/file/d/1VMFqKNHc_Cvdn107IB_GYLlQeppb_2hM/view?usp=sharing)
Besides the report from HWinfo, I also added several EC readings from RWEverything and I also attached my decompiled DSDT.dsl from Clover Bootloader.
As I would like to know more about these EC values can you please tell me where those RPM readings are in the EC (offsets) and in the DSDT.dsl (which device and which formula)? I need this for some patching on Linux and Mac, but I find it hard to get into.
Thanks in advance!

Edit: I now something more, I think, specifically with only one fan. Then only 0xCE/ 0xD0/ 0xD1 is giving a value more than zero, I think this must be the CPU fan. When more fans are on, the EC offsets at 0xCF/ 0xD3/ 0xD4/ 0xE0/ 0xE1 are also becoming more than zero. When everything is off, both (0xCE/0xD01/0xD1) and (0xCF/0xD3/0xD4/0xE0/0xE1) return 0. However, I cannot find anything interesting in the DSDT.

Also the EC offsets at 0xF9/ FA/FB/FC seems like for custom fan settings which can be configured by the manufacturer's 'Clevo control center' program.

 

[Martin]

Thanks for the data, support of your model will be added in the next build.
Fan speeds in this case seem to be in the EC at 0xD0/0xD1, 0xD2/0xD3 and 0xE0/0xE1. These are the RPM1, RPM2 and RPM3 values in the DSDT.
0xCE, 0xCF might be the fan duty cycles.

 

[MegaCookie]

Thanks for the data, support of your model will be added in the next build.
Fan speeds in this case seem to be in the EC at 0xD0/0xD1, 0xD2/0xD3 and 0xE0/0xE1. These are the RPM1, RPM2 and RPM3 values in the DSDT.
0xCE, 0xCF might be the fan duty cycles.

Thanks a lot! What is the actual equation to get the fan speeds in RPM? Is there also some kind of method in the DSDT, where these values are set and get? For example Asus laptops are using the TACH device, or FRDC for HP.

 

[Martin]

Conversion formula is: 1966080 / RPMx, where RPMx is the 16-bit value read from EC.
In your DSDT I don't see the RPM being read/set, I think that some other systems were reading this only.

 

[MegaCookie]

Conversion formula is: 1966080 / RPMx, where RPMx is the 16-bit value read from EC.
In your DSDT I don't see the RPM being read/set, I think that some other systems were reading this only.

Thanks, I know a lot more now! So RPMx are the two combined EC values, right? And where did you get the 1966080 from then? Just curious

 

[doofus]

As per other thread, I own a Clevo P17SM-A c.2014, here is the report

any help greatly appreciated

 

[Martin]

As per other thread, I own a Clevo P17SM-A c.2014, here is the report

any help greatly appreciated

Your model should be already supported. If you don't see fan speed, please provide the HWiNFO Debug File for analysis.

 

[doofus]

Yes it works, all good now, thanks.