Sensor readings are off

[Raz0r]

Hey guys,

somehow my VR readings are way off..
Im currently on version 5.92-3580 and also tried the newest beta with the same results.
Is this fixable?

 

[Martin]

It looks like the VOUT value is off. But I'm not sure if this is fixable, because judging by the data displayed it seems that the VR returns such invalid data.
Is the "Power (POUT)" value also shown?
Can you please attach the HWiNFO Debug File with sensor data?

 

[Raz0r]

It looks like the VOUT value is off. But I'm not sure if this is fixable, because judging by the data displayed it seems that the VR returns such invalid data.
Is the "Power (POUT)" value also shown?
Can you please attach the HWiNFO Debug File with sensor data?

No, POUT is missing.
Also the value of VOUT sometimes turns grey.

 

[Martin]

I'm sorry, this is an issue of the VRM - it returns invalid data.

 

[Raz0r]

I'm sorry, this is an issue of the VRM - it returns invalid data.

Hmmm sad to hear, but alright. Thanks for your help

 

[Raz0r]

Just for infos, I figured out something superweird:
If I launch my official AsRock A-Tuning tool, the sensors are getting read out super accurate. No invalid data at all. I just have to launch the tool.

 

[Martin]

That's interesting. Does A-Tuning need to be running all the time, or does it work also after you close it?

 

[Raz0r]

Seems like it's also working after I close it. Will keep an eye on it.

 

[Martin]

Please let me know the result.
It might be helpful if you could attach a HWiNFO Debug File with sensor data - one before you launch A-Tuning (when you get erratic values) and another one right after launching A-Tuning. I need to compare what's the difference.

 

[Raz0r]

So I restarted my PC, launched HWINFO and as usual, the sensors didnt even show up (every vrm out)

Then I closed HWINFO, started A-Tuning and launched HWINFO and voila, everything there with correct readings. Might be some kernel driver of A-Tuning...

 

[Martin]

Thanks.
It looks like A-Tuning does some adjustments to the VRM settings (probably sets some offsets), but there are too many differences and since I don't have full documentation for that VRM I won't risk breaking something by attempting to change its configuration.

 

[Raz0r]

Alright, thanks for looking at it.

 

[Martin]

I have an idea, but not sure if that will work. Would you be willing to run a test using a special test build?

 

[Raz0r]

I have an idea, but not sure if that will work. Would you be willing to run a test using a special test build?

If nothing gets permanently wrecked, sure

 

[Martin]

Don't worry, it should be safe
Please try this build: www.hwinfo.com/beta/hwi64_601_3646.zip
Now you should see 2 VRM sensors, each reporting different data, the second one should be for Vcore.
Please let me know how it works...

 

[Raz0r]

Restarted my PC and tried it... Seems all good also looks like the bottom temperature reading is right, because I checked afterward with the official build and A-Tuning running. 

 

[Martin]

Thanks for the feedback. That's expected - the first sensor is for the integrated GPU rail and second one for the main CPU core (IA).
Looks all right from my side too

 

[Raz0r]

So HWINFO read from the iGPU VRM at the beginning and A-Tuning fixed that somehow?

 

[Martin]

This VRM supports 2 rails and one needs to switch them before reading.
By default the iGPU rail is visible and that's what HWiNFO was reading before.
Now I updated HWiNFO to switch the rails and so can report both.

 

[Raz0r]

Awesome! Thanks Martin