RAK WisBlock 4631 reports incorrect temperature in telemetry starting with firmware v1.11.

[Emmpunkt]

Description:
My RAK 4631 is sending incorrect temperature values in the telemetry data. The reported temperature is consistently about 8°C higher than the temperature measured by the attached BME sensor.
This issue appears to start with firmware v1.11.0.
When I downgrade to v1.10.0, the temperature readings are correct again.
Observed behavior:

• Firmware 1.11.0 and newer → temperature is ~8°C too high
• Firmware 1.10.0 → correct BME temperature
• It looks as if the device is reporting the CPU temperature instead of the BME sensor temperature
  Question:
  Is anyone else seeing incorrect temperature readings with firmware 1.11.0 or later?
  Could there be a regression where the wrong sensor source is used?

[IoTThinks]

If there is BME sensor attached to the repeater, the temperature of the BME sensor will overwrite the temperature of the MCU temperature.

Do you see humidity and other telemetries of BME sensor?

[Emmpunkt]

Yes , i can see humidity, pressure, altitude...
But i have to correct myself. With 1.10. its not working.
The values shown are static, they don´t change.

[IoTThinks]

Yes , i can see humidity, pressure, altitude...

Then BME sensor is attached and working.
By right, the temperature of the BME sensor should override the value of MCU temperature.
I have tested it fine before on repeaters.

Let me check again in a day or two.
May due to false positive or fimrware issue or change of mobile app.

BTW, is your RAK a repeater or companion?

[Emmpunkt]

Its a repeater

[weebl2000]

Its a repeater

Try building & flashing this one: https://mcimages.weebl.me?commitId=fix-temp-telemetry - I think this should fix the issue with CPU temperature being preferrred over external sensor.

[Emmpunkt]

Sorry for the late answer. I did not try yet because of #1878
Will it boot from battery if i build it with your fix?

[weebl2000]

Sorry for the late answer. I did not try yet because of #1878 Will it boot from battery if i build it with your fix?

If you merge this branch and the other one then I think yes. Or build from here, it contains both fixes: https://mcimages.weebl.me?commitId=rak4631-boot-and-telem

[IoTThinks]

@Emmpunkt The temperature from BME280 overrides the MCU temperature correctly in my testing just now.
I implemented this feature and have tested it fine before.
Usually, if you have the BME sensors in the same box or same room, the temperature for NRF52 can be different from BME sensors a few degrees only.

This is RAK4631 and BME280. I plugged and unplugged the cable of RAK4631 during the testing.

This is with BME280 sensor attached.
When I flow some "human air" to the BME280, the temperature is jumping. You can test also.

This is without BME280 sensor. Usually in room temperature, MCU temperature is higher than BME280 temperature.

So it seems to work for me.
At least, the temperature from BME280 override the MCU temperature correctly.

This issue appears to start with firmware v1.11.0.
When I downgrade to v1.10.0, the temperature readings are correct again.

The MCU temperature was added since repeater 1.12.
Could you help to check if the issue start with firmware 1.11 or 1.12?

And are you using BME280 or BME680?
BME280 has some fixed to mitigate BME280 self-heating causing inaccurate readings.
cc28b1a

Before jumping to a fix, I guess we need to confirm where the error is first.

[Emmpunkt]

Its a BME280.
Sorry, because of my work shift i´m not able to make bigger tests.
The BME sits outside of the housing in some sort of Stevenson screen.
I flashed the firmware by @weebl2000 , but i didn´t see any difference.
Maybe i did the build wrong or the sensor is just broken. In the moment the temperature difference is around 2°C, wich is lower than before. But i need the time and a replacement sensor for further testing.
Humidity and pressure are valid, only temperature is off.

[IoTThinks]

@Emmpunkt It may due to a fix in BME280 on its own instead of the sensor source.

BME280 has some fixes to mitigate BME280 self-heating causing inaccurate readings.
cc28b1a

No rush.
Update us if you have some new test results.

[IoTThinks]

@Emmpunkt How is your testing?
You can blow some hot air into the BME280 alone (not the RAK) and see if the temperature jumps.

[Emmpunkt]

I made myself a test unit today for testing the BME.
Just an ESP32-C3 with Esphome on it.
The shown temperatures where the same on Esphome and MC. So my my BME might be broken.
Humidity and Pressure are correct, but the temperature is to high...

[IoTThinks]

@Emmpunkt Please blow some hot air into BME sensor to see if the temperature go up.
If it is, then the temperature is from BME sensor (not MCU sensor)