LIS3DH Temperature calibration and drift over reset cycles

[ICunn.14]

Understanding that the temperature sensor on the LIS3DH provides a relative value rather than absolute... Is it possible to calibrate this value for an individual sensor against a known measured temperature and then use the relative changing value to calculate changes that would persist through multiple reset cycles? Might this value drift over time and lose accuracy from the calibration or will it stay relatively accurate, say within 1 or 2C?

Thanks!

[Eleon BORLINI]

Hi @ICunn.14​ , unfortunately there is no register for temperature offset and/or gain calibration in LIS3DH. The relative value is referred to 25°C: if you test the device in a controlled environment, and you read a value different from 0h in the temperature output register, you could store that value and reuse it for your further calculations. If you want to be more accurate in temperature, I suggest you to use the LIS2DTW12 sensor. Regards

[ICunn.14]

Thanks Eleon,

That sounds like it confirms my approach then, we'll test the value at a known temperature during initial configuration and save that value as a reference point in firmware. When we power the LIS3DH and take a reading we'll compare it against the calibration value we stored to determine deviations. We have the ability to periodically update this calibration if there is drift, but I hope it won't change much over the lifetime.

I'll look into the LIS2DTW12 for future but it looks like a different footprint so that will wait for a future design iteration.

Regards,

Ian