Python • Re: very noisy SoC temperature measurement (RPi5)

After reading all this, I just round the temps to the nearest degree, which I think is good enough for my purposes.

This is a perfectly valid choice. However the way you do it doesn't address the noise issue.
Remember that there's a constant +/- 3LSB noise, and that 1LSB = 0.55°C
Sampling temperature without taking this into account leads to data having poor physical relevance.Look at min and max values. Do you think that over each of these 10sec intervals there was truly a 4°C excursion?
I'd suggest you observe min and max not on raw temperature but on denoised temperature.
There's some python code for this in a previous post.

Statistics: Posted by pragma — Tue Feb 06, 2024 9:24 am

---