I recall divergence from expected were edge cases and I am not sure they would be significant or observed by most. Nearly all the maths tests for correctness had passed previously, presumably they now all do. So I wouldn't expect it to be changing most calculations.A contributing factor to the possible differences in thermistor readings might be related to recent changes in the Pico SDK's floating point handling.
I am not sure the Pico SDK itself has changed; I believe MicroPython moved to using their own functions. It did beg the question for me whether they were saying Pico SDK is flawed for the edge cases but I haven't been following that saga closely.
Statistics: Posted by hippy — Wed Jan 24, 2024 12:01 pm