Help requested for temperature sensor calibration

[MarkE]

A two point calibration gives you slope and offset for each sensor.

The reliability of the values depends on how repeatable the raw calibration values are.  Measure carefully, measure often, hold conditions as pristine as possible.  To the extent that you have truly random noise, averaging more measurements will help remove the noise from the calibration.  Ditto for live measurements.  A problem that is often encountered is that things that are not random seem random because they are not recognized as being associated with something else going on.  Averaging lots of measurements do not reliably attenuate noise from deterministic sources.  When measuring temperature in a room, air currents of all kinds can disturb measurements, (ventilation on/off, doors opening and closing, people walking by ... ) as can the sun shining through a window, etc.

[s3370389]

A two point calibration gives you slope and offset for each sensor.

The reliability of the values depends on how repeatable the raw calibration values are.  Measure carefully, measure often, hold conditions as pristine as possible.  To the extent that you have truly random noise, averaging more measurements will help remove the noise from the calibration.  Ditto for live measurements.  A problem that is often encountered is that things that are not random seem random because they are not recognized as being associated with something else going on.  Averaging lots of measurements do not reliably attenuate noise from deterministic sources.  When measuring temperature in a room, air currents of all kinds can disturb measurements, (ventilation on/off, doors opening and closing, people walking by ... ) as can the sun shining through a window, etc.

This calibration will be done in a clean room. But I will give this some thought.

After I have redone the ice-bath I will post the graph to show the effects of stirring, if anybody else is interested.

Thanks MarkE for taking the time.

Regards,
s3370389