Barometer affected by charging

The barometric reading has to be temperature-compensated (air varies in density with temperature, of course) and charging warms up the battery and rest of the watch's internals beyond ambient.

Ergo, it will move the barometer reading slightly while it is going on. It should go back to where it was after the charging is complete and the watch returns to ambient temperature.

(Note that there is also an error band on the sensor; I don't know how many bits of resolution it has or what the endpoints are on its readout.)

The barometric reading has to be temperature-compensated (air varies in density with temperature, of course) and charging warms up the battery and rest of the watch's internals beyond ambient.

Ergo, it will move the barometer reading slightly while it is going on. It should go back to where it was after the charging is complete and the watch returns to ambient temperature.

(Note that there is also an error band on the sensor; I don't know how many bits of resolution it has or what the endpoints are on its readout.)

Ok, but also body temperature could affect barometer reading, but with body temperature barometer doesn't change...

If it's temperature compensated, wouldn't that keep the reading steady as temperature changed? But 1 or 2 mbars might just be "noise".

If it's temperature compensated, wouldn't that keep the reading steady as temperature changed? But 1 or 2 mbars might just be "noise".

Right , pressure won't change even if temperature changes... However, 1 or 2 millibar are about 10-20 meters, so not so minor variations...

If it's temperature compensated, wouldn't that keep the reading steady as temperature changed? But 1 or 2 mbars might just be "noise".

Yes, if the temperature sensor is in the same place as the barometric sensor. This is why wearing the watch doesn't hose you in this regard, since the heating is pretty general.

But charging puts specific heat gradients into the watch that are not uniform (the battery and charging circuit) and the delta is potentially fairly large. I can see that accounting for a fair bit of error while charging is going on, and added to the "one count" uncertainty (whatever it is) likely accounts for the rest.

BTW I manually calibrated at sea level a couple of days ago, have autocalibrate off, and now am at a place that the watch says is 1424'. The "official survey" altitude is 1447 here. That's pretty close....