Would also be curious to know if Davis is using the Heater in the SHT31 in later firmware revisions to improve temperature accuracy.
But since you've mentioned it, I do find myself curious as to how heating a temperature sensor can enhance its accuracy?
Yeah, there's probably a few of us trying to figure out this one....
Although the whole device is tiny, (2.5 x 2.5mm), maybe the heater is very localised to the humidity sensing element & isolated from the temp sensor part or the device.
Also the data sheet states the heaters power at 4.5 to 33mW, and being that it's a completely digital device, any heating input influencing the temp element could be calculated & adjusted for within the device. The data sheet says it's for "plausibility testing only".
But I've read other stuff discussing it's relevance to improved humidity readings in high humidity circumstances when condensation is present.
Dunno, but in loose terms you could say turn the heater on for a bit, ignore the temp for a bit & focus on the humidity; turn it off; go back to reading the temp again.
Also plausible is a device (as in board) with two STH31's present. One with the heater on the other heater off.
Bit Academic, but interesting anyway.
Phil