So this is a relatively technical question.
I am currently playing around with adding NTC 10k thermistors to my weather station, alongside other types of thermometers I already use.
Right now I have a test setup with a Wemos D1 Mini (similar to the Arduino) and an Adafruit ads1115 as my analog to digital converter. I am using the sketch suggested by Adafruit:
https://learn.adafruit.com/thermistor/using-a-thermistor All this works quite well, and during daytime the thermistor temperature values line up nicely with the PT100 sensors in the same actively ventilated radiation shield.
[ You are not allowed to view attachments ]
However, when the temperature drops, I see a deviation. See this chart incuding last night and this morning.
[ You are not allowed to view attachments ]
The thermistor reads about 0,3-0,4 higher at around 9 degrees celsius. When temperatures rise, all sensors come into agreement again. I have not yet been able to see what happens when temperatures rise well above 20. Am waiting for this to happen.
In the mean time, my question: what could cause this? Specifically the thermistor reading 'too high' when temperatures decrease. Imprecision of the thermistor itself? Since the resistance increases when temperatures drop, I was thinking perhaps the inaccuracy of the thermistor becomes apparent? Looking for clues. Anyone have experience with this?
Will be adding additional thermistors with higher precision reference resistors soon. Will be interesting to see the results.