Hi,

I already was lurking towards PT100, but afraid of the effort - The Wheatstone bridge, the Operational Amplifiers. All has to be precisely set up, probably equal or better as 0.1% to get finally the accuracy of 0.1 C.

So - thanks for the hint to MAX31865. It reduces effort remarkably!

Martin

Hello everybody,

I would like to measure temperature in the area of -30 to +30 with a precision of 0.1 degree.
If my thermometer shows 22.5, is should be 22. 5, neither 23 nor 21.5.
What is the best way to get this precision at Arduino?
DHT22. Hm.
DB18B20 - not bad, but precision is not there, +-0.5 degrees.
HDC1000. hm.
BME280 on a break out board: All for a toss.
Maybe, somebody soldered it a bit too tough and against specification on the board.
In the area of 10 to 26 degrees, I'm getting results scattered across 6 degrees.
The sensor OK at 10 C, did not deliver properly at 25 C.
My best results i obtained with NTCs. I have 3 different ones (10K, 20K, 100 K Ohm), and they scatter a little only from each other. What is great, when I compare this to other couples of sensors.
However, taking into account the precision of the NTC and series resistor and the ADC resolution of Arduino, I do not think I'll ready for a clear .1 degree precision.

What is the best way to get precision of 0.1 degrees?

Thanks!
Martin