I think I’ll be happy with just two closely-matched probes, rather than absolute accuracy…
In that case just put them next to each other in a room, wait half an hour and measure.
Absolute value would matter if swapping with someone else though.
1 Like
I filled my silicone collapsible coffee cup (silicone is a decent insulator) with ice cubes, then filled it with cold water from a bottle I keep in the fridge, then left it about half an hour. My IR thermometer recorded the water as 0.7°C and the surface of a melting ice cubes as 0.0. Probably the ice couldn’t quite bring the water down nearer to zero.
I reckon this is the best way to get a reproducible temperature. Any cup or mug would do as the melting ice should hold the water at a constant temperature for some time (the latent heat of fusion effect, iirc).
It took at least five minutes for the sensor to stabilise once in the water (I floated them between ice cubes).
On my multimeter I needed to change the range from 20 kΩ to 200 kΩ after putting them in ice water; there was a moment when I thought putting one in water had broken it 
.
1 Like
From my experience at the NPL, calibration at 0 °C should used crushed, wet ice, like in a slush machine. It should look grey rather than white. Put the sensors in plastic bags and immerse the leads in the ice.
Check results by repeated measurements. For an amateur it’s possible to get an uncertainty around 0.1 °C.
It’s a surprisingly difficult measurement to make!
1 Like
I found that as well. Good point on using slush ice, I’ll remember that for the next time. The plastic bags sure will complicate matters I guess. I only used this method for waterproof sensors before and just completely immersed them.
From your description the sensor is not a PRT but a negative temperature coefficient thermistor.
There are standards for inferring temperature from resistance with PRTs with uncertainty at a level of 0.01 °C or better, but there is no agreed formula for inferring the temperature from the resistance of an NTC sensor. At the level of 0.1 °C, the standard equations are all nominal. And presumably the Vaillant software has picked one of the formulas, but we don’t know what it is.
1 Like
We have Vaillant manuals which contain tabulated data. I attached the full file.
For reference, here are screenshots of the tables. German only, unfortunately.
installations-und-betriebsanleitung-hydraulikstation-meh-97-6-2567180 (1).pdf (1.8 MB)
1 Like
I’ve got the same problem. Any ideas where I can get sensors or if my installer needs to sort? Hp is only about 9 months old.
See above.