Valid for use with hash command: Thermistor

Theta is a measure of the amount of self-heating in the thermistor. Because, unavoidably, there must be some current flowing through the thermistor, it will heat up. The heating up will cause it to give a slightly higher temperature reading than it should.

The Theta hash function allows for some compensation for self-heating. Note that it will never be dead accurate, and if mis-applied may do more harm than good. You don’t have to use Theta.

The Rt parameter is the thermal resistance of the thermistor, in ºC/mW. For example, if Rt is 1ºC/mW, it means that if 1mW of heat is dissipated in the thermistor it will self-heat by 1ºC. If 1.23mW is dissipated it will self-heat by 1.23ºC, etc. The SPLat/PC is able to calculate the temperature rise for all measurement temperatures (it will vary) and compensate for it in the generated polynomial.

Thermistor data sheets may specify thermal resistance, dissipation factor, or neither (in the latter case you can’t really compensate for it).

Dissipation factor is the inverse of thermal resistance. Here’s an example from a data sheet:

This is given in units of mW/K. The K simply denotes Kelvin, which for this purpose is the same as Celsius. So you can read that as 2.2mW/ºC, which means it takes 2.2mW of dissipation to raise the temperature by 1ºC. Now, the Theta hash function wants the inverse figure, so the number we want is 1/2.2 or 0.455ºC/mW.

(One of the reasons we did it that way around was to allow you to specify 0 as the parameter value, and not hit divide by zero problems later)

Theta can be a slippery customer!

The thermal resistance of a thermistor will vary wildly with its operating environment. If you put a fan on it, it will get rid of heat more easily and therefore not self-heat so much (a fact than can be used to measure air flow with two identical thermistors dissipating different amounts of heat). If you put the thermistor in water or oil, the heat will also be removed more quickly. If you attach a heatsink to it it will self-heat less. Even a blob of Blu-Tack will act as a heatsink, because it increases the surface area. A consequence of the last observation is that if you package a thermistor bead in, say, a brass protective ferrule or immersion probe, you will also alter (almost certainly lower) its thermal resistance.

In conclusion, the figure provided in a data sheet is useful only if you use the thermistor without additional packaging, and in still air.