[time-nuts] temperature sensor
brooke at pacific.net
Mon Jul 21 15:09:51 EDT 2014
A number of decades ago I wanted an easy way to get precision temperature measurements when using HP Rocky Mountain Basic.
HP was selling thermistors specified to be accurate to 0.1 deg C when you interchanged thermistors. The readout can be
much more sensitive for applications like oven temperature control where you are concerned with changes in temperature.
But they are very nonlinear and at first that looked like a show stopper, but when you use the Steinhart-Hart equation
it turns out that the temperature is given by a linear equation that's straight forward to compute. See:
AFAICR both leads were orange orange (the color code for accuracy was lead colors) and cost about $40 each a long time ago.
Modern low cost (pennies each) thermistors are typically 10k Ohms at room temperature and the Steinhart-Hart equation
coefficients are not given. Once you know the equation applies it's not too difficult to make some measurements and
derive the coefficients.
More information about the time-nuts