[Openec] temperature sensor

[Richard A. Smith]

Frieder Ferlemann wrote:

> 
> We need to know the beta of the NTC. It makes little sense to

I've asked and I don't have it.  You will have to measure it.

http://www.specsensors.com/ntc-engineering.asp

We are not using 1% parts and have no model for what the self-heating is 
so you are _still_ going to have to calibrate it if you want accuracy.

Steinhart-Hart is more accurate over a wider range than beta is anyway.

> invest time into inaccurately measuring the voltages at a subrange of the
> temperature span while endangering the XO and still not being sure that this
> sample would be representative.

Endangering the XO?  Who told you that? We don't care about the 
accuracy. +/- 10C is fine.

The thermistor was an after thought just in case we thought we could do 
something clever with that info.  2 resistors and an unused A/D channel 
is practically free.  It has zero influence on the thermal protection 
circuits.

-- 
Richard Smith  <richard at laptop.org>
One Laptop Per Child