Quoting "M. Adam Davis" : > Is this the reverse biased method: > http://www.freepatentsonline.com/5955793.html > > If so, need a bit longer for this patent to run out (unless you want > to invalidate it, which takes time and money). > > It's hard to beat a cheap thermistor for cost, though. Even Digikey > has $0.57 leaded thermistors in quantities of one, and when you get > down to surface mount you easily halve that cost - in volume they are > very cheap. > > Can you give more implementation details on the below method, if it's > not the patented one above? > > -Adam The method involves measuring the difference in Vf of a diode (more typically a diode-connected transistor) at two currents-- typically something like 10:1 ratio between the two. You get accuracy that can be guaranteed to be within a few degrees C, typically better, and interchangeability. Transistors with low Rb are more suitable, or you can use more complex methods to correct for resistance. The downsides are that your signal level goes down by 10:1, so you've got more like -200uV/K than -2mV/K, and you're still measuring voltage, so you need an accurate reference or you need to calibrate for the reference voltage. Obviously you need some kind of amplification (0.1C resolution requires ADC resolution of 10 or 20 microvolts) for most applications. This sort of thing is not too difficult to do in a mixed-signal ASIC, and it's exactly how the temperature of the CPU in your computer is measured (of course the "transistor" is integrated into the CPU chip, not an external device). Best regards, Spehro Pefhany -- "it's the network..." "The Journey is the reward" s...@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist