diode sense for temperature

I think it sounds like a great idea! Unfortunately, I am a hobbyist, and not an EE. PLEASE post the schematic, if you come up with anything.

Thanks,

Dave

Not sure about how well 1N4148s fare in such applications but I've used the Vbe of transistors very sucessfully in temperature monitoring applications.

Do bear in mind that a couple chosen at random may not match very closely.

Graham

Look at those, NTC and PTC, perfect for what you want to do. you'll need to have a pot on one of them to balance the circuit.

--
"I\'d rather have a bottle in front of me than a frontal lobotomy"
http://webpages.charter.net/jamie_5

lobotomy"

Hide quoted text -

Could you stick them in some sort of 'Wheatstone' bridge circuit. Balance the bridge while both diodes even if they different are at same temperature and then calibrate as one thinks you will have to anyway for actual differences in their respective tempearures? Just an idea?

only looking for "gross" differences in temp. im not looking for anything resembling laboratory accuracy

difference span may be 5-10-20 degrees or more,

repeatabilty is non-critical, within a few degrees is ok.

tolerance can be +- 2-4 deg either way to allow for built in hysteysis (hysterical)

u know what i mean , oh jeez i'm getting my friend al's hymers disease!

or maybe i'll just buy some warm wool sweaters and toss this project ;-)) .

usage,

state

difference

lobotomy"

Hide quoted text -

Diodes make exquisitely linear temperature sensors, but note that each unit will have a different calibration factor. When fed with constant currents, the (negative) voltage slopes will be different. (You can use an ordinary ohms range on a DMM to test.) You can manually adjust for this, or there are ways to compensate using current mirrors with transistor pairs instead of diodes (look up "band gap reference"), but if that is too much trouble, consider using temperature measurement chips that already do that for you.

Best regards,

Bob Masta DAQARTA v3.50 Data AcQuisition And Real-Time Analysis

Scope, Spectrum, Spectrogram, FREE Signal Generator Science with your sound card!

Sure, that will work, but an IC temperature sensor, or a thinfilm RTD, would be more repeatable. Or a pair of series-wired, back-to-back thermocouples, but you'd only get about 30-40 microvolts per deg C.

Put ceramic caps right across the diodes so they don't pick up rf from cell phones and such.

John

It's not that hard to do, so don't give up yet. You can use execute your original idea using two diodes and a comparator. You need some way to trim (calibrate) the circuit; a single potentiometer will adjust for the diode difference and the comparator offset, all in one fell swoop.

Maybe if you were going to make 1000 of them and wanted to save a few bucks. Why not use something like LM34? The datasheet even has "remote" application examples.

--
Ben Jackson AD7GD

http://www.ben.com/

You'd have to either match the diodes or calibrate them for Vf.

But once you know the offset, they both follow the same curve, so sure, why not?

FWIW, I did a little experiment not too long ago with just one, and found that they work quite well. :-)

Good Luck! Rich

I'm guessing: A: he's already got the diodes B: they don't need a third wire for a power source C: he's only looking for a differential reading

At least, those are the reasons I'd use diodes.

Cheers! Rich