Thermistor help

Hey guys,
A friend of mine gave me a couple of thermistors to play around with and I’ve never used them. I’m having a bit of trouble and am hoping someone could help me out.

First problem: The thermistors are a 5k ohm and a 2.2k ohm thermistor. I have a 5.1k ohm resistor to pair with the first, a proper 2.2k ohm resistor for the 2nd. Is not having a precise pairing an insurmountable problem?

Second: Neither thermistor has a data sheet that includes the constants for the steinhart equation which seems to be the go-to route for computing temp. Is there a way for me to figure out the constants from the information provided in the data sheet?

Third: All of the arduino tutorials on thermistors use a 10k ohm thermistor. Why is that?

I know that I could make a fairly accurate guess at temp using the data tables provided in the data sheets with the two thermistors but I’m concerned about memory in the program I’m inserting this into. It’s a fairly long and complex oven control system and memory is somewhat of a concern.

The thermistors I was given are:
2.2k Ohm thermistor:http://www.newark.com/jsp/search/productdetail.jsp?SKU=21C9790
5k Ohm thermistor:http://www.newark.com/jsp/search/productdetail.jsp?SKU=86R9240

The datasheets are accessible through the Newark pages if you want to take a peek.

I haven't followed the links you gave, but since the resistance of a thermistor varies with temperature it must be obvious to you that the exact resistance you connect to it doesn't matter.

I assume you're using the thermistor and a constant resistor as a voltage divider and using an analog input to read the voltage - then you will calculate the thermistor resistance from the voltage and the temperature from the resistance.

Not much of a problem at all. Resistor tolerances, thermistor tolerances, and AD noise all come into play. There are plenty of stock thermistor libs that will do the job none of those tables are needed. I like to use a long term average rather than trusting any one reading when it comes to thermistors.

Thanks to both of you.

Yea I was planning on doing a 2 readings per second and working with a 5 reading average and including a test to discard readings that are more than 10% off from any other reading.

I understand that I can easily compute the current resistance of the thermistor, it's less clear to me, since resistance per degree changes with temp, how I account for that in computing the current temperature without referencing a specific thermistor's datasheet.

It could be I'm missing the forest for the trees here.

I would google Steinhart-Hart calibration to see if there is a program or method to determine the constants given a few data points.

Quick google found this. I didn't get a chance to read it, but it may help: http://www.mstarlabs.com/sensors/thermistor-calibration.html

Arrch I think that's perfect! Huge thank you. I'm not sure how I missed it.

Sometimes you hit a stumbling block and just can't see what's right in front of you. Thanks again.

nbm326: Arrch I think that's perfect! Huge thank you. I'm not sure how I missed it.

Sometimes you hit a stumbling block and just can't see what's right in front of you. Thanks again.

No problem. I've been messing with one lately for my Sous Vide cooker and have been going back and forth on whether to put in an equation or a table full of values.

Another thing to keep in mind is that the sensitivity to temperature changes is greatest when the thermistor resistance equals the passive resistance, eg 5.1K or 2.2K.

IOW, you probably need to use a 50K or 100K thermistor rather than a 10K thermistor to get max sensitivity for readings around room temperature. But then, max sensitivity may not be an issue here.

Yea so right now I'm messing around with a Sous Vide cooker as well, just as a platform to play around with heating elements a bit before building this oven for a friend. Eventually I'm going to be using a thermocouple at which time I'll probably be hitting the forums again for advice :)

The problem is that I don't want to rely on an implementation that won't be applicable once I plug in the pieces to the larger oven program. So in this case room temperature sensing isn't really a big deal for me, and I think a 5k should be plenty since the water temps will all be 110+ degrees