Thermistor Linearization

Hi, Two recent projects involved industrial sensors that required temperature compensation. The budget dictated thermistors.

The sensor's temperature response was linear, and of course thermistors are very non-linear.

I accomplished the first project with hours and hours of modeling in Excel and my other favorite math tools.

For the second project I wrote Thermistor Cal design software.

If you'd care to take a look it is here -->

formatting link

  Cheers, Tim
formatting link
Reply to
Tim
Loading thread data ...

Tyco Electronics sells some linear thermistors.

formatting link

Reply to
Arlet

I don't see any mention of the Steenhart-Hart three parameter fitting function, nor any sign of a third fitting parameter.

--
Bill Sloman, Nijmegen
Reply to
bill.sloman

Hi, Excellent comments, Thank you.

I should have clarified. The program is for NTC thermistors

The program uses the Exponential Model.

The formula is: R = R0 * exp( B * (1 / T - 1 / T0 ))

where - R is the thermistor's resistance at temperature T in degrees Kelvin .

R0 is the thermistor's nominal resistance at reference temperature T0 in degrees Kelvin

B is the thermistor's B Constant as specified by the manufacturers data sheet

T0 is thermistor's the reference temperature as specified by the manufacturers data sheet, normally 298.16 degrees K (273.16 + 25 C)

T is temperature of interest in degrees Kelvin ( K = 273.16 + degrees C)

Thanks, Tim

formatting link

Reply to
Tim

Sorry. That should have been Steinhart-Hart

formatting link

A few years ago the Review of Scientific Instruments published a paper advocating the simple exponential model for restricted temperature ranges or when a certain amount of inaccuracy could be tolerated. The three-paramater Steinhart-Hart relationship is also imperfect, but is good for errors of a few thousandsth of a degree Kelvin, where the two-parameter exponential relationship tends to produce errors about ten times larger.

--
Bill Sloman, Nijmegen
Reply to
bill.sloman

Same old world! Early 1960 and working in test group we were concerned by providing accurate results of our measurements to the "curious" clients. Slide rules not good enough, 7 places logaritmic charts - a lot of work, thermocouple equations with 8 segment equations need computers to start comprehending what it is about. In the end you arrive on solutions: a) Asking if a child is sick or not, you get a mother of 6 kids and in one simple touch she gives proper answer. b) On a (geometrical) plane you can draw a straight line through _any_ three points if the line is thick enough!

Have fun.

Stanislaw.

Reply to
Stanislaw Flatto

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.