[SOLVED] Interpolation/function identification

LaszloF · Oct 23, 2007

Dear Sirs,

I want to measure temperature with some NTC's.
The nature of these isn't linear, in the datasheet there is a table for resistance values in 5C(degree Celsius) steps. From that data i need to get a function, or make some type of interpolation, the measurement needs have 0.01C resolution.

What do you suggest? Is it even possible to make this?

Thank you in advance.

elvis09 · Oct 30, 2007

Hello LaszloF,

The best way to do tthat with the finest accuracy is to use a spline interpolation of your points. Take a look at :
http://en.wikipedia.org/wiki/Spline_interpolation

The algorithm to use for it is simple to implement. You can take an example in the Numerical Recipies. See in: http://www.nrbook.com/a/bookcpdf.php
the chapter "3.3 Cubic Spline Interpolation 113"

Good luck!

Cheers,

Elvis

LaszloF said:
Dear Sirs,

I want to measure temperature with some NTC's.
The nature of these isn't linear, in the datasheet there is a table for resistance values in 5C(degree Celsius) steps. From that data i need to get a function, or make some type of interpolation, the measurement needs have 0.01C resolution.

What do you suggest? Is it even possible to make this?

Thank you in advance.

LaszloF · Oct 31, 2007

First of all, thanks Elvis, i will look into it, meantime i found a linearization method for NTC's, it's called the Steinhart-Hart equations.

But for a general solution I will use the spline interpolation , i will try to compare the two methods on NTC's and post a feedback.

Regards,
Laszlo

[SOLVED] Interpolation/function identification

LaszloF

Member level 2

elvis09

Junior Member level 3

LaszloF

LaszloF

Member level 2

Similar threads

[SOLVED] Interpolation/function identification

LaszloF

Member level 2

elvis09

Junior Member level 3

LaszloF

Member level 2

Similar threads

Privacy & Transparency

Privacy & Transparency