Channel Calibration information

[FTI_Newton]

I apologize if the question is not in the appropriate board - I could find none better.

I haven't found any official NI documentation on Virtual Channel Calibration that answers my questions.

How is the data interpolated between points in a table? Straight Line segments? A curve fit? 

Can I force a linear best fit instead of a table? Is it possible to get stats back from the fit (R-value)?

I currently don't use it because I can't report on the uncertainty in my per-channel system calibration. I'm using Global scales with global virtual channels right now. It works, but the built-in channel cal would be much cleaner and simpler all around.

[HeatherBaker]

Hi FTI Newton,

Either a table or polynomial calibration is performed, depending on the number of data points entered.  You can read back which method is used, as well as, the values with the DAQmx Channel Property node.  Hopefully this helps!

[FTI_Newton]

Not entirely.

What is the criteria NI uses (# of points) to decide between polynomial and table?

Can I force it to be table or polynomial?

If a table, are datapoints between table values linearly interpolated?

For a polynomial, can I get any information on goodness of fit? MSE or Rvalue?

Can I force the order of the polynomial, in particular, force a linear best line fit?

[HeatherBaker]

After looking more into the behavior and settings, a table is always what is used from the Calibration tab within MAX.  See the attached VI as a workaround and to get the information you need.  The channel will be calibrated in MAX, then the values will be extracted and run through the polynomial fit VI where you can specify the order and retrieve the R value.  This information can be written to the property node and force the polynomial calibration.  A programmatic implementation in MAX/DAQmx would be worth a suggestion on the Idea Exchange and I recommend posting there.

[FTI_Newton]

No good for stored calibrations - changing it runtime is not traceable. 

Thanks for looking into it - appreciated.