Calibration problems

I am using a NI-6036E board connected to a simple op-amp amplifier circuit to 
measure forces from FlexiForce resistive force sensors.  I have Comedi version 
0.7.70.  My data acquisition code is probably most similar to the sample code 
tut2.c.  I am having some trouble with the calibration.

I took a small data set to give an idea of the magnitude of the error that I'm 
getting, and here are my results:
At zero load, I measure between 0.058V and 0.075V at the terminal board, but 
my reading from Comedi is anywhere in the range of 0.008V to 0.071V, and the 
larger voltages at the terminal board don't correspond to larger values from 
Comedi.  When I load the sensor such that I measure between 4.12V and 4.18V at 
the terminal board, I get readings between 3.50V and 3.51V from Comedi.  I'm 
set on the -10V to +10V range.

I am running the calibration function at startup.  I have submitted some 
questions about this before.  I appreciate the replies, but none of the 
suggestions that I tried seemed to help.  Does anyone know what kind of error 
I should expect for this kind of measurement?  Is this reasonable?  I could 
deal with a DC offset if it was a consistent offset across the entire 
measurement range, and if it was the same every time I re-boot and 
re-calibrate.  Any more ideas?

Has anyone tried bypassing the Comedi calibration and working directly from 
the integer raw data output?  Would it likely be more reliable than what I 
have now?

Thanks,

Heather

Received on 2006-07-13Z19:37:12