Re: comedi_calibrate

On Wed, Jun 05, 2002 at 09:43:09AM +0000, Caleb Tennis wrote:
> 
> Just curious about the effects of using comedi_calibrate on my NI E series 
> board.  My main concern is that after running comedi_calibrate that my 
> calibrations in my program for the equipment connected will be modified, 
> somewhat substantially.  
> 
> Dave, can you give a reason why it is beneficial to run comedi_calibrate 
> (particularly for this series of boards) at each bootup like you say in the 
> docs?  With the industrial equipment that i have connected to my board, it's 
> not feasible to calibrate each piece of equipment very often (bi-annually or 
> annually is what's generally regarded as acceptable). 

The board boots into a state where all the trim DACs are set
to the end of their range.  So if you measure a grounded
signal, you'll typically see -0.1 volts.  comedi_calibrate
adjusts the caldacs so that the measurement is relatively
close to the actual signal.  In the optimal case, the error
is a lot smaller than 1 bit, e.g., 10.0/4096 volts for a
12 bit A/D converter, and typically better than the accuracy
of the A/D converter.

Note that except for the 0 offset, comedi_calibrate depends on
knowing the precise voltage of the high-stability voltage reference
on the boards.  This still needs to be measured annually for
proper calibration.  The actual value of the reference is stored
in EEPROM, and I'm not sure that comedi_calibrate reads it
correctly on all boards.

One might assume that if you don't run comedi_calibrate, you'll
still have stability of measurements, i.e., measuring X volts
will return the same value over the course of a year.  But the
boards are not speced for this behavior, and I have no data to
support or refute it.

comedi_calibrate is currently designed to fit the role of
quiet autocalibration at boot time.  It's missing a number of
features that would be useful in an industrial setting, such
as decent logging, and being able to save and restore calibration
settings.



dave...

Received on 2002-06-05Z17:27:30