[Gb-ccb] Calibrations of the CCB inputs

[Martin Shepherd]

I have now run calibrations of all of the CCB input ports, using
Randy's test detector circuit fed with voltages generated by the GPIO
card. To do this I am generating a staircase ramp of 819 differential
voltages, covering the full input range of the ADCs, of -2.5V to 2.5V,
and at each voltage I am then measuring the mean of 16384 ADC samples
(using dump mode), along with the RMS deviation of these samples from
the mean. I am then performing a weighted least squares fit of a
straight line to these means, versus the differential
input-voltages. This yields a calibration factor, a zero-offset, and a
correlation factor, which reflect the characteristics of the
combination of the test detector circuit, the GPIO digital-to-analog
converters and the CCB ADCs.

The results for the 16 ADC inputs are as follows:

  Port     Cal factor          Offset      Correlation coefficient
  name     (ADC counts/V)    ADC counts         (unitless)
  ----     --------------    ----------    -----------------------
  J1         -3256.6           8208.1             0.999998
  J2         -3257.1           8216.2             0.999998
  J3         -3237.9           8195.5             0.999998
  J4         -3260.0           8210.7             0.999998
  J5         -3262.7           8186.0             0.999999
  J6         -3253.4           8206.7             0.999999
  J7         -3249.6           8204.6             0.999998
  J8         -3263.0           8202.6             0.999998
  J9         -3236.2           8199.7             0.999999
  J10        -3252.5           8217.2             0.999999
  J11        -3251.1           8196.1             0.999999
  J12        -3256.0           8193.4             0.999999
  J13        -3247.7           8214.3             0.999997
  J14        -3242.8           8213.0             0.999999
  J15        -3254.7           8206.3             0.999998
  J16        -3246.3           8225.8             0.999998
 	    -------           ------
  Mean:      -3251.7           8206.0

The theoretical calibration factor is (2^14-1)/5 = 3276.6
The theoretcial zero-offset is (2^14-1)/2 = 8191.5

Apart from the negative sign in front of the measured scale factors,
the numbers are reasonably close to the expected values. Given that
the magnitudes of the scale factors and offsets are both consistently
a bit low, the test detector circuit, which is common to all of the
tests, is probably responsible for most of the discrepancies.

The negative sign in front of the scale factor, was unexpected, and
will require some investigation.

Martin