[Gb-ccb] CCB tests with DC inputs.

[Martin Shepherd]

Using a breakout cable that JD made for me on Friday, over the weekend
I managed to use one of the GPIO-card's miscellaneous analog outputs
to drive known DC voltages into the differential inputs of the
CCB. This should make the tests that involve inserting known DC
voltages at the CCB inputs, much less tedious and more versatile than
setting these voltages by hand, as we had previously planned.

Using this new feature, I now have a test program that generates a
staircase ramp of known voltages at one of the CCB inputs, and at each
step in this staircase, measures the mean and standard deviation of a
frame of dump-mode samples. At the moment, I'm just printing these
numbers to the screen, in tabular form, but my next task will be to
include code to compute a weighted least squares fit of the numbers to
a straight line, in order to work out the ADC gain, zero-offset and
goodness of fit. I'll also try plotting the generated values in
gnuplot.

Note that in order to be able to take over control of the CCB from the
manager (which was left connected over the weekend), I reconfigured
the CCB server not to accept connection requests from computers in the
192.33.116.* IP address range. So I imagine that there will be lots of
error messages in the manager's log regarding failed connection
attempts. I have reenabled access to this block of addresses now.

Martin