[evlatests] Referenced Pointing Tests

[Ken Sowinski]

some comments about the referenced pointing test.

The executor routinely used still does not do the
azimuth correction properly, omitting dividing the
offset by cos(El).  This may explain some of the
differences between 14 and 16 and why secondary
ref pointing improved things.

14B failed because levels seem to have changed for
K band and the attenuators are still set for what
used to be.  IF B is the worst; TP is so low that
data is being flagged.  Has equipment been swapped
or has something broken?

We should try to find out why 13 and 18 appeared not
to work.