[evlatests] K-band PDif effectiveness

[Rick Perley]

    All is very good at this band. 
   
    1)  All antennas are giving good, solid, reliable measures of PDif 
and PSum.
    2)  The application of PDif always corrects system gain changes 
correctly -- with nearly all exceptions clearly due to pointing errors.  
(More on this separately). 
    3) ea10 is a possible exception to the rosy picture.  Corrected 
visibilities are showing variations at the ~3% level, which are not due 
to PDif values.  These variations are proportional to the frequency, and 
are the same between opposite polarizations -- a strong indicator of 
pointing errors.  Yet the referenced pointing solutions appear to be good. 

    4) Tant on 3C84 is expected to be ~2.7K.  This is seen, as expected, 
on all IFs.  There is some variation in Tsys and Tant amongst antennas 
and IFs which are very likely due to errors in Tcal -- these can be 
calibrated against the sky across the bands -- a project for a different 
day.