[evlatests] PDif at Ka-band

[Rick Perley]

    For this band, like at K-band, the application of PDif almost 
invariably gives much improved gain performance.    The only exceptions 
to this are where it is clear that the pointing is incorrect. 

    The only exception is:

    ea01:  On the RCP side, the PSum and PDif values are large (factor 
of 10 or more higher than normal), random numbers (both positive and 
negative) for more than 75% of the time.  There are some good scans with 
reasonable values.  The situation is the same for IFs A and B.   The LCP 
side is fine. 
    Clearly something in the receiver, eh?