[evlatests] A couple of PSum, PDif issues

[Rick Perley]

    In reviewing a very lovely 2-hour C-band dataset, I find a small 
number of undesirable oddities in the PDif values:

    1) ea02 has a rapidly oscillating value in PDif in LCP only -- in 
both IFs (C and D).  The period is about 20 seconds, and the amplitude 
is a few percent.  The oscillation is not in PSum. 
    2) ea19 has the same effect as ea02, but in RCP (both A and B).  
Similar period, with  ~6% pk-pk amplitude.  This is not in PSum. 
    3) ea23 has a highly variable (10%) PSum and PDif in all IFs.  The 
character of the variations is different in LCP from RCP, but is the 
same for all subbands within a single polarization.  The is very likely 
a real gain effect, as the Tsys is unaffected. 

    All other antennas gave beautiful switched power data.