[evlatests] Two more PDif problems

[Rick Perley]

    1) ea08, at Q-band only, as zero-mean PDif values.  (all noise).  
PSum if fine.  Evidently, the cal is either on or off all the time.

    2) ea10, in IF 'A' only, ** at all bands **, is showing two or three 
dramatically different total power values -- differing by a factor of 
two or three between the two 'states'.   The average value for 'PDif' is 
the same for each power state -- hence this is likely not a gain issue.  
But the scatter in the PDif values is hugely different:  For the lower 
power state, the PDif scatter is about right.  For the higher power 
state, the scatter is about ten times higher -- far too much to 
attribute simply to the increased power level.