[evlatests] PDif problems at L-band

[Rick Perley]

    Finally ...

    1) Bad PDifs:

       -- ea28 in LCP (IFs C and D) gives rubbish values of switched 
power, most (but not all) of the time.  Some scans have stable, 
reasonable values.  (This antenna has *never* given us stable PDif in 
that polarization.  It's got to be a receiver issue). 
    
    2) PDifs not reflecting real gain changes.  A longish list here ...
      --  ea01, all IFs:  Three scans had false values of PDif, (not 
seen in PSum) which caused incorrect gain adjustment (about 15%).  No 
other clues.  Unlikely to be RFI. 
     --  ea13 in RCP.  The antenna gains are changing -- each scan 
differently -- by up to a factor of 3!  Application of PDif removes most 
-- but not all -- of these changes.  I deduce that the failure to 
completely remove these changes is due to nonlinearities.  Perhaps the 
new (lower) input power settings will fix this?  This still doesn't 
explain the huge changes in PSum (factors of 3 or more).  Whatever 
happened to 'set and remember'?
     --  ea18 in LCP.  Similar to ea13, but not nearly as severe. 
     --  ea23 in LCP -- ditto above. 
    
    3) Tsys issue:
     -- ea07 has an 80K system temperature -- and I think it's been that 
way for a long time.