[evlatests] Erratic switched power values at K-band

[Rick Perley]

     K-band data were taken over the weekend, as part of the 'primary 
beam' measurement program.  Both observations had no change in 
frequency, *except* that referenced pointing was done on occasion. These 
pointing observations were done at Ku-band.

     Examination of the visibilities, and subsequently the switched 
power tables, showed that a number of antennas exhibited very erratic 
gains following the Ku-band pointing observations. 'Erratic' means gain 
variations on timescales of seconds, and with amplitudes exceeding 10% 
in power.  These periods of erratic behavior typically lasted ~10 
minutes before stability was restored.
     These fluctuations are common in all spectral windows in one 
polarization, but differ between the two polarizations.  Some antennas 
are much worse than others.   Some seem nicely immune.

     I'm applying these gains now, with the expectation that the 
visibility fluctuations will be corrected.  (If this doesn't happen, 
another note will follow).

     The good antennas (no fluctuations noted beyond noise) are:

     3, 6, 8, 9, 10, 11, 18, 19, 21, 24, 26, 27, 28.

     ea21 in RCP has a special problem -- hugely noisy PDif values (20% 
or so) on second timescales.  Pure noise -- no trends or changes.  But 
the effect was absent in the beginning of the run, then slowly grew to  
maximum badness partway through, with no change thereafter...

     Rick