[evlatests] New failure mode!

[Rick Perley]

    I ran detailed sensitivity tests yesterday evening.  Detailed 
reports will
be made later today, after I've had the time to calibrate the data. 
    Here I report on a curious new (to me) failure mode.

    The observations were made in single-IF spectral line mode -- four
observations for each band -- 20 observations in all on the calibrator,
and 20 more on blank sky. 

    There were actually 23 calibrator observations -- three band/IFs
were repeated (i.e. mode 1D at X, K, and Q bands was observed
twice -- all other combinations were taken only once). 
  
    Of these 23 trials, 6 showed no fringes on the calibrator for
either 14 or 16!  Two of these were 'repeat' observations, for which
the other observation showed fringes. 

    A curious sequence was seen in the failures:

Time(IAT)   Band/IF     Good?
----------------------------------------------
1:01               X-1D            Y
1.04               K-1D            Y
1:06               Q-1D            Y
1:09               L-1A            Y
1:12               C-1A            X
 1:15              X-1A            X
1:18               K-1A            Y
1:21               Q-1A            Y
1:24               L-1B            Y
1:26               C-1B            Y
1:29               X-1B            Y
1:32               K-1B           Y
1:35               Q-1B            X
1:38               L-1C            X
1:41               C-1C            Y
1:43               X-1C            Y
1:46               K-1C            Y
1:49               Q-1C            Y
1:52               L-1D            Y
1:54               C-1D            Y
1:58               X-1D            X
2:01               K-1D            X
2:03               Q-1D            Y
----------------------------------------

    Note that failures (emphasis:  both antennas fail!) are in
pairs, separated by 6 successful observations.  I claim this
distribution is not random. 

    The blank sky observations always follow the calibrator,
but I can't tell if the system was working or not on these scans.