[evlatests] Three things are wrong ...

[Rick Perley]

    We have taken a few minutes of fast dump continuum (0.416 second 
integration) data to diagnose our current ills.  Three remarkable 
phenomena are seen:

    1)  Antennas 13, 21 and 25 each have dual levels on all IFs.  The 
patterns are familiar: 
                Antenna 13 alternates between the two states, whose 
amplitude ratio is 1.05.
                Antenna 21 has three 'down' states and one 'up', ratio 
is 1.20
                Antenna 26 alternates states, ratio is 1.26. 
          There is no effect on the phase. 

    2) Most VLA antennas are showing single-record amplitude drops 
*exactly on 10-second ticks*.  The antennas for which this is obvious 
are:  7, 8, 9, 12, 15, 20, 22, 27, and 28. 

    3) All VLA antennas have a square-wave phase behavior, such that the 
phase jumps by a large amount (typically 100 degrees), exactly on a 
10-second tick (the change is probably responsible for the amplitude 
dropout noted above), and is held in that state for 10, or 20, or 
occasionally a longer multiple of 10 seconds.  It then jumps back to the 
originating state.  All IFs seem to be doing this in an identical 
manner.  (I'll confirm this shortly). 

    In addition, the VLA and EVLA are incoherent in IFs A and C.  Ken 
has already diagnosed this as due to the Flukes being set to the wrong 
frequency.  Why this happened is something I'll leave to him to explain.