[evlatests] 10-second phase jumps

[Rick Perley]

    A holography test was run last night which clearly indicated that 
our phase problems have at least two different origins.  One is induced 
by the holograph raster line, the other associated with single 10-second 
pointings.  I'll report separately on these two phenomena in an upcoming 
report.

    In meeting with Ken and Walter this morning, it was agreed that the 
latter phenomenon (phase changes on individual 10-second pointings) was 
likely not due to holography mode itself, and an attempt to prove this 
was made during empty dynamic time.  To do this we:

    submitted a VLA script which simply sat on a calibrator, with no 
source changes.
    submitted an EVLA script which re-initialized the source every ten 
seconds. 

    AC IFs were set to 1485 MHz.  BD IFs to 1275 MHz.  BW = 12.5 MHz. 
    This reproduced the ten-second phase jump quite well.  We got 16 
minutes of data, which were dumped at the fast 410 millisecond rate.   
Unfortunately, all IFs for antenna 13 were non-operative, and 14B and D 
were also dead. 

    Two phase 'events' occurred, both *exactly* ten seconds long. 

    Event A:  All EVLA antennas jumped together at 15:43:32.1
       14A jumped by 105 degrees
       14C by  -253
       16A by 183
       16C by 183
       16B by -112
       16D by -111
       18A by -249
       18C by 110
       18B by  248
       18D by -111

    Event B:  EVLA antennas 16 and 18 jumped together at 15:51:42.  *** 
Note that antenna 14 did not!***
        The phase  jumps in these event are *identical* to those in EVent A.

    Also note that the phase jump is actually the same for all IFs on a 
given antenna, (since +105 = -253 in phase rotation) and that antennas 
16 and 18 have the same jump.  Antenna 14's jump is of a different 
magnitude (but A and C jump by the same quantity). 

    The phase discontinuity is very sharp -- far less than 400 
mseconds.  There is no phase transition joining the two states, and the 
amplitude drop is negligible at the beginning of the 10-second jump, and 
very small (but measureable) on the other end of the 10-second period.