[evlatests] WIDAR phase continuity and closure

[Rick Perley]

    Barry made some fixes yesterday which have resulted in a complete 
removal of the phase jumps which accompanied scan boundaries.  We should 
be 'good to go' for more demanding phase continuity tests -- changing 
source, changing band, and changing frequency. 

    We also have identified a good 'naked' (i.e., perfect point source) 
for closure testing -- 0217+738.  A thirty minute observation made this 
morning on this object at C-band shows amplitude closure at better than 
0.1%, and phase closure at better than 50 millidegrees.    I think this 
is about as good as we can expect, given the strength of the source, and 
the presence of background structure.  (I used ~100 MHz BW in each IF, 
and 1 minute averages for the determination of the amplitude and phase, 
following standard self-calibration). 

    But not all things are perfect.  Some noted issues:

    1) A very few (perhaps 5 at most) visibilities are identically zero 
in both amplitude and phase.   Each sub-band has a different pattern of 
these zeros -- different baselines showed zeros at different times.  
However, when one sub-band showed this, all the others had similar zeros 
within a few seconds. 

    2) All antennas shows low amplitudes in the first 22 seconds of the 
run -- but the phases were fine.  Antenna motion, perhaps? 

    3) Antenna '1' (actually 18) showed variable 'gains', of order 1%, 
on a 10-minute timescale.  A faster timescale was also noted.  All 
others were stable to much better than 0.1%! 

    4) Antennas 2 and 4 suffered a small phase jump -- 4 degrees, at 
13:38:47.  (Antenna 1 was the reference here).  This was about 1 minute 
into the observation.  No other discontinuities in phase were seen. 

    5) Antenna 4 had a large drop in amplitude for exactly one minute:  
13:40:52 to 13:41:52.  Note that this is not coincident with a scan 
boundary -- for some reason, these occur 'on the 7s' (i.e., at 13:40:57 
-- offset by 7 seconds from a 'ten-second tick' in the scan listings). 

    Overall -- simply superb data!

    I'm analyzing a similar dataset at X-band now.