[evlatests] Friday data, X-band

[Rick Perley]

    The Friday, 24 March dataset is of 6-hour duration, so even though 
one EVLA antenna
was absent (13), there is much useful information. 

    The dominant effect in the monitor data was steady drop in Tsys for 
nearly all
antennas.  This is undoubtedly due to the aftereffects of the site-wide 
power failure which
occured shortly before the run began. 

    Antenna gain stability is very good for all antennas -- other than 
slow effects clearly
related to the continuously changing Tsys.  The overall stability (rms 
stability is better than
0.5 % for nearly all antennas) is in marked contrast to the peculiar 
gain changes (occuring
on timescales of minutes, with amplitude of a few to over 10 %) seen in
both the Saturday and Monday databases.   I have no ready explanation 
for why. 

    The ~5 hours' observation of 3C84 gave a very nice map -- after 
standard self-cal
and editing, a dynamic range of nearly 100,000 was obtained.  The 
resulting model
(the source is slightly resolved) allows a careful look at closure.  The 
results are:
       *  The intra-EVLA antennas (14, 16, 18 -- antenna 13 was 
unfortunately out) show
consistent closure errors between themselves of ~1.0% in amplitude 
only.  They give
too high a visibility, when calibrated against VLA antennas. 
       *  VLA baselines show a maximum closure of about 0.3%, probably 
due to thermal
noise. 

    We might speculate that the EVLA closure is due to bandpass 
mismatches, as seen in
continuum.  But these data were taken in spectral line mode, with a 
careful bandpass
calibration.  We should see no errors due to bandpass mismatches!