[evlatests] Early comments on June03 WIDAR-0 data

[Rick Perley]

    An hour of data was taken on the point-source 0217+734. 

    Things which aren't right:

    1) Significant delays remains (except for antenna 1, which Ken fixed 
yesterday).   Today's values are (in nsec):

    1   0
    2   ref
    3   8.6
    9   -2.1
    18   12.3
    19   8.8
    23   2.0
    24   5.8
    25   9.5
    28   10.0 

    2) As before, five antennas have regular 180 phase jumps on all 
IFs.  The 'jumpers' are:  1, 3, 9, 23, and 25.  Yesterday, the period 
between jumps was 11 minutes and a bit.  Today it is 14 minutes.  The 
jump times are:

    21 29 57
    21 44 37
    21 58 37
    22 12 37

    ... hence, the interval is now 14 minutes plus some seconds.  Note 
that whereas yesterday the jumps all occured on a time ending in '8', 
today it's a '7'.  None of these jumps times aligns with a scan 
boundary.  They all occur within the 40 second scan durations. 

    3) Antenna 1 has unstable amplitude gain -- many different levels 
were seen.  About half-way through, it stabilized, and was steady 
afterwards.  All other antennas were fabulously stable, except antenna 
24, which (after a short outage due to wind) returned to a level which 
was about 10% lower fringe amplitude than before. 

    4) Amplitude gains (the value needed to correct the observed 
visibilities to the known flux density) vary rather widely -- far more 
than they should, since we think these antennas all have about the same 
sensitivity.  Required gains vary by a factor greater than 2.0 between 
the antennas (antenna 19 is the lowest, with 8.1, antenna 2 the highest 
with 20.4).  If the numbers I'm getting are proportional to correlator 
coefficient, then we are seeing a range of over a factor 5.5 in antenna 
SEFD -- which I don't believe. 
   
    5) I did a quick calibration, an looked at the histogram widths for 
a single channel.  The rms was 0.74 Jy (1 sec integration, 125 kHz 
channelwidth) -- close to the 0.63 Jy I expect.  The 'real' part of the 
visibility was perfectly gaussian.  The 'imaginary' part had an enormous 
surplus at zero.  Perhaps a result of calibration ...More work will be 
done on this. 

    6) Bandpass solutions, maps, etc. are in progress. 

    In general, very nice data.  But lots of issues remaining...