[evlatests] Freq offsets & phase transfer / better.

[Vivek Dhawan]

The fix for the frequency offset in the correlator effective LO
was put in last Wednesday, and an X-band test dataset acquired
on Thursday (X1925 MJD 55119.94, ~1.4hr duration). This had a
loop of cal+3 sources spread over 5deg RA, ~11deg Dec. X band.

The phases on all sources show no sign of phase wind or phase
offset with sky position, to about 20deg max error, limited by
atmospheric variability. At least a factor 5-10 better. So phase
transfer looks good, with minor glitches which may just be weather
around sunset.

However, the images still show a systematic stretch in the radial
direction, and indicate a frequency offset < 75 KHz (vs. 275-375
in Ed's email from Oct 14, based on pre-fixed data). A factor
5 better, roughly. Again, weather may limit the accuracy - the
offset is 0.35" (1/15 beam) for the target 10.8deg away in dec.

So, I guess we gather data and do more stringent tests. At some
convenient time, I would like to do a real all-sky baseline run
for a couple hours at C or X band.

Vivek.

PS:
Since some of the WIDAR antennas were moved, I should adjust the
positions of those (8, 18, 24) from the latest VLA baseline run
and image again. The maximum shift was 2cm, ~20 degrees error on
those baselines on the source farthest from calibrator. It might
make some difference because the position shift corresponds to
about 20deg of phase. I started on this, but messed up the
antenna numbers. I'll report later.