[evlatests] Phase transfer + 'deep' images, new data

[Vivek Dhawan]

This data was 2.5 hours, 2100-2340 C band, 2 stokes, 4subb of 128MHz.
5 sources all tracked at their phase center. (See end for exact file).

1. Frequency offsets in the phase model:
Ed already reported on the image stretch - much improved.
Phase (slow) drifts are gone, phase connection between sources is pretty
good, but there are times when the disconnect is upto 60 degrees over 1
minute over 10deg of sky. Could just be weather (Monday afternoon), but
it seems a bit much for C band in D array. A good test would be to use
the VLA and WIDAR in parallel on the same sources and compare similar
baselines.

2. Image dynamic range: The data are very clean, no obvious problems.
Its easy to get 10,000 to 1, but still having trouble getting down to
thermal noise when averaging over channels. (Expected noise for time on
source= 3600sec; 12 antennas; robust=-1)

RR, by subb, 120MHz   0.20   0.17   0.16   0.16   rms, mJy/beam.

LL, by subb, 120MHz   0.19   0.20   0.19   0.17   rms Expect ~0.1mJy

I, 120MHz   0.15  Expect 0.065  (central 32 MHz of 4 subb co-added)

I, 480MHz   0.12  Expect 0.03

V, 480MHz   0.04  Expect 0.03

I image shows a couple background sources at 1 & 2 mJy, plus low-level
rumble. The V image is pretty good off-center, but there are artifacts
close to center at the +_0.2mJy level.

----------------------

Data is here:
From: Michael Rupen <mrupen at aoc.nrao.edu>

    I downloaded and converted yesterday's 5-hour C-band phase transfer
test.  See /home/casa-dev-14/mrupen/WIDAR0,
   C1310plus19_001.55123.876379050926
There's .ms, split.ms (post-split in CASA), and .uvfits.
This was a series of 1-minute scans on 5 calibrators at the phase center.
Should be pretty obvious, I hope...

       Michael