[evlatests] Friday's instr pol test - initial impressions

[George Moellenbrock]

I've had a quick look at the data from Friday's ~2-hour L-band instr pol 
run, and the data look pretty good.  The observation covered 1300-1550 MHz 
in 20 12.5 MHz bands each with 15 channels and full polarization.  Walter 
helped me get the first LO set more optimally than the conventional 
obs2script default (as of last Friday).

1. EVLA antennas 14, 16, & 13 all participated and all appear to have
delivered decent data.

2. There are no specific tunings which appear to have failed on the EVLA 
antennas as had happened on Rick's earlier polarization run.  There are a 
few selections where RFI seems to be a problem, as expected, but there are 
no outright drop-outs of the visibility signal.

3. Given rapid frequency changes (1/min), it is a bit hard to say much at 
this stage about phase-stability of the EVLA antennas with time, although 
it is pretty clear that it is at least slipping very rapidly with freq 
change (~0.5 cycle/change or worse), and may be changing arbitrarily with 
frequency changes.  VLA antennas appear to return to the same phase for 
each frequency setting (a 20-min timescale); EVLA antennas do not.  More 
on this later.

4. It is immediately clear the instrumental polarization on EVLA baselines 
is substantially better, in both level and variability with frequency than 
on VLA-only or hybrid baselines.  Much more on this later.

5. I can't say much about relative sensitivity just yet; stay tuned.

-George