[evlatests] Instr. pol test Thurs night (4/6)

[George Moellenbrock]

I had another L-band instr. pol test last night.  Nominally 3 hours 
(LST=0500-0800), the first 1:40 minutes totally lost to high winds. About 
80 minutes of decent data.  I am in mode PA (IFs A & C only, 4 
correlations), 12.5 MHz bandwidth (15 chans), 20 different tunings to 
cover 250 MHz from 1300-1550 MHz.  Observed 3C84 for good narrow-bandwidth 
SNR, and 3C286 to attempt an on-the-sky R-L phase registration among 
tunings.

Looks the the EVLA ants took a few extra minutes to get going after the 
wind stopped.

ON BASELINES TO ANT 8 (W16):

1.  13, 14, 16 had parallel-hand amps (AA, CC) and SNR indistinguishable 
from VLA baselines, EXCEPT 14 was dead (flagged online) in LCP until the 
last 25 minutes, and dead in RCP (unflagged) for 40 minutes in the middle. 
Both pols came back on at the same time, @ 7:45pm local time.  This is 
after the tinkering Jim reported via email last night; maybe he did some 
more after that?   Sorry, no info for IFs B & D, which seem to be 
frustrating Rick and Barry more today on 13.

2.  13, 14, 16 all show significant ~half-cycle monotonic phase variation 
across the 12.5 MHz bands.  This looks about as it alway does: most of the 
change is in the low 25% of the band.  In the ~linear 75% of the band, the 
delay is no worse than 20nsec, about 4X worse than typical of VLA 
baselines.

3. 13 clearly has higher residual phase rate than the other EVLA antennas, 
of order 1 mHz.  This rate differs sign and magnitude for the two sources, 
as might be expected from a position error for this antenna.  14 and 16 
also have rates a bit worse than is typical for VLA ants, but not as bad 
as 13.

4. 13 didn't lock up for the setup centered at ~1419 MHz (!).  This is
just as it was a week ago in my last instr. pol run.  Something about
that frequency.  (I am using the nominal LO setups.)

-George