[evlatests] EVLA Polarimetry, continuum

[Rick Perley]

    I've reduced the continuum polarimetry test from last night.  The 
results are excellent.

    The source was 3C147, which provides only 5 mJy of polarized flux -- 
with Stokes ' I = 22 Jy, this is a negligibly low fractional 
polarization.   Another advantage of this source is that we can 
completely ignore any ionospheric faraday rotation effects -- as there 
is no plane of polarization to suffer rotation. 

    The gain-calibrated cross-hand correlations are -- with a single 
exception -- nearly perfectly steady.  The variations seen last fall are 
completely absent, except in antenna 11, which shows a large variability 
in both amplitude and phase.  More on this later. 
   
   When referenced to an EVLA antenna, I find that EVLA antennas are 
only slightly more highly polarized than VLA antennas.  When referenced 
to a VLA antenna, VLA antennas are considerably better than EVLA 
antennas.  The highest cross-polarizations are seen on VLA -- EVLA 
baselines.  But there is no indication that EVLA antennas are more 
variable in cross-polarization than VLA antennas.  Some slight trends in 
cross-hand response (at a fraction of 1%, or a few degrees) are seen on 
timescales of hours -- these deserve some more investigation. 

    EVLA antenna 11 is spectacularly different than all others.  It 
appears fairly clear that this antenna has two polarization states -- 
one steady, and another which is variable, and which lasts about one 
hour.  The variable component is seen in both amplitude and phase. 

    Images were made in all polarizations, following polarization 
calibration, but omitting antenna 11.  The 'Q' and 'U' images are not 
noise-limited -- showing a 'lumpy' appearance near the center of the 
image of amplitude about 2 mJy (which is only 0.01% of the maximum in 
I!)  The localized 'lumpiness' tells us that there are slowly varying 
cross-polarization residuals, but at a very low level. 

    All in all -- a very satisfying result.  We must find what is wrong 
with antenna 11's polarization.  Otherwise, it now seems clear that the 
EVLA antennas are close to the project requirements in polarization.  It 
seems quite likely that the 4-band dipoles are responsible for the 
variable cross-polarization seen in last fall's observations.  Direct 
proof of this (via an on-off experiment) is probably justified.