VLA Test/Observation Coordination Meeting
                           B.G. Clark
                        February 24, 2005


1.  EVLA tests

B. Clark noted that a problem has developed with the Digitizer and
Data Transmission System module, necessitating a redesign.  Prototypes 
of the new design are expected to be available for installation on an
antenna in April or May, and quantity production in July.  There 
are four modules of the old design in existence, one of which does
not work at all, and one of which works so poorly that it suffers a
significant loss of SNR.  Electronics does not want to invest additional
effort in fixing the old design modules, so the EVLA is effectively 
limited to two IFs (total, not per antenna) until the new design 
prototypes are available.  The two IFs have now been installed on 
antenna 14, and antenna 13 is no longer useful for testing.

B. Clark reported on a long track at L band, taken with one IF on
antenna 13, on January 27.  The antenna was suffering the well known
phase glitches.  To (mostly) suppress them, the antenna 13 phases
were passed through a three-point median filter.  Examined after that,
the phase behavior of 13 looked quite comparable to other antennas.
However, the amplitudes showed a great deal of 10s to 10s jitter,
not exhibited on other antennas, about 5% typical scatter between
adjacent points.  There was also a slower variation; after running
through a 50s boxcar filter, there was a cyclic variation with a six
or seven minute periodicity and three or four percent in amplitude.  
The cause of neither effect is known, though there is some possibility 
that the fast variation could come from delay cogging, arising because 
the delay is not very well set (it was not checked before the observation 
was started).

2.  L band polarization

G. Moellenbrock has been looking at the polarization characteristics 
across L band.  To look at things in the simplest possible way, he 
has been looking at RL and LR fringes on the unpolarized source 
3C 84.  This is a direct measure of the feed orthogonalities.  Most
baselines show cross power of a few percent in midband, rising to 
something of order 10% below 1300 MHz or above 1600 MHz.  The only
systematic effect easily recognized is the standing wave between 
the feed and the subreflector, with about a 25 MHz periodicity.  
Although this effect is small compared to the general rise at the
ends of the band, it does probably set the fineness needed in 
frequency dependent polarization calibration software.

Fringe amplitude and phase seemed to repeat well from day-to-day,
but there were more appreciable differences with data taken a couple
of months later.  The reason is unknown.  Two antennas have the
hybrid polarizer which it is planned to use with the EVLA system.
Although these antennas have different characteristics from the rest
of the array, they are comparable in degree, and seem to be about as
stable, so the usual sort of polarization calibration should work about
as well.