[evlatests] Sensitivity

[Rick Perley]

    A quick check of sensitivity was done at L and C bands.  For this, I 
used
channel 8 only (790 kHz resolution), and the calibration machinery to 
correct
errors in listed Tcals.  Finally, Eric's new program ANBPL was used to solve
for the antenna-based weights from the calibrated baseline-based values. 

    In this form the weights are proportional to (antenna 
efficiency/Tsys)^2. 

    1) L-Band.

    The 'weight histogram' for VLA antennas is fairly tight -- from 4.9 to
7.6.  EVLA antennas 14 and 16 are at the tail end of the distribution -- a
weight of about 4.5.  Antenna 13 trails by a long ways -- a weight less
than 3.0.  We thought we had corrected for pointing errors, but I suspect
there is more wrong with antenna 13 than pointing. 
    The poor sensitivity of the EVLA antennas, as judged by this parameter,
is seen in the scatter of a UVPLOT.  The apparent sensitivity of 14 and 16,
as judged by weights, is about where we would expect them to be based
on the observed efficiency, with their Tsys being the same as the median
VLA antennas.  We know the efficiency is lower (at 1465 Mhz), but had
expected this to be offset at least partially by an improvement in Tsys. 
    Or -- extra noise is being introduced from elsewhere. 

    2) C-Band. 

    The weight histogram for VLA antennas is remarkably broad --
roughly from 2.5 to 9.5.  The distribution within these bounds is also
remarkably flat -- nearly uniform. 
    Antenna 14 is at the top end in both IFs A and C, with a weight of
9.5.  Antenna 16 is at the ~33% percentile -- weight of about 7.5 
Antenna 13 is by far the worst, with a weight of ~1.5.  If the antenna
were on the half-power point, this value would rise to ~6.0 -- still
far below what it should be. 

    Although 14 and 16 are very good at C-band, they are not good enough. 
They should be at least 25% to prehaps 50% better (in this weight parameter)
than any other VLA antenna. 
    The same worries expressed above for L-band apply here.