[evlatests] EVLA Characteristics from last evening

[Rick Perley]

    From last night's long blank-field integration at L-band:

    All 10 EVLA antennas up and running.  However:

    1) 21C gave all zero visibilities.
    2) 18 B and D gave only noise.
    3) 19B was only barely alive -- very poor sensitivity.
    4) 19D was only slightly better than 19B. 
    5) 14B and D are still doing the 40 degree bi-stable phase bit that 
I described yesterday. 
    6) 24A and C have some data drops. 

    In addition, amongst the VLA antennas:

    1) 28D gave spectral with oscillating amplitude -- a sawtooth 
pattern.  A correlator problem which I understand is being corrected.
    2) Antenna 4, all IFs, shows brief data drops.  This problem has 
been going on for a couple weeks, at least. 

    Otherwise, data quality was very good. 

    Sensitivity:

    I calculated the antenna-based sensitivity, via ANBPL.  In general, 
the EVLA antennas are all significantly poorer than the VLA antennas, as 
has been reported many times.  These observations were taken at very 
high elevation (above 70 degrees) where the discrepancy between EVLA and 
VLA is maximized.  
    Antenna 24 is the only one whose sensitivity is comparable to the 
average VLA antenna.  All others are from 15 to 30% lower in G/Tsys than 
the median VLA. 
    The worst EVLA antennas (in sensitivity at L-band) are:  21, 17, 13, 
and 14.    In general, it seems that the problem lies with Tsys.  For 
all four antennas, the reported (back-end) Tsys values are 40 to 80 K.  
We can correct for errors in Tcal by noting the change in Tsys when 
going onto the calibrator -- presuming the antenna efficiencies are 
about 50%, we should get about 1.5K extra.  The values observed are 
close, although in some cases, a downward correction is suggested.