[evlatests] EVLA X-Band gain stability

[Rick Perley]

    I am now beginning the data reduction of the ~400 GB of data taken 
during the 30 hour 'flux densities' run.  Data quality is excellent 
overall.  I've completed basic reduction of the X-band data, using a 
single subband centered at 8457 MHz.    Clean switched power was 
obtained for all antennas, so a full and proper calibration utilizing 
the PDif values was executed, and a fit for elevation gain dependency 
was made. 
    Antenna gain stability is outstanding on all antennas except two -- 
noted below.  The antenna gain dependency is at the expected level 
typically 0.5% (in amplitude) from high to low elevations.  The 
solutions for RCP and LCP are always similar, lending credence to an 
interpretation that the dependency is truly due to the antenna, rather 
than the electronics. 

    There are two antennas that show notable -- indeed, radical! -- 
departures from this happy picture:

    1) Antenna 1:  Shows a 24-hour sinusoidal dependence in gain, with 
+/- 2% amplitude variation.  The low gain is seen at 0h IAT (5PM in the 
afternoon), the high gain 12 hours later.   The effect is identical in 
RCP and LCP. 

    2) Antenna 22:  Shows large (+/- 10%) and apparently random changes 
in gain throughout, which are not corrected by the switched power 
monitoring.  The switched power looked normal, so these changes are not 
due to the electronics.  Curiously, the RCP and LCP gains are 
approximately anticorrelated -- when one is up, the other is down ...  
No obvious explanation comes to mind...

    Antenna 19, RCP only, is the only other discrepant antenna -- and 
the effect here is quite small, or order 1% gain variations.  This 
antenna/polarization also showed major changes in bandpass shape 
throughout the run, so there is likely a correlation. 

    All other antennas are at least as stable as the VLA antennas were 
'back in their prime'.