[evlatests] X-band Switched Power

[Rick Perley]

    I've nearly completed the full X-band calibration from 'flux 
densities'.  A detailed report will come later today.  But here I'd like 
to point out a few issues concerning application of 'switched power' at 
X-band. 

    In general, the switched power system is a spectacular success.  
Both the diurnal gain variations, and attenuations due to the poor 
switches are nearly completely removed via application of 'PDif'.  These 
strong statements will be supported later via graphs and plots. 

    But there are two spectacular exceptions, deserving special notice. 

    1) ea01.  The switched power plot vs time shows the expected diurnal 
variation -- but shows it is completely inverted!  When all other 
antennas go up, ea01 goes down!   And, sure enough, when we apply the 
PDif to the visibility data, the resulting CALIB solutions show a 6% 
amplitude diurnal residual (12% in power).  So -- something  is 'upside 
down' here. 

    2) ea22.  This antenna's PDif shows the expected diurnal variation, 
but on top of this are large variations from scan to scan, of the type 
normally associated with 'switch problems'.  For other antennas (notably 
ea23), these PDif changes are matched by the data, so the application of 
PDif restores the amplitudes.  But this is not the case for ea22 -- the 
resulting visibilities after application of PDif lie scattered over a 
range of 20% in amplitude, with no trends in time or elevation. 
I suspect the problem lies in the visibilitiy data, as the range and 
behavior of PDif lies well within the norms, while the variation in the 
visibilities lie far outside the norms.