[evlatests] P-Band Tsys measurements

[Rick Perley]

    I reported yesterday that some antennas had very erratic 
visibilities, which were likely due, at least in part, to wild 
variations in the measured system temperature. 

    AIPS has a nice feature now which allows the data to be loaded as 
raw correlation coefficients -- exactly as the correlator produces them 
-- without the application of gain changes from the Tsys measurements.  
Doing this, and comparison to the normal process allows a determination 
of how much of the visibility variations are due to Tsys errors, and how 
much due to other sources of variability. 

    I have done this for yesterday's P-band test data.  The results show 
that for VLA antenna 8, EVLA antenna 21, and EVLA antenna 23, IF 'A', 
*all* the instabilities in visibility are due to the Tsys calculation.  
The raw correlations for all these are a picture of perfect 
tranquility.   Something is seriously wrong with the synchronous power 
meters for those antennas.