[evlatests] Switch Issues

[Rick Perley]

    One more report, and I promise to go away and be quiet for a couple 
of weeks ....

    The original purpose of the SST is to check for faulty switches.  In 
the 2-hour test, there are a total of about 2700 switch change commands 
(13 scans * 8 bands * 26 antennas).  (We could multiply this by 4 if we 
considered  each IF as separate, which we probably do). 

    For all these changes, only 2 'hard' failures occurred (no fringes 
at all):  ea16 at L-band on all IFs at 10:03, ea16 on IFs B and D at 
L-band at 10:12, and ea16 on IFs A and C at Ka band at 9:59. 

    The usual set of hard-to-summarize behavior was seen in the 
amplitude stability amongst the various bands.  I note here only the 
most egregious ones:

    1) 9CD, 25AB and 18 on all IFs at L-band.  The situation with 18 at 
L-band is particularly terrible!  As usual, these gain changes are 
accompanied by bandpass changes of a few percent. 

    2) 25CD at X-band.   In general, only X-band shows instability on 
single IFs.  At all other bands, the changes in amplitude are nearly 
always the same on both IFs of a single polarization, and the two 
polarizations are quite independent. 

    3) It is hard to make definitive statements for Ku, K, Ka, and Q 
bands, as the SNR is lower, and (more importantly), pointing variations 
are dominant. 

    4) 7AB and 18D at S-band.

    5) 7AB, 9CD, and 11CD at C-band.  11CD is particularly bad, with 
fast changes within the scans!