[evlatests] Bandswitch troubles (?)

[Rick Perley]

       A 7.5 hour observation of 3C273 (a very strong calibrator-like 
object) was made at four bands (L, S, C, and X) last evening.  
Approximately 245 band change commands were made throughout. 
    Review of the C-band observations reveals:

    Of the 60 observations * 27 antennas * 2 IF  pairs = 3240 'events', 
all but 21 provided good amplitudes without any sign of a switch failure 
(where 'failure' means a total, or very large, change in attenuation).  
That's a success rate of 99.4%. 

    The 21 'failures' come in two types:

    A) Total failure.  Zero fringe amplitudes.  This happened on 11 
events, (where I count 2 failures if both AC and BD streams fail to 
fringe).  The antennas involved were:  ea11, ea25 , ea19 (twice) , ea08 
-- for all of these both IF pairs failed at the same time, and on ea02, 
on the 'AC' side only. 

    B) Partial failure -- fringe amplitudes low by at least a factor of 
100, but still recognizable.  This occurred 10 times -- always on both 
IFpairs together.  The antennas involved were:  ea05, ea19 (three 
times!), and ea17. 

    One other characteristic is noted:  On occasion, an antenna is slow 
to fringe -- in all cases there was no antenna motion involved, only a 
change of band, and the fringe amplitude recovery clearly indicates that 
a slow/sticky subreflector is at fault.  The antennas involved are:  
ea17 (four times), and ea19 (once).  I suggest that servo check out 
antenna 17, which is certainly a 'repeat offender'.