[evlatests] Failures to Fringe, upon band changes

[Rick Perley]

    Joe ran the 'band change stress test', where we cycle amongst the 
eight bands, with about 45 seconds per band, for two hours, looking for 
switch failures. 

    These come in two varieties:  Those which cause no fringes (total 
failure -- the worst kind, especially since these are not caught by the 
flagging system), and those which cause a change in amplitude (possibly 
benign, since the switched cal should detect and adjust). 

    Here I report only on the 'first kind' -- total failures.  The 
second kind requires more effort, which will be done later. 

    There were 13 observations of the source at each of the eight bands 
-- 104 band changes.   27 antennas fringed (most of the time), so we had 
a total of 27*104 = 2808 chances for a failure. 
    A quick review shows that there were only 3.5 clear failures, plus 
one event which must have an origin in the T302, and two others which 
are likely receiver-related.  All of the issues listed below were not 
flagged by the system:

    1)  At L-band, antenna 21 never provided any fringes at any IF.  I 
presume this is because there is not a receiver.
    2) At S-band, antenna 9 gave no fringes, although it is listed as 
having a receiver installed since April 2009. 
    3) At X-band, antenna 19 failed to provide fringes for a single 
scan, on all four IFs (at 11:08:00)
    4) At Ku-band, antenna 10 failed to provide fringes for a single 
scan, on IFs A and C only.  (at 11:02:30)
    5) At K, Ka, and Q bands, antenna 13 gave zero amplitudes in RCP 
only (IFs A and B), for all scans *except the first*.  This must surely 
be caused by the T302 (UX Converter). 
    6) At Ka band, antenna 10 failed to provide fringes for a single 
scan, on all four IFs (at 11:00:50). 


    All antennas claimed to have a receiver at Ku-band fringed (other 
than the single failure listed above).  At S-band, all antennas listed 
as being outfitted with a receiver fringed, except antenna 9, as noted 
above.