[evlatests] Band Switch Test

[Rick Perley]

     I ran a very demanding band switching test last Friday.  This 
consisted of rapidly switching between all possible (cassegrain) band 
pairs -- with 8 bands, that is 28 pairs.  For each pair, I observed 5 
cycles, each observation at each band is 1 minute long.  A single source 
was observed.

     The detailed results will take some time to fully process. The 
procedure is to first review the switched power data, then compare this 
to the actual visibility data (best done with calibrating with, and 
without the switched power applied).  This has been completed for L-band 
(results in next email) only, so far.

     Three top-level conclusions can be made here:

     1) For the low-frequency bands (L, S, C and X), the switched power 
behavior is very good, with none of the multi-level or 'shredded' values 
present.  There are individual, antenna-specific issues, which will be 
noted separately in succeeding messages.

     2) For the high-frequency bands, (Ku through Q), all sorts of awful 
behaviors are seen.  Most are consistent with the bad switches in the T303.

     3) The good news is that ea10, which has a new type of switch in 
its T303, there are *no* observed instances of 'bad behavior'.

     Details will follow, over time ...