[evlatests] Referenced Pointing Status

[Rick Perley]

     In preparation for the upcoming 'flux density 2019' project, I ran 
a test script last night.  In general, the system is working quite 
well.  One mysterious problem (to bizarre to try explain here) is being 
investigated now.   It involves X-band however ...

     The project requires extensive use of referenced pointing. With the 
troubles I've reported in the recent past in mind, I carefully reviewed 
the referenced pointing data from last night's test.  There were three 
referenced pointing executions, all using the standard X-band setup.  
All three worked beautifully -- with a single exception.

     The one exception was for ea17.  For one of the three executions, 
there were no solutions.  In perusing the flag table, I see that the 
data were flagged for 'focus error'.  However, in examining the 
unflagged data, the *amplitudes* of the flagged data are perfectly fine 
-- the system was certainly fringing beautifully.  The 'problem' was in 
the phases -- they were off by about 80 degrees, indicating that indeed, 
the subreflector was not quite in the correct position -- I figure it's 
about 4 mm off (or more, in multiples of ~18 mm -- phases are circular ...)

     For phase-stable interferometry, this small (essentially trivial) 
error may be important.  But for the amplitude-only analysis (such as 
referenced pointing) or for strong sources ( when self-cal can rescue 
phases), such a small error is of no consequence.

     Missing a referenced pointing will usually be very important for 
high frequency observing. And, unlike regular observing, the flags can't 
be undone after the fact.

     So is it possible that the program which determines the referenced 
pointing be taught that it should pay no attention to 'focus error' flags?

     This issue may be of some importance -- 'focus error' flags are 
disturbingly common in multi-band observations such as these.