[evlatests] Pointing Issues -- all resolved?

[Rick Perley]

    I reported this morning on a very peculiar, and inexplicable, test 
whereby application of referenced pointing apparently caused a steady 
rise in amplitude, despite the absence of any solutions to apply!  The 
effect was especially large at X-band, with an apparent rise in source 
power by 15% on all antennas over 90 seconds.  There are many ways to 
lose amplitude in an interferometer, but very few to do the inverse.

    I repeated last night's test on the very strong calibrator 3C273.  
The file utilized was identical to the one last night, except that the 
dwell times were increased from 1 to 6 minutes (to see if there was any 
end to the amplitude rise), and to add an additional 'dummy' scan at the 
beginning. 

    The remarkable result is that these new data behave exactly as 
expected, and as desired:

    The initial X-band referenced pointing worked, and the application 
of the gains resulted in a steady, and improved amplitude.
    The referenced pointing at K and Q bands also worked (with some 
exceptions at Q-band), and the application of the results also provided 
higher amplitudes which did not vary. 
    Disabling the 'subreflector rotation trick' significantly degraded 
the K and Q band results (both in amplitude and in steadiness), and had 
little effect at X-band. 

    On the basis of these results, all is well with referenced pointing, 
and with the application of subreflector rotation to offset 2nd-order 
pointing loss. 

    The overnight results will have to be placed in the 'unresolved 
mysteries' bin.