[evlatests] More Holography Troubles

[Rick Perley]

    Walter found a problem in the code which he felt would fix the 
'off-by-one' raster problem.

    Sort of. 

    1)  L-band rasters are now correctly lined up.

    2) X-band rasters are now off by one step.  They used to be off by 
two steps (or maybe more). 

    3) I messed up the C-band raster file -- we'll have to try that 
again alter.

    A new problem has been found -- with more serious implications.  
Occasionally, a
single, 10-second pointing has the wrong phase.  This is easy to detect 
when oversampling
within the main beam (which we are doing here).  Ken tells me the system 
reinitializes the
phase for each pointing position (why?!), so this newly found 'phase 
jump' might be
caused by the L302. 
    The characteristics are interesting:  All three working antennas 
'jump' at the same time,
and in all IFs.  The magnitudes are different between antennas, but the 
same for oppostie polarizations.  For example, at one instance of this, 
the phases changed by:

    13 AC   14 degrees
    13 BD    85 deg
    14 AC   32
    14 BD   102
    16 AC   49
    16 BD   116

    The 'jump' lasted only for one ten second pointing position, and the 
phase returned to
the 'correct' values at the next step.   Only the EVLA antennas show this. 
    Very few steps show the phase jump -- perhaps one in 50.  But this 
messes up
holography.