[evlatests] EVLA-VLA incoherency statistics

[Rick Perley]

    As has often been reported, we occasionally see the VLA and EVLA 
being mutually incoherent -- EVLA-EVLA baselines are fine, VLA-VLA 
baselines appear fine (but in fact, are not), and cross-baselines give 
no fringes at all.  This phenomenon *always* occurs for whole scans, and 
*only* occurs in IFpair 1 (AC side).     Presumably, the VLA antennas 
have been slightly mis-tuned. 

    One of my polarization tests, done early this month, involved 
numerous changes of mode and frequency settings -- alternating between 
50 MHz continuum, and 12.5 MHz mode PA, at X-band.  I find that:

    7 of 199 continuum observations showed the incoherency (only on the 
AC IFs),
    12 of 216 spectral line observations showed the incoherency. 

    About 5% of the scans are affected.  This is uncomfortably high, 
given that these data are not caught by the flagger.  The problem is 
easy to catch in observations of strong sources, but cannot be seen in 
observations of weak sources. 

    Is there a fix in the works?