[evlatests] More on Modcomp-free flagging

[Rick Perley]

    My test at noon stepped through the bands in the usual (going 
upwards in frequency) manner.  The transition from X to U band provided 
some interesting characteristics of the present flagging. 

    While at X-band,

       * The VLA antennas were flagged as bad (although the data are 
fine) 10 seconds before the end of the scan (as defined by the end of 
good VLA data). 
   
    Then at U-band,
      
       * Unflagged VLA data appeared one second after the last good EVLA 
data at X-band.  The data, although unflagged were not good.
       * Good (and unflagged) VLA  data showed up 14 seconds later.  
(About right, if the command to change occurred at the end of the X-band 
data). 
       * Eight of the EVLA antennas were correctly flagged.  But two -- 
19 and 21 -- were not.  These two antennas are not working at *any* 
band, yet the flagger has missed this fact -- even for U-band!

    As there are no EVLA U-band antennas, there is no 10-seconds of 
flagged, but good, VLA data at the end of the U-band scan.  However, the 
beginning of the next band's data -- K -- is within one second of the 
end of the U-band data.  The VLA data in the K-band file are flagged as 
good, but are not.  As at the other bands, the good data begins about 15 
seconds later.