[evlatests] Flagging Issues, continued

[Rick Perley]

    I ran a 'fast dump' (0.417 second integration) observation at all 
bands last evening. 

    In general, the array is operating very well.  But there are some 
curious residual issues, mostly connected to flagging, which I report on 
below. 

    The program ran through the bands, using standard frequencies and 
bandwidths, from P through Q, twice, with two minutes on each band. 

    1)  Beginning of scan problems. 

    When a band change is made, the new band database shows a single 
record, listed by AIPS as a separate scan, which is in some cases 
flagged, and in other cases not.  In all cases, the data are not good.  
Following this single record scan is the new scan, which contains a 
single unflagged (and bad) record, followed by ~10 seconds of flagged 
data, followed by the good (unflagged data). 
    The sequence described above occurs *every* time upon band change.  
It does not happen when one does not change bands. 

    In the end, this is a minor problem, as a `QUACK' operation will 
always remove these initial records.   I certainly recommend that users 
do this operation as a matter of course. 

    2) End of scan problems. 

    We are still occasionally seeing a few bad, and unflagged, records 
at the ends of some scans.  In my test, this occurred at X-band (last 
five records on one scan), K-band (last 3 records on one scan), and at 
Q-band (last 3 records on one scan).  This problem was not seen on any 
of the low-frequency bands. 

    3) Peculiar antenna-IF flagging. 

    We had very curious flagging at two bands. 

    a) P-Band:  There were three observations made at this band -- an 
initial pair (to see if initial scan issues are still with us), and a 
third, 20 minutes later (between which the other bands were observed). 
       Three EVLA antennas were flagged in whole or in part, but in all 
cases, the data actually look to be fine:
          o  Antenna 23, all IFs, was flagged on the first scan. 
          o  Antenna 25, all IFs, was flagged on the third scan.
          o  Antenna 24, IF 'D' only, was flagged on all three scans.
          o  Antenna 23, IF 'B' only, was flagged on the third scan. 

    b) C-Band:  There are two scans at this band, with very curious 
flagging behavior:
          Antenna 16, IFs A and B, gave no fringes on either scan, and 
were correctly flagged.
          Antenna 16, IF C and D, gave fringes on both scans, BUT:  IF 
'D' was flagged, while IF 'C' was not. 
          Antenna 23 was flagged on all four IFS on the 2nd scan, but 
the data appear to be fine. 

    At all other bands, flagging appeared to be correct, other than the 
initial/final record issues listed above.