[daip] [!4879]: aips - CALIB & DOFIT - strange behaviour of REFANT is included in DOFIT

[Michael Bietenholz]

Michael Bietenholz updated #4879
--------------------------------

CALIB & DOFIT - strange behaviour of REFANT is included in DOFIT
----------------------------------------------------------------

           Ticket ID: 4879
                 URL: https://help.nrao.edu/staff/index.php?/Tickets/Ticket/View/4879
           Full Name: Michael Bietenholz
               Email: mbieten at yorku.ca
             Creator: User
          Department: AIPS Data Processing
       Staff (Owner): Eric Greisen
                Type: Issue
              Status: Open
            Priority: Default
                 SLA: NRAO E2E
      Template Group: Default
             Created: 06 May 2014 01:28 PM
             Updated: 31 July 2014 11:48 PM
      Resolution Due: 08 August 2014 03:54 PM (7d 16h 6m)



I think what happens is that the SN-table entries for some antennas which were *not* in DOFIT wind up with SNR=0 in the SN table, and then the corresponding visibility data gets dropped when CALIB writes out the output uv-file (DOAPPLY=1).  What I don't get is why some of the SN-table solutions for antennas for which no solution was desired wind up with SNR=0

Here's a random chunk of my SN-table (sorry about the column alignment - they should have used a fixed-width font for this thing).  I used DOFIT=2,5,8,9 and REFANT=7.  Notice e.g., antenna 1, not in DOFIT, mostly has SNR=5 (its 5.000) but at 5:46:33 it goes to 0.   There is data on all the baseline to antenna 1 around 5:46.  There is not enough SNR on baselines to ant 1 to usefully self-cal so it seems likely if one *tried* to solve for antenna 1, one would get SNR < 5 - but since its not in DOFIT it shouldn't be trying to solve for ant 1.

I think I was wrong when I said that  31DEC13 *not* doing this, it also does it. It did *less* of it on my particular data set, but that could just be some fluke to do with my particular data set.

The workaround cure is to TACOP your SN table so as to get rid of any 0 weights, and then run SPLIT to apply the copied SN version.  But I think the thing that is going wrong is that the solutions for antennas not in DOFIT, which should just be amp=1,phase=0,SNR=[something_nominal] seem to sometimes wind up with SNR=0.

LISTR, OPTY=GAIN, DPARM=8,0
Stokes = R    IF =  1 Freq =  8.393125000 GHz
 
  Time   Source    --01--02--03--05--06--07--08--09--10--11
 
Day #   0
05:15:18.0 SN1996CR   5  44   5       5   5  33       5   5
05:31:33.0 SN1996CR   5  24   5       5   5  40       5   5
05:46:33.0 SN1996CR   0 316   5   5   5   5 202       0   5
06:00:26.0 SN1996CR   5  17   5   0   5   5  14       5   5
06:17:54.0 SN1996CR   5 114  23  12   5  29  72       5   5
06:32:56.0 SN1996CR   0  63   0   0   5  37  87       0   5
06:47:58.0 SN1996CR   0 134   5   7   5   5 274       5   5
07:02:49.0 SN1996CR   0  64   8  36   0  26  60       5   0
07:19:06.0 SN1996CR   5  72   0  61   5   5 163       5   0
07:34:08.0 SN1996CR   0 129  22  20   5   5 110       0   0
07:47:59.0 SN1996CR   0  52   0   5   0   5 100       0   5


------------------------------------------------------
Staff CP:  https://help.nrao.edu/staff