[daip] Question and Idea re: BLCAL

[Samuel Conner]

To Whom it may concern,

  I am trying to develop a highly accurate Clean component model of the diffuse emission in 3C84 in an archival B-Array observation with excellent UV coverage (it was a VLBI full aperature synthesis with almost all antennas) which I hope to be able to use in other observations for the purpose of a BLCALing observations which have only a short scan on 3C84 (after correction for change in the point source flux).

  Given that the diffuse emission is faint and diffuse, it is hard to map accurately without prior BLCAL on 3C84, since the 1% closure errors at the longer spacings introduce a lot of high frequency noise. I have tried using UVTAPER to suppress this, but then the beam swallows the close-in emission and I don't get a particularly good diffuse emission model either.

  It occurred to me that if I could BLCAL only the longer spacings, which are pretty insensitive to the diffuse emission, that would partly solve my problem by really cleaning up the noise in the diffuse part of the map. I could then map and self-cal all spacings and do further iterations of BLCAL to correct the short spacing correlator gains. I have experimented with this and it "seems" to work well.

  BLCAL does not admit a UVRA choice, so I have been cutting the data into long and short pieces and BLCALing the long, then merging the two. I am trying to hack BLCAL to introduce UVRA and to pass (1,0) through as "solutions" on the correlators excluded by the UV range, so that I can keep all my data. The coding was simple but I haven't tested it yet. This might be a useful enhancement to BLCAL in the Classic AIPS distribution.

The question:  I don't see anything obviously wrong with this proposed procedure, and it seems to work, but I am reluctant to rely solely on my own judgement in a doubtful and novel situation. Do you see a problem here?

  Thanks for your time.

Respectfully yours,

Sam Conner