[daip] PEELR suggestion

[Eric Greisen]

Hanno Spreeuw wrote:
> 
> We are running PEELR on a P-band VLA-A dataset of the Galactic Center.
> We ran default UVPLT on the output (u, v) dataset of PEELR.
> The visibility amplitudes reach zetaJansky levels!
> My guess is that the model we are feeding PEELR is not good enough, with 
>    too few unresolved sources to calibrate the long baselines.
> CALIB then finds gain amplitudes for the outer antennas very close to 
> zero. And CLINV will invert that, making very large gain amplitudes.
> It would be nice if we could run CALIB in PEELR phase only.
> That is something that might work in this case.

I have asked Fraser about this suggestion and he feels that allowing 
solmode 'P' as well as default 'A&P' should be okay.  He cautions that 
it is essential to examine CALIB's  solutions closely (APARM(6)=4 
perhaps and SNPLT) to guard against simply bad solutions - which can 
come from facets without adequate flux for self-cal in that facet alone.

The MNJ tomorrow will pick up the change.

Eric Greisen