[daip] Problem in implementation of MINPATCH in AIPS task IMAGR

[Subhashis Roy]

Hi,

I notice a problem in the way MINPATCH is implemented in the IMAGR
(31DEC04), which becomes important while using Mutil-resolution
Clean(MRS). What I find with Imagr after a few iterations is that it
discards the value of Minpatch specified in the Adverb, and chooses its
own value depending on the size of the Clean Box. I believe this is done
to reduce computation in Minor cycle (Minpatch bigger than the Box size
is not required). However, once the Minpatch gets small, it does not
increase it while Clean proceeds to a different Facet (size of the Box
could differ). 

This causes major problem in Multi Resolution Clean. While iterating
between high and low resolution image, once it finds a small Box in
higher resloution it starts using a small Minpatch (e.g., 21 cell).
However, when it goes to Clean the lower resolution image of the same
field (where the chosen Clean Box size was much bigger), it uses the
same Minpatch. Unfortunately in many cases, the main lobe of the low
resolution Beam extends to a larger area. Using a Minpatch smaller than
the main lobe in the Minor cycle results in diffraction pattern of a
source to remain on the map plane, which are again picked up as source
in that Minor cycle creating deep negatives in that larger Box in low
resolution.

The present solution what I use is to use 'Tell Imagr Minpa=A_large_no',
ones such things happen. However, this requires me to personaly monitor
the MRS all the time, which takes many hours to complete. Also this is a
problem for DOTV=-1.

Therefore, could you please look into it ?

Thanks.
Subhashis