[alma-config] alma configuration simulations

[Ed Fomalont]

I agree with Mark's point unreservedly.  There is a good chance that
the ultimate limit of most ALMA images may not be caused by the
defects from imperfect u-v coverage, except for the zero-spacing hole.

Perhaps as important as pointing errors are the expected gain and phase
errors.  Those from receiver noise are usually not too detrimental
since they average away and their effects can be reasonably well
estimated.  Systematic phase and gain errors, after calibration from a
nearby source and other apriori calibrations, tend to be more persistent
in time and, thus, even a small error can produce significant defects
in the image, at least at the level of one in one thousand.

It is important to try to assess the relative importance among u-v
coverage, pointing errors, random phase and gain errors and systematic
phase and gain errors.  If any one tends to be dominant, then that
area should receive the greatest attention in how to minimize the
effects (receiver noise is obvious!) in hardward and software.

    Cheers, Ed