[alma-config] A few comments on John's simulations

[Tamara Helfer]

Hi John and everyone,

I think your simulations are a nice start.

Just a few comments:

1) As others have remarked, a true ring array is currently the
strawperson model for the largest arrays only.  The D and C
and even B arrays are tapered arras.  Leonia has achieved as
good sidelobes in his optimized tapered arrays as in his
filled arrays (of which your spiral is a subset).  An
outstanding issue is whether it's better to optimize the arrays for
overall rms sidelobe level or for smallest "worst" sidelobe.
Another issue is if having an array with low sidelobes inside
the primary FWHM but having a high sidelobe lurking outside the
"optimized" region is a problem.  

2) It's important to match the array to the source appropriately
for simulations.  It's really unlikely that sources like CygA
or M51 will be observable with the largest arrays.  Certainly
these won't be common observations.

3) Are your simulations mosiaced?  If you proceed with a
"joint" imaging/deconvolution method, the mosaiced dirty beam
will have significantly lower sidelobes than a single-pointing
dirty beam.

4) Whether you test donut or ringlike arrays, I think it would be
good to use arrays that have been run through Leonia's optimizer.
I can't remember the numbers exactly, but I think he was getting
sidelobes that were a factor of 2-3 lower than sidelobe levels
I was getting for "randomly" generated arrays.

Best wishes,

	Tamara