[alma-config] BOUNCE alma-config at majordomo.cv.nrao.edu: Non-member submission from [Barry Clark <bclark at aoc.nrao.edu>]

[Min Yun]

------------- Begin Forwarded Message -------------

Received: from cv3.cv.nrao.edu (cv3.cv.nrao.edu [192.33.115.2]) by 
zia.aoc.NRAO.EDU (8.9.3/8.9.3) with ESMTP id KAA27628 for <myun at aoc.nrao.edu>; 
Thu, 10 Feb 2000 10:03:50 -0700 (MST)
Received: from kochab.cv.nrao.edu (kochab.cv.nrao.edu [192.33.115.108]) by 
cv3.cv.nrao.edu (8.9.3/8.9.3/CV-SOL-3.0) with ESMTP id MAA17395 for 
<myun at nrao.edu>; Thu, 10 Feb 2000 12:03:48 -0500 (EST)
Received: (from majordom at localhost) by kochab.cv.nrao.edu 
(8.9.3/8.9.3/CV-SOL-3.0) id MAA23217; Thu, 10 Feb 2000 12:03:47 -0500 (EST)
From: owner-alma-config at kochab.cv.nrao.edu
Date: Thu, 10 Feb 2000 12:03:47 -0500 (EST)
Message-Id: <200002101703.MAA23217 at kochab.cv.nrao.edu>
To: owner-alma-config at kochab.cv.nrao.edu
Subject: BOUNCE alma-config at majordomo.cv.nrao.edu:    Non-member submission from 
[Barry Clark <bclark at aoc.nrao.edu>]   
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Length: 3989

>From myun at nrao.edu  Thu Feb 10 12:03:43 2000
Received: from cv3.cv.nrao.edu (cv3.cv.nrao.edu [192.33.115.2])
	by kochab.cv.nrao.edu (8.9.3/8.9.3/CV-SOL-3.0) with ESMTP id MAA23212
	for <alma-config at kochab.cv.nrao.edu>; Thu, 10 Feb 2000 12:03:43 -0500 
(EST)
Received: from zia.aoc.NRAO.EDU (zia.aoc.nrao.edu [146.88.1.4])
	by cv3.cv.nrao.edu (8.9.3/8.9.3/CV-SOL-3.0) with ESMTP id MAA17389
	for <alma-config at cv3.cv.nrao.edu>; Thu, 10 Feb 2000 12:03:41 -0500 (EST)
Received: from bclark.aoc.nrao.edu (bclark [146.88.1.119])
	by zia.aoc.NRAO.EDU (8.9.3/8.9.3) with ESMTP id KAA27479;
	Thu, 10 Feb 2000 10:03:39 -0700 (MST)
Received: (from bclark at localhost) by bclark.aoc.nrao.edu (8.7.3/8.6.10) id 
KAA09481; Thu, 10 Feb 2000 10:03:38 -0700 (MST)
Date: Thu, 10 Feb 2000 10:03:38 -0700 (MST)
From: Barry Clark <bclark at aoc.nrao.edu>
Message-Id: <200002101703.KAA09481 at bclark.aoc.nrao.edu>
To: alma-config at nrao.edu
Subject: Re: [alma-config] History
Cc: bclark at zia.aoc.NRAO.EDU
X-Sun-Charset: US-ASCII

It's always fun to comment on history.  And the ALMA configuration discussion
does often give me a bit of a feeling of deja vue.

In particular, we were quite uncertain whether Clean or other non-linear
deconvolutions would be appropriate for the VLA.  Only after the instrument
was built were we sure that large improvements would result from their
application.

Yes, the VLA was designed mainly on the basis of examining beam shapes.
(Making a beam was a fairly costly process those days, so it was mainly
designing on the basis of looking at u,v planes, and making beams of the
most promising ones.)  We were aware of the first rumblings of Clean,
and of one or two other techniques that have since vanished.  I remember
specifically at the Decade Review panel of 1970ish, at a sub-panel meeting,
being asked if these techniques would not let us use a smaller number of
antennas, and giving a rather Rylesque answer, that incompleteness in the
u,v plane really did represent missing information, and there was no 
guarantee that "invisible distributions" were always non-physical, and
even if they were, the closer to the "truth" we started, the easier job
we would have.

Sensitivity was not really a consideration in the VLA design.  We looked
at how much sensitivity was lost in going from "natural" weighting to 
"uniform tapered", but it did not seem to much depend on details of the
configuration (that is, within Y configurations, for example), so we ignored it.

Cost was a real consideration.  We took a quick look at ring configurations,
and noticed that, although their u,v planes were a little nicer looking
than than Ts or Ys, their beams did have the big sidelobes due to the sharp
cutoff in the u,v plane unless, you tapered by a lot.  And they, being much
less centally condensed, did not encourage the common VLA trick of tapering
your map to match the object to maximize sensitivity.  We regarded them, in
any case, not enough better to justify the additional pi/2 factor in track 
cost.

With the VLA, the u,v coverage is so centrally condensed that apodization is
no problem; there is no sensitivity out there anyway.  Clean and whatever
are used to bridge the annoying inner holes in the coverage.

The only time I heard the term "Cambridge Clean" was in a much more vicious
sense - the accused practice of observers who skipped configurations, so that
the ring lobes fell within the element beam, of applying judicious amounts
of white-out to contour plots being prepared for presentation.

I think the idea of designing to a criterion on the beam, or even to a 
criterion on the size and location of holes in the u,v plane is still
probably the right way of doing the design; then you start the simulations
with Clean or MEM or whatever to see what performance you expect from that
design.  Since there is no guarantee these are the best you can do with
the given instrument, they remain only a semi-quantitative demonstration of 
what deconvolution can do for you.

------------- End Forwarded Message -------------