[alma-config] CoDR agenda

[John Conway]

Hi,

 Have to leave to pack my bags in few minutes, since
I fly out tomorrow morning. Just a few comments on 
todays discussion.

1) I've been doing some MEM simulations to complement
the CLEAN ones in my last memo, which I'll talk about
in Tuscon. I find for the two tests images I've looked at
so far long track that the imaging errors are about 3 times 
larger for uniform or nearly uniform coverage as compared to 
the highly tapered coverage from zoom spirals.
I therefore tend to agree with Leonia (and disagree
with Keto) that its down to the 
bad sidelobes and not just the alogirthm (admittedly MEM 
does better for both uv coverages than CLEAN at 
extrapolating beyond the uv edge but the tapered coverage 
still gives  significantly better reconstructions).


2) If the imaging simulations is a 'draw' I would argue that 
zoom arrays are preferable to fixed arrays, first  because it
gives the option of extra science capabilities (adjustable resolution, 
matched uv coverage in arrays, avoidance of tapering), and  secondly it 
requires fewer pads (and more importantly number of antenna moves to 
cover through all  resolutions). Finally if one bulids a zoom geometry and  
the continuous moving mode proves impractible then one can decide
to run it as a set of 5 (or whatever number one chooses) fixed arrays.  



3) I thought we had discussed the real estate question
a few weeks ago. Assuming there is a limited 3km area then the largest 
array should obviously have all the antennas on the perimeter to get max
resolution. The conceptual spiral in my memo 283 moved gradually from a 3km 
spiral to a 3km ring to give the last factor of 2 in resolution. As I
note in my strawman layout memo (292) an extra 20-30 pads are probably
needed to be placed around the perimenter of the 3km 5deg gradient 
area to get  max resolution for this largest array. Given the shape
of the terrain this largest array will look like some form of 
triangle.

See everyone in Tuscon!

   Cheers
      John.