[alma-config]A complete design

[Mark Holdaway]

Ah!  I think I just realized a problem, though it might be only
a trivial problem, or possibly no problem at all.

People do simulations with the antenna locations specified
either in terms of earth centered coordinates or a local tangential
coordinate system.  

The MASK is specified in UTM.

Now, as the Chatnantor maps are not too close to the local UTM
meridian, there will be a misalignment between "local UTM 'north'"
(an arbitrary term I just invented defined by a vector along
UTM-X = constant) and true geographic north.  In fact, 
UTM-X = constant on opposite sides of the mask will deviate from
"geographical north" by different amounts.

I suspect that we have not taken proper account of the true coordinate
transformations between UTM and the astronomical coordinates used
in the simulation or (u,v) calculation packages.  
(Or, a more precise statement: when I used to do this work, I did it
wrong WRT this point, and I am assuming other people may have gotten it
wrong as well, or more likely, not cared about the problem.)

I think the main effect will be a rotation, but there will be some
other coordinate distortion as well.

SO: to the extent that people have optimized configurations away
from a zenith snapshot, we will be somewhat in error.  For a zenith
snapshot, the orientation of the array on the ground doesn't matter,
but the symmetry is broken by earth rotation synthesis, so a
rotation of the coordinate systems will be of some consequence.

PROBABLY MORE IMPORTANT than going from simulations to plopping the
antennas down on the mask is going the opposite direction:  given the
antenna positions, we will need to calculate accurate astronomical u,v,w,
initially for simulation and eventually for astronomy. 

At the very least, it should be important to find out what the various
workers have done with coordinate conversions to the MASK.

For what it is worth,

	-Mark


> 
> Frederic,
> 
> The "exact" definition of UTM coordinates involves reference ellipsoids,
> coordinate projections, map datums, and other topics of little interest
> to us. In practical terms, UTM coordinates measure offsets in meters
> North and East from a reference point (far) outside our area of
> interest. The GPS receivers are setup for these coordinates. 
> 
> Bryan's mask and the maps are all based on UTM coordinates (SAm56). (Look in the FITS headers.)
> 
> The adopted center of the array is Chajnantor South:
>        7453228.830 N      627807.166 E
> 
> Where is the center of your coordinate system? If you used Chajnantor South, just add those offsets to your locations. If you subtracted some offset from Bryan's mask, add it back.
>