[daip] Re: [alma-config] New DOT and D resolution results

[Leonia Kogan]

John,

We have decided here that such an artificial feature would lead 
UVCON too far aside of the main road of AIPS.

However I have modified UVCON which you can use as your private version.
Some comments: 

1. BPARM(8) switches the options
2. IF BPARM(8) .GT. 0 then UVCON interpretes the input file data 
   as baselines. But UVCON recalculates the input positions to the equatorial 
   coordinate system. Threfore to get baselines equaled to the input files 
   data, the input file has to have zero(not 1) at the first column.
3. If you want simulate 2016 baseline (corresponded to 64 antennas), 
   you need to have 2016 lines in the input files(not 2017).
   The private version of UVCON does not subtruct the first line.
   Instead it consider each line including the first one as a baseline.
4. Baseline are coded in AIPS as 256*IA1+IA2, so there is a confusion 
   in the baselines read as a random parameter. You'll see that 
   using PRTUV for example. Fortunately IMAGR (as I understand) does not 
   pay attention on the baseline number.
5. I have checked your UVCON and it works with my data. But I can not 
   guarantee that it will works always.
6. All changes are labeled by the line: CJOHN. It can help you if you 
   want to continue modification.

Leonia
----- Begin Included Message -----

>From jconway at ebur.oso.chalmers.se Thu Feb  1 02:22 MST 2001
Date: Thu, 1 Feb 2001 10:21:58 +0100 (CET)
From: John Conway <jconway at ebur.oso.chalmers.se>
To: "[Leonia Kogan]" <lkogan at zia.aoc.NRAO.EDU>
cc: jconway at oso.chalmers.se
Subject: Re: [alma-config] New DOT and D resolution results
In-Reply-To: <200101311852.LAA00416 at bonito.aoc.nrao.edu>
MIME-Version: 1.0
X-Lines: 57
Status: RO


Hi Leonia,

 I was wondering whether it would be possible to change UVCON to 
take >2017 antennas (say 2500).
 
 What I'd like to to is to test the mapping qualities of different 
distributions of uv  points (say uniform within a circular boundary 
and gaussian tapered) without worrying about the problem of how to create 
these uv distributions from an antenna array.

If the input allowed 2017 station inputs the first staion could 
be at position at 0,0 and the other 2016 'station' positions  
in the UVCON input file would have  coordinates of the uv 
points I wanted.

UVCON could then have a  switch which generated only uv data from the
input model for the
baselines between antenna 1 and the other antennas. We could then 
hopefully use IMAGR or  VTESS to do deconvolution tests. Would there
be any problem CLEANING such data? It would just look like a conventional
uv file with data only on baselines to antenna 1, right?

Perhaps changing the number of stations is a big job, since it may have 
implications for the rest of AIPS.  In that case we could live with 
1000 stations and just implement the switch in one of the APARMs of
UVCON to just generate uv data on basleines to antenna 1, then we can 
do simulations of 999 uv point arrays, which is fine to test the
different imaging qualities of different uv point distributions.

    John.




> Just for information I have fixed the problem and I am going to put the
> fixed version to the system soon.
>  
> Additionnaly I have increased the number of elements(antennas) to 1000
> to make UVCON work for SKA. The new UVCON does not plot both UV
> coverage and the configuraion, if the number of antennas more than 100
> plotting only configuration.
>  
> As I know you are not served by midnight job.
> So the new version of UVCON will not be available to you imidiately.
> 
> John, 
> 
> Explain me please once more why do you want to have even more ellements 
> (how many?).
> 
> Thanks
> 
> Leonia
> 
> 
----- End Included Message -----