[daip] AIPS antenna numbers

[Eric Greisen]

Colin Lonsdale wrote:
> Hi Eric,
> 
> In another rare circumstance, I find myself trying to analyze some MWA data.  It has 128 antennas so I have instantly discovered that I can't get to first base without getting below the apparent AIPS limit of 90 antennas.  I can afford to throw away 38 of the antennas (since they are in useless locations for the investigation I am trying to do) but I can't find a way to get a consistent dataset that refers only to antenna numbers of 90 or less - I would need to renumber both antennas and baselines.  INDXR and various other tasks simply choke on the file.
> 
> What I am trying to do is use trusted and well understood AIPS calibration tasks to see if the DR limitation in some MWA solar images is a CASA problem.  There is tons of SNR and superb uv coverage, but Divya Oberoi has been reducing this in CASA and can't seem to get better than 500:1.
> 
> Is there any way to deal with this from within AIPS, or do I have to find a way to edit the dataset before generating the uvfits file so that AIPS sees only what it likes?
> 
> Thanks,
> 
> - Colin

If you are able to compile AIPS rather than use the binary distribution, 
then you could change $INC/PUVD.INC parameter MAXANT before doing so and 
then build with a higher allowed limit.  The 90 was based on a Japanese 
solar interferometer with 80 some.  There will be some annoyances from 
formats that know that I2 is enough but it should work up to MAXANT of 
200 or so.  There is a deeper limit at 255 in the encoding of baseline 
numbers.

I use gcc 4.1.2 and g77 3.4.6 (which are quite old) on my personal Linux 
machine.  Intel compilers work fine as do some gfortran versions (4.2.1 
worked but 4.0 and 4.1 did not).

Eric Greisen