[daip] forwarded message from Subhashis Roy

[Chris Flatters]

Subhashis Roy <roy at ncra.tifr.res.in> writes:

> Aips 'help file' for the task 'PCAL' says that it can handle data taken
> with linearly polarised feed. However, the ATCA manual describes 
> (http://www.atnf.CSIRO.AU/computing/software/atca_aips/node12.html)
> 
> that AIPS does not handle data taken with linearly polarised feed
> properly.
>  
> As I want to analyse some ATCA data regarding polarisation, I want to
> clarify the above confusion.
> (I find AIPS to be much advanced as compared to MIRIAD in several aspects
> and want to stick to it rather than in MIRIAD.)
> 
> Hence, can you kindly help me in clarifying the status of calibration of
> data 'taken with linearly polarised feed' in AIPS.

There is a problem with the calibration of linear polarization data in AIPS
but it is mainly in regard to the handling of self-calibration rather than
the actual calibration of the polarization response of the antennas.

The two "parallel hands" of linear polarization will measure the following
Stokes parameters

	XX = I + Q
	YY = I - Q
	
CALIB will assume that XX = YY = I when determining self-calibration solutions
for amplitude and phase. This will probably introduce errors into the gain 
solutions for X and Y since radio sources can have linear polarizations at
the 10% level; in effect CALIB will treat the contribution from the Q Stokes 
parameter as a gain error. Any damage that this causes will be done before
you get to the stage where you want to run PCAL.

Note that this is not a fundamental problem with AIPS. AIPS merely lacks a
task that will use a full-polarization model for self-calibration.


	Chris Flatters
	cflatter at nrao.edu