[daip] Fxpol troubles..

[Chris Flatters]

Jean-Francois Desmurs <desmurs at oan.es> writes:

> I'm starting the reduction of my project ED009 (Nov98, line experiments
> observed with the EVN in dual polarization and correlated at Socorro)
> and I have a problem with the labeling of polarization axes.
> 
> As usual when I download the data, the polarization appears to be two
> separated IFs, normally this is corrected running the task FXPOL.
> 
> Well in my case, after running FXPOL I obtained 1 IF and two
> polarization axes but these axes are label: LL and RL (and not LL and
> RR).  I tried different version of AIPS and the problem persists.  I
> tried to corrected by hand using GETHEAD/PUTHEAD but all the CTYPEn are
> empty.
> 
> I know that one antenna (Onsala) observed just one polarization, is
> this can be the source of the problem? How can I correct the label of
> the second polarization axes?

Versions of FXPOL prior to 24 May 2000 can only correct the polarization
labels for experiments in which odd numbered bands are assigned RR
polarization while even bands are assigned to LL. You have two choices.

The preferred choice is to use the 31DEC99 version of FXPOL with an
appropriate choice of BANDPOL for your data (BANDPOL='*(LR)').

Your only other choice is to go through the following painful procedure

1  Use AXDEFINE to change the reference value for the STOKES axis from
   -2 to -1.
    
2  Use UVCOP to copy IFs 1 and 2 to separate files.

3  Use VBGLU to join the IFs in reverse order.

4  Run FXPOL on the data set created by FXPOL.

Note that trying to skip the second and third steps will leave you with a
data set in which LL is labelled as RR and RR is labelled as LL. Since you
will be using external calibration data for EVN telescopes, this would lead
to the wrong calibration data being applied to each polarization as well as
a reversal in the sign of the circular polarization (if relevant).

	Chris Flatters
	cflatter at nrao.edu

	Chris Flatters
	cflatter at nrao.edu