[daip] Polarization Calibration

[Eric Greisen]

Doug Roberts wrote:
>   Hi Eric,
> 
> We have purchased a fancy new server and I'm getting the latest versions 
> of AIPS installed on it.  We have a set of scripts that automate 
> calibration of many datasets to do the same systematic calibration on 
> each data set.  There is a part of the script that does polarization 
> calibration (PCAL, RDLIF and applied with CLCOR).  The same script on 
> the same data is now getting a warning about no PCAL solutions:
> 
>     CLCOR3: Task CLCOR  (release of 31DEC12) begins
>     CLCOR3: Copied CL file from vol/cno/vers  1    1   3 to  1    1   4
>     CLCOR3: CL version input   3 output   4
>     CLCOR3: WARNING: PCAL SOLUTION UNKNOWN, SO NOT CORRECTED
>     CLCOR3: ANTCOR: UNKNOWN POLN. SOLN. TYPE =
>     CLCOR3: CLPARM( 1)=   66.00 / Phase(deg) to rotation for IF   1
>     CLCOR3: CLPARM( 2)=   66.00 / Phase(deg) to rotation for IF   2
>     CLCOR3:   3975 Records modified
>     CLCOR3: Appears to have ended successfully
>     CLCOR3: jansky 31DEC12 TST: Cpu=      0.0  Real=      0  IO=         3
> 
> The old version of the program that didn't give this lack of solution 
> warning was 31DEC11.  The new version that gives the warning is 31DEC12. 
> I looked at the help files for the two versions and the both seemed 
> exactly the same (I was thinking that there was a new default or that a 
> new variable was introduced to PCAL or something). 
> 
> Do you recall what might have changed in PCAL between the two versions? 
> I can send you more information if you need on the inputs etc.

I don't see anything that could account for this - the code in the two 
versions of CLCOR is the same and PCAL has not changed in any way that 
would affect the antennas table between the end of 31DEC11 and now.  It 
is possible that your 31DEC11 dates from some time prior to its being 
frozen.

However, the modern versions of PCAL have many more options and RLDIF 
has some too.  PCAL may have made a PD (spectral solution) table in 
which case it would not write the required keyword in the header.  And 
RLDIF has the option to correct the position angles of the D term 
solutions in either the PD table or AN table.  CLCOR would not normally 
be used in either 31DEC11 or 31DEC12.  Furthermore the values you are 
using are very suspicious - the correction is exactly equal to the 
source polarization angle?????

CLCOR is looking for the keyword POLTYPE in the header of the AN table.
PRTAB (DOCRT=1) will show you if it is present.

You may have to run your script a piece at a time and pay attention to 
what it is doing (compare TGET PCAL in new and tst) and use PRTAB on the 
AN file etc.  I attach the CookBook appendix covering modern PCAL.

Eric Greisen
-------------- next part --------------
A non-text attachment was scrubbed...
Name: COOKE.PDF
Type: application/pdf
Size: 488462 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/daip/attachments/20120522/a45aa233/attachment.pdf>