VLA/VLBA Polarization Calibration Page
Steve Myers & Greg Taylor
NRAO, Socorro
Last update: 10 Jun 2005
ALERT! On 2005-06-09 I discovered a bug in our
AIPS procedures that affects Q-band A and B configuration data calibrated off
of 3C48. The effect of this error is that the Q-band polarization angles
(last column of the tables) can be significantly in error, effectively shifted
against 3C48. This error can be as high as 30deg (smaller in B-config, very
small in C and D so far).
I have fixed the data from 2004-06-27 onward and I am working
backwards to determine the extent of the error. Note that in general, my
guess is that the Q-band polarization angles in A and B config are only good
to about 10 degrees or so due to resolving the primary calibrators. This
is under investigation, possibly a better procedure can be devised.
If you have any questions or concerns about specific data, contact me.
- Steve Myers
The polarization calibration program is still being observed, though as you
see we have a large backlog in data in be processed (going back to 2003). We
are working on this backlog, and you can contact the analysts or
smyers at nrao if there are specific dates you are interested in.
General note on data quality: This program is observed as a series
of snapshots of the main calibrator plus various targets, and several scans
of a leakage calibrator. The data is reduced semi-automatically using
AIPS (procedures) with some scientist intervention. The goal of the program
is to calibrate the polarization electric vector position angle (EVPA) for
a network of sources suitable for VLBA observation (compact structure).
Flux densities for these sources are produced and posted but THESE FLUX
DENSITIES ARE NOT SUFFICIENTLY ACCURATE TO BE USED AS PRIMARY FLUX DENSITY
SCALE CALIBRATORS. The fluxes are provided as a check on data quality (errant
fluxes may indicate problems due to weather for that observation or other
problems) and as a guide to how strong that calibrator is. Note also you
should be careful in just picking one observation to use, be sure to compare
to nearby observations to make sure that this day is not errant.
The EVPA and flux densities are reference to the primary VLA calibrators
3C48 (0137+331) and 3C286 (1331+305). There is an entry in the HTML tables
for which one was used on a given day.
If you want to see what data is available from this program, you can
use the new
NRAO Data Archive Tool typing in POLCA as the project name [
query]
for data since Sept 2004 (when the program became POLCA) or TESTT
[query]
for data from Sept 1999 to Sep 2004 (look for the CXKQ data as there is some
ambiguity with other programs with similar name in the archive). You can
download the data using this archive tool and analyze it yourself if desired.
Note that the POLCA data is immediately publically accessible.
- Steve Myers
New: As of July 2004, I Stokes
source models for primary calibrators 3C48, 3C138, 3C147 and 3C286 are
available with AIPS. These can be found in your AIPS distribution in the
TEXT/STARS directory and can be loaded into your catalog within
AIPS using the CALRD task. See the
30Jun04 AIPSLetter for details. These models are also available from
this site in the Images/Current subdirectory.
(2005-5-26)
- These pages are being overhauled, especially this page, the Gains pages
and the Images pages. Please bear with us while we chase broken links.
(2003-12-16)
- The K-band subreflector rotations have
been reoptimized as of Mon 13 Jan 2003, so current posted gain curves are
invalid. See note on the Gain Curve Page.
- A list of VLA memoranda dealing with polarization issues is
available here. This is somewhat out of date and unfortunately
most memoranda are not available on the web, only in hardcopy form
at NRAO libraries. (2002-9-18)
- See previous news items for older items.
Starting in September of 1999 we began a VLA polarization monitoring
program for a number of compact calibrators, widely distributed in Right
Ascension, primarily to enable better calibration of the VLBA. Each source
is observed about once or twice a month and reduced semi-automatically. Flux
densities and polarization angles are tied to observations of 3C48, 3C138,
or 3C286. The calibrators in this dataset should
be useful as EVPA (polarization E-vector angle) calibrators for VLA
and VLBA polarization observations. Note that the angles given in
this table are R-L Phase differences as CLCOR desires, not the EVPA
itself. The "Flux" column is derived from GETJY after using SETJY
versus 0137+331 (3C48) or 1331+305 (3C286) to establish the scale, and
thus can be used in SETJY. Note that the angles derived for 0521+166 (3C138)
in B-configuration agree with those given in the
polarization calibration guide, and the flux density seems stable.
Be sure to read the table notes for cautionary
information. Currently only C (5 GHz), X (8.5 GHz),
K (22 GHz), and Q (43 GHz) bands are being monitored.
At K and Q bands, there are significant excursions in the flux density
calibration due to problems with pointing, gain curves, or poor
weather for a given session. We are looking into improving the
reliability of these measurements, so bear with us.
In addition to the selected sources, other possible calibrators
were observed during the course of the program. The entire database
can be found in the Master Calibration
Database.
So far, we have identified five stable polarization
angle calibrators (up to Q-band) other
than 0137+331 (3C48) , 0521+166 (3C138)
and 1331+305 (3C286) :
0423-013 ,
0854+201 ,
1310+323 ,
1751+096 , and
2136+006 .
Some of these are relatively flux-stable also. In addition, the source
2355+498 , although essentially unpolarized,
has proved to have a stable flux density
and thus could serve as a secondary flux calibrator. Surprisingly,
the very bright sources 3C273 (1229+020) and
3C279 (1256-057) seem to have relatively stable EVPA
at high frequency despite being highly variable in flux density!
CAUTION: There are significant systematic deviations in the derived
flux densities for some of these sources, and use of these for flux
calibration could be done with extreme care. Check with us if you plan
to do this and are concerned.
Some further info:
The most recent versions of AIPS runfiles are available for B-config
(May 2005), with info in these files to change for different configs
and calibrators. These are:
CALBC,
CALBX,
CALBK,
CALBQ for each of CXKQ bands.
In the AIPS versions 31DEC01 and later,
FILLM will by default apply gain curves and approximate opacity corrections,
and thus you should not use manual gain curve corrections unless you defeat
these (check BPARMS). You can check your CL 1 table to see if you have
done this. If so, and you want to change this, you can delete CL 1 and remake
it with INDXR to start over with either unity gains [BPARM(2)=-1] or default
gains [BPARM(2)=0]. The default gain curves can be found in your AIPS
installation in the TEXT/IONS/VLA.GAINS file. I recommend using
these, so by default set BPARM=-1,0 in FILLM, where BPARM(1)=-1 defeats
the WX-based opacity correction (see below).
If you want to use TIP curve derived opacities in CLCOR,
set BPARM(1)=-1 to turn off the automatic application of WX-based opacities
when using the 31DEC01 or later AIPS. Because the WX opacities are not very
accurate, and are time variable over the run, I do not recommend using these
at this time if better TIP data is available. I will investigate this
further.
VLBA Scientific Memo 26 detailing this program is
available.
If you have scheduled tipping scans in order to determine
atmospheric optical depths, you can find the results on
Bryan Butler's Tip Scan Archive page.
Steve Myers & Greg Taylor
Tables for both the selected calibrators and for all sources
observed during the program are available together or broked down
by band or by epoch:
Using VLBA test data taken in November 1999, we have compiled
some polarization images of selected calibrators from the database:
In particular, note the polarization structure in potential
calibrators from this list, as the VLA polarization angles will be
emission weighted integrals over these structures.
These test observations were analysed in AIPS using an
automatic mapping procedure. Example of
RUNFILES which load the PROCEDURES into AIPS are available here, indexed by a
two-letter code designating the Configuration (A-D) plus 2 hybrids (CnB and
DnC), and frequency band
(CXKQ), and with the Date of the RUNFILE and the Version of AIPS it was
written for:
To modify these for your own use, change the variable that are
loaded in the procedure SETIT to correspond to the primary and
secondary calibrators that you are using in your run. The procedures
functions should be reasonably self-explanatory.
To run your RUNFILE, copy it to the directory you start AIPS from,
in AIPS set the VERSION='PWD', and invoke it (eg. RUN CALAK). The
sequence of procedures to run is explained in the header to the
RUNFILE.
Questions about this page should be directed to either analysts
or smyers at nrao.edu.
Modified on
Thursday, 28-Feb-2008 10:59:52 MST
|
