[daip] [Fwd: Re: CLCOR position shifting]
Craig Walker
cwalker at nrao.edu
Wed Jan 27 15:12:13 EST 2010
Mark,
I finally got a chance to look at your data, which I had downloaded.
Clearly something very weird is happening in CLCOR. The rate of change
of the derived delay abruptly changes in the middle of the scan, and
then takes a step a bit later. The phases reflect something of the
sort, but the rates do not. We tried rerunning CLCOR with your
parameters and got something similar. We then tried with smaller
shifts, near 0.1 arcsecond, and don't see a big problem.
I showed all this to Leonia and gave him a copy of the data set. He
will try to figure out what is wrong.
Eric and Leonia believe that an arcsecond shift is too big for CLCOR to
do accurately. It will ultimately be interesting to see just how true
this is and your data set is a good one for such tests.
Note that one problem with this testing is that, each time CLCOR is run,
the source position in the SU table is changed. Thus with many repeated
runs, the source is marching off in some direction. I just ran several
tests, and now the source position is off by several arcseconds. I
don't know if this is having any effect on the immediate problem, but I
doubt it. I do see from the history that you ran a very large number of
CLCORs, as I would have if I'd seen something of the sort.
Cheers,
Craig
Mark Reid wrote:
> Craig,
>
> I finally got around to the test of CLCOR you suggested (looking at
> phase versus time) and the results are bizarre.
>
> Attached is a .tar file with 2 uvplt's. Both are phase versus time for
> a single 16 MHz IF (averaging channels across the band to incrase SNR)
> on a short baseline (PT-LA). The data were correlated with a position
> offset of (RA*cos(Dec),Dec)=(0.81",0.77")
>
> 1824+10X_CL5.ps applying standard calibration (CL5)
>
> 1824+10X_CL5.ps after shifting in CLCOR (CL6)
>
> Something really bad is happening in CLCOR. The parameters I used were
> as follows:
>
> AIPS 1: CLCOR Task which applies various corrections to CL tables.
> AIPS 1: Adverbs Values Comments
> AIPS 1: ----------------------------------------------------------------
> AIPS 1: INNAME '2009NOV16-A1' Input UV file name (name)
> AIPS 1: INCLASS 'CALDAT' Input UV file name (class)
> AIPS 1: INSEQ 1 Input UV file name (seq. #)
> AIPS 1: INDISK 1 Input UV file disk unit #
> AIPS 1: SOURCES '1824+10X' Source list ' '=>all.
> AIPS 1: *rest ' '
> AIPS 1: STOKES ' ' Stokes type to process
> AIPS 1: SELBAND -1 Bandwidth to select (kHz)
> AIPS 1: SELFREQ -1 Frequency to select (MHz)
> AIPS 1: FREQID -1 Freq. ID to select
> AIPS 1: BIF 0 Lowest IF number 0=>all
> AIPS 1: EIF 0 Highest IF number 0=>all
> AIPS 1: TIMERANG *all 0 Time range to use.
> AIPS 1: ANTENNAS *all 0 Antennas to correct.
> AIPS 1: SUBARRAY 0 Subarray; 0 => 1.
> AIPS 1: GAINVER 5 Input CL table 0=>high
> AIPS 1: GAINUSE 6 Output CL table: not =
> AIPS 1: GAINVER -> high+1
> AIPS 1: OPCODE 'ANTC' Operation code.
> AIPS 1: CLCORPRM 0 0 Parameters (see HELP CLCOR).
> AIPS 1: 0 0 0.814 0.77
>
> FYI, The .tar file also contains a runfil (TEST1.08D) which documents
> the steps taken to do the "standard" calibration, making CL5. (It
> starts with CL2 which was generated by VLBATECR.)
>
> I have also put the entire data set as a FITS file on my website
>
> http://www.cfa.harvard.edu/~reid/2009NOV16-A1.FITS
>
> The file is about 300 MB in size.
>
> Finally, in case it is of any use, I've attached a printout of my
> position fitting program which lists the correlated positions and fitted
> position offsets of sources (those with an "X" at the end of their names
> were correlated with arcsec-ish errors).
>
> Mark
>
> -------------------------------------------------------------
> Mark J. Reid Phone: 617-495-7470
> Harvard-Smithsonian CfA Fax : 617-495-7345
> 60 Garden Street Email: reid at cfa.harvard.edu
> Cambridge, MA 02138, USA Web : www.cfa.harvard.edu/~reid
> -------------------------------------------------------------
>
> Craig Walker wrote:
> > I was waiting for an answer from the AIPS group, but haven't seen
> one. Only they can answer your direct question.
> >
> > But one diagnostic question - is the SNR good enough to see the data
> phases? If so, what do they look like when you try to correct the
> position. They should be reasonably flat and, if CLCOR is messing
> things up badly enough that FRING can't find anything, they should be
> all over the map.
> >
> > Cheers,
> >
> > Craig
> >
> > P.S. Mark, Will you be at the AAS?
> >
> >
> > Mark Reid wrote:
> >>
> >> Dear DAIPser:
> >>
> >> I have been trying to debug some procedures (runfiles) for our
> VLBA Key Project for parallax measurements. As a test I
> observed/correlated some calibrators at their correct positions and
> offset by ~1 arcsecond.
> >> I then FRINGe fit the data, getting delays and rates, and then solve
> for position offsets. This seems to work well.
> Plenty of SNR, but I've not done that check yet. Actually, that data
> set (BR145A1: 2009Nov16) is a good one for checking software. It has
> many strong QSOs processed at two coordinates.
>
> I have run POSSM and one can easily see the phase-slopes across the
> bands, even for the sources with the ~1" position shifts. After
> correction of the position errors the SNR drops considerably. (And I've
> tried both sign conventions for the shifts.)
> I'll look at the phases as a function of time on Monday.
>
> Mark
>
--
---------------------------------------------------------------------
R. Craig Walker Array Operations Center
cwalker at nrao.edu National Radio Astronomy Observatory
Phone 575 835 7247 P. O. Box O
Fax 575 835 7027 Socorro NM 87801 USA
---------------------------------------------------------------------
More information about the Daip
mailing list