Correcting antenna positions: LOCIT

V.2 Correcting antenna positions: LOCIT

The data taken for pointing analysis after antennas have been moved is now also used to determine corrections to the assumed antenna locations. The data are loaded into using BDF2AIPS in the usual way, but DOKEEP = 1 is required in order to load pointing data. Then FRING is run on one scan from a strong source with SOLINT = 5 minutes and with no rates (DPARM(9) = 1). Then CLCAL is used to apply the SN table to all times in the CL table. Then the data set is reduced in size by running SPLAT, averaging in frequency and time:

> DEFAULT SPLAT C R

to initialize the adverbs

> INDISK n ; GETN m C R

to specify the data set.

> OUTDISK INDISK C R

to keep on same disk.

> APARM 3, 1, 0, 0, 1 C R

to average data in frequency, but not over IF. The input data are taken at 1-second interval.

> DOCAL 1 ; GAINU 0 C R

to apply the calibration.

> CHANNEL 64 ; CHINC 0 C R

to average all spectral channels together.

> SOLINT 5/60 C R

to average over time to 5-second intervals.

> INP C R

to check the inputs.

> GO C R

to make a rather smaller data set.

> INDISK n ; GETN j C R

to specify the averaged data set.

> INEXT ’NX’ ; INVERS 0 ; EXTDEST C R

to delete the index table which has each direction of the pointing as a separate scan.

> DEFAULT ’INDXR’

to initialize everything.

> INDISK n ; GETN j C R

to specify the averaged data set.

> CPARM 0 , 10 C R

to set maximum scan length to 10 minutes.

> GO C R

to make the new index table.

It might be wise to make a backup copy of the data set at this point in case later machinations go wrong. Use UVCOP.

A solution table is now needed containing the phases of every scan in the data set. Use

> DEFAULT CALIB C R

to initialize the task and adverbs.

> INDISK n ; GETN j C R

to specify the averaged data set.

> SOLINT 10 C R

to average full scans.

> APARM 3, 0 C R

to require at least 3 antennas, but do not average polarizations or IFs.

> SOLTYPE ’L1R’ ; SOLMODE ’P’ C R

to solve for phases only with recursive L1 methods.

> DOCALIB 1 ; GAINUS 0 C R

to apply calibration, a no-op at this stage.

> SNVER 0 ; REFANT nn C R

to create a new solution table and to select a reference antenna (nn)

> INP C R

to double check.

> GO C R

to make the solution table needed next.

Finally, fit the antenna positions. It is best to do each IF separately and also try the mode that uses the phase difference between IFs.

> DEFAULT LOCIT C R

to initialize things including DOTV false, and REFANT 0.

> INDISK n ; GETN j C R

to specify the averaged data set.

> BIF 1 ; EIF BIF C R

to examine one IF at a time.

> STOKES ’RR’ C R

to examine one polarization at a time.

> DPARM 0, 1, 10, 120 C R

to use each scan by itself (rather than differencing consecutive scans), do not fit the K term, and limit the fit to elevations between 10 and 120 degrees. All hour angles and declinations are used. The total phase is used.

> BPARM 3, 0, -180, 180, 3, 4, 1, 3 C R

to plot 3 antennas per page between -180 and 180 degrees as boxes with connecting lines, and, if doing more than 1 IF or polarization, plot them in the same plot.

> DOOUTPUT -1 C R

to ignore any values contained in pre-existing files named OUTPRINT.arrayname. If this file exists before running LOCIT and DOOUTPUT > 0, it will be read and the corrections contained within it added to those found before they are written to a new OUTPRINT.arrayname file and also to the other output files. See HELP LOCIT C R for details.

> OUTPRINT ’FITS:date’ C R

to set the output file base name used for output text files.

> INP C R

to check that all is well.

> GO C R

to run the task.

Then do

> STOKES ’LL’ ; GO C R

to redo with the other Stokes.

> BIF 2; EIF BIF C R

to redo with the second IF.

> STOKES ’RR’ ; GO C R

to redo with the other Stokes.

Now compare the answers and uncertainties found in each of the four runs. Also it may be good to do

> BIF 1; EIF 2 C R

to include both IFs.

> STOKES ’RR’ C R

to do one Stokes at a time.

> BPARM(9) 1 C R

to select the mode which uses the phase difference between the 2 IFs.

> GO C R

to get a fifth estimate of the antenna position corrections.

Repeat for STOKES ’LL’.

Figure V.1: Example of the residual phases after a LOCIT solution is applied to the input SN-table phases. The solutions for antennas 14 and 15 are fine, but antenna 16 did not converge to a correct solution. The data from both IFs and both polarizations were used in this run of LOCIT and are represented by different colors in the plot.

Now comes the hard part. LOCIT will have created a great many plot files, an example of which is shown in Figure V.1. It also writes text files named OUTPRINT with extensions:

FIT	This file contains all solutions with their uncertainties in a form that is easy to read. It may be used to accumulate corrections depending on the setting of DOOUTPUT.
EVLA	This file contains only significant corrections in a form that can be hand edited to put in the dates when the antennas move and when the database is fixed. The edited result may be edited to a system file on the web to enable VLANT to apply the appropriate corrections.
PAR	This file, containing only significant corrections, is read into Parminator to update the database.
001	This file contains only significant corrections in a script form to read into CLCOR. Do this, to make a new CL table and then rerun CALIB to see if the resulting phases really end up near zero.

You will need to review the plots (TVPL is helpful) and study the .FIT file to decide which antennas have significant corrections that need to be made. You will almost certainly need to edit the .001 file and then apply it to your averaged data set. Repeat CALIB and then examine the resulting SN table with EDITA or SNPLT. If things worked well, then edit the .PAR and .EVLA files to match.

[prev] [prev-tail] [front] [up]