P-band imaging in <img src="cmsy10-41.png" alt="A" class="10x-x-41" /><img src="cmsy10-49.png" alt="I" class="10x-x-49" /><img src="cmsy10-50.png" alt="P" class="10x-x-50" /><img src="cmsy10-53.png" alt="S" class="10x-x-53" />

The following steps relate to imaging your calibrated data. As the aim of this appendix is calibration, the imaging will be discussed only briefly here. The following items refer to continuum imaging. Make a dirty image of the center with IMAGR

> default imagr C R

> indi d; getn 10; outdi indi C R

> bchan 5; echan = n - 4 C R

to omit edge channels in each IF, where n is the number of channels in an IF.

> nchav = echan-bchan+1; chinc nchav C R

to average all spectral channels into the image.

> cellsize 10 C R

to set a cell size which will depend on the VLA configuration.

> imsi 1024,1024 C R

> niter 0 C R

to do no Cleaning

> go C R

Clean image with faceting, auto-boxing, and TV display; we use DEFAULT IMAGR here to initialize all of IMAGR’s many parameters

> default imagr C R

> indi d; getn 10; outdi indi C R

> task ’setfc’ C R

> cellsize 0 ; imsize 0 C R

to have SETFC determine the correct cell and image size for these data.

> bparm 2.2, 5, 2, 8, 0.3 C R

to image fully out to a radius of 2 degrees and to include 0.3 Jy NVSS sources out to a radius of 8 degrees.

> flux 1 C R

to include exterior sources above 1 Jy only.

> boxfile ’MYAREA:Target1.box C R

to list the facets found.

> go C R

The following does not use default since we need adverbs set for, and by, SETFC, such as inname, imsize, and cellsize

> task ’IMAGR’ C R

> flux 0 C R

> bchan 5; echan = n - 4 C R

to omit edge channels in each IF, where n is the number of channels in an IF.

> nchav = echan-bchan+1; chinc nchav C R

to average all spectral channels into the image.

> niter 5000 C R

to Clean 5000 iterations.

> do3dim true ; overlap 2 C R

to use true 3-D geometry, Cleaning one facet at a time in each cycle (recommended).

> dotv true C R

to control the imaging with the TV.

> oboxfile boxfile C R

to save all changes to Clean boxes.

> im2parm=2,0 C R

to allow up to 2 Clean boxes to be created automatically at each cycle.

> go C R

Note, the field may have enough spectral index variation to require imaging (and self-calibration) of each spectral window (IF) individually.

There is normally enough flux in a P-band field to do phase self-calibration and we recommend that you now do that. Use CALIB with the output of step 39 as the model including all NFIELD facets as the NMAPS images.

> default calib C R

> indi d; getn 10; outdi indi C R

> nmaps = nfield C R

to include all facets

> in2d = indisk; get2n M C R

to set the 2^nd file name to the output image of class ICL001 (catalog number M).

> soltyp ’l1r’ ; solmode ’p’ C R

to start with phase-only solutions.

> solint = nx C R

to set a short integration time as n times the sampling interval.

> go C R

Examine the output SN table with EDITA to see if there are times with bad solutions. If not, repeat steps 39 and 40 using

> tget imagr; docal = 1; go C R

to make new images.

> tget calib; in2seq = 2 C R

to point at the second set of images.

> solmode ’a&p’ C R

to do amplitude as well if there is enough flux in the field.

> go C R

Re-examine the new SN table with EDITA. In particular, watch for periods of time when the amplitude gain of an antenna gets rather larger than normal. It has been found that the ionosphere can defocus sources on scales as small as the antenna diameter. It is best to delete such data.

Spectral-line imaging is both simpler and more complicated. One may image each spectral channel individually. However, even for absorption experiments, it may be best to image all channels free of the spectral line together to make a continuum image and model. This will almost certainly require multiple facets and extensive cleaning but it will produce a single continuum model that can be applied to all spectral channels with UVSUB. Spectral index in the continuum may force you to do this one spectral window at a time. You can add the continuum image back to the cube with COMB.

At this point, run IMAGR on the output of UVSUB one spectral window at a time over all spectral channels of interest with NCHAV=1; CHINC=1. You may be able to limit the field of view imaged to one facet surrounding the target source.

The task TVSPC is a lovely way to explore the output image cube. To run this task, we must first transpose the cube

> default trans C R

> indi d; getn C; outdi indi C R

where C is the catalog number of the output cube from IMAGR.

> transcod = ’312’ C R

to put frequency first.

> go C R

to make a transposed cube.

One also needs a “meaningful” single-plane image, perhaps the central facet of the continuum image made above. Or we can compute the first moment image.

> default xmom C R

> indi d; getn T ; outdi indi C R

where T is the catalog number of the output cube from TRANS.

> flux = f C R

to compute only with brightness > f Jy/beam.

> go C R

to amke a first moment image.

Finally

> default tvspc C R

> indi d; getn M; outdi indi C R

where M is the catalog number of the meaningful image, perhaps the output plane from XMOM.

> in2di d; get2n T C R

where T is the catalog number of the output cube from TRANS.

> in4di d; get4n C C R

where C is the catalog number of the output cube from IMAGR.

> go C R

Have fun!

This allows you to look at spectra from the cube at individual spatial pixels as well as the cube as an image at individual frequencies. Enter EXPLAIN TVSPC or see Memo 120 for details.

Make image(s) and final calibrated data sets into FITS files with FITTP. Exit AIPS and

> setenv FIT /lustre/mmao/Pband C R

for tcsh and c shells

> export FIT=/lustre/mmao/Pband C R

for bash shell

Then start AIPS again

> default fittp C R

> indi d; getn M C R

to select an image or data set

> dataout ’FIT:VirA.FITS C R

to make an appropriate output name

> go C R

And repeat with new M and DATAOUT values.

L.2 P-band imaging in