Making Cubes at Jodrell

Livedata

1. Running livedata on oddjob

  • rlogin -l multi oddjob -log into oddjob from lyra
  • xdisplay -set the display to lyra
  • rm -r /home/multi/aips++/aips++.log -Get rid of old log file (can mess things up).
  • livedata
  • Change the 'Read Directory' to /arthur_data/rpfits/data to read in real time, or otherwise where the file is.
  • Change the 'Write Directory' to /scratch/lyra_2/HIJASS/backups/sdfits
  • Click on Bandpass Calibration, Monitor Output, Statistics and Write Data
  • Make sure 2 monitor windows come up - if not, get out of livedata, delete the aips++.log file and log into oddjob again.
  • Under queue selection, choose the available files, then click the 'Auto Queue' box for real time reduction
  • Press the green start button.


    2. Running livedata on lyra

    Its faster to run livedata on lyra, so if you are not processing in real time, this might be a better option. All the above steps apply, except the data will have to be ftp'd from arthur to lyra, as lyra cannot see arthurs disks at this stage.

  • Remember to still delete the aips++.log file
  • You will need to tell lyra where aips++ is: source /home/vkilborn/krun.csh
  • Thats about it, just run livedata as above.



    gridzilla

    To run gridzilla on either lyra or oddjob just type gridzilla If you are currently running livedata on oddjob, then you may have to run gridzilla on oddjob as well, as I'm not sure if they both use the aips++.log file.
  • Once in gridzilla, set the 'search path' to where the sdf files are.
  • Change the Beam FWHM to 12.0 arcmin
  • Change the kernal FWHM to 12.0 arcmin
  • click the 'autosize' button if you don't know where your cube center is.
  • Use the mouse to highlight all the sdfits files you want to cube.
  • Change the output FITS file name, and check the output FITS directory (don't want to write it to home!)
  • press 'go'

    Its a good idea to watch the sdf files load into the gridder (the first thing the gridder will do), as sometimes there is a problem and you can pick it up straight away.



    Concatenating the cubelets from gridzilla: cubecat & build_cube.csh

    After gridding you will be left with 20 or so 'cubelets' which need to be concatenated. The aips++ routine cubecat is quite unreliable, so try if you like. But build_cube.csh is reccommended for general use.

    1. cubecat

  • At either oddjob or lyra prompt, type cubecat
  • Press the 'select input' button, and highlight the cubelets to concatenate
  • Leave the processing options as default unless you want to hanning smooth or compress the output cube
  • Check the output directory and output name
  • Press 'go'

    The cubelets can be deleted after the big cube is made.

    2. build_cube.csh

  • This routine is stored in /home/multi/bin
  • At either oddjob or lyra prompt, type /home/multi/bin/build_cube.csh <cube name>, where the cube name is the stem of the cube name, before the _001 etc assigned by gridzilla. Lots of .mir cublets will be left once this has finished, which can then be deleted.


    Continuum subtraction: polycon

    This is done to the merged miriad cube. There are 2 versions, both kept in /home/multi/bin.
    For SUNs, its called polycon_sparc
    For linux pcs, its called new_polycon

    To run, type
    /home/multi/bin/new_polycon in=< cube name>.mir out=< cube name>_poly.mir
    It takes 10-20 mins to complete, depending on the speed of your computer.

    General Notes

    If any of the aips++ routines give funny errors or won't run, firstly delete the aips++.log file and try again, then get a new xterm and try again.