Data Reduction - Scanning

Livedata and Gridzilla - Reducing rpfits to mscals, then gridding

Livedata

To get livedata to run, the first thing we have to do is check the environment variables KARMAROOT and KARMABASE have been set to the working version of karma at /home/rms/karma This can be done by typing

printenv  KARMAROOT   
printenv  KARMABASE

To set or change their current settings type:


	setenv KARMAROOT /home/rms/karma
	setenv KARMABASE /home/rms/karma

You can then start livedata.

Type livedata
Note - do not run it in the background (i.e. do not use the &), livedata will not work if you do this.
You can now move the windows around a bit. (Minimise the two multibeamview windows, as these will not be used at present).
Now click the Monitor Output Button on the Livedata GUI, this will switch it off, otherwise Livedata will run EXTREMELY slowly.
Next we have to set-up the directories for reading and writing the data. Under Read Directory type /arthur_data/rpfits/data Then click Update.
A list of the .mbf (raw data files) currently available on ARTHUR should appear. The file names all start with the observation time coded as YY-MM-DD_HHMM. ARTHUR usually only holds data from the last few days. If the observations are older than this they should be found on oddjob's disc in a directory under /scratch/oddjob_1/rpfits
Select the mscal directory by typing /scratch/oddjob_1/mscals into the Write Directory box. If you wish, you could create a sub directory in this directory called 'p74' or 'virgodeep' and set this as the directory, depending on which data you are reducing.
We now have to select which .mbf's we want to reduce.
If Chris has cleared out all the old data (i.e. if there is just data with 00-03-30_......mbf), then we can just select all data in the directory - just click on the Auto-Queue button. Otherwise, highlight the files you want to reduce (they are prefixed with the date of observation, so you should be able to work out which ones you've just made), then click the Queue Selection button to queue the files for reduction. Even if you've clicked the Auto-queue button, you MAY have to manually queue the first file.
You can then start livedata reducing by clicking the Start button. The list of reduced files should gradually appear in the Processed files list. If it hangs, just shut everything down and start again. Note that the two 'Multibeambiew' windows will not close when you shut down livedata - you will have to close these down manually.
Note also that during reduction you may occasionally get an error - something like TAC leap table out of date, regenerate or something similar, this is normal. You may also get a few Ignoring this scan as Tsys = ..., these are also common.

Gridzilla

Once you have the reduced mscals for your cube, you can grid them by typing :

Gridzilla
Again, do not use the ampersand.
First, click on the Autosize button, this means that gridzilla will automatically calculate the RA and DEC of your cube.
Click on the Truncated 16-bit integer button.
In the Search path box, put the path to your mscals, eg
```
	/scratch/oddjob_1/mscals
```
Or whatever subdirectory you've decided to write your mscals to.
Clicking on Search will then display the list of mscals in this directory.
Assuming all these mscals are for the same cube, use the mouse to highlight them all. Then give the cube a name in the bottom box, and a directory to write to in the box above (eg. set this to the same one as the mscals so that we have all the data for each set in the same place).
You can then start the gridding by clicking Go. Gridzilla will tell you when it's finished (look in the terminal window which you started gridzilla from, it will tell you how many sub-cubes it's created, and when it finishes).
Gridzilla MAY or MAY NOT have created sub-cubes rather than a one single cube. If it's created many small cubes (which will have the name you gave the cube in gridzilla, with _001 _002 etc on the end eg. virgop16_001.fits virgop016_002.fits etc.), then you can do the following :
- Copy /home/multi/dan/build_cube.csh to the directory which has all your sub-cubes in.
- Type build_cube.csh virgop16
  ( i.e. replace virgop16 with the name of the sub cubes, but WITHOUT the _???.fits extension). This should then use a set of miriad commands to create a single cube. If this doesn't work, you may have to :
  - ```
   source /usr/local/miriad/bin/MIRRC 
```
  (to set-up the new version of miriad)
- textedit build_cube.csh and replace all references to the fits and imcat commands with the full paths to the new versions ie.
```
  	fits  -> /usr/local/miriad/bin/sun4sol/fits
  	imcat -> /usr/local/miriad/bin/sun4sol/fits
```

And that should be it!
DCZ, JAH Mar 30 2000

Last modified: Wed Jun 14 17:32:28 BST 2000