Transfer of data from Arthur to the Unix machines

This operation is normally carried out by Mark Roberts or Chris Jordan , but in a case of emergency the following is the course of action required.

1. Log onto a SUN machine and set the directory to /psrdata/timing
2. ftp arthur
3. cd disk$data:[pulsar.obs.psrdata_1]
4. binary
5. get 970102_1.bin 970102_1.VBIN (the uppercase is important) if there is more than one file for a particular date then transfer the file with ;1 first then each of the others in order giving a generation letter (note position of it) i.e.
   get 961217_1.bin;1 961217_1.VBIN
   get 961217_1.bin;2 961217_1b.VBIN
   get 961217_1.bin;3 961217_1c.VBIN etc
   
6. quit
7. vms2unix 970102_1
   vms2unix 961217_1b etc
   note that the .VBIN extension must not be used
   This will create a file 970102_1.BIN
   961217_1b.BIN (yes uppercase BIN)
   
8. MOST IMPORTANT.
   In the directory $PSRHOME/data/obsinfo there is a file called OBSFREQUENCIES.DAT. To avoid problems later on this file should be updated with the details of the observing frequencies etc. before any files are sorted.
9. psrprof
10. set in1f 970102_1.BIN
11. sort
    At this point the program looks at the 7th and 8th characters of the file name to decide whether to create or append to files 97a.bin, 97a_2.bin or 97a_3.bin. This is why the extension must come after the '1'
12. JUST AS IMPORTANT.
    In the directory $PSRHOME/data/obsinfo there is a file called

    There is a file in $PSRHOME/data/obsinfo called filessorted. It is most important that this file is updated with the files that have been sorted, as the next time this will be used as the reference as to which files still need to copied over from Arthur to be sorted.
    

    If the files are NOT sorted in chronological order then the command in psrprof 'seek time' will not work and there will be data that CANNOT be acessed or analysed.

More information about the further processing of the data see Use of the Pulsar Timing Programs