Errors associated with reducing filterbank data

About the only thing that can go wrong is not having the pulsar details correct in the catalogue OR not integrating for long enough.

If the data reduction has failed because the integration was not long enough then fch3 will generate the message Not enough data for a subint which will be written either to the screen, when processing manually, or to the log file, when processing automatically. There is nothing that can be done to fix this problem apart from integrating longer.

The second most common problem is a bit more complex. You can check what the catalogue thinks it knows about the pulsar by using psrinfo and typing

psrinfo -c "P DM" 0437-4715 (for example)

When creating archives the parameters that you must have set to something in the catalogue are:
  • TDMP: - ideal time to integrate for before dumping a profile
  • TDMPMIN: - time below which the data is thrown away.
  • NSUB: - Number of frequency sub-bands
  • NBIN: - Number of bins in the profile
  • NSPAN: - Number of minutes for tempo polynomial
  • NCOEF: - Number of coefficients in the polynomial

    • So it is best to try:

    psrinfo -c "TDMP TDMPMIN NBIN NSUB NSPAN NCOEF" psrname

    and see if you like what it is about to use. If not change these parameters in /psr/soft/psrinfo/fch3.db on kepler and then remake the catalogues by typing:
    make catalogues
    in the /psr/soft/psrinfo directory. Note if you are reducing this data at Parkes you will have to copy psrcat.unformatted back up to $PSRHOME/soft/psrinfo at Parkes before rerunning tarchfb_driver.
    Back to the Filterbank data reduction page.
    Back to the main page.