• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name johna

    Log entry time 17:13:32 on March 8,2001

    Entry number 58055

    keyword=Shower calibration code.

    Now that the shower counter calibration code is working, I have put the necessary scripts into the e94104 espace directory. The current calibration (shower_calib.dat.final) should be pretty good for all energies, so it should not be necessary to calibrate the shower counter again for the online analysis, but I wanted to document the procedure. To calibrate a run, you need to do the following:

    1. Make sure that shower_calib.dat (and shower_calib_r.dat if it exists) has correct pedestal values.

    2. Put the correct HRS-Right momentum (i.e. dipole field) into header/hdr_shcal_test

    3. Run espace_kumac/sh128_cal.kumac (or use the script replay_shcal). I found
    that 100k events was adequate for a clean run (high e/pi ratio), but many more are required in some cases (run 2435 has a high pi/e ratio, and 200k was not enough - I did the final fit with 800k).

    *Keep in mind that shower_calib_r.dat will be taken as input if it exists, so make sure it has the correct pedestals (or get rid of it before doing the calibration).

    *ALWAYS REMEMBER that shower_calib.dat is overwritten by the calibration procedure, so make sure you have a saved version.

    *Note that if the fitting routine fails, you get a little error message in the espace window, and it does not write out the output files. This error message is easy to miss, so look for it to make sure everything worked alright.

    *Finally, the pedestals have to be correct in order to get a calibration that will work well at mulitple energies. I fit the pedestals from a run with the ADC thresholds set to zero, and they are in the file shower_calib.dat.newpeds. These are also in the current version of shower_calib.dat, but since this is overwritten by the calibration procedure, it is a good idea to check the pedestals against the 'newpeds' file (there are at least two other sets of pedestal values, and they give substantially worse results).

    After the fit, shower_calib.dat is automatically updated, and so this calibration gets used for all runs. If you're just testing the calibration, you should save the new calibration, and restore the old version to shower_calib.dat (if you're working in a shared directory).

    Finally, I find that the fitting code works fine when I run espace in my account, but sometimes fails when run by adaq (code is identical, as far as I know, and the espace kumac is identical, but at the very end the fitting routine fails). It may have something to do with a difference in the code or the version of some library. I cannot check this because I don't know where the source code is for the online version. My version is in /work/halla/e94104/disk1/johna/shcal/espace, if any of the experts want to take a look at the source or Makefiles.