GAMMASPHERE information, calibration BRIEF

This is a brief description of how the germanium calibration of the GAMMASPHERE array is done, both off-line and on-line. The descriptions are given is the order the tasks normally are executed.

1. Detector selection, etc.:

Starting with a raw un-calibrated array, the first thing to do is to decide which detectors will be on during the experiment. Use the EPICS cfd screens to toggle on/off detectors. Also, make sure the gain calibration option is enabled on the EPICS control screen.

Go into the gecal1 script and specify the 'master' germanium selection acording to what detectors are on. Store this selection using the 'store master' option (red button). You must also specify a list of the EFFs that you are using.

Use the 'dmp to spe' file option in the script to dump the EFF detector spectra to disk and use the Radford gf2 utility to display the spectra of those detectors. You may want to clear the spectra so you don't catch some old data. This step ensures that all detectors that are 'on' are in good working order and worthy of going to tape.

2. Array calibration:

To do the first level gain calibration, place a 207Bi source in the array. Zero the germanium spectra in the EFFs using the zapping option in the gecal1 script. Wait for enough statistics, and use the auto-bi-cal option in the script to align the two peaks of this source. The script will automatically zero the germanium spectra. Thus, after a short while the germanium spectra can be dumped again - the peaks should be perfectly aligned. Use the 'x1.cmd' or 'x2.cmd' gf2 command files to quickly display all the spectra in overlay mode or one-by-one, respectively. (You execute the command file by e.g., 'cf x1'. Use 'x0' and 'nx' to set the range, 'rd1' to display full range and 'ex' to expand, click here for further details on commands in Radfords 'gf2' package)

If very high precision of the energy calibration is desired, it is recommended to use the multi-line calibration in the script too. This calibration should only be done AFTER the 207Bi calibration has been done. An Eu-152 and Ta-182 source in the array will give a wide energy range calibration. [However, this option is not quite debugged yet, it should be ready soon (tl/95/3/16)].

All the steps above should be done just before beam is put on target so that the array is perfectly aligned when the beam arrives. Use the dump utility periodically to check that the calibration is still good - if the beam is delayed.

Save the current calibration parameters so they can be reloaded if the EFFs have to be rebooted. The save file also documents what calibration parameters were used in the experiment.

3. On-line Doppler correction:

If it is a thick target experiment, the off-line calibration task is completed at this point - go on to the 'on-line calibration' section. However, if a thin target is used, then the following additional steps should be followed in order to also on-line Doppler correct the germanium detectors:

With the un-shifted source calibration parameters still in the EFFs, allow the beam on target and wait for a some statics to build up (you don't need much, this is singles data). Use the 'dump ring' utility in the gecal2 script to dump the sum of spectra in each of the 17 rings of the GAMMASPHERE array. Look for a 'moving' peak in the spectra: r17.spe, r32.spe, ..., r162.spe. (there is a gf2 command file: 'rdisp.cmd' to help with this task).

With gf2, find the mean position of a selected 'moving' peak in some or all of the ring spectra and input these values to the Doppler correction calculation routine in the gecal2 script. Specify '0.0' for the mean values of the peak in the rings you did not determine - the program will then ignore those rings in the calculation. By pushing the 'calc' button the v/c value is found. In the gecal1 script this v/c value must be specified next to the 'Doppler correction' button. By pushing the button, the current (source) calibration parameters are modified and the spectra are cleared. Note: Calculated v/c values are generaly not quite accurate enough. The above experimental determination is strongly recommended.

Save the Doppler corrected calibration coefficients so they can be reloaded in case the EFFs have to be rebooted. You should now be able to start the taping of the experiment. Use the 'dump rings' utility to make sure the Doppler correction went well - the moving peaks should be lined up perfectly.

4. On-line calibration (drift correction)

The gecal1 script contains a tool to correct for any drifts of the detectors during the run. After the beam has been on for a while (and after a possible Doppler correction); but before any detector can drift appreciably, a set of reference spectra must be stored. We will force the germaniums to perfectly match this master set of spectra, so it is wery important to make these reference spectra perfect. (Some potential problems are discussed here)

Use the 'INITIALIZE' button at the bottom of the script to do this. It is important to do this at the right time. We want to get enough data to have a good spectrum, yet, we do not want to wait so long that any detectors can drift. The reference spectrum is used to compare the spectra from the germaniums later in the run. If the current spectrum does not match, the calibration can be changed to readjust the detectors back to this 'perfect' condition (where both the calibration and v/c correction is unambiguously known).

Be careful not to hit the 'initialize' button during the run.

At any given time, the EFF germanium spectra can be cleared and a snapshot of the current spectra dumped. The 'dmp EFF spectra' procedure will write several command files for use in Radfords gf2 utility. The one appropriate for a check for any drifts is: 'ref.cmd'. Use 'cf ref' in gf2 to execute it. The command file will go through all the detectors you specified in the gecal1 script and one by one display the reference spectrum (in blue) and the current spectrum (in red) overlayed. Mark down the ones that need correction (if any).

Specify the detectors that need correction in the top of the gecal1 script. Also specify the gain and offset search grid (if not already done) and push the 'drift cor' button. There are two versions of this button, the blue one will do everything - except deposit the new calibration parameters (thus, a dry run). The detectors that were corrected will have their EFF spectra cleared - so a new dump of the EFF spectra will allow you to see how the correction went. If the new gain coefficients were inside the search grid, the correction should be perfect. If not, then it might be necessary to use the man_cal (manual calibration) function in the gecal2 script to correct for the gain/offset drift.


web master: torben@anl.gov