GLAST Calorimeter Software Weekly Report
Week of 26 April 2004

NRL

Sasha:  Preparation for APS talk.  Double checking muon calib for "antiquenching".

Andrey:  No GLAST work

Mark:  Short week.  Continued work on I&T testing and calibration document.  Worked on task list.

Zach: Finished getting GR working, started to look at CalibGenCal

LLR

Pol:  this last week, I finalised my note on calibration.

Berrie:  See presentation on recent recon work

Bordeaux

Thierry, Benoit, Johan: our weekly report is the following:

• we installed the rh9 GlastRelease (version v4r2) on our machines and incorporated the EPAX hadronic model;
• in the GSI analysis area, we found a bug in the generation of the root TTree used in the analysis, for the runs with "incomplete" data readout, ie. with zero suppression and/or 1-range data readout. This bug appears to be responsible for the problems encountered for these runs as reported during the face-to-face meeting.
• Susanna Svensson, an undergraduate student from Sweden, has joined our group to work on the analysis of the CERN data.

GAM

Agnieszka and Eric:

• study of the impact of the selection criteria on the MIP peak position for C,N,O,F ions,
• stability of the MIP peak position in time for C,N,O,F ions,
• study of the selection criteria for heavy ions.

Frederic: continuation of last week (GSI runs with "auto-range" and/or "zero-suppression")

SLAC

Tracy:  See presentation on recent clustering work.

Additional Notes from 5 May 04 CAL S/W VRVS

Benoit reports that pulse timing is probably not the cause of the "antiquenching" phenomenon.  Shaping time on the minical is significantly different from that on the EM, yet the same effect is observed

Analysis of stopping Ni beam (1 GeV/nuc) disagreement with simulation may be due to upstream material not being simulated

Bill emphasized that at 100 MeV incident, thin conversion, the shower incident on the CAL looks like a "rain" of secondary photons.  This results in more than predicted losses in cracks.  Accounting for this may correct the anomaly in hits/energy when comparing TKR hits in thin and thick parts of TKR.  This may also be the source of 20% bias in CAL recon.  He has included corrections in his latest CalValTools

Bill also emphasized the importance of accounting for vertex location in low energy hit correlation.  Applying this to Berrie's existing low energy hit correlation algorithm (as Berrie has suggested here) may result in an improvement in resolution.