GLAST Calorimeter Software Weekly Reports
Week of 12 January 2004
NRL
Sasha: This week I tried to install in
my desktop computer the recent version of GlastRelease used for DC1 data sets
generation as well as EngeneeringModule package created for EM calibration data
analysis (both requiring VC7), while keep working the existing calibration
software based on older GlastRealese compiled with VC6. I have installed both
compilers, they are working without conflicts. The previously reported problems
were due to incompatibility if vcmt v17 with VC6: it causes this compiler to
crash. So I keep two versions of vcmt in my computer: vcmt v16r1 to work with
VC6 and v17 to work with VC7.
Plan for next week: to study the difference in Cal energy resolution between DC1
data and the simulations done at the end of 2002 ( the problem reported by Bill
Atwood and Steve Ritz).
Andrey: The last couple of weeks I've
been working on another
set of CAL simulations:
In particular, we had a meeting (Eric, Sasha and me) where
we reviewed and discussed first simulation results, for
0.3, 2, 10 GeV photons incident at 5, 25, 50 degrees on the
GLAST with four "fail options" for each energy/angle setting:
1) No fail
2) 2 randomly picked columns in the same (randomly
picked) layer per tower fail
3) 8 randomly picked columns per (random) "tower side" fail
4) 1 whole (random) layer per tower fails
As a result of discussion we decided to have one more
set of simulations, with smaller "incident area" (0.5 m^2
instead of original 6.0 m^2) to make sure that we get rid
of side-effects, when EM shower leaks out the CAL when
original photon hits the detector closer to the edge.
These side effects contribute to the growing tail in the lower
part of energy spectrum, which is hard to fit. Using smaller
incident surface (~ 4 most central towers) allowed virtuallyeliminate the tail in the simulated energy distribution.
Another thing we decided - is to add one more fail mode -
that is a "sum" of modes (3) and (4), and to add one more
energy setting, 50 GeV.
Right now I almost finished "production" of digi files for
all energy and angle settings mentioned above (using 0.5 m^2
hitting area).
The next thing I'll be doing is producing results for various
fail modes. Also Sasha made a suggestion to use a correction
on energy deposited in Tracker, when calculating the energy
in CAL. This is the next step I will work on.
Mark: Mostly admin. tasks (meeting prep and attendance, organizing weekly reports). Telecon with Eric G., Eduardo, Xin to discuss future EM, FM CAL ground calibration activities and requirements, future of EM analysis at SLAC. In addition, installed vc++ 7.0 in preparation to investigate heavy ion calibration techniques. Also started to upgrade to latest DC1 science software.
LLR
Hosting a CENBG-GAM-LLR-CEA GLAST meeting on Thurs-Fri where we'll discuss recon, GCRcalib, beam tests and other topics.
Pol: further
deepening his knowledge on calrecon/digi. He found that pedestals in the EM are
correlated within diode channels and between diodes from a same end. The
correlation can be well described by a bidimensionnal gaussian distribution
(with a main direction that has also to be defined) that he is now estimating.
This will require modifications in CalDigi to sample bidimensional distributions
for the pedestals from an xml file. Results in the next CAL SW meeting.
Also has been analyzing tomography measurements with Oscar on pre-flight
structures. Might show results one rainy day.
Berrie: looking at a
fraction of anomalous events in MC gammas where the reconstructed deposited
energy is much larger than the McIntegrated deposited energy and yields data
that have to be corrected in recon methods. This seems to occur in events close
to diodes and the diode giving a large pulse has to be eliminated in the sum,
and the signal that is valid on that CDE has to be corrected for attenuation.
Results in the next CAL SW meeting.
Looking at a correction method for energy estimations < 500 GeV. Results
in the next CAL meeting.
Bordeaux
Benoit, Johan: A
first-round calibration for the Minical data has been obtained, using the proton
and helium spectra. The ionisaton peaks have been fitted, and their energies
compared with the predictions from GEANT4, allowing the quenching magnitude to
be established. The charge-changing cross-sections have also been determined
experimentally. More comparison with the simulation results will be made next
week.
We had several face-to face meetings in Paris with the LLR and GAM colleagues
over the status and plans of the beam-data analysis.
Thierry: 3 days on duty on the CELESTE experiment, and compiling/testing the new JQMD interface with Geant4_v6. On this last point I got it running under rh7.2, not rh9. Now starting to test...
Montpellier (as reported by Benoit)
The GSI data analysis has started. The first steps towards the energy calibration have been performed, following a procedure similar to that used at Bordeaux for the Minical. Contribution to the calibration of the FRS data will start out next week.
SLAC
Xin: This week I tried to reconstruct energy of VDG gammas. I sent you some slides yesterday I think we need to correct threshold effects for CAL, especially at low energies.