February 6, 2007 Software
- Review minutes from January 30, 2007 Software meeting
- FCAL status (Mihajlo)
- Action items
- Golden Channels (Alex)
- Rates in CDC layers added inside ring 1 (Richard)
- Update on the background normalization factor (Richard)
- Results on tagger background from electron beam halo (Richard)
Attendees: David L., Elke A., Simon T., Elliott W., Elton S., Mark I., Eugene C. On the phone: Mijhalo K., Richard J., Alex D.
Mihjalo reported on some recent work regarding the FCAL reconstruction. The work focussed on the rate of multi-cluster backgrounds generated from single photon events. These tend to be generated from "pe-shower"-ing off of another part of the detector upstream. It appears to come primarily from the FDC cables which result in a large increase of multi-cluster events at around 9-11 degrees.
A linearity correction has been applied to remove most of the descrepancy between the reconstructed and thrown cluster energies. This descrepancy can be seen in the plots shown at the December 11, 2006 Software meeting. With the energy correction, the distribution appears to have a floor term of almost 5%. Alex pointed out that Rad-φ had a floor term closer to 3.2%. Richard suggested another non-linearity correcting iterator algorithm may not be enabled in the current reconstruction. Mihjalo will look into it.
For next time, Mihjalo was going to look into doing the following:
- Throw photons distributed evenly over solid angle rather than just angle.
- Look at the splash profiles of the multi-cluster events. In particular:
- How close are the clusters (can we merge them?).
- How is the energy distributed between the clusters.
<deferred until next week>
Alex gave a brief report on a discussion he'd had with Elton regarding the "golden channels" we should look at with the simulation. The channel looks to be a good choice since it is a close match to the expected decay path for hybrids.
For the purposes of the upcoming reviews, they thought of combining data from the following 2 channels:
The cross-section for the channel is known to be 9 times smaller than for the . By combining data, we can show that the detector resolution is good enough to distinguish the from the and it will use both charged and neutral particles.
Rates in CDC layers added inside ring 1
Richard gave a report on a brief backgound rate study he did where 4 additional CDC layers were put into the HDGeant geometry. The rates are shown in the picture below. The yellow area represents the new layers while the white are the layers from the nominal design. Note that in this simulation, the start counter was removed since it overlapped in space with the new layers. The rates in the nominal design look to have increased slightly due to the larger amount of material in the new layers as compared to the start counter.
Update on the background normalization factor
Richard reported that he found a bug in the simulation of the electron beam source in the upstream simulation program (not hdgeant). This program is used to find the factor by which the rates obtained using the built-in coherent bremstrahlung generator in hdgeant must be scaled in order to get the true rates. This factor is nominaly 1.1 for 10^7 running (see src/programs/Simulation/HDGeant/control.in). After the bug fix, the simulation was rerun to see if it affected the rates in GlueX recently obtained from Monte Carlo. The rates were not affected by the change.
Results on tagger background from electron beam halo
Richard reported that he is currently writing a note detailing the results of a study he has done on the beam halo. In this study, Richard has created a model of the beam halo using data generated by ray tracing by the CASA group. With the model, he was able to estimate the rates in the tagger (both microscope and whole hodoscope) due to the beam halo. These indicate the halo-induced rates are below the 1% level. It was noted however, that the CASA simulation stops tracking particles once they get a certain distance from the beamline. It may be that particles generated from scattering off the beamline will somehow contaminate the beam in the form of addition halo which may force the background tagger rates to exceed 1% of the primary rate.
- Freezing tags in subversion through hooks (David)
- Put tarballs of external packages on Wiki(David)
- SVN web interface for browsing change logs (David)
- Design and implement exception and logging scheme for JANA
- Discuss integration issues
- Experiment with shared pointers