Minutes of the UVic testbeam meeting on July 19, 2002
---------------------------------------------------------
(present: Ian G., Tayfun, Tamara)

Ian has finished the AFS install, with clients on chimera and dragon. 
 The 4.0.0 release of the ATLAS software has been mirrored, but there 
are problems with it that are beyond our control.  The 3.2.0 release is 
on troll, including the data challenge, testbeam, and Athena software. 
 The data challenge software is also installed on dragon.  Ian has been 
working on getting Athena to work on dragon, but there are many problems 
impeding his progress.  He has been investigating installing it on muse 
instead.

Tayfun has started to look at the electron data.  He has done some 
clustering and found alpha_em using the chi-square method.  He has 
decided that the 6-cell cluster is optimal.  Plots can be found at 
http://particle.phys.uvic.ca/~tince/electron.html.  Tayfun will be 
looking at using different sized clusters for different energies.  Some 
discussion rose as to the pros and cons of this method.

I have made E/E_0 plots for the clusters I used to find mu and sigma as 
a function of cluster size.  These can be found at 
http://particle.phys.uvic.ca/~starke/response.html.  Clusters 5a and 5b 
provide fairly good results, with five of the seven data points with the 
1% range.  I will be trying the chi-square method of calculating 
alpha_em after this.  

Tayfun and I have found some problems with the way that 
PlotAllCellsAlg.cxx calculates the mean histograms for different cells. 
 If the noise is very large (for example, in cells 296 and 300 of some 
of the electron runs, 11380 (100 GeV) in particular) the mean is 
calculated to be very large, despite the roughly equal number of 
negative signals to positive signals.  I have taken out the cuts on the 
data, but it still seems to be calculating the means too high.

Ian plans to continue working on the software and trying to get it to 
work better on more systems.  Tayfun will be trying different cluster 
sizes for different energies to see how to achieve the best results.  I 
will be investigating the way in which root calculates the mean of 
histograms, and will hopefully get rid of the problems with the means.

Tayfun asked why E/E_0 for lower energies (ie. 20 GeV) was generally 
smaller than for other energies.  I thought that this had been because 
the lower energy particles lost a greater fraction of their energy in 
the metal plates between the LAr gaps than higher energy particles, but 
we would like to get a more definite answer (from someone who knows what 
they're talking about!)

We also we wondering what the highest energy of a particle would be that 
the HEC would see once the LHC was in operation.  If anyone knows 
offhand, let us know!