Tamara
reported that she has completed the study for her HV corrections.
She produced plots of (Eh2inner/Eh2outer) vs. ((Eh1-Eh3)/(Eh1+Eh3)) - that's
the `asymmetry plot' - for 3 different energies. Only limited number
of runs are suitable to do produce this plot, and the data of 150 GeV pions
has been chosen as the sample best suited to provide the base for the
correction. A fit was applied to the asymmetry plot of this 150 GeV data.
The fit is of the form: p0 * exp (-p1 ( x+1 ) ^ p3) + p2.
Tamara applied this correction on an event-by-event basis to all runs in
her analysis and found that for the runs at higher energies the resolution
improves significantly, while the response is lowered by about 2% for all energies.
This correction is different from the correction in the NIM paper,
largely because Tamara uses fixed-size clusters, which are bound to
contain more noise than event-by-event clusters.
This leads to the next point of Tamara's efforts, which will be the
study of noise. She is planning to use muon runs for that.
At this point Michel explained in a bit more detail how Sven went about to
subtract the noise in his analysis, which is based on event-by-event clustering:
The cells numbers of all the cells included in the cluster are written to file
for each event. Then these cell numbers are read back and are applied to
form a `cluster' each event of a muon run. Over an entire run, this yields
the background due to electronic noise.
Michel
reported that he has implemented the ATLAS clustering algorithm, which
was originally written by Sven, into TBRootAna. At the moment it is
implemented for the 2-d case, and will be committed to the TBRootAna repository.
Michel has tested it out on a pion run and it works beautifully.
He explained to us the basic logic of this topological algorithm.
[TBRootAna now has the systemAlg ClusteringAlg which clusters in 2d or 3d.
See
recent email for more info, and visit the
TBRootAna example web page.
This procedure could help find local hot spots of hadronic showers in the finely segmented
electromagnetic part of the calorimeter.
Ashok
told us that he managed to compile the Athena 8.0 source code at CERN.
His next aim is to try and implement this on the local machines!
The Athena releases from CERN do not contain the source code.
Ashok has been in contact with David Quarrie about this.
The most common way to do the transfer would be via kserver,
but for that, CERN requires kerberos 4. However, at UVic we have
kerberos 5. The pserver access is not really recommended, because
is is not secure. David and Ashok will try to do the transfer via a
password-less ssh protocol. This seems to be a bit of spearheading
undertaking, and we are curious to see how it will work out.
There is also some talk of including the source-code as a tarball
in future releases.
Tayfun
has now investigated the position resolution in terms of phi rather than x.
He had found a distinct structure when plotting sigma(x_calo) vs. x_mwpc.
When he plots sigma(phi_calo) vs phi_mwpc the structure disappears, at least for
the high energies. However, this introduces the problem of
making use of the y-information of the wire-chamber, which is not extremely reliable.
(Richard suggested that a nominal y position could be used instead.)
When plotting (phi_mwpc - phi_calo) vs. (phi_mwpc), Tayfun still sees structure.
The effect cannot be explained by leakage of energy. A possible cause could
be shrinkage of the detector. In order to find out whether shrinkage
could be showing its effects, emails have been sent to Rob and Peter Schacht,
and Margret sent an email to Fares to find out whether the EMEC geometry
files for the cell z and eta positions he had sent her last summer were
for the warm or cold geometry. Fares as not yet replied, but Peter
sent a brief explanation of how some of the alignment in the beam has been
done and explained that there is one fixed point in the EMEC module
and all the shrinkage happens with respect to that one fixed point.
Tayfun also showed the position resolution as a function of Ebeam for various
impact points.
Margret
briefly reported that she now knows where in the LArG4TB code the
values for the energies in the cells are filled into the Ntuple
and the LArG4 code provides her with an example of how to tackle
the separation of electromagnetic and non-electromagnetic energy.
What exactly classifies `hadronic' energy is not a simple thing,
and Margret has contacted Fred Jones from the TRIUMF Geant4 group
to find out what sort of philosophy and thought has to go in making
these classifications.
Warren
was unable to attend the meeting but Richard contributed
an email from Warren about his activities for the minutes:
Hi Richard,
I wanted to send a quick email to let you know where I'm at as far a work
goes before I head out later this morning for Calgary.
- read many webpages/documents on Athena, CERN computing, HEC/EMEC
Testbeam, Atlas Unix sevices, tutorial transparencies, etc...
- received my scratch space on my CERN account and have been setting up and
checking out packages, and learning through playing around with the files.
- was able to check out packages relating to the beam chamber code. I've
started looking over the code and will try to spend some time this weekend
going over it as well.
- received an email from Rob with a LArMonitoring example this morning.
I've gone through it and did get it running, and I will look into some of
the code to see if I can figure out how it all works. (Michel also
received this email example).
I'll still be in email contact this weekend in case you have any
comments/suggestions/info to pass on from the group meeting.
Thanks again and see you next week.
Warren
Current known beam test related travel plans
Margret: 4 weeks in September
Michel: May 9 to May 22
Michel: June 13 to July 3
Michel: August 2 to August 14
Richard: June 19 to June 30
Tayfun: August 1 to August 29
Warren: May 16 to June 29