Status Report of PRS/µ
D.Acosta
University of Florida
Current U.S. activities
PRS/µ Activities
New PRS organization
USCMS Physics, May 2001
1
Darin Acosta
EMU Software Workshop
Workshop held at UCDavis in late February helped
focus EMU software and physics activities
Brought together software and electronics experts to close
the loop on ORCA validation
è Reviewed detailed CMS Notes describing CSC digitization
and trigger simulation
è Found that much of the simulation is right!
è Also found that some of the modeling is naïve (noise &
crosstalk)
è Scrutinized neutron background parameterization
p Some problems discovered (and later fixed…)
è Discussed future areas of work
p Improved neutron parameterization
p Geometry description & DB, detector misalignment,…
è Plenty of opportunities for new collaborators
è
USCMS Physics, May 2001
2
Darin Acosta
Neutron Background Simulation
Biggest factor affecting CSC performance
High rate of large amplitude pulses seen in GIF data
è Can overlap and distort CSC pulses from muons
p spatially and temporally
è e.g. LCT trigger logic builds in 4-fold coincidence to reduce
background rate
è
Current simulation:
UCDavis parameterization based on CMSIM 112 creates
GEANT hits (JUNK flag), which in turn are fed to ORCA
è New simulations done by UCDavis based on CMSIM 121
corroborate earlier results
è However, a bug was found and subsequently fixed by Tim Cox
affecting distribution of hits
è Nicely documented here:
p http://ucdcms.ucdavis.edu/~ptc/Cmsim/Junk/mc_junk_bug.html
p Tim’s CPT plenary talk
è Impact on L1 CSC trigger studied:
è
USCMS Physics, May 2001
3
Darin Acosta
CSC Event at GIF
USCMS Physics, May 2001
4
Darin Acosta
L1 CSC Trigger Efficiency
Increasing
background
Should crosscheck absolute
level of
background
Also, what is
effect on HLT
and offline?
USCMS Physics, May 2001
5
Darin Acosta
PRS/µ News
MC production
~1.2M events (high lumi) + 0.3M events (low lumi) currently in
production in Italy for L2/L3 studies
è Problems running ORCA 4.4.0 (or 4.5.1) analysis jobs on digis
è Expect ~2M events for 2002–2003
è
Group Activities
è
è
è
è
è
Optimum weighting procedures for high MC statistics
Punch-through and neutron background studies
Conversion to OSCAR
L2/L3 studies:
p Muon reconstruction, muon-track matching, various muon
propagators being developed
Standard analysis package being developed:
p MuonReco/MuonAnalysis has Ntuple code
New Organization
è
Looking for additional manpower and L3 coordinator names
USCMS Physics, May 2001
6
Darin Acosta
L2 Efficiency
Losing
efficiency
in DT/CSC
overlap
w.r.t. L1
(need RPC)
L2 uses L1 candidates as seeds, refines PT measurement
USCMS Physics, May 2001
7
Darin Acosta
L2 PT Resolution
Compare
to L1:
σ ~ 23%
High PT
tail in
DT/CSC
overlap
USCMS Physics, May 2001
8
Darin Acosta
L2 Trigger Rate
Most rate comes from low PT muons, which are already isolated!
USCMS Physics, May 2001
9
Darin Acosta
L2 Reconstruction in Overlap
Low PT muons reconstructed as high PT in DT/CSC overlap
USCMS Physics, May 2001
10
Darin Acosta
Partly caused by lost Segments
Lose MB1 segment in L2
Track-Finding algorithm
⇒ Poor PT resolution
Should re-tune cuts
But what about tracks
without any MB1
segment? They are kept
in L1, but will HLT be able
to do anything with them,
and will we be able to do
physics with them?
USCMS Physics, May 2001
11
Darin Acosta
L3 Muon Trigger Studies
Neumeister, Dorigo, Loreti, Bellucci
Start with Muon-Tracker matching
è
è
“Connection Machine” builds seeds in tracker
L2 provides seeds in muon system
Propagate from muon system to outermost tracker layer
è
è
GEANE, or
Fast Muon Propagator (Olga Kodolova)
Some work on developing common Muon-Tracker
navigation methods
Lacaprara
è
Builds on propagators developed for Tracker
USCMS Physics, May 2001
12
Darin Acosta
Muon-Tracker Matching Efficiency
W → µ with pile-up, old tracker geometry
Leonardo Bellucci
USCMS Physics, May 2001
13
Darin Acosta
L3 PT Resolution with Tracker
W → µ with pile-up, old tracker geometry
Leonardo Bellucci
USCMS Physics, May 2001
14
Darin Acosta
New Structure of PRS/µ Group
Level-3:
Detector
Hit
Simulation
Response
Simulation &
Reconstruction
CMSIM→OSCAR
Muon detector
conversion
simulation and
(Geant3 → Geant4) reconstruction in
ORCA
Liaison to SPROM
Liaison to RPROM
Alignment,
Calibration,
& Database
Development
of geometry
and calibration
DB, impact of
misalignment
Liaison to DB
and detector
groups
Ugo’s manpower estimates:
3 FTE
4 FTE
2FTE
4 FTE
Most likely underestimated
USCMS Physics, May 2001
15
Online &
Control
Software
Testbeam data
management,
analysis tools,
online monitoring
HLT &
Physics
Objects
HLT selection
for L2, L3
(isolation, mip,
track matching)
What about
Analysis tools
commissioning &
What previous
standalone DAQ?
HLT group was
How much of
for
muon/trigger
2 FTE
control software
done in PRS?
Liaison to detector
groups
Darin Acosta
Some Possible Contributions
All the boxes shown, and the tasks listed in the PRS
section of the CMS SW&C Deliverables document
need help. So take your pick!
If you want ideas, here are some:
Conversion from G3 to G4. Pressure to do this by year’s end
(How much will TeV muons bremsstrahlung?)
è Development of L3 algorithms for DAQ TDR
(no U.S. involvement so far)
è Normalization of flux rate estimates to CSC hit rates
è Validation of CSC detector simulation to data
(cosmic rays @ FAST site & testbeam)
è Integration of FAST site analysis tools and event display into
ORCA and Visualization packages
è Misalignment studies. Introduce offsets in reconstruction
and simulation and study consequences (corrections)
è Development of FAST simulation for Physics TDR
è Development of procedures for calibration, and how to apply
to reconstruction
è
USCMS Physics, May 2001
16
Darin Acosta
Conclusions
Much more is being demanded of the PRS groups than
ever before
èFull
chain from detector control and data storage to offline
reconstruction (including trigger and simulation) is included
We’re still learning what some of these demands are!
Trying to define them for the CMS Software and Computing
Tasks and Deliverables document
è
This will take the work of the entire collaboration
Almost everything outside of detector construction and core
software is covered by the PRS groups
è
There is no shortage of areas to work on!
USCMS Physics, May 2001
17
Darin Acosta