The Level-1 Endcap Muon Trigger
D. Acosta
University of Florida
Outline
• Requirements
• Organization
• Design status
• Test beam results
• Software status
• Simulation results
Requirements
• High efficiency with as low a PT threshold as possible
• Single muon trigger rate < few kHz at L=1034 cm-2s-1
– Should have large safety factor
• PT resolution ≈ 20%
– Require ϕ, η information from 3 muon stations because
of the small and non-uniform magnetic bending
– Require ϕ precision ~1 mm ( ~0.1 CSC strips )
• Multi-muon capability
– ≤ 3 muons per 60° azimuthal sector
– Best 4 muons overall sent to Global Level-1 Trigger
• Rejection against beam halo muons, cosmics,
punch-through
• Correct B.X. assignment
• Programmable
– Trigger algorithms must be tunable
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
2
Muon Rate dN/dηdt (Hz)
Effect of Resolution on L-1 Rate
10 7
10
L = 1034 cm-2s-1
6
Resolution
10 5
10 4
50%
10 3
40%
10 2
30%
10
1
10
10
CSC 2-station resolution
-1
-2
20%
10%
0%
PYTHIA6
1
10
10
2
PT Threshold (GeV/c)
Require PT resolution ~20% so that rate does not plateau
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
3
Level-1 Trigger Scheme
Strip FE cards
Strip LCT card
LCT
CSC Track-Finder
Motherboard
Port Card
Sector Receiver Sector Processor
OPTICAL
FE
SR
TMB
PC
2µ / chamber
3µ / port card
SP
LCT
FE
Wire FE cards
3µ / sector
Wire LCT card
36µ
In counting house
CSC Muon Sorter
RPC
On chamber
On periphery
4µ
DT
4µ
4µ
Global µ Trigger
Global L1
4µ
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
4
Coverage of CSC Trigger
• CSC trigger primitive coverage for η > 0.9
• CSC Track-Finder coverage for η > 1
• DT/CSC overlap from 1 < η < 1.2 covered by CSC
Track-Finder
• DT/CSC overlap from 0.8 < η < 1 covered by DT
Track-Finder
Drift Tubes
MB/1/4
MB/2/4
ME/1/3
YB/0/3
MB/1/3
MB/0/3
YB/2/2
YB/1/2
YB/0/2
MB/2/2
YB/2/1
MB/1/2
YB/1/1
MB/0/2
YB/0/1
MB/2/1
MB/1/1
MB/0/1
ME/1/2
YE/2
5.975 m
4.905 m
4.020 m
2.950 m 2.864 m
2.700 m
HB/1
1.9415 m
HE/1
1.711 m
HF/1
1.811 m
EB/1
EE/1
YE/1
3.800 m
CB/0
ME/1/1
ME/2/1
η = 3.0
MB/2/3
7.380 m
7.000 m
MB/0/4
YB/1/3
YB/2/3
ME/2/2
ME/3/2
ME/3/1
ME/4/1
YE/3
10.86 m
ME/4/2
η = 1.479
η = 2.4
η = 0.5
η = 1.1 η = 1
Cathode Strip
Chambers
1.290 m 1.185 m
SB/1
0.440 m
η = 5.31
0.00 m
0.000 m
2.935 m
3.90 m
4.332 m
5.68 m
6.66 m
6.45 m
7.24 m
8.495 m
9.75 m
10.91 m
10.83 m
10.63 m
14.53 m
14.96 m
14.56 m
SE/1
CMS - PARA- 003 - 14/10/97
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
5
PP
/pg/hr
1998 Trigger Performance
Prototype Modules
Comparator
ASIC
Comparator
Board
Trigger
Motherboard
Anode LCT
Cathode LCT
CSC Chamber
TriDAS Week at CERN, Nov. 1999
2
Anode LCT Bunch Timing vs
vs..
Background Rate
Frank Petriello
• Nominal max.
wire rate 20kHz
• Observe eff >99%
up to ~140kHz
TriDAS Week at CERN, Nov. 1999
3
Cathode LCT Resolution
Si telescope tracks
96.0% “High-Pt” tracks:
Resolution
(half-strips)
3.7% “Low-Pt” tracks:
Position
(di-strips)
0.25% no Cathode LCT:
Position
(di-strips)
TriDAS Week at CERN, Nov. 1999
Resolution
(di-strips)
Alon Attal
Net CLCT
efficiency: 99.75%
4
LCT Efficiencies
From 1998 Test Beam
• ALCT efficiency:
• 98.7+-0.2% for exact match with TDC position
• 99.5+-0.1% for +-1 wire match
• CLCT efficiency:
• 85.4+-0.6% “perfect”: high-Pt pattern and half-strip match
• 93.1+-0.5% if add high-Pt patterns +-1 halfstrip. Some of these
may have CLCT correct but SCA wrong (noise, delta rays, peds.)
• 97.0+-0.3% if add low-Pt patterns with di-strip match
• 97.1+-0.3% if add low-Pt patterns +-1 di-strip match
• 99.2+-0.2% with addition of “bad matches”, mostly high-Pt
patterns with 2 half-strips separation
TriDAS Week at CERN, Nov. 1999
5
Total LCT Efficiency
• High-momentum (>10 GeV) muons, using 1998 test beam
results
• Lower limit of total LCT efficiency:
• Take no correlations between mismeasurements
• 0.970[cathode]*0.995[anode]*0.98[coincidence]*0.99[BX ID]=93.6%
• Anticipated improvements:
• 0.992[cathode pattern] expected with some tuning
• 0.99[coincidence] with addition of LHC phase correction
• 0.992[cathode]*0.995[anode]*0.99[coinc]*0.99[BX ID]=96.7%
TriDAS Week at CERN, Nov. 1999
6
LHC Conditions
• Simulated by
Gamma
Irradiation
Facility
photons
• Cathode data
from CSC DAQ
• 6 layers x 80
channels x 16
time samples
TriDAS Week at CERN, Nov. 1999
Bryan Smith
9
CLCT99 High-Rate Results
200 (K=5) is highest CSC
rate expected
Lisa Gorn
LCT Layers
CLCT time
resolution (ns)
Missing:
1.8% K=5
5.0% K=2
Time walk (ns)
Single layer:
Position within
1/2-strip:
•Evts with LCT
98% ~flat
Within +-1 half-strip
Exact half-strip
N.B depends on
beam chamber
accuracy:
~1mm (0.2 half-strip)
TriDAS Week at CERN, Nov. 1999
10
Pt-AssignTF-C++
Comparison of PT
30
25
20
15
10
5
0
0
5
10
15
20
25
30
Entries
Pt-AssignTF-Fortran
10 4
10
3
10 2
10
1
-15
-10
-5
0
5
10
15
Pt-AssignTF-C++ - Pt-AssignTF-Fortran
Difference due to slightly different Track Assembly
procedures for multiple trigger primitives in one station
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
11
L1 CSC Simulation Results
• CSC software in ORCA is not fully validated yet
– This is being addressed now
• C++ Track-Finder is shown to be equivalent to Fortran
version
• Therefore, use trigger primitives from CMSIM and a
standalone Fortran Track-Finder to extract background
rates and single muon efficiencies
• Limitations:
– Does not test system in ORCA
– Trigger primitive simulation in CMSIM (and ORCA) has
no timing simulation
– Rate results will be “conservative”
• Florida is producing large MC samples for Muon HLT
studies, so switch back on digitization for some
samples
– Samples generated using same parameters as for HLT:
Pythia and CMSIM116
– Use unweighted Pythia Min Bias events, or events
weighted by PThard
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
12
Halo Muon Rejection
• Track-Finder logic rejects extrapolations in which
two track segments are not within an η road
projecting to the interaction point
Z
• Logic also rejects extrapolations in which two track
segments are both parallel to the beam axis
• Out of a sample of 9142 halo muons simulated in
CMSIM, only 1 passed the Track-Finder cuts
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
13
Single Muon Efficiency vs. PT & η
Eff
Single Muon Track Finding Efficiency (OL + EC)
1
0.75
0.5
Pt = 3 GeV/c
0.25
Eff
0
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
ηgen
1
0.75
0.5
Pt = 5 GeV/c
0.25
Eff
0
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
ηgen
1
0.75
0.5
Pt = 50 GeV/c
0.25
0
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
ηgen
Track-Finder algorithm efficiency only
Excludes chamber and trigger primitive inefficiency
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
14
Track-Finding Efficiency vs. ϕ
Eff
1.05 < η < 1.2
1
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
60
Mod(Φgen,60) (deg)
FSU track (Phi Gen)
Eff
η > 1.2
1
0.8
0.6
Pt = 50 GeV/c
0.4
0.2
0
0
50
100
150
200
250
(Pt 50 GeV η gt 1.2) FSU track (Phi Gen)
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
300
350
Φgen
15
PT Measurement from 3 Stations
∆Φ23 (deg)
Pt=3 GeV
Pt=5 GeV
Pt=7 GeV
Pt=10 GeV
5
4
3
2
1
0
-1
-2
-3
-4
-5
-2
1.6 < η < 2.0
0
2
4
6
∆Φ23 vs ∆Φ12
8
10
∆Φ12 (deg)
• Low PT muons are well separated from high PT muons
• Parameterize vs. η, ∆ϕ12, ∆ϕ23
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
16
1 / PT Distribution
9000
Entries/GeV
-1
Pt Assignment (for Pt=3 GeV 1.6 < η < 2.0)
mean=0.340851
8000
7000
sigma=0.0596905
6000
5000
4000
3000
2000
1000
0
0.1
0.2
0.3
0.4
0.5
0.6
-1
1/Pt (GeV )
1/Pt assign Pt=3 GeV
Entries/GeV
-1
Pt Assignment (for Pt=10 GeV 2.0 < η < 2.4)
45000
mean=0.101545
sigma=0.0325276
40000
35000
30000
25000
20000
15000
10000
5000
0
0.05
0.1
0.15
0.2
1/Pt assign Pt=10 GeV
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
0.25
0.3
-1
1/Pt (GeV )
17
σ( 1/Ptrec - 1/Ptgen )/( 1/Ptgen )
PT Resolution
0.7
Pt = 5 GeV (2 Stn)
(ME1-ME2)
0.6
0.5
ME1/3
MB1
Pt = 5 GeV (3 Stn)
(ME1-ME2-ME3)
ME1/2
ME1/1
0.4
0.3
0.2
0.1
0
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
ηrec
Note: ME1 is required, otherwise resolution is ~70%
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
18
Single Muon Trigger Acceptance
Efficiency
2-Station PT Assignment with 3 or 4 CSC Stations
1
1.2<η<2.4
0.8
0.6
4 CSC Stations
3 CSC Stations
Pt cut = 10 GeV/c
0.4
Pt cut = 20 GeV/c
Pt cut = 30 GeV/c
Pt cut = 40 GeV/c
0.2
Pt cut = 50 GeV/c
0
0
20
40
60
80
100
Pt (GeV)
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
19
Single Muon Trigger Acceptance
Efficiency
3-Station PT Assignment with 4 CSC Stations
1
1.2<η<2.4
4 CSC Stations
0.8
0.6
Pt cut = 10 GeV/c
0.4
Pt cut = 20 GeV/c
Pt cut = 30 GeV/c
0.2
Pt cut = 40 GeV/c
Pt cut = 50 GeV/c
0
0
20
40
60
80
100
Pt (GeV)
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
20
Single Muon Trigger Acceptance
Efficiency
3-Station PT Assignment with 3 CSC Stations
1
1.2<η<2.4
0.8
3 CSC Stations
0.6
Pt cut = 10 GeV/c
0.4
Pt cut = 20 GeV/c
Pt cut = 30 GeV/c
0.2
Pt cut = 40 GeV/c
Pt cut = 50 GeV/c
0
0
20
40
60
80
100
Pt (GeV)
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
21
Min Bias Trigger Study
Entries
500K unweighted Pythia Min Bias events with Pthard > 0
put through CMSIM116 and Track-Finder simulation
200
175
150
125
100
75
50
25
Entries/GeV
0
10
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
ηrec
2
10
1
10
-1
0
20
40
60
80
100
Pt assign
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
120
140
Pt (GeV)
22
L1 CSC Single µ Trigger Rate
(per unit rapidity)
760K unweighted Pythia Min Bias events, Pthard = 0,
with pile-up, put through the Track-Finder simulation
Rate is for 1034 luminosity assuming σ = 55 mb
Rate dN/dηdt (kHz)
Single µ Rate (Min Bias sample 14 collisions in 1 BX)
10 4
3-Stn Pt 4 CSC Stns
3-Stn Pt 3 CSC Stns
10 3
2-Stn Pt 4 CSC Stns
2-Stn Pt 3 CSC Stns
10 2
|η| > 1.2
10
1
10
10
-1
-2
34
-2 -1
L = 10 cm s
10
-3
1
10
10
2
Ptmin (GeV)
Target rate of 1 kHz for single muons is achieved with
3-station sagitta measurement for a threshold of ~15 GeV
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
23
L1 CSC Single µ Trigger Rate
Restricted η Coverage
Rate dN/dηdt (kHz)
Single µ Rate (Min Bias sample 14 collisions in 1 BX)
10 4
3-Stn Pt 4 CSC Stns
3-Stn Pt 3 CSC Stns
10 3
2-Stn Pt 4 CSC Stns
2-Stn Pt 3 CSC Stns
10 2
1.2 < |η| < 2.1
10
1
10
10
-1
-2
34
-2 -1
L = 10 cm s
10
-3
1
10
10
2
Ptmin (GeV)
Significant rate reduction by excluding the high η region
Lowers threshold to about ~10 GeV
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
24
L1 CSC Single µ Trigger Rate
No Pile-up
260K Pythia Min Bias events with Pthard > 7 GeV
Weighted samples in Pthard provide better statistics
at high PT
Rate dN/dηdt (kHz)
Single µ Rate (Min Bias sample)
10 4
3-Stn Pt 4 CSC Stns
3-Stn Pt 3 CSC Stns
10 3
2-Stn Pt 4 CSC Stns
2-Stn Pt 3 CSC Stns
10 2
|η| > 1.2
10
1
10
10
-1
-2
34
-2 -1
L = 10 cm s
10
-3
1
10
10
2
Ptmin (GeV)
Slightly higher rates predicted than Pthard=0 samples
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
25
L1 CSC Single µ Trigger Rate
No Pile-up, Restricted Coverage
Rate dN/dηdt (kHz)
Single µ Rate (Min Bias sample)
10 4
3-Stn Pt 4 CSC Stns
3-Stn Pt 3 CSC Stns
10 3
2-Stn Pt 4 CSC Stns
2-Stn Pt 3 CSC Stns
10 2
1.2 < |η| < 2.1
10
1
10
10
-1
-2
34
-2 -1
L = 10 cm s
10
-3
1
10
10
2
Ptmin (GeV)
3-station sagitta measurement gives large safety factor.
2-station measurement is acceptable at 1033 luminosity
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
26
L1 CSC Double µ Trigger Rate
(per unit rapidity)
760K unweighted Pythia Min Bias events, Pthard = 0,
with pile-up, put through the Track-Finder simulation
Rate is for 1034 luminosity assuming σ = 55 mb
Rate dN/dηdt (kHz)
Di-µ Rate (Min Bias sample 14 collisions in 1 BX)
10 4
3-Stn Pt 4 CSC Stns
3-Stn Pt 3 CSC Stns
10 3
2-Stn Pt 4 CSC Stns
2-Stn Pt 3 CSC Stns
10 2
|η| > 1.2
10
1
10
10
-1
-2
34
-2 -1
L = 10 cm s
10
-3
1
10
10
2
Ptmin (GeV)
3-station measurement significantly reduces rate.
No di-µ triggers survive η < 2.1 cut for both muons
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
27
Conclusions
• Current CSC trigger prototypes worked and gave
good results at testbeams
• Design of future prototypes is going well
• Should have full system test through to the TrackFinder by next summer
• The CSC trigger shows good background rejection
– Surviving candidates are real muons
– PT resolution drives the trigger rate
• Low luminosity:
– Can efficiently trigger with 3 CSC stations using a
2-station PT measurement. Threshold would be
acceptable at L=1033 for η< 2.1
• Low luminosity, high η:
– Efficient triggering at L=1033 for η> 2.1 requires
4 CSC stations and a 3-station PT measurement
– Need ME4/1
– Might be useful for b-physics acceptance
• High Luminosity:
– Efficient triggering at L=1034 with acceptable
thresholds requires 4 CSC stations and a 3-station PT
measurement for all η
– Need ME4/1 and ME4/2
D. Acosta, University of Florida
TriDAS Review, November 9, 1999
28