L0 trigger and related detectors

1
L0 trigger and related detectors

• Alessia Satta
• Universita di Roma
• on behalf of the collaboration

LHC2003 International Symposium Fermilab, 3 May
2003
2
Input figures at L0

• Bunch crossing frequency 40MHz
• Non empty bunches 30MHz
• 80mb of non elastic interactions
• 60mb in the acceptance of the spectrometer
• sbb/sin 6x10-3
• nominal luminosity 21032 cm-2s-1
• 8 (1.7) MHz of single (double) interactions
• GOAL L0 output rate 1MHz

3
Strategy
B decay signature high PT (ET) particles m e h
g p0
L0 Pileupveto reduces rate to 9MHz. L0 CALOMUON
must provide reduction factor9 gt medium Pt cuts
ETh 3.5 GeV, ET? 3 GeV, PTµ 1.2 GeV
4
High Pt signature
Pion transverse momentum (MeV/c)
5
Calorimeter detectors

• SPD PS 15 mm scintillating detectors
  interspersed with
• 2.5X0 lead
• Electromagnetic cal shashlik 2mm lead 4mm
  scintillator - 25Xo
• Hadronic cal. iron scintillating tiles
• - 5.6 lI

6
Calorimeter (II)

• SPD/PS/ECAL 3 zones
• Cell
• 40.4 / 60.6 / 121.2 mm
• The smallest cell size Moliere radius
• - s(E)/E10/vE 1.5
• 5952 channels each
• HCAL 2 zones
• Cell
• 131.3 / 262.6 mm
• - s(E)/E80/vE 10
• 1468 channels
• SPD/PR/ECAL/HCAL fully projective - HCAL
  granularity doesnt match the others

7
Calorimeter trigger principles

• Goal select the candidate of h, e, g, p0 with
  highest Et
• shower has a 'small' size ( contained in 2x2
  cells)
• search for a high energy releases in 2x2 tower
  in ECAL and HCAL
• in each calo FE (4x8 cells) card the highest
  candidate is selected
• process further only these candidates
• Reduced complexity and cabling 200 candidates
  for ECAL and 50 for HCAL starting from 6000 and
  1500 cells.
• e, g local candidates validation
• Electromagnetic nature of ECAL maximum is
  validated using the PreShower , charge using
  the SPD

8
Calorimeter trigger principles cntd

• Hadron local candidates validation
• ideally add the energy lost in ECAL in front of
  the candidate
• expensive different granularity gt complex
  connectivity
• useful only if the ECAL contribution is important
• look only at ECAL candidates !
• Manageable number of connections
• The Calorimeter gives also global information to
  the trigger
• total ET in HCAL gives interactions trigger
  (reject elastic, diffractive, m-halo)
• hits multiplicity in SPD potentially useable to
  reject too crowded events

9
Performance of L0Calo
Assuming a trigger rate of 600kHz for h, 100kHz
for e , 25kHz for g L0Calo efficiency () for
events selected by offline analyses
e g h
B-gtpp 9 3 55
Bs-gtDs K 5 2 37
Bs-gtJ/?(ee)f 36 4 24
Bs-gtKg 28 47 30
All triggers important !!!
10
Muon system

• 5 stations with calorimeter and iron shielding
  between them
• Technology MWPC with 4 ORed gas gaps (2 in M1)
• 1380 chambers
• Efficiency gt 99 per station
• Total absorber lI 20 gt minimum momentum 8GeV

11
Muon system

• 4 Regions, with different pad granularity
• Y full projectivity
• Pad dimension
• Min 6.3x31.3 mm2
• Max 25x31 cm2
• optimized for constant PT resolution
• 55k pads combined in strips-gt 26k channels to
  L0/DAQ

12
L0 Muon basic principle

• Search tracks in M1-M5
• 192 projective towers in parallel
• Required hits in all stations
• Assuming origin interaction point
• Exploit B-kick to calculate PT (magnet PT kick
  1.2 GeV/c)

• up to 8 m candidates
• 2/quadrant with
• highest pT

13
Performance

• PT resolution 20
• High efficiency
• Very robust against high background level in the
  detector
• Halo muon negligible in nominal conditions

Nominal bkg Max bkg (nominalx5) Max bkg Max halo muon
B-gtmX () 46 41 36
B-gtJ/y(mm)Ks 90 83 79
Neutron induced background
Halo muon x10 0.1/x-ing
Normalized to events with m in Muon system
14
PileUp veto detector

• 4 R-sensor half detectors upstream of
  interaction region
• Coverage -4.2lt ? lt-2.9
• Sensors active area 8mmltRlt42mm
• Pitch 40µm to 103µm
• 45o sections
• OR of 4 neighbouring strips
• 2048 channels towards L0

PileUp stations
Half station
15
Pile Up veto motivation

• LHCb designed for single interactions
• Easiest to reconstruct and tag
• More robust input for L1 and HLT
• Multiple interactions fill bandwidth of L0 ( 2x
  probability to pass L0).

16
Working principle of PU veto
True combinations
All combinations
RB cm
RA cm
RA cm
If hits are from the same track
ZPV cm
build a ZPV histogram, search highest peak, to
remove combinatorial background mask the hits in
the peak , repeat the algo , find a second peak
(signature of multiple interactions)
17
Performance
B-gtpp
Minimum bias
Height of second peak
If cut of second peakgt3 retain gt98 of single
and reject 60 of multiple
Height of second peak
possible to populate the 1 MHz with preferably
single interactions
18
L0 hardware implementation

• Custom electronics using commercial components
• Synchronous system and pipelined
• No dependence on occupancy and on history
• Latency 4.0ms (1.0ms for algorithms)
• Part of L0Calo near the detector
• Use SEU immune components
• L0Muon PU veto far from detector

19
Summary

• L0 uses calorimeter muon and dedicated silicon
  vertex detector
• Reduces to 1MHz the input rate
• Robust and flexible
• Sends L0 candidates to L1 for further processing

L0 efficiency () L0 efficiency () L0 efficiency () L0 efficiency () L0 efficiency () L0 efficiency ()
m e g h all
B-gtpp 7 9 3 55 61
Bs-gtDsK 8 5 2 37 44
Bs-gtJ/y(ee)f 7 36 4 24 52
Bs-gtJ/y(mm)f 90 5 3 30 93
Bs-gtKg 6 28 47 30 82
20
Robustness

• The L0 efficiencies of various channels show a
  large region of very stable performance
• Decreasing the L0 bandwidth to 750KHz
• results in loss15

PT m cut