The Hades second level trigger: from the concept to the first results

1
The Hades second level trigger from the
concept to the first results
Alberica Toia Jörg Lehnert II Physikalisches
Institut Justus-Liebig-Universität Gießen,
Germany for the HADES Collaboration

• Motivation and Requirements of the 2nd Level
  Trigger
• Functionality
• Components
• Algorithms
• Performance
• Suppression factor
• Efficiency
• Comparison Online-Offline Analysis

2
HADES Physics Program

• Modification of vector meson properties due to
  interaction with the medium
• Access
• normal nuclear matter p, p beam
• dense nuclear matter heavy ion beam

Electromagnetic form factors of hadrons in time
like region w transition form factorfew
existing data not in agreement with VMD models
SPECTRAL FUNCTION OF r MESON
w TRANSITION FORM FACTOR
ee- -gt p0 w
Model M.Post et al., Nucl.Phys.A 689 (2001)
753 University of Giessen European Graduate School
w -gt p0mm-
need for electromagnetic probes(no strong final
state interaction) r,w -gt ee-(branching
ratio 10-5 10-6 )
Model F.Klingl, et al., Nucl.Phys.A 624 (1997)
527 Data R.Dzhelyadin et al., Phys.Lett.B 102
(1981) 296 V.Druzhinin et al., INP84-93
Novosibirsk
3
HADES Detector

• Components
• MOMENTUM RECONSTRUCTION
• Mini Drift Chambers (MDC) in front and behind a
  supreconducting toroid magnet
• LEPTON IDENTIFICATION
• Hadron-blind Ring Imaging Cherenkov (RICH)
  detector
• Electromagnetic Shower detector
• Time of Flight (TOF) wall

• Features
• 106 events / s
• 45 geometrical acceptance
• up to 200 charged particles
• 1 mass resolution

TOF
SHOWER
MDC
RICH
MAGNET
4
Trigger System
First Level Trigger multiplicity -gt reaction
centrality
1100

• Second Level Trigger I (Image Processing Units -
  IPU)Identification of electron candidates
• Cherenkov rings
• Electromagnetic showers
• Time-of-flight
• Second Level Trigger II
• (Matching Unit - MU)
• Selection of lepton pairs
• Correlation of electron candidates in front of
  (RICH) and behind magnetic field (Shower, TOF)
• Invariant mass determination

• 60 VME boards
• FPGA, CPLD, DSP implemented
• decision 10 ms
• process 3 GByte/s raw data

5
Ring Recognition Requirements

• Ring Recognition
• constant ring diameter
• small ring size (8 pads diameter, 32 pads on
  circumference)
• incomplete rings (short radiator length of
  36-65cm)
• smeared out rings (optical distortions, wire
  chamber response)
• detector background
• scintillation-/Cherenkov light (radiator/window)
• direct ionization in the detector
• electronic noise
• background electron sources
• low mass dalitz pairs (p0-dalitz)
• conversion pairs

• Technical Boundary Conditions
• 105 decisions / s
• short latency (some 10 ms)
• flexible algorithms
• flexibility in input data order
  

6
The Ring Recognition Hardware

• VMEBus Cards
• system speed 40 MHz
• Pattern Reconstruction Card
• pipelined operation, three units in parallel
• interfaces to readout and matching unit (16/20
  MHz)
• configuration, status, test
• Ring Recognition Unit
• 12 FPGA
• all 96 columns in parallel,rows sequential_at_12
  MHz

• Pattern Reconstruction Card

• Ring Recognition Unit

• Cards for one HADES sector

• Hardware Operation
• Tests with HADES Daq-Trigger System
• with selected subsystems 17 kHz
• Standalone Tests up to 48 kHz

• Operation in CC beamtimes
• approx. 450 Mio events
• rates up to 11 kHz (current system limit)
• Operation in different modi
• ring recognition, pad recognition

7
Ring Recognition Algorithm
Task Recognition of asympotic Cherenkov rings on
96x96 pad plane

• The Algorithm
• Based only on hit pattern information
• Fixed ring diameter of 8 pads
• Evaluation of 2 regions on 13x13 pad square
• ring circumference (32 pads)
• inner/outer veto region (48 pads)
• Logical OR in groups of 3-4 pads (at least
  one pad)
• different grouping of pads
• Number of valid pad groups in both regions
• Threshold conditions
• Local maximum search (Sum in ring region)

8
Performance 1- Suppression Factor
Ring Recognition Thresholds
fraction of accepted events for different sets of
thresholds

• suppression factor depends on
• reaction conditions
• detector conditions

Matching Unit Selectivity

• fakes at low pad multiplicity
• efficiency loss for higher thresholds

9
Performance 2- Efficiency
Single Lepton Recognition Efficiency

• Efficiency studied with simulations of
• pure e-
• e- embedded in CC

p (MeV)

• No momentum dependence
• Increases with polar angle (due to longer
  radiator path length)

q (degrees)

• SIMULATIONS
• Uncertainty of event generator
• Uncertainty of hit digitization
• Need for an absolute reference system pp
  -gt pph

Efficiency
Fakes/evt pure e- (10 lt ? lt 90
100 lt p lt 1000 MeV 88.7 0.17Noise
1) CC(1.9AGeV) e- 87.6 0.4
10
Performance 3 Comparison Online-Offline
Analysis 1

• Good identification of particles
• Good separation between positive and negative
  tracks

electrons
p 1/DqMETA - MDC
Low magnetic field better understanding of
background sources (p0 Dalitz, conversion)
positrons
MDC opening angle
11
Performance 3 Comparison Online-Offline
Analysis 2
Compatible results at large opening angle
satisfactory reconstruction of open
pairs Discrepancies at small opening angle due to
different screening mechanism close pairs ?
efficiency loss ?
12
Beam Time 2002 Trigger Operation
Trigger condition 1 lepton (1
ring matched by TOF / Shower) Event Reduction
8 accepted events
Matching Unit

• no polar cut
• azimuthal match ( correction due to
  inhomogeneous field)

13
Beam Time 2002 Lepton Enhancement

• Enhancement of correlated rings in triggered
  events

Trigger condition
LVL2 Triggered Events
LVL1 Minimum Bias Events
0.2 lept / evt
0.025 lept / evt
--gt lepton candidate enhancement 8.25
14
Summary and Outlook
Jörg Lehnert Alberica Toia I.Fröhlich,
A.Gabriel, D.Kirschner, W.Koenig, W.Kühn,
E.Lins, T.Pérez, M. Petri, J.Ritman, D.Schäfer,
M.Traxler II. Physikalisches Institut Justus-Lie
big-Universität Gießen, Germany GSI Darmstadt

• Characterization of second level trigger and its
  components
• Estimation and methods for
• Suppression factor
• Single lepton efficiency
• Pair ratios online/offline and vice versa
• Low magnetic field
• possibility of resolving close pairs
• better understanding of the ring properties
• Production run background suppressed by a
  factor gt 10