Detector for GlueX

Detector for GlueX

JLab PAC 23Jan 20, 2003

Physics Beamline Hall D

GlueX Detector

Software Trigger

ComputingEnvironment PRL

What is needed?

9-GeV polarized photon beam Coherent bremsstrahlung beam

Hermetic detector for multi-particle charged and neutral final statesCharged Solenoid-based detector

Select events of interest with high sensitivity High DAQ rate capability with software trigger

Analysis environment for successful PWA

Construction Site

Status of Civil Design

Credible optics design Layout that provides room for detectors and

access to equipment Beam containment proposal Concept for civil design GEANT Calculations show that the shielding

satisfies radiation protection guidelines

Hall D site layout

Coherent bremsstrahlung beam

Photon beam energy (GeV)

Flux

Delivers the necessary polarization, energy and flux concentrated in the region of interest

P = 40%

Photon beam energy (GeV)

Line

ar P

olar

izat

ion

Gluonic excitations transfer angular momentum in their decays tothe internal angular momentum of quark pairs not to the relative angularmomentum of daughter meson pairs - this needs testing.

GlueX will be sensitive to a wide variety of decay modes - the measurements of which will be compared against theory predictions.

To certify PWA - consistency checks will be made among different final states for the same decay mode, for example:

b1 0 3

0 2

Should givesame results

X b1For example, for hybrids: favored

not-favoredX Measure many decay modes!

Hybrid decays

GlueX detector

Solenoid ships from Los Alamos

Unloading at IUCF

Exploded view of the GlueX detector

TARGET VTXCDC

FDC

CERENKOV Pb-GLASS DET

12.0

0

The components are extracted by “4 ft” from each other for maintenance.

TOF

Particle kinematics

All particlesMost forward particle

p → X p → K+K─+─ p

GlueX Detector

Central trackingStraw tube chamber

Vertex Counter

Forward drift chambers

Charged particle resolution

CalorimetryBuilt by IU for BNL Exp

852

Pb GlassBarrel

Mass

0

Pb/SciFi detector based on KLOE/E = 4.4 %/ √E, threshold = 20 MeVt = 250 ps

Particle identificationTime-of-flight, Cerenkov counter, and constraints for exclusive events

p → K*K*p, E = 9 GeV

Hall D Prototype (IHEP Run 2001)Time Resolution (2cm by 6 cm Bar) vs. Position

0

10

20

30

40

50

60

70

80

90

-100 -80 -60 -40 -20 0 20 40 60 80 100

Position (cm)

Tim

e R

esol

utio

n (p

s)

Predicted resolution for one bar, beam thru 2 cm.Predicted resolution for one bar, beam thru 6 cm.Predicted resolution for two bars, beam thru 6 cm.Measured resl'n, 6 cm, two bars, no weighting.Measured resl'n, 6 cm, two bars with weighting.

Poor PM was on this end.

42 ps avg. resolution

R. Heinz / IU

PID with Cerenkov and forward TOFTOF =100 ps resolution

n= 1.0014 n= 1.0024

p → K+K+p, E = 9 GeV

Acceptance is high and uniform

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

5 GeV

Mass(X) = 1.4 GeVMass(X) = 1.7 GeV

Mass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

8 GeV

Mass(X) = 1.4 GeVMass(X) = 1.7 GeV

Mass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

12 GeV

Mass(X) = 1.4 GeV

Mass(X) = 1.7 GeVMass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

p -> n

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

5 GeV

Mass(X) = 1.4 GeVMass(X) = 1.7 GeV

Mass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

8 GeV

Mass(X) = 1.4 GeVMass(X) = 1.7 GeV

Mass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

Cos(GJ)

12 GeV

Mass(X) = 1.4 GeV

Mass(X) = 1.7 GeVMass(X) = 2.0 GeV

-3 -2 -1 0 1 2 30

0.2

0.4

0.6

0.8

1

GJ

p -> p p Xn n

p Xn 00n

Acceptance in

Decay Angles

Gottfried-Jackson frame:In the rest frame of X

the decay angles aretheta, phi

assuming 9 GeVphoton beam

Mass [X] = 1.4 GeVMass [X] = 1.7 GeVMass [X] = 2.0 GeV

Trigger Rates

Output of Level 3software trigger

Luminosity limits

GlueX raw rates will bewell below currentlyrunning CLAS electronbeam experiments

DAQ architecture 40 VME front-end crates

Gigabit switch

200 Level 3 Filter Nodes

8 event builders

4 event recorders

4 tape silos

8 100-Mbit switches

Fully pipeline system of electronics

Flash ADCs: 13000 channelsTDCs : 8000 channels

DeadtimelessExpandableNo delay cables

(Non-pipeline: Limits photon flux < 107/s, incurs deadtime, requires delay cables)

Data Volume per experiment per year (Raw data - in units of 109 bytes)

100

1000

10000

100000

1000000

1980 1990 2000 2010

E691

E665

E769

E791

CDF/D0

KTeV

E871

BABAR

CMS/ ATLAS

E831

ALEPH

J LAB

STAR/ PHENIX

NA48

ZEUS

But: collaboration sizes! Ian Bird

Data Handling and ReductionCLAS GlueX

Event size 5 KB 5 KBData volume 100 TB/year 1000 TB/yearcpu speed 0.4 GHz 6.4 GHzcpu time per event 100 ms 15 mscpu count 150 225Reduction speed 7 MB/s 75 MB/sThroughput limit cpu speed cpu speedReduction time 0.5 year 0.4 year

Data rate up by 10, computer costs down by > 5

GlueX Computing Effort ~ 2 x CLAS

Computing Model

DAQ Level 3 Farm

EventReconstruction

Calibration

Physics Analysis

Physics Data

Tier “2” Centers

Tier “1” Center (Jlab)

Tier “2” Simulation Center

1 PB/year 0.2 PB/yr

0.2 PB

100 MB/s

70 MB/s

20 MB/s

Physics Analysis

20 MB/s

Monte Carlo

Detector designed for PWADouble blind studies of 3 final states

pn

X

Linear Polarization

m [GeV/c2]

GJ

a2

LeakageAn imperfect understanding of thedetector can lead to “leakage” of strength from a strong partial waveinto a weak one.

STRONG: a1(1260) (JPC=1++)

Break the GlueX Detector (in MC).

Look for Signal strength in Exotic 1-+

Under extreme distortions, ~1% leakage!

Ongoing R&D effort

Solenoid – shipped to IUCF for refurbishment Tracking – testing straw chamber; fabricating endplate prototype

(CMU) Vertex - study of fiber characteristics (ODU/FIU) Barrel calorimeter – beam tests at TRIUMPF; fabricated first test

element of the Pb/SciFi matrix (Regina) Cerenkov counter – magnetic shield studies (RPI) Time-of-flight wall – results of beam tests at IHEP show ps (IU) Computing – developing architecture design for Hall D computing Electronics – prototypes of pipeline TDC and FLASH ADC (Jlab/IU) Trigger – Studies of algorithm optimization for Level 1(CNU) Photon tagger – benchmarks of crystal radiators using X-rays

(Glasgow/UConn) Civil – beam height optimized; electron beam optics shortens length of

construction; new radiation calculations completed (Jlab)

7

R.T. Jones, Newport News, Mar 21, 2002

Benchmarks of Diamond Crystals

Stone 1407 Slice 1 (4mm x 4 mm X-ray rocking curve)

Richard Jones / Uconn

Stone 1482A Slice 2 (10mmx10mm X-ray rocking curve)

High Quality Poor Quality

Straw Tube chamber workGraduate Students: Zeb Krahn and Mike SmithBuilt a -gun using a 10 Ci 106Ru Source

Getting coincidences with both cosmics and ’s

gun

cosmics

ArCO2 90-10ArEthane 50-50

Carnegie Mellon University

Building a Prototype Endplate

Build endplates as 8 sections with tounge and groove.

Checking achievable accuracy

Barrel calorimeter prototyping

Hybrid pmts can operate in fields up to 2 Tesla

University of Regina

Pb/SciFi prototype

FLASH ADC Prototype

Paul Smith / IU

250 MHz, 8-bit FADC

Pipeline TDC

= 59 ps

TDC counts

First prototype results: high resolution mode

Jlab DAQ and Fast Electronics Groups

Cassel review of Hall D concluded“The experimental program proposed in the Hall D Project is well-suited for definitive searches for exotic states that are required according to our current understanding of QCD”

“An R&D program is required to ensure that the magnet is usable,

to optimize many of the detector choices,

to ensure that the novel designs are feasible,

and to validate cost estimates.”

Working with input from many groups on electronics, DAQ, computing, civil, RadCon, engineering, and detector systems.

Hall D BudgetSolenoid 1075

Detectors 11149

Tracking 4837

Calorimetry 4374

Particle ID 1938

Computing 2102

Electronics 3770

Beamline 7188

Infrastructure 1950

Civil 8859

Total 36093

Budget assumes that inaddition to the constructioncost, the Physics operations Budget will increase by 30% to support a Hall D group.

Budget estimates are under review as we update the GlueX Design report. No substantialchanges since 3/01.

What is Needed? PWA requires that the entire event be identified - all particles

detected, measured and identified. The detector should be hermetic for neutral and charged particles,

with excellent resolution and particle ID capability.

The beam energy should be sufficiently high to produce mesons in the desired mass range with excellent acceptance.

Too high an energy will introduce backgrounds, reduce cross-sections of interest and make it difficult to achieve above experimental goals.

PWA also requires high statistics and linearly polarized photons. Linear polarization will be discussed. At 108 photons/sec and a 30-

cm LH2 target a 1 µb cross section will yield 600M events/yr. We want sensitivity to sub-nanobarn production cross-sections.

Missing mass resolutionp → X p → K+K─+─0 p

Hall D schematic

Outline

Site and civil construction Detector Requirements Recent developments

DAQ Electronics Offline Computing

Pipeline TDC based on COMPAS F1 chip Resolution

4 channels per chip of high resolution (60 ps least count) 8 channels per chip of low resolution (120 ps least count)

Density per board 32 channels of low resolution inputs 64 channels of high resolution inputs

Packaging in VME64x Dynamic range – 16 bits (~ 4 s) Development by JLab Fast Electronics and DAQ

groups (F. Barbosa and E. Jastrzembski)