Department of Energy
Budget Justification

DOE F 4620.1 U.S. Department of Energy OMB Control No.

(04-93) Budget Page1910-1400

All Other Editions Are Obsolete

(See reverse for Instructions) OMB Burden Disclosure

Statement on Reverse

11/1/01 - 10/31/02 Year 3 DE-FG02-95ER40899

ORGANIZATION Budget Page No: 1

The Regents of the University of Michigan, Ann Arbor, MI 48109

PRINCIPAL INVESTIGATOR/PROJECT DIRECTOR Requested Duration: 12 (Months)

PI: Bing Zhou, Associate Professor Task A A. SENIOR PERSONNEL: PI/PD, Co-PI's, Faculty and Other Senior Associates DOE Funded (List each separately with title; A.6. show number in brackets) Person-mos. Funds Requested Funds GrantedCAL ACAD SUMR by Applicant by DOE1. PI: Bing Zhou, Associate Professor 0.00 0.00 1.50 $14,083 2. Co-PI: J. Wehrley Chapman, Professor 0.00 0.00 1.00 $10,278 3. Co-PI: Homer A. Neal, Professor 0.00 0.00 1.50 $27,098 4. Greg Tarle, Professor 0.00 0.00 1.00 $11,489 5. Rudolf Thun, Professor 0.00 0.00 2.00 $21,356

6. ( 5 ) OTHERS (LIST INDIVIDUALLY ON BUDGET EXPLANATION PAGE) 56.00 0.00 0.00

$246,657

7. ( 10 ) TOTAL SENIOR PERSONNEL (1-6) 56.00 0.00 7.00 $330,961

B. OTHER PERSONNEL (SHOW NUMBERS IN BRACKETS)

1. ( 0 ) POST DOCTORAL ASSOCIATES 0.00 0.00 0.00 $0

2. ( 4 ) OTHER PROFESSIONAL (TECHNICIAN, PROGRAMMER, ETC.) 36.00 0.00 0.00 $130,535

3. ( 0 ) GRADUATE STUDENTS 0.00 0.00 0.00 $0

4. ( 0 ) UNDERGRADUATE STUDENTS 0.00 0.00 0.00 $0 5. ( 0 ) SECRETARIAL - CLERICAL 0.00 0.00 0.00 $0

6. ( 0 ) OTHER

TOTAL SALARIES AND WAGES (A+B)$461,496

C. FRINGE BENEFITS (IF CHARGED AS DIRECT COSTS) estimated 28% of Total Salaries and Wages $129,219

TOTAL SALARIES, WAGES AND FRINGE BENEFITS (A+B+C) $590,714

D. PERMANENT EQUIPMENT (LIST ITEM AND DOLLAR AMOUNT FOR EACH ITEM.)

TOTAL PERMANENT EQUIPMENT

E. TRAVEL 1. DOMESTIC (INCL. CANADA AND U.S. POSSESSIONS) $25,000

2. FOREIGN $50,000

TOTAL TRAVEL $75,000

F. TRAINEE/PARTICIPANT COSTS

1. STIPENDS (Itemize levels, types + totals on budget justification page) $0

2. TUITION & FEES $0

3. TRAINEE TRAVEL $0

4. OTHER (fully explain on justification page) $0

TOTAL PARTICIPANTS ( 0) TOTAL COST $0

G. OTHER DIRECT COSTS

1. MATERIALS AND SUPPLIES MATERIALS AND SUPPLIES Overnight packages, postage, telephone, general consumables, software, visitors $33,132

2. PUBLICATION COSTS/DOCUMENTATION/DISSEMINATION $0 3. CONSULTANT SERVICES $0 4. COMPUTER (ADPE) SERVICES $0 5. SUBCONTRACTS $0

6. OTHER CERN running costs

$20,000

TOTAL OTHER DIRECT COSTS $53,132

H. TOTAL DIRECT COSTS (A THROUGH G) $718,846

I. INDIRECT COSTS (SPECIFY RATE AND BASE)

$718,846 x 26%

TOTAL INDIRECT COSTS $186,900

J. TOTAL DIRECT AND INDIRECT COSTS (H+I)

$905,747

K. AMOUNT OF ANY REQUIRED COST SHARING FROM NON-FEDERAL SOURCES $0

L. TOTAL COST OF PROJECT (J+K)$905,747

Department of Energy

Budget Justification

FY 2001 (Year 3 of 3)

Personnel:

Bing Zhou, Associate Professor, PI, 1.5 months summer salary $14,083

J. Wehrley Chapman, Co - PI, Professor, 1.00 months summer salary $10,278

Homer Neal, Co - PI, Professor, 1.5 m onths summer salary $27,098

Greg Tarle, Professor, 1.00 months summer salary $11,489

Rudolf Thun, Professor, 2.00 months summer salary $21,356

Edward Diehl, Research Scientist, 12 months salary $50,400

Dan Levin, Research Scientist, 12 months salary $5 8,190

Shawn McKee, Research Scientist, 12 months salary $50,900

Zhengguo Zhao, Research Scientist, 8 months salary $36,667

Steven Goldfarb, Research Scientist, 12 months salary $50,500

Helmut Schick, Technician, 12 months salary $50,660

Curtis Weaverd yck, Technician, 10 months salary $39,941

John Reece, Technician, 12 months salary $30,600

Eric Myers, Programmer, 2 months salary $9,334

Salaries are consistent with those paid for similar positions within the department of Physics and the

University. Fringe Benefits estimated at 28% of salary request.

Equipment: none

Travel:

Airfare/Transportation Accommodations Per diem

(25 trips) Domestic $400 $400 $200

(25 trips) Foreign $1000 $700 $300

Travel estimates are based on current airline tic ket quotes and past trips of a similar nature.

Other Direct Costs:

Materials and Consumables – includes postage, overnight express shipping, freight, copy -

ing, software, computer manuals, books, misc. electronic parts, general office consumables,

phone/v ideo charges, ATLAS computing/collaborative R&D development tools

$33,132

Other – CERN running costs $20,000

Estimates based on past experience of a similar nature.

Indirect Costs:

The indirect cost rate is 26% of MTDC, as negotiated with the Universi ty of Michigan and the DHHS on

June 29, 1995.

T askA-ATLAS

FundingYear2002

Faculty

M.Campbell,J.Chapman,H.Neal,J.Qian,G.Tarle, R.Thun,andB.Zhou

ResearchPh ysicist

E.Diehl,S.Goldfarb,D. Levin,S.McKee,Z.Zhao

Technical Support

P.Binchi,T.Dai,J.Reese,H. Schick,andC.Weaverdyck

Graduate Students

C.Han,G.Mikus,andQ.Xu

DatafromMichigan’sEMS5

R andallLaboratoryofPhysics

UniversityofMichigan

AnnArbor,MI48109-120September2001

Con tents

1Introduction1

2PrecisionMuonDetectorConstruction1

2.1MDTTubeWiring................. ................2

2.2PrecisionCham berAssembly...........................6

2.3PrecisionToolingforthe2ndSeriesChambers. ................9

2.3.1Wiring StationModi?cation........................9

2.3.2Re-buildToolingforChamberAssembly..... ............10

2.4ChamberServices. .................................12

2.4.1Newgasbardesignandprototype........... ..........13

2.4.2Chambergassystem installationandtest.................14

3ChamberReadout17

3.1MDTReadoutMultiplexerDevelopment........ .............17

3.2ChamberCosmicRa yTest.............................18

4Computing21

4.1Overview................................ .......21

4.2LeadershipRoleinUSA TLASNetworking....................22

4.3ATLASMuonDatabaseCoordination........... ............22

4.4ATLASDetector DescriptionActivities......................23

4.5HEP-WideDetectorDescriptionActivities... .................24

4.6University ofMichiganNetworkInfrastructure..................24

4.7MDTProductionDatabase................... .........25

4.8MDTSystemSimulation Studies.........................26

4.9CSCStagingSummary....................... ........27

4.10MDTAlignmentStudies ..............................27

4.11MDTAlignmentConclusions...........................28

4.12UniversityofMichigan-CERNNetwork ResourceReservationStudy.....28

4.13TriggerDatabase..................................29

T askA-ATLAS

FundingYear2002

Faculty

M.Campbell,J.Chapman,H.Neal,J.Qian,G.Tarle, R.Thun,andB.Zhou

ResearchPh ysicist

E.Diehl,S.Goldfarb,D. Levin,S.McKee,Z.Zhao

Technical Support

P.Binchi,T.Dai,J.Reese,H. Schick,andC.Weaverdyck

Graduate Students

C.Han,G.Mikus,andQ.Xu

1Introduction

FY2001hasbeenaveryproductiveyearfortheATLAS projectattheUniversityofMichigan.

WereportoursubstantialprogressandachievementsinFY2001inthisdocument.The

Michigangrouphastakenonsigni?cantresp onsibilitiesfortheATLASexperiment.Our

majortasksinclude:making31,000longprecision drifttubesandassemblingthesetubesinto

80largemuonchambers;coordinatingthemuon detectorfront-endelectronics;andleading

themuonsoftwaredevelopmentandtheUSATLAS computingnetworkR&D.Thedetailsof

theprogressintheseareaisdescribedinthefollowingsections.

2PrecisionMuonDetectorConstruction

InthepastyearMichiganhasmovedfromtheA TLASmuondetectorR&Dandpreparation

phasetomassproductionoftheMDTchambers.Themajorachievementsinthepastyear

aresummarizedbelow.

† Aftertheconstructionofavery successfulfull-sizedprototypechamber(Module0in

June,2000),wefurtherenhancedourproduction facilities,addingtheElectromagnetic

Micrometerforwirepositionmeasurement,andsev eralopticalmonitoringsystemsfor

c hamberproductionandchamberglobalalignment.Thesenewfacilitieshelpedussig-

ni?cantlyspeedupourproductionrate andimproveonourhighqualitystandards.

† After?nishingthe?rstchamberseries (EMS5)inearlyJunethisyear,werebuiltour

tubeassemblylineandteststations,andchamber productionfacilitiesforthesecond

c hamberseries(EMS4).Theaccuracyofthetoolingandsetuphasreachedalevelof10

microns,whichismuchbetterthantheATLASc hambertoolingspeci?cations.

† 10,000precisiondrifttubeshavenowbeen assembled.Ofthesetubes,98.2%passedall

thequalityassurancetests.Therecenttubetest rejectionrateisabout1%,whichisa

factorofthreelowerthantheATLASMDTproductionspeci?cations.

† TwentythreelargeMDTchambershavebeen assembled,includingsixteenof?rstseries

chambers(EMS5),andsevenofsecondserieschamb ers(EMS4).Thechamberproduc-

tion rateis15%higherthanweoriginallyestimatedforthe?rstyearofproduction.

† Wehavehad2chambersmeasuredinthex-ray tomographyfacilityatCERN,andboth

metthe25„ mRMSprecisionrequirementonwirep ositions.

† Weplayedaleadingroleintheredesign, prototypingandtestofthesecondtypegas-bar

(chambergasmanifold)forATLASendcapmuonchamb ers.Thisnewgas-bardesign

isgreatly improvedinitseaseofassembly,installationandsealing.Inaddition,the

machiningcostofthenewdesignismorethana factorof2lowerthantheoriginal

design.

† Eightchambershavebeenequippedwithgas systems,andsevenchambershavepassed

theATLASchamberleakcerti?cation.This progressisespeciallynoteworthysincewe

receivedthechamberservicepartsverylate (September,2001).

† Acosmicrayteststationhasbeenbuilt, andanMDTchamberisunderactivetestwith

alltheelectronicsandreadoutsystems.

† Wehavedesignedandconstructedchamber shippingandstoragecrates.Thesehave

beenusedtoshipall16EMS5chamberstoCERNvia sea-borncontainertransport.All

crateshavearrivedwithnovisibledamage.Wehaveopened8ofthe16cratestocheck

forbrokendrift-tubewires(eachchamberhas 384tubes).Sofar,nowireshavebeen

brokenintransitorinstorage.

Figure 1showstheMichigantubeandchamberproductionratesoverthepastyear.As

thechartshows,afteraninitialramp-upperiod, ourproductionratesstabilizedto200-250

tubesperweekand2-3chamberspermonth.

2.1MDTTubeWiring

WebeganchambermassproductioninSeptember, 2000assoonastuberawmaterialsarrived,

andquicklyrampeduptofullproductionspeed.MDT chamberproductionrequirestwo

main operations:stringing/testingofindividualdrifttubesandconstruction/testingofen tire

chambers.Ourwiringlaboratory consistsofawiringstationonwhichatubeisassembled

including:settingtubelength;wiringthetube; insertingandcrimpingtheendplugs;setting

wiretension.Onceatubeisassembled,thetube lengthandwiretensionaremeasuredonthe

samestation.Alltheoperationsareunder computercontrolanddatalogging.Twooperators

workonthewiringstation(seeFigure2), completingonetubeinabout7minutes,about50

perday.Signi?cantpreparationworkisalso requiredtoproduceatubesincealltheendplug

parts(endplugbody,wirelocator,twister, groundspring,ando-ring)mustbecleanedinan

ultrasonicbathandassembledbeforeuse.

Tube and Chamber production

100

150200250300

Sep OctNovDecJanFebMarAprMayJunJulAug

tubes/week

2000 2001

12345678910

Sep OctNovDecJanFebMarAprMayJunJulAug

chambers/month

Tubes/week

Chambers/month

Figure 1:Tubeandchamberproductionrates.Pro ductionwasslowerat?rstaswedebugged

oursystem.ProductionstoppedinJune,2001toc hangetoolingfromtheEMS5chamber

typ etotheEMS4chambertype.

Figure 2:Twooperatorsareassemblingtubesin theMichiganwiringstation.Anotheroper-

atorisperformingthetubeleaktestinatest station.

Wire tension (g)

Tube QA Test Results

500

1000

330 340350360370

EntriesMeanRMS

8964

349.3 3.672

D Tube length (mm)

250

5007501000

-1 -0.500.51

EntriesMeanRMS

8964

-0.8414E-01

0.9917E-01

Leak rate (10

-8

bar-liter/sec)

500

1000

1500

2000

0 0.250.50.751

EntriesMeanRMS

8964

0.7249E-01

0.1007

Dark Current (nA/m)

500

1000

1500

0 1234

EntriesMeanRMS

8964

0.5941 0.4571

Wire offset ( m m)

500

1000

-40-2002040

EntriesMeanRMS

25958

0.8332

9.067

Figure 3:PlotsofQAmeasurementsonMDTtubes. Redarrowsindicatetolerancelimits.

Aftertubewiringthetubemustpassthroughaseriesofstrictqualitycontroltests:wire

tensionmeasurement(twice- immediatelyafterwiring,andjustbeforechamberassembly);

tubelengthmeasurement;tubeleakrate measurement;wirepositionmeasurement;anddark

Tube Failures

0.1

0.20.30.40.5

Wiring Len.Tens.LeakPos.DCMisc

Tube failure rate by type

Percentage failure

Tube Failures by month

24681012

Sep OctNovDecJanFebMarAprMayJunJulAug

Tube failure rate by month

Percentage failure

Figure 4: Upperplot: TubefailureratesbyQAtesttype(wiring= failureintubewiring;

Len=badtube length;Tens.=badwiretension;Leak=badleakrate;Pos.=badwire

position;DC=baddarkcurrent;Misc=other failures.Lowerplot: Totaltubefailurerate

bymonth.Thisplotshowsthetubefailurerate fallingasproductionprogressesandimproves.

currentmeasurementwithdriftgasunderHV. Figure3showstheresultsfromallthese

measurementsdemonstratingtheveryhighquality ofMDTtubesproducedatMichigan.As

mentionedpreviously,theoverallfailurerateoftubesisabout1.8%,thoughtheratehas

graduallyfallenovertimeandisnowab out1%.Figure4showsfailureratebyQAtesttype,

andtotalfailurerateovertime.

Wehavehonedourtubeoperationoverthepastyear. Thelogisticalaspectsarenot

trivial.Wemuststoreandtransporthundredsoflong,fragiletubesthroughthevarious

stagesofproduction.Westoremostof therawtubesataremotewarehouse(duetolack

storagewithinthephysicsdepartment),and transporttubesbacktoourlocalstoragefor

productioneverymonth.Wehavebuiltseveral storageracksinthewiringandtestingrooms.

Wetracktubesviaabarcodewrittenoneachtubeand anextensivecomputerdatabase,

which logsallproductionandqualityassurancedataforindividualtubes.Theproduction

statussummarytableisupdateddaily, andcanbeviewedontheweb.[1]

Figure 5:Precisioncombslayoutonthegranite table.Opticaldevicesareusedtoalignthe

combsto10micronstoplacethetubelayersonthem. Ourlargeautomaticgluingmachine

is alsoshowninthepicture.

2.2 PrecisionChamberAssembly

The progressinchamberassemblyatMichiganisremarkable.Overthepastyearwehavegone

fromhavingproducedjustoneprototyp echamber,Module0 inJune2000,tofullchamber

productionspeedofonechamberevery8days,which ismuchfasterthantheoriginalplanned

rateof10daysperchamber.Serieschamberpro ductionbeganinOctober,2000.Theproce-

dureforchamberproductionhasworkedwell, thoughwehavemademanyre?nementstothe

operationovertime.

Chamberassemblyincludesgluingtwomulti-lay ers(eachconsistingofthreetubelayers

of64tubeseach);buildingaspacerframe;and gluingthespacerframebetweenthetwo

multi-layers.Eachlayergluingtakesoneday.Inaddition,opticsforin-planealignment, and

globalchamberalignmentmustbe installedwithprecisionjigging,andreferenceimagesfor

thesesystemsonthegranitetablemustberecorded aspartofthechamberproduction.

The mechanicalprecisionisprovidedbytheelaboratejiggingonthechamberassembly

table.Theassemblytableisaverylargeandhea vygraniteplate,forûatnessandstability.

Thetubesareplacedonnotchedaluminumbars calledcombs topositionthem.Thecombs

havebeencarefullymadewithnotchespositioned toanaccuracyof5„ m.Thecombs,of

which7-13maybeuseddependingonchambersize,m ustbecarefullyalignedandclamped

on thegranite(SeeFigure5).Tubelayersarepositionedwithaliftingframecalleda sti?back

duringgluing.Thesti?backispositionedby nestingintosphereblocktowers,whichcomein

di?erentheightsusedforthevariouslayersofthe chamber.Figure6showsthegluingofthe

spacerframetothetwomultilayers,illustrating themainfeaturesofthechamberjigging.

Chamberassemblyprecisioniscarefully monitored.Themostimportantfactoristhe

mechanicalprecisionofthegluingprocess.The MDTchambersaredesignedforatracking

Figure 6:Gluingthespacerframetothetwom ultilayers.Thesti?backrestsonthesphere

blocktowersandpositionstheuppermultilayer. Thelowermultilayerrestsonthecombs.

precisionof80 „ m,whichrequirestubepositionsaccurate to25„ m.Thisisanimpressive

precisiontoachieveforchamberswhichare3-6m long.

Wemadegreate?ortstosetupand monitorthejiggingonthegranite.Westrivetotake

redundantindependentmeasurementswheneverp ossibletoremovesystematicerrors.For

example,toalignthecombsweusebotha wire-microscopemethod(alignmenttoastretched

wireviewedwithamicroscope),andalaser-CCD method(alignmenttolaserviewedbyCCD,

calledtheBCALmethod).Inaddition,wemonitor thepositionsofthecombsperiodically

throughoutproductionwiththeBCAL.Figure7sho wsaseriesofBCALmeasurementsmade

duringEMS5chamberconstructiondemonstratingtheaccuracyandstabilityofthecombs

duringthisperiod.Thecombswerep ositionedtoanaccuracyof8„ mRMSforEMS5

construction.

Wemonitorthepositionofeachtubelayervia opticalsensorsonthesti?back.Oneach

cornerofthesti?backare2 RASNIK masks[2]whichareviewedbyCCDcamerason the

granitetable.ARASNIKmaskisac heckerboardpatternmaskwithbinarypositiondata

encodedinthesquares.Byreadingthebinarycode, andanalyzingthelight/darktransitions

ofthecheckerboardpattern,aposition measurementaccuratetobetterthan1„m maybe

obtained.Themeasurementsarehighlyaccurate andfastsinceimagescanbeobtainedand

analyzedinamatterofseconds.Thesti?back monitoringsystemisknownastheOptical

PositionSystem(OPS).Eachcornerofthesti?back hasanOPSsystemwhichgivesredundant

measurementsofthesti?back(andhencetubelayer)positions.Thuswemonitortheposition

ofthesti?backasweglueeachlayerto ensurethatthejiggingissetupcorrectlyineach

assemblystep.Inaddition,thereareRASNIK systemssetupinsidethesti?backtomonitor

distortionsofthesti?back.Thesemeasurements aresomewhatredundantwiththeOPS

measurements.Figure8showsOPSmeasurementsdoneontheEMS5andEMS4chambers.

Thesemeasurementsdemonstratethehighmec hanicalprecisioninchamberconstruction.The

smalltailsintheEMS5distributionsaredueto instabilitiesintheOPSsystemwhichwere

?xedduringtheEMS4re-toolingprocess.

-30

-20-100102030

-200 -150-100-50050100150200

Notch 15 comb position (cm)

Deviation m m

4/28/00 8/1/00 10/18/00 1/4/01 3/7/01 5/18/01

Comb Height

-30

-20-100102030

-200 -150-100-50050100 150200

Notch 15 comb position (cm)

Deviation m m

Comb Lateral Position

Figure 7:PlotsofBCAL measurementsoftheEMS5combheights(upper)andlateralposition

(lower)madeduringEMS5 construction.Thesemeasurementsweremadeinnotch15onthe

9combsusedintheEMS5setup. Thecombpositionisthepositionofthecombonthe

graniteinchambercenterco ordinates.

The?nal chamberprecisionassurancetestandapprovalforATLASMDTchambersisthe

3Dx-raytomographatCERN.This devicecanmeasurewirepositionsinall6layersofacom-

pletedchamber.Michiganhas haditsmodule0chamberandoneoftheEMS5serieschambers

measuredintheCERNx-ray tomograph.Theaveragewiredeviationfromspeci?cationwas

foundtobe16 „ mand20 „ mforthe2chambersresp ectively,bothcomfortablywithinthe

25 „ mwirepositionsp eci?cationofATLAS.Timeconstraintsdonotallowallchambersto

bemeasuredwiththex-ray tomograph.Duetothecarefulmonitoringourchamberjigging

duringconstruction(i.e.the BCALcombmeasurementsandOPSdiscussedearlier),weare

con?dentthatallofourchamb ersarewithinspeci?cations.

Basedontheexperienceofthe ?rstyearchamberproduction,Michiganiscon?dentthat

wecancompleteallourplanned basechambersinatimelyandcoste–cientmannerandmeet

theATLASspeci?cationsoncham berquality.

EMS5 OPS Lateral Dev. m m

123

-40 -2002040

EntriesMeanRMS

120

-5.944

12.61

EMS5 OPS Vertical Dev. m m

2.5

57.51012.51517.52022.5

-40-2002040

EntriesMeanRMS

240

3.200 7.442

EMS4 OPS Lateral Dev. m m

0.5

1.5

22.5

3.5

-40 -2002040

EntriesMeanRMS

5.562 7.850

EMS4 OPS Vertical Dev. m m

2468101214

-40 -2002040

EntriesMeanRMS

4.781 2.798

Figure 8:PlotsofOPS measurementsforEMS5andEMS4serieschambers.TheEMS4

chambermeasurementsare somewhatbetterduetoimprovementsmadeinthesystemover

time. 2.3 PrecisionTooling forthe2ndSeriesChambers

2.3.1WiringStationMo di?cation

TheEMS5 tubeproductionwas?nishedonMay15,2001.Re-toolingofthewiringfacilities

included:

† Movingthewiring platformclosertogethersincethesmallestEMS4tubelengthisbelow

whatcouldbeaccommodated withintherangeofourmovableplatform.Alaserwas

usedtoalignthetwoplatforms tobetterthan25microns.

† Re-calibratingthetub elengthstandard:aspecialnewdrifttubebuilttoserveasa

lengthstandard,andmeasured withaonemeterMitutoyodigitalscalewithanaccuracy

andrepeatabilityof0.03mm. Thesedevicesarecarefullycalibrated,andusedto?ndthe

\home" locationofthemovingplatform,andto measurethetubelengthsinproduction

andQAprocesses.

† Re-tuningthecrimpinggapforthewiring pinstoensurethewirecrimpquality.

†Modifyingtheproductioncontrol electronics,andthecomputerprogramsforEMS4

tubes.

† Modifyingthedynamicrangeofthetube leaktesterandtheEMMIstationfortubeQA

tests.

ThenewstationsbeganproducingEMS4tubesonJune 1.

2.3.2Re-buildToolingforChamb erAssembly

DuringJune5-July11we re-builtthe2ndserieschamber(EMS4)assemblyprecisiontooling.

Themajorworkislistedbelow:

† AsurveyalltheEMS5jigging,andthe granitetableûatness.WefoundthattheEMS5

jigginghadbeenstableto10microns.Wealsofound thatthegranitetableûatness

changed from15micronsto25micronsintheareaofthechamberconstruction.

† Re-buildingthelargesti?backframeand theassociatedopticalmonitoringdevices.

† Modi?cationofthesti?backholding structureontheoverheadcraneandthevacuum

manifold.

† Re-positioningtheprecisioncombsand sphereblocks.Combswereshimmedtoarelative

heightbetterthan20microns(afactorof2better thanthe?rstserieschamberjigging

set).Comblateralpositionswerealignedto10microns.

† Re-buildingthecombvacuumsystemonthe granitetable.

† Re-buildingthetableforchamberspacer frameassembly,whichconsistsoftwolarge

supportstructuresandthreejig-plates.After shimming,theoverallûatnessofthetable

was60microns(betterthantheATLASsp eci?cation).

† Re-installationandcalibrationthe OpticalPositionSystem(OPS).Thissystemislo-

catedoneachcornerofthesti?backtogive redundantmeasurementsofthesti?back

(andhencetubelayer)positions.

† Upgradingthelargegluemachineby increasingtheheightofthegantry,soitwould

passovertheOPStowerstoavoidtheneedofmoving theOPStowersduringchamber

gluing.

† Upgradingthein-planealignmentreadout system,andthegluecontrolcomputerpro-

grams.

† Surveyingthenewtoolingusingoptical devicesandmicrometerstoensurethatthe

toolingisbuiltcorrectly,andalignedtothe requiredaccuracy.Ourcarefulattentionto

detailpaido?aswediscoveredthattheanglecombs (combsfortheendsofthechamber

which areangledtomatchthetrapezoidshapeofendcapchambers)haddeveloped

anunexpected30 „ mbowduetothecombclampingdevice.We madeagreate?ort

tounderstandand correctthisbow.Finally,byaddingshimstopushuptheends

oftheanglecombs,wewereabletocorrectthecomb deformation.Figure9shows

measurementsoftheheightoftheanglecombsbeforeandafterthecorrectionwas

made.

-30

-20-100102030

0 102030405060

Notch

Dev. mm

Notch

Dev. mm

Combs Bowed

-30

-20-100102030

0 102030405060

Notch

Dev. m m

Notch

Dev. m m

Combs Straight after fix

Figure 9:PlotsofBCAL measurementsofanglecombheights,beforeandafteranglecomb

bowwas?xed.A1andA2referto the2anglecombs.

There-constructedchambertoolingandfacilitieswentbackintooperationonJuly12.

Sincethenwehave producedEMS4chambersatarateofonechamberper8days.Detailsof

there-toolinginformationcan befoundonourWeb.[3]

Service DateAvailable

F aradaycageSept.,2001

Gasmanifold Aug.,2001

TemperaturesensorsSept., 2001

HVhedgehogcardsJan.,2002

SignalhedgehogcardsJan.,2002

SignalmezzaninecardsMay,2003

Magnetic?eldsensorsSept.,2002

ChamberServiceModuleMay,2003

Table1:ChamberServicesComponents

2.4ChamberServices

Afterthe\basechambers"areassembled, considerableworkremainsfortheinstallationof

chamberservices.Thisworkincludes:

† installationofthechambergasmanifold andFaradaycagebaseplates;

† performingaverystrictchamberleak test;

† installationoftemperatureandmagnetic sensors;

† installationofthehighvoltagehedgehog cards,andperformingthechamberHVtest;

† installationofthesignalhedgehog cards;

† installationofthesignalmezzanine cards;

† installationofthecompleteFaraday cages;

† installationofthechamberservicemo dules(readoutmodule);

† installationofthechambersurvey targets;

† overallchamberoperationtestwith cosmicrays.

Amajordi–cultywehave facedinourchamberproductionandtestingisthedelayofthe

parts,particularly,thechamberservices.These partsareunderactivedevelopmentandwill

beinstalledonchambersassoonastheybecomeav ailable.Thefulllistofchamberservices

andexpectedavailabilityisgiveninTable1.

Ourmajorachievementsinchamber servicesinthepastyeararethere-designofthe

chambergas-barandinstallationofgassystemson chambers.Wereportourprogressbelow.

2.4.1 Newgasbardesignandprototype

Inthepastyear,theMichigangrouphas activelyparticipatedinthedesign,prototyping,and

teststoimproveand?nalizethegassystem.Whenwe installedthegassystemonourmodule

0,wefoundthatthegas-bar(chambergasmanifold)designhadmajorproblems,particularly

tubeletinsertionandadjustment di–cultiesandthecreationofmetalchipsduringassembly

causingchamberleaks.Weproposedanewdesign, andmadeseveralgas-barprototypes

to demonstratethatthisavoidedtheproblemsexperiencedintheoriginaldesign.TheUS

ATLASmuoncollaborationhasaccepted thenewdesignandfabricateditatamuchlowercost

thantheoriginaldesign.Usingthenewgas-bars,w ehavesofarassembledandcerti?edone

EMS5and?veEMS4chambers.AllthesechambersmettheATLASgasleakspeci?cations.

Figure10showsthecomparisonofthe\old"and the\new"gas-bardesigns.

Figure 10:Comparisonofthe’old’(lefttwo)and the’new’(righttwo)gas-bardesign.The

oldonesconsistof2piecetubeletretainerwith splittaperedferruleandthreadedstopper,

theo-ringholesonthegas-bararethreadedholes (withangles).Thenewonesonlyhavea

simplecylinderwithonesimpleretainerbarper24tubelets,theo-ringholesonthegas-bar

arestraightholes.Inaddition,the o-ringsizeincreasedforsealrobustness.

2.4.2 Chambergassysteminstallationand test

AsindicatedinTable1,theb ottleneckofinstallingthechambergassystemistheFaraday

cage(FC)baseplate,whichmustbeinstalledonc hamberbeforethegassystem.Sincethe

arrivalinMichiganoftheFConSeptember10,wehaveworkedveryintensivelytoretro?tall

theEMS4chambersmadesinceJuly,2001 (7chambers),andwillhenceforthinstallchamber

gasservicesduringchamberproduction.EMS5cham bershavealreadybeenshippedtoCERN

(duetolackofstorageatMichigan)andwillberetro?ttedbefore?nalinstallationatCERN.

Gassystemshavebeeninstalledonsev enEMS4chambers,and?vehavepassedthecham-

berleakcerti?cationtestinthelastthreeweeks ofSeptember.Wehavebeenabletomake

suchsigni?cantquickprogresssincewehavefullydevelopedtoolingandhaveexperiencein

assemblyandtestingofthegassystem.

Wedesignedandbuiltagas-bar pre-assemblyandteststationshowninFigure11.The

jigprovidesanmock-upoftheendofasinglem ultilayer.Thetubeendplugsarerepresented

bymachined,solidbrassplugsarrayedaccording tothegeometryoftheglued-uptubesina

chamber.Thesedummyendplugsduplicatethe O-ringgroove,aswellasthethreadedrod

whichengagesthesignalcap.Themotivationforus todevelopthepre-assemblystationis

basedonthefollowingconsiderations:

† Itallowsforconvenientaccesstothe gas-barandalltubelets.

† Leakdebuggingisexpedited.Thesmall (700cc)volumeofthegas-barplustubelets

canbecerti?edforgas-tightnessinacoupleof hours.

† Gas-barscanbeassembledasaparallel tasktootherworkthatmustprecedechamber

mounting,suchasFaradaycageinstallation. Assemblyofgas-barscanthereforeproceed

accordingtoanunsteady(student)laborsupply andcanbedoneinlocationsotherthan

thechamberassemblysite.

† Pre-productionofpre-assembledand testedgas-barscaneasethealreadystringent

scheduledemandsontheproductionofacompletely instrumentedchamber.

Whenwewerew aitingforthedeliveryoftheFC,weusedourpre-assembly/teststationto

assembleandtest30gas-barsinthesummerof2001. Hence,wewereabletoretro?tthe

EMS4c hambersquicklyinSeptember.

After thegassysteminstallationthechamberasawholemustbeleakchecked-the

mostdi–cultprocessinchambertests.Duetothe largevolumeofachambers,meetingthe

ATLASleaktightnessspeci?cationrequiressigni?cante?ortandtime.Overthepastyear,

wehavebeenabletodevelopasensitive leakcheckingschemeusingprototypesofthegas

system.Tocertifyagastightchamber,itmustbe pressurized,scannedwithgasdetectors

alongthetubeends,andmonitoredforpressure dropsformanyhours.Themethoddeveloped

inMichiganusesaniterativeleakcheckingpro cedure,of\passes"thatemployincreasingly

sensitivedetectiontechniques.Ourtools includeinstrumentsforquicklydetectingcrudeleaks

toveryminuteones,lessthan10

¡6

mbarlsec

¡1

.Forthelatterweconsiderahelium mass

spectrometertobeessential. Figure12showsthe?nalleakdebuggingofanEMS4chamberusingaheliummassspectrometer sni?er.

Figure 11:Anoperatorassemblesagas-barina pre-assemblystation.Thissystemincludes

thegastestsystemtocheckthepre-assembled gas-bars.

Aftertheiterativeleakc heckingprocedure,theoverallchamberleakratemustbemea-

suredtocertifythatthechambermeetsATLASsp eci?cations.Theapproachweadoptis

tomeasurechamberpressureoveranextendedperiod(1-2days)usingeitherabsoluteor

di?erentialpressuresensors,orboth (forredundancy).Thechamberneedstobeisolatedin

thermallystableenvironmentwherethetemp eratureûuctuationsarelessthan1

–

K,andthe

temperatureismonitoredatseveralpointsonthem ultilayer.Figure13showsthechamberleakcerti?cationstationwithbothabsoluteor di?erentialpressuretransmittersandrefer-

Figure 12:Finalchamberleakdebuggingusinga heliummassspectrometersni?er.

ence chambersattachedtothetestchamber.Temperatureandpressurearemonitoredand

recordedbyacomputereveryminute.Usingthis station,weareabletocertifyachamberin

twodays.

Figure 13:Chamberleakcerti?cationstation. Thetwocylindersattachedtothetestchamber

arethereferencechambers.

Withourrecentintensiveexperienceinchamber leaktesting,wefeelthatweshouldbe

abletoincludethechamberserviceinstallationinourproductionandmaintainourpresent

rate of8daysperchamber.Nevertheless,it remainsachallengetomaintainfullproduction

ratenextyear.

3ChamberReadout

3.1MDTReadoutMultiplexerDevelopment

Thedevelopmentoftheon-chamber readoutmodulefortheMDTisoneoftheresponsibilities

ofMichigan’selectronicsgroup.Theprimary individualsengagedinthisworkareJ.Chapman,

PietroBinchi,andnumerousstudents.Ourdesignw orkisafollow-onfromsimulationsdone

inpreviousyears.TheVerilogcodeusedinthesim ulationwasadoptedandaprototype,the

CSM-0,wassynthesizedfromthiscode.Theelemen tsoftheCSM-0areshowninFigure14.

Theunitcontains18channelsofserialtoparallelreceivers,bu?eredstorageofdatafrom

theindividualchannels,andan18c hannelpollingmultiplexerthatassemblesevents.The

fullyassembledeventsaretransferredtoanother FIFOforreadoutbythedataacquisition

system.Ahardwareimplementationfollowedandtw enty?vecommerciallyfabricatedCSM-

0boardswerebuiltandtested.Fifteenofthesemodulesareinthe?eldatlocationsin

Japan,Europe,andtheUS.Successwith theCSM-0forreadoutoftheMDTchambersvery

satisfyingandaworkshoptoexhibitthechamberp erformancecharacteristicsasexaminedby

thesemanygroupsisplannedinNovember2001.Mic higanhasalsoprovidedadataacquisition

package,MiniDaq,foroperationoftheCSM-0and con?gureditwithtrackreconstructionand

displaysoftwarewrittenatHarvard.TheMiniDaq systemisimplementedwithauser-friendly

GraphicalUserInterface(GUI).Thissystemhas provenattractiveandseveralinstitutions

notoriginallyexpectingtoestablish electronicstoreadMDTchambersarerequestingCSM-0

units.Wehavebeenaskedtofabricatean additional10CSM-0boards.Wepreparedworkshop

trainingmaterialdescribingtheCSM-0,leda CSM-0workshopatCERNthispastyear,and

expecttoo?eranothercourseduringtheupcomingy ear.

AsecondprototypeCSMcalledthe CSM-1willbedesignedduringtheupcomingyear.It

willemploythenewestFPGAtechnologytocompress thedesignintoasinglechip,shrinking

itssizeandloweringitspowerconsumption.This newunitwillalsomoveclosertothe?nal

CSMwhichwilluseasingleFPGA.Withthisnew designtheelectricalenvironmentofthe

?nalsystemwillberealized.J.Chapmanis planningtoresideatCERNnextsummerto

guidetheuseoftheCSM-1intheH8testbeam.

Theambitiousdesignspeci?cations forthe?nalon-chamberCSMwillbecompletewhen

theCSM-1isoperational.This?naldesignisexp ectedto?tintoafootprintof80mm£

130mm £ 40mm.Itwillmultiplexupto18sourcesof datafrom432TDCchannelsat

25ns/32bitdataword,consumelessthan8wattsofpowerincludingthepowerneededby

the820Mb/sopticallink.Anotherimp ortantaspectofthedesignisthatitmustinclude

self-monitoringofpotentialradiationinduced singlebitupsetsandincludeamechanismto

reloadtheFPGAcodeshouldanyupsetsbefound. Thislatterrequirementimpliestheuse

ofûash-ramforfastrecon?gurationshouldan upsetbediscovered.

Worktodateonthe ?nalCSMincludesthedesignandfabricationofapassiveinterconnect

thatlinkssignalsfromtheASD/TDCboardstothe CSMandlinkstheCSMtotheELMB

en vironmentalmonitor.Inaddition,theVerilogHDLcodedevelopedfortheCSM-0isbeing

T ubes

H edgehog

M ezzanine

3 6

T TCem

Mux

JTAG

18 Mezzanine

Cards Maximum

CSM-0

Serial to Parallel

F IFO

V ME

Interface

P C

A dapter 36 pin to RJ45

P ower

T rigger/Clock

L ite & AMT-1

Figure 14:Ablockdiagramof theprototypeChamberServiceModuleforreadoutofafull

chamberofdrifttimes.

recon?guredtobeusedinthe ?nalCSM.ThisinvolvessettingallconstantsviaJTAGinstead

oftheVMEbasedinitialization usedintheCSM-0.Theothermajorchangefocuseson

compressingthe4FPGAdesignof theCSM-0intothenewlineofXilinxchipswherethe

functionscanbeprovidedbya singlechipoftheVirtexseries.Thissinglechipcanalsoassume

thefunctionofoutputdriver sendingdatatotheMRODviaa?berlink.Triggertimingand

controlisincludedontheCSMin theformoftheCERNdesignedchiptheTTCrx.Its

dataarrivesviaasecond?ber. Withthisdesignthemanylargechamberswithsigni?cant

electronicschannelscanbe fabricatedasself-containedunitsconnectedtothetriggeranddata

acquisitionsystemby2small?b erconnectors.Ablockdiagramofthe?nalCSMisshownin

Figure15whereitsconnections totheTTC,ELMB,andMRODcanbeseen.FiveMichigan

undergraduatesparticipatein theATLASelectronicsdevelopmentprojectsinsummer2001

withgreatsuccess.

3.2ChamberCosmicRayTest

Togainbetter understandingofMDTchamberperformanceandtoprovidedirectfeedback

intotheproductionprocesswe haveestablishedacosmicrayteststation.Thisfacilityal-

lowsachambertobeoperated withnominalATLASparametersusingcosmicraymuons.To

facilitatethesetestsa dedicatedTDCreadoutelectronicshasbeenfabricated.Theseelectron-

icsarereferredtoas \Mezzanine-lite"TDCcardswhosedataarereadoutviaaprototype

ChamberServiceModule,CSM-0. DataacquisitionisprovidedbytheMiniDAQprogram

developedlocally.

Theteststationisshownin Figure16andschematicallyinFigure17.Twolargescin-

T ubes

H edgehog

Mezzanine

4 0

T TCrx

M ROD

1 06.7 Mbyte/s

S -Link/G-Link

TTC

T TC Fibre

C lk, L1A, C alib

M ux

JTAG

M onitor

E LMB

Mux

1 8 Mezzanine

Cards Maximum

6 4 ADCs

E LMB

C AN

Bus

R OB

TTCvi

Central

Control

CSM

S erial to Parallel

X tmr

O ctal ASD

& AMT-2

Power

N ote: All connecti ons to the CSM are through a passive interconnect

except the TTC & MROD sy stems

I nterconnect board

Figure15:Ablock diagramofthe?nalChamberServiceModulewhichincludesconnection

totheTriggerTimingandCon trol,theMuonReadoutDriver,andtheenvironmentalmonitor

oftheDetectorControlSystem.

tillatorpaddles spanthedimensionofthechambertransversetothetubes(de?nedasthe

Ycoordinate).Astackoflead blockscenteredoverthebottomscintillatorprovidearegion

whereelectronsandsofterm uons(< 100MeV)canberejected fromthetrigger.Thescintil-

latorsareinstrumentedateach endwithphoto-tubes.Thetypicalfour-foldcoincidencerate

(bothPMTsfromeachend?ringin a50nswindow)is50Hz.

Thechambersareoperatedbyûo wingagasmixtureofAr(93%CO27%)at3bar

pressure,inparallelthrough eachdrifttubeatnominalrateof1/2volumeexchangeperday.

Thiscorrespondsforatypical EMS5chambertoaûowrateoforder250cc/minute.By

comparison,theleaktightness anMDTchamberisspeci?edtobe

VdP

=2 £ N £ 10

¡ 8

barlsec

¡ 1

HereNisthenumberoftubesinam ultilayerandV isthevolumebeingev aluated.Fora

m ultilayeritisgivenby V = N £ V

avg

,and V

avg

istheaveragetubevolume.Thec hamber

pressureis monitoredattheinputandoutputofthegasûowsystemwhilethetemperatureis

measuredat4locationsonthec hamber.Theoutputûowrateissetbyanelectronicmassûow

controller.Thepressure,or morepreciselyP=T issetbymodulatingthe inputûowcontroller.

Thispressureregulationisbya feedbackloopwhichusesthe?rstandsecondderivativesof

P=T .

TheHVismaintainedinparallel acrossalltubesinbothmultilayers.Itissetat3080V

whichnominallycorrespondsto againof2£ 10

forourgasmixture.Typical currentsare

onthe < 1microamppermultilay er.

Figure 16:EMS5MDTChamberinCosmicRayTest Setup

0 -87 Td(ET Q0 0 1 rgq 10 0 0 10 0 0 cm BT/A 1 Tf0.1 0 0 -0.1 231.2 276.5 Tm(

LEAD BRICKS

PMT SCINTILLATOR PADDLE

MDT CHAMBER

Figure 17:SchematicofCosmic RayTestSetup

Wehav econductedpreliminarytestrunswiththeobjectiveofunderstandingnoisesources

andthebehaviorof thereadoutelectronics.Wehavebeenabletooperatethechamberat

thespeci?edATLASMDT thresholdcorrespondingtoabout20electrons.Atthisthreshold

mostchannelshaveamodestorlo wnoiserate-muchlessthan1Hz.

Figure 18:MuonTrackobservedintestofEMS5MDT Chamber

Theanalysispackageusedtop erformpreliminary(andcrude)trackingis MiniDana writ-

tenbyGregNovackatCERN.Itprovidesadefault lineartime-to-spacefunction,butcontains

thetoolsnecessarytoderivecorrect R ¡ T relationsfromthedatagivenuniform irradiation

ofthetubes.

Fora15minutetestrunwehavebeenabletoobserve the?rstmuontracksinanEMS5

chamber. AtypicaltrackisshowninFigure18.Theradiiofthecirclescorrespondtothe

measuredimpactparameterofthetrack.Thisimage demonstratesthebasicfunctionalityof

thechamber.

Weanticipatecompletingasequenceofcosmicray runswiththistestapparatus.Our

primaryobjectiveswillbetocharacterizethechambertubee–ciencyversusresolutionand

resolutionasafunctionofimpact parameter.Further,weintendtostudytuberesolutionas

afunctionofdistancefromtubeends.Towardsthe tubecenter,wherethecatenarysagittais

greatest,isfoundthegreatestdisplacementof thewirefromthetubeaxis.Thisinturnmay

causeaslightasymmetryinthe R ¡ T relationforobliquecosmicrays.Wewill investigate

anye?ectonthe resolutiontothise?ect.

4Computing

4.1Overview

TheUniversityofMichiganATLASGrouphasbeeninv olvedinseveralcomputinginitiatives

overthepastyear.Thesehavebeendirectedtoward supportofUSATLASprioritiesin

GRID testbedstudies,thedevelopmentanddeploymentofcollaboratorytoolsfordesignated

USATLASprojects,andenhancingourlo calnetworkingcapabilitiesinsupportofATLAS

andTevatronexperiments.

Inadditiontoastrongcollectionofindividuals alreadyintheMichiganATLASgroup,

including JayChapman,BingZhou,HomerNeal,Edward Diehl,ZhengguoZhao,Myron

Campbell, ShawnMcKee,DanLevin,StevenGoldfarb,TieshingDai,andEricMyers,Bill

Martinisarecentaddition.Heisaprofessorofn uclearengineeringattheUMandhasdeep

researchinterestsandexperiencein computationalphysics.Martinhasbeenveryactiveover

thepasttwodecadesinMonteCarlomethodsdev elopmentandtheuseofhighperformance

computersinscienti?ccomputing.Heisthe foundingdirectoroftheUM’sLaboratoryfor

Scienti?cComputationandwasthefounding directoroftheUniversity’sCenterforParallel

Computing(CPC).Heiscurrentlydirectorofthe UM’sNPACI(NationalPartnershipsfor

AdvancedComputationalInfrastructure)grant,whichwillplayamajorroleinassuringthe

successofourATLAScomputinge?orts atMichigan.

GridcomputingforATLAS- UMisoneofthetestbedsitesforGridComputinginthe

USATLASGride?ortandhasrecentlyhostedagrid testbedmeetingforUSATLAS.Asour

w orkevolveswehopetomergethee?ortsoftheUMATLASphysicistswiththoseinvolved

inthegride?ortattheCPC.Theresulting synergybetweenPhysics,CPCandNPACIwill

resultinverysigni?cantgride?ortforATLAS. Thissynergyshouldformthebasisforan

eventualUSATLASTier2centeratMichiganandpro videacriticalconnectionbetweenTier

2developmentandthegridcomputingmodel.

EricMyersandShawnMcKeehavebeencen trallyinvolvedintheongoingworkofthe

USATLASGRIDTestbedWorkingGroup.AtMichiganwe havedevelopedanddeployed7

testbed relatednodes.TheGlobusandCondorworkingenvironmenthasbeenestablished.

Ourworkto-datehasinvolvedthedevelopmentand testingofsoftwaretomonitoranddeploy

asecuregridenvironment.Extensiveworkhasbeen doneonestablishingpropersecurityfor

ourgridtestbednodesaswelladevelopingnetwork testingandmonitoringtoolsfortheUS

ATLASgridtestbed.

4.2 LeadershipRoleinUSATLASNetworking

ShawnMcKeehasbeenaskedtotakeonaprojectmanagementroleasleaderoftheUSATLAS

networkinge?ort.Shawnwillserveas primarycontactfornetworkissuesinUSATLASand

willde?nethenetworkingneedsfortheproject. The?rstdraftofnetworkingissuesandplans

hasbeendevelopedaspartoftheUSATLASGridF acilitiesPlan.

LeadershipRoleinthe LHCCommonProjectsAnewworkinggroupwithinInternet2is

beingformedaspartoftheLHCCommonProjectsw ork,andwillbeco-chairedbyHarvey

Newman(CalTech)andShawnMcKee(UniversityofMichigan).The?rstorganizational

andstatusmeetingwillbeheldonOctob er1aspartoftheInternet2Membermeeting.The

primarygoalistoinsurethatInternet2willpro videtheneedednetworkingcapabilitiesfor

highenergyandnuclearphysics.HomerNeal(Univ ersityofMichigan)andHarveyNewman

(CalTech)servesasmembersoftheInternet2ApplicationsStrategyCouncil.

4.3ATLASMuonDatabaseCoordination

StevenGoldfarbhasactivelycontinuedhisp ositionastheATLASMuonDatabaseTask

Leader,coordinatingthedevelopmentofthedatabasesoftwareinfrastructureandthein-

tegrationofthissoftwarewiththe o†ineapplicationsandtheATLASsoftwareframework.

Overthepastyear,signi?cantprogressasbeen madeinthedomainsoftheEventDataModel

and theDetectorDescriptionandworkhasb eguntowardde?ningtherequirementsofthe

conditionsdatabase.

RegardingtheEventDataModel,softwareclasses providingaccesstothesimulateddigits

andhitsfortheMDT,RPCandTGCdetectorsarenow allcomplete.Theseclassesde?ne

anin terfaceforthesoftwareapplicationstotheentiresetofATLASPhysicsTDRdataand

arecurrentlyinusebytherecentlydevelopedm uonobject-orientedreconstructionprogram,

calledMOORE.Accesstothehits,aswellasthe digits,allowsfordigitizationstudieswithin

thesoftwareframework,ATHENA.Inaddition,p ersistentcyclassesandschemahavebeen

developedforthestorageandretrievalofthe digitstoandfromanObjectivitydatabase.

TheseclassesarecompletefortheMDTdigitsand nearingcompletionfortheRPCand

TGC digits.Followingthis,alloftheATLASPhysicsTDRdatawillbeaccessiblefroman

Objectivitydatabaseforfuturestudiesandev aluationofthetechnology.

Forthe DetectorDescription,e?ortshavefocusedonthedevelopmentofacompletede-

scriptionoftheMuonSpectrometerusingtheATLAS GenericDetectorDescription(AGDD).

TheUniversityofMichiganhasplayedaleadingrollnotonlyinthee?orttoporttheexisting

muondetectordescription(AMDB)toA GDD,butalsointhedevelopmentofthesyntax

ofthelanguageandinthedesignanddevelopmentof theunderlyingsoftwaremodel,which

providestheinterfacetotheapplications.Thesee?ortsaredetailedinthefollowing section.

4.4ATLASDetector DescriptionActivities

Steven GoldfarbandShawnMcKeehavemadeasigni?cantcontributiontothedesign,de-

velopmentandtestingofAGDDfortheMuonSp ectrometer.Thegoalofthise?ortisto

provideasinglesourceofdetectordescriptionforallapplicationsinacoherentandmain tain-

ablemanner.Theurgencyofthis e?ortisunderlinedbythefactthattheATLASSoftware

ManagementandDatabaseCoordinationhaverecen tlyidenti?eddetectordescriptionasbeing

onthecriticalpathforthedevelopmentofthe o†ineapplicationsandthesoftwareinfras-

tructure.Theire?ortshavefocusedonthe translationoftheexistingAMDBdatabaseto

AGDDandontheevaluationofthesoftwarein terfacetothemuon-speci?capplications,such

assimulationandreconstruction.

Thispastyear,ShawnMckeedevelopedasetof scriptstoautomaticallyconverttheMDT,

RPCandTGCdescriptionsfromtheAMDBdatabaseto thebasicAGDDsyntax.Thiswork

ledto severaldetailedstudies,coordinatedwiththeMuonSimulationandReconstruction

GroupsontheusageofAGDDforthe applications.Itwasdeterminedfromthesestudiesthat

AGDDissu–cientforgeneralapplications,suchas visualization,andiscapableofgenerating

GEANT4geometriesforsimulation.However,the genericsyntaxisnotoptimalforthe

futuremaintenanceofthedescriptionsbydetectorexpertsandthesoftwareinterfacedoes

notprovidethesimulationdevelopers withtheabilitytoexploittheoptimizationmechanisms

providedwithGEANT4.Toaddressthese conclusions,StevenGoldfarbproposedanextension

mechanismtoAGDDallowingforthedevelopmentof detector-speci?csyntaxandproviding

asoftwareinterface,whichcanbetunedtothe needsoftheapplications.

E?ortsare nowconcentratingonthedevelopmentofthenewsyntaxandtheevaluation

ofthesoftwareinterfacewiththeapplications. StevenGoldfarbhasdevelopedthesyntaxfor

theMDT,RPCandTGCdetectorsandthesoftwarein terfacefortheMDTchambers.Je?

Giansiracusa,aUniversityofMichiganREUSummerStudentatCERN,developedthesyntax

and softwareinterfaceforthecomplete BarrelToroidsystem,asasummerproject(avery

positiveindicationofthesimplicityande?ectiv enessoftheextensionmechanismtoAGDD).

Evaluationoftheresultingdescriptionsfor GEANT4arebeingcoordinatedwiththeMuon

SimulationGroupandarebasedonsimulatingthe geometryoftheMDTEndcapTestBeam

tob eheldthisSeptemberatCERN.E?ortsarealsounderwaytotestthedescriptionsinthe

MuonboxandMOOREreconstruction programsandarebeingcoordinatedwiththeMuon

ReconstructionGroup.[BillMartinandhis colleagueshaveexpressedinterestincontinuing

thedevelopmentoftheinertmaterial description,includingtheremainingmagnetandsupport

structures,inthecomingyear.Thiswillprovidea valuableservicetotheongoingevaluation

ofAGDD,whileprovidingtheirsoftwareteamto withabetterunderstandingoftheATLAS

o†inesoftwareenvironment.]Inparalleltotheev aluations,wearecontinuingwiththetrans-

lationofAMDBtoAGDD,andacompletedescription oftheMuonSpectrometer,including

thesurroundingmagnetsandsupportstructureistargetedfortheendof2002.

4.5HEP-WideDetectorDescription Activities

InadditiontotheATLAS activitiesdescribedabove,StevenGoldfarbhascontributedtothe

developmentofcommonHEP-widesoftware strategiesfordetectordescription.In2000and

2001,hehostedHEPDD(HEPDetectorDescription)w orkshopsatCERNandLBNL.Par-

ticipan tsincludedsoftwarecoordinatorsanddevelopersfromthefourLHCexperiments,the

GLASTandLCDexperiments,andthe GEANT4,ROOT,andWIREDsoftwarecollabora-

tions.Theseworkshopshavebeenanimportant forumforthesharingofdetectordescription

softwaremethodsanddatabasedesign,andhavepro videdvaluableinputtotheATLASef-

fort.Inaddition,theyhaveledtoseveraljointe?orts,includingthedevelopmentofa common

syntaxforde?ningmaterials andthede?nitionofgeometrymark-uplanguage,calledGDML.

4.6UniversityofMichiganNetwork Infrastructure

TheUniversityofMic higanhasagoodnetworkinfrastructure,butitisnotsu–cientinits

currentformtomeettheanticipatedneedsofour programforthedataintensiveTevatron

andLHCexperiments.Weareactivelyinvolvedin seeingthatthenecessaryinnovations

aredevelopedtoinsurethatourneedsaremet.Wearepresentlyexploringagigabit?ber

backboneasapossiblesolution.Wehav eacquiredahighendlayer3switchandexpectto

connecttothegigabitbackboneinthenextfeww eeks.Thiswillallowhighbandwidthgrid

testingbetweensimilarlycapabletestbedsites, andwillimmediatelyincreasethee?ective

bandwidthbetweenUMandFermilab.Also,thiswill enableQoStestingandmonitoringto

CERN.Weviewthisworkasbeingcriticaltotheevolutionofourlocalcapabilitiesandto

advancingthetechnologiesneeded ultimatelyforGRIDcomputing.

LeadershipintheDevelopmentandDeploymentofCollaboratoryToolsforUSATLAS

HomerNealandEricMyers,withthesupportofUS ATLASandtheNSF,undertookthe

taskof demonstratingthevalueofwebbasedarchivingfortargetedsoftwaretraining.The

initialATLASpilotprojectwasonedesignedto deliveraweb-basedcourseonGEANT4.This

newapplicationwillbecriticaltoallsimulation studiesinATLAS,butitisunderstoodby

onlyaverysmallfractionofthecollaboration. Andreadell’Acquaistheprincipalexpertin

ATLASonGEANT4andhedeliveredinAnnArboraweek longworkshoponthisapplication

whic hwerecorded,processedandhavenow releasedtotheentirecollaboration.Itsusefulness

isbeingevaluated.Ournextprojectwilllikelybe conductingasimilare?ortfortheoverall

ATLASsoftwarearchitectureATHENA.Inaddition, wehavebeenaskedbyATLASleadersto

consideredprovidingwebarchivedlecturesonC++,CMT,andanATLASMuonElectronics

tutorial.Wehavebeenactivelyinvolv edinthedevelopmentofthesoftwareapplications

requiredforhighqualitywebarchiving,aswellas deployingtheseapplicationsforpurposes

suchasGEANT4training.

4.7MDTProductionDatabase

MuonChamberProductionDatabaseThelastyearhas beenatransitionfromtubeand

chamber constructionprototypingtoproduction.Thedatabasesystemhassimilarlygrown

toencompasschamberassemblyandtestingdata collectionaswell.Wenowhave10new

primarystationscomprisingourMDTchamberproductionsequenceaswellasanumberof

extrameasurementstations.

† Gluingstation

† Sti?-backreadoutstation

† Spacerassemblystation

† Spacerreadoutstation

† PMOinstallationandcalibrationstation

† Jiggingsetupandmeasurementstation

† Chamberleaktestingstation(indev elopment)

† Chambercosmic-rayteststation(indev elopment)

† Chamberelectronicsteststation(indev elopment)

† Finalchambercheckliststation(indev elopment)

Alongwiththenewstations areanumberofnewcapabilitiestoourproductiondatabase

system:

† Productionandpartsinventory automaticallytrackedandprovidedviaWWW

† Inventoryrecording,modi?cation,and displayviaWWW

† Detailedproductiongraphshotlinkedto detailedinformation

† Additionalseparatebackupofindividual SQLqueriesandviews

† Perchamberdatabasescreatedandav ailablefordownload

† ChamberchecklistimplementedviaWWWin terface

† Automatedemailnoti?cationofproblems duringdatabaseupdateprocessing

† Primaryproductionmanagersnoti?edof productionproblemswhichoccurredduring

thepreviousdayautomaticallyviaemail

† Data\reconcile"interfacepageviaWWW, ?ndspotentialdataproblemsduetofaulty

entry

Inadditiontothedevelopmentworkduringthelast year,ShawnMcKeehasbeeninvited

to helpanumberofotherATLASinstitutionsimplementcomponentsofthesystemhehas

developed.Thisincludesbeinginvitedto Rome,theinstitutionresponsiblefortheglobal

ATLASMDTproductiondatabase,Cosenzaand,in September2001,Beijing.

Nextyearwe willcontinuetoimproveandupdatetheproductiondatabase,especiallyas

we?nalizethechambertestingprocedures. Emphasiswillbeonchamberlevelinformation.

AsmentionedintheTaskHsummarywehaveupgraded allproductionstationstoWindows

2000andnewhardware:AllprimaryproductionstationsnowruninexpensiveIDERAID1

arraysastheirprimarybootanddata disks.Allprocessorsnow600-900Mhzandmemory

atleast128MB.Partofthefocusforthecomingyear willbetoestablishmoresecureand

reliablenetworkconnectionsbetweenallproductionstationsandimprovethebackup system

wehavedeveloped.

Animportantpointtorememberisthatthepro ductiondatabasesystemiscloselytied

toourproductionprocess.Itisuseddailyto monitorproductionandtoidentifyproblems

asearlyaspossibletoavoidwastingtimeor material.Ofcourseitalsoservestoprovide

long-termdocumentationforallcomponentsthatw ewillbuildforATLAS.

4.8MDT SystemSimulationStudies

Overthe pastcoupleofyearstheUMgrouphasconductedaseriesofsimulationstudies

whoseobjectivehasbeentobeoptimizeaspectsofm uondetectorconstruction,stagingand

alignment.ArecentLehmannreview(March2000) forwardedarecommendationthatUS

A TLASconstructfewerthanthe320MDTchambersand328-layerCSCchambersdescribed

intheMDTTechnicalDesignReport.Anearlier studycompletedlastyearbytheUMgroup

focusedonoptimalmeanstostagetheMDTchambers. Thisyearweproceededwithtwo

newsim ulationstudies:The?rstinvestigatestheimpactonCSCsystemperformancewhen

itisreducedinscopefromdoublelayerto singlelayer.Asecondstudyexploresthemuon

momentumresolutionfornon-optimallyalignedc hambers.

ToaddressCSCstagingwehave generatedaseriesofMonteCarlosimulationsandtrack

reconstructionsundertwoscenarios,namely: withthefullbaselinedesignemployingthefull

CSC(doublet)andusingonlyaonelayerCSC (singlet).ForthesingletcasethefrontCSC

layerwasexcised.

Duetotheneedtorapidlyproduceanddisseminate thisstudywehavechosentorunwith

standaloneversionsoftheMDTsimulation,LHCTOR[4]andreconstruction,MUONBOX

[5]codes.TheformerisaGEANTbasedprogram whichsimulatingtheMDTsensitivede-

tectorsandthemasspro?lesoftheinternaldetectors,magnetandotherstructuralelemen ts.

TherunparametersofLHCTORareset toenable– -rayproductionandallstandardphysics

processes.Gammasandelectronswere trackedto0.5Mev,muonsto1MeV.MUONBOX,

after itprovidesdigitizationoftheLHCTORpro ducedhits,performstrackandmomentum

reconstruction.Theanalysisparametersof MUONBOXareintrinsictuberesolution85 „m;

randomnoisehitsturnedo?.Theactualtube resolutioncodedinMUONBOXisparame-

terizedtoyieldanimpactparameterdependentvaluewhichisthensampledfromanormal

distribution.Thegeometrical databaseusedisVersionKalongwithVersion1ofthemagnetic

?eld.Thisgeometricaldescription,althoughnot themostrecent,iscompletelyadequatefor

thepurposesofthisdi?erentialtypestudy.A crosscheckofthebaselinecon?gurationus-

ingDICEandATRECON(release39)(whichcalls MUONBOX)employingthegeometrical

descriptionVersionM(amdb

simrec.M2.8)andmagnetic?eldVersion2 (bmagatlas02.data)

yieldedvery similarreconstructionratesandresolution.

TheresultsofthisworkwillbepublishedasanA TLASmemo.

Thesimulationruns consistedof5000generatedmuons(„

and „

inequalnumbers)with

transversemomenta, P

,setat20,100or500GeV.Inallrunstheangular rangespanned360

degreesazimuthanda rapidityinterval1: 9 <·< 2 : 7

Asecondsetofrunswasperformedinwhichinwhich SMH ! 4 „ andSUSY A ! 2 „

decayswereallowedtopropagate.The reconstructedHiggsmasswasthenusedasbenchmark

togaugetheimpactofthedi?erentstaging con?gurations.

4.9CSCStaging Summary

Forsinglemuonswehavefound that1)comparedtothebaselinedetectorcon?guration(dou-

bletCSC),the50%reducedCSCdetectorhas » 1-3%lowerreconstructione–cienciesov er

muontransversemomentumrange20 GeV-500GeV;2)Thereconstructedmuonmomentum

resolutionisnotsigni?cantlydegraded, particularlyformuonP

lessthan100GeVforthesin-

gletCSCcon?guration;3)Thenon-Gaussiantails inthereconstructedtransversemomentum

distributionsareslightlyworseforthesinglet CSCcon?gurationcomparedtothebaseline.

Forreconstructionof4muondecaysof150GeV Higgs,we?ndthatthedoubletCSC

con?gurationyieldsa4„ reconstructione–ciencyof69.2%whereas thesingletCSCcon?g-

urationhasan e–ciencyof68%.ForSUSYHiggs(2muondecays)theoveralle–ciencyis

higher-butthedi?erencebetweenthedoubletand singletdi-muonreconstructione–ciency

remainssmallat90.6%and89.5%respectively.The changeinmassresolutions(2and4muon

decays)isvirtuallynil.

Weconclude thattheUSCSCmuondetectorstagingplandescribedinthisstudywill

notseverelyimpacttheHiggsbosondetectionson the?rstdayoftheLHCphysicsrun.

How ever,signi?cantmuondetectione–ciencydegradation(upto3%)withthereducedCSC

chambercon?gurationimpliesthatfor theLHChighluminosityruns,particularly,forheavy

newgaugeboson(forexample, W

, Z

)detections,the2ndCSC-singletupgradeshould have

highpriority.

4.10MDTAlignmentStudies

TheALTASMDTsystemisdesignedtobeinternally monitoredbyover10,000Rasniklight

b eams(codedtargets)thatwillprovidecontinuousfeedbackonchamberdistortionsandpo-

sitioning.Thissystemisessentialto maintainthebestresolutionduringperiodsofdata

acquisition.Yet,thereareanumberofwaysinwhic huncertaintiesinchamberplacementcan

o ccur:Theremaybeuncertaintyintheinitialchamberplacement,surveying,andorien tation;

TheRasniksystemmayhaveanyn umberofpartialmalfunctionsdoto?rmwareand/orme-

chanicalfailures;therecanbechambermotions thatarenotproperlytrackedbytheRasnik

imagers;andtheremaybeunrecoverabletracking uncertaintiesduetonon-parallelchambers

evenwhenthechamberorientationiswellknown.

Toevaluatethesepossibilities,we conductedsinglemuonresolutionandreconstruction

e–ciencystudiesoveralimitedrangeofuncertain tyinMDTchambersangularorientation.

Thetoolsusedinthisworkaresimilartothose utilizedfortheCSCstagingstudies.

F orthisstudywegeneratedasequenceofrunscontaining5000singlemuonsofPt=20

and100GeVoverrapidity1 <·< 2.Thedigitizationofthetracksassumed thetrue,

thatis,misalignedchamb ers.Thereconstructionassumed,conversely,thatthechambers

wereintheiridealizedlocationsandproperly orientedwithrespecttotheirneighbors.The

misalignmentsweredonebyrandomlyrotatingeach chamberaboutaspeci?edaxes.We

allow edformaximumangulardisplacementsof0.33,0.67,1.0and1.33mradaboutthree

rotationalaxes.Theseaxes,S,T,Zcorrespond respectivelytothetubes,alongthebeamand

transversetothetubes.

4.11MDTAlignmentConclusions

AtPt=20GeVtheresolutiondegradationforamaxim um0.33mraduncertaintyintheS,T

andZ axisisnegligible.At1.0mradaboutSandTthereisob