CERNLHCC 2001-010LHCb TDR 428 May 2001

LHCb

Muon SystemTechnicalDesignReport

Printedat CERNGeneva,2001

ISBN 92-9083-180-4

ii

The LHCb Collaboration1

Brasilian Center for particles physics,CBPF, Rio de Janeiro, BrasilP.R.BarbosaMarinho,I.Bediaga,G.Cernicchiaro,A.Franca Barbosa,J.Magnin,J.MarquesdeMiranda,A.Massafferri, A.Reis,R.Silva

University of Rio de Janeiro, UFRJ, Rio deJaneiro, BrasilS.Amato,P.Colrain,T.daSilva,J.R.T.deMello Neto,L.dePaula,M.Gandelman,J.H.Lopes,B.Marechal,D.Moraes,E.Polycarpo

University of Clermont-Ferrand II, Clermont-Ferrand, FranceZ.Ajaltouni, G.Bohner, V.Breton,R.Cornat,O.Deschamps,A.Falvard1

, P.Henrard,J.Lecoq,P.Perret,

C.Rimbault,C.Trouilleau,A.Ziad

CPPM Marseille, Aix University-Marseille II, Marseille, FranceE.Aslanides, J.P.Cachemiche, R.Le Gac, O.Leroy, P.L.Liotard, M.Menouni, R.Potheau,A.Tsaregorodtsev, B.Viaud

University of Paris-Sud,LAL Orsay, Orsay, FranceG.Barrand,C.Beigbeder-Beau, D.Breton, T.Caceres,O.Callot, Ph.Cros,B.D’Almagne, B.Delcourt,F.Fulda Quenzer, A.Jacholkowska1

, B.Jean-Marie,J.Lefranc¸ois, F.Machefert, V.Tocut, K.Truong,

I.Videau

TechnicalUniversity of Dresden,Dresden,GermanyR.Schwierz,B.Spaan

Max-Planck-Institute for Nuclear Physics,Heidelberg, GermanyC.Bauer, D.Baumeister, N.Bulian, H.P.Fuchs, T.Glebe, W.Hofmann, K.T.Knopfle, S.Lochner,M.Schmelling,B.Schwingenheuer, F.Sciacca,E.Sexauer2

, U.Trunk

PhysicsInstitute, University of Heidelberg, Heidelberg, GermanyS.Bachmann,P.Bock, H.Deppe, F.Eisele, M.Feuerstack-Raible,S.Henneberger, P.Igo-Kemenes,R.Rusnyak,U.Stange,M.Walter, D.Wiedner, U.Uwer

Kir chhoff Institute for Physics,University of Heidelberg, Heidelberg, GermanyV.Lindenstruth,R.Richter, M.W.Schulz,A.Walsch

Laboratori Nazionali dell’INFN, Frascati, ItalyM.Anelli, G.Bencivenni, C.Bloise,F.Bossi,P.Campana,G.Capon,P.DeSimone,C.Forti, A.Franceschi,F.Murtas,L.Passalacqua,V.Patera(1),A.Saputi,A.Sciubba(1)(1) alsoat Dipartimentodi Energetica,Universityof Rome,“La Sapienza”

University of Bolognaand INFN, Bologna,ItalyM.Bargiotti, A.Bertin, M.Bruschi,M.Capponi,I.D’Antone, S.deCastro,P.Faccioli, L.Fabbri,D.Galli,B.Giacobbe, U.Marconi, I.Massa, M.Piccinini, M.Poli, N.Semprini-Cesari,R.Spighi, V.Vagnoni,S.Vecchi,M.Villa, A.Vitale,A.Zoccoli

1This list includesadditionalcolleagueswhomadeparticularcontributionsto thework presentedin this TDR.

iii

University of Cagliari and INFN, Cagliari, ItalyS.Cadeddu,A.Cardini,M.Caria,A.Lai, D.Marras,D.Pinci,B.Saitta(1)(1) alsoat CERN

University of Ferrara and INFN, Ferrara, ItalyV.Carassiti,A.CottaRamusino,P.Dalpiaz,F.Evangelisti,A.Gianoli, R.Malaguti,M.Martini, F.Petrucci,M.Savrie

University of Florenceand INFN, Florence,ItalyA.Bizzeti, M.Calvetti, G.Collazuol,E.Iacopini,G.Passaleva,M.Veltri

University of Genoaand INFN, Genoa,ItalyS.Cuneo,F.Fontanelli,V.Gracco,P.Musico,A.Petrolini,M.Sannino

University of Milano-Bicoccaand INFN, Milano, ItalyM.Alemi, T.Bellunato(1),M.Calvi, C.Matteuzzi,M.Musy, P.Negri, M.Paganoni(1) alsoat CERN

University of Rome,“La Sapienza”and INFN, Rome,ItalyG.Auriemma(1),L.Benussi,V.Bocci,C.Bosio,G.Chiodi,D.Fidanza(1),A.Frenkel,K.Harrison,I.Kreslo,S.Mari(1),G.Martellotti,S.Martinez,G.Penso,G.Pirozzi(1),R.Santacesaria,C.Satriano(1),A.Satta(1) alsoat Universita dellaBasilicata,Potenza

University of Rome,“Tor Vergata” and INFN, Rome,ItalyG.Carboni,D.Domenici,G.Ganis,M.Gatta,R.Messi,L.Pacciani,L.Paoluzi,E.Santovetti

NIKHEF , The NetherlandsG.van Apeldoorn(1,3), N.van Bakel(1,2), T.S.Bauer(1,4),J.van den Brand(1,2), H.J.Bulten(1,2),M.Doets(1), R.van der Eijk(1), I.Gouz(1,5), V.Gromov(1), R.Hierck(1), L.Hommels(1), E.Jans(1),T.Ketel(1,2),S.Klous(1,2), B.Koene(1),M.Merk(1), M.Needham(1),H.Schuijlenburg(1), T.Sluijk(1),J.vanTilburg(1),H.deVries(1),L.Wiggers(1)(1) Foundationof FundamentalResearchof Matterin theNetherlands,(2) FreeUniversityAmsterdam,(3) Universityof Amsterdam,(4) Universityof Utrecht,(5) on leave from Protvino

Institute of High Energy Physics,Beijing, P.R.C.C.Gao,C.Jiang,H.Sun,Z.Zhu

Research Centre of High Energy Physics,Tsinghua University, Beijing, P.R.C.M.Bisset,J.P.Cheng,Y.G.Cui,Y.Gao,H.J.He,Y.P.Kuang,Y.J.Li, Q.Li, Y.Liao, J.P.Ni, B.B.Shao,J.J.Su,Y.R.Tian,Q.Wang,Q.S.Yan

Institute for Nuclear Physicsand University of Mining and Metalur gy, Krak ow, PolandE.Banas,J.Blocki,K.Galuszka,L.Hajduk,P.Jalocha,P.Kapusta,B.Kisielewski, W.Kucewicz, T.Lesiak,J.Michalowski, B.Muryn, Z.Natkaniec,W.Ostrowicz, G.Polok,E.Rulikowska-Zarebska,M.Stodulski,M.Witek, P.Zychowski

iv

Soltan Institute for Nuclear Physics,Warsaw, PolandM.Adamus,A.Chlopik, Z.Guzik,A.Nawrot, M.Szczekowski

Horia Hulubei-National Institute for Physicsand Nuclear Engineering (IFIN-HH), Bucharest-Magurele,RomaniaD.V.Anghel3

, C.Coca, A.Cimpean, G.Giolu, C.Magureanu,S.Popescu,T.Preda, A.M.Rosca(1),

V.L.Rusu4

(1) alsoat HumboltUniversity, Berlin

Institute for Nuclear Research (INR), Moscow, RussiaV.Bolotov, S.Filippov, J.Gavrilov, E.Guschin,V.Kloubov, L.Kravchuk, S.Laptev, V.Laptev, V.Postoev,A.Sadovski, I.Semeniouk

Institute of Theoretical and Experimental Physics(ITEP), Moscow, RussiaS.Barsuk,I.Belyaev, A.Golutvin, O.Gouchtchine,V.Kiritchenko, G.Kostina,N.Levitski, A.Morozov,P.Pakhlov, D.Roussinov, V.Rusinov, S.Semenov, A.Soldatov, E.Tarkovski

Budker Institute for Nuclear Physics(INP), Novosibirsk, RussiaK.Beloborodov, A.Bondar, A.Bozhenok, A.Buzulutskov, S.Eidelman, V.Golubev, S.Oreshkin,A.Poluektov, S.Serednyakov, L.Shekhtman,B.Shwartz,Z.Silagadze,A.Sokolov, A.Vasiljev

Institute for High Energy Physics(IHEP-Serpukhov), Protvino, RussiaL.A.Af anassieva, I.V.Ajinenko, K.Beloous, V.Brekhovskikh, S.Denissov, A.V.Dorokhov,R.I.Dzhelyadin, A.Kobelev, A.K.Konoplyannikov, A.K.Likhoded, V.D.Matveev, V.Novikov,V.F.Obraztsov, A.P.Ostankov, V.I.Rykalin, V.K.Semenov, M.M.Shapkin, N.Smirnov, A.Sokolov,M.M.Soldatov, V.V.Talanov, O.P.Yushchenko

Petersburg Nuclear PhysicsInstitute, Gatchina, St.Petersburg, RussiaM.Borkovsky, B.Botchine,S.Guetz,V.Lazarev, N.Saguidova, E.Spiridenkov, I.Velitchko, A.Vorobyov,An.Vorobyov

University of Barcelona,Barcelona,SpainR.Ballabriga(1),S.Ferragut,Ll.Garrido, D.Gascon,S.Luengo(1),R.Miquel5

, D.Peralta,M.Rosello(1),

X.Vilasis(1)(1) alsoat departamentd’EngineriaElectronicaLa Salle,UniversitatRamonLlull, Barcelona

University of Santiagode Compostela,SantiagodeCompostela,SpainB.Adeva, P.Conde, F.Gomez, J.A.Hernando, A.Iglesias, A.Lopez-Aguera, A.Pazos, M.Plo,J.M.Rodriguez,J.J.Saborido,M.J.Tobar

University of Lausanne,Lausanne,SwitzerlandP.Bartalini, A.Bay, B.Carron,C.Currat,O.Dormond,F.Durrenmatt,Y.Ermoline, R.Frei, G.Gagliardi,G.Haefeli, J.P.Hertig, P.Koppenburg, T.Nakada(1), J.P.Perroud, F.Ronga, O.Schneider, L.Studer,M.Tareb,M.T.Tran(1) alsoat CERN,on leave from PSI,Villigen

v

University of Zurich, Zurich, SwitzerlandR.Bernet,E.Holzschuh,P.Sievers,O.Steinkamp,U.Straumann,D.Wyler, M.Ziegler

Institute of Physicsand Technologies,Kharki v, UkraineS.Maznichenko, O.Omelaenko, Yu.Ranyuk

Institute for Nuclear Research, Kiev, UkraineV.Aushev, V.Kiva, I.Kolomiets,Yu.Pavlenko, V.Pugatch,Yu.Vasiliev, V.Zerkin

University of Bristol, Bristol, U.K.N.H.Brook,J.E.Cole,R.D.Head,A.Phillips, F.F.Wilson

University of Cambridge, Cambridge, U.K.K.George,V.Gibson,C.R.Jones,S.G.Katvars,C.Shepherd-Themistocleous,C.P.Ward,S.A.Wotton

Rutherford Appleton Laboratory, Chilton, U.K.C.A.J.Brew, C.J.Densham,S.Easo, B.Franek, J.G.V.Guy, R.N.J.Halsall, J.A.Lidbury, J.V.Morris,A.Papanestis,G.N.Patrick,F.J.P.Soler, S.A.Temple,M.L.Woodward

University of Edinburgh, Edinburgh, U.K.S.Eisenhardt,A.Khan,F.Muheim,S.Playfer, A.Walker

University of Glasgow, Glasgow, U.K.A.J.Flavell, A.Halley, V.O’Shea,F.J.P.Soler

University of Li verpool, Li verpool, U.K.S.Biagi,T.Bowcock,R.Gamet,P.Hayman,M.McCubbin,C.Parkes,G.Patel,S.Walsh,V.Wright

Imperial College,London, U.K.G.J.Barber, D.Clark, P.Dauncey, A.Duane, M.Girone(1), J.Hassard,R.Hill, M.J.John6

, D.R.Price,

P.Savage,B.Simmons,L.Toudup,D.Websdale(1) alsoat CERN

University of Oxford, Oxford, U.KM.Adinolfi, G.Damerell, J.Bibby, M.J.Charles, N.Harnew, F.Harris, I.McArthur, J.Rademacker,N.J.Smale,S.Topp-Jorgensen,G.Wilkinson

CERN, Geneva, SwitzerlandJ.Andre, F.Anghinolfi, F.Bal, M.Benayoun(1),W.Bonivento(2), A.Braem, J.Buytaert,M.Campbell,A.Cass, M.Cattaneo, E.Chesi, J.Christiansen, R.Chytracek7

, J.Closier, P.Collins, G.Corti,

C.D’Ambrosio, H.Dijkstra, J.P.Dufey, M.Ferro-Luzzi, F.Fiedler, W.Flegel, F.Formenti, R.Forty,M.Frank, C.Frei, I.Garcia Alfonso, C.Gaspar, G.GraciaAbril, T.Gys, F.Hahn, S.Haider, J.Harvey,B.Hay8

, E.van Herwijnen, H.J.Hilke, G.von Holtey, D.Hutchroft, R.Jacobsson,P.Jarron,C.Joram,

B.Jost,A.Kashchuk(3),I.Korolko(4),R.Kristic,D.Lacarrere,M.Laub,M.Letheren,J.F.Libby, R.Lindner,C.Lippmann,M.Losasso,P.MatoVila, H.Muller, N.Neufeld,J.Ocariz9

, S.Ponce,F.Ranjard,W.Riegler,

F.Rohner, T.Ruf, S.Schmeling, B.Schmidt, T.Schneider, A.Schopper, W.Snoeys, V.Souvorov(3),

vi

W.Tejessy, F.Teubert,J.Toledo Alarcon, O.Ullaland,A.Valassi,P.VazquezRegueiro, F.Vinci do San-tos(5),P.Wertelaers,A.Wright10

, K.Wyllie

(1) on leave from Universite deParisVI etVII (LPNHE),Paris(2) on leave from INFN Cagliari,Cagliari(3) on leave from PNPIGatchina,St.Petersburg(4) on leave from ITEP, Moscow(5) on leave from UFRJ,Rio deJaneiro

1now at Grouped’AstroparticulesdeMontpellier(GAM), Montpellier, France

2now at DialogSemiconductor, Kirchheim-Nabern,Germany

3now at OsloUniversity, Oslo,Norway

4now at PennsylvaniaUniversity, Philadelphia,USA

5now at LBNL, Berkeley, USA

6now at CollegedeFrance,Paris,France

7now at IT Division,CERN,Geneva,Switzerland

8now at SWX SwissExchange,Geneve,Switzerland

9now at UniversitedeParisVI etVII (LPNHE),Paris,France

10

now at LancasterUniversity, Lancaster, UK

vii

Acknowledgements

The LHCb Muon groupwould like to expressits gratitudeto many people,at CERNandin thehomeinstitutes,whohaveparticipatedatvariousstagesto thedesign,testingandprototypeactivitiespresentedin this report. We areaware that the successfulconstructionandcommissioningof the LHCb muonsystemwill alsoin thefuturedependon their skills andcommitment.

We would like to thankA.Mazzari,V.Brunner, M.Grygiel andin particularN.Grub for their helpinthepreparationof thisproposalandtherelatedLHCb Notes.

viii

Contents

1 Intr oduction 11.1 Physicsrequirements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Generaldetectorstructure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2.1 Logical layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2.2 Physicallayout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 EvolutionsincetheTechnicalProposal. . . . . . . . . . . . . . . . . . . . . . . . . 41.3.1 Modificationsto thelogical layout . . . . . . . . . . . . . . . . . . . . . . . 41.3.2 Modificationsto thedetectortechnologies. . . . . . . . . . . . . . . . . . . 4

1.4 Structureof thisdocument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 DetectorRequirementsand Specifications 62.1 Backgroundenvironment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Muonsystemtechnologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2.1 MWPCdetectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82.2.2 RPCdetectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.2.3 Technologystudiesfor theinnerpartof stationM1 . . . . . . . . . . . . . . 11

2.3 Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.4 Detectorlayout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122.4.2 Chamberarrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 PhysicsPerformance 163.1 Simulationprocedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.1.1 Eventgeneration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.1.2 Trackingof particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.1.3 Backgroundandspillover . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.1.4 Digitisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183.1.5 Futuredevelopments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.2 Level-0muontrigger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193.2.1 Triggerdesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193.2.2 Triggerperformance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.3 Muon identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233.4 Reconstructionof muonicfinal states. . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.4.1 B0d J ψ µ µ K0

S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253.4.2 B0

s µ µ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263.5 Muon tagging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

4 Prototype Studies 284.1 MWPCstudies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

4.1.1 Front-endelectronics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284.1.2 Resultsof MWPCprototypetests . . . . . . . . . . . . . . . . . . . . . . . 304.1.3 Ageing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.2 RPCstudies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.1 Front-endelectronics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.2 Resultsof RPCprototypetests . . . . . . . . . . . . . . . . . . . . . . . . . 364.2.3 Ageing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

ix

5 TechnicalDesign 405.1 TheMWPCdetector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

5.1.1 Detectoroverview andrequirements. . . . . . . . . . . . . . . . . . . . . . 405.1.2 Chambercomponentsanddesign . . . . . . . . . . . . . . . . . . . . . . . 405.1.3 Chamberconstructionandtooling . . . . . . . . . . . . . . . . . . . . . . . 435.1.4 HV andFE interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465.1.5 Qualitycontrolandtesting . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

5.2 TheRPCdetector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485.2.1 Detectoroverview andrequirements. . . . . . . . . . . . . . . . . . . . . . 485.2.2 Chamberdesign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495.2.3 Chamberconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.2.4 Qualitycontrolandtesting . . . . . . . . . . . . . . . . . . . . . . . . . . . 515.2.5 Front-endelectronics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

5.3 Readoutelectronics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525.3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525.3.2 Front-endboards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535.3.3 Intermediateboards. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.3.4 ECSinterface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555.3.5 Serviceboards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 565.3.6 Off-DetectorElectronicsboards . . . . . . . . . . . . . . . . . . . . . . . . 575.3.7 Systemsynchronisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575.3.8 Radiationlevels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.3.9 Coolingof theFE-electronics . . . . . . . . . . . . . . . . . . . . . . . . . 59

5.4 Servicesystems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.4.1 Gassystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.4.2 High-Voltagesystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.5 Chambersupportstructuresandmuonfilter . . . . . . . . . . . . . . . . . . . . . . 625.5.1 Chambersupportstructures . . . . . . . . . . . . . . . . . . . . . . . . . . 645.5.2 Muonfilter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5.3 Beampipeshielding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

5.6 Safetyaspects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

6 Project Organisation 686.1 ScheduleandMilestones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

6.1.1 Chambers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686.1.2 Electronics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 686.1.3 Infrastructure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 706.1.4 Installationandcommissioning . . . . . . . . . . . . . . . . . . . . . . . . 70

6.2 Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 706.3 Divisionof responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

x

xi

!" #$%&

'()*+

,--

./0 12

345 67

89:;

<=>?

@ ABC A@

DFEGHIKJL JML GN

OKPQR P S

TKUVWR UQ

XZY[\] [^_`Za[Y

bKcdefg

hijlkmn

oKpqrsltp

uv

wx

yz|

Figure 1 TheLHCb spectrometerseenfrom above (cut in thebendingplane),showing the locationof thefiveMuonStations.

xii

1 Intr oduction

Muon triggering and offline muon identificationare fundamentalrequirementsof the LHCb ex-periment. Muons are presentin the final statesof many CP-sensitive B decays,in particularthetwo “gold-plated” decays, B0

d J ψ µ µ K0S

and B0s J ψ µ µ φ. Moreover, muonsfrom

semi-leptonicb decaysprovide a tagof the initialstateflavour of accompanying neutralB mesons.In addition, the study of rare B decayssuch asthe flavour-changingneutralcurrentdecay, B0

s µ µ , may reveal new physicsbeyond the Stan-dardModel.

TheLHCb muondetectorusesthepenetrativepower of muonsto provide a robust muon trig-ger. Theheavy-flavour contentof triggeredeventsis enhancedby requiring the candidatemuonstohave high transversemomentum,pT. The sameuniquepropertiesareutilisedoffline, to accuratelyidentify muonsreconstructedin the trackingsys-tem and to provide a powerful B-mesonflavourtag.

In this introduction,the physicsrequirementsare discussedand an overview of the muon de-tector systemis presented.A brief summaryoftheevolution sincetheTechnicalProposalis thengiven,followedby anoutlineof therestof thedoc-ument.

1.1 Physicsrequirements

The main requirementfor the muon detectoristo provide a high-pT muon trigger at the earli-est trigger level (Level-0). The effective LHCbLevel-0 input rate is about 10MHz, on aver-age,at ~ 2 1032 cm 2 s 1 , assuminga non-diffractive inelastic p-p cross-sectionof 55mb.This input ratemustbereducedto 1MHz within alatency of 4 0 µs, while retaininggoodefficiencyfor eventscontaininginterestingB decays. Themuon trigger provides between10% and30% ofthis trigger rate. In addition, the muon triggermustunambiguouslyidentify the bunchcrossing,requiringa time resolutionbetterthan25 ns.

Themuondetectorconsistsof fivemuontrack-ing stationsplacedalongthebeamaxisandinter-spersedwith a shield to attenuatehadrons,elec-

tronsandphotons.The muontrigger is basedona stand-alonemuon track reconstructionand pT

measurementwith a resolutionof 20%.To trigger,amuonmusthit all 5 muonstations,giving alowermomentumthresholdfor efficientmuontriggeringof about5GeV c. Hits in thefirst two stationsareusedto calculatethe pT of thecandidatemuon.

Thepolarangleandmomentumof particlesarecorrelated,suchthat high-momentumtrackstendto be closerto thebeamaxis. Multiple scatteringin theabsorberincreaseswith thedistancefrom thebeamaxis,limiting thespatialresolutionof thede-tector. Thegranularityof thedetectorvariessuchthatits contribution to the pT resolutionis approx-imately equalto the multiple-scatteringcontribu-tion. Thevariouscontributionsto thepT resolutionareshown in Figure2.

05

10

15

20

25

30

9

10 20 30

40 50

60

momentum (GeV/c)

p T r

esol

utio

n (%

)

TotalGranularityMultiple scattering between M1 and M2Magnet parameterization andmultiple scattering before M1

Figure 2 Contributionsto the transverse-momentumresolutionas a function of the muon momentum,av-eragedover the full acceptance.The pT resolutionisdefinedas prec

T ptrueT ptrue

T , andis shown for muonsfrom semi-leptonicb decayhaving a reconstructedpT

closeto thetriggerthreshold,between1 and2 GeV/c.

The muon systemmust also provide offlinemuon identification. Muons reconstructedin thehigh-precisiontracking detectorswith momentadown to 3 GeV/cmustbecorrectlyidentifiedwithan efficiency of above 90%, while keeping thepionmisidentificationprobabilitybelow 1.5%.Ef-ficientmuonidentificationwith low contaminationis requiredboth for taggingand for the cleanre-constructionof muonicfinal stateB decays.

1

Table 1 The logical pad size in the four regions of eachstationprojectedto M1 (scaledas zM1 zMi ). Thisindicatestheexactprojectivity in y betweenstationsandthedoublingof sizein bothdirectionsbetweenregions.Theregion innerdimensionsof M1 arealsoshown.

Pad Dimensionsat M1 ( cm2) RegionInnerM1 M2 M3 M4 M5 Dimensionsat M1 ( cm2)

R1 1 2.5 0.5 2.5 0.5 2.5 2 2.5 2 2.5 24 20

R2 2 5 1 5 1 5 4 5 4 5 48 40

R3 4 10 2 10 2 10 8 10 8 10 96 80

R4 8 20 4 20 4 20 16 20 16 20 192 160

1.2 Generaldetectorstructure

The positionsof the muonstationswithin LHCbcan be seen in Figure 1, which shows a topview of the experiment. The first station(M1) isplacedin front of the calorimeterpreshower, at12.1m from the interactionpoint, and is impor-tantfor thetransverse-momentummeasurementofthe muon track used in the Level-0 muon trig-ger. The remainingfour stationsare interleavedwith themuonshieldat meanpositionsof 15.2m(M2), 16.4m (M3), 17.6m (M4) and18.8m (M5).The shield is comprisedof the electromagneticand hadronic calorimetersand three iron filtersand hasa total absorber-thicknessof 20 nuclearinteraction-lengths. The chamberswithin the fil-ter areallocatedabout40cm of spaceandaresep-aratedby threeshieldsof 80cm thickness. Theinner andouterangularacceptancesof the muonsystemare20(16)mradand306(258)mradin thebending(non-bending)plane,similar to thatof thetrackingsystem. This providesa geometricalac-ceptanceof about20% for muonsfrom b decaysrelative to the full solid angle. The total detectorareais about435m2.

1.2.1 Logical layout

Thelogical layoutdescribesthex andy logicalpadgranularityin eachregionof eachmuonstation,asseenby the muon trigger and offline reconstruc-tion. The physicalimplementationof the logicallayout is presentedin the next subsection.Giventhedifferentgranularityrequirementsandthelargevariation in particleflux in passingfrom the cen-

tral part, close to the beamaxis, to the detectorborder, eachstationis subdividedinto four regionswith different logical-paddimensions(Figure 3).Region andpadsizesscaleby a factor two fromoneregion to thenext. Thelogical layoutsin the5muonstationsareprojective in y to theinteractionpoint.

Thexdimensionsof thelogicalpadsin stationsM1 –M3 aredeterminedprimarilyby theprecisionrequiredto obtaingoodmuonpT resolutionfor theLevel-0 trigger. The pady dimensionsin all fivestationsare determinedby the requiredrejectionof backgroundtriggerswhich do not point to theinteractionregion. The resultinglogical pady xaspectratiosare2.5in stationM1 and5 for stationsM2 andM3. StationsM4 andM5, which areusedto confirmthepresenceof penetratingmuons,haveaspectratiosof 1.25. The logical paddimensionsaresummarizedin Table1. The total numberoflogicalpadsin themuonsystemis 55,296.

1.2.2 Physical layout

The physical layout refers to the physical im-plementationof the logical pad layout describedabove. Startingfrom thelogicalpads,two stepsintheimplementationcanbedistinguished:

1. Logical pads– logical channels:Logical padsare obtainedfrom the cross-ing of horizontal and vertical strips wher-ever possible. Thesestrips, referedto aslogical channels,areformedfrom thefront-endchannels.This allows a reductionin thenumberof channelsto behandledby theoff-

2

Region 4

Region 3

Region 2

4804

2002

1001

500

250

250

3

00 3

00 6

00 1200 2

402

4003

Beam Pipe Sheilding

5

0mm x 250mm

25mm x125mm

12.5mm x 63mm

6

.3mm x 3

1mm

Logical channel

Logical channel

Logical pad

Reg 1

Sector

Figure 3 Front view of onequadrantof muonstation2, showing the dimensionsof the regions. Insideeachregion is shown asector, definedby thesizeof thehorizontalandverticalstrips.Theintersectionof thehorizontalandvertical strips,correspondingto the logical channels,are logical pads. The region andchanneldimensionsscaleby a factortwo from oneregion to thenext.

chamberelectronicsandthe trigger proces-sor from 55,296to 25,920.Thecrossingofthestripsis performedby thetriggerproces-sor. Stripsareemployedin stationsM2–M5.Thedimensionsof thehorizontalandverti-cal stripsarelimited by the logical channeloccupancy.

Becauseof theveryhighparticleratesin sta-tion M1, strip readoutis not possiblethere.Similarly, in region R1 of stationsM4 andM5 the intermediatestepof stripsis not re-alized,as the reductionin logical channelswouldbeinsignificant.

2. Logical channels– physicalchannels:Thechamberelementreadoutby onefront-endchannelis referedto asphysicalchan-nel in the following. Themaximumsizeofthe physicalchannelsis constrainedby therequirementsfor efficient operationof thechambers,discussedin Section2. Thereforein generalphysicalchannelsaresmallerthanlogical channels.Physicalpadreadoutpro-viding aspacepoint is preferred,wherepos-sible,asthis doesnot requiretheadditionallogical AND of the two coordinatesin anxandy strip readoutof thechamberandsub-sequentloss of efficiency. This schemeis

3

followedin mostof theregions.

As anexception,regionsR1 andR2 of sta-tionsM2 andM3 arecharacterizedby a y xaspectratio of 5 andhigh granularityin thebendingplane(cf. Table1). Therefore,x yreadouthasbeenchosenin this regionsandthe logical channelsare formedfrom inde-pendentphysicalchannels.

Redundancy is ensuredby having two detectorlayerswith independentreadoutper station. Thesignalsof correspondingphysicalchannelsin thetwo layersare logically OR-edon the chambers.The total numberof physicalchannelsin the sys-temis about120,000.

Thereadoutschemeoutlinedabovehadanim-portantimpactonthedetectordesign.Easeof con-structionandminimizationof thenumberof differ-ent chambersizesandtypeswereotherconsider-ationsin the choiceof the detectorlayout. As aconsequence,the chambersizesare rathersmall,varying from 0 1 m2 in region R1 to 0 5 m2

in region R4. Eachregion of eachstationis im-plementedwith chambersof onesize,all usingthesamereadout,resultingin a total of 20 differentchambertypesin thewholemuonsystem.

1.3 Evolution since the Technical Pro-posal

With respectto the Technical Proposal[1] themuon detector has retained its basic structure,namelyfive measuringstations. However, manymodificationshavebeenmadeto thedesign.Someof thesechangesareminor, suchasan increaseintheabsorberdimensionsfrom 70cmto 80cm,andtheremoval of thefirst absorberof 30cmthicknessbehindthe hadroncalorimeter. Othershave beenintroducedasaconsequenceof acarefullayoutop-timization study [2] andextensive R&D work ondetectortechnologies.

1.3.1 Modifications to the logical layout

The most importantchangesto the logical layoutarethefollowing:

The dimensionsof the four regions havebeenmodifiedin orderto minimisethenum-

berof differentchambertypesrequired.Thegranularity in the bendingplane has beendrasticallyimproved for muonstationsM2(by 100%) and M3 (by 200%), while thegranularity in the non-bendingplane hasbeenreducedby 25%everywhere.

The original logical layout, basedon pads,hasbeenreplacedby a layoutbasedon ver-tical andhorizontalstripsin stationsM2 toM5, as outlined before. In addition, onlytwo detectorreadoutlayers are envisagedperstation,insteadof four layersconsideredin the TechnicalProposalfor somepartsofthe system. This allowed a substantialre-ductionof thenumberof physicalchannels,from 236,000to 120,000.Moreover, whilethegranularityof thesystemincreasedfrom45,000 to about 55,000 logical pads, thenumberof logicalchannelscouldbereducedfrom 45,000to 26,000. Despitethis largereductionof readoutchannels,the perfor-manceof theL0 muontrigger is very simi-lar to thatreportedin theTechnicalProposal(for detailsseeSection3).

1.3.2 Modifications to the detector technolo-gies

In theTechnicalProposal,MultigapResistivePlateChambers(MRPC) andasymmetricCathodePadChambers(CPC)weresuggestedas technologiesfor the muon system. During a period of exten-sive detectorR&D work other technologieshavebeenstudiedaswell: singleanddoublegapResis-tive PlateChambers(RPC)[3], Thin GapCham-bers (TGC) [4], and simple Multi Wire Propor-tional Chambers(MWPC) [5, 6]. The first testsconductedshowedthattheMRPCs[7] do notper-form betterthansinglegapRPCs,whicharemuchsimpler to construct. Thereforethe MRPCshavenot beenconsideredfurther as a technologyforthe LHCb muon system. For the final choicebetweenthe other technolgyoptions,an internalLHCb review panelsupplementedby externalex-pertswassetup to arrive at a decision.Basedonperformancestudies,morerefinedestimatesof thebackgroundrate [8, 9] and ageingconditionsinthe detectors,andfinally cost, risk andresources

4

considerations,the following conclusionhasbeenreached:

Most regions of the detector(52% of thetotal area) should be instrumentedwithMWPC chambersoperated in low gainmode. From the point of view of ageingthey aresuperiorto TGCsoperatedin lim-itedspacechargemode.

RPCsshouldbe employed in the outer re-gions (R3 and R4) of the last two stations(M4 andM5), wheretheparticlerateis be-low 1kHz cm2(48% of thetotalarea).

Finally, the technologyfor the inner part (R1andR2)of thefirst stationM1, whereratesof upto460kHz cm2areconsidered,muststill beselected.This amountsto 2.9m2, i.e. lessthan1% of thetotal area.

1.4 Structure of this document

This TechnicalDesignReportis intendedto be aconcisebut self-containeddescriptionof theMuonsystem.Furtherdetailscanbe found in themanyTechnicalNotes,referencedthroughout.

Detectorspecificationsaregiven in Section2.This is followed by a descriptionof the physicsperformanceof thesystem,determinedusingsim-ulated events, in Section3. Section4 containsan overview of resultsobtainedin the laboratoryandat the testbeamusingprototypes,which giveconfidencethat the requiredperformancewill beachieved. The technicaldesignof the detectorsis presentedin Section5, andfinally the issuesofproject organisation,including the scheduleandcost,arediscussedin Section6.

5

2 Detector Requirements andSpecifications

The basicfunction of the LHCb Muon systemisto identify and trigger on muonsproducedin thedecayof b hadrons.The trigger logic is designedin sucha way that informationfrom all five muonstationsis required. In order to achieve a muontriggerefficiency of at least95%,thesingle-stationefficiency hasto be higherthan99%. Redundantsingle-stationefficiency of 99%canbeensuredbyhaving two independentdetectorlayersperstationandtakingthelogicalOR.

Thedetectorefficiency is mainlylimited by theintenseflux of chargedandneutralparticlesin theangularcoverageof theLHCb experiment.Theseflux levels exceedthoseexperiencedby the AT-LAS [10] andCMS [11] muonspectrometersandposeadifferentchallenge.

Becauseof the importanceof the backgroundconditionsin determiningthedesignof the muonsystem,thissectionopenswith abrief overview ofbackgroundsrelevant to theb µ X detectionintheLHCb experiment.Theimpactof theoperatingconditionson the muon-chambertechnologiesisthendescribedandthemaindetectorandelectron-icsparametersarelisted.Prototypetestshavebeenundertakenanddemonstratethatthelistedspecifi-cationscanbeachieved. Thesetestsaredescribedin Section4 of thisdocument.

2.1 Background envir onment

High particle fluxes in the muon system im-posestringent requirementson the instrumenta-tion. Theserequirementsincludetheratecapabil-ity of the chambers,the ageingcharacteristicsofthe detectorandredundancy of the trigger instru-mentation.Thehigh hit ratesin thechamberalsoeffect the muon transversemomentumresolutiondue to incorrecthit association. Four classesofbackgroundscanbedistinguished:

1. Decay muons: The large number of π/Kmesonsproducedin the p-p collisionscon-tribute mainly to the backgroundin themuonsystemthroughdecaysin flight. Suchdecaymuonsform themainbackgroundfortheL0 muontrigger.

2. Showerparticles: Photonsfrom π0 decayscan interact in the area around the beampipe andgenerateelectromagneticshowerspenetratinginto themuonsystem.Hadronsemerging from theprimarycollision canin-teractlatein thecalorimetersandcontributeto the background in the muon systemthrough shower muons or hadron punch-through.

3. Low-energy background: Another impor-tant backgroundis associatedwith low-energy neutronsproducedin hadroniccas-cadesin the calorimeters,the muon shieldor in acceleratorcomponents.They createlow-energy radiative electronsvia nuclearn-γ processesand subsequentCompton-scatteringor via thephoto-electriceffect inthedetectormaterialof themuonchambers.Thephotonshave a probabilityof a few permil to generatedetectableelectronsvia theseeffects,whicharein generalonly affectingasingledetectorlayer. Moreover, thehits dueto the low energy backgroundoccurup to afew 100ms after the primary collision (seeFigures10 and11).

4. Beamhalomuons:Thecharged-particlefluxassociatedwith thebeamhaloin theacceler-ator tunnelcontainsmuonsof a ratherwideenergy spectrumandthelargestflux atsmallradii (seeFigure17). In particularthosehalomuonstraversingthedetectorin thesamedi-rectionasparticlesfromtheinteractionpointcancauseaL0 muontrigger.

Backgroundcausedby real muonstraversingthe detectoris well simulatedwith the availableMonte Carlo packages.The uncertaintyattachedto the total p-p cross-sectionandto themultiplic-ity producedin theprimarycollisionsis estimatedat the level of 30%[12]. Thereis a limited know-ledgeof the showering processesin the absorbermaterialanda significantuncertaintyin theback-grounddueto low energy neutrons.An estimatefor theratein thevariousregionsof themuonsys-temhasbeenobtainedfrom a detailedstudy, com-paring two simulation packages,GCALOR [13]andMARS [14]. MARS predicteda countingrate

6

Table 2 Particle rates in the muon system. The first row gives the maximal particle rate in each regionandstationper interactionas obtainedfrom GCALOR; the secondgives the calculatedrateat a luminosity of

5 1032 cm 2 s 1 assuminga total p p cross-sectionof σ=102.4mb; andthe last row the rateincludingthesafetyfactors.

Station1 Station2 Station3 Station4 Station5

8.3 10 3 cm2 2.7 10 4 cm2 7.2 10 5 cm2 4.7 10 5 cm2 3.2 10 5 cm2

Region1 230kHz cm2 7.5kHz cm2 2 kHz cm2 2.3kHz cm2 880Hz cm2

460kHz cm2 37.5kHz cm2 10kHz cm2 6.5kHz cm2 4.4kHz cm2

3.3 10 3 cm2 1.9 10 4 cm2 2.3 10 5 cm2 1.6 10 5 cm2 1.3 10 5 cm2

Region2 93 kHz cm2 5.3kHz cm2 650Hz cm2 430Hz cm2 350Hz cm2

186kHz cm2 26.5kHz cm2 3.3kHz cm2 2.2kHz cm2 1.8kHz cm2

1.4 10 3 cm2 4.7 10 5 cm2 7.3 10 6 cm2 5.4 10 6 cm2 4.7 10 6 cm2

Region3 40 kHz cm2 1.3kHz cm2 200Hz cm2 150Hz cm2 130Hz cm2

80 kHz cm2 6.5kHz cm2 1.0kHz cm2 750Hz cm2 650Hz cm2

4.5 10 4 cm2 8.3 10 6 cm2 3.0 10 6 cm2 1.8 10 6 cm2 1.7 10 6 cm2

Region4 12.5kHz cm2 230Hz cm2 83Hz cm2 50Hz cm2 45Hz cm2

25 kHz cm2 1.2kHz cm2 415Hz cm2 250Hz cm2 225Hz cm2

twicethatof GCALOR[15]. A conservativesafetyfactorof 5 hasthereforebeenappliedto theback-groundratesof the GCALOR simulationfor sta-tions M2–M5 beforeconsideringthe detectorde-sign. For stationM1, which is positionedin frontof the calorimetersandthereforelessaffectedbytheuncertainties,asafetyfactorof 2hasbeenused.The hit rate in stationM1 is stronglyaffectedbytheLHCb beampipe.

A combinationof the expectedratesfrom thevarious processesyields the angular-dependenceof the expectedcountingrate in eachof the fivestationsof themuonsystem.Therateshave beencalculatedusinga simulationbasedon GCALORwith low trackingthresholds[8, 9]. Detailsof thesimulationstudiesand the obtaineddistributionsarediscussedin section3.1. The resultingmaxi-mal ratesperregion aresummarisedin Table2 fora luminosityof ~ 5 1032 cm 2 s 1 , atwhichtheLHCb experimentshouldbeableto operateforshortperiods. The raterisesfrom a few hundredHz cm2 in the outer regions of stationsM4 andM5 to a few hundredkHz cm2 in the innermostpart of stationM1. The rateshave beenobtainedwith anAl-Be beampipe,recentlyadoptedfor theLHCb experiment.

2.2 Muon systemtechnologies

Thecombinationof physicsgoalsandbackgroundconditionshave determinedthe choiceof detec-tor technologiesfor the various stationsand re-gions.Thefollowing threeparametersparticularlyaffectedthetechnologychoiceandthelocationoftheboundarybetweenthem:

1. Ratecapabilityandageing: Thehigh parti-cle fluxesin themuonsystemaffect the in-stantaneousefficiency of RPCs,becauseofa voltagedropon theelectrodes.Moreover,the materialsfor the chambersshouldhavegood ageingproperties,allowing 10 yearsof operation.In thecaseof wire chambers,the accumulatedcharge shouldnot exceed1C cm, which in generalis consideredasanupperlimit for safeoperation.

2. Time resolution: The muon systemmustprovide unambiguousbunchcrossingiden-tification with high efficiency. The require-mentis atleast95%efficiency within a20nswindow for eachof thetwo layersin thesta-tion.

3. Spatial resolution: The spatial resolutionmust allow the determinationof the pT of

7

triggeringmuonswith a resolutionof 20%.This requiresa granularity in stationsM1and M2, usedfor the measurementof pT,asgiven in Table1. Sincetwo detectorlay-ers are employed in eachstation, inclinedtrackstraversing the detectorcan hit morethanonelogicalpad.Thiseffect is describedas geometricalclustersize. Dependingontheaveragecrossingangle,thepadsizeandthe layer separation,the geometricalclus-ter sizevariesbetween1.1 in theouterpartand1.3 the inner part of the muonsystem.To minimiseany additionaldeteriorationofthe intrinsic detectorresolution,crosstalkbetweenreadoutchannelsshouldbelimitedsuchthat it doesn’t addsignificantly to thegeometricalclustersize.

Basedontheaboveconsiderations,RPCshavebeenadoptedfor regions R3 and R4 of stationsM4 and M5, where the expectedrate is below1kHz cm2, andtherequirementson crosstalk arerathermodestdueto thelargesizeof logical pads.This typeof detectorhasanexcellenttime resolu-tion of 2nsandis well adaptedfor fasttrigger-ing.

MWPChavebeenadoptedfor all otherregionsexceptthe innerpartof M1. A time resolutionofabout3ns is achieved in a doublegap by usinga fast gasand a wire spacingof 1.5mm. Spacecharge effectsdueto theaccumulationof ionsarenotexpectedfor ratesof up to 10MHz cm2.

Thechoiceof thesetwo technologies,andthelocationof theboundarybetweenthem,alsotakesinto accountthe robustnessof the demonstratedperformance,discussedin Section4, andconsid-erationsof resourcesandschedule.Thespecifica-tionsof MWPCsandRPCsaresummarisedin thefollowing sections.

For the inner part of stationM1 ( 3 m2) atechnologystill hasto be selected.Varioustech-nologiessuch as asymmetricMWPCs or triple-GEM detectorsareunderinvestigation.Thestatusof theR&D for themis briefly described.

Cathode pads

2.5mm

2.5mm

1.5mmAnode wires

Detector ground

Guard trace

Figure4 Schematicdiagramof onesensitivegapin aMWPC.

Table 3 Main MWPCparameters

Parameter Designvalue

GasGap 5 mmWire spacing 1.5mmWire Diameter 30 µmOperatingvoltage 3.0-3.2kVNo. of gaps 4Gasmixture Ar / CO2 / CF4

(40:50:10)Primaryionisation 100e /cmGasGain 105

Threshold 3fCCharge/ 5mmtrack 0.8pC

2.2.1 MWPC detectors

2.2.1.1 General description

A schematicdiagramof an MWPC is given inFigure 4, and the principal chamberparametersare summarisedin Table 3. The chambershavea symmetriccell with an anode-cathodedistanceof 2.5mm andan anode-wirespacingof 1.5mm.The MWPC gas is a non-flammablemixture ofAr CO2 CF4 (40:50:10). The fact that this gascontainsno hydrogenresultsin a low sensitivityto neutronbackground.

A muon crossingthe 5mm MWPC gasgapwill leave anaverage50 electronsthatdrift to thewires in the electricfield. The electronsandionsmoving in the avalanchecloseto the wire inducea negative signalon thewire anda positive signalwith thesameshapeandabouthalf themagnitude

8

on eachof the cathodes.Eachchambercontainsfour sensitive gapswhich are connectedas twodoublegapsto two front-endchannels.This pro-videsredundancy andensuresthattheefficiency ofeachdoublegapis about99%.

Chambersare readoutdifferently, dependingon their position in the muon system. In regionR4of stationsM1 – M3 thechambershave anode-wire readout(throughdecouplingcapacitors).Inregion R3 of stationsM1 – M3 and regions R1andR2 of stationsM4 andM5 cathodepadsarereadout.In regionsR1 andR2 of stationsM2 andM3 a combinedreadoutof wire andcathodepadsis usedasa consequenceof therequiredgranular-ity. Anodewiresaregroupedinto verticalstripstomeasurex whereasthey coordinatesareprovidedby the coarsergranularityof the horizontalcath-odepads.Wiresaregroupedin padsof 4 to 42 tomatchtherequiredgranularity, varyingfrom 6mmin region R1 of stationM2 to 62mm in region R2of stationM5. Thesegroupsof wiresarereferredto aswire-padsin this document.The muonsys-tem requires864 four-gap chambers,with 2.5 106 wiresandabout80,000front-endchannels.

2.2.1.2 Specialconditions and requirements

Ageing: High rates of up to 70kHz cm2

raise concerns about chamber ageing dueto gas polymerisation on wires or cathodes.The chambersmust survive 10 years of op-eration (108 s) at the nominal luminosity of~ 2 1032 cm 2 s 1 with the operationalgasgain of 105. Theaccumulatedcharge undertheseconditionswill be about 0.5C cm on the wiresand 1.7C cm2 on the cathodesin the regions ofhighestparticleflux.

Crosstalk: Two sourcesof crosstalk canbedistinguished:(1) directly inducedcrosstalk, if aparticlecrossesthechambercloseto theedgeof apad,and(2) electricalcrosstalk, dueto capacitivecouplingbetweenpads.

The amountof directly inducedcrosstalk be-tweencathodepadsis givenby thewire pitch andcathode-wiredistance.To keepthecross-talkratefrom direct inductionbelow 20%thecathodepaddimensionsshouldbe larger than2.2cm [16], as-suminga thresholdof six primary electrons. In

caseof wire-padreadout,crosstalk dueto directinduction is negligible, as it occursonly if a par-ticle crossesa gap at less than 200µm from thecentrebetweento neighbouringwire pads.On theotherhand,theaveragegeometricalclustersizein-creasesasthewire-padsizedecreasesfor thesameaveragetrackinclination,giventhe4.6cm separa-tion of the outergapswithin a chamber(seeFig-ure 46). To keepthe averagegeometricalclustersizebelow 1.2,thewire-padsizeshouldbeat least6mm.

Electrical cross talk betweenpads dependson several factors, including the amplifier inputimpedance,the padcapacitancesandmutualpadto padcapacitances.Betweenwire padsit canonlybe limited by a low amplifier input impedance.Electricalcrosstalk betweencathodepadscaninadditionbelimited by guardtracesin betweenthepads.It wasshown [16] thatfor anamplifierinputimpedanceof 50Ω togetherwith the pad lay-outdescribedin Section5.1.2.2theelectricalcrosstalk doesnot addsignificantly to the geometricalclustersizedescribedearlier.

2.2.2 RPC detectors

RPCs[17] areparallelplatechamberswith a gasgap betweentwo resistive electrodes. They arecharacterisedby excellent timing properties.Thereadoutoccursvia capacitive couplingto externalstrip or pad electrodes,which are fully indepen-dentof the sensitive element.Furtheradvantagescomparedto othertechnologiesaretherobustnessandsimplicity of construction.They arealsowelladaptedto inexpensive industrialproduction.

2.2.2.1 General description

The operatingprinciple of the RPC is shown inFigure 5. Ionising particlescreateelectron-ionclustersin thegas,whereanintenseconstantelec-tric field is presentbetweentwo parallel plates.Multiplication in the gas,averagedover the gap,is typically 107. Thelargegasgainis necessaryin orderto havehighefficiency evenfor thoseclus-tersthatarecreatedcloseto thepositive electrode.Densegaseswith high Z arepreferredsincetheyhave a large ionisationprobability. Theplatesare

9

pick-up electrodes(s)

signal

+

-

HV d

x

q0

Q

Qqe

¡ ¢¢

Figure5 Principleof theRPC.Primaryelectronclus-ters, moving in the gas over a distancex, createanavalancheof total charge Q. A charge qe is inducedon thereadoutelectrode(s).

Table 4 Main RPCparameters

Parameter Designvalue

Gasgap 2 mmOperatingvoltage 9-10kVGasmixture C2H2F4 / C4H10 / SF6

(95:4:1)Gasgain 107

Avalanchecharge 30 pCtimeresolution £ 2 nsNo. of RPC/chamber 2Threshold 40fCBakelite resistivity 8 ¤ 2 109Ωcm

madeof bakelite, a phenolicresinwith high vol-umeresistivity (typically 109 1011 Ωcm).

Giventhecharacteristicexponentialgrowth ofthe avalanche,the output signal inducedon thereadoutelectrodeshasa very fast rise time (τ 1ns), yielding the excellent time resolution ofthesedetectors.A thresholdof 40fC is adequateto achieve full efficiency.

If the electricfield is too large, the avalanchemay be followed in time by a streamer. Workingin streamerregimestronglyreducestheratecapa-bility of thesedetectors.Therefore,in LHCb theRPC will be operatedin avalanchemode,wherethecharge is limited to tensof pC [18]. Themaincharacteristicsof the RPCdetectorsof LHCb aresummarisedin Table4.

Thegasmixtureensureshigh efficiency andawideoperatingplateau(about500V) in avalanchemode. The readoutelectrodesare strips with a

pitchof about6cm,andastrip-to-stripdistanceof2mm. To copewith therequestedspatialgranular-ity thestrip lengthin regionR3is half thelengthinregion R4 (respectively 15 and30cm). The clus-ter size(dueto particlescrossingtheboundarybe-tweentwo stripsandto electro-magneticcross-talkfrom a strip to another)is lessthan2. To achievethe requiredredundantefficiency of 99% per sta-tion, eachchamberwill consistof two RPCsin ORwith independentHV supplyandreadout.

2.2.2.2 Specificrequirements

In the RPCdetectors,oncethe avalanchehasde-veloped,thebakelite platesarelocally dischargedandthedetectorremainsinsensitive in a small re-gion around the hit spot. The time neededtorecharge theplatesandrestorethefield limits therateperformanceof theRPC.Sincethe time con-stant for recharge is directly proportionalto thebulk resistivity ρ of the electrodes(being of theorderof milliseconds),it is desirableto have low-resistivity bakelite for the electrodes. Bakeliteplatesarenow availablein awiderangeof resistiv-ities. ATLAS andCMS,whichoperateatrelativelylow rate, useρ ~ 1010 1011 Ωcm [10, 11]. InLHCb the specificationsrequirea maximumratecapabilityof 750 Hz cm2(seeTable2). This canbe achieved by usingbakelite with ρ in the range 8 ¤ 2 109 Ωcm which hasgiven very satisfac-tory resultsin our tests[19, 3].

Detailedsimulationstudiesof RPCshavebeenperformed,in particularon crosstalk, which arereportedin Ref. [20].

The relatively large rate in LHCb also raisesthe question of possible ageing effects, whichcould reducethe ratecapability. The relevant pa-rameteris thetotal integratedchargeper cm2. Thisamountsto about1.1 C cm2 in 10 yearsof oper-ation,a valuewhich is 3 timeslargercomparedtothe valuesachieved in previous testsby ATLASandCMS.Thereforemeasurementsat theGammaIrradiationFacility (GIF) facility atCERNarecar-ried outatpresentto testtheoperationof RPCde-tectorsfor severalLHCb years(seeSection4.2fordetails).

The intrinsic noiseof theRPCsshouldbelowenoughnot to spoil the performanceof the muon

10

trigger. Simulationsof theeffect of RPCnoiseontriggerperformancehavebeencarriedout[21] (seealsosection3.2.2),which show that noisecanbetoleratedprovided its rate is limited to less than1kHz cm2. However, ageingconsiderations,asdiscussedin Section5.2.2.1,leadto a morestrin-gent requirementandlimit the RPCnoiseto lessthan100Hz cm2.

2.2.3 Technologystudiesfor the inner part ofstation M1

MWPCs with an anode-cathodedistance of2.5mm have a limitation on the cathode-paddi-mensionof about2.2cm due to directly inducedcrosstalk. With therequiredpadgranularityin re-gion R1 of 1 2.5cm2, the clustersizewould beabout1.4. Moreover, theaccumulatedchargedur-ing 10 yearsof operationat nominal luminosityin theMWPCswould be2.9C cm, which is wellabove theupperlimit generallyconsideredsafeforoperation.

With the new Aluminium-Beryllium beampipe, recently adoptedas the baselinefor theLHCb experiment,thebeforementionedproblemswith symmetricMWPCsarediminishedfor regionR2. Due to the reducedrate in this region thephysical channelgranularity has been increasedto 2 5 cm2, which leadsto a clustersizeduetodirect induction of 1.22. The total accumulatedchargewouldbe1.2C cm.

Theclustersizeduetodirectinductioncanalsobereducedto 1.2for regionR1if theanodes-wiresareplacedasymmetricallyin thegasgapat 1mmfrom thecathode-pads.Theinducedchargeon thecathodpadswouldallow to operatethechamberata factor two lower gasgain, which would reducethe total accumulatedcharge to 1.4C cm. Suchchamberswereproposedin theTechnicalProposalfor this partof themuonsystemandremaina vi-ableoption[22].

Another technologyunder considerationforthis part of the systemare triple-GEM detectorswith padreadout.Concernsaboutthetime resolu-tion of suchdeviceshave beenaddressedin recentbeamtests. An efficiency of 96% (93%) withina 25ns (20ns) time window hasbeenmeasuredwith a chamberprototypeusinga CF4 basedgas

mixture anda fast pre-amplifier, which is an en-couragingresult [23]. The measuredclustersizeon a 18 16 mm2 (6 16 mm2) padwas1.1 (1.5).Studieson the ageingand high-rateperformanceof suchadetectorareongoing.

2.3 Electronics

Themuonsystemfront-end(FE)electronicshastopreparethe informationrequiredby the L0 muontriggerasquickly aspossibleandmustconformtotheoverall LHCb readoutspecifications[24]. Thereadoutelectronicschaincomprisesthe followingelements(seeSection5.3for details):

FE boards on the chamberswith ampli-fier, shaper, discriminator(ASD) chips,andchips to combinethe output signalsof theASD to form logical channels;

Intermediate(IM) boards on thesideof themuon system,to generatelogical channelsfor thoseregions where this has not beenpossibleon the chambers,becausethe log-ical channelsaremadeof physicalchannelsbelongingto differentchambers;

Off-Detector Electronics (ODE) boards,also located on the side of the detector,where the data is synchronisedand dis-patchedto theL0 trigger. It comprisesalsothe L0-pipelines,L1-buffers and the DAQinterface.

Several stringentrequirementsmust be satis-fiedby theFEelectronics,in particularby theASDchip. The requirementsfor the ASD chip in thehighestrateregionsaresummarizedin Table5. Inthe following someimportantaspectsarepointedout.

In the highestrate regions the maximal totaldoseis about1MRad, which requiresthe useofradiationhardchipson the FE board. The layoutof thechipsis thereforedoneusing0.25µm CMOStechnologyor similar processes,which areknownto beradiationhard.

High signal rateshave a large impact on thedetectorefficiency due to deadtimegeneratedbytheoutputpulsewidth of theASD chip. Figure6shows how thepulsewidth affectsthesinglelayer

11

Table 5 Front-endchip requirementsfor the muonsystem.Someparametersapplyonly to thehighestrateregions.

Parameter Specification

Detectorcapacitance 40-250pFMaximumsignalrate 1MHzMaximumtotaldose 1MRadInput resistance £ 50ΩAveragepulsewidth £ 50ns(ASD output)Peakingtime 10ns

PW=20 ns

P¥

W=40 ns

P¥

W=60 ns

Figure 6 Efficiency as function of signal rate forpulsesof differentwidth.

efficiency, assumingthatthereadoutelectronicsisonly sensitive to the trailing edgeof the signal.Sinceanoutputpulsewidth below 50nsis unreal-istic, a maximalsignalrateof 1MHz perphysicalchannelhasbeendefined,which keepsthe ineffi-ciency below 5%. The maximal rateper channelsetsan upperboundon the size of the FE chan-nels. In orderto minimisethedeadtime, unipolarpulseshapinghasbeenchosen.A unipolarlinearsignal processingchain also requiresa BaseLineRestorationcircuit (BLR) to avoid baselinefluc-tuations.Moreover, thechambersignalhasaniontail whichrequiresadedicatedion tail cancellationnetwork to filter it.

The input resistanceof the pre-amplifierhasto besmallerthan50Ω in orderto limit thecrosstalk dueto capacitive coupling.Performancetestsof pre-amplifiers,summarisedin Section4, showthat input capacitancesof up to 250pF providea satisfactoryperformance.This valuesetsanad-ditional limitation on thesizeof thereadoutchan-nels.

Finally, the power dissipationof the FE chipshould be low to keep thermal gradientson thechambersto aminimum.

The particle ratesto which the electronicsofthemuonsystemareexposedaffectalsothesingleeventupset(SEU)behaviour of thesystem.Wher-everlogicaloperationsareforeseenin theelectron-ics,specialprocedureslike triple votingwill there-fore be implementedto ensuretheSEUimmunityof thesystem.

2.4 Detector layout

Therequirementsandspecificationsdescribedpre-viously areanimportantingredientto thedetectorlayoutsummarisedhere.Detailsof themuonsys-temconfigurationcanbefoundelsewhere[25].

An importantconstraintto thedetectorlayoutcomesfrom theL0-muontriggerdesign,whichre-quiresprojectivity at eachstationfor the correctexecutionof thealgorithm.

From several view pointsthe situationof sta-tion M1 is special.While M1 is notusedfor muontrackidentification,it playsanimportantrolein thetransversemomentummeasurementof the muontrack in theL0-trigger andis thereforepositionedin front of the calorimeters.This positionmakesit not only subjectto thehighestratesin themuonsystem,but impliesalsothatspecialcarehasto betaken in the detectordesignto minimise its radi-ation lengthX0. The performancestudiesfor thecalorimeter[26] have beendonewith a simplifiedmuonchamberdescriptionincluding0.1X0 of ma-terial andshowedgoodperformance.Therequire-mentin termsof radiationlengthfor M1 hasthere-fore beenset to 0.1 X0 for the chambersensitivearea.

2.4.1 Overview

Theallocatedspacefor themuonstationsis 40cm,except for station M1, where the spaceis only37cm. The thicknessof the four iron absorbersis 80cm. The inner and outer acceptanceof themuonsystemis 20mrad 306mrad in the bend-ing plane(x) and16mrad 258mradin y. Theto-tal detectorareais about435m2, of which228m2

will be equippedwith MWPCs and 207m2 withRPCs. The inner part of stationM1 amountsto2.9 m2. All basic muon systemdimensionsaresummarisedin Table6.

12

Figure 7 Side view of the muon systemin the y¦ zplane

2.4.2 Chamber arrangement

Thechosenchamberheightsof 20–30cmmatchesthe requiredy-granularity in region R4. In thecaseof MWPCs,this allows theuseof theanode-wire readout,which provides a factor two largerpulseheight than the cathodesignaland thus al-lows larger input capacitances.Moreover, con-structingall chamberswith thesameheightin onestationavoids complicationsin theboundaryareabetweendifferent regions. The chamberlengthsarea consequenceof theaim to have thesmallestpossiblenumberof differentchambertypeswithineachstation.For region R4,chamberconstructionissueslimited thechamberlength.

Thesamechamberdimensionshave beencho-senfor MWPCsandRPCs. Although the bound-arybetweenthetwo technologieshasbeendefined,having the samechamberdimensionsmakes onetechnologythe backupfor the other in somere-gionsof themuonsystem.

At the centreof eachstationwill be a 44mm(42mmin M1) thick Aluminium chamber-support

structure.Thechambersarepositionedat four dif-ferent z positionsrelative to this support,z1§ 2 infront and z3§ 4 behind. All chambersin the samehorizontalrow areeitherin front of, or behindthesupport. The chamberswill be arrangedas indi-catedin Figures7, 8 and9, wherethreeviews ofthemuonsystemareshown.

The positioningof the chambersin the x yplanewithin a stationis donein sucha way astopreserveasmuchaspossiblethefull projectivity ofthelogical layout. This is mandatoryfor a correctexecutionof theL0-muontriggeralgorithmandtominimisethegeometricalclustersizeandgeomet-rical inefficienciesat the boundaryof the cham-bers. The logical layout is definedat the centralplaneof the stationandthesensitive areaof eachchamberis sizedas if it were at this plane. Thex- andy-positionsof the centresof eachchamberwithin astationareobtainedsimplyby positioningeachchambercentreso that it projectsfrom theinteractionpoint (IP) to its positionin the logicallayoutat thecentralplaneof thestation. In doingso, the chambersat positionsz1§ 2 will overlap inx with their neighbours.The overlap however isalwayslessthanhalf of onelogical channel.Sim-ilarly, theholesintroducedbetweenthechambersat positionsz3 § 4 aresmall, andarefurther limitedby thethicknessof thechambersof £ 75mm in z.Viewedfrom the interactionpoint the total lossinangularacceptanceis lessthan0.1%. The corre-spondingy-overlapsarenegligible dueto thesmally-dimensionsof thechambers.

13

R1

R2

R3

R4 Chamber

Chamber

Chamber

Ch.

Z2 Z1 Z2 Z1 Z2

Z4 Z3 Z4 Z3 Z3Z4

Figure 8 x¦ y view of a quarterof stationM2, onechamberin eachregion is highlighted.Therows of chambersmarkedin a darkershadearein positionsz3 ¨ 4 behindthesupportstructure,thosenot markedarein z1 ¨ 2 in front ofthesupportstructure.

Z3

Z4

Z2

Z1

Z©

3

Z4

Z©

2

Z1

Figure 9 Partial view of the muonsystemin the x¦ z planeat y

0. Therearetwo setsof chamberpositionsindicatedin differentcolours,beforeandafterthechambersupport,in eachstation.Eachsetindicatesthepositionof the chambersin a horizontalrow, the othersetof positionscorrespondto the chambersin the rows directlyabove andbelow this row. Theprojectivity of thechambersto the interactionpoint hasbeenindicated.Thefoursensitivegapsin eachchamberarealsoindicated.

14

Table 6 Summarytableof theMuonSystemparameters

StationM1 StationM2 StationM3 StationM4 StationM5 Sum

Stationz-pos.(m) 12.10 15.20 16.40 17.60 18.80Stationdimensions(m2) 7.7 6.4 9.6 8.0 10.4 8.6 11.1 9.3 11.9 9.9 435

SurfaceRegionR1 ( m2) 0.6 0.9 1.0 1.2 1.4 5.1Technology t.b.d. MWPC MWPC MWPC MWPCReadout t.b.d. Combined Combined Cathode CathodeNo. of chambers 12 12 12 12 12 60Chamb.sens.area( cm2) 24 20 30 25 32.4 27 34.8 29 37.1 30.9Physicalchannels 4608 2688 2688 2304 2304 14592Logical channels 2304 1536 1536 1152 1152 7296Logical unit size( cm2) – 3.75 25 4.05 27 – –Logical padsize( mm2) 10 25 6.3 31.3 6.7 33.7 29 36 31 39

SurfaceRegionR2 ( m2) 2.3 3.6 4.2 4.8 5.5 20.5Technology t.b.d. MWPC MWPC MWPC MWPCReadout t.b.d. Combined Combined Cathode CathodeNo. chambers 24 24 24 24 24 120Cham.sens.area( cm2) 48 20 60 25 64.8 27 69.5 29 74.3 30.9Physicalchannels 9216 5376 3840 2304 2304 23040No.logicalchannels 2304 1536 1536 672 672 6720Logical unit size( cm2) – 15 25 16.02 27 17.4 29 18.6 30.9Logical padsize( mm2) 20 25 12.5 62.5 13.5 67.5 58 72 62 77

SurfaceRegionR3 ( m2) 9.2 14.4 16.8 19.3 22.1 81.8Technology MWPC MWPC MWPC RPC RPCReadout Cathode Cathode Cathode Electrode ElectrodeNumberof Chambers 48 48 48 48 48 240Sensitive area( cm2) 96 20 120 25 129.6 27 139 29 148.5 30.9Physicalchannels 9216 9216 9216 4608 4608 36864No. logical channels 2304 1344 1344 480 480 5952Logical unit size( cm2) – 60 50 65.54 27 70.4 58 74 62Logical padsize( mm2) 40 100 25 125 27 135 116 145 124 155

SurfaceRegionR4 ( m2) 36.9 57.7 67.2 77.4 88.3 327.3Technology MWPC MWPC MWPC RPC RPCReadout Anode Anode Anode Electrode ElectrodeNo. chambers 192 192 192 192 192 960sensitive area( cm2) 96 20 120 25 129.6 27 139 29 148.5 30.9Physicalchannels 9216 9216 9216 9216 9216 46080No. logical channels 2304 1344 1344 480 480 5952Logical unit size( cm2) – 120 100 130 108 139 116 149 124Logical padsize( mm2) 80 200 50 250 54 270 131 290 248 309

15

3 PhysicsPerformance

The performanceof the muon detector, for trig-gering and for physicsanalysis,hasbeenevalu-atedusingsimulatedevents.This sectioncontainsanoutlineof thesimulationprocedure,anddetailsof the studiesundertaken. The efficiency of themuontriggerundernominalconditionsis reported,andinformationis given on the trigger sensitivityto changesin the detectorcharacteristics.Offlinemuon identification is discussedboth in relationto flavour taggingand in relation to channelsofphysicsinterestwith µ µ in thefinal state.

3.1 Simulation procedure

Theeventsimulationconsistsof severalbasicop-erations,performedin series: generation,track-ing of particlesthroughthe experimentalappara-tus,additionof background,anddigitisation.

3.1.1 Event generation

Proton-protoninteractionsat the LHC centre-of-massenergy of 14TeV aregeneratedwith Pythia6.1 [27], usinga multiple-interactionmodelchar-acterisedby varying impactparameteranda run-ning pT cut-off. The model parametershavebeentuned [12] to reproduceresultsfor proton-antiprotoncollisionsat centre-of-massenergiesinthe range50GeV to 1.8TeV, the highestenergiesfor which dataare available. Parton distributionfunctions are taken from the set CTEQ4L [28].Eventgenerationis performedtakinginto accountthe angulardispersionof the beamparticlesandthe spatialdistribution of the primary-interactionpoint.

3.1.2 Tracking of particles

TheLHCb apparatusis describedin thecontext ofGeant3.21[29], following thelayoutsgivenin therelevant TechnicalDesign Reportsfor calorime-ters [26], RICH [30] andvertex locator[31], andusingthe layout of theTechnicalProposal[1] forthetrackingdetectors.Thedescriptionsof thefivemuon stationsand the iron shieldingclosely fol-low the designspresentedin this report. Eachof

the276chambersforming amuonstationis simu-latedin detail [32], taking into accountthediffer-ent materiallayers,the appropriategasmixtures,and the aluminium frames. There are four gasgapsfor an MWPC and two for an RPC.Cham-bersarepositionedasdescribedin Section2.4.2,sothatthey areprojective to theinteractionregionanddistributed over four planesper station. Thechambers’aluminiumsupportsarenot simulated,but areexpectedto havenegligible effectonmuon-systemperformance.

Geant transport routines track particlesthrough the experimentalsetup, taking into ac-count all secondaryprocesses,and allowing fortheeffect of themagneticfield in theregion of thetrackingdetectors.ParticledecaysarehandledbytheCLEOcollaboration’s QQpackage[33], whichreliesasmuchaspossibleon measuredbranchingfractions, and includes detailed information ofdecaykinematics. Except in the muon shields,electronsand photons are tracked down to anenergy of 1MeV, whereashadronsand muonsare tracked down to 10MeV. Inside the muonshields,and to keepthe event-simulationtime atan acceptablelevel, the tracking thresholdsare10MeV for muonsand500MeV for otherparticletypes. A hit is recordedin a muon chamberwhenever a gas gap is traversedby an ionisingparticle. The coordinatesfor the pointsat whichtheparticleentersandexits thegasgaparestored,togetherwith theparticle’s timeof flight.

3.1.3 Background and spillover

In stationsM2 to M5, the relatively high track-ing thresholdsfor the muon shielding suppresshits that would result from the generationoflower-energy shower particles,or from the emis-sionandconversionof photonsfollowing thermal-neutroncapture. The loss is correctedby addingbackgroundaccordingto a set of parameterisa-tions [8, 9]. Theseare obtainedby comparingthe minimum-biashit distributions from the stan-dardprogramandthe correspondingdistributionsfrom a more completesimulation, basedon theGCALOR package[13].

The trackingcut-offs in GCALOR aresetev-erywhereto 0.1MeV for electronsand photons,

16

0ª

10

20«30

40

50¬60

0ª

1 2 3 4 5¬

60ª0.5ª 1

1.5

2

2.5

3

3.5

4

0ª

1 2 3 4 5¬

60ª

0.1ª0.2ª0.3ª0.4ª0.5ª0.6ª

0ª

1 2 3 4 5¬

60ª

0.1ª0.2ª0.3ª

0ª

1 2 3 4 5¬

60ª

0.1ª0.2ª0.3ª

0ª

1 2 3 4 5¬

6

Tra

cks

per

met

re p

er e

vent

Radial distance (metres)

M1 M2 M3 M4 M5

Low-threshold simulation High-threshold simulationand parameterised backgroundHigh-threshold simulation

Figure 10 Radial distribution of tracks at muon stations,as given by low-thresholdsimulation (solid line),high-thresholdsimulation (shadedhistogram)and high-thresholdsimulation with parameterisedbackground(shadedcircles).

0ª5¬

10

15

20«25

-1 2 5¬

8

10-1 102 105 108 0ª0.2ª0.4ª0.6ª0.8ª 1

1.2

1.4

1.6

-1 2 5¬

8

10-1 102 105 108 0ª

0.05ª

0.1ª

0.15ª

0.2ª

-1 2 5¬

8

10-1 102 105 108 0ª0.02

ª0.04ª0.06ª0.08ª 0.1ª0.12ª0.14ª

-1 2 5¬

8

10-1 102 105 108 0ª

0.02ª0.04ª0.06ª0.08ª

0.1ª

0.12ª

-1 2 5¬

8

10-1 102 105 108

Par

ticle

s pe

r bi

n pe

r ev

ent

Time of flight (ns)

M1 M2 M3 M4 M5

Low-threshold simulation High-threshold simulationand parameterised backgroundHigh-threshold simulation

Figure11 Particleflight timesatmuonstations,asgivenby low-thresholdsimulation(solid line), high-thresholdsimulation(shadedhistogram)and high-thresholdsimulationwith parameterisedbackground(shadedcircles).ParticlesproducedbeforeM1 andparticlesfrom showersin thecalorimeterandmuonshieldingarerecordedwithamaximumflight timeof about100ns.Latertimes,peakingataround105ns,aredueto particlesemittedfollowingthermal-neutroncapture.

10 6 eV for neutronsand 1MeV for other parti-cles. The GCALOR simulationalso includesin-frastructureelementsnear the muon systemandrelevantfor low-energy processes.

Parameterisationsareextractedbothfor theto-tal multiplicities from GCALOR and for the dif-ferenceswith respectto the standardsimulation.Theparameterisationsprovidedescriptionsof spa-tial andtemporaldistributions,andcorrelationsbe-tweenmultiplicities for high and low thresholds.With parameterisedbackgroundadded,the stan-dardsimulationfor stationsM2 to M5 reproducesthe GCALOR results to better than 10% (Fig-ures10 and11). No parameterisedbackgroundisaddedfor stationM1. Distributions for this sta-tion from thestandardsimulationdo not perfectly

match those from GCALOR, but disagreementsare small comparedwith the intrinsic GCALORuncertainties.

Possibleinaccuraciesin thesimulationat highandlow energiesareaccommodatedby usingtheparameterisationsof totalmultiplicities to increasethe numbersof hits in all five muonstations,ac-cordingto chosenscalefactors.Two setsof scalefactorsareconsideredin theperformancestudies.A scalefactorof 1 for eachstationis takentocorre-spondto asituationof nominalbackground.Scalefactorsof 2 for M1 and 5 for M2 to M5 definemaximalbackground.In this lattercase,thescalefactorfor M1 is lessthanfor theotherstationsbe-causeof thesmallercontribution from low-energyprocesses.

17

Low-energy particlesassociatedwith a singleinteractionarrive at the muonstationswith flighttimes extending to milliseconds(Figure 11), or-dersof magnitudegreaterthan the 25 ns intervalbetweenbunch crossings. This meansthat hitsrecordedin a readoutwindow containingjust oneinteraction can include contributions (spillover)from interactionsthatoccurredmany bunchcross-ingspreviously. Thefractionof hits from spilloverin a 20 ns readoutwindow rangesfrom about2%in M1 to around20% in M5. Sincespillover hitsin different stationsare essentiallyuncorrelated,their effect on occupanciesis more pronouncedthantheireffecton triggerrate.

Giventhata readoutchannelis deadfor sometensof nanosecondsafterreceiving a signal,parti-clesarriving shortlybeforethestartof thereadoutwindow cangive rise to inefficienciesfor detect-ing particlesinsidethe window (dead-timeineffi-ciency, seeSection2.3).

The effectsof spillover anddeadtime on thehitsdetectedin themuonsystemduringeachread-out interval is allowedfor by explicitly generatinghits from earlierinteractions.Thishit generationisbasedon theparameterisationsdefinedin relationto thebackgroundsimulation.

3.1.4 Digitisation

Thedigitisationmodelcloselyfollowsthebaselinechoicefor themuon-systemreadout(Section5.3).Designvaluesareusedfor the numbersof physi-calchannelsperchamber–with readoutof cathodeonly, anodeonly, or bothcathodeandanode– andfor the mappingsof physicalchannelsto logicalchannels.Basicresponsecharacteristicsof cham-bers(efficiency, crosstalk, noise,time jitter) andreadoutelectronics(noise,timing) are taken intoaccount. All responseparameterscan be freelyvariedwithin thesimulation.Nominalvalues(Ta-ble7) arebasedontestbeammeasurements,whereavailable,or on conservative estimates.

For eachhit recordedin a muon-chambergasgap,all physicalchannelslying on theline joiningtheentryandexit pointsof thehit-generatingparti-cle areinitially takenasfired at thefront end.Thenumberof gasgapsgiving input to a singlephysi-cal channelis two for theMWPCsandonefor the

Table7 Nominalvaluesfor responseparametersusedin the simulation. Cross-talkprobabilitieshave a de-pendenceon particlepositionwithin a channel,sothatvaluesshown areindicativeonly.

Single-gapjitter(root-mean-squarevalue)

MWPC anode 5.7nsMWPC cathode 6.6nsRPC 1.4ns

Single-gapefficiency 0.95

Cross-talkprobabilityMWPC 0.08RPC 0.15

ChambernoiseMWPC NoneRPC 100Hz cm2

Electronicnoise 100Hz/channel

Synchronisationimprecision 3ns

Channeldeadtime 50 ¤ 10ns

Width of readoutgate 20ns

RPCs. In regionsR1 andR2 of stationsM2 andM3 physicalchannelsarecalculatedfor thetwo in-dependentreadouts,whereasfor otherregionsonlyasinglereadoutis present.Thenumberof physicalchannelsfiredpergasgapperreadoutis a functionof the particle’s track angleandposition,andef-fectively includesthe contribution of geometricalclustersize.

A small fractionof the initially fired channelsis randomlysuppressedto simulatesingle-gapin-efficiencies.

Physicalchannelsnot crossedby any particle,but adjacentto a fired channel,are taken to bethemselvesfiredwith acertainprobability, soastosimulatedirectly inducedcrosstalk andcrosstalkdue to capacitive coupling The cross-talkmodelusedis basedon the experimentalmeasurementsreportedin Sections4.1and4.2.

Additional physicalchannelsarefired at ran-dom to simulate the effect of noise, two formsof which are considered. Chambernoise is ex-pectedto benegligible for theMWPCsbut presentatsomelevel in theRPCs.Thistypeof noiseis ex-pressedin termsof acountrateperunit area.Elec-tronicnoise,associatedwith thereadoutsystem,is

18

pp

µ−

µ­ +

M1 M2 M3 M4 M5

Muon stations

B

Figure12 Trackfindingby themuontrigger. For eachlogical-padhit in M3, hitsaresoughtin M2, M4 andM5,in afield of interest(highlighted)arounda line projectingto theinteractionregion. Whenhitsarefoundin thefourstations,anextrapolationto M1 is madefrom thehits in M2 andM3, andtheM1 hit closesttheextrapolationpointis selected.Thetrackdirectionindicatedby thehits in M1 andM2 is usedin the pT measurementfor thetrigger,assuminga particlefrom the interactionpoint anda singlekick from themagnet.In theexampleshown, µ® andµ crossthesamepadin M3.

specifiedasacountrateperphysicalchannel.At this stagein thedigitisation,asinglephysi-

cal channelmayhave receivedmorethanonesig-nal. This canoccurbecausetwo or moreparticlescrossthe channel,becauseof crosstalk or noise,or becausea particle inducessignalsin two gasgapsattachedto the samefront end. The arrivaltime of eachsignal is evaluated,taking into ac-count the bunchcrossingof origin, particleflighttimes,chamberjitter, andsynchronisationimpre-cision. Signal lossesat a channelbecauseof thedeadtime associatedwith thearrival of an earliersignalareintroduced,thenafixed-lengthtimegateis applied. Datafor thephysicalchannelsrecord-ing a signal insidethe gateare transmittedto thenext stageof thereadout.

The two front endsof a stationsareORedto-gether. Surviving physicalchannelsare mappedto logical channels,in the form of stripsor pads.For thestrip regions,padsaredefinedfrom thein-tersectionsof horizontaland vertical strips. Padinformationfrom all regionsis usedin the triggersimulationandin themuon-identificationstudies.

3.1.5 Futur edevelopments

The simulationandall resultsreportedbelow arebasedon softwaredevelopedin Fortran.In thefu-ture, high priority will be given to the migrationto object-orientedC++ softwarewithin theframe-work of Gaudi[34]. Theobject-orientedsoftwareis expectedto be at a stagewhere it can com-

pletelyreplacetheFortransoftwareaboutmidwaythrough2002.

3.2 Level-0 muon trigger

TheLevel-0 (L0) muontriggerandits implemen-tationaredescribedin detailin severalLHCb notes[35, 36]. Thearchitectureis fully synchronousandpipelined,and the algorithm run is very closetotheonereportedin theTechnicalProposal[1], andusedin theperformancestudiesdescribedbelow.

The L0 muon trigger looks for muon trackswith a largetransversemomentum,pT. It searchesfor hits defining a straight line through the fivemuonstationsandpointingtowardstheinteractionpoint (Figure12). The position of a track in thefirst two stationsallows determinationof pT.

Thissectioncontainsashortdescriptionof themuon-triggerdesign,and a summaryof the ex-pectedperformance.

3.2.1 Trigger design

TheL0 muontriggeris implementedwith thefourquadrantsof the muon systemtreatedindepen-dently. Trackfinding in eachregion of a quadrantis performedby 12 processingunits, arrangedonprocessingboardsin groupsof four for regionsR1,R3 andR4,andin pairsfor region R2 (Figure13).

A processingunit collectsdatafrom the fivemuonstationsfor padsandstripsforming a towerpointing towards the interactionpoint, and also

19

Figure 13 Organisationof themuontriggerprocess-ing boards,delimitedby thick lines,andof theprocess-ing units,shown ashatchedandwhitesquares.

receives information from neighbouringtowers.Track finding in a processingunit startsfrom theinformationfor the96 logical padsdefinedby theintersectionsof the4 horizontalstripsand24verti-cal stripsrepresentingtheunit’s input from stationM3. The tracksearchis performedin parallelforall pads.

For eachlogical-padhit in M3 (trackseed),thestraightlinepassingthroughthehit andtheinterac-tion point is extrapolatedto M2, M4 andM5. Hitsarelookedfor in thesestationsin searchwindows,termedfieldsof interest(FOI), approximatelycen-tredon thestraight-lineextrapolation.Thesizeofthefield of interestis dependentonthestationcon-sidered,thedistancefrom thebeamaxis,thelevelof background,andtheminimum-biasretentionre-quired. Whenat leastonehit is found insidethefield of interestfor eachof the stationsM2, M4andM5, amuontrackis flaggedandthepadhit inM2 closestto theextrapolationfromM3 isselectedfor subsequentuse.

Thetrackpositionin stationM1 is determinedby making a straight-lineextrapolationfrom M3andM2, andidentifying in theM1 field of interestthepadhit closestto theextrapolationpoint.

Sincethe logical layout is projective, thereisa one-to-onemappingfrom padsin M3 to padsinM2, M4 andM5. Thereis alsoa one-to-onemap-ping from pairsof padsin M2 andM3 to padsinM1. This allows thetrack-findingalgorithmto be

implementedusingonly logicaloperations.Once track finding is completed,an evalua-

tion of pT is performedfor a maximumof 2 muontracksper processingunit. The pT is determinedfrom the track hits in M1 andM2, usinglook-uptables.

Thetwo muontracksof highestpT areselectedfirst for eachprocessorboard,and then for eachquadrantof the muon system. The informationfor up to eightselectedtracksis transmittedto theLevel-0decisionunit.

3.2.2 Trigger performance

The performanceof the muon systemhas beenstudiedand optimisedusing a simulation of thetrigger algorithm [35]. This algorithm is fullyspecifiedby giving the horizontalandvertical di-mensionsof the FOI for eachstation,andthe cuton pT. The muon-systemperformanceis quanti-fied by evaluatingthetriggerefficiency for select-ing muonsfrom b-hadrondecaysasa functionofthe minimum-bias(MB) retentionlevel. Possiblelimitationsintroducedby thehardwareimplemen-tationof thetriggeralgorithmarenot considered.

3.2.2.1 Trigger efficiency

Theb µ X acceptanceis definedasthenumberof b µX eventswhereapromptmuon(µb) com-ing directly from thesemileptonicb-hadrondecaysatisfiesthemuontriggerselection,dividedby thetotal numberof b µ X eventsin the full solidangle. The b µ X acceptanceis the productoftwo contributions: thegeometricalacceptance,εG,of the detector, definedas the fraction of eventswherethepromptµb hits stationM3, andthetrig-gerefficiency, εµTr .

The value of εG is fixed for a defineddetec-tor layout, whereasthe value of εµTr dependsonthe bandwidthassignedto the L0 muon trigger(the MB retention)and also on the assumptionsmadeconcerningbackgroundlevel and detector-responseparameters.For the setupconsideredinthis report,a Monte-CarlocalculationgivesεG ~0 19 ¤ 0 01.

The rateof the L0 muontrigger is dominatedby the presenceof true muons,mainly from de-

20

Table 8 Performanceof theL0 muontriggeralgorithm,with fieldsof interestasgivenin [37]. Statisticaluncer-taintiesfor thedifferentrunningconditionsarecorrelated,asthesameeventsampleswereusedin all studies.

Nominal Maximalbackground background

MB b µ X pT cut b µ X pT cutRetention εµTr [%] [GeV c] εµTr [%] [GeV c]

1% 41 8 ¤ 1 1 1 40 ¤ 0 04 35 5 ¤ 1 8 1 69 ¤ 0 112% 55 2 ¤ 0 9 1 02 ¤ 0 02 49 7 ¤ 1 4 1 17 ¤ 0 073% 61 4 ¤ 0 7 0 75 ¤ 0 02 56 9 ¤ 1 2 0 97 ¤ 0 04

cays in flight of pions and kaons. The contri-bution from accidentalcombinationsof particletracksandbackgroundhits in stationsM2 to M5variesfrom 3%to 17%,dependingon backgroundscalefactor.

A procedurehasbeendevelopedto determinethe trigger FOIs and pT thresholdso as to max-imiseεµTr for agivenlevel of MB retention[2, 37].If two setsof parametersgive compatibleperfor-mance,the set with highest pT thresholdis se-lected.

3.2.2.2 Event samples

Performanceestimateshave beenobtainedusing104 b µ X eventsand105 non-diffractive inelas-tic interactions.Eacheventin thetwo samplescor-respondsto a singleproton-protoncollision, con-sistentwith theuseof apile-upvetoin theL0 trig-ger.

In the b µ X events, the b hadron is re-questedto be inside a 600mrad forward cone,aconditionsatisfiedby about40%of b hadronsgen-eratedover thefull solidangle,andis forcedto de-caytoamuon.Theaccompanying bhadrondecaysaccordingto branchingfraction.

Non-diffractive inelastic interactions(55mbcrosssection)are chosenas a good approxima-tion of MB events contributing to the triggerrate. A separatestudy [38], using a pT cut of1GeVc, shows that the muon-triggeracceptancefor diffractive interactionsand elasticscatterings(combinedcrosssectionof 47mb) is about2% oftheacceptancefor thenon-diffractive inelasticin-teractions.Thecontributionof diffractiveandelas-tic eventsto thetriggerrateis suppressedevenfur-ther if the muon trigger is placedin coincidence

0¯

0.2¯0.4¯0.6¯0.8¯

1

1.2

0¯

0.5¯

1 1.5 2°

2.5°

30¯25

50±75²100

125

150

175

200

225°250°0

¯0.5¯

1 1.5 2°

2.5°

3

B0 →³ J/ψ (µ´ +µ-) KSµ0

b ¶

→ µ´ XMB interactions

Sig

nal e

ffici

ency

Min

imum

-bia

s su

ppre

ssio

n

Cut on muon pT· (GeV/c)

Figure 14 Minimum-biassuppressionandtriggeref-ficiency for decaysindicated,for nominalbackground.Efficiency estimatesarerelativeto muonsthatareinsidetheacceptanceof M3 andcomedirectlyeitherfrom theb-hadrondecay(b ¸ µ X events)or from the J ψ de-cay (B0

d ¸ J ψ ¹ µ® µ»º K0S events). The FOIs usedare

optimisedfor a minimum-biasretentionof 2%.

with aninteractiontrigger, requiring,for example,aminimumenergy in thecalorimeters.

3.2.2.3 Resultson trigger performance

Several runningconditionshave beenconsidered,correspondingto different levels of MB retentionandbackground.

The performanceof the L0 muon trigger fornominal backgroundis illustrated in Figure 14,wheretrigger efficiency andMB suppressionareplottedas function of the pT cut using FOI opti-misedto give anMB retentionof 2%. Thetriggerperformancefor nominalandmaximalbackgroundis summarisedin Table8, andis discussedin moredetail in [37].

Comparisonshave beenmadewith previousstudies,performedwith simplifiedsimulations,not

21

0.2¯0.3¯0.4¯0.5¯0.6¯0.7¯

0¯

1 2°

3 4¼

5±

60.2¯0.3¯0.4¯0.5¯0.6¯0.7¯0

¯1 2 3 4 5

±6

TDR muon systemTP muon system

b ½ → µ

X e

ffici

ency

Background scale factor in M2-M5

1%

2%

3%

Figure 15 Trigger efficiency for b ¸ µ X decayswithin M3 acceptanceasafunctionof backgroundrate,for threevaluesof MB retention.Foreachsetup,theup-peredgeof a bandcorrespondsto nominalbackgroundin M1, andthe lower edgeis obtainedwith theM1 hitratedoubled.Bandedgeshave a statisticaluncertaintyof about2%.

includingdetailsof thechambergeometryandde-tector response[39], aswell aswith resultsfromsimulationsperformedatvariousstagesduringtheevolution of the detectorlayout [2]. Since theTechnicalProposal,the numberof logical chan-nelshasbeendecreasedby 40%,andarealisticde-tectorsimulationhasbeenintroduced. Neverthe-less, the improved detectoroptimisationensuresthat theb ¾ µ X triggerefficiency is slightly bet-ter thanthat quotedin the TechnicalProposal[1](Figure15).

Systematicstudieshavebeencarriedoutof theeffect on the L0 trigger of any variation in themuon-detectorefficiency andnoiselevels relativeto thedesignspecifications(Figure16), andshowthat the systemis robust [21]. In thestudies,oneof theresponseparametersis varied,while theoth-ersarekept constant.The decreasein trigger ef-ficiency, with MB retentionfixed at 1%, 2% or3%, remainsless than 5% for a single-gapeffi-ciency as poor as 80%, for RPC noiseof up to1kHz¿ cm2, and for electronicnoisein excessof1kHz/channel.

The effect of the signal time spreaddue tochamberjitter andsynchronisationimprecisionhasbeeninvestigated[40], andthetriggerefficiency isfound to be stableunderany reasonablevariationin themuon-systemtiming characteristics.

Studiesusinginclusive b-hadrondecays,gen-

0.2À0.3À0.4À0.5À0.6À0.7À

0.7À

0.75À

0.8À

0.85À

0.9À

0.95À

1

0.2À0.3À0.4À0.5À0.6À0.7À

102Á

103Â

104Ã

3% MB retention2% MB retention1% MB retention

3% MB retention2% MB retention1% MB retention

a)

b)Ä

b →

µX

effi

cien

cyb

→ µ

X e

ffici

ency

b →

µX

effi

cien

cy

Single-gap efficiency

RPC noise (Hz/cm2)

Figure 16 Trigger efficiency for b ¸ µ X decayswithin M3 acceptanceas a function of (a) single-gapefficiency and(b) RPCnoise,for threevaluesof MBretention.Curvesarefits of second-orderpolynomialsto thepointscalculatedwith thesimulation.

eratedover thefull solidangle,havealsobeenper-formed[38]. With pT cutandFOI optimisedat2%MB retention,thetriggeracceptancefor theinclu-sive b-hadroneventsis Å 5Æ 7 Ç 0Æ 2È %. Thefractionof casesin which thetrigger is satisfiedby a µb isÅ 34 Ç 2È %.

3.2.2.4 Machine-relatedbackground

Muons from the beamhalo have beensimulatedusingdistributions calculatedin studiesof beam-gasinteractionsbeforethedetectorhall. The dis-tributionsof energy andradialpositionfor muonscrossingaplane1m upstreamfrom theinteractionpoint, andtravelling in the directionof the muonsystem,are shown in Figure 17. The beam-halomuonsare concentratedat small radii, and havea wide energy spectrum.About 80% have a mo-mentumof lessthan5GeV¿ c, andso will be un-ableto penetratethefull depthof themuonsystem.Muonsenteringthe experimentalhall behindM5will notcausehits in thesamebunch-crossingtime

22

0À24É68

10

12

14

0À

50Ê

100 150 200 250 300

0À5Ê10

15

20

25Ë30

35

10-1

1 10 102Á

103Â

104Ã

a)

b)Ä

Flu

x of

muo

ns (

Hz/

cm2 )

Radial distance (cm)

Muo

ns (

MH

z)

Ì

Energy (GeV)

Figure 17 Distributionsof (a) radialpositionand(b)energy, for muonsfrom thebeamhaloenteringthede-tectorcaverntravelling in thedirectionof themuonsys-tem.

window in the different muon stations. Prelimi-narystudiesindicatethatbeam-halomuonswouldbe presentin only 1.5% of the bunch crossings,andthefractionof bunchcrossingsin whichahalomuonwould causea trigger is lessthan0.1%. Atthe presentlevel of understandingof themachinebackground,halo muonsshouldnot significantlyaffect theL0 muontrigger.

3.3 Muon identification

Efficientmuonidentificationis importantin there-constructionof physicschannelswith muonsin thefinal state,aswell asfor taggingtheflavour of theinitial b quarkin decaysrelevant to studiesof CPviolation. For rare decays,suchas B0

s ¾ µÍ µÎit is essentialto have high muon-identificationefficiency while keepingthe misidentificationofother particles,mainly pions,as low aspossible.Theperformanceof a realisticmuon-identificationalgorithm has been tested using reconstructedchargedtracksin 10000b ¾ µX events,simulatedwith nominal responseparametersfor the muon

0À

0.2À0.4À0.6À0.8À

1

2 4 6 8 10 12 14 16

M2+M3M2+M3+M4 or M5M2+M3+M4+M5µ

iden

tific

atio

n ef

ficie

ncy,

εµ

µ momentum (GeV/c)

Figure 18 Muon-identificationefficiency as a func-tion of momentum,for different requirementson thenumberof hits. At eachmomentumvalue,muonshavebeengeneratedwith uniformly distributedpolarangle.

system(Table7). Full detailsof thestudyaregivenelsewhere[41].

Well-reconstructedtracksthathaveat leastonehit in the vertex detector, and are within the ge-ometricalacceptanceof stationsM2 andM3, areextrapolatedfrom trackingstationT10to themuonsystem.The minimum momentumfor a muontopenetrateto M3 is about3GeV¿ c. A trackis iden-tified asamuonif hitsarefoundinsiderectangularsearchwindowscentredonthetrackextrapolation.Thenumberof hits requiredis dependenton trackmomentum:

Ï M2 Ð M3 for p Ñ 6GeV¿ cÏ M2 Ð M3 Ð M4 or M5 for 6 Ñ p Ñ 10GeV¿ cÏ M2 Ð M3 Ð M4 Ð M5 for p Ò 10GeV¿ cThe search-window dimensionsare parame-

terised as a function of momentumin the fourregions of each station. The parameterisationis obtained using single muons, generatedattheprimary-interactionpoint with momentumflatovertherange1 to 150GeV¿ c, andwith uniformlydistributed polar angle. The search-window di-mensionstake into accountthemultiple scatteringin thecalorimetersandmuonfilter, aswell asthefinite granularityof themuonsystem.Themomen-tum rangesindicatedabove werechosento avoiddecreasesin muon-identificationefficiency whena higher numberof stationsis required. For ex-ample, as shown in Figure 18, requiring hits inthreestationsratherthan two has little effect on

23

Table 9 Muon-identificationefficiency (%), andpar-ticle-misidentificationprobabilities(%), for b ¸ µ Xevents.

Nominal Maximalbackground background

εµ 94Æ 0 Ç 0Æ 3 94Æ 3 Ç 0Æ 3Ó e 0Æ 78 Ç 0Æ 09 3Æ 5 Ç 0Æ 2Ó π 1Æ 50 Ç 0Æ 03 4Æ 00 Ç 0Æ 05Ó K 1Æ 65 Ç 0Æ 09 3Æ 8 Ç 0Æ 1Ó p 0Æ 36 Ç 0Æ 05 2Æ 3 Ç 0Æ 1

0À0.01

À0.02À0.03À0.04À0.05À0.06À0.07À0.08À

0À

5Ê

10 15 20Ë

25Ë

30 35 40É

45É

Nominal backgroundMaximal background

π m

isid

entif

icat

ion

prob

abili

ty,

πÔ

πÕ momentum (GeV/c)

Figure 19 Pion-misidentificationprobability as afunctionof momentum,for b ¸ µ X events.

efficiency for muonshaving a momentumgreaterthan6GeV¿ c.

Table9 shows resultsobtainedwith b ¾ µ Xevents for the muon-identificationefficiency, εµ,and for the misidentificationprobability,

Ó, for

electrons,pions,kaonsandprotons.Valuesquotedarerelative to all trackssatisfyingthecriterialistedabove for being extrapolatedto the muon sys-tem.Bothnominalandmaximalbackgroundshavebeenconsidered.

It should be noted that the number of pi-onspassedto themuon-identificationalgorithmisabouta factorof 10higherthanthenumberof par-ticles of other types,so pionsareby far the mostimportantsourceof misidentification. As wouldbe expected, the misidentificationprobability ishighestat low momentum(Figure19), wherethenumberof hits requiredis smallestand,to accom-modatemultiple scattering,searchwindows arelargest.Also, pionsof lower momentumaremorelikely to decayin flight beforereachingthemuonstations.

0À0.1À0.2À0.3À0.4À0.5À0.6À0.7À0.8À0.9À 1

0À

0.02À

0.04À

0.06À

0.08À

0.1À

0.12À

0.14À

0.16À

0.18À

0.2À 0À0.01À0.02À0.03À0.04À0.05À0.06À0.07À0.08À0

À0.02À

0.04À

0.06À

0.08À

0.1À

0.12À

0.14À

0.16À

0.18À

0.2À

εÖ µ, nominal backgroundεÖ µ, maximal background×× π, nominal background

π, maximal background

µ id

entif

icat

ion

effic

ienc

y, εµ

π m

isid

entif

icat

ion

prob

abili

ty,

πÔ

Cut on ∆Sx

Figure 20 Muon-identification efficiency, andpion-misidentificationprobability, asa function of thecuton ∆Sx.

For nominal background, the muon-identification efficiency in b ¾ µ X events is94% andthe pion-misidentificationprobability is1.5%. In 61% of casesthe pion is misidentifiedbecauseit decaysin flight to a muon,andin 27%of casesthe misidentificationoccurs becauseamuonfrom anotherhadrondecayis presentcloseto the pion. In both of thesecasesthe hits foundby the identificationalgorithmaredueto muons.The remaining 12% of pion misidentificationsresultfrom somecombinationof backgroundhits,ghosthits, and hits generatedby punchthrough,alignedby chancewith thepion track.

The numberof randomalignmentsincreaseswith thebackgroundhits, so that for the situationof maximalbackgroundthemisidentificationprob-abilities aresignificantlyhigher. Improvementispossibleby requiringthatthetrackslopemeasuredin the x–z planeby the muon systembe consis-tent with the valuemeasuredin the trackingsys-tem. Figure20 shows the muon-identificationef-ficiency and pion-misidentificationprobability asa functionof thedifference,∆Sx, in the two mea-suredslopes. Undermaximalbackgroundcondi-tions, requiring ∆Sx to be lessthan 0.07 gives amuon-identificationefficiency of 90%, while thepion-misidentificationprobabilityfalls from 4%toabout2%.

In the physicsanalysespresentedbelow, onlyabout1% of the muonsrelevant for decayrecon-structionor flavour tagginghave a momentumoflessthan6GeV¿ c. As a result,misidentificationprobabilitiescanbe reduced,with negligible loss

24

of usefulmuons,by introducinganexplicit cut onmomentum.After this cut is applied,otheridenti-ficationcriteriacanbere-optimised.For particleswith momentumgreaterthan6GeV¿ c, andthesit-uation of maximal background,requiring hits inall of the stationsM2 to M5 andaskingfor ∆Sx

lessthan0.05givesa pion-misidentificationprob-ability of 1.2%,while themuon-identificationeffi-ciency remainsat90%.

Information from the RICH systemandfromthe calorimeterscanbe usedto decreasestill fur-ther the misidentificationrates. A preliminarystudy[42] indicatesthat,evenin thecaseof maxi-mal background,misidentificationprobabilitiesoflessthan1%, anda muon-identificationefficiencyabove 90%,arereadilyattained.

3.4 Reconstructionof muonic final states

The analysis of b decays with muonic finalstatesconstitutesa significant part of the LHCbphysics program. Of particular interest isthe well-establishedCP-violating decay B0

d ¾J¿ ψ Å µÍ µÎÈ K0

S, from which theangleβ of theuni-tarity trianglecanbedetermined.ThedecayB0

s ¾µÍ µÎ involvesa flavour-changingneutralcurrent,andis stronglysuppressedin theStandardModel.New physicsmight be detectedas a significantenhancementof the branchingfraction. The de-caymodesindicatedgive riseto muonswith quitedifferent kinematics,andso have beenchosentodemonstratethe physicspotential offered by themuondetector. Thestudiespresentedareprelimi-nary.

The two analysesdiscussedhave been per-formedconsideringonly single-interactionevents,95% of which areassumedto be acceptedby thepile-up veto. The combinedefficiency of triggerlevelsL1, L2 andL3 for eventsthatcontainafullyreconstructeddecay, and are passedby the pile-up veto, the L0 trigger and the offline selection,is takento be80%.

3.4.1 B0d ¾ J¿ ψ Å µÍ µÎ È K0

S

Reconstruction of the decay B0d ¾

J¿ ψ Å µÍ µÎÈ K0S Å πÍ πÎÈ is described in detail

in [43]. The first stepis to selectpairsof oppo-

0À0.1À0.2À0.3À0.4À0.5À0.6À0.7À0.8À0.9À 1

0À

1 2Ë

3 4É

5Ê

6 7Ø

8 9Ù

10

0À0.1À0.2À0.3À0.4À0.5À0.6À0.7À0.8À0.9À 1

0À

10 20 30 40 50Ê

60 70Ø

80 90Ù

100

a)

b)Ä

Di-m

uon

iden

tific

atio

n ef

ficie

ncy

J/ψÚ transverse momentum (GeV/c)

Di-m

uon

iden

tific

atio

n ef

ficie

ncy

J/ψ momentum (GeV/c)

Figure21 Probabilityto identify bothmuonsof aJÛ ψdecayas a function of (a) JÛ ψ transversemomentumand(b) JÛ ψ totalmomentum.Valuesarecalculatedrel-ative to eventsthat satisfy the L0 trigger andanalysisselectioncriteria,andhave thetracksof thetwo muonsreconstructedin the spectrometer. Lossesfrom recon-structedtracksoutsidethe geometricalacceptanceofthemuonsystemareincluded.

sitely charged tracksoriginating from a commonvertex and identified as muons by the muonsystem(seeSection3.3). To ensuregoodvertexresolution, each reconstructedmuon track isrequired to have at least one hit in the vertexdetector. A J¿ ψ candidateis obtainedwhen themassof a di-muon pair is consistentwith thatof the J¿ ψ. The resolutionon the J¿ ψ massis10MeV ¿ c2. Figure 21 shows the probability toidentify bothmuonsof theJ¿ ψ asa functionof theJ¿ ψ transverseandtotal momentum.Thedi-muonefficiency is almostflat as a function of the J¿ ψtransversemomentum,and hasan averagevalueof Å 79Æ 7 Ç 0Æ 8È %. The geometricalacceptanceofthe muon systemfor muonsreconstructedin thespectrometerdecreasesfor J¿ ψ momentabelow30GeV¿ c, becauseof theincreasedbendingof themuontracksin themagneticfield.

TheK0S candidatesarereconstructedfrom two

25

oppositelychargedtracksforming a commonver-tex andgiving amassconsistentwith theK0

S mass,the resolution for which is 3.5MeV ¿ c2. Bothtracksarerequiredto beidentifiedaspionsby theRICH system. The J¿ ψ and K0

S candidateshavetheir momentarefittedwith massconstraints,andarethencombinedto identify B0

d decays.Themo-mentumvectorof any B0

d candidateis requiredtopoint to the reconstructedinteractionvertex. Theresolutionon the B0

d massis 7MeV ¿ c2, and theproper-time resolutionis 36fs.

In the sampleof signal events, a low-levelcombinatoricbackgroundis found uniformly dis-tributedin a masswindow of Ç 60MeV ¿ c aroundthe peak. All of the backgroundeventsare dueto the combinationof a true J¿ ψ from a b-hadrondecay, and a true K0

S, either from fragmenta-tion or from the decay of the other b hadron.The total numberof backgroundeventscan thenbe obtainedby multiplying the numberof back-groundeventsfound in the signal sampleby theratio BRÅ b ¾ J¿ ψX ÈÝÜ BRÅ J¿ ψ ¾ µÍ µÎ ÈÞ¿ßÅ fb ÜBRÅ B ¾ J¿ ψ Å µÍ µÎÈ K0

S Å πÍ πÎÈÞÈÞÈ , wherefb is theprobability that a b quark will fragment into aB0

d. The resultingbackgroundis assumedto re-tain a flat distribution within the masswindow,andto have thesametriggeracceptanceasthesig-nal. Any combinatoricbackgroundarising fromthe misidentification of pions as muons in themuonsystemis negligible. The resultingsignal-to-backgroundratiounderthemasspeak( Ç 3 stan-darddeviations)is 3 Ç 1.

The L0-trigger efficiency for eventsthat havea fully reconstructeddecayand satisfy the anal-ysis selectioncriteria is Å 98Æ 1 Ç 0Æ 3È %. The L0-muon trigger efficiency is Å 95Æ 2 Ç 0Æ 5È %, andÅ 70 Ç 1È % of the fully reconstructeddecaysareselectedonly by the muontrigger. With a B0

d ¾J¿ ψ Å µÍ µÎ È K0

S Å πÍ πÎ È visiblebranchingfractionof1Æ 8 Ü 10Î 5, thepresentanalysisshows thatLHCbwill fully reconstructmore than 105 decaysperyear.

3.4.2 B0s ¾ µÍ µÎ

The B0s ¾ µÍ µÎ decayis mediatedby a flavour-

changingneutralcurrent. In the StandardModel,it occursvia loopdiagramsandthebranchingfrac-

0À0.1À0.2À0.3À0.4À0.5À0.6À0.7À0.8À0.9À 1

0À

3 6 9Ù

12 15

0À0.1À0.2À0.3À0.4À0.5À0.6À0.7À0.8À0.9À 1

0À

30 60 90Ù

120 150

a)

b)Ä

Di-m

uon

iden

tific

atio

n ef

ficie

ncy

Bsà transverse momentum (GeV/c)0

Di-m

uon

iden

tific

atio

n ef

ficie

ncy

Bs momentum (GeV/c)0

Figure 22 Probabilityto identify bothmuonsof a B0s

decayasafunctionof (a)B0s transversemomentumand

(b) B0s total momentum.Valuesarecalculatedrelative

to events that satisfy the L0 trigger and analysisse-lection criteria, andhave the tracksof the two muonsreconstructedin the spectrometer. Lossesfrom recon-structedtracksoutsidethe geometricalacceptanceofthemuonsystemareincluded.

tion is calculatedto beabout3Æ 5 Ü 10Î 9 [44]. Thestudyof LHCb sensitivity to B0

s ¾ µÍ µÎ reportedbelow is preliminary, and is currently limited bythe low statisticsavailablefor thebackgroundes-timate.Detailsof thereconstructionprocedurearereportedin [45].

The analysisusespairs of well-reconstructedoppositelychargedtracksformingasecondaryver-tex, andinconsistentwith anorigin at the interac-tion point. Thecombinedmomentumvectorof thesecondarytracksis requiredto point back to theprimary vertex. In the absenceof muon identifi-cation, the µÍ µÎ massspectrumfor the pairs ofsecondarytracksin simulatedB0

s ¾ µÍ µÎ eventsis dominatedby thecombinatorialbackground.AB0

s masspeak,with awidth of 18MeV/c2, emergeswhen the criteria for muon identification (Sec-tion 3.3)areapplied.

The geometricalacceptanceof the muonsys-

26

tem for B0s ¾ µÍ µÎ decayswith both muonsre-

constructedis Å 92Æ 9 Ç 0Æ 7È %. Theefficiency of themuon-identificationprocedurefor pairsof recon-structedmuontracksinside the muon-systemac-ceptanceis Å 88 Ç 1È %. Figure22shows theproba-bility of having bothmuonsinsidethegeometricalacceptanceof themuonsystemandcorrectlyiden-tified, asa functionof the B0

s transverseandtotalmomentum.Theefficiency is essentiallyflat for aB0

s momentumabove 25GeV¿ c.The L0-trigger acceptancefor decayswith

both muonsidentified is Å 97Æ 7 Ç 0Æ 5È %, the L0-muonacceptanceis Å 97Æ 3 Ç 0Æ 5È %, and Å 74 Ç 1È %of eventsare selectedonly by the muon trigger.After oneyearof LHCb operation,the estimatednumbersof signal andbackgroundeventsare10and3.3respectively.

3.5 Muon tagging

The ability to tag the flavour of the initial stateof B mesonsusing the muon systemhas beenstudied. For this study, fully reconstructedB0

d ¾J¿ ψ Å µÍ µÎÈ K0

S events selectedby the L0 triggerhave beenused. The B0

d ¾ J¿ ψ Å µÍ µÎ È K0S recon-

structionis asdescribedabove (Section3.4.1).Thefollowing pre-selectioncutsareappliedto

reconstructedcharged tracksnot associatedwiththeB0

d ¾ J¿ ψ Å µÍ µÎ È K0S candidate,to obtainasam-

pleof trackswith ahighprobabilityof beingdecayproductsof theaccompanying b hadron:

1. pT Ò 1Æ 2GeV/c;

2. numberof hits in vertex detectorÒ 1;

3. impact-parametersignificanceÒ 3;

4. impactparameterÑ 2mm.

Tracksarethenselectedasmuonsif they areiden-tified assuchby themuon-identificationalgorithm(Section 3.3). An averageof 0.053 tracks perevent passthis selection,of which 89Ç 3% aremuons. Some81Ç 4% of thesemuonsare fromsemi-leptonicb-hadrondecaysand15Ç 4 arefromb ¾ c ¾ µ decaychains. If morethanonetrackpassesall cuts, the track with the highesttrans-versemomentumis used. The charge of the se-lectedtrack is usedto tagtheflavour of the initial

stateof thereconstructedB meson.Theoverallef-ficiency, ε, of the muon tag is Å 5Æ 3 Ç 0Æ 5È %, andthemistagrate,ω, is Å 27 Ç 5È %. For perfectmuonidentification the efficiency is Å 5Æ 1 Ç 0Æ 5È %, andthemistagrateis Å 25 Ç 5È %. Thesetwo quantitiescanbe combinedinto a measureof the statisticalpower, á , of thetag:

áãâåä ε Å 1 æ 2ωÈÞÆForperfectmuonidentification,áçâèÅ 0Æ 11 Ç 0Æ 02È ,whereasin thepresentstudy áåâãÅ 0Æ 10 Ç 0Æ 02È .

27

4 PrototypeStudies

An intensiveprogrammeof developmentwork hasbeenundertakenfor theLHCb muonsystem.Pro-totypesof MWPC and RPC detectorshave beenconstructed,allowing thestudyof anumberof im-portantpropertiesin testbeamsandin the labora-tory. A summaryof the work is describedin thissection.

4.1 MWPC studies

Generalfeaturesof the MWPC designandoper-ation aresummarisedin section2.2.1. Prototypechambershavebeenevaluatedin severaltest-beamstudiesover the pasttwo years,and someof thekey results are reportedhere. A full presenta-tion of resultscanbe found in Refs. [46, 47, 48,49]. Beforediscussingthechamberperformance,the front-endspecificationsaredescribed,andanoverview is given of the different front-endchipstested.

4.1.1 Front-end electronics

4.1.1.1 Front-end specifications

Parameterscharacterisingthe readoutelectronicsareshown in Table10, andarediscussedin detailin Refs.[16] and[50].

Since both cathodepads and wire pads areread, the front end must be able to handlebothnegative andpositive polarities. For a double-gapchamberwith agasgainof 105, theaveragesignalchargein thefirst 10nsis 40fC on theanode-wirepads,andhalf thisvalueonthecathodepads.Lan-daufluctuationsresultin thesignalhaving a largedynamicrange,sothatascaleextendingto 150fCis requiredto guaranteeoptimumtail cancellationfor more then95% of the signals. The optimumamplifier peakingtime tp is 8ns, the time reso-lution degradingfor both shorterand longer val-ues. Figure 23 shows the simulateddependenceof the time resolutionon amplifier peakingtimeandthreshold.Thenoiserateperchannelrequiresa thresholdof at least6 primary electronsat theworkingpoint.

In the regions of highest occupancy, signalrates of up to 1MHz/channel are anticipated.

Table 10 ElectronicsParameters.

Parameter

Av. chargein 10ns 40fC (doublegap)Input polarity positive andnegativeSignaltail t0=1.5nsDetectorcapacitance 40-250pF(doublegap)Maximumsignalrate 1MHzMaximumtotaldose 1MRadDecouplingcapacitors 1nF(doublegap)Loadingresistors 100kΩCoupling AC for wire signals

DC for cathodesignalsSpecifications

Peakingtimeatdisc. é 10nsEquiv. noisecharge Ñ 2fC (Cdet=250pF)Linearrange 150fCInput resistance Ñ 50ΩShapingcircuit unipolar2 Ü pole/zeroAveragepulsewidth Ñ 50ns(ASD output)Baselinerestorer é 1µs responsetime

0ê12

3ë45ì6í7î8ï9ð10

0ê

5ì

10 15 20 25 30ë

tñ

p=3ns

tñ

p=10ns

tñ

p=15ns

Measurement tò

p=8ns

Threshold (Primary Electrons)

Tim

e R

MS

(ns

)

Figure 23 Simulateddependenceof the time resolu-tion on amplifierpeakingtime andthreshold,for a sin-gle MWPCgap.At a gainof 105 thenoiselevel allowsa thresholdof ó 6 primaryelectrons.

Unipolar pulse shapingis usedto minimise thedeadtime, which is the largest sourceof ineffi-ciency. The signal has an ion tail, with a con-stantof t0 ô 1Æ 5ns[50]. A dedicatedcancellationnetwork is neededfor tail suppression,the stan-dard solution beinga doublepole/zerofilter net-

28

Table 11 Parametersof front-endchips tested. Thepeakingtime tp is givenfor smallvaluesof Cdet.

tp (ns) ENC(eÎ ) Rin Å Ω ÈPNPI 8 1250+50eÎ /pF 25ASDQ++ 10 1740+37eÎ /pF 25SONY++ 10 1962+37eÎ /pF 25CARIOCA 10 750+30eÎ /pF õ 10

work. The averagearrival time of the last elec-tron is about30ns,sothefront endshouldnot in-creasethe deadtime to morethan50ns. The in-put resistancemustbe smallerthan50Ω, to limitcrosstalk from capacitive couplingbetweenadja-centchannels.Thenoiselevel shouldbeassmallaspossible.Sincethedetectorcapacitancerepre-sentsa seriesnoisesource,thenoiselevel is com-pletely determinedby the front-enddesign. TheASDQ++ front end (seebelow) hasa good per-formanceacrossthe entire capacitancerange,sothat the 1740+37e/pF noise of this front end isspecifiedas the maximum tolerated. The wiresignalsareAC coupledandthe HV loadingtimeis τ â RLCdec â 100µs. This implies large base-line fluctuationsat high ratesfor a unipolarlinearsignal-processingchain. The front end must in-cludea baseline-restorationcircuit (BLR) to avoidthisproblem.

Variousfront-endchips have beenstudiedinorder to find the optimal solution for the muonsystem. Their characteristicparametersaresum-marisedin Table11. The aim of the studieshasbeento find a single chip satisfyingthe require-mentsfor all regions of the system. Resultsob-tainedarediscussedbelow.

4.1.1.2 Comparison of fr ont-end chips

ThePNPIelectronics[6], built from discretecom-ponents,consistsof an on-chamberpreamplifierandanoff-chambermainamplifier. Themainchar-acteristicsarea peakingtime of 8ns,an input re-sistanceof 25Ω, anda tail-cancellationnetwork.Theequivalentnoisecharge(ENC) hasbeenmea-suredto be 1250+50e/pF. This frontendwas de-velopedespeciallyfor theMWPCsandservesasareferencefor all otherfront-endtests.

TheSONY chip [51], originally developedfor

the ATLAS TGCs, has beentestedon a proto-type with readoutof both wire padsandcathodepads[5]. The drawbacksof this chip area largedeadtime,dueto themissingtail-cancellationcir-cuit; a poor time resolution,due to the low sen-sitivity; a largepeakingtime;andprovenradiationhardnessof upto only 50kRad.It is, therefore,notpossibleto usethis chip without additionalcom-ponents. With additional components,this chipwould be viable only for the chamberswith largewire pads,for which therequirementson radiationhardnessarelessstringent.For thelargewire pads,theconnectionof eightchannelsto onechipresultsin readouttraceswith a lengthof up to 25cm. Itis, therefore,advantageousto have a preamplifiercloseto thepad,andto sendtheamplifiedsignalsto a multi-channelmainamplifieranddiscrimina-tor. Sucha schemehasbeenimplementedwith adiscretecomponentpreamplifierandshapercloseto thepad,andtheSONYchipservingasmainam-plifier anddiscriminator(SONY++). This schemehasa performancevery similar to thatof thePNPIelectronics[49].

TheCMSelectronics,originally developedforthe CMS cathode-stripchambers,hasbeentestedonawire-padchamber, but gaveunsatisfactoryre-sults, probablybecauseof the long peakingtimeof 30ns. Anotherdrawbackof thechip is thelongdeadtime, in excessof 200ns. This chip is notappropriatefor theLHCb muonsystem.

The ASDQ chip [47], developedfor the COTchamberat Fermilab, is an offspring of the AS-DBLR chip developedfor the ATLAS TRT. Theprinciple characteristicsare a peaking time of8ns, a tail-cancellationnetwork, an input resis-tanceof 280Ω andanENC of 1100+70e/pF. Ex-cept for the large input resistanceand the noiseslope, this chip is well matchedwith require-ments. The weakpointshave beenovercomebyaddinga common-basetransistorexternal to thechip (ASDQ++). This lowers the impedancetoabout25Ω (ASDQ++), andresultsin an ENC of1740+37e/pF[47]. This front endis apossibleso-lution for MWPCsin all regions.

The prefered solution, however, is theCERN And RIO Current-modeAmplifier (CAR-IOCA) [52], a0.25µmCMOSchip,which is at anadvancedstateof development.Samplesbuilt ac-

29

Detector Capacitance [pF]

Equ

ival

ent N

oise

Cha

rge

[e]

Figure 24 Dependenceof equivalentnoisechargeondetectorcapacitance,for CARIOCA chip. The mea-surednoiseslope(closedcircles) is 30eö /pF, in goodagreementwith thecalculation.

cordingto the final designshouldbe deliveredinthesecondhalf of 2002.Theattractive featuresofthe technologychosenareradiationhardnessandlow cost. Moreover, the measuredENC is only750+30e/pF. Noise studiesfor a prototypechipshow agreementwithin 10% betweensimulationandmeasurement(Figure24).

4.1.2 Resultsof MWPC prototype tests

4.1.2.1 Wire-padreadout

In region R4, only wire padswill be read, andcathodeswill be grounded.Efficienciesof singlegapsandof a doublegapfor an 8 Ü 16 cm2 wirepad,with readoutusingPNPIelectronics,is shownin Figure25. The correspondingADC andTDCspectrafor thedoublegap,at a voltageof 3.05kV,areshown in Figure26.

Eachlogical channelis to receive inputsfromtwo doublegapsperstation.Theefficiency plateaufor a single double-gapchamberstartsat about2.95kV (95%efficiency) andendsat 3.35kV, giv-ing a comfortableoperatingrangeof 400V. Theplateauendis definedby thevoltageabove whichthe dark count rateof a padexceeds1kHz. Theintendedworking point is 3.05kV, which is 100Vabove the start of the plateau. This ensuresthatthereis goodredundancy, andthat the muonsys-

80

82.5

85

87.5

90

92.5

95

97.5

100

102.5

105

2.9÷

3ø

3.1ø

3.2ø

3.3ø

3.4ø

HV (kV)ù

Effi

cien

cy (

%)

R.M.S. Double Gap

R.M.S. Single Gaps

Eff. Single Gaps

Eff. Double Gapú

0

1

2÷3ø45

6

7

8

9

102.9÷

3ø

3.1ø

3.2ø

3.3ø

3.4ø

Tim

e R

esol

utio

n (n

s)

Figure 25 Efficiency (20ns time window) and timeresolutionmeasuredfor two single gaps(shadedandunshadedcirclues) and for one double gap (shadedsquares),for an 8 û 16cm2 wire padreadusingPNPIelectronics[49].

0 50

100 150 200 250 300 350 400 450

0 5 10 15 20 25 30 Tü

ime (ns)

0 10 20 30 40 50 60 70 80

0 20 40 60 80 100

Amplitude (ADC channel)

Figure 26 TDC spectrum(top) and ADC spectrum(bottom)measuredfor a doublegap,for an8 û 16cm2

wire pad at an operatingpoint of 3.05kV [49]. Thetail in the TDC spectrumis dueto primary-ionizationstatisticsandto electron-drifteffects.

tem has low sensitivity to environmental varia-tions. The electricalcrosstalk – the probabilitythata particlecrossingthecentreof onewire padcausesthe firing of a neighbour– hasbeenmea-suredto be5%for 3.15kV and10%for 3.25kV.

30

90ð92ð94ð96ð98ð

100

2.8 2.9 3ë

3.1ë

3.2ë

3.3ë

3.4ë

HV (kV)

Effi

cien

cy (

%)

15ns

20nsý25ns

Figure 27 Double gap efficiency (time windows of15, 20, 25ns) for a 4 û 16 cm2 wire pad, readusingtheASDQ++front end[47].

97.5ð97.75ð 98ð98.25

ð 98.5ð98.75ð 99ð99.25

ð 99.5ð99.75ð 100

0ê

100 200 300ë

400 500ì

600í

Rate (kHz)

Effi

cien

cy (

%)

EfficiencyþSigma

1

1.5

2

2.5ý3ë3.5ë4

0ê

100 200 300ë

400 500ì

600í

Tim

e R

esol

utio

n (n

s)

Figure 28 Double-gapefficiency (20ns time win-dow) andtimeresolutionasafunctionof rate,for awirepad. The fact that the time resolutionis unchanged,andtheefficiency dropswith a slopeof 0.4%/100kHz,showsthattheinefficiency is dueonly to pile-upof sig-nalswith 50nsaveragepulsewidth [47].

TheSONY++optionhasalsobeentestedwiththe wire-padchamber, and shows a performancevery similar to that of the PNPI electronicswithrespectto efficiency andcrosstalk.

The ASDQ++ performancefor a 4 Ü 16 cm2

wire pad(Cdet ô 100pF) is shown in Figure27,fortime windows of 15, 20 and25ns. Theefficiencyin 20nsis 94%at 2.9kV and99%at 3.15kV. Theplateaulengthis 450V.

88ï90ð92ð94ð96ð98ð

100

102

2.9 2.95 3ë

3.05ë

3.1ë

3.15ë

3.2ë

3.25ë

3.3ë

3.35ë

3.4ë

HV (kV)

Effi

cien

cy (

%)

PNPIÿ

ASDQ++

Figure 29 Double-gapefficiency (20ns time win-dow) for a 2 û 8 cm2 cathodepad,readusingthePNPIelectronicsandusingASDQ++[47].

Theelectricalcrosstalk hasbeenmeasuredtobe2.5%at3.15kV and5%at3.25kV.

The resultof a high-ratetest is shown in Fig-ure 28. The efficiency lossdue to signalpile-upis about0.4%/100kHz for a 20ns time window,compatiblewith themeasuredaveragesignalpulsewidth of 50ns. The fact that the measuredroot-mean-squarevalue for the time resolutionis in-dependentof the rate indicatesthat, asexpected,thereareno detectoreffects.

4.1.2.2 Cathode-padreadout

In region R3,cathodepadswill beread,andwireswill be groundedat AC. The efficiency for a 2 Ü8 cm2 cathodepad(Cdet â 40pF), readusing thePNPIelectronicsandusingASDQ++,is shown inFigure29.

The start of the plateauis shifted by about100V relative to thatfor thewire pads,becausethesignalon a cathodepadis only half thesignalat awire pad. This still leaves a comfortableplateauof 350V. Theeffect of the input resistanceon thecrosstalk is shown in Figure30. The pad-to-padcrosstalk at the working point of 3.15kV is 22%for ASDQ (Rin â 280Ω) and2.7% for ASDQ++(Rin â 25Ω). As expected,thecrosstalk is propor-tional to theinput resistance.

In the chambertested,the cathodepadswere

31

0ê5ì10

15

20

25ý30ë35ë4045

50ì

2.9 3ë

3.1ë

3.2ë

3.3ë

3.4ë

HV (kV)

Cro

ssta

lk (%

)

R

in=280Ω

R

in=25Ω

Figure 30 Cross talk between two cathode padssharing an 8cm edge, for input resistancesof 25Ω(ASDQ++) and280Ω (ASDQ). The crosstalk is pro-portionalto theinput resistance[47].

0.6ê0.8ê 1

1.2

1.4

1.6

1.8

2ý2.2

2.4ý

0ê

2 4 6í

8ï

10 12 14 16 18 20

X-coordinate (mm)

Clu

ster

Siz

e

MeasuredòSimulated

Figure31 Measuredandsimulatedcrosstalk for par-ticlescrossingthechambercloseto theborderbetweentwo cathodepads[47].

separatedby only 0.4mm, resultingin a crossca-pacitanceof 1pF/cm. In thefinal design,thepadsareseparatedby 1.3mm,with anadditionalguardtrace. This will reducecoupling, and thereforecrosstalk, by a factor 4. The crosstalk due todirect induction(Figure31) is in goodagreementwith thesimulation.

4.1.2.3 Combined readout

In regionsR1 andR2, cathodepadswill be readfollowing a chess-boardpattern. Wire padswillalsobe readin stationsM2 andM3. The criticalissuesin theseregionsare, therefore,the readouttraceson the cathodeboards,and the combinedreadoutof cathodesandwires.A prototypecham-ber containingall of the cathodestructuresof R1andR2 hasbeentestedwith theASDQ++chip. Aschematicview of this prototypeis shown in Fig-ure32. Thechambercontainstwo doublegaps:itis a full-scaleprototypewith four sensitive gaps.

Small cathodepadswith an edgein common(for example,1s and 3s in Figure 32) showed amutual capacitanceof 1pF. The capacitancebe-tweenneighbouringsmallcathodepadswherethereadouttraceof onepadrunsunderneaththeotherpad(for example,1sand2sin Figure32)hasbeenmeasuredto be 4pF. Both measuredcapacitancesmatchthesimulationwell, showing that thesepa-rametersarewell understood.

The efficiencies of small cathodepads andlarge wire pads,for combinedreadout,areshownin Figure33. If thethresholdfor thecathodepadshadbeenthe sameasfor the wire pads,the cath-odeefficiency plateauwould be shiftedby 150V.Theshift is smallerin practicebecausethethresh-old couldbesetlowerfor thecathodepadsthanforthewire pads,theformerhaving lessnoise.

Wire-padcrosstalk hasbeenstudiedby evalu-ating the clustersizefor a chamberinclination of4

with respectto thebeam.Clustersizesof 1.12at 3.15kV and1.23at 3.3kV have beenmeasuredfor 1.2cm wire pads.

Electrical cross talk betweencathodepadsthroughwire pads,a crucial numberfor thecom-binedreadout,hasbeenmeasuredby focusingthebeamon onecathodepadandcountingcross-talkhits in its vertical neighbours. Only out-of-timecrosstalk hasbeenfoundfor this coupling,asex-pected.Thecross-talkprobabilities(padpositionsasshown in Figure32) are0.62%(5s)and0.57%(3s)at 3.15kV, 1.9%(5s)and2.1%(3s)at 3.3kV,well within designspecifications.Sincethe padstestedcorrespondto the largest wire padsto beusedwith combinedreadout,thiskind of crosstalkwill notposeaproblem.

32

121 24 36 48 60

15cm7.5cm

3.15cm1l2l3l

5l4l

6l

8l

9l10l

12l

14l15l16l

11l

13l

1s

3s2s

4s5s6s7s8s9s

11s12s

14s13s

15s16s

10s

7l

Large Wire Pads (1.2cm) Small Wire Pads(0.6cm)

Figure32 Schematicview of theprototypefor regionsR1andR2 of M2 andM3.

90ð92ð94ð96ð98ð

100

2.9 3ë

3.1ë

3.2ë

3.3ë

3.4ë

Voltage (kV)

Effi

cien

cy (

%)

Wire Pad Gap1Cathode Pad Gap1

Wire Pad Gap2Cathode Pad Gap2

Figure 33 Efficienciesfor the two doublegapsof achamber, for wire padsand cathodepads,with com-bined readoutusing the ASDQ++ and an input resis-tanceof 25Ω. The thresholdfor the cathodepad islower thanthat for the wire pad,andso the curvesareverysimilar.

The electricalcrosstalk due to direct capaci-tive couplinghasbeenevaluatedby focusingthebeamon pad7s andcountingthe hits on pad8s.The crosstalk is entirely in time, with probabili-tiesof 1.9%at3.15kV and6.5%at3.3kV (mutualcapacitanceof 4pF).Sincethemutualcapacitancebetweenpadswill be õ 4pF in the entire muonsystem,theproblemof readouttracesisconsideredsolved.

4.1.2.4 Conclusions

Test-beammeasurementsfor prototypesindicatethat the MWPCs satisfy all requirementsfor theLHCb muon systemwith sufficient redundancy.All major potentialproblemshave beenshown tobe solved. The importanttaskfor the nearfutureis to test full-size prototypeshaving all channelsequippedwith electronics,to evaluatethe perfor-manceof acompletesystem.

4.1.3 Ageing

Theperformanceof chambersafter intenseirradi-ation is a major concernof the experiment. Lo-cal ageingtestshave beencarriedout at PNPI,us-ing thesamegasmixture andsimilar materialstothoseplannedfor theMWPCsof themuondetec-tor. Thesetestsshow no ageingup to anaccumu-latedchargeof 13C/cm[53].

Globalageingtestsarebeingperformedat theCERN GammaIrradiation Facility (GIF), whichprovides a very intense(740GBq) 137Cs source.The CMS EMU group hasrecentlyconductedaglobalageingtestof this typewith achamber[54]similar to the LHCb chambers. This shows noseriousageingeffects up to collectedchargesof0.4C¿ cm on thewiresand0.5C¿ cm2 on thecath-odes.

A globalageingtestof a prototypeof a LHCbMWPC,constructedwith thecomponentsandma-terials to be employed in the experimentis cur-rently in progress. Unlike the CMS chambers,

33

5

0

100

150

200

2

50

3

00

0

2

0 4

0 6

0 8

0 100

Time (days)

I (µA

)

9

30

9

50

9

70

9

90

1010

1030

P (m

bar)

Gap Currents Pressure

Figure 34 Current(I) drawn in eachof four gapsintheMWPCprototype,asa functionof time. Theatmo-sphericpressure(P) during this period is also shown.Thegapsindicateperiodswheretheageingtestwasin-terruptedandthedip at72 dayswasdueto a largedropin gastemperature.

this chamberhas gold-patedcathodes,as fore-seenin the regions of highestirradiation, to im-prove the ageingperformance.The charge accu-mulatedso far is é 0.32C/cm2 on the cathodes,and about 0.1C/cm on the anodewires, corre-spondingto abouttwo LHCb years.The testwillcontinueuntil the endof 2001,whenan accumu-latedcharge of 0.5C¿ cm canbe expectedfor thewires.Thiscorrespondsto theirradiationof nearlyten LHCb yearsin the regions of highestinten-sity for which MWPCsareconsidered.Similarly,theaccumulatedchargeon thecathodesshouldbeabout1.7C¿ cm2. With thechargeaccumulatedtodate,no ageingeffectshave beenobserved. Cur-rentsrecordedfor eachsenstivegapduringthefirst97daysof testaredisplayedin Figure34,togetherwith thevariationin pressure.Theincreasesin cur-rent after 23 daysandafter 62 daysaredue to a50V increaseof highvoltage,to acceleratethetest.The currentfluctuationsaremainly due to varia-tionsin temperatureandatmosphericpressure.

4.2 RPC studies

SeveralRPCprototypeshavebeenbuilt andtested,to select the most appropriatesolution for theLHCb muonsystem.Mostof theresults,in partic-ular thoseon ratecapability, have alreadybeenre-portedelsewhere[3, 19]. In this section,previoustestsarebriefly summarised,thenrecentdevelop-ments,concerningthefront-endelectronics,widerreadoutstripsandageingtests,arediscussed.

Chambersmeasuring50 Ü 50 cm2, and con-sisting of one or two gasgaps,have beenbuilt.A single gas gap has a depth of 2mm, and iscontainedbetweentwo bakelite plates,each2mmthick and characterisedby a resistivity of 9 Ü109 Ωcm. The basicstepsin the gapconstructionare the following. First, a polycarbonateframe,7mm wide, is gluedto oneof the bakelite plates.Next, rows of disc-shapedpolycarbonatespacers,Å 2Æ 00 Ç 0Æ 01È mm highand10mmin diameter, areglued to the sameplate. The row separationandthe centre-to-centredistancebetweenspacersareboth 10cm. The gasgapis closedby the secondbakelite plate,thenthebakelite’s internalsurfacesaretreatedwith linseedoil. Theexternalsurfacesof thebakelite arecoatedwith a spray-onlayerofgraphite,which serves to distribute voltages,andare covered with a PET insulating film, 200µmdeep. Four gasinlets/outletsarepositionedcloseto thestructure’s corners.

The readoutelectrodesarestripswith a widthof 3cm or 6cm, and lengths varying between25cm and 50cm. Adjacent readoutstrips areshieldedby 0.5mm guardstrips, to reducecrosstalk.

The gasmixture usedis C2H2F4 ¿ C4H10¿ SF6

(95:4:1) [55]. The main constituent,C2H2F4, isa non-flammable,environmentallysafegas[55],characterisedby high density and large primaryionisation( Ò 60 primary ion pairsper cm). Thepercentageof isobutaneusedis below theflamma-bility thresholdof 5.75%. The admixtureof SF6

reducestheformationof streamers.

4.2.1 Front-end electronics

In Ref. [3], two possibilitiesfor the readoutchiparepresented:theGaAs-basedsolutionadoptedbyATLAS [10], andthe BiCMOS-basedchip devel-opedby the Bari group for CMS [56]. Unfortu-nately, theproductionprocessonwhichtheformersolutionwasbasedis now obsolete,andsotheAT-LAS chipmustberuledout.

Readouthas beenperformedusing two dif-ferent typesof front-endelectronics:(1) a “stan-dard” chainconsistingof hybrid fastvoltageam-plifiers,with a gainof about300,followedby dis-criminators; (2) the 8-channelintegratedcharge

34

I n

V

test Preamplifier

D

ummy Preamplifier

V

Bias

G

ain Stage

V

th

Z

ero-crossing Discr iminator

V

mon V

dr ive

M

onostable D

r iver

O

utp O

utm

Figure 35 Block diagramfor a singlechannelof theCMSfront-endchip [56].

Table 12 Nominal propertiesof the CMS readoutchip.

Technology 0.8µm BiCMOSDimensions 2.9Ü 2.6mm2

Input impedance é 15ΩInputpolarity negativeDynamicrange 20 fC – 20 pCThresholdrange 20 fC – 500fCChargesensitivity é 1 mV/fCPreamp.bandwidth 116MHzEquivalentNoiseCharge 4fCOutputpulsewidth 50ns– 300nsOutputlevel LVDSPowersupply +5 V, GNDPowerconsumption é 45mW/channel

preamplifier-discriminator chipusedby CMS.TheASIC chipfor theCMSfront endhasbeen

designedandmanufacturedusing0.8µm BiCMOStechnology. Thecircuit compriseseightchannels,eachconsistingof an amplifier, a zero-crossingdiscriminator, a mono-stableand a differentialLVDS line driver. A block diagramfor a singlechannelis shown in Figure35. The preamplifierhasan input impedanceof about15Ω at the sig-nal frequencies.This valuehasbeenchosenin or-derto matchthelowestvaluefor thecharacteristicimpedanceof the CMS strips,and is closeto thecharacteristicimpedance(13Ω) of the 6cm widestripsconsideredfor theRPCsof theLHCb muonsystem.

The propertiesof the CMS readoutchip aresummarisedin Table12.

An 8-channelfront-endboard(Figure36) hasbeenbuilt andtestedin thelaboratory. This proto-typeis designedto readeightstripsof 3cm width,with the channelinputs directly solderedon thestrips,i.e. without intermediateconnectors.It in-

Figure 36 Photographof the first version of thefront-endboard.

V

threshold =100mV

q in (fC)

effi

cien

cy

q eff =81.6188fC

d=0.241154

0.5

q eff

RMS 0.1068

delay(ns)

entr

ies

0

0.2

0.4

0.6

0.8

1

70 72 74 76 78 80 82 84 86 88 90

0

100

200

300

400

500

600

108 109 110 111 112 113 114 115 116

Figure37 Typicalresultsfor (top)responseasafunc-tion of input chargeand(bottom)time jitter, from chipcalibrationat anominalthresholdof 100mV.

cludesan adjustablevoltagefor remotethresholdsetting. Theboardis poweredby a singlevoltagesupply.

The performancemeasuredfor the readoutchip is in agreementwith that seen by CMS,with small variations among different channelsand chips. The threshold calibration is about1Æ 1fC/mV. Figure 37 shows the responseas afunction of the signal charge, and the time jitter(root-mean-squarevalueof lessthan0Æ 11ns). Theequivalentnoisecharge at the input is about4fC.Thevariationin time delayamongtheeightchan-nels is lessthan 0Æ 35ns, well below the require-mentfor theapplicationconsidered.

35

S

ource

S

cintillator S

cintillator H

odoscope Muon Beam

3

Independent RPCs

Pre-Amp Electronics

105 cm

Figure38 RPCtestsetupatGIF

0

0.2

0.4

0.6

0.8

1

8500 8750 9000 9250 9500 9750 10000 10250 10500

HV (V)

Effi

cien

cy

500 Hz/cm 2

900 Hz/cm 2

1.8 kHz/cm2

Figure 39 Efficiency curvesfor a single-gapRPCinGIF, for theirradiationratesindicatedon theplot.

4.2.2 Resultsof RPC prototype tests

4.2.2.1 Rate capability

The maximumrateconsideredin the RPCdetec-tor is 750Hz¿ cm2(seeTable2). TheRPCperfor-mancefor sucha ratehasbeeninvestigatedusingparticlebeamsandthe CERNGammaIrradiationFacility (GIF) [57]. The latter is designedto ex-poselarge-areadetectorsto a continuousphotonload,with fluxescomparableto thoseexpectedatLHC. The photonflux canbe attenuatedandad-justedusinga systemof filters. Thesetupfor theGIF testsis shown in Figure38.

Severalsingle-gapRPCshave beentested.Allhad the same design, with 3-cm-wide readoutstrips. A double-gapRPC, following the CMSconfiguration(single central readoutplane) has

eff. .OR. vs HV

clus. size vs HV eff. theoretical .OR. vs HV

HV (KV)

effic

ienc

y

clus

ter

size

SRPC2 .or. SRPC3

0.8

0.825

0.85

0.875

0.9

0.925

0.95

0.975

1

9!

9.5!

10 10.50"0.5"1

1.5

2#2.5#3$3.5$4%

Figure 40 Efficiency andclustersize for the logicalOR of two single-gapRPCswith readoutstrips 3cmwide.

also beentested. The chamberswere readusingthe“standard”electronics(seeSection4.2.1).

All single-gapRPCs gave an efficiency ofmorethan95%for an irradiationrateof up to 1.8kHz

&cm2 over the entire detectorsurface. Fig-

ure39 shows an exampleof themeasureddepen-denceof theefficiency on theappliedvoltage,forthreedifferentirradiationrates.

To obtaintheefficiency perstationrequiredinthemuonsystem,pairsof RPCswill beoperatedinOR.Thisensuresanoverallefficiency greaterthan99%, with a meanclustersizebelow 2, asshownin Figure40. Thedouble-gapRPCalsohadahighefficiency, but is not considereda valid optionbe-causethemeanclustersizeis significantlygreaterthan 2 [3]. This is attributed to the fact that thereadoutplaneshave a different electrical layout,favouringcrosstalk.

4.2.2.2 Comparison of fr ont-end electronics

The BiCMOS chip is usedby CMS for double-gapRPCs, which differ from theLHCb chambersin terms of the layout of the readoutplane andthe compositionof the gasmixture (no SF6). Acheckthatthechipperformssatisfactorilywith theLHCb setuphasthereforebeencarriedout. Forthis check,a single-gapRPC was equippedwithBiCMOS electronicsto readthe signalscollectedon16stripsof 3 ' 25 cm2. Thesestripsweretermi-

36

0

0.25

0.5

0.75

1

8600 8800 9000 9200 9400 9600 9800 10000 10200 10400

HV (V)

Effi

cien

cy Standard electronics

CMS electronics

1

1.5

2

2.5

0.86 0.88 0.9 0.92 0.94 0.96 0.98 1

Efficiency(

Clu

ster

Siz

e

Standard electronicsCMS electronics

1

1.5

2

2.5

0 1 2 3 4 5 6 7 8 9 10

X position (cm)

Clu

ster

Siz

e

)HV = 9800 VStandard electronics

CMS electronics

Figure 41 Comparisonof testresultsfor two identi-cal chambers,oneequippedwith “standard”electron-ics, andonewith CMS electronics.Resultsaregivenfor (top)efficiency asa functionof theappliedvoltage;(centre)efficiency as a function of clustersize; (bot-tom) clustersizeasa function of position in the strip.Thereadoutstripshada width of 3cm.

natedwith their characteristicimpedanceof about27Ω. The efficiency andclustersizeof this RPCwerethenmeasuredin a testbeam.

A secondsingle-gapRPC, equippedwith the“standard”electronics,hasbeentestedin thesameconditionsand served as a reference. The striplengthfor this referenceRPCwas50cm. For tech-nical reasons,the effective thresholdsof the twoRPCswere different: about120fC for the refer-enceRPC,andabout180fC for theRPCwith BiC-MOS electronics.

Theresultsof thetestareshown in Figure41.The top plot shows theefficiency asa functionofthe appliedvoltage,the plateaustartingat highervoltages for the BiCMOS RPC becauseof thehigher threshold.The centreplot shows the rela-tion betweencluster size and efficiency, not ex-pectedto dependon the thresholdvalue. Finally,the bottomplot shows the clustersizeasa func-tion of thepositionat which a particlecrossesthestrip. Thepeakscorrespondto particleshitting theboundarybetweentwo strips.Therelative shift inthe peakpositionsfor the two detectorstestedisdueto thedifferentalignmentswith respectto thebeam.Thesmalldifferencesbetweenthetwo sets

24.46 / 18

Constant 99.42Mean -114.0

Sigma 1.137

0*

20

40+60,80-

100

-180 -160 -140 -120 -100 -80

Figure 42 Time resolutionof a single-gapRPCwithreadoutstripsof 6cm width andBiCMOS electronics(triggerresolutionunfolded).

of resultscanbeunderstoodasbeingdueto differ-encesin theappliedthresholdsandtiming proper-tiesof thetwo electronicchains.

The conclusionfrom thesetests is that theCMS BiCMOS electronicsperformssatisfactorilyandcanbechosenasthebaselinesolutionfor theRPCreadout.

4.2.2.3 Performancewith largestrips

The measurementsdescribedabove have all beenobtainedusingreadoutstrips3cm wide. Thepos-sibility of reducingthenumberof readoutchannelsby doublingthestrip width hasbeenput forward,and is now the preferredsolution (seeTable16).First testsof the detectorperformancewith read-outstripsof 6cmwidth havebeencarriedoutwithcosmicrays,usingtheBiCMOS electronics.Thereadoutstripsusedwere30cm long andwereter-minatedwith their characteristicimpedance.Thiswasabout13Ω, or half thevaluefor stripsof 3cmwidth. Thetime resolutionmeasuredis betterthan1. 2ns,asshown in Figure42,andthenoiserateislessthan0.5Hz

&cm2.

Figure43shows theefficiency asa functionofappliedvoltagefor thresholdvaluesof about70fCand100fC. Theefficiency in theplateauwaslargerthan97% in both cases.A first measurementin-dicatesan averageclustersizeof about1.2 up tothe highestvoltagesapplied. However, a preciseassessmentof thecapacitive andgeometricalcon-tributionsto theclustersizerequiresfurthermea-surements,usingparticlebeams.

37

HV(V)/

ε

V0

th = 90mV

Vth = 60mV

HV(V)/

ε

V0

th = 90mV

Vth = 60mV

0*

0.2*0.4*0.6*0.8*

11

94002

96002

98002

10000 10200 10400 10600

Figure 43 Efficiency asa functionof appliedvoltagefor asingle-gapRPCwith 6cmreadoutstripsandBiC-MOS electronics,for two thresholdvalues. Measure-mentswereperformedwith cosmicrays.

4.2.2.4 Conclusions

Theprototypetestsbriefly describedin thissectionhave shown that:

1. single-gapRPCscanbeoperatedat ratesof1.8kHz

&cm2without degradationof perfor-

mance;

2. theBiCMOS readoutelectronicsdevelopedby the CMS collaborationcan be usedfortheRPCreadout;

3. readoutstrips with a width of 6cm showgoodperformance.

4.2.3 Ageing

RPCdetectorshave not previously beenoperatedundersuchheavy backgroundconditionsasareex-pectedat theLHC. Ageingtestsperformedby theRPCgroupsof ATLAS andCMS have shown thattheir detectorscan withstandthe radiationdosesexpectedin thoseexperiments[58, 59]. However,thebackgroundlevelsexpectedat LHCb polaran-gles are an order of magnitudemore severe [8].An ageingtestis thereforebeingperformed,whichshouldrealisticallyreproducetheconditionsof theLHCb experiment.

It hasbeendemonstratedthat theresistivity ofbakelite isunaffectedby radiationdosesseveralor-dersof magnitudehigher than thoseexpectedinLHCb [3]. The otherparameterrelevant to char-acteriseageingis thechargetransportedacrosstheRPC.

hours

Cin

t(C/c

m2 )

hours

gif d

uty

cycl

e

0

0.05

0.1

0.15

0.2

0 200 4003

600 800 1000 1200 1400 1600 1800

0

0.25

0.5

0.75

1

0 200 400 600 800 1000 1200 1400 1600 1800

Figure 44 Summaryof the GIF ageingtest duringthefirst threemonthsof 2001,showing (top) total inte-gratedchargedensityfor the irradiatedRPC, and(bot-tom) theGIF activity (timewith source“on”).

Assuminga maximumparticle flux of Φ0 4375Hz

&cm2 in region R3, and taking 30pC as

theaverageavalanchecharge in the RPC(seebe-low), thecurrentdensityis found to beaboutJ 411nA

&cm2, andthetotalchargeaccumulatedin 10

LHCb years(108 s) is aboutQ10y = 1.1C&cm2. In

region R4 (Φ0 5 100Hz&cm2), thecurrentdensity

andthetotalchargeareabout70%lower. Forcom-parison,an accumulatedcharge of 0.3C

&cm2 has

beenreachedin theATLAS test[58].In the LHCb test, startedin January2001, a

singlegapRPCwill beirradiatedattheGIF facilityfor at leastone year. The chamberundertest ispositionedat a distanceof aboutonemetrefromthesource.Thisgivesacurrentdensityaboutthreetimeslargerthanin theexperiment.

Theperformancecharacteristicsof anRPCde-pendon severalparametersotherthanirradiation,including, for example,temperatureandpressure.To isolatethe effect of the irradiation, a second,similar, chamberis operatedoutsidetheirradiationarea,and is usedasa reference.The two cham-berswerethoroughlytestedin theT7 beambeforeinstallationat theGIF. During irradiation,thecur-rent, temperature,pressureand countingratesofthetwo chambersarecontinuouslymonitored.

The averagecharge generatedby backgroundhits hasbeenmeasuredundertheassumptionthatall currentdrawn by theirradiatedchamberis due

38

0.6560.76

0.7560.86

0.8560.96

0.956

1

1000 2000 30007

4000 50008

60009

7000:

8000;

9000<

Max Rate (Hz/cm= 2)

>

Effi

cien

cy

Max Rate (Hz/cm= 2)

>

Effi

cien

cy

Rate (Hz/cm? 2)

>

Effi

cien

cyHV= 9.8 kV

Irradiated chamberReference chamber@

Figure 45 Efficiency asa function of ratefor irradi-atedchamberand referencechamber. The integratedcharge densitycollectedby the irradiatedchamberisabout0.2CA cm2 (seetext).

to the observed hits, taking into accounttheclus-ter size. The valuemeasuredin this way is about30pC.

Figure44 summarisestheGIF testduring thefirst threemonthsof 2001. The upperplot showsthe integratedcharge per cm2. The lower plotshowsthedutycycle. Thishas,unfortunately, beenratherpoor, so thatonly 0.2C

&cm2 have beenac-

cumulatedsofar. A first checkof theperformanceafter irradiation with this integrated charge hasbeenperformedby measuringthechambercharac-teristicsat theCERNT11 beam.Figure45 showsthemeasuredefficiency asa functionof rate2. Nodifferencein performancebetweenthe irradiatedandthereferencedetectorhasbeenobservedup toratesof at least3.5kHz

&cm2. Thetotal chargeac-

cumulatedcorrespondsto abouttwo yearsof oper-ationfor thedetectorsin region R3 andfour yearsin region R4. At thepresentrate,it shouldbepos-sibleto collectabout0.7–0.8C

&cm2by theendof

2001, correspondingto abouteight LHCb yearswith theconditionsof regionR3.

2The differencebetweenthe radiation load of photons,producedat the GIF, andpions,usedin the testbeammea-surements,inducesadifferentrateperformance[3, 7].

39

5 TechnicalDesign

5.1 The MWPC detector

TheLHCb muonsystemwill useMulti-Wire Pro-portionalChambers(MWPC) for the four regionsof stationsM2 andM3, for regionsR3 andR4 ofstationM1 andfor regionsR1 andR2 of stationsM4 andM5. Thefull systemcoveredby MWPCsconsistsof 864 chambersand about80,000FE-channels.An overview of the MWPC detectorisgivenin Table13.

5.1.1 Detectoroverview and requirements

Therequiredgeometricaltolerancesfor thecham-ber constructionhave beendeterminedbasedonGARFIELD [60] simulations, practical consid-erationsand prototyperesults. For the simula-tions[50] amaximaltolerablevariationin gasgainof B 20% hasbeenassumed,andonly individualparametershave beenvaried while the other pa-rametershave thedesignvalue. Themaximalde-viationsof singleparametersfrom thedesignval-uesaregivenin Table14 togetherwith theaccept-ablegeometricaltolerances.The geometricaltol-erancescombineall singleeffectsandarethereforemorestringentthanthemaximaldeviationsfor sin-gle parameters.

Table 14 Geometricaltolerancesfor theMWPCs.

Parameter Maximal AcceptableDeviation Tolerance

PanelThickness B 200µmPanelflatness B 75µm B 50 µmGasgapsize B 80µm B 70 µmWire planeoffset B 300µm B 100µmSinglewire offset B 250µm B 100µmWire pitch B 80µm B 40 µm

5.1.2 Chamber componentsand design

Thegeneraldesignandconstructionis thesameforall chambersandis discussedin detailin Ref.[61].They have four separategasgaps,eachwith anan-odewire planeand, in regionsR1 to R3, a planeof cathodepads.In Figure46 onecanseea cross

Figure46 A crosssectionof thewire chambersshow-ing thefour gapsandtheconnectionto thereadoutelec-tronics.

sectionof one chamberwherethe four gasgapsareshown, togetherwith theconnectionto theFEelectronics.

Themaincomponentsfor theMWPCsarethefollowing:

C Structuralpanelswith FR4 laminates(totalthickness10.2 B 0.2mm);

C Wire fixation bars with 2.40 B 0.08mmthickness;

C Gold-platedtungstenwiresof 30µm diame-ter;

C Sidebarsof 4.9 B 0.08mmontheshortsideof thechamber, wherethegasinletsarelo-cated;

C Gap barsof 2.40 B 0.08mm thicknessontop of thewire fixation barsto closethegasgapover thelong sideof thechamber;

C Spacersalongthechamberborderto ensureprecisionof thegasgapof 5.00 B 0.05mm.

5.1.2.1 Panels

The panelsarethe basisof the chambermechan-ical structure.The requirementon the flatnessof

40

Table 13 Summarytableof theMWPC detector

StationM1 StationM2 StationM3 StationM4 StationM5 Sum

ChambersRegionR1Numberof Chambers 12 12 12 12 48Sensitive area( cm2) 30' 25 32.4' 27 34.8' 29 37.1' 30.9Anodechannels 96 96 2304Wire padsize( mm2) 6.3' 250 6.7' 270Numberof wires 800 864 928 992 4.4' 104

Cathodechannels 128 128 192 192 7680Cathodepadsize( mm2) 37.5' 31.3 40.5' 33.7 29' 36 31' 39ChambersRegionR2Numberof Chambers 24 24 24 24 96sensitive area( cm2) 60' 25 64.8' 27 69.5' 29 74.3' 30.9Anodechannels 96 96 4608Wire padsize( mm2) 12.5' 250 13.5' 270Numberof wires 1600 1728 1856 1984 1.7' 105

Cathodechannels 128 64 96 96 9216Cathodepadsize( mm2) 75.0' 31.3 162' 33.7 58' 72 62' 77ChambersRegionR3Numberof Chambers 48 48 48 144Sensitive area( cm2) 96' 20 120' 25 129.6' 27Numberof wires 2560 3200 3456 4.4' 105

Cathodechannels 192 192 192 27648Cathodepadsize( mm2) 20' 100 25' 125 27' 135ChambersRegionR4Numberof Chambers 192 192 192 576sensitive area( cm2) 96' 20 120' 25 129.6' 27Anodechannels 48 48 48 27648Wire padsize( mm2) 40' 200 50' 250 54' 270Numberof wires 2560 3200 3456 1.8' 106

B 50µm is of critical importancefor gasgainuni-formity andconsequentlyfor thewidth of theop-erationalplateau.

A panel consists of two copper clad FR4(fire-resistantfibreglass epoxy) laminates,inter-leaved with a core. For the core various mate-rials are under investigation. Besidesthe panelsbasedon Nomex honeycomb,othermaterialslikepolyurethanicfoam and Chempir [62] are underconsideration.Thechoicefor thecorematerialisstill to bemade.

Honeycomb panels: It has been demon-stratedin several prototypesthat Nomex honey-comb panelswith the requiredspecificationcan

beproduced.Therefore,FR4-laminatesof 1.6mm(0.8mm) thicknesswith D 30µm copper, inter-leaved with 7mm (8.6mm) honeycomb are thebaselinepanelfor thechamberconstruction.How-ever, the productionof suchpanelsis rathertimeconsumingand expensive. Thereforeother solu-tions like theonebasedon polyurethanicfoamorchempircoreareunderinvestigation.

Polyurethanic foam panels: Thepanelsarecomposedof two sheetsof FR4filled with a rigidpolyurethanefoam,which is theresultof a chem-ical reactionbetweentwo components:thepolyoland the isocyanate. A panel of 20E 20E 1 cm3

with arigid polyurethanefoam(ESADUR120)has

41

SPACER HOLE CARD I/O GAS

GUIDE HOLE WIRE BAR GUIDE HOLE LATERAL FRAME

Figure47 Cathodestructureandchambercomponentsfor anMWPC designedfor regionR3.

beenproducedusinganonprecisemould,showinga very high mechanicalrigidity. Small and largeprecisionmouldsof 30E 30 cm2 and40E 150cm2

are under preparationto test the requestedpla-narityandto verify theconstructionsequence.Themouldshave to sustaina pressureof 5kg/cm2 dueto theexpansionof thefoam.

Chempir panels: Similar to the other op-tions, the panelsare composedof two sheetsofFR4, in this caseinterleaved with a polyisocianu-rate core (ChempirCore 75 [62]), which can berectifiedwith veryhighprecision( B 10µm) at lowcost. First testshave shown that suchpanelspro-videtherequiredrigidity. Thelongtermbehaviouris underinvestigation.

5.1.2.2 Cathodeplanes

DetailedPSPICE[63] simulationshave beencar-ried out in orderto minimisethe capacitanceandthecrosstalk inducedby thereadouttracesrunningunderthecathodepads[16] (seealsoSection4.1).Thestudiesshowed that, in orderto minimisethecathodecapacitance,it is preferableto have twopanelswith cathodepadson bothsides,insteadofhaving four panelswith cathodepadson only oneside(seeFigure48). Moreover, sucha configura-tion provides bettershieldingto the cathodes,asthey arealwayssurroundedby detectorground.

In all regionsthecrosstalk canbekeptbelowthe5% level, which is within therequirementsforthemuonsystem.

In region R3 thecathodepadscanbeaccessedfrom thetopandbottomof thechambers(seeFig-

HV

10k

100k

100k

100kWire-pad OR

100k

LVDS

LVDS

Figure 48 Cross section of the chamber with aschematicof thereadoutchain.

ure 47). This avoids the use of a double layerboardin PCBtechnologyfor thecathodestructurein this regions, which are difficult to produceintherequireddimensionsof 140E 35 cm2. First in-vestigationsof usinga milling machineto realisethecathodestructurearepromising,but a full testhasstill to becarriedout. Guardtracesof 0.5mmwidth betweenthe cathodepadsare foreseentominimize thecrosstalk. Thewidth of the insulat-ing surfacebetweenthe padsandthe guardtraceshould not exceed0.4mm to avoid problemsofchargeupat high rates.

In regions R1 and R2 the cathodeshave achessboardstructure,as indicated in Figure 49.Only a fraction of the cathodepadsof region R1and R2 can be accessedfrom the border of thechamber. Most of the padshave to be read bytracesrunningon thebottomof thecathodeboardto the edgeof the chamber. A doublesidedPCBwill be usedto implement this structure. Spe-cial carehasto be taken to minimize the capaci-

42

1A

2A

3A

4A

5A

6A

8A

7A

1B

6B

7B

8B

5B

Rin

Rin

4B

C

3B

2B

1C

2C

w

h

Figure 49 Cathode-andwire-padstructureandread-out.

tancebetweenthereadouttracesandthepads.Thereadouttracesof 0.25mm width areseparatedby0.25mmgroundtraceswith agapof 0.25mm. Thepadsareconnectedthroughmetallizedholesto thereadouttraces.

5.1.2.3 Wire fixation bars and gapbars

On the long sidesof the panelswire fixation barsareglued. The barshave a thicknessof 2.4mm,0.1mmlessthantheanode-cathodedistance.Theywill be madeaccordingto standardprinted cir-cuit boardtechnology. A patternof fingerpadsisetchedon the barswhich will be usedfor solder-ing thewires. They areinterconnectedin groups.Thegroupingof wiresis determinedin thecaseofanodewire readoutby the requiredgranularityinthe x-coordinate. In order to minimize the crosstalk in the caseof cathodepadreadout,wires aregroupedtogetheraccordingto thex-dimensionofthecathodepads.

The wire fixation bars and the other frameswill be correctly positionedon the panelsusinga set of guide postswhich are insertedinto the

frames. Cylindrical spacers,(5.00 B 0.05)mmthick, guaranteetheexactgapalongtheperimeterof thechamber. In this casethewire fixation barsandtheotherframescanhave standardtolerances.Thesidebarshave additionalholesfor thegasin-lets/outlets. This solution is both lessexpensiveanddoesnot requireaccurateglueingto maintaintherequiredtolerences.

5.1.2.4 Wire

The total number of wires in the chambersisabout2.5 E 106, with a total wire lengthof about1200km. Therefore,a great deal of effort hasbeenexpendedto developanefficient andreliableschemeof winding and attachingwires, as dis-cussedin section5.1.3.

Gold-platedtungstenwire of 30 µm diameterhasbeenchosenfor thechambers.Measurementshave shown a lineardependenceof theelongationon the appliedweight up to 140g. At the wirespacingof 1.5mm,afreewire lengthof 30cmandwith nominal HV of 3.15kV, the wires becomeelectrostaticallyunstableif their tensionis below30g. The chosenwire tension is (60 B 10)g,controlledwith a standardwire tensionerduringwiring.

5.1.3 Chamber construction and tooling

Two possiblewaysto build thechamberswith theabove parametershave beenconsidered. One isproducingpanelswith wiresonbothsides(doublegaps),theotherwith wiresononesideonly (mono-gaps). The two optionscanbe seenin the draw-ing of Figure50. Both methodshave their merits.Themainadvantageof monogapsis that thepan-elscanbehandledmoreeasilyduringthedetectorconstruction.Moreover, in caseof incurableprob-lemsduring theglueingor solderingprocessonlyonegapwould be lost. Doublegapwiring, on theotherhand,is betteradaptedto thecathodedesign,which is basedon two doublegapsinsteadof foursinglegaps,aspointedout in section5.1.2.2. Allprototypeshavebeenbuild sofarwith adoublegapstructure.Hence,this designis alsobackedup byexperienceandpositive resultsfrom tests.Thefi-nal choiceof the constructionmethodhasstill tobemade.

43

Wire fixation bar

Wire planes on onesF ide of the panels

Wire planes on bothsF ides of the panels

Gap bar

Gap barSide Bar

Side Bar

Figure50 Explodedview of chambersreadyto beassembled.Theleft sideshowspanelswith singlesidedwiresandtheright sidpanelswith wireson bothsides.Thepositionof thevariousbarsis indicatedaswell.

Figure51 Schematicdrawing of theframecrosssec-tion.

Prior to thewiring a panelis assembledin thefollowing way:

C thesidebarsareinsertedin theguidepostsandgluedto thepanels;

C the wire fixation bars are also positionedusing the guide posts(theseare especiallyneededif thebarsaremadeof severalpiecesdueto thechamberlength)andgluedto thepanels;

C thegapbarson top of thewire fixation barsaregluedto thepanels.

Thefinal assemblycanbeseenin Figure50.

5.1.3.1 Wiring

Double sidedwiring is doneby winding directlyaroundthe honeycomb panels. In this way sym-metrically loadedpanelswith wire planeson bothsidesareproduced. The panelis fixed to a rigidframe wherethe positioningcombsare mounted

(seeFigures51 and52). To achieve the requiredprecision,the wire spacingis determinedby thecombswhile theanodeto cathodedistanceis givenby theadjustmentbars.

C Grooved combs: The grooved combshavea diameterof 15mm andaremachinedin aprecisewaytohaveapitchof 1.5mm,whichdeterminesthe wire spacing. The groovedepthis of about0.25mm, hencethe innerdiameterof the combsis smaller than thedistancebetweenthetwo wire planes.

C Adjustmentbars: The wire heightwith re-spectto thecathode-planeisadjustedby pre-cision barsmountedto the frame. On oneside the bars are fixed, and on the other,they canbeadjusteddependingon thepanelthicknessto achieve a wire to cathodeplanedistanceof 2.5mm.

Thewiring procedurewastestedfor a700mmlong detectorpanel, and the averagepitch mea-suredwas 1.5 mm with a root meansquareof14µm. This precisionis well within the specifi-cationsof B 40µm. For biggerchambers,thepanelshouldbefixedto theframealongits longsideev-ery500mmtoavoid differencesin sagbetweenthepanelandtheframe.

Theproductionof panelswith wireson a sin-glesidecouldbedoneusingthesamewindingma-chine,but with adifferentframe.Basedonacalcu-lation, no deflectionof thepanelsis expecteddueto theasymmetricloadof 60g perwire. Testsdone

44

Figure52 Wiring of apanelmountedto thealuminiumframe.

Figure53 Thelasersolderingandthesolderdispenser.

45

on bothdeflectionandtorsionof a panelwired inthiswayconfirmedthis. Thewiring guidelinesarethesameasmentionedfor doubledsidedwiring.

5.1.3.2 Glueing

Oncea frameis wired, it canberemovedfrom thewinding machineto have thewiresgluedandsol-dered.This separatesthe threeimportantstepsofthechamberconstructionandallowsaparallelpro-duction.

The wires areglued to the wire fixation barsbeforesoldering. This procedureguaranteesthatthewiresarekeptin placewith afixedheightwithrespectto thecathodeplane.Thegluingalsokeepsthe wire tensionto its nominal value. Standardepoxygluelike Adekit A145 [61] polymerizinginabout24hoursat roomtemperatureis foreseenforthewire glueing.

5.1.3.3 Soldering

Oneof the cleanestsolderingmethodsis the useof a laserbeam. Due to the large numberof sol-dering points in the constructionof the MWPCs( D 5 E 106), theuseof an automatedandreliablemethodis desirable.A testwasmadewith anau-tomaticsolderingstation[61] (seefigure53).

The result of the test wasvery cleanand thecontrolof local heatwasvery good.Thewire suf-fers lessheatstresswith respectto conventionaltechniques.

Usingaconservativevalueof 3sfor thesolder-ing of onewire andassumingthat the final setupwill be equippedwith two laserheads,the timeevaluatedfor the wire solderingof all MWPCswould be 2100 hours. This numberdoesnot in-clude the time spentfor the layer settingup andfor the requiredchecks,which is proportionaltothenumberof chambers.

5.1.3.4 Final assembly

To proceedwith the final assemblyof the cham-ber, five panelsshouldbeready:two doublesidedwired panelsandthreegroundpanels(or four sin-gle sidedwired panelsandonegroundpanel).Allthe panelsare alreadyequippedwith side bars,wire fixation barsand gap bars. The cylindrical

precisionspacersareinsertedin the holesaroundthechamberandthepanelsareassembledmakinguseof the guide postsat the four corners,as in-dicatedin Figure54. For the final closingof thechambersthe five panelsare kept togetherwithscrews. In the side bars, the gas inlets/outletsfor eachgap are mounted. The gas tightnessisobtainedby gluing the five panelstogetherwithepoxyglue.

5.1.4 HV and FE interfaces

Several constraintsdeterminethe available spacefor andthelocationof thetheHV-interfaceandthereadoutelectronics:

C Densityof channelsin innerregions;

C Spaceproblemsdueto theproximity to thebeampipein regionR1;

C Connectivity requirementsdueto logic ORson FE-electronicscards.

As a consequence,the border region of thechambershave the spacerequirementssumma-rized in Table 15. Theseparametersensuresuf-ficient spacebetweenthechambersfor theroutingof cablesfor readout,highandlow voltage,andgaspipes. A detailedstudyof thesecombinationsforthevariousstationsandregionsleadto theschemeshown in Figure46.

Table15 Spacerequirementsaroundthesenstiveareato theborderof thechamber.

Spacerequirements

Topwith anode/cathodereadout 85 mmBottomwith cathodereadout 70 mmBottomwith no readout 35 mmSidewith no readout 50 mmSidewith cathodereadout 60 mm

5.1.4.1 HV interface

The HV connectionis realizedby interfacecardswhich carry the loading resistorsand the decou-pling capacitors.TheHV boardswill alsocarryalarge amountof groundconnectionson the other

46

WG

IRE BAR

I/O GAS

FRONTAL SPACER HOLE

GH

UIDE POST

GH

UIDE POST

LATERAL SPACER

LATERAL FRAME

CI

ARD

FRONTAL SPACER

Figure54 A sketchof theMWPC assembly.

side to reducepickup noise. Samplesof thesecardswill betestedin realconfigurationswith ex-isting wire chamberprototypesto checkthevalid-ity of this solution. The valueof the decouplingcapacitorshouldbe much larger than the capaci-tanceof the groupof wires connectedto it. Thisensuresa low impedanceto groundor to the am-plifier. A valueof 0.5-1nF satisfiesthis conditionin all cases.

The upperlimit on the HV loadingresistorisgivenby themaximalallowedvoltagedrop,whilethe lower limit is set by the introducedparallelnoise.Thechoicefor theresistorsis therefore100kΩ.

5.1.4.2 FE interface

The FE electronicswill be implementedin twostages;the first stageasa sparkprotectionboard(SPB) and the secondas the Amplifier-Shaper-Discriminator(ASD) chipboard(ACB). TheACBis mountedasshown in Figure46ontheSPB.Thisdesignkeepsthe distancethe signalsmust pro-pogatefrom thewire padsandcathodepadssmall.The dimensionsfor theseboardsaregiven by thethicknessof the chamber(70mm) and the maxi-malallowedspacebetweenadjacentboardson thechamber(50mm). The50mm aredeterminedbyregionR1,wherethehighestgranularityof readoutchannelsoccurs. Eachboardreceives the signals

of 8 readoutchannelsfrom eachdoublegap,thusa total of 16 channels.

The SPB will be a 50 E 70 mm2 two layerboardthatcontainsasystemof resistorsanddiodesfor eachchanneldesignedto limit the voltageintheeventof a sparkor discharge. Thedesignusesatwostagedoublediodescheme.Thisdesignfullyprotectsthereadoutchannels.

The ACB is a 50 E 70 mm2 four layer PCBcontainingtwo ASD chips,for which thepreferedsolutionis theCARIOCA chip (seeSection 4.1),andthechip,whichprovidessomebasiclogic anddiagnosticfunctions(seeSection5.3.2).

5.1.5 Quality control and testing

Quality testsof the individual chambercompo-nentsandfor theassembledchamberareforeseen.Thekey itemsto becheckedduringchambercon-structionarethefollowing:

C Panelplanarity, which canbe verifiedwitha properapparatus;

C Wire quality, with optical inspection andtestsof mechanicalpropertieson samples;

C Wire trimming, to checkthat thewire is cutsufficiently closeto thesolder;

C Soldering quality, to check the electricalcontinuity after the semi-automaticsolder-ing andtrimming of thewires.

47

C Wiretensionchecks, usinganautomatedres-onancemethod;

C Wire positioning, to be verified on a smallsetof pointsoneachwire plane;

C Gastightness, to be verifiedusingstandardproceduressuchasapplyinga small under-pressureto the chamberin an He-bagandlooking for possibleleaks;

C HV training and tests. This testwill be themost important as it will certify the qual-ity of the productionfrom eachcentre. A“good” chambershouldbeableto sustainavoltagewell into theoperationalplateaufora minimum numberof hours,after trainingwith an automatedprocedurehasbeenper-formed.

Afterwardsthe chamberwill be inspectedforuniformity in its response.A veryefficientmethodis to perform a scanof the wire planeswith asource,checkingthatthemeanvalueandthestan-dard deviation of the measuredcurrent is withinspecificationsover the chamber. In addition, acompletetest with cosmic rays to determinetheefficiency plateauandtime resolutionwill beper-formed.As faraspossible,thesetestswill bedonewith thefinal electronics.

Important information about individual com-ponentsandthe final chamberwill be storedin adatabase.Thisallowsto retrieveatany timethere-sultsof all quality control measurementsandwillaid in understandingpossibleproblems.

5.2 The RPC detector

The prototypetestsdemonstratethat the require-ments of efficiency, redundancy, rate capabilityand clustersize are optimally met by a solutionbasedon chambersmadeof two identicalsingle-gap RPCs,eachwith its own strip readoutplane(Figure55). This allows two independentdetectorlayersperstation.Thetwo stripplanesarereadoutindependentlyandthesignalsfrom correspondingstrips in the two layersarelogically OR-ed. Thetwo gapsof theRPCareconnectedto independentHV supplies.Thisallowstheadjustmentof theHVfor slightly differentplateauvoltages.In addition

RPC 1J

RPC 2J

HV 1K

HV 2

Figure55 Schematiclayoutof thechamberwith twoindependentsinglegapRPCs.

oneof thegapscouldbeswitchedoff, if necessary,at thepriceof a smallreductionof theoverall effi-ciency.

5.2.1 Detectoroverview and requirements

Themainparametersof theRPCsystemaresum-marisedin Table 16. The minimal number ofchambertypesper stationis obtainedwith cham-bersizeshaving anactiveareaof 139 E 29cm2 and148 E 31 cm2 in stationM4 andM5, respectively.In region R4 the stripshave full length(about30cm), with the readout at one end. In region R3they aresplit in themiddleandarereadoutatbothends. In both casesthe electronicsboardsarein-stalledon thechamberhorizontalsides.Thestripwidth in stationM4 (M5) is 5L 6 M 6L 0N cm, with apitch of 5L 8 M 6L 2N cm and interleaved with narrowgroundstripsof 0L 5mm.

Maximumstandardisationhasbeenan impor-tantgoalin thedetectordesign:

C thedimensionsof theRPCgapsarethesameeverywhere;

C chambersof region R3 andR4 in the samestationhave thesamesensitive areabut dif-fer in theverticalstrip dimension,which inregion R3 is half thatin region R4;

C the sensitive area(different in station M4andM5) is definedby thestrip readoutandby thegraphitelayer(seelater);

Uniform operationwithin each chamberre-quirescontrolling the planarityof the gasgapsatthelevel of B 10µm.

48

Table 16 Main parametersof the LHCb muonsystemin the regionsequippedwith RPCs.Whererelevant thehorizontaldimension,i.e. theonein thebendingplane,is givenfirst.

Physicalquantities: RegionR3 RegionR4StationM4 StationM5 StationM4 StationM5

Detectorsurface(m2) 19.3 22.1 77.4 88.3Horizontaldimension(cm) 139 149 278 297Verticaldimension(cm) 116 124 232 248Chambersensitive area(cm2) 139E 29 148.5E 30.9 139E 29 148.5E 30.9Channelsize(cm2) 5.8E 14.5 6.2E 15.5 5.8E 29.0 6.2E 30.9Maximal rate(Hz/cm2) 750 650 250 190Numberof chambers 48 48 192 192Total 480Physicalchannelsperchamber 48E 2 48E 2 24E 2 24E 2Total 27648

1

2O 3

P

1 56

4Q

7

Figure56 Chambercrosssection:Al-polystyreneex-ternalpanels(1); HV contacts(2); RPCdetectors(3);multi-pin connectors(4); polycarbonatespacers(5);strip planes(6); aluminiumexternalprofiles(7).

5.2.2 Chamber design

A detailedcrosssectionof the basicchamberisshown in Figure 56. The chamberis madebytwo independentRPCdetectors,eachwith its ownreadout strip plane. All componentsare sup-ported and kept togetherby meansof two sup-portpanelswhichalsoprovidethenecessaryrigid-ity to the chamber. The panelsare madefrompolystyrenefoam (density40 kg/m3) sandwichedbetweentwo aluminiumsheets(0L 5mm thick) andgluedwith epoxyadhesive. Theoverall thicknessof eachpanelwill be about10mm. Thesepanels

arerigidly connectedon the sidesby shapedalu-minium profiles,0L 5mm thick. Sinceit is impor-tant that the panelsprovide an adequateanduni-form pressureon the detectorassembly, the pos-sibility to have them pre-loadedis underconsid-eration. This could be obtained,for example,byglueing them on a templatein sucha way as tohave a non-zerosagitta. Wherepossiblethe useof standardcommercialmechanicalitemswill bepursued.

The gas gap of eachRPC detectoris madeof two bakelite plates,2mm thick, andlaminatedwith a thin melaminefoil on thesurfacein contactwith thegas.Thepurposeof themelaminelayeristo improvethesurfacequalityof thebakeliteplatesoverthatof purelyphenolicplates[11]. Thechoiceof thebakelite volumeresistivity is drivenby con-siderationsaboutratecapability. As mentionedintheprevioussection,satisfactoryresultshavebeenobtainedwith volumeresistivitiesslightly lessthan1010Ωcm. Hence,bakelite plateswith volumere-sistivities in therangeof M 8 B 2NRE 109 Ωcmwill beused.Ontheoutersurfaceof thegaps,thebakeliteis paintedwith a conductive graphitelayer (resis-tivity about100 kΩ/square). The graphitelayerdistributestheHV andgroundon thebakelite,andis electricallyinsulatedby meansof a200µm thickPolyethylene-Teraphtalate(PET) film glued ontothegraphiteitself by a “hot melt” process.

To ensureapreciseandconstantgapwidth, the

49

bakelite platesarekept parallelby polycarbonatespacers,10mm in diameter, and 2L 00B 0L 01mmin height. Theseare positionedwithin eachgapto form a squarearray with a 10E 10cm2 basicrepeatingcell. The topology for spacerposition-ing is different in the two RPCsof a chambertoavoid correlatingthe insensitive areas.A polycar-bonaterectangularframe, 7mm wide and of thesameheightandtolerancesasthespacers,is usedto closethegapat theedges.Thecirculationof thegasis insuredby thefour gasinlets/outletsplacedat the four cornersof the detector. They arealsomadeof polycarbonateandare3mm in diameter.All theseparts(bakelite plates,spacers,frame,gasconnections)aregluedtogetherusingepoxyadhe-sives.

The two strip readout planes are locatedon the inner part of the chamber(Figure 56).Theseplanesare madeby a dielectric substrate(polystyrenefoam 3L 8mm thick, relative permit-tivity εr 4 1L 4), sandwichedand glued betweentwo copperfoils. Oneof the foils constitutestheground,theotheris milled onaspecialmachinetoform the strips. The 6cm wide stripswill have acharacteristicimpedanceof about13Ω. The sig-nalswill bebroughtoutsideto the front-endelec-tronicsvia standardmulti-pin connectors.

An alternative schemeis underinvestigation,in which thestrip readoutplaneis etchedon a PCboardmadeof a0L 8mmthick FR4layer. ThePCBwould thenbegluedon thepolystyrenefoamsub-strate.Thebasicstrip modulewill beasshown inFigure57. This solutionwould additionallysim-plify theconstructionby avoiding interconnectingcablesfrom the strips to the the externalconnec-tor. Its feasibility, in termsof costandelectricalproperties,hasto bechecked.

An exploded view of the RPC chamberisshown in Figure58.

5.2.2.1 Oil tr eatment

It is well known thatdepositinga thin layerof lin-seedoil on thebakelitesurfacehastheeffectof re-ducingthenoiseandthedarkcurrentof theRPC.This occurspartly becauselinseedoil hasa largerconductivity thanthebakelite.

The treatmentof the bakelite plateswith lin-

Region 4 Region 3

GND plane

GS

ND plane GS

ND plane

Strip Strip

290

/ 309 14

5 / 1

55

Figure57 Printedcircuit for strip planes:basicmod-ulesfor R3 andR4; dimensionsin mm (the first indi-catedis for M4, thesecondfor M5).

seedoil was recently much debatedin the RPCcommunity. This was triggeredby the problemsfacedby BABAR [64], which were attributed tohigh-temperatureoperationcoupledwith improperoil polymerisation.In a “sticky” layer theoil canform “stalagmites”becauseof electrostaticattrac-tion, that caneventuallyshort-circuitthe gapandinhibit properoperationof theRPC.

Sincetheoil is appliedafter thegapis assem-bledandsealed,qualitycontrolof thelayeris verydifficult. Thereforeit wouldbepreferableto avoidthe oil treatment,which would also simplify theconstructionof the detectors.However, this willunavoidablyincreasebothdarkcurrentsandnoise.

The increaseddarkcurrentwill resultin extraageing.In orderto controlthiseffect,thedarkcur-rent shouldbe kept below the currentdue to theflux of real particlestraversing the detector(seeSection4.2.3). Consideringan increasedageingof 25%for chambersin Region R3 asacceptable,RPCswithout oil treatmentwill be a viable solu-tion if the darkcurrentdensitycanbe kept below3 nA T cm2 (30µA T m2). Assumingthattheaveragecharge of thenoisehits is 30pC, this correspondsto a maximumnoiserateof 100HzT cm2 perRPCgap,whichis well below therateacceptableby thetrigger(seeFigure16).

Preliminary studiesperformedby the CMScollaborationwith non-oiledchambers[65] haveshown that with the melaminetreatmentof the

50

Al profilesGU

ap 1

SV

trip PlanesGU

ap 2

Figure58 Explodedview of theRPCchamber.

bakelite platesit is possibleto reachfull efficiencykeepingthenoiselevel below 50HzT cm2.

A test of RPCswithout linseedoil treatmentis scheduledfor the summerof 2001, in order todefinitely assessif thesedetectorscanbe usedintheexperiment.

5.2.3 Chamber construction

The industrialcapabilitiesfor producingRPCde-tectors is by now well established. The totalamount of melaminelaminatedbakelite neededfor the LHCb RPCsis about1000m2andcanbeproducedindustrially in a few weeks. Sincethecommercialplatescomein ratherlargedimensions(about3L 2 E 1L 3 m2) they will have to becut to therequiredsize. The possibility of having theshap-ing performedby the firm producingthe bakeliteis underinvestigation.

Oncethe bakelite platesare ready, the wholenumberof gapsrequiredbyLHCbcanbeproducedin a few months. The gapswill be producedfol-lowing the LHCb specifications,using the sametechniquesdetailedin Refs. [10, 11]. The spec-ifications refer to the sizeof the gapsandof the

graphitepaint defining the active area,the posi-tioningof thespacersandthelinseedoil treatment(if applied).

5.2.4 Quality control and testing

The individual chambercomponents,as well astheassembledchambers,will undergo somebasicquality controlandtests.

Quality control of the bakelite platesis per-formedat thefactoryby measuringthevolumere-sistivity andthesurfaceroughnessatseveralplaceson theplates3. Platesnot satisfyingthespecifica-tions will be rejected,the otherswill be groupedaccordingto the measuredresistivity values. Asecondselectionstepwill be performedin RomeII, looking in particularfor possiblesurfacedam-age(scratches)thatcouldhaveoccurredin thestor-ageandshippingphase.Herethematchingof thetwo platesfor agivengapwill alsobemade,whichcould possiblyrequirea new measurementof the

3TheCMSgroupdevelopedaspecialtool for amulti-pointresistivity measurements[66]; we areinvestigatingthepossi-bility to usethesametool or asimilaronebetterspecifiedforourcase.

51

resistivity.Thegapproductionwill becloselymonitored

atthefactoryby techniciansandphysicistsbelong-ing to theresponsibleinstitutes.In thecasetheoiltreatmentwill benecessary, a procedurehasto bedefinedto control its quality. A possibility wouldbe to performa routinecheckby openingup andinspectinga fraction of the gasgaps(e.g. one inten)duringtheproductionphase.

Theproducedgapshave to becheckedfor gastightnessand their capability to standhigh volt-ages.This will bedoneby theproducersusingAror N2.

Pairs of gapssatisfyingthe basicquality con-trols mentionedabove will be used to producechambers.This assemblyphasewill take placeinRomeII andFirenze.Thequalityof thegapsdeliv-eredby theproducingfirm will becheckedat firstby performinganautomatedI-V measurement,inparallel for aboutten gaps. For this purpose,thegapswill beequippedwith connectorsto distributethehighvoltageandmonitorthedarkcurrent.Thetengapswill beflushedwith gasat a rateof about3 l/h for about15 hours;sincethe gapvolumeisabout1 litre, this will insurethatat leastfour vol-umesarechangedduring the flushing procedure.Gapswith dark currentsthat aretoo large will berejected. The tolerancesto acceptor reject gapswill be definedwith the next prototypes. Thesequality checkswill start assoonas the first gapswill beproducedin orderto assurea promptfeed-backto thefactoryin caseof problems.

Chambers with two good gaps will beequippedwith strip planesandcompletelyassem-bled. At this point the chamberswill be readyfor the final testswith cosmicrays, using the fi-nal electronics.The testshouldprovide the basicperformanceparametersof thechambers,suchasthe plateaucurve, efficiency andclustersize,andthe degreeof uniformity over the chamber. Forthispurposeatrackingtriggertelescopewill bede-signed,mostprobablyalsomadeof RPCs.

Thesuccessfullytestedchamberswill thenbeshippedto CERN wherethey will be storedandflushedwith nitrogenuntil installationin the ex-periment.

Eachchamber/gapwill be identifiedby a barcode which will allow to retrieve all the impor-

tant factsaboutthe chamber/gap,the history andresultsof all the measurementsdoneby the com-paniesandin theinstitutelaboratories.

5.2.5 Front-end electronics

As discussedin Section4.2.2.2, the CMS BiC-MOS electronicssatisfiesour basicrequirementsandhasbeenchosenasbaselinefor theRPCread-out. Dependingon theavailability, we planto testalsotheASDQ[47] andCARIOCA [52] chipsde-velopedfor the MWPC, which potentially havesomenice featurethat could allow operationofRPCsat lower gain.

The Front-endboards(FEB) will be mountedon the upperside of the chamber(in region R4)or on the upperand lower sides(in region R3),and will readout 16 channels,eight channelsinthe first layer of the RPC and eight in the sec-ondlayer. Threeboardswill benecessaryto com-pletely readouta chamberin region R4, and sixin region R3. A first prototypewith eight chan-nelsonly hasbeendevelopedfor thetestchambers(Figure36). The secondprototype,presentlyun-derdevelopment,shouldbecloseto thefinal form,housingtwo chipsandthe connectorsto take thesignal out of the chambers. The operationalre-quirementsof theCMSreadoutchipareaLV sup-ply of +5V and a variablevoltage to adjust thethreshold.Thepower andthresholdsettingshouldbeprovidedvia ECS.

The structureof the FEBs has not yet beenfrozen. The Front-Endarchitecture[67] requiresthatpartof thelogic gatesaremovedon thedetec-tor. This is discussedin detail in section5.3.2.

5.3 Readoutelectronics

5.3.1 Overview

The main taskof the electronicsis to preparetheinformationneededby the L0 muontrigger [36].This correspondsto:

C Formation of around26,000 logical chan-nelsignals,startingfrom thearound120,000physicalchannels,correspondingto theout-putsfrom theASD chips.

52

C Eachlogical-channelsignal is taggedwiththe numberof the bunch crossing(BX) towhich it belongs(BX identifier).

As far aspossible,thefirst stepshouldbeper-formedon the chambersin orderto minimisethenumberof LVDS links exiting thedetector.

The secondsteprequiresa time alignmentofchannels.This is necessarybecausesignalsfromdifferent channelstake different pathsbeforebe-ing sentto theL0 pipelinesandtaggedwith theirBX identifiers. Moreover, eachfront-end chan-nel hasa time behaviour thatdependson theread-out schemeandthechamberoperatingconditions.Signalshave a width, including tails, comparableto the BX cycle (seeSection4.1). Correcttimealignmentis thereforenecessaryto avoid ineffi-ciencies.

On-chamberformation of logical channelsfrom physicalchannelsis achieved using a cus-tom integratedcircuit for DIagnostics,timeAlign-mentandLOGics(DIALOG). TheDIALOG chipalsoallows programmingof a delayfor eachsin-gle input channel,andcontainsfeaturesusefulforsystemset-up,monitoringanddebugging(seesec-tion 5.3.2). The detectorlayout and the logicalchanneldistribution allow generationof logicalchannelsin the front-endboards(FEB) in a largepart of the system. However, in regions R3 andR4 of stationsM2 to M5 and region R2 of sta-tionsM4 andM5, physicalchannelsfrom differentFE-boardsandchambersmustbecombined,andasimple local solution is not feasible[67]. In thiscase,logical channelformationrequiresa furtherstepof logical operations,performedin the inter-mediateboards(IB), discussedin section5.3.3.

Oncegenerated,the logical channelsaresentto theOff-DetectorElectronics(ODE),wheretheyareassignedthe correspondingBX identifier anddispatchedto the L0 trigger. The ODE containsalsotheL0 pipelines,theL1 buffersandtheDAQinterface (Section5.3.6). Both the intermediateboardsandtheODE boardsarelocatedinsidethecavern,on eithersideof thedetector.

TheExperimentControlSystem(ECS)of themuon detectoris a distributed systembasedonCAN bus [68]. The ECS performsbasiccontrolandmonitoringof theODEboardsandcontrolsthe

Table 17 Totalnumberof unitsin thesystem

Item Numberrequired

Front-endboards 75368-channelASD chips 15072DIALOG chips 7536Serviceboards 144Serviceboardcrates 12Intermediateboards 168Intermediateboardcrates 12SYNCchips 4032ODE boards 168ODE boardcrates 12Data-concentratorboards 12

Table 18 L0 andL1 parametersfor LHCb FE elec-tronics

L0

Maximumrate 1.1MhzLatency 4.0µsConsecutive triggers Max. 16Derandomiserdepth 16 eventsDerandomiserreadouttime 900ns

L1

Rate 40–100kHzBuffer size(latency) 1927eventsDerandomiserbuffer Min. 15events

FEBs throughspeciallydesignedServiceboards(SB) (seesection5.3.5). Furthertechnicaldetailsof themuonarchitecturearefoundin Ref.[69] and[67]. A simplifiedschemeof themuonarchitectureis shown in Figure59.

Table17 givesa summaryof thetotal numberof units in the system. The Level-0 and Level-1parameters,definedin [24], aresummarizedin Ta-ble 18. Thesetableswill bereferredto in the fol-lowing sections.

5.3.2 Front-end boards

The on-chamberelectronicsis basedon a two-stagescheme. The main functionalitiesare per-formedby two chips.

53

On C

HA

MB

ER

SIn C

RA

TE

S

OW

ff Detector Electronics (XODE)

PhysicalChannels

(120k)

DAQ

BX SynchronizationFine Time measurement

L0 & L1 buffersTY

rigger & DAQ interfaces

TY

rigger

Intermediate Boards (IB)

In CR

AT

ESOut

Logical channelGeneration

(IB)

TZ

TCRx(clock)

DIALOG

Front End Boards (FEB)

S[

ervice Boards

ECS

Front-end controlsDCS nodes

Low voltagePulse system

A\

SD A\

SD A\

SD

Programmable Delays

Logics, DAQs, I2C Node

42k channels

25k channels

8.6k logical

channels

17.3k logical

channels

Figure59 Simplifiedschemeof MuonArchitecture.

DIG S]

hape

ASD’s Thresholds

Configuration IN (~150 bits)

1^6

LVDS I_N

1^6

Adjustable D`

elays

Ma

UX 16:1

Pb

hysical channel rate I_ 2 Cc

Inter face

EC

S

DAC 1^0 bits x 2

2 Adjustable Delays To ASD S]

ync Calibration Signal (from SB)

S]

tar t/Stop (from SB)

S]

caler

Od

R/AND4 x 4

Od

R/AND2 x 8

MUX

8e LVDS Od

UT (fto ODE or

IB)

Od

R/AND8 x 2

MASK REG 1^6 bits

Delay REG 5g bits x 18

Mux REG 2h bits

InMux REG 4i bits

MA

SK

Figure60 DIALOG chipschematic.

54

5.3.2.1 ASD chip

The first stageconsistsof the Amplifier-Shaper-Discriminator (ASD) chip. The characteristicsof these chips are describedin Sections4.1.1and4.2.1. This stageoutputsdigital signalsusingtheLVDS standard.Thesignalshaveapulsewidthof 50B 10ns,dependingontheshapeof thesignalsbeforethediscriminator. Moreover, theASDQandCARIOCA chipsoptionallyprovidedigital signalswith alengthproportionalto thepulseheightof thechambersignal. This canbe usedfor testingandmonitoring purposes. EachASD chip integrateseightchannels.

5.3.2.2 DIALOG chip

The secondstageis the DIALOG chip, describedin detail in Ref. [70]. EachDIALOG chip dealswith sixteenphysicalchannels. It processesthedigital outputsof the first stageas illustrated inFigure60.

Signalsenter the programmabledelays(oneper signal),wherethey canbe delayedfrom 0 to25ns in stepsof 1.5ns. Physicalchannelscanbemasked,in orderto make themindividually acces-sible throughthe logical-channelreadout. Signalarrival times aremeasuredin the ODE boardsattheendof thechain.

Thesignalsin theDIALOG chipareshapedtoa fixed width lessthan 25ns. Dependingon thespecificpositionof physicalchannelsin thedetec-tor, thesignalsarecombinedwithin thesamede-tector layer accordingto the logical channelsize,andwith thecorrespondingsignalsin theotherde-tectorlayer.

Individual physicalchannelscanbeexaminedandthe numberof accumulatedhits canbe mon-itored via a dedicatedcounter. The circuit is ac-cessedvia the ECS,which takescareof configu-ration andprogramming.DIALOG alsocontainstwo 10 bit DACs for settingthe thresholdsof twoASD chips.

A singlefront-endboard(FEB) dealswith 16physicalchannels. It consistsof two ASD chipsandoneDIALOG chip. Eachboardcanoutputamaximumof eightlogicalchannels.Dependingonthe local topology, eight, four or two outputsareused. The digital outputsof variablepulsewidth

Table 19 Numberof IBs in eachof the five half sta-tions. The numberof Input/Outputsignalsper boardaregivenin brakets.

M1 M2-M3 M4-M5

R1 – – –R2 – – 12(96/28)R3 – 24(192/56) 12(96/40)R4 – 24(192/56) 12(96/40)

canalsoexit theFEBsfor specialtestsof thecham-bers,andof theFEBsthemselves.It is notforeseento usethis featureduringnormaldatataking.A to-tal of 7536FEBsis foreseen.

5.3.3 Intermediate boards

In regionR2of stationsM4 andM5 andin regionsR3 andR4 of stationsM2 to M5, a further levelof logicalcombinationis required,becauselogicalchannelsareformedfrom signalswhich originatefrom differentchambers.Thelogical combinationis carriedout in theintermediateboards,whichareplacedto thesidesof thedetector. EachIB allowsa maximumof 192 inputsanda maximumof 64outputs. Both the inputsand the outputsusetheLVDS standard.TheIBs only performtheresiduallogic (ORs)neededto completethegenerationoflogical channelsin theregionsmentioned.

A total of 168IBs is foreseen.Table19 showsthe numberson inputsandoutputsper IB for thedifferentdetectorregions.

5.3.4 ECS interface

A distributed control systembasedon a specificfield-bus provides the basiccontrol and monitor-ing functions of the muon detector. The archi-tectureof this systemis basedon ELMB (Em-beddedLocal Monitor Box [71]) a specialCANnode board designedby the ATLAS collabora-tion to operatein a moderaterate environment.The ELMB is a general-purposesmall plug-onmodule,comprisingtwo commercial8-bit micro-controllers and one CAN-Controller chip. Themain processoris a 4 MHz micro-controllerwith128kbytesflashmemoryto storetheprogram. Ithandlescommunicationwith the CAN controller

55

ELMB CAN node

ELMB CAN node

ELMB CAN node

2x I2C ELMB CAN node

2j

x I2C

2x I2C

I2C

I2C

I2C

SB

LVDS Long cables

LVTTL Short cables

FEB FEB FEB

10x FEBs

PC 4x 6 CAN branches per PC

5 CAN branches per PC

Each branch tko 6 Service

Boards

Each SB has 4 CAN nodes

PC 2x ODE

control

ODE

ELMB CAN node

Each branch to 1/2 muon station 10–24 CAN nodes

Counting room

FEB and SB control

Sl

YNC Cm

hips

FPGAs

JnTAG

Figure 61 TheECSsystemfor theMuon Chambers.Shown aretheECSCAN interfacesto theServiceBoards(SBs)andthefanoutsto theFrontEndBoards(FEBs).TheCAN branchesto theOff DetectorElectronicsboards(ODEs)arealsoshown andtheinternalconnectionswithin theODEboards.

and performsuser-chosenlocal tasks. A secondsmallermicro-controlleractsasdedicated“watchdog”, prompting a refreshwhen the first micro-controllerblocks, for examplein the caseof Sin-gle EventUpset(SEU).The ELMBs aregroupedin differentCAN bus branches,eachbranchcon-tainingupto 24ELMB CAN nodes.TherearetwoECSsubsystems,sketchedin Figure61, oneser-vicing the front-endboardsthroughthe SBs andcomprising24 CAN branches,the other servic-ing the ODE boardsvia 10 CAN branches.TheCAN branchesarecontrolledby six PCsplacedinthecountingroom. TheECSstructureis given inRef. [67].

5.3.5 Service boards

The Service Boards (SBs) are 9U size VMEboards.EachSB housesfour ELMB CAN nodes.EachELMB CAN nodecanhandletwo I2C buses.Thesebusesareextendedupto 10m (longdistancebranches)using an LVDS driver. At the end ofthe long distancebrancha LVDS receiver placed

on a chamberwill drive tenDIALOG chipson theFEBswith thestandardLVTTL I2C bus(shortI2Cbranch) [67]. At start-up,a remoteCAN com-mandwill triggera localprocessin theELMB thatwill write the registersinside the DIALOG chip.The ELMB periodically checksthe consistencyof the DIALOG registers. In caseof errors, theELMB senddiagnosticsinformationto the CAN-busPCinterfaceandcorrecttheerror. In addition,theELMB regularly monitor the ratecountersin-sidetheDIALOG chip andreturnthevalueto theCAN-busPCinterface.Therearetwo waysto ac-cessandcontroltheregistersof theDIALOG chip:thefirst requirestheuseof specialtasksrunningin-sidetheELMB, whichperformsall theoperations,the secondinvolvesrunningfrom the control PC,usingI2C instructions.Theuseof theELMB localintelligencereducesthe loadof thePCprocessorsandof network communications.

Anotherfunction of the SBsis to provide thepower supply generationcircuitry for the FEBs.Theappropriatevoltagelevelsaregeneratedontheserviceboardsandtransmittedto theFEBs.As the

56

distancebetweenthe SBs and the FEBs is 10 to15m, a voltageregulationfacility is envisagedfortheFEBs.

TheSBsarealsousedto sendandreceive sig-nalsuseful for front-endcalibrationandto moni-tor the correctoperatingconditionsof the cham-bersandfront-endelectronics(e.g. temperature).A systemto pulsethe front-endchannelsfor testanddiagnosticsis alsoenvisaged.

The SBs are placedto either side of the de-tector, in the rackshousingboth the intermediateboardsandtheODE boards.In total 144SBsareforeseen.

5.3.6 Off-Detector Electronicsboards

TheODE boardssynchronisesignalsanddispatchthemto the L0 trigger. They alsocontainthe L0pipelines,theL1 buffersandtheDAQ interface.

A schematicof the ODE boardsis shown inFigure 62. Eachboardreceives up to 192 logi-cal channels(LVDS) and outputsdatato the L0muon trigger and the DAQ system. The signalsto the L0 muontrigger aresentdirectly via opti-cal links. The datafor the DAQ aremultiplexedonto the Data Concentrator(DC) board(one perODE crate),wherethey areformattedandsenttothe readoutunits (oneS-link percrate). Up to 18ODE boardsandoneDC arehousedin oneODEcrate.TheDC alsocontainstheTTC-Rx chip andmanagesthe distribution of the systemclock andtriggersignalto theODEboardsin thecrate.

IncomingsignalsareassignedtheappropriateBX identifierandsentto theL0 pipelines.In par-allel, thedataandthe four leastsignificantbits ofthe associatedBX Id are sent to the L0 trigger.Dataresideson theL0 pipelinesfor 4µs beforere-ceiving the L0 yes/nosignal. Triggereddataarethenmoved to the L1 buffer, wherethey wait forthe L1 decision. If this is positive, dataarezero-suppressedandmultiplexedto theDC ontheback-planebus.

The averagedatasizefor b o µ X eventshasbeenestimatedto about1.6kbytes,assumingthatfull information for eachhit would be a 32-bitword. About 75% of the hits originatefrom sta-tion M1.

TheLVDS receivers,theL0 pipelinesandthe

L0 derandomizerareintegratedin asinglecompo-nent,theSYNC chip [72], presentlyunderdevel-opment.A customchiphasbeenpreferredoveranFPGAsolutionfor severalreasons,fully explainedin reference[72]. This solution is morecompactandreliable,allowing greaterflexibility in theim-plementationof functionscrucial for systemoper-ation:

p Systemtime alignment: The relevance ofmeasuringthe phaseof logical channelswithin theBX periodhasbeendiscussedinreference[40]. A low-resolution(3bit) TDCis adequatefor thispurpose.Thetime infor-mationassociatedto the hits is sendto theL0 pipelineandentersthenormaldatapathto theDAQ.

p Systemremotecontrol: It is of fundamentalimportanceto equipthesystemwith a num-ber of error-detectionfeatures,allowing re-motecontrolanddiagnosisof possiblemal-functioningin theboards.

Theothermain boardcomponents(L1 buffer,trigger interface and bus interface) are basedonFPGAs. EachODE boardalso containsa CANnode.

A total of 168ODE boardsand12 DC boards(onepercrate)is foreseen.

5.3.7 Systemsynchronisation

A systemsynchronisationprocedureunderconsid-eration involves two phases,briefly describedinthefollowing:

1. When the whole systemis powered-up,acalibrationrun is performed. This consistsof a completescan of all physical chan-nels. In order to adjust a single channel,other physicalchannelsnormally logicallyORedwith it are masked within the DIA-LOG chips. The signals’BX identifier andthephaseinsidetheBX cycle aremeasuredat the level of theL0 pipelines.Time spec-tra for eachchannelareacquiredinsidetheSYNC chips.

57

L1 Buffer

ECSnode

TriggerInterface

Oq

pticalLinks (r12)

Bus InterfaceBoardCs

ontroller

OD

E B

oard

Logical Channels

BXst ync

L0 p

ipel

ine

Phasemeasure

Su

ignal phase(r3 bits

Resolution)Hits

Derandomizer

TimeSu

pectraRAM

LVDS Receivers

MultiplexerL0

yes

BX

clo

ckSync Chip

Res

et

L1

yv esL0

Trig

ger

EC

S s

yste

m

Fine Time

TTC-Rx

ZeroSu

uppression

Cs

rateCs

ontrollerSu

-link

DAQ Interface

DAQ (RU)

Data

Concentrator

Data from boards in the crate

Hits + BX Id

BXcw ounter

Rou

t Con

trol

s

Figure62 Schematicof theODEboardsandthecrateinterface

58

2. Thedetectionof the relative phasebetweentherising edgeof thesystemclock andthatof thetime spectraallows thecalculationofthe delay to be addedto eachchannelforcorrecttime alignment.This delayconsistsof two numbers.The first is the phaserel-ative to theclock rising edge(fine-timeoff-set), the secondis the differencein BX cy-cles(bunch-crossingoffset). Thedelaycor-rection is realizedby setting the fine-timecorrectionontheFEB(insideDIALOG) andthe BX correction inside the ODE boards(SYNCchip).

Comparedto otherschemesnot usingsignal-phaseinformationof channels,the procedurede-scribed above is rather fast, although sufficientstatisitcsalso for the lower-rateouter regionsareneccessary. Moreover, it must be executedonlyonce,when the systemis set-upbeforedatatak-ing. The fine-time and bunch-crossingoffsetsconstitutecalibrationconstants,to be saved in adatabase.Thedetectoralignmentis monitoredon-line usingtheSYNCchip.

5.3.8 Radiation levels

Theradiationlevel estimateconsideredfor thesys-temdesignis givenin Ref. [73]. This givesanac-cumulateddosein 10LHC yearsof about1 MRadin theinnerpartof stationM1. Thedosedecreasesradiallyandis reducedto about1kRadat thesidesof the detector. This meansthat radiation-hardelectronicsmustbeusedfor theFEBs,while com-mercialdevicescanbeplacedin theIBs andODEboards. Both the CARIOCA and the DIALOGchip are designedin radiation hard-technologies(0x 25µm CMOS). It is also importantto considerparticle rates,which affect the singleevent upset(SEU)behaviour of thesystem.In this respecttheuseof FPGAs, containingimportantamountsofSRAM, mustbe consideredcarefully. Wheneverpossible,complex logical functionswill be imple-mentedusinganti-fusebasedFPGA technologies(e.g. Actel family). Thenumberof SRAM-basedFPGAs(e.g. Xilinx andAltera families) is to beminimised,even in the ODE boards. Wheneverused,a specific procedureis envisagedto com-pare the FPGA configurationwith a local copy.

This function is performedby theCAN-nodemi-cro controller, which alsotakescareof uploadingtheFPGAconfigurationwhenmismatchesarede-tected[74]. SEUimmunity of theDIALOG chip,also placedin the higher-rate regions, is ensuredthroughtheuseof triple voting for all internalreg-isters and the implementedlogical functionality.Moreover, thecontentof the registersis refreshedregularly via theI2C interface.

5.3.9 Cooling of the FE-electronics

The inner regions of the Muon Systemare char-acterisedby a large amountof electronicslocatedin a smallspace,in particularin region R1 of sta-tions M2 and M3 where eachchambercontains224 readoutchannels. An estimateof the dissi-patedpower hasbeenperformedin thesecases,assuminga Faradaycagecontainingthe chamberandtheelectronicswith thermalcontactonly alongtheouterperimeter, while the front andrearfacesareconsideredinsulating.A nominalconsumptionof 15mW/channelhas been assumedfor CAR-IOCA. To this number, the consumptionof theDIALOG chip and the presenceof someserviceelements(local regulators, controls, etc.) mustbe added. The amountto be addedhasnot yetbeenquantified,and is thereforenot consideredin this evaluation.UsingtheapproximateformuladT = 900y (P/S)0 z 8, wheredT is the internalther-malgradient( C),Pis thetotaldissipatedpower inthebox (W) andS is thesurfaceavailablefor heatexchange(cm2), the inner chambersarefound tobe subjectto an increasein temperatureof 10 C.Althoughthisdoesnotrepresentasignificantprob-lem, several possibilitiesare under investigation,for examplepumpingair throughthe box with asimple pipe network. A betterunderstandingofthesituationwill comewhena morerealisticesti-mateof thepowerconsumptionbecomesavailable.In addtion,a testwith a realbox is plannedin thenearfuture.

5.4 Service systems

5.4.1 Gassystem

The design of the gas systemunder considera-tion for the muonsystemis describedin detail in

59

Figure63 Schematicpipeandcomponentdrawing of themuongassystem.

60

Ref. [75]. As a consequenceof the two detectortechnologiesemployed, two independentgassys-temshave to be designedand constructed. Thegas volume of the MWPC amountsto 4.5m3.Test resultshave shown that this detectoris suit-ably operatedwith a gasmixtureof Ar | CO2 | CF4

(40:50:10).For theRPCdetectorwith a total vol-umeof 0.83m3 a mixtureof C2H2F4 | C4H10| SF6

(95:4:1) non-flammablecompositionis foreseen.A compilationof thebasicparametersof bothde-tectorsconcerningthe gassystemis listed in Ta-ble 20.

Gas componentswill be mixed with the ap-propriatecompositionin the mixer in the surfacegas building at the experiment. In addition theRPCgasmixture will be monitoredcontinuouslywith an infrared analyser. If the C4H10 ratio ex-ceedstheflammabilitylimit thegassupplywill bestoppedimmediately. Both gassystemswill runin a closedloop (Fig. 63). The expectedcircula-tion flow rate will be between6 to 12 hourspervolume exchange. A pump in the return line al-lows thegasto becompressedbeforeenteringthegasbuilding at the surface. To stabilisethe pres-surein themuondetector, a back-pressureregula-tor in parallelwith thepumpcontrolsthepressureto 0.5mbarbelow atmosphericpressureattheinletof thepump.Usinganinline purifiersituatedin thegasbuilding at the surface, the regenerationratewill be kept above 90%. The purifier consistsoftwo cleaningagents:a molecularsieve (3A) to re-move watervapourandactivatedcopperasreduc-ing agentfor oxygenremoval. A humidity andanoxygenmeterwill measurethe impurity concen-trationbeforeandafter thepurifier. Eachhalf sta-tion of themuonsystemis connectedto onesupplyline andonly in stations4 and5, whereboth gasmixturesareneeded,two input lineswill beused.Distribution racks,mountedon the detectorsup-port,will distribute thegasto themuonchambers.Onehalf stationis dividedin tengasregions,con-trolledby gasflow metersattheinputandoutputofeachseparatedgasvolume.A safetyrelief bubbleris incorporatedinto every gaschannelto preventover pressureof thechambermodules.Theexist-ing DELPHI supplyandreturnpipesbetweentheSGXbuilding andtheUX cavernwill bereusedbytheLHCb experimentandhencefor thetwo muon

gassystems.Thegascontrolwill follow thegen-eralrecommendationsof theJoint-Control-Projectof thefour LHC experiments(JCOP)[76].

5.4.2 High-Voltagesystem

The HV distribution systemof the chambersisbasedon the assumptionthat for the safeandre-liable operationof the detectorsthereshouldbeindependentoperationof the gapsinsidea cham-ber. However, for theMWPCs(andpossiblyalsofor theRPCs)this leadsto a very largenumberofHV channels(3456for theMWPC) that couldbeprohibitively expensive. Alternativesolutionshavealsobeenstudied.

5.4.2.1 MWPC detector

For the MWPC, assumingthat the variouscham-bersexhibit similargainresponses(i.e. the”knee”of the efficiency plateau is within 50V) andthereforewill beoperatedwell insidetheregionoffull efficiency, a schemeis foreseenwhereseveralgapsareseriallychained.For thedeterminationoftheoptimalconfiguration,the following consider-ationsshouldbetakeninto account:

~ Chainsshould comprisehomologousgaps(i.e. gap1 of chamber1 chainedwith gap1 of chamber2, etc...),to minimise,in caseof HV problems,the effect on geometricalefficiency of thetrigger;

~ Due to high local rate, chains shouldequalisethe amount of current drawn byMWPC, alsothecharacteristicsof HV sys-temscommerciallyavailableshouldbecon-sidered.

In this schemethenumberof channelsneededwould be 1200. Assuminga currentof 3µA perMHz of incomingradiation(atagainof 105), eachHV channelwould draw lessthan0.5 mA at theoperatingvoltage.

TheMWPC HV systemwill have a maincon-trol unit, fully remotelyprogrammable,in thecon-trol room and detachedsubunits, built in radia-tion toleranttechnology, locatedin theoff-detectorelectronicsracks, which will provide HV to the

61

Table 20 Basicgasparametersof MWPCandRPC

Detector MWPC RPC

Gas Ar | CO2 | CF4 C2H2F4 | C4H10| SF6

Volume(m3) 4.5 0.83Flow Rate(m3 | h) 0.375–0.75 0.07–0.14Regenerationrate(%) 90 90Impurities:O2 (ppm) 50 10Impurities:H2O (ppm) 50 10Pressure(mbarabove atm.) 1 1Pressuremax.(mbar) 5 3

chambers.Eachchannelshoulddeliver up to 4kVand 1mA. Models with thesecharacteristicsarecommerciallyavailable.

A carefulschemeof HV-grounding,compati-blewith thegroundingof thedetectorandFEelec-tronicsis understudy.

5.4.2.2 RPC detector

The RPCsystemrequires960 HV channelseachable to deliver up to 12kV and 1mA. The solu-tion envisagedatpresentfollowsthedevelopmentsmadein thisfield by theATLAS andCMS collab-orations. In sucha schemethe RPC power sup-ply systemis madeof two main blocks: a mainboard(GeneratorBoard,GB), locatedin thecount-ing room,supplyingthepower at a mediumvolt-ageandmanagingthe remotecontrolsandmoni-toring,andaremotesystem(DistributorBox,DB),located in the electronicsracks, consistingof aradiationtolerantboardhostinga transformationstagegeneratingboththehighandthelow voltagesneededby thesystem.In this way theproblemofdistributing thehighvoltageover longdistancesisminimised,while keepingthe main power supplycontrollers,which areradiationsensitive, in a safeenvironment.

The GBs supply a mediumvoltagewhich iseasyto distributeover longdistanceswith conven-tional cables(the foreseenvoltage is 48V), andsupportsa communicationlink for remotecontrolandmonitoringpurposes.The GBs would be in-terfacedwith thegeneralLHCb controlsystemandwould beplacedin thecountingroomallowing aneasyandsafeaccessto thepower supplysystem.

In theDBs the input mediumvoltageis trans-formed into two HV channels,which are float-ing, allowing an optimal grounding configura-tion. Moreover, the HV channelsare fully pro-grammableremotely. The voltagesand currentscanberemotelymonitoredthroughthecommuni-cationlink drivenby theGBs.Thecommunicationline is optically decoupledto preserve thefloatinggroundof thevoltagechannels.

Finally, the feasibility of a solutionwith a re-ducednumberof HV channelsto beappliedin thecasethe costbecomesprohibitive is understudy.In this configurationtwo gapsin a singlechamberwouldstill bepoweredby differentHV lineswhilethetwo gv tdr.pscorrespondinggapsbelongingtotwo adjacentchambersin thebendingplanesharethe sameHV. This configurationwould allow tokeepa high detectionefficiency even in caseoffailure of a high voltagechannel,while reducingby a factorof two the numberof DBs neededbythesystem.

5.5 Chamber support structures andmuon filter

Two independentsupportstructuresareenvisagedfor themuonsystem:

~ A supportstructurefor the chambersandelectronic racks, suspendedfrom a gen-eral structure(gantry)on the top, which isusedaswell by othersub-detectorsinclud-ing the scintillating pad detector, and thepreshower [26].

~ A supportstructurefor the muon filter on

62

Figure64 Sideview of theMuonsystemsupportstructures

63

movableplatforms.

The two support structureswill be briefly dis-cussedin this section. Details can be found inRef. [77].

A common requirementfor all LHCb sub-detectorsis that provision hasto be madefor ac-cessto theLHC beampipe, for maintenancepro-cedures,in particularfor bake-out. For this rea-son the muon systemwill be constructedin twohalves,so that onesidecanbe withdrawn, allow-ing accessto thebeampipe. Thecompleteretrac-tion of eachhalf stationis alsoimportantfor instal-lation andmaintenanceof the chambersandtheirFE-electronics.

5.5.1 Chamber support structures

Eachmuonstationconsistsof 276muonchambers.They arearrangedin four layersto provide full an-gular coverage. Two layersare in front and twobehinda wall-like supportstructurehangingfromthetop. Themuonstationsmustfit into the40cmspaceavailablebetweentheabsorberwalls,exceptfor stationM1, for which the allocatedspaceisonly 37cm.

5.5.1.1 Requirementsand constraints

The muon stations will be constructedin twohalves. Thepossibility to completelyretracteachhalf station should be ensuredfor installationand maintenanceof the chambersand their FE-electronics,andto provideaccessto thebeampipe.

The chambershave to be positionedwithineachmuonstationwith a precisionof about1mmin the x andy directionsand5mm in the z coor-dinatealongthebeamaxis. Gaspipesandcablesto the chambersfor readoutand HV/LV have tobesupportedwithin theallowedthicknessfor eachstation. Gaspipeswill be routedvertically fromthetop to thedistribution system,while signalandHV/LV-cableswill be routed horizontally to thesides,wherethe readoutelectronicrackswill beplaced.

Table 21 summarisesthe overall dimensionsandweightsof thefivemuonstationsandthesup-portingstructuresfor thefivemuonstations.

Figure 65 Fixationof chambersto thesupportstruc-ture

5.5.1.2 Wall structure

Thesupportwallsaresuspendedby meansof rect-angular linking piecesfrom an iron beam. Onthe bottom, the walls areguidedto prevent acci-dentalcontactswith iron absorbersduring open-ing/closingoperationsand to improve the stabil-ity of thestationposition. Two racksfor theelec-tronicsareplacedat half heighton both sidesofeachmuonstation.Theracksaresupportedby anindependentlinking pieceandrigidly fixed to theexternalsideof the chambersupportingstructureto preventany relative displacement.A schematiclayoutis shown in Figure64.

Each chambersupport wall consistsof two1-2mm thick aluminium sheets,interleaved by alayerof 0.7mm thick corrugatedaluminium. Thebenefitsof this designarea high degreeof mod-ularity andfull flexibility in chamberpositioning.The total thicknessof the wall is 42mm for M1and44mm for M2–M5. Rivetsareusedto con-nect the threesheetstogether. Eachhalf wall iscomposedof several panelswhich areconnectedtogetherby rectangularaluminiumtubes.Special

64

Table 21 Muon stationsupportstructuredimensionsandweight. The indicatedchamberweight is basedonhoneycombpanelsfor theMWPCs.

StationM1 StationM2 StationM3 StationM4 StationM5

Activearea( m2) 49.0 76.6 89.2 102.7 117.27.7 y 6.4 9.6 y 8.0 10.4 y 8.6 11.1 y 9.3 11.9 y 9.9

Wall structurearea( m2) 52.1 80.4 94.3 108.3 121.07.9 y 6.6 9.8 y 8.2 10.6 y 8.9 11.4 y 9.5 12.1 y 10.0

Totalweightof chambers(kg) 3700 5400 6100 4900 5500Wall structureweight(kg) 471 1223 1369 1564 1742

carehasto betakenin designingthejunctionof thetwo halves,avoiding any mechanicalinterferencedueto theoverlappingof chambers.

5.5.1.3 Chamber positioning

The Chambersareattachedto horizontalrails onthe walls, which allow to move and position thechambershorizontally. The correctvertical posi-tioning is fixed by the position rails. Referencepinswill beaddedto fix thecorrecthorizontalpo-sition of thechambersandto beableto reproduceit. Thesystemis sketchedin Figure65.

5.5.1.4 Maintenance

Accessto the chamberscloseto the beampipe isonly possibleif a muonstationis completelyre-tractedbetweenthe iron absorbersby about4mfor stationM1 andabout6m for stationM5. In-cluding about3m spacefor electronicsracksandaccessto them,the requiredspacefrom thebeamline extendsto 11m in stationM1 andto 15mfor stationM5 (seeFigure66)4.

5.5.2 Muon filter

The muon filter is comprisedof the electromag-netic and hadronic calorimetersand three ironshields,interleavedbetweenthemuonstations.Anadditionalshield immediatelybehindstationM5

4The presentlayout of the LHC cryogenicsaccommo-datedat the end of the LHCb experimentalareadoesnotleave sufficient spacefor the maintenanceof the muonsys-tem.Otheroptionsfor thepositioningof theLHC cryogenicsarethereforeunderinvestigationin acollaborative effort withtheacceleratorgroupsconcerned.

Table 22 Muonfilter composition

DetectorElement DepthCalorimetry 267cm 6.8λI

Muonfilters1–3 80cm 4.4 λI

Total thickness 20 λI

protectsthis station from machinerelatedback-groundand back splashfrom nearbyLHC beamelements.

The total weight of the muon shield is about2100tons. The compositionof the muonfilter issummarisedin table22.

Thematerialfor themuonfilter will comefromtheiron blocksof theWestAreaNeutrinoFacility(WANF). The blockswill becomeavailableearlyin 2002 and have a density of about7.2kg/cm3

anda sizeof 80cm and160cm in length, 40cmand80cmin width and20cm,40cmand80cminheight. Thesedimensionsmatchthe requiredab-sorberthicknessof 80cm. A detailedlayoutof theabsorberwalls with the existing blocks hasbeendone[77]. The areaaroundthe beampipewill beequippedwith specialblocks,whichgivealsosup-port to theplugsbetweenthewalls.

The movable platforms for the muon filtermake use the existing rails in the experimentalarea,formerly usedfor theendcapof theDELPHIdetector. It is not foreseenthat the filters have tobe moved idependently. Openingof the iron ab-sorbersimplies openingthemuonstationsof thatside,sincethe electronicsracks,which are60cmwide and permanentlyconnectedto the stations,do notallow theopeningof theiron wallsalone.

In orderto ensurestaticstabiltyin caseof seis-

65

Figure66 Top view of theMuonsystemwith thestationsmovedout for maintenance.

66

mic activity, theiron blocksformingonemuonfil-ter areconnectedby weldedjoints. Following thisprocedure,monoblocksof absorberswill be cre-ated. In addition, the iron absorberson the samesideof thebeampipewill beconnectedto onean-other to enhancefurther the static stability. Themechanicalconnectionshave to beremovedeverytime a stationwill be retractedfor accessto thechambers.No moving of the iron wall is foreseenfor maintenanceof themuondetector.

5.5.3 Beampipe shielding

Specialplugsareforeseencloseto thebeampipetoprotectthechambersfrom particleswhich emergefrom thebeampipe.An optimizationof thisshield-ing hasbeenperformed[78], taking into accountthe 1.5cm requiredspacefor bakeout instrumen-tion of the beampipe. The plugs extend from12mradout to 18mradin x projections,andfrom12mradto 15mradin y projection.

5.6 Safetyaspects

The Muon detectorsof LHCb will follow theCERN safetyrulesandcodes,CERN safetydoc-umentSAPOCO42 andEuropeanand/orinterna-tionalconstructioncodesfor structuralengineeringasdescribedin EUROCODE3.

In thefollowing thespecificrisks,andactions,aresummarised,asdiscussedin the Initial SafetyDiscussion(ISD) with the CERN Technical In-spectionandSafety(TIS) Commission.

1. The chambersdiffer in dimensionsaccord-ing to theirpositionin themuonsystem.De-pendingontheirsize,theweightof MWPCsdiffers between 5kg in region R1 and 30kg in region R4. Theweightof a RPCis 20kg. The chamberscanbe handledby1–2 personsduring the installation,follow-ing thesafetyinstructions5.

2. The final choice of the chambermaterialswill bedoneaccordingto thesafetyinstruc-tions6, and the use combustible materials

5SafetycodeA56SafetyInstructions41

will beminimised.Alternative materialsre-placingthepolystyrenein theRPCsareun-derinvestigation.

3. The gas mixtures used in the muonchambers,Ar | CO2 | CF4 (40:50:10)for theMWPC detector and C2H2F4 | C4H10| SF6

(95:4:1) for the RPC detector are notflammable.As thedetectorswill beoperatedwith a maximumoverpressuretowardstheatmosphereof 300 Pa setby high through-putbubblers,thechambersarenotclassifiedaspressurevessels7.

4. The RPC detectorswill be run at 10kV8.The total current for eachsupply line willbe 1mA. The low voltagesupply to thedetectorread-outis below 15V9.

5. Thehigh pressurepartof thegassystemsislocatedin the surfacebuildings. The sys-temswill bebuilt accordingto theappropri-aterules10.

6. The effectsof seismicactivity will bestud-ied in collaborationwith TIS.

7SafetycodeD2 Rev.28H.V.A. asdefinedin SafetyInstructions339SafeExtra Low Voltage(S.E.L.V.) asdefinedin Safety

Instructions3310SafetyInstruction42andSafetyCodeD2 Rev.2

67

6 Project Organisation

6.1 Scheduleand Milestones

Theoverallwork programmeandscheduleis sum-marisedin Figure 67. It is split into threeparts,the chambers,the electronicsand the infrastruc-ture. The schedulecovers the period up to theend of 2005 and ensuresthat the Muon systemis fully commissionedandoperatingtogetherwiththeotherLHCb sub-detectorsby this time.

6.1.1 Chambers

6.1.1.1 MWPC detector

The engineeringdesign of the MWPC detectorshouldbefinalisedandfrozenby theendof 2001,basedon the experienceacquiredwith the proto-type chambers. Moreover, by the end of 2001the MWPC ageingtestshouldhave provided sig-nificant results,providing additional informationaboutthelongtermbehaviour of thechambercom-ponentsunder irradiation. In the year 2002 the“module0” of eachof thechambersin thevariousregionswill beconstructedandtestedwith thefinaltoolsdevelopedduring theyear2001. In parallel,the productionlines will be set up. In casepan-els basedon honeycomb will be used,the panelpreparationis rathertime consumingand shouldthereforestart about1| 2 year in advanceof thechamberproduction. It is foreseento have fourcentresfor chamberconstruction,assemblingandtesting.Thetime estimatedfor chamberconstruc-tion is two years. Therefore,productionshouldhave startedby January2003 at the latest. De-pendingon the complexity of a chambertype, afully testedchambershouldbeproducedwithin 1–5 working days. Chamberinstallationand com-missioningof the muon systemshouldstart mid2004 and take aboutone year. The major mile-stonesfor theMWPC detectorsaresummarisedinTable23.

6.1.1.2 RPC detector

Developmentson prototypechambersshould befinishedby the end of 2001, when the engineer-ing designshould be finalised and frozen. Thetestsplannedin 2001 will allow a final decision

on the oil treatmentof RPCsin LHCb. In addi-tion, significantresultsfrom the RPCageingtestshouldbe available by that time. A “module 0”for eachchambertype will thenbe preparedandtestedwith atimescaleof ninemonths.In parallel,theproductionlineswill beprepared,consistingofthe assemblytools andtestsetupsin the institutelaboratories.A periodof threemonthsfor thepro-ductionandselectionof thebakelite plates,andaperiodof nine monthsfor the gapproductionhasbeenestimated.Both productionswill bedoneinindustry, receiving promptfeedbackfrom thequal-ity checksperformedin the institutes. Chamberassemblyshouldtake about100 working daysatthe rate of 5 chambers/day. Measurementswithcosmicrays will be performedin parallel in twocentresandshouldbecompletedwith a time scaleof oneyear. The chamberswill be transportedtoCERNin thefirst half of 2004,followed by a pe-riod of aboutoneyearfor installationandcommis-sioning. Themajormilestonesfor theRPCdetec-tor aresummarisedin Table23.

6.1.1.3 Chambersfor the inner part of M1

The detectortechnologyfor the inner part of sta-tion M1 shouldnot be chosenlater than January2003. This leavestwo yearsfor finalising thede-tectordesignandtheconstructionof thechambers.Installationandcommissioningshouldtake placein thefirst half of 2005.

6.1.2 Electronics

The main tasks for the Muon System FE-electronicsandthemajormilestonesarealsosum-marisedin Figure67 andTable23. A few impor-tantaspectsarepointedout in thefollowing.

6.1.2.1 Front-end chip

A critical taskis thedesignof theCARIOCA FE-chip. Besidesthepreamplifier, which shows verysatisfyingresults,the designincludesalso a fastshaper, a baselinerestorer(BLR) and a discrim-inator. In particular the latter two requiremorework during the year2001. In orderto obtainfi-nalproductsto bemountedontheFE-boardsearlyin 2003,thedesignandtestsof CARIOCA should

68

Table 23 Muon ProjectMilestones

Date MilestoneMWPC detectors

January2002 EngineeringdesigncompletedJanuary2003 Begin chamberconstructionandtestsDecember2004 Chamberconstructioncompleted

RPC detectorsDecember2001 Decisionon useof linseedoilJanuary2002 RPCengineeringdesigncompletedMay 2003 Begin RPCassemblyandtestsDecember2004 Chamberconstructioncompleted

Chambersfor the inner part of M1January2003 TechnologychoiceDecember2004 Chamberconstructioncompleted

ElectronicsMarch2002 CARIOCA designandtestcompletedMarch2002 DIALOG designandtestcompletedJune2002 SYNCdesignandtestcompletedOctober2002 Full chainelectronicstestcompletedJanuary2003 Begin FE-boardproductionOctober2003 Begin IM- SBandODE-boardproductionDecember2004 Electronicsassemblyandtestcompleted

Muon filter and support structur esDecember2003 Iron filter installationcompletedJune2004 ChambersupportstructuresinstalledJuly2005 MuonSystemcommissioningcompleted

be completedby March 2002,when the FE-chipstatuswill bereviewed. Assumingthatthetestsofthefull CARIOCA prototypearesatisfying,aboutninemonthsareleft for theengineeringrunandtheproductionof thechips. In casethetestsof CAR-IOCA revealsevereproblems,enoughtime wouldbeleft to switchto thebackupsolution,theadaptedASDQ chip. The viability of the backupsolutionwill be maintaineduntil the final decisionon theFE-chiphasbeentaken. ThenecessarystudiesfortheASDQ chip will becarriedout duringtheyear2001.

6.1.2.2 Readoutelectronics

Thebasicdesignof thevariouscomponentsof thereadoutelectronicschainshouldbecompletedbe-tweenmid 2001andmid 2002,asindicatedin Fig-ure67. This includesin particularthedesignof theDIALOG andSYNCchips.In orderto avoid ade-lay to the testof the full readoutchain,scheduledfor autumn2002,abackupsolutionintegratingtheessentialfunctionalitiesof the SYNC chip inside

anFPGAis envisaged.In parallelto thetestof thefull electronicschain,radiationtestsfor theDIA-LOG andSYNC chipswill becarriedout to provetheir SEU immunity. The testsarefollowed by aperiodof two years,duringwhich theengineeringdesignof thevariousboardswill befinalised,andthe tendering,production,assemblingandtestingtakesplace.

6.1.2.3 Monitoring and control

Thepreferredsolutionfor theECSinterfaceof themuon systemis basedon the CAN-ELMB stan-dard. In orderto prove theSEUimmunity of thisinterfaceandthebackupsolution(SPECS[79]11)extensive tests are in progressto allow a finalchoiceof the ECSinterfacefor the muonsystemin autumn2001. This will be followed by a pro-totyping and test phasebeforeproductionof theserviceboardsstartsbeginningof 2003.

11Serial Protocol for ECS, based upon the ATLASCalorimeterserialprotocolSPAC

69

6.1.3 Infrastructur e

6.1.3.1 Muon Filter

The iron blocks for the muon filter will becomeavailableearlyin 2002.As themuonfiltersoccupya ratherlarge amountof space,their constructionin the experimentalareawill be delayeduntil theinstallationof theLHCC cryolineshasbeencom-pletedin September2003. The installationof theshieldsshouldtake aboutthreemonths.

6.1.3.2 Chamber support structure

The muonsystemsupportstructuresareattachedto gantries,which limit theclearanceof thecranein theexperimentalarea.In orderto maximisethetimeduringwhich full useof thecraneis possible,the installationof the gantriesshould only beencarriedoutat thebeginningof 2004,whenthecon-structionof themuonfilters andtheinstallationofotherheavy equipmentis completed.Theinstalla-tion of thesupportstructureswill take aboutthreemonths.Themuondetectorinstallationcouldstartin parallel with the LHCC octanttest, scheduledfrom April to September2004. Sincebeamtestswill be donemainly in the nights, only minimalinterferenceis foreseen,except for the hindrancecausedby thebeampipe.

6.1.4 Installation and commissioning

Both MWPC andRPCdetectorswill undergo in-stallation and commissioningduring the secondhalf of 2004andfirst half of 2005. Commission-ing with other LHCb sub-detectors,using com-monDAQ will begin in thesummerof 2005. Sixmonthsof operationin this modeareforeseentoensurethe muon detectorswill be ready to takedataatnominalLHCb luminosityearlyin 2006.

6.2 Costs

The total cost for the Muon systemis estimatedto be 10,830kCHF. Table 24 shows the cost es-timatesplit accordingto the systemcomponents.For thechambersandelectronicboardsabout10%of spareshave beenincluded. Wherever possi-ble, thecostestimationof componentsis basedon

quotesfrom industryor recentpurchasesof similaritems.

The iron for the muon filter is a special in-kind contribution from CERN, for which thecosthasbeenestimatedto 4,000kCHF. Theremainingcostof 6,830kCHF for themuondetectorsystemhasbeenestimatedunderthe assumptionthat theCARIOCA FE-chip can be usedfor the MWPCdetector. The total costwould behigherby about600kCHF in casethe adaptedASDQ chip wouldberequired.

6.3 Division of responsibilities

Institutescurrently working on the LHCb Muonproject are: CentroBrasileiro de PesquisasFisi-cas CBPF, and UniversidadeFederaldo Rio deJaneiroUFRJ,Rio deJaneiro(Brazil),UniversitiesandINFN of Cagliari,Ferrara,Firenze,Roma“LaSapienza”,Potenza,Roma“Tor Vergata”,Labora-tori Nazionalidi FrascatiLNF (Italy), PetersburgNuclearPhysicsInstitutePNPI,Gatchina(Russia)andCERN.Work on the Level 0 Muon trigger iscarriedoutby CPPMMarseillein closecollabora-tion with theMuongroup.

The sharingof responsibilitiesfor the mainMuon Projecttasksis listed in Table25. It is notexhaustive, nor exclusive. Detailsof theresponsi-bilities for thevarioussystemcomponentswill befinalisedby thetime of theengineeringdesignre-ports,andwhenpreciseinformationof all fundingagenciesis available.

Softwareis a major taskin theprojectandnotlisted in Table25. It is understoodthat theMuongroupassuchis responsibleandwill have the re-sourcesof 6 FTE to provide all muonsystemspe-cific software. This includesalgorithmsfor simu-lation, reconstructionandmonitoringsoftwareforDAQ andcontrols.

The studiesrequiredto maintainthe viabilityof the backupsolution for the FE-chip are car-ried out by theCERNandRomeI groupsin closecollaborationwith theUniversityof Pennsylvania,wheretheASDQchiphasbeendesigned.

70

¡¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À ÁÂ Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ ÒÓ Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ïð ñ ò ó ô õ ö ÷ ø ù ú û ü ý þ ÿ ! " # $ % & '( ) * + , - . / 0 1 2 3 4 5 6 7 89 : ; < = > ? @ A B C D E F G H I J K L

M N O P Q R S T U V W X Y Z[ \ ] ^ _ ` a b c d ef g h i j k l m n o p q r s t u v w x y z | ~ ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ±² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á Â ÃÄ Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æç è é ê ë ì í î ï ð ñ ò ó ô õ ö ÷ ø ù úû ü ý þ ÿ ! " # $ % & ' ( ) * + , - . / 0

1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G H I J K L M N OP Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d e f g h i j k l m n op q r s t u v w x y z | ~ ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ²³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð ÑÒ Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß à áâ ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö÷ ø ù ú û ü ý þ ÿ !" # $ % & ' ( ) * + ,- . / 0 1 2 3 4 5 6 7 8 9 :; < = > ? @ A B C D E F G H I J K L M NO P Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d e f g h i j k l m n o p qr s t u v w x y z | ~ ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ²³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á Â Ã Ä Å Æ Ç È É Ê Ë ÌÍ Î Ï Ð Ñ Ò Ó Ô Õ Ö× Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é êë ì í î ï ð ñ ò ó ô õ ö ÷ø ù ú û ü ý þ ÿ ! " # $ % & ' ( ) * + , - ./ 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G HI J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _ ` ab c d e f g h i j k lm n o p q r s t u v w x y z | ~ ¡ ¢ £¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ² ³ ´ µ ¶ · ¸ ¹ º »¼ ½ ¾ ¿ À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì ÍÎ Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û ÜÝ Þ ß à á â ã ä å æ ç è é ê ë ì íî ï ð ñ ò ó ô õ ö ÷ ø ù ú û ü ý þ ÿ ! " # $% & ' ( ) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; <= > ? @ A B C D E F G H I J K L M NO P Q R S T U V W X Y Z[ \ ] ^ _ ` a b c d e f g h i j k l m n o p q r s t u v wx y z | ~ ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ®¯ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á Â Ã Ä Å ÆÇ È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × ØÙ Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ òó ô õ ö ÷ ø ù ú û ü ý þ ÿ ! " # $ % & ' ( ) * + , - . / 01 2 3 4 5 6 7 8 9 : ; <= > ? @ A B C DE F G H I J K L M N O P Q R S T U V W X Y Z [ \ ]^ _ ` a b c d e f g h i j k l m n o p q r s t u v w x yz | ~ ¡ ¢ £ ¤¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿ À Á ÂÃ Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û ÜÝ Þ ß à á â ã ä å æ ç è éê ë ì í î ï ð ñ ò ó ô õ ö ÷ ø ù ú û üý þ ÿ ! " # $ % & ' (

) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B CD E F G H I J K L M N O P QR S T U V W X Y Z [ \ ]^ _ ` a b c d e f g h i j k l m n o p q r s t u v w xy z | ~ ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ ­ ® ¯ ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾¿ À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö ×Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ÷ ø ù ú û üý þ ÿ

! " # $ % & ' ( ) * + , - . / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _ ` a b c d e f g h i j k l m n op q r s t u v w x y z | ~

Figure67 Scheduleof Muonproject,up to startof LHCb datatakingearly2006

71

Table 24 Muonprojectcostin 2000prices(kCHF).

Item Unit Number sub-totalof units (kCHF)

MWPC detector: 1220Panels m2 1400Specialcathodes m2 220Wire Fixationbars 0.5m 15000Frames 0.3m 6000Spacers piece 25000Wire km 1215HV-boards board 13000Tooling prod.lines 4Assembly chamber 900

RPC detector: 260Bakelite m2 1300Gasgapproduction chamber 500Strips m2 500Foampanels m2 500HV-Connectors piece 1000Mechanics chamber 500Assembly chamber 500

Electronics: 4040CARIOCA-chip piece 12500DIALOG-chip piece 8000RPCFE-chip piece 3500FE-boards board 8000Spark-Protection-boards board 8000LVDS links link 1800IM-boards board 180Off-Detector-Elec.-boards board 180Service-boards(ECS) board 160Data-Concentratorboards board 12Opticallinks to L0 trigger link 1250Crates crate 36LV-powersupplies/cables system

Services: 1310MWPCGasSystem system 1RPCGasSystem system 1MWPCHV System system 1RPCHV System system 1SupportStructures module 10

Muon filter: 4000Muon SystemTOTAL 10830

72

Table 25 Muonproject:Sharingof responsibilities.

Task InstitutesMWPC detectors:

StationsM1 – M3, outerpart Ferrara,LNF, PNPI,RomeI / Potenza(ConstructionandTesting)StationsM2 – M5, innerpart CBPF, CERN,Ferrara,LNF, UFRJ(Constructionandtesting)

RPC detectors:StationsM4 – M5, outerpart Firenze,RomeII(Constructionandtesting)

Inner part of station M1:(Constructionandtesting) Cagliari,LNF

Readoutelectronics:CARIOCA chip design,productionandtesting CERN,UFRJDIALOG chipdesign,productionandtesting CagliariSYNCchip/FPGAdesign,productionandtesting CagliariMWPCFE-boards(productionandtesting) CBPF, PNPI,RomeI / Potenza,UFRJRPCFE-boards(productionandtesting) Firenze,RomeIIIM-boards,design,productionandtesting LNFServiceboards,design,productionandtesting RomeIODE-boards,design,productionandtesting Cagliari,LNF

Services:Gassystems(Design) CERNMonitoring,Control(ECS) RomeI

Experimental areainfrastructur e:Chambersupportstructures CERN,LNFMuonfilter supportstructures CERNMuonfilter installation CERN

73

References

[1] LHCb TechnicalProposal,CERN/LHCC98-4.

[2] P. Colrain et al., “Optimisationof theMuonsystemLogical Layout”, LHCb 2000–016Muon.

[3] M. Adinolfi et al., “Proposal for the RPCMuon Detectorof LHCb”, LHCb 2000–53Muon.

[4] G.Auriemmaetal., “Thin GapChambersfortheLHCb Muon detector”,LHCb 2000–113Muon.

[5] B. Bochin et al., “Wire Pad ChambersandCathodePadChambersfor theLHCb MuonSystem”,LHCb 2000–114Muon.

[6] B. Bochin et al., “Wire Pad Chambersforthe LHCb Muon System”,LHCb 2000–003Muon.

[7] P. Colrainet al., “Performanceof a multigapRPCprototypefor theLHCb MuonSystem”,NIM A 456(2000)62-66

[8] N. Saguidova et al. “GCALOR StudiesofBackgrondin the LHCb Muon chambers”,LHCb 1998-059Expt;A. Tsaregorodtsev, “Muon Systemparame-terisedbackground– algorithm and imple-mentation”,LHCb 2000–011Muon.

[9] P. Colrain, “Upgrade of the Muon Systembackgroundparameterisation”,LHCb 2001–029Muon.

[10] ATLAS muoncollaboration,“ATLAS muonspectrometer, Technical Design Report”,CERN/LHCC97–22

[11] CMS muon collaboration, “CMS Themuon project, Technical Design Report”,CERN/LHCC97–32

[12] P. Bartalini et al., “Tuningof multiple inter-actionsgeneratedby PYTHIA”, LHCb 1999–028; LHC SM working group,“B decaysattheLHC”, CERN-TH/2000–101

[13] C. Zeitnitz and T.A. Gabriel, The Geant-CalorinterfaceUser’s Guide(1999), ¡¢£¤¢¤¢¥¦ §¨©¥¦ §¨©ª§

[14] I. Azhgirey, I. Kurochkin and V. Talanov,“Developmentof MARS CodePackageforRadiation Problems Solution of Electro-NuclearInstallationDesign”,in: Proc.of XVConferenceon Charged ParticlesAccelera-tors,Protvino,October22–24(1996)

[15] R.Le Gac,A. Tsaregorodtsev andV. Talanov,“Particle fluxesin the LHCb Muon System.Comparisonsof GCALOR and MARS cal-culations”,LHCb 1999–036Muon.

[16] W. Riegler, “Crosstalk, cathode structureandelectricalparametersof theMWPCsforthe LHCb muon system” LHCb 2000–061Muon.

[17] R. Santonicoand R. Cardarelli, NIM 187(1981)377.

[18] R. Cardarelli,A. Di CiaccioandR. Santon-ico, NIM A 333(1993)399;C. Baccietal., NIM A352(1995)552;R. Cardarelli,V. Makeev, R. Santonico,NIMA 382(1996)470.

[19] M. Adinolfi et al., “Performanceof low-resistivity single and dual-gap RPCs forLHCb” NIM A 456(2000)95

[20] W. Riegler, “RPC Simulations”, LHCb2000–112Muon.

[21] J.H. Lopes, “Simulation of chamberineffi-ciency andnoisein the LHCb muonsystemandstudyof its effecton theL0 muontriggerperformance”,LHBb 2001–060Muon.

[22] B. Bochin,et al., “Beam testsof theMCSCprototypes for the LHCb inner tracker”,LHCb 1998–008IT;O. Fedin, et al., “Study of MCSC with 36keV monochromaticX-rays”, LHCb 1998–010IT.

[23] G. Bencivennietal., “A triple-GEMdetectorwith padreadoutfor the inner region of the

74

first LHCb muonstation”, LHCb 2001–051Muon.

[24] J. Christiansen,“Requirementsfor the L0front-end electronics”, LHCb 1999–029Trigger.

[25] B. Schmidt,“LHCb Muon Systemby Num-bers”,LHCb 2000–089Muon.

[26] LHCb CalorimetersTechnical Design Re-port,CERN/LHCC2000–036(2000).

[27] T. Sjostrand, “High-energy physics eventgenerationwith Pythia 5.7 and Jetset7.4”,Comput.Phys.Commun.82 (1994)74.

[28] H.L. Lai et al., “Improved parton distribu-tions from global analysis of recent deepinelastic scatteringand inclusive jet data”,Phys.Rev. D 55 (1997)1280.

[29] GEANT DetectorDescriptionand Simula-tion Tool, CERN Program Library LongWriteupW5013(1994).

[30] LHCb RICH Technical Design Report,CERN/LHCC2000–037(2000).

[31] LHCb VELO Technical Design Report,CERN/LHCC2001–011(2001).

[32] S.Amatoetal, “LHCb muonchambergeom-etrysimulation”,LHCb 2001-033Muon.

[33] QQ- theCleoeventgenerator,§«¦¨©¤§§¤£ª¬§¦­®¯°«¨±²²

[34] G. Barrandet al., “GAUDI: Thesoftwarear-chitectureandframework for building LHCbdataprocessingapplications”,Proc.Interna-tionalConferenceonComputingin High En-ergy andNuclearPhysics,Padua,Italy, 2000,Ed. M. Mazzucato (INFN, Padua, 2000),p. 92.

[35] G. Corti et al., “LHC-B muontrigger”, NoteLHCb 97–001(1997);M. Borkovsky et al., “Study of the LHCbmuontrigger”, LHCb 1997–007Muon;

E. Aslanideset al., “An alternative architec-ture of the L0(µ) processor”,LHCb 1997–024Trigger;E. Aslanideset al., “The L0(µ) processor”,LHCb 1999–008Trigger.

[36] E. Aslanideset al., “A synchronousarchitec-ture for the L0 muontrigger”, LHCb 2001–010Trigger.

[37] E. Aslanides et al., “Performanceof themuontrigger algorithmwith the muonTDRdetector”,LHCb 2001-059Muon.

[38] G. Martellotti et al., “Monte-Carlo samplesandefficiency definitionfor themuonsystemoptimisation”,LHCb 2001–007Muon.

[39] E. Aslanideset al., “L0(µ) trigger perfor-mancewith the March 2000 muon systemlayout”, LHCb 2000–101Trigger.

[40] V. Boccietal., “Timeresolutionof themuondetector: consequencesfor trigger perfor-manceanddetectorsynchronisation”,LHCb2001–058Muon.

[41] E. Polycarpo and J.R.T. de Mello Neto,“Muon identification in LHCb”, LHCb2001–009Muon;J.R.T. deMello NetoandE. Polycarpo,“Up-date on muon identification”, LHCb 2001-061Muon.

[42] A.C. AssisJesuset al., “Multi variatemeth-odsfor muonidentificationatLHCb”, LHCb2001-084Muon.

[43] P. Ball etal., “B decaysattheLHCb”, ReportCERN–TH–2000–101.

[44] A. Ali, “Flavour ChangingNeutral Currentprocessesin B decays”,Nucl. Phys.B (Proc.Suppl.)59 (1997)86.

[45] H. Dijkstra et al., “Study of the sensitivityof the LHCb experimentto the rare decayB0

s ³ µ µµ ”, Proc.XXI Brazilian NationalMeeting on Particle and Field Physics,SaoLourenco, Brazil (2000).

75

[46] B. Bochin et al., “Beam tests of WPC-7 prototype of the wire pad chambersforthe LHCb Muon System”,LHCb 2000–102Muon.

[47] A. Kashchuket al., “Performancestudyof aMWPC prototypefor the LHCb Muon Sys-temwith theASDQ chip”, LHCb 2000–062Muon.

[48] D. Hutchcroftet al., “Resultsobtainedwiththe first four gap MWPC prototypecham-ber”, LHCb 2001–024Muon.

[49] B. Bochinet al., “Beamtestsof WPC-8andWPC-9prototypesof thewire padchambersfor the LHCb Muon System”,LHCb 2001–025Muon.

[50] W. Riegler, “Detector physics and perfor-mance - simulations of the MWPCs forthe LHCb muon system”,LHCb 2000–060Muon.

[51] O. Sasaki,“Amplifier-Shaper- DiscriminatorICs andASD Boards”,¨§¤¤¶· «¨« ¸¹º»¼¼£±

[52] F. Anghinolfi et al., “CARIOCA - a fastbi-nary front-endimplementedin 0.25 micronCMOS using a Novel current-modetech-nique for the LHCb muondetector”,LHCb2000–093Muon.

[53] T. Fergusonet al., “Ageing investigationofCMSEMU prototypechambers”,CMSNote1999/011

[54] V. Souvorov et al., “CMS cathode stripchamberageingtestat GIF”, CMS Note (tobepublished)

[55] P. Camarrietal. NIM A 414(1998)317.

[56] M. Abbresciaetal., NIM A 456(2000)143.

[57] S. Agosteo et al., “A facility for the testof large areamuonchambersat high rates”,NIM A 452,(2000)94.

[58] G. Aielli et al., “RPC ageingstudies”,Pro-ceedingsof the9th ViennaConferenceon In-strumentation,Vienna2001,(to bepublishedin NIM).

[59] M. Abbresciaetal., “Long termPerformanceof DoublegapResistive Platechambersun-der gammairradiation”, Proceedingsof the5th International Conferenceon Position-Sensitive detectors,London,Sept.1999

[60] R. Veenhof, “GARFIELD, a drift-chambersimulationprogram”,W5050(1999);¦¨§«¦¤©«¦©¤¥ ©±¤§£

[61] G. Auriemmaet al., “Design andConstruc-tion of the Wire Chambersfor the LHCbMuonSystem”,LHCb 2001–026Muon.

[62] G. Auriemma,D. Fidanzaand C. Satriano,“Testresultsof ChempirCorepanelsfor theMWPC of the LHCb muonsystem”,LHCb2001–008Muon.

[63] PSPICEanaloguecircuit simulator, version8.0x(1998).

[64] Seefor details:«±«¬± ½¦¨©¾«¿¼À¨©«¨Á ÂÃÃÄ«©¨¥© ª§

andlinks therein.

[65] G. Iaselli,Privatecommunication.

[66] Seefor detailsÅ«±«

Å«§¨¦«¦«¨¤Ä§and links

therein.

[67] A. Lai et al., “Muon Front-endArchitecture:anupdate”,LHCb 2001–030Muon.

[68] Protocoldescription:¦« ¬¨¦¦«¨£¨¦¦ «¤¦£±

[69] A. Lai etal., “Muon Front-endArchitecture”,LHCb 2000–017Muon.

[70] S.Cadeddu,A. Lai, “DIALOG: achipfor themuon detectorfront-endand logic”, LHCb2001–062Muon.

76

[71] ELMB technicaldocumentation: § «±¨¦«¤©¦«¸Æ®¸¹Ç¼°È»¹º¸²Æ¼¾Çº­¹®Éʹʫ£¤Ë§

[72] S. Cadeddu,L. Impagliazzo,A. Lai, “TheSYNCchipin themuonODEsystem”LHCb2001–063Muon.

[73] V. Talanov, “Estimationof absorbeddoselev-els at possiblelocationsfor LHCb detectorelectronics”,LHCb 2000–015Muon.

[74] “The LHCb ExperimentalControlsSystem”,§¦«¬¦«¤©¦«¦¨ª«¥ª¦¨©¨§«£¤± §

[75] F. Hahn,S. Haiderand R. Lindner, “LHCbMuon Gas systemproposal” LHCb 2001–023Muon.

[76] S. Haider, R. Barillere, “Proposal for GasControl Systemsfor the LHC experimentsgassystems”,«¦«¨«¦¤©«¦¶¦¨¬©¨¶¤¦Ç¸¹­¨©ª¨§§«£¤¥¦«©¨¨ §£±

[77] J. Andre et al., “LHCb Muon DetectorMe-chanical Support Structure”, LHCb 2001–032Muon.

[78] W. Bonivento,“Beampipeandshieldingopti-misationfor theLHCb muonsystem”,LHCb2001–028Muon.

[79] R. Bernieret al., “SPAC: SerialProtocolfortheATLAS Calorimeter”,ATL-AL-ES-00051999.

77