La partecipazione del gruppo di Bologna (luminometro) alle attivita’ di ATLAS A. Bertin, M. Bruschi, S. De Castro, L. Fabbri, P. Faccioli, B. Giacobbe, F. Grimaldi, I. Massa, M. Piccinini, M. Poli, C. Sbarra, N. Semprini-Cesari, R. Spighi, M. Villa, A. Vitale, A. Zoccoli M. Bruschi-CSN1 Napoli 21 Settembre 2005 1 Contenuto Le attività del gruppo di Bologna in ATLAS: Test delle schede di elettronica del LVL1 per le camere a Muoni Luminometro (LUCID): elettronica di lettura e di trigger + simulazioni Presentazione del progetto L’elettronica di LUCID e stime (preliminari) dei costi Altri interessi del gruppo: Partecipazione alla fisica & trigger di alto livello: fisica di alto Pt e diffrattiva Partecipazione al computing M. Bruschi-CSN1 Napoli 21 Settembre 2005 2 Attività sul LVL1 a Bologna Il test delle schede di elettronica del LVL1 (prodotte da Roma1) per le camere a Muoni, verrà effettuato a Bologna. Inizio: Test di: Settembre 2005. ~800 Pad-OR ~800 mother boards con test JTAG e ELMB Rate di test richiesto: 7 board/day Tempo previsto per il test del sistema: >6 mesi M. Bruschi-CSN1 Napoli 21 Settembre 2005 3 Attività sul LVL1 a Bologna - II Prima riunione e trasporto del materiale il 7 Settembre. La scorsa settimana testate le prime 20 schede OR Le rimanenti schede OR (~800) in viaggio verso Bologna. Inizio lavoro sistematico sull’intero sistema (OR+mother) ad inizio Ottobre. 3 persone dedicate full-time al lavoro (2 su fondi universitari) + altre a rotazione. Schedula dettagliata del test a Ottobre. M. Bruschi-CSN1 Napoli 21 Settembre 2005 4 ATLAS -LUMINOSITY Importance of Luminosity measurements: Cross sections for “Standard “ processes t-tbar production W/Z production …… Theoretically known to better than 10% ……will improve in the future New physics manifesting in deviation of x BR relative the Standard Model predictions Important precision measurements Higgs production x BR tan measurement for MSSM Higgs …… M. Bruschi-CSN1 Napoli 21 Settembre 2005 5 ATLAS –Luminosity (cont.) Some examples Higgs coupling Relative precision on the measurement of HBR for various channels, as function of mH, at Ldt = 300 fb–1. The dominant uncertainty is from Luminosity: 10% (open symbols), 5% (solid symbols). tan measurement Systematic error dominated by luminosity (ATLAS Physics TDR ) (ATLAS-TDR-15, May 1999) M. Bruschi-CSN1 Napoli 21 Settembre 2005 6 ATLAS Luminosity Measurement Program • Relative luminosity a DEDICATED luminosity monitor is needed LUCID LUCID will provide the luminosity per BX as well • Absolute luminosity – Goal: • measure L with ≲ 2-3% accuracy – How: • LHC Machine parameters • Use ZDC in heavy ion runs to understand machine parameters • rates of well-calculable processes:e.g. QED, QCD • optical theorem: forward elastic rate + total inelastic rate: Roman Pots – needs ~full |η| coverage-ATLAS coverage limited – Use tot measured by others (TOTEM) – Combine machine luminosity with optical theorem • luminosity from Coulomb Scattering Roman Pots ATLAS pursuing all options M. Bruschi-CSN1 Napoli 21 Settembre 2005 7 Motivations for LUCID Requirements: A very radiation hard detector to be used as luminosity monitor Good time resolution to resolve individual beam crossings Insensitive to soft background particles Pointing capability A large dynamic range and no saturation at the highest luminosity A simple, robust and cheap construction Solution: LUCID: LUminosity measurement using a Cherenkov Integrating Detector - The design is based on the Cherenkov Luminosity Counter (CLC) that is operating successfully at CDF. - Gas filled tubes around the beampipe act as a Cherenkov detector and detects particles from the I.P. that are above the Cherenkov threshold (2.7 GeV for pions and 9 MeV for electrons) M. Bruschi-CSN1 Napoli 21 Settembre 2005 8 Basics of the detector 2 detectors x 200 Al tubes filled with C4F10 or Isobutane at atmospheric pressure Beampipe Winston cones at the end of the tubes focus the Bruschi-CSN1 Napoli Cherenkov light onto M.quartz fibres 21 Settembre 2005 9 The fibre read-out M. Bruschi-CSN1 Napoli 21 Settembre 2005 10 General Considerations-I The purpose of this talk is to describe a possible baseline for the design of the ATLAS luminometer (LUCID) readout electronics and trigger scheme. • • • • Our aim (since last June) is to achieve, as soon as possible, the following points: Define the general scheme of the electronics Tune, by MC simulation and test beam, the final design parameters Provide a cost estimate per readout channel and time schedule for the realization of the electronics Start as soon as possible with the design in order to be ready in 2007 M. Bruschi-CSN1 Napoli 21 Settembre 2005 11 General Considerations-II Main Goals of the LUCID electronics: 1. For each triggered event (ROD level): Number of tracks Tracks time of arrival 2. Monitor level Number of tracks per bunch Tracks time of arrival per bunch 3. Trigger Level Provide a fast trigger on “properly” filled bunch or on-time events Provide a RAP-GAP vetoing for forward physics Strategy: exploit available FED solutions where possible M. Bruschi-CSN1 Napoli 21 Settembre 2005 12 Known Facts Signal from IP Background from sec. vtx Background from particles crossing fibers •Reject off - BX background sources (part of beam gas interactions, satellites BX, interactions originating off-IP) •Provide a detailed BX structure monitor •Guarantee the selection of events in time with the readout electronics of all the ATLAS detector IMPORTANT TOOL FOR THE WHOLE ATLAS DATA TAKING 1) Amplitude meas. (2+1 bit @ L1) 2) Signal time-of-arrival measurement (1 bit @L1) 3) Hit fibers pattern (7x2 bit @L1) • High Occupancy ~30% (at max. luminosity) • Max. 3 tracks/tube (at max. luminosity) •The amount of information to be handled @L1 can be encoded in 18 bit •MAPMT and FE are in a low level radiation area for electronics (5 Gray/year) M. Bruschi-CSN1 Napoli 21 Settembre 2005 13 M. Bruschi-CSN1 Napoli 21 Settembre 2005 14 System Architecture VME BUS: TTCvi, CTRL sign.,etc FRONT END (FEPCB) 22x2 M A P M T Similar to Roman Pot FE: OPERA/MAROC chip Input: MAPMT Output: DIGITIZED INFORMATION on LVDS Links (~0.5 Gb/s) L V D S L I N K S ROD T R I G G. ROD ROD C A R D ROD PC HV FE CONTROL M. Bruschi-CSN1 Napoli 21 Settembre 2005 ROS 15 THE OPERA/MAROC CHIP BLOCK FUNCTIONALITY DIAGRAM 9 1 Photons Preamp. Photomultiplicator Bipolar Fast Shaper 63 Unipolar Fast Shaper Gain correction 6 Bits n=0..5) SUM_OUT 9 Signal current outputs (sum over 7 out) MUX_OUT 1 Multiplexed current output (for channel monitor) TRIG_OUT Trigger 63 Trigger outputs x 2 bit/output/BX on 9 LVDS TX (2n-4, 4 discriminator thresholds 4 x 10 bits DACs Modifications for Lucid in blue text M. Bruschi-CSN1 Napoli 21 Settembre 2005 16 OPERA/MAROC chip : LAYOUT – Technology: SiGe 0.35 m – Submitted mid-June, expected mid September – Chip area : 12 mm2 (3.5mm *3.9mm) – 64 channels, 3.5V power supply – Power consumption : 350 mW – 228 pins –A lot of flexibility: • Gain adjustment per channel (6 bits) • 4 thresholds • Multiplexed currentmeasurement 1 tube 7 readout channel 1 mapmt 9 tubes Output for LUCID: 9 current (sum over 7 channels) 1 current (multiplexed for channel control) 63 x 2 bit (80 MHz clock, for trigger) M. Bruschi-CSN1 Napoli 21 Settembre 2005 17 1/9 data from the MAROC CHIP Single Tube Readout Unit (7 channels) LVDS S/P 2 2 #1 2 1 2 TUBE 2 Fan Out Discr. (Prog. Thr +NR) MAROC CHIP SUM_OUT LHC Clock TDC START STOP TDC 2 Window Programmable Comparator GI 8 Multiplicity + per Tube ADC LUT 1 2 LUT #7 18 (260 kB) 3 to the trigger unit 3 ADC GATE STRU 25 ns LHC Int. Time time ADC GATE 8 2 15 ns int 10 ns reset TDC START time RAW DATA TO READOUT per STRU ~ 5-6 Bytes/BX M. Bruschi-CSN1 Napoli 21 Settembre 2005 18 LUCID ROD BOARD (22 units + spares) – VME 9U s-LINK Busy to ROS SUM_OUT 1_1 SUM_OUT 1_2 160 MB/s ~200 Bytes/ev s-LINK 6 Bytes stru 1 stru 2 Analog_In 1 6 Bytes Analog_In 2 from CTRL LOGIC from CTRL LOGIC stru 20 SUM_OUT 2_10 LVDS 1_1 LVDS 1_2 LVDS 2_9 EVENT BUFFER DPRAM DPRAM 6 Bytes Analog_In 20 VME I.F LVDS_In 1 LVDS_In 2 LVDS_In 18 LVDS 2_1 (to trig. unit) 3 3 LVDS S/P LVDS 2_2 (to trig. unit) opt. lnk from TTCIX VME P1 LVDS S/P LVDS 1_1 (to trig. unit) LVDS 1_2 (to trig. unit) DPRAM TTCRQ i.f. 3 CTRL LOGIC M. Bruschi-CSN1 Napoli 21 Settembre 2005 19 LUCID TRIGGER BOARD (1 unit + spares) – VME 9U s-LINK Busy to ROS 160 MB/s 1 LVDS 1_1 43 D e t e c t o r 44 1 2 LVDS 1_2 LVDS 22_1 LVDS 22_2 1 43 D e t e c t o r 44 2 LVDS 23_1 2 LVDS 23_2 LVDS 44_1 LVDS 44_2 opt. lnk from TTCIX to the L0 trigger s-LINK ~200 Bytes/ev Signal Buffer 44 ser. Inp 594 bit/BX 44 ser. Inp 594 bit/BX FPGA based TRIGGER PROCESSING UNIT VME I.F VME P1 Signal Buffer TTCRQ i.f. CTRL LOGIC M. Bruschi-CSN1 Napoli 21 Settembre 2005 Algorithm: MC simulations are needed 20 Description of the main building blocks of the readout electronics OPERA MAROC CHIP • Adapted to LUCID needs from the RP design (ATLAS Orsay group) Gated Integrator + ADC • 8 bit for the total sum should be enough • Contacts have been taken with the LHCb preshower group (Clermont) for adapting their GI+ADC solution • • • TDC 25 ns full range; 300 ps MAPMT resolution 7-8 bit CERN HPTDC will be used (32 channel at 2MHz 2 channels at 32 MHz, but still convenient) • • • LOGIC Mainly Based on LUT (but still to be optimized) IMPLEMENTED on FPGAs Flexible and robust ENGINEERING • Integration has to be studied but standard VME 9U will be probably preferred M. Bruschi-CSN1 Napoli 21 Settembre 2005 21 First Costs Estimate Item Q.ty Units MAPMT MAROC 2 unit/ROD 2 unit/ROD UNIT COST TOTAL COST (EURO) (EURO) MIN MAX MIN MAX 1600 1700 200 400 Comments 3200 400 3400 800 a multiproject of 20 keuro (shared?) is assumed for 50 chip 400 100 360 200 1060 400 suhner gx 02272 + SMB connectors 100 2x20 twisted pair cable 360 200 1060 150 10 35 100 200 50 75 100 20 20 500 100 350 100 200 50 75 1500 360 80 3315 750 purchased directly by LHCb outlet (2500 chip) 200 investigate LHCb PS FE solution 350 CERN HPTDC 100 200 50 75 2000 360 80 4165 100 100 200 200 50 50 75 75 1500 1500 100 200 50 75 1500 1925 100 200 50 75 1500 1925 FRONT END (FE) coax cables (incl. connectors) flat cables (incl.connectors) LVDS TX CTRL FPGA TOTAL FRONT END (MATERIAL)/ROD 20 1 18 1 25m/ROD 25m/ROD unit/ROD unit/ROD 20 100 20 200 20 100 20 200 READ-OUT BOARD (ROD) LHCb PS Gated Integrators 8 bit ADC 40 MSPS TDC TTCRQ S-LINK (incl. cable) VME IF FPGA CTRL LOGIC FPGA STRU LUT,DPRAM,EVENT BUFF. FPGA LVDS RX LVDS TX TOTAL ROD (MATERIAL) 5 20 10 1 1 1 1 20 18 4 unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD unit/ROD 100 5 35 100 200 50 75 75 20 20 TOTAL ITEM MAPMT MAROC COSTS SUMMARY UNITS PROD MIN MAX FACTOR (EURO) (EURO) 50 80000 85000 50 10000 20000 FE ROD TB 3 25 2 VME+PC 1 TOTAL 2 2 2 6360 165750 7700 6360 208250 7700 40500 40500 310310 367810 TRIGGER BOARD (TB) TTCRQ S-LINK (incl. cable) VME IF FPGA CTRL LOGIC FPGA TRIGGER PROCESSING UNIT FPGA TOTAL TB 1 1 1 1 1 unit/TB unit/TB unit/TB unit/TB unit/TB VME+PC VME 9U CRATES VME CPU PC TOTAL SYSTEM 3 unit 3 unit 2 unit 7500 7500 4000 4000 3000 3000 22500 22500 2+1 spare 12000 12000 2+1 spare 6000 6000 40500 40500 M. Bruschi-CSN1 Napoli 21 Settembre 2005 22 Tentative Time Schedule 2005 2006 2007 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 FE CARDS (3 units) Tests on the available version of the MAROC chip Design (incl. mechanics) frozen Layout submitted (company will provide fully mounted prot+prod.) Prototype board received Prototypes tested Electronics ready (full production) Electronics installed Installation tested Electronics commissioned ROD CARDS (25 units) Design (incl. mechanics) frozen Layout submitted (company will provide fully mounted prot+prod.) Prototype board received Prototypes tested Electronics ready (full production) Electronics installed Installation tested Electronics commissioned TB (Trigger Board) (2 units) Design (incl. mechanics) frozen Layout submitted (company will provide fully mounted prot+prod.) Prototype board received Prototypes tested Electronics ready (full production) Electronics installed Installation tested Electronics commissioned M. Bruschi-CSN1 Napoli 21 Settembre 2005 23 Time Profile Goals of the project Detector ready beginning 2007 FED and ROD ready in 2007 in useful time before the start of LHC operation Trigger Board before the end of 2007 Funding Profile according to these goals: 2006 ITEM % 2007 kEuro % kEuro MAPMT MAROC 100 100 85.0 20.0 0 0 0.0 0.0 FE ROD TB 100 20 0 6.4 41.7 0.0 0 80 100 0.0 166.6 7.7 0 0.0 100 40.5 VME+PC TOTAL 153.0 214.8 M. Bruschi-CSN1 Napoli 21 Settembre 2005 24 Next Steps: System Optimization (Test Beam & MC) 1. Study the best solution for the tube readout (gas pressure, number and type of fibers per tube, MAPMT type) Test Beam at DESY October Main goals of the Test beam (1/10-14/11): Photon generation, transfer and losses Generation, processing and transmission of electrical signals Baseline Readout Device: MAPMT H 7546 (64 ch-UV glass win.) 2. MC Study to refine design parameter (TDC & ADC resolution), occupancy (# of channels), trigger algorithm M. Bruschi-CSN1 Napoli 21 Settembre 2005 25 TEST BEAM: the MAPMT-FIBERS connection M. Bruschi-CSN1 Napoli 21 Settembre 2005 26 Richieste finanziarie 2006 (variazioni rispetto ai moduli in colore violetto) Richieste finanziarie 2006: M.I. : 21 k€ metabolismo M.E. : 142.5 k€ metabolismo + test beam + C&I Consumo: 21 k€ metabolismo (+5 k€ per l`attivita` sul LVL1) Inventario: 17.5 k€ farm di computing per analisi (Tier 3 like) in comune con il gruppo BO-RPC (responsabile per I 2 gruppi: F. Semeria) Costruzione apparati: 270k€ (<350k€ ) per elettronica LUCID Comprendenti PM, Cavi, crates etc. Trasporti: 3 k€ (per l`attivita` sul LVL1) M. Bruschi-CSN1 Napoli 21 Settembre 2005 27 Conclusions • The Bologna Group activities in ATLAS are started I many different areas. • The group will test the LVL1 boards (~830 Pad-OR and mother boards) in Bologna. The setting-up is started. The systematic work will begin in October. • The group is involved in the LUCID detector (lumi monitor). Main activities are concerning: – MC simulations – Electronic design M. Bruschi-CSN1 Napoli 21 Settembre 2005 28 BACKUP M. Bruschi-CSN1 Napoli 21 Settembre 2005 29 Conclusions - II • We are developing, based on the present knowledge of the detector, a baseline design of the LUCID electronics • We consider this baseline being in a quite advanced stage and capable to fulfill the detector requirements (more: with the TDC options LUCID become an IMPORTANT DETECTOR for the ATLAS DATA TAKING) • Forthcoming test beam and MC simulation will be valuable inputs to improve (in case, simplifying) the design • We have already solutions at hand for critical devices (frontend chip, G.I.,ADC, TDC) that make us confident for a start of data taking together with LHC M. Bruschi-CSN1 Napoli 21 Settembre 2005 30 More infos - Groups involved in LUCID: University of Alberta, University & INFN Bologna, CERN, University of Lund, University of Montreal, Max Planck Institute, University of Manchester (?), SACLAY Italy would represent ~50% of the group - Total cost of the project for INFN : ~ 400 k€ Would represent ~50% of the total cost M. Bruschi-CSN1 Napoli 21 Settembre 2005 31 General Considerations-III • For the description of the readout electronics I will refer essentially to the baseline of the detector described in the LOI • Detector: formed by two parts each one consisting of 200 Cherenkov counters (tubes) 5 layers/section x 40 tubes/layer x 7 fibers/tube x 2 sections = 2800 fibers • Signal: Prompt particles coming from the IP (primaries) will traverse the full length of the counter and generate a large amplitude signal in the photo-detector • Background I: Particles originating from secondary interaction of prompt particles in the detector material and beam-pipe (secondaries) are softer and will traverse the counters at larger angles (multiple reflections), with shorter path lengths Background I significantly smaller than signal Background II: Particles crossing the readout fibers will produce light only on the crossed fibers Background II will have different pattern of hit fibers wrt signal • M. Bruschi-CSN1 Napoli 21 Settembre 2005 32 Signal amplitude measurement • CLC have the important feature to guarantee a proportionality between the number of primary particles traversing a single tube and the resulting signal amplitude. • These detectors response is not subjected to Landau fluctuations (present in scintillators) and the counter’s amplitude distribution will show distinct peaks for the different particle multiplicities hitting the counters. LUCID, with an appropriate readout and trigger system can provide the Total Tracks multiplicity per BX M. Bruschi-CSN1 Napoli 21 Settembre 2005 33 Signal time-of-arrival measurement • • • A “precise” measurement of the arrival time of a track in the LUCID detector will help to: Reject off - BX background sources (part of beam gas interactions, satellites BX, interactions originating off-IP) Provide a detailed BX structure monitor Guarantee the selection of events in time with the readout electronics of all the ATLAS detector IMPORTANT TOOL FOR THE WHOLE ATLAS DATA TAKING Hit fibers pattern Important to reject Background (essentially of type II) M. Bruschi-CSN1 Napoli 21 Settembre 2005 34 M. Bruschi-CSN1 Napoli 21 Settembre 2005 35 M. Bruschi-CSN1 Napoli 21 Settembre 2005 36 Location of the detector Situation when the forward shielding is removed: M. Bruschi-CSN1 Napoli 21 Settembre 2005 37 M. Bruschi-CSN1 Napoli 21 Settembre 2005 38