Upgrade tracker of the HL-LHC Sergio Dez Cornell, Berkeley Lab - PowerPoint PPT Presentation

Silicon strip staves and petals for the ATLAS Upgrade tracker of the HL-LHC Sergio Díez Cornell, Berkeley Lab (USA), On behalf of the ATLAS Upgrade strip tracker Collaboration HSTD-8, Taipei, Taiwan, Dec 5th-8th, 2011

Motivation: ATLAS Phase II Upgrade (HL-LHC)  Numerous challenges for silicon sensors on ATLAS Phase-II Upgrade  Higher granularity to keep same low occupancy  Higher radiation tolerance to deal with increased radiation environment  Novel powering solutions to power efficiently x7.5 more channels  Maintain low cable count to keep detector performance  Reduce cost per sensor to cover larger area (~ 200 m 2 )  Replacement of ATLAS Inner detector by an all-silicon tracker: Strips tracker : Si tracker (Utopia Layout) 3 layers of short strips (2.5 cm) 300 cm staves 2 layers of long strips (9.6 cm) staves 75 cm 10 disks of endcap petals 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 2

Stave concept layout and current prototypes Barrel strip stave (short strip version): 1.2 m 12 cm  Designed to minimize material Stave cross-section: • Shortened cooling paths Readout ICs Kapton flex hybrid Cu bus • Module glued to stave core with embedded pipes tape • No substrate or connectors, hybrids glued to sensors  Designed for large scale assembly Ti coolant tube • Simplified build procedure Carbon fiber facing  All components testable independently  Aimed to be low-cost Si Strip sensor Carbon honeycomb High T conductivity foam • Minimize specialist components Short strip module: “Stavelets”: • Stave prototype with 4 modules per side • 1 n-in-p strip sensor with • Single-sided stavelets (serial and DC-DC powered) 4 x2.5cm strips already built and under test at RAL [1] • 2 hybrids, each with 10 ABCN130 (256 ch) + 1 HCC/hybrid • Binary readout • Current prototypes: ABCN250 (128 ch/chip) + BCCs 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 3

Stave/petal powering  LV: Two powering distributions under study for n hybrids, each with current I 1 2 3 4 5 6 n-1 n  Serial powering • Total current = I Constant • Different GND levels per hybrid …… current • AC coupling of data lines source • Bypass protection required 1 2 3 4 5 6 n-1 n  DC-DC powering • Total current = n·(I/r * ) • Switching system …… Constant • Can be noisy voltage • + High mass source - *r = voltage conversion ratio  HV: Parallel power limited by cable reuse and/or material limitations  HV rad-hard switching for multiplexing under study recently (early stage) [2]  Current module and stave prototypes have proven to be a powerful test bench for the different powering options considered 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 4

Other components of the stavelet prototypes  Basic Control Chip (BCC) boards for data I/O (1 per hybrid)  AC coupled multi-drop system LVDS reception  Generates 80 MHz DCLK and handles 160Mb/s multiplexed data from each hybrid  Serial powering: Power Protection Board (PPB) [3] 39x6 mm 2  Fast response and slow-control bypass of modules within an SP chain  Allows alternate SP shunt circuits  Excellent performances demonstrated on SP stavelet All hybrids on Slow control disables odd hybrids  SPP ASIC submitted Aug 2011 V = 22.7 V, I = 5.09 A V = 12.7 V, I = 5.09 A  DC-DC powering: buck DC-DC converter AMIS4  Custom low-mass inductor and shield [4]  AMIS 4 ASIC: • Over current, over temperature, input under- voltage, and soft start state machine for reliable start-up procedure [5] 13x28 mm 2  New prototype circuits underway 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 5

Stave modules production and tools  Scalability for large scale production even at prototyping stage  Panelization of laminated hybrids • Designed for machine placement of passives and solder reflow • Tools developed for controlled gluing and wire bonding of ABCNs • Conservative design rules for high yield and volume, and low cost • Final hybrids testable on panels, ready for module assembly  Diverse tools developed for uniform gluing of hybrids to sensors • Numerous options investigated: glue spread on sensor or hybrid backplane, different glue stencils,… • Optimized glue thickness for best module performances: ~ 120 μ m  Automated wire bonding of ASICs to sensor and hybrids to test frames  Fully testable modules, ready for stave assembly 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 6

Stave module testing  PCB test frames: cheap and flexible test benches for testing  Different power configurations, G&S, added circuitry …  DAQ system for stave modules and stavelets: HSIO  Generic DAQ board (ATCA form factor) with single (large) Virtex-4 FPGA for data processing & connection to controller PC  Interface board: connectors & buffers for connectivity to FEE  Currently supports up to 64 streams (>64 streams with larger FX100 FPGA in future)  Upgraded sctdaq software  Allows standard 3ptGain, Response Curve, Noise Occupancy, DT Liverpool Noise,… on ABCN -250 modules  Expected noise performances for parallel, serial, and DC-DC powered modules  Similar ENC noise performances obtained at the different sites Berkeley, serial Freiburg, serial 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 7

Stave module construction and test  Numerous institutes involved in the construction and test of stave modules and stavelets [6] Up to 31 modules built so far (Nov 2011) 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 8

Proton irradiations of stave modules  Irradiated at CERN-PS  24 GeV proton beam scanned over inclined modules  Module biased, powered, and clocked during irradiation  Up to 2x10 15 cm -2 reached  Sensor and module behave as expected • Noise increase consistent with shot noise expectations Slide borrowed from T. Affolder, TIPP2011, June 2011 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 9

Stavelets  Stave prototypes with 4 modules per Power and PPBs side SP stavelet  Sensors directly glued to bus tape with EOS “soft” glue for easy module replacement board or removal  Key test bed for electrical testing  Powering, protection, G&S, …  Single-sided serial and DC-DC powered stavelets built and tested so far Custom Cu bus tape BCCs  SP stavelet tested with custom constant Power and Buck DC-DC converters current source (0-6A, OVP), excellent performances [7] DC-DC stavelet EOS board Custom Cu bus tape BCCs 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 10

Stavelet bus tape layout SP shield Layer (Al) SP Trace Layer (Cu) HV SP Current Return LVDS Clock/Command/Data & NTC 100 μ m track/gap over 40cm (1.2m) 11 For DC-DC, the power section of the SP tape is cut off and replaced by a custom section Slide borrowed from P. Phillips, 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 11 TWEPP2011,Sept2011

Electrical tests on stavelets  ENC noise close to noise on individual modules for both stavelets  Approximately ~ 20e higher in both cases  SP stavelet: PPB and bypassing hybrids does not affect noise performances  Double Trigger Noise clean at 1 and 0.75fC with appropriate current routing  Slightly better DT Noise performances at 0.5fC for DC-DC stavelet  Still work in progress [1] Serially powered stavelet H0 H1 H2 H3 H4 H5 H6 H7 Column 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ENC 661 623 628 675 650 636 697 760 687 646 640 666 680 661 624 656 DTN @1.0fC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DTN @0.75fC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DTN @0.5fC 130 40 1 58 3 1 255 1181 32 4 56 102 50 26 50 237 DC-DC powered stavelet H0 H1 H2 H3 H4 H5 H6 H7 Column 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 dENC 8 1 27 26 11 2 17 26 -10 -9 28 31 -26 -23 -2 -2 DTN @1.0fC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DTN @0.75fC 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DTN @0.5fC 0 1 6 36 18 5 12 38 12 2 4 9 0 0 0 4 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 12

Stave material estimates  Stave material estimates for 130 nm stave [8, 9] :  Based on as-built stavelets Adhesives %X0 3% Stave core 0.55% Stave core 28% Bus tapes 0.30% Modules Modules 1.07% Stave core 54% Tapes Bus tapes Module to stave adhesives 0.06% 15% Modules TOTAL 1.98% Module to stave adhesives  Titanium cooling tube: 2.2mm OD x 0.14mm wall  Tapes contribution could be significantly reduced (~50%) by removing Al screen + one glue layer: under investigation  Sensor dominates module material (~ 63%)  Power components will add 0.03 - 0.15 %X0, depending on power scheme (first approximation: changes in bus tape not considered) 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 13

Endcap petals: Petalet program  The endcap petal follows closely the barrel stave design  First petal cores already been produced  First endcap hybrids (ABCN-250 ASICs) produced and tested  Petalet prototype underway  Combines innermost radius sensors and region where petal splits in 2 sensor columns [10] Endcap hybrid “Petalet” 06/12/2011 S. Díez Cornell, HSTD-8, Taipei (Taiwan) 14