Activités 65nm / 130nm - Indico
Transcription
Activités 65nm / 130nm - Indico
Activités 65nm / 130nm Marlon Barbero - CPPM ATLAS-IN2P3 15 octobre 2013 Plan • Question de la tenue aux radiations des pmos étroits en 65nm. • Nouveaux tests sous irradiation. • Organisation de l’activité au sein de RD53. • Priorités ATLAS-France 65nm/130nm pour 2014. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 2 Dose effects on Shift register Tests carried out on the ATPIX65 chip Small array of pixels 16x32 Designed at Berkeley with the 65 nm CMOS process PS tests : 24 GeV protons (SEU tests) At high levels of dose, failures appear from 400 MRad in the shift register Tests with 24 GeV proton beam SEU tolerant memory design for the ATLAS pixel readout chip CERN Xray Tests : 10 keV The chip was placed to 1cm of the source Tests with Xray beam The dose rate is 6.24 Mrad/hour. A dose level of 950 Mrad reached after one week and a half The same behavior as for the proton tests DFF Failures appear from a dose level of 400 Mrad 3 CPPM’s previous studies Rows 0-255 Rows 0-255 • Test SEU in LBNL-dvp’ed 65nm proto (CERN PS) Observed localized bits stuck in shift register used for pixel config. loading. 390 MRad to 420 MRad 260 MRad to 310 MRad • Hypothesis: TID effect / Pattern 1111 Pattern 1111 leaking. • Narrow pmos dose effect. Columns 0-15 ATLAS-France – 65nm studies and RD53 Columns 0-15 4 Banc de test pour caractérisation de transistors Test board: - 4 layers, - class 4 (insulation min. 150µm) - chip package LCC44 cts (not a good idea) - better in JLCC44 - delay: 12 days ( wait for it for the week 31) - components already received Supply for the commands of ESD protections Triaxial2coax connexions for IV measurements (Vhigh, Vlow) KEITHLEY 2634B (sourcemeter) 6 KEITHLEY 7002 (switches matrix) slots available (Keithley 7158) (each slot contains 10 inputs to 2 outputs) GPIB links NVDD, PVDD, VSS LabView software ATLAS-France – 65nm studies and RD53 5 Etude d’effet de dose au CERN • Irradiations sous X au CERN du 7 au 16 octobre transistor pmos de petites tailles ( tenue aux radiations des librairies digitales?). • Quelques problèmes techniques sur cartes de test véritable démarrage mercredi matin. • >700MRad en ce moment. ATLAS-France – 65nm studies and RD53 6 Ids (A) Résultats 300MRad D012 nmos :120nm/60nm (@Vds=50mV) ATLAS-France – 65nm studies and RD53 7 (V) Vgs Ids (A) PD012 pmos :120nm/60nm (@Vds=50mV) Vgs (V) ATLAS-France – 65nm studies and RD53 8 D024 nmos :240nm/60nm (@Vds=50mV) Ids (A) Ids (A) Résultats 300MRad PD024 pmos :240nm/60nm (@Vds=50mV) Vgs (V) Vgs (V) • Maintenant: Finalisation d’irradiation 1GRad? Démarrage de la période d’annealing. • Bilan dans quelques semaines. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 9 RD53 • Nouveau groupe de RD approuvé par LHCC (fin juin). • “Development of Pixel Readout Integrated Circuits for Extreme Rate and Radiation” == 65nm ATLAS / CMS. • Période initiale de 3 ans. • 2 spokemen: Christiansen (CERN/CMS) & Sciveres (LBNL/ATLAS). • Instituts: Bari / Bonn / CERN / CPPM / Fermilab / LBNL / LPNHE / NiKHEF / New Mexico / Padovia / Pavia & Bergamo / Pisa / Perugia / PSI / RAL / UCSC / Torino (~100 auteurs) • 6 Groupes de Travail / Thématiques: – WG1: Radiation (qualification de la techno pour 1GRad / 1016 neq.cm-2, évaluation des librairies digitales après irradiation, modèle transistors après radiation). Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 10 RD53 -suite– WG2: Top Level Design (méthodologie, vérification -ASIC contenant 0.5Gtrans!-, analog integration, power distrib., clock distrib. etc…). – WG3: Simulation Test Bench (system verilog simulation, optimization -global/regional/pixel architecture, lien avec les événements physiques…). – WG4: I/O (interface definition, protocols, blocs de readout et de contrôle…). – WG5: Analog Design (évaluation de différents ampli / d’ADC techniques -ToT, shared ADC…-). – WG6: IP Blocks (definition des besoins, librairie de blocs IP, s’accorder sur les specs communes, …). • Groupes Français: CPPM (WG1-MB convener-/WG6) - LPNHE (WG1/WG3/WG6) - LAL/Orsay (en cours d’approbation) Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 11 LPNHE RD53 LPNHE est engagé à participer à la collaboration. Une attention particulière à la possibilité d'utiliser des technologies telles que les mémoires associatives pour le formatage / compression des données Tracking trigger upgrade ATLAS: FTK LPNHE est impliqué dans le développement des puces de mémoire associative 65 nm (leadership en collaboration avec INFN) 2012/2013 – mini@sic prototype 2013/2014 – AMchip05: ~12 mm2 MPW prototype 2014/2015 – AMchip06: puce pour installation finale de FTK AIDA LPNHE est impliqué dans la réalisation du common blocks a 65 nm (WP3): - readout logic blocks - SET tolerant logic Et dans les sub-projects (WP3): - “Readout ASICs in 65nm technology interconnected using the CEA-LETI or EMFT process” (LAL/LAPP/LPNHE/MPP) - “Interconnection of ATLAS FEI4 chips to sensors using SLID interconnection and ICV (high density TSVs) from EMFT” (MPP/GLA/LAL/LIV/LPNHE) mini@sic prototype reçu en Juin, en cours de test, premiers résultats très satisfaisants LAL/OMEGA participation in RD53 LAL/OMEGA decided to joint RD53. Request to RD53 IB was sent 16 september 2013 AIDA overlap: common blocks development (one block of the list: ramp generator -pour ADC Wilkinson-, 10bit DAC, analog probe buffer, PLL 40-500 Mhz) Common work with CERN on IP blocks on-going AIDA: 65 nm chip for EMFT(SLID) prototype to be decided. Request for 2014: 23 kEuros Blocks CPPM • SEU Tolerant Memories (physical node separation, TRL, error correction….). • Analog MUX + 10/12-bit DAC (avec designer LAPP Annecy). • SAR ADC. • Comparator. • Pixel charge ADC (avec designer LPSC Grenoble). • Bandgap Voltage Reference. • Temperature Sensor. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 14 65nm requests 2014 • CPPM: – Design test transistors, blocks… – 23k – Irradiation and tests – 10.5k – Licences – 1.5K • LAL: 23k. – OmegaPix. – Block dvp. – puce 65nm pour EMFT (SLID). • LPNHE: 15k. – 5k production IC (au sein de RD-53???). – 10k contribution pour puce OmegaPix. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 15 Manpower • LAL omega: A. Lounis, G. Martin-Chassard, D. Thienpont, J. Tongbong. • LPNHE: G. Calderini, M. Bomben, F. Crescioli (D), J.F. Genat (D), O. Le Dortz (D), G. Marchiori. • CPPM: P. Breugnon, M. Menouni (D), D. Fougeron (D), J. Liu (PhD), M. Barbero, A. Rozanov, P. Pangaud (D), S. GodiotBasolo (D), A. Wang (D), F. Gensolen (D) Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 16 Résumé: en ce moment • RD53 ATLAS + CMS ; organisation de / momentum pour l’activité 65nm (CPPM / LPNHE / LAL impliqués). • Premier but: Validation de la technologie 65nm aux doses du futur HL-LHC. • Pour le moment: Questions sur la tenue aux radiations des pmos étroits en TSMC 65nm. • Irradiation d’une ASIC de test sous rayons X. • 2nd meeting RD53 WG1 la semaine prochaine. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 17 Conclusion • Priorités 2014: – Irradiations transistors 65nm TSMC Xray / protons (PS fin 2014???) question de la tenue des librairies digitales. – Irradiation autres technologies 65nm. – Travail sur design blocks 65nm. • Financement 2014: – Demandes IN2P3: CPPM 35k / LAL 23k / LPNHE 15k. – Comment utiliser l’argent d’Aida WP3.3? Seulement pour design blocks? (caractérisation de transistor de tests??? matrices de pixels???) Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 18 backup Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 19 65nm irradiation (CPPM/Mohsine) • Narrow PMOS dose effect. Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 20 SEU tolerant memory A.Rozanov IN2P3 15.10.2013 ESD study A.Rozanov IN2P3 15.10.2013 Bandgap reference A.Rozanov IN2P3 15.10.2013 Generic ADC for monitoring A.Rozanov IN2P3 15.10.2013 Temperature sensor A.Rozanov IN2P3 15.10.2013 A.Rozanov IN2P3 15.10.2013 A.Rozanov IN2P3 15.10.2013 A.Rozanov IN2P3 15.10.2013 A.Rozanov IN2P3 15.10.2013 Sensors • Probably different in different layers • Planar, 3D, diamond – classical sensors (realistic, but high costs of sensors and bumpbonding) • CMOS sensors (HV CMOS, HR CMOS, T3 CMOS) • In case of 2D MAPS 65nm design can be reused for digital part A.Rozanov IN2P3 15.10.2013 Pixels chip • Building blocks: connection with AIDA, but specific pixel specification • Reliability, SEU tolerance • High rate • Large chip 2x2 cm, 1 billion transistors • Complex: close to Pentium processor, 256K channels of DAQ system per chip • Low power (0.3 W/cm2) • CERN frame contract with TSMC and IMEC, design tool set, libraries, design exchange (including AIDA blocks) A.Rozanov IN2P3 15.10.2013 Pixel chip generations Generation Current FEI3, PSI46 Phase 1 FEI4, PSI46DIG Phase 2 Pixel size 100x150um2 (CMS) 50x400um2 (ATLAS) 100x150um2 (CMS) 50x250um2 (ATLAS) 25x100um2 ? Sensor 2D, ~300um 2D+3D (ATLAS) 2D (CMS) 2D, 3D, Diamond, MAPS ? Chip size 7.5x10.5mm2 (ATLAS) 8x10mm2 (CMS) 20x20mm2 (ATLAS) 8x10mm2 (CMS) > 20 x 20mm2 Transistors 1.3M (CMS) 3.5M (ATLAS) 87M (ATLAS) ~1G Hit rate 100MHz/cm2 400MHz/cm2 1-2 GHz/cm2 Hit memory per chip 0.1Mb 1Mb ~16Mb Trigger rate 100kHz 100KHz 200kHz - 1MHz Trigger latency 2.5us (ATLAS) 3.2us (CMS) 2.5us (ATLAS) 3.2us (CMS) 6 - 20us Readout rate 40Mb/s 320Mb/s 1-3Gb/s Radiation 1MGy (100Mrad) 3.5MGy (350Mrad) 10MGy (1Grad) Technology 250nm 130nm (ATLAS) 250 nm (CMS) 65nm Architecture Digital (ATLAS) Analog (CMS) Digital (ATLAS) Analog (CMS) Digital Buffer location EOC Pixel (ATLAS) EOC (CMS) Pixel Power ~1/4 W/cm2 ~1/4 W/cm2 ~1/4 W/cm2 A.Rozanov IN2P3 15.10.2013 3rd generation pixel architecture • • • 95% digital (as FEI4) Charge digitization ~256k pixel channels per chip • • Pixel regions with buffering Data compression in End Of Column A.Rozanov IN2P3 15.10.2013 CPPM contributions to 65nm chip • Test 65nm pixel matrix designed by Abder Mekaoui at LBNL, pixel size 25x125 um, 3x4 mm, 16x32 pixels • Pixel configuration: Triple Redounded Latch (TRL) 12.5µm × 4.3µmwith error correction. Data loaded with shift register (SR) DFF 6µm × 2.4µm • Irradiation at Los Alamos, CERN PS (CPPM), X-ray CERN (CPPM) A.Rozanov IN2P3 15.10.2013 SEU test results • DFF in Shift register Pattern 0000 Pattern 1111 All patterns Pattern 0000 Pattern 1111 Rows 0-255 Rows 0-255 All pattern s • TRL memory Columns 015 Columns 015 Columns 015 Columns 0-15 A.Rozanov IN2P3 15.10.2013 Columns 0-15 Columns 0-15 SEU test results • DFF 260-310 Mrad • DFF 390-410 Mrad A.Rozanov IN2P3 15.10.2013 Proton PS 24 GeV tests • Test 16 columns of 256 TRL and 256 DFF. Design from ARM standard library. • σTRL = 2.6 10-16 cm² • σDFF = 4.5 10-14 cm² σDFF / s TRL = 170 After 400 Mrads the state of some DFF cells remain at 1 A.Rozanov IN2P3 15.10.2013 Narrow pmos dose effect • High increase of the threshold voltage, transistor completely off at 850MRad X-ray A.Rozanov IN2P3 15.10.2013 Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 39 Marlon Barbero CPPM, ATLAS-France 15 octobre 2013 40 Organization • 17 institutes, 100 authors (50 designers) • Two co-spokepersons: Maurice Garcia (ATLAS), Jorgen Christiansen (CMS) • Institutes Board • Initial plan for 3 years • May finish in common ATLAS-CMS chip, but exchange of experience and blocks will already big success • Next IB 4 October 2013 at CERN after Aix-LesBain HL LHC workshop A.Rozanov IN2P3 15.10.2013 Working groups WG Domain WG1 Radiation test/qualification Qualification of technology to 10 MGy TID, 1016 n.eq./cm2. Transistor simulation models after irradiation. Evaluation of logic cell libraries after irradiation. Expertise on radiation effects in 65nm WG2 Top level design Design methodology, verification and test of ~5 × 108 transistor IC. Analog integration in large digital chip. Power distribution Synthesis constraints. Clock distribution and optimization WG3 Simulation and verification test bench System Verilog simulation and Verification framework. Optimization of global architecture/pixel regions/pixel External system and external physics data. Verification of test chips and evolving designs WG4 I/O Definition of readout and control interfaces (e.g. LPGBT). Definition of standardized I/O protocols and performance Implementation of readout and control interface blocks. Standardized interfaces: Control, Readout, etc. WG5 Analog design Evaluate and compare alternate amplifier designs. Evaluate and compare charge ADC techniques vs. number of bits (TOT, shared ADC, etc.) WG6 IP blocks Define common requirements for IP block design. Evaluate, document, and keep library of IP blocks Generate overview and recommendations. Each block will have its own prototyping milestones A.Rozanov IN2P3 15.10.2013 Participation matrix Institute WG1 Radiation WG2 Top level Bari C Bergamo-Pavia A Bonn C A CERN B(*) CPPM WG3 Sim./Ver WG4 I/O WG5 Analog A WG6 IPs A C A B A B B A (*) A C(*) A B(*) A B C C B A Fermilab A B LBNL B A B LPNHE Paris A B A A A A A NIKHEF New Mexico A Padova A Perugia B Pisa PSI RAL B A A A A B B A A A B A C B B Torino C B C UCSC C B C A: Core competency, B: High interest, C: Ability to help B A.Rozanov IN2P3 15.10.2013 A A A A C A A A (*): General CERN support for 65nm Conclusions • 3D Tezzaron-GF on hold due to financial constrains, wait and see • CMOS sensor RD on good track (chip 3D techno). Potentially may bring very large cost saving and increase in physics reach • ATLAS and CMS give good example of cooperation in order reduce costs by RD53 in technology 65 nm. Start with TSMC, but radiation studies will be critical for final choice. • RD53 compatible with all sensors, including HV CMOS A.Rozanov IN2P3 15.10.2013