Single-Event Upsets in the PANDA EMC

Similar documents
Towards Trusted Devices in FPGA by Modeling Radiation Induced Errors

A Practical Look at SEU, Effects and Mitigation

Self-Test and Adaptation for Random Variations in Reliability

Irradiation Resistivity and Mitigation Measurement Design for Xilinx Kintex-7 FPGAs

Single Event Characterization of a Xilinx UltraScale+ MP-SoC FPGA

Soft Errors re-examined

The ATLAS Pixel Detector

Self Restoring Logic (SRL) Cell Targets Space Application Designs

Neutron Irradiation Tests of an S-LINK-over-G-link System

THE ATLAS Inner Detector [2] is designed for precision

Using Geant4 in the BaBar Simulation. CHEP03 25 March 2003 Dennis Wright (SLAC) on behalf of the BaBar computing group

Product Update. JTAG Issues and the Use of RT54SX Devices

An Introduction to Radiation-Induced Failure Modes and Related Mitigation Methods For Xilinx SRAM FPGAs

Diamond detectors in the CMS BCM1F

Sensors for the CMS High Granularity Calorimeter

arxiv: v1 [physics.ins-det] 1 Nov 2015

Quick Report on Silicon G-APDs (a.k.a. Si-PM) studies. XIV SuperB General Meeting LNF - Frascati

DAQ Systems in Hall A

A flexible FPGA based QDC and TDC for the HADES and the CBM calorimeters TWEPP 2016, Karlsruhe HADES CBM

M. Alderighi/F. Casini

SVT DAQ. Per Hansson Adrian HPS Collaboration Meeting 10/27/2015

Status of readout electronic design in MOST1

The CALICE test beam programme

The Silicon Pixel Detector (SPD) for the ALICE Experiment

The ATLAS Tile Calorimeter, its performance with pp collisions and its upgrades for high luminosity LHC

Reconfigurable Communication Experiment using a small Japanese Test Satellite

A pixel chip for tracking in ALICE and particle identification in LHCb

Studies of large dynamic range silicon photomultipliers for the CMS HCAL upgrade

Impact of Intermittent Faults on Nanocomputing Devices

FRANCO MEDDI CERN-ALICE / University of Rome & INFN, Italy. For the ALICE Collaboration

Scintillation Tile Hodoscope for the PANDA Barrel Time-Of-Flight Detector

New gas detectors for the PRISMA spectrometer focal plane

Time Resolution Improvement of an Electromagnetic Calorimeter Based on Lead Tungstate Crystals

The hybrid photon detectors for the LHCb-RICH counters

Single Event Upset Hardening by 'hijacking' the multi-vt flow during synthesis

Beam test of the QMB6 calibration board and HBU0 prototype

ECAL LED system update. A. Celentano

Local Trigger Electronics for the CMS Drift Tubes Muon Detector

The Readout Architecture of the ATLAS Pixel System

HARDENED BY DESIGN APPROACHES FOR MITIGATING TRANSIENT FAULTS IN MEMORY-BASED SYSTEMS DANIEL RYAN BLUM

Versatile Link. Radiation Qualification

SLHC- PP EU DELIVERABLE: SLHC-PP v1.0. End of Month 36 (March 2011) 23/03/2011. Integration in full-scale detector modules

LHCb and its electronics.

The Status of the ATLAS Inner Detector

Concept and operation of the high resolution gaseous micro-pixel detector Gossip

Commissioning and Performance of the ATLAS Transition Radiation Tracker with High Energy Collisions at LHC

Updates on the Central TOF System for the CLAS12 detector

Commissioning and Initial Performance of the Belle II itop PID Subdetector

Measurement Results of Multiple Cell Upsets on a 65nm Tapless Flip-Flop Array

Double Screening Tests of the CMS ECAL Avalanche Photodiodes

Soft Error Resilient System Design through Error Correction

The CMS Detector Status and Prospects

Status of CMS and preparations for first physics

A Tool For Run Time Soft Error Fault Injection. Into FPGA Circuits

Performance of a double-metal n-on-n and a Czochralski silicon strip detector read out at LHC speeds

Trigger Report. Wesley H. Smith CMS Trigger Project Manager Report to Steering Committee February 23, 2004

PROCESSOR BASED TIMING SIGNAL GENERATOR FOR RADAR AND SENSOR APPLICATIONS

HAPD and Electronics Updates

The Pixel Trigger System for the ALICE experiment

DESIGNING AN ECU CPU FOR RADIATION ENVIRONMENT. Matthew G. M. Yee College of Engineering University of Hawai`i at Mānoa Honolulu, HI ABSTRACT

in Xilinx Devices each) Input/Output Blocks XtremeDSP slices (DSP48) System Monitor Block

The Readout Architecture of the ATLAS Pixel System. 2 The ATLAS Pixel Detector System

Surface Screening with the BetaCage

Reduced Triple Modular Redundancy for Tolerating SEUs in SRAM-based FPGAs

Understanding Design Requirements for Building Reliable, Space-Based FPGA MGT Systems Based on Radiation Test Results

The TORCH PMT: A close packing, multi-anode, long life MCP-PMT for Cherenkov applications

The temperature management of photo cathodes at MAMI and MESA

Threshold Tuning of the ATLAS Pixel Detector

ILC requirements Review on CMOS Performances: state of the art Progress on fast read-out sensors & ADC Roadmap for the coming years Summary

A Soft Error Tolerant LUT Cascade Emulator

Data Acquisition System for Segmented Reactor Antineutrino Detector

Development of beam-collision feedback systems for future lepton colliders. John Adams Institute for Accelerator Science, Oxford University

Beam Loss Detection for MPS at FRIB

arxiv:hep-ex/ v1 27 Nov 2003

TitleLarge strip RPCs for the LEPS2 TOF. Author(s) Chu, M.-L.; Chang, W.-C.; Chen, J.- Equipment (2014), 766:

Pixel Detector Control System

FRONT-END AND READ-OUT ELECTRONICS FOR THE NUMEN FPD

Compact Muon Solenoid Detector (CMS) & The Token Bit Manager (TBM) Alex Armstrong & Wyatt Behn Mentor: Dr. Andrew Ivanov

PIXEL2000, June 5-8, FRANCO MEDDI CERN-ALICE / University of Rome & INFN, Italy. For the ALICE Collaboration

Screen investigations for low energetic electron beams at PITZ

1 Digital BPM Systems for Hadron Accelerators

Sensors for precision timing HEP

Professor Lloyd W. Massengill

Short summary of ATLAS Japan Group for LHC/ATLAS upgrade review Liquid Argon Calorimeter

Architectural Consequences of Radiation Performance in a Flash NAND Device

Dark current and multipacting trajectories simulations for the RF Photo Gun at PITZ

Synchronization Voter Insertion Algorithms for FPGA Designs Using Triple Modular Redundancy

A fast and precise COME & KISS* QDC and TDC for diamond detectors and further applications

Frontend Electronics for high-precision single photo-electron timing

Tests of the boards generating the CMS ECAL Trigger Primitives: from the On-Detector electronics to the Off-Detector electronics system

G. Pittá(*), S. Braccini TERA Foundation, Novara, Italy (*) Corresponding author.

Performance of the MCP-PMT for the Belle II TOP counter

Status of GEM-based Digital Hadron Calorimetry

Simulations on Beam Monitor Systems for Longitudinal Feedback Schemes at FLASH.

PERFORMANCE ANALYSIS OF AN EFFICIENT TIME-TO-THRESHOLD PWM ARCHIECTURE USING CMOS TECHNOLOGY

A prototype of fine granularity lead-scintillating fiber calorimeter with imaging read-out

SciFi A Large Scintillating Fibre Tracker for LHCb

First LHC Beams in ATLAS. Peter Krieger University of Toronto On behalf of the ATLAS Collaboration

Single-Event Upset Technology Scaling Trends of. Unhardened and Hardened Flip-Flops in Bulk CMOS. Nelson J. Gaspard III.

Status of CMS Silicon Strip Tracker

Transcription:

Single-Event Upsets in the PANDA EMC Results from a neutron irradiation of the front-end digitiser board M. Preston, P.-E. Tegnér (Stockholm University) H. Calén, T. Johansson, K. Makònyi, P. Marciniewski (Uppsala University) M. Kavatsyuk, P. Schakel (University of Groningen) SFAIR Meeting, 27 th of October 2016, Göteborg

Single-Event Upsets in the PANDA EMC 2/14 The PANDA Electromagnetic Calorimeter Figure from EMC Technical Design Report, PANDA. First step of readout electronics: FPGA-based sampling-adcs.

The front-end digitiser board 2 Kintex-7 FPGAs 2 SFPs 8 ADCs, 64 channels 32 light sensors Single-Event Upsets in the PANDA EMC 3/14

Single-Event Upsets in the PANDA EMC 4/14 Front-end electronics in the Electromagnetic Calorimeter 600 front-end digitiser boards in the EMC 375 in Barrel 217 in Forward Endcap

Single-Event Upsets in the PANDA EMC 4/14 Front-end electronics in the Electromagnetic Calorimeter 600 front-end digitiser boards in the EMC 375 in Barrel 217 in Forward Endcap Figure courtesy of C. Schnier. Distributed over several crates placed around the detector perimeter (average 1.3 metres from the beam at the Forward Endcap). Exposure to a high flux of particles Radiation effects Of interest here: neutrons (but also other hadrons, photons, electrons... could cause problems)

Single-Event Upsets in the PANDA EMC 5/14 What does radiation do to electronics? Wide range of effects, but generally: Transient effects (device not permanently damaged) Permanent effects (device permanently damaged) A subset of transient effects is Single-Event Upsets (SEUs). Radiation-induced bit-flip (0 1) in memory cell.

What does radiation do to electronics? Incoming neutron not directly ionising n + A (e. g. 28 Si) ionising particle(s) Charge deposit change of state in cell. Figure from Introduction to Single-Event Upsets, Altera. Single-Event Upsets in the PANDA EMC 5/14

Single-Event Upsets in the PANDA EMC 6/14 Single Event Upsets in FPGAs Such a process affects an FPGA semiconductor device with several sensitive parts (Look-Up Tables, routing, etc.). Several types of memory: 1. Configuration Memory 2. Block Memory 3. Distributed Memory 4. Flip-Flops

Single-Event Upsets in the PANDA EMC 6/14 Single Event Upsets in FPGAs Such a process affects an FPGA semiconductor device with several sensitive parts (Look-Up Tables, routing, etc.). Several types of memory: 1. Configuration Memory 2. Block Memory 3. Distributed Memory 4. Flip-Flops # of bits (1) contains Function of design (2-4) contain State of design We looked only at neutron-induced errors in the Configuration Memory.

Single-Event Upsets in the PANDA EMC 7/14 Neutron irradiation The board was irradiated in June 2016 at the The Svedberg Laboratory (TSL) in Uppsala.

Single-Event Upsets in the PANDA EMC 7/14 Neutron irradiation The board was irradiated in June 2016 at the The Svedberg Laboratory (TSL) in Uppsala. Board placed in the Standard User Position (SUP), beam perpendicular to the board. One FPGA read out. Beam diameter: 130 mm Φ Figure from Prokofiev, A. V. et al. (2009) 2009 IEEE Radiation Effects Data Workshop. 166-173. Readout

Single-Event Upsets in the PANDA EMC 7/14 Neutron irradiation The board was irradiated in June 2016 at the The Svedberg Laboratory (TSL) in Uppsala. Board placed in the Standard User Position (SUP), beam perpendicular to the board. One FPGA read out. Neutron flux between 5 10 5 and 1 10 6 s 1 cm 2 (>10 MeV). 1.0 10-1 TSL PANDA, p pbar = 15 GeV/c Normalised Neutron flux 1.0 10-2 1.0 10-3 1.0 10-4 1.0 10-5 10 100 1000 10000 Neutron energy [MeV]

Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. Single-Event Upsets in the PANDA EMC 8/14

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC).

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. n 1 X Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC). 1. Incident neutron ionising particle SEU?

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. 2 Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC). 1. Incident neutron ionising particle SEU? 2. Single-Bit Upset (SBU). Corrected.

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. 3 Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC). 1. Incident neutron ionising particle SEU? 2. Single-Bit Upset (SBU). Corrected. 3. Intra-Frame Multi-Bit Upset (MBU).Corrected.

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. 4 Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC). 1. Incident neutron ionising particle SEU? 2. Single-Bit Upset (SBU). Corrected. 3. Intra-Frame Multi-Bit Upset (MBU).Corrected. 4. Inter-Frame MBU. NOT corrected

Single-Event Upsets in the PANDA EMC 8/14 Error detection and correction Xilinx provides the Soft Error Mitigation (SEM) Controller, an IP core which may be implemented in the firmware. Correction capability depends on the structure of the error and SEM mode. 5 Memory divided into frames (1 frame = 101 32-bit words) SEM protected each frame with an Error Correction Code (ECC). 1. Incident neutron ionising particle SEU? 2. Single-Bit Upset (SBU). Corrected. 3. Intra-Frame Multi-Bit Upset (MBU).Corrected. 4. Inter-Frame MBU. NOT corrected 5. Inter-Frame MBU + extra. NOT corrected

Single-Event Upsets in the PANDA EMC 9/14 Measurement process Start Configure FPGA Beam on Beam off Log file SEU time + address Yes Detect SEU Correctable? No

Single-Event Upsets in the PANDA EMC 10/14 Results SEU cross section The cross section for an SEU is given by Number of SEU σ SEU = N SEU T meas Φ n N bits Measurement duration Neutron flux Total number of bits

Single-Event Upsets in the PANDA EMC 10/14 Results SEU cross section The cross section for an SEU is given by Number of SEU σ SEU = N SEU T meas Φ n N bits Measurement duration Neutron flux Total number of bits Upset rate, r SEU : r SEU = σ SEU Φ n N bits Mean Time Between Failures, MTBF : 1 MTBF = σ SEU Φ n N bits

Results SEU cross section 1.2 10-14 1.0 10-14 σ SEU [cm 2 bit -1 ] 8.0 10-15 6.0 10-15 4.0 10-15 2.0 10-15 This measurement Wirthlin et al. [1] 4.0 10 5 5.0 10 5 6.0 10 5 7.0 10 5 8.0 10 5 9.0 10 5 1.0 10 6 1.1 10 6 1.2 10 6 Neutron flux Φ [cm -2 s -1 ] [1] Wirthlin, M. J. et al. (2014) JINST. 9 C01025. Larger Kintex-7 @ TSL Single-Event Upsets in the PANDA EMC 11/14

Results SEU cross section 1.2 10-14 1.0 10-14 σ SEU [cm 2 bit -1 ] 8.0 10-15 6.0 10-15 4.0 10-15 2.0 10-15 Average σ SEU = 7.48 10-15 (± 7%) cm 2 bit -1 4.0 10 5 5.0 10 5 6.0 10 5 7.0 10 5 8.0 10 5 9.0 10 5 1.0 10 6 1.1 10 6 1.2 10 6 Neutron flux Φ [cm -2 s -1 ] Cross section is independent of the neutron flux. Single-Event Upsets in the PANDA EMC 11/14

Single-Event Upsets in the PANDA EMC 12/14 SEU rates in the Electromagnetic Calorimeter A simulation of the neutron flux in the EMC has been made using PandaRoot. p pbar = 15 GeV/c, L = 2 10 32 cm 2 s 1 Φ n at position of digitisers is 150 cm 2 s 1. Assumptions: (i) neutron energy spectra are identical, (ii) only neutrons > 10 MeV cause SEUs, (iii) all boards are placed at same position (1 m from centre at FW endcap). MTBF 22000 s per FPGA.

Single-Event Upsets in the PANDA EMC 12/14 SEU rates in the Electromagnetic Calorimeter A simulation of the neutron flux in the EMC has been made using PandaRoot. p pbar = 15 GeV/c, L = 2 10 32 cm 2 s 1 Φ n at position of digitisers is 150 cm 2 s 1. Assumptions: (i) neutron energy spectra are identical, (ii) only neutrons > 10 MeV cause SEUs, (iii) all boards are placed at same position (1 m from centre at FW endcap). MTBF 22000 s per FPGA. But, 600 boards in the EMC, with 2 FPGAs per board. The results scale: rseu = 0.056 s 1 MTBF = 18.1 s. MTBF for errors not correctable by SEM = 584 s 9.7 min. That is, every 10 minutes reconfiguration would be needed somewhere in the EMC (300 ms from Flash).

Single-Event Upsets in the PANDA EMC 13/14 Summary and Outlook Soft errors in the Configuration Memory of the Kintex-7 FPGA have been studied.

Single-Event Upsets in the PANDA EMC 13/14 Summary and Outlook Soft errors in the Configuration Memory of the Kintex-7 FPGA have been studied. The cross section for SEU in the Kintex-7 Configuration Memory was determined to be 7.48 10 15 cm 2 bit 1, in agreement with previous results.

Single-Event Upsets in the PANDA EMC 13/14 Summary and Outlook Soft errors in the Configuration Memory of the Kintex-7 FPGA have been studied. The cross section for SEU in the Kintex-7 Configuration Memory was determined to be 7.48 10 15 cm 2 bit 1, in agreement with previous results. The Mean Time Between Failures in the Configuration memory in the whole EMC would be 18.1 seconds (any error type) 9.7 minutes (uncorrectable errors) The uncorrectable errors could be mitigated with other methods (for example TMR of the Configuration Memory).

Single-Event Upsets in the PANDA EMC 13/14 Summary and Outlook Soft errors in the Configuration Memory of the Kintex-7 FPGA have been studied. The cross section for SEU in the Kintex-7 Configuration Memory was determined to be 7.48 10 15 cm 2 bit 1, in agreement with previous results. The Mean Time Between Failures in the Configuration memory in the whole EMC would be 18.1 seconds (any error type) 9.7 minutes (uncorrectable errors) The uncorrectable errors could be mitigated with other methods (for example TMR of the Configuration Memory). Upcoming proton irradiation at KVI, Groningen (November). Thermal-neutron irradiation in Stockholm is planned.

Thank you for your attention! Single-Event Upsets in the PANDA EMC 14/14