Radiation Effects and Mitigation
Insert cool group name here Josh Mackler Amanda Quan Joe Tostenrude
CURRENT EVENTS Why do we care? • Damage
o Satellites o Power grids o Cellphone networks
• Flights re-routed (travelling high altitude) • Electronic components reducing in size
Recent Solar Event (Jan 22, 2012) Halloweenstorm Phobos-Grunt
Jun 7, 2011
SPACE RADIATION OVERVIEW
Nikkei Science, Inc. of Japan, by K. Endo
Radiation Sources: • Solar events - Solar wind
o 9-13 year cycles o Solar flares o Coronal mass ejections
• Galactic Cosmic Rays o Energies in GeV o Protons, Alpha particles, heavy ions o Found everywhere in interplanetary space
• Van Allen Belts - Magnetosphere o Earth's magnetic field interaction with
ionized solar wind o Deflect majority of particles o Traps protons and electrons o Varies with solar events and within regions
South Atlantic Anomaly o Aurora Borealis and Australis
TERRESTRIAL RADIATION • Nuclear reactors
o Gamma and neutron radiation o Interfere with sensors and control circuits
• Nuclear explosions o Electromagnetic radiation o Electromagnetic pulse (EMP) o High dose of ionizing particles
• Medical devices o Radiotherapy interaction
RADIATION METRICS AND EFFECTS • Dose o Semiconductor lattice displacement o Trapped charge in thin gate oxide
o Increased gate leakage o Units in Rad(Si) - 0.01 J per kg
• Total Ionizing Dose o Total deposited energy due to protons and electrons (primarily) o Voltage threshold shift o Cumulative long term damage
• Dose Rate Upset o Over shorter intervals (20-50ns) o Power rail collapse, metal burnout
RADIATION METRICS AND EFFECTS
• Single Event Effects o Categories
Destructive Unrecoverable Recoverable Data Integrity
o Types Single Event Upset Single Event Transient Single Event Latch-up
RADIATION METRICS AND EFFECTS o Single Event Upsets (SEU)
Flip-flop or memory bit cell state changes due to single ion strike Multi-bit upsets in adjacent memory cells State machine control logic upset (SEFI)
o Single Event Transients (SET) Charge from ion deposited in circuit Glitches on combinational logic
o Single Event Latch-up (SEL) Heavy ion or high energy proton induced latch-up Low impedance path from power rail to transistor
MITIGATION APPROACHES Why doesn't shielding work?
• Some sensors need to be exposed • Impractical for some applications
• Alpha particles – sheet of paper • Beta particles – few mm of aluminum • Xray/Gamma rays – several cm of lead • High energy protons in space – good luck!
• Shielding can slow down ions and increase the energy transfer of high-energy protons
Approaches
• Radiation Hardening by Process (RHBP) • Radiation Hardening by Design (RHBD) • Software
RHBP • Radiation Hardening By Process
o Advantages: Immune to latch-up Mitigates SEU
o Disadvantages: Expensive Low yield Process instability Technology lag
RHBD TECHNIQUES RTL code techniques • Fully specified case statements • Error Detection and Correction (EDAC) • Memory scrubbing • Triple modular redundancy (TMR)
Custom digital logic standard cell library • Dual Inter-locked Storage Cells (DICE) • Temporal (pulse) filters • High drive single stage clock drivers
Custom memory • Bit and word level interleaving • High capacitance bit cell
MAESTRO Example of RHBD design in practice
• 49 core processor based on Tilera TILE64 architecture • 2-D mesh with low-latency high-bandwidth register-
mapped networks • 500 Krad TID Artisan memories • EDAC and scrubbing to meet SEU rate requirements • Cache parity • RHBD Digital logic flow • Software suite for parallel processing
SOFTWARE MITIGATION OF RAD - OVERVIEW
Fault-tolerance in software domain is not as well understood as fault-tolerance in hardware domain
• Controversial opinions exist on whether reliability can be used to evaluate software. • Software failures are mostly due to the activation of design faults by specific input
sequences. • This makes the reliability of a software module dependent on the environment that
generates input to the module over the time. Software fault-tolerance techniques:
• Single-version o Fault detection o Containment o Recovery mechanisms
• Multi-version o Redundant software modules o Design diversity rules
SOFTWARE MITIGATION OF RAD - SINGLE-VERSION
Acceptance testing/fault detection • Timing checks • Coding checks • Reversal checks • Reasonableness checks • Structural check
Techniques for fault containment
• Modularization • Partitioning • System closure • Atomic action
Fault Recovery techniques
• Exception handling • Checkpoint and restart
SOFTWARE MITIGATION OF RAD - MULTI-VERSION
Redundancy • Doesn't work quite the same way since duplicate software systems will be exactly the
same and do not have the same induced errors from the environment that hardware does (i.e. software will have all the same inputs and therefore all the same outputs)
• Recovery blocks use the concept of retrying the same operation in expectation that the problem is resolved after the second try.
• Allocation - at which level is redundancy most useful? (procedure, process, whole system)
Design diversity • Triplicating a software module cannot tolerate faults because the same fault is
replicated in all blocks. • The goal becomes diversity rather than similarity. This is widely accepted to be a
difficult task.
CONCLUSION
NASA/GSFC Eng Seminar –Radiation Effects 101 presented by Kenneth A. LaBel – Oct 28, 2005
REFERENCES http://www.latimes.com/news/science/la-sci-solar-flare-20120124,0,2353551.story?track=icymi http://www.space.com/14404-mars-rover-solar-radiation-storm-curiosity.html http://radhome.gsfc.nasa.gov/radhome/papers/Core_Tech.pdf http://en.wikipedia.org/wiki/Radiation_hardening http://en.wikipedia.org/wiki/Silicon_on_sapphire http://www.honeywellmicroelectronics.com/radiation-hardened-technology.php https://nepp.nasa.gov/mapld_2009/talks/083109_Monday/03_Malone_Michael_mapld09_pres_1.pdf http://holbert.faculty.asu.edu/eee560/see.html http://www.aeroflex.com/ams/pagesproduct/datasheets/leon/ut699leon3ftdatasheet.pdf http://web.mst.edu/~umrr/cf147.pdf http://blogs.mentor.com/jvandomelen/blog/2011/01/14/radiation-mitigation/ http://techbites.com/201006032772/myblog/blog/z0002-yes-mentor-introduces-rad-tolerant-fpga-synthesis.html http://en.wikipedia.org/wiki/Triple_modular_redundancy http://en.wikipedia.org/wiki/Memory_scrubbing http://en.wikipedia.org/wiki/Radiation_hardening http://press.xilinx.com/phoenix.zhtml?c=212763&p=irol-newsArticle&ID=1448836&highlight= http://www.ieee.org/organizations/pubs/newsletters/npss/0305/rhbd.html http://www.aeroflex.com/ams/pagesproduct/articles/COTSJournalRH.pdf http://webcache.googleusercontent.com/search?q=cache:4qbA8y8E9KQJ:holbert.faculty.asu.edu/eee560/tiondose.html+what+is+%22total+ionizing+dose%22&cd=5&hl=en&ct=clnk&gl=us http://news.discovery.com/space/cosmic-rays-killed-russian-mars-probe-120131.html
• http://radhome.gsfc.nasa.gov/radhome/papers/LaBel_FPGA04.pdf • http://webcache.googleusercontent.com/search?q=cache:hxS9cFg33rYJ:www.aero.org/capabilities/seet/SEupset.html+seu+b
it+flips&cd=1&hl=en&ct=clnk&gl=us • http://webcache.googleusercontent.com/search?q=cache:hxS9cFg33rYJ:www.aero.org/capabilities/seet/SEupset.html+seu+b
it+flips&cd=1&hl=en&ct=clnk&gl=us