Soft Errors & Transient Faults
A neutron or an alpha particle can flip a bit in a safety-relevant microcontroller and leave the silicon undamaged. Follow the path from radiation physics and single event effects through ECC and hardening to a transient FMEDA under ISO 26262.
- Chapters
- 12
- Chapters
- Figures and interactive visuals
- 15
- Figures and interactive visuals
- Worked transient FMEDA
- 1
- Worked transient FMEDA
Quick answers
- What is a soft error in ISO 26262?
- A soft error is a transient fault in which an ionizing particle, typically an atmospheric neutron or an alpha particle from package materials, deposits enough charge to flip the state of a storage node without damaging the transistor. The wrong value disappears when it is rewritten or the device is power-cycled, unlike a permanent stuck-at fault. ISO 26262 treats such transient faults as random hardware faults, and ISO 26262-11 gives guidance on analyzing them in semiconductors.
- What causes soft errors in automotive chips?
- Three terrestrial sources matter. Atmospheric neutrons from cosmic-ray showers dominate at ground level and cannot practically be shielded in a vehicle. Alpha particles come from trace radioactive impurities in package and chip materials, which is why low-alpha materials are used. Thermal neutrons can trigger upsets through capture in certain materials. Altitude and location change the neutron flux, so the environment belongs in the mission profile.
Why it pays for itself
A transient FMEDA you can audit
Follow a complete worked transient FMEDA for a fictional microcontroller, element by element, so your own analysis has a structure an assessor can trace.
Size ECC and scrubbing correctly
Choose between error codes, use interleaving so clustered upsets stay correctable, and set scrub intervals that support the coverage your FMEDA claims.
Derating without cheating
Apply masking and safe-fraction arguments while keeping base rates raw and every derating factor disclosed, the discipline that separates a defensible claim from an assessor finding.
What you’ll be able to do
Classify single event effects
Separate soft from destructive effects and treat each correctly in the analysis.
Judge a soft error rate
Read measurement reports critically and keep raw rates apart from derating.
Choose codes and scrubbing
Match ECC, interleaving and scrub intervals to the faults they must handle.
Apply ISO 26262 to transients
Analyze transient faults alongside permanent ones with clear rationales.
Build a transient FMEDA
Carry elements from raw rate to residual contribution with auditable steps.
Chapter by chapter
- 01
When a Bit Flips
See why a fault that leaves no trace is different from permanent and intermittent faults, and why it is rare per chip but constant per fleet.
- A fault without a trace
- Permanent, transient, intermittent
- Rare per chip, constant per fleet
- 02
Where the Particles Come From
Follow particles from space to the ECU, the terrestrial sources that matter, and how deposited charge flips a cell.
- Atmospheric cascade
- Terrestrial sources
- Critical charge
- 03
The Single Event Zoo
Sort single event effects into soft and destructive branches and learn the distinction ECC depends on.
- Family tree of effects
- Soft branch
- MCU versus MBU
- 04
Silicon Sensitivity & Scaling
Map which parts of a chip are sensitive, how scaling changes the picture, and what the operating point does.
- Chip sensitivity map
- Scaling trends
- Operating point effects
- 05
Measuring the Soft Error Rate
Learn the units, the ways to obtain a rate, why accelerated testing works, and how to read a vendor report critically.
- FIT and FIT per Mbit
- JESD89 test methods
- Reading a vendor report
- 06
From Raw SER to Effective Failure Rate
Filter raw upsets down to safety-relevant failures through masking and vulnerability windows, with every derating visible.
- Logic masking
- Architectural vulnerability
- Derating bookkeeping
- 07
Detecting & Correcting: Codes and Scrubbing
Climb the ladder of error codes, follow a SEC-DED round trip, and see why scrubbing and interleaving matter.
- Ladder of error codes
- What ECC misses
- Scrubbing
- 08
Hardening the Design
Combine protection layers within one budget, examine lockstep and recovery, and choose mechanisms with a selection matrix.
- Six layers, one budget
- Lockstep pair anatomy
- Selection matrix
- 09
What ISO 26262 Says About Transient Faults
Find where the standard addresses transients, how fault injection backs the claims, and how supplier and integrator split the work.
- Transients in the standard
- Fault injection evidence
- Supplier and integrator split
- 10
Worked Example: A Transient FMEDA
Carry a fictional microcontroller subsystem through a per-element transient FMEDA and roll the results up.
- Per-element pipeline
- Safe-fraction rationales
- Transient roll-up
- 11
The System View & Emerging Trends
Look at large ADAS compute, bit flips in neural networks, lessons from avionics and field telemetry as evidence.
- ADAS compute
- Bit flips in neural networks
- Field telemetry
- 12
Pitfalls, Checklist & Assessor Questions
Test your analysis against claims that do not survive assessment and a lifecycle checklist for transient faults.
- Claims that fail assessment
- Assessor questions
- Transient fault checklist
Read all 12 chapters with Expert.
Start with a free account: the complete Functional Safety Concept guide and 5 practice exams a month, then upgrade when you need this guide.
Expert plan pricingNot just text: the visual toolkit
The atmospheric cascade
How a primary cosmic ray becomes neutrons at ground level.
The race inside the cell
Deposited charge against the critical charge of a storage node.
Strike footprint and interleaving
Why bit interleaving turns clustered upsets into correctable ones.
The scrubber race
Periodic correction against the arrival of a second upset.
An FMEDA you can push on
An interactive transient FMEDA for the worked example.
Transient FMEDA for a fictional lockstep microcontroller
A fictional microcontroller subsystem is sized so every step can be followed, from raw upset rates per element to a transient roll-up.
- The three-factor pipeline applied per element
- Safe-fraction rationales, the part assessors examine
- What the mitigation stack achieves per element
- Rolling up the transient metric with the RAM and logic split
- The contribution to the PMHF
Unlock in course
Who this guide is for
- Hardware safety engineers who must add transient faults to their FMEDA
- SoC and memory designers choosing between ECC schemes, interleaving and scrubbing
- FMEDA reviewers asked to judge a supplier soft error rate and its derating
- Engineers hearing "cosmic rays" in a design review and wanting the real physics
Frequently Asked Questions
Common questions about Soft Errors & Transient Faults
Start the course today
A free account unlocks one full concept guide, 3 work product templates, 1 guided process, the Markov and RBD simulators, and 5 practice exams per month. The Pro and Expert plans unlock more of the 93-guide library. No credit card required.
Continue with
Where this topic leads next.
- HardwareISO 26262 and SemiconductorsPractical handbook for applying ISO 26262 Part 11 to digital, analog, programmable logic, multi-core SoCs, and sensor silicon. Covers SEooC and DIA for IPs and chips, on-chip safety mechanisms, FMEDA at silicon, DFA at die level, fault injection campaigns, and EDA tool qualification.
- HardwareFPGA SafetyHardware whose circuit is data: configuration upsets and essential bits, scrubbing against the FTTI, TMR in the fabric, HDL discipline, tool confidence, and the FPGA FMEDA that closes the safety case.
- SoftwareSoftware Architectural & Unit DesignLearn architecture patterns for fault isolation, testability, and diagnostic coverage in safety-critical software. Covers modularity, design-by-contract, complexity metrics, and structural coverage requirements per ASIL.