01 / ENVIRONMENT

Start with mission dose and event conditions

The relevant environment depends on altitude, inclination, shielding, solar conditions and mission duration. Total ionising dose can shift device behaviour over time. Displacement damage can degrade materials and sensors. Single-event effects can corrupt a bit, interrupt a processor, damage a power device or trigger a destructive condition.

A credible radiation requirement therefore states the orbit and duration, environment model, shielding geometry, uncertainty, component derating and acceptable rate of corrected, interrupted and permanent failures.

02 / DEFENCE IN DEPTH

No single protection layer is enough

01Parts and layout

Select characterised components, derate them and isolate power domains so one event cannot propagate through the node.

02Physical mitigation

Use shielding where analysis shows net benefit; mass and secondary-particle effects prevent “more shielding” from being a universal answer.

03Error control

Error-correcting memory, scrubbing, checksums and redundant state detect and repair recoverable corruption.

04Containment and recovery

Watchdogs, current limiting, reset paths, checkpointing and workload migration keep faults bounded.

Radiation-aware software assumes transient errors will occur. It validates data before use, limits command authority, records anomalies and makes degraded modes explicit.

03 / MEMORY AND STATE

Protect the data path from write to recovery

Candidate non-volatile memory technologies may reduce sensitivity to particular upset mechanisms, but supplier characteristics are not mission proof. The complete path includes controller logic, buses, caches, power loss, firmware and replica management.

  • Attach integrity metadata to stored blocks and checkpoints.
  • Maintain recoverable replicas across independent fault domains.
  • Scrub state on a measured schedule rather than assuming permanence.
  • Record corrected errors as health telemetry and retirement signals.
  • Test recovery from corrupted, stale and partially written state.
04 / EVIDENCE PROGRAMME

Test the parts, board, software and mission together

Qualification should combine radiation analysis, available component test data, targeted exposure and fault injection. Proton, heavy-ion or gamma testing is selected according to the failure mechanisms and mission environment. Board-level tests must verify power protection, latch-up response, memory correction and autonomous recovery under representative load.

Software fault injection is complementary, not a substitute for physical exposure. It can demonstrate detection and recovery logic while hardware testing characterises whether the assumed faults and thresholds are realistic.

05 / CLAIM CONTROL

Use precise language about tolerance

PUBLIC-SAFE
  • Radiation-aware architecture
  • Candidate shield stack
  • Error-correcting design target
  • Modelled mission dose
  • Tested to a named method and limit
NOT YET EVIDENCED
  • Radiation-proof
  • Immune to bit flips
  • Guaranteed five-year life
  • Flight-proven
  • Zero data loss

Proof of Pulse currently documents mitigation concepts and validation requirements. Final orbit, part selection, dose margin and integrated qualification remain open engineering gates.