Chapter 2 · Lesson 6 of 6 · about 10 minutes
The transport handoff
Transport connects the environment outside the spacecraft to the hardware inside it. Its outputs feed dose requirements, displacement-damage assessments and single-event rate calculations. The next step also needs the part response and the consequence in the circuit or the system. A shielding value by itself does not close an RHA case.
The exposure the next engineer needs
Accumulated effects
The dose and displacement-damage exposures go into an end-of-life performance assessment, together with device data that apply to the part and its operating conditions.
Single events
The local particle field has to be combined with a measured response and a rate method that applies to it. What the event means for the mission depends on the circuit protection, the recovery and how the function is used.
Work product
The work product of this stage is an analysis request and a set of geometry diagnostics. The radiation results stay blank until a suitable transport analysis has actually been run and reviewed.
A reproducible transport handoff
Deliver the dose in each material of interest and the local spectra the next step needs, each tagged with its source case, location, time interval and uncertainty. Record the versions of everything and the boundary definitions. Keep the predicted exposure, the device-response evidence and the project's design margin traceable as three separate items.
TID. Report the absorbed dose with its material, location, configuration and time interval. Keep the contribution of each source and the statistical case you chose. rad(Si), rad(SiO₂) and crew equivalent dose are different quantities or different materials, and the handoff should say which one it gives.
Displacement damage. Keep the local fluence by species and energy, the target material and the damage-function convention. A DDD or an equivalent fluence has to be paired with a device endpoint that applies to the part, including the relevant annealing and operating conditions.
SEE. Keep the species, the energy or the material-specific LET, the angular convention and the boundary the response calculation assumes. Say whether the package is inside that boundary, and whether the response is defined per bit, per device or per function.
Reproduction. Archive what someone would need to run the calculation again: the source files, the geometry and material revisions, the tool and data releases, the physics and settings, the sampling weights, the normalization, the score definitions, the checks that were done and the known limitations.
The local LET spectrum. The precomputed integral LET flux examples in the reading list are for silicon behind 100 mil Al. At that fixed shielding you can compare orbits or locations, read off F(>L), or difference the curve into finite LET bins. A local LET spectrum is still only an exposure. Before it becomes an SEE rate it needs a device response and an angular convention that match it.
Exercise
Produce a transport handoff whose exposure quantities and response conventions another engineer can reproduce and judge.
Use the same quantities and conventions from the environment specification through the requirements and the response analysis, so that nothing is redefined along the way.
Going deeper
Response folding and assurance limits
A transported field becomes an engineering exposure or an event rate when it is folded with a response model that matches it. Collapsing the spectrum, the angles or the time history can throw away information that response needs. In the assurance argument, the uncertainty in the exposure, the uncertainty in the measured response and the consequence for the system have to stay distinguishable.
Response functional. A linear spectral exposure can be written schematically as Q = Σᵢ∫Kᵢ(E,Ω)Φᵢ(E,Ω)dE dΩ, with the area and angular conventions of K and Φ matched. K may be a deposited-dose response, a material damage function or an event response. Each endpoint has its own K; there is no single function that serves all of them.
Applicability. The SEE cross section and the rate method have to agree on the event definition, the device conditions and the angular or sensitive-volume assumptions. LET by itself can hide the species, the range and the geometry. Make sure the package is transported once, either in the local-field stage or in the response stage, and not in both.
Time reduction. For an instantaneous event-rate model that matches the exposure, the expected count is N = ∫R(t)dt. Two exposures with the same count can produce different outages when recovery, protection or correlated events matter. In the same way, an integrated dose loses the dose-rate, bias, temperature and annealing history that the device interpretation may depend on.
Assurance claim. Keep the environment statistic you selected, the transport and model uncertainty, the measured-response uncertainty, the project margin and the acceptance criterion as separate items. A margin factor is not automatically a confidence percentile. The controlled package should make visible every assumption that could change the decision.
Takeaway
Hand over the local exposure together with what it means, what it assumes and what assessment remains to be done.
Check yourself
What has to accompany the transported spectrum when it is handed to the rate and requirement work?
Its basis: the environment model and epoch, the geometry revision, the depth and material of the scoring point, whether the result is omnidirectional, the shielding it already includes, and the cases that were not run. Without that basis the next engineer cannot tell what the number covers.
A later design change thickens one wall. Which handoff products have to be redone?
Those whose scoring points see that wall: the doses and LET spectra at parts behind it, and any requirement derived from them. A record that links each product to the geometry revision and the scoring points makes the reanalysis small and specific.
A transport report gives the local dose at every part, and the project treats it as closing the TID case. What is missing?
A shielding value by itself does not close an RHA case. The dose has to go into an end-of-life performance assessment together with device data that apply to the part and its operating conditions, and the predicted exposure, the device-response evidence and the project's design margin have to stay traceable as three separate items. The work product of the transport stage is an analysis request, a set of geometry diagnostics and an archive from which the calculation can be run again; the assessment that remains to be done is part of the handoff.
Two exposures give the same expected event count, N = ∫R(t)dt. Why can they produce different outages, and what does the same kind of collapse do to a dose?
The count loses the time history: two exposures with the same count can produce different outages when recovery, protection or correlated events matter, which is why the consequence is worked at the function level in chapters 9 and 10. An integrated dose loses the dose-rate, bias, temperature and annealing history that the device interpretation may depend on, such as the dose-rate sensitivity of a bipolar part (lesson 3.3). Collapsing the spectrum, the angles or the time history can throw away information that the response needs, so the handoff should keep what the next step will fold.
A precomputed integral LET spectrum for silicon behind 100 mil of aluminum is available for the orbit, and a colleague folds the part's cross-section curve with it and reports a rate. What has been skipped?
A local LET spectrum is only an exposure; before it becomes a rate it needs a device response and an angular convention that match it, and the cross section and the rate method have to agree on the event definition, the device conditions and the angular or sensitive-volume assumptions. The precomputed curve is for a fixed shielding, so it supports comparing orbits or reading F(>L), and it stands in for the part's own shielding only when that shielding is the same. LET by itself can hide the species, the range and the geometry, and the package has to be transported once, either in the local-field stage or in the response stage, not in both.
Further reading
- NASA RHA Guidelines: transport and CAD review, Section 4, printed pp
- ESA/SPENVIS: MULASSIS
- Hugh Evans: space environment and calculation methods
- Jun et al. (2020): transport-code intercomparison
- Leif Scheick: environment and shielding course, slides 1–19
- Allan Johnston: Space Radiation Effects on Microelectronics, slides 4–29 (environment and ionization) and 19–20 (MER)
- IEEE NSREC 2017 Short Course, "RHA for Satellite Systems, From Macro to Nano", §3
- IEEE NSREC 2016 Short Course, "Techniques for SEE Modeling and Mitigation", §2
- IEEE NSREC 2018 Short Course, "Variability in Environments, Devices and Radiation Effects", §B
- AFRL: AE9/AP9/SPM run modes
- IEEE NSREC 2024 Short Course, "Electronic and Photonic Technologies, Basic Concepts to Advanced Mechanisms", §3–4, especially displacement-damage discussion pp
- ESA/SPENVIS: Geant4 source normalization
- NIST STAR: silicon density
- IEEE NSREC 2015 Short Course, "Practical Problems in Spacecraft Design", §2
- IEEE NSREC 2022 Short Course, "Multi-Scale, Multi-Physics of Radiation Effects", §3
- Vanderbilt: CREME96 workflow
- IEEE NSREC 2013 Short Course, "Evolution of Radiation Studies on the Path of Moore’s Law", §4
- SR-NIEL: spectral-fluence damage calculation
- IEEE NSREC 2011 Short Course, "Back to the Basics", §3
- ESA: GRAS
- NASA Avionics RHA Guidelines, Sections 3–8, 10 and Appendix B: mission context, environment, transport, effects, SEECA, monitoring and requirements
- Koontz (2017): Space Radiation Environments, NASA NESC, slide deck reviewed: slides 4–6 (units, shielding and counting), 10 (July 2000 event), 12–14 (GCR), 29 (cycle and neutron proxy), 31–32 (acc
- Vanderbilt: CREME96 HUP
- Vanderbilt: CREME96 PUP
- Space-RHA SEE tool: precomputed CREME96 environment examples, p.
- SILSO / Royal Observatory of Belgium: annual sunspot number V2