ECSS Radiation Hardness Assurance — RDM Single-Event Effects (space-systems/ecss/e1012-rdm-see)
Use when the task is to apply the radiation design margin (RDM) approach
for single-event effects (SEE) on spacecraft EEE components under
ECSS-E-ST-10-12C §5.1.3: each SEE-sensitive part is first categorized
by effect type (destructive or non-destructive), then its margin is
computed and checked against the required RDM factor.
Domain quick reference
- §5.1.3 divides single-event effects into two assessment tracks based
on whether the effect is destructive or non-destructive. Destructive
effects — single-event latchup (SEL), single-event burnout (SEB),
and single-event gate rupture (SEGR) — can permanently damage the
device and require a LET-threshold margin check. Non-destructive
effects — single-event upset (SEU), single-event functional interrupt
(SEFI), and single-event transient (SET) — are recoverable and
require a rate-based margin check. Each part is assigned to exactly
one track before its margin is evaluated.
- For destructive effects the RDM is the ratio of the device's measured
LET threshold (the minimum linear energy transfer, in MeV·cm²/mg,
that triggers the effect) to the worst-case environment LET at the
device location after shielding. The LET threshold must come from
heavy-ion test data for the actual part type; a generic datasheet
figure is not an acceptable substitute. The minimum required RDM is
2.0 unless a higher value has been imposed by the project.
- For non-destructive effects the RDM check is applied to the predicted
upset rate: the predicted rate (events per device per day) multiplied
by the required RDM factor must not exceed the allowable rate set by
the system-level upset budget. Equivalently, the ratio of the
allowable rate to the predicted rate must be at least the required
RDM factor. The predicted rate is derived from the environment LET
spectrum and the device's measured cross-section versus LET curve.
- A part with no heavy-ion test data must be treated as worst-case
(LET threshold = 0 for destructive track; infinitely high predicted
rate for non-destructive track) — the absence of data is itself a
non-compliant finding that must be resolved before design acceptance.
- The minimum required RDM factor is 2.0 for both tracks in the
standard case. Project- or mission-specific requirements may impose
a higher value.
Workflow
- Inventory every SEE-sensitive EEE component and assign each one an
effect type (SEL, SEB, SEGR, SEU, SEFI, or SET). Reject any part
with an unrecognized or missing effect type before it enters the
assessment — an unknown type cannot be routed to the correct track.
- Categorize each part as destructive (SEL, SEB, SEGR) or
non-destructive (SEU, SEFI, SET). Parts on the destructive track
proceed to step 3; parts on the non-destructive track proceed to
step 4.
- For each destructive-track part: retrieve the measured LET threshold
from heavy-ion test data, retrieve the worst-case environment LET at
the device location (from the shielding analysis), and compute the
LET margin = LET_threshold / LET_environment. Flag the part if either
value is missing or non-positive — no margin can be asserted without
both values.
- For each non-destructive-track part: retrieve the predicted upset
rate (events/device/day) from the rate calculation and the allowable
rate from the system upset budget. Compute the rate margin =
allowable_rate / (predicted_rate × rdm_factor). A predicted rate of
zero always passes; a missing or zero allowable rate is a finding.
- Compare each computed margin against the required RDM factor: a
destructive-track part passes when LET_margin >= rdm_factor; a
non-destructive-track part passes when rate_margin >= 1.0 (which is
equivalent to predicted_rate × rdm_factor <= allowable_rate). Parts
that fall short are flagged as non-compliant.
- Aggregate the findings across all parts. The design is SEE-RDM
compliant only when the findings list is empty. Non-compliant parts
must be dispositioned (shielding increase, part substitution, or
system-level upset-budget relaxation with substantiation) before
design acceptance.
Pitfalls
- Routing a part to the wrong track — for example, treating SEL (which
can latch the supply rails and destroy the device if current is not
cut promptly) as a non-destructive rate effect omits the LET-threshold
check that guards against catastrophic single-pass failure.
- Using a LET threshold from a different die revision or process node
than the flight lot — LET sensitivity is a function of the physical
construction; a threshold measured on a predecessor device is not a
valid substitute for the actual flight-lot type.
- Accepting a predicted rate of "essentially zero" as an automatic pass
without reviewing the cross-section data — a very low predicted rate
can result from an underestimated LET spectrum or a saturation
cross-section that was not measured at the high-LET tail.
- Applying the RDM factor as a divisor on the allowable rate rather than
as a multiplier on the predicted rate — both formulations are
equivalent, but mixing them in a spreadsheet produces a factor-of-RDM²
error in the effective margin.
- Treating the absence of heavy-ion test data as a "data gap to address
later" — under §5.1.3 the lack of test data is itself a non-compliant
finding; the part cannot be accepted into the design without either
supplying the data or substituting a qualified part.
Behavior contract (gate 3)
The effect-type categorization, LET-margin computation, rate-margin
computation, per-part compliance check, and multi-part aggregation logic
are exercised by the gate 3 contract test:
scripts/test_e1012_rdm_see.py against
scripts/e1012_rdm_see_logic.py (stdlib unittest, offline). Run:
python3 scripts/test_e1012_rdm_see.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and
paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
1---2name: e1012-rdm-see3description: "Use when determine the radiation design margin for single-event effects (SEE) on spacecraft EEE parts under ECSS-E-ST-10-12C §5.1.3: categorize each SEE-sensitive device as subject to destructive effects (latchup, burnout, gate rupture) or non-destructive effects (upset, functional interrupt, transient), compute the LET-threshold margin for destructive parts and the upset rate margin for non-destructive parts, verify each margin meets the required RDM factor, and flag any part with insufficient margin before design acceptance. Trigger: ecss, e-st-10-system-scope, e-st-10-12c, radiation-design-margin, single-event-effects, see, let-threshold, upset-rate, see-hardness-assurance."4license: Apache-2.05compliance: STANDARDS-REF6standards:7 - id: ecss8 reference-only: true9gated: false10domain: space-systems11pack: space-systems12compatibility: "agentskills.io SKILL.md; any SKILL.md host (Claude Code, Hermes, OpenClaw)"13metadata:14 source: https://skillmd.com/skills/ashfordeou/e1012-rdm-see15 domain: space-systems16 subdomain: ecss17 tags: [ecss, e-st-10-system-scope, e-st-10-12c, radiation-design-margin, single-event-effects, see, let-threshold, upset-rate, see-hardness-assurance]18 version: 0.1.019 author: Aero Agent Skills20---2122# ECSS Radiation Hardness Assurance — RDM Single-Event Effects (space-systems/ecss/e1012-rdm-see)2324Use when the task is to apply the radiation design margin (RDM) approach25for single-event effects (SEE) on spacecraft EEE components under26ECSS-E-ST-10-12C §5.1.3: each SEE-sensitive part is first categorized27by effect type (destructive or non-destructive), then its margin is28computed and checked against the required RDM factor.2930## Domain quick reference3132- §5.1.3 divides single-event effects into two assessment tracks based33 on whether the effect is destructive or non-destructive. Destructive34 effects — single-event latchup (SEL), single-event burnout (SEB),35 and single-event gate rupture (SEGR) — can permanently damage the36 device and require a LET-threshold margin check. Non-destructive37 effects — single-event upset (SEU), single-event functional interrupt38 (SEFI), and single-event transient (SET) — are recoverable and39 require a rate-based margin check. Each part is assigned to exactly40 one track before its margin is evaluated.41- For destructive effects the RDM is the ratio of the device's measured42 LET threshold (the minimum linear energy transfer, in MeV·cm²/mg,43 that triggers the effect) to the worst-case environment LET at the44 device location after shielding. The LET threshold must come from45 heavy-ion test data for the actual part type; a generic datasheet46 figure is not an acceptable substitute. The minimum required RDM is47 2.0 unless a higher value has been imposed by the project.48- For non-destructive effects the RDM check is applied to the predicted49 upset rate: the predicted rate (events per device per day) multiplied50 by the required RDM factor must not exceed the allowable rate set by51 the system-level upset budget. Equivalently, the ratio of the52 allowable rate to the predicted rate must be at least the required53 RDM factor. The predicted rate is derived from the environment LET54 spectrum and the device's measured cross-section versus LET curve.55- A part with no heavy-ion test data must be treated as worst-case56 (LET threshold = 0 for destructive track; infinitely high predicted57 rate for non-destructive track) — the absence of data is itself a58 non-compliant finding that must be resolved before design acceptance.59- The minimum required RDM factor is 2.0 for both tracks in the60 standard case. Project- or mission-specific requirements may impose61 a higher value.6263## Workflow64651. Inventory every SEE-sensitive EEE component and assign each one an66 effect type (SEL, SEB, SEGR, SEU, SEFI, or SET). Reject any part67 with an unrecognized or missing effect type before it enters the68 assessment — an unknown type cannot be routed to the correct track.692. Categorize each part as destructive (SEL, SEB, SEGR) or70 non-destructive (SEU, SEFI, SET). Parts on the destructive track71 proceed to step 3; parts on the non-destructive track proceed to72 step 4.733. For each destructive-track part: retrieve the measured LET threshold74 from heavy-ion test data, retrieve the worst-case environment LET at75 the device location (from the shielding analysis), and compute the76 LET margin = LET_threshold / LET_environment. Flag the part if either77 value is missing or non-positive — no margin can be asserted without78 both values.794. For each non-destructive-track part: retrieve the predicted upset80 rate (events/device/day) from the rate calculation and the allowable81 rate from the system upset budget. Compute the rate margin =82 allowable_rate / (predicted_rate × rdm_factor). A predicted rate of83 zero always passes; a missing or zero allowable rate is a finding.845. Compare each computed margin against the required RDM factor: a85 destructive-track part passes when LET_margin >= rdm_factor; a86 non-destructive-track part passes when rate_margin >= 1.0 (which is87 equivalent to predicted_rate × rdm_factor <= allowable_rate). Parts88 that fall short are flagged as non-compliant.896. Aggregate the findings across all parts. The design is SEE-RDM90 compliant only when the findings list is empty. Non-compliant parts91 must be dispositioned (shielding increase, part substitution, or92 system-level upset-budget relaxation with substantiation) before93 design acceptance.9495## Pitfalls9697- Routing a part to the wrong track — for example, treating SEL (which98 can latch the supply rails and destroy the device if current is not99 cut promptly) as a non-destructive rate effect omits the LET-threshold100 check that guards against catastrophic single-pass failure.101- Using a LET threshold from a different die revision or process node102 than the flight lot — LET sensitivity is a function of the physical103 construction; a threshold measured on a predecessor device is not a104 valid substitute for the actual flight-lot type.105- Accepting a predicted rate of "essentially zero" as an automatic pass106 without reviewing the cross-section data — a very low predicted rate107 can result from an underestimated LET spectrum or a saturation108 cross-section that was not measured at the high-LET tail.109- Applying the RDM factor as a divisor on the allowable rate rather than110 as a multiplier on the predicted rate — both formulations are111 equivalent, but mixing them in a spreadsheet produces a factor-of-RDM²112 error in the effective margin.113- Treating the absence of heavy-ion test data as a "data gap to address114 later" — under §5.1.3 the lack of test data is itself a non-compliant115 finding; the part cannot be accepted into the design without either116 supplying the data or substituting a qualified part.117118## Behavior contract (gate 3)119120The effect-type categorization, LET-margin computation, rate-margin121computation, per-part compliance check, and multi-part aggregation logic122are exercised by the gate 3 contract test:123scripts/test_e1012_rdm_see.py against124scripts/e1012_rdm_see_logic.py (stdlib unittest, offline). Run:125python3 scripts/test_e1012_rdm_see.py126127## Compliance128129- ECSS standards are freely downloadable (ESA); cite the source and130 paraphrase per standards-map.yaml.131- compliance: STANDARDS-REF, gated: false.