space-rha.comShort courseChapter 11 · The assurance decision

Chapter 11 · Lesson 3 of 3 · about 12 minutes

Change triggers and the operations handoff

A mission extension, a new lot, a hotter operating mode or a revised recovery path can invalidate a small but important part of the argument. The response is neither a full restart nor an assumption that the earlier closure still holds. Find the assumptions the change touches and revisit the requirements, analysis and evidence that depend on them.

What you need to know

The assurance argument holds only under its assumptions, and a change to any of them reopens the part of the argument that depends on it. The usual triggers are a new lot or die revision, a higher operating temperature, a changed bias or load, a revised recovery path, a firmware image that alters the scrub interval, a mission extension, a new orbit phase, or new knowledge about a mechanism from another mission's anomaly. The response is a bounded impact assessment rather than a restart. Compare the new baseline with the assessed one, trace which exposure, response, consequence, mitigation and verification items depend on the changed premise (the structure of lesson 11.1 makes this traceable), assign the actions, and record the disposition before the change is credited. Some changes reach further than they look. A change to a common controller or supply can touch several functions whose part data did not change. A die revision changes the part's response without changing the catalogue number. An unchanged calculation is valid only if its unchanged inputs still apply.

A mission extension is the most common change and the easiest to underestimate. It scales the cumulative exposure by the ratio of the durations, which can move a lot bound from acceptable to not acceptable. It scales every expected count the same way. It moves the mission into a different phase of the solar cycle, which changes the galactic and the event terms in opposite directions. And it generally finds the hardware with less margin than it had at launch: the parts have accumulated dose and the recovery paths have a history. The extension decision needs the same argument as the original, with the flown configuration and the flight record (chapter 12) as new evidence.

The operations team inherits the argument, and they can act on it only if it was handed over in a form they can use. The handoff gives them the expected radiation signatures by mission phase (corrected-error counts, interrupt rates, current steps, parameter trends) with their predicted ranges, the detection definitions and counters that will show those signatures, the recovery steps with their timing and side effects, the limits beyond which an observation needs attention, the configuration identifiers that the evidence applies to, and the route by which anomalies, extensions and software changes reach the radiation lead. The monitoring should also watch the assumptions the argument depends on. If the argument assumes a scrub interval of one hour, the telemetry should report that the scrubber runs. Alert thresholds are set from this mission's own rates and detection capability, not copied from a teaching case.

Worked example 11.3. Extending the mission from three years to five

QuantityThree yearsFive yearsEffect on the argument
Mission dose at U315 krad(Si)25 krad(Si)the required dose with margin rises to 50 krad, so the 90/90 lot bound of 25 krad falls further below it; the low-dose-rate offset error grows beyond 12 mV
Expected U2 interrupts2847the uncovered fraction scales with it; the monthly window allocation is unchanged, so the availability claim holds if the watchdog coverage holds
P(U1 high-current event), best estimate0.14 %0.23 %the residual risk statement is reissued with the new band and the solar-cycle phase of the added years
U4 proton fluence1.6 × 10¹⁰ per cm²2.6 × 10¹⁰ per cm²the CTR margin of 2.00 at five years was already computed at that fluence (10.4) and holds
Solar-cycle phasedeclining phasethrough minimumthe galactic terms rise by a factor of two to three and the event term falls; the recorder's heavy-ion SEFI share grows but stays below the proton share

Two items depend on the change: the U3 claim (A1), which was already open, and the U1 residual (A2), which is reissued. The U2 availability claim (A3) holds on its own evidence, and the three-year flight record (chapter 12) now serves as evidence for the watchdog coverage that was previously estimated by injection. The handoff to operations is updated with the predicted ranges for the added years.

Exercise

Record in one traceable place the change, the assumptions it affects, the evidence that depends on them, the actions required, and who has the authority to reopen or close the requirement.

Deliver to operations a monitoring baseline, the expected radiation signatures with their ranges, the escalation route and the list of changes that reopen the argument.

Going deeper

The dependencies set the scope of the reanalysis

A change to a common controller or supply can affect several functions whose part data did not change. A die revision can change the part's response without changing the catalogue number. If the dependency structure of the argument has been kept, the reanalysis can follow the affected premises and stop there. An unchanged calculation is useful only if its unchanged assumptions still apply.

Recomputing the numbers is one task; revalidating the physical and system model behind them is another, and a change can call for either or both.

Flight evidence can challenge the argument

Flight rates can confirm a prediction within its limits, or they can expose a gap in the model or in the detection coverage. Reading them still requires the exposure, the live time, the detection efficiency and event definitions that have not changed since launch. A quiet record from coarse telemetry may say little about the rate of the events it cannot see. Decide before launch how new flight evidence will be reviewed, and do not read a run of successful operations as proof that the parts are immune.

Closure at launch is conditional on the mission staying within the assessed configuration and on the assumptions handed to operations.

The common mistake

A checklist alone does not find every effect of an architectural or process change; an engineer has to review the change against the argument.

The absence of reported anomalies is weak evidence when the detection coverage and the telemetry coverage are unknown.

Takeaway

  • Change control limits the reanalysis to the parts of the argument that the change touches.
  • The argument stays useful after launch only if operations can see its conditions and limits in the telemetry and act on them.

Check yourself

A firmware update changes the scrub interval from one hour to one day to save power. Which claims reopen?

The uncorrectable-word claim (A3.1). The accumulation term rises by about 24 times, while the interleaving term is unchanged. The operations handoff is updated as well, because the expected corrected-error counts per telemetry interval change.

Why does a mission extension change more than the cumulative dose?

It scales every expected count, moves the mission into another phase of the solar cycle, and finds parts with accumulated dose and recovery paths with a history. The argument is rebuilt with the flight record as evidence; scaling the old numbers is not enough.

What does the operations handoff contain that a test report does not?

The expected signatures and their ranges by mission phase, the counters and definitions that will show them, the recovery steps with their timing, the alert limits, the configuration identifiers the evidence applies to, and the route back to the radiation lead for anomalies and changes.

For the extension from three to five years, a colleague scales the U3 mission dose from 15 to 25 krad(Si), notes that the 90/90 lot bound is 25 krad, and concludes that the claim holds. What is missing?

The design margin of 2. The required dose with margin rises to 50 krad, so the 90/90 lot bound of 25 krad, which was already below the 30 krad required for three years (lesson 11.1), falls further below it, and the low-dose-rate offset error grows beyond the 12 mV that already exceeded the 10 mV budget. The U3 claim (A1) was open before the extension, and the extension widens the gap; the dispositions of lesson 11.2, calibration in flight or a replacement part, are the ones on the table.

The extension takes the mission from the declining phase of the solar cycle through solar minimum. Why do the single-event terms move in opposite directions, and what does that do to the recorder's interrupt budget?

The galactic flux is highest at solar minimum, and large solar particle events are more frequent in the active years around maximum, though large events have occurred close to minimum (chapter 1 lesson 3); so the galactic terms rise by a factor of two to three while the event term falls. For the recorder the heavy-ion SEFI share grows but stays below the proton share, so the expected U2 interrupts scale with the duration, from 28 to 47, and the availability claim holds if the watchdog coverage holds. The U1 best estimate rises from 0.14 % to 0.23 %, and the residual risk statement is reissued with the new band and the solar-cycle phase of the added years.

Why does the handoff ask operations to watch the scrubber and the interrupt counter, and not only to report anomalies?

Because the argument holds only under its assumptions, and operations can see those assumptions only through telemetry. If the uncorrectable-word claim assumes a scrub interval of one hour, the telemetry has to report that the scrubber runs (the scrub cycle counter of lesson 12.1). The expected signatures by mission phase, such as the 28 interrupts over three years, are handed over with their predicted ranges so that an observation outside them is noticed, and the alert thresholds are set from this mission's own rates and detection capability rather than copied from a teaching case. The flight record then becomes evidence for the watchdog coverage that was previously estimated by injection.

Further reading

  • K. A. LaBel et al., "Emerging radiation hardness assurance (RHA) issues: a NASA approach for space flight programs," IEEE Trans. Nucl. Sci., vol. 45, no. 6, pp. 2727-2736, 1998. doi:10.1109/23.736521.
  • NASA, "Avionics Radiation Hardness Assurance (RHA) Guidelines," NASA/TM-20210018053, 2021. NTRS 20210018053.
  • ECSS-Q-ST-60-15C, "Space product assurance: radiation hardness assurance, EEE components," 2012.