The 'deployable = risky' assumption conflates motorized mechanisms with passive stored-energy springs
- Tape-spring (STEM) hinges store strain energy when flattened and spring to deployed position when released—no motors, no latches, no wear mechanisms.
- They've achieved >98% success across 100+ antenna deployments.
- The mental model that 'any deployment is unreliable' is based on 1990s motorized mechanism failures, not 2020s passive deployment heritage.
- Multiple proven paths exist—the constraint interpretation determines which path, not whether a path exists.
The critical question: Does 'no deployable radiators' mean 'no complex motorized mechanisms' or 'absolutely no stowed-then-deployed elements'? If the former, tape-spring panels are your answer. If the latter, sublimation plus thermal mass is your ceiling without architectural changes.
Tape-Spring Deployable Radiator Panels
55-80W in 250g using proven antenna deployment heritage—blocked only by constraint interpretation and dispenser compatibility
Porous Plate Water Sublimator
100W peak capacity using Apollo/ISS EVA heritage—blocked by mission duration vs. water mass tradeoff
- If this were my project, I'd spend the first week on constraint clarification, not hardware design.
- The 'no deployables' constraint is the branching point that determines everything else.
- I'd compile the tape-spring reliability data (>98% across 100+ missions), schedule 30 minutes with whoever owns that constraint, and directly ask: 'Is this about motorized mechanism failure risk, or are you telling me I can't have any stowed-then-deployed elements?' If the answer is 'we're worried about reliability,' I'd show the data and get sign-off for tape-spring radiators that week.
- If deployment is truly forbidden—and sometimes constraints are genuinely firm for good reasons I don't know about—I'd implement eclipse ops scheduling and PCM thermal battery immediately ($0 and $15-30K respectively) to establish a 30-35W baseline.
- Then I'd do the duty cycle analysis for sublimation: how many peaks per day, how long, how predictable? If the math works (and for most imaging/downlink missions it does), sublimation at 283 kJ per 100g is elegant.
- If peaks are too frequent or unpredictable, I'm stuck at 30-35W without architectural changes.
- The one thing I'd avoid is parallel development of multiple hardware solutions.
- Pick the path after constraint clarification, then execute.
- Analysis paralysis in thermal design leads to generic solutions that don't optimize for anything.