Rotational stability: intermediate-axis and long-axis instability

The most debated physics topic in the thread. The results below are the ones that survived correction.

Key results

Two different instabilities (do not conflate them)

  1. Intermediate-axis (Dzhanibekov, "tennis-racket") instability. A body rotating about its intermediate principal axis is unstable even when rigid. Rotation about the maximum- or minimum-inertia axis is stable in the rigid-body approximation Paul451, 2019-04-07.
  2. Long-axis (minor-axis) instability. A body spinning about its minimum-inertia axis (a long cylinder spinning lengthwise) slowly transfers momentum to the maximum-inertia axis through energy dissipation (flexing, sloshing), and ends up tumbling end-over-end. This is what happened to Explorer 1 edzieba, 2019-11-05; InterestedEngineer, 2023-10-26. It needs a flexible or liquid medium. Globus's modelling showed it for long cylinders, including O'Neill pairs Paul451, 2026-09-18; Paul451, 2025-04-22.
  3. Paul451 repeatedly complained that "wobble" was being used for both instabilities, and for nutation, precession and CoR offset, "making people dumber" Paul451, 2023-11-08. "1" proposed the terms: ΔL = 0 → "Dzhanibekov motion"; ΔL ≠ 0 (external torque) → "wobble" 1, 2025-04-22.

Design rules that emerged

  • Spin about the maximum-inertia axis, with margin.
  • Rule of thumb from spin-stabilized spacecraft: the spin-axis MoI should be at least 1.2× any other axis (Brown, Elements of Spacecraft Design, AIAA 2002) lamontagne, 2023-09-28.
  • For drum and cylinder habitats, Globus (Kalpana One, revised 2007) requires length along the spin axis < 1.3 × radius; a 100 m diameter drum can be at most 65 m long lamontagne, 2026-09-19; lamontagne, 2023-09-28.
  • Flat endcaps destabilize a cylinder at about 0.85 diameter length Shevek23, 2019-10-25. Inverted endcaps are the most stable and give more land Twark_Main, 2026-09-19.
  • Tori and wheels spin about their maximum axis and are stable. They remain passively stable even as the other two (equal) axes swap, provided something (flexing, water) transfers momentum internally Paul451, 2023-11-08; Paul451, 2025-04-22. N ≥ 3 spokes with habs has the same stability as a disc Twark_Main, 2025-10-25.
  • Batons, barbells and sticks tumble end-over-end about an axis perpendicular to the bar. For a symmetric bar that axis has the same inertia as the other perpendicular axis, so any mass on the spin axis (a docked capsule, a node) creates a "T" that prefers flat spin KelvinZero, 2015-11-12; Paul451, 2019-03-23. Fixes:
  • Put mass in the rotation plane at 90° to the bar: solar arrays, radiators, a third arm. This makes a "frisbee"-stable body Paul451, 2015-11-13; Paul451, 2019-01-07.
  • Use X (4 arms) or Y (≥3 arms) layouts Coastal Ron, 2019-03-23; Coastal Ron, 2019-03-24.
  • Twark_Main's list: hang equipment on one axis, elongate the habitat in one direction, redistribute internal mass, or use ≥3 bolas masses Twark_Main, 2024-08-02.
  • A 3-arm station with one short arm is passively stable (Iz/Iy ≥ 1.2) if the short arm is at least ~60% of the long arms, or ~50% with density-sorted storage and panels. Balance angle: short arm = 2·sin((180° − angle between long arms)/2) Twark_Main, 2024-09-02; Twark_Main, 2024-09-06; Twark_Main, 2024-09-08.
  • Internal mass distribution cannot rescue a thin stick. With all mass at a 3.5 m offset on a 100 m stick the ratio is only 1.0147, because the d² term is small Twark_Main, 2024-09-10.
  • To be passive a 100 m stick would need about 44.7 m of width, "halfway to a torus" mikelepage, 2024-09-12.
  • Twark_Main's estimate for Vast's 100 m stick with its panels: ratio ≈1.01–1.03, not passively stable Twark_Main, 2024-09-20. Expected solution: CMGs applying restoring torque on a sporadically spun stick Twark_Main, 2024-09-19; mikelepage, 2024-09-19.
  • Starship configurations Paul451, 2019-04-03; Paul451, 2019-04-06:
  • Long-axis "hotdog" spin: unstable long-term.
  • End-over-end "tumbling pigeon": stable.
  • Parallel side-by-side pair: flat spin at highest inertia, stable unless the Raptors are ≥80–90% of the mass.
  • A 2025 Starship V4 render spun about the wrong (unstable) axis Twark_Main, 2025-09-14.
  • Non-rotating axial sections (Peter_GST's GST-01 design) add inertia along the spin axis. The rigid-body ratio came out at only 1.05, and 0.99 when ring "soil" was replaced by the same mass of shielding, which is a potential catastrophic mode. The rev2 redesign, with smaller and closer industrial modules, reached 1.2 Peter_GST, 2025-04-21; Peter_GST, 2025-08-31. Paul451's advice: move non-rotating mass close to the rotation plane Paul451, 2025-04-22.

Things that do not change angular momentum

  • No internal process, such as flexing panels or "free spin" tricks, changes total angular momentum. Only external torques (propellant, magnetorquers, gravity gradient, photon pressure) can edzieba, 2019-11-05; 1, 2019-11-07.
  • Large structures are not rigid; they behave like stiff springs. Coupled segments exchange momentum with delay, and underdamped oscillations are possible. A zero-L flexible system can reorient itself 1, 2026-09-19; 1, 2025-04-22.

Tools used

Open questions

  • Can a slightly-too-long drum with active mass shifting deliberately precess, for zero-propellant sun tracking? Paul451, 2026-09-18
  • How much stability margin is needed when large visiting ships dock on the axis of a small station?
  • Do squat, fat drums with nearly equal inertias risk axis exchange during mass movements? Coastal Ron, 2026-09-19

Sources

Forum posts

  • KelvinZero, 2015-11-12 (reply #755): Docked capsule makes T shape preferring flat spin. Post.
  • Paul451, 2015-11-13 (reply #760): Arrays at 90° in spin plane give stability. Post. Attachments: Station solar arrays.png.
  • Paul451, 2019-01-07 (reply #1267): Hang arrays in plane at 90° (frisbee-stable). Post. Attachments: Station solar arrays - 2.png.
  • Paul451, 2019-03-23 (reply #1358): Capsule docked on dumbbell axis is unstable. Post.
  • Coastal Ron, 2019-03-23 (reply #1370, 3 likes): Intermediate axis theorem: X or Y layouts. Post.
  • Coastal Ron, 2019-03-24 (reply #1380, 1 like): Tennis racket theorem explanation. Post.
  • Paul451, 2019-04-03 (reply #1467, 3 likes): Single-Starship spin configurations (3 likes). Post. Attachments: BFS side-docked.png.
  • Paul451, 2019-04-06 (reply #1477, 1 like): Parallel config stable unless engines ≥80–90% of mass. Post.
  • Paul451, 2019-04-07 (reply #1499, 2 likes): Stability = relative moments of inertia; only min/max stable. Post.
  • Shevek23, 2019-10-25 (reply #1679): Cylinder stable if L ≤ D; flat endcaps ~0.85 D. Post.
  • edzieba, 2019-11-05 (reply #1707): Energy dissipation changes axis but not L. Post.
  • edzieba, 2019-11-05 (reply #1710): No net L change without external torque. Post.
  • 1, 2019-11-07 (reply #1721, 4 likes): Potential energy term in rotating frame (4 likes). Post.
  • lamontagne, 2023-01-25 (reply #3096): FreeDyn. Post.
  • Coastal Ron, 2023-01-27 (reply #3132): Wolfram model. Post.
  • Twark_Main, 2023-06-09 (reply #3814, 1 like): highfrontierblog stability posts (1 like). Post.
  • lamontagne, 2023-09-28 (reply #4047): Kalpana One revised paper appendix A. Post.
  • lamontagne, 2023-09-28 (reply #4049, 1 like): Stability ratio ≥1.2 (Brown 2002) (1 like). Post.
  • InterestedEngineer, 2023-10-26 (reply #4235): Explorer 1 lecture video. Post.
  • Paul451, 2023-11-08 (reply #4335, 1 like): 'Wobble' misuse; torus passively stable. Post.
  • Lampyridae, 2024-08-02 (reply #4685, 3 likes): Jensen 2024 stability paper (3 likes). Post.
  • Twark_Main, 2024-08-02 (reply #4694): Four ways to restore passive stability. Post.
  • Twark_Main, 2024-09-02 (reply #4779): Balance formula for 3-arm station. Post.
  • Twark_Main, 2024-09-06 (reply #4783, 3 likes): Iz/Iy vs arm angle; ≥60% short arm stable (3 likes). Post. Attachments: moment_of_inertia_ratio_by_arm_angle.png.
  • Twark_Main, 2024-09-08 (reply #4784): Stability ratio vs short-arm length. Post. Attachments: moment_of_inertia_ratio_by_short_arm_length.png.
  • Twark_Main, 2024-09-10 (reply #4791): Internal mass alone can't stabilize stick: 1.0147. Post.
  • mikelepage, 2024-09-12 (reply #4792): Stick needs ~45 m width → halfway to torus. Post.
  • mikelepage, 2024-09-19 (reply #4800, 2 likes): Mass distribution >1.0 + momentum wheels. Post.
  • Twark_Main, 2024-09-19 (reply #4801, 1 like): Vast likely uses CMGs for restoring torque. Post.
  • Twark_Main, 2024-09-20 (reply #4804): Vast stick MoI ratio ≈1.01–1.03. Post.
  • Twark_Main, 2025-04-18 (reply #4884, 4 likes): High Frontier game as stability simulator (4 likes). Post.
  • Peter_GST, 2025-04-21 (reply #4889, 1 like): GST-01 ratio 0.99 → shelved. Post.
  • 1, 2025-04-22 (reply #4891, 1 like): Terminology proposal. Post.
  • Paul451, 2025-04-22 (reply #4893): Tori stable; IA only when rotating about intermediate axis. Post.
  • Paul451, 2025-04-22 (reply #4894): Move non-rotating mass near rotation plane. Post.
  • Paul451, 2025-04-22 (reply #4897): Long-axis ≠ intermediate-axis; timing of discovery. Post.
  • Peter_GST, 2025-08-31 (reply #4928, 3 likes): GST rev2 reaches 1.2 (3 likes). Post. Attachments: GST01_rev2_small.jpg.
  • Twark_Main, 2025-09-14 (reply #4958): Starship V4 render spins wrong axis. Post.
  • Twark_Main, 2025-10-25 (reply #5026): N ≥ 3 spokes as stable as torus. Post.
  • Paul451, 2026-09-18 (reply #5145): Long-axis instability needs flexible medium; O'Neill pairs affected. Post.
  • Paul451, 2026-09-18 (reply #5146, 1 like): Designed-unstable drum for zero-propulsion sun tracking?. Post.
  • Twark_Main, 2026-09-19 (reply #5151, 2 likes): Inverted endcaps best (2 likes). Post.
  • 1, 2026-09-19 (reply #5152): Large stations behave like springs; zero-L can reorient. Post.
  • Coastal Ron, 2026-09-19 (reply #5157): Squat station worry. Post.
  • lamontagne, 2026-09-19 (reply #5159, 2 likes): Kalpana rule: L < 1.3 R; Kalpana 2 violates it (2 likes). Post. Attachments: 2007KalpanaOne (1).pdf.

External references

  • Brown, C. D. (2002). Elements of Spacecraft Design. AIAA Education Series.
  • Globus, A., Arora, N., Bajoria, A. & Strout, J. (2007). The Kalpana One Orbital Space Settlement Revised. AIAA.
  • Jensen, D. W. (2024). Space Station Rotational Stability. arXiv:2408.00155; Jensen, "Design Limits on Large Space Stations", arXiv:2302.12353.
  • highfrontierblog.com (2014): "Rotational dynamics"; "Inverted endcaps are better"; "Energy loss makes colonies tumble".