Orientation: sun-pointing vs edge-on

Which way the spin axis should point, and what it costs to keep it there. This topic was debated in five episodes between 2015 and 2026 Habitant, 2026-09-17.

The two options

Option 1: spin axis points at the Sun ("sun-facing", "SSPO") Option 2: spin axis perpendicular to the ecliptic / orbit plane ("edge-on", "rolling around the Sun")
Solar arrays Fixed in the rotation plane, always lit Paul451, 2015-08-30; Paul451, 2022-10-27 Body-mounted on the rim (only ~1/π, i.e. 32–40% effective) Paul451, 2022-10-27, or counter-rotated at station rpm (~2.1 million rev/yr at 4 rpm) Paul451, 2022-10-27
Radiators Out the back, in permanent shadow Inside the ring, shaded by the torus mikelepage, 2022-11-26
Thermal Stable Paul451, 2022-11-28 Predictable; rim panels reduce heating Coastal Ron, 2022-11-26
Pointing cost Continuous torque (see below) Essentially zero; only station-keeping and reboost Twark_Main, 2025-05-01

Key results

The cost of sun-pointing (corrected physics)

  1. 2015–2022 belief: keeping the axis on the Sun needs about 1°/day, "a tiny amount of Δv" Paul451, 2022-10-27.
  2. November 2022 correction (mikelepage, citing Walter Lewin's gyroscope lecture): a spinning body's axis can only be made to follow the Sun by continuous torque. This must be applied parallel to the spin axis, so thrusters are in the right position only once per revolution mikelepage, 2022-11-28. His worked estimates:
  3. a 50,000 t, r = 100 m torus at 3 rpm needs ~2,200 t of propellant per year;
  4. a 2001-size ring (~2 Mt) needs ~216,000 t per year.
  5. So above some size, Option 2 wins. JohnFornaro conceded JohnFornaro, 2022-11-28.
  6. Final correction (Twark_Main, 2025 and 2026): the angular momentum needed for a 360° precession is 2π × the station's spin angular momentum per year, not 4× the spin-up. The vector "walks" around a circle Twark_Main, 2025-04-29; Twark_Main, 2026-09-19. With one firing per revolution at 4 rpm the path is a ~2.1-million-gon. Correcting for the Equation of Time needs thrust adjustments of about 1 part in 3000 Twark_Main, 2026-09-19.

Worked propellant budgets

Station Result
JohnFornaro's ring, 47 billion lb, 2000 yd diameter 408 kN continuous; 815 kN with sinusoidal control; ~3.5 Mt/yr of CH4/LOX (~17,000 Starships) Twark_Main, 2023-09-09
Same ring, with 1000 yd outrigger masts and ±45° bang-bang firing 3.6× less Twark_Main, 2023-09-09
Same ring, reshaped into three 260 yd spheres a further 3× less
John's 6 Mt station, 427 m radius, 1.4 rpm L = 1.6×10¹⁴ kg·m²/s → 3.2×10⁷ N·m → 71 kN at 450 m → ~500,000 t/yr at Isp 450 s Twark_Main, 2025-04-29
6 Mt station, 500 m arm (lamontagne's spreadsheet, corrected) 91 kN continuous, "quasi-infinite" propellant lamontagne, 2023-09-21; Twark_Main, 2023-09-25
~3000 t "tomahawk" with 3:1 mast, r = 300 m, Isp 2500 s 11 N, 14 t/yr, ~270 kW Twark_Main, 2025-04-30
1000 t modular station ~10 t/yr lamontagne, 2026-09-19
20,000 t station ~400 t/yr (3–4 tanker launches): "better to use an equatorial radiator band"
  • Torque scales with the sine of the tilt angle between the rotation plane and the Sun direction. It is zero when the Sun lies in the rotation plane mikelepage, 2022-11-29; mikelepage, 2023-09-11.
  • mikelepage's negative result: a "spinning top" deliberately precessing 10° per day to bounce sunlight into built-in reflectors cost more than yearly sun-pointing mikelepage, 2025-04-28.
  • There is no free precession. A gyroscope floating in zero g does not precess by itself; tidal torques cancel over 360° Twark_Main, 2023-03-09; Twark_Main, 2023-10-11.

Ways to torque without propellant

Axis choices in Earth orbit

  • In LEO, point the axis normal to the orbital plane so no change of axis direction is needed, with panels tracking Barley, 2022-10-27. Nodal regression complicates this for inclined orbits.
  • Peter_GST's GST-01 at i = 23.4° uses an axis normal to the ecliptic, with altitude boosts about every 50 days tied to the J2 regression reset; gravity-gradient torque ≈ 3×10⁻⁴ N·m Peter_GST, 2025-04-10; Peter_GST, 2025-04-16.
  • "1" suggested an axis normal to the orbit inclination instead 1, 2025-04-14.
  • Gravity gradient at 500 km is significant for sun-pointing LEO stations and shifts quickly as the station orbits (1988/1991 NASA studies) lamontagne, 2023-10-31.
  • ISS rotates once per orbit (about 4°/min) to keep its belly to Earth. A spinning station would pay heavily for that Coastal Ron, 2019-04-02.
  • For large stations, a sun-synchronous orbit gives power stability Paul451, 2026-09-17.
  • Lighting geometry with Option 2: use a conical 45° mirror with the axis perpendicular to the ecliptic. This gives solar elevation >0°, and resizing the mirror adapts the design to its distance from the Sun Twark_Main, 2023-02-25.

Consensus that emerged

Open questions

  • Is a counter-rotating shield (Kalpana-type) cheaper over life than propellant for pointing?
  • What is the best pointing strategy for a small LEO station, given gravity gradient, drag and eclipses?

Sources

Forum posts

  • Paul451, 2015-08-30 (reply #574): Arrays across interior, radiators behind, one axis rotation/yr. Post.
  • Coastal Ron, 2019-04-02 (reply #1464): ISS rotates once per orbit. Post.
  • Paul451, 2022-10-27 (reply #2227, 2 likes): 1°/day tiny Δv; gyros don't cancel constant torques. Post.
  • Barley, 2022-10-27 (reply #2229, 4 likes): Axis normal to orbital plane (4 likes). Post.
  • Paul451, 2022-10-27 (reply #2237): Arrays must counter-rotate 2.1 million rev/yr. Post.
  • Paul451, 2022-10-27 (reply #2238): Sun-pointing: fixed arrays, radiators out the back. Post.
  • Paul451, 2022-10-27 (reply #2240, 1 like): Edge-on: rim cells 1/π efficient; radiators inside ring. Post. Attachments: Spin station not pointing at sun.jpg.
  • mikelepage, 2022-11-26 (reply #2447, 1 like): Option 2: radiators shielded by torus. Post.
  • Coastal Ron, 2022-11-26 (reply #2461): Option 2 predictable heating, no orientation motors. Post.
  • Paul451, 2022-11-28 (reply #2475): Option 1 stable thermal environment. Post.
  • mikelepage, 2022-11-28 (reply #2477, 2 likes): Sun-pointing needs continuous torque; ~4× spin-up/yr (2 likes). Post.
  • JohnFornaro, 2022-11-28 (reply #2478, 1 like): Rim thrusters; counter-rotating regolith ring. Post.
  • mikelepage, 2022-11-29 (reply #2493): Torque ∝ sin(angle); zero in Option 2. Post.
  • Twark_Main, 2023-02-25 (reply #3409): Conical 45° mirror with ecliptic-perpendicular axis. Post.
  • Twark_Main, 2023-03-09 (reply #3533): No natural precession without external force. Post.
  • Twark_Main, 2023-09-09 (reply #3910): John's ring precession: 408–815 kN, 3.5 Mt/yr. Post.
  • mikelepage, 2023-09-11 (reply #3914): Near-edge-on shallow angles cheaper. Post.
  • mikelepage, 2023-09-12 (reply #3924, 1 like): Photon-pressure torquing (1 like). Post. Attachments: Photontorque_annot.mp4.
  • lamontagne, 2023-09-21 (reply #4013): Torque math link. Post.
  • lamontagne, 2023-09-21 (reply #4015, 2 likes): Torque spreadsheet 66 t thrust (2 likes). Post. Attachments: Torque calculation.xlsx.
  • Twark_Main, 2023-09-25 (reply #4017, 1 like): Corrected to 91 kN (1 like). Post. Attachments: Torque calculation_revised.xlsx.
  • lamontagne, 2023-09-28 (reply #4042): Avalon stabilizing solar sail 15 km behind. Post.
  • Twark_Main, 2023-10-11 (reply #4153): Tidal torques cancel. Post.
  • lamontagne, 2023-10-31 (reply #4254): Gravity gradient torque at 500 km (Queijo/Garrett). Post. Attachments: 19880010196.pdf, 19910010864.pdf.
  • Peter_GST, 2025-04-10 (reply #4844, 5 likes): GST-01 orbit and orientation (5 likes). Post.
  • 1, 2025-04-14 (reply #4876, 2 likes): Axis normal to orbit inclination; despin thrusters (2 likes). Post.
  • Peter_GST, 2025-04-16 (reply #4878): Gravity-gradient torque 3e-4 N·m. Post.
  • mikelepage, 2025-04-28 (reply #4903, 1 like): Spinning-top precession experiment negative (1 like). Post.
  • Twark_Main, 2025-04-29 (reply #4910): Switchable reflective panels. Post.
  • Twark_Main, 2025-04-29 (reply #4911): Yo-yo pendulum torquer. Post.
  • Twark_Main, 2025-04-29 (reply #4917, 2 likes): 2π × L per year; John's ring 71 kN, 500 kt/yr (2 likes). Post.
  • lamontagne, 2025-04-29 (reply #4918): Back to counter-rotating for sun-facing. Post.
  • Twark_Main, 2025-04-30 (reply #4919): Tomahawk: 11 N, 14 t/yr. Post.
  • Paul451, 2025-04-30 (reply #4921, 2 likes): Waste panels rather than build complex pointing (2 likes). Post.
  • Twark_Main, 2025-05-01 (reply #4923, 1 like): Ecliptic-perpendicular pointing cost ≈ 0. Post.
  • Habitant, 2026-09-17 (reply #5127, 1 like): AI-compiled history of the sun-pointing debate (1 like). Post.
  • Paul451, 2026-09-17 (reply #5128, 1 like): Large stations: sun-sync orbit; VHEO/BEO. Post.
  • MickQ, 2026-09-18 (reply #5129): Tidal tether sinker. Post.
  • redneck, 2026-09-18 (reply #5137): Only far from planets; fragile. Post.
  • Twark_Main, 2026-09-19 (reply #5153): 2π not 4×. Post.
  • Twark_Main, 2026-09-19 (reply #5154): 2.1-million-gon; Equation of Time. Post.
  • lamontagne, 2026-09-19 (reply #5158, 2 likes): ~10 t/yr (1000 t); ~400 t/yr (20,000 t) (2 likes). Post.

External references

  • Walter Lewin, MIT 8.01 Lecture 24 (gyroscopes), youtube N92FYHHT1qM.
  • physics.stackexchange: force to change a flywheel's axis lamontagne, 2023-09-21.