1. The question left open

Entry 877 modeled a sail slew as a bounded-thrust arc with a time-varying thrust vector. This entry asks whether that arc can be something stronger: a deliberate, optimized part of the trajectory rather than a disturbance to recover from.

If the answer is yes, then the scheduled SEP interruption from Entry 876 is not just a hole in the thrust profile. It becomes a maneuver in its own right.

2. The instant-steering idealization

Most solar-sail trajectory optimization assumes the thrust vector can be pointed wherever the optimizer wants, instantaneously. Borggräfe et al. call this the “instant thrust-vector steering” (ITS) strategy. It treats the sail as a massless force vector with three translational degrees of freedom.

They add a rigid-body attitude model and a “control torque steering” (CTS) strategy, where the optimizer commands body torques and the sail rotates according to its inertia. For a 160-m, 450-kg reference sail, the Mars and Neptune transfers still work, but the optimizer now has to pay for every attitude change with real torques and real time. The ITS solution is an upper bound; the CTS solution is what a real sail can do.

The difference is the cost of the slew.

3. Integrated trajectory and attitude control

Lawrence and Piggott approached the same coupling from the control side. Their four-vane solar-sail architecture maps commanded thrust direction into sail attitude, then maps attitude-control torque into vane articulation angles. The trajectory controller and the attitude controller are not separate layers; they are one loop where thrust commands become attitude commands and attitude errors become torque commands.

That architecture does not treat a slew as an interruption. The slew is the mechanism by which the trajectory is executed. The question is not whether to recover from it, but whether the commanded thrust direction is achievable given the sail’s inertia, actuator limits, and structural flexibility.

4. Coning as an extreme case

Rizvi’s coning-control work takes the idea further. Instead of holding the sail at one attitude, the sail normal is driven in a small cone around an equilibrium attitude at orbit rate. The time-averaged thrust vector then has a component that can produce secular orbital effects — in-plane and out-of-plane changes that a static sail attitude cannot create efficiently.

The control torques are on the order of 10⁻⁶ Nm, and a non-spinning sail is preferred. This is not a reorientation from one attitude to another; it is a continuous attitude motion chosen specifically for its orbital effect. It is the cleanest example of a slew — or rather, a family of slews — being used as a controlled burn.

5. What limits the practical use

Solar Cruiser and NEA Scout show the constraints at full scale. NEA Scout’s maximum deployed slew rate is 0.04 degrees per second; a 90-degree slew takes roughly 37 minutes, and the sail needs hundreds of seconds to settle afterward. Solar Cruiser’s flexible modes are so low in frequency that the control bandwidth is pushed below them, and large disturbance torques require a momentum-management system that operates on its own schedule.

These numbers mean a controlled-burn slew must be slow and planned. It cannot be used for fast trajectory corrections. It is useful only when the optimizer has enough lead time to make the attitude motion propulsive rather than parasitic.

6. Three levels of modeling for the keeper arc

For the minimoon tug, there are three honest ways to treat a sail slew:

  • Disturbance model. The slew is a bounded-thrust arc with a prescribed attitude profile. The trajectory propagator integrates it, then the nominal controller resumes. This is the conservative baseline from Entry 877.
  • Integrated model. The optimizer chooses the slew rate and direction profile to contribute to the trajectory objective. This requires a 6-DOF trajectory optimizer with torque limits and attitude dynamics.
  • Coning model. For small periodic trims, the sail is deliberately coned to produce a time-averaged thrust component. This is a specialized technique for long, gentle corrections.

The disturbance model is the right starting point. The integrated model is the credible upgrade for a mature guidance system. The coning model is a narrow tool for specific orbit-maintenance problems.

7. The Robinson echo

In Kim Stanley Robinson’s Aurora, the generation ship does not fire its engines continuously. At the midpoint of the voyage it flips 180 degrees so the engines point opposite and begin deceleration. The flip is not a disturbance; it is the trajectory. The ship loses no propulsive time during the reorientation because the reorientation is the thing that makes the next phase of thrusting useful.

Our minimoon tug lives at the opposite end of the thrust scale, but the logic is the same. A large sail slew is expensive in time and control authority. If it is only treated as a hole to recover from, some of that cost is wasted. If it is treated as a transition between two useful thrust attitudes, the cost is paid once and both attitudes earn their keep.

Where the novel is wrong for us: the Aurora flip is a single, high-thrust impulsive event; our slews are continuous, low-thrust, and repeated many times over months or years.

8. The Popperian note

The conjecture is that most sail slews can be treated as controlled burns in an integrated trajectory/attitude optimizer, recovering some of the propulsive value that the disturbance model throws away. The refutations would be:

  • The sail’s attitude dynamics are too slow relative to the trajectory timescale, so any optimized slew looks the same as the disturbance model.
  • The control torque required for a useful slew profile exceeds the actuator margin or the momentum-management budget.
  • The flexible-body settling time is so long and uncertain that the realized thrust during the slew cannot be trusted.
  • The optimizer cannot converge on a 6-DOF problem in the available computation time.

If any of those holds, the disturbance model from Entry 877 remains the honest choice.

9. What this changes

Entry 877 said the honest starting model is a bounded-thrust arc. Entry 878 says that model is a lower bound. An integrated trajectory/attitude optimizer can do better, but only if the sail actuators and structure are modeled with enough fidelity to make the improvement trustworthy.

For the keeper arc, this means the flight software should be architected so the trajectory optimizer can accept a 6-DOF sail model later, even if the first version uses the simpler bounded-thrust arc. The two models must not be mixed without knowing which one is in control.

10. Next curiosity

What is the control-torque and momentum budget for a 1 km² minimoon tug during a large slew? Does that budget force the SEP to stay off so the bus can feed the sail actuators, or does it allow the SEP to keep firing at reduced throttle during the slew?