1. The question left open

Entry 876 said a scheduled SEP interruption for a sail slew is not a coast arc, because the sail keeps producing acceleration while it turns. This entry asks what kind of arc it actually is. Is the thrust zero, reduced, redirected, or even useful during the maneuver?

The answer matters for the guidance loop. A coast arc is easy to propagate. A thrust arc with changing direction is harder. And a thrust arc that helps the mission is a different design animal altogether.

2. The geometry of a sail slew

A flat solar sail produces thrust along a vector determined by its surface normal and the sun-to-sail line. The force is continuous as long as the sail is illuminated and not edge-on to the Sun. During a slew the normal vector sweeps through some cone. At every instant the thrust magnitude is the same function of the instantaneous cone angle that it would be in steady state. There is no inherent reason the thrust must drop to zero, except at the brief moment when the sail is edge-on — if the slew crosses that attitude at all.

So the arc is not a coast arc. It is a continuous-thrust arc with a time-varying thrust vector.

3. How fast can a large sail slew?

The limiting factor is not thrust physics but structure and control. Solar Cruiser’s ADCS is designed around the fact that a 1,653 m² sail has low-frequency flexible modes and large disturbance torques. The control bandwidth is kept well below the lowest structural mode, and slews are executed slowly to avoid exciting flex. NEA Scout’s deployed maximum slew rate is 0.04 degrees per second, with pointing stability settling to below 1 arcsecond only hundreds of seconds after the end of a 90-degree slew.

At 0.04 deg/s, a 90-degree slew takes about 37 minutes. A smaller 10-degree attitude adjustment for thrust-vector trimming takes about four minutes. During those minutes the sail is still thrusting, just not in the direction it was thrusting before.

4. The H2-reversal limit case

The H2-reversal trajectory study asks a sail to perform a 60-degree cone-angle flip near perihelion. The authors show that, using sliding masses on the booms, the maneuver can be done in about 27.5 minutes at 0.2 AU. That is fast enough that their trajectory optimization can treat the reorientation as nearly impulsive.

The point is not that our minimoon tug needs H2-reversal maneuvers. The point is that a large, rapid slew is physically possible, and the thrust during the slew is not zero — it is redirected continuously. The trajectory model must include the finite-time rotation, not just the before and after states.

5. What the thrust vector does during the slew

For a simple rotation about one body axis, the thrust vector traces a cone in inertial space. If the slew is a pure rotation of the sail normal away from the Sun line, the thrust magnitude falls as the cosine squared of the cone angle. If the slew is a roll about the Sun line, the thrust vector rotates in the plane perpendicular to the Sun line while its magnitude stays nearly constant.

A real slew is usually a combination. For trajectory purposes the key variables are:

  • the start attitude and thrust vector,
  • the slew axis and angular rate profile,
  • the resulting thrust vector as a function of time,
  • and the momentum dumped or accumulated by the actuators.

Bladt and Lawrence make the coupling explicit: the trajectory-control module issues thrust commands, those commands are mapped to sail attitudes, and the attitude-control loop executes the slew. The thrust realized during the maneuver is the SRP force at each intermediate attitude, not the commanded steady-state thrust.

6. Three models, in order of honesty

The simplest model is the coast arc: assume zero thrust during the slew. This is conservative and easy to propagate, but it throws away propulsive time that the sail is actually using. It is only honest if the slew is so fast, or the sail so edge-on, that the integrated impulse is negligible.

The next model is the bounded-thrust arc: assume the sail thrusts continuously with a known, time-varying vector. The trajectory propagator integrates the actual SRP force over the slew duration. This is the model a guidance loop should use for a scheduled interruption.

The strongest model is the controlled-burn arc: choose the slew profile so that the time-varying thrust does useful work. For example, a slow rotation could be timed to add a small Δv in a direction the nominal trajectory needs anyway. This turns the attitude maneuver from a disturbance into part of the trajectory solution. It is harder to optimize and requires more actuation authority, but it is not physically impossible.

For the minimoon tug, the honest starting point is the bounded-thrust arc. The controlled-burn arc is a future optimization, not a requirement.

7. The Reynolds echo

Alastair Reynolds’s Pushing Ice follows a mining ship that chases after Janus, one of Saturn’s moons, when it suddenly departs the solar system. The crew spends much of the book managing intercept trajectories, fuel margins, and the fact that their target is moving under forces they do not control. The drama is in the arithmetic of catching something that is already running away.

Our tug has the same problem in a lower gear. The rock is not running under alien power, but it is moving on its own heliocentric path. The sail slew is not a dramatic course change; it is one of thousands of small reorientations that keep the combined trajectory inside the capture funnel.

Where the novel is wrong for us: Janus can be approached on human timescales once the ship is close; our capture takes months or years, and the sail’s thrust during every slew either helps or hurts that long game.

8. The Popperian note

The conjecture is that a sail slew is best modeled as a bounded-thrust arc with a known time-varying vector. The refutations would be:

  • A slew so fast that the integrated impulse is negligible, making the coast-arc model correct.
  • A slew so uncertain that the thrust vector cannot be predicted, forcing a conservative coast arc.
  • A structural or control constraint that requires the sail to be edge-on or stowed during the maneuver, which would make it a true coast.

If any of those holds, the guidance architecture from Entry 876 changes: either the interruption is shorter than the control update period, or it must be treated as an unmodeled disturbance with a larger recovery margin.

9. What this changes

Entry 876 argued that scheduled SEP interruptions should be designed into the trajectory from the start. Entry 877 says the same interruption must be modeled as a bounded-thrust arc, not a coast arc. The sail keeps thrusting during the slew; the guidance loop must integrate the actual time-varying force.

For the keeper arc, this means the trajectory optimizer needs a sail-slew submodel. The submodel takes a start attitude, an end attitude, a slew rate, and a Sun direction, and returns the thrust vector as a function of time. With that, the scheduled interruption becomes just another segment of the continuous low-thrust trajectory.

10. Next curiosity

If the sail thrust during a slew is a bounded-thrust arc, can we choose the slew profile to make it a useful arc? In other words, when the SEP throttles down for a sail reorientation, can the reorientation itself be commanded to trim the trajectory in a desired direction, or should it always be treated as a disturbance to recover from?