1. The Epstein drive and the small ship

James S. A. Corey’s Leviathan Wakes gives the solar system the Epstein drive, a fusion engine that lets ships cross the inner planets in weeks instead of months. The desktop will not have an Epstein drive, but the same logic applies: a ship that cannot move is not a ship. It is a habitat, a target, or eventually debris. The propulsion attachment exists to prevent those fates.

Entry 252 closed the structural arc and announced that propulsion is next. This entry defines what the propulsion attachment does.

2. The three jobs

The propulsion attachment has three jobs:

  • Orbit maintenance: counter atmospheric drag and keep the desktop in its assigned orbit.
  • Manoeuvring: perform collision avoidance, phasing, rendezvous, and disposal burns.
  • Attitude assist: provide torques or translation for attitude control when reaction wheels and magnetorquers are not enough.

These jobs overlap with the attitude control system, but the propulsion attachment owns the hardware that expels mass to produce force. The ADCS owns the decision of when and how much to fire.

3. Why propulsion is an attachment

Propulsion could be treated as a platform function hidden inside the body. Calling it an attachment makes several things clearer.

First, it exposes the propellant budget as a consumable resource that other attachments and customers can see. A payload that requires frequent manoeuvring pays for the extra propellant it consumes.

Second, it allows the propulsion module to be replaced or upgraded. A first desktop might use chemical thrusters; a later one might add electric propulsion or receive propellant resupply.

Third, it separates the hardware from the control logic. The ADCS commands thrust; the propulsion attachment provides it.

4. Performance metric: delta-v

The value of a propulsion system is measured in delta-v: the total velocity change it can produce. Delta-v determines what manoeuvres are possible over the desktop’s lifetime.

For a desktop in LEO, the main delta-v consumers are:

  • Drag make-up: roughly 5–25 m/s per year, depending on altitude and solar activity.
  • Collision avoidance: a few metres per second per event, perhaps several events per year.
  • Disposal: roughly 50–100 m/s to lower perigee enough for re-entry within 25 years.
  • Operational manoeuvres: phasing, rendezvous, or constellation management, which can add tens of metres per second.

A practical total delta-v budget for a 5-year LEO desktop is roughly 150–300 m/s. This is enough for station-keeping, a few avoidance burns, and controlled disposal.

5. Propulsion architectures

Three architectures are common for small spacecraft:

Architecture Thrust Isp Best for
Cold gas Low Very low Fine pointing, small delta-v
Chemical monopropellant Moderate Moderate Orbit maintenance, disposal
Electric Very low Very high High total delta-v, slow manoeuvres

For the desktop, a chemical monopropellant or bipropellant system is the right starting point. It provides enough thrust for collision avoidance and disposal without the long burn times of electric propulsion. Cold gas is too low in performance for orbit maintenance. Electric propulsion is attractive for future versions that need large total delta-v.

6. Green propellants

Traditional hydrazine monopropellant systems are reliable but toxic and difficult to handle on the ground. Green propellants such as AF-M315E and LMP-103S offer similar performance with lower toxicity and simpler ground operations. They are increasingly common for small spacecraft.

For the desktop, a green monopropellant system is a sensible choice. It avoids the operational overhead of hydrazine while providing the thrust needed for LEO operations.

7. Integration with structure and power

The propulsion attachment needs:

  • A propellant tank or tanks mounted to the structure.
  • Thrusters mounted on the exterior with clear plume paths.
  • Feed lines routed through the structure.
  • Heaters to keep propellant within its operating temperature range.
  • Valves and pressure sensors monitored by the platform avionics.
  • Power for heaters and valve actuation.

These interfaces must be designed alongside the structure. A tank that fits in the bay but blocks the robot arm’s reach is not a good design. A thruster that fires across a radiator panel will contaminate or overheat it.

8. Safety and failure modes

Propulsion is one of the few subsystems that can destroy the spacecraft if it misbehaves. A stuck-open valve can empty the tanks. A thruster that fires continuously can tumble the desktop. A leaking propellant line can deposit corrosive chemicals on optics and radiators.

The attachment must include:

  • Isolation valves upstream of each thruster.
  • Limit switches or encoders to confirm valve state.
  • Software limits on burn duration and total impulse.
  • Physical guards around thruster nozzles to prevent damage to adjacent surfaces.
  • A safe mode that closes all valves on detection of an anomaly.

What this changes

  • The propulsion attachment provides orbit maintenance, manoeuvring, and attitude assist.
  • It is a replaceable module with a visible propellant budget.
  • A first desktop needs roughly 150–300 m/s of delta-v over its lifetime.
  • Chemical monopropellant, preferably green propellant, is the baseline.
  • The design must integrate tanks, thrusters, feed lines, and heaters with the structure.
  • Safety features include isolation valves, state confirmation, and burn limits.
  • The next entry can estimate the desktop’s delta-v budget in detail.