Belt Corridors & Transfer Windows

The arrival problem

No Atmosphere, No Oberth Effect: Why Capture at Ceres Costs 4.5 km/s

Arriving at a big planet is nearly free. Arriving at Ceres costs almost everything you brought.

A ship approaching a body carries some excess velocity above escape, written v_∞. The speed it actually reaches at closest approach is

v_p = √(v_∞² + v_esc²)

and the burn needed to capture into orbit is that speed minus the local circular or escape speed. Because the two velocities add in quadrature, a deep gravity well does most of the work for free — the ship falls in, gains speed, and a small burn at the bottom converts a hyperbolic pass into a bound orbit. This is the Oberth effect, and every inner-system mission is designed around it.

Ceres has an escape velocity of 516 m/s. There is nothing to fall into. An arrival excess of 4.86 km/s produces a periapsis speed of 4.89 km/s, so the ship must cancel essentially all of it under power. The recorded figure for the operational geometry is 4.53 km/s of capture against 4.86 km/s of excess — 93% of the approach velocity, paid entirely out of the propellant tank. There is no atmosphere to aerobrake in either, which removes the trick Mars and Earth both allow. Observed

Computed at the surface for a common arrival excess, to isolate the effect of well depth. Real missions arrive with different excesses and capture into orbits rather than at the surface; the operational Ceres figure of 4.53 km/s reflects that geometry. The ordering and the magnitude of the penalty do not change.

Why electric

Specific Impulse Decides the Belt: 21% Propellant Against 79%

The same trip, on chemical propulsion, spends four-fifths of the ship on fuel.

Run a 7 km/s transfer through the rocket equation. At the 3,000 s specific impulse of a gridded ion thruster, the propellant mass fraction is 21%. At the 450 s of a hydrogen-oxygen engine it is 79%. Nearly the entire ship is fuel, and what is left cannot carry a useful payload across a 15-month trip with the spares it needs on arrival.

That is not a preference for efficiency. Every other destination offers an escape route from the arithmetic — an atmosphere to brake in, a deep well to fall down, a short trip that lets a chemical stage stay small. The belt offers none of them, so high specific impulse stops being an optimization and becomes the entry requirement. Electric propulsion is the belt’s price of admission.

Dawn is the ground truth and the whole proof of concept: 11.5 km/s of ion delta-v on 416 kg of xenon, three gridded thrusters, and 7.5 years from launch to Ceres capture. No chemical mission has ever entered orbit around a main-belt body and none has been proposed that closes. A Mars gravity assist is worth about 2.6 km/s on the Earth-to-Ceres route, which Dawn used and which anything larger will use too. Observed

Transit time

Earth to Ceres Takes 15.5 Months, and the Four-Month Trip Does Not Exist

The fast-transit numbers in circulation are off by two orders of magnitude in the thing that matters.

A minimum-energy transfer from Earth to Ceres rides an ellipse with semi-major axis (1.00 + 2.77) / 2 = 1.885 AU. Kepler's third law gives its period as 1.885^1.5 = 2.59 years, and the trip is half of that: 15.5 months. Ceres’s eccentricity moves this between 432 and 513 days depending on where in its 4.6-year orbit the arrival falls, and the delta-v with it, between 10.56 and 11.69 km/s.

Proposals for a four-month crossing are common and none of them survives contact with power-limited propulsion. A 120-day Lambert solution needs roughly 88 km/s of impulsive delta-v. For a power-limited vehicle the payload fraction goes as (1 − Δv / v_ch)² with a characteristic velocity v_ch = √(2ηt / α), where α is the power plant’s specific mass in kilograms per kilowatt. Closing that trip requires α between 0.4 and 1.3 kg/kWe. The best paper studies claim 20. Flight-proven hardware sits near 150.

What does close, at a specific mass around 10 kg/kWe, is a 300 to 360-day transit on a vehicle of 10 to 22 MWe and 670 to 970 tonnes, running at 4,600 to 5,500 s. And the gate on that vehicle is not the reactor. It is rejecting tens of megawatts of waste heat at 800 K, which sets the radiator area, which sets the mass. The belt’s fast ship is a heat-rejection problem wearing a propulsion costume. Modeled

Specific mass α covers the reactor, power conversion and radiators together. The four-month case is listed because it appears repeatedly in mission concepts, and because the gap is a factor of 15 to 50 against the best paper designs rather than an incremental shortfall.

Launch cadence

The Belt Gets More Launch Windows Than Mars, Because It Is Farther Away

7.8 windows per decade against 4.7. Mars is the worst-placed body in the inner system precisely because it is close.

The interval between launch opportunities is the synodic period, set by how fast the two bodies drift apart in angle:

1 / S = | 1/T₁ − 1/T₂ |

Mars orbits in 1.881 years, so 1/S = 1 − 0.532 = 0.468 and windows come 2.14 years — 25.6 months — apart. Ceres orbits in 4.60 years, so 1/S = 1 − 0.217 = 0.783 and windows come 1.28 years — 15.3 months — apart. As a target’s period grows, its synodic period converges on one year, so distant bodies get close to an annual window while nearby ones do not.

Over a decade that is 7.8 Earth-to-Ceres windows against 4.7 Earth-to-Mars. It partly rescues the belt from its own lead-time problem: a program that can only iterate once every 25.6 months learns half as fast as one that iterates every 15.3, and in a regime where reliability has to be bought through iteration, window cadence is a learning rate.

The result is Earth-relative and does not transfer. A route between two non-Earth bodies has its own arithmetic, and the Ceres–Mars pairing is worse than either Earth route: 1/S = 0.532 − 0.217 = 0.314, giving 38.2 months between windows. A tug on that run completes 3.1 deliveries per decade against 4.7 on the Earth–Mars run. Any claim that a Ceres-supplied Mars enjoys better cadence than an Earth-supplied one is backwards. Observed

One coincidence is worth recording because it sits almost exactly at Ceres: the synodic period equals the Hohmann transit time at 2.75 AU. Inside that radius a ship can be caught by the next window before it arrives; outside it, it cannot. The belt straddles the line.

The worst-case spares queue for a belt operation is one window wait plus one transit: 15.3 + 15.5 = 30.8 months. Cutting the transit to 300 days with a high-power electric tug brings that to about 25 months, which is why the tug is uncrewed-program infrastructure before it is a crewed-program enabler.

Departure

Leaving Ceres Is Nearly Free, and That Is the Belt’s Only Real Advantage

516 m/s to escape, against 9.4 km/s from Earth’s surface.

The shallow well that makes arrival expensive makes departure trivial. Escape velocity is 516 m/s, low enough that an electromagnetic launcher — a 610-meter track at 30 g — can throw cargo off Ceres entirely without propellant, at 0.037 kWh/kg against the Moon’s 0.882 for the equivalent job. From the surface to a Mars transfer costs about 2.83 km/s all in, using the launcher to depart and Mars’s atmosphere to arrive.

This is where a claim in circulation needs correcting. Ceres is not closer to Mars than Earth is — not on closest approach (54.5 against 132.0 million km, Earth 2.4 times nearer), not on transit time, and not even on orbit-to-orbit transfer velocity, where Earth is cheaper because the jump from 1.00 to 1.52 AU is larger than the fall from 2.77 to 1.52. Ceres is farther from Mars than Earth is, and cheaper to ship from anyway, because it has no gravity well to climb out of.

The honest form of the advantage is a launch-capacity argument rather than a distance or energy one. Earth spends 1.5 to 2.6 tonnes of mass lifted to low orbit for every tonne that reaches a Mars transfer. Ceres spends none, because nothing was ever lifted. That matters exactly as long as Earth launch cadence is the scarce resource, and stops mattering when it is not. On distance, transit, window cadence and fleet utilization, the Earth route wins every row.

Coming home is the unretired problem. A return from belt distance enters Earth’s atmosphere at 12.84 km/s — the Stardust regime, demonstrated in January 2006 on a 46-kilogram capsule. Scaling that heat shield to tonne class is a qualification program nobody has costed. Projected

The link

Light Time to the Belt: 15 to 31 Minutes, and a Blackout Every 15 Months

Not hours. The figure repeated in most sources is wrong by a factor of three.

One-way light time to Ceres runs from 14.7 minutes at opposition to 31.4 minutes at conjunction. A round trip never exceeds about 70 minutes anywhere in the main belt. Descriptions of belt latency as "hours" overstate it substantially, and the correction matters because the design consequence sits at a threshold: any latency above roughly a second rules out direct teleoperation, so the belt is an autonomy regime regardless, but the difference between 30 minutes and 3 hours decides whether a human supervisor can hold a work session together at all.

Communications also stop entirely when Ceres passes behind the Sun. The outage runs 5.0 days within 3° of solar conjunction and 8.3 days within 5°, recurring every 1.28 years — the same synodic period that sets the launch windows, because it is the same alignment. Every operation must run unsupervised across that gap, with a store-and-forward link either side and no possibility of a command in between.

Local supervision changes the picture and is the reason the crewed architecture puts people in Ceres orbit rather than on Earth: from a station at synchronous altitude, latency to a surface machine is milliseconds. See the presence question on the hub. Observed

Corridor budgets

Belt Corridor Δv and Transit Reference

Indicative budgets for the routes into, out of and across the belt. Full corridor detail on the corridors page, and the Δv calculator takes arbitrary pairs.

The Vesta figure is stated with the 4.93° mutual inclination priced in, which is 1.6 times the number a coplanar transfer atlas gives. Coplanar figures for intra-belt routes understate them consistently and by large factors.

Only one vehicle is currently funded to arrive anywhere in the main belt on any of these corridors. China’s Tianwen-2, after sampling the near-Earth object Kamoʻoalewa in 2026 and releasing its return capsule at Earth in November 2027, continues to the main-belt comet 311P/PANSTARRS for insertion in January 2035. The Psyche orbiter reaches its target in August 2029 but Psyche is an inner-belt body on a separate errand. There is no scheduled traffic on the Earth–Ceres corridor at all.

Record

Provenance and Data Sources