Orbit Insertion at the Giant Planets
Capture Δv and periapsis choice, moon-assisted capture, finite-burn losses, rings and radiation, and aerocapture
Saturn, 1 July 2004
Cassini had been travelling for six and a half years. In the early hours of 1 July 2004 (UTC) it climbed through the plane of Saturn's rings, through a gap between the F and G rings, using its high-gain antenna as a shield against dust. Then its main engine fired for 96 minutes, slowing it by 626 m/s.
The burn was centred on the closest approach to Saturn of the entire mission, about 18,000 km above the cloud tops — a periapsis of 1.3 Saturn radii, just outside the main rings. Light took 84 minutes to bring the news to Earth; the engine had stopped before anyone on the ground knew it had started. If it had failed, Cassini would have flown past Saturn and never come back.
Arriving at a giant planet, a spacecraft has an excess speed of typically 5–6 km/s at Jupiter and Saturn, and anything from 4 to 15 km/s at Uranus and Neptune depending on how fast it came. To be captured, it must remove enough energy at periapsis to make the orbit elliptical. The burn is
Hyperbolic periapsis speed minus elliptical periapsis speed. The semi-major axis a is set by the period of the first orbit.
💡Two rules for a cheap capture
Burn as close as you dare. Both speeds grow as the periapsis shrinks, but the difference shrinks: the Oberth effect again. At Jupiter with v∞ = 5.6 km/s, capture into a 100-day orbit costs 456 m/s at 1.1 R_J, 614 at 2 R_J, 968 at 5 R_J and 1,920 at 20 R_J.
Capture loosely. The first orbit should be as large as the planet's sphere of influence and the solar perturbation allow. Circularising instead would cost 16.9 km/s at Jupiter — more than the entire trip from Earth. Everything after the capture burn is done with moon flybys (the lesson on moon tours), not propellant.
📐Worked Example: what Juno's and Cassini's burns tell you
Run the capture equation backwards (first lab). Juno's insertion on 4 July 2016 (5 July UTC) was a 35-minute burn of 542 m/s, at a perijove about 4,200 km above the cloud tops, into a 53.5-day orbit. The arrival speed that implies is v∞ = 5.63 km/s. Cassini's 626 m/s into a 148-day orbit with periapsis 80,230 km implies v∞ = 5.63 km/s as well.
Juno's plan was to fire again at the next perijove to shorten the orbit to 14 days. A valve problem in the propulsion system led the team to stay in the 53.5-day orbit instead. Going straight to 14 days at insertion would have taken 931 m/s — 389 m/s more than the burn Juno actually made.
Loading interactive…
Let a moon do part of the braking. A spacecraft falling towards Jupiter crosses the orbits of the Galilean moons on the way in. Passing behind a moon — on its trailing side — turns the spacecraft's moon-relative velocity backwards and lowers its energy relative to Jupiter. The turn is only a degree or two, but the relative speed is about 15 km/s, so the velocity change is hundreds of metres per second.
The second lab models Galileo's arrival on 7 December 1995: an Io flyby at 892 km a few hours before insertion. It slows the spacecraft by 198 m/s relative to Jupiter — 'nearly 200 m/s' in Galileo's own mission description — and cuts the insertion burn by 169 m/s. The figure quoted in the mission literature is 175 m/s, about 90 kg of propellant. Europa Clipper uses Ganymede the same way, about twelve hours before its insertion burn on 11 April 2030; in the lab a 200 km Ganymede pass saves about 196 m/s.
🛰Galileo: insertion with a broken antenna
Galileo's main antenna never opened, so every byte from Jupiter came through a low-gain antenna at a few tens of bits per second. On arrival day the spacecraft also had to receive the atmospheric probe's data — the probe had been released five months earlier and entered Jupiter's atmosphere at about 47 km/s — record it, and then fire its engine for orbit insertion. A tape-recorder fault weeks earlier forced the team to cancel the close-up images of Io planned for that day; the Io flyby itself, needed for the propellant budget, went ahead.
A long burn is not an impulse. The capture formula assumes all the Δv is delivered at periapsis. A real engine takes time: Juno's 645 N main engine burned for 35 minutes; Europa Clipper, which has no single large engine, burns for about six hours on smaller thrusters. During a long burn the spacecraft moves away from periapsis, where each metre per second does less good, and gravity has time to work against it.
The third lab integrates the burn for a 6,000 kg spacecraft arriving at v∞ = 5.6 km/s with perijove 4 R_J. The impulsive answer is 739 m/s. With 1,000 N the burn takes 1.1 hours and costs 1 m/s extra; with 250 N, 4.5 hours and 19 m/s; with 150 N, 7.75 hours and 49 m/s (6.7%). The loss grows rapidly once the burn spans a large part of the periapsis passage. The burn should be centred on periapsis: starting it at periapsis instead wastes 30 kg of propellant.
⚠Where you may burn is not where you would like to
The cheapest periapsis is just above the atmosphere, and nobody uses it. At Jupiter the radiation belts between roughly 2 and 6 R_J would destroy unshielded electronics within a few passes; Galileo inserted near 4 R_J after its Io flyby; Juno dives inside the belts at 1.06 R_J on a polar orbit designed to spend as little time in them as possible. At Saturn the rings extend from 1.1 to 2.3 R_S: Cassini burned at 1.3 R_S, above the ring plane, after crossing it through a gap. The capture equation sets the cost; the environment sets the periapsis.
💡Aerocapture
A planet with an atmosphere can do the braking for free: fly once through the upper atmosphere, protected by an aeroshell, lose the excess energy to drag and exit on a captured orbit, then raise periapsis with a small burn. The Δv saving is the whole insertion burn, which at Uranus or Neptune can be several kilometres per second for a fast arrival (3.1 km/s for v∞ = 12 km/s at Neptune in the first lab). The difficulty is the corridor: too shallow and the spacecraft does not slow enough, too steep and it burns up or is captured too deeply. Aerocapture has been studied for Titan, Uranus and Neptune for decades but has not yet been flown; aerobraking, the slow version, has been used routinely at Mars and Venus.
↓But there is a problem
The capture orbit is a long, loose ellipse — 53 days for Juno, 148 for Cassini — that reaches far beyond the moons the mission came to study. To shrink it with the engine would cost hundreds or thousands of metres per second.
Instead, every one of these missions shrank it by flying past moons, dozens of times, on a carefully chosen sequence of orbits. Choosing that sequence is a design problem of its own, with its own graph and its own rules.