O'Neill Slowboat Simulator

• Hull. O'Neill's 1974 cable model: hoop bands carry air pressure plus the spin weight of the interior, the longerons and themselves; longerons carry the pressure on the end caps. Working stress and density: steel 1,034 MPa and 7,850 kg/m³, titanium 750 and 4,430, basalt fibre 800 and 2,100, carbon fibre 1,500 and 1,600, nanotube 30,000 and 1,300. For a 3.2 km steel cylinder with 1.5 m of soil the model gives 7.6 t/m² of wall; O'Neill quoted 7.5. End caps are hemispheres carrying the same interior load.
• One cylinder. O'Neill paired two counter-rotating cylinders, so double everything for his Island Three. A lone cylinder longer than about 1.3 radii is not passively stable in spin and needs active control or a counter-rotating twin.
• Air. A 20 °C atmosphere turning with the hull, thinning toward the axis.
• Shielding. Globus and Strout put the wall needed for 20 mSv a year in deep space at 6 to 7 t/m² of water or 10 to 11 t/m² of lunar rock. Interstellar space is somewhat harsher than that.
• People. Floor area divided by the area you allow each person. Farms have to fit inside that figure unless you stack decks, which adds load. Marin's generation-ship study needed about 900 m² of farmland per person on a mixed diet.
• Drive. Constant thrust, one stage, and the same engine brakes at the far end, so on its own the ship needs twice the cruise speed in total. The rocket equation and both burns are relativistic, and burn times are ship time. No magnetic sail and no staging. Jet power is half the propellant mass times exhaust velocity squared, divided by burn time. That formula returns 40 TW for the first stage of Daedalus, which matches the published figure.
• Laser sail. A perfect mirror pushed by a laser at home. The force is twice the power over the speed of light, falling as the sail recedes. The formulas reproduce Forward's crewed design: 75,000 TW pushing 78,500 t to half of light speed in 1.6 years. The sail is sized by the light it can take. The transmitter is the wider of two sizes: the one that keeps a 1-micron beam on the sail until the boost ends, and the sail's own width, so that its optics take no more light per square metre than the sail does. The sail is cut loose when the boost ends. The ship's own drive still has to brake, because nobody is at the far end to shine a laser. Forward's answer, a detached outer sail that reflects the beam back, needs a ring about as heavy as the ship when the payload is this massive, and is not modelled.
• Specific power. Drive mass is jet power divided by this number, covering reactor, nozzle and radiators. It is the most speculative input here and spans six orders of magnitude between what exists and what Daedalus assumed.
• Energy sources. Each has a share of the reacting fuel's mass that becomes energy, and a default share of that energy that reaches the jet: deuterium with helium-3, 0.39% and 70%; deuterium alone, 0.38% and 45%; deuterium with tritium, 0.38% and 15%; proton with boron-11, 0.08% and 50%; pulse units, 0.38% and 15%; fission, 0.09% and 60%; antimatter, 100% and 10%. Propellant that does not react rides along as reaction mass, so the share that has to react is half the exhaust velocity squared over the energy available per kilogram.
• Antimatter. Catalysed fusion is calibrated on AIMStar's 15 kJ of fusion per 500 million antiprotons, a gain of 100,000, which reproduces that design's 28.5 micrograms. A beamed core has nothing to dilute its exhaust, so the nozzle's effective exhaust velocity fixes its efficiency: 33% of light speed means about 5% of the energy does useful work. Storing antimatter is not costed.
• Reference designs. A named design's specific power is its thrust power over its dry or engine mass, worked out from published thrust, exhaust velocity and mass. NASA's fission-fragment engine is averaged over all the fuel it fissions, not only the fragments that escape. For pulse units the default assumes the best weapons ever built, 6 kilotons per kilogram, and Dyson's pusher-plate geometry.
• Waste heat. Radiators run at 1,200 K and radiate from both faces. The share of wasted energy that lands on the ship is a guess for every drive.
• Gravity assists. A flyby is free fall, so it does not strain the hull: at Jupiter's cloud tops the tide across a 38 km ship is about a thousandth of a g. What falls short is speed. The most a planet can add is its own cloud-top orbital speed, 42 km/s for Jupiter, against the 3,000 km/s of a 1% c cruise. A powered dive past the Sun needs hard thrust for a few hours, which no drive here can give.
• Lights. With no sun, the interior needs about 100 W of lighting per square metre of floor plus 10 kW per person.
• Left out. Whether a closed ecosystem, the machinery and a society can last the trip; dust erosion at speed; and the cost of building any of it.

Sources: O'Neill, The Colonization of Space (1974); Globus and Arora, Kalpana One; Globus and Strout, radiation shielding; Hein and others, World Ships: Feasibility and Rationale (2020); Project Daedalus; Discovery II, Firefly and Hyde figures; Schmidt and others, Antimatter Requirements and Energy Costs (NASA, 1999); ICAN-II and AIMStar; Keane and Zhang, beamed core nozzle; Antimatter rocket; Forward's antimatter factory; Werka, fission-fragment engine (NASA, 2012); Dyson's interstellar Orion; Forward's laser sails; Breakthrough Starshot; weapon yield per kilogram; Rama's dimensions.