Thirteen lessons, four missions. Can you design a propulsion system that passes every test at once?
There's no multiple choice here. You get a mission, a set of limits, and every trade-off from the course, all fighting each other at the same time.
How a propulsion engineer thinks
Propulsion design doesn't start with the engine. It starts with the mission: how much velocity change, for how much payload, within what mass, over what time, and in what environment. Sutton's closing chapter on selecting a propulsion system follows a consistent chain of reasoning, and you now know every link of it:
- How much Δv, and how fast? The rocket equation turns Δv and payload into a propellant fraction, and the time available decides whether a high-thrust burn is needed or a months-long spiral will do (Lessons 1, 3 and 12).
- Which propellant? Isp against density against storability. Hydrogen gives the most performance per kilogram, needs the biggest tanks and boils away the fastest. Hypergolics keep for years (Lessons 4 and 5).
- Which family? Liquids throttle and restart, solids are simple and dense but can't be switched off, hybrids fall in between, and electric thrusters swap thrust for exhaust velocity (Lessons 5, 9, 11 and 12).
- How do the propellants get into the chamber? Pressurized tanks are simple but heavy at high chamber pressure. Turbopumps are light but have to balance their own power (Lessons 5 and 8).
- Will the chamber survive, and stay quiet? The cooling limits and combustion stability have a veto over every other choice (Lessons 6 and 7).
- Does the nozzle fit, and can it steer? Area ratio buys vacuum Isp, but the bell has to fit inside the vehicle (Lessons 2 and 13).
The studio
Choose a mission, then shape your propulsion system. The drawing updates as you design: the tanks stretch with propellant volume, the bell grows with area ratio and turns red if it won't fit, and the gauges compare your Δv and mass with the mission's needs. Each failing test says why, and links to the lesson that holds the fix. Your designs are saved in this browser.
Design reviews
Stuck? These reviews explain one route to a mission-ready design for each mission. Try on your own first: the struggle is where the understanding forms.
Going deeper
Upper stage: performance in a hurry
The upper stage needs 5.6 km/s in about ten minutes, with enough thrust to keep climbing at 0.7 g. That rules out electric propulsion immediately: no power system could feed that thrust (Lesson 12). The default kerosene gas-generator engine comes close, but the rocket equation punishes its 330-odd seconds.
Hydrogen changes the game. At around 450 s vacuum Isp, a smaller propellant load does the job, even after paying for the bulky hydrogen tank. Keep the area ratio high enough for vacuum performance, but watch the exit diameter against the 2.4 m interstage. A few minutes of storage is no problem even for hydrogen. Methane and kerosene work too if you accept a heavier stage: count how many extra tonnes each costs. This is why so many upper stages, from the Centaur to Ariane 6's upper stage, burn hydrogen.
Going deeper
Lunar lander: control over performance
The lander's hard problems are deep throttling and restarts, not Isp. It has to hover and touch down gently, so it must throttle by 5:1 or more. That counts against solids, which can't be throttled or restarted, and against turbopump engines, whose pumps and turbines are hard to run far from their design point. A pressure-fed engine throttles simply by closing valves. Keep the chamber pressure modest, because the tanks must hold more than that pressure (Lesson 5).
Then look at storage: a week in space is too long for hydrogen, but LOX with methane, kerosene or a hypergolic pair will keep. A hybrid is a genuine option too, because it throttles and restarts on its oxidizer valve (Lesson 11). The Apollo lunar module's descent engine was a pressure-fed, deeply throttleable hypergolic engine for exactly these reasons.
Going deeper
GEO satellite: two ways up
Fifteen years in orbit decides the propellant: only a storable bipropellant or an electric thruster's xenon survives that long. Cryogenic propellants boil away, and a solid motor can't do the station-keeping burns.
- Chemical route: NTO/MMH, pressure-fed, with a small engine of a few hundred newtons to a couple of kilonewtons and a generous area ratio. Most of the launch mass ends up being propellant.
- Electric route: a Hall or ion thruster on a few kilowatts, with an Isp of 1,500–2,000 s. The propellant mass collapses to a fraction of the chemical figure, but the orbit raising takes months. Push the power and Isp too high and the trip time or the power limit fails: that's the Stuhlinger optimum from Lesson 12, and it's why the "all-electric" satellites of the 2010s could launch on smaller rockets.
Going deeper
Asteroid-belt probe: the exponential wall
At 9 km/s, even NTO/MMH's roughly 320 s needs a mass ratio of about 17. There isn't room in 3.5 t for the tanks, let alone 900 kg of probe. The only way out is a much higher exhaust velocity: electric propulsion. At 2,500–3,500 s, the mass ratio drops below 1.5.
Then the constraints are power (at most 25 kW available) and trip time (four years). Raising the Isp saves xenon, but at fixed power it lowers the thrust and stretches the trip. Find the Isp where both fit. That's the same power-limited trade Dawn's designers made.
The whole picture
Follow the propellant one last time, from tank to plume, and notice that you now know why each thing happens.
Propellants are stored as dense liquids or a solid because tank mass is dead weight (Lessons 3 and 5). They are pushed or pumped into the chamber by a cycle that must balance its own power (Lesson 8), and sprayed through an injector into a chamber whose walls survive only because the fuel cools them on the way in (Lesson 6). They burn to a temperature and molecular weight set by thermochemistry (Lesson 4), in a chamber that must not sing (Lesson 7), and expand through a converging–diverging nozzle that turns heat into directed velocity (Lesson 2). The resulting thrust is tilted to steer (Lesson 13), and the rocket equation turns its exhaust velocity into mission Δv (Lessons 1 and 3). A solid does it all with one programmed lump of propellant (Lessons 9 and 10), a hybrid splits the work (Lesson 11), and an electric thruster supplies the energy from outside the propellant (Lesson 12).
All of these pull against one another. That isn't a flaw in the technology. It is what engineering at the edge of physical limits looks like, and you have just done it yourself.
- Momentum is everything. Thrust is mass flow times exhaust velocity, plus a pressure term, and specific impulse measures how well propellant is used.
- The rocket equation is exponential. Every second of Isp and every kilogram of dry mass matters, and staging is the only escape.
- The nozzle converts heat into velocity. It chokes at the throat and expands best to the ambient pressure. Hot, light exhaust is what makes a fast jet.
- Each propulsion family trades the same things differently. Performance, density, storability, control, simplicity and cost.
- Everything couples. Chamber pressure touches the feed system, cooling, stability, nozzle size and mass all at once.
If you can look at a launch vehicle, a lander or a satellite and see the physics behind its propulsion, and the price each choice pays somewhere else, the course has done its job. Ad astra.