Rockets of Today

DRAGON 2 — USA, 2019/2020

The Dragon is, broadly speaking, a traditional capsule with a conical top and a heat shield on the bottom. The side taper is mild, only 15 degrees, making it rather blunt at the top. This gives it a tall profile, which makes sense because it incorporates into its interior a lot of parts which other capsules relegate to an attached service module, so they can be reused. It’s pretty close to the size and shape of an engine bell from a Saturn V.

Relative to SpaceX’s original uncrewed Dragon, it has one major enhancement besides the passenger accommodations: a set of “SuperDraco” rockets around the base, firing at an angle over the rim of the heat shield. (Regular “Draco” is the name of the small thruster motors dotted around the capsule for maneuvering.) A SuperDraco is actually more powerful than the cheap non-reusable Kestrel engine they used for the second stage of their early Falcon 1 rocket, and the Dragon has eight of them. These would function as an escape system in the event of a launch mishap, and in the original plan, would allow the capsule to soft-land on pavement at any desired location, on hydraulic legs which pop out of the heat shield, thereby saving all the costs of having to send out a recovery team to pick it up. They burn hypergolics (monomethylhydrazine), which is unfortunate because this means that after such a landing, you would have to wait a while before opening the door or approaching the capsule from outside, due to the toxicity. (Someone commented that “Hazmat suits might be standard attire for ground crew.”) They are pumpless, requiring the fuel to be pressurized with helium to about 70 bar. The parts are made from the same inconel alloy that most rocket engines use, but are 3D-printed — a technique which is becoming a trend in the industry.

With these built-in engines, the original plan was for it to land without needing parachutes at all, keeping them onboard only as a backup. But NASA thought this sounded too risky for live passengers, and gradually others agreed, and the propulsive-landing idea was shelved. They are omitting the hydraulic legs that were originally part of the design, so the heat shield will be seamless, and now land it with an old-school splashdown into water, despite the inconvenience. Unfortunately a water landing means that the capsule needs more refurbishment than would be necessary if it landed dry, because the salt water might damage many parts, though they do have covers over the SuperDraco nozzles. But a rinse with seawater at least might help mitigate the issues with the toxicity of residual hypergolic propellant... though even after a dunk in the drink, the first astronauts to splash down in a Dragon had to wait extra time inside the capsule because the recovery crew detected a whiff of leaking N2O2 when the capsule was hoisted onto the boat. (The astronauts described riding in the Dragon as smoother than the Shuttle on ascent but rougher on descent, with the worst bit being the opening of the chutes.)

The refurbishment already includes work on the exterior. After a splashdown, the whole outside looks scorched, and even the windows appear smoked and rather opaque. They don’t bother to put a fresh white coat on their used boosters, or even wash much of the soot off as far as I’ve ever seen, but they do make the crew capsules look shiny and fresh each time. How deeply into the protective coating that process goes, I have no idea, but it looks like it might be significant.

This disappointing increase in the refurbishment cost may be one reason why SpaceX is now eager to leave the Dragon design behind, and switch to an entirely reusable spacecraft which can land on pavement, or ideally, land directly on its own launchpad. (Their current plan is to have movable arms on the launch tower catch both the booster and the Starship as they hover... but that sounds pretty risky: a mishap could put the whole launch complex out of commission.)

Even with the popout feet omitted, having these motors built-in gives one key advantage: because they aren’t jettisoned halfway up like a traditional launch escape tower, they can be used for an emergency abort at any time, bringing the crew safely back from any point in the flight, launchpad to orbit. They also weigh a lot less than an external tower. This is consistent with SpaceX’s general philosophy of putting into the reusable craft a lot of stuff which would ordinarily go into an expendable part. But this approach bit them pretty hard when the first Dragon 2 to visit the station was undergoing a static-fire test of its Draco and SuperDraco thrusters after being fished out of the water, as the entire capsule was destroyed in an explosion. This is the third time a SpaceX rocket has exploded; the other two were both due to failures of internal pressure vessels. This one, it turns out, was due to corrosive hypergolic oxidizer leaking through a valve, then being shot out at high speed when the engine was starting up. The fix is to add single-use burstable seals, which means that the SuperDraco engine may now be limited in its ability to shut off safely with fuel still in the tanks. With this fix, they were able to complete extensive retesting without incident, and fly live astronauts.

SpaceX hoped to get about ten flights out of a Dragon 2 capsule before it would need heavy refurbishing. What its lifespan might be with refurbishment is unstated — perhaps they won’t know until they wear some out. But if the water landing refurbishment cost is as substantial as some fear, there may be no point in keeping them in service after a single-digit number of uses. After SpaceX decided to put all of its effort into the Starship and have no further ambition for the Dragon 2 beyond bridging the transition, they dropped the number of expected flights per capsule to five — a number that the Endeavour reached in 2024. At that point NASA and SpaceX got busy on certifying the fleet for more flights. NASA wants them to be available through the decade.

NASA’s early plan was that refurbished Dragons would be used only for launching cargo, not crew, but once it started actually carrying astronauts, they relented and said a reflown one could be used by astronauts. (They said this much earlier for Starliner despite its poorer testing record.) The cargo flights are now handled by separate capsules that do not have seats or SuperDracos. There are so many unneeded parts stripped out that its cargo capacity ends up being about double that of the first-gen Dragon, though unfortunately their capacity for large bulky items is reduced, because since the old Dragon was berthed by the Canadarm rather than actively docking, it had a much wider hatchway.

Maybe there could be a compromise, like using the superdracos to splash down into a pool of fresh water when there are no people onboard. That would save some corrosion and some ocean travel. Or, as Musk has loosely speculated, perhaps it could land in the big shipborne net that they try to catch fairings with. But I doubt any of these will happen, because SpaceX is not interested anymore. And also, they now may not have enough fuel. Somewhere in the transition to a legless design it looks like the tanks got smaller or something. They use some of it for orbit raising — why, I don’t know — so they’re arriving at the station already low. This also leaves them unable to offer orbit-raising services to the space station, so the job instead has to be done by a Cygnus (which is losing its Antares rocket soon), a Starliner (which is struggling to work at all), or a Progress from Russia (which may cease cooperation).

Until re-entry, the Dragon has a rudimentary service module on the back, which SpaceX calls a “trunk”. This is not much more than an interstage — a carbon fiber tube that you can look straight through when it’s detached — but it has solar panels and heat radiators, and can carry 14 cubic meters of unpressurized cargo tied to the inner walls. It has no propulsion. But the difference between a trunk and a true service module isn’t just rocket engines; a typical service module also contains a lot of tanks and batteries and other equipment which does not need to share breathable air with the crew. In the Dragon, this stuff is put under the floor, above the heat shield. The amount of stuff which gets thrown away with the trunk is kept to a minimum. For missions beyond low orbit, they have talked about a plan to build a stretched trunk with added capacity... though I don’t think they have any existing way yet to connect plumbing into it for extra air and water, which would be the main reason you’d want such a thing. The only fluid connection to the current trunk is for coolant. At this point, they are no longer pursuing any interplanetary expansion of the Dragon’s role, leaving that to the Starship.

The trunk has fins on it. If the launch escape rockets have to be used, the trunk stays attached in order to keep the nose of the craft stably pointed forward. In the atmosphere, once the trunk is jettisoned, the capsule tends to promptly swing around so it’s flying heat shield first. (All traditional space capsules behave this way, for safety during reentry.)

In the capsule, the available interior space for passengers is 9.3 cubic meters. (The Apollo command module had 5.9, with some of it inaccessible.) With a Falcon 9 under it, the cargo and passengers for a space station flight can weigh up to 3.3 metric tons. The limiting factor is probably the thrust of the escape rockets; for cargo flights I suspect they could lift a lot more if there was any need to. For returning to the ground, the limit is 2.5 tons, and it will probably be rare to bring back even half of that. ISS garbage can be put into the trunk to burn up, with a limit of 0.8 tons. The capsule’s dry mass is given as just 6.4 tons; with fuel for the Dracos it’s at least eight tons, and all the minor supplies bring it up to 8.9... but that doesn’t include the trunk, which pushes it to 12.5 tons at launch. That’s a good low number for something this roomy.

In passenger mode, it was originally supposed to seat up to seven, as NASA had originally asked for a capacity to equal the Shuttle, but space station flights will only seat two to four, which means that the back row of seats is omitted to make room for cargo. For now, all NASA flights will use this setup, and it’s hard to think of a scenario where all seven would be used. I kind of doubt that they could be added back now, after some later changes to the interior. (Even if only two are flying, there will always be four seats so the capsule can be used as an emergency lifeboat. There are always at least two crew capsules attached to the space station, so everyone will have a seat. Speaking of lifeboats, an Apollo capsule with two extra seats was once built in case a rescue flight to Skylab was needed.)

Dragon 2 is the first crew capsule to ever make a sales point of being designed for passenger comfort. Compared to most spacecraft, the interior looks as slick and shiny as an Apple store. Even the space suits that the passengers would wear for safety are designed for looking slick as well as for functionality. They are so tailored that I suspect there are compromises in freedom of movement.

in operation

The original design had five oval windows in the sides, including one in the door, but NASA allowed only two to stay. There’s also a little window in the docking hatch. This is covered during launch and reentry by a hinged fairing dome. They wanted five windows so they could use the Dragon for orbital tourism flights. On such flights they replace the docking hatch with a glass dome for wide views. The hatch is then reinstalled for space station flights. The second flight of the Resilience, which was only the fourth crewed Dragon flight overall, was the first such tourist flight, dubbed “Inspiration 4”. So SpaceX put paying tourists into orbit just a couple of months after Virgin Galactic and Blue Origin managed their first brief suborbital rides. Another flight, “AX-1”, took a private group to the ISS on the Endeavour, which was the same Dragon to first carry live people (Bob Behnken and Doug Hurley). AX-1 rode the same booster that Inspiration 4 used. Then they added two more craft to the fleet, Endurance and Freedom. They figured that four should be all they need, and didn’t plan to make any more for the time being, but then changed their mind and said they would build a fifth, as yet unnamed, but we haven’t seen anything yet — that might just be intended as a spare. With this fleet they can do flights about six times a year, which is well above current demand, but hardly an order of magnitude leap. This implies that they’re not interested in pursuing a major expansion of private orbital flight in the Dragon, and aren’t going to do all that much beyond what NASA is paying them for. They clearly expect the Starship to be human-rated before there is a need for a larger fleet, which probably would not happen anyway until costs drop quite a bit.

Jared Isaacman, the billionaire who paid for Inspiration 4 (and used it to raise a ton of donations for St. Jude’s), is hiring more Dragon flights. On the next one, they’re going to try a spacewalk. Does this mean replacing the glass dome with an airlock? No, they couldn’t make one fit. This means they have to do like Gemini and pump the air out of the entire cabin so the EVA suit can float out on the end of a hose. SpaceX had plenty of work to do on making a suit that’s usable for an EVA, particularly as this flight will go up to high altitudes where the radiation starts to get rather spicy.

Does this mean that the Dragon will be able to do the kind of satellite maintenance missions that the Shuttle used to do? No, without a proper airlock, and the sorts of heavy untethered suits that would never fit through the Dragon’s hatch, such ideas will have to wait for the Starship.

They had said that once they got crewed flights going, they would put one of these atop a Falcon Heavy and take a pair of unnamed billionaires on a flight around the moon! That would be quite a showoff stunt, as well as proving that space tourism might sometimes be able to make real money. But then they dropped the idea because of the expense of getting the Heavy rated for human flight, when they hope to obsolete it in a few years anyway. They then said they’d put paying moon passengers in a Starship instead. A Japanese fashion billionaire named Maezawa paid for the flight, and apparently this is the same guy who had signed up for the original offer to go there in the Dragon. He recruited artists to come along, and dubbed the mission “Dear Moon”. But after several years, seeing how slow the progress was in comparison to Elon’s promises, he backed out. Anyway, yes, the Dragon 2 was designed to operate well beyond low Earth orbit. And speaking of stunts, SpaceX had also announced plans to land an uncrewed “Red Dragon” on Mars, using the SuperDracos with extra fuel tanks, as a preparatory test flight for their planned Interplanetary Transport System. But they dropped this idea too, for essentially the same reason: they want to put all their ambition into the Starship instead of into their older product lines. Since the Starship is large and light enough to do plenty of aerobraking in the thin Martian atmosphere, they felt that it would not be a useful learning exercise to attempt a more or less purely propulsive landing on Mars, as the Red Dragon would have done.

Musk really doesn’t want to invest any more into anything other than the Starship. I’m not convinced that this is wise. I suspect there are still going to be lots of cases where a little Dragon makes a lot more sense than an airliner-sized Starship.

Records per individual spacecraft through March 2024:

Endeavour: 5 flights with 18 people
(Demo-2, Crew-2, Axiom-1, Crew-6, Crew-8)
Resilience: 2 flights with 8 people
(Crew-1, Inspiration 4... planned: Polaris Dawn)
Endurance: 3 flights with 12 people
(Crew-3, Crew-5, Crew-7)
Freedom: 3 flights with 12 people
(Crew-4, Axiom-2, Axiom-3)

That’s 50 occupied seats on 13 flights (or 14 if you count the empty Demo-1 flight), all successful. There’s a long way to go to catch up to the Space Shuttle:

Columbia: 28 flights with 158 people (7 lost)
Challenger: 10 flights with 60 people (7 lost)
Discovery: 39 flights with 251 people
Atlantis: 33 flights with 196 people
Endeavour: 25 flights with 166 people

That’s a total of 831 shuttle rides on 135 flights, if I’ve counted right — more than twice as many people as have flown on Soyuz, though fewer crewed launches. I’m counting seats occupied at launch, not distinct human beings... many astronauts did several flights each, of course.

The NASA craft of the Space Race era have counts which the Dragon is now passing. Mercury put just four people into orbit, Gemini had 10 flights with 20 seats, and Apollo capsules had 15 flights with 45 seats.