Imagine standing among the small crowd in a viewing gallery aboard a command ship floating 35 miles off the coast of Cape Canaveral. Five miles away you see the upper portion of a rocket bobbing gently, waves lapping at the fuselage. Though isolated at sea there’s a buzz of activity in the air. The conversation on board is punctuated by a loudspeaker booming status updates on the rocket in the water. Finally, an announcement crackles saying that the Cape has given a GO for launch. With NASA’s blessing, the launch director on board begins the final countdown. You hear a distant rumble as a massive engine roars to life under water and gets progressively louder as the rocket starts to rise. In just seconds the staggering, 400-foot behemoth leaves the ocean and disappears from view into the sky. In less than a half hour its 1.1 million pound payload will be in orbit around the Earth.

This was the future Aerojet foresaw with its 1963 proposal for the Sea Dragon rocket.

John Glenn’s launch on an Atlas in 1962. NASA.

Rockets in 1963

In 1963, rockets, particularly those rated for spaceflight, were still very much in their infancy. Though both the United States and the Soviet Union had launched a handful of manned and unmanned missions by this point, both were relying on rockets that were more or less on par with one another in terms of size and ability. NASA, finishing up the Mercury program at the time, was using a version of the Atlas ICBM called Atlas D. The rocket stood 94.3 feet (28.7 metres), had a diameter of 10 feet (3 meters), and a payload capacity of 3,000 pounds (1,360 kg). The Soviets were still using rockets in the R-7 family that had launched both Sputnik in 1957 and Yuri Gagarin in 1961. There were multiple versions for different missions, but the ones in use around time time were equivalent to the Atlas D — they stood between 98 and 112 feet (between 30 and 34 metres), had a diameter of about 9.8 feet (2.9 meters), and could put the same 3,000 or so pounds (1,360 kg) in to low Earth orbit.

Against this technological background the Sea Dragon was a massive leap. This two-stage rocket had only one engine in each stage and stood a staggering 400 feet with a diameter of 75 feet at its widest point. It was heavy, weighing in at about 40 million pounds. It was also more powerful, capable of putting 1.1 million pounds into low Earth orbit. And as one might guess from the name, the Sea Dragon was designed to launch vertically out of the ocean. A final advancement She Dragon had over its contemporaries was reusability; it was designed to be almost entirely reused in an effort to keep launch costs down.

By comparison, the three-stage Saturn V that would make its debut four years later in 1967 stood 363 feet (110.6 meter), had a diameter of 33 feet (10.1 metres), and could put about 310,000 pounds (2.97 million kg) into LEO. It also had more engines: five in the first stage, five in the second, and one in the third. It also used a conventional, ground-based launch pad. This remains, for the time being, the biggest rocket that’s ever flown.

But in spite of the Sea Dragon’s scale, nothing about the project was a deal breaker for Aerojet. In the 1963 proposal the company maintained that the rocket could be made reliable and simply taking advantage of existing technology and shipbuilding techniques.

Artist concept of the Sea Dragon in the water. Aerojet.

Sea Dragon from the Bottom Up

Though it represented a whole new way of launching into space, the Sea Dragon used conventional chemical rocketry. From the bottom up it had a first stage, an interstage, a second stage, a transition stage, a payload stage, then an Apollo-type command-service module that would be responsible for guidance, navigation, and control.

Both the first and second stages used a single, pressure-fed engine. Rather than letting gravity push the fuel and oxidizer into the combustion chamber, these engines used a smaller tank of a pressurizing gas to force both fluids into the combustion chamber, keeping the pressure higher. The first stage, measuring 262 feet (79.8 meters), burned a mix of Kerosene (RP-1) and liquid oxygen (LOX) at a 2.3:1 ratio pressurized with methane to produce an incredible 80 million pounds of thrust.

Atop the first stage was an interstage made of aluminum with reinforcing stiffeners. This served to transmit loads between the first and second stages, and it also housed the separating elements that facilitated staging.

Next came the second stage. It measured 275 feet (83.82 meters) and burned a mix of LOX and liquid hydrogen (LH2) at a ratio of 5:1, a combination chosen for its high performance and a reasonable cost per pound. The second stage also had four auxiliary LOX-LH2 engines pivoted around a single inclined axis that were controlled by electronic actuators. These were used for directional and roll control in flight as well as the extra kick need to get the payload into orbit.

On top of the second stage was a second interstage that would separate it from the payload stage. Finally, an Apollo command module stood at the very top of the rocket.

So at its core, the Sea Dragon was a basic if extremely large and powerful rocket. But its launch environment — 40 miles off the coast of Cape Canaveral in the Atlantic Ocean — presented some unique challenges. But before it could launch at sea the Sea Dragon had to be built.

The Sea Dragon assembled in the lagoon. Aerojet.

Building the Dragon

Since it would be launching from the ocean, it didn’t make sense to build the Sea Dragon on land then pop it in the water. To keep things simple, the Sea Dragon was designed to be built at sea as well, which presented its own challenges. Salt water is both corrosive and electrically conductive, sufficiently dangerous to complex systems like spacecraft that NASA spent USD $1.65 million in the early 1960s — about USD $2.1 billion today — developing a paraglider wing system to get away from splashdown landings. But just because it’s difficult and potentially dangerous doesn’t mean it’s impossible. Submarines are the perfect example — complex vehicles with electrical components that can survive immersed in salt water for months at a time. So Aerojet took inspiration from  military warships. According to the proposal, it was a simple matter of adequate protection. All electrical equipment and components were housed inside a waterproof skin with internal access only granted via portions of the rocket above the waterline. All external wiring would be contained in conduits pressurized with dry nitrogen gas. Once the LOX was loaded in, insulation offered protection against icing.

To physically build the Sea Dragon, Aerojet proposed using classic shipbuilding methods. Large parts like the main stage components would be constructed in existing shipyards and dry docks while smaller parts could be made in offsite factories. Every completed piece would then be moved into the assembly facility near Cape Canaveral, a specially dredge assembly lagoon.

The lagoon was designed to offer “quiet waters” for assembly. It was sort of the wet equivalent of the VAB, the building designed to offer a round-the-clock assembly site for the Saturn V free from weather constraints. The Sea Dragon could come together in this lagoon. A special piece of the launch puzzle would be added at this point as well: a ballast unit.

The ballast unit was the secret to launching the rocket at sea. It was comprised of six cylindrical tanks, support struts, and an opening on the aft end that would fit over the main engine providing a more controlled environment for the thrust chamber firing underwater. This also helped minimize issues of the engine nozzle fluttering with side loads from the ocean environment. But before launch, it helped keep the rocket steady during constriction, and used in conjunction with a floatation assembly and mooring cables, each piece of the rocket could be well controlled while it was mated horizontally.

Aside from it being in the water, the mating and final checkout was pretty standard for a rocket. Systems were checked, and once it was fully together the tanks would be filled from RP-1, LOX, and LH2 storage units nearby. As the final checkout progressed, the tanks could be topped off as needed; LOX boils off at a rate of about 1 or 2 percent per day and LH2 at a rate of 5 percent per day.

Assembled Sea Dragon in the lagoon. Aerojet.

Off to the High Seas

Once the Sea Dragon was fueled and checked out, it was time to leave the safety of the calm lagoon and head for the open ocean. This was a bit of a tricky exercise. The ocean meant waves, and that meant stressed on the rocket’s long frame. There was some inherent protection. Both the first and second stage boasted propellant tanks with strong walls, which, coupled with the tanks being filled and pressurized to about 30 psi, provided the bulk of the vehicle’s structural support. But there was still some danger from the stress of from sagging (waves lifting both ends of the rocket simultaneously) and hogging (a swell lifting only the mid portion of the rocket). Having the rocket fueled and the tanks pressurized during towing helped offset this risk.

Once the Sea Dragon was at its launch point, the ballast unit was slowly filled with fluid — something like drilling mud — to kill its buoyancy. As it filled it sank, and weighing its full 10 million pounds it forced the rocket from an horizontal into an upright position. The manoeuver put the most stress on the rocket. Internal pressure of the fuel and oxidizer tanks as well as the interstage helped keep it strongly rigid, but there was still the danger of sea swells coupling with the heaving motion of the vehicle. But again, a simple rule stating that uprighting was safe with waves smaller than 12 feet high and wind below 25 knots kept the Sea Dragon safe.

Once erect, permanent rails on the side of the rocket gave technicians any last minute access they might need via a small service car that ran up and down the length of the vehicle. This also doubled as a way for crews to board the spacecraft atop the rocket should it be a manned mission; Aerojet imagined that the Sea Dragon would eventually be man-rated. This access car was removed only when the rocket was fully checked out and for launch.

Once everything was ready, the same tug that had brought the Sea Dragon from the lagoon out to sea would manage launch operations working in tandem with the Cape. The rocket’s inertial guidance system would be aligned using a Loran installation on the shore after which point its own self-aligning stable platform keeps the azimuth and vertical directions stable. Though it’s out at sea, the Cape is still in control. But once everything has been aligned and checked out, the Cape gives the command and the tug boat executes the launch of the Sea Dragon.

Sea Dragon’s path from the lagoon to the launch point. Aerojet.

Sea Dragon Rising

The auxiliary engines on the second stage were the first to fire. Then the first stage engine ignited, completely immersed under water. This brought an initial penalty in thrust, but it didn’t last long. The control system could handle any launch dispersions as it left the water in the seconds after the main engine ignited. It was a tricky way to launch. The external pressure on the rocket’s base and engine were far greater than anything a rocket sitting on a launchpad at the Cape would be exposed too. At the moment of ignition the first stage engine was exposed to 150 psi; a rocket at sea level is exposed to about 5psi. Not to mention the ocean provides far less rigid support than a concrete launch pad and hold down arms. But again, the pressurized RP-1 and LOX tanks helped counteract this external pressure from water density, as did pressure inside the combustion chamber, ensuring it fired without flooding. And so the Sea Dragon roared to life and rose out of the ocean.

From there the launch was fairly standard. When the Sea Dragon reached a speed of 280 feet per second (85 mps) it began the pitch manoeuver, nudging the nose downwards by 7 degrees. It retained that orientation as the first stage burned through its available fuel. After 81 seconds of flight, the rocket reached an altitude of 125,000 feet (38.1 km), a speed of 5,800 feet per second (1,767 meters), and had been subjected to about 4.2 gs. At that point the first stage engine cut out and the rocket immediately went through the three-second, pressure-driven staging manoeuver fast enough there wouldn’t be any damaging contact between rocket parts or significant perturbation of the flight path. And because the second stage engine ignited almost immediately after the first shut down there was very little loss of velocity from coasting.

The second stage engine, aided by the four auxiliary engines, burned for an additional 260 seconds at which point the rocket would reach an altitude just over 911,417 feet (172.6 miles, 150 nm, 277.8 km*) and a velocity of 17,630 feet per second (5,373 meters). From there it would all come down to the auxiliary engines. They would burn for an additional 22.4 minutes, a low thrust period but one sufficient to raise the payload’s altitude to 345.2 miles (300 nm or 555.6 km) and tweak its trajectory to reduce eccentricity. Orbit injection would happen about 4,100 nm (4,718 miles or 7,593 km) downrange from the launch points and some 1,344 seconds after the Sea Dragon left the water. When all the engines shut down, the Sea Dragon’s payload would be in a nice, circular orbit.

Though the Sea Dragon was designed for both manned and unmanned missions, the basic payload was imagined as an aluminum tank containing more than a million pounds of LH2 for orbital refueling. The Apollo-style command and service modules would provide guidance, control, and communications between the payload and the Cape. But it could even be simpler. If the mission was just getting that LH2 tank into orbit a Mercury or Gemini-type vehicle could serve as the command pod, providing adequate rendezvous and docking support for a fuel transfer.

Schematic showing Sea Dragon’s flight. Aerojet.

Recover, Recycle, Reuse

Unlike the single-use Atlas of the day, the Sea Dragon was designed to be almost entirely reused. This means that as the rocket was flying and staging, recovery of each piece began as soon as its useful portion of a launch was complete.

Beginning with the ballast unit. Immediately after launch the ballast unit would sink, but once the internal flotation bags were remotely inflated it would start to rise. Once it surface it would be towed back to the lagoon for its next mission.

Next to end its useful life was the first stage. After the engine shut down the propellant valves closed, trapping pressure in the tanks, about 100 psi in LOX tank and 290 psi in RP-1 tank. Empty but pressurized it would coast to a peak altitude of 335,000 feet (102 km) before beginning its fall back to the ocean. To slow the stage for impact, Aerojet studied a few options. A parachute was ruled out because a chute large enough to protect the stage from smashing would have to be 2,700 feet in diameter (822 meters) chute with a 45-foot (13 meter) drogue deploying supersonically somewhere between Mach 1.8 to 2.0, so the odds of a failed deployment were high. A nozzle skirt was similarly passed over because while it was a passive solution — literally adding a flared portion to the base of the rocket stage to slow it for impact — the weight penalty was too high.

The perfect solution was an inflatable aerodynamic decelerator, a highly reliable and easy system. It was a large, conical flare 300 feet (91 meters) in diameter with a half angle of 55 degrees. The flare was a torus 30 feet (9.1 meters) in diameter made rigid with smaller inflatable tubes just 10 feet (3 meters) in diameter. The assembly was made of a rubberized nylon-dacron reinforced fabric with outer skin of ablating rubberized asbestos fabric that could be burned away, sacrificed for the sake of thermal protection up to 1,000 degrees Fahrenheit. Pressurization of the flare comes from the same methane tank as the first stage fuel tanks, and it only needed 30 psia to be appropriately rigid. This would also ensure the stage would hit at the right, apex-down orientation. Between the deceleration, internal pressure, and correct impact orientation, the first stage would survive its reentry. It would experience just 6.5 gs and an impact velocity of 300 feet (91 meters) per second, hitting about 170 nm (195.5 miles or 314 km) downrange from the launch point. Battery operated radio beacons would help recovery crews find it. Once recovered, the stage was vented to a safe stabilizing pressure, the flare partially deflated, and the stage towed back to the lagoon.

Second Stage recovery would be similar, though its return to Earth would begin with small retrorockets in the nose since it would be coming home from orbit. After atmospheric reentry, a skirt similar to the one on the first stage would slow it to the desired impact speed of 210 feet (64 meters) per second. Again, pressure inside the tanks — 50 psi in this case — would protect it from impact. It would also be recovered and towed to the lagoon for refurbishment and reuse.

What a Sea Dragon static fire test at sea would look like. Aerojet.

The Apollo (or Mercury or Gemini) spacecraft guiding the mission could also return to Earth for a safe splashdown, exactly like they did on manned missions.

Once in the lagoon, triage would determine what needed to be replaced and what could be reused. Refurbishment on the tanks would be done with the stages floating in the lagoon, though one-shot valves, paint, insulation, ablative material, and the whole interstage structure would need to be completely replaced after each flights.

After being checked out and cleared for flight, the pieces would be mated again in the lagoon. With this high degree of reuse, Aerojet imagined a single Sea Dragon stage could fly 10 times, potentially completely as many as 100 missions. 

Dragon Slain

Amazingly, even though this Sea Dragon concept was studied after only a handful of humans had left the Earth, no part of the system was deemed so complicated, so big, or so wildly futuristic that a solution couldn’t be engineered. It was a behemoth, yes, but a viable one.

Aerojet planned to use as many existing facilities as possible from shipyards to seaways, though some new infrastructure would certainly be necessary. The Sea Dragon would demand construction of very specific rocket parts as well as cryogenic storage facilities near the assembly point. And the assembly lagoon would also have to be an entirely new facility, though it could use existing dockside equipment and midsize ship service vessels with some special installations added. As a point of comparison, the VAB was built in a little under a year years for $117 million or about $924 million today. Building the lagoon would be equivalent.

Sea Dragon being serviced before flight. Aerojet.

When the report was published in 1963, Aerojet predicted the systems could be operational in just 68 months (a little over five and a half years) at a total cost of $2.8 billion, so about $22.4 billion today. But even though the initial price tag sounds high, it would be more efficient overall. Over an expected 240 flights, the total cost per pound would even out to between $10 and $30 per pound into orbit. And all the systems were designed with that goal in mind. Like the inflatable recovery system. The weight penalty was just 2 percent of the payload ability, negligible in the scheme of things. The only costs associated with reusing the first stage were the costs of towing it 170 miles back to the lagoon and the refurbishment, which worked out to less than the $24 million needed to build a new one. Though if the first stage was expended the new configuration could see the Sea Dragon take an additional 30,000 pounds (13,607 kg)  into orbit.

Even launching at sea wasn’t damaging. Analysis showed that the payload penalty was about 3.2 percent compared to a land launch, so again fairly minor if the rocket could be reused.

But as we know the Sea Dragon never flew. It was ultimately a concept study that didn’t have a place the 1960s landscape. NASA was dominating the space game with support from the military and its focus was firmly on getting to the Moon, not a rocket that could make spaceflight routine without immediate lunar applications.

*The report lists this as 750,000 feet, which I think is erroneous since that doesn’t line up with the nm measurement.Source: Sea Dragon Concept reports (contract NAS8-2599) vol 1 and 3.