by Tim Fernholz
SpaceX sourced valves, critical for controlling the plumbing of a rocket—the piping that carries propellant and the gases used to cool and pressurize the engine—from a company that primarily created the valves that inflated most of the life rafts used by the US Navy. The first high-pressure tanks SpaceX’s rockets used for storing fuel were subcontracted to a Wisconsin company called Spincraft that made metal silos for storing dairy products. Both companies were praised by Musk for their work, but the ever-impatient entrepreneur would soon bring as many tasks in-house as possible when he found that subcontractor delays were putting the Falcon 1 behind schedule. On one inspection trip, he reacted angrily to news of delays, telling a nearby worker that “you’re fucking us up the ass, and it doesn’t feel good.” It was a blast of temper that SpaceX employees would become familiar with when deadlines neared and their boss’s patience was running out.
Yet Musk’s directness, whether he was pleased or displeased, was also a prized asset at SpaceX. Executives and technicians alike soon learned that the CEO of their new company rarely played favorites when it came to business decisions. Early engineers like Bjelde described an atmosphere where the best idea won and “physics defined what was possible and what was not”—not status, price, or politics. Conversely, those who burned out or caught the bad side of a Musk tantrum might say that physics, while important for rockets, wasn’t the right guidebook for managing people. Outside observers who worked with SpaceX would quickly notice a unique culture taking root at the company, where eighty-hour weeks were the norm.
“At Lockheed Martin, it would never occur to a subcontract manager to say, ‘I don’t like these bids. I think we could build this in-house for a third of the price,’” a veteran aerospace executive told me. “They’d be met by a panel of engineers that would say, ‘This company does nothing but make rivets, and you’re telling me we can make an individual single rivet for half the price?’” Conversely, a NASA executive who spent years working with SpaceX remembered that employees would say, “‘Well, we could go buy this from this vendor, but it’s like $50,000. It’s way too expensive; it’s ridiculous. We could build this for $2,000 in our shop.’ I almost never heard NASA engineers talking about the cost of a part.”
At SpaceX, however, Musk expected every single cent of his personal investment to be spent wisely, with speedy results. The engines are the most important and expensive part of any rocket, and at SpaceX they were the responsibility of Tom Mueller. “When I started developing the Merlin engine, the conventional thinking was ‘Only governments can develop rockets,’” he told a group of students later.
Mueller quickly found that, leaving aside price, aerospace vendors just moved too slowly for his boss. “If it takes them two weeks or a month to give you a quote, you’ve already got the wrong vendor,” he said, recalling an estimate he received of “hundreds of thousands of dollars for the part, [and] it’s going to take eighteen months to develop it. I’m like ‘No, I need it in three months,’ and they kind of laugh at you.” Mueller’s team began building parts in-house. They even considered powering the rocket engine with jet fuel, which, due to its wide availability, cost $4 per gallon less than the ultra-refined kerosene—RP-1, or Rocket Propellant-1—they would eventually settle on. The jet fuel just didn’t run right in the engine.
The company began its engine test efforts out in the Mojave Desert, naturally, borrowing an experimental setup from a company called XCOR Aerospace. It was one of the companies competing for the X Prize. SpaceX’s early focus on a small, expendable rocket meant they wouldn’t be competing, but their teams shared a try-anything spirit. The fledgling engineers at SpaceX, however, quickly outstripped the generosity of their compatriots and the patience of local officials with the pace of their testing, and it was clear they would need a place of their own to fine-tune the propulsion systems.
After some consideration, they settled on three hundred acres of land in McGregor, Texas, a remote town outside Waco. It had some history: a failed space start-up, Beal Aerospace, founded by a billionaire Texas banker with a mathematical bent, had set up shop there in the nineties. It had been another attempt to profit from the planned satellite constellations that made it seem that a private launch company was a smart idea. Beal gave up in 2000, having tested a new rocket engine but never launching a vehicle. SpaceX would be able to save money by modifying some of the existing equipment to suit its purposes. Eventually, it would construct its BFTS—Big Fucking Test Stand. One hundred feet tall, with concrete legs ten feet in diameter and extending seventy feet under the earth, the BFTS would test engines that can generate fifteen hundred tons of thrust.
At TRW, Mueller had tested rocket engines at Stennis Space Center, a major NASA facility in Mississippi since the Apollo days. When firing engines, Stennis boasted a crew of a hundred workers in two shifts; SpaceX would test the Merlin rocket with ten people on hand. “It doesn’t take an army of people to run an engine like that,” Mueller said. “I think the government contractors have convinced themselves it does.”
The challenge of building a rocket engine isn’t in the theory but in the practice: optimizing a system that can handle the power needed to lift a rocket without destroying itself is no easy task. Rocket engines rely on components called turbopumps, akin to jet engines, to force propellant into combustion chambers at enormously high pressures and temperatures. The Merlin’s combustion chamber operates at a pressure of more than a thousand pounds per square inch, and the temperature reaches six thousand degrees Fahrenheit—more than three times hotter than the melting point of steel. To avoid a catastrophic meltdown, the engineers lined the chamber with a coating of resin and silica fibers that would absorb heat and flake off, protecting the engine just long enough for it to drive the rocket into space—exactly 160 seconds.
The propulsion team spent ten-day shifts in Texas, bookended by long nights driving back and forth between the McGregor test site and their homes in Los Angeles for intermittent breaks. Sometimes they borrowed Musk’s corporate jet, with an extra passenger cramming himself in the bathroom for the short hop between the two facilities. The effort progressed with frustrating slowness, and Musk didn’t hesitate to let his engineers know that he wasn’t happy with their pace. And in addition, things were going wrong: test stands were blowing up, engines were melting themselves well before their theoretical rocket would arrive in space, and cows were stampeding in circles in nearby fields, in natural abhorrence of such unearthly activities.
The heart of the engine showed SpaceX’s total obsession with simplicity, iteration, and efficiency. It was designed around something called a pintle injector. Mueller is credited with inventing the mechanism, which is the subject of SpaceX’s first and only patent. It’s a technology for mixing the liquid oxygen and kerosene in the engine’s combustion chamber. Rather than push the fuel through a complex part with many small holes—“a showerhead”—the pintle is a simple bolt. It screws into place inside a pipe that injects the fuel into the combustion chamber. When the high-pressure flow of fuel hits the pintle, it sprays out to mix with the oxygen. In their quest to optimize the mix of propellant and fuel for maximum thrust, SpaceX’s engineers could simply swap in and adjust new pintles, rather than remachining the entire engine. That meant they could iterate faster and test new designs to find the most efficient one.
The test program for the engine would run for fifteen months and not be completed until fall 2004. Fine-tuning turned out to be vital to finally reaching full duration. To prevent the engine from melting in the heat of the exhaust despite the protective covering, SpaceX engineers dialed up the amount of oxygen in the combustion mix. This led to a cooler-burning, if less powerful, engine. In the first successful full-duration test, journalist Michael Belfiore joined the team as they watched from a dirt-covered bunker where they could safely monitor the tests (and stampeding cows) with closed-circuit cameras. Mueller gave the go-ahead to fire up the engine and it kicked into flame, shaking the bunker for th
e entire amount of time it takes a Falcon 1 flight to orbit. Amid the cheers, Belfiore recalls Mueller turning to an aide and shouting, “Call Elon! Tell him we just ran a full duration!”
The Merlin engine would go through four major design upgrades in the years ahead, retaining its basic construction but becoming more powerful and reliable. In its current evolution, the engine is one of the most efficient ever designed, using 98 percent of its propellant and boasting a vacuum thrust-to-weight ratio of 180—that is, while it weighs about half a ton, it can generate well over ninety tons of force. If that measurement doesn’t convey the power created by the flame-spouting engines, try, as one rocket engineer suggested I do, imagining a machine hurling ninety tons of bricks out the back. Lockheed did not build a new engine for its Atlas V; Boeing had spent five years on the engine for its Delta IV. But Mueller’s team built a new American rocket engine from the ground up in just over two years.
Once the engine was capable of carrying a rocket into space, the growing team at SpaceX would need to finish building it. This meant the dreaded task of systems integration—putting the engine, the avionics, and everything else together inside the rocket’s structure and making sure it all played nicely. This process began almost a year after Musk had predicted the first rocket would fly.
Meanwhile, though his Life to Mars scheme was in the past, Musk still sought to make a big, public statement about the possibilities of spaceflight. He had decided he would display a full-scale mock-up of the Falcon 1 in Washington, DC, outside of NASA headquarters, in December 2003. This would remind the world that private enterprise was taking on space, and hopefully would help attract media attention—and investors. He was already telling reporters that he was working out the funding needs for another, bigger rocket. The task of building the full-scale mock-up alongside the work of developing the working rocket put serious stress on the company’s employees, who didn’t share their boss’s appreciation for the magical wonder of publicity.
They also didn’t always appreciate Musk claiming credit for the development of the rocket. “At this point I can say I know a great deal about rockets,” Musk told an Australian journalist, who described the entrepreneur as a “self-educated” space expert in the summer of 2003. “I know a lot about Falcon. I could redraw the entire vehicle without the blueprints.” In 2017, he would tell an audience that “we started off with a few people who really didn’t know how to make rockets. The reason I ended up being the chief engineer or chief designer, it’s that I couldn’t hire anyone. Nobody good would join.”
The event in Washington didn’t make the publicity splash that Musk might have expected. On a cold December night outside the National Air and Space Museum, on the Washington Mall, the invited guests—congressional staffers and interns, government officials from NASA and the Federal Aviation Administration, which licenses commercial spaceflights—didn’t spend much time admiring the cylindrical rocket mock-up before heading inside for warmth. Musk read a congratulatory letter from the kooky congressman Dana Rohrabacher, a California Republican known for his interests in space privatization, improved Russian relations, and legal marijuana. An op-ed he published that week had rubbed salt into NASA’s wounds.
“I have witnessed time and again NASA’s overpromising, overmarketing and underestimating costs,” the congressman wrote, calling on more reliance on private space. “NASA goes for the grandiose, ignoring doable, more affordable alternatives.”
In the wake of Columbia, many in the space community were distressed with the slow pace of NASA activity and the lack of anything new. The space shuttle and the construction of the ISS had dominated virtually everything else at the agency since the early nineties. Inside NASA, thousands of talented engineers and researchers pointed the finger back at lawmakers, who, after all, set NASA’s priorities and issued its budget, which were often contradictory. But others saw the new generation of entrepreneur-driven space companies as potential saviors for US aspirations in orbit. In congressional testimony a month prior, a commercial space booster had hailed “the first ripples that will be caused by the new Alt.Space ‘barons’ and their own rocketship projects,” among them Musk and the amusingly erroneous “Scott Bezos of Amazon.com’s Blue Horizons.”
Yet if overpromising was a cardinal sin in the space world, it was also an original sin: everyone shared it. The space shuttle would not fly again until more than two years after the Columbia disaster. Most of the projects attributed to the so-called alt.space barons would be defunct or delayed for more than a decade. Musk, now past his first flight goal, told reporters that the Falcon 1 would launch just four months later, in March 2004. It’s not clear how many visitors at Musk’s sideshow knew the “rocket” in front of them was a mock-up and not a functioning vehicle. Moreover, he wasn’t done raising expectations. Musk unveiled his plans for a new rocket, dubbed the Falcon 5 because it would fly on five Merlin engines, not just one. That rocket, Musk said, would be ready two years later, in 2005.
The next year would turn out to mark a major turning point in the history of private spaceflight, a milestone that would benefit both Musk and his competitors.
The only problem? Musk and SpaceX had nothing to do with it.
7
Never a Straight Answer
The whole culture of program management in the US aerospace and defense industry is today enormously biased toward excessive conservatism. To me, this is an unintended consequence of representative democracy.
—Michael Griffin, former NASA administrator
On October 4, 2004, Paul Allen’s largesse and Burt Rutan’s know-how put the first privately funded, reusable, human-carrying vehicle into space twice in one week. The space shuttle program was still on the ground.
SpaceShipOne was constructed by Scaled Composites, Rutan’s Mojave-based experimental aircraft company. He was the kind of guy who sported silver mutton chops, lived in a hexagonal pyramid house of his own design, and did his own research into the JFK assassination. His company specialized in pushing limits: building a plane capable of a nonstop flight around the world and designing kits for hobbyists to build their own lightweight aircraft.
Scaled will be remembered for SpaceShipOne and winning Peter Diamandis’s $10 million X Prize, demonstrating that a private company could put people into space without government help. None of the other competitors had come close to creating a vehicle that could carry a human being one hundred kilometers up, much less do so twice in a week.
Rutan realized early that anyone fooling around with vertical rockets and space capsules was wasting their time. The Apollo program and traditional human rocketry were simply the wrong inspiration for this contest. But a generation of Mojave astronauts—pilots who tested rocket planes for the US Air Force—offered a different approach. Rutan had begun his career as a flight test engineer at Edwards Air Force Base, helping those pilots figure out how to push the envelope and still survive.
Among the most storied planes in the program was the X-15, a joint NASA/Air Force project that looked more like a ballistic missile than a fighter jet—its pilots had to jettison one of its four tail feathers in order to land the vehicle on a runway. It was carried aloft on the bottom of a B-52 bomber until about 8.5 miles above the ground. Then it would be dropped, fire its rockets, and shoot past the speed of sound. Between 1959 and 1968, eight of its pilots—including future lunar pioneer Neil Armstrong—crossed the invisible line marking the edge of space and were awarded their astronaut wings.”
Rutan decided that the X-15 was the right prototype for a vehicle to win the X Prize. He could build his own space plane, taking advantage of nearly a half century of technological advances as he did so. He could make it safer: the X-15 had claimed the life of one test pilot, Mike Adams, when it entered a spin and broke apart during a 1967 flight. Rutan had an idea for how to improve the design. The greatest danger was experienced while reentering the atmosphere at high speed, when the rocket motor was exhausted but the air too thin for the
wings to keep the vehicle under control. Rutan’s key innovation was a large rotating wing that would “feather” upward as the space plane entered the atmosphere. This would force SpaceShipOne to fly belly first, akin to a falling badminton shuttlecock.
In 2000, after several years of talk, Rutan convinced Allen that this design would succeed. The billionaire funded a joint venture to build SpaceShipOne and win the X Prize. Allen would eventually put $20 million into the project; he hoped to jump-start a new age of commercial space. In the summer of 2004, the space shuttle was still grounded when Scaled Composites’ sixty-four-year-old test pilot Mike Melvill pulled the lever that dropped SpaceShipOne from its mother ship. He rocketed into space for the first time, reaching an altitude of just over a hundred kilometers. They were the first and indeed only of the rag-tag band of X Prize competitors to actually get to space, but winning the prize required they do so twice in fourteen days. They chose the last week in September, just three months later, to make a double attempt, with the requisite media hoopla and VIPs on hand.
