by Paul Allen
My hands-on rocket experiments began with old-fashioned kitchen matches wrapped in foil, perched on bent paperclip launch pads. I played with Jetex glider kits powered by guanidine nitrate pellets: minimal thrust, lots of hissing exhaust. For more excitement, Doug Fullmer and I strung a string across the street and fixed a bottle rocket to one end with a lit cherry bomb attached. If we timed it just right, the cherry bomb would explode over a passing car, startling the motorist and sending us diving into the bushes.
I had less success when I tried to launch the arm of an aluminum lawn chair by stuffing it with powdered zinc and sulfur and setting it atop a coffee pot. With my cousin Chris in rapt attendance, I lit the fuel. My rocket sputtered and shuddered before toppling over and melting. My ultimate Wernher von Braun moment came in Doug’s basement, when we tried to make a rocket fuel called Grandma’s Fiery Molasses. We heated potassium nitrate and sugar into a slurry, using Doug’s father’s blowtorch instead of a Bunsen burner. That might have been too hot, because the slurry ignited until the flames licked the ceiling. To our relief, the fuel burned itself out without torching Doug’s house. We kept it a secret and didn’t repeat the experiment.
Other enthusiasms came and went, but my obsession with rocketry endured. At age sixteen, I sat with my mother and sister to watch Apollo 11’s lunar module land in the Sea of Tranquility. Six hours later, we saw Neil Armstrong walk on the moon. That night I went outside as I had when Yuri Gagarin flew, my eyes fixed on the pale disk above. In wonderment I mused, There are people from earth up there, walking around.
After Apollo, NASA shifted to unmanned probes. Space lost its cachet, but it never lost my interest. In the spring of 1981, in the middle of Microsoft’s frenetic work on Project Chess, Charles Simonyi suggested that we fly to Florida for the maiden launch of the space shuttle Columbia. Neither of us had seen a live liftoff; Charles had missed Apollo 8 by half an hour after making a wrong turn in Georgia. I was all for going until we realized that the launch was set for Friday, April 10, the date for Microsoft’s first-ever companywide meeting. Instead of joining the throng at the Kennedy Space Center, we’d be stuck in a Red Lion Inn in Bellevue.
Then some kismet: The launch was postponed to Sunday because of a software glitch. We reached the NASA causeway before dawn. At seven, there began a deep rumbling that became more than sound, until the air was literally vibrating. We could see the orange glow as the rocket fired, and then I could feel its heat on my face. As the Columbia headed straight up, and the crowd chanted “Go!” I got a lump in my throat. (Rocket takeoffs get me every time, even on television.)
Bill wasn’t happy that we’d taken the weekend off with IBM’s deadlines still pending. But I didn’t second-guess going. I had seen a rare thing; I’d seen history.
IN SEPTEMBER 1996, I flew to Mojave, California, to meet Burt Rutan, president of Scaled Composites and the renegade genius of modern aerospace engineering. Burt was tall and thick-chested, dressed in denim. He had a mane of silvery hair, muttonchop sideburns, a desert tan, and a fanatic’s glint in his eye. He’d designed more than thirty unorthodox experimental aircraft, mostly with lightweight carbon composites, and had set a raft of world distance and duration records. In 1986, his Voyager became the first plane to be flown around the world without landing or refueling.
Burt had already begun thinking about a supersonic plane that could fly above the atmosphere. Two years later, over lunch in Seattle, he floated his plan for a manned rocket flight into suborbital space. Burt wanted to demonstrate that you didn’t need NASA-level resources to create a commercial space tourism industry and bring ordinary people to the same black sky that once greeted Alan Shepard.
I had a narrower goal. I wanted to do something in rocketry that no one had done before. I wanted to do it with Burt because none of his designs had crashed during testing. In government-funded spaceflight, there was a historical 4 percent fatality rate. For space tourism to succeed, that risk needed to be cut to no more than one in five thousand, comparable to the early airline industry.
Burt had a lot going for him: a crack engineering team and a brilliant body of work, plus the aura of confidence that marks great innovators. But the right design still eluded him, and so our project was tabled.
* * *
IN 1965, Burt was hired as a flight-test project engineer at Edwards Air Force Base just south of Mojave, the place where Chuck Yeager first broke the sound barrier. Burt’s stay overlapped with the government’s research program for the X-15, the only winged aircraft to fly into space. It was launched not from the ground, but from a B-52 mother ship, high in the atmosphere. The idea was to avoid the most dangerous phase of a ground launch, the first seconds, when there’s no way to abort after liftoff without creating a fireball. (If a rocket starts instead at 50,000 feet, and then something goes wrong, you can dump propellant and still have time to glide down for a safe landing.) Beyond being safer, the two-plane concept avoided the conventional one-and-done missiles that would have made the rocket’s price prohibitive. Yet another plus: In thinner atmosphere, you use less fuel, which allows for a smaller rocket.
But if the going up seemed relatively straightforward, the coming down was something else again. In 1967, during Burt’s stint at Edwards, an X-15 test pilot named Mike Adams was killed during reentry after a spin at Mach 5 (five times the speed of sound) broke his plane into pieces. Burt was committed to finding a more dependable design—something at least as safe as NASA’s Mercury model, where the astronaut sat inside a capsule at the tip of a missile and parachuted back to earth. His first big idea, which he called a “carefree reentry,” was to “feather” the capsule like a badminton shuttlecock. The resulting drag would decelerate the craft faster and minimize heat buildup as it returned to the atmosphere.
But the capsule-shuttlecock approach had drawbacks. You couldn’t land the spacecraft without a parachute. And you couldn’t market flights for tourists if you needed search-and-rescue missions to find errant capsules. Besides, parachute mishaps were all too common; a Russian cosmonaut died in the first Soyuz flight in 1967 after his chutes failed. Burt needed a craft that could reliably survive reentry and yet enable a pilot-controlled, horizontal landing on a runway. Which brought him back to square one: a winged airplane.
One morning in 1999, Burt had his eureka moment: The wings themselves could act as the feather! SpaceShipOne was effectively two planes in one. During the boost phase and ascent, the wings would be configured normally. But for the supersonic phase of the descent, their rear halves would fold upward along a hinge at a 65-degree angle, creating the high drag needed for a carefree reentry. Back in the atmosphere, the plane would revert to its original configuration and become a pilot-controlled glider, along the lines of the X-15. From a safety perspective, this was the best of both worlds.
Burt tested several feathered-wing designs by tossing Styrofoam and balsa-wood models off the Mojave control tower. In the summer of 2000, after he found one that passed his tests, we reached an agreement. Under its terms, Burt would develop the design and build the two planes in return for a minority equity stake. He’d later call me the perfect customer because I deferred to him on all mission-critical decisions. That’s a core element of my management philosophy: find the best people and give them room to operate, as long as they can accept my periodic high-intensity kibitzing.
At the start, my goal was to get the first privately supported astronauts to space and back, thereby kicking off a new era of commercial spaceflight. Though the X Prize for this feat had been announced four years earlier, its financing seemed shaky and we weren’t counting on it. But in 2002, not long after Burt and I signed our contract to form Mojave Aerospace Ventures (MAV), the Ansari family purchased a “hole-in-one” insurance policy to guarantee the $10 million prize. To meet its criteria, Burt modified his design from a solo pilot craft to a pilot plus two passengers. The cost estimate more than doubled, from $9 million to $19 million, and I knew it was unlikely to stop there
. Based on what I’d heard about bleeding-edge aircraft, I expected SpaceShipOne to come in overweight, underpowered, over budget, and behind schedule.
The remarkable thing about MAV was that it built a manned space program from scratch with a staff that averaged around thirty people. And they didn’t just engineer a spacecraft; they also built the launch airplane, flight simulator, avionics system, and rocket motor test facility. I flew down periodically for project reviews as White Knight, the mother ship, was prepped for its first go in August 2002. Looking like a catamaran with wings, White Knight was both launch vehicle and pilot training platform. All of its control systems, down to the pattern of porthole windows in the cockpit, were identical to those on our rocket ship.
My anticipation grew as SpaceShipOne’s carbon-fiber design took shape like something out an old sci-fi magazine. With a mere twenty-seven-foot wingspan, it weighed less than a Honda Civic. As X Prize Chairman Peter Diamandis would say, it was “a spaceship that could fit in a two-car garage.” Unlike the space shuttle (but like the X-15), SpaceShipOne would fly through the sound barrier without active computer assistance. Early on in his career, Burt couldn’t afford wind-tunnel testing, and so he relied on carefully designed flight test programs, computational analyses, and supremely skilled pilots. Although SpaceShipOne’s avionics would monitor trajectory and projected altitude and warn about potential problems, the pilot commanded the controls: the manual stick and rudder pedals, the gas jets that adjusted attitude in space, even the rocket boost cutoff. “This,” Burt said, “is really out there.”
Scaled Composites’ first effort to build a rocket was a tricky proposition. For safety and flexibility, Burt employed a hybrid rocket engine, the first ever used in a manned spacecraft. At ignition, liquid nitrous oxide would flow from a tank into a motor filled with synthetic rubber, causing a controlled burn. As the rubber was consumed, high-speed gases would spurt out of the motor’s nozzle and propel the ship. If the pilot needed to shut down the motor, he could simply cut off the nitrous flow and dump what was left.
On February 1, 2003, five months before our first flight test, the space shuttle Columbia broke up on reentry over Texas. All seven people aboard were killed, a harsh reminder of the risks we faced.
GREAT LEAPS IN aviation have long been spurred by cash incentives, going back to the $25,000 Orteig Prize won by Charles Lindbergh in 1927. Competitions tend to invigorate people’s ambitions. They’re also laced with cautionary tales like Admiral Byrd, the Orteig favorite, who spent a then extravagant $100,000 on his entry. He crashed during a practice takeoff, giving Lindbergh the opening he needed.
The X Prize rules were strict. We needed to reach 100 kilometers (or about 62 miles) above the earth, the edge of space—not once but twice within a two-week period, and no later than December 31, 2004. No more than 10 percent of the craft could be replaced between the two flights, and it had to return from the second one intact. Finally, no government aid was permitted.
The contest attracted twenty-seven teams from seven countries, though Burt took none of them seriously. The Canadian da Vinci Project, which counted on a gigantic helium balloon to carry their rocket plane, seemed especially far-fetched. We had some concern about rumored covert efforts in Eastern Europe, but our main adversaries were the clock and the unpredictability of Mach 3 flight.
AS A RULE, test pilots have a rare combination of intelligence, motivation, and emotional stability. As I came to know the men chosen to fly SpaceShipOne, I was impressed by their fearlessness but also their attention to detail. All three were in the Lindbergh tradition, exceptional people.
Brian Binnie was a lanky military test pilot with degrees from Brown and Princeton. A veteran of more than thirty combat missions during Desert Storm, he left the navy after they tried to make him a desk jockey. He was thoughtful and highly organized, a quiet guy whose intensity burned beneath the surface.
Pete Siebold was a young aeronautical engineer who’d designed the avionics for the planes and the flight simulator, which he operated like a pinball wizard. He knew every line of the software’s code and brought tremendous authority to the pre-and postmortems for each flight. Round-faced and curly-haired, he also had impressive cockpit skills. In his one powered flight, his trajectory was absolutely on point.
Mike Melvill was the outlier, a high school dropout and balding grandfather who wore baseball caps and wire-rim glasses. At sixty-three, Mike was three years past the retirement age for commercial pilots. But he was incredibly fit, a world-class kayaker who biked a hundred miles a week. Since hiring on in 1978, he’d twice been Burt’s best man. He believed in Burt’s planes without question and would do anything not to let his friend down.
Mike could struggle in the simulator. But for seat-of-the-pants improvisation and that mysterious quality called feel, there was no one you’d rather have in a critical test flight. Burt would call him “the best stick-and-rudder pilot I’ve ever seen.” In a project with more than its share of calculated risks, Mike’s makeup would be a key to our success.
A WELL-DESIGNED FLIGHT test program expands the envelope cautiously, incrementally. Burt was a master at moving forward in baby steps. First, SpaceShipOne would be tethered to the belly of White Knight in captive-carry flights. If all went well, the ship would progress to independent glides and finally to a series of six rocket-powered boost flights. Each step presented tougher g-forces or airspeeds or altitudes, or some mix of the three.
On December 17, 2003, the hundredth anniversary of the Wright brothers’ first flight, SpaceShipOne was set for its first powered outing—and the first privately funded flight to break the sound barrier. After consulting with Burt, Doug Shane, the director of flight operations, tapped the one pilot who had previously flown at supersonic speeds, Brian Binnie. Having managed the project’s rocket development program, Brian was also more familiar with the rocket’s motor. He was the obvious choice.
I flew to Mojave early that morning to join the five A.M. preflight meeting in Burt’s hangar. After SpaceShipOne was wheeled out on a dolly and secured beneath White Knight, I followed Burt and Doug to the control room, where team members donned headsets and sat before their networked desktop or laptop computers. (I got a kick out of seeing how modern PCs were being put to use.) As they scanned their screens, Doug ran them through a checkoff.
“Aerodynamics.”
“Go.”
“Propulsion.”
“Go.”
“Systems.”
“Go.”
This flight was new territory for Burt, with his first plane designed to go faster than Mach 0.7. The thrust of a rocket motor is basically a controlled explosion. If a leak led to some structural failure, it was a good bet that we’d lose our pilot. The other nightmare scenario was that the motor might not light at all, which happened several times in early ground testing: a puff of smoke, then nothing. With eight hundred pounds of rubber molded into the rocket casing, a lot of it needed to burn away before the pilot could land SpaceShipOne in anything close to its normal center-of-gravity envelope. If too much rubber remained, he’d come in so hard and fast that he might have to divert to a longer runway at the air force base, and even then there was no telling whether the spaceship’s small tires could handle the weight. As White Knight climbed into the sky, with the spaceship strapped beneath it, I had plenty of time to wonder, Is it really going to work?
Nearly an hour after departure, more than eight miles high, SpaceShipOne was released. A moment or two later, Brian lit the motor. It felt, he’d say later, “like a tsunami” coming through the cabin. Unlike a jet, a rocket accelerates to full power instantaneously, like a slam in the back. Back at mission control, I peered at the video monitor. To me, nothing can compete with a rocket boost for pure exhilaration: the bright orange glow, the vertical contrail of ice crystals like an arrow reaching for space. I was awed and grateful to be a part of it.
Then came something unforeseen, a wave of dread. I’d been deeply affected w
hen the Apollo 1 crew lost their lives in a preflight fire; I’d felt sick watching Challenger disintegrate a minute into its flight. I knew intellectually that someone might die in SpaceShipOne, but that was Burt’s familiar territory, not mine. (In the software business, after all, your worst outcome is an error message.) Now I knew the person whose life hung in the balance, and I found that hard to handle.
A fifteen-second burn got Brian twelve miles high at a maximum speed of Mach 1.2, and we had made our first bit of history. All systems were in order, so I went outside to watch the landing with Dave Moore, my liaison to MAV, where he served as managing director. (Along with his deputy, Jeff Johnson, he was instrumental in keeping the project on track.) Dave briefed me as I followed Brian’s final descent. Then he stopped, because my face had gone white.
Navy pilots are notorious for hard landings, a product of their aircraft-carrier training. Their heavy-handed style is a running joke among test pilots, but today it wasn’t funny. Brian hit the runway so hard that one of his landing struts snapped, and his plane careened off the pavement in a cloud of red desert dust. As we raced to the site, my heart was in my throat. Was Brian hurt? When they opened the cockpit, I was thankful to see that he was safe, visibly cursing himself out. My attention turned to SpaceShipOne—how badly was it damaged, and how much time would we lose? Our plan had been to complete the X Prize flights by the following summer. Any delay could push us up uncomfortably against our deadline.
Burt was consoling Brian, stressing the positives. He’d flown a good boost, gone supersonic, feathered and defeathered, had a nice glide. As for the plane, Burt said, “All we’ve got there is real minor stuff.” He came over to me: “I think we can get it back.” Once the craft was towed back to the hangar, we had a clearer picture. The ship’s hull was scraped but not damaged; the nitrous oxide tank was intact; the landing gear was fine, just ripped out of the ship. The accident would set us back about two months. There was still time to get in the three remaining powered flights on our test schedule.