by Tim Fernholz
GPS, as it is known, began in the late seventies as a military enterprise operated by and for the US Air Force. A dozen satellites orbited the earth, allowing US military units with the correct equipment to triangulate their position on the ground with astonishing accuracy. There was no civilian use of this technology until tragedy mandated it. In 1983, a Korean Air passenger liner accidentally wandered into Soviet airspace and was shot down by a fighter jet, killing 269 innocents on board. In a show of Cold War munificence, Ronald Reagan offered civilian airliners the use of GPS to avoid future deadly navigation errors.
By 1989, the Air Force began planning an upgraded set of GPS satellites, and the first consumer GPS receiver was on the market. It was expensive and, in truth, not all that good, since the military had intentionally degraded the accuracy of the civilian system to prevent abuse by criminals or terrorists. Yet the usefulness of the service was becoming clear as pilots around the world adopted satellite navigation to replace and augment older, radar-based techniques. The Clinton administration okayed the continued use of the GPS system, as well as another set of expansion satellites. Vice President Al Gore began a push to expand the transmissions from the system to two new, dedicated civilian channels, effectively creating a digital infrastructure for entrepreneurs to build around. Today the system is vital to the global economy, with nearly every financial transaction coordinated by GPS timing signals.
Arthur C. Clarke had first envisioned satellites creating global communications networks in 1945. As governments demonstrated that the technology had arrived for this futuristic vision, private companies announced plans for their own constellations—for mobile communications, and to broadcast television signals. Rocket companies took note: beyond GPS, there would be a growing abundance of lucrative cargo to be launched into space.
The workers who showed up on launch day that January, then, could only feel as if they were at the beginning of a new leap forward in space commerce.
As the countdown moved forward, propulsion engineer Brian Mosdell’s job was to prepare the rocket for launch, opening and closing valves with push-button electronics as propellant was loaded. He worked with his team from a control room about six hundred feet from the launchpad, and the countdown proceeded without issue. This Delta II rocket had one main engine and—typical of the designs of the time—was strapped with nine additional solid fuel rocket boosters to drive the vehicle up into space. Strip charts—paper on easels marked by a mechanical pen—tracked data from the spacecraft. As the engines ignited on time, Mosdell strolled into the next room to stand with the launch director and the rest of the team. Console operators were at their stations to track the flight.
Alas, it wasn’t a long one. Seconds after liftoff, the metal case enclosing one of Delta’s solid fuel rocket boosters cracked. What had been a controlled chemical reaction, launching the rocket by generating more than one million pounds of pressure, broke loose. The force tore open the rocket’s engines and triggered the vehicle’s onboard self-destruct system. “We have an anomaly,” the launch announcer said. Amazingly, the upper stages of the rocket were able to ignite and fly free of the initial explosion before their own self-destruct mechanisms were activated. The rocket was barely sixteen hundred feet above the earth when it exploded “like a giant fireworks display.” Instead of a chrysanthemum of burning paper, it sent enormous pieces of white-hot metallic debris cascading onto the launch center directly below.
The fortified concrete blockhouse where Mosdell and his colleagues were watching the launch was barely a quarter mile from the launchpad, and surrounded by a thick earthen berm. Still, Mosdell’s boss, a veteran launch director known for his sangfroid, took one look at the monitors, said, “Boys, this is bad,” and dove under a computer console. Other engineers followed suit, scrambling for cover. Mosdell stood dumbstruck until debris began crashing down around the blockhouse, causing dust to fall from the ceiling and knocking over furniture. He, too, finally dove for cover as the massive blast wave shook the building.
Safe for the moment thanks to their fortified position amid the conflagration, the seventy people in the launch house quickly realized they had a new problem. A chunk of the rocket had fallen onto the conduit protecting the data cables connected to the blockhouse. As the cables began to burn, acrid smoke from the melting plastic and metal began to penetrate the building. Their fortress, now filling with smoke, threatened to become a tomb as oxygen ran out. Hunkered down and waiting for rescue, they slipped on plastic oxygen masks to extend their air supply. Just as the situation was becoming completely untenable, they heard firefighters banging on the door to escort them to safety.
Though Mosdell and his colleagues escaped without serious injury, the technicians’ vehicles, parked just behind the control center where the blockhouse was located, hadn’t been so lucky. Struck by falling debris, Mosdell’s truck had caught fire. The windows had melted into glass waterfalls that ran down inside the doors. Calling his insurer to report the loss, Mosdell couldn’t help but deadpan when asked what had happened to his vehicle. “A rocket hit it,” he said. The baffled insurance agent repeated the question, and he told her to switch on CNN, which was just breaking into its regular reporting with video of the tremendous explosion. She got the picture. The twenty cars destroyed would cost more than $400,000, a tiny fraction of the cost of replacing the rocket, satellite, and launchpad consumed in the conflagration.
This was the first and only total failure of a Delta II rocket in some fifty launches. But it was also the first in a series of failures that would show how optimism about the US space program had been misplaced. The years ahead would throw rocket makers into panic and ultimately force the government to sign off on a $60 billion monopoly.
Calamity always comes first. Rocketry, especially developing new rockets, is an extremely fraught and expensive process. Beginning in the late 1950s, the Mercury program, which launched the first American astronauts into space, blew up rockets left and right—sometimes with their future passengers on hand to observe— before engineers worked out the kinks. The average rocket development program delivers its product twenty-seven months late. Because the machinery necessary to deploy millions of pounds of force with precision is so complex, and failures tend to be total, rocket designers traditionally focus on lots of up-front design work to “buy down” risk, and favor approaches that have proven reliable in the past.
The Delta II rocket that exploded in January 1997 was derived from the technology behind Cold War nuclear missiles. It had been resurrected only as a result of tragedy.
At the dawn of the space shuttle program, in the 1970s, the government had aimed to create a one-stop shop for space access. The shuttle orbiter, with its enormous cargo-lifting capacity, its ability to convey astronauts, and its versatile maneuvering capabilities, was intended to be the workhorse of the US space program. The vision of the 1979 James Bond thriller Moonraker—which included half a dozen space shuttles conveying reinforcements to Agent 007 as he fought his way through a villain’s space lair—wasn’t a flight of fancy. It reflected the expectations of the US government. NASA hoped to fly its five-shuttle fleet as often as twenty, thirty, or even sixty times a year, doing everything from scientific exploration to satellite repair.
That dream, deferred by slow technology development, came to an end in 1986. That year’s flight of the space shuttle Challenger was intended as a public relations coup that would demonstrate that the United States was bringing not just test pilots and astrophysicists, but also regular citizens, into space. A nationwide search for an educator to fly on the mission had selected Christa McAuliffe, a New Hampshire social studies teacher, from among thousands of applicants. She trained for a full year before the flight.
Amid great excitement and heightened anticipation, teachers wheeled televisions into classrooms so that schoolchildren around the country could watch the glorious launch of the space shuttle. Instead, they witnessed a space disaster when the vehicle exploded just o
ne minute and thirteen seconds into its flight. It was the worst NASA accident since a harrowing 1967 episode in which an Apollo space capsule burst into flame during a routine pre-launch test, killing the three astronauts on board. Challenger led to serious reconsideration of US space policy. The loss of the seven astronauts who perished in that venture convinced the top brass at NASA that it was too risky to put humans on every flight into space. That was especially true of those missions intended to launch satellites or space probes, with no real human exploration component. Since the shuttle was designed expressly for human spaceflight, this conclusion required the government to find new rockets to launch satellites without people on board.
Noble Prize–winning physicist Richard Feynman, a member of the distinguished panel investigating the disaster, famously demonstrated the tiny flaw that had led to Challenger’s loss. He dunked rubber O-rings, used in the construction of the solid fuel rocket booster, into a pitcher of cold water. At lower temperatures, the tight-fitting seals became brittle and stiff, making them likelier to crack under pressure and leak superheated gas. On launch day, temperatures had fallen below freezing—well below acceptable conditions for the rocket boosters—and warnings from the engineers who built the boosters apparently never made it up the chain to NASA management.
The experiment illustrated the tiny margin of error in rocketry. Something as simple as a colder-than-average day could mean total failure. The space shuttle orbiter was NASA’s first attempt at a reusable spacecraft, so the Challenger investigations also shone a spotlight on the obstacles to launching the same vehicle into orbit multiple times. As NASA engineers worked to ensure that careful inspection and additional refurbishment prevented accidents in the future, the benefits of reusability appeared to decrease. The shuttle’s hoped-for ability to fly many times each year did not materialize as NASA had originally foreseen. Rather than a simple turnaround, preparing a used shuttle orbiter for flight required more than 1.2 million different procedures. Not only did this increase spending directly, but the extra time made it difficult to spread the costs of the program over a high rate of flight.
Seeing the results of its experiment with reusability, the American space brain trust decided that it needed an expendable rocket if it was going to put satellites into orbit without risking lives or breaking the bank. But there were few existing alternatives when President Reagan told the world that “NASA will no longer be in the business of launching private satellites.” The shuttle had, in effect, closed out the market for expendable launch vehicles. The promised rate of shuttle flights, as well as government subsidies of $50 million per launch to rent out the shuttle’s spacious cargo bay, had convinced most US rocket makers to mothball their operations at the beginning of the decade. It had also, in 1980, helped convince a consortium of European countries to fund Arianespace, a rocket maker that would guarantee their own access to space.
Indeed, the US reliance on just one launch vehicle for space access had worried some Americans, especially as delays and cost overruns in the shuttle program mounted, but it was not until Challenger that the government was forced to reckon with the consequences of its policy.
“The government put all their eggs in one basket,” John Garvey, a veteran aerospace engineer who began his career the year of the Challenger disaster and spent the following decades developing rocket technology at McDonnell Douglas, Boeing, and a series of space start-ups, told me. “The shuttle was flawed because it tried to do everything for everybody, and it ended up not satisfying anybody. The government tries to do this every ten years.”
With the United States now looking around for expendable rockets to fill the gap, a few reliable defense contractors were called on to resurrect their dormant supply chains.
McDonnell Douglas stepped up first, reopening production of the Delta II, a rocket adapted from an intercontinental ballistic missile designed to deliver nuclear weapons across oceans. Its second stage was inherited from the first US rocket to launch satellites in the fifties. And Lockheed Martin put together the Titan IV, a rocket also derived from Cold War ICBMs. These “heritage” rockets would allow the United States to launch GPS and intelligence-gathering satellites without risking astronauts on routine missions. Yet it quickly became clear that relying on old technology wasn’t the answer. In 1994, an Air Force study found that the government was spending $300 million per year on rocket failures and delays. The 1997 Delta II explosion was simply the most visceral example of this trend.
The US space community realized that what it needed was a new expendable rocket system. Developing and testing a launch vehicle doesn’t come cheap; the government estimated that building on its heritage technology would require an investment of at least $1 billion, whereas an entirely blank-slate approach might cost more than $5 billion. Since “new” was too costly, the government settled on “evolved”—that is, based on reliable technology rather than starting from scratch to create a new space vehicle. Financing a multi-billion-dollar development program, however, didn’t really fit into the government’s wish list. This plan evolved against the backdrop of vicious budget fights between President Clinton and a Republican Congress that would end with a controversial government shutdown.
Still, the government found a way to come up with the money—always the biggest hurdle in the space business. It bet that any new rockets would have plenty of private satellites to launch, creating an incentive for space companies to kick in their own capital to cover some of the development costs. Yet in 1994, defense policy analysts wrote that “there are limited opportunities to significantly expand the space launch market.” By 1998, however, the government was awarding contracts based on the premise of a booming demand for rocket launches in the private sector. What changed? In a word: the internet.
The hunt for a new rocket was also taking place against the backdrop of a bigger revolution: the rise of digital technology. Computers were becoming more of a consumer tool, and the internet more of a public platform. Private-sector technologists were starting to become cultural heroes in a way that astronauts had been decades before, combining the steel allure of modern technology with aspirations for a bold new future. In 1994, as the government reevaluated its space technology strategy, Microsoft founders Bill Gates and Paul Allen were dominating the business world with the promise of greater productivity through digital tools. Elon Musk and his brother, Kimbal, fresh out of college, were renting Silicon Valley office space for their first start-up, Zip2, an early attempt to put local information online. Jeff Bezos, meanwhile, was in the process of leaving his work at the Wall Street firm D. E. Shaw to realize an idea for a far-out business that he and his colleagues were calling “the everything store.”
Indeed, as the US government was dreaming up the Evolved Expendable Launch Vehicle program, or EELV, as it would be widely known, Marc Andreessen was delivering the first-ever graphic web browser, Netscape, to the world. Compare the speed of action in these respective institutions: in the four years it took the government to finally settle on a strategy to build the new rockets and put pen to paper, Netscape debuted its browser, went public, was purchased by AOL in a $4.2 billion deal, and in the process laid the groundwork for an antitrust lawsuit that would end Microsoft’s desktop computer monopoly. In that same period, Bezos started Amazon and took the company public in 1997, beginning the future e-commerce giant’s path to domination.
This fast-moving digital revolution had brought together the two key ingredients in space exploration: huge amounts of capital and plenty of eager dorks. Computers were beginning to dominate the economy completely, at least in the minds of the people in the industry. So why not start hurling them into space?
The proliferation of satellite schemes—Teledesic, Iridium, SkyBridge, Globalstar—implied rising future demand for rockets to get the satellites into orbit. Eyeing these schemes in the midnineties, as they developed proposals for new rockets, Lockheed Martin, McDonnell Douglas, and ultimately Boeing were able to promise ve
hicles that, by the weird pricing standards of high-explosive space vehicles, were fairly cheap. The architects of the EELV program imagined rockets capable of carrying at least ten metric tons to low earth orbit and five metric tons to geostationary orbit. They thought the cost of each rocket launch would range between $50 and $150 million, in 1994 dollars. Lockheed Martin and McDonnell Douglas won the development contracts, at $500 million apiece, and Boeing soon snapped up McDonnell.
The next year, 1998, it was Lockheed’s turn to see one of its old reliables fail, this time during an attempt to launch a missile early-warning satellite from Cape Canaveral. Its trusted Titan IV rocket broke up forty-one seconds after liftoff, resulting in total destruction of the satellite and the vehicle; no automobiles were lost this time around. But the satellite and the rocket had together cost $1.3 billion, an enormous waste. Investigators determined that an exposed wire had shorted out the guidance computer, causing a sudden change in direction that the rocket couldn’t withstand. Nervous Air Force generals and NASA scientists were now openly fretting about their launch program.
The good news was that Boeing was prepared to unveil a new generation of the Delta rocket. It would be the first sign of progress in the new US strategy of outsourcing rocket development to private firms. The Delta III, as it was known, would make its inaugural launch from Cape Canaveral just over two weeks later, on August 27. In a surprisingly risky decision, the first flight of this new rocket would carry a real communications satellite for PanAmSat, rather than a simulated payload.
The new rocket exploded seventy-one seconds after liftoff.
As it flew, the vehicle began shaking as three solid fuel rocket boosters strapped to the main body rocked in unison, a situation its designers hadn’t foreseen. Attempting to correct and maintain a straight course, the rocket’s guidance computers battled the vibrations valiantly, like a race car driver frantically spinning his steering wheel back and forth to avoid losing control on a tight corner. This worked, but only temporarily: the hydraulic system that steered the rocket ran out of fluid after so many unexpected maneuvers. Like the race car driver simply letting go of the wheel, the rocket lost control. It entered a high-altitude wind shear moving faster than the speed of sound, twisted, and was torn apart.
