What The New Yorker Gets Wrong About the SpaceShipTwo Accident

SpaceShipTwo debris in storage. (Credit: NTSB)

by Douglas Messier
Managing Editor

Psychologists have identified five stages of grief: denial, anger, bargaining, depression and acceptance. These stages are clearly on display in Virgin Galactic’s Rocket Man, Nicholas Schmidle’s profile of Mark Stucky in The New Yorker. A substantial part of the story chronicles how the test pilot dealt with the death of his close friend, Mike Alsbury, in the breakup of SpaceShipTwo Enterprise during the vehicle’s fourth powered flight four years ago.

It’s a touching portrait of Stucky’s grief for his fellow Scaled Composites pilot, with whom he had flown while testing the suborbital spacecraft being developed for Richard Branson’s Virgin Galactic. (Stucky later moved over to Virgin, which took over the SpaceShipTwo program after the accident, to test the second SpaceShipTwo, Unity.)

However, Schmidle tells only half the story in his otherwise insightful profile. He places nearly all the blame on Alsbury, while ignoring the findings of a nine-month federal investigation that identified systemic flaws in the development program and the government’s oversight that contributed to the accident.

It’s similar to the flawed, self-serving narrative that Branson used in his latest autobiography, “Finding My Virginity,” complete with a not-entirely-fair jab at the press coverage of the crash. The billionaire uses pilot error to obscure a decade of fatal mistakes and miscalculations.

An Inexplicable Mistake

SpaceShipTwo begins to break apart as its tail booms deploy prematurely with the engine still firing. (Credit: Virgin Galactic/NTSB)

SpaceShipTwo broke up after Alsbury released the locks that held the vehicle’s twin tail booms in place as the spaceship made a powered ascent during its fourth powered flight on Oct. 31, 2014.

The movable tail booms are designed to reconfigure the ship into the form of a shuttlecock to provide a gently reentry from space. Unlocking them during ascent is necessary to make sure the locks release properly. If they do not, the pilots would have to abort the flight and return to the Mojave Air and Space Port because the spacecraft would probably not survive a feather-down descent.

SpaceShipTwo feather controls (Credit: Scaled Composites)

Alsbury was supposed to unlock the feather when the vehicle reached Mach 1.4; at that point, the aerodynamic forces on the ship would keep the tail booms in place. Instead, he released them just after SpaceShipTwo reached Mach 0.8. Aerodynamic forces caused the tail booms to reconfigure, flipping the ship over and causing it to break up.

Alsbury was killed as SpaceShipTwo broke up. Pilot-in-command Pete Siebold parachuted to safety with serious but survivable injuries. He later returned to flight status with Scaled.

Disbelief, Depression & Anger

SpaceShipTwo breaks up after the premature deployment of its feather system. (Credit: MARS Scientific/NTSB)

As a former U.S. Air Force fighter and test pilot, Stucky was used to violent death in the air. He had lost dozens of colleagues over the years. But, this accident was different.

Alsbury had been both his wingman and his co-pilot, and during the previous year he’d become Stucky’s closest friend. “Nothing’s affected me quite like his death,” he told me. He couldn’t stop thinking about his friend’s mistake—why Alsbury had “skipped a chapter in the hymnal.”

Stucky’s disbelief over the seemingly inexplicable mistake was mixed with anger. On the day of the accident, he lashed out at his estranged son who had reached out to him in sympathy. And then there was this a few days later on Parabolic Arc.

Doug Messier, an aerospace blogger based in Mojave, published posts accusing Virgin Galactic of being cavalier about safety. In the comments section, Stucky replied angrily: “In the last couple of weeks I’ve lost a great friend and a great spaceship. Almost immediately you and other pseudo-journalists were quick to lay blame on the rocket motor and perceived programmatic pressures, both of which couldn’t be further from the truth. You act horrified about unconfirmed testing schedules and yet never compare them to other manned rocket programs—either historical or planned.”

Schmidle does his readers no favors by summarizing my coverage in single sentence, and then letting Stucky define it further. The passage lacks context about the detailed coverage of the program and the accident that Parabolic Arc provided.

SpaceShipTwo debris on the desert floor. (Credit: Jason_DiVenere)

I did get the cause of the accident wrong. Stucky was right to be angry about that. I deeply regret that mistake to this day.

But, it was what I and two photographers who recorded the accident that morning believed based on what we saw and what the photos appeared to show. It wasn’t until Sunday evening — more than two days after the Friday morning crash — that federal investigators revealed the truth and the photos began to make sense.

Even if we had gotten it right, it wouldn’t have changed the outcome. Alsbury would still have been dead, and the ship he was piloting spread out over 35 miles of the High Desert. The press wasn’t responsible for that.

As for schedule pressures, the federal investigation found plenty of evidence of them. But, we’ll get to that later.

Bargaining with the Past

Mike Alsbury (Credit: Scaled Composites)

The ill-fated Halloween flight was the most challenging test of SpaceShipTwo to date. The vehicle used new type of hybrid engine that had never been fired in flight before. The plan was to fire the motor for 38 seconds — nearly double the 20-second firing times on the two previous flights — to reach a record altitude of 135,000-138,000 ft. (The previous record was 71,000 ft.) SpaceShipTwo would then make its first supersonic feathered reentry.

Stucky had commanded SpaceShipTwo Enterprise‘s first two powered flights, the first one with Alsbury as co-pilot, and then served as co-pilot for the third test that was commanded by Virgin Galactic Chief Pilot David Mackay. He expected to command the fourth powered test. However, Stucky was replaced Siebold, who had returned to flight status after recovering from an earlier paragliding accident.

Stucky came to believe the change contributed to the accident. Schmidle’s story recounts a presentation that Stucky and Mackay at the Society of Experimental Test Pilots’ annual symposium in 2015.

On the first three powered flights of SpaceShipTwo, Stucky said, he had imposed “sterile cockpit rules,” because he “did not want extraneous calls during boost and other high-workload portions.”

But, Mackay noted, “Scaled gave their pilots-in-command the leeway and authority to run their cockpit as they saw fit,” and this failure to “impose a rigid set of standard operating procedures” had “created an environment where callouts and actions could be missed both in the cockpit and in the control room.”

Peter Siebold, who had recovered from his injuries, was in the back of the banquet hall. Neither Stucky nor Mackay mentioned his name, but Stucky privately faulted him for his unnecessary shout of “Yeehaw!” during the boost, and for not stopping Alsbury from unlocking the feather.

After the presentation, Stucky and Siebold ran into each other by the coffee urns. “It’s too early,” Siebold said. They have hardly spoken since.

Stucky and his wife, Cheryl Agin, came to believe that the flight would have ended differently if only he had been in command that day.

One morning, Stucky said to me, “If I had been flying with Mike that day and heard him say, ‘Unlocking,’ I think I could have stopped him in time.” Agin said, “I will always believe that if Mark had been in that spaceship that they wouldn’t have crashed.”

It was a classic case of the bargaining phase of grief. If only I had been there. If only we had a chance to do it all over again…

Maybe. Maybe not. Everything happened very quickly, and the pilots were an extremely handling a heavy workload as the vehicle approached Mach 1. It took Alsbury about a second to unlock the feather. Siebold later told federal investigators he didn’t remember hearing Alsbury make the call that he was unlocking the system. It is not clear Siebold missed it because he had earlier shouted “Yeehaw!” two seconds earlier.

Ultimately, you can’t change the past, you can only learn from it. In this case, however, the bargaining has a subtext not mentioned in Schmidle’s story that people in Mojave have whispered about since the accident but have never said publicly.

Stucky was a military pilot before joining Scaled. Siebold and Alsbury were civilians who learned test piloting at the company. Some feel there is a significant difference in skill levels and abilities between the two classes of pilots. They question whether Siebold and Alsbury were the best combination of pilots to put in the cockpit for such a complex test. Perhaps Stucky could have prevented the accident.

The debate over who should have flown that day is academic at this point. It is notable, however, that Virgin Galactic has only hired military test pilots to fly SpaceShipTwo. Stucky’s military background allowed him to come over the Virgin after the accident.

Acceptance

Centrifuge used for pilot and passenger training. (Credit: NASTAR)

Stucky seems to have found some acceptance over the loss of his friend after conducting centrifuge training for SpaceShipTwo flights at the National Aerospace Training and Research Center (NASTAR) outside of Philadelphia.

Stucky conveyed what he had learned to Alsbury’s widow, Michelle Saling, during a sad visit to SpaceShipTwo’s crash sites north of the Mojave Air and Space Port. Saling has struggled raising two young children since the loss of her husband.

In the car, he told her about some of the training he’d done in the centrifuge. “I know you don’t want to hear that something good came of all this,” he said. But one lesson he’d taken from the crash was the importance of scripting every word and action on a test flight. While in Philadelphia, he and the other pilots had each committed the same blunder: calling out “trimming” when they were still subsonic. Stucky explained to Saling that if the pilots trimmed—went sharply nose up—before passing Mach 1, “it would aggravate the transonic pitch-up.” After returning from Pennsylvania, Stucky had asked the engineers to adjust the altitude and airspeed dials so that blue lights began glowing after the spaceship went supersonic; this color prompt would indicate that it was safe to begin trimming.

“Maybe Mike’s death saved eight people from dying on a later flight,” Stucky told Saling. “I’m not telling you that to try and make you feel better. Just that we’ve made some serious changes. Because ninety-nine per cent isn’t good enough.”

Saling thanked him. It was oddly comforting to hear examples of other pilots’ fallibility. Her husband was not the only one prone to skipping a chapter in the hymnal.

It’s a touching scene. And Stucky is probably right; if Alsbury hadn’t made the mistake when he did, another pilot would likely have done so on a commercial flight filled with six millionaire passengers.

Those flights had been fast approaching by October 2014. After a decade of backing the program, Virgin Galactic was short of funds and under pressure to begin commercial service. Abu Dhabi’s sovereign wealth fund, aabar Investments, had poured $390 million into Branson’s space company. It was impatient with delays and unwilling to provide additional funding until it saw a flight to the ship’s maximum altitude, sources told Parabolic Arc.

Richard Branson rolls out Virgin Galactic’s Spaceship Unity in Mojave. (Credit: Virgin Galactic)

Prior to the accident, there were plans for two more powered flights — for a total of only six — before Scaled ended the test program and turned SpaceShipTwo over to Virgin Galactic to begin commercial service. Branson and his son, Sam, were scheduled to be aboard the inaugural flight from Spaceport America in New Mexico in the first quarter of 2015.

Some experts worried that six powered flights would be insufficient to uncover all of SpaceShipTwo’s flaws. It was especially worrisome that only one of those flights would have included a full engine burn to get the spacecraft to its maximum altitude. Enterprise was also a proof-of-concept prototype that was not originally intended to enter commercial service.

The crash made those concerns academic. It also set back Virgin Galactic’s space program about three years. The company has addressed the issue of premature feather deployment with modifications to the system and cockpit procedures for the second SpaceShipTwo, Unity. It also examined other potential flaws in the spacecraft.

Be Careful What You Wish For….

Success is a lousy teacher. It seduces smart people into thinking they can’t lose.”
— Bill Gates

Prototypes are, by nature, flawed vehicles. The job of the test pilot is to find all the flaws without destroying the vehicle or getting himself killed. That doesn’t always happen.

In this case, there were other factors at work. Alsbury wasn’t the first person to die in the SpaceShipTwo program; he was the fourth. There were systemic problems in Scaled’s safety culture and government oversight during the 10 years the company developed and tested the spacecraft. The roots of the problem can be traced back, ironically enough, to Scaled’s greatest achievement.

Mike Melvill stands atop SpaceShipOne after a suborbital flight on Sept. 29, 2004. (Credit: RenegadeAven)

On June 21, 2004, Scaled and its maverick founder, Burt Rutan, launched SpaceShipOne into space and the history books from Mojave. For the first time, a privately-built, crewed vehicle had achieved what had been the sole purview of government space agencies. Pilot Mike Melvill was awarded civilian astronaut wings for the feat.

Three months later, Rutan and his backer, Microsoft Co-founder Paul Allen, claimed the $10 million Ansari X Prize after SpaceShipOne flew to space twice within a two-week period. The entire project, which had been accomplished with a small team and cost only $28 million, was held out as a model for future projects and as the beginning of the end of expensive, overpriced government space efforts.

As he stood on a podium that day in Mojave before a cheering crowd, Rutan couldn’t resist trolling NASA, whose human space program had been grounded since February 2003 after the loss of the space shuttle Columbia and her seven-member crew.

Peter Diamandis and Burt Rutan on stage after SpaceShipOne won the Ansari X Prize on Oct. 4, 2004.

“I was thinking a little bit about that other space agency, the big guys,” Rutan told the crowd. “I think they’re looking at each other and saying, ‘We’re screwed.’”

Rutan was not alone in his disdain for a space agency whose name he derisively pronounced as ‘nay-say’. A similar attitude permeated Mojave, even within space companies that had never flown — and, in some cases, would never fly — anything anywhere near space. It was difficult for many people in this libertarian, get-government-off-our-backs enclave to believe that NASA or any other federal agency had much to teach them about spaceflight, safety or much of anything else.

The commercial spaceflight regulations approved two months after SpaceShipOne’s final flight incorporated a light touch approach. Oversight was given to the FAA Office of Commercial Space Transportation (FAA AST), whose experience with human spaceflight was limited to overseeing three brief SpaceShipOne flights. NASA, which had been flying astronauts in space for more than 40 years, was not involved.

FAA AST’s authority was limited to protecting the “un-involved public,” i.e., people on the ground who were not involved in the flight, as well as property. The office had no authority to promulgate regulations to protect pilots and passengers until after there was an accident or a close call. Ticket holders would fly at their own risk under an “informed consent” doctrine. Subsequent state laws would codify this approach, holding flight prividers blameless except in cases of gross negligence or intentional harm.

The idea was to give Scaled and other commercial space operators a learning period during which to experiment with different designs and techniques before the government came in with strict safety regulations that might drive these entrepreneurial startups out of business with excessive costs and expensive lawsuits.

What Rutan and Scaled learned was they weren’t as smart as they thought they were. They assumed that scaling up SpaceShipOne’s small, hybrid rubber-nitrous oxide engine for its much larger successor would be a relatively straightforward task. And that nitrous oxide was safe and benign. Neither assumption was true.

On July 26, 2007, Scaled engineers conducted a test to flow nitrous oxide through a newly designed valve. The test was known as a cold flow because no rubber fuel was present. Believing nitrous oxide posed no threat, 11 people were allowed to stand near the test rig when the operation began.

Seconds into the test, there was a massive explosion that killed three engineers – Eric Blackwell, Todd Ivens and Glen May – and landed three others in the hospital with serious injuries. They were the first fatalties Scaled had every experienced in its history.

A memorial plaque to three Scaled Composites engineers killed in 2007 is displayed in Mojave’s Legacy Park. (Credit: Douglas Messier)

Rutan would claim there was no way to know that nitrous oxide could explode on its own. However, nitrous oxide is a mono-propellant — meaning it can do precisely that under certain conditions. There had been previous incidents in other industries, including the massive explosion of a nitrous oxide trailer in 2001 due to an overheated pump in Einhoven, The Netherlands. That accident would later be incorporated in Scaled Composites’ nitrous oxide safety guide.

For all of Rutan’s boasting, it’s difficult to imagine such an accident occurring at one of the NASA’s test facilities. Or on a test stand run by any experienced rocket company. They would have done their homework on nitrous oxide, understood the risks, and taken the standard safety precaution of clearing the test stand of personnel before beginning any hazardous activities.

Engineers would spend years redesigning the nitrous tank to prevent a recurrence of the accident. Meanwhile, Scaled was dealing with a different problem: the larger engine’s vibrations and oscillations were so severe that they threatened to excessively shake up the crew and passengers and cause damage to the ship. It would take a decade to produce an engine that could be fired for more than 30 seconds in flight.

Siebold and Alsbury were testing that engine when SpaceShipTwo was brought down by another danger that Scaled and FAA AST failed to properly address.

A Failure of Imagination and Oversight

The remains of SpaceShipTwo’s cockpit. (Credit: Douglas Messier)

The National Transportation Safety Board’s (NTSB) accident report, released nine months after the crash, did not place all of the blame on Alsbury. Instead, the safety board concluded the probable cause of the accident was

Scaled Composite’s failure to consider and protect against the possibility that a single human error could result in a catastrophic hazard. This failure set the stage for the copilot’s premature unlocking of the feather system as a result of time pressure and vibration and loads that he had not recently experienced.

It had been 17 months since Alsbury had flown as Stucky’s co-pilot on SpaceShipTwo’s first powered flights. The vehicle was a hand-flown ship that created an enormous workload for both pilots. A FAA safety expert said SpaceShipTwo had the highest pilot workload he had ever seen for any flying vehicle.

In order to begin powered flights, Scaled required an experimental permit from FAA AST. The NTSB found that the system safety analysis (SSA) the company submitted as part of its permit application in 2012 was deficient.

“Although Scaled’s Composite system safety analysis correctly identified that uncontrolled feather operation would be catastrophic, the SSA process was inadequate because it allowed an analysis that failed to identify a single human error could lead to a problem in the boost phase and failed to boost the effectiveness of mitigation measures,” the board concluded.

Pete Siebold’s pilot seat from SpaceShipTwo Enterprise. (Credit: NTSB)

The NTSB found Scaled’s mistake was compounded by federal regulators. The FAA AST’s “evaluations of Scaled Composites’ initial and first renewal of the SpaceShipTwo experimental permit application were deficient because the evaluations failed to recognize that Scaled Composites’ hazard analysis did not meet regulatory requirements to identify hazards caused by human error.”

The SSA had been evaluated by an agency official without a lot of experience who concluded that it essentially met the safety requirements even though it did not meet the exact letter of the law. Scaled had evaluated the risks more as it would for the aircraft it was accustomed to building than for a supersonic space plane that would soar out of the atmosphere.

FAA AST issued an one-year, renewal permit for powered flights in late May 2012. The first powered flight took place nearly a year later on April 29, 2013.

In the year that followed the issuing of the permit, other FAA AST safety officials who had come over from NASA’s recently ended space shuttle program reviewed the SAA and found it to be deficient in its analysis of safety hazards.

These experts had spent years training astronauts and knew they made mistakes all the time. The shuttle’s systems and procedures were designed to prevent a single mistake from bringing down the vehicle.

The experts also knew there was a big difference between jets and spacecraft. “Above 50,000 feet, the game changes,” a safety expert later told NTSB investigators.

The experts proposed a get plan under which Scaled and FAA AST would have addressed the deficiencies in the SSA. The additional work would have delayed flight tests in a program was already running years behind schedule and many millions over budget.

Former FAA AST chief George Nield

FAA AST management under then Associate Administrator George Nield rejected the plan to improve the safety analysis. He sided with agency officials who argued the SAA was sufficient for the FAA AST’s limited regulatory role of protecting the un-involved public and property on the ground.

The office renewed the permit in May 2013 and again one year later, allowing Scaled Composites and Virgin Galactic to continue with flight tests without any external delays.

Agency safety experts who argued the SSA was deficient would later complain that management had succumbed to political pressure to keep testing on schedule. They said such pressure was common to all programs the office issued permits for, not just for SpaceShipTwo flights. No one identified exactly who was responsible for the pressure.

Two months after the May 2013 permit renewal, FAA AST published a waiver signed by Nield that exempted Scaled Composites from requirements to identify and mitigate pilot error and software error. SpaceShipTwo would crash 14 months later due to pilot error.

The waiver mentioned a series of operational steps that Scaled was taking to protect people and property on the ground. NTSB investigators discovered that agency employees in the field who oversaw SpaceShipTwo’s flight tests took no steps to enforce the waiver because it lacked the authority of a formal federal regulation.

FAA AST personnel “did not ensure that Scaled Composites complied with the waiver or required whether mitigations would adequately address human errors with catastrophic consequences,” the safety board concluded.

There were also communications problems between Scaled and FAA AST officials that combined with other factors to limit the effectiveness of reviews. In order to lessen the burden on Scaled, FAA AST safety officials had to submit questions through their management for review. Those questions were sometimes rewritten, resulting in responses that didn’t always address the concerns, or rejected all together meaning experts received no answers at all.

“The lack of direct communications between staff and Scaled Composite’s technical staff, the pressure to approve experimental applications within the 120-day review period, and the lack of mission safety assurance interfered with the FAA’s ability to evaluate applications,” the NTSB concluded.

And there was one more crucial failure the NTSB found. “Scaled Composites did not ensure that pilots correctly understood the risks of unlocking the feather early,” the accident report stated.

We’ll never know if an additional safety review would have prevented the accident. In the end, FAA AST’s regulations did achieve their principle goal of protecting people and property on the ground — but just barely. When the cockpit containing Alsbury’s body slammed into Cantil Road, it barely missed hitting a couple of truck drivers in their vehicles.

The End of the Beginning

Richard Branson (l) and George Whitesides (r) walk with SpaceShipTwo pilots David Mackay and Mark Stucky after a successful glide flight. (Credit: Kenneth Brown)

The NTSB report tells a much more complicated story of the accident than the one Schmidle conveys in his profile of Stucky. Years of failures and missed opportunities put him in a cockpit where a single catastrophic human error could destroy the vehicle.

With better supervision, Scaled might have avoided both fatal accidents. However, it never really received it from FAA AST or its business partner, Virgin Galactic.

At the time of the test stand accident in 2007, Branson could offer little more than sympathy to a devastated Rutan and his team. Virgin Galactic was largely a marketing operation at that point with no real in-house technical expertise.

By the time powered flights began in 2013, Virgin Galactic and its production arm, The Spaceship Company, had built up their own internal capabilities. However, Virgin officials were pushing Scaled and FAA AST to keep flight tests moving forward at that point so it could begin commercial service.

Virgin continued with powered flights even though experienced FAA AST safety experts had identified serious flaws in the safety analysis. It was willing to fly under a waiver that exempted Scaled from identifying and mitigating pilot and software error. That waiver was primarily focused on protecting the peopleo on the ground than ensuring the safety of the ship and its crew. Finally, Virgin pushed for a relatively short powered flight test program so it could begin flying paying passengers in early 2015.

Despite company officials proclaiming that safety was the north star of Virgin’s operations, they left a key safety position open as the test program entered a critical stage and they prepared to fly their founder and his son. Jon Turnipseed retired as vice president of safety in December 2013 amd decamped to Idaho. Virgin did not name a replacement until after the accident.

Branson mentions none of this in his latest autobiography. Instead, he pins all the blame for the crash on the co-pilot. It’s as if the rest of the NTSB report doesn’t exist. Branson’s other target is the media.He devotes about six pages to criticizing the press coverage of the accident before spending less than one page on the tragedy of Alsbury’s death and the widow and two children he left behind.

Following the crash, Branson’s ambitious space program was grounded, just as NASA’s program had been 10 years earlier when Rutan had imagined space agency officials shaking in its boots following the success of SpaceShipOne. Four years later – and 14 years after Branson announced the program – Virgin Galactic has still not reached space.

The fatal flight was the tragic swan song for Scaled Composites’ involvement in the SpaceShipTwo program. It would be up Virgin Galactic and The Spaceship Company to complete the second SpaceShipTwo Unity and qualify it to begin commercial flights.

During the decade it worked on the program, Scaled managed only four powered flights, one of which crashed and none of which got anywhere near space. Four people had died in a program that had likely cost $600 million to that point. It’s a sad record in which neither Scaled nor Virgin can take any pride.

The very success Scaled had achieved with SpaceShipOne had taught it all the wrong lessons. Oversight proceeded on flawed assumptions. You can blame Alsbury for pulling a lever too soon, but in the final analysis, the systemic problems in the program are ultimately what led to that fatal mistake in the skies over the Mojave.