SpaceShipTwo: Lessons Learned on the Commercial Space Frontier

SpaceShipTwo disintegrates as its two tail booms fall away. (Credit: Kenneth Brown)
SpaceShipTwo disintegrates as its two tail booms fall away. (Credit: Kenneth Brown)

SpaceShipTwo had exploded.

At least that’s what it looked like from our vantage point at Jawbone Station on that fateful Halloween morning ten months ago. And that’s what it looked like in Ken Brown’s photos. Ken had been standing next to me, training his telephoto lens on the small spacecraft nine miles overhead.

It appeared that something had gone wrong with the motor, the nitrous oxide tank, or some other part of the propulsion system. There was a flash of red, with a giant cloud of nitrous oxide escaping from the breached tank. Then pieces of the ship started falling all over the High Desert.

As I watched clouds of dust thrown up as the debris hit, it seemed that everyone’s worst nightmare had occurred. Two weeks earlier, I had sat in an office at the Mojave spaceport and predicted exactly this would happen to two friends visiting from Los Angeles.

A Gnawing Fear

This flight test of SpaceShipTwo was different from the previous three powered ones. They were testing a new type of hybrid engine for the first time, one that burned nylon instead of rubber. It wasn’t simply a change in the fuel grain – there had been significant alterations in the propulsion system and the ship. People were on pins and needles about how the new system would perform in a planned burn of 38 seconds, nearly twice that of any previous flight test.

There was also the aggressive flight test schedule. After more than nine years filled with setbacks and delays, Virgin Galactic suddenly seemed to be in a big rush for Scaled Composites to complete the flight test program so commercial flights could begin. This was to be done by the end of 2014. There would be a handful of additional flights, then Richard Branson and his son, Sam, would climb aboard to inaugurate commercial service.

Richard Branson
Richard Branson (Credit: Douglas Messier)

It struck me as a dangerous approach. There weren’t enough powered flights to really wring out the ship, and to find all the flaws in it that could destroy the vehicle, its crew and passengers. It didn’t appear that the nylon hybrid – which was only announced at the end of May – had been sufficiently tested on the ground. I genuinely feared they would push the technology, the pilots or both too far and too fast, resulting in a tragedy.

I knew I was not alone in feeling this way. Others had the same concerns. I couldn’t do anything to slow things down. I hoped that they could.

Over the summer, I had seriously considered leaving Mojave. I wasn’t sure I wanted to be here if the worst happened. From a journalistic standpoint, it would be one helluva story. Emotionally, it would be very difficult to take.

As October wore on amid signs that another powered flight was imminent, I had trouble sleeping at night and concentrating during the day. By the afternoon of Oct. 30, it was clear they would be flying the next day. Ken and I made plans to meet up at Jawbone Station, a visitors’ center 20 miles north of Mojave, where we would have a good view of the drop. I drifted off to an uneasy sleep around midnight.

So, knowing all that I knew, and seeing what we had seen on that awful fall morning, I concluded SpaceShipTwo had suffered a catastrophic engine failure. I would turn out to be very wrong. It would take a while to learn the truth.

The Truth Emerges

SpaceShipTwo fuselage. (Credit: NTSB)
SpaceShipTwo fuselage. (Credit: NTSB)

The first clues came when news photos were shown of the main section of the fuselage with the nitrous tank. It didn’t show any evidence of an explosion or fire. That didn’t rule out a failure of the propulsion system, but it raised questions about whether SpaceShipTwo had blown up. And if it hadn’t exploded, what had brought it down?

We had tried to reach this piece of debris after the crash. Ken had managed to follow it as it crashed down miles away to the north near Koehn Lake. But, our efforts were for naught.

First, we were blocked by paving work on the Redrock Randsburg Road. A construction crew directed us toward a detour. We took a right onto Neuralia Road, then a left after the railroad tracks onto Cantil. About halfway down toward the tiny crossroads village of Cantil, we came across a grisly scene.

The spot where SpaceShipTwo's cockpit crashed. (Credit: Douglas Messier)
The spot where SpaceShipTwo’s cockpit crashed. (Credit: Douglas Messier)

SpaceShipTwo’s cockpit had slammed into the road, narrowly missing two truck drivers who had passed that spot seconds before. One of the first things I noticed after getting out of the car was a boot with a foot in it. Not far away was the co-pilot’s seat with what remained of co-pilot Mike Alsbury.

Traumatizing is an accurate description for the experience. I had arisen in the pre-dawn darkness four hours earlier at 6 a.m. I was so shaken I would not sleep until 10 p.m. the following day – a stretch of 40 hours. Ken was likewise shaken up; he later told me later that he got maybe an hour of sleep the first night after the crash.

I had thought that anticipating an accident would somehow make it easier to deal with when the time came. In some ways, maybe it did. I wasn’t at all surprised. The anticipation probably infused the accident with layers of meanings and emotions that were difficult to process given the shock I felt and the uncertainty over what happened.

With our path to the main wreckage blocked, we headed back toward Mojave after photographing the crash site. Ken had to file his pictures, and I had reporting to do. We couldn’t do either of those things out there. The area was so remote we barely had cell service.

The only official news that Friday came during a brief press conference at 2 p.m., about four hours after the accident. Scaled Composites and Mojave spaceport officials indicated the ship didn’t blow up. They did not go any further than that, undoubtedly in deference to the National Transportation Safety Board (NTSB) investigation that would start the following morning.

They already knew the cause of the accident from the data they had received from the ship, including cockpit video. But, they were unable to reveal what they had learned. The rest of us were left to speculate as to what had caused the crash.

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

It wasn’t until Sunday evening – more than two days after the accident – that NTSB officials revealed that Alsbury had unlocked SpaceShipTwo’s feather device too early during powered ascent. The spacecraft’s twin tail booms deployed due to aerodynamic forces. The ship hadn’t blown up; it had broken up.

It was jaw dropping. I couldn’t believe it. An experienced, highly skilled test pilot like Alsbury making such an elementary mistake. How could that have happened? I would have never imagined it.

Nor, it appears, did the engineers at Scaled Composites. They had built the feather system without a fail-safe system. As the NTSB put it in its investigative report released last month:

Probable Cause

The National Transportation Safety Board determines that the probable cause of this accident was Scaled Composites’ failure to consider and protect against the possibility that a single human error could result in a catastrophic hazard to the SpaceShipTwo vehicle. 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, which led to uncommanded feather extension and the subsequent aerodynamic overload and in-flight breakup of the vehicle.

And therein lay the real tragedy of SpaceShipTwo’s feather device. It was a safety system designed to deploy during re-entry to keep the spacecraft and its occupants from burning up in the Earth’s atmosphere.

Burt Rutan (Credit: Douglas Messier)
Burt Rutan (Credit: Douglas Messier)

Scaled Composites Founder Burt Rutan had designed it in response to an X-15 accident that had killed pilot Mike Adams in 1967. The rocket plane had entered the atmosphere sideways and broken up in hypersonic flight; Rutan was determined to make sure that would not happen to any of his spacecraft.

That singular focus, however, led to less attention on what could happen if the feather was deployed earlier than planned. The designers just assumed it would never happen. They were very wrong. And the result was tragic.

In flight test, you can worry immensely about the obvious things. But, it’s sometimes the things you don’t see coming that can kill you.

A Learning Period’s Painful Lessons

Back in 2004, an emerging commercial human spaceflight industry fresh off the triumph of Rutan’s SpaceShipOne convinced Congress to approved an eight-year moratorium (since extended) on government regulations. The industry needed a learning period during which to experiment with different technologies without the burden of government requirements. The Federal Aviation Administration’s oversight would be strictly limited to responding to accidents and close calls.

Inherent in this approach was a central conceit: industry had little to learn from government’s experience in these matters. Bureaucrats would be a millstone around industry’s neck – slowing things down, driving up costs, and contributing little or nothing toward making systems efficient or safe. There was an arrogance that said: get out of our way, we know what we’re doing. We can do this better than you can.

Eleven years later, the results of this approach are decidedly unimpressive. SpaceShipTwo has had as many casualties – four – as powered flights, with none of them coming anywhere near space. In addition to Alsbury, three Scaled Composites engineers died during a ground test in 2007. From that tragedy, Scaled Composites learned that nitrous oxide flowing under high pressure can explode on its own, with no fuel present.

Remains of Scaled Composites test stand after a nitrous oxide explosion in July 2007.
Remains of Scaled Composites test stand after a nitrous oxide explosion in July 2007.

Scaled also learned that it is best to design your rocket engine first and then design your spaceship around what it can do rather than the other way around. That lesson has resulted in years of delay and cost Richard Branson and Aabar Investments hundreds of millions of dollars.

Last year, Virgin Galactic and The Spaceship Company also learned that storing flammable fuel grain next to nitrous oxide tanks is a very, very bad idea. The companies narrowly avoided a catastrophe when a fire broke out and threatened to rupture the tanks. Such a blast could have taken out a good section of the Mojave Airport and the firefighters who battled the blaze.

The NTSB report on the SpaceShipTwo accident is chock full of lessons about how to build a space plane, the FAA’s failures to properly perform its limited oversight responsibilities, and the shortcomings in emergency response planning that left injured pilot Pete Siebold lying in the desert for half an hour before the first helicopter arrived.

Some people were – and still are – quite angry over my errors after the loss of SpaceShipTwo. I get that. It wasn’t my finest hour. I could, and should, have done much better. But, things were chaotic and uncertain. And mine wasn’t the worst error made that day – or during the decade-long life of the SpaceShipTwo program.

I learned some valuable lessons from my failures. First, what you think you see is not necessarily what actually happened. Your eyes can deceive you, especially during sudden, fast-moving events in which things happen literally in the blink of an eye.

I also learned that in aviation accidents, your first instincts about what caused a crash can be wrong. It is best to reserve judgment and let all the facts come before rendering a verdict.

These were painful but necessary lessons. The pain paled in comparison to that felt by Alsbury’s family, friends and colleagues. I always keep them foremost in my thoughts when I think about this accident. This program has had a terrible human cost.

  • Hemingway

    Have lessons been learned? Virgin Galactic chief executive George Whitesides just said things are on track now.

  • Terry Rawnsley

    Great article Doug. It is always difficult to acknowledge your own mistakes, even when they are ultimately inconsequential.

  • Hemingway

    Excellent commentary!

  • tdperk

    ” Inherent in this approach was a central conceit: industry had little to learn
    from government’s experience in these matters. ”

    I believe the main conceit there is the idea that industry will learn anything from government except how to keep the cost of space access upwards from $5k/lb.

    That there should be no single point of catastrophic failure is already an established aerospace norm with no government regulation involved. That government will itself fall prey to institutional pressure to do ultimately disastrous things is seen in the decision making process leading to the two Shuttle losses.

    Regulation can not prevent humanity–thinking it can is a conceit in and of itself.

  • Douglas Messier

    It’s interesting to contrast and compare commercial human spaceflight with the commercial cargo and crew initiatives by NASA. Very different approaches. The latter involves partnerships in which the space agency is trying to transfer as much knowledge and expertise as possible so the private sector can perform these functions in a safe and much more affordable manner. The success of COTS and CRS have been mixed, but the cargo failures have been relatively low risk. Best to learn the lessons there before flying astronauts.

    There’s an argument that commercial human spaceflight companies will fly safely because they have an incentive to do so. Otherwise, they’re out of business. This will make them somehow superior to government.

    This assumes that private companies have the skills and ability to fly safely. That’s a tricky thing to do; safety is a lot more complicated than it looks. Especially with first-generation spacecraft with little or no flight history.

    The other assumption is that private companies won’t fall into the same traps as NASA did with its human spaceflight failures. That schedule pressures and the “hey, we’ve flown with this before and nothing happened, it’s safe enough” mentality won’t creep in and adversely affect operational decisions. I’m not convinced that is true.

  • windbourne

    Great article, to a sad moment in flight.

  • windbourne

    NASA really is a brainshare that so many private companies can and should learn from. The accumulated knowledge amongst its engineers are second to none.
    Sadly, many in private space, want to to ignore it. Likewise, we have politicians and various political groups, working hard to kill off NASA. To do so, would be a great loss not only to America, but to Humanity.

    If nothing else, I see this as the same lesson that SpaceX learned with F1-1, except for the fact, that this involved human life: That is, there is still good and useful procedures and steps that must be taken and can best be learned from experienced groups such as NASA.

  • ThomasLMatula

    Good article. It appears that one major lesson is still apparently not being learned, namely that you need to limit risk to those on the ground when testing rocket vehicles. I still find it amazing there was a open road with traffic and a road crew near the flight path that had no idea about what was going on. When rocket vehicles, not just sounding rockets, but air to surface and air to air, are tested at WSMR all roads in the area are closed. It is surprising that is not being done in Mojave and I expect is another example of the lack of serious regulation by the FAA.

    Hopefully they will look to some of the military’s experience in this regard. In the old X-15 days they plotted its flight path carefully to minimized its time off range and risk to those on the ground from any accidents. And the area was much more empty in those days. They also had multiple chase planes and helicopters for security and safety as well as to assist the pilots of the X-15. In one incident a light plane was headed for a lake bed used landing by the X-15. It cleared out fast after being buzzed by a F-100 chase aircraft. They also had a C-130 with a full rescue team aboard, including a crash truck, for rapid response to emergency landing sites. It saved John McKay’s life in the roll over crash in 1962.

    Although those precautions would likely be far too expensive for VG, they should at least look at how to better respond to any future accidents. And also how reduce the element of luck in protecting those of the ground from being harmed.

  • ThomasLMatula

    There are still folks who think the Challenger orbiter blew up, not that the external tank rupturing from the stress of the O-Ring leak resulted in the breakup of the Orbiter from the stress of over 20G from the resulting attitude change of the vehicle. So it’s a natural mistake given the limited data available.

  • ThomasLMatula

    The problem is that folks working on a vehicle become blind to failure modes from being too familiar with it. It’s like a writer trying to proof read their own books, they will miss things that will jump off the page to a fresh pair of eyes seeing it.

  • Vladislaw

    we accept 35,000 deaths on highways every year. Planes crash, boats sink, trains derail. If government was running those transportation systems and a congressional committee would have be to formed after ever accident and mishap the Nation would grind to a halt.

    No one really bats an eye over commercial transportation deaths. We may force, through legislative action to improve safety, like seat belts or air bags, but we still accept a lot of mayhem in transportation. I believe it will be just like this once space transportation becomes just another routine transportation system.

  • Dave Salt

    My understanding is that commercial companies are working with government agencies (e.g. XCOR’s use of wind tunnels), just not under big contracts like CRS.

    Basic design analysis techniques (Hazard Analysis, FMECA, etc.) that would have highlighted such single point failures are employed by all aerospace vehicle production companies, including Scaled, so a lack of federal oversight is no excuse for ignoring this hazard.

  • Henry Vanderbilt

    Not to directly contradict your point; there is indeed a huge amount of useful expertise in NASA.

    But, a major caveat: That expertise is embedded within a bureaucracy that (depending on which bit of NASA we’re talking about) ranges from mildly dysfunctional to destructively pathological. And the worst parts of it (naming no names) are (for good reason) chronically underemployed, and thus eager to lamprey on to any project that comes too near them.

    In other words, you have to be really careful in seeking help from NASA to keep the net effect positive.

    Comercial Cargo did an outstanding job of steering through these rocks. Commercial Crew, the jury’s still out, but signs are, less so. Some of which was inevitable – after Commercial Cargo’s success, the lampreys were alerted.

  • I’ve already said (and warned about them) ALL these things FIVE YEARS AGO in this article

    but YOU (and the admins of other space blogs) have asked me to stop to post here, a thing that I’ve done, and that I will done again, after this lone post

    now, great part of the same space bloggers say that THEY have always said that and cry crocodile tears for the pilot dead on a veichle clearly dangerous BY DESIGN, from day 1

  • SpaceTech

    LoL, I always love the look I get when I try to explain the fact that Challenger didn’t “Blow Up”. When I explain the aerodynamic break up scenario I have been told “You’re crazy” more times than I can count.

  • Douglas Messier

    They’re always on track. Just seems to be the same track over and over again. Shortly after the crash, they said they’d complete the second SpaceShipTwo in six months. There were people in Mojave looking at where they were at the time and saying that wasn’t going to happen.

    They’re not the only ones to underestimate schedules. XCOR is years behind on their estimate for Lynx flights. Virgin just seems to be more driven by PR and marketing. I’m sure six months sounded good at the time.

  • HeftyJo

    The Challenger orbiter tried as hard as it could to maintain proper attitude. The telemetry showed that the SSME’s were gimbaling to their maximum range to try and hold course.

  • Douglas Messier

    Thanks, Terry. And thanks to the others who praised the post.

    I debated whether to even publish it. Some folks urged me not to publish it. They told me I wouldn’t win over any of my critics. They also felt my mistake was relatively inconsequential in the scheme of things.

    But, others don’t feel that way. I’ve gotten a lot of criticism over this. So I wanted to put it all down on the record.

  • Vladislaw
  • Terry Rawnsley

    I think that the most important thing that I took from this article is that journalists are not cold, impassive recorders of events. Although they try very hard to not become a part of the story they cover, they still feel its effects on themselves and the story they tell. You needed to exorcise a demon that was affecting you. Sometimes the only way to do that is publicly. Congratulations for having the courage to do what other, more widely known journalists, have failed to do.

  • Richard

    I do find it somewhere between sad and deeply ironic that the two touted ground breaking safety features of SS2, the feather and the ‘benign’ engine are the two things that have caused deaths in the program. Doug as for your immediate interpretation of events, I cannot understand how you could have come to any other conclusion than you did, when I look at the videos released at the time, it looks like an explosion. I still think most people on the street would think the same thing if you played the videos today, even the onboard ones, it happens so fast.

  • Doug Weathers

    Great article, Doug. Thanks for writing this.

  • Ray Baker