Virgin Galactic pilot recalls colleague's crash | Hacker News Reader

Virgin Galactic pilot recalls colleague's crash

bbc.co.uk

44 pointsRobadob29 comments

There is a really good piece from Newsnight on Virgin Galactic: https://www.youtube.com/watch?v=M-gxmqUfQ80

I would encourage you to watch all of if you're interested in Virgin Galactic.

It is pretty clear to me that they have one last chance to get this right. If they suffer another catastrophic failure any time in their next 4-5 flights, they will need to shut down. The fact they do not seem to be addressing design flaws concerns me. If I were a customer I'd be seeking a refund.

I hope my pessimism proves to be unfounded and they usher in a new era, but right now it must be hard for all of them over there.

I am not an engineer, but it occurs to me that high-velocity aerodynamic flight is still a seriously difficult problem; and one best avoided if not actually neccessary. The resemblance to an aicraft hides the fact that it is something completely different at various stages of its flight. And in all these stages there are different problems that could occur, and limited experience of those issues.

Looking at it from another angle, configuration changes are often a source of problems.

So here the plane has two aerodynamic configurations, aircraft mode and shuttlecock mode. (Technically being carried is also one, and being a rocket vs glider is debatable.)

Now the wrong mode was entered at the wrong time, more precisely at the wrong dynamic pressure (altitude and speed) and disaster struck.

It'd be nice to avoid having to have such modes but with current power sources it's probably unavoidable to have some. Staging is a mode change as well. (If you were powered by say, antimatter, you could make the rocket a big stainless steel ball with just a powerful engine that could take anything and fly in any attitude, almost at any speed at any altitude without breaking up. Aerodynamic moments would be tiny.)

Airplanes also have mode limitations, like that you can't extend the landing gear at high speeds.

I'm sceptical how the co-pilot somehow did the single most wrong thing to this aircraft possible (unhinging the wing).

There must be some interesting report to be had on how design can kill if that is what went wrong.

It's a fun exercise think about how this kind of failure can be compensated while always having to ask, what if that compensation also fails.

The flight computer could easily work out that the wing can only be unlocked AFTER the violent acceleration only during a complete free-fall, but then it's no problem thinking of a hundred ways that computer could fail. Plus you probably want a manual override anyway to the lock etc.

If that was actually my job I'd never sleep easy.

Its an experimental aircraft doing test flights, I don't think they spend a lot of time trying to make it blunderproof. Theres a million things the test pilots could do to endanger the craft. I imagine they mostly just assume that the pilots will not mess up.

Exactly. A pilot could also fly it into the ground, or into a building, or into WKII. Do we need physical safety mechanisms to prevent this? Or can we depend on pilot training?

More importantly, there are a bunch of scenarios where the wing needs to swing that don't involve the pre-conditions you've specified. No one knows what they are, but we know they exist--because they always do--and the last thing needed is for the computer to tell the pilot, "I'm sorry I can't do that Dave" when the pilot tries to do something it has been programmed to not allow. At that point a software developer's imagination is flying the aircraft, which is never a good thing.

That we can't imagine those scenarios right now is irrelevant to this point: we can't imagine the correct physics beyond the standard model, either. That doesn't mean we won't discover it as we gather more data. But in the aircraft industry "more data" often involves dead people.

We generally find it easy to reason like, "In context X do (or don't do) Y until condition Z." The hard part is figuring out what consquences that will have in context P, Q or R... each of which is likely to have a very low probability, but one of which is almost certain to happen eventually.

No new info there. We've known for ages that one of the pilots unlocked the feathering system at the wrong time.

It's a tough problem to solve but Virgin Galactic have claimed they will be ready for passengers "in less than 2 years) every year since they launched. In 2004.

I'm curious what's the government restriction to photography of the assembly process. Is this an export controls thing?

First, you can already see this stuff on youtube. And second, there are some advancements here but they seem of more commercial interest than military. And third, they're still having engine trouble (they just did a major tech pivot) so it's not clear you'd want to copy it anyway.

My guess is it's Richard Branson's knack for self promotion.

It's the ITAR:

https://www.pmddtc.state.gov/regulations_laws/itar.html

I had not realized how delayed this project had become, crash or no crash. It does seem odd that something so vital to the safety of the craft wasn't protected by systems to warn the pilots. It might be, but I am thinking along the lines of treating it like stall warnings or such or having an interlock that cannot be thwarted easily

What's with the title? No details have emerged (although the full report is due in a few months). Even BBC's title is misleading. "Virgin Galactic pilot recalls colleague's crash"- he said he couldn't see anything.

We changed the title from "Virgin Galactic crash details emerge" to that of the article.

That was the original title at the time of submission, they appear to have updated the title (and maybe the article) since.

Reading this I can't help comparing it against SpaceX.

SpaceX is actually 2 years older than Virgin. But SpaceX is Getting Things Done at warp speed for some years now, while Virgin Galactic keeps having multiple crashes. Despite SpaceX having a harder mission.

Is it all due to the Elon Musk effect?

What's the secret?

Why is the SpaceX missions harder? Only a couple of aeroplanes have ever gone more than mach 3, and with significant difficulty. In comparison there are numerous rocket launch vehicles with thousands of successful launches. The atmosphere is a very hostile place and aerodynamic control is not trivial.

Getting stuff into a suborbital trajectory is basically a subset of getting stuff into orbit. It's basically like, if you can drive from NYC to LA, you can drive from NYC to Chicago.

It could very well be that Virgin's chosen approach is more difficult because they are using airplanes instead of rockets. But nothing forced them to choose that approach, and it's not the mission. If you find it harder to get to Chicago than I do to get to LA because I'm driving a car and you're using a kayak, that doesn't mean your mission is harder.

Except a car costs $30,000 and a Kayak costs $300 with much lower running costs. Virgin are not trying to be the first to get into Space, they're trying to reduce the cost and make it more affordable.

Lower cost of access to space is SpaceX's big goal as well. So far they're doing much better at it. I wouldn't be surprised if SpaceX could meet Virgin's price for a suborbital trip, if they felt it was worth their while. But they have vastly larger fish to fry.

I just figured that Virgin Galactic doesn't have nearly the resources that SpaceX had or has. Getting stuff into orbit with hundreds of millions of dollars to spend would be easier than getting stuff into sub-orbit with peanuts to spend.

But then I checked Wikipedia and:

"After a claimed investment by Virgin Group of US$100 million, in 2010 the sovereign wealth fund of Abu Dhabi, Aabar Investments group, acquired a 31.8% stake in Virgin Galactic for US$280 million, receiving exclusive regional rights to launch tourism and scientific research space flights from the United Arab Emirates capital. In July 2011, Aabar invested a further US$100 million...."

On the SpaceX side, according to this page, SpaceX spent $390 million developing Falcon 1 and Falcon 9, total:

http://www.parabolicarc.com/2011/05/31/nasa-analysis-falcon-...

I don't know what the difference is. Maybe this is another example of how people should stop trying to use airplanes to get to space.

The difference probably has to do with the fact that SpaceX has a functional revenue model beyond what's essentially a pre-order. Further down in the Wikipedia article for SpaceX, as of 2012 they had taken in over $4 billion in lifetime revenue. Also, they got a $1 billion investment from Google and Fidelity in exchange for 8.333% of the company this past January.

I think the answer really is money, SpaceX has more of it because it has built a product it can actually sell right now.

[1] http://en.wikipedia.org/wiki/SpaceX#Funding

That seems backwards. They didn't have ongoing revenue until they proved themselves capable. Their early days were a similar situation to Virgin Galactic now, but, apparently, with even less money and doing harder stuff.

Space X and Virgin Galactic are building completely different products. The only commonality is that they both go into Space (and Virgin Galactic barely even that).

SpaceX is not actually doing much in the way of 'new', to date. Their main product is an improved rocket engine design, but it's just an improvement on an existing product. (The self-landing rockets are amazingly impressive, but they haven't managed to achieve success with it yet).

Virgin Galactic is a completely new design (launching a passenger rocket plane in mid flight), at least in the non-military sector. That's a whole world of new unknowns they are having to overcome and, when you consider that, it's incredibly ambitious what they are hoping to achieve. They are less likely to achieve success as a result unfortunately. But well done for being the first to try.

The other big difference of course is money. SpaceX has way more funding behind it, and a much greater commercial potential.

The X-15 was a rocket plane that went into space in the sixties. It was operated by the air force and NASA. I don't know if there's a big difference in the novelty compared to more traditional rockets.

Slightly different requirements.. X-15 was single seat, no provision for sight seeing, trained test pilots that could handle high G forces, etc..

Also, X-15 had much more tolerance for failure, frankly. When you're doing something for the first time with military test pilots, you can be 95% reliable. Virgin has to be nearly 100%, and that last 5% is a bitch.

In 2008 SpaceX had a terrible year (3 launch failures in a row) and almost went bankrupt http://www.space.com/5693-spacex-falcon-1-falters-time.html

edit: along with Tesla. interesting story: http://inspiremore.com/in-3-days-elon-musks-tesla-motors-spa...

> Virgin Galactic keeps having multiple crashes.

Uh, what? This is the only crash they have ever had. They did have an accident on a test stand in 2007 when a tank blew up, but that wasn't a crash.

Reply on news.ycombinator.com