Virgin Galactic Spaceship 2 NTSB Crash Report

[Kilmeister]

NTSB has finally released a report that has basically confirmed what everyone already suspected. Co-pilot activated a breaking system too soon which resulted in loss of vehicle. Not too much too see here except that the NTSB sends a few harsh words towards Scaled Composites about training and engineering.

http://www.nytimes.com/2015/07/29/science/space/virgin-galactic-spaceshiptwo-crash-traced-to-co-pilot-error.html?_r=0

[SpaceXray]

I'm surprised how the pilot managed to pull the wrong lever at the wrong time. How exactly do you confuse the control stick with the feather release handle? Very strange, but that's what happened judging from the video they released.

[Robotengineer]

I'm surprised how the pilot managed to pull the wrong lever at the wrong time. How exactly do you confuse the control stick with the feather release handle? Very strange, but that's what happened judging from the video they released.

That's what I'm wondering. Perhaps Scaled Composites and/or Virgin Galactic were ignorant/negligent enough to put the feather release handle right next to the control stick? Or put it next another stick/handle?

[PakledHostage]

Human factors is a complicated thing. What may seem obvious to someone judging from the comfort of their armchair may not be so obvious to people in the heat of the moment.

As an example, the gear lever on commercial transport aircraft, even in this day and age of glass cockpits, are mandated to have a tire shaped knob on the end. The allowable profiles are right in the FARs. Clearly there have been problems where highly trained but fallible people have grabbed or failed to grab the landing gear retract lever at some critical time.

[Kilmeister]

That's what I'm wondering. Perhaps Scaled Composites and/or Virgin Galactic were ignorant/negligent enough to put the feather release handle right next to the control stick? Or put it next another stick/handle?

When you have to consider how many controls and indicators there are in a modern cockpit, you can imagine the difficulty engineers face when factoring in so many possible scenarios. It's not hard to see something confusing for a pilot being overlooked by an engineer designing the thing. Reminds me of an episode of Air Disasters I watched once. An autopilot instrument to set a vertical descent rate had two modes: Descent Angle, Descent Rate. The display was small and only was 2 digits to save space on the crowded dash. The difference between the display was "3.3°" and "33". The Pilot attempted to set to 3.3° angle of descent, but set it at a hefty 3,300 feet/min rate which ended up plowing them into a mountain many miles short of the runway. It wasn't the only factor into the crash, but it was a significant one.

https://en.wikipedia.org/wiki/Air_Inter_Flight_148

[Kryten]

This wasn't the 'wrong lever' as such, it was a lever activated a little while (14 seconds) before it was necessary. It doesn't seem the pilots were aware activating it that early would have this effect (remember it's merely an unlock), and they were both under a very high workload.

EDIT: Or to directly quote the report;

3. The copilot was experiencing high workload as a result of recalling tasks from memory while performing under time pressure and with vibration and loads that he had not recently experienced, which increased the opportunity for errors.

and

8. Scaled Composites did not ensure that the accident pilots and other SpaceShipTwo test pilots adequately understood the risks of unlocking the feather early.

[VirtualCLD]

When you have to consider how many controls and indicators there are in a modern cockpit, you can imagine the difficulty engineers face when factoring in so many possible scenarios. It's not hard to see something confusing for a pilot being overlooked by an engineer designing the thing. Reminds me of an episode of Air Disasters I watched once. An autopilot instrument to set a vertical descent rate had two modes: Descent Angle, Descent Rate. The display was small and only was 2 digits to save space on the crowded dash. The difference between the display was "3.3°" and "33". The Pilot attempted to set to 3.3° angle of descent, but set it at a hefty 3,300 feet/min rate which ended up plowing them into a mountain many miles short of the runway. It wasn't the only factor into the crash, but it was a significant one.

https://en.wikipedia.org/wiki/Air_Inter_Flight_148

To add to that, this is what I've heard from hearsay:

The co-pilot was supposed to unlock the feather system at Mach 1.4, but unlocked it too early at Mach 0.8. It's not hard to imagine someone who accidentally unlocks or activates a system a little too early when operating a complicated piece of machinery in a busy situation. The reason for unlocking the feather "braking" system at Mach 1.4 is due to the fact that aerodynamic forces are strong enough to keep it down. If the system doesn't unlock , then they abort the remainder of the rocket burn, dump oxidizer and land. If they wait to unlock the feather system once they are "in space" and find out it won't unlock, then they'll just break-up on re-entry.

EDIT: Ninja'ed by PakledHostage

[Robotengineer]

Why wasn't the feather release automated? It seems like a no-brainer to automate such a crucial part of the mission and craft.

[Dkmdlb]

Wired reported that doing things manually was part of a larger plan to keep things simple.

The design philosophy behind the two aircraft was to keep everything as basic as possible. “A simple system is less likely to fail,†chief pilot Dave Mackay told us during a recent visit to Virgin Galactic’s HQ in Mojave, California. Mackay, a former RAF test pilot, has flown 130 kinds of aircraft. He tested the Harrier hover jet and flies a 1909 Bleriot for fun. The man has skills, and they’re put to use on WhiteKnightTwo and SpaceShipTwo, which have no flight control computers. All the systems are manually operated. No autopilot here: When all your pilots are among the best in the world, you don’t need one. Better to let them just fly on their own.

Source

[PakledHostage]

If your wired quote is accurate, then it sounds like the problem may run deeper than just some unfortunate engineering decisions. An organizational culture of arrogance and complacency has killed more than a few highly skilled people who were among the best in the world at what they did...

[Bill Phil]

Wired reported that doing things manually was part of a larger plan to keep things simple.

Source

The problem is that even if the aircraft is mechanically simple, humans are not.