Slashdot Mirror
SpaceShipTwo's Rocket Engine Did Not Cause Fatal Crash
astroengine writes It wasn't SpaceShipTwo's hybrid rocket motor — which was flying on Friday with a new type of fuel — that caused the fatal crash, the head of the accident investigation agency said late Sunday. The ship's fuel tanks and its engine were recovered intact, indicating there was no explosion. "They showed no signs of burn-through, no signs of being breached," Christopher Hart, acting chairman of the National Transportation and Safety Board, told reporters at the Mojave Air and Space Port in Mojave, Calif. Instead, data and video relayed from the ship show its hallmark safety feature — a foldable tail section designed for easy re-entry into the atmosphere from space — was deployed early, causing the in-flight break-up.
Normally, the feather system wouldn't be unlocked until the rocket-powered spaceship is moving about Mach 1.4, or 1.4 times faster than the speed of sound.
Instead, the co-pilot moved the lever from locked to unlock when the spaceship was traveling at about Mach 1, Hart said.
Modern cars have a feature where you can't take the key out of the ignition if you forget to put the thing in park. This feature saves me from stupid about twice a year. Humans are error-prone - there's a whole field, poka yoke dedicated to preventing these sorts of errors.
Test pilots are the best of the best. If one of them can make a catastrophic mistake then so can any commercial pilot.
Now, they may have figured that that sort of safety gear was "for later" and test craft are often bare-bones, and test pilots are often relied on to not make those kinds of mistakes. Assuming the premise here, we might see more automation early in the design process going forward. Virgin might be able to survive a year-long investigation but that kind of delay is an ongoing liability. It may turn out to be faster and cheaper in the long run to add in those costs up front, if delays are calculated into the cost.
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
What the reporter states doesn't make sense.
If there was a structural failure related to the tail, one would imagine it would be because it was deployed while the vehicle was traveling faster than the design speed for the tail, causing greater than planned aerodynamic stresses. Saying it failed because they were going slower than the normal deployment speed just doesn't make sense. Things break when you overstress them, not when you under-stress them.
Am I missing something?
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
This factor, not mere ideology or efficiency of free markets, is the reason we need to privatize risky technologies. The problem with a government effort is not that it is marginally less 'efficient' than a private one, but that in a Luddite-dominated culture a government effort, unless we can make it military and secret, will be doomed by its inevitable first accident. The Challenger crash caused a two-year delay of NASA's most advanced manned system, and the Columbia crash killed it for good.
Nope - under ICAO, for air crashes, preliminary report is supposed to be within one month.
Plus, this was a test flight, and was probably instrumented to death - very likely they have full real time telemetry before they even need to go near recorders in the wreckage. Apparently they have cockpit video too - something the aviation industry has always resisted (at least the unions).
Normally, the feather system wouldn't be unlocked until the rocket-powered spaceship is moving about Mach 1.4, or 1.4 times faster than the speed of sound.
Instead, the co-pilot moved the lever from locked to unlock when the spaceship was traveling at about Mach 1, Hart said.
Modern cars have a feature where you can't take the key out of the ignition if you forget to put the thing in park. This feature saves me from stupid about twice a year. Humans are error-prone - there's a whole field, poka yoke [wikipedia.org] dedicated to preventing these sorts of errors.
Test pilots are the best of the best. If one of them can make a catastrophic mistake then so can any commercial pilot.
Now, they may have figured that that sort of safety gear was "for later" and test craft are often bare-bones, and test pilots are often relied on to not make those kinds of mistakes. Assuming the premise here, we might see more automation early in the design process going forward. Virgin might be able to survive a year-long investigation but that kind of delay is an ongoing liability. It may turn out to be faster and cheaper in the long run to add in those costs up front, if delays are calculated into the cost.
In test aircraft, you start with highly trained pilots and allow them to do what they need to do with the controls, since you don't know when an unusual use of the controls might save the crew or the ship. In later iterations, you build automation to model what the pilots did successfully.
Even in light airplanes, with only a few seats, it is frequently possible to set the controls in a way that will damage the engine or airplane, with no interlocks. It's why learning to fly takes time and has some expense. It's also why there are additional ratings for more complicated airplanes. On most business jets that are 20 years old or more, you can damage the engines by starting them incorrectly. You can certainly set the engines to a fuel burn that will overheat the engines during flight, the pilots can choose 115 to 120% of normal power for a few minutes during takeoff (safely, and routinely) but cannot use that for cruising flight.
Also, automation works well when all variables are known and controlled. In this case, the ship is not meant to feather while the engine is running, and also not while the ship is in the thicker atmosphere. It's supposed to only feather when the ship is in the higher, thinner air and the engine is shut down. At that higher altitude, Mach 1.4 could be under 200 miles per hour of indicated airspeed. Mach changes with altitude, which is why it's used instead of indicated airspeed. Reentering from the edge of the atmosphere would likely be a bit complicated for automation, but can be taught to a test pilot in a few days in the simulator.
As much as we may not like to admit it, the human brain has a much more advanced sensor pack and faster processing power than dedicated computers, plus the ability to handle edge cases very well.
I'd point out that if they had some kind of automated interlock, and that system prevented feathering (incorrectly) when feathering was needed, that would also cause loss of the vehicle due to letting it get too fast upon reentry.
If Orrin Hatch had not used his power to keep the shuttle booster manufacturing in Utah then the crew of Columbia would be alive today.
Since Utah is nowhere near where the shuttle is launched they had to be made in pieces so they could be shipped across the country. A less safe design. The original design called for the boosters to be made completely in one piece. O ring seals can't fail when there aren't any O rings to begin with.
The crew of Columbia died because of pork barrel politics.
Ha! Rather than merely point out that you have confused the names "Columbia" and "Challenger" everybody has decided you needed a new arsehole tearing.
All I want is a secure system where it's easy to do anything I want. Is that too much to ask ~~ Randall Munroe