Spaceflight Now has the video released by the NTSB.
[Update a while later]
Wayne Hale: Pilot error is never the root cause.
[Afternoon update]
Dough Messier has broken down the crucial seconds frame by frame.
9 thoughts on “SpaceShipTwo’s Final Moments”
Comments are closed.
I’ll start by admitting some ignorance of this topic, because I’m not that interested in Spaceship 2, because if successful it seems to be more of a hobby effort than a progression of useful spacedlight. I see nothing wrong with that effort, but it doesn’t excite me either.
But I did have a strong emotion when I first heard of the crash and what the pilot did, and it can be expressed like this: “what flight envelope would support the rocket engine firing and the wings being feathered simultaneously?” There may be an answer and I’m ignorant of it, but I can’t think of one. So with this ignorance, I wonder, “why would the system allow for the feathering while the rocket motor was firing?”
I don’t know if this is what Rand or Wayne are thinking. I know George in a previous thread mentioned interlocks, but this to me, in my ignorance seems simple. In fact, for a short while, I thought perhaps it was a test objective to verify the interlock worked to prevent feathering with the motored powered, but there are safer ways to test such a system, and it turns out it wasn’t a test objective.
I have no idea why the pilot would feather the wings, but I also can’t come up with one reason an engineer would allow them to do so. Maybe one, an engine lit malfunctioning sensor, but an override more complicated then just moving the feather switch would be a practical engineering solution.
Oh well, I’m sure someone will fixate on my ignorance, and I’ll be dismissed as an idiot. So I’ll return to not caring about aerospace and making money elsewhere…
I think they want to feather while still under thrust so they can abort if there is a deployment failure. But it’s not supposed to happen until they’re well past max q, and through
subtransonic.I think I’ve noted in the past that this was an example of a “safety” feature (feathering wings) killing people. It was a poor choice, IMO, made because Burt didn’t understand or trust rocket engines. Same reason he went with hybrids.
I’m thinking here that MaxQ should not just be a forward velocity, but add an altitude. Nominal MaxQ for shuttle entry was at HAC, but punch a hole in the leading edge, and dynamic forces become pretty violent at 200,000 feet.
Max Q is function of both velocity and altitude. It increases with former, and decreases with latter. That’s why it peaks as you ascend.
Or, rather, Q (dynamic pressure) is a function of those things. Max Q is the peak.
To be precise, dynamic pressure, usually denoted by the lowercase letter “q” with a a bar over it and pronounced “qbar”, is defined as 1/2 * rho * V^2, where rho is the atmospheric density and V is the true airspeed. Note that the bar is used over the q to distinguish dynamic pressure from body axis pitch rate, which is denoted by just “q”.
For a given angle-of-attack, sideslip, Mach number, and for a rigid vehicle, the aerodynamic forces and moments scale linearly with qbar, if you can ignore Reynolds number effects (Re increases linearly with airspeed, and has a strong effect on drag, as Re largely determines where the boundary layer will transition from laminar to turbulent flow, but is less important for non-drag terms). Flexible effects are much more complicated; some terms scale with qbar, some don’t.
I understand the Q, so at the very least training should not have been “feather at certain Mach number”, but feather at certain “altitude and Mach number”. I realize the job of the test pilot is to learn what the envelope is for the right altitude and Mach number, but the guidance I’ve read to date seems poor.
I struggle with some of this, because I’m happy to pin the error on the test pilot, assume the vessel is clear to fly, and move on. Because that lowers the cost of spaceflight and again, is the job of the test pilot. At the same time, this seemed utterly ridiculous in terms of steps. I get it, the test pilot commanded feathering early, but I just don’t get the procedures.
Anyway, the folks involved are far more pained about this, I’m sure. Hopefully, they’ll learn and add a means to prevent this from so easily occurring in the future.
Assuming they were on the correct trajectory, the Q would be inferred from the Mach number.
A clarification: Alsbury didn’t feather the tails early; he unlocked the feathering mechanism early, before the vehicle had gone through the transonic regime, where the loads are going to be high because of the combination of aerodynamics and qbar. If you look at, say, the nondimensional stability derivative CLalpha, aka the lift curve slope, as a function of Mach number, it will peak in the transonic region – roughly Mach 0.9 to 1.3 or so. Other key stability derivatives do the same. So, the procedure was to unlock the feathering mechanism after passing through the transonic region, to confirm that the feathering mechanism was usable prior to committing to altitudes and Machs where the feathering is essential for safe descent. When Alsbury unlocked the mechanism early, the transonic airloads overpowered the feathering actuators, and the tail feathered, with catastrophic results.
I agree with Rand that Rutan’s “breakthrough” was never really a good idea; the failure modes are tough to deal with. How do they test the feathering actuators prior to flight? Every aircraft I’ve worked on did a full range sweep of every actuated surface prior to flight; can SS2 do that with the feathering mechanism? What is the redundancy level of the actuation system? It becomes a single-point failure; to my knowledge, it is very unusual for civilian aircraft large enough to require powered actuation to have a primary flight control surface be a single point failure (military aircraft are a different story). The feathering mechanism isn’t a traditional control surface, but it is definitely safety critical.