NTSB: Uncommanded Feather of SpaceShipTwo
NTSB Releases New Clues in SpaceShipTwo Accident Investigation (with video), SpaceRef
“Acting Chairman Christopher Hart’s of the National Transportation Safety Board (NTSB) released new clues at the third media briefing Sunday evening.
1) The vehicle had a nominal release followed by a nominal Ignition.
2) Approximately 9 seconds after ignition, the “feather” parameters changed from lock to unlock.
According to Hart the “normal procedures are to command the feathers to unlock after Mach 1.4 so aerodynamic forces do not extend feathers prematurely.”
Found fuel tank, oxidizer tank, and engine – no sign of burn through or being breached.
— NASA Watch (@NASAWatch) November 3, 2014
Hart: Review of cockpit forward looking camera shows that the feather was unlocked by the copilot. #SpaceShipTwo
— NTSB (@NTSB) November 3, 2014
Hart: Normal procedures are to unlock feathers after Mach 1.4 so aerodynamic forces do not extend feathers prematurely. #SpaceShipTwo
— NTSB (@NTSB) November 3, 2014
Hart: Engine burn was nominal up until feather extension. #SpaceShipTwo
— NTSB (@NTSB) November 3, 2014
Hart: Approximately 2 seconds later, just above Mach 1.0, feathers moved toward the extended position….#SpaceShipTwo
— NTSB (@NTSB) November 3, 2014
Folks:
Well, there’s your side loads. Feather at Mach 1 would rip the tails off just like we saw.
That’s good news and bad news.
The good news is that SpaceShipOne was a perfectly good spacecraft flown in a way that was known would destroy it.
The bad news is that, if it were simply a procedural error, a life was unnecessarily lost.
tinker
So, perhaps a silly question Tinker, and people here will know far more about it than me: think of the way some aircraft have non-powered leading edge slats that automatically deploy as aerodynamic forces reduce below a specific airspeed.
Is their a way to design the feathers to similarly deploy only below the required Mach number and otherwise remain retracted?
There is no below a required Mach number. There is something called a critical Mach number which happens just before all the regions of airflow across the vehicle become supersonic. That’s typically about Mach 1.2. Between Mach 0.8 and 1.2 is typically where one would see the highest drag loads. Reentry is at a higher Mach number than that, it’s just that drag loads are considerably less at that altitude because of air density. So, there isn’t really a number below that which it would be safe for the feathers to be deployed. It’s a range of Mach numbers and altitudes that feathers shouldn’t be deployed in.
But to answer your original question, the answer is yes, many different ways. The question is whether or not they can afford the weight increase required to implement that kind of a solution? SpaceShipTwo is a pretty bare bones design without the level of redundancy or complexity of the Space Shuttle. And just like the shuttle, every pound is critical as it takes away from performance. It becomes a question of how many seats are they willing to give up which is a big issue for a commercial concern.
Not being an engineering-type I always thought that the ‘feather’ design totally creative and the work of SC. Yes?
Um, kinda dumb/obvious question — in normal flight, does the feathering occur while the engine is still firing?
.
The unlock lever was moved, but the feather lever was not. Sounds like something was hooked up wrong. It is not clear to me if the feather would happen automatically. I would think if there is a feather lever, it would only be done manually. Glad it was not the engine. I wonder if they have another Space Ship or will have to build another.
Unlocking to keep from feather sounds backwards. If locked it could not feather. Maybe it needed to float. One theory then is that because of the thicker air it forced the tail into feather. So what moves the tail. If an actuator, then it would keep it locked unless forces overcame the actuator. Up elevator, actually down elevator, if it was floating, could be used to move the tail up and lock it. But they have video of the cockpit. If a jack screw is used, it could strip. An airliner was lost because the elevator trim jack screw was not greased and it wore enough to strip. My theory was that the elevator control would be reversed because it was free floating. Moving the elevator up would cause the stab to nose up, giving full down. The pilot pulled on the yoke, got full down and could not understand why it was diving and kept holding it. The airplane did do an outside loop, which supports my theory. NTSB did not agree.
Thanks NTSB for finding and reporting what happened so quickly. Now what happened to Antares?
VSS Voyager is ~60% complete.
VSS Enterprise was the one destroyed.
My understanding is that VG and contracted with SC for 4 or 5 SS2’s and 2 WK2’s
I get the impression that aerodynamic loads at certain speeds and altitudes can force the feather open, overpowering the actuators. That’s why there is a lock for the mechanism.
…but you don’t know. The thought crossed my mind too, but I don’t know how it works.
One possible explanation is an instrumentation fault. The crew were meant to unlock at around Mach 1.4 so, if the airspeed indicator was reading significantly fast, the unlock could have been carried out dangerously early.
Only speculation at this point. NTSB would have to verify the instrumentation readings through the Flight Data Recorder (if extant) and interviewing the surviving crew member.
Listen to the early presser from Hart. The vehicle was very heavily instrumented with a slew of cameras inside and out.
From the way the craft broke up, it would seem to be highly likely that the vast majority of the instruments survived. The lack of a destructive explosion would increase the likelihood that the instrumentation would make it to the ground in a condition that was conducive to recovering the data.
Don’t everyone jump on “pilot error” before looking at high vibration.
If this were done by error and not by fault, then were there warnings? Overrides?
If something is a real no-no should there not be a buzzer, flasher, override release requirement, whatever, when it is attempted?
Is it in a place where it is easily bumped or confused with another mechanism?
Personally I’ve never been a fan of this feather system SpaceShip 2 uses. It is just another moving part that can go wrong, either by computer error, human error, or mechanical problem.
I would probably feel safer in a vehicle that doesn’t require the feathers and the significant re-configuration of the vehicle mid-flight.
The reason for the variable geometry is that it permits stable atmospheric entry even under manual control, while shifting to an (also stable) high lift configuration for a low speed landing. Any fixed geometry spaceplane with reasonable lift (i.e. the Shuttle) is dynamically unstable during entry and would not be controllable if the guidance navigation and control system or the primary flight control surfaces experienced a failure.
The cockpit video shows the co-pilot moving the feather lock lever at around Mach 1.0 and that the spacecraft started to feather 2 seconds after that. The vehicle then broke up.
Supposing that the feathering of the tail caused the breakup we still don’t know the following:
o Why did the co-pilot unlock the feather at Mach 1.0 instead of the 1.4 required in the flight profile? I can think of all sorts of reasons including:
– The co-pilot made a mistake.
– The air-speed display was incorrect.
– The vibration in the cockpit distorted the co-pilot’s vision so he thought that the airspeed was 1.4
– It only took 9 seconds for the spacecraft to go from drop speed to Mach 1.0. The co-pilot was trying to anticipate when the craft was going to reach Mach 1.4 and jumped the gun by only a few seconds.
o Why did the tail start to feather without actuating the feather lever?
– The aerodynamic forces at Mach 1.0 are sufficient to force the tail to feather. (I guess at Mach 1.4 the aerodynamic forces are different such that feathering won’t happen)
– The feather lever in the spacecraft might only be a wired control signal and the on-board computer decided to actuate the feather for some reason.
– The tail position sensor feeding the feather drive could have been broken or wired wrong. (I’ve done this. It’s quite exciting to see a robot arm suddenly swing in the opposite direction at full speed.)
I bring up these possibilities primarily to illustrate that the list of things that contributed to the accident can be very large and/or subtle. With only a few sentences of information from the NTSB, it is impossible to figure out what really happened or to lay blame anywhere.
User interface or human factors probably contributed to the problem. The function of the “lock” lever could be ambiguous. Does it act as an “interlock”, which enables or disables the action of the feather lever? Or is it controlling a latch, which implements a physical lockdown of the tail booms?
If you think of the function of the lock lever as more like an interlock or arming switch, then you wouldn’t think about uncommanded deployment. It might have been better to call it “latch/unlatch” rather than “lock/unlock”.
Of course, the best thing would have been to have the designer tell the pilots that unlocking (or unlatching) the feather function could be like unlocking (or unlatching) a convertible’s folding top at freeway speed. You wouldn’t forget the mental image of the possible consequences.
–
The new national sport of space geeks, unbridled speculation based upon the flimsiest of information.
We saw how that worked out for the engine hysteria didn’t we…
Seems to be design error. The wings did not support the weight of making a move. The error is repeated if it makes another ship. The maneuvering system is very fragile. Do not bore the opposing forces of buoyancy and air resistance. Every project is doomed, except if avoiding maneuver uphill, which completely undermines the business. Or, make a patch for reinforcement …
If the NTSB is on the right track about the feathers early deployment breaking SS2 apart then the fire is probably just a side effect.
Residual burn in the remains of the engine, probably.