My thoughts, over at PJMedia.
I should note that I since I wrote it yesterday, I’m starting to think that perhaps a chunk of nylon at cold temperatures aloft broke off and blocked the nozzle, because I’m hearing that the oxidizer tank itself was intact, meaning that it was a combustion-chamber explosion (which would be consistent with the pictures). So perhaps it was a problem with the new fuel. Either way, we won’t know until the NTSB completes its investigation, but either way, I think they have to (finally) take a new approach.
[Evening update]
This article at The Telegraph is pretty devastating.
I think that the biggest issue at this point is how to stave off demands that the FAA start regulating, and to somehow still extend the learning period.
There was one early witness account that said the engine seemed to light, go out for a brief moment, and relight. As I understand it, compressed nitrous oxide drops to -127 F when it vaporizes. If you started a burn and it went out, the flow of cold oxidizer would chill the surface of the nylon, making it very brittle and inducing thermal stresses while the nylon underneath it is still soft and warm. Could that crack off the surface layer, perhaps like shattering a thin pain of glass? And if this happened right as combustion recommences, you might have hunks or chards of extremely cold nylon (of varying sizes) with an immense surface area in a combustion chamber that was already filled with pressurized oxidizer.
You wouldn’t see this effect in a normal shutdown and restart because you cut the oxidizer flow, so the burning surface never gets chilled. You could probably check for the potential failure mode in a firing test by rapidly switching from oxidizer to a spray of extremely cold, inert gas.
A nitrous oxide “hard start” seems possible to me. A chunk of nylon breaking off not so much unless the engineers did something stupid with the design.
Although it’s probably not a good time to second-guess the investigation, I’m pretty certain the root cause of the accident was engine related and tend to agree with your scenario.
I too first suspected N2O ‘exothermic decomposition’ but images of the wreckage seem to show the N2O tank reasonably intact. However, as reports suggest (tbc) they were around 12 seconds into the burn when disaster struck, my best guess would be an over-pressure in the ‘combustion chamber’ resulting from a blocked nozzle due to a loose chunk of fuel that was fractured during a ‘hard start’ (i.e. even though ignition seems to have worked as expected, it may well have caused damage that took some time to bite… with catastrophic consequences).
I’ve never understood what the harm is in “second guessing” the investigation. The investigation results will be what they are, regardless.
I can only agree, Rand. However, speculating on what caused the accident can lead to all sorts of ideas that may undermine knowledgeable people’s credibility in the eyes of the public when the full facts are known.
One example is that, having now seen the NTSB video showing what appears to be the motor casing intact (it’s part buried but shows no evidence of a burst), I’m not sure the chamber ‘over pressure’ scenario holds true. Another example is the letter to the Daily Telegraph warning of the dangers of N2O…
http://www.telegraph.co.uk/news/worldnews/northamerica/usa/11203634/Branson-spaceship-explosion-The-missed-warnings.html
…that may well be true but, if the images of the apparently intact N2O tank are considered, appears not to be relevant to this accident.
What worries me is that there seems to be numerous flaws in the SpaceShipTwo design but that this accident will only uncover one, with the others being labelled as scaremongering… until the next ‘incident’.
I think that it’s all right to speculate, as long as one is clear that it is only speculation.
The harm is that it is usually the first report that sticks in the memory, not the best report. The formal investigation has all the data and the best tools to deal with it, is likely to be the best report. If the first report gets it wrong, the subsequent response will be not to what actually happened, but to some weighted average of what actually happened and what the early speculation thought had happened. Worst case, the official, accurate report is suspected of being a cover-up, because we already “know” what happened and here’s an official bureaucratic committee trying to say otherwise.
This is most likely to happen when you have someone with at least a semi-official connection to the incident speaking where non-technical decisionmakers are listening, which is why you won’t find anyone from Virgin or Scaled (or Orbital or Aerojet) speculating in a press conference. Doing that sort of thing will rightly get you fired. Harm is minimal with an essentially closed, technically minded group far from the action, so I don’t mind participating here or on arocket or the like, but it is something to keep in mind.
Honestly, I don’t see how Virgin Galactic survives this. After all this time they essentially have nothing that works except the carrier aircraft.
That carrier aircraft is plenty useful. It wasn’t just designed for SS2. There were also plans to drop other rockets, including a version of Falcon. Those plans will be getting another look. Orbital has lots of experience with Pegasus, and suddenly their first stage doesn’t look so good…
At the beginning of the Aeronautical revolution, there was a long period where aircraft designers struggled between the idea of building the aircraft to fit the available engines, or build engines to fit the aircraft. It took a few years for the two businesses to sync up. It seems to me that we have the same problem today with commercial space. We have spacecraft being designed with the hope that an engine platform will magically arrive to save their spacecraft ideas at a later date.
Aviation history has many examples of failed aircraft that were designed around immature engine designs. This may well become another one.
Three retired Space Shuttle OMS engines could replace the hybrid and only require about half the propellant mass, as long as everybody is okay with using toxic hypergolics in a passenger vehicle. There are so many other options out there.
That Daily Telegraph article made me want to punch someone, preferably the buzzards who aren’t even waiting until the wreckage has cooled to get a little attention from the press. (No, I don’t mean you, Rand.)
It might be worth remembering that Virgin Galactic is not even remotely in the space flight business.
It’s building airplanes capable of climbing to the Kármán altitude of 100 km to provide an interesting several-minute ride and a certificate to people with the money to afford it. Nothing wrong with that, I’d like to do it myself and might even put it on the bucket list if they succeed.
But it has zip to do with actual space flight.
But it has zip to do with actual space flight.
Another Internet “expert” trying to define spaceflight out of existance.
Egad, indeed.
He has a point. Getting to orbital altitude (and 100km isn’t orbital altitude, by the way – not for long anyway) is a far easier proposition than getting into orbit. I haven’t done the maths, but it’s probably at least an order of magnitude easier.
On the other hand, getting to orbital altitude (probably around 200km would be reasonable) is a good start on spaceflight – but only if you have an engine that works up there but can’t be used on the ground for some reason, to get you the rest of the way. An example might be an Orion-drive craft; ground launch is NOT a good idea for obvious reasons, but using such a drive in a craft that’s been lofted to height with chemical rockets might be.
Or perhaps a DPF (dense plasma focus) fusion drive, if the characteristics of such a drive turn out right. Or a NERVA rocket. Or…
I haven’t done the maths,
Then you have no basis for an opinion.
but it’s probably at least an order of magnitude easier.
Someone forgot to tell NASA that. Mercury-Atlas wasn’t orders of magnitude larger than Mercury-Redstone.
Even if you were correct, so what? Orbital spaceflight is orders of magnitude easier than a trip to Alpha Centauri — which is much easier than a trip to the Galactic Core — which is easier than the Andromeda Galaxy. You can always imagine a harder mission.
Crossing the Atlantic is orders of magnitude harder than the first Wright Brothers flight. So, I guess that wasn’t “true” aviation, by your definition?
And we should stop flying Cessnas, because they don’t have the range or capacity of a Boeing 747?
Under John F. Kennedy, the US abandoned development of cost-effective suborbital flight for the unaffordable Apollo program, just “because it was hard.” Because no other mission would be more impressive to mankind or more expensive to accomplish. The result is that the United States has hardly any spaceflight capability at all.
And why should we settle for building your Dense Plasma Fusion drive, when Warp Drive would be so much better?
If we followed your logic, we would never do anything. There’s always something better, and harder and more expensive and further in the future.
That way lies madness.
“And we should stop flying Cessnas, because they don’t have the range or capacity of a Boeing 747?”
No. We should stop flying Cessnas because Van’s RV’s are better…especially the rV-4 and RV-8 🙂 ***
*** Just a joke Ed.
Very well. I have the time to do the maths now; didn’t before. The energy required to loft a kilogram of material to a given height is its gravitational PE, governed by the relation E=mgh which simplifies to E=gh if one assumes a 1kg mass – useful for comparison. Assuming g=10 (roughly correct, and the lower value at 100km isn’t significant) this works out to 1E6 J/kg.
Orbital energy at LEO (200km, for comparison) is governed by the relation E=mv^2/2, which simplifies to E=v^2/2 for the comparison 1kg mass. Orbital velocity at that height is around 7.8 km/s. Plug in the figures and one gets a figure for orbital velocity KE of approximately 3E7 J/kg. Of course, using an equatorial launch eastwards takes a little off this; but on the other hand, the gravitational PE of 2E6 J/kg for 200km altitude has to be input as well, which (probably) approximately cancels this out.
So my guesstimate of an order of magnitude difference was correct. In fact, orbital insertion at 200km LEO is roughly 30 times as difficult in energy terms as lofting an object to 100km. The suborbital lob used as the first Mercury flight is somewhere between the two cases, because the capsule had to be given a transverse velocity. I simply don’t know what this was, except that it obviously wasn’t orbital velocity.
Having done the maths, I now recall having done this calculation a few months ago; must have unconsciously remembered the result. I suppose a professional rocket scientist (which I’m not!) would know this as a matter of course.
That won’t work in the real world. If you launch straight up like that, you won’t survive reentry when you hit the atmosphere straight on.
You also overlook drag losses, gravity losses, reduced engine performance at low altitudes…
If what SpaceShipTwo is designed to do is “space flight” then my floating in the surf 20 feet from the beach on a inner tube is “ocean travel”.
Another armchair astronaut, who’s never been above the stratosphere, thinks it isn’t “real” spaceflight unless it looks like Star Trek. 🙂
Get back to us when your logbook shows some hours in that “inner tube.”
In the meantime, you would do well to remember the words of Anton Ego:
In many ways, the work of a critic is easy. We risk very little, yet enjoy a position over those who offer up their work and their selves to our judgment. We thrive on negative criticism, which is fun to write and to read. But the bitter truth we critics must face, is that in the grand scheme of things, the average piece of junk is probably more meaningful than our criticism designating it so. But there are times when a critic truly risks something, and that is in the discovery and defense of the new.
And if I wanted to answer you snark for snark, I could say that your writing is worthless because responding to a blog post is orders of magnitude easier than writing a book.
“But it has zip to do with actual space flight.”
Has just as much to do with space flight as Goddard’s test flights in Woostah.
I can imagine a chunk of nylon breaking loose and blocking the throat, but would that cause a catastrophic failure? It could with a solid, but with a hybrid it’s still limited by the supply of oxidiser. Even if the combustion chamber ruptured and air got in, you would still expect it to burn instead of exploding. Then again, something must have caused it to explode, and if the nitrous tank was still intact, that cannot have been the cause. Also, the fuel was new, which makes it an obvious suspect.
Clogging the nozzle doesn’t shut off the flow of oxidizer, or eliminate the rest of the fuel. It burns and pressure builds up until the nozzle becomes unclogged, one way or another…
Thank you, Rand, for all the discussion and updates. The Telegraph article was interesting. Sir Richard makes the obligatory obeisance to safety being the “number one priority”, but then it transpires that his engineers had been warning of the dangers of the propulsion system for years. While some will be saying that this proves that spaceflight is especially difficult and only space agencies should be allowed to do it, the problem is clearly one of project management, not of underlying technological difficulty.
Stephen
Oxford, UK
Yes, as with Challenger and Columbia, this was primarily a management problem, and says nothing about the intrinsic difficulty of spaceflight.
Or until the pressure drop between the combustion chamber and nitrous tank has been eliminated, though that doesn’t appear to have happened here. But even if the oxidiser continues to flow, you still can’t get all of the fuel to react at once the way you can with a pure solid. Oxidiser flow is still a limiting factor.
Another new thing they had on this flight is a methane tank used during ignition.
Yes, that “simple” hybrid engine seemed to be getting more and more complicated, just like sometimes “fins” turn into wings on a lifting body.
Come to think of it, a mere rupture of the combustion chamber might be catastrophic, even if the propellant doesn’t detonate. I read some speculation SS2 had flipped 180 degrees. Could an asymmetric rupture be enough to flip the plane by 180 degrees?
Mother nature can be very ingenious at times, making something that would at first seem impossible only all too evident… ever seen a complete combustion chamber (i.e. one that hasn’t burst) fly away at right angles to its thrust-line?
“Flipped 180 degrees” sounds like an exaggerated speculation. Asymmetric thrust can do nasty things at high speed, though. Challenger’s structural breakup was due to asymmetric aerodynamic forces, which in turn were caused by asymmetric thrust. Flying sideways is hard.
I’ll assume that neither the spacecraft’s control systems, flight control computers, nor pilots were prepared to deal with asymmetric thrust. Flying outside the envelope is hard, and often fatal.
Flight control computers? Scaled don’t need no stinking flight control computers…seriously, they have a phobia about it, plus they have nobody with the requisite skills.
This is the daftest (most daft?) thing I’ve read all day.
What’s daft about it? It is a known fact that Scaled will only build aircraft with 100% mechanical flight controls. I admit my use of the word “phobia” was hyperbolic, but I was making a joke at the previous commenter’s mention of flight control computers, apparently not knowing that Scaled refuses to build nothing other than 100% mechanical flight controls. And it is completely accurate that nobody at Scaled has the requisite skills and experience to either (a) specify, design, integrate, test, etc. a flight control computer, or (b) analyze, design, implement, and test control laws to reside in such a computer. I am somebody who has such skills and experience; I am familiar with Scaled’s capabilities; I feel quite confident in my statement. Note that nowhere did I say or imply this topic is particularly relevant to the SS2 accident.
What’s daft about it? Either your statement is daft, or you’re right and Scaled is daft. Either way, daft. Take your pick.
One final comment before abandoning this thread: There’s nothing wrong with mechanical flight controls; lots of my favorite airplanes use mechanical flight controls :-/. I fully subscribe to Saint-Exupery’s philosophy “perfection is finally attained not when there is no longer anything to add, but when there is no longer anything to take away” (see http://en.wikiquote.org/wiki/Antoine_de_Saint_Exupéry); it was hammered into me by my design professors. And I haven’t tried to analyze SS2 (or SS1) to see for myself what kind of tradeoffs Scaled made by going mechanical-only. What I object to is Scaled’s dogmatic position that mechanical flight controls are inherently simpler and safer than fly-by-wire. That position takes away a major part of the design trade space in one fell swoop, for no sound reason – only dogma.
It’s true that Scaled doesn’t have the personnel or institutional expertise to do FBW, or possibly to realistically include it in their trade space. But they are a wholly-owned subsidiary of Northrop Grumman, and NG most definitely knows FBW on all kinds of configurations (at least three Collier trophy winners, for example). So Scaled has access to those resources, if they so choose to ask for them.
Those photos appear to show the tail surfaces broken free from the fuselage. When that happens, the plane will swap ends. It’s like trying to push a rope. Did the explosion sever the tail surfaces or did structural failure happen first? If the reports of no rupture in the engine casing or oxidizer tank are true, then perhaps it wasn’t the engine at all.
The severed tail booms appear in one photo to be alongside, not behind, the main fuselage at a time when the main fuselage appears to be already tail-forward. I’d want decent video to be sure, but that suggests the vehicle flipped first, and aerodynamic forces then tore off the booms. Which leaves us with the question of what could flip the vehicle with the booms intact.
A partial explanation is that venting the N2O supply (the white cloud in the photos) with the fuel unburnt would shift the CG well to the aft, resulting in the least-stable configuration possible. But, least-stable rather than actually unstable, for any sensible design, as losing your oxidizer early is an obvious failure mode that you’d have to be able to deal with.
From an Aviation Week article:
Although not commenting on whether the uncommanded deployment led to the over-stressing of the structure and the subsequent in-flight breakup, Hart says the data is a “statement of fact rather than a statement of cause.” The feathering system was first used on a powered flight during the second rocket-propelled sortie on Sept. 5, 2013, when SS2 reached Mach 1.43 and an apogee of 69,000 ft.
The NTSB also reports that the fuel and oxidizer tanks as well as the hybrid rocket motor were all intact and showed no signs of burn through or of “being breached.” The findings support photographic evidence of the mishap which clearly indicated a successful ignition and continuing rocket burn before the catastrophic structural breakup.
There is apparently no indication of a breech in the nitrus oxide tank or the engine casing. It’s possible – despite all that was said and written since the accident – that the new engine had nothing to do with causing the accident. It’s too soon to tell. It could have been a malfunction in the feathering mechanism that caused an uncommanded feather at too high a speed in too dense air. Feather deployment under full power would cause a violent pitch up that, combined with the thrust vector no longer being aligned with the vehicle trajectory, could cause very violet G forces.
The WSJ is reporting that the tail feather pivoting was deployed prematurely. I’m not sure whether they are saying that the rocket was still burning at the time, or only describing the design.
http://online.wsj.com/articles/ntsb-cites-improper-pilot-command-in-virgin-galactic-crash-1414993698
Well, that’s not good. Whether commanded or uncommanded, it means the engine we don’t like hasn’t failed, so we might go through all this again with a subsequent vehicle. But at least it explains the flip.
As long as we’re speculating, what would cause an un-commanded tail feather deployment? Maybe excessive engine vibration broke the locking mechanism?
That’s a possibility. We’ll have to see the g-profile. But I don’t understand why you’d unlock prior to MECO.
I’m wondering if the process is:
Drop and zoom climb, accelerating to Mach 1.4 as the atmosphere becomes thin enough to safely deploy the feather. If the feather doesn’t deploy for some mechanical reason, then cut the engine and you’ll get only a Mach 1.4 re-entry that can be flown without the feather system. If the feather does deploy then the burn can continue as normal.
That would mean there’s a go-no go decision at around Mach 1.4 as they zoom out of the atmosphere, and that makes me wonder whether any automatic systems were added that would deploy the feather shortly after unlocking in case the pilots remembered to unlock it but forgot to deploy it, as a “safety” feature, necessary because if they continue the climb and only belatedly try to deploy the feather, the ship is doomed. If the engineers did something like that, it would change the “unlock” lever to an “unlock and deploy” lever, and if that didn’t get communicated to the pilots, they’d do just what they did to keep ahead of things on the checklist. Of course, the other possibility is some kind of fault that that deployed the feather once it was unlocked.
I’d say the preliminary report from the NTSB would indicate that the whole foofraw over the hybrid engine is just vapor.
Perhaps some mea culpas are called for?
It was a reasonable assumption with the information we had at the time. I for one was certainly always clear that it was speculation.
For many years, I’ve noticed that whenever a sensational news event happens (plane crash, mass shooting, etc.), about 90% of what is said in the first few days turns out to be wrong. With the 24 hour news cycle, networks have a lot of airtime to fill. When it’s a technical matter, they’ll latch onto any piece of speculation without regard to the source and endlessly speculate without the first shred of actual information. Years from now, people will swear it was the engine that brought down SpaceShipTwo (and it’s still too soon to completely rule out anything as a contributing factor) just like some people swear TWA 800 was shot down by a missile. In a week or so, the Press will move on to other matters so that by the time the NTSB releases any solid information, few people will hear about it.