NASA, CNES Warn SpaceX of Challenges in Flying Reusable Falcon 9 Rocket
aviationweek.com
aviationweek.com
Additionally, SpaceX isn't dumb, they've already tested running their engines through multiple cycles on test stands, they know the issues they'll run into and they have a reasonable idea how long they can expect to use them (spoilers: lots). And SpaceX has been operating reusable rockets over the course of several flights in the form of the the Grasshopper and F-9R-Dev vehicles which they've used for test flights.
They're saying the NASA people had a working model for decades, and hence people should trust NASA's expertise more, not the new guy taking big risks trying new things. In reality, they can see that SpaceX knows what they're doing, and NASA has been falling behind for a long time, but that's really hard for NASA to accept. Still makes me sad to see this kind of attitude. Risk taking is necessary for innovation, and calculated risks often pay out big time, but the establishment is still always too afraid to take the said risks... And then when somebody else is there to take those risks, they try to discourage them.
The damning bit is that what SpaceX ends up doing will be the sort of thing that could have been done back in the '80s or maybe even the '70s. But nobody had the right combination of pragmatism and vision to make it happen. But they could have, and there were bread crumbs all along the way.
So why didn't they? A lot of it ultimately comes down to a few factors. Group think and bureaucracy. And fads. Yup, fads. One problem that orbital aerospace has been constantly afflicted with is the coolness factor problem. When pushing new projects it can be difficult to do so unless there is some sort of coolness factor, some way in which the new project is demonstrably more advanced than previous generation efforts. This is anathema to actual progress because most progress comes from copying, perfecting, and incremental innovation. The coolness factor is how we ended up with Shuttle (so cool), and X-33/VentureStar (much, much cooler). But look at the Falcon 9, it's just a 2 stage LOX/Kerosene booster, they've been building those since the '50s/'60s, it's the definition of uncool. But what it lacks in coolness it more than makes up for in usefulness.
In terms of fads you can look to the Hydrogen rocket fad that still has a grip on a lot of engineers. LOX/LH2 undoubtedly has a very high Isp, and Isp is supremely important in the rocket equation. But it's not everything, and LH2 has a crapload of very serious downsides when it comes to using it in rocketry, many of which add significant cost to the system. The Hydrogen fad is why the Shuttle had SRBs (because the SSMEs couldn't generate enough thrust on their own to get off the pad) and thus is partially responsible for the Challenger disaster. It's also responsible for the Delta IV, which is an extremely expensive launcher.
But back to breadcrumbs. They were there. In the early '90s the DC-X program demonstrated vertical takeoff/vertical landing was a feasible technology, and it did so using LOX/LH2 even. But the program was largely ignored because NASA was too dumb to see the promise of a reusable stage, since it wasn't cool enough relative to the idea of an SSTO RLV (in the same year as the last flight of the DC-X NASA got the X-33 program rolling).
If NASA was just a little bit smarter and more practical it would have been easy to see the promise of the DC-X, easy to see the practicality of partial reusability, and easy to see the advantages of simpler launchers. But they didn't. Instead they pursued cool idea after cool idea, dumping around a billion dollars into X-33 alone. And when their overly ambitious cool ideas didn't pan out they ended up with nothing. And it was only through the comparative pragmatism of the defense department and the EELV program (which developed a couple boring 2-stage launchers) that the US had any reasonable launch capability whatsoever after the Shuttle program ended. Now it will be through the comparatively superior pragmatism (not to diminish the genius of SpaceX) that we'll end up with even greater launch capability and unprecedented reusability.
Errr, not to my memory.
Back when it was being thought about, Single Stage To Orbit (SSTO, https://en.wikipedia.org/wiki/Single-stage-to-orbit) was beyond the foreseeable state of the art, and the initial plans were to have to complete, reusable stages, a 1st stage big monster that would get the 2nd stage high and fast enough, and then would land "normally".
But NASA just didn't get enough money to do it right (the '70s were a particularly bad time to be trying to do such a thing in the US, and in things like the SSMEs just plain screwed up (https://news.ycombinator.com/item?id=7711061). So at practically every turn, they sacrificed true reusability, operating costs, and ultimately safety (not a coherent design, but a series of compromises, and of course no provision for failure) to keep development costs within their budget.
The point is, NASA flew the Shuttle and knew that it was a stretch to call it reusable, but they had hopes.
SpaceX has a lot more data on their side and they know that there's nothing fundamentally not reusable about their stages. They've cycled the engines through enough simulated launches to know how long they'll last, and it's more than enough flights (dozens) to translate to massively reduced launch costs if they pull it off.
That's just one reason why NASA's experience with the Shuttle is utterly irrelevant to the Falcon 9. Frankly, there's just not enough to do to a Falcon 9 to make it costly enough to reuse to end up with the problems of the Shuttle. It would have to be completely non-reusable for that to be an issue, and that's extremely unlikely given the design and the data we have so far (for example, the reuse of the sub-orbital testing vehicles in Texas).
Sour grapes, much?
This is why it's important to have different groups of engineers trying different things--you get a chance to pull off things previously thought impossible or prohibitive.
There is also a question of will though. IANARS (I am not a rocket scientist), but it seems to me a lot of political investment behind rockets in USA and USSR was military in origin. This meant the rockets and rocket engines they spent most effort on was for single delivery of a warhead with no chance of reuse. Civilian rocketry was just a byproduct of efforts to deliver the bomb. We see this with Russian commercial satellite launches using repurposed ICBMs. If military didn't care about the reuse, the civilian programs would have had to work on it on their own.
I believe that with enough will and advances in technology, engineering problems for reusable rocket engines could be overcome.
SpaceX is intending to break that cycle. If launch costs are low, then you can afford to put cheap payloads into orbit. If the payload fails, just launch another one. If the best chance of a successful launch matters a lot to you, pay a premium to launch on a virgin booster.
However re-usability of the first stage is only going to get them so far. If half the total cost of a launch (pure guess) is in the booster hardware, and they get just 10 launches out of each one, then the total cost per launch will be cut to close to half of what it is now. That's great, amazing even, but it's not 10x cheaper and certainly not 100x. To get down even close to 10x cheaper they need to recover the second stage as well.
I could go into the topic in depth but suffice it to say that politics and bureaucracy have had a stranglehold on NASA manned spaceflight and launch vehicle development since the beginning, and that has had a very negative impact on progress and capabilities. In contrast, unmanned spaceflight at NASA has been, comparatively, much better managed. And there you see a much clearer and sharper continual increase in capabilities, ambition, missions, and so on over time. The missions that would have been acceptable in the 1970s (planetary flybys, for example) would today not return enough science to be justified. Today we have multi-year rover missions on Mars, and we have spacecraft which rendezvous and orbit multiple asteroids using ion engines. Meanwhile, in manned spaceflight we haven't been out of Earth orbit since the early '70s.
Meanwhile, the Falcon 9 stage is primarily just engines, tanks, and frame, with a sprinkling of avionics and RCS plus landing legs for the reusable version. The idea that a system so monstrously simple in comparison to the Shuttle could possibly have the same levels of difficulty of reusability is either extraordinarily naive or propaganda.
To compare, let's see what SpaceX is actually doing: The last, second attempt to recover a booster of Falcon 9 was:
http://aviationweek.com/space/spacex-plans-multiple-reusable...
"While SpaceX remains coy about the exact fate of the modified Falcon 9 first stage, the overall results appear to be encouraging despite the apparent breakup of the booster after it landed in heavy seas."
Previous attempt:
"The latest flight test follows an unsuccessful recovery attempt on Sept. 29, 2013, when the first upgraded Falcon 9 v1.1 was launched from Vandenberg AFB, Calif. Although three of its nine Merlin 1D engines fired to slow the initial descent ahead of a single engine firing for final braking, the vehicle’s spinning motion could not be controlled. As a result of aerodynamic torque, the fuel centrifuged inside the tank, flaming out the engine and sending the stage crashing into the sea."
Oh, also, I should mention how difficult it is to optimize an engine for vacuum as well as sea level operation. In a vacuum you want a high expansion ratio nozzle because that maximizes Isp, but it also lowers pressure. That doesn't work if you have to fight against 1 atm of sea level pressure. That alone is actually a big reason why rockets have multiple stages, because the Isp boost of the high-expansion ratio nozzle on the upper stage is an enormous performance win. One way around the problem is to boost the exhaust pressure, enabling a higher expansion ratio nozzle which can still operate at sea level, but that makes the engine even more complex and difficult to manufacture, let alone reuse.
So you have the SSME which is a high pressure staged combustion regeneratively cooled LOX/LH2 engine optimized for both sea level and vacuum operation. It's a wonder they were reusable to any degree whatsoever within the first several flights. Even so it took until the mid 1990s for the engine design to fully mature and enter a truly operational phase (vs. prototype phase).
Meanwhile, let's look at what SpaceX has on their plate. They have a low pressure (1/3 the psi of the SSME) gas-generator cycle (non-staged combustion) LOX/Kerosene pintle injector engine which uses two different nozzles for sea level and vacuum operation. It's a design with heritage back to an era before color television. It's such a tried-and-true design with so much history that the patents on it expired decades ago. Additionally, in contrast to all the rigmarole an SSME has to go through for each flight the Merlin-1D merely accelerates to a fraction of orbital speed, never leaves the atmosphere, and, in a reuse scenario, is back at the launch site within minutes. No hypersonic atmospheric flight. No re-entry. No bumpy landing. No cold soak. No transition from atmospheric to vacuum operation.
It's a vastly simpler engine which is exposed to vastly less strenuous conditions.
Again, if you think that the Merlin-1D will face anything approaching the problems of the SSME in reusability then you are either extraordinarily ill-informed or you are peddling propaganda.
To be honest it would take something coming out of left field to throw off their reusability plans at this point. At the end of the day the engines just haven't been exposed to very significant adverse conditions during a normal flight, there's no reason to assume the durability would be so much less than on the test stand. More so, the stage is mostly just fuel tanks, engines, and landing gear. There just aren't that many components, and there isn't much opportunity for something to go wrong. They fill and drain the fuel tanks repeatedly, there's no reason to assume they'll suddenly stop working after actual flights. And the overall structure of the vehicle is more than strong enough to withstand several launches and returns.
In terms of results SpaceX achieving reusability will be a big deal. But in terms of the prospects for reusability, the vast majority of the unknowns have now been filled in, at this point it would be more surprising if it didn't work than if it did.
That downplays the viciousness of that environment, "The atmosphere in low Earth orbit is comprised of about 96% atomic oxygen." (http://www.nasa.gov/topics/technology/features/atomic_oxygen...), it's truly corrosive stuff.
I wouldn't be surprised to learn at least one component of the SSME was, or was initially, susceptible to it. A look at more hits from https://www.google.com/search?q=atomic+oxygen+space+shuttle could be instructive.
The Shuttle was required to be reusable as a political goal. However, it was not given the funding it needed to actually accomplish it in any reasonable fashion.
The Air Force put some tough requirements on it, such as the requirement for enough cross-range capability when landing to be able to launch into a polar orbit from Vandenberg, deploy a satellite, and land back one orbit later. This is purely a wartime requirement. No peacetime mission needs this. This required much bigger wings (and thus much more weight), even though the Shuttle never once launched out of Vandenberg, let alone on one of these crazy single-orbit missions.
Politics screwed up other stuff too. The solid rocket boosters, for example, were manufactured on the other side of the country, largely because of Congressional pork. The long and difficult transport process required the boosters to be built segmented, and that design directly lead to the loss oF Challenger.
Similarly, budget constraints dictated the cheaper SRB-plus-external-tank design rather than a more sophisticated and more reusable (and more expensive) reusable booster. This design directly lead to the loss of Columbia.
The net result was that the Shuttle had a very narrow engineering margin. The SSMEs are engineering marvels in terms of how light they are and how much power they put out. But putting out a lot of power for little weight is a really bad plan for something you want to reuse. Reusability means robustness, and robustness means extra weight and less power.
The Shuttle is what you get when you say, "How much will it cost for a reusable launch system? $10 billion? Too much, we'll give you $5 billion." (Numbers in the ballpark but mostly made up for illustration.)
On the other hand, the Falcon 9R is what you get when a for-profit space launch company wants to make more money. It's not a prestige project, it's not a political project, it's not a military project dictated by thoughts of all-out war with the USSR, it's not subject to Presidential grandstanding or Congressional pork. It's just a bunch of smart people thinking of ways to save their customers money and turn a better profit while doing it.
The Shuttle was an extremely ambitious design that did everything differently. Ultimately, everything was too ambitious to be reusable.
Falcon 9R, on the other hand, is deeply unambitious. The rocket is a pretty standard design. They don't even try to use liquid hydrogen for any stage, even though it's theoretically more efficient, instead favoring easier-to-handle kerosene. Their rocket engines are high performance but not on the bleeding edge. And when they decided to go for reusability, instead of building a whole new system with wings and ceramic tiles and all sorts of other extreme innovations, they just said, let's slap some legs on the first stage and give it some extra fuel. It's a completely different approach that pretty much only shares the word "reusable".
I don't think the Shuttle really has many lessons to offer when evaluating the potential of Falcon 9R.
"The Air Force put some tough requirements on it, such as the requirement for enough cross-range capability when landing to be able to launch into a polar orbit from Vandenberg, deploy a satellite, and land back one orbit later."
That's not what I remember at the time this was happening, nor was it e.g. the story in a good study published in Science. Sure, the DoD/intelligence community wouldn't have minded such a capability, but NASA insisted on it to make the shuttle more attractive to them and to help justify the monopoly NASA obtained on these sorts of launches.
Something the DoD didn't like at all, and was partly freed from after the Challenger disaster. Which also prompted a closer look at the safety of Vandenberg launches and the canceling of them.
http://history.nasa.gov/SP-4221/ch5.htm
Quote: "However, it meant that if a shuttle was to execute a one-orbit mission from Vandenberg, it would return to the latitude of that base after 90 minutes in space only to find that, due to the earth's rotation, this base had moved to the east by 1100 nautical miles. Air Force officials indeed expected to launch the Shuttle from Vandenberg, and they insisted that the Shuttle had to have enough crossrange to cover that distance and return successfully."
It goes into quite a bit of detail on the Air Force's involvement and reasoning.
I'd be interested to see them if you had sources that say otherwise.
My sources are what I contemporaneously learned and remembered in the '70s, and that paper in Science (the world's #2 science journal after Nature, not that it wasn't back then very political, just that this claim cuts against their politics). It would have been published in the '80s, very possibly after the Challenger disaster prompted a lot of second looks, but I can't be more specific.
On the other hand, you've given me no reason whatsoever to believe you over them. Your ??? is completely out of line. I merely gave a source for what I said and asked if you had one. I understand if you don't, but don't be surprised if I then don't necessarily believe your claim.
My "???" indicates I think it's quite possible the author's sources lied to him. You're welcome to believe that NASA people, when talking about their greatest and most expensive in every way failure---these decisions killed two complete crews of 7 each---are more honest than I am, but I'm not entirely sure how much company you'll have, the world has seen rather a lot of post-Apollo NASA by now....
(This alone should give one serious pause: https://en.wikipedia.org/wiki/Rogers_Commission_Report#Role_...).
I don't understand your incredulity at my reaction. What exactly did you think was going to happen when you showed up making an effectively anonymous comment saying, "no, that's wrong, believe my 30-year-old memories over the official history"?
"The Space Shuttle was a political football whose design was dictated by the whims of Congress and the meddling of the military first...."
Then you note the military didn't want it, but wanted it to be useful if forced. I would add the Science article specifically addressed the single orbit polar mission, saying the Air Force didn't see a great need for that and would have been happy with gentler flight profiles. It explicitly said NASA insisted on it, and that's consistent with their post-Apollo institutional biases, much the same produced the hot-rod SSMEs.
"Meddling", when it wasn't, qualifies as bashing in my book, especially in the full context of your sentence.
As for my incredulity, maybe start here: https://en.wikipedia.org/wiki/Historiography
I was a History minor in college....
However, if you insist on taking offense at my comment, I'm not going to fight it that hard. I'm not really interested in what offends you.
What I am interested in is some sort of external source for the stuff you're saying. Pseudonymous internet comments from "hga" and vague references to "the Science article" don't count.
Partly true, but keep in mind that was a problem NASA brought on themselves. They chose an overly ambitious design which congress, rightfully in my opinion, balked at. They tried to skip technological generations, which is a nono in R&D and always a very expensive and risky prospect. Instead of coming up with a new idea they doubled down on their fantasy, turning the Shuttle idea into a political juggernaut. And they got their way, they got massive buy in into the Shuttle, by selling it to everyone as the be-all-end-all of everyone's space dreams. It would be all things to all customers: NASA manned spaceflight, NASA space science, commercial satellite companies, even the military. But part of the cost of that political juggernaut was imposition of additional design elements, which brought more compromises. But on the other hand with so many people on board the program had political inevitability and was almost impossible to shut down. The Shuttle would already have been a bit of a turkey relative to its promises but those additional compromises made it one of the most expensive and troublesome launchers in history. But those compromises came about because of the unwillingness of the Shuttle's original backers to consider other options, they made those deals and those compromises and they were forced to live with the consequences.
No all contemporary Russian engines have a much higher Isp. Usually because they use high chamber pressures in the 3000+ psi range. The merlin series runs a mere 1000 psi.
It does win one performance metric of thrust to weight. Little surprise because they might give up 10% of thrust (err, whatever) by running at a lower pressure but obviously running a third of the chamber pressure the engine walls are going to be about a third the thickness and a third the mass, so its a huge win for thrust to weight, knock down the nominator by 10% and the denominator by a factor of 3 and you win. So they do hold a record in this category.
Also lower pressures imply lower temps imply lower thermal load, lower heat shock, lower thermal distortion and thermal related stress on startup/shutdown. It inherently should always outperform a Russian engine WRT restart / reusability unless something has gone very wrong.
Two things make engines un-reusable. One is simple runtime. It seems intuitively obvious that a 1000 psi turbopump is cheaper and more reliable than a 3000 psi turbopump, all things being equal. The other thing that kills reusability is thermal shock effects where need something to be "X" thick to handle the forces and pressures, but the exterior or one side heats up much faster than the inside creating enormous stresses. So low pressure means low temps
THANK YOU! 16 years ago I visited Vandenberg and saw SLC-6, and it has always confused me why they spent so much money building a shuttle launch complex they never used. The argument I always heard was, "well a polar orbit wouldn't make sense for the shuttle" ... but as a wartime option as you describe, now I can see why they did it.
I'm not saying I agree with them, but finally it makes sense why they built and maintained the place, without ever using it.
Without wandering into the weeds of the other subthread, NASA managed to get a mandate that the shuttle would be the sole method of putting US satellites into orbit. Forget about the wartime requirements or desires, this meant that anything the military wanted to put into a polar orbit would have to be done by the shuttle, naturally launching from Vandenberg.
Going from memory of what I've read in the last few days, and of what went into the decision, one factor was the shuttle ended up 20% overweight, and the higher energy requirements for polar orbit launches resulted in a plan for lighter filament-wound instead of steel SRBs: https://en.wikipedia.org/wiki/Space_Shuttle_Solid_Rocket_Boo... and of course the Challenger disaster helped people to realize just how dangerous the SRB concept was to begin with.
Post Challenger, during the Shuttle fleet's 3 year grounding, a hard look was taken at what was needed to bring SLC-6 up to spec (see end of https://en.wikipedia.org/wiki/SLC-6#Space_Shuttle), and NASA was no longer in a position to force the Air Force to use the shuttle. Which the latter had always thought unwise and the disaster proved to all who could see.
Can the method used by Scaled Composites, jet lifting a rocket to higher altitude, be employed as a better solution? Why not rocket power the plane portion and go for even higher based launch? Granted some cargo would be lost to weight for control surfaces and wings but recovery should be simpler, a glide path cannot be as hard as landing upright.
That's what the first stage of the Falcon 9 is - a rocket powered boost to enable the second stage to launch from a high altitude.
Terminal velocity in atmosphere is proportional to density and engines are surprisingly dense so you'd need to pack a simply enormous parachute. On the other hand, that big mostly empty tank will slow down the engine for you at very little cost, and its nice to reuse it anyway.
At the terminal area like 500 feet off the ground you could eject the engines and let the tanks explode but thats messy and bad PR and something you haven't added can't fail, so no parachute means no parachute failures, and you'd already navigated under power to within 500 feet of landing, so its simple, cheaper, more reliable, and much safer to just land under power.
Merlin family vacuum ISP a sedate 311 seconds. COTS and evolutionary based on existing engineering practice. Designed for low cost and reliability, not utter extremes of high performance.
(edited to correct, not COTS. Kept thinking they adapted a RD series design at the start, but I checked and they didn't. Same general story though, lower Isp and lower chamber pressure and very conventional pintle injector design etc its good engineering aka very boring, all their engine design competitors are far more competitive, which is dumb because fuel and tanks are cheap, engine development isn't)
Its very much like comparing the long term reliability of a 2024 specs F1 series racing engine none the less manufactured today in 2014 (or at least they did their best), against the engine in my new Toyota that's in it's 10th mfgr year and is pretty much bulletproof.
In another subthread, someone commented on the Space Shuttle being designed for reuse. Well, sort of, especially at the beginning, at the top, but various ambitions that weren't strictly needed (e.g. the SSMEs, single polar orbit missions) and insufficient budget (turning the reusable first stage every one envisioned before into the kludge of the throwaway tank and massively dangerous SRBs) meant by the time it launched it had no hope whatsoever of meeting those goals.
Add bogus safety calculations, and it was never going to be a "space truck". Space X, on the other hand, seems to understand what this is all about. And unlike NASA isn't burdened by implicit requirements like employing lots of people at the maximum number of sites, or the explicit mission of boosting the South's technical prowess.
That said, I think it's pretty obvious that even after SpaceX successfully lands a Stage 1, they are going to do lots of simulated and real-world testing before using one for any customer cargo.
I can see the day in the future, though, when a first stage that has been launched 2 or 3 times is considered safer than a first stage that has been launched 0 times. I anticipate a bathtub curve of failures.
Some buyers who are risk averse may opt to only use first-launch rockets, while others who are cost sensitive will buy rockets on their last legs. This could revolutionize the entire launch market, and enable many more low-cost satellites to be launched.
You could easily swap the quoted line with this (related to Tesla) and it would still make sense:
"We tried to make an electric car before," ... "Look how long and how much money it took for us to do that, and we still weren't successful for all parts."
Here's the story:
This business of taking software written for the 8086 processor and porting it to the 80286 normally required completely rewriting the software by hand, often taking years of painstaking effort. It wasn’t just a matter of recompiling the software, of having a machine do the translation, because Microsoft staunchly maintained that there was no way to recompile 8086 code to run on an 80286. Bill Gates swore that such a recompile was impossible. But Drew Major of Superset didn’t know what Bill Gates knew, and so he figured out a way to recompile 8086 code to run on an 80286. What should have taken months or years of labor was finished in a week, and Novell had won the networking war. Six years and more than $100 million later, Microsoft finally admitted defeat.
http://www.cringely.com/2013/03/24/accidental-empires-chapte...
Accidental Empires is a good book.
I am not a rocket scientist. So someone please explain to me why this quote isn't the single stupidest thing I've read all day.
F22's would be a lot cheaper to make (than today) if everyone had one in their backyard, like a Honda Civic.
If they keep component X too long, they'll need to reimplement / rebuild X from scratch in a few years when the original factory/production line doesn't exist anymore. This would require a lot of up-front cost that may be higher than just producing a new element of that type for every launch.
I don't think the message was completely negative though: "depending on the level of reusability" sounds more like "this is hard, but if you pull it off with cost of restarting production << cost of reusability over N years, that's an achievement"
Edit: many words
If the cost of those launches drops by a significant margin (such as a 90% reduction as suggested), there will be many more launches as every country with two pennies to rub together launches communications and land monitoring satellites.
Now what? What do you do with all your expensive stuff? Maybe you can just build a whole bunch of them at once, and leave them stockpiled? But that's not very efficient is it? What if it turns out there's a major flaw that has to be corrected, or what if it turns out you're not needing as many rockets as you need?
Maybe you can stretch out your production time more - but since 4 months is the longest 'efficient' time you can spend on a rocket, then you're just increasing your costs.
Rocket (and airplane, and tank, and ships, and other expensive hardware things) production lines can't just spin up and spin down. This is why military procurement is partly so screwed up - because it's not really aimed at producing stuff at the best rate, it's aimed for the dual goal of getting Congressmen in office, and maintaining the manufacturing base for the future.
But if you are SpaceX, you've brought all your manufacturing in-house, so you can produce things at cost. If you need 4 more nozzles just like the last 100, you walk to the nozzle guy inside the SpaceX facility and ask him for 4 more nozzles, and he makes them for you just like he made the last 100.
I'm sure it will still cost SpaceX money, but not as much money as it would cost someone else.
My favorite SpaceX story:[1]
> Significantly, the Merlin engines—like roughly 80 percent of the components for Falcon and Dragon, including even the flight computers—are made in-house. That’s something SpaceX didn’t originally set out to do, but was driven to by suppliers’ high prices. Mueller recalls asking a vendor for an estimate on a particular engine valve. “They came back [requesting] like a year and a half in development and hundreds of thousands of dollars. Just way out of whack. And we’re like, ‘No, we need it by this summer, for much, much less money.’ They go, ‘Good luck with that,’ and kind of smirked and left.” Mueller’s people made the valve themselves, and by summer they had qualified it for use with cryogenic propellants.
> “That vendor, they iced us for a couple of months,” Mueller says, “and then they called us back: ‘Hey, we’re willing to do that valve. You guys want to talk about it?’ And we’re like, ‘No, we’re done.’ He goes, ‘What do you mean you’re done?’ ‘We qualified it. We’re done.’ And there was just silence at the end of the line. They were in shock.” That scenario has been repeated to the point where, Mueller says, “we passionately avoid space vendors.”
[1] http://www.airspacemag.com/space/is-spacex-changing-the-rock...
LATE EDIT Apparently at a Q&A SpaceX said their new factory will make 400 Merlin engines a year. That feels way enough to get out of "prototype" territory and into "economy of scale" territory.
If you have a market for 400 engines a year and build 400 engines a year, you will get good at building engines and you will have a lot of opportunity to improve your ability to build all those engines.
If you have a market of 400 engines a year but reuse each engine 40 times, your build rate will converge on building 10 new engines a year. Once your initial production is done and you are in the sustaining level, you will be back in the boat of all the other rocket companies that are only building ~10 engines a year. The inference is that 10 engines a year doesn't give you the volume to sustain (and refine) your low cost build process.
Where I think the argument goes wrong the worst is that Christophe Bonnal is mapping traditional rocket making experience - low rate production of very complex, very expensive hardware - onto SpaceX. SpaceX is rewriting the rules. They appear to have rewritten the rules on the cost of expendable rockets and, with reuse, will rewrite those rules again. That should rewrite the rules on the number of launches since they will be much more affordable.
The (counter) premise is that dropping prices of reusable rockets will allow more launches which will absorb what would otherwise be excess SpaceX rocket building capacity.
If you're doing six launches a year, then building six new rockets a year is not really much more expensive than building one rocket once and refurbishing it for a new launch six times a year. Especially when you complicate the design to an extreme degree to achieve reusability, thus causing that one rocket to cost a ton of money to build and a ton of money to refurbish.
Where he gets it wrong, I think, is that Falcon 9R is not a pure reusable system. It's an expendable system that's been lightly modified for reusability. This is important because it means that the rocket is still cheap to build and cheap to refurbish.
Look at the Shuttle program: they built four orbiters initially, then a fifth as a replacement after one was destroyed, and that's it. They operated these things for decades.
Compare with the Falcon 9R program: they built, launched, and deliberately threw away one already. They'll do this several more times. The vehicles are cheap enough that they can intentionally lose hardware as part of the test program. The vehicles are capable enough that they can get paid to test them by still using them to launch payloads into orbit while doing it.
The Shuttle had to fly dozens of times a year just to break even compared to the cost of building new expendable rockets. They could never come anywhere close to that. This is the sort of thing that M. Bonnal is referring to, and it makes sense in that context. The difference is that Falcon 9R costs little more than a regular Falcon 9 to build, and on the second launch it will already be past breakeven and saving money. Perhaps it can be reused 40 times, but unlike the Shuttle, it doesn't need to be reused 40 times to justify an extremely high initial cost and extremely high ongoing overhead.
There is so much more margin of error with the Falcon 9, the industry veterans have no idea what they are talking about. Their experiences are built on a generation of excessive complexity.
Anyway, you are right. The amazing thing about Falcon 9R is just how little was added to the standard Falcon 9. Slap some legs on it and give it some fancy software and call it a day, pretty much. It's a completely different world from custom-designing a reusable spaceplane from scratch.
I don't know that too many people make that distinction, but it does seem to be the accepted definition of "sailplane". There's nothing wrong with calling, say, an ASW-20 a glider, but calling the Shuttle a sailplane isn't quite right.
Shuttle operations were so complicated they required a standing army of engineers and technicians just to keep the fleet operational and perform all the refurbishments in between flights. The cost of that standing army plus the limitations of flight rate imposed by the duration of refurbishment operations meant that ultimately every Shuttle flight ended up costing about $1.1 billion. Contrast that with the cost of building a new orbiter, which was around $2 billion.
This was a huge problem for the Shuttle program from day one, because the huge ongoing costs meant that NASA manned spaceflight were locked into the Shuttle system and incapable of budgeting for developing improvements or alternatives. And indeed, they ended up stupidly flying more or less a prototype design for 3 decades.
But that sort of insanity isn't a universal problem for reusability. Indeed, the idea that any sane entity outside of deeply entrenched government bureaucracy would embrace a "reusable" system where each flight cost 1/2 as much as brand new hardware is a bit of a stretch on its own. However, that's a problem which is not even remotely possible for SpaceX, because the cost of the vehicle is so much lower. SpaceX already manufactures reusable F9 first stages, they are just ordinary F9 first stages plus landing legs. They cost a few million dollars to produce, not billions, and SpaceX pumps out several of them a year. SpaceX is planning for a very low cost of operations for reusability. They'll land the stage, retrieve it with a vehicle, bring it into a warehouse where it can be run through some tests then assembled into a full launcher, fueled, and relaunched. There is a very low cap on the costs for that operation, because if it starts costing more than pumping out new stages then SpaceX will stop doing it. So there's no "trap" possible for SpaceX. If they get reusability working then they'll just have that much more manufacturing and R&D capacity plus profit margin for working on the next generation of rocket.
It really seems to me that only companies with the "California tech startup mentality" (whether they're in Cali or not) can actually do anything today. I'm starting to refer to the rest of the economy as the "legacy economy," "legacy industry," etc.
I understand some of the reasons: bottom-up engineer-led engineering teams, lean operations, little or no bureaucracy compared to large older institutions, etc. But there seems to be something else too, some kind of difficult to quantify attitude or mentality. I get it, but I find it hard to describe except to say it's about believing in the future, believing that things are actually possible. Everyone else everywhere else seems to be looking backward and thinking of all the reasons things can't be done.
Like this, for example. "Here, we're NASA, and here's why it can't be done." The NASA of the 50s and the 60s is rolling in its grave!
Legacy economy indeed. Disrupt everything. Ditch it all and start over.
As to the NASA of the 1960s, that was a NASA who developed 4 different manned spacecraft in a 10 year period (Mercury, Gemini, Apollo CM & Apollo LM) and flew them on 4 different rockets (Atlas, Titan, Saturn I, Saturn V). That's the sort of activity that looks similar to what SpaceX is doing today and very dissimilar to what NASA is doing today (sinking billions per year into masses of bureaucratic BS and very little hardware).
Followed by the Atlas that used kerosene and LOX like Space X, the Titan which used fuels that could be stored in the rocket for long terms, making it a practical ICBM (taken out of service in the mid-80s).
All those were adopted from military missiles, and were followed by the purpose built Saturn I and V, and both of their first stages also used kerosene and LOX.
Am I seeing a pattern here ^_^?
(Quibbles: the "kerosene" used in rockets is actually highly refined stuff: https://en.wikipedia.org/wiki/RP-1 and the Saturn upper stages used liquid hydrogen and LOX.)
The big advantage for SpaceX in using the same propellant and mostly the same engine in the 2nd stage on the Falcon 9 is that the 2nd stage is dirt cheap. In any other rocket they'd see the opportunity to build a small little LOX/LH2 stage and improve the payload tremendously. But then you'd end up with a lot of cost in the 2nd stage, which is precisely what SpaceX doesn't want. With the Falcon 9 design 3/4 of the manufacturing cost is in the first stage. Which means that reusing just the first stage (by far the easiest reusability prospect) nets huge gains. This is what you call low lying fruit. And it's fruit that would have been there for anyone else to pick except they kept making the tree taller without thinking about it, instead of just picking the fruit.
So now instead of a more expensive rocket with more payload SpaceX ends up with a vastly less expensive rocket with slightly less payload (easily worked around with the Falcon Heavy configuration). Even just 2 reflights of the first stage translates to an overall reduction of 50% in the capital cost of each flight (or 60% in the case of the FH).
I'm a little concerned about the propellant crossfeed aspect of the Falcon Heavy (the liquid boosters on the sides feed propellant to the core, which will be mostly full when they separate ... and that's much sooner than a normal Falcon 9 first stage is finished boosting, might make reusability easier).
If they pull it off they'll be the first ever, right?
Not saying they can't, but unlike almost everything else they're doing, this is new "rocket science" to my limited knowledge.
The scheme, as described in more details by you, is a lot simpler, and less concerning than I was envisioning (figures, given this is Space X). There's no cross tanking, that is, feeding the tanks of the central core from the sides.
Hmmm, although as I remember, when they had to shut down one engine during an ascent, it shed debris because of the pressure imposed by the adjacent engines. But I note the Merlin 1D, while producing significantly more thrust at sea level (nearly 50%), can throttle between 70-100 percent. Obviously they've thought through these stresses, and of course there are structural implications in making them highly reusable, they're willing to trade off weight penalties for durability, and therefore likely strength.
Anyway, thanks for the details, and all the contributions you've made to this topic!
And there's the rub. SpaceX is working on having reusable components land on solid ground.
The meat of the 'argument' (and really, I don't believe it's an argument) is that the Shuttle Main Engines (the ones mounted on the tail of the Shuttle, the ones that land like a plane) had similar expectations of reusability, and ended up being absolutely terrible at reusablility. That's the point to address, not the throw away line.
Another consequence of the LH2 choice for the Shuttle is that it didn't have enough thrust at sea level to get off the pad, so it needed high-thrust, low-cost boosters as a 0th stage. Solid rockets being the obvious go-to solution, those got used.
In a very real way the choice to use LH2 resulted not just in much of the cost and complexity of the Shuttle system but also the death of 2 crews and the loss of 2 orbiters.
The methane direction SpaceX is taking going forward strikes me as almost a no-brainer. It gives you a higher isp than kerosene, is about as easy to handle as LOX, and is super-abundant and cheap. In addition to being basically 95% of natural gas it can also be made electrochemically/thermochemically from water and CO2 or trivially obtained from biomass digestion. The latter options are all possible on Mars. It's also a renewable fuel if you get it from non-fossil sources on Earth.
Also, while some comparisons to the shuttle's main engine are warranted, refurb costs of the shuttle's solid boosters will not be comparable as they landed in the ocean rather than on dry land. Ocean water is nasty stuff for hitech equipment.
They are definitely planning on doing boost-back to the general area of the launch site though, not aiming for a target down-range. The launch this weekend will have them bringing the stage back closer to land than the previous attempt.