What do you mean by lower performance? When people talk of reducing cost, they usually mean dollar per kilo. So, what reduction of performance could improve the $/kg ratio?
Would using ICBMs for LEO non-human insertions make sense? Has the Air Force been using those now that the Shuttle isn’t shuttling?
So far, using a plane to get the rocket up to 10-20KM seems to only have been used for small rockets like SpaceShipOne. Would using a large bomber/cargo plane to get a rocket into thin atmosphere reduce costs?
I realize that this can vary a lot but how much do rocket engines tend to go for? I’m looking for a ballpark for liquid and solid rockets.
All of the standard costing models (in the United States, at least) for predicting the cost of a rocket system are based upon inert and carried mass, hence the focus on maximing the propellant mass ratio to obtain the highest payload for a given vehicle configuration as the quality metric. However, this drives both development costs (to use exotic materials and processes) and minimum margins (which results in reduced reliability and complexity in handling and integration operations). Since the cost models are essentially developed from the emperical data on ICBM-based launch vehicles where performance for a given size enveloped is key, processing is labor intensive by virtue of vertical integration, and cost is not a driver they really don’t provide good guidance on novel ways to reduce costs. However, there are plenty of studies on low cost access to space (search on “big dumb booster”, “double bubble rocket”, and “Traux Sea Dragon” for examples) which clearly indicate that going to a lower performing, heavier vehicle with reduced complexity (pressure fed or expander cycle propellant feed) has positive cost reductions by as much as an order of magnitude or more compared to a comparable staged combustion or gas generator cycle, especially in the heavy lift and above category.
Nearly all orbital rocket launch systems (aside from the STS and SLS) are based on ICBMs, although the current EELV vehicles (Delta IV, Atlas V) are sufficiently distant from their progenators that they are essentially new vehicles. The Orbital Sciences Minotaur family use the LGM-30F ‘Minuteman II’ and LGM-118A ‘Peacekeeper’ lower stack as the basis for their orbital launch vehicles.
The advantages of air launch are not to much the propulsive performance for operation at higher altitude as being able to launch directly into the optimal azumith without being constrained by launch sites or worry about overflying populated areas. There is also some advantage to being about ground winds and weather, but the tradeoff is considerably greater complexity and practical limits on the size of vehicle that can be lifted.
The actual cost of an engine is, while not insignificant by itself, almost irrelevent in terms of overall cost to orbit. That’s why reusability isn’t the great trade it may seem to be. Reducing the procesing and checkout cost by reducing complexity is the key to realizing order of magnitude cost reductions.
Looking for something else I stumbled upon this old chestnut. Somehow it reminded me of some of the issues faced in building spacecraft. (I don’t know the attribution, it seems to have been around forever. It is shall we say, rather old tech - but the point remains true.)
The Designer
The designer sat at his drafting board
A wealth of knowledge in his head was stored
Like “What can be done on a radial drill
Or a turret lathe or a vertical mill?”
But above all things a knack he had
Of driving gentle machinists mad.
So he mused as he thoughtfully scratched his bean
“Just how can I make this thing hard to machine?”
If I make this perfect body straight
The job had ought to come out first rate
But would be so easy to turn and bore
That it would never make a machinist sore
So I’ll put a compound taper there
And a couple of angles to make them swear
And brass would work for this little gear
But its too damned easy to work I fear
So just to make the machinist squeal
I’ll make him mill it from tungsten steel
And I’ll put these holes that hold the cap
Down underneath where they can’t be tapped
Now if they can make this it’ll just be luck
Cause it can’t be held by dog or chuck
And it can’t be planed and it cant be ground
So I feel my design is unusually sound.
And he shouted in glee, "Success at last!
This goddam thing can’t even be cast.’
While I would not want to underestimate the difficulty of the task, orbital rocketry has been slowly inching away from “nearly impossible” for some time now. Any organization that makes a serious attempt can build a 2-stage kerosense+LOX rocket that brings a useful payload to orbit, and almost certainly succeed. It’s well past the point of “black magic” engineering.
You can take any engineered system, name some uncommon characteristics of that particular system, and say “X is inherently unlike any other highly engineered system”. It may be true but it doesn’t convey much information. At best, it enumerates the set of things that can’t be approached with “rule of thumb” type engineering–useful but not an argument in and of itself.
Rocket engines don’t need to be tested for hundreds of hours–they just need an equivalent “proportionality constant” for their test time. An engine designed to fire for a couple hundred seconds may only need a few thousand seconds of testing to achieve a similar level of confidence.
The same engine can be tested multiple times, so there’s no limit to the timeframe.
A full static test of a Falcon 9 first stage would take 170 seconds. Remove all but one engine and the tank lasts 1530 seconds.
In fact, all the way back to 2008, SpaceX had tested a single engine for 27 minutes.
More flights and more commonality means more amortization of testing, and can thus justify a higher cost.
I hope I didn’t imply anything about commoditization inevitably resulting in lower costs–as you note, there are many paths in that direction and some of them are clearly the wrong ones.
I am fully on board with your arguments about SSTOs. Even on paper they do not look good. Perhaps one day we’ll be able to build composite hydrogen tanks that weigh a couple percent of the mass they hold, but not today. Reusability will have to take another route for the foreseeable future. SpaceX is taking their own route, and while it remains to be seen if they succeed, I look forward to the results.
That was certainly the case on the Shuttle, but I would claim that’s because the engines never quite reached a level of reliability where they could get away without full refurbishment after each flight.
But aside from things like ablative nozzles, as far as I know there are no elements in a rocket engine which must wear significantly after each firing. Sure, the conditions are extreme, but unless there is an element that is known to degrade at a certain rate, an engine that lasts 200 seconds should be able to last 2,000 or 20,000 seconds with sufficient refinement.
One problem I’ve heard of is coking of the combustion chamber and turbopump. It is, from what I read, a problem specific to kerosene engines. I wonder if SpaceX’s interest in methane (a fuel which is otherwise not particularly compelling next to RP-1) is related to a lifetime limit for kerosene engines.
There is always a finite envelope in which redundancy is effective. An aircraft that loses an engine on takeoff will not succeed in carrying the passengers to their destination, although there is a high probability that they will survive. A rocket that loses an engine may not complete its secondary mission, or not be able to put the payload into a sufficiently high orbit, or otherwise. Sure, rockets have narrower margins than other craft, but that doesn’t mean they are fundamentally different.
I don’t disagree with any of this, and yes, what you describe is generally what I have in mind when I say “commoditization.” Using the same fuels for upper and lower stages, nearly the same engines, large numbers of engines for greater amortization of design and test costs, and so on all go in this direction.
I also agree that the costs will never come down to that of commercial travel, but that we are so far from that that there is a lot of room for improvement. AFAIK, for commercial air travel, fuel is something like 1/2 or 1/3 of the total operating costs. For a Falcon 9–already fairly cheap–it’s more like 1/250. So there are two orders of magnitude to go, and if SpaceX can achieve even a single order of magnitude it will be a stunning achievement. Even just a factor of 2 will put them in a dominant position and (modestly) improve the range of things for which space is an economically viable destination.
Stranger : when I read your posts, I sense an attitude of defeatism.
We’ll NEVER solve these problems. We’ll always be at the mercy of long engineering schedules, ridiculous costs to reach space, slow technological advances in all areas, and so on and so forth.
I feel like if I talked to you in person, you would pretty much deny that anything other than marginally better than the status quo was flat out impossible.
Do you feel this way overall? How difficult will space travel be in 50-100 years? Do you think that technologies that would make this easy are even possible?
I personally feel that, based on the evidence, the bottleneck is actually the speed of a human brain. Human brains run at about 1 kHz and most of the time a neuron takes is wasted, during which neurotransmitters diffuse across synaptic clefts.
A principle called “Amdahl’s Law” is the reason why engineering takes so long and is so difficult : TLDR the reason is that since human minds are so slow, when you split any problem between multiple people you cannot speed up a process any faster than the non-parallelizable portion of the task.
Designing a rocket has many non-parallelizable portions. What I mean is, in mathematics, a calculation that depends on a previous step cannot be performed until the previous step is done. For instance, if you wrote the statement : A + B = C, if C>X do X else do Y you cannot solve the conditional until you do the addition operation.
If you are designing a rocket, there’s enormous inter-dependencies between all the different systems, and so you end up with a large number of steps that must be done in a serial manner. It doesn’t matter whether you have enough budget to afford 1000 engineers or 100 million, you cannot do it any faster than the serial steps can be performed by merely human minds, which run at about 1 kHz as mentioned earlier.
This bottleneck is not a law of physics. There’s straightforward proposals to build massive computing centers using specialized hardware that will, in 20-50 years, be able to emulate an entire human mind. Reasonable estimates of speedup achievable with one of these machines is somewhere between 1000 and 1 million times. So, in the short term, if you had an emulated human engineer who thought 1000 times faster, you could have them working the critical path portions of your engineering project and get things done in a fraction the time.
In the longer term, you would build vast arrays of extremely high speed robotics, and I think you could build a factory that could churn out newly designed machines in a matter of hours. That is, the super-intelligent emulated humans would design a new machine in a few minutes and have a full scale prototype within hours.
…I don’t sense that at all. Remember we are posting in General Questions: and his posts are based on facts and realism: not optimism, and considering he is the closest thing to an expert in this field on this subject that this board has I’m not sure why “his attitude” is important. Stranger appears to have comprehensively answered the OP’s question, and done a great job of “fighting ignorance”, which is what this board is about.
Oh, okay. That is an amazing revelation you have there. So you’re going to call up the various extant and former private spaceflight companies such as Rotary Rocket, American Rocket Company, Interorbital Systems, Blue Origin, Reaction Engines Ltd, Copenhagen Suborbitals, PlanetSpace, t/Space, Truax Engineering, Rocketplane Kistler, Starchaser Indistries, XCOR, SpaceDev, Armadillo Aerospace, and others, tell them to take their hands off their dicks and start doing this simple task that “any organization that makes a serious attempt” can do. In fact, I’m not sure why you just don’t go out and do it yourself seeing as how your expertise and ability to understand and solve the almost trivial problems of space launch eclipses the current lame-brained contenders in the field. Show us all how it’s done.
Yes, words from the expert. It’s all just a bit of coking that is the problem and elsewise we could operate engines more reliably and for longer durations than a Cessna 172. O’ Enlightened One, please educate me and the rest of the aerospace community about these trivial tasks that we’ve all conspired to make so difficult.
I don’t know where you are getting this but you are bringing it into this discussion all on your own. In this thread and elsewhere I have repeatedly and with great length and explication, described areas of improvement and specific design features which could improve reliability and reduce costs substantially. For instance, [POST=16428446]here[/POST] is a thread where I describe a feasible path to a sustainable, reduced cost, high rate launch infrastructure. It is true that I address the absurdly optimistic claims about how easy all of this should be with the harsh cold shower of reality, but that is because I am attempting to bring some actual knowledge and experience with rocket launch systems instead of just repeating a bunch of ill-informed bullshit that space nerds like to spout to each other to prove how much smarter they are than the engineers who actually work on these systems. This is like the guy who goes into a restaurant claiming that he could prepare the meal so much better than the cookstaff even though he’s never worked in a kitchen in his life, or the self-professed football expert who won’t shut the fuck up about how Payton Manning should get off the field and make way for a real player even though he couldn’t make it halfway down the field without running out of breath and having his flabby gut pop over his belt.
How difficult will space travel be in the next century? It really depends on how much effort we put into building a sustainable infrastructure to support launch and in situ resource exploitation of space. Space enthusiasts tend to get wrapped up over singular accomplishments like sending a crewed mission to Mars or individual technologies like an air-breathing multimode spaceplane without understanding the lessons we should have learned from the last century; to wit, that if you don’t have both a reliable method of getting to orbit (and beyond) AND a solid, long term plan and set of goals which justify the cost and risk of being there, no individual technology or flags-and-footprints mission will sustain a space program with enough momentum to accomplish useful and enduring goals. I’m sorry, but space programs don’t exist strictly to excite and delight space nerds; they exist because they a) create high paying technical jobs in critical Congressional districts, b) help us develop a greater understanding of the world around us and the ways in which it can sustain or will destroy us, and c) to provide material benefits to the public at large, e.g. Earth surveillance (weather forecasting, climate prediction, agricultural management for higher crop yields, defense, et cetera), telecommunications (satellite television, handheld worldwide communications, et cetera), space surveillance (solar flare warnings, meteor flybys), and navigation (Global Positioning Systems, EPIRB beacons, et cetera). If we’d had a sustainable program to follow the Apollo Manned Lunar program (which was essentially cancelled almost as soon as Armstrong set foot on the Moon) we could have regular space exploration, large orbiting habitats, resource extraction and utilization from Near Earth Asteroids, wider robotic exploration of the Solar System and probably even some amount of crewed exploration of Mars and the asteroid belt. Instead, we got…the Shuttle, a delivery truck to nowhere that was too expensive to operate commercially and too useless to provide any genuinely advanced capability. The hurdles to increased space exploration and human habitation in space are not fundamentally technical (though there are some substantial technical challenges to be met in order to make that a reality); the basic problems are planning, commitment, fiscal and scientific justification, and an understanding that pursuing the most bleeding edge technologies does not equate to effective or cheap access to space.
This has to be the weirdest argument I’ve ever seen for why the space industry is so slow to make progress in advancing the technology and reducing costs. You’ve essentially asserted that this is all just a problem of scale, and like solving a computational fluids dynamics matrix or making cheese sandwiches, if we just break up the problem to smaller pieces and distribute it out to this hypothetical artificial intelligence-enabled super-beowulff cluster we could design a magic carpet spacecraft in minutes. After that, it can then go about resolving the world’s cultural issues, provide a self-consistent statement of Skolem’s paradox, and show how the Laffer curve actually makes some kind of sense.
The problem is, we aren’t making fuckin’ cheese sandwiches, and the bottleneck isn’t just the number of people working the problem who aren’t/can’t/won’t talk to one another. In fact, we have an entire systems engineering and system requirements analysis methodology to aid in codifying and communicating the complex technical detail in large highly engineered systems. And as much as space enthusiasts and upstart software tycoons like to claim that if we ignore all of that nonsense and just throw smart people and a barrel of money at the problems they’ll figure it out in no time, those companies tend to fail, and fail badly (e.g. the sequential failure of the first three Falcon 1 flights to stupid problems that would have been caught by applying industry standard practices and independent review), and most of those that don’t have an 'angel investor’ who will continue to pour on the money like sugar in a child’s cereal go out of business even if they have a good fundamental concept (see the Kistler K-1 as an example of good engineering and a solid concept being undermined by bad planning, communication, and having too many cooks in the kitchen).
There are some basic and fundamental hurdles to making space launch cheaper, faster, and more reliable, and some of the limiting factors are material capability and the basic physics of space flight. None of them are showstoppers, provided your goals aren’t totally unreasonable (e.g. daily flights to the Moon or some other science fiction nonsense with no basis in reality) but they don’t get solved by just trying to break the problem into smaller pieces; they will be solved by advances in either basic propulsion and materials science, or by recognizing that there are ways within the existing technology or ready extrapolations thereof to reduce the amount of overhead and labor in the entire design/test/build/integrate/fly cycle, and particularly in the testing and integration phases by building in higher margins in lieu of intensive test and inspection so as to automate the processes.
It is absolutely true that established companies like Boeing and Lockheed Martin have almost no interest in making space launch cheaper because it will just reduce their margins; it is also true that many of the upstarts have proposed methods and designs which offer the potential of substantially improving launch capabilities. But just because Elon Musk comes on stage doing his Hammer dance and boasting that we’re all going to be retiring to Mars in a decade don’t fuckin’ make it so! Musk isn’t personally designing anything and frankly appears to know fuck all about aerospace engineering or the small details that make a rocket launch vehicle succeed or fail; he has a team of hundreds of engineers and technicians working long hours and performing labor intensive build and inspection all so he can brag in his Walken-esque manner about how easy this entire business is. Uncritically accepting this kind of marketing speak and using it as the basis to argue that everybody else working in the field is either an idiot or is part of a gigantic conspiracy to keep space enthusiasts from spreading their seed across the galaxy and have sex with blue-skinned Martian princesses three a time is infantile and obtuse. The actual reality of achieving a sustainable space logistics and exploitation infrastructure requires applying critical thinking skills to understand how to plan and advance the technology and capabilities in a practicable manner.
You’re building a strawman here. I can’t build a skyscraper or bridge either, and yet I can say confidently that they are straightforward engineering problems and that anyone with enough money can assemble a team of engineers and build one.
Most of the companies you mentioned are directed at suborbital flight. Others are going for advanced, unproven propulsion methods, or had some form of reusability. All have limited budgets. None are seeking to develop a “boring” kerosense-LOX rocket because, well, it’s been done before.
So no, I would not include any of them among organizations making a serious attempt at making a minimum baseline rocket. Blue Origin would be the closest but even then it’s not obvious how aggressive their orbital timeline is, and they are working with LH2, which ups the difficulty significantly.
Also, I want to be clear that I was not referring to an economically viable rocket, which is a necessary condition for most organizations that want to stick around. I was referring solely to a machine that solves the problem of getting some useful payload into orbit, which is the task that you have said is “almost impossible”.
I was actually hoping for some feedback there, but I guess snark is cool too.
People are wondering about SpaceX’s announcement about their methane development plans. Methane doesn’t look all that stunning from an Isp point of view, but it’s cleaner and may make a full-flow cycle easier if coking is a bottleneck.
I was also curious about progressive failure modes. Coking, I guess, is one. The Challenger report mentioned cracking of the turbine blades, but it sounded like a manufacturing defect. What other modes are there? Erosion of the combustion chamber? Breakdown of lubricants? I’m genuinely asking here.
'Course, that ain’t free. Sometimes it’s better to fail often than to tie yourself to a review bureaucracy (and sometimes it isn’t).
Do you know of any postmortems on what went wrong with the K-1? Bad management can kill any project but I’d be interested in the specifics.
It is aptly clear that you do not have enough information beyond very cursory knowledge to make any kind of realistic estimate about the effort it does or should take to develop and operate an orbital launch system. You’ve asserted that it is an easy and routine effort for “any organization that makes a serious attempt”, when the empirical reality is that it is a very difficult task that even with the legacy knowledge requires hundreds or thousands of person-years of direct effort, and even the companies and people who have been doing this for decades frequently experience failures, as demonstrated by the long list of companies which have failed to successfully develop and launch a rocket system. Arguing that some of them are focused on suborbital launch or using LOX and hydrocarbon propellants, and therefore are not valid for comparison is disingenuous; in fact, if they fail at even more modest goals and using ostensibly more simple solid motors or hybrid engines, how hard to you think it must be to launch a staged combustion-powered engine using cryogenic oxidizer into orbit?
Methane is a desirable fuel because of its storability (it doesn’t gel up the way kerosene/RP-1/RG-1 does or expand as LH2 will), its ease of use (can be readily vaporized just by bringing it up to temperature), has relatively low molecular weight combustion products (good specific efficiency), has relatively benign corrosive properties, and most importantly for interplanetary exploration, can be readily synthesized from commonly found elements. Methane, and other gas-phase propellants such as propane, diethyl and dimethyl ethers, et cetera, have not been seriously considered because of their relatively low performance compared to cryogenic fuels and are not as storable as hydrazine-based propellants, but they have a lot of nice features for long duration use and in-situ synthesis, and incidentally, are almost ideal fuels to use in continuous wave detonation rocket engines which are probably the next phase in chemical propulsion.
Coking (the deposition of solid carbonaceous material on critical surfaces such as injectors) can be a problem with all hydrocarbon fuels but especially heavy hydrocarbons such as RP-1/RG-1/kerosene because of the high carbon content and the tendency of incomplete combustion. This is normally addressed by ensuring good mixing and running oxidizer rich (also desirable to get better delivered I[SUB]sp[/SUB]), but can be a problem when throttling the engine or resulting from combustion instabilities. In normal operation significant ‘coking’ should not occur as it represents a loss of efficiency and may result in operating problems in the engine.
As for progressive damage, engines, thrust vector control system, and the propellant feed system see a lot of stress over their short operating duration, including pressure cycling, wear on bearings in pumps and joints, thermal cycling and thermal degradation, seals (especially cryogenic seals of any kind), and whether you design the nozzle for ablative cooling or not, the reality is that the nozzle throat and upper wall will experience some degree of ablation and degradation due to heating (more with solid propellants than liquids, but there is still progressive damage). It isn’t impossible, of course, to design and built an engine that can operate for minutes or even hours of total duration provided that you don’t overstress it into a limit failure, but to combine both long operating life and high performance is extremely challenging from both a design and manufacturing quality standpoint. This is an area where trading performance for life could have significant cost reductions (e.g. the Truax Sea Dragon concept) as well as provide reliability through robustness and simplicity, but that isn’t the route that companies such as SpaceX are taking.
The essential problems that doomed the Kistler K-1 rocket were not technical in nature, and in fact they passed many technical hurdles including extensive testing of the recovery system at Yuma Proving Grounds. What they failed to account for was the optimism in the schedule which assumed a ‘success-oriented program’ (one of my favorite bullshit phrases) along with an ability to manage a bunch of subcontractors without firm control over the program. They had companies like Aerojet, Scaled Composites, Lockheed, Orbital Sciences, and others all working on various systems with minimal coordination, and so of course all of these companies went off and did what they pleased without any consideration for the overall project or goals, which resulted in Kistler not being able to demonstrate technical milestones despite a strong (in my opinion) operational concept and vehicle design. This is a worthwhile caution for anyone electing to enter into the field of orbital launch systems; even if you think you understand the technical problems and have a good crew of engineers and technicians, the organizational problems of getting everything in a highly complex system to come together when you need it to is enormously challenging.
Adding even one or two technologies that are not on a firm path of technical maturity can throw the entire schedule out of wack and result in pouring tens or hundreds of millions of dollars down a blind hole. This isn’t just a lesson learned from Kistler; it is a general observation about all attempts to develop multistage launch vehicles. The ICBM and FBM programs, as well as the Apollo/Saturn program, succeeded as well as they did because of the extensive planning, the rigorous systems engineering practices, the firm management of various subcontractors and systems, and not just a small amount of working people into exhaustion and divorce, and occasionally into a grave. Regardless of your personal opinion based upon whatever you’ve read from press releases and space enthusiast sites, this is not a routine or trivial engineering endeavor as even a cursory examination of the previous actors who have attempted and often failed will show.
But perhaps I’m incorrect in my assessment of your level of practical knowledge about the field. Please feel free to outline in detail the various phases and specific efforts necessary to design, test, and fly a launch vehicle, and the ways in which “order of magnitude” reductions in operating costs can be achieved.
Please, for the sake of the children, do some kind of minimum fact-checking before making claims that are almost entirely wrong.
For the record, the Apollo lunar program was first conceived in early 1960 as a follow-on to Project Mercury. Kennedy announced the intention to Congress on May 1961 for the program to place a man on the surface of the Moon by the end of the decade, but the program was already well into concept selection and system design phases. The first uncrewed flight of the Saturn V rocket (needed for Lunar missions) was the Apollo 4 flight in November 1967. Apollo 6, the first uncrewed flight to perform a trans-Lunar injection maneuver, flew in April 1986 and was very nearly lost due to destructive POGO oscillation in both the S-II and S-IVB stages. The first crewed launch of the Apollo/Saturn system was Apollo 7 in October 1968 using the Saturn IB vehicle to Earth orbit. Apollo 8 in December 1968 was the first Lunar orbit mission. The first crewed mission to land on the surface of the Moon was, of course, Apollo 11 in July 1969. There were six more attempted Lunar landing missions using the Apollo/Saturn system (Apollo 12 through 17) from November 1969 to May 1973, with five successes and one recoverable mission failure (the infamous Apollo 13 in April 1970). There were three more planned ‘J-class’ lunar missions which were cancelled, and surplus Apollo and Saturn IB systems were used for the three Skylab missions (SLM-1 through -3) in 1973 and the Apollo Soyuz Test Project in 1975 (largely a goodwill mission to make to a Soyuz capsule).
So it was a minimum of about 7 years from the inception of Apollo to the first (uncrewed) flight of a vehicle capable of sending a crewed payload to the Moon, and the Moon program continued until 1973, almost thirteen years after the project inception. All of this information is readily accessible from a variety of sources on the internet including NASA’s own historical archives that are publicly accessible, so please do make the effort to educate yourself before making baldly incorrect claims.
I never said it was easy; I merely said the problem had inched back from “almost impossible”. You listed a bunch of companies that hadn’t made it to orbit, but there is a long list of countries and companies that have made it. Yes, they mostly depended heavily on legacy knowledge, but that hardly invalidates my point, which is that many of the hard problems and dead ends have been identified already. If only one or two organizations had ever made orbit I’d agree that it remains a “black magic” technology reserved only for the “lucky” or exceptionally well-funded. But there are at least a dozen, and perhaps more, depending on how much indigenous technology you require to count as a separate effort.
I find it interesting that you would use Armadillo Aerospace as a counterexample, since it was started by a software engineer that perhaps underestimated the realities of orbital spaceflight (they nevertheless persisted and have had some success in non-orbiting vehicles). You can’t have it both ways. Either AA was always a joke and not a legitimate counterexample; or they are a “real” company, in which case a guy with zero experience (Carmack) can, in fact, just decide to make a rocket company.
You misread me–I argued that using LH2 puts an organization in a class outside of the claims I was making. It is kerosene+LOX that is relatively easy (not absolutely easy, but relatively easy).
So my point, again, is that building a simple rocket is not nearly as hard as it used to be. LH2 is challenging, as you noted, so for a LH2 effort to fail is hardly surprising, especially for the smaller companies. And obviously, building an air-breathing rocket is an enormous challenge for anyone, let alone an underfunded British firm, so it’s no surprise that Reaction Engines haven’t made orbit either. Likewise for anyone trying to build an SSTO of some kind.
These things are valiant efforts but have nothing to do with doing the minimum necessary to get to orbit. And since doing the bare minimum is also not particularly economic, the only remaining reason to build a minimum-tech orbiter is national pride. Perhaps you have a long list of national efforts that failed. There are certainly a few–Brazil has fared particularly poorly. But as best I can tell, far more have succeeded than failed, and that tells me that persistence is an effective strategy.
Modest goals but extremely modest budgets. I think it’s clear that you can’t build an orbiter for $10M, or even $100M, but that $1B is probably enough unless you have a spectacularly ineffective organization.
Thanks for the info. Gotta read up on detonation engines. I know from other contexts that it is theoretically more efficient than deflagration, but have little knowledge beyond that.
Okay. I wonder if throttling is another factor in SpaceX’s decision. Obviously throttling is an important component of any landing operation.
I suppose this is the key. It does have some positive synergies with redundancy: if you can handle an engine failure, then under normal conditions you can throttle back a bit and save some wear and tear.
Maybe, but they aren’t going for the absolute bleeding edge, either. The Merlin engine uses a gas generator cycle, which is less efficient but requires less advanced metallurgy than (say) the oxidizer-rich staged cycle of some Russian engines. They have also claimed that they have a relatively low parts count.
Of course, this is still lot more complex than a pressure-fed Sea Dragon (a rocket which I’d really like to see built).
It’s interesting that you say that, because when I read the Astronautix page of the K-1 project, I saw their list of partners, and immediately thought “holy shit, that’s a lot of subcontractors to manage”. I guess my intuition was not wrong in that case…
No denying that, but I hardly think it’s unique to launchers. I’ll grant that rockets are probably more coupled than most other engineered systems, so that if one subcontractor fails you it’s far more difficult to find a replacement, but still: big projects fail all the time for basically the same reasons.
Sure, and that’s why I limited my argument above to a relatively simple kerosene+LOX rocket. I would never suggest anything straightforward about technologies that have not already been put into full production multiple times unless it is a very incremental improvement.
I certainly don’t claim any practical knowledge. I only know what I read and can compute from first principles. That said, while I do not deny some utility of practical knowledge, I also think it’s of limited use in developing new technologies–almost by definition, a new technology must discard the old lessons.
At any rate, I certainly don’t have any specific plan in mind to achieve two orders of magnitude reduction in flight costs, except by use of non-rocket technology (which is an entirely different topic). Achieving a single order of magnitude reduction probably requires “perfect” reusability; I’ll define this as a rocket than can be launched twice in a row with only refueling and restacking (even if, in practice, you would generally do more stringent testing). It also requires reductions in ground costs; a major factor in this is a reusability program that flies the stages back to their launch point.
All of that is probably a long way off, but I’d consider a factor of two a legitimate medium term goal, achievable with a moderately reusable first stage and refinements to their stack to reduce complexity and integration costs.
Okay, forget about Apollo redux. WHat about a simple TLI leading to a lunar flyby? Is that doable in 12-18 months. No need to do all up, you can build components in Earth orbit if it is practicable.
You get a Soyuz/Shenzou type craft to mate with something that can given it the ommph to get it into Lunar trajectory. Hmmmmmm…is there anything like the old SIV B, which is around and launch able.
…Stranger makes the following estimates based on his intimate knowledge of the industry:
You don’t appear to be posting anything in contradiction to this, in fact it appears you are arguing for the sake of arguing. You admit you have no practical knowledge: so do you actually dispute the timeframes that Stranger estimates, or are you simply putting faith in the “human ability to overcome adversity?” This is General Questions: if you have real world information that shows that things could be done faster I would love to see it.
Admittedly, we did get off on a tangent, but I’m not sure where you get the implication that I disagreed with his estimates for an Apollo-like mission. They seem reasonable to me. I had a fairly narrow disagreement with a particular statement of his and it generated some interesting (to me) discussion.
In post #41 I sketched out a plan for what amounts to a suicide mission, but would probably have some chance at working. Stranger more or less agreed in post #42, with the caveat that the mass estimates should be doubled. That would not ruin the mission, as the plan was to have two landers (for better odds), each 1000 kg, but the same launch vehicle could support a single 2000 kg craft. My outline used an existing launch vehicle and components which are already designed or at least mostly fleshed out on paper.
He estimated a “couple of years” for that plan, which is fair enough, and a reasonable improvement over an Apollo-like program. It’s only interesting in the context of a sci-fi “aliens will destroy the Earth if we don’t get to the Moon in X timeframe” scenario, of course, but that’s the OP’s hypothetical.
Stranger : you’ve made a lot of criticism of SpaceX, pretty much calling it a cult of personality.
Now, the news articles on the company show photographs of their main facility. There’s the factory, where the manufacturing is done (they cut a lot of the parts on site), and, overlooking the factory is the offices where the engineers and the executives meet. Their director of propulsion personally designed and built a rocket engine in his garage years earlier, so presumably their upper management actually know something about rockets.
Their mission control is on the ground floor, with floor to ceiling windows that look out on the factory.
Allegedly, they use a minimal number of subcontractors, doing most of the rocket-specific design and manufacturing in house.
Now, this to me sounds like a new way to do things. Allegedly, the “cost-plus” payment model of government encourages you to have a dozen companies in between the top level integration company and the firms actually making the parts. Has this model been tried before?
With the “ticking bomb will kill us all” scenario, could you speed up solving the problem by funding a dozen parallel efforts? You’d have 12 or more separate efforts racing each other and the clock to solve the problem. This would obviously cost 12x as much, or more, but in this case it’s get to the Moon and cut the red wire or we lose the planet. Apollo didn’t have this level of urgency, and if we lost an astronaut it would cause tremendous negative publicity.
Again, parallelization does not equal efficiency or compressing schedule, as the old aphorism of Brook’s Law, “Nine women can’t make a baby in one month,” succinctly indicates. Of course, you can perform independent tasks, such as developing the service module and habitat module in parallel, provided that you have sufficiently well defined in interfaces and matching requirements. But developing and testing, say, twelve different super-heavy lift launch systems in parallel? Not only is it not going to be any faster; they’re going to end up competing with each other for budget, attention, and probably materials and qualified vendors. Performing development in parallel paths is an approach to take with developing new technologies in order to assess which contractor is making the most progress or to avoid having a single decision point result in a show stopping limitation (which is why this was done on the Manhattan Project). But it isn’t the case that we don’t know the general steps to develop such a system; we just have to go through those steps in sequence, and that takes a finite amount of time. Even if we cut out the supposedly unnecessary (according to some) steps of system requirements analysis and trade studies, component-level development testing, documentation, design reviews, et cetera it is still a massive set of tasks requiring thousands of people to provide products in a certain order. We can’t just chop the problem into infinitesmial bits and have it all crunched out by dinnertime.