Worst Haiku ever.
I have a short list of people I would volunteer to go to Mars. Available upon request.
If it happened all at once, sure. Selling over the course of many years wouldn’t be a problem. Especially since he’s already said several times that he plans to do this.
A decent chunk of that $172B is in SpaceX itself (worth $74B, of which Musk owns 54%). They’re private and not prone to as wild swings as public companies.
The remarkable thing about social media is that we can link to what people actually say. For instance:
They won’t let a ship go to sea unless there’s enough lifeboats for everyone on board, and you think they’ll let a ship go to space without a way for everyone to come home?
Of course they will. Because it’s not a common carrier. And because a lifeboat for a spaceship is impractical.
Airliners don’t carry spare wings. If one of the ones they have fail, you’re screwed. Why aren’t airliners required to carry spare wings? The darn things do fail from time to time. Because it’s excessively impractical to do so.
Space exploration, and space tourism, are not at the technological level where we can demand “8 9s” reliability like we do for wings. So some much lesser standard is mandated and the appropriate regulators try to ensure the engineers live up to the mandate. But very clearly the mandate, whatever it is, is very highly informed by what the engineers say they can accomplish and what the businessdudes say they can afford to have their engineers accomplish.
If I qualified for a Mars mission, I’d volunteer in a minute. Even if it were 100% one-way.
It sounds like all of the knowledge you have about how architecture studies are done and development costs has come from Elon and/or space enthusiasts who assume that NASA and “Old Space” are some kind of intrinsic evil. If I’m wrong about that you can illustrate your experience for me, but claiming that “complex of Congress, NASA, and traditional aerospace contractors is explicitly designed to make things as expensive as possible” might be a legitimate (if somewhat hyperbolic) criticism of the acquisition process, but that that is explicitly not what an architecture study does. In fact, the costs that come out of an architecture study are actually just secondary figures which are well understood to be only rough estimates specifically because of the uncertainties in both the costs of developing and maturing nascent technologies and also the likelihood that a detailed system requirements analysis will indicate a need to increase the scope of some part of the architecture, e.g. the payload you assumed would be carried to orbit in six launches will actually require nine.
The idea that “NASA’s cost models aren’t based on any underlying economic facts” is a nonsense statement. It is true that historical baselines are used to validate individual models of things like engine or thermal protection system development, but they are fundamentally based on the estimate of how much labor and testing cost goes into developments and explicitly include estimates of labor rates, materials costs, facilities modification and maintenance, transportation, et cetera, and are reviewed by an independent review team that draws both from industry (typically retired ‘greybeards’ with no fiscal interest in the project), academic scientists and engineers, and finance people familiar with this type of project.
Part of the problem of trying to make a refined estimate of an architecture for sending a crewed mission to Mars is that a large portion of the architecture is dependent upon technologies that are at a low state of technological maturity, generally evaluated in terms of Technology Readiness Level (TRL). Because those technologies require significant development before they are even ready to go into a qualification or certification process which would be required for any high criticality (e.g. crewed) application. This is actually one of the issues that adjunct studies like the ones I worked on are intended to address by taking a more detailed look at the assumptions that went into a segment of the system and assessing whether they are really valid and/or how much work would really be required to even get to a point of being ready for application.
Because of how much a crewed Mars mission is beyond current experience in crewed space exploration most of any potential architecture (even a “Mars Direct” program hypothetically using existing technology) most of the segments of the architecture are at a relatively low level. In fact, except for the Earth-to-Orbit segment, everything else on the architectures we looked at evaluated out at TRL 4 or lower. The idea that you can somehow maintain costs at some hypothetical minimum baseline while doing fundamental technology development is not realistic; nor is the notion that many ‘space enthusiasts’ have that the crew will somehow MacGyver their way through problems throughout the mission.
By the way, regarding the disparity between SpaceX reported development cost and the Department of Defense (not NASA) estimate using the NAFCOM costing model, it should be noted that the Merlin engine design did not come out of nowhere; it started as very much a heritage of the TRW Low Cost Pintle Engine (so much that SpaceX attempted to sue Northrup Grumman for ‘stealing’ engine design information that turned out to have gone the other way), although it has gone on to deviate pretty far from that engine family. It is fair to note that SpaceX did manage to develop the entire vehicle for less cost than the ULA contractors historically have, but there are a couple of reasons for that; one is that SpaceX has been able to be certified for EELV under FAR Part 12 (commercial acquisition) which has dramatically reduced requirements for unit qualification and data deliveries to the government than the FAR Part 15 development that vehicles like Atlas V and Delta IV were developed under. The other is that those historical models have been largely based upon ICBM-type development where there is a premium on reliability, performance, and verifiability are considered key figures of merit versus a more balanced tradeoff of cost versus these metrics.
You can argue that conventional development costs are too high, and in fact I also worked on a study that indicated that the ULA EELV launches should only cost ~60% of the contract costs including full mission assurance, and our work was vindicated when ULA CEO Tory Bruno actually admitted they could launch for about half of the cost after SpaceX won EELV contracts and started eating ULA’s lunch. But, again, the Earth-to-orbit is the easiest and most develop segment of any crewed interplanetary architecture, and success in that area does not translate in being able to develop all of the necessary technologies on the cheap.
There is literally no basis for that assertion other than “Elon said so”, and like the vast majority of things Elon says, the source is PDOOMA. Saying, “ISRU is the long pole” really understates the issue; in fact, to date no one has extracted so much as a single drop of water from an extraterrestrial source, and manufacturing propellants (even one as simple as methane and oxygen) are technologies that would require a decade or more of concerted effort to get to a maturity level to not only have reasonable confidence that they work but are sufficiently reliable to pin the safety of a crewed mission upon them.
The idea that “full second-[stage] reuse on Earth” is somehow going to result in massive reductions in the overall mission costs misses the essential fact that that segment of the architecture is in no way a major cost driver. In fact, even assuming that a hypothetical conjunction-class crewed Mars mission is using a really expensive launch system like an upscaled Delta IV, the cost still only works out to be two or three percent of the overall estimated mission cost, and while saving a few billion dollars isn’t an opportunity to pass up it is basically a rounding error compared to the unknowns in other segments of the architecture for which the cost estimates are just optimistic guesswork. The reality is that we don’t actually have better than a rough order of magnitude guess as what the ultimate cost of a crewed Mars mission would be, and that US$500B is almost certainly a significant underestimate once all of the actual development costs get folded in.
As an aside, one of the curious things you learn in doing studies like these (and playing around with ‘toy models’ of space launch vehicles) is that designing a vehicle to go to orbit isn’t actually all that difficult; it’s not Kerbal Space Program easy, but provided you aren’t trying to achieve maximum theoretical performance and are prepared to trade off larger size for operational efficiencies you can come up with a concept that ought to be capable of orbital flight (and for the most part, they look like variations of a ‘Big Dumb Booster’ or Philip Bono’s big SSTO and stage-and-a-half concepts because that is where the trade study drives you). The technical problems are really being able to build the vehicle, transport it, and operate it with sufficient reliability, but the fundamental stumbling block is just being able to do it in a way that recoups your costs. Really giant rockets–like the Nova, the rocket that was to follow the Saturn V, or Bob Truax’s Sea Dragon–have never been built, not because they aren’t feasible but because there just isn’t a commercial or operational need for them. You can build a giant rocket with a 500 ton payload capacity, but good luck figuring how to transport it on the ground or get someone to pay you to carry some hypothetical giant payload over and over to the point that you can pay back your investors and keep your suppliers from sending you notices that if you don’t pay the invoice they’ll stop delivering parts.
O’Neill was big on solar power satellites being microwave frequency power down to ground states of hundreds of square kilometers each, which was supposed to be the industry that would put space habitats on a fiscally sustainable. The problem with this is that solar power satellites are fundamentally just collectors of a resource that is already available on the surface of the Earth. It is true that they can get the full 1360 W/m2 at Earth orbit full time instead of a fraction of that for a peak at a few hours out of the day, but once you start looking at the inefficiencies in the system and the costs of maintaining both the ground stations and satellites, it just makes more sense to collect solar energy where it is needed on the ground, and this was before the more recent drop in highly efficient PV solar panels.
Solar power satellites might make sense in terms of providing a more consistent base loading but there is no reasonable estimate of the costs that is favorable even compared to nuclear fission much less natural gas. There was a spurt of investment in this concept in the mid-Nineties during the second commercial aerospace boom, and then another one in the mid-to-late 2000s so a fairly large amount of capital went to startup companies promising to demonstrate a proof of concept but thus far nobody has ‘beamed’ more than a really powerful ham radio from space, and the space junkies have mostly moved on to helium-3 as the golden goose that will make lunar bases viable under the mistaken notion that if we were able to extract it (at a gram to several thousand tons of regolith) and bring it to Earth that we would have magical nuclear fusion, apparently unaware that 3He fusion is even orders of magnitude more difficult than the D-T fusion that we are still perpetually twenty years from mastering.
A handful of people willing to pay millions for a space flight makes for great press but is not a basis for a sustainable industry where development and ongoing operating costs run into nine or ten figures. I know a fair amount about Blue Origin and while the space tourism angle is what they are selling to be public behind the curtains they are working on much larger orbital vehicles. Virgin Galactic, on the other hand, is intrinsically limited to suborbital hops with a few minutes technically “in space”, and the idea that there are enough people willing to pay tens of thousands of dollars to recoup the long development costs and delays due to flight failures isn’t really rational. It exists because it is being backed by a flamboyant billionaire entrepreneur who doesn’t actually know anything about space flight, and once the novelty of it wears off for the small cohort of people willing to shell out the costs for a flight it doesn’t really have a path to profitability. And this is the general problem with “space tourism”; the assumption that you can build a tourism industry that leads technology and infrastructure has resulted in one ambitious project after another going bankrupt and abandoned, and that is for projects on Earth (or undersea) where there is established transportation infrastructure and necessary resources.
Stranger
You neglected the other stuff about the multi-billion dollar tourism and recreational industries that exist and thrive to serve the very rich. As I mentioned, the Supercar and Superyacht businesses are worth billions annually, and there is a thriving market for ultra-expensive vacations.
For example, Richard Branson has a private island in the Caymans that you can stay on - for $62,500 per day. I was in St. Baarts a few years ago, and the area offshore was festooned with giant yachts, the cheapest of which would have been several million dollars, and the most expensive, the Eclipse, is worth 1.5 billion dollars.
If you could offer people a luxury suite on the Moon, there are tens of thousands of people who could afford to pay $200 million to stay in it for a couple of weeks. And thousands who probably would. A guy like Robert Knok could afford it easily - he just paid 4.8 BILLION dollars for a new yacht.
The Moon would be the ultimate tourist destination. Hell, I could see Arab Shieks actually paying to help build out the place for the prestige.
No, it’s based on pretty straightforward things we know about what things cost in general, how much money SpaceX has, and basic physics.
The basic Starship architecture for sending stuff to Mars consists of launching to LEO, refueling there, and then heading to Mars. We know it’ll take somewhere around 10-12 launches for the refueling.
Stating with some assumptions:
- Can a vacuum Raptor engine achieve ~380 s Isp? Yes, this is a reasonable number based on full-flow staged combustion and methane propellant.
- How about sea level Isp of 330 s? Again, yes.
- The Starship dry mass needs to be around 120 tons. This is harder to be sure about since it depends very much on the internals. But if you do some napkin math with stainless steel and the few-millimeter thickness they need, it seems plausible.
Working backwards from Mars: is it reasonable for a Starship to land there? Sure, if they can pull off the bellyflop. 1000 m/s is probably around what they’ll need for the final powered landing assuming aerobraking is effective. 70 t of propellant would be sufficient for 120 t dry mass and 100 t cargo.
How about simply getting to Mars? A fully-fuelled Starship has around 6.9 km/s of delta V. We need 1 km/s for the landing, leaving 5.9 km/s for departure and insertion. A typical transfer needs ~3.9 km/s, so it would seem there’s enough margin.
So, how do we get a fully fuelled Starship in LEO? If the first launch leaves the tanks dry, we’d need another 12 launches at 100 t each to refuel. That’s probably pessimistic since the refueling will use a dedicated tanker with a better mass fraction, but let’s go with 13 to be conservative.
So can a full-stack Starship make orbit with 100 t? Again, this seems plausible. The second stage will need to reserve some propellant to land again, so we won’t get a full 6.9 km/s out of it. However, it probably won’t need to reserve much, since it sheds the majority of velocity in the atmosphere. Also, it doesn’t land with cargo. Let’s say 6.0 km/s is achievable.
That leaves around 3.4 km/s of work to do for the first stage, considering gravity and aero losses. Plugging some conservative numbers in, the first stage can get around 3.9 km/s delta V, leaving 500 m/s for the landing. Again, plausible.
What we have so far is a system that can land 100 tons on Mars. It requires around 13 launches of Starship. What does it cost?
We aren’t talking yet about ISRU or anything, so we’ve expended one upper stage by stranding it on Mars. It needs some structure, some engines, avionics, etc. But we’ve seen that the structure is pretty cheap, as are the engines. They couldn’t be that expensive, because we know more or less how much money SpaceX has, and they can’t afford to have spent billions on the program so far. The engines can’t cost more than a few million each, and the structure again in the millions (just consider how much they’ve blown up already). Avionics can’t be that different from the Falcon 9 upper stage, and that costs <$10M total. I think $100M for an upper stage is a pretty pessimistic estimate (I’m excluding life support, since for now I’m just talking about cargo).
The booster will cost significantly more, but will also be more reusable. SpaceX has already more or less mastered first-stage recovery, and there’s little reason to believe the Superheavy booster will be any different. The marginal cost should be a few million at most, if the booster lasts >100 launches.
Propellant? Basically free. Again, easy to do the math and it comes out to <$1M per launch.
So we have a $100M upper stage, maybe $10M amortized cost per launch for the refueling trips, and I’ll add another $10M per to run their spaceport, refurbishment and other incidentals. That comes to $360M total to land 100 t on Mars, or $3600/kg.
I think these numbers are pretty conservative, and that with a bit of work might bring it down even further. But $3.6k/kg is a pretty good start.
Again, you are assuming that all of the cost of the mission is in propulsion, and everything else is just noise. This is fundamentally incorrect. I would detail this out but you seem to be convinced in the authority of your own beliefs, and I’ve long understood the futility of factual argument over conviction and faith.
Stranger
Well, space tourism might be big, but obviously the main industry for a lunar settlement will be ZAFO.
You italicized the part of my post specifically talking about launch+landing costs, and so that’s what I responded to. And Musk has said almost nothing about the cost of the landed equipment anyway, so there’s really nothing to say on that front. We obviously agree that ISRU is a difficult problem. However, it’s made less difficult if launch is cheap.
Considering just the power generation alone, you need thousands of square meters of solar panels to generate enough power to produce the propellant for the trip back within a couple of years. But how much does that weigh? It makes a difference if we are talking ultra-thin-film panels at a low TRL level, or something closer to what we can already produce easily. How about the deployment mechanism? Do you spend mass on a robust boom system, or somehow deploy sheets between more widely-spaced anchors? The latter might also require some advanced robotics as compared to a simple motorized cable system. Etc.
So if your budget for the power supply is a few billion, it matters greatly if it has to mass 100 t vs. 1000 t. 100 t might require a multi-billion development program all by itself, while 1000 t might be relatively easy.
That’s just one aspect, of course. There are some problems for which the difficulty is not so dependent on the mass. But it is often the case that you can trade mass for development and production cost.
If you really are optimizing for cost, then the marginal cost of mass-reducing a component in the system will be the same as the launch cost. If launch costs $50k/kg, then you will spend $49,999 to use a more expensive material or redesign the part or whatever if it shaves off a measly 1 kg. But if launch costs $3.6k/kg, then there are fewer opportunities to make that kind of trade, and so it brings down the cost of the entire thing.
I’m in between the two of you. The details can be killer, and we’re glossing over a lot of them. Just deploying the solar panels you mention would be a very difficult task that will take a huge amount of engineering. Trying to manufacture cryogenic fuels and pump them into tanks on a ship on Mars is a huge undertaking. Musk hasn’t even addressed how to keep people alive for a year and a half on Mars, and our track record with long-term enclosed habitats is not good.
There are many things we haven’t even begun to price such as the cost of money, launch facilities costs, insurance, depreciation… The systems costs for sending people to Mars (air, water, food, radiation shielding, heating, cooling, elevators, yada yada) will be expensive and Musk hasn’t even started building that stuff out yet.
On-orbit refueling has never been done before. It will be hard to test on Earth be ause the space environment is so different. And if uou have to refuel a dozen times before youngo anywhere, that entire process has to be incredibly reliable. If it takes 12 launches at $20 million per launch, that’s a quarter of a billion dollars to put a fully fueled Starship in orbit. SpaceX can’t afford to lose many fully fueled Starships, and their landing tests on Mars will require one and if it fails it’s a year and a half until the next attempt, while they burn through a lot of fixed costs.
On the other hand, Government numbers like $500 billion for a Mars mission are ridiculous. This is the same organization that’s building a rocket that will cost close to $2 billion per launch, so they aren’t exactly models of cost containment. I’d take their estimates and divide by ten. But that would still be a $50 billion cost, which is almost an order of magnitude more than Elon is claiming.
On-orbit refueling is definitely a hard problem. However, they’ll be in LEO and will have plenty of time to practice. They can put two Starships up there and practice docking, ullage control, pumping, etc. over and over until they are comfortable. No need to come back down, just ping-pong the fluids between two Starships, starting small and working their way up. And they can stay in a relatively low orbit so that if something does go kaboom, it’s unlikely to produce any long-lived debris.
As another example of a mass/cost tradeoff, consider the hydrogen part of ISRU.
The fully-fledged version of Martian ISRU needs water as a feedstock. You split it into H2 and O2, then use the Sabatier reaction to turn it into methane.
But getting the water is probably the single most difficult aspect of ISRU. You have to get it from the poles, or dig underground, or somehow condense it out of low spots with water fog… all difficult. The CO2 processing is very straightforward in comparison. It’s actually being tested as we speak on Perseverance with the MOXIE experiment.
So why not bring your own hydrogen? A Starship with 20 t cargo launching from Mars and landing on Earth needs around 860 t propellant. 190 t of that is CH4, and 47 t of that is hydrogen. So it’s actually not much compared to the total mass.
If mass is cheap, then, just bring your own hydrogen. Then the ISRU unit doesn’t have to deal with mining or anything, just pulling in the atmosphere. In fact, you could built an ISRU-specialized Starship that carries 60 t of hydrogen (bring a little extra for margin), 40 t of machinery, and dumps the result back in the tanks. There’s nothing to unload and it just needs a power input. The explorers can dump the machinery overboard and ride that back, or transfer the propellant and take their original ride (probably a better idea).
Certainly not NASA; of the three parties I mentioned, NASA is the only one to have made any effort to break out of that vicious cycle. And in fact it took only a few key programs (namely, COTS, CRS, and various related programs and offices), done almost as cover, that got SpaceX through their darkest period. But that was the exception, not the rule.
Nor do I think old space is any kind of intrinsic evil or anything; the worst you can say is that you got reliable hardware and paid through the nose for it.
No, the problem is with the system. The interests of the parties are not aligned in a way which drives down prices, unlike typical markets. Congress doesn’t just allocate money to NASA; they dictate which technologies they must use. Because voters are their customer, they choose technologies not for their cost or even suitability for purpose, but because they are made in their home districts.
To the extent that NASA is complicit in the system, it’s because they choose to develop launch capability in-house. Congress is of course all too happy to go along with this for the aforementioned reasons. And for the most part, so is NASA because it earns them a bigger budget even if it doesn’t make sense.
How about big aerospace? One problem is with the so-called revolving door, where the same people circulate between NASA and aerospace companies. It doesn’t take much imagination to think that lowering costs might not be high on their minds. And there is also the problem of cost-plus accounting, which made sense during a war but makes much less sense when you’re just trying to put some mass into orbit or build a generic satellite. It gets even worse when you realize that the people charged with awarding the performance bonuses may be on track to use that revolving door.
Not explicitly. But the cost models are based on what things cost historically, and those costs are affected by the complex as it exists. The NAFCOM analysis of the Falcon 9 actually included a “commercial adjusted” number of $1.7B. It’s not clear to me if they did anything here besides multiply by 45%, but regardless, it’s an implicit admission that the numbers come from a particularly wasteful acquisition process.
Indeed. The Fastrac engine was also a key ancestor, to the point where SpaceX tried to just buy the leftover Fastrac turbopumps. NASA actually agreed but couldn’t provide them quick enough, so SpaceX went to Barber Nichols to get a similar design made.
At any rate, I don’t think our positions here are really as far apart as they might sound. I agree that many of the enabling technologies for a Mars mission are at a fairly low TRL. Where I am more optimistic than you is that I think cheap launch is itself an enabler for reducing development costs. It makes no sense to only spend 3% of your budget on launch. If it costs that little, your architecture is wrong–it means you’ve failed to make a useful trade somewhere. Maybe that is something obvious like using cheaper materials, or maybe it is a bit more indirect like using a technology that’s less efficient but has already been developed. Maybe it means rewriting your mission plan completely, such as my previous example of whether to bring the hydrogen or not.
SpaceX can and does make these trades, and also between risk and cost (they can get away with blowing up test vehicles; NASA less so). Whether that’ll get them to Mars remains to be seen.
here’s a question …why not build an iss type of thing above mars and some sort of shuttlecraft that can go between it and mars … that way whatever needs to be built (if even can be( on mars can be?
and you can run whatever test and experiments that need to be ran ? and once everything sorted out you can start building on mars its self?
This just isn’t right. Polar exploration back when it was truly hard ran a substantial risk of lethal mishaps to the crew of the ships that got them to their starting point, for example.
Your point only goes to quantity not quality. And what is a reasonable quantity of risk is inherently subjective.
They die. Same as happens to any number of people who do risky things and reach a point where they don’t want to be doing it any more but can’t reasonably be saved.
Why? I don’t know why I bother to ask, because I already know you aren’t going to be able to come up with an answer that satisfies me and you won’t agree with me that there is no rational reason. I expect you’re going to say it’s immoral for reasons I very much doubt will make sense to me.
There are two sorts of people in this world, I think.
You may be right about what the government might prevent. But it would not be particularly rational.
MVA’s cause vast amounts of death and injury and the costs are spread across the community to a degree that rationally justifies insisting on everybody taking a degree of care for themselves even against their will. No matter what people might say beforehand, if someone is severely injured in an MVA in a way that would have been prevented by a seatbelt that they chose not to wear, they will nonetheless expect to be assisted afterwards, and it’s human nature to offer that assistance.
The same rational justification for imposing safety on people does not hold true for relatively one-off stunts or risky explorations particularly when there is virtually no chance that, if something goes wrong, any rescue is going to be possible. Let alone one-way trips to Mars
The US government allows stunt people to perform ridiculously dangerous stunts all the time. Recently someone jumped out of a helicopter at 25000 without a parachute into a net. The US government allowed that. You may say that the risks were less (and of course they were) but it’s only a question of degree, not principle.
And conversely they say no to a lot of them as well. My point wasn’t that the government wouldn’t allow it, but rather that it shouldn’t be mystifying if they went that way. None of the stupid thing the government does mystifies me anymore.
Which government agency approves which stunts can be done?