NASA lays out how SpaceX will refuel Starships in low-Earth orbit
arstechnica.com
arstechnica.com
[1] Talking about the Artemis program to the American Astronautical Society https://www.youtube.com/watch?v=OoJsPvmFixU
I see Artemis as an attempt to industrialise the process of getting humans and large volumes of materials to the Moon, ultimately as a stepping stone to Mars. This isn't a proof of concept self-contained "mission" like Apollo was. It felt a bit like saying the global aviation system is overengineered compared to the Wright Flyer. I don't see why it really matters if it needs 15, 20 or even 50 reusable vessels for orbital refuelling as long as that process is repeatable and safe.
That said, it's been a few months since I watched that video so I may not have done a good job here of remembering what my criticisms even were.
The shocking thing for me from that video was learning how useless the Lunar Gateway is most of the time due to its near-rectilinear halo orbit. This leads it to be only within range for rescuing those on the moon, for a couple days a week.
Why does Gateway have this type of orbit? Because the Senate Launch System doesn't have the delta-v to get the capsule to circularized low-lunar orbit. We designed a moon rocket, which can't get to the moon.
It's more a logistics problem than anything else. If a Superheavy and Tanker takes 3 weeks to relaunch, you build 3 to launch weekly and 21 to launch daily. Now you need to have 21 processing facilities and 21 times more people to do that. Launch pads, landing pads, and Mechazillas don't need that long to recover, so you can build just a couple extra ones with each one serving a group of processing facilities.
I genuinely never heard about the Gateway being there for rescue until Destin's talk. (Not saying he invented it. Just that it wasn't what it was advertised for.) Compared to just landing a spare Earth ascent vehicle, it seems like a stupid way to solve for our lack of low-gravity trauma medicine experience.
> Why does Gateway have this type of orbit? Because the Senate Launch System doesn't have the delta-v to get the capsule to circularized low-lunar orbit
NRHO also minimises communication black-outs with Earth. With the Gateway in orbit, other craft in circular orbit can maintain continuous contact with the Earth.
But why would you set that as a pre-requisite?
The Gateway is also, as I understand it, a staging ground for testing habitation for a vehicle to Mars.
Well, the argument in his video is that there are multi-day gaps in the ability to rescue astronauts off the moon. Balancing that out with a few very small relay satellites, I would pay for a couple Electron launches for the relay, and have a lifeboat capability 24/7.
> The Gateway is also, as I understand it, a staging ground for testing habitation for a vehicle to Mars.
I have heard that as well, but what I don't understand is how a higher orbit simulates a Mars trip much better. I mean, either way you are not inside a magnetosphere, right? I guess in low-lunar orbit you are more shaded from solar wind by the moon half the time?
Right, which is where he loses me--this is not why we have a Gateway.
> Balancing that out with a few very small relay satellites, I would pay for a couple Electron launches for the relay, and have a lifeboat capability 24/7
Building a communications relay constellation isn't a simple task. It's on the roadmap for Artemis.
> what I don't understand is how a higher orbit simulates a Mars trip much better
The high orbit is irrelevant to this aim. You'd get the same deep-space habitation science from a low, circularised orbit. The high orbit is for creating a one-and-done communication relay.
I am just an everyday armchair astronaut, so please correct me where I'm wrong, but I figured:
1) A lunar lander does not have the delta-v for direct return to earth, from the lunar surface. (or the heat shielding)
(2) Part of the Gateway's function is similar to Apollo's Command Module. A place for the lander to dock upon return, and facilitate return to earth.
I can see how Gateway differs from Apollo's Command Module, in that it does not return to Earth, but stays in lunar orbit. This does test Mars-like endurance, and is cool. Orion will detach from Gateway, and return independently, right? But why the near-rectilinear halo orbit, which puts the lifeboat capability out reach sometimes? If an astronaut needs something like an appendix removed (random example,) then a few days waiting for Gateway to swing by could make a difference.
edit: Oh, near-rectilinear halo orbit also makes it easier for Orion to return to Earth, right? So maybe he said that Orion lacks the delta-v, not SLS. Maybe that helps explain it.
Starship HLS is the lunar lander [1]. That said, it is not designed to survive Earth re-entry at lunar velocities.
> why the near-rectilinear halo orbit, which puts the lifeboat capability out reach
I guess I'm sceptical of any Gateway serving as a lifeboat in an emergency with a day-to-day sensitivity where aid can be provided on Earth but not on the Moon.
> Orion lacks the delta-v, not SLS
Correct. SLS is expensive, Orion underpowered.
Or any velocity. It doesn't have thermal tiles.
Criticize the parts of the system that are both expensive and prevent us from taking advantage of Starship's ability to land 100 tons on the moon. SLS.
They're aiming for a turnaround time of eventually less than a day. I'm not sure how ships are going to be caught and stacked, but even if ground ops took hundreds of people it's still less than a man-year effort as far as cost when it's done in a day.
Fuelling, systems checks, clearing the launch area, launching, fuel transfer, re-entry, recovery, diagnostics, repairs, recertification, snacks for the repair crew, lawyers, drinks for the lawyers, et cetera.
The goal is automating everything. But even the Falcon family is currently around 3 weeks' sustainable turnaround.
They pump tons of fuel per second, obviously automated. Connecting/disconnecting pipes are with that movable platform. Granted, today it's different, and some more operations might be needed, like, say, more careful stage positioning on the launch system after it landed previously, but they don't seem problematic to include, right?
> systems checks,
SpaceX vastly increased number of channels collecting data on rocket performance. They have much better understanding of the rocket state after landing because of that. Some checks could be done on earth, but we surely don't do system checks every time we sit in a car, and this comparison is only somewhat stretched, agree?
> clearing the launch area,
For Super Heavy the landing happens some minutes after launch; don't think there's going to be many people near the pod. The zone below the trajectory will probably be permanently closed, at least for period of intensive pad use, just like it's done for construction projects now.
> launching,
Of course.
> fuel transfer, re-entry, recovery,
All in-flight operations are automated, even today, because there's no crew around. Problems and misses today, but we're just testing.
> diagnostics,
What's the (crucial) difference with system checks? Seems like we're talking about the same goals.
> repairs,
No repairs for the nominally working system. If checks show repairs are needed, the system - rocket stage - goes to the factory for fixes, that's much longer than the time between flights, but we have spares to use while the repairs happen.
> recertification,
What do we need here to do that is costly or takes a long time? How long does recertification take for a commercial airliner between flights?
> snacks for the repair crew,
Well, repairs are supposed to be a relatively exceptional thing, so done in the factory, with full lunches if not dinners.
> lawyers, drinks for the lawyers, et cetera.
All that for customer's money and time, not on a launch pads.
> The goal is automating everything. But even the Falcon family is currently around 3 weeks' sustainable turnaround.
Don't you have a suspicion that Falcons aren't being improved that much now because it doesn't make sense for the current mass flow to orbit and given that Starship is planned to become available soon?
Sure. Easier said than done, particularly with rapid reuse and when all your equipment is being repeatedly thermal and pressure shocked.
> Some checks could be done on earth, but we surely don't do system checks every time we sit in a car, and this comparison is only somewhat stretched, agree?
We don't send cars to space and back again. (We've also had cars for a century.) Hopefully by the time we're sending humans to Mars, this will be automated. But not in time for Artemis.
> zone below the trajectory will probably be permanently closed
You're describing major shipping lanes in the ship-surveillance area. KSC and CC also tends to close off a few popular flight routes into Orlando.
> in-flight operations are automated, even today, because there's no crew around
Still need people monitoring, including--by law--people ready to abort or order redirect.
> What's the (crucial) difference with system checks?
You have to open the rocket up.
> No repairs for the nominally working system
This has never been demonstrated. (Repairs/refurbishment.)
> Don't you have a suspicion that Falcons aren't being improved that much now because it doesn't make sense for the current mass flow to orbit and given that Starship is planned to become available soon?
No. Falcon launches are currently pad limited.
Note: I think Starship is vastly superior to SLS. But that's in part because the ground operations, including infrastructure, are reused and thus amortised across many flights.
the system is unable to get to the orbit, its not tested. Its all high-hopes and conjectures as it is now.
All of those plans are opinions, that's the problem.
Its like planning your infrastructure around hyperloop, but it doesnt exists. So all of the plans are founded on vaporware.
That is highly problematic thinking.
That's more than SLS has managed.
The flight also did a fuel transfer test.
The larger question is whether it will be able to reenter. However 14 fully expendible starships are still cheaper than a single SLS launch.
SLS did get to earth orbit with the Orion payload. While the latest Starship (IFT-3) barely reached "orbital speeds" and failed on re-entry. How is that "more than SLS has managed"?
You could argue that since SLS is not reusable, it kind of also failed on re-entry by design. But the Starship failure was an accident and it has never actually reached orbit, nor has it delivered a meaningful payload successfully. And Starship has ambitions do a lot more than SLS, acting as a full lunar lander and being reusable.
All in all, Starship is in a much earlier stage of development. SLS already fulfilled the purpose it was designed for back in 2022.
SLS burned an absurd amount of money and spent over a decade making sure that the first launch went relatively well, because they can't afford to have visible failures, in part because people are stupid when it comes to testing, and in part because they can only build them fast enough to launch once every other year (unless they're given another ridiculous sum of money and another decade to build more infrastructure).
Starship is testing iteratively, so they spend less money up front to figure out everything before flying, and instead put in a reasonable effort ahead of time and then just fly to figure out what needs improvement. They can afford to do this, because they are designing for mass prduction and because as a private company, they don't have to answer to the ignorant public. Building a new Starship stack is a matter of months, and they're more limited by the rate at which they can test and get through paperwork to launch the things than the rate at which they can build them.
For reference, SLS launched in 2022 and next might launch in 2025, Starship had a close to orbital velocity launch in March (it effectively proved they can get to orbit, it was not actually orbital so as to avoid leaving debris stuck in orbit until they can prove the ability to deorbit) and is expected to launch again in late May/early June.
Indeed SLS is very inefficient, Starship certainly has more potential and its achievements are more impressive already in terms of innovation, speed of development and costs. But it's not nearly as capable yet, there are still many unknowns.
SLS burned a lot of money because it needed to have massive production in politically-needed districts.
It had orbital velocity, but it wasn't on an orbital trajectory. At least based on the numbers provided and Scott Manley's description of events.
I think the NASA award for pumping tonnes of oxidizer required them to be at or above orbital velocity.
You correct. Artemis 1 put the Orion capsule in a distant retrograde lunar orbit [1] before successfully splashing down, all unmanned. I believe this is the same orbit astronauts are going to fly in Artemis 2 in 2025.
I don't think you can compare Hyperloop and Starship, Hyperloop is just a wild bullshit idea, Starship is a far more realistic project and there are some real resources behind its development.
I would view the claims about the cost of reuse with a huge question mark though. SpaceX has been much better than previous efforts in making reuse feasible, but I think only looking at the fuel is likely to be hugely misleading in terms of actual costs.
hindsight 2020, there have been bilion of $ thrown at it. Musk was promoting it as easy thing to do. And yet its "wild bullshit idea" how come reusable rockets is not a "wild bullshit idea"?
I dont understand how people can eat anything up Musk wildly claims and forget all BS the moment it turns out it was BS.
According to his claims, he will have people on the way to mars last year, for sure.
How come people who hired bunch of grannies to knit a nav computer memory got to moon without loosing a single rocket. And now loosing 3 rockets and still not orbiting is an amazing achievement?
> How come people who hired bunch of grannies to knit a nav computer memory got to moon without loosing a single rocket.
NASA did lose a good couple rockets. They also lost the Apollo I and its crew even before it flew, as well as a number of near misses that were recovered by the brilliance of astronauts (Armstrong saved the day TWICE) and ground crews alike. They also lost two shuttles precisely because they couldn't afford bad news from propagating back to the senate.
> And now loosing 3 rockets and still not orbiting is an amazing achievement?
NASA-style development is very risk averse, which is natural, as it's primarily a government agency tasked with keeping the aerospace and rocket industry happy and healthy, and preventing talented engineers and scientists from moving to Europe, India, Russia, or China. Or North Korea.
NASA can't afford to lose three rockets, because each small piece is expensive and exhaustively simulated and tested before being assembled and launched. It does that to advance our theoretical understanding of how to build spacecraft.
SpaceX iterates quickly and cheaply, blows up engines, rockets, tanks, launch pads and anything else, because it makes them cheaply and use them to learn what NASA does by careful simulation and lab testing.
Well :) Saturns were simpler, and more modest by performance. And also tested more on test stands, not like Russians used to do - with more emphasis on flight testing. Look how much more successful the Starship testing program is in comparison to Soviet N-1 flights - and remain assured, N-1 was made by serious professionals.
I don't understand the question. Could it be the fact that a Falcon 9 just landed after its 20th trip to space? Seems pretty reusable to me.
> How come people who hired bunch of grannies to knit a nav computer memory got to moon without loosing a single rocket.
Apollo 1 wasn't the whole rocket, but I'd say the three astronauts who died were worth more than that. How many astronauts has SpaceX lost? ETA: Not one of those first stages returned to earth intact, so in that sense, all the rockets that launched were lost. ETAA: Apollo 1 wasn't on a Saturn V, so my initial response doesn't really follow.
Is that, I have no arguments so I call people names reply? There are so many of the posts using it like a 'I win ticket'
Its just a stupid Ad hominem attack.
The stuff i can make up and call it probable future that never materializes but then nobody remembers it as I move on to another grift, facts wink
All Musk weirdness aside - SpaceX revolutionized the launch business for the planet, and for quite some time already, so it makes sense to watch what they're doing. Results tend to follow, even though many are delayed - but delayed only in comparison to announcements, not to how it's done elsewhere.
It couldn’t even get its attitude right for rentry.
> the system is unable to get to the orbit, its not tested.
Yes. Just like every other rocket that is under development.
> Its all high-hopes and conjectures as it is now.
Engineering plans from the company that built the most reliable rocket in human history while pioneering first-stage reusability are "hopes" and "conjectures". They're just spitballing, you guys.
> All of those plans are opinions, that's the problem.
Totally agree. Engineering is more of an feeling than a science.
What's the problem here which is going to stop the progress on the Starship program?
And having many launches if you actually get a lot out of it isn't really a problem.
His whole approch seem to be 'this more complex then apollo' and therefore bad.
He spent a lot of time going after SpaceX but spent almost non going after the real problem with the moon program. The incredible expensive lagecy technology forced by congrsss that is dragging the whole programs potential down.
So I think in totalility message is really just reinforcing the Johnson center ultra conservative 'money is irrelevant' approch to engineering.
Maybe it's somewhat outdated point of view - it's about 10 years old. But, if you think in general, you might still agree that for complexities above certain point robots could become very, very expensive.
Also, we shouldn't forget that human flights became vastly cheaper than they used to be.
A lot of the manned space stuff - even a lot of the unmanned space stuff - is someone deciding that we need to use a hammer to make something, and therefore making nails part of every plan they have. Even if the job is inflating a balloon.
> A lot of the manned space stuff - even a lot of the unmanned space stuff - is someone deciding that we need to use a hammer to make something, and therefore making nails part of every plan they have.
Used to be more, now it's less and less, with decisions being less centralized. In USSR space science initially was discussed on the level of national academy, the highest existing. Now a small group of students can invent a mission to study a particular aspect and get to results.
I do agree that we need to do small steps before big steps. But overall this "lets save money with reusing Shuttle stack" has backfired spectacularly. And delayed both small step and bigger step for decade(s).
There are real questions about LEO versus deep space habitation, specifically in respect of solar radiation. Gateway will help settle those questions.
> this "lets save money with reusing Shuttle stack" has backfired spectacularly
100% agree. And I honestly don't see us using SLS and Orion to go to Mars. But Gateway will be proving real science and hardware. And SLS, Orion and even the Gateway are more easily deprecated from the Artemis framework than e.g. Starship.
Could we have a better system? Sure. Is there a better path forward if we want boots on Mars in the 2030s? No.
Those kind of missions are complex tasks and complicating things make them more complex and error prone exponentially
Again, the criticisms are valid. But it's unlikely we can get to Mars--repeatedly and sustainably--without mastering in-orbit re-fuelling. So while it isn't technically necessary for a Moon mission, nor even repeated Moon missions, it is for a Moon mission that's building to Mars. (Again, repeatedly--we can do without it [1]. But it's far more expensive in the long run.)
> How long would it take to re-certify starship for a launch?
Falcon 9 has been turned around in 3 weeks [2]. The bottleneck has thus far been launch pads. If you can cut that to 2 weeks, you need 5 to 7 refuelling Starships ready per launch. Once again, the bottleneck would be launch pads.
You and I are on the same page about that re-fuelling overhead being bonkers. It's mostly because SLS and Orion suck. But if the only thing we learn from Artemis is fuel depoting, that's a huge win.
> You are going to need dozens of them launching everyday for months to prep one mission and hoping nothing will go bad
The timeline for the first propellant-transfer test is three to four weeks. 14 launches over 30 days comes to a launch every other day. Ambitious, but only about twice as frequently as the Falcon family launched in 2023, and for a shorter interval.
[1] https://en.wikipedia.org/wiki/Mars_Direct#Second_launch
[2] https://twitter.com/SpaceX/status/1520155013661659137?ref_sr...
Not sure I understand your numbers here in relation to the quoted part?
The unknown variable is in-orbit propellant boil-off. How much propellant will poof away for each day in orbit, and during the propellant transfer. The test next month will observe this for 3 to 4 weeks. So for the actual mission, we will presumably have 3 to 4 weeks to put all the fuel in orbit. Assuming it takes 14 launches to launch that fuel (we won't know the final number until we do the tests), that gives us a required average launch cadence of about 2 days.
It should be, but we haven't done it yet [1].
> active cooling systems can be used, they don't need to be big or complex
Cooling in space is always complicated given the only way to dump heat is radiating (or dumping into ejected mass, i.e. boil off).
[1] https://en.wikipedia.org/wiki/Orbital_propellant_depot#Heavy...
Yes, but there were some calculations for this - http://astronautix.com/n/n1blocksr.html - which hints on boiling off to be manageable for large stages.
Re-certify for flight? I can envision an automated procedure which would analyze the data obtained from the rocket in flight and after flight which makes a decision in a split second, just like some systems in cars routinely do.
Hoping nothing will go bad? You can get much more risks on unmanned missions. You'll also get numbers on your side, when if something goes wrong you fix it, systematically, for future attempts. That's not exactly just hoping, right?
Except there isn't any data to back that up. It took off empty this last time, and didn't have enough fuel to complete it's mission.
There's seemingly no physical - or economical, for that matter - reason for this to be impossible, so why not? Starship definitely didn't complete testing at this point, many things are still ahead of us, but to talk about impossibilities - why? Yes, of course we didn't do that yet, but we're talking about possibilities in the future.
"By this point". Well, we have testing program here (which to me looks pretty successful so far), the capabilities of the system - demonstrated, that is - are increased with each flight. Load test - that is, bringing to orbit the full payload the Starship is designed for - has not happened yet, and, as far as I remember, Falcon also didn't have such a test for some record mass. I'm sure we'll see some reasonable payload flights at some point "soon". What's the problem here?
> Especially considering that they should be long past that point according to their last updated roadmap.
Ah :) maybe you're talking about the thing which SpaceX refers jokingly as "we transform impossible things to late". Yes, "Musk's time" is always very aggressive, and I even suspect to know why he's, as an engineer and a visioner, does that. And why Shotwell's estimates are more accurate - she excels at operations, after all.
But I don't see a principal problem here either. Yes, the - rather aggressive, by aerospace standards - schedule slips to the right, while still producing results, years in a row. Big deal.
https://www.nasa.gov/news-release/nasa-publishes-artemis-pla...
I suspect this isn’t happening this year and this article is only a few years old. On traditional launches in the past, at least a 25% load is put in the hold, even on the first launch. You generally want to show off that you can at least lift off the ground, unless you can’t.
Everything on Artemis is late; Starship is less late than pretty much everything else.
Would note that to date, Artemis has been solely delayed by Orion not being able to get its shit together.
It was an intentional decision not to go into orbit.
Well, peak ~4% of US budget, not GDP, in 1966. A lot, but not that much.
I think the average figure is 2.5% yearly for 10 years, so that would be as if it were 25% in one year for the entire programme.
But it’s kinda hard to define, since it also grew GDP tremendously (in the US). It may have been a much higher percentage with 1945 costs measured against 1939 GDP, if that makes sense, but I’m not sure it’s appropriate to make that comparison.
It's like asking, who, Miss Universe what her opinion on the US Space Program is.
However, IIRC he has an aerospace masters, is currently pursuing a PhD and used to work on testing missiles, so he's not a complete rando.
Sure, the debates aren't fully settled yet, but they're certainly comprehensive enough to make it obvious that the criticisms are kind of weak, such that even if Starship isn't the one to establish all the necessary tech for meeting Artemis' stated goals, it's by far the most promising attempt to do it, as the others weren't even trying.
Starship is in testing, which was chosen to be done with test flights. Those test flights are of course something very different than assumed regular launches later on, after the system would be complete in design, development and testing. Right now Starship isn't even planned to immediately come back to Mechazilla. Starship is obviously incomplete now, but that doesn't mean it can't be completed and successfully used.
Comparing to that the existing plans and results with NASA's Artemis parts are more bleak. Expensive, slow, underperforming - and even dangerous. Smarter Every Day's video points to gaping problems, and he's even explaining why he chose to talk about it - there and then. It shouldn't matter much that he's an engineer - though he couldn't perhaps to talk about it without enough understanding. In his case, it was more important to articulate important points aloud.
Some indicators are that starship's payload capacity keeps getting revised down, and the launch system still needing weeks of repairs between launches, which means even pessimistic estimates of the number of tanker launches and their cadence are probably too optimistic.
> iterative engineering approach to something and complex and full of hard dependencies as Starship will achieve rapid reusability sometime between much later than promised and never
To the opposite, you can't have a complex system working without having first simpler system working, and then iterating from that. Let me know if you need the quote. For rockets in particular, the approach "build a little, test a little" is the current best practice. Starship complexity of course is splittable into engine complexity, set of engines, body, simple control and then other functions - actually, with the exception of Raptors, it's quite an iterateable system.
> This is a project with an unknowable worst-case completion date because there are way too many issues with too many unknowable times to resolve issues, many of which can only be worked on once previous issues are resolved, and way too many issues continue to be discovered.
No, this is totally wrong. And the current pace of existing testing shows just that.
"unknowable worst-case completion date" - strictly speaking, you can never be sure 100% about anything, and if you understand this statement "reasonably", then it's just wrong. Starship will be flying more or less successfully in 3 years, that's rather conservative, and in this decade, very conservatively. Otherwise it will probably be dropped, but chances for that are really low.
"there are way too many issues with too many unknowable times to resolve issues" - no, there are just several rather well-known areas - like, attitude control on the orbit (done with somewhat smaller orbiting bodies), thermal protection during return (done with similar materials, spaceplanes, so both materials and aerodynamics are not completely new), active bracking with engines (done extensively with Falcon-9 first stages), getting caught with Mechazilla (fine controls for Falcon-9 landings and even helicopter catches of payloads from space are relevant) and then quite a number of progressively smaller issues. Will work no problem in the end, high probability.
"many of which can only be worked on once previous issues are resolved" - not exactly. While you can't return with orbital speed unless you got that speed in the first place, other issues could be tested - and Starship flights to 10 km altitude a few years ago were just that, tests for the last 10 km of flight. Not exactly identical to how it's going to happen according to today's plans, but not irrelevant either.
"way too many issues continue to be discovered" - why do you think that? Rockets are something which was done with 50+ years old technology, which for this area is quite a bit; there's just not too much of complexity here. "Rocket science is not a rocket science", as they say in the industry.
"Converging on completion is not in sight." - again, where did you get this from? You're way off in your estimates.
It's not known about number of tankers needed to refuel one Starship. But this is not a particularly important question. Even with 30 tankers required the system still could be made to work. And likely there will be much less of tanker flights than that, so what's the problem?
I wonder what would happen if this Hollywood star - https://en.wikipedia.org/wiki/Hedy_Lamarr - would be asked about, say, spread spectrum communications. Or a certain patent bureau clerk - about the world-filling ether.
Edit: aha, found - "Then, SpaceX will fine-tune tank pressures and fire propellant settling thrusters." And they are using free flow. Wonder how long the transfer will take and how much of a propellant settling acceleration is needed?.. Also :) the relative positions of ships become important if no pumps are used...
Maybe not in all variants, but definitely in the crew version, as having the ability to transfer propellant even if the pressure differences and settling cannot be achieved seems worth the lost performance.
Pumps are way more complicated than pressure feeding.
It shouldn’t if they’re using pressure to feed propellant. Do you mean relative to the Sun in respect of boil off?
That's how I'm reading it, analogous to a pressure-fed engine [1]. A separate, expendable gas supply pressurises the donor tank to force the propellant out and over.
> A separate, expendable gas supply pressurises the donor tank to force the propellant out and over.
But what if the fuel is a mix of liquid bubbles and the gas freely floating in the tank?
(BTW, on the different point, some pumps applicable here could be really simple, like analogous to https://en.wikipedia.org/wiki/Vacuum_ejector#Water_aspirator)
I believe the plan is to impart a gentle spin.
> some pumps applicable here could be really simple
The Venturi effect reduces pressure. It's definitely worth exploring. But you'd need to test it to explore where it causes volumetric effects, e.g. by exacerbating boil off.
That would be something new. And for a partially refueled stage, that's a lot of mass to spin. How fast? the speed of spinning could be important too.
No clue! But I think they were targeting mili-g acceleration--could probably be achieved by the boil off alone.
Now we just need the Attenborough voiceover about the mating habits of starships as they spin through space.
That would send fluids to the back of the tanks.
At this stage it might be easier to build fueling station in the orbit with actual transfer pumps.
As for spinning two Starships of significantly different mass... that should be interesting to see. Doable, yes :) but interesting nevertheless.
Pushing the liquids is not the hard part. Settling the liquids in 0g is the hard part.
But unless I am mistaken, the biggest problem with fuel is that it obviously costs fuel to get fuel up. It has a weight so its not like you can just say, I want 100 lbs of extra fuel, you need that amount of fuel plus the amount of fuel to carry that amount (and then the amount to carry that amount, and so on and so forth).
So in reality what is the difference between say, 100lbs of fuel across 10 Starships vs 100 lbs of fuel on a single ship?
Yeah you would need a bigger ship to handle all of the fuel necessary, but something about this being necessary for the moon seems off? Mars I can understand more.
I just feel like I am missing something here, until reading this I did not realize that refuel in space was necessary for the Moon.
Also to be clear I am not discounting the work on doing this, but I was just always under the impression that this was a thing that was necessary for Mars where now it seems like we are talking 11 launches to get too the moon?
Hopefully you can see how that's kind of unrealistic and unscalable to do in one launch.
Orbital refueling is a core technology necessary for sending useful amounts of payload around the solar system because the rockets themselves can't really get much larger without becoming completely uneconomical. Congress understood that orbital refueling would make their obscenely expensive SLS program unnecessary, which is why they effectively banned NASA from using the word "depot" (until Senator Shelby retired) under threat of canceling the entire space technology program, and strongarmed ULA into not pursuing an upper stage capable of carrying Orion and performing orbital refueling (which again, would've replaced a $2B per launch rocket with a couple of launches of a ~$150M rocket).
The number of launches sounds like a lot, but consider that F9 is already launching 2-3 times each week, and Starship is being designed explicitly to not have the factors that are limiting F9 cadence from increasing further. The engines are better designed for reuse, the launch infrastructure is easier to build many copies of, the boosters always come back to land, the second stage integrates onto the booster at the launch pad etc.
Also, as a side note, they'll need roughly the same number of refueling flights for Mars. The fuel requirements for going to the Lunar surface and Martian surface are actually pretty similar, since on Mars you can save a lot on using the atmosphere to slow down, while you can't on the Moon.
Exponential.
> something about this being necessary for the moon seems off? Mars I can understand more
It's not necessary for the Moon. The reason we're doing it is so once Artemis is done on the Moon, we have a system ready to send astronauts to Mars.
[1] https://en.wikipedia.org/wiki/Tsiolkovsky_rocket_equation#Ex...
The first paragraph:
> ambitious refueling demonstration, a technical feat that will put the Moon within reach.
Again here:
> By the time NASA and SpaceX are ready for a lunar landing, multiple Starship tankers will take off from at least two launch pads to aggregate propellants at a depot to supply the Moon-bound Starship.
Then again here:
> On future Starship missions targeting destinations in deep space, like the Moon, this docking maneuver and refueling process will be repeated multiple times to give the vehicle enough gas to boost itself out of Earth orbit.
Another one:
> The results of this refueling demonstration will allow SpaceX and NASA engineers to calculate how many refueling tankers they will need to fill up a Starship heading for the Moon.
All of this makes it seem like this is necessary just to get to the moon with Starship which really makes me wonder about the practicality of this when later it mentions the possibility of 10 refueling missions.
Or am I misreading what this is saying?
The Artemis architecture requires in-orbit refuelling to go anywhere interesting. (More due to the limitations of SLS and Orion than Starship, but I digress.) The Artemis architecure isn't necessary to go to the moon--Apollo had no in-orbit refuelling.
We've chosen an architecture that requires in-orbit refuelling to go anywhere because it better scales to going to Mars.
The point of refueling is getting to orbit and then topping up all the fuel you lost getting there (i.e. most of the fuel because getting out of the atmosphere takes a lot of energy). Starship when it starts flying will be by far the largest amount of mass we can get to orbit. Like upto 150 tonnes or whatever the maximum payload is going to be. So, that amount of fuel would enable large amounts of mass to be accelerated to a path to the Moon, Mars, or wherever way beyond what would otherwise be feasible.
And with a reusable launch vehicle, it actually gets doable from a price point of view as well.
This is the same goal with SpaceX. To drive the price of ton to orbit down a lot. That is the only way a city on Mars happens.
And how much is that going to pollute our atmosphere?
CBC made this nice interactive graph highlighting the issue.
https://newsinteractives.cbc.ca/features/2023/rocket-polluti...
I'm a huge proponent of space flight, but my main concern is how the drive towards commoditization, again, seems to gloss over potential hidden costs with a complex, global impact.
And it emits about the same amount of CO2 as a trans-pacific jet flight.
But then it's much less of passenger * mile with the given definition.