55 years ago, a Saturn V rocket put men on the moon and returned them to Earth after having been fueled up prior to departure - no mid-flight refueling was required.
Artemis 2 looks like once it gets into its initial Earth orbit, it will need to supplied with more fuel from Starship vehicles that will rendezvous with it in orbit. How many Starships? Depending on who you talk to, anywhere from 8 to 20.
Why the need for orbital refueling (and why so many other vehicles to refuel it)? The SLS is 5.7M pounds, and the Saturn V was 6.2M pounds, so comparably sized. Does that missing 500K pounds make the difference between needing refueling and not needing it?
The vehicle they intend to send to the Moon is going to be re-used for multiple missions. So that accounts for it.
It’s a far more ambitious program than Apollo ever attempted. As much of a success as each mission was, there was literally nothing left for the next missions to re-use. Building up in-space infrastructure that we can re-use is a key component of any serious expansion of in-space capabilities.
And if this were actually “building up in-space infrastructure” it might be worth it. However, this system requires a massive number of flights for refueling every flight (the “4 to 8” figure cited by SpaceX as aspirational at best), and does not create any kind of permanent propellant depot or in-situ resource processing. It is not an expandable space infrastructure and in the current form is predicated upon the success and viability of a single commercial entity doe not have a demonstrated a record of on-time delivery of advertised capability.
As for why NASA isn’t just using a version of Apollo/Saturn, the intent of the Artemis program is to land a significant amount of hardware in the Lunar south polar region. The poles are the most difficult part of Earth’s Moon to land on, and the Apollo J-class missions were at the maximum payload capability to land a lightweight rover and a very small habitat with an ascent system. So to land something larger and with more capability—and in particular, enough energy storage or production to remain on the Moon over the ~two week dark period per month, as well as enough mass to protect astronauts from solar charged particle radiation—requires a much larger transport and landing vehicle.
Of course, this begs the question of the value of sending astronauts to the Moon and establishing some kind of outpost at all. It is often brought up that there is substantial ice in the shadowed craters of the lunar poles, which is true enough but it is also literally rock-hard and would take an enormous amount of energy and a fantastical amount of infrastructure to melt. Other resources are speculative at best (don’t get me started with 3He) and would require sifting through or mining below the fine, abrasive, highly charged dust of the lunar regolith which was a significant problem on the Apollo missions and for which there are no broadly applicable solutions to mitigate.
The Apollo missions landed in the ballpark of 2 tons to the lunar surface. Astronauts, suits, rover, tools, etc. Nothing left over for anything like a permanent habitat or anything you’d want for a long-term presence.
Starship will land roughly 100 tons to the surface. Even NASA hasn’t really figured out what to do with all this capacity.
The refueling flights should be thought of as a third stage for Starship. The Saturn V needed three stages, because the extra delta V needed to get to the moon is just a bit too much for two stages to handle efficiently. But refueling has the same effect: it raises the mass ratio of the orbiting stage. So even though Starship is a two-stage vehicle, it acts as three when you include refueling.
It remains to be seen how many refueling flights will be required. The baseline mass to orbit is about 100 tons, and has 1200 tons total propellant capacity. This implies 12 flights, but some will be lost to evaporation unless they can re-chill it. 20 flights is possible but seems very pessimistic.
It could be much lower as well–Starship itself is expected to get upgrades, perhaps to 200 tons, and it may not need a full propellant load. It’ll most likely require several flights, regardless.
But since it’s fully reusable and uses only cheap propellants, they expect the flights themselves to be cheap. Even if they don’t work out second-stage reusability in time for Artemis, flights will probably cost less than $100M (Falcon 9 costs <$30M). So even a full complement of propellant flights will cost less than a single flight of SLS. If they do figure out second-stage reusability, we’re probably talking $10-20M per flight.
SpaceX intends to create a fueling hub in orbit. Tankers will fill the larger depot, so when an actual mission happens it won’t have to wait around to be fueled by six launches or whatever.
Apollo went straight to the Moon, but it was a ‘flags and footprints’ mission that only returned a tiny capsule with three people, and which needed an ultralight lander to pull it off. We need a different architecture if we want to send serious mass to the Moon.
The Starship system is designed to be orbitally refueled, and there was a propellant transfer demo on the last flight.
It seems complex, but intermediate fuel dumps are standard logistics. If we want to seriously do stuff in soace we have to learn to refuel vehicles in space.
Starship is designed for rapid turnarouund and full reusability, so six launches hopefully isn’t a big deal in the future. Starship currently has two launch towers and they are building a third at the Cape, so if they can’t hit a two-hour turnaround on a rocket like Musk hopes, they can still launch at least three per day. There are already four more Starships being prepped for the next flights. Reusability plus a running production line means there will be a growing fleet of the things once they stop blowing up.
Regarding this specifically:
The SLS is not being refueled and the weight has nothing to do with it. Starship is a different vehicle completely, and is what will get refueled.
The SLS carries the Orion capsule to the vicinity of the moon. Orion is not remotely capable of landing on the moon–there are many reasons why, but they basically come down to Apollo having trimmed every ounce of fat from the Lunar Module and Service Module, whereas Orion is rather bloated in comparison.
Astronauts will ride Starship down to the surface and back after rendezvousing. And as said, it’s a massively larger and more capable craft than the Apollo system was. And so it needs the equivalent of a rocket 10x as large as the Saturn V, which is impractical. But you can get the same effect with a vehicle that’s of similar size, but gets refueled 10x times (or whatever the number ends up being).
One advantage to this architecture is that the refueling is the complicated bit, and it can be done entirely in Low Earth Orbit with no one onboard. If something fails or delays the procedure–no big deal, because you just delay the astronaut launch. One that step is complete, the remaining parts are much more straightforward than Apollo was.
Also, having prbital refueling capability will greatly expand whatnwe can do elsewhere in space. Curiosity is the heaviest payload we have sent to Mars, and it was 2200 lbs. A fully fueled Starship can send 100 tons to Mars orbit.
Future missions to the outer planets should not need to make circuitous ‘slingshot’ orbits around other planets. We can build missions that can be even larger and yet fly direct to the outer planets.
A fully fueled Starship can, I believe, put 100 tons into a Hohmann transfer orbit for every body in the solar system except Mercury and Pluto. This will greatly accelerate future planetary science.
More or less, but with a few things added and deleted. Maybe stretched a little, some solar panels added, a cryocooler, some kind of insulation. But also no heat tiles or control flaps, since it’ll never land. It easily gets to orbit because Starship has 100 tons of payload capacity in the first place.
Are they just reusing the now dry starship tanks for fuel storage? Plus maybe more tanks in the payload area? How many tons of fuel does a starship hold in non payload tanks?
We haven’t seen a design yet for the depot, but yes, it uses the dry tanks for storage. After all, it has the same capacity as the craft that’s going to land. Maybe they want a little bit more for margin, but they can stretch the tanks a bit for that. There are a few tens of tons in the header tanks, but I suspect they’ll delete those since they’re only crucial for landing.
I believe the plan is for a stretched Starship to act as a tabker. It will orbit either empty or with a normal payload of fuel, then other Starships will replenish its tanks 100 tons at a time. A full Starship has about 1200 tons of fuel. The confusion over how many flights will be needed is probably due to the uncertainty over how much fuel Starship will need for the moon mission, whether there is boil-off, etc.
This architecture also allows for new generations of space telescopes. JWST’s biggest complication was the origami folding needed to fit it into a standard payload fairing. The mirror was 6.5m. Starship has a 9m diameter, so you could put an 8m mirror on a scope and ship it up in one piece. On-orbit refueling means we could place that scope anywhere we want. Luvoir-class telescopes are now much more feasible and should be substantially cheaper to build.
Yup! Say hello to Luvoir, the telescope I’ve been jonesing for:
It will have either an 8.1m mirror, or 15.1m. The larger version had been killed, but with Starship flying it’s back on the rable. This telescope will make detailed measurements of the atmospheres of many exoplanets.
I would like to see an actual cost breakdown that demonstrates that. Not Elon-ic bombast or Gwynne Shotwell making an offhand claim, but an actual cost breakdown statement that could be audited by a third party financial reviewer.
This was transfer from one tank on the vehicle to another tank, and although SpaceX has been very cagey about the amount of fluid transferred it is clearly not very much given the duration. This is an almost completely difference scale of problem from transferring fluid from one full-sized tank to another on-orbit.
First of all, an 8 meter diameter parabolic reflecting mirror is close to the largest size of unitary optic ever produced for even a terrestrial telescope that is only subject to a constant 1 gee loading and minor thermal fluctuations. Producing an 8 meter diameter mirror (much less one 20 meters in diameter) that could withstand the launch environment (acceleration, shock, vibration) would be an order of magnitude more difficult than just making unfolding mirrors notwithstanding the essential impossibility of testing such a monster to flight-like environments in a terrestrial laboratory. You might even have difficulty getting it into the Space Power Facility for thermal vac testing much less any dynamics. Nor is there actually a benefit to doing so; at some point the additional light gathering area is of marginal utility and it would actually be better to have multiple spacecraft flying in formation to create a long baseline interferometer than trying to build a giant unitary telescope.
But all of this discussion about ‘Starship’ (regardless of the damn-fool logistical nightmare proposed) misses the point of the over impact of the Artemis program in how dogmatically supporting the crewed program in an era where money for science research is being restricted means that a multitude of actually useful and cost-effective uncrewed planetary research programs are being sidelined or cancelled, and experienced staff is being RIF’d en masse in order to pay for Artemis. For instance, the 25 year old Chandra X-Ray Observatory, which has done and continues to provide data for groundbreaking research in high energy astrophysics around the world, is going to be shut down despite the fact that it continues to provide new data for pennies and potential cancellation of Mars Sample Return (which took many years of promotion to get approved) and a multitude of Earth observation missions supporting climate monitoring and model development, all so we can stick a flag at the Lunar South Pole.
Probably not going to happen. We only know some of their finances from various leaks, and very little otherwise. However, if you want someone that’s not Musk or Shotwell, there’s Christopher Couluris (director of vehicle integration):
In a briefing earlier this year, SpaceX director of vehicle integration Christopher Couluris said the company can “bring launches down to below $30 million per launch.”
″[The rocket] costs $28 million to launch it, that’s with everything,” Couluris said, adding that reusing the rockets is what is “bringing the price down.”
That is consistent with what we know. We know that the boosters themselves can be reused >15 times, so the amortized capital cost is very low, assuming it didn’t cost more than $30-40M in the first place.
The second stage probably costs around $10M. Fairings get reused too.
So, if $10M is the cost of an upper stage, and amortization is $3M, then they need refurb and incidentals to come in under $15M to hit the $28M target. That doesn’t seem unreasonable, especially given their demonstrated turnaround time (21 days at a minimum) and obvious high cadence. Not to mention a strong push toward low-cost operations, like horizontal integration and autonomous flight termination.
That article was also from 4 years ago, and they’ve likely optimized more since then.
We do know a little about how much money they bring in from investors. It’s hard to make the numbers work without cheap flights. If you think they’re fudging the numbers on one thing, it just means they’re doing better in some other respect, like the cost to build satellites. No way to prove any of this without access to the financials, but we can say that they have very low costs averaged across the business.
There are still major uncertainties in the Starship program that could severely limit it. For example, if they can’t manage the precision landing of the booster they’d have to land it and transport it back, destroying the rapid turnaround idea. If they can’t get Starship back through re-entry it will have to be flown expendable. That would still make it the largest, most capable rocket ever, but it would be really expensive to refuel with expendable tankers. It might work for a major NASA mission, but not commercially.
I expect that they’ll eventually get it working close to what was advertised, although a 2-hour turnaround seems a bit of a stretch. But it could take many flights before they are comfortable enough to try precision landing back at the tower.
$5200M expenses in 2022. 61 launches in that year. So, a max upper bound of $85M/launch.
But of course that’s silly since they have so many other things going on. 34 of those launches were Starlink. We don’t know what their satellites cost either, but it’s probably at least $1B. Which itself is “unreasonably” cheap compared to what satellites usually cost.
And of course they have to keep the Dragon program running (which they get paid for, but still counts as expenses). And Starship development was also in full swing. It’s hard to believe that is less than a billion, given the amount of hardware and ground equipment we’ve seen. And a bunch of other miscellaneous stuff.
So you can believe that Falcon 9 flights aren’t really that cheap, but if that’s the case then all the other stuff has to be compressed into an unreasonably small amount. On the other hand, if F9 flights are ~$30M, then that makes them roughly 1/3 of total expenses… which seems about right given everything else.
No, I don’t want to take a SpaceX employee’s word for what it costs; I’d like to see an actual cost breakdown which is something that any launch contractor should easily be able to do. I have personally worked on studies for reusability as well as reviewed historical studies, and from that I know that the marginal cost savings per flight vehicle from reusability are actually not that great because most of the cost is not in the hardware but the ‘touch labor’ involved in integration and testing. What does come from reuse is the ability to increase flight tempo and to amortize the relatively fixed costs of maintaining a launch facility, and there is probably both a significant savings and increased revenue stream there. But I’ll note that SpaceX has continued to increase both their bare manifest costs and the all-up costs of full payload integration and specialty services, so the idea that a Falcon 9 launch ‘costs’ <$30M is not realized by the customer.
SpaceX has done regular funding rounds (at least two per year for the last three years) from private equity and taken in several billion dollars. No doubt this is done to bankroll Starship and an expectation (as yet unrealized) that Starlink will be a tens-of-billion-dollars-per-year ongoing revenue stream but the claim that SpaceX is internally funding all of this development from their profits is manifestly false.
)
