totm dec 2023 Artemis Discussion Thread

[Barzon]

3 hours ago, tater said:

There is no alternate future where SLS is launching at a rate where a Mars mission is possible with SLS.

I remember some Boeing guy saying recently that if they got the funding they could build 6 core stages in a year for a Mars program in like the 2030s/40s

[tater]

29 minutes ago, Barzon said:

I remember some Boeing guy saying recently that if they got the funding they could build 6 core stages in a year for a Mars program in like the 2030s/40s

With infinite money, anything is possible. I will continue to assume a basically constant NASA budget. At ~1 B$ per SLS (marginal), then about the same for literally any payload sitting on top (900M$ for Orion, and at least as much for any Mars ship parts), that results in 12 B$/yr, not counting program costs (~2.5 B$). So to do 6, NASA would need to spend almost 15 B$/year of their budget just on that. They'd pretty much have to cancel everything else, including just keeping the lights on for programs that already have spacecraft observing other planets, or the Earth.

And even if they did, they'd need those stages launched to assemble all the propellant parts rapidly, before boil off. Doing so at NRHO makes little sense.

[sevenperforce]

46 minutes ago, tater said:

I'd expect that LockMart would use a stripped down version of that Mars lander they were talking about (for single stage reusable). Haven't seen any real specs on that, though.

LockMart has proposed the use of the Orion pressure vessel, ECLSS, and avionics for their notional single-stage lunar lander. https://www.lockheedmartin.com/content/dam/lockheed-martin/space/documents/ahead/LM-Crewed-Lunar-Lander-from-Gateway-IAC-2018-Rev1.pdf

Quote

The lander is an aggregation and evolution of existing systems and capabilities. The crew pressure vessel is based on a lightweight version of the Orion structure components, which can be tailored and optimized for lander design drivers where launch abort and water landing loads are not dominant. The avionics systems, life support systems, controls and displays, and crew systems are common and interchangeable with Orion. This degree of commonality supports safety while leveraging existing deep space human rated systems in the most cost-effective way possible.

The two-tank LOX/LH2 cryogenic propulsion system reflects the background with the Centaur upper stage and its descendants. One possible technology for the main engines are RL-10 derivatives with deep throttling capability as demonstrated during the Delta Clipper program. Four engines were selected to provide engine out capability. A three engine solution would have had difficulty with off-axis thrust with one engine out. This ensures that there are no black zones in descent or ascent, and that a self-powered abort to the Gateway safe haven is always possible.

 

[tater]

Ars article:

https://arstechnica.com/science/2019/10/as-nasa-tries-to-land-on-the-moon-it-has-plenty-of-rockets-to-choose-from/?utm_source=nl&utm_brand=ars&utm_campaign=aud-dev&utm_mailing=ARS_Rocket_101019&utm_medium=email&bxid=5d89234dfc942d47888527e2&cndid=&esrc=&utm_term=ARS_RocketReport

 

Diagram is useful.

 

[jadebenn]

I'm gonna shill my reddit post here, because I strongly disagree with Berger's take that nobody's going to propose an SLS-launched lander.

 

[tater]

12 minutes ago, jadebenn said:

I'm gonna shill my reddit post here, because I strongly disagree with Berger's take that nobody's going to propose an SLS-launched lander.

Then you need a 37 ton lander. NASA thinks a single stage lander is 50+ tons, and a 2 stage is 41-50t, and a 3 stage is 36-42.

So SLS alone means that LockMart or Boeing needs to come up with a 3 stage lander for SLS that doesn't use hydrolox.

Not impossible, and good for Boeing as it throws vast piles of money back at Boeing.

The price should matter, though. 2024 isn't magic, the Moon isn't disappearing any time soon. So if it costs an extra few billion to fly on SLS vs commercial LVs, then it should not happen. Better to build more landers, and send them up in the least expensive way possible.

[jadebenn]

14 minutes ago, tater said:

Then you need a 37 ton lander. NASA thinks a single stage lander is 50+ tons, and a 2 stage is 41-50t, and a 3 stage is 36-42.

So SLS alone means that LockMart or Boeing needs to come up with a 3 stage lander for SLS that doesn't use hydrolox.

Not true. Check out the links in my argument; I took some screenshots from the architecture analysis pdf. SLS block 1 cargo allows a two-launch architecture 

Edited October 10, 2019 by jadebenn

[tater]

3 minutes ago, jadebenn said:

Not true. Check out the links in my argument; I took some screenshots from the architecture analysis pdf. SLS block 1 cargo allows a two-launch architecture 

That IP03 image shows 2 stages, but gives no numbers.

If we take what the other (NASA) slide shows, min Ascent stage is 9t. The Descent is exactly as you suggest, doable with SLS. But not the 2 combined unless the whole thing is made smaller, and more LM like (it's flags and footprints, at that point).

I liked Artemis as a goal for sustainable lunar missions to the same location. I like this rush to footprints a lot less. Any landing system that will not be used again is a waste of time. I don't mean reused. I mean any DESIGN that is not used again is a wast of time, it's the same as ICPS instead of EUS was always stupid. "Save money" by building a MLP and VAB bays to only be used once (now not once, but that was not the plan).

If they then make a different lander, it needs all the testing of the first one. Do it right the first time, IMHO, then make them as needed. So if they put a crappy, 2 person sortie lander so they can say they did it. Yuck.

A second SLS launch makes no sense unless the whole mission is just the 2 launches. Not 2 + commercial.

Why? Because as soon as you have to dock the lunar lander together at Gateway, you've added all that complexity to the mission profile, and you can't then say "docking 2 stages together at the Moon is easy, but having a tug dock to that is impossibly complex."

At the point the lander is not sent all up, assembled on Earth, then SLS cost ruins any distributed launch architecture. I'll buy simplicity, etc as an argument for 1 SLS launch of a complete lander, 1 of crew. Go to the Moon, add anything else, and then might as well send lander the cheapest way possible.

[sevenperforce]

1 hour ago, tater said:

Diagram is useful.

Very useful.

If the transfer vehicle is supposed to ferry the ascent module back to the gateway, then there should be variation in the mass of the ascent module between the two-stage and three-stage options. Specifically, if a 9-12 tonne ascent module is using hypergolics with a notional Isp of 320 s, it would need 5.1-6.7 tonnes of propellant to get back to the gateway, putting its dry mass at 3.9-5.2 tonnes. Going only to LLO to meet the transfer vehicle is much cheaper and would cost only 4.2-5.4 tonnes of propellant, making the total wet mass only 7-10 tonnes.

Establishing the expected dry mass of the ascent module at ~3.9-5.2 tonnes is interesting. Represents a ~100% mass growth over the Apollo ascent module and lunar hab, which is not THAT much given that it is supposed to have 100% more crew, and roughly half the weight of a full Orion plus tankage, as I indicated above.

Edited October 10, 2019 by sevenperforce

[tater]

3 minutes ago, sevenperforce said:

If the transfer vehicle is supposed to ferry the ascent module back to the gateway,

I think the nominal architecture is that the transfer vehicle brings the stack to LLO, and is disposed. Ascent stage is built with dv to take it to Gateway.

[jadebenn]

17 minutes ago, tater said:

That IP03 image shows 2 stages, but gives no numbers.

If we take what the other (NASA) slide shows, min Ascent stage is 9t. The Descent is exactly as you suggest, doable with SLS. But not the 2 combined unless the whole thing is made smaller, and more LM like (it's flags and footprints, at that point).

Right. When I say two-launch, I mean just for the lander. Descent goes up on SLS. Ascent goes up on a commercial vehicle. Crew goes up on another SLS.

The IP 03 proposal is exactly what I was referring to.

Since the ascent module is meant to be reusable after the first incarnation, that would mean you'd only need one additional SLS launch for each lunar sortie, at least until the ascent module needs to be replaced.

17 minutes ago, tater said:

If they then make a different lander, it needs all the testing of the first one. Do it right the first time, IMHO, then make them as needed. So if they put a crappy, 2 person sortie lander so they can say they did it. Yuck.

The limiting factor (or at least the one NASA anticipates) is life support. The HLS solicitation con-ops state that the lander must be able to support a crew of 4 with pre-deployed surface assets. Without those, it only supports 2.

17 minutes ago, tater said:

A second SLS launch makes no sense unless the whole mission is just the 2 launches. Not 2 + commercial.

Why? Because as soon as you have to dock the lunar lander together at Gateway, you've added all that complexity to the mission profile, and you can't then say "docking 2 stages together at the Moon is easy, but having a tug dock to that is impossibly complex."

At the point the lander is not sent all up, assembled on Earth, then SLS cost ruins any distributed launch architecture. I'll buy simplicity, etc as an argument for 1 SLS launch of a complete lander, 1 of crew. Go to the Moon, add anything else, and then might as well send lander the cheapest way possible.

I've had this argument with you before, so for brevity I'll just summarize it.

The bidders have incentive to bring down lander development costs through simplifying the design. SLS allows a simpler design without giving-up capability. The question for the bidders, then, is whether they can save enough in development to justify launching on SLS. That is the million-dollar question that will decide what route they take.

Edited October 10, 2019 by jadebenn

[sevenperforce]

1 minute ago, tater said:

I think the nominal architecture is that the transfer vehicle brings the stack to LLO, and is disposed. Ascent stage is built with dv to take it to Gateway.

That's not what I read at the above article.

Quote

Along the way, the crew rides in an 'ascent module,' where they live during the lunar surface stay and in which they launch from the Moon's surface back to the waiting transfer vehicle.

However, when I look at the HLS Concept of Operations linked here, it suggests otherwise. Transfer vehicle is disposed by crashing into the moon; ascent module is disposed in the vicinity of Gateway. Initial surface ops are expected to have only two crew once again.

Dry mass estimate for the ascent module still holds.

It's interesting that in the three-stage configuration shown in the Concept of Operations, it explicitly has a Falcon Heavy silhouette launching all commercial payloads.

[jadebenn]

14 minutes ago, sevenperforce said:

It's interesting that in the three-stage configuration shown in the Concept of Operations, it explicitly has a Falcon Heavy silhouette launching all commercial payloads.

It's not surprising. The 3-stage lander will already be strapped for mass, and FH is the most-capable non-SLS launcher that will be available in the relevant timeframe.

Edited October 10, 2019 by jadebenn

[tater]

34 minutes ago, jadebenn said:

The bidders have incentive to bring down lander development costs through simplifying the design. SLS allows a simpler design without giving-up capability. The question for the bidders, then, is whether they can save enough in development to justify launching on SLS. That is the million-dollar question that will decide what route they take.

No, it doesn't simplify anything. The hard part is assembling the 2 stage lander at the Moon. If that is doable, then docking to a tug stage---exactly as the LM was docked to the CSM---is trivial.

If they can't routinely dock something automatically, Artemis isn't a thing.

If they can't do the much more complex docking of ascent to descent stage of the lander (much stronger connection required, and yet it needs to let go instantly for abort/ascent), Artemis is not a thing.

If the Ascent stage cannot dock to Gateway/Orion, Artemis is not a thing.

Adding a single, trivial docking maneuver of "transfer stage" to the lander is not a long pole. Docking fails (hard to imagine), and they are still at Gateway. Leave lander, try again next time (storable props). Tug fails later? Abort to Gateway with either lander stage, easily.

Also, an all commercial lander launch means it will cost far less to test. I think using this without a complete all up test is insane in a NASA where they are so careful about CST-100 and Crew Dragon.

22 minutes ago, sevenperforce said:

That's not what I read at the above article.

However, when I look at the HLS Concept of Operations linked here, it suggests otherwise. Transfer vehicle is disposed by crashing into the moon; ascent module is disposed in the vicinity of Gateway. Initial surface ops are expected to have only two crew once again.

Dry mass estimate for the ascent module still holds.

It's interesting that in the three-stage configuration shown in the Concept of Operations, it explicitly has a Falcon Heavy silhouette launching all commercial payloads.

Yeah, Berger's text vs what NASA slides have been showing is off.

Note that the ascent stage masses listed are the same for 2 and 3 stage options, so clearly it returns to Gateway.

Edited October 10, 2019 by tater

[sevenperforce]

16 minutes ago, tater said:

Note that the ascent stage masses listed are the same for 2 and 3 stage options, so clearly it returns to Gateway.

Right, that's what I initially had issue with. Thought it could have been an error on NASA's part. Which would not have been improbable....

[tater]

Ascent stage is a crew vehicle, and as such should end up costing a few hundred million a pop (since that's what crew vehicles seem to cost). The descent stage is just a stage, but with legs. Wait, some stages already have legs. OK, a stage with 1 engine, and legs. That's a stage 2. About 25% of the cost of a rocket launch, so the ascent stage should cost about 15 million bucks (or we're being ripped off). Nice thing is that testing should be pretty reasonable with a boilerplate crew cabin for mass.

3 minutes ago, sevenperforce said:

Right, that's what I initially had issue with. Thought it could have been an error on NASA's part. Which would not have been improbable....

A 3.75 ton (dry) ascent vehicle has 2747 m/s dv at 9t wet. 3km/s at 3.45t. Very doable (half again larger than LM)

[sevenperforce]

Here's a novel thought.

If fifteen tonnes is the maximum anything but SLS will be throwing to TLI, what about a fifteen-tonne landing module?

Hear me out. Put the loaded hab at 3.8 tonnes: half the landed mass of Orion. Suppose crew, cargo, and consumables come to 0.8 tonnes, making the empty hab 3 tonnes. I'll use a mass ratio approximation from the Apollo descent module. Propellant mass was 8200 kg; dry mass 2134 kg. That's a structural fraction of 20.1% which can definitely be beaten by modern materials, shared components, and the square-cube law. Allowing ourselves 15 tonnes of total vehicle mass without crew, cargo, or consumables gives us a 12 tonne propulsion module; if we only use the square-cube law then that gives us 2358 kg of propulsion module dry mass and 9642 kg of props.

At a notional isp of 320 s, the lander would have 2.95 km/s of dV. What's that good for? Well, it's good enough to provide the last 350 m/s of terminal descent, plus ascent, plus return to the gateway. It would need a transfer stage to ferry it, not just from the Gateway to LLO, but from the Gateway to LLO to the initiation of terminal descent: about 2.25 km/s.

You need about 15 tons of hypergolics to push 15 tonnes of payload through 2.25 km/s.

So you don't need three stages at all. All you really need is to beef up the transfer vehicle and let the ascent vehicle perform the terminal landing burn, and split your vehicles into two 15-tonne chunks. At that point, you can do two commercial launches and be done with it. You'll need a separate logistics launch to refuel both vehicles after they rendezvous with the Gateway, but those props are not super significant and you need the logistics launches anyway. If the ascender/lander can be reused, then you'd refuel it and the transfer vehicle from a dedicated logistics launch, which would mean two launches per sortie, ad infinitum.

Alternately, you can stack both transfer and lander on SLS Block 1B Cargo and do it Constellation-style (albeit with rendezvous at Gateway rather than in Earth orbit). With that approach you have enough margin to not need a separate logistics launch.

Edited October 10, 2019 by sevenperforce

[tater]

^^^so the Ascent stage has legs, does the terminal landing burn, and the other stage is a Gateway---> LLO---->deorbit crasher?

[jinnantonix]

@sevenperforce I like your idea, attractive in it's simplicity. 

The AV loses some efficiency by needing to carry around ladders and landing legs.  This to be balanced around not needing the equipment for undocking the AV from the DV at lunar launch.  Also, any excess weight, logistics packaging, science instruments, etc would need to be manually dropped from the AV prior to launch.

Assuming the AV is re-usable, what engine configuration would you use?  I would suggest an array of 4 engines with N+1 redundancy.  Modified RL10 or AJ10-190 plus an array of hypergolic thrusters for fine tuning would seem suitable.  I am not sure of the re-usability of RL10s?  I would suggest the transfer vehicle would use SuperDraco's plus DRACO docking thrusters as it will only need to fire main engines twice, first during insertion into NRHO, then to de-orbit the AV.

Delivery would require two Falcon Heavy launches (would this give SpaceX a monopoly?).  The Orion with crew launched on a SLS Block 1B.  Probably the refuelling could be delivered using a Cygnus piggybacking on the SLS?  Or at worst a third commercial launch on a Vulcan or Falcon 9.  The Cygnus would need to be beefed up for the insertion into NRHO.

 

Edited October 10, 2019 by jinnantonix

[tater]

The problem with RL-10s is cryos.

Storables make so much more sense for this application, even with the grossly lower I_sp.

[sevenperforce]

5 hours ago, tater said:

The problem with RL-10s is cryos.

Storables make so much more sense for this application, even with the grossly lower I_sp.

I will try and put together a table. There are a few major considerations, including whether you have no propellant transfer, slight propellant transfer, or full propellant transfer. Also whether you have cryos, ZBO cryos, a combo of storables and cryos, or pure storables.

I need to do some stage mass calculations to nail down a better idea of what sort of mass fraction is possible. My gut is that the best combination, all-around, will be a hypergolic ascent stage and a cryogenic transfer stage. But we will see.

7 hours ago, tater said:

^^^so the Ascent stage has legs, does the terminal landing burn, and the other stage is a Gateway---> LLO---->deorbit crasher?

Yes, exactly. Basically you transfer extra dV to the ascent stage and thus end up with roughly equivalent-mass stages.

6 hours ago, jinnantonix said:

The AV loses some efficiency by needing to carry around ladders and landing legs.  This to be balanced around not needing the equipment for undocking the AV from the DV at lunar launch.  Also, any excess weight, logistics packaging, science instruments, etc would need to be manually dropped from the AV prior to launch.

Assuming the AV is re-usable, what engine configuration would you use?

The degree of reuse is a big deal as well. Unless you are able to do propellant transfer of some kind, reuse is of limited value. There are still a few options. For example, you can make the actual crew capsule reusable and expend the ascent stage and tanks in the vicinity of the Gateway after each flight. Saves a good bit of mass between sorties. Or, if you can do only limited propellant transfer, you can put RCS on the crew capsule and give it just enough tankage to get from LLO to Gateway, and expend the ascent stage in LLO.

[tater]

ZBO, and other stuff is certainly possible, I just don't see it as very likely in the next couple years. I operate under the assumption that any notion that the very first flight of this thing will be landing on the Moon, with "the first woman and next man" is absurd. They will have to test it, and since this is not the Apollo era with Saturns launching every few months, it means the thing has to be ready to fly in maybe 2023 at the latest so it can be tested.

Edited October 11, 2019 by tater

[sevenperforce]

The most critical question in evaluating possible architectures is fluidics management. On one extreme you have proposals like the LockMart sixty-tonne fully-reusable elevator monstrosity, which depends on a cislunar hydrolox depot supplied by a solar electrolysis plant, or Starship itself. If you have total propellant transfer capability, all kinds of unique architectures open up. Of course, that does not play well with pressure-fed solutions, so it excludes a lot of current/COTS planning and requires ZBO technology.

The other extreme is to have no propellant transfer whatsoever. The only way to do reuse under these circumstances is to have a reusable crew cabin with an expendable bus, or at most a system that uses expendable, replaceable drop tanks which can be attached using a Canadarm to structural umbilicals.

The middle ground is to have limited propellant transfer, slightly more flexible than what we have already demonstrated with Progress, etc. In this regime, it is possible to have a reusable vehicle with small pressure-fed tanks that are vented to vacuum and then filled via umbilical from much larger, expendable tanks. This has some advantages over a zero-transfer approach, though they are not necessarily overwhelming advantages. Since the larger tank has to be expended anyway, the launch supply requirements aren't dramatically different. Still, it can be a useful way to achieve certain solutions. For example, if you used a crasher stage for transfer and descent and an expendable landing/ascent stage but wanted to have a reusable crew cabin, you could give the crew vehicle ~800 m/s of onboard tankage, refilled each sortie from the landing/ascent stage. Then you can leave the spent ascent stage to decay in LLO and take the crew vehicle back to the Gateway on RCS.

Another question is the number of permissible separation events. Being KSP players, we are highly inclined to shed any unnecessary mass with extreme prejudice. For example, in a combined lander/ascender architecture, we would be strongly inclined to jettison the ladder, landing legs, and any compressive structures at liftoff. Such a design then thoroughly trounces the Apollo descent+ascent solution, because not only do you shed unnecessary mass before liftoff, but the surface structures are far lighter since they have less mass to support. However, separation events add risk, which may not be acceptable in real life.

Several ground rules and assumptions need to be established. The mass fraction of tankage should be estimated by comparison to a launch vehicle stage of commensurate size and engine cycle as well as identical propellant. An expander-cycle hydrolox stage will have a much poorer tankage mass fraction than a gas-generator kerolox stage, which will have a much poorer tankage mass fraction than a hypergolic turbopumped stage, which will have a much better mass fraction than a pressure-fed hypergolic stage. Pressure-fed stages are not quite able to get square-cube advantages, while turbopumped stages generally are.

Estimating the additional mass of landing legs and associated ground structure is best done by reference to the excess mass of the Apollo descent module. The descent module had a dry mass of 1,983 kg and carried 8,200 kg of hypergolic propellants. 179 kg of its dry mass was, of course, its engine, but pressure-fed tanks for that much propellant should have only massed around 1,400 kg (compare to Transtage, with pressure-fed tanks massing 1,750 kg and 10,297 kg of hypergolic propellant). Thus we can estimate conservatively that the ground structural mass of the Apollo descent stage was no more than 590 kg. Since it supported a landed mass of 6,834 kg, we can add a 9.4% structural fraction to any lander.

[sevenperforce]

Using the more precise GR&As above, I'll try to nail down the numbers for the above architecture a little more cleanly.

Before, I said a 3-tonne crew capsule with 800 kg for crew, cargo, and consumables. Assuming that the lander/ascent vehicle comes off of TLI at exactly 15 tonnes (since that's our peak throw) and develops 316 seconds of isp during the powered lunar flyby and rendezvous with the Gateway (summing 428.5 m/s), it will arrive at Gateway having burned 1.9 tonnes of propellant. It will take on 800 kg of crew, cargo, and consumables; its engines will next fire for the terminal landing burn. 

I earlier estimated it would do about 350 m/s of the terminal burn. That may go up or down, but it's a good starting point, so at touchdown it should mass around 12.4 tonnes. By the above metric, we'll add about 1.1 tonnes for the landing legs and ground structure. We have better materials and construction nowadays so I feel like that bakes in some conservatism.

Assuming we want at least a third of a gee for efficient takeoff, we will need at least 45 kN. We could go with a single vacuum-expanded Superdraco, but I don't think it would be able to throttle low enough for hover. Better to do a tight cluster of three OMS-heritage throttleable AJ-10 engines at around 26.7 kN each and 319 s; with a decent gimbal range that should give single-engine-out capability early in the flight and double-engine-out capability later in the flight. This tells us our engine mass will be 0.3 tonnes. Going back to our original 15-tonne vehicle coming off TLI, that's 10.6 tonnes of tankage and propellant. Using Transtage's 14.5% tankage mass fraction gives us 1.5 tonnes of tankage and 9.1 tonnes of propellant, dry (loaded) mass of 6.7 tonnes. Now we can start at the end and work backward. It burns just under 1.8 tonnes to get from LLO to Gateway and around 7 tonnes to get from the surface to LLO, leaving nothing for the terminal landing burn or the rendezvous with Gateway. So we are not that far off, but it doesn't quite work.

However, there's one puzzle piece that I think should make everything work. All other things being equal, it really makes sense to use a reusable crew cabin/capsule. In fact, it may well be the precise puzzle piece that makes everything work. Consumables can be easily swapped out (bottled air, CO2 scrubbers, and so forth) via Orion trips. And it can't be expended en route; it must make it back to Gateway...so why not keep it at Gateway? Saves having to build and send a new one every trip.

The Zvezda module has been refueled numerous times by Progress and Progress M1 resupply spacecraft and has remained on orbit for two decades. Accordingly, it is not a huge leap to develop the capacity to have the crew cabin/capsule with RCS thrusters (think Dracos) that can be refueled from a docked or berthed vehicle. The mass of the RCS thrusters is negligible, and the tankage mass ratio would probably be similar to the 17.3% of the AVUM upper stage on Vega. The crew capsule would be able to make the journey from LLO to the Gateway on its own, entirely on internal propellant. To get 730 m/s on likely lower-efficiency RCS thrusters (I will use 300 s, since that's what the Dracos develop), you burn just over 1.1 tonnes of propellant to get from LLO to Gateway and the dry loaded mass of the crew capsule is right at 4 tonnes. Note that this also means the crew capsule can deliver itself to the Gateway easily, as Falcon 9 can throw five tonnes to TLI without even expending the core.

Of course, you have to get that 1.1 tonnes of propellant from somewhere. The easy solution is to get it from the expendable lander/ascent vehicle, which will berth or dock to the base of the crew capsule rather than docking directly to Gateway. Without those three tonnes of extra mass being thrown from TLI, its same 15 tonnes includes 2 tonnes of tankage and 11.6 tonnes of propellant. It arrives at Gateway with 9.7 tonnes of propellant and transfers 1.1 tonnes to the capsule, leaving it with 8.6 tonnes. It will still take it around 7 tonnes of propellant to get off the surface and into lunar orbit, but that leaves 1.6 tonnes of propellant for the terminal landing burn: around 300 m/s.

So, what are the requirements on the transfer vehicle? Well, the transfer vehicle needs to deliver itself from TLI to the Gateway (428.5 m/s), and then it needs to deliver 17.2 tonnes from the Gateway to LLO (730 m/s), and then provide the first 1,570 m/s of the descent and landing burn.

This is where the RL-10 can really shine. A 15-tonne hydrolox transfer stage, launched onto TLI by anything capable of throwing it, will burn 1.4 tonnes of propellant getting from TLI to the Gateway. Pump-fed hydrolox tanks boast a mass fraction of around 9%, meaning the stage will dock to the lunar sortie stack with a dry mass of 1,529 kg and 11,855 kg of propellant...giving it 2,165 m/s of dV. That's enough to get it from Gateway to LLO and from LLO to about 435 m/s short of the lunar surface...within 135 m/s of where we need to be.

So it's doable. Two commercial launches, one reusable "space taxi" crew capsule.

[Barzon]

your rank seems  very accurate right now, Sevenperforce .