A thought experiment, to which I’ve gotten little response so far.
So, there's an argument going on in Space X (as opposed to SpaceX) about why we have to continue to waste money on SLS/Orion if we want to beat the Chinese back to the Moon. When I propose alternatives, I get objections like "Dragon can't come back from the Moon," "Nothing else…
— Not-So-OK Boomer (@Rand_Simberg) February 18, 2026
Go back about a decade and develop refueling for the Falcon stage 2 in orbit. The rest simplifies. Falcon Heavy and Starship when and as available.
On-orbit refueling is one of those technologies that Space-X needs to move to the top of the pile…..
I think it is more critical to opening up space than Starship. I suspect that after it has been in use for a couple of years, people will wonder what the big deal was. IOW, I expect it to be easy, after it’s developed, and frequent. I also expect it to be copied far faster than booster recovery.
Booster recovery is “Democracy, Whiskey, Sexy” (what an Iraqi thought liberation from Saddam’s rule by the US would bring). Something tells me on-orbit refueling is not only not as glamorous but intrinically more challenging.
Otherwise, why has NASA and SpaceX and everyone else put it on “the back burner” for decades?
Because if successfully demonstrated, it would politically threaten SLS?
Was this a trick question?
yup
John Hare writes:
“develop refueling for the Falcon stage 2 in orbit. ”
Interesting idea. What would that buy you? I ask only for information – not to oppose the idea. How could you use a refueled F9 second stage?
A fully fueled Falcon 9 upper stage in orbit would be a very capable space tug. Remember, the Falcon 9 stages relatively early during the launch, typically about T+ 2:15 to 2:45. The upper stage provides about 2/3rds or more of the velocity needed to orbit. A fully fueled upper stage in orbit could boost a significant mass to the moon or beyond, even more than a Falcon Heavy. Or it could be used as a reusable tug to change the orbits of satellites. You’d need to put in rechargeable batteries and some solar cells to maintain electrical power.
And a propellant heater to keep the RP-1 from turning into an RPopsicle.
A not so yummy Popsicle, that smells like kerosene.
What Grok has to say…
Pour point depressants (PPDs) are commonly used additives in petroleum-based fuels like kerosene to lower the temperature at which the fuel begins to gel or solidify. These polymeric compounds, such as polymethacrylates, ethylene-vinyl acetate copolymers, or alkylated naphthalene polymers, work by interfering with the formation of large wax crystals in the fuel as it cools. By mixing a small amount (typically 0.01–0.5% by weight) of a suitable PPD with RP-1, the pour point can be reduced by up to 40°C, depending on the specific formulation and base fuel composition. This helps maintain fluidity in extremely cold environments like cis-lunar space, where temperatures can drop below -100°C.
RP-1 itself is already refined to have a relatively low pour point (around -51°C per military specifications), but for extended exposure to space conditions, PPDs could provide additional margin without significantly altering other properties like density or combustion performance. Examples of commercial PPD products include those from Afton Chemical, Clariant, or TotalEnergies, which are designed for similar hydrocarbon fuels. Always test compatibility, as additives must not compromise rocket engine performance or introduce contaminants.
The main problem is that all of the objections are bald-faced lies. You have to be willing to walk in the footsteps of Jerry Pournelle, who would argue until his opponents had not only turned blue in the face, but died from apoplexy. Even in the face of an equally stubborn opponent. I had an argument with him that lasted 20 years. And, in the end, he won. Every objection you cite has a factual counter argument that shows your opponent is a liar, a fool, or both. If they won’t concede, fuck ’em.
Honestly it seems an obvious choice to me. Artemis II alone will use 4.2 Billion dollars. Dragon Crew has a design limit of around 10 days, and has been used for 5 (Polaris Dawn). Searches list Orion to be somewhat larger although actual numbers ~330 cu ft for both are essentially identical. I suspect the Delta V for the Trans Lunar injection for Dragon Crew as well as for the Trans Earth injection are a bit short. Of course SLS/Orion is no great chakes using some bizarre orbit instead of the ones used for the Apollo flights.
So My proposal:
Cancel All of artemis
Take a 1-2 of that and issue a fixed cost contract to
1) Extend and test Dragon Crew time on orbit
2) Provide for sufficient power/delta V for a reasonable TLI
3) Extend/enhance the Dragon crew service module and thrusters to be able to regularize to a useful appolo like lunar orbit and provide decent TEI
4) If needed (likely) man rate Falcon 9 Heavy
5) Test and update the Dragon Crew PICA 3 heat shield. It should be acceptable to replace said shield after a lunar run
Add a billion or so in early delivery bonuses.
Ultimately, I suspect the Lunar Starship will be doing the Earth to Moon part of the run with the Dragon initially acting as ferry.
Why do I say Dragon/Falcon 9 ? Because right now, short POSSIBLY the Chinese, no other hardware that is man rated is even close. ULA and Boeing have proved themselves utterly inept and far too glacier like in their actions. Blue Origin has no capsule, New Glenn is NOT man rated, and they’ve launched New Glenn precisely twice. Ariane 6 likely hasn’t got the oomph snd ESA has no capsule. SLS BARELY has the oomph and as far as I can tell Orion is an accident waiting to happen. This leaves you with SpaceX no matter how you fiddle if you want this before say 2040.
1-2 billion of that, somehow the Billion with a B went missing
I think items 3 and 5 are interesting. But only if you are in a rush to get to the moon using only a partially reusable vehicle before Starship HLS supplants it.
What follows is pure speculation. I have no numbers to back this up. (Dick Eagleson please note lol).
There is plenty of empty space in the Dragon trunk (anent: service module) that could be utilized for propulsion. The question is mass/weight. Would it make it too heavy for F9 launches? This could perhaps be compensated for by using an F9H. OR more importantly, would it screw up the CG of the Dragon/F9 Second Stage combo (putting it too far forward to be stable)? If so, maybe using F9H without a 2nd stage and just sacrifice the core booster.
Note that the Apollo Service Module engine was over-designed for historical reasons that escape me at the moment. A Dragon replacement would likely use smaller rocket(s), perhaps a cluster. Could it leverage the existing Dragon LES? Or use it as a basis for rocketry in the trunk?
Really though I don’t see the point of modifying Dragon for TLI/TEI. Starship HLS is being designed for exactly that and also serves as a lander, which you otherwise still wouldn’t have. If you want to continue to use Dragon as a LEO ferry vehicle, should be no problem as long as there are no strange delta-v requirements. But again a Starship based ferry would be totally reusable.
For item 5, note that this could complicate the Dragon design in ways not anticipated. Not that it isn’t worth thinking about. The question I have is whether or not the heat shielding being developed for Starship could be leveraged for a Dragon capsule and would it be sufficient to handle a re-entry from a lunar return trajectory rather than ablating PICA? The latter I presume would have to be replaced for every mission? Once/if we get a successful recovery of Artemis III we’d know if one can leverage the newer Orion heat shield formulation. Isn’t that PICA as well?
An interesting thought experiment, but mostly a distraction from the Starship w/orbital refueling baseline. I’d say the timeline we should focus-on/accelerate. Trying to shoe-horn in lunar missions for Dragon reminds me of how, by trying to re-use shuttle hardware, we got into the SLS mess in the first place.
To me (and likely Elon), after ISS, it seems Dragon/F9’s days are numbered. Sure they may win taxi contracts for the new commercial space stations being proposed and that might keep it going beyond what Musk anticipates is Dragon’s service life. Nonetheless, it is not SpaceX’s focus these days.
> would it screw up the CG of the Dragon/F9 Second Stage combo
> (putting it too far forward to be stable)?
It would put the CG of the Dragon+trunk too far *aft* to be stable during an ascent abort. Dragon currently keeps the trunk attached during ascent aborts (despite the mass penalty) for the light trunk to move the CP aft of the CG for stability.
> Note that the Apollo Service Module engine was over-
> designed for historical reasons that escape me at the moment.
At the time the SPS was sized, the mission mode was Direct Ascent, which required the SPS to perform both the lunar landing and the ascent from the lunar surface. So the SPS was designed to provide T/W > 1 in lunar g.
The Crew Dragon doesn’t have a service module. It has a trunk, which is basically an empty cylinder. It houses the heat radiators on the inside and solar cells on the outside. There are small fins on the bottom of the trunk to provide stability in a launch abort situation. Cargo Dragon can carry additional payload in the unpressurized trunk because it doesn’t have a Launch Escape System (LES). I seriously doubt they’ll ever attempt to launch a Crew Dragon with cargo in the trunk. That would add mass to the escape stack, reducing the possible acceleration at a time when you need it the most.
It would require a major redesign of the LES to convert the trunk into a true service module. It would be far too heavy to perform an abort dragging the SM along for the ride, so you’d need a way to maintain stability with the capsule alone. Perhaps some form of ballute would work, but that would require a lot of testing.
It might be possible to develop a SM insert to go inside the trunk. In the case of an abort, explosive bolts could separate the SM from the trunk so the abort can proceed as designed while the SM stays attacked to the upper stage. This would introduce at least one additional failure mode of the separation not happening, or if it fails to separate cleanly. You would also need additional wiring interfaces between the capsule and the SM for command and control. That interface would need to be severed during the abort and also before reentry. It seems a small thing, but those can sometimes bite you.
You would almost certainly need the FH to launch a Crew Dragon with a SM.
True ultimate goal is Starship L(unar) taxi with Dragon providing the ride to orbit until Starship C(rew) is ready to go. The thing is the Starship L needs orbital refueling (untested) and there is nothing other than an earthbound mockup for crew quarters. And of course no one has placed any environmental system in any starship so it’s in the same spot as Artemis II. Looking at other responses uprating the Dragon Crew might have some issues. Although I had thought work was going on to uprate the Super Draco’s and extend the trunk for the Dragon, which will be used to deorbit the ISS. Although I don’t THINK that was intended to be manned, so the escape issues aren’t there. I presume if things go south and you have an issue with the Dragon Tug you just let it RUD. .Given the comments, it looks more and more like pushing through to Starship L is the way to go.
Wasn’t the Apollo Service Module engine sized for Direct Ascent of the Apollo spacecraft from the lunar surface? The decision to use Lunar Orbit Rendezvous was late-in-the-game of President Kennedy’s (Ted Sorensen’s?) end-of-the-decade challenge. Too late to redesign many hardware components?
Tregonsee314 writes:
“Honestly it seems an obvious choice to me. Artemis II alone will use 4.2 Billion dollars. ”
Yes..it’s SO obvious that there has to be another explanation – or set of explanations – as to why people support paying (wasting) all those billions. Evidently just pointing out the cost isn’t enough to move people from their positions.
The best answer I see is that SLS spread its contracts all over a bunch of companies and locales, so support is heavy on both sides of the congressional aisles. It is pork on the hoof. It also supports several large corporations, which, although only moderately skilled in making space-going hardware, are very skilled in lobbying.
That can’t be the only reason.
would it screw up the CG of the Dragon/F9 Second Stage combo (putting it too far forward to be stable)?
All my CG/CP experience was with model rockets well inside Earth’s troposphere. I should include a caveat that I don’t have a clue if this matters much in vaccum.
It’s not much of a thought experiment:
Will we do the only viable approach? Or get ourselves blown up? It’s more a detection of thought experiment.