You are not logged in.
Neither moon is geosynchronous with the planet. You cannot build a proper space elevator to either of them, stretching to the surface. The lower ends would drag around the surface of the planet, Phobos very rapidly indeed. All you could do is build partial elevators bottoming out above the atmosphere, meaning well above the entry interface altitude of 135 km at Mars.
GW
To my posting from 2022, #5 above, I would add only the laminar and turbulent flame speeds of hydrogen gas premixed in air: about 4 ft/s (1.3 m/s) laminar, and maybe 4 times that turbulent. Which is how we know the Hindenburg's fire was propagated by the skin, not the unmixed gas in the lifting gas bags. 800+ feet long burned in about 30 seconds from tail to nose. That's 25-30 ft/s. Faster than turbulent flame speed, and the gas was NOT premixed!
The fabric skin was doped with nitrate dope, inherently much more flammable than the later butyrate dopes. That's all they had back then. And the light gray pigment was a mixture of, get this, aluminum powder and iron oxide. In the wrong ratio to be real thermite, but "good enough" to be an accelerant in a fire.
BTW, the insurer knew about this right after the crash, but kept the cause secret to avoid any liability for their client.
Methane is also notorious for leaks, but has narrower flammability limits, more like propane and stove gas. ~5%-10% is an educated guess. I'd have to go look it up.
Biggest problem with both is leaking seals, especially with the materials in the gas phase. That always leaks more easily.
GW
Here is what I posted under the ISS thread in "human missions"
From CBS News/Space:
Space
Space station crew briefly moves to "safe haven" amid new concerns over leak
By Miles Doran, Alex Sundby
Updated on: June 5, 2026 / 11:45 AM EDT / CBS News
Out of an abundance of caution, NASA on Friday briefly directed five of the seven crew members aboard the International Space Station to wait inside the docked SpaceX Crew Dragon "Freedom" spacecraft — known as a "safe haven" — amid new concerns over some cracks in a transfer tunnel in the Russian module.
NASA then told the crew members they were comfortable with backing out of the safe haven configuration and returning to normal operations after the Russian space agency Roscosmos paused repair efforts and decided to take measurements and assess data.
"We look forward to working with Roscosmos on a collaborative approach to address the leaks," NASA said in a statement.
The cracks have been the source of a small air leak that has persisted on and off for the last six years or so. The leak is considered a top safety risk. There have been several attempts over the years to repair the cracks.
"The cracks have always been a concern that NASA watches very closely," the agency said.
Friday's effort comes after Roscosmos noticed a slow pressure drop in the transfer tunnel last month following the arrival of a Russian cargo ship.
NASA said it and Roscosmos have been trying to determine the cause of the cracks while the Russian space agency addresses the issue through "operational mitigation measures and periodic partial-repair efforts."
After new leaks appeared, NASA initially said Roscosmos decided to move forward with a more extensive repair operation Friday before the Russians decided only to perform measurements.
"We continue to work with our Russian counterparts, along with the rest of the international community that supports the space station, to arrive at a more permanent resolution," NASA said.
----
my opinion: they bloody well know what the problem is, metal fatigue from thermal cycling day/night every 90 minutes. The tunnel wall was too thin to support such a long life while cycling. The cracks are "everywhere", but are very hard to see. And they keep lengthening, until something breaks apart.
The way you stop a fatigue crack from lengthening further is to find its very end, which is very, very hard to determine without the right kind of X-ray or other detection equipment, something very hard to do on a pressure vessel in space, because your source is on one side pf the shell, and the receiver must be on the other. Plus, in this case, the tunnel is buried inside the outer structure of the module. Then you drill a hole at that end of the crack, to relieve the stress created by the sharp corner at the end of the crack. Miss the end even slightly, and your repair fails. In vacuum, your hole makes the leak worse, until you can patch it with something, along with the length of the crack. Such cracks are rarely straight, by the way.
GW
From CBS News/Space:
Space
Space station crew briefly moves to "safe haven" amid new concerns over leak
By Miles Doran, Alex Sundby
Updated on: June 5, 2026 / 11:45 AM EDT / CBS News
Out of an abundance of caution, NASA on Friday briefly directed five of the seven crew members aboard the International Space Station to wait inside the docked SpaceX Crew Dragon "Freedom" spacecraft — known as a "safe haven" — amid new concerns over some cracks in a transfer tunnel in the Russian module.
NASA then told the crew members they were comfortable with backing out of the safe haven configuration and returning to normal operations after the Russian space agency Roscosmos paused repair efforts and decided to take measurements and assess data.
"We look forward to working with Roscosmos on a collaborative approach to address the leaks," NASA said in a statement.
The cracks have been the source of a small air leak that has persisted on and off for the last six years or so. The leak is considered a top safety risk. There have been several attempts over the years to repair the cracks.
"The cracks have always been a concern that NASA watches very closely," the agency said.
Friday's effort comes after Roscosmos noticed a slow pressure drop in the transfer tunnel last month following the arrival of a Russian cargo ship.
NASA said it and Roscosmos have been trying to determine the cause of the cracks while the Russian space agency addresses the issue through "operational mitigation measures and periodic partial-repair efforts."
After new leaks appeared, NASA initially said Roscosmos decided to move forward with a more extensive repair operation Friday before the Russians decided only to perform measurements.
"We continue to work with our Russian counterparts, along with the rest of the international community that supports the space station, to arrive at a more permanent resolution," NASA said.
-----
my take on it:
They were alarmed enough over opening the hatch to enter the module that 5 of the 7 aboard took shelter in the capsule. 2 went into the module to do some sort of job that may have gotten changed.
All I can offer is this: the sooner this worn-out facility is replaced, the safer the crews will be. Replace it with one in a low-inclination orbit to support going to the Moon and elsewhere.
GW
Here it is, Tom:
Went to a picnic in McGregor yesterday (Sat 5-30-2026) sponsored by SpaceX, and got a bus tour of their site. First one of those sponsored picnics in about a decade, and the first one I had been to. Most of the old WW2-vintage buildings that I knew were gone, but the old propellant lab ruins and the old machine shop buildings were still there. So was the underground bunker control room for the big-motor stands, although the associated big earthen revetment and firing stands are gone now. My old ramjet test facility was near that propellant lab.
Spacex now has most (but not all) of its Raptor and Merlin test stands rigged with water deluge, and located below the surface, to deaden the noise. I had noticed the noise reductions over the last year or so.
There was a Raptor-1 sea level engine on display outside by one building, and the surviving "Grasshopper" Falcon-core landing test vehicle outside by another. It so happened there was a Falcon-9 core wrapped in plastic coverings and fitted to two sets of truck wheels as its own trailer, waiting to go on the road for delivery. The tour guide did not know if that was Canaveral or Vandenburg. The old tower stand they inherited from Andy Beal is still there, but they don't use it anymore. Tests up in the air atop it are really loud, even if only Merlins.
There is a firing stand facility specifically devoted to Raptor vacuum engines. I did not have the chance to question the tour guide about it, but I saw no sign of any sort of exhaust diffuser equipment. So the area ratio on the vacuum engines is limited to one that won't quite separate, when fired in the open air, at a high throttle setting. Pretty much like I laid out in the rocket performance estimating lesson of the orbits+ course on the forums. My empirical separation criterion is probably a bit too conservative for curved bell nozzles. It was developed for the simpler conical bells typical of tactical solid rockets. With a curved bell, there are games you can play with nonuniform exit plane pressure distributions (higher at the edge, lower near the centerline), to delay separation to slightly-higher backpressures than my old criterion would predict.
I did notice they are very careful to tout the benign nature of the rocket exhausts with methane and kerosene, and that there were questions about that from the audience on the bus. With big rockets, those exhaust clouds do often leave the reservation. Accordingly, I noticed they say nothing about the exhaust fumes from the Draco and Super Draco thrusters. Those are very small in comparison, and their exhaust clouds do not leave the reservation, although they are quite toxic. (I remember similar public concerns about fumes from the solids long ago, which were well-known to everyone to be toxic.)
GW
To make NASA moon base construction feasible in terms of budgets, you need landers sent from Earth that have high payload fractions, so that vehicles will be affordably smaller at departure, and yet still set tons of cargo on the lunar surface. Simple as that.
You do NOT want to launch things directly from the surface of the Earth to do that, because of the high dV to LEO (near 9.5 km/s as corrected for gravity and drag losses), and the low payload fractions of any conceivable launcher to LEO. Such a rocket that could do that "in one go" to LEO, with something carrying tons to the moon as its payload, would dwarf even Starship/Superheavy.
You want to use the existing heavy-lift launchers (hopefully soon to include New Glenn and Starship/Superheavy, along with Falcon-Heavy and ULA's Vulcan) to send pieces to LEO that can be assembled by docking, the same way we built ISS. You assemble your lunar vehicles in LEO and fuel them there, and then depart for the moon from LEO.
But that still has lousy payload fractions, because it costs you about 3 km/s dV to depart LEO, plus another comparable dV at a minimum, to land where you want to go on the moon. Still near dV = 6 km/s or more! Still low payload fractions. It'll still be a budget-buster!
What you want, to get high payload fraction in the things that reach the moon, is to eliminate almost all of the 3 km/s departure dV from LEO. You do that with a tug stage that takes you from LEO to an extended elliptical departure orbit, one with a perigee speed very near escape. Artemis-2's departure just demonstrated the feasibility of doing it that way! They used the ICPS stage as their tug. The service module had a "next-to-nothing" fraction-of-a-km/s dV requirement for departure. Most of its capability could go toward course corrections!
Now with elliptic departure, your vehicles going to the moon have a dV requirement nearer 3 km/s, not 6, so they have room for high payload fraction, and can be much smaller and less expensive to build. In turn, that means less stuff sent up to LEO to be assembled into these vehicles, and less propellant needed to fuel them. You save there, too!
And if you reuse the tug stage and base it at that station in LEO, you save even more! It needs to meet a dV near 3 km/s pushing a heavy load, and another 3 km/s getting back to LEO, but that second burn pushes NO LOAD! It's like an overall dV under 4 km/s all at full payload, so its effective payload fraction can be high, and the tug is nowhere near as big as you might otherwise think!
Moon, Mars, the final departure speeds are comparable. The same mission architecture and the same station and tug, that works for Mars. Or most anywhere else!
THAT is how you create a cost-effective space program that will be able to accomplish astounding things! NOT by what we have been doing, or even contemplating up to now!
This would work for both the planetary probe programs at JPL, and the manned mission programs at the rest of NASA.
GW
I think it is silly to ballyhoo this into a race. The Chinese do good work. They will probably land a crew and get them home. But that does NOT mean they are ready to build a base on the moon.
I'm unsure that NASA is really ready to build a base on the moon. They as yet have no crew lander, and no large 1-way cargo landers, both of which are prerequisites for building a base at all! And thinking they can have either SpaceX or Blue Origin landers ready for crews in only a year is utter nonsense! I know better! Flight test does NOT go that smoothly! It never has, and it never will!
They also still have no inexpensive (cost-effective) way to send things to the moon. That requires assembly by docking and fueling at a space station in the right orbit. ISS is in the wrong orbit, and is wearing out to the point of endangering its crews. Gigantic budgets are going to bust any of their plans.
I've seen this budget-busting thing before, multiple times. The Space Shuttle was a kluged-up cluster instead of a clean and fully-reusable two-stage airplane, because of money. NASA was forced to use an idea from outside the agency (regarding lunar orbit rendezvous), in order to save money by getting down to 1 Saturn 5 launch per mission. The USAF space plane efforts of the late 50's and early 60's were shut down in favor of NASA's space capsules, because the country could not afford both approaches simultaneously, and the capsules were ready to fly first.
GW
Sorry to see that happen to them. LOX-LH2 can be quite the powerful explosive (I saw a hint of a shock wave). I do hope the initial assessment that all personnel are accounted-for proves correct.
Bear in mind that this was to be only the 4th flight of this New Glenn rocket. Technically speaking, it is still in experimental flight test, as is SpaceX with its Starship, although I see that few want to believe that statement.
Back in the 1950's and 1960's a new rocket in flight test like that, blew up more often than it actually flew at first, meaning the first few dozen flights! NOT the first few flights! Despite all the computers, why should things be any different today? For anyone?
This sort of thing is to be expected during experimental test flights of a new rocket, or jet-powered missile, even high-performance airplanes. It's why you locate the testing facilities near oceans or way to hell-and-gone out in the desert. There were so many failures with the old jet-powered Snark cruise missile about 1960 that those waters were referred to as "Snark-infested waters", as a sort of gallows humor by the crews that flight-tested it.
GW
Well, I got the grid fin missing alignment wrong in my "exrocketman" posting. I had the missing fin lined up with the belly mid-line. It was really lined up with the dorsal midline. I'll have to go fix that sketch. But Scott Manley and I agree that the booster flip was 90 degrees away from where it should have been.
My suspicion is that whatever caused that may be related to, or even caused, the engine failures. And THAT is what they are going to have to find out for the FAA.
Anyhow, the video Oldfart found does a really nice job explaining why you want liftoff T/W 1.5 or higher. And where the gravity loss comes from. Once you have the right flip-over gravity turn, higher liftoff T/W is your main influence on reducing gravity loss. There are limits: as you burn off propellant at the same thrust, acceleration rises, perhaps too high for what you are trying to accomplish, in terms of structural design and/or what your payload can endure.
GW
FAA grounds Starship/Superheavy V.3.
As well they should! This must be understood and dealt with. Notwithstanding Musk's impatience.
As close as they came on Flight 12, understanding and actually fixing what went wrong, will only make Flight 13 even better.
And most of SpaceX knows that, even if Musk does not.
Go Shotwell! Get it done!
GW
I put my observations and speculations about Flight 12 into an article posted on "exrocketman", including a couple of illustrations that make what I am trying to say a lot clearer and easier to understand. The site is http://exrocketman.blogspot.com, and the title is "Kudos to SpaceX for Flight 12", dated today at 5-27-2026. It's top of the page right now, but later you can use search code 27052026, the archive tool with date and title, or the search keywords launch or space program.
GW
I do not know, but I suspect there was some sort of hot staging problem that induced the skirt damage and the failure of 1 Raptor aboard the Starship second stage. I do not know, but I suspect that the hot staging problem ultimately induced the engine failures seen just after hot staging in the Superheavy booster. Those engine failures occurred somewhere just after it started the flip-around on some engines that were burning for the flip. The second stage rocket blast assists that thrust-vectoring flip-around.
One possibility is that propellants did not stay settled enough during the flip to feed only liquids to the engines. That staging looked "abnormal" to me in the sense that its flip plane was "horizontal" instead of the more "vertical" orientations seen before, measuring "vertical" by where the grid fin was missing.
I do know that they saw grid fin damage due to stage 2 Starship rocket blast in the version 2 hot-staging flights, and I would suppose that the stage 2 rocket blast was supposed to impact the new version 3 Superheavy along where the grid fin was deleted, and not along the "lateral" side where a grid fin actually was located.
The "smoke" around the inoperative vacuum Raptor, and the reddish glows on its bell edge and on the engine bay skirt, of stage 2 Starship, are strongly suggestive of ongoing quite-significant propellant leaks and a near-vacuum-pressure fire of some kind, in that bay. Oxygen might be reactive enough to support a fire like that, despite the near-vacuum conditions.
That last is just speculation on my part about ongoing leaks, but it is supported by their deletion of the restart-in-space experiment. They may not have had enough propellant left after reaching the trajectory, to support both the restart and the landing. It did take a longer burn to reach trajectory on 5 engines. And if my suspicion is correct about a significant ongoing leak, it was dumping propellants overboard all during that longer interval.
I'm unsure about the interaction between hot staging not going quite right, and whatever damage caused what looks like a propellant leak on the inoperative vacuum Raptor. But the transient mechanical shocks and heating may be a little too much. And I know they use a lot of 3-D-printed parts in that engine. That 3-D printed metals technology obviously gets the strength of forged parts, but as best I understand (which may not be up-to-date), still not quite the elongation-to-failure of forged parts. In other words, such 3-D printed parts are more brittle and thus more susceptible to mechanical shocks, than forged parts.
One thing that I am reasonably sure of, is that they need a a way to fully cut off propellant to an inoperative engine, one that failed because of leaking cracks in the feed lines upstream of the pumps. It does not look like Starship can yet do that.
GW
I saw the SpaceX live video starting about 40 minutes into the flight, through landing. Then I gave them a couple of hours to get the video up for re-run, and watched from launch to the end. This was ship 39 v.3 and booster 19 v.3. I could see 3 larger grid fins on the booster, and the upper stage heatshield is not all the same hex tiles all over. Away from the belly mid line, the lateral tiles are large rectangular-plan sections, curved to match the shell. No clue what the hex tiles or these large rectangular heat shield plates are made of, but both are black in color. The lateral ones show a charred appearance after entry, as if they were ablatives. The hex tiles do show any change in appearance, so I think those are either refractory, or else a very slow ablative indeed.
This one rose off the pad faster than any of the v2's did, so liftoff T/W looked to be greater than 1.5, which is a good thing that lowers gravity losses significantly. Hot staging looked good with two exceptions: there were fewer engines than were planned to be burning, for the booster flip, and one of the Starship vacuum Raptors shut down immediately after staging. The booster made no boost-back burn at all, with all engines off, once flipped. I think there may have been 1 engine burning at touchdown, downrange, so it it hit the sea pretty close to Mach 1. There were no clues in the video about what the engine-out problems might have been.
Starship upper stage made a longer-than-planned burn to reach the suborbital transfer ellipse energy a bit later than planned, on 5 engines instead of all 6. More or less surface-grazing with about a 200 km apogee over the South Atlantic, and entry off the west side of Australia. They pretty much decided to delete the in-space ignition test at about the time of cutoff and coast onto the transfer ellipse. The satellite simulators deployment was all good, and faster than before. Entry looked rather good, and I saw nothing to suggest any hinge line burn-throughs on the aft flaps. No sheet metal damage either, near as I could tell. Belly flop descent and turn maneuvers went well. They had planned to make a 2-engine touchdown instead of the nominal 3-engine touchdown, and it looked pretty good to me. Once down in the water, it toppled over and exploded, just like expected.
Except for the engine failures, especially on the booster, it looked like a pretty good flight test to me. Most of the objectives were achieved, near as I could tell. Better than I really expected, for a significantly-changed configuration with new 3rd-version engines.
GW
I added a 4th worksheet to the "stage studies.xlsx" spreadsheet file that looks at TSTO sizings vs staging speed. I think that's the version Tahanson43206 has in the drop box now.
It works good, but be aware that those effects are down in the weeds compared to Isp and inert fraction effects.
Also be aware that this spreadsheet does not look at stage thrust requirements, at re-sizing the engines to fit such selections, or at determining whether such engines will fit behind the stage.
And also be aware that the inert fractions are merely assumed. No inert build-up activities were included.
GW
From AIAA's "Daily Launch" email newsletter for Friday 22 May 2026:
Ars Technica
Uh-oh, the International Space Station is leaking again
NASA confirmed Thursday that the Russian segment of the International Space Station has begun leaking atmosphere into space again. It’s an old problem that NASA recently hoped was resolved. For more than half a decade, engineers from Roscosmos and NASA have been tracking the leak rate from a small Russian module attached to the space station that leads to a docking port.
-----
My take on it:
I looked at the article linked by the newsletter headline. For the first time, I saw "microscopic structural cracks" that are hard to detect and treat, listed as the cause of the leaks. That is very, very suggestive of fatigue cracks in the wall of the transfer tunnel inside the Zvezda module. That would seem to indicate the module is already well past its fatigue life. The risk is sudden explosive decompression, without any advance warning (the leaks have been the advance warning!). The "fix" has been keeping the hatch to the module closed, unless it is being used. But if that decompression occurs while the hatch is open, the whole station decompresses suddenly, and all aboard die.
GW
Some unidentified issue was unresolvable. Multiple attempts and recycles to the 40 sec hold point, using up the 5 minute cold time window they had. No launch today. Maybe tomorrow.
GW
Cleese was pretty much spot-on.
GW
The propellant is settled for transfers by using what is called "ullage thrust" to slightly-accelerate the vehicle (or docked vehicles). That dos change the orbit a bit. SpaceX does this ullage thrusting with attitude thrusters. At least some of those are cold gas thrusters that use boiloff vapors from the tanks as their cold gas supply. It's done by regulating the boiloff pressure within the tanks. What the thrusters do not use is vented. Long-term cryo storage (meaning weeks+) is not something they have yet addressed.
GW
Re post 194: the gold statue of a politician referred to in that post, I believe to be the gold statue of Trump that was very recently installed at his Mar A Lago estate. Something like 12 to 16 feet tall, if memory serves.
GW
This one is basically a reprise with the new block 3 vehicle of Flight 11 done successfully with the block 2 vehicle. Given the past track record (of everybody, not just SpaceX, when flying new configurations), expect some shortfalls or failures.
GW
Regarding post 20:
quote 1: certain Gallium eutectic metal mixtures convert CO2 into pure Carbon powder and O2 at room temperature without any thermal or electrical power input.
quote 2: Energy is required to separate Carbon from Oxygen, since energy is released when those two elements are combined.
The statements in quotes 1 and 2 cannot be both be true.
GW
From AIAA’s “Daily Launch” email newsletter for Thursday, 14 May, 2026:
SPACENEWS
SpaceX sets date for first Starship version 3 launch
SpaceX has set a date for the long-delayed first launch of its next-generation Starship vehicle, which is critical to the company’s ambitions
SPACE
Will Starship launch from foreign shores? SpaceX 'constantly exploring' options for megarocket liftoff sites
SpaceX just revealed that it's on the hunt for additional launch sites for its Starship megarocket, eyeing locations both inside and beyond
SPACENEWS
SLS to launch without upper stage for Artemis 3
NASA plans to fly the Space Launch System on Artemis 3 without an upper stage as the agency begins to define revised plans for the mission. NASA, in a May 13 update on Artemis 3, said the SLS will launch with an inert “spacer” in place of the Interim Cryogenic Propulsion Stage on the mission. The spacer, being built at Marshall Space Flight Center, will have the same dimensions and interfaces as the ICPS.
-----
My take on these things:
Spacenews article: date is 19 May, 6:30 PM EDT. Or possibly later. Similar suborbital flight to Indian Ocean splashdown as before, no attempted booster recovery. Block 3 booster + block 3 Starship, which are more-or-less a prototype of the production orbital transport version. Starlink simulators to deploy, one or more with imaging capability to examine heat shield. Restart one Raptor in space. Maneuvers during entry. Missing tile test. Basically a reprise of Flight 11, but with a block 3 vehicle.
Space article: SpaceX is adding pads at Canaveral, and looking at purchasing land in Louisiana for another launch site for Starship. SpaceX is also looking at foreign launch sites, but will have to get around ITAR to do that.
Second Spacenews article: Artemis-3 will launch without an ICPS upper stage, but with an inert dummy taking its place. Being only an LEO mission, ICPS is not needed. The core and SRB’s are enough to send the capsule and service module to low orbit, as they did on Artemis-2.
GW
Re post 43 above: I have since realized that what everyone on these forums really needs is a 1st-look spreadsheet that roughs out the bounds on what can be done with both SSTO and TSTO, reusable or not. You do that first, with assumed Isp and inert fraction values, just to bound what might be feasible or not. The better your assumptions, the better your results. The more ideas you screen this way, the fewer you must put more effort into, trying to better estimate inerts, and trying to size engines and see if they fit.
I have come up with such a spreadsheet, and will get Tom to put it in the drop box for download, and post links to it here. I want it added to the orbits+ course materials, as part of lesson 8's materials.
GW
I got distracted, and did not notice the time. When I looked up, it was already past 8:00 central. Sorry.
GW
Most 2-stagers today, stage down near 60-some km altitude and a speed near Mach 3 to 6 (which is 1 to at most 2 km/s). That's also near 60-some km downrange, with the trajectory bent over to around only 10 degrees above local horizontal. The bulk of the dV to orbit (the other 6 km/s) is best supplied by the second stage at its higher Isp with a real vacuum nozzle design on its engine(s). The first stage has to have sea level or at best compromise-ascent nozzles on its engines.
If there are SRB's, these are usually staged of a little before the first stage core, so they can give a higher thrust over a shorter time interval, at a fixed packaged total impulse (propellant mass times its ascent-averaged Isp = to the time integral of thrust). It would not make sense trying stage all that stuff away all ay once. Way too risky!
The regular rocket launcher can be made reusable by recovering its first stage for a rather slow re-entry at around Mach 3 to 6-ish. But, with a decelerating entry burn to hold entry to Mach 3 at most, that relieves you of needing a heat shield for it, even if it is made of aluminum alloys. It does a powered vertical landing, as both SpaceX and Blue Origin have demonstrated. That first stage can have a stage inert fraction of around 5%, without that heat shield, and restricted to hard pad landings with minimal landing legs.
The second stage must be a fully-qualified re-entry vehicle for 8 km/s entries at the least, which means you must cover half or more of its surface with heat shielding. That right there pushes you to an inert fraction in the 7-10% range. Then it must have some means to land: (1) vertical powered landing, (2) horizontal dry lake bed landing if a lifting body shape, or (3) horizontal runway landing with wings.
Choice (1) means you must add some sort of minimal landing legs, restricted to hard pad landings. That will push you to prey near 10% inert fraction, which is just about where SpaceX has been with its Starship test vehicles.
Choice (2) has more surface area and thus more surface to protect with heat shielding. It will also need suitable high-speed landing gear to land on that lake bed. That will drive to to inert fractions somewhere in the 12-15% range. And depending upon weight during the return (it may have payload aboard), touchdown speed for lifting bodies was demonstrated long ago to fall in the 250+ mph range, perhaps close to 300 mph if still loaded with payload. That's why you need a huge dry lakebed for your landing field! All such high-sped landings will be inherently hazardous! It is easy to lose control while rolling that fast, and tumble the vehicle.
Choice (3) has yet more surface area and the mass of the wings, which will also need heat shielding. It will also need landing gear, but with wings, your touchdown sped can fall in the 150-200 mph range, which is much less hazardous, and can be feasible at airports with runways longer than about 10,000 feet. That sort of buildout is going to put you in the 15-20% range of inert fraction.
Your second stage will not have to fight much in the way of losses, so that 6 km/s second stage is its required dV. You have to carry enough payload to be worthwhile, and you will have to deliver that dV at an Isp high enough for inert fractions in the 10-20% range. You'll do a lot better with the higher Isp of LOX-LH2, although SpaceX is sort-of making Starship work at about 10% inert and LOX-LCH4.
If instead you decide to do your stages as some kind of vertical-launched cluster of spaceplanes, there are 2 things you MUST deal with: (1) your first stage inert fraction is going to be in the 10-15% range, because you need very little heat shielding for it, but you must be a lifting body or have wings. It only shoulders 1-2 km/s of delivered dV, but it shoulders almost all of the 1-2+ km/s worth of losses (there will be more drag loss! So that's something like 2-4 km/s worth of ideal dV capability, where its payload is the second stage (much smaller) spaceplane. And its inert fraction is 10-15%, depending upon the details. The second stage spaceplane faces the same design values and choices as listed above for (2) lifting body or (3) winged spaceplane.
I see no value to attempting 3 stages as all spaceplanes.
The best choice might actually be a real spaceplane atop a Superheavy-like booster.
GW