Starship Expendable Upper Stage and other FrankendRocket Concepts

[meekGee]

It's really non-intuitive that landing a Starship on the Moon and returning it to EDL takes more delta-v than escaping the solar system, but that is the case.  So if a fully-fueled pusher can manage that (and it can, as long as the target Starship is refueled as well), I don't think you need anything fancier than that.

This can't be emphasized enough.
Everyday people say "but the moon is closer". But the fact of the matter in terms of fuel the rest of the solar system(one way) is closer. The moon round trip takes more.

Every time someone asks about LEO starship: "Can't we exchange the mass of the heat shield for more fuel and just slow down propulsively?"

You think the moon is nearby?  LEO is only 200 km away!  It's not intuitive, and few people are willing to put their trust in math over intuition.

[InterestedEngineer]

You think the moon is nearby?  LEO is only 200 km away!  It's not intuitive, and few people are willing to put their trust in math over intuition.

Intuition can be trained - that's why I encourage every to play KSP for at least 30 hours.  It completely changes your intuition about space travel.

[meekGee]

You think the moon is nearby?  LEO is only 200 km away!  It's not intuitive, and few people are willing to put their trust in math over intuition.

Intuition can be trained - that's why I encourage every to play KSP for at least 30 hours.  It completely changes your intuition about space travel.

Can happen, but it's always easier to be lazy and use one-line slogans, so it doesn't

[TheRadicalModerate]

You think the moon is nearby?  LEO is only 200 km away!  It's not intuitive, and few people are willing to put their trust in math over intuition.

Intuition can be trained - that's why I encourage every to play KSP for at least 30 hours.  It completely changes your intuition about space travel.

My intuition about space travel is that my intuition is almost always wrong.  Akin's Law #1 is the Prime Directive.

[Overtone]

On reflection, the drop tank concept may still make sense if the mission profile requires a higher propellant mass fraction than a recoverable pusher can provide.

You don't have to recover the pusher.  There's no question that a drop tank will result in a lower dry mass, but you need really high energy (like, well past heliocentric escape speed) before you care very much.  So you'd need a very compelling case before figuring out how to do a drop tank.  And it would likely be easier just starting from a very energetic HEEO.  That would take a buttload of tankers, but there's no real development involved.

Fast transfer to Jupiter?

FWIW, a v3 Starship with no tiles, flaps, header tanks, and a jettisonable nose fairing can do a direct transfer to Jupiter from VLEO (200 x 200).  For that matter, a stripped-down v3, even with fairly conservative dry mass estimates, can get more than 40t of payload from VLEO to heliocentric escape.

It's really non-intuitive that landing a Starship on the Moon and returning it to EDL takes more delta-v than escaping the solar system, but that is the case.  So if a fully-fueled pusher can manage that (and it can, as long as the target Starship is refueled as well), I don't think you need anything fancier than that.

ETA:  HEEO refueing can obviously also be made to work.  The pusher would just make things all LEO-based, which is handy.  I've more-or-less given up on the idea of a combined pusher/depot, simply because the refueling interface is kludgy.  It'd lovely to remove an RPOD from the system, but it's not necessary.  Both the pusher and an HEEO refueling require the same number of RPODs for the target system.  (A dedicated pusher would require one more RPOD with the depot than just having the depot boost to HEEO.)

Concur.  Thanks for educating me/us on this.

[Vultur]

FWIW, a v3 Starship with no tiles, flaps, header tanks, and a jettisonable nose fairing can do a direct transfer to Jupiter from VLEO (200 x 200). 

Hmm. What kind of payload could it carry?

I'm sure there are planetary protection issues with hanging out around Jupiter's moons*, but a Starship dive into Jupiter releasing a ton of atmospheric probes ahead of itself (like Galileo's, but modern and lots of them sampling different points) might be interesting. The Starship itself could relay data before crashing into Jupiter itself shortly afterwards, which would remove any planetary protection issues.

The Starship "bus/mothership" could carry huge solar panels, like the ones demonstrated at Jupiter on Juno but larger, while the atmospheric probes could be battery powered since they'd have only short lifetimes once separated - no RTGs needed.

If the communications issues from Jupiter could be handled, this might be doable on a fairly small budget as science missions go - the probes would be simple and mass-produced. This feels like it could be a combined Rocket Lab (probes) and SpaceX (Starship) effort.

*Except Callisto, I guess. The Starship could probably deliver a Callisto lander, then remove itself as a planetary protection concern by crashing into Callisto itself (sufficiently far from the lander to not interfere). But a Callisto lander probably would need plutonium - the large solar arrays needed at Jupiter's distance might not be practical for a lander, and even if so, it'd probably need at least RHUs to survive the night.

[TheRadicalModerate]

FWIW, a v3 Starship with no tiles, flaps, header tanks, and a jettisonable nose fairing can do a direct transfer to Jupiter from VLEO (200 x 200). 

Hmm. What kind of payload could it carry?

I'm sure there are planetary protection issues with hanging out around Jupiter's moons*, but a Starship dive into Jupiter releasing a ton of atmospheric probes ahead of itself (like Galileo's, but modern and lots of them sampling different points) might be interesting. The Starship itself could relay data before crashing into Jupiter itself shortly afterwards, which would remove any planetary protection issues.

The Starship "bus/mothership" could carry huge solar panels, like the ones demonstrated at Jupiter on Juno but larger, while the atmospheric probes could be battery powered since they'd have only short lifetimes once separated - no RTGs needed.

If the communications issues from Jupiter could be handled, this might be doable on a fairly small budget as science missions go - the probes would be simple and mass-produced. This feels like it could be a combined Rocket Lab (probes) and SpaceX (Starship) effort.

*Except Callisto, I guess. The Starship could probably deliver a Callisto lander, then remove itself as a planetary protection concern by crashing into Callisto itself (sufficiently far from the lander to not interfere). But a Callisto lander probably would need plutonium - the large solar arrays needed at Jupiter's distance might not be practical for a lander, and even if so, it'd probably need at least RHUs to survive the night.

The easiest thing?  Detach the payload after TJI, and send the Starship to anywhere but Jupiter.

Juno had a dry mass of 1.6t, bus and payload.  Wet mass was 3.6.  For Isp=310s, that's 2500m/s of delta, which had to support multiple flyby maneuvers.  The jovian insertion burn cost 540m/s.

Imagine what you could do with 5t bus, a 10t payload, and 25t of MMH/NTO.  That's about 3000m/s of delta-v, with no flyby maneuvers needed.

No need for a Starship at Jupiter.  Sure, it's less cool than a Starship that is at Jupiter, but it's very cool to have payload that 10x as big.

[Paul451]

Cross-post from the Surface Habitats on the Moon thread:

Q: Would starship be able to lay down on the surface of moon / mars to make a permanent habitat if needed? Linear seems better than vertical for long term use.

A: No

Note this is from 5 years ago and he may have changed his mind since then.

The Mars plan has changed in the intervening years, previously they wanted to get ships back from Mars, but now planning to leave most of the ships at Mars to be used locally. If you look at the 2024 presentation, there's a slide showing rows of horizontal cylindrical habitats looking suspiciously like laid down Starship tank sections.

"but now planning to leave most of the ships at Mars to be used locally"

Does that change the optimum design for a Mars ship?

AIUI, the original Starship concept was based on the observation that an SSTO vehicle capable of getting back to Earth-entry if refuelled on the surface of Mars, was also able to get to Mars surface with a significant payload if refuelled in Earth orbit, was also a suitable size for a reusable second stage for Earth launch to LEO.

But if you don't need to get them back (or the vast majority of them), then does the math change. For example, it takes less propellant to get from HLEO to Mars-entry than from Mars surface back to Earth-entry. If you don't need the return, a smaller/shorter version of Starship for Mars, with the same cargo.

In which case, would changing the proportions of the two Starship stages make a more efficient launch vehicle. Eg, moving more of the propellant to the first stage, having a shorter, lighter upperstage; the same total take-off mass (and hence T/W) being the same, but ending up with a better payload ratio. (Hence why most launchers have proportionally smaller upperstages than Starship.)


Aside: This still allows the shorty Starship to return small payloads to Earth (for eg, a few crew, surface samples, bio-samples), by doing refuelling in Mars orbit. One shorty as an SSTO Mars tanker, another as the launch-then-refuel-then-return vehicle. Mirroring Earth to Mars.

[meekGee]

Cross-post from the Surface Habitats on the Moon thread:

Q: Would starship be able to lay down on the surface of moon / mars to make a permanent habitat if needed? Linear seems better than vertical for long term use.

A: No

Note this is from 5 years ago and he may have changed his mind since then.

The Mars plan has changed in the intervening years, previously they wanted to get ships back from Mars, but now planning to leave most of the ships at Mars to be used locally. If you look at the 2024 presentation, there's a slide showing rows of horizontal cylindrical habitats looking suspiciously like laid down Starship tank sections.

"but now planning to leave most of the ships at Mars to be used locally"

Does that change the optimum design for a Mars ship?

AIUI, the original Starship concept was based on the observation that an SSTO vehicle capable of getting back to Earth-entry if refuelled on the surface of Mars, was also able to get to Mars surface with a significant payload if refuelled in Earth orbit, was also a suitable size for a reusable second stage for Earth launch to LEO.

But if you don't need to get them back (or the vast majority of them), then does the math change. For example, it takes less propellant to get from HLEO to Mars-entry than from Mars surface back to Earth-entry. If you don't need the return, a smaller/shorter version of Starship for Mars, with the same cargo.

In which case, would changing the proportions of the two Starship stages make a more efficient launch vehicle. Eg, moving more of the propellant to the first stage, having a shorter, lighter upperstage; the same total take-off mass (and hence T/W) being the same, but ending up with a better payload ratio. (Hence why most launchers have proportionally smaller upperstages than Starship.)


Aside: This still allows the shorty Starship to return small payloads to Earth (for eg, a few crew, surface samples, bio-samples), by doing refuelling in Mars orbit. One shorty as an SSTO Mars tanker, another as the launch-then-refuel-then-return vehicle. Mirroring Earth to Mars.

I think the most important observation is that a terrestrial second stage that leaves a low dV (reusable) booster also has the ability (once refueled) to inject from earth orbit to TMI.

There's a large range of trade-offs though.

What the vehicle has inherently is impulse, not dV.  You can trade payload mass, for example, with transit speed (dV and heat shield)

The ability to fly back directly from the surface of Mars is useful too of course, but IMO won't come into play immediately.

To maximize the trade-off space, I think you want to maximize the stored impulse.

Also from a risk management perspective, for every successful landing, you want the most stuff on the ground, both payload and empty ship. So that's another reason to maximize everything.

Earth-sode refueling will be relatively low risk, and occurs pre-mission, so IMO the more refills per ship the better.

[Twark_Main]

To maximize the trade-off space, I think you want to maximize the stored impulse.

Also from a risk management perspective, for every successful landing, you want the most stuff on the ground, both payload and empty ship. So that's another reason to maximize everything.

Clearly payload and "empty ship" aren't equal in value, however, otherwise why even try to maximize payload? Just maximize empty ship and have zero payload and that should be just as good.

So if Starship's tanks (and impulse) are oversized for the amount of payload mass Starship can actually EDL on Mars, then you're losing flexible payload mass and replacing it with inflexible empty tank mass.


In other words, if we assume Starship is expendable than the "correct" design should be matching the throw mass to Mars with the entry mass at Mars.

[meekGee]

To maximize the trade-off space, I think you want to maximize the stored impulse.

Also from a risk management perspective, for every successful landing, you want the most stuff on the ground, both payload and empty ship. So that's another reason to maximize everything.

Clearly payload and "empty ship" aren't equal in value, however, otherwise why even try to maximize payload? Just maximize empty ship and have zero payload and that should be just as good.

So if Starship's tanks (and impulse) are oversized for the amount of payload mass Starship can actually EDL on Mars, then you're losing flexible payload mass and replacing it with inflexible empty tank mass.


In other words, if we assume Starship is expendable than the "correct" design should be matching the throw mass to Mars with the entry mass at Mars.

We should treat tankage and some other components as useful payload, but clearly there is a desired ratio.  You don't want all tanks and no cargo, or all cargo and no tanks.

And then some flights prefer shorter transit, so you trade lower overall empty mass (cargo and ship) for more delta V.

I don't think they'll make different variants of ships just for this, but they might play with the other variables.  And they might have different internal configurations such crew+cargo or just cargo, dedicated ISRU units designed to operate as-landed vs ones that off-load cargo, etc.

When you're talking about hundreds of ships, many things are possible.

[Twark_Main]

To maximize the trade-off space, I think you want to maximize the stored impulse.

Also from a risk management perspective, for every successful landing, you want the most stuff on the ground, both payload and empty ship. So that's another reason to maximize everything.

Clearly payload and "empty ship" aren't equal in value, however, otherwise why even try to maximize payload? Just maximize empty ship and have zero payload and that should be just as good.

So if Starship's tanks (and impulse) are oversized for the amount of payload mass Starship can actually EDL on Mars, then you're losing flexible payload mass and replacing it with inflexible empty tank mass.


In other words, if we assume Starship is expendable than the "correct" design should be matching the throw mass to Mars with the entry mass at Mars.

We should treat tankage and some other components as useful payload, but clearly there is a desired ratio.  You don't want all tanks and no cargo, or all cargo and no tanks.

I can imagine scenarios for both of these actually, but yes I think in general we're in agreement here.

Unfortunately, that doesn't give a direct answer to the question of whether the optimum size of the tanks has become smaller with expendability. 

Could we imagine scenarios where the optimum tank size gets bigger, even? Eg a shortage of tanks/modules, such that you're just delivering oops-all-tank Starships to convert?

[meekGee]

To maximize the trade-off space, I think you want to maximize the stored impulse.

Also from a risk management perspective, for every successful landing, you want the most stuff on the ground, both payload and empty ship. So that's another reason to maximize everything.

Clearly payload and "empty ship" aren't equal in value, however, otherwise why even try to maximize payload? Just maximize empty ship and have zero payload and that should be just as good.

So if Starship's tanks (and impulse) are oversized for the amount of payload mass Starship can actually EDL on Mars, then you're losing flexible payload mass and replacing it with inflexible empty tank mass.


In other words, if we assume Starship is expendable than the "correct" design should be matching the throw mass to Mars with the entry mass at Mars.

We should treat tankage and some other components as useful payload, but clearly there is a desired ratio.  You don't want all tanks and no cargo, or all cargo and no tanks.

I can imagine scenarios for both of these actually, but yes I think in general we're in agreement here.

Unfortunately, that doesn't give a direct answer to the question of whether the optimum size of the tanks has become smaller with expendability. 

Could we imagine scenarios where the optimum tank size gets bigger, even? Eg a shortage of tanks/modules, such that you're just delivering oops-all-tank Starships to convert?

If we consider non-expandable ships, ones that have to come back, then the question becomes just how fast we want them back. Because the dV for the return leg determines the size of the ship.

However, I never understood how that was going to work. Not even when it was Carbon Composite ship. The timelines just don't add up.

With one way ships I think it almost always rolls to make them as large as possible. So the Booster determines just how large the empty ship can be. The bigger the better, and then as many refueling trips as possible.  Then it forks: Fat or Fast, which you control with the payload configuration.

(The tankage part remains tankage on the surface. The payload probably becomes hab space in some fashion, or some other structure.)

The rest (batteries, motors, pumps, actuators, engines and bells) are mostly configuration independent.

[wes_wilson]

Single drop tank inline and ahead of the ship coupled using the same hardpoints as the chopsticks and the megabay lifter.  Coupling structure may look a lot like the megabay lifter though would be structurally different because stress is compression not tension (plus needs to unfold after riding to orbit inside a ship payload bay).

Propellant flows down into the ship due to acceleration thrust so no pumps needed - keeps the discarded tank simpler and cheaper. Plumb the connections into the header tank lines just below the header tanks. Pierce the leeward side then cross inside the ship to the propellant lines on the windward side.

Would be an interesting analysis to figure out if an additional hardpoint on the leeward side is needed to adequately handle thrust vectoring stresses.

This seems like it's a lot harder to dock than a pusher would be.  Mind you, a pusher isn't easy to dock either, but its hard-capture system is just the existing Starship staging latches.  Soft-capture is still a problem, but it's less of a problem than guiding something on top of the pointy end.

Furthermore, a drop tank is inherently expendable, since it has no propulsion to get home.  A pusher, on the other hand, is reusable:  it just shuts down and stages with enough prop to return to LEO at the next perigee.  Fill it up at a depot and it's ready for the next mission.

Both are excellent points.  Thanks TRM.  I was reacting to crandles57.  A more complete introduction to my post would have been "if you're going to do a drop tank at all, use a single tank inline rather than trying to balance/fly several tanks arranged around the ship."

On reflection, the drop tank concept may still make sense if the mission profile requires a higher propellant mass fraction than a recoverable pusher can provide. Interesting option would be to 3D print the drop tank in orbit.  It's just a big lightweight tank with minimal rigidity except in the base plate, no insulation needed.  Building it in orbit eliminates structure needed to survive launch stress, enables making a tank bigger than you can launch, and using a dumb tank avoids the need to launch/maintain the engines electronics and valves in the pusher.   CONOPS: dock the ship to an empty tank while the tank is still held by the orbital factory, which provides the attitude thrusters and control needed for RPOD. Ship carries the tank away.  Depot/ship interact and transfer propellant as before, but the ship routes propellant into the drop tank in addition to the ship's internal tanks so it can accept several depot loads of propellant before departure from LEO.

Fast transfer to Jupiter?

A Super Heavy booster would make a quick/dirty inline pusher as it already has all the configuration needed for Starship to sit atop it.  SSTO a booster, refuel in leo, attach a starship, push it on out.   Am sure there are objections around docking/berthing sh/ss in orbit, sh not having refueling ports, inefficiencies in ISP, rockets aren't legos, etc.  But perhaps those challenges are less than designing a pusher from scratch, especially for infrequent use.

 Bonus Frankenconcept:  Could you align multiple SH's end to end?  SH/SH/SH/SS.  That's as franken as it gets.

[spacenut]

Well the booster has sea level engines not vacuum engines.  Also, once in orbit, you don't need 33 engines to push Starship out to somewhere.  This is why a lot of upper stage engines are lower thrust hydrolox engines for in space use.