Why hasn't there been an "all-in-one" (same vehicle) Mars architecture before?

[Pipcard]

The rationale behind an "all-in-one" Mars transportation architecture like Starship is that it is supposed to reduce development costs and that SpaceX can't afford to design and develop multiple, specialized vehicle types.

SpaceX chose the one-vehicle-does-everything (really one base vehicle with multiple variants) because it's easier - particularly in terms of financing. If your constellation and your regular launch customers are launching on your Mars vehicle, you can justify more development money for the Mars vehicle and can build and iterate it faster.

If it turns out that there is an easier way to get to Mars, SpaceX will certainly choose it. I don't see an easier way. Do you? The NASA way of many vehicles each with a special purpose is so expensive that even with 10x or 100x more money than SpaceX they will never make any real progress toward putting a person on Mars.

But every other Mars mission concept has always assumed the launch and assembly of various specialized components: Mars Transfer Vehicles, Mars Ascent Vehicles, Earth Return Vehicles, etc.

I see this discourse on Spaceflight Twitter ("Spitter"), and other spaceflight discussions all the time. So if it's such a sensible idea, why hasn't this been accepted before, and why is it still not accepted by some? Is it about technical risk? The risk of aerocapture? Doubts about the economics of reuse and refueling, or the effectiveness of ISRU? (to be fair, I was also thinking like this six years ago)

https://twitter.com/defconfuck/status/1500229783334100997

https://twitter.com/cowboytragedy/status/1485082240342183937

[Coastal Ron]

The rationale behind an "all-in-one" Mars transportation architecture like Starship is that it is supposed to reduce development costs and that SpaceX can't afford to design and develop multiple, specialized vehicle types.

SpaceX chose the one-vehicle-does-everything (really one base vehicle with multiple variants) because it's easier - particularly in terms of financing. If your constellation and your regular launch customers are launching on your Mars vehicle, you can justify more development money for the Mars vehicle and can build and iterate it faster.

If it turns out that there is an easier way to get to Mars, SpaceX will certainly choose it. I don't see an easier way. Do you? The NASA way of many vehicles each with a special purpose is so expensive that even with 10x or 100x more money than SpaceX they will never make any real progress toward putting a person on Mars.

But every other Mars mission concept has always assumed the launch and assembly of various specialized components: Mars Transfer Vehicles, Mars Ascent Vehicles, etc.

Remember though, the assumption before Elon Musk and SpaceX was that reusable launchers were impractical, and reusable spaceships that can land on Earth AND Mars were impossible.

I see this discourse on Spaceflight Twitter ("Spitter"), and other spaceflight discussions all the time. So why hasn't this been accepted before, and why is it still not accepted? Is it about technical risk? The risk of aerocapture? Doubts about the economics of reuse and refueling, or the effectiveness of ISRU? (to be fair, I was also thinking like this six years ago)

Until now no one thought that one vehicle could take off from Earth, transit to Mars, land on Mars, and then take off again, transit to Earth, and then land back on Earth. That is a LOT of transportation segments to do with one vehicle, but that is what the Starship transportation system is planned to be able to do.

Add on top of that the amount of cargo that it can carry while doing all of that, and it becomes clear that this approach could drive down the cost of getting humans and cargo to Mars by simplifying the transportation system.

However SpaceX has not proved it can do all of those transportation segments, but the reasoning for trying seems sound.

As venture capitalist John Doerr is famous for saying:

Ideas are easy. Execution is everything. It takes a team to win.

So another way to answer your question is that until now (or recently), SpaceX hadn't shown the world it was possible to use one reusable vehicle for a Mars transportation architecture.

[su27k]

Because all the previous Mars architecture were thought up by governments, they don't care about cost, and they can't easily sell the excess launches/services due to various reasons (laws preventing government from competing with private companies, unable to expand market using something like Starlink, turf war between different parts of the government i.e. NASA vs DoD, etc)

[sdsds]

Because SpaceX saw that ISRU propellant could be used to fill a single-stage methalox vehicle capable of getting from the Martian surface to Mars orbit. Once you believe that, most of the remaining aspects of the original "MCT" re-usability architecture fall into place.
Of course it also helps if you have an iron-clad belief that lifting commodity propellant to LEO is going to be super-cheap.
Underlying that is the observation Musk made in 2019 that it would take around one million tons of cargo to build a self-sustaining settlement on Mars. No one else cared about that, because no one else thought making life multi-planetary was the goal of Mars missions, or at least that it could be anywhere on the planning horizon.

[M.E.T.]

Because previous missions were for small scale exploration, while SpaceX’s goal is mass colonization.

So flight numbers are orders of magnitude higher, hence the need for uniformity and economies of scale.

[volker2020]

It is no question, that from a technical perspective, you could build vehicles better suited for a certain part of the mars mission. But the one important thing that is missing in his analysis is dollar.

SpaceShip will allow to transport 150t of mass to mars for good money. And in the end, if you are serious about colonizing mars, that is the only metric interesting.

I admit, I would really love to see a nuclear cycler, grabbing 10 SpaceShips and move them to mars, but while I really like the idea, this stops the moment I think about short term costs. If we had some governments supporting this idea and adding maybe 30 Billion to development, well maybe. Until than ...

[mikelepage]

In principle, I agree with the posters you've quoted - that Starship will eventually be best used as an atmospheric launch/landing craft at Earth/Mars, and for interplanetary cargo transport.

The problem as I see it is that inert cargo is fine over interplanetary transit timeframes because it doesn't need much power and doesn't generate much heat, but the large, space-optimised solar arrays and thermal rejection systems required for long-duration human spaceflight will take up way too much space on Starships that will spend the majority of each mission on the ground at Mars. Remember that Starship hasn't even tried to address the solar power/thermal problems yet, and the animations that depict these massive (but still way too small) solar arrays folding up neatly into tiny parts of Starship's cargo bay aren't believable as yet.

What you could have as well as Starship, would be a Starship-derived and constructed interplanetary vehicle that uses Methalox/Raptor propulsion and can use the Starship refuelling infrastructure, but can take a much greater number of people at a time due to extra power and heat management capacity, and provides significant comfort/safety improvements over and above Starship, like spin-gravity and better radiation protection. I'm imagining something like the baton structure depicted in the Vast company materials. You could imagine these travelling with the fleet when it goes - say one crewed ship for every ten cargo Starships. People would split up and transfer to the starships only in the days before EDL at Mars, then refuel it and take it with them when they return to Earth.

The key sticking point is that in order for it to be reusable, you need it to be able to do aerocapture at the end of each transit, and I don't know enough about how much heat shielding is required for something of those dimensions to brake into orbit.

EDIT: It just occurred to me that, as long as the crew transfer to the Starships and do a direct entry, the interplanetary vessel could spend the best part of an Earth-Mars synodic period aerobraking into low Mars orbit in preparation for the return journey. So maybe the heat-shielding for aerobraking doesn't need to be all that robust - just the bare minimum to get into high Mars orbit.   

[M.E.T.]

It is no question, that from a technical perspective, you could build vehicles better suited for a certain part of the mars mission. But the one important thing that is missing in his analysis is dollar.

SpaceShip will allow to transport 150t of mass to mars for good money. And in the end, if you are serious about colonizing mars, that is the only metric interesting.

I admit, I would really love to see a nuclear cycler, grabbing 10 SpaceShips and move them to mars, but while I really like the idea, this stops the moment I thing about short term costs. If we had some governments supporting this idea and adding maybe 30 Billion to development, well maybe. Until than ...

Forgive my ignorance. How does a cycler collect the 10 Starships in Earth orbit? Does the cycler slow down or do the Starships speed up to match the cycler’s velocity? If the latter, zero fuel is saved compared to the Starship flying to Mars independently.

And since the Starships still need to atmospherically brake and propulsively land on Mars, no heat shield or landing fuel mass is saved either.

So what is the benefit of using a nuclear cycler for the middle part of the journey?

[dglow]

Because SpaceX saw that ISRU propellant could be used to fill a single-stage methalox vehicle capable of getting from the Martian surface to Mars orbit. Once you believe that, most of the remaining aspects of the original "MCT" re-usability architecture fall into place.
Of course it also helps if you have an iron-clad belief that lifting commodity propellant to LEO is going to be super-cheap.
Underlying that is the observation Musk made in 2019 that it would take around one million tons of cargo to build a self-sustaining settlement on Mars. No one else cared about that, because no one else thought making life multi-planetary was the goal of Mars missions, or at least that it could be anywhere on the planning horizon.

Mars Direct had the same realization re: an ISRU-propelled return vehicle yet didn’t follow the all-in-one architecture. Heck, Zubrin continues to rail against landing a full Starship on Mars. What Musk and SpaceX realized is that a fully-reusable vehicle reduces how aggressively you need to play the mass-optimization game.

That realization, plus the desire to mass-produce these vehicles, opened them up to interesting tradeoffs – like stainless steel.

[eriblo]

As soon as you are doing aerocapture you need a proper heat shield. Peak temperature occurs before peak acceleration so the temperatures will be similar to direct entry (although duration is shorter).

[kevinof]

The only company to step up with a Mars program is SpaceX - there is no other alternative design (only those on paper) and SpaceX is going with something that suits their needs both in LEO and for Mars. So a one design platform tweaked to work for all situations fits their business and engineering needs. If mars was the only destination and they were spending someone’s else’s money then then design might have been different. Their design reflects their commercial needs elsewhere but is till a fantastic design for mars.

Nothing stopping others from spending their $$$ and building their own platforms based on their architecture tradeoffs but I have my doubts we will see much of anything.

[mikelepage]

In principle, I agree with the posters you've quoted - that Starship will eventually be best used as an atmospheric launch/landing craft at Earth/Mars, and for interplanetary cargo transport.

The problem as I see it is that inert cargo is fine over interplanetary transit timeframes because it doesn't need much power and doesn't generate much heat, but the large, space-optimised solar arrays and thermal rejection systems required for long-duration human spaceflight will take up way too much space on Starships that will spend the majority of each mission on the ground at Mars. Remember that Starship hasn't even tried to address the solar power/thermal problems yet, and the animations that depict these massive (but still way too small) solar arrays folding up neatly into tiny parts of Starship's cargo bay aren't believable as yet.

What you could have as well as Starship, would be a Starship-derived and constructed interplanetary vehicle that uses Methalox/Raptor propulsion and can use the Starship refuelling infrastructure, but can take a much greater number of people at a time due to extra power and heat management capacity, and provides significant comfort/safety improvements over and above Starship, like spin-gravity and better radiation protection. I'm imagining something like the baton structure depicted in the Vast company materials. You could imagine these travelling with the fleet when it goes - say one crewed ship for every ten cargo Starships. People would split up and transfer to the starships only in the days before EDL at Mars, then refuel it and take it with them when they return to Earth.

The key sticking point is that in order for it to be reusable, you need it to be able to do aerocapture at the end of each transit, and I don't know enough about how much heat shielding is required for something of those dimensions to brake into orbit.

EDIT: It just occurred to me that, as long as the crew transfer to the Starships and do a direct entry, the interplanetary vessel could spend the best part of an Earth-Mars synodic period aerobraking into low Mars orbit in preparation for the return journey. So maybe the heat-shielding for aerobraking doesn't need to be all that robust - just the bare minimum to get into high Mars orbit.

Just realised this specialisation of functions helps another way too. If your starship fleet travels with an interplanetary transfer-only vessel that aerobrakes into orbit and doesn't have to land, it means that the fleet has one ship that doesn't have any propellant to keep cold either. Being in the sun constantly, with big solar arrays, means it can cast a shadow over the other vessels in the fleet, meaning much less work to keep the propellant cold in the Starships which do have to land. In this scenario, since your crew is in the transfer vessel, you wouldn't even need to keep the temperature in the landing Starship crew cabins all that warm, just warm enough not to damage anything in storage. 

So to take this a bit further, let's imagine this interplanetary transfer vessel is - for simplicity's sake - a standard-sized Starship - but instead of the smaller "Elonerons" and a full complement of heat shield tiles of the standard Starship, it has much larger, inflatable "Elonerons" that - during launch - are held tightly on either side of the ship. Once the TMI burn has been performed, the Elonerons inflate to butterfly-like proportions relative to the rest of the ship, with a heat shielding surface on one side (a la LOFTID), and the solar arrays on the other. The transfer ship is then ready to take on crew, to commence spin-G for the cruise phase, whilst also casting a shadow over the rest of the fleet. 

Days or weeks prior to EDL at Mars, the spin is stopped and the crew transfers back to the landing Starships and prepare for landing. The transfer Starship leaves the Elonerons out to perform the aerobrake maneuver, but takes a much higher path through the Mars atmosphere, just enough to brake into a highly elliptical orbit.

As a side note: this scenario would make it easier to specialise this transfer Starship for spin gravity, because unlike previous debates around tail to tail spin gravity, this craft would never be used "the right way up" on Mars, so crew quarters would be designed for nose-down operations.
 

As soon as you are doing aerocapture you need a proper heat shield. Peak temperature occurs before peak acceleration so the temperatures will be similar to direct entry (although duration is shorter).

Okay. But heating is proportional to the density of atmosphere you're moving through right? So the bigger your surface area (e.g. inflatable Elonerons) the higher you can pass through in the atmosphere, so lower overall heating sustained for equivalent delta V, yes?

[jdon759]

It is no question, that from a technical perspective, you could build vehicles better suited for a certain part of the mars mission. But the one important thing that is missing in his analysis is dollar.

SpaceShip will allow to transport 150t of mass to mars for good money. And in the end, if you are serious about colonizing mars, that is the only metric interesting.

I admit, I would really love to see a nuclear cycler, grabbing 10 SpaceShips and move them to mars, but while I really like the idea, this stops the moment I thing about short term costs. If we had some governments supporting this idea and adding maybe 30 Billion to development, well maybe. Until than ...

Forgive my ignorance. How does a cycler collect the 10 Starships in Earth orbit? Does the cycler slow down or do the Starships speed up to match the cycler’s velocity? If the latter, zero fuel is saved compared to the Starship flying to Mars independently.

And since the Starships still need to atmospherically brake and propulsively land on Mars, no heat shield or landing fuel mass is saved either.

So what is the benefit of using a nuclear cycler for the middle part of the journey?

The typical idea is that vehicles will speed up to match the cycler.  The advantage is that the vehicles can then be designed to be lightweight (relative to the cycler) or conversely, the cycler can be very very massive; not needing to worry about spending propellant at each arrival/departure.  This also provides a very large volume for the comfort of the crew during the long voyage, and transfer vehicles do not waste power and consumables - as these are provided by the cycler.

[Pipcard]

Of course it also helps if you have an iron-clad belief that lifting commodity propellant to LEO is going to be super-cheap.

Starship critics/skeptics expect it to cost something like $100 million a launch, or that launch cadence will not be rapid, or that it can be cheap but that cryogenic propellant transfer won't be practical. That is why they may view "the need to refuel multiple times to leave LEO" as a disadvantage and advocate for high-energy third stages and dedicated MTV assembly.

[InterestedEngineer]

Why hasn't there been an "all-in-one" (same vehicle) Mars architecture before?

Here's why

Elon:

The extreme difficulty of scaling production of new technology is not well understood. It's 1000% to 10,000% harder than making a few prototypes. The machine that makes the machine is vastly harder than the machine itself.

Creating a custom Mars looper, a Mars Gateway, a Mars Lander that could handle the volume needed for colonization would be three times harder than creating one machine that produces variants of a single proven implementation.   

There are only a few system engineers at Elon's level capable of architecting the machine that builds the machine,  So it probably wouldn't even happen if the task was three times harder.

[DanClemmensen]

Why hasn't there been an "all-in-one" (same vehicle) Mars architecture before?

Here's why

Elon:

The extreme difficulty of scaling production of new technology is not well understood. It's 1000% to 10,000% harder than making a few prototypes. The machine that makes the machine is vastly harder than the machine itself.

Creating a custom Mars looper, a Mars Gateway, a Mars Lander that could handle the volume needed for colonization would be three times harder than creating one machine that produces variants of a single proven implementation.   

There are only a few system engineers at Elon's level capable of architecting the machine that builds the machine,  So it probably wouldn't even happen if the task was three times harder.

There are many, many brilliant system architects that can do this. The problem is that there are no projects of this complexity that are under control of a single top-level system architect, except SpaceX. That's because the funding level required for these projects is so high that the project must be subdivided at the political level and parceled out to separate organizations without a top-level architecture being finalized by a competent system architect. The best the resulting pool of system architects can do then is negotiate among themselves to create proper formal interfaces among the top-level subsystems, and this process is intensely political as it affects the separate responsibilities and interests of the various stakeholders instead of being focused on the best top-level architecture.

[InterestedEngineer]

Why hasn't there been an "all-in-one" (same vehicle) Mars architecture before?

Here's why

Elon:

The extreme difficulty of scaling production of new technology is not well understood. It's 1000% to 10,000% harder than making a few prototypes. The machine that makes the machine is vastly harder than the machine itself.

Creating a custom Mars looper, a Mars Gateway, a Mars Lander that could handle the volume needed for colonization would be three times harder than creating one machine that produces variants of a single proven implementation.   

There are only a few system engineers at Elon's level capable of architecting the machine that builds the machine,  So it probably wouldn't even happen if the task was three times harder.

There are many, many brilliant system architects that can do this. The problem is that there are no projects of this complexity that are under control of a single top-level system architect, except SpaceX. That's because the funding level required for these projects is so high that the project must be subdivided at the political level and parceled out to separate organizations without a top-level architecture being finalized by a competent system architect. The best the resulting pool of system architects can do then is negotiate among themselves to create proper formal interfaces among the top-level subsystems, and this process is intensely political as it affects the separate responsibilities and interests of the various stakeholders instead of being focused on the best top-level architecture.

You are more optimistic about system architects than I am.   I am a system architect, and very humbled by Elon's ability, and have rarely met a system architect who amazes me.

But yes, the system architect having the money to do the project also puts him in charge, instead of the politicians and the program managers.   And that is what it takes, and that combo is what is extremely rare.   Now go find three of those, each with their own shareholder profit motives that have to be satisfied.

Will never happen.

[Vultur]

Of course it also helps if you have an iron-clad belief that lifting commodity propellant to LEO is going to be super-cheap.

Starship critics/skeptics expect it to cost something like $100 million a launch, or that launch cadence will not be rapid, or that it can be cheap but that cryogenic propellant transfer won't be practical. That is why they may view "the need to refuel multiple times to leave LEO" as a disadvantage and advocate for high-energy third stages and dedicated MTV assembly.

And to be fair to the skeptics/critics, the very low cost and high cadence probably requires rapid, cheap (low refurbishment) tanker reuse, ie not like Shuttle TPS etc. That's kind of undemonstrated.

I mean I doubt it's going to be 100 million a launch even with expendable tankers, but maybe 50 million or something.

[DanClemmensen]

Why hasn't there been an "all-in-one" (same vehicle) Mars architecture before?

Here's why

Elon:

The extreme difficulty of scaling production of new technology is not well understood. It's 1000% to 10,000% harder than making a few prototypes. The machine that makes the machine is vastly harder than the machine itself.

Creating a custom Mars looper, a Mars Gateway, a Mars Lander that could handle the volume needed for colonization would be three times harder than creating one machine that produces variants of a single proven implementation.   

There are only a few system engineers at Elon's level capable of architecting the machine that builds the machine,  So it probably wouldn't even happen if the task was three times harder.

There are many, many brilliant system architects that can do this. The problem is that there are no projects of this complexity that are under control of a single top-level system architect, except SpaceX. That's because the funding level required for these projects is so high that the project must be subdivided at the political level and parceled out to separate organizations without a top-level architecture being finalized by a competent system architect. The best the resulting pool of system architects can do then is negotiate among themselves to create proper formal interfaces among the top-level subsystems, and this process is intensely political as it affects the separate responsibilities and interests of the various stakeholders instead of being focused on the best top-level architecture.

You are more optimistic about system architects than I am.   I am a system architect, and very humbled by Elon's ability, and have rarely met a system architect who amazes me.

But yes, the system architect having the money to do the project also puts him in charge, instead of the politicians and the program managers.   And that is what it takes, and that combo is what is extremely rare.   Now go find three of those, each with their own shareholder profit motives that have to be satisfied.

Will never happen.

If Elon needed to accomplish a task that took thrice the resources, he would start by accumulating the resources. But more likely he would start by finding a cheaper way to do the task, or find an achievable task. His self-selected task is to colonize Mars. his approach appears to be feasible, but is far cheaper than any other realistic approach that I have seen. It also helps to work on a problem in a simple domain that you can truly understand. He's not trying to solve world hunger, world peace, etc. He did take a shot at global warming with Tesla, but it's beyond the reach of a single self-funded system architect.

[pathfinder_01]

Of course it also helps if you have an iron-clad belief that lifting commodity propellant to LEO is going to be super-cheap.

Starship critics/skeptics expect it to cost something like $100 million a launch, or that launch cadence will not be rapid, or that it can be cheap but that cryogenic propellant transfer won't be practical. That is why they may view "the need to refuel multiple times to leave LEO" as a disadvantage and advocate for high-energy third stages and dedicated MTV assembly.

Refueling when it comes to Mars is a non-issue
.  The reason NASA preferred MTV is because the wanted an  Apollo  like Approach and thought that going into Mars Orbit and not direct landing was the safest way to Mars. Mars pretty much demands refueling if you are going to do it via MTV and single launch architectures with high energy third stages are expensive and limited.

 NASA's plans tended to be build a MTV big enough to hold everything the crew would need for a 2 year mission to Mars and back. The lander may use ISRU or be carried with the MTV or another MTV sent ahead.
By going into orbit you take some risk out such as bad weather at landing site and you don't need to land 2 years of supplies at once.

Elon can't go this way. There is no commercial use for a MTV or expendable three stage high energy rockets. This would be too costly for him. He needs to go another way.

With Starship you could preposition supplies on the surface and he is willing to take the risk of needing ISRU to get the crew back. Most importantly Starship must have commercial uses or else it will be too expensive for Elon.

NASA is not in the business of launching anything other than manned missions and something like starship wouldn't be attractive.