Possible cost-reduction possibilities for the NASA portions of MSR

[deltaV]

Regardless of the exact cost of Altair, we know that commercial crew, HLS, and MSR have costs of the same order of magnitude. Commercial crew has already succeeded. HLS attracted several bidders and seems to be on track to finish successfully (albeit likely a little late). So it seems probable that fixed-price MSR would also work.

If people aren't convinced why not give it a try and see. NASA can do an RFP that allows both fixed-price and cost-plus bids like the ISS deorbit RFP did. NASA should put in the RFP that it will be a lot more skeptical when judging the cost estimates in cost-plus bids than those in fixed-price bids and if this isn't fair to cost-plus that's working as intended. NASA should allow bidders to submit two bids, one cost-plus and one fixed-price, or structure it as separate fixed-price and cost-plus procurements, so bidders who can't predict how NASA will compare the two types of bid can play it safe and submit both types.

[thespacecow]

There is no such thing as commercial mars anything. Its 100% government funded.

Wrong, it's not 100% government funded. For one thing, SpaceX self funded Red Dragon and early Raptor/Starship R&D.

And Red Dragon landed on -- oh wait, it went nowhere beyond powerpoint.

And? So? You're telling me 100% of government funded Mars mission landed on Mars? What happened to NASA's Mars DRM 1 to 5? Oh wait, "it went nowhere beyond powerpoint"...

BTW, Red Dragon went far beyond powerpoint, they built propulsive landing into Dragon 2, and now we know it's even somewhat functional such that NASA/SpaceX has enabled it as a backup to parachutes.

[thespacecow]

It's totally ridiculous to claim only cost+ can do MSR. Fact is, MSR is no harder than HLS, given MSR is estimated to be $10B, while NASA's estimate for Altair is $12B. And we have not one but two companies building HLS under fixed cost service contract.

If you actually believe that NASA could execute an Altair lander today for $12 billion, I have a few prime bridges in New York harbor to sell you.

Yes, if NASA uses centers and cost plus contractors, they'd be lucky to get it done for $24B. Which just re-enforces my original point that MSR is just not that hard comparing to projects commercial companies are already doing.

[thespacecow]

Could SpaceX do it more cheaply than JPL? Only if they could use the SpaceX development method, which involves rapid iterative development and lots of failed flight tests from which data is collected. There are real obstacles to doing that at Mars, because of the time it takes to get there and the difficulty of returning telemetry.

The obstacles can be overcome:

1. SpaceX's iterative development process doesn't have to have lots of failed test flights, they didn't need it for F9 and Dragon for example.

And now we have another great example in SpaceX being able to catch SuperHeavy booster successfully on the first try.

If you think SpaceX has to have multiple Mars windows just to get Mars EDL right, think again.

[Eric Hedman]

Could SpaceX do it more cheaply than JPL? Only if they could use the SpaceX development method, which involves rapid iterative development and lots of failed flight tests from which data is collected. There are real obstacles to doing that at Mars, because of the time it takes to get there and the difficulty of returning telemetry.

The obstacles can be overcome:

1. SpaceX's iterative development process doesn't have to have lots of failed test flights, they didn't need it for F9 and Dragon for example.

And now we have another great example in SpaceX being able to catch SuperHeavy booster successfully on the first try.

If you think SpaceX has to have multiple Mars windows just to get Mars EDL right, think again.

Over confidence is one of the deadly sins of engineering.  However well they prepare for the first Mars EDL, there is never a hundred percent guarantee something this complex will work perfect on the first try.

[deltaV]

Over confidence is one of the deadly sins of engineering.  However well they prepare for the first Mars EDL, there is never a hundred percent guarantee something this complex will work perfect on the first try.

SpaceX sometimes succeeds on the first try, e.g. Falcon 9, Falcon Heavy, cargo Dragon, crew Dragon, and Starship booster catch, and sometimes takes multiple tries, e.g. Falcon 1, Falcon 9 booster recovery, Starship launch, and Starship upper stage reentry. So "this test will definitely succeed" and "this test will definitely fail" are both unreasonable opinions.

[TrevorMonty]

Over confidence is one of the deadly sins of engineering.  However well they prepare for the first Mars EDL, there is never a hundred percent guarantee something this complex will work perfect on the first try.

SpaceX sometimes succeeds on the first try, e.g. Falcon 9, Falcon Heavy, cargo Dragon, crew Dragon, and Starship booster catch, and sometimes takes multiple tries, e.g. Falcon 1, Falcon 9 booster recovery, Starship launch, and Starship upper stage reentry. So "this test will definitely succeed" and "this test will definitely fail" are both unreasonable opinions.

They have two different engineering approaches depending on project and customer.
1) Suck and see approach which is learn by failure eg SS. SpaceX internal project.
2) More cautious standard aerospace approach with careful and thorough design eg Dragon. NASA was customer and paying for this project.

[DanClemmensen]

Over confidence is one of the deadly sins of engineering.  However well they prepare for the first Mars EDL, there is never a hundred percent guarantee something this complex will work perfect on the first try.

SpaceX sometimes succeeds on the first try, e.g. Falcon 9, Falcon Heavy, cargo Dragon, crew Dragon, and Starship booster catch, and sometimes takes multiple tries, e.g. Falcon 1, Falcon 9 booster recovery, Starship launch, and Starship upper stage reentry. So "this test will definitely succeed" and "this test will definitely fail" are both unreasonable opinions.

They have two different engineering approaches depending on project and customer.
1) Suck and see approach which is learn by failure eg SS. SpaceX internal project.
2) More cautious standard aerospace approach with careful and thorough design eg Dragon. NASA was customer and paying for this project.

SpaceX appears to be using a hybrid approach for Starship HLS, and this would likely apply for MSR: build a major new system for internal use using "suck and see", and then build custom derivatives using an approach that meets customer requirements as specified in the contract.

[thespacecow]

Could SpaceX do it more cheaply than JPL? Only if they could use the SpaceX development method, which involves rapid iterative development and lots of failed flight tests from which data is collected. There are real obstacles to doing that at Mars, because of the time it takes to get there and the difficulty of returning telemetry.

The obstacles can be overcome:

1. SpaceX's iterative development process doesn't have to have lots of failed test flights, they didn't need it for F9 and Dragon for example.

And now we have another great example in SpaceX being able to catch SuperHeavy booster successfully on the first try.

If you think SpaceX has to have multiple Mars windows just to get Mars EDL right, think again.

Over confidence is one of the deadly sins of engineering.  However well they prepare for the first Mars EDL, there is never a hundred percent guarantee something this complex will work perfect on the first try.

Of course not, there's never a 100% guarantee something will work for anything humans build, that's not what's being discussed here. The point here is that SpaceX is no less capable than old space - e.g. JPL - when it comes to getting novel technology working on the first try, if they wanted to. What they just did is no less complex than skycrane for example.

[thespacecow]

https://twitter.com/SpcPlcyOnline/status/1848497511104655772

Note that SpaceX proposal is listed under "End-to-End Mission Architecture", this proves those who claim SpaceX only proposed Starship as launch vehicle are quite wrong, as I have been saying all along.

[deltaV]

https://twitter.com/SpcPlcyOnline/status/1848497511104655772

That slide says "Final reports received October 15". I wonder if that includes Rocket Lab, who got their award late on August 24 (source: RASMSR selections doc with abstracts) with contract signed September 27 or so (source: https://spacenews.com/nasa-awards-rocket-lab-study-contract-for-mars-sample-return/). Maybe RL was sufficiently interested that they didn't let not getting an award stop them from beginning the study? Or maybe they squeezed a 90 day study into ~50 days.

[Spiceman]

https://twitter.com/SpcPlcyOnline/status/1848497511104655772

Note that SpaceX proposal is listed under "End-to-End Mission Architecture", this proves those who claim SpaceX only proposed Starship as launch vehicle are quite wrong, as I have been saying all along.

And it has been debunked, all along. https://forum.nasaspaceflight.com/index.php?topic=59133.msg2601282#msg2601282

[VSECOTSPE]

It's totally ridiculous to claim only cost+ can do MSR. Fact is, MSR is no harder than HLS, given MSR is estimated to be $10B, while NASA's estimate for Altair is $12B.

MSR is much harder than HLS in terms of the distances involved (200x farther), mission duration (years versus weeks), landing precision (almost 2x more precise), planetary protection (the highest possible categories versus practically none), EDL (atmosphere versus none), operations (20+ minutes of comms time delay versus a couple seconds), etc.

The only way in which HLS is harder than MSR is that HLS involves astronauts and the systems and limits necessary to keep them alive.

None of this means that a particular contract can or cannot be used.

But it’s ignorant of the physical facts to argue that two engineering projects are about the same difficulty because they cost about the same to develop.  That’s like saying that a Bugatti hypercar and a McMansion involve the same difficulty of engineering because they both cost a few million bucks.  They don’t.

Which just re-enforces my original point that MSR is just not that hard comparing to projects commercial companies are already doing.

I’m sorry, but this is just baloney.  These companies for the most part are revisiting and remaking capabilities that the government proved out decades ago.  The private sector is doing it much more efficiently and at larger scales than the public sector did or could do.  But we’ve had suborbital human space flight, medium-lift and heavy-lift launch, crew capsules, and satellite constellations before.  Even seemingly novel capabilities like boost-back, vertically-landed, reusable launch stages have precursors in government work like DC-X, any number of planetary landers, various civil and defense entry vehicles, SLI, and most recently the Lunar Lander Challenge.  (I was involved with the last two.)

MSR requires capabilities that have never been done before by anyone.  The specifics depend on the architecture, but MSR will involve some combination of the first supersonic retropropulsion at another planet, the first launch from another planet, the first rendezvous/docking/transfer around another planet, and/or the first propellant production on another planet.  None of these are trivial things. 

That doesn’t mean that a private entity is not capable of doing them.  But if NASA asks a company or companies to do some or all of these, NASA has to be cognizant of, and deliberate about, what it is asking the private sector to do.  It’s not asking the private sector to reinvent the wheel — which is what we asked the private sector to do when we started commercial cargo and commercial crew.  It’s asking the private sector to invent the wheel for the first time.  That is a much bigger ask, and it should be reflected in terms of the procurement strategy, funding, program development and testing, technical support, and schedule.  Otherwise, we’re setting up MSR and the companies involved for another failure.

The obstacles can be overcome:

1. SpaceX's iterative development process doesn't have to have lots of failed test flights, they didn't need it for F9 and Dragon for example.

This is woefully ignorant of the Falcon 1 launch failures, the Dragon Spacecraft Qualification Unit, the partial failure on the first commercial Falcon 9 launch, all the Grasshopper testing, and the six reusable first stage losses before the first successful landing.

If you think SpaceX has to have multiple Mars windows just to get Mars EDL right, think again.

SpaceX thinks they need multiple windows.  Their proposal to NASA includes a test landing of Starship at Mars in 2029 before undertaking the MSR mission.  See:

https://wccftech.com/spacexs-planned-mars-landing-timeline-revealed-in-nasa-document/

The point here is that SpaceX is no less capable than old space - e.g. JPL - when it comes to getting novel technology working on the first try, if they wanted to. What they just did is no less complex than skycrane for example.

Again, this is woefully ignorant of the testing that went into both systems.

Note that SpaceX proposal is listed under "End-to-End Mission Architecture", this proves those who claim SpaceX only proposed Starship as launch vehicle are quite wrong, as I have been saying all along.

This is not news.  We’ve known SX has proposed to use Starship for all segments of the mission since June.  See link above.

[VSECOTSPE]

That slide says "Final reports received October 15". I wonder if that includes Rocket Lab, who got their award late on August 24 (source: RASMSR selections doc with abstracts) with contract signed September 27 or so (source: https://spacenews.com/nasa-awards-rocket-lab-study-contract-for-mars-sample-return/). Maybe RL was sufficiently interested that they didn't let not getting an award stop them from beginning the study? Or maybe they squeezed a 90 day study into ~50 days.

Whatever happened with RocketLab’s proposal, SpaceNews is reporting that RocketLab delivered their report on October 15 with the other seven industry studies:

NASA received Oct. 15 the final reports of 12 studies exploring different approaches to MSR, said Jeff Gramling, MSR program director at NASA Headquarters, at an Oct. 21 meeting of the Committee on Astrobiology and Planetary Sciences of the National Academies’ Space Studies Board.

Eight of the studies were from industry, seven awarded in June and an eighth to Rocket Lab made in August but only announced Oct. 7.

https://spacenews.com/nasa-on-track-to-decide-new-approach-for-msr-by-end-of-year/

RocketLab is advertising $2B for a seemingly conventional MSR architecture that launches in 2028 and returns samples 2031-2033.  It involves entirely new hardware — two Neutron launches and apparently their own lander, MAV, and ERO. But they’ve proven they can get new LV and interplanetary spacecraft to the pad, and this kind of architecture would seem to avoid unconventional approaches to issues like planetary protection that could be a showstopper for other architectures.  I wouldn’t bet the entire MSR effort on RocketLab’s proposal.  But if they pass technical review, for $2B, I’d make a side bet on them, and if RocketLab gets to Mars first, then they get to bring the samples home.  I’d especially make that side bet if it involved test flights before touching the samples, even if that cost a billion or so more and a couple more years.

[thespacecow]

https://twitter.com/SpcPlcyOnline/status/1848497511104655772

Note that SpaceX proposal is listed under "End-to-End Mission Architecture", this proves those who claim SpaceX only proposed Starship as launch vehicle are quite wrong, as I have been saying all along.

And it has been debunked, all along. https://forum.nasaspaceflight.com/index.php?topic=59133.msg2601282#msg2601282

And that debunk has now been proven to be completely wrong by reality, just as I predicted back then.

Poptip: You can't debunk reality.

[VSECOTSPE]

And that debunk has now been proven to be completely wrong by reality, just as I predicted back then.

I don’t mean to be overly critical, but your old post wasn’t making a prediction about SX.  It just argued that all the industry studies had to address all segments of the MSR mission:

The study is for an entire MSR mission from start to finish, you can't just tell NASA "here's one element, you figure out the rest". This should be obvious since NASA needs to compare the cost/schedule/etc for the entire mission, they can't do this if you just gave them one element without the rest. So if you only propose one element, then the rest will be filled in by elements from POR, and this is not competitive against POR if your one element is just a bigger LV.

This was untrue back then and it still is today.

Moreover, it became public that same month (the same week, in fact) that SX proposed to use SS for all MSR mission segments:

https://wccftech.com/spacexs-planned-mars-landing-timeline-revealed-in-nasa-document/

So even if you had made a prediction, it had a shelf-life of at most a few days.

Poptip: You can't debunk reality.

The English idiom is “pro tip”, not “poptip”.

[thespacecow]

It's totally ridiculous to claim only cost+ can do MSR. Fact is, MSR is no harder than HLS, given MSR is estimated to be $10B, while NASA's estimate for Altair is $12B.

MSR is much harder than HLS in terms of the distances involved (200x farther), mission duration (years versus weeks), landing precision (almost 2x more precise), planetary protection (the highest possible categories versus practically none), EDL (atmosphere versus none), operations (20+ minutes of comms time delay versus a couple seconds), etc.

The only way in which HLS is harder than MSR is that HLS involves astronauts and the systems and limits necessary to keep them alive.

None of this means that a particular contract can or cannot be used.

But it’s ignorant of the physical facts to argue that two engineering projects are about the same difficulty because they cost about the same to develop.  That’s like saying that a Bugatti hypercar and a McMansion involve the same difficulty of engineering because they both cost a few million bucks.  They don’t.

The physical facts are already taken into account by the estimate, as long as the estimate is relatively accurate and the teams on the projects are of similar competence. If a project is more difficult, it just meant it'll need more engineering hours, more testing, etc, all of which are reflected in cost estimates.

And yes, the human factors in HLS makes it harder than MSR, but it's not just ECLSS ("keep crew alive"), it's also the payload mass. HLS has orders of magnitude more payload mass than MSR, which necessitate a much bigger spacecraft, bigger engines, and in case of HLS also includes bigger launch vehicles and/or orbital refueling.

As for Bugatti hypercar and McMansion, that's comparing apples to oranges. Each Bugatti car only carries a small fraction of the overall development cost, since they amortize the cost over many cars. If you actually bother to do some research, you'll see the development cost for a new Bugatti model is well over $1B.

For McMansion, only a small portion of the cost of building it is the Architect Fee which is analogous to development cost, typically it's 6% to 20%, so maybe $1M at the very top end. There's 3 orders of magnitude difference between the development cost of the two projects, which does reflect the engineering difficulty difference between them.

Which just re-enforces my original point that MSR is just not that hard comparing to projects commercial companies are already doing.

I’m sorry, but this is just baloney.  These companies for the most part are revisiting and remaking capabilities that the government proved out decades ago.  The private sector is doing it much more efficiently and at larger scales than the public sector did or could do.  But we’ve had suborbital human space flight, medium-lift and heavy-lift launch, crew capsules, and satellite constellations before.  Even seemingly novel capabilities like boost-back, vertically-landed, reusable launch stages have precursors in government work like DC-X, any number of planetary landers, various civil and defense entry vehicles, SLI, and most recently the Lunar Lander Challenge.  (I was involved with the last two.)

False. For one thing, only major SpaceX detractors like thunderf00t would use DC-X to belittle what SpaceX has accomplished. DC-X is not at all comparable to what SpaceX did with F9 first stage, for example it didn't demonstrate in-air engine restart, nor did it demonstrate Supersonic Retropropulsion, both required by F9 first stage landing. And F9 has many other design features that DC-X doesn't have, such as various forms of heat shield, and grid fins.

Also it's funny you use Lunar Lander Challenge as an example of "government work" when it's literally new space companies designing and building things.

Finally, government also "for the most part are revisiting and remaking capabilities that the government proved out decades ago", for example all the Mars parachute designs can trace back to Viking.

MSR requires capabilities that have never been done before by anyone.  The specifics depend on the architecture, but MSR will involve some combination of the first supersonic retropropulsion at another planet, the first launch from another planet, the first rendezvous/docking/transfer around another planet, and/or the first propellant production on another planet.  None of these are trivial things. 

For starters, nobody is proposing propellant production on Mars for MSR. As for the others, they're not inherently difficult, just haven't been demonstrated yet. For example there's nothing inherently difficult in launching rocket from Mars, we have demonstrated engine start and flight in atmosphere with similar density as Mars, and we have also launched rocket from low gravity world (i.e. the Moon) before, MSR just needs to put all these together, which is exactly what commercial entities are good at.

And HLS also requires capabilities that have never been done before, cryogenic in orbit refueling for example.

That doesn’t mean that a private entity is not capable of doing them.  But if NASA asks a company or companies to do some or all of these, NASA has to be cognizant of, and deliberate about, what it is asking the private sector to do.  It’s not asking the private sector to reinvent the wheel — which is what we asked the private sector to do when we started commercial cargo and commercial crew.  It’s asking the private sector to invent the wheel for the first time.  That is a much bigger ask, and it should be reflected in terms of the procurement strategy, funding, program development and testing, technical support, and schedule.  Otherwise, we’re setting up MSR and the companies involved for another failure.

Again, this is not the first time NASA asked private sector to invent the wheel for the first time. HLS, in fact the entire Artemis program, is relying on cryogenic in orbit refueling, which has never been done before.

Also it's funny you make MSR sounds so difficult, yet you have no problem with Rocket Lab - a startup with ~2k employees - proposing to do this for just $2B. That's not what a hard problem looks like. You know what program RL didn't even bother to bid on? HLS.

The obstacles can be overcome:

1. SpaceX's iterative development process doesn't have to have lots of failed test flights, they didn't need it for F9 and Dragon for example.

This is woefully ignorant of the Falcon 1 launch failures, the Dragon Spacecraft Qualification Unit, the partial failure on the first commercial Falcon 9 launch, all the Grasshopper testing, and the six reusable first stage losses before the first successful landing.

You didn't bother to read what I write. I specifically didn't include F1 or F9 landing. F1 is just Elon Musk learning how to run a launch company, so it doesn't count. For F9 landing, they're doing the tests basically for free, so there's no pressing need to get it right the first time.

As for Dragon Spacecraft Qualification Unit and Grasshopper, I never said SpaceX didn't need to do any testing. In fact the rest of my original post included several ways SpaceX can do tests for Mars EDL without going to Mars, which is exactly analogous to Dragon Spacecraft Qualification Unit and Grasshopper.

If you think SpaceX has to have multiple Mars windows just to get Mars EDL right, think again.

SpaceX thinks they need multiple windows.  Their proposal to NASA includes a test landing of Starship at Mars in 2029 before undertaking the MSR mission.  See:

https://wccftech.com/spacexs-planned-mars-landing-timeline-revealed-in-nasa-document/

They need multiple windows to execute MSR, not for getting a successful Mars EDL.

SpaceX proposed a 2029 demo landing before doing MSR, this is similar to the unmanned test landing in HLS or unmanned test flight in Commercial Crew. This is not some test that you want to fail.

The point here is that SpaceX is no less capable than old space - e.g. JPL - when it comes to getting novel technology working on the first try, if they wanted to. What they just did is no less complex than skycrane for example.

Again, this is woefully ignorant of the testing that went into both systems.

What does testing have anything to do with this? It's a non sequitur.

Note that SpaceX proposal is listed under "End-to-End Mission Architecture", this proves those who claim SpaceX only proposed Starship as launch vehicle are quite wrong, as I have been saying all along.

This is not news.  We’ve known SX has proposed to use Starship for all segments of the mission since June.  See link above.

Apparently some people didn't get the memo, like Mr. Spice above, so it's a good idea to repeat this sometimes.

[thespacecow]

And that debunk has now been proven to be completely wrong by reality, just as I predicted back then.

I don’t mean to be overly critical, but your old post wasn’t making a prediction about SX. 

I did, it's in the comment preceding that discussion, where I made it clear that it's absurd to think SpaceX would only propose Starship as LV in MSR.

It just argued that all the industry studies had to address all segments of the MSR mission:

The study is for an entire MSR mission from start to finish, you can't just tell NASA "here's one element, you figure out the rest". This should be obvious since NASA needs to compare the cost/schedule/etc for the entire mission, they can't do this if you just gave them one element without the rest. So if you only propose one element, then the rest will be filled in by elements from POR, and this is not competitive against POR if your one element is just a bigger LV.

This was untrue back then and it still is today.

That argument is in support of my original prediction. And no, I didn't argue that "all the industry studies had to address all segments of the MSR mission", just that if an industry study only addressed a single element, the rest of the elements will come from POR. It's not at all wrong, it's literally how the RFP is written.

Moreover, it became public that same month (the same week, in fact) that SX proposed to use SS for all MSR mission segments:

https://wccftech.com/spacexs-planned-mars-landing-timeline-revealed-in-nasa-document/

So even if you had made a prediction, it had a shelf-life of at most a few days.

So? What does this have to do with anything?

Besides, the absurd idea that SpaceX only proposed Starship as LV only came up after SpaceX was selected to do the MSR study, which is only ~2 weeks before the source selection document is released and we know SpaceX proposed demo landing in 2029. So I couldn't make a prediction about the absurdity of this idea very long before even if I wanted to.

Poptip: You can't debunk reality.

The English idiom is “pro tip”, not “poptip”.

Funny you claim that you "don’t mean to be overly critical" then go after my spelling. You do realize some of us don't have English as our native language?

[zubenelgenubi]

Moderator:
Whoa.  A native American English speaker offering friendly advice on English idiom: I would not take it as an attack.

Maybe a few hours off of the Internet would help? 🤔

[TheRadicalModerate]

In a no doubt futile effort to end the "Is not! Is so!" dialogue that we seem to have fallen into, I did a bit of mass-hacking, and I'm pretty sure a v3 Starship can do ES-LEO-refuel-MTO-MarsEDL-receiveSamples-LMO without any refueling on the martian surface, while carrying a 5t payload to the surface and a 1.5t payload away from the surface.

A 1.5t kick stage + EEV + OS ought to be more than enough to get the EEV+OS into an Earth transfer orbit and do the necessary course corrections before and after releasing the EEV+OS.  You'd have to have a fairing that could both deploy rovers / helicopters / whatever to the surface, as well as provide chomper-like functionality to deploy the return vehicle in LMO, but that doesn't sound like an insurmountable problem.  You'd also need to have a reliable zero-boiloff cryocooler that worked on the martian surface, to keep the LMO ascent prop cold while the samples got collected and transferred.

This of course doesn't solve the problem that Starship can't be made Cat IVb-compliant, but the conops is indeed really simple, once Starship can reliably land on Mars.