totm nov 2023 SpaceX Discussion Thread

[SOXBLOX]

13 hours ago, kerbiloid said:

The more likely is that some well-known company is a paper screen with a painted Mars and a shiny Tintin rocket, behind which some serious men from absolutely another price category are managing numbers with ten-eleven zeroes to push Boeing and LockMart from the space market.

And a smiling person standing in front of the screen, facing towards the audience, is an anthropomorphic personification of that, and a skilled showman making you believe that the painting is real.

He owns an electric company, manufacturing accumulators and one of electrocar models, and everything one can read from the press about that company, is that its owner is a normal, typical "greedy capitalist shark", a skilled businessman never letting a cent pass by, and even (iirc) never changing it work for the covid measures.
The electrocar factory descriptions are also far from the romantic space communism, it's a pure money-making enterprise, without impurities.

One can remember than when SpaceX finished some stage of Falcon works, he had fired a significant part of the workers, even when currently he keeps building, so this was an act of temporary economy, not of hard financial state.

So, everything we know about this person is that he never lets a billion fall from hands on Martian ground.

And the easiness of spending his own billions on obviously blurry Martian projects, makes to think that he doesn't spend them, but runs the show in front of the screen, while the real investors put (tens of ?) billions on such strange purpose.

And as probably there are not so much billionaires ready to spend their money on Mars, it looks like the only such billionaire can be only Mr. Budget (probably a French surname, idk) and the money are actually spent for mundane purposes. Say, to support with money a bunch of companies against a couple of others, supported by another influential group of investors.

***

About Proton and other premature happiness.

As it was stated not once, most part of world launches don't depend on Starship, Proton, or whatever.
They are pure national activities, and it doesn't matter whose rocket is bigger.

What about the Southe-East Asia and other custom customers, that part of market was anyway unavailable for Roscosmos:

1. After the known events and following sanctions even if Proton was ten times cheaper than Falcon, the NASA owners would just put it under sanctions and push from the market.

2. Even if there were no sanctions, the men behind the screen would throw another bag of money to decrease the Falcon launch price even below the cost, and anyway this part market was lost.

So, the decision to dismiss Proton was absolutely practical, because Angara and Soyuz-5 use the modified engines from Energy (rocket), many times tested, being manufactured, and enough mature to use and upgrade.
While Proton (an ICBM from 1960s, btw) uses engines which aren't used on other space rockets, and now on other ICBM.
So, while RD-253/276 is a good engine, it's now a one-rocket actor, so as the existing part of market was anyway lost, the rocket row could be dismissed in whole.

To the date there is no sense in mass production of Angara, because most of light sats and OneWeb are being launched by Soyuz, while 20 t heavy modules and PTKNP will be required by 2028 when the lunar program starts.
The Angara engines are tested, usual, debugged, manufactured.
The Angara fuel compartments have flown several times, they are OK.
The Angara in whole (different versions) consists of the same Universal Rocket Module which has been tested.

So, it makes sense to start manufacturing and using Angara whenthe lunar program (including the ROSS station) starts, by 2028.

Before that it's no need in it.

 

Actually, the Russian space program is on transfer stage of Hohmann orbit from the previous ignition (ISS building)  to the next one (lunar program and ROSS).

If look with attention, absolutely all parts of it are intact and many of them at least once tested, just it's too early to ignite the engines.

 

Also, that's why I seriously believe that Rogozin is a figure not lesser than Musk.
The former is an officer on the watch, keeping the cruise ship on course between seaports., and a PR person
The latter is a showman for press and a PR person.

The first few paragraphs reduce to "Elon Musk isn't rich enough to pay for all this" and then expounding a conspiracy theory to explain the fictitious discrepancy.

As for the second part...

1) The current tally of launches planned for or completed this month is China: 5; USA: 6; and the RF, EU, and...Romania? one each. Of the USA's six, half are SpaceX. Of those three, two are Starlink. i.e. not a "purely national activity".

2) This isn't the Roscosmos thread. If you wanna praise and glorify Roscosmos while excusing its failures, at least do it there? 

[wumpus]

6 hours ago, magnemoe said:

The +10 time reuse for profitability is ULA numbers and its probably correct for them as they will have to design and build an new rocket for reuse. 

Musk has thrown around $1 billion in R&D to add reusability to Falcon 9.  Most other swags appear to agree with this number.  Total Falcon 9 booster reuse is standing at 66 (my quick count on the infallible wiki, 54 block 5, 12 full thrust).  That comes to around $15 million for each booster they haven't needed to manufacture, and it would be hard to claim that they can make them for less.  I'd put the break even point somewhere around 20-30 recovered boosters.  It isn't 10, and it certainly isn't 240 (they've made 24 block 5 boosters).

Granted, this also implies that it is worth >$15 million + an expended upper stage + launch costs for every starlink launch, but I suspect that this is true.  I won't claim they really want to cover the full $60 million they charge customers.

1 hour ago, Jacke said:

So flying people on Starship before that better safety record is demonstrated without adding in abort modes is condemning 1% of them to die.

Falcon has had a single event in 124 launches that would require an abort mode (the other destruction of cargo could be avoided by loading fuel before crew).  The Shuttle had 135 missions and lost two orbiters.  While abort modes are a good idea, I strongly suspect that the rentry and especially landing is where spacex will lose the majority of Starships (there's no reason to avoid riding a Dragon.  You really wouldn't want to bet your life on a safe Falcon 9 booster landing).  Also note that the issues with the first failure (and presumably the fueling LOEverything) were fixed going from Falcon 1.1 to FT to Block 5 (there were a few fixes in the tiles in post Columbia shuttles, but not much other improvements).

I'd be fairly surprised if the first crewed Starship didn't have a separate capsule with ascent aborts and either parachutes or pressure fed retro-rockets.  What we are seeing is likely the "minimum viable Starship", not the end product (which may well end up in only 2 parts.  But I don't think they will get there in a single step).

Edited August 16, 2021 by wumpus
found the FT reuse number but didn't edit the original phrase

[Silavite]

This is still a long way off, and odds are likely that nothing comes of it... but how much protection does Starbase have against high winds, surge, and rain?

[MKI]

2 hours ago, Jacke said:

I see no reason for the reliability of Starship in the short term to be better than the current best for other launch vehicles, about 1% failure rate.  It may get better, but that will have to be demonstrated.  That is about 5 orders of magnitude worse (100,000 times worse) or greater than for current aircraft.

...

That's where those 5 orders of magnitude of safer performance for aircraft come from, a whole lot of sweat and blood over time.  I don't think Starship will get that for rocketry in under 10 years, as other launch vehicles flying for much much longer haven't gotten to that point.

 

Starship will be nothing like any existing launch platform, providing a 1% example failure rate out of thin air to push the idea 1% of those people that get on Starship are going to die is a fallacy.

Its also true the best launch vehicles have a 0% failure rate, but you might argue sample size, which is where I'm making my point. The idea Starship could have 100+ perfectly executed launches within a few years is actually feasible. Scale that up to potentially hundreds of successful launches continuously, limited to payloads and integration. This is the whole damn point of the rocket, which is why Elon is moving so fast for this 1 goal. If Starship can sustain rapid reusability I don't see how it can't out fly every single launch platform ever created within a few years of getting fully operational. 10 years is where I'd assume it could out perform every single launch platform ever made combined, assuming there's that much to actually do with the system.

Obviously this sort of scale is unprecedented, incredible and plainly insane. Which is why its so hard to actually believe it will work. Hell even Elon believes stuffs gonna blow up on their way, which is why they are moving so fast, which is also why I don't see them "slowing down". I actually see them speeding up as things get more stable. The question is more of how fast can you iterate and how quickly you can fail. Right now SpaceX is flying while also learning a lot. But they are still in the early stages, if they can make a system that can reliably deliver payloads to LEO with full reusability (that is semi-rapid) they will already bypass Falcon 9 capabilities, and all other capabilities with an insane price point.

 

2 hours ago, Jacke said:

Starship has no abort modes for a lot of the potential failures of a launch vehicle, in a similar way as the Shuttle lacked those same abort modes.

The issue with Shuttle isn't so much its "lack of abort modes", but rather the fragileness of the entire system. Mind you 50% of all destroyed Shuttles wouldn't of been saved by "gods hand" abort modes, as Columbia broke apart during descent* (yes I'm abusing small sample sizes again). Focusing only on abort modes during ascent misses all the other insane stuff Starship will have to accomplish. Here's a rough list of what it needs to accomplish to get to Mars and back:

- Ascent (abort mode could go here)

- Re-fuel in orbit (using technology never tried before)

- Trans-Mars ejection burn

- Trans-Mars coasting for 6 months

- Mars injection burn/entry

- Mars atmospheric entry/glide/landing (victory for mankind)

- Stay on Mars for years

- Re-fuel on Mars (somehow)

- Trans-Earth ejection burn (I assume no orbital re-fueling??)

- Trans-Earth coasting for another few months (is it 6?)

- Earth atmospheric entry/glide/landing (welcome heroes home)

 

So the idea "it needs an abort mode to work" ignores the rest of the completely insane requirements, to the point the first ascent should be the least of your worries since that's the part that should be tested the most, forward backwards upside down, and inside out.  

 

2 hours ago, Jacke said:

An aircraft always has the abort mode of becoming a glider, even if it's a poor one and perhaps in a very bad place with respect to its altitude and airspeed.  And aircraft are designed and redesigned to have less chance of catastrophic and lesser failures as lessons have been learned from building and operating them, as well as investigating incidents and accidents on thousands of aircraft over millions of flight hours over decades.

The thing with abort modes are they only cover specific situations, outside of those situations they are useless, or potentially even dangerous (see Gemini ejection seats). For example, "planes can glide" only covers the situation when you have enough airspeed, altitude, and control surfaces available, if your missing any one of your those your just as dead with or without engines. Passengers don't have ejection seats for when a pilot loses control of the aircraft, nor can a plane activate its launch tower to get out of a flat spin, or does a plane magically gain the ability to defy gravity when in a stall, nor do passengers have much chance of surviving if there is a high altitude structural breakup of the aircraft. All these scenarios could result in complete loss of the aircraft with no hope of survival and we are totally fine with it.

Suddenly we sprinkle in the idea of "propulsion powered descent" and "no escape tower" and suddenly its a death trap. 

I see no reason why investigating incidents over thousands of aircrafts over millions of flights isn't possible with Starship. That's the whole point of rapid reusability right now. To get the flight time, experience, and push the limits to see where they break.

 

 

 

 

Edited August 16, 2021 by MKI

[OrdinaryKerman]

2 minutes ago, MKI said:

ascent

descent, unless you're talking about Challenger

[Nothalogh]

15 hours ago, kerbiloid said:

push Boeing and LockMart from the space market

Anyone who does that is doing the lord's work

[RyanRising]

53 minutes ago, MKI said:

Here's a rough list of what it needs to accomplish to get to Mars and back:

- Ascent (abort mode could go here)

- Re-fuel in orbit (using technology never tried before)

- Trans-Mars ejection burn

- Trans-Mars coasting for 6 months

- Mars injection burn/entry

- Mars atmospheric entry/glide/landing (victory for mankind)

- Stay on Mars for years

- Re-fuel on Mars (somehow)

- Trans-Earth ejection burn (I assume no orbital re-fueling??)

- Trans-Earth coasting for another few months (is it 6?)

- Earth atmospheric entry/glide/landing (welcome heroes home)

So the idea "it needs an abort mode to work" ignores the rest of the completely insane requirements, to the point the first ascent should be the least of your worries since that's the part that should be tested the most, forward backwards upside down, and inside out.  

The Apollo moon landings also had a lot of stuff going on that wasn't ascent, and could have easily killed the astronauts, yet they still put an LES on the Saturn V cause it was something that could be ameliorated with it. The existence of other failure modes does not mean you shouldn't mitigate those you have. And I very much doubt they'll be able to prove out an airliner-like safety record by the time they want to launch crew on Starship, whether for Mars or other purposes. They'll be able to do a ridiculous amount of launches, sure, but you need millions of flights to show a safety record like that. For context, there have been about six thousand orbital launches ever attempted.

Edited August 16, 2021 by RyanRising

[Beccab]

1 minute ago, RyanRising said:

The Apollo moon landings also had a lot of stuff going on that wasn't ascent, and could have easily killed the astronauts, yet they still put an LES on the Saturn V cause it was something that could be ameliorated with it

The Saturn V was an very dangerous rocket that flew more than 50 years ago, so not really a good comparison. Yes, it didn't explode mid flight, but that doesn't change the overall safety of it: had Shuttle stopped at STS-13 like the Saturn V did it would have had a 100% safety record as well

[RyanRising]

15 minutes ago, Beccab said:

The Saturn V was an very dangerous rocket that flew more than 50 years ago, so not really a good comparison. Yes, it didn't explode mid flight, but that doesn't change the overall safety of it: had Shuttle stopped at STS-13 like the Saturn V did it would have had a 100% safety record as well

How about Soyuz? It's flown a lot, so there's lots of data on it, and should be understood extremely well. It's had changes and upgrades made to it over the years, yet still has an escape system. Needs it, too.

Edited August 16, 2021 by RyanRising

[MKI]

11 hours ago, RyanRising said:

The Apollo moon landings also had a lot of stuff going on that wasn't ascent, and could have easily killed the astronauts, yet they still put an LES on the Saturn V cause it was something that could be ameliorated with it.

The Saturn 5 was also not suppose to be reusable what-so-ever, it was also designed using the method of "trial and error" and over engineered simply due to the fact they didn't really know what they needed to engineer. Hence the completely overpowered design to meet the requirements of getting a man on the moon and back safely. Everything else is thrown out the window.

Apollo killed 3 people (Apollo 1), and almost killed 3 more (Apollo 13). Neither of which had anything to do with the LES. Having an LES didn't make the overall system much safer. Better engineering made it safer.

 

11 hours ago, RyanRising said:

The existence of other failure modes does not mean you shouldn't mitigate those you have. And I very much doubt they'll be able to prove out an airliner-like safety record by the time they want to launch crew on Starship, whether for Mars or other purposes. They'll be able to do a ridiculous amount of launches, sure, but you need millions of flights to show a safety record like that.

I never said Starship wont mitigate other failure modes, I only wanted to point out that the set of requirements are too wide to cover with any dedicated LES, and that the idea "it will never be safe because it doesn't have one" is impractical and misses the other large amounts of risk. I also didn't directly compare Starship to airliner safety because there are a number of differences between the two fields.  With airliners you have a vastly larger interconnected fleet, with multiple places human error could screw things up. SpaceX is one company, focusing on 1 launch at a time (or a few)

The scale just isn't there, nor does it need to be. If I wanted Starship to get to "airline safety levels of risk" it would take a long time to get anywhere near the scale and percentage of airliners. It also could just keep being perfect in a specific scenario and much more risky in another.

Simply put, getting to LEO in Starship 10000 times perfectly is a totally different level of risk as carrying out a full Mars mission and getting back. Its never been done, and the risk is incredible, as the unknowns are vast. What is known however is getting to LEO is just the first part of the entire journey. Dominating that scenario is peanuts for SpaceX, and only a matter of time, but also only the first leg of what Starship is designed for.  

 

edit

It should be noted that there are a few contingencies already planned around Starship during ascent to deal with possible failure modes. Namely executing a RTL, except more reliably than the Shuttle, where Starship fires all of its engines to escape a potentially failing SH booster. I have no idea where current numbers are on such a concept, but I can't see why it couldn't be done in a pinch even if it requires firing the vacuum raptors. I also don't really have an idea of what sort of scenario you'd have where a SH is failing to the point you need to execute a RTL. Unlike the Shuttle, you can turn off the engines/de-throttle and potentially detach and fly the entire stack back when its safer/possible, or lose the SH and fly just Starship back.

On the topic of powered decent, redundancy has already been demonstrated where the 3 gimbled engines all fire to see which is in best shape for landing. I personally think this is important in the previous RTL scenario, where potential damage on multiple engines might be possible. However, if your flying Starship+Super Heavy with quick turnaround, and able to land it on Mars reliably and get it back. Raptor should be pretty reliable during normal operation, especially so if you only need 33% of them to be working.  

Edited August 16, 2021 by MKI

[Beccab]

5 minutes ago, RyanRising said:

How about Soyuz? It's flown a lot, so there's lots of data on it. It's had changes and upgrades made to it over the years, and yet still has an escape system. Needs it, too.

That's a better comparison, yup. Let's look at the soyuz launch aborts:

-Soyuz 18a: failed second stage ignition, abort using the service module engines since the LES was already ejected. Starship here is either irrelevant or able to do the same, because should superheavy suffer issues that require separating from it its engines can be shut down and Starship should be able to have a TWR > 1 even alone. I said irrelevant because second stage ignition failure should be almost impossible, given that you'd need to have at least 1 RVac failing as well as 2 Raptors failing to ignite at the same time.

-Soyuz T-10a: pad abort following the rocket taking fire. This one is a lot harder, starship abort would mean that Superheavy is blown up by the exhaust of the Starship engines, which makes survivability very uncertain. Absolutely best to prevent this in any way possible with redundancies on the tower as much as possible

-Soyuz MS-10: abort following side boosters colliding with the core. Luckily Starship doesn't have side boosters that can cause this and every other related issues goes in the first category

 

So, out of the Soyuz aborts in his history only 1 would probably have been impossible with Starship. There is another risk reduction at this point: uncrewed starships. For at least the first half of this decade and possibly until 2030 there's going to be a huge disparity between the crewed launches and uncrewed ones, in the order of 100 uncrewed for 1 crewed. Every possible issue is much more likely to happen in one of those launches than the ones where the life of someone is at risk, so if that 1 failure starship is unlikely to survive happens every 150 launches there's a 99% chance it will happen in an uncrewed launch.

[kerbiloid]

3 hours ago, SOXBLOX said:

The first few paragraphs reduce to "Elon Musk isn't rich enough to pay for all this" and then expounding a conspiracy theory to explain the fictitious discrepancy.

As for the second part...

1) The current tally of launches planned for or completed this month is China: 5; USA: 6; and the RF, EU, and...Romania? one each. Of the USA's six, half are SpaceX. Of those three, two are Starlink. i.e. not a "purely national activity".

2) This isn't the Roscosmos thread. If you wanna praise and glorify Roscosmos while excusing its failures, at least do it there? 

The first few paragraphs have nothing common with what you told.

The Roscosmos part is a reaction on another one's talk about Roscosmos in the SpaceX thread.

1 hour ago, Nothalogh said:

Anyone who does that is doing the lord's work

Not sure if the lord's competing to the launch providers and spacecraft manufacturers.

Edited August 16, 2021 by kerbiloid

[RyanRising]

53 minutes ago, MKI said:

I never said Starship wont mitigate other failure modes, I only wanted to point out that the set of requirements are too wide to cover with any dedicated LES, and that the idea "it will never be safe because it doesn't have one" is impractical and misses the other large amounts of risk. I also didn't directly compare Starship to airliner safety because there are a number of differences between the two fields.  With airliners you have a vastly larger interconnected fleet, with multiple places human error could screw things up. SpaceX is one company, focusing on 1 launch at a time (or a few)

The scale just isn't there, nor does it need to be. If I wanted Starship to get to "airline safety levels of risk" it would take a long time to get anywhere near the scale and percentage of airliners. It also could just keep being perfect in a specific scenario and much more risky in another.

Well, of course you can't mitigate other failure modes with a launch escape system. Clue's in the name: It's for escaping the launch. Something I think will be incredibly important until Starship launches  have a level of risk comparable to... maybe we could use that oft-quoted 1/270 LOC ComCrew metric - I think that was including the LES, right? To empirically prove that out, you'd need 270 flights without a failure for each one with.  They could probably make that work in a reasonable timeline if they launched a Starship a day for several years. But until they have flown that much, it would be needlessly risky to launch crew without an escape system. The other risks are still there, and by no means made safer by it, but they're risks you do need to take to make the mission work. This one isn't.

 

1 hour ago, Beccab said:

That's a better comparison, yup. Let's look at the soyuz launch aborts:

-Soyuz 18a: failed second stage ignition, abort using the service module engines since the LES was already ejected. Starship here is either irrelevant or able to do the same, because should superheavy suffer issues that require separating from it its engines can be shut down and Starship should be able to have a TWR > 1 even alone. I said irrelevant because second stage ignition failure should be almost impossible, given that you'd need to have at least 1 RVac failing as well as 2 Raptors failing to ignite at the same time.

-Soyuz T-10a: pad abort following the rocket taking fire. This one is a lot harder, starship abort would mean that Superheavy is blown up by the exhaust of the Starship engines, which makes survivability very uncertain. Absolutely best to prevent this in any way possible with redundancies on the tower as much as possible

-Soyuz MS-10: abort following side boosters colliding with the core. Luckily Starship doesn't have side boosters that can cause this and every other related issues goes in the first category

So, out of the Soyuz aborts in his history only 1 would probably have been impossible with Starship. There is another risk reduction at this point: uncrewed starships. For at least the first half of this decade and possibly until 2030 there's going to be a huge disparity between the crewed launches and uncrewed ones, in the order of 100 uncrewed for 1 crewed. Every possible issue is much more likely to happen in one of those launches than the ones where the life of someone is at risk, so if that 1 failure starship is unlikely to survive happens every 150 launches there's a 99% chance it will happen in an uncrewed launch.

I'm not sure what that second point is trying to make - Soyuz does a similar thing, but that didn't stop those three launch aborts from happening, it just means there are other launches that could fail. As for the breakdown of failure modes, it seems to show just as much that the Soyuz escape system isn't necessary as it does to show the Starship one isn't. MS-10 probably could have tried firing the second stage to get them out of there, in 18a it wasn't used, and T-10a... well, could redundancies on the tower have provided an alternate means of escaping? Perhaps. But I don't see anyone clamoring to try that.

45 minutes ago, kerbiloid said:

The first few paragraphs have nothing common with what you told.

 

Quote

The first few paragraphs reduce to "Elon Musk isn't rich enough to pay for all this" and then expounding a conspiracy theory to explain the fictitious discrepancy.

Quote

And a smiling person standing in front of the screen, facing towards the audience, is an anthropomorphic personification of that, and a skilled showman making you believe that the painting is real.

And the easiness of spending his own billions on obviously blurry Martian projects, makes to think that he doesn't spend them, but runs the show in front of the screen, while the real investors put (tens of ?) billions on such strange purpose.

You claim it's not for Mars, but instead the "real investors," separate from Musk, are putting billions of money towards something else. While the outward show of the program is a cover-up, "making you believe that the painting is real." Sounds a lot like a conspiracy theory to me, which is exactly what SOXBLOX described.

[kerbiloid]

13 minutes ago, RyanRising said:

You claim it's not for Mars

I claim, I had a plastic ruler with animated Saturn and orbital station.

A reacher guy has animated an Mars picture and a Tintin rocket.

I'm still not at Saturn...

 

[RyanRising]

10 minutes ago, kerbiloid said:

I claim, I had a plastic ruler with animated Saturn and orbital station.

A reacher guy has animated an Mars picture and a Tintin rocket.

I'm still not at Saturn...

I don't think a plastic ruler can be evolved into a Saturn orbital station. But that Tintin rocket can be evolved into a Mars lander - at least, it's a little more plausible, wouldn't you say?

But that's still kinda avoiding my point. Whoever gave you that ruler did not claim it would take you to Saturn. The person showing us this rocket claims it, or something like it, will take people to Mars. I can get that you don't believe it, but to say that guy's knowingly lying and there are bigwigs behind the scenes doing the "real" planning for the rocket for some other purposes is, indeed, a conspiracy theory.  And so it does have something to do with what SOXBLOX said?

Edited August 16, 2021 by RyanRising

[kerbiloid]

7 minutes ago, RyanRising said:

Whoever gave you that ruler did not claim it would take you to Saturn.

I bought it for ten kopecks in the school.

And it still looks for me that it was cheaper than the loudy Martian claims, with same chances to get there.

Edited August 16, 2021 by kerbiloid

[Beccab]

33 minutes ago, RyanRising said:

Perhaps. But I don't see anyone clamoring to try that.

I think we agree then, we are just focusing on different points. Nobody is trying to get the LES off Soyuz because it is practical there: not a lot of mass, not a lot of cost compared to the rocket, easy to make. On starship nothing of this is true: how would it even work? Expulsion seats like on gemini, which are more likely to kill you than the rocket and that won't work on pad aborts? A huge SRB thrown away every launch, probably bigger than the ones on shuttle to even work, paired with similarly huge parachutes that would eat half of the payload at least? There isn't a single launch escape system that can work on Starship without being extremely impractical and little more effective than just separating it from SH and going away

[RyanRising]

5 hours ago, kerbiloid said:

I bought it for ten kopecks in the school.

And it still looks for me that it was cheaper than the loudy Martian claims, with same chances to get there.

Their engineers seem to believe its lineage can. But maybe they’ve been fooled by the grand conspiracy too. 

 

5 hours ago, Beccab said:

I think we agree then, we are just focusing on different points. Nobody is trying to get the LES off Soyuz because it is practical there: not a lot of mass, not a lot of cost compared to the rocket, easy to make. On starship nothing of this is true: how would it even work? Expulsion seats like on gemini, which are more likely to kill you than the rocket and that won't work on pad aborts? A huge SRB thrown away every launch, probably bigger than the ones on shuttle to even work, paired with similarly huge parachutes that would eat half of the payload at least? There isn't a single launch escape system that can work on Starship without being extremely impractical and little more effective than just separating it from SH and going away

I’ve been partial to the idea of a smaller separable crew section at the top of the ship, where the crew would sit for launch. Should be within the range of what we can bring down on parachutes, if cramped (I think they managed to cram 10 people into an Orion test article in a useable launch config), and would use a hybrid motor with the header tank’s LOX as oxidiser and some kind of rubber or paraffin for fuel. Very hard to have ignite mistakenly, weighs less than SRBs would, but should have comparable thrust. Should be useful in case of pad abort, in-flight, and maybe even landing burn issues - you’d have the capsule pitch up during its escape burn and the high TWR + low ship terminal velocity means it should be able to get out of that situation too.

The main issues with developing something like that would probably be the hybrid motors (tricky beasts, those are the mechanism to separate the launch module from the rest of the crew section cleanly, and of course the chutes. I haven’t run the numbers on this construction, though - I need to do that to work out the extra mass for hybrid motor + casing and see if we’ve developed any parachutes that could feasibly handle something more massive than Orion. 

Still unnecessary mass for something like a Mars mission, but I imagine the ship for landing on Mars would not be the same one used to launch the crew, so no need for it to be installed there.

 

EDIT:  Hmm, Orion weighs 20,500 pounds (ew, pounds) and JPADS was seen as having a clear path to airdrop stuff up to 60,000 pounds. Dunno what the impact velocity of that would be, though, so it might not be relevant. Orion’s parachutes are 1,144 pounds, though I’m not sure if that’s counted in its landing weight. Probably not, but if they are, that tells me parachutes will have to be ~5.6% of the capsule’s mass. Orion only seats four, but that’s cause it has to take them all the way to the moon and back. Our abort capsule is probably going to be cramped and comparable to the Soyuz descent module, which is 2.9 tonnes for three people. They can use the rest of the ship to stretch their legs once they get to orbit. Using the same ratio for the ten I was imagining for Starship (cause 100 people is ridiculous. dearMoon is supposed to have 10 people), that’s just 9.7 tonnes. Call it 10. As for the structure of that section of the ship, the LOX header tank should weigh about 640 kg, and I’m going to approximate the section we’re chopping off for our capsule to be a come 6 m tall and 7 m wide, made of 4mm stainless. For that cone, including the base, that works out to ~0.417 m^3 (does this forum do superscriptes?), which at 7930 kg/^3 is 3.32 tonnes. I have no idea how much the heat shield will weigh, on that section of the ship, so I’m gonna call it half a ton, bringing us up to 14460 kg. I have no idea what the dry mass of hybrids is like, especially ones optimised for thrust rather than impulse. Probably similar to solids? A GEM-63XL would have a dry mass of 5,400 kg, bringing our dry mass to 19,860. (Orion seems like a bad comparison here, seeing how its LAS is able to be thrown off the vehicle and includes that fairing.) Add 5.6% for parachutes, and what our LAS needs to pull away from the ship is 20,972 kg, plus whatever’s left of the LOX header’s 21,300 kg. Also, I’m going to assume a similar mixture ratio for hybrids as there is for kerosene - 2.7. Most hybrid fuels I know of are hydrocarbons, so that seems reasonable. 
Put those numbers into the rocket equation, you get 280 = 250 * 9.8 * ln((20,972 +21,300 +MFuel)/(20,972 + 21,300 - 2.7MFuel), which works out to a fuel mass of about 1,270 kg.

All told, this 10-person hybrid Starship escape module should weigh about 22,250 kg, including the chutes, capsule, structure around it, hybrid fuel and motors, but not the LOX in the header tank. I’ve tried to make sure all my errors trend towards overestimating the mass rather than underestimating it, but that still leaves ~77,000 kg of payload for the orbital accommodations and the actual mission. And if your ship blows up the crew gets to live and fly another day. Seems like a good trade to me.

P.S. I just noticed: after all that waffling about, we’re at ~2.2 tons per person. Pretty close to Orion’s 2.3 tons per person. Our capsule has an integrated abort system, theirs has appropriate space and heatshielding for lunar flights and reentry, both are for wildly different purposes, but that one number came out pretty close. Weird.

Edited August 16, 2021 by RyanRising
Did things in a more sensible order.

[RCgothic]

Starship won't have LES because the Mars version can't have LES.

Crew-rated Earth version with a LES? Removed the LES for Mars? That's a change and now it's not crew-rated anymore.

[Rakaydos]

So I'm seeing a disconnect between "Near term there's no WAY starship is reliable enough for crew" and "SpaceX will fly a million times before crew, so it'll be safe."

Lets focus on the middle term. How much iteration can SpaceX do before december 2026? (in preparation for the 2027 mars launch window) Note that reuse doesnt actually figure into the iteration rate, only the expirimentation rate. They can fly, say, SN37 a total of 50 times, but they cant make any improvements based on those flights until they build SN38.

So Starbase seems to have a starship production rate of about 1/month. They're actually a bit faster, but building boosters before they qualify landing will slow them down at the beginning, so we'll call 1/month an average.

SN20 will be finished and stacked  before the end of august 2021. That gives 4 months (and thus 4 starship iterations) in 2021, 12 in 2022, 12 in 2023, 12 in 2024, 12 in 2025, and 12 in 2026.

That's (approximately, see assumptions earlier) 64 starship iterations before the first crewed mars flight. Not exactly edison's "I have found a thousand ways NOT to make a lightbulb", but it's also not counting computer simulations for the obvous pitfalls, and software changes that do not require a new-build starship. It also doesnt figure in Raptor upgrades- Raptor v2 is in the pipeline, and Elon doesnt expect to be satisfied until at LEAST Raptor v5, and raptors have their own iteration rate, independent from starships.

[Elthy]

30 minutes ago, Rakaydos said:

So Starbase seems to have a starship production rate of about 1/month

Thats the current rate. 2 Years ago it was 0, in two years it could as well be 1/week. Its impossible to say how fast this will progress, even SpaceX will change their plans several times in the next months.

[Rakaydos]

12 minutes ago, Elthy said:

Thats the current rate. 2 Years ago it was 0, in two years it could as well be 1/week. Its impossible to say how fast this will progress, even SpaceX will change their plans several times in the next months.

I'm taking the pessimistic, plateau'ed number for now. Remember that I'm not including booster production. Also, there has to actually be some testing between revisions, and then they have to incorporate the changes into the next model. But you're right, this is not a maximum number of revisions, or a minimum, merely an expected number.

[RyanRising]

2 hours ago, RCgothic said:

Starship won't have LES because the Mars version can't have LES.

Crew-rated Earth version with a LES? Removed the LES for Mars? That's a change and now it's not crew-rated anymore.

I don’t believe the crewed Mars version of Starship and the crewed Earth version of Starship are going to be the same, no matter if we choose to include an LES on the Earth version or not. 

1 hour ago, Rakaydos said:

So I'm seeing a disconnect between "Near term there's no WAY starship is reliable enough for crew" and "SpaceX will fly a million times before crew, so it'll be safe."

Lets focus on the middle term. How much iteration can SpaceX do before december 2026? (in preparation for the 2027 mars launch window) Note that reuse doesnt actually figure into the iteration rate, only the expirimentation rate. They can fly, say, SN37 a total of 50 times, but they cant make any improvements based on those flights until they build SN38.

So Starbase seems to have a starship production rate of about 1/month. They're actually a bit faster, but building boosters before they qualify landing will slow them down at the beginning, so we'll call 1/month an average.

SN20 will be finished and stacked  before the end of august 2021. That gives 4 months (and thus 4 starship iterations) in 2021, 12 in 2022, 12 in 2023, 12 in 2024, 12 in 2025, and 12 in 2026.

That's (approximately, see assumptions earlier) 64 starship iterations before the first crewed mars flight. Not exactly edison's "I have found a thousand ways NOT to make a lightbulb", but it's also not counting computer simulations for the obvous pitfalls, and software changes that do not require a new-build starship. It also doesnt figure in Raptor upgrades- Raptor v2 is in the pipeline, and Elon doesnt expect to be satisfied until at LEAST Raptor v5, and raptors have their own iteration rate, independent from starships.

That’s plenty of iteration, which is great for optimising and improving the design, but it doesn’t help with proving reliability through flight rate. We can’t just make one that works to do that, we have to make one that works and then fly its design a couple hundred times successfully with only one or two failures to be able to responsibly claim it’s safe enough for crew without an escape system. The first flight of Starship design iteration 58 has exactly as much proven reliability as the first flight of Starship design iteration 3.

Edited August 16, 2021 by RyanRising

[Rakaydos]

28 minutes ago, RyanRising said:

That’s plenty of iteration, which is great for optimising and improving the design, but it doesn’t help with proving reliability through flight rate. We can’t just make one that works to do that, we have to make one that works and then fly its design a couple hundred times successfully with only one or two failures to be able to responsibly claim it’s safe enough for crew without an escape system. The first flight of Starship design iteration 58 has exactly as much proven reliability as the first flight of Starship design iteration 3.

Ah yes- it's one thing to actually improve the design, another to PROVE it's been improved.

Still, that's where cheap rapid reuse WILL come into play.   If by iteration 60 they've nailed "3 flights a day per pad" reuse, that's 90 flights per month before the next iteration comes out of starbase. 3 months to get 270 flights, the commercial crew calculated LoC number.

Edited August 16, 2021 by Rakaydos

[RyanRising]

37 minutes ago, Rakaydos said:

Ah yes- it's one thing to actually improve the design, another to PROVE it's been improved.

Still, that's where cheap rapid reuse WILL come into play.   If by iteration 60 they've nailed "3 flights a day per pad" reuse, that's 90 flights per month before the next iteration comes out of starbase. 3 months to get 270 flights, the commercial crew calculated LoC number.

I think managing to prove it out like that before flying crew is the best solution. I just worry they’ll try to put people on it before it’s had enough time to do those >270* successful flights for every launch failure.

*or whatever number you find acceptable for flights between accidents. Doesn’t have to be that one, just has to be high. 

Edited August 16, 2021 by RyanRising