totm nov 2023 SpaceX Discussion Thread

[tater]

22 minutes ago, Jacke said:

Jumping right to orbital assembly and refueling is making a similar mistake to what was done with the Space Shuttle: trying to make an operational vehicle and process when there's not been enough experimental vehicles and process testing.  Look at what went wrong with changing the modules on the ISS just now.  And how easily that could have turned out so much worse.  This is an area that needs more research, development, testing, and practice.

I disagree entirely.

The way to do it is to do it. Failure is an option. Try, fail, iterate, try again.

Or I suppose we can continue to talk about prop depots and refilling as we have done literally since Apollo, and maybe our great, great grandchildren will grow a pair and actually try it?

Spoiler

Was proposed in the 1960s, too, but:

1970:

 

1971:

 

2011:

 

 

Edited August 9, 2021 by tater

[Spaceman.Spiff]

8 minutes ago, tater said:

The way to do it is to do it. Failure is an option. Try, fail, iterate, try again.

Exactly;

Space Shuttle didn’t have that option because it had to be crewed. 

[sevenperforce]

Looking closely at the tile inspection process on SN20.

Many of the tiles with red-orange tape are visibly cracked or damaged. None of the tiles with green tape are visibly damaged but many look misaligned. If you blow up the image you can see “OK” scrawled on many of the other tiles.

My guess is that red means broken and green means loose.

It does seem like a lot of inspection work. But when you consider that it took sixteen hours to replace a single Shuttle tile....

11 minutes ago, tater said:

The way to do it is to do it. Failure is an option. Try, fail, iterate, try again.

Or I suppose we can continue to talk about prop depots and refilling as we have done literally since Apollo, and maybe our great, great grandchildren will grow a pair and actually try it?

Agreed. Unwillingness to fail is unwillingness to succeed. 

[tater]

BTW, the problems with orbital assembly are also different than they were years ago. The ability to do things without humans on site is improving constantly, which is an infinite risk mitigation in terms of human risk. No humans, no risk of life and limb.

With small payloads and expensive launch costs, it's obviously still extremely risky WRT dollars spent. If costs can come down, then the ability to tolerate failure increases, because the cost of an experimental failure is far lower.

If one were to envision the orbital assembly of a few elements that are ~8m in diameter in a world without Starship, we have to assume SLS. So now our first experiment of EOR assembly costs several billion without even counting the cost of the modules. It's too big a risk to bother trying without certainly of success ahead of time. The ground based test rigs alone would cost millions—likely more than it would cost with SS to just fly the things and try it in space.

Edited August 9, 2021 by tater

[mikegarrison]

10 minutes ago, sevenperforce said:

Unwillingness to fail is unwillingness to succeed. 

You know, it's interesting to read this SpaceX thread and see how many people laud SpaceX for learning through failures, and then compare and see how many of the same usernames lash out at the Starliner whenever it has a problem.

(Not saying you are one of them, @sevenperforce)

Edited August 9, 2021 by mikegarrison

[Meecrob]

4 hours ago, kerbiloid said:

A cross-section is a rounded triangle with extended aside flat bottom.

Like the cross-section of Apollo, too.

  Reveal hidden contents

 

No, its not. The Apollo capsule is obviously different from a vehicle with wings. They might have served the same purpose as far as making astronauts not die in the heat of re-entry, but even a 5 year old can tell the difference.

[mikegarrison]

2 minutes ago, Meecrob said:

No, its not. The Apollo capsule is obviously different from a vehicle with wings. They might have served the same purpose as far as making astronauts not die in the heat of re-entry, but even a 5 year old can tell the difference.

They both had blunt, nearly flat surfaces they presented upon re-entry. Most crewed re-entry vehicles have shared this feature. Many uncrew-ed RVs too.

Edited August 10, 2021 by mikegarrison

[sevenperforce]

1 minute ago, mikegarrison said:

You know, it's interesting to read this SpaceX thread and see how many people laud SpaceX for learning through failures, and then compare and see how many of the same usernames lash out at the Starliner whenever it has a problem.

I think it is a fundamentally different question. Willingness to fail is great when you are working with dev articles. You design a test that will push a disposable test article to its limits in order to figure out the unknown failure modes.

Repeated unforced errors in a system that’s supposed to be in its final configuration, ready to fly human beings? Particularly when the company claimed more flight tests weren’t needed because they had validated everything on paper? That’s...disconcerting.

[mikegarrison]

2 minutes ago, sevenperforce said:

I think it is a fundamentally different question. Willingness to fail is great when you are working with dev articles. You design a test that will push a disposable test article to its limits in order to figure out the unknown failure modes.

Repeated unforced errors in a system that’s supposed to be in its final configuration, ready to fly human beings? Particularly when the company claimed more flight tests weren’t needed because they had validated everything on paper? That’s...disconcerting.

Call me -- unconvinced. Unconvinced that it's anything other than "SpaceX good; anybody else bad".

[Jacke]

33 minutes ago, tater said:

I disagree entirely.

The way to do it is to do it. Failure is an option. Try, fail, iterate, try again.

Yes, but fail cheaply in a test program, not in the main program.  Again, look what happened to Shuttle.  Look what happen at the ISS last week.  Both could have turned out much worse.  And if there's something major wrong in the design, it's a lot cheaper to learn in a test program than in a main program.

Like the N.A.C.A. and NASA have been doing since 1915.  106 years.  And which private industry does too from time to time.

 

33 minutes ago, tater said:

Or I suppose we can continue to talk about prop depots and refilling as we have done literally since Apollo, and maybe our great, great grandchildren will grow a pair and actually try it?

I'm a trained tank gunner.  Drop me in any tank I was qualified on and I'll be up to speed in a few minutes.  Give me about an hour talking to someone else qualified on a vehicle's gunnery system and I'll be fair, but I should have a dry practice and a range shoot to get the skills right.

I put you in the Gunner's seat in a turret.  Do you know how to do the safety precaution?  Do you know how to go to action?  Load the weapon (if no other Loader)?  Traverse your arcs?  Take fire orders?  Ready the weapon and lay it on target?  Note the fall of shot and report your expected correction?  Take the Crew Commander's interrupt with a new correction if the CC gives one.  Apply the correction?  Do you know the Gunner's 4 responses, cold?

Why not?  Tanks have been around since 1916.  105 years?

But you haven't done this.  Could have done it in a game.  Or maybe you're a tank gunner in your real life too.  But you need to have been trained.  And that training came from years of experience and work on the vehicle and to make a training program.

I've seen all about doing stuff in space since the 1960's.  And some of it has been done, basic assembly and refueling of the ISS with hypergols.

But again, I say, you gotta walk before you run.  Because jumping into a full blown program means everything better be exactly right and nothing better go wrong.  'Cause that's where big cost failures and lose of life comes from.

 

5 minutes ago, sevenperforce said:

I think it is a fundamentally different question. Willingness to fail is great when you are working with dev articles. You design a test that will push a disposable test article to its limits in order to figure out the unknown failure modes.

Repeated unforced errors in a system that’s supposed to be in its final configuration, ready to fly human beings? Particularly when the company claimed more flight tests weren’t needed because they had validated everything on paper? That’s...disconcerting.

Sorry, @sevenperforce, I'm out of likes today.  That's spot on.

Edited August 10, 2021 by Jacke

[Meecrob]

9 minutes ago, mikegarrison said:

They both had blunt, nearly flat surfaces they presented upon re-entry. Most crewed re-entry vehicles have shared this feature. Many uncrew-ed RVs too.

Yeah, I respect your opinion of it, but one was ballistic, the other was aerodynamic. Of course we could talk for ages on lifting body designs and all that, but literally one is a plane and one is hunk of metal. One has a parachutes, the other lands like a plane. I think we are on the same page...If I get you right, you just want to make sure people talk accurately, so if I have not been accurate, please tell me.

Edit, I'm on this board to learn, not be a know-it-all even though I probably sound like it at times

Edited August 10, 2021 by Meecrob

[tater]

1 minute ago, Jacke said:

Yes, but fail cheaply in a test program, not in the main program. 

SS/SH is a cheap test program.

For the taxpayer, it's functionally free, as cheap as it gets. From the standpoint of taxpayer, the "main program" is SLS.

SpaceX is building their own GSE tanks, which means their costs are in the same ballpark (possibly more, but with internal benefits of custom needs) as buying those same tanks which are an off the shelf item—and dirt cheap compared to rocket stages in the general market. SS and SH are cheap. the engines are the bulk of the cost, and the engines are cheap compared to everyone else. This is a dev program, that;s how they roll.

4 minutes ago, Jacke said:

But again, I say, you gotta walk before you run.  Because jumping into a full blown program means everything better be exactly right and nothing better go wrong.  'Cause that's where big cost failures and lose of life comes from.

SpaceX is not filled with clueless people. Musk said in that interview/tour that they would not even attempt refilling until they sorted SS entry, descent, and landing ops. They have an idea what they are up to. Validate booster. See how TPS works if possible. More launches to validate both, and establish landing accuracy on booster so they can then attempt catches. Having gained experience, only then will they even start cutting holes in the thing to make it hold cargo. Only then will they start messing with prop transfer—and they will do it like they did F9 booster recovery, launch a payload (Starlink, probably), leave SS then try a refilling. Any loss has already paid for itself.

[Silavite]

1 hour ago, sevenperforce said:

I think it is a fundamentally different question. Willingness to fail is great when you are working with dev articles. You design a test that will push a disposable test article to its limits in order to figure out the unknown failure modes.

Repeated unforced errors in a system that’s supposed to be in its final configuration, ready to fly human beings? Particularly when the company claimed more flight tests weren’t needed because they had validated everything on paper? That’s...disconcerting.

 

1 hour ago, mikegarrison said:

Call me -- unconvinced. Unconvinced that it's anything other than "SpaceX good; anybody else bad".

Tim Dodd's interview with Elon talks about this almost explicitly.

Spoiler

5:10 to 6:25 is the relevant bit

Dragon requires extreme conservatism, Falcon is a bit less conservative, and Starship requires rapid testing/failure/iteration.

[tater]

1 hour ago, mikegarrison said:

Call me -- unconvinced. Unconvinced that it's anything other than "SpaceX good; anybody else bad".

Most of the complaints I have about other providers (launch or spacecraft) have to do with cost, I suppose, if they are competing for the same services.

WRT Commercial Crew, Boeing bidding more is just SpaceX's mistake of bidding too low. Boeing getting paid $287M extra a few years ago because they were clearly going to finish first and do an extra crew launch (already part of their existing contract) kinda bugged me even before it became backwards.

WRT launch providers, I like ULA, they are extremely reliable, and with some cost competition finally existing (due to SpaceX) they came down some in price. BO? It;s not a matter of SpaceX good, BO bad—I desperately want to see BO succeed, and past that I want them to have had some sort of Skunk Works such that they hit the ground and the next thing you know they have a crew vehicle, lander, whatever already made. I'd love to see that, but it seems pretty unlikely.

FWIW my attraction to SpaceX is largely just the transparency of it, combined with their (entirely unneeded) sense of urgency. I'm not getting any younger, and I'd like to actually see something cool happen WRT spaceflight, unlike the last few decades (Shuttle was an awesome thing to see, but it just didn't move the ball down the field much, ya know?).

[sevenperforce]

2 hours ago, mikegarrison said:

Call me -- unconvinced. Unconvinced that it's anything other than "SpaceX good; anybody else bad".

I know you weren't talking about me personally, but for my part I was very disconcerted by both the AMOS-6 disaster and the exploding crew capsule. Doing WDRs and static fires with the payload integrated is exactly the sort of unforced error I'm talking about (even though the spontaneous combustion of solidified crystalline oxygen and carbon fiber had never happened before in history). They learned their lesson, yes, but that was not a "test to failure" scenario, so they don't exactly get credit for it.

I was slightly less worried about the exploding crew capsule once I learned that it WAS a "test to failure" scenario -- that they were intentionally subjecting it to stresses well outside the flight envelope in order to look for that exact kind of unknown failure mode.

[southernplain]

3 hours ago, Jacke said:

Somewhat true.  What was wrong was considering the Shuttle to be an "operational space truck" when it was a first and very experimental design put into production that flew and pushed so much so far.  There was a lot of poor design, especially for crew safety considering its abort modes (and Starship has those same safety lackings and should never launch with crew).  Its supply and maintenance chain and handling were far from optimal, partly due to politics.  And there wasn't sufficient safety-over-operations attitude at the start, which led to Challenger.  That it flew so many flights with so low a failure and incident rate was partly luck and a lot of hard work, especially considering the faults.  Its replacement should have been started in design no later than the early 1990's.

Sure, absolutely Shuttle was an important part of spaceflight history. It taught us many valuable lessons and contributed mightily to our understanding of crewed spaceflight technologies.

However, we also got stuck in LEO for 30 years, it cost approximately $200 billion dollars, and the lives of 14 people.

A Shuttle program that proved the concept of refurbishable spaceplanes and then was retired for Shuttle 2.0 without killing anyone would have been an unqualified success. That wasn't what we got, due to the politics of it all.

3 hours ago, Jacke said:

Jumping right to orbital assembly and refueling is making a similar mistake to what was done with the Space Shuttle: trying to make an operational vehicle and process when there's not been enough experimental vehicles and process testing.  Look at what went wrong with changing the modules on the ISS just now.  And how easily that could have turned out so much worse.  This is an area that needs more research, development, testing, and practice.

We don't start with orbital assembly right now, my point was merely that we should not tailor our LV development specifically to avoid it at all costs. A 100t+ class reusable LV gives us enormous capabilities regarding orbital assembly. If Starship is as cheap as SpaceX say it could be, we can afford to mess around with "research, development, testing, and practice." Without a fully reusable SHLV, that probably isn't going to be feasible. 

2 hours ago, mikegarrison said:

Call me -- unconvinced. Unconvinced that it's anything other than "SpaceX good; anybody else bad".

I would enjoy seeing "anybody else" develop, and fail, fast like SpaceX. Visible progress massively increases my interest in their programs. 

Also, slick launch coverage helps. SpaceX leads the industry in launch day coverage by a wide margin if you ask me.

[kerbiloid]

7 hours ago, southernplain said:

Why though? Orbital assembly works and we are going to have to do more of it if we want to become a spacefaring civilization.

Because it's much easier and more effective to assemble and check the whole thing in the factory.
It's much better to use large blocks to assemble things in space.

Skylab was ~80 t, and it contained personal cabins, bath, toilet, kitchen, etc in a single block.
Mir consisted of 20 t blocks and they had to split these basic systems between the base module and Quantum-2, remove the gyrodynes from the base module and deliver them additionally on Quantum-1, and place them outside, because there was no room for them. That's just because they had to fit a smaller rocket.

What's better: one trip of a reusable 500 t super-heavy or several tens launches of Shuttle todeliver every chunk separately?

What's cheaper and makes the orbital station more available for customers: when you build it here and launch in one piece or keep doing this for 20 years, so while the very first modules start leaking, the very last need repair on ground, and you always attach a new module to the battered station?

***

Lunar base.

As we can see, the minimal size of a reasonable long-term space base is something at least like Skylab (~80 t) or Mir (~120 t).
I.e.  a space base should start from a 100 t base module to make it a full-featured submarine rather than a lego flea circus.
(And thus the projected orbital stations of Glushko bureau were 100 t heavy, like "Polyus").

The lunar base is the same as orbital base, but on ground.
It's also much more reasonable yo deliver a 100-t full featured base module ready to use, the add other modules from time to time.

A rocket able to deliver a 100 t module to the Moon surface is ~500 t capable to LEO.
I.e. the whole ISS in a single launch.

Any serious technological equipment starts from tens of tonnes.
So, to build something industrial on the Moon, still ~100 t delivery is required.

Let alone the lunar harvesters .

***

Custom customers.

The current space base requires space-rated equipment, which is much more expensive than the trash from the closest warehouse.

This limits everything: experiments, space tourism, space industry, etc.
Only richest ones can do this, and still do only 3d presentations.

Launching 500 t stations with thick insulatiion and huge pressurized internal volume means ability to use the cheap trash from the supermarket, and it will be working in orbit like on ground, because it's protected from radiation and inside a normal atmosphere.

This in turn may allow mass orbital building of customer stations
(and on-ground building of customer shelters in case if the customer station deorbits).

Also this can allow to place hundreds of superheavy communication platforms with heavy equipment instead of seeding the LEO with hundreds of thousands of the mosquito fleet like Starlink.

***

Let alone a 500 t orbital telescope and Salyut-sized Jupiter orbiters and Voyagers.

***

The biggest airplanes weight about 500+ tonnes.
Bigger ones are unlikely possible just because they will have different weather at the nose and at the tail when ascending, lol.
I mean, unlikely bigger wings can work effectively.

So, ~500 t (5 times heavier than shuttle) is probably close to the upper limit of a spaceplane.

That means than a 500 t rocket can launch a 500 t spaceplane and allow mass passenger traffic to LEO and back.

***

So, ~500 t payload is good.

7 hours ago, sevenperforce said:

You are correct that a primitive cylinder moving with its axis at 90° to the direction of flight will have the same L/D ratio as a sphere—zero—and will descend ballistically.

You are incorrect in supposing that Starship will perform re-entry with its long axis at 90° to the direction of flight. Starship will fly with a significant angle of attack and have an L/D ratio significantly greater than a capsule like Apollo.

Any proof?

All I can see, exactly none of spaceplane was a cylinder, even with winglets.

Since 1960s they have experimented with cones, cut cones, spaceplanes, other type of lifting bodies.

But noone even tried to use a cylinder.

I guess, the aero- and thermodynamics haven't changed since then.

[kerbiloid]

8 hours ago, sevenperforce said:

Peroxide is, incidentally, hyperbolic. It is also what the Mercury capsule used.

And of course Dragon 2 vents its residuals while descending under chute  as well. All capsules do.

Yes, Mercury used HTP. And this just makes it even safer.

Soyuz ~30 kg of peroxide onboard, vented out after aerobraking.

Dragon - 1+ t of MMH+NTO in the capsule on launch, they are literally sitting on the fuel barrel.

Its early presentations presented rocket landing on same pad as Falcon. The parachute was mentioned as "too heavy".

Then a backup chute appeared, but they still keep the fuel tank and 8 (!) thrusters in the capsules. Each one of them can explode and we could see how it looks like. At least, they won't suffer.

8 hours ago, MKI said:

I think Starship could still have the awesome James-Bond level capability of hijacking a Satellite in orbit right? 

Until the satellite self-destruction system got engaged.

7 hours ago, Deddly said:

And designed to land horizontally.

Yes, what's the problem?

Don't the planes do?

6 hours ago, Meecrob said:

Dude, its a spaceplane, not a capsule.

Who said it's a capsule?

6 hours ago, sevenperforce said:

But this is only if there is no angle of attack (or, depending on your frame of reference, a 90° AoA). With a nonzero angle of attack, lots of things start changing. And that is what someone seems intent on ignoring.

The wings are flat for purpose.

  

4 hours ago, Meecrob said:

Yeah, I respect your opinion of it, but one was ballistic, the other was aerodynamic.

No.
The conical capsules are aerodynamic (in sense of controlled AoA flight, you mean).
The ballistic mode is only an emergency mode when they can't manage the flight, with overloads 8 up to 20 g.

All conical capsules, since Gemini and Soyuz, enter the atmosphere with one edge forward, not bottom-to-flow.

They control their AoA either by engines, or by a mechanical CoM position management, to not become ballistic.

Their normal flight has absolutely no difference from a spaceplane, only the final phase differs.

Btw, that's exactly why you can think it's a flying saucer moving edge-forward.
Because it's a flying saucer moving edge-forward.

Edited August 10, 2021 by kerbiloid

[Beccab]

18 minutes ago, kerbiloid said:

Don't the planes do?

You are going to ignore it like the previous 50 times, but:

- starship is not shuttle

- starship is not buran

- starship is not a plane

- starship is starship

[kerbiloid]

1 minute ago, Beccab said:

You are going to ignore it like the previous 50 times, but:

- starship is not shuttle

- starship is not buran

- starship is not a plane

- starship is starship

Starship is just another one spacecraft following the same aerodynamic and thermodynamic laws, whatever you call it.

[Beccab]

3 minutes ago, kerbiloid said:

Starship is just another one spacecraft following the same aerodynamic and thermodynamic laws, whatever you call it.

And it doesn't land on a runaway

[OrdinaryKerman]

8 hours ago, kerbiloid said:

Starship is just another one spacecraft following the same aerodynamic and thermodynamic laws, whatever you call it.

Same physical laws, yes. Same flight profile? Definitely not. There isn't a reason for Starship to look like something else if it isn't supposed to behave like that something else.

Edited August 10, 2021 by OrdinaryKerman
i hate typos

[kerbiloid]

2 minutes ago, Beccab said:

And it doesn't land on a runaway

To land, it should first survive the aerobraking.

The landing technics plays absolutely no role.

1 minute ago, OrdinaryKerman said:

Same physical laws, yes. Same flight profile? Definitely not. 

On aerobraking - definitely yes.

(Unless you want to take out 8 km/s + 400 km by engines)

Its mechanical energy should be turned into heat and dissipated by the bottom.

When you have a fusion engine and delta-V 100 km/s per flight, no problem, you even don't need a heat protection.

But while you have only chemicals , you should aerobrake.

Edited August 10, 2021 by kerbiloid

[tater]

7 minutes ago, kerbiloid said:

To land, it should first survive the aerobraking.

The landing technics plays absolutely no role.

Yes, and as we all know, SpaceX has zero experience with flying long tubes at a few km/s down towards Earth, and they also have zero experience with EDL from orbit. They also own no computers, and so don't model anything.

They literally have no idea what they are doing, they are just welding some tubes together, and doing it Kerbal style.

Edited August 10, 2021 by tater

[Beccab]

3 minutes ago, tater said:

Yes, and as we all know, SpaceX has zero experience with flying long tubes at a few km/s down towards Earth, and they also have zero experience with EDL from orbit. They also own no computers, and so don't model anything.

They invented nothing, Roscosmos had even drawings of that already. SpaceX is so behind