Do y'all think the Space-X Super heavy/Star ship would work out?

[kerbiloid]

1 hour ago, Rakaydos said:

Sea Dragon was designed to be built 100% in a naval shipyard in the 60s/70s by unskilled workers. 

And that's very good regardless of the workers' skills.

(Also, they anyway have a coastal launch site like both Nexus and Sea Dragon, but of course I don't insist they should splash.)

1 hour ago, DDE said:

The BFR is probably not at a point where the square-cube law avenges the notorious mass inefficiency of high-thrust pressure-fed rockets.

All of them are/were going to lift ~500+ t.
So, they could have thicker walls instead of 40 turbopumps.

Another advantage: on Mars they don't have a service facility.
The dumber is a rocket - the better it flies there.

Edited February 5, 2019 by kerbiloid

[sevenperforce]

1 hour ago, DDE said:

The BFR is probably not at a point where the square-cube law avenges the notorious mass inefficiency of high-thrust pressure-fed rockets.

The problem is that pressure-feeding is not necessarily square-cube agnostic. When you figure in issues like buckling moments, etc., you can very easily run into a scenario where the tank dry mass grows faster than the square.

1 hour ago, Rakaydos said:

Mueller isn't building Raptor in a marsh in texas, though Elon may for the rest of the rocket. It's small enough to be put on a truck, so there's no problem building it elsewhere and installing it there. He can design it to be the best reusable rocket engine, not just in the world, but the best that an experienced early 21 century rocket engineer is capable of designing with effectively unlimited design funding.

"Merlin has the best TWR in the world, but Raptor is coming."

Exactly. Don't let the "sheet-metal" rocket fool you. The Raptor engine is Elon's real "heart of gold" here; everything else is just window dressing.

 

Edited February 5, 2019 by sevenperforce

[Rakaydos]

29 minutes ago, kerbiloid said:

And that's very good regardless of the workers' skills.

(Also, they anyway have a coastal launch site like both Nexus and Sea Dragon, but of course I don't insist they should splash.)

All of them are/were going to lift ~500+ t.
So, they could have thicker walls instead of 40 turbopumps.

Another advantage: on Mars they don't have a service facility.
The dumber is a rocket - the better it flies there.

The "40 pumps" (it's actually 76 pumps, 2 pumps per engine) let Starship be less than 1/4 the size  of Sea dragon. Those pumps are expensive, but that doesn't actually matter when you can just land it and refuel.

Edited February 5, 2019 by Rakaydos

[kerbiloid]

5 minutes ago, Rakaydos said:

Those pumps are expensive

I don't mean their cost, just the 76 roaring mechanisms frighten me.

While a thick tank is just gently saying "Pshhh!..."
Also on Mars they would probably want something really simple.

Edited February 5, 2019 by kerbiloid

[Starman4308]

23 hours ago, sh1pman said:

The reason why I don't think LES is a good idea is because it only covers a tiny fraction of possible failure situations. It can only save lives if a failure happens during the burn of Super Heavy or shortly after the stage sep. LES is useless later into the launch. It won't help if a failure happens in orbit, during tanker randezvous, docking, refilling, departure burn, in transit to Mars/Moon, during atmospheric entry and landing, on the surface of another body, during liftoff from that body, burn to Earth and Earth atmospheric entry. It needs to be safe in all of these situations, not just first few minutes of launch. Airplanes are as safe as they are now not because passengers have escape pods.

So I think that instead of investing their time and money into the implementation of Starship LES, SpaceX should focus on making failures less likely.

Literally the exact same logic could have been applied to the Mercury, Apollo, Gemini, Soyuz, Orion, Crew Dragon, and Starliner launch escape systems... and yet those vehicles had them.

Why? Because launch is one of the highest-risk operations. In most other situations, you have time to diagnose and troubleshoot issues. During ascent to orbit, you sometimes have less than a second to react to problems that can arise from a very wide selection of machinery. Launch is the most important time to have a robust abort system; the probability of failure is (relatively) very high and the time permitted to troubleshoot is very low.

Furthermore: let's assume the LES is basically a Crew Dragon placed near the tip of the Starship, possibly with extra SRMs if a speedy exit from the Starship is required.

If a failure of the Starship happens in LEO, you just separate from the Starship and, at your leisure, make your reentry burn. If a failure happens during lunar injection or most of a Martian injection, you just wait until apogee and make a reentry burn. During EDL, you can just have the two vehicles reenter separately.

[magnemoe]

22 hours ago, sh1pman said:

The reason why I don't think LES is a good idea is because it only covers a tiny fraction of possible failure situations. It can only save lives if a failure happens during the burn of Super Heavy or shortly after the stage sep. LES is useless later into the launch. It won't help if a failure happens in orbit, during tanker randezvous, docking, refilling, departure burn, in transit to Mars/Moon, during atmospheric entry and landing, on the surface of another body, during liftoff from that body, burn to Earth and Earth atmospheric entry. It needs to be safe in all of these situations, not just first few minutes of launch. Airplanes are as safe as they are now not because passengers have escape pods.

So I think that instead of investing their time and money into the implementation of Starship LES, SpaceX should focus on making failures less likely.

Landing is probably the most dangerous part., followed by takeoff, reentry is also dangerous. 
yes an LES will not help you outside of takeoff and landing and only on Earth but this is the most dangerous part also far more common than mars missions. 

For the first manned versions of the Starship i would probably make cargo hold a bit larger, reduced crew capacity down to say 20 or less, add an simple escape pod on top. Double duty as an usable room so more volume than it has to be 

Later on they probably drop it I agree but not with the aggressive timeline they are running. 

[coyotesfrontier]

When I think of a LES for BFR, I think of something like this:

This would result in the lower half of the spaceship being destroyed, but 20-60 lives are more important.

[Spaceception]

44 minutes ago, Starman4308 said:

Literally the exact same logic could have been applied to the Mercury, Apollo, Gemini, Soyuz, Orion, Crew Dragon, and Starliner launch escape systems... and yet those vehicles had them.

Why? Because launch is one of the highest-risk operations. In most other situations, you have time to diagnose and troubleshoot issues. During ascent to orbit, you sometimes have less than a second to react to problems that can arise from a very wide selection of machinery. Launch is the most important time to have a robust abort system; the probability of failure is (relatively) very high and the time permitted to troubleshoot is very low.

Furthermore: let's assume the LES is basically a Crew Dragon placed near the tip of the Starship, possibly with extra SRMs if a speedy exit from the Starship is required.

If a failure of the Starship happens in LEO, you just separate from the Starship and, at your leisure, make your reentry burn. If a failure happens during lunar injection or most of a Martian injection, you just wait until apogee and make a reentry burn. During EDL, you can just have the two vehicles reenter separately.

Not really. In five of those cases, they never left LEO, in two they can go out to the Moon (or Mars in Orion's case), but at that point, the LES has already been detached. Plus they're all significantly smaller than Starship, and don't double as the second stage.

How would you design an LES for Starship that doesn't cut too much in the payload on ascent, and can be detached to avoid dead weight for the rest of the journey? Or is useful enough to keep the whole journey? Such as what @magnemoe suggested with making it a usable room. But that may only work when the crews are small. The interior is already somewhat larger than the ISS, you don't want to cut that down too much. And it's supposed to eventually carry a lot of people. While that may be reason enough to want an LES, it may just make it harder to have one in the first place. I think @sh1pman's points are something to consider. Is a stripped down Dragon with extra Draco's or SRMs even enough to carry the starship away from an exploding superheavy fast enough? Fuel and all?

I suppose launching the crew separately on crewed Dragons may be a good idea for the first missions, but it becomes infeasible if they get to the point of launching many of them to Mars at once.

Edited February 5, 2019 by Spaceception

[Starman4308]

4 minutes ago, Spaceception said:

Not really. In five of those cases, they never left LEO, in two they can go out to the Moon (or Mars in Orion's case), but at that point, the LES has already been detached. Plus they're all significantly smaller than Starship, and don't double as the second stage.

How would you design an LES for Starship that doesn't cut too much in the payload on ascent, and can be detached to avoid dead weight for the rest of the journey? Or is useful enough to keep the whole journey? Such as what @magnemoe suggested with making it a usable room. But that may only work when the crews are small. The interior is already somewhat larger than the ISS, you don't want to cut that down too much. And it's supposed to eventually carry a lot of people. While that may be reason enough to want an LES, it may just make it harder to have one in the first place. I think @sh1pman's points are something to consider. Is a stripped down Dragon with extra Draco's or SRMs even enough to carry the starship away from an exploding superheavy fast enough? Fuel and all?

I suppose launching the crew separately on crewed Dragons may be a good idea for the first missions, but it becomes infeasible if they get to the point of launching many of them to Mars at once.

I don't see your point about the LES for other launch vehicles. All of them had LES systems, even the borderline useless ejection seats of Gemini, for the most dangerous part of the mission: launch.

Once again: "It has to work this way for Mission X, therefore it will work this way" is a rather poisonous attitude. There's no need for early Starship missions to carry more than a few astronauts at a time, and speculating beyond that is premature. Airliner-level reliability is nice to talk about, but none of that has yet been demonstrated. For immediately forseeable missions, Starship can easily afford to keep an LES around.

[DDE]

50 minutes ago, Starman4308 said:

Once again: "It has to work this way for Mission X, therefore it will work this way" is a rather poisonous attitude.

It’s an attitude that gets missions cancelled over safety concerns.

1 hour ago, Spaceception said:

How would you design an LES for Starship that doesn't cut too much in the payload on ascent

You probably don’t.

Which is hardly surprising for a vehicle that gets most of its benefits by ditching certain aspects that added passive safety in most stages of flight, such as multiple-stage Mars orbiters.

[kerbiloid]

1 hour ago, Starman4308 said:

Literally the exact same logic could have been applied to the Mercury, Apollo, Gemini, Soyuz, Orion, Crew Dragon, and Starliner launch escape systems... and yet those vehicles had them.

Small capsules for 1...6 packed into a "spherical cone".
Easy to jettison, easy to protect from heat, no problem with orientation, no problem with overturn on landing.
Small chute(s), neither a bunch of twenty cupola, nor a stadium-sized cupola.

Crews of specially trained sporty people who can withstand 8..20 g on the ejection or the ballistic re-entry.

Every 1 of 70 fllights resulting in a catastrophe.

So, not exactly the same. 

1 hour ago, Starman4308 said:

Why? Because launch is one of the highest-risk operations. In most other situations, you have time to diagnose and troubleshoot issues. During ascent to orbit, you sometimes have less than a second to react to problems that can arise from a very wide selection of machinery. Launch is the most important time to have a robust abort system; the probability of failure is (relatively) very high and the time permitted to troubleshoot is very low.

4 catastrophes, 3 aborted flights.

Soyuz-1, -11, Columbia: reentry
Challenger: ascending, T+74.
Soyuz-18-1, Soyuz MS-10: ascending.
Soyuz T-10: lift-off.

So, just 1 of 7 on liftoff, 3 - on climbing, 3 - on aerobraking.

If add the flights when they have happily avoided a catastrophe, almost all of them were having troubles during aerobraking (heat protection leakage, disorientation, chute problems, etc).

One Soyuz - in orbit (1 of 2 engines malfunction)

One Apollo - just in flight.

1 hour ago, coyotesfrontier said:

This would result in the lower half of the spaceship being destroyed, but 20-60 lives

would stay on Earth because almost total payload capacity will be occupied by the heatshield and LES fuel.

1 hour ago, Starman4308 said:

All of them had LES systems, even the borderline useless ejection seats of Gemini

For the professional military pilots.

[tater]

2 hours ago, Spaceception said:

How would you design an LES for Starship that doesn't cut too much in the payload on ascent

Who cares about payload? If payload matters, send the cargo version.

Crew Starship can bring more people that we'd ever want to send someplace (forget colonies), and still have many 10s of tons of excess capacity. LES will cost a fraction of that mass.

 

 

1 hour ago, kerbiloid said:

So, just 1 of 7 on liftoff, 3 - on climbing, 3 - on aerobraking. 

So 4/7 on liftoff (ascent, all the same, crew on huge bomb). LES mitigates all 4 of these. That's 4/7 crews possibly alive for little cost.

Of the 3 remaining reentry problems, 1 was a parachute issue, one a valve that opened to space, right? Challenger was a legit reentry failure. So a valve could happen to Starship only if an airlock is somehow left open (there's no SM or OM to disconnect from which was the issue, right?), but such a problem will always be an issue for every crew vehicle, forever, so off the table. A parachute problem is a problem with the nominal landing system. That's exactly what LES use on landing would be for on Starship. Failure of landing system (propulsive in this case), results in LES firing, then they only die if those chutes also fail. better than certainly becoming a hole in the ground.

So basically 2/7 are left assuming the LES has no heat shield. If the LES has a heat shield, only 1/7 is left, and that's a problem that literally every spacecraft ever will always have.

[kerbiloid]

1 hour ago, tater said:

rew Starship can bring more people that we'd ever want to send someplace (forget colonies), and still have many 10s of tons of excess capacity. LES will cost a fraction of that mass.

Usually LES costs up to a half of the mass to save.
So, 10 t of LES means ~20..30 t of the capsule to save (including its heatshield).

1 hour ago, tater said:

So 4/7 on liftoff (ascent, all the same, crew on huge bomb).

Somebody (maybe Lockheed?) said that if an engine fails, it happens in the first 2 seconds after the ignition.
As we can see, with the crewed ships this happened only once (S.T-10), though happened.
(The abort case when the escape pod has to leave the fireball and jump to 1-2 km to open a chute or start gliding, at 12...20 g acceleration making even trained pilots feel bad).

Twice there was a structural failure (Challenger, S. MS-10) when the booster engines were still working, so the escape pod had to leave the mad rocket which is still chasing the pod.
So, still T/W.

Twice (S.MS-10, S.-18-1) it was aerobraking from the upper atmosphere at ~8 g, and at least once required a heatshield, not just the chutes.

Several times it was entering the atmosphere wrongly oriented and performed a ballistic descent, including a chaotic one (Volynov's). 

If read the book, almost 20-30 times the aerobraking of Shuttle and Soyuz had problems with heatshield (including the tail surface almost burnt through) of Shuttle and chutes (Soyuzes and one of Apollos).
Also, if, say, Columbia had an escape pod, is it for sure that it could safely separate at the hypersonic speed without killing the civilians inside?

Several times there was a situation when the deorbiting was possibly impossible.
S.-33, if the backup engine fails, too.
Columbia if the hole was found before its deorbiting.
One of Mercuries (when the pilot had spent almost all fuel in the RCS acrobatics).

At least twice the ship had almost sunk with the crew: one of Mercuries and one Soyuz when it had splashed into a winter lake, also overturned.
A shuttle would probably sink or crash.

Apollo-13 had unexpected crash in mid-space with no visible reason and had to continue flying, happily in a proper orbit.
If name this flight with its own name, it was a catastrophe with the full loss of the ship and the crew.
The only reason why they have survived was the feverish work of the ground specialists .
If the crew stayed with their ship face-to-face with their own skills and controls but without antennas and that cheating, they would die in several hours.
So, from this pov we can treat Apollo-13 as a ship loss. 

 

1 hour ago, tater said:

1 was a parachute issue, one a valve that opened to space, right?

Soyuz-1 (and almost all early Soyuzes until 12 when it had been totally redesigned) flight was one whole "my name is disaster".
Almost all of them had troubles in orbit and after due to poor design and implementation.
(Afair, the Soyuz-1 parachute had been kicked inside the compartment along the smooth veneer bands with a sledgehammer to fit the volume when the inner walls had been covered with unproper color, and the post-mortem tests have shown that the chute requires twice as greater force to pull out than by design).

The infamous Soyuz-11 valve caused the leakage paranoia still going on.
No other ship ever had problems with decompression, nobody needs those suits, but they still keep wearing them like talismans.
If S.-11 had no problems with that valve, everybody still would be laughing at the spacesuits in cabin.

***

So there were tens flight accidents close to a catastrophe, and almost all of them were inevitably requiring the controlled aerobraking of the escape vessel.
If the escape vessel was a plane, not a dropship, that means the escape pod should be an intact plane with heat protection, properly working control surfaces, and ability to perform basic orbital maneuvers to change the orbit.

As any spaceplane fuselage is too narrow to place a "spherical cone"/"conical sphere" for crew > 6..7 inside, so any spaceship with > 6..7 crew should be a reliable and redundant plane which is itself its escape pod.

 

Only 4 times there was a problem on start. Any of them was including a descent from high altitude, sometimes with a heat shield. So, the escape pod still should be a full-featured vessel.
Say, if the Challenger's head would be falling from near-orbit altitude and speed, without a heat protection it would burn even with parachutes onboard.

And as the LES main impulse causes accelerations inappropriate for a mass hiring, while it was used at least twice, the only way for a mass use spaceplane is to keep flying intact whatever happens.
No escape pods, they are just for small spaceplanes like DynaSoar, LKS, Hermes, etc..

Edited February 5, 2019 by kerbiloid

[tater]

2 minutes ago, kerbiloid said:

Usually LES costs up to a half of the mass to save.
So, 10 t of LES means ~20..30 t of the capsule to save (including its heatshield).

Dry mass of Starship is something like 100t, with no cargo. Add at least 100t to that. There's a lot of mass to play with. LES the entire nose, I don't care if it costs an extra 40 tons, it's worth it, certainly for early flights, and it still brings all the crew, AND 60+ tons to LEO. And it is by design refillable on orbit.

The nose of this beast is a capsule shape, seems like a no-brainer to me. Heatshield with a hatch in the bottom. Crew launches and reenters in the top capsule, on orbit they open hatch and use the rest of the volume (capsule can be 50 m^3, that still leaves 950 m^3 of crew space (minus the LES system)). Abort is any catastrophic failure on launch to orbit (heatshield in case it's a suborbital spaceflight that needs shielding). Abort also throughout EDL, the specifics of previous failures don't matter much (obviously if parachutes on LES fail, everyone dies).

[kerbiloid]

1 minute ago, tater said:

LES the entire nose,

Equip the entire nose with its own heatshield (from where? from back? from bottom?)
Make it self-stabilizing on aerobraking. Also if it aerobrakes with its back first, it will cause up to 10 g overloads killing your passengers.
Carry tens tonnes of the LES fuel every flight up and down, or eject not just the cabin, but the whole ship which weights much more than the cabin itself. Also, if you have ejected the whole whip, you don't need to separate its nose.

5 minutes ago, tater said:

The nose of this beast is a capsule shape, seems like a no-brainer to me.

Heatshield is on its back? Or on its bottom?
Also you have to place several tonnes of never used abort motor between the cabin and the tail, so between the wings. And suddenly you see that you have not just to eject the cabin, but burst the whole plane around, as the abort motor is right where the wings are attached to the fuselage.

8 minutes ago, tater said:

Crew launches and reenters in the top capsule,

And the passengers need a longer cabin, so the nose gets elongated.
Or vice versa the fuselage gets L/D of a cabbage.

[Rakaydos]

6 hours ago, kerbiloid said:

I don't mean their cost, just the 76 roaring mechanisms frighten me.

While a thick tank is just gently saying "Pshhh!..."
Also on Mars they would probably want something really simple.

Luddite. :p

[ment18]

Ejecting an entire capsule doesn't make sense, it requires changing the outer shape of the Spaceship to accommodate it, and it will never be able to be even close to off the shelf, you can't just throw a Dragon 2 onto a rocket and expect it to work.  Honestly, LES doesn't matter for Spaceship because its reliability is unprecedented.  We've never been able to fly the exact same vehicle (both booster and spaceship) five times a month, so you can prove the system works just by flying it ten times.  Failures are not random, they happen for reasons and a lot of those reason are manufacturing mistakes and failures that only happen in extreme situations.  Both of these are heavily mitigated by flying the same rocket many, many times.  By the time people are put in the Spaceship, the system will have flown hundreds of successful missions, making the cost of a LES completely pointless.  At some point, adding a LES is just increasing the number of parts that can fail catastrophically.  The Spaceship will be highly redundant and highly tested, LES is just a bandaid to the real problem.

[sevenperforce]

6 hours ago, kerbiloid said:

I don't mean their cost, just the 76 roaring mechanisms frighten me.

While a thick tank is just gently saying "Pshhh!..."
Also on Mars they would probably want something really simple.

Apart from the transpiration heat shield, the actual framework of the vehicle is ridiculously simple. They're packing all the complicated bits into the powerplant and letting autogenous pressurization and sheet metal take care of the rest. So if you need to repair on Mars or on LEO, NBD. 

Going pressure-fed would actually be MORE challenging. Harder to transfer propellant, harder to refuel, harder to repair if something goes bad.

8 minutes ago, ment18 said:

Ejecting an entire capsule doesn't make sense, it requires changing the outer shape of the Spaceship to accommodate it, and it will never be able to be even close to off the shelf, you can't just throw a Dragon 2 onto a rocket and expect it to work.  Honestly, LES doesn't matter for Spaceship because its reliability is unprecedented.  We've never been able to fly the exact same vehicle (both booster and spaceship) five times a month, so you can prove the system works just by flying it ten times.  Failures are not random, they happen for reasons and a lot of those reason are manufacturing mistakes and failures that only happen in extreme situations.  Both of these are heavily mitigated by flying the same rocket many, many times.  By the time people are put in the Spaceship, the system will have flown hundreds of successful missions, making the cost of a LES completely pointless.

While I only agree to a degree, I think this definitely expresses the SpaceX view. They can test and retest and re-retest the same hardware over and over, and "flight-proven" launch vehicles will ultimately end up being considered more reliable than anything off the shelf.

I mean, technically that's already true. SpaceX has reused launch vehicles 18 times with a 100% success rate.

But yeah, I wager that people will never fly on a brand new Starship. Every vehicle needs to go "there and back again" once, to show everything is working properly.

17 minutes ago, ment18 said:

At some point, adding a LES is just increasing the number of parts that can fail catastrophically. 

This is the part that people miss. At some point, the possibility of RUD falls so low that the possibility of being killed by your escape system is high by comparison, and then LES no longer makes any sense. Would I like a parachute if I'm flying on a brand new plane for the first time? Yes, I would...but not if it contains a grenade with the pin millimeters away from popping out unexpectedly.

[Starman4308]

58 minutes ago, ment18 said:

Honestly, LES doesn't matter for Spaceship because its reliability is unprecedented.

Oh, there's plenty of precedent. There's all sorts of experimental vehicles that have never flown with identical reliability records... no reliability record.

This is what I have a huge gripe with: people who assume that because the Starship/Superheavy is meant to be this wonderful thing that can fly over and over without fail, it will be this wonderful thing that can fly over and over without fail.

Reminds me a whole lot of the Space Shuttle. Granted, SpaceX has advantages here (such as no Thiokol insisting his SRBs be used), but this is wholly speculative.

31 minutes ago, sevenperforce said:

This is the part that people miss. At some point, the possibility of RUD falls so low that the possibility of being killed by your escape system is high by comparison, and then LES no longer makes any sense. Would I like a parachute if I'm flying on a brand new plane for the first time? Yes, I would...but not if it contains a grenade with the pin millimeters away from popping out unexpectedly.

Which is why, in peacetime, fighter jets have their ejection seats removed. Oh, no, wait, they don't. Because fighter jets still crash even in peacetime, and those are vastly more reliable than any demonstrated orbital launch vehicle.

It will be a very, very long time and tens of thousands of flights before I could be convinced flying to orbit without an LES would be a good idea.

Give it ten years, and maybe I'll be eating my words. Maybe Starship/Superheavy will work as intended and be the wave of the future; maybe it will achieve reliability on par with airliners. Until then, I'm reminded very heavily of the Space Shuttle and of the more conventional designs by much more experienced companies that still occasionally fail.

More likely, I suspect flaws will come out of the woodwork, and things will not work out as well as hoped. Probably still successful... but not "airliner-like" levels of reliability and reusability. It's just too different from anything that's been done before for there not to be skeletons in the design closet.

Edited February 5, 2019 by Starman4308

[sevenperforce]

12 minutes ago, Starman4308 said:

Oh, there's plenty of precedent. There's all sorts of experimental vehicles that have never flown with identical reliability records... no reliability record.

This is what I have a huge gripe with: people who assume that because the Starship/Superheavy is meant to be this wonderful thing that can fly over and over without fail, it will be this wonderful thing that can fly over and over without fail.

I think the argument is more -- when Starship/Superheavy starts flying people, it will have enough of a reliability record to justify it.

If it does not have a reliability record, it will not fly people.

12 minutes ago, Starman4308 said:

Which is why, in peacetime, fighter jets have their ejection seats removed. Oh, no, wait, they don't. Because fighter jets still crash even in peacetime, and those are vastly more reliable than any demonstrated orbital launch vehicle.

This is a conflation of reliability and accident risk. Fighter jets operate in an envelope where the risk of accident is far far greater than the reliability of the vehicle itself. Most fighter jet ejection scenarios happen because of operator error and/or pushing a vehicle beyond its limits, not because of a design or vehicle component failure.

A more apt description would be a CH-47 troop transport. Are accidents possible? Yes, they are. However, any ejection seat mechanism capable of safely ejecting the pilots and passengers would prevent the Chinook from doing its job and introduce greater risk than benefit, and the Chinook is not typically pushed into a regime beyond its design limits. 

12 minutes ago, Starman4308 said:

It will be a very, very long time and tens of thousands of flights before I could be convinced flying to orbit without an LES would be a good idea.

Starship and Superheavy are designed to be able to sustain the failure of any single major system and still recover. Engine-out? No worries, just push the other engines harder. Tank breach? No problem, just abort using the header tanks. Hydraulics lockup? Abort and then use engine gimbal to recover and land. Superheavy RUD? The Starship boosts itself to safety.

That's not to say that I would want to fly without an LES right away, either. But it's an issue of math. If the risk of your LES system malfunctioning is greater than the chance of needing it, then you don't want it.

[Starman4308]

7 minutes ago, sevenperforce said:

I think the argument is more -- when Starship/Superheavy starts flying people, it will have enough of a reliability record to justify it.

If it does not have a reliability record, it will not fly people.

This is a conflation of reliability and accident risk. Fighter jets operate in an envelope where the risk of accident is far far greater than the reliability of the vehicle itself. Most fighter jet ejection scenarios happen because of operator error and/or pushing a vehicle beyond its limits, not because of a design or vehicle component failure.

A more apt description would be a CH-47 troop transport. Are accidents possible? Yes, they are. However, any ejection seat mechanism capable of safely ejecting the pilots and passengers would prevent the Chinook from doing its job and introduce greater risk than benefit, and the Chinook is not typically pushed into a regime beyond its design limits. 

Starship and Superheavy are designed to be able to sustain the failure of any single major system and still recover. Engine-out? No worries, just push the other engines harder. Tank breach? No problem, just abort using the header tanks. Hydraulics lockup? Abort and then use engine gimbal to recover and land. Superheavy RUD? The Starship boosts itself to safety.

That's not to say that I would want to fly without an LES right away, either. But it's an issue of math. If the risk of your LES system malfunctioning is greater than the chance of needing it, then you don't want it.

On the first point: I suppose. I just get annoyed by people who assume that it will obtain that record as a matter of fact. There's very little evidence I can see to support it will reach that point. I'm not even sure we really disagree all that much: I am just more annoyed by the evidence-less claims that the Starship will get there.

On the conflation of reliability and accident risk: does it matter? Orbital launch vehicles, even with very cautious flight plans, still fail much more frequently than fighter jets intentionally pushed to their limits. There is an enormous gulf between where orbital launch vehicles are and where they would need to be for skipping the LES to be a good idea.

I agree in the abstract that at some point, a super-reliable vessel would actually be safer without LES. In the practical, orbital launch vehicles are assemblages of lots of high-energy propellants combined with complicated parts built to the extreme edges of engineering, exposed to extreme environments that are far less-well explored than the low atmosphere that most of humanity operates in. There's a few orders of magnitude missing in between where OLVs are now and where they would need to be for no-LES to be a good idea.

Redundancy has been a feature of every orbital launch vehicle ever. Even the Space Shuttle had a fair deal of redundancy. Not enough of it, to be certain, but a fair chunk of it. Redundancy is not a get-out-of-jail-free card. It is in SpaceX's favor that they are doing their best to have redundant parts as often as possible, but that does not guarantee success.

Overall, the thing Starship/Superheavy resembles most is other orbital launch vehicles. Most of the things that distinguish them from other OLVs are untested approaches like actively-cooled stainless-steel heat shielding. This does not fill me with overwhelming confidence.

[magnemoe]

6 hours ago, Spaceception said:

Not really. In five of those cases, they never left LEO, in two they can go out to the Moon (or Mars in Orion's case), but at that point, the LES has already been detached. Plus they're all significantly smaller than Starship, and don't double as the second stage.

How would you design an LES for Starship that doesn't cut too much in the payload on ascent, and can be detached to avoid dead weight for the rest of the journey? Or is useful enough to keep the whole journey? Such as what @magnemoe suggested with making it a usable room. But that may only work when the crews are small. The interior is already somewhat larger than the ISS, you don't want to cut that down too much. And it's supposed to eventually carry a lot of people. While that may be reason enough to want an LES, it may just make it harder to have one in the first place. I think @sh1pman's points are something to consider. Is a stripped down Dragon with extra Draco's or SRMs even enough to carry the starship away from an exploding superheavy fast enough? Fuel and all?

I suppose launching the crew separately on crewed Dragons may be a good idea for the first missions, but it becomes infeasible if they get to the point of launching many of them to Mars at once.

My idea was an usable room, the main room with the big window.  And no it will not work with an huge crew. However the dear moon crew is not above 20 either. 
I'm thinking short term, not mars colonization or use it for point to point. 
And it would be an escape pod not an spaceship. You can dock and rescue in orbit, during reentry you can use all the methane for cooling during deorbit, abort during reentry would be very hard anyway. 
Might reduce performance on escape system a bit you have cargo hull and crew quarters as an buffer. 

[tater]

1 hour ago, sevenperforce said:

A more apt description would be a CH-47 troop transport. Are accidents possible? Yes, they are. However, any ejection seat mechanism capable of safely ejecting the pilots and passengers would prevent the Chinook from doing its job and introduce greater risk than benefit, and the Chinook is not typically pushed into a regime beyond its design limits. 

I think this is a really poor analogy to Starship, however.

The use case for people in space is pretty much nonexistent already. Science? Robots are better. Manufacturing? Ditto. None the less, they want a crew version---for Mars. I'm not a Mars person, but that's their goal, so they'll work towards it, regardless. In the meantime, financial support for a crew version comes from #dearmoon, and if they want repeat business with crew, NASA, likely to the Moon (since that's the new, and more achievable focus). Unless they fly hundreds of times with no incidents before they consider putting people on it, I can only imagine NASA would demand some abort modes.

So they fly hundreds of times... How many times are they in fact likely to fly it even for Starlink when part of the point is to cheaply deploy Starlink (as few launches as possible), and they can likely do that in 10s of launches, not hundreds (I've seen some people say 20 fit on F9, others said 40, but even at 20, you can fit many, many more in Starship).

We all know that Shuttle really had no abort modes, but they acted as if they did.

Edited February 5, 2019 by tater

[magnemoe]

1 hour ago, sevenperforce said:

I think the argument is more -- when Starship/Superheavy starts flying people, it will have enough of a reliability record to justify it.

If it does not have a reliability record, it will not fly people.

This is a conflation of reliability and accident risk. Fighter jets operate in an envelope where the risk of accident is far far greater than the reliability of the vehicle itself. Most fighter jet ejection scenarios happen because of operator error and/or pushing a vehicle beyond its limits, not because of a design or vehicle component failure.

A more apt description would be a CH-47 troop transport. Are accidents possible? Yes, they are. However, any ejection seat mechanism capable of safely ejecting the pilots and passengers would prevent the Chinook from doing its job and introduce greater risk than benefit, and the Chinook is not typically pushed into a regime beyond its design limits. 

Starship and Superheavy are designed to be able to sustain the failure of any single major system and still recover. Engine-out? No worries, just push the other engines harder. Tank breach? No problem, just abort using the header tanks. Hydraulics lockup? Abort and then use engine gimbal to recover and land. Superheavy RUD? The Starship boosts itself to safety.

That's not to say that I would want to fly without an LES right away, either. But it's an issue of math. If the risk of your LES system malfunctioning is greater than the chance of needing it, then you don't want it.

Problem is Dear Moon and other projects they want to push. 
They don't want to wait to 2030 before doing manned missions. 

Rockets are more dangerous than fighter jets who again is so dangerous ejection seats are standard. System is so hard stressed and you operate on the edge of envelope. 
US has lost an number of fighter jets and at least 3 bombers the last 20 years. Just a couple has been in combat as in shot down. 

[sevenperforce]

31 minutes ago, Starman4308 said:

On the conflation of reliability and accident risk: does it matter? Orbital launch vehicles, even with very cautious flight plans, still fail much more frequently than fighter jets intentionally pushed to their limits. There is an enormous gulf between where orbital launch vehicles are and where they would need to be for skipping the LES to be a good idea.

As I said before, fighter jets are not necessarily the best analogue here, for several reasons.

That being said, there are still some points to be made. If my math is right, most non-combat jet ejections take place during air shows, which are NOT what the vehicle was designed for (and in which ejections are more likely because the vehicle is operating closer to the ground than typical, giving less time for correction). And ejections have killed pilots by mistake. RAF fighter pilot Sean Cunningham was killed in 2011 when his (poorly-installed) ejection seat handle was jostled and he was launched from the cockpit during preflight checks. That's not to say that ejection seats are a bad idea; they objectively save more lives than they take. But that's math. The equation (vehicle risk vs ejection system risk) is the same for an F-22 as it is for a Chinook; it's just a different solution.

Contrast should be drawn between Starship and STS. The Shuttle would have been safer with full-envelope LES, but did not have the mass budget for it. Starship has the mass budget for full-envelope LES (seriously, it can carry multiple Dragons to orbit easily) but that would most likely make it LESS safe, not more safe. 

31 minutes ago, Starman4308 said:

I agree in the abstract that at some point, a super-reliable vessel would actually be safer without LES. In the practical, orbital launch vehicles are assemblages of lots of high-energy propellants combined with complicated parts built to the extreme edges of engineering, exposed to extreme environments that are far less-well explored than the low atmosphere that most of humanity operates in. There's a few orders of magnitude missing in between where OLVs are now and where they would need to be for no-LES to be a good idea.

And this is where indefinite reuse comes in handy, because you can launch the same vehicle dozens of times at a fraction of conventional LV cost in order to prove complete reliability. That's where those orders of magnitude evaporate.

31 minutes ago, Starman4308 said:

Redundancy has been a feature of every orbital launch vehicle ever. Even the Space Shuttle had a fair deal of redundancy. Not enough of it, to be certain, but a fair chunk of it.

The Space Shuttle had ample redundancy in a handful of systems, but zero redundancy in the systems that really mattered.

1 hour ago, Starman4308 said:

Most of the things that distinguish them from other OLVs are untested approaches like actively-cooled stainless-steel heat shielding. This does not fill me with overwhelming confidence.

Oh, I have low confidence in an actively-cooled transpiration heat shield. 

But we can test it.

Once it tests, then we can talk.

20 minutes ago, magnemoe said:

My idea was an usable room, the main room with the big window.  And no it will not work with an huge crew. However the dear moon crew is not above 20 either. 
I'm thinking short term, not mars colonization or use it for point to point. 
And it would be an escape pod not an spaceship. You can dock and rescue in orbit, during reentry you can use all the methane for cooling during deorbit, abort during reentry would be very hard anyway. 
Might reduce performance on escape system a bit you have cargo hull and crew quarters as an buffer. 

Elon recently said that the Starship will now use cold-gas thrusters for simplicity rather than meth-oxy RCS thrusters.

If not for this, I would almost say "what if we put head-pressure tanks and RCS thrusters in the crew compartment...operational during a nominal mission, but always full so enough to pull it free during aborted launch/landing and enough for separation and pointing in an orbital RUD situation?"

The problem is that if you make the entire crew compartment its own capsule, then at some point it just makes more sense to re-enter separately, which means you have a three-stage vehicle rather than a two-stage vehicle. And then complexity overwhelms safety.

25 minutes ago, tater said:

I think this is a really poor analogy to Starship, however. The use case for people in space is pretty much nonexistent already.

No argument on use case. I desperately want one, but...yeah. I'm talking more about the vehicle itself. Ejection seats (and all the pyros required to make them work) on a Chinook introduces more failure modes than the likelihood of needing the ejection seats; the same is true for an operational Starship.

25 minutes ago, tater said:

Unless they fly hundreds of times with no incidents before they consider putting people on it, I can only imagine NASA would demand some abort modes.

So they fly hundreds of times... How many times are they in fact likely to fly it even for Starlink when part of the point is to cheaply deploy Starlink (as few launches as possible), and they can likely do that in 10s of launches, not hundreds (I've seen some people say 20 fit on F9, others said 40, but even at 20, you can fit many, many more in Starship).

We all know that Shuttle really had no abort modes, but they acted as if they did.

I think we all systematically overestimate the cost-per-launch with a fully-reusable LV.