totm nov 2023 SpaceX Discussion Thread

[CatastrophicFailure]

18 minutes ago, wumpus said:

You mean that real rocket scientists share one of my most annoying KSP design failures?

As a wise man once said: Check. Yo. Staging. 
Also don’t go drilling random holes in important bits...

[magnemoe]

On 6/18/2020 at 4:32 PM, sevenperforce said:

I'd be much happier riding a passively-stable capsule. Although chutes are a harder problem than KSP would lead us to believe.

When it comes to passive aerodynamic stability, a broad truncated cone (Apollo CM, Orion) is more stable than a narrow truncated cone (Crew Dragon), which is more stable than a blunted hemisphere (Soyuz), which is more stable than a blended lifting body (X-37), which is more stable than a spaceplane (Orbiter), which is more stable than whatever Starship is.

SpaceX has demonstrated that they can do retropropulsive landing, so I am not worried about that. I am worried about passive stability from re-entry interface until landing burn ignition. Just look at CRS-16. The booster had enough passive aerodynamic stability to survive entry and to maintain roughly consistent pointing up to landing burn ignition, and the engine burn successfully dropped it into the drink intact (although its subsequent, unavoidable tipover was not pretty). But Starship does not have enough passive stability to do this. A stuck hinge could mean Starship goes nose-down, tail-first, or tailspins...all of which are unrecoverable.

An falcon 9 stage is stable as the heavy engines is down and the grind fins generate lot of drag in the rear. 
Not sure if something could stuck the hinges with the force it is on them, now the electrical motors hold system has some complexity and probably need gearing, yes you will need multiple motors for redundancy, same for power system. One problem with motors over hydraulic is that the motors will use lots of power even then not moving if they counter air pressure on the flaps. 
On the other hand aerodynamic stability, if you designed an space shuttle style craft today its an good chance it would also be aerodynamic unstable simply because it would work better. 

And yes they do retropropulsive landings as routine but track record is not stellar even after they stopped calling them experimental. Yes landing is an secondary objective for falcon 9 and many of the landings has constrains. 
Still its one of the reason an escape system makes sense especially as I doubt Musk will wait 5 years before launching people on starship. 

[magnemoe]

On 6/18/2020 at 5:13 PM, tater said:

(the Orion capsule + LES is ~16t, and the LES is entirely external, so very costly from a mass standpoint)

Yes but you drop the nose LES with the first stage or boosters who reduces the mass penalty a lot, 

[tater]

49 minutes ago, magnemoe said:

Yes but you drop the nose LES with the first stage or boosters who reduces the mass penalty a lot, 

 

I'm not comparing SLS/Orion to Starship. The point is that the LES for Orion masses about what the capsule does, including a bunch of structure. A LES with more crew capability for Starship would have less mass than the Orion version, because it needs no external tractor/shroud/etc. The point is that you have margin to burn on SS. For the foreseeable future the only crew use case needs a relatively small number of crew, even given the huge size of SS. 10 people would be plenty. Put all 10 in a crew compartment that with integral LES and chutes masses the same as the Orion capsule + LES +extra, and we're only talking 20 tons for a crew vehicle that like most large crew habs is mostly empty anyway.

[RealKerbal3x]

1 hour ago, tater said:

 

I'm not comparing SLS/Orion to Starship. The point is that the LES for Orion masses about what the capsule does, including a bunch of structure. A LES with more crew capability for Starship would have less mass than the Orion version, because it needs no external tractor/shroud/etc. The point is that you have margin to burn on SS. For the foreseeable future the only crew use case needs a relatively small number of crew, even given the huge size of SS. 10 people would be plenty. Put all 10 in a crew compartment that with integral LES and chutes masses the same as the Orion capsule + LES +extra, and we're only talking 20 tons for a crew vehicle that like most large crew habs is mostly empty anyway.

Agreed. And 10 people is plenty for initial crew flights. If they're going to build a colony on Mars, it's going to have to be a small outpost with a few dozen people for a while before it gets big enough for a large population. By the time they can fly 100 people (it's going to be a long time) they might not need an LES anymore. I don't know how they would build one for 100 people anyway

I feel like we're going to see SS getting a launch escape system at some point, maybe with a detachable crew module at the top replacing that giant window. We have no idea what the layout of the crew cabin is and I don't know if SpaceX does either.

[tater]

I was chatting with a friend in the phone early this morning, and we were talking a bit about things that were just discussed here (failure modes, etc), and he went on some tangent and I had to remind him, "That wouldn't work on Mars." At which point he remembered the whole Mars requirement (like me, he's not a Mars colonization fan).

[Nightside]

1 hour ago, tater said:

I was chatting with a friend in the phone early this morning, and we were talking a bit about things that were just discussed here (failure modes, etc), and he went on some tangent and I had to remind him, "That wouldn't work on Mars." At which point he remembered the whole Mars requirement (like me, he's not a Mars colonization fan).

But is that a reasonable requirement? As this thing gets closer to reality I think we will see a lot more compromises. We already know there will be a Lunar variant, a Cargo variant, and Fueling variant. Why not add An Earth crew shuttle variant and a Mars Lander/Return variant - could make a lot more sense than an all-in-one package capable of landing on Earth and Mars. If the cost savings of a reusable 2nd stage are really as economical as SpaceX claims, we can afford to add more launches for safety.

An interplanetary mission will require several fueling trips before it is ready, will the full crew really be sitting around on the ship in orbit while fueling? You could make things a lot easier if you fuel the rocket in LEO before loading crew. Then send the crew shuttle, lets call it CrewShip. This variant only needs to transfer a certain number of people (~10- 20?) so it can have more redundant safety systems, extra flaps, 10 parachutes, emergency detachable crew capsule or whatever - it isn't carrying any major cargo, so it can afford to 

Design it for crewed landing on Mars

 

1. Launch uncrewed  MarsShip to LEO

2. Launch x FuelShips  to autonomously refuel (I don't recall how many fueling trips would be required).

3. Launch CrewShip to rendezvous with Marship

4.  Launch CargoShip with any other cargo not included on Launch #1.

5. Depart LEO with MarsShip, go to Mars, avoid 10000 possible deaths, then

6. Return from Mars to LEO, transfer crew back to CrewShip and descend in a safe manner. 

7. MarsShip also returns to Earth, but uncrewed so if it fails no one dies.

 

Meanwhile CrewShip can be used for tourists, Moon Flybys, or whatever. Also, a few  Dragon2 missons could fill the place of a CrewShip.

 

Thoughts?

[mikegarrison]

There are basically four known and proven kinds of emergency crew escape that work even if the main source of propulsion is unavailable:

1) individual bail out / ejection seat

2) crew pod ejection

3) whole vehicle parachute (this could include the standard capsule LES)

4) whole vehicle aerodynamic glide

I guess capsule LES being 2 or 3 kind of depends on whether you consider the capsule to be the "whole vehicle" or just a "crew pod".

Edited June 19, 2020 by mikegarrison

[tater]

I think alternate types would work, but I think part of their rationale is to get flight experience by having the same type.

[magnemoe]

21 hours ago, sevenperforce said:

A passenger airliner has multi-fault tolerance and multiple landing modes (runway landing, belly landing, splashdown). If the rudder becomes stuck, for example, the pilot can use differential engine thrust to counteract it. If an aileron locks up, the pilot can use leading edge flaps, slats, and elevators to balance and provide supplementary control. If the landing gear locks up they can belly land. If they lose both engines on ascent they have an aux power unit to provide enough power to operate the control surfaces, and on and on. There are contingency modes for practically any set of failures. 

Starship can ONLY land safely if all of its flaps work with sufficient precision to successfully execute the tail-flip AND if the tanks have sufficient pressure AND if the legs all lock out. There are no other options. It cannot glide. It cannot splash down. If a flaperon gets locked, it might be able to feather the others and survive re-entry but the tail-flip is going to be extra dicey.

Starship does not need an runway, you can land it in any fields downrange, if the legs are not able to hold it weight on soft soil it simply sink to the skirt and vacuum engines. Starship main constrain is fuel for powered landing. 
Yes it has to do the flip but it does not have to hit an target. 
Note that this is an issue for ocean landing and launch platforms or as I call them carriers as their size, complexity and functionality will be much more like an aircraft carrier than an barge. 

[tater]

 

[kerbiloid]

Early Buran had detachable LES lateral solid motors.
So, the upper part of the Spaceship could be implemented as a detachable compartment with lateral or internal LES motors, either liquid or solid.

Normally they could be used either to finally reach the orbit (like in Clipper), or to deorbit.
In this case solid ones should contain several fuel charges to be burnt either at once (LES) or one-by-one (deorbit or near-orbit LES).
 

The separated cabin with pax should have a slanted back with heat protection and a heat shield from one side, to form a sneaker shape when detached.
So, its adapter should be adapted for the slanted cabin back.

The cabin should have two extendable telescopic back trusses with small triangle fins, for passive stabilization during the aerobraking, like in Spiral escape capsule late design.
(Like not in the ogive one, but in the spherical one).
Or two folded long narrow winglets, for the same purpose.

After  aerobraking, the heatshield, together with flammable RCS, should be separated, releasing the single-use emergency landing damper (truss or legs).

Then it should follow the PTKNP way.
Release a drogue chute from top, to slow down to 20..30 m/s and get more or less horizontal.
(A large chute is bad because it's a good sail, the cabin will roll).

At 50 m or so altitude it should ignite the single-use solid-fuel landing engine.
On success - cut the chute.

The landing engine should have a common toroidal gas chamber with nozzles attached in several places,
and a set of fuel charges to be ignited one by one and exhaust through all nozzles at once.
So, if a charge fails, the next one should be ignited immediately.

The landing nozzles should be equipped with swinging shutters.
So, the landing engine uses its nozzles as both landing retropulsor and landing RCS, by redistributing the gas pressure between the nozzles and by swinging the nozzle shutters.

The cabin should hover at several meters altitude with zero velocity and get horizontal.
Then it should swing the nozzle shutters up, to make the net T/W <~ 1, and to land at 1..2 m/s speed on the damper (deformable legs or truss).

On touchdown the shutters should get full up, pressing the cabin to the ground until the current fuel charge stops burning.

***

As the cabin may have a specific shape, far from a cylinder, it can be put into the cylindric ship hull onto a radial ejection pad.
(Like a pad for a toy seaship to put it on flat table, but with solid ejection motors.)

Then, on the active abort situation, the pad should ignite its short-burn ejection motors, to push the cabin radially from the hull.
Once the cabin is out, the pad should detach from the cabin and ignite retroengines to separate from the pad and return into the crashing hull.

The cabin should radially fly away from the exploding ship for a cabin length distance and ignite nose and rear quick turn soild engines.
If the ship was falling nose down or flying horizontally (and so the cabin is still flying), the quick turn engines should quickly turn the capsule nose-up retrograde and cancel its rotation in this direction.
They should have high-thrust radial nozzles for that.

After that both quick turn engines should be separated and thrown radially from the cabin's course and exhaust.

Then LES engines ignite and things follow same scheme.

***

This is also how a spaceplane ejectable cabin should be implemented.

***

The idea of a long cylindric passenger ship is nice for retro-sci-fi, but absolutely sick in real world, from any pov.
Its irl implementation makes no sense even to discuss.

Though, if the passengers are cheap and expendable, they can just chain them with several axial strings along the payload bay, and short chains attached to collars,
and implement the LES system just for the crew cabin like described above.

(Anyway, highly likely, it will be the main way to send human resources to Martian colony, after one or two starship accidents).

***

The idea of having same capsule for both Earth and Mars is same absolutely sick, until you get thermonukes with enough delta-V budget to slowly land or lift without aerobraking.

Edited June 20, 2020 by kerbiloid

[RCgothic]

A new thrust puck has appeared. First sighting of SN8 or 9?

 

[magnemoe]

15 hours ago, Nightside said:

But is that a reasonable requirement? As this thing gets closer to reality I think we will see a lot more compromises. We already know there will be a Lunar variant, a Cargo variant, and Fueling variant. Why not add An Earth crew shuttle variant and a Mars Lander/Return variant - could make a lot more sense than an all-in-one package capable of landing on Earth and Mars. If the cost savings of a reusable 2nd stage are really as economical as SpaceX claims, we can afford to add more launches for safety.

An interplanetary mission will require several fueling trips before it is ready, will the full crew really be sitting around on the ship in orbit while fueling? You could make things a lot easier if you fuel the rocket in LEO before loading crew. Then send the crew shuttle, lets call it CrewShip. This variant only needs to transfer a certain number of people (~10- 20?) so it can have more redundant safety systems, extra flaps, 10 parachutes, emergency detachable crew capsule or whatever - it isn't carrying any major cargo, so it can afford to 

Design it for crewed landing on Mars

 

1. Launch uncrewed  MarsShip to LEO

2. Launch x FuelShips  to autonomously refuel (I don't recall how many fueling trips would be required).

3. Launch CrewShip to rendezvous with Marship

4.  Launch CargoShip with any other cargo not included on Launch #1.

5. Depart LEO with MarsShip, go to Mars, avoid 10000 possible deaths, then

6. Return from Mars to LEO, transfer crew back to CrewShip and descend in a safe manner. 

7. MarsShip also returns to Earth, but uncrewed so if it fails no one dies.

 

Meanwhile CrewShip can be used for tourists, Moon Flybys, or whatever. Also, a few  Dragon2 missons could fill the place of a CrewShip.

 

Thoughts?

They could launch one tanker, then launch more tankers to fill up that tanker, then launch the crewed ship who take most of the fuel from the tanker leaving it just enough to land. no extra launches needed 
This could also be practical if launching expensive stuff like an space telescope to L2 as you just do one docking operation. 
You also avoid the problem of getting stuck in orbit if something interrupt launches like problems on launchpad or you need to inspect the tankers for some issue. 

Now you could technicality abort on Mars, assuming you have both header tanks in the nose and just the top 1 or 2 floors of crew module being ejected you will have pretty decent dV.
Yes this require that you land at an base who has an rover who can dock to the escape module somehow also that the escape module is stable without atmosphere. 
 

[sevenperforce]

16 hours ago, Nightside said:

Why not add An Earth crew shuttle variant and a Mars Lander/Return variant - could make a lot more sense than an all-in-one package capable of landing on Earth and Mars.

You forget that the vehicles which land on Mars need Earth return capability. Earth return is also highest pucker-factor, so they want to use the Earth EDL as the baseline and have the most experience with them. Which means they need to design in Mars EDL. 

[wumpus]

19 hours ago, RealKerbal3x said:

Agreed. And 10 people is plenty for initial crew flights. If they're going to build a colony on Mars, it's going to have to be a small outpost with a few dozen people for a while before it gets big enough for a large population. By the time they can fly 100 people (it's going to be a long time) they might not need an LES anymore. I don't know how they would build one for 100 people anyway

I feel like we're going to see SS getting a launch escape system at some point, maybe with a detachable crew module at the top replacing that giant window. We have no idea what the layout of the crew cabin is and I don't know if SpaceX does either.

Wouldn't a LES for Starship just need enough thrust for Starship to detach and leave Super Heavy Booster (or whatever they call stage 1)?  Enough high-boost short-duration SRBs should do the trick, and then Starship should be able to land on its own (it might have to dump propellant before landing).

[Nightside]

 

1 hour ago, magnemoe said:

They could launch one tanker, then launch more tankers to fill up that tanker, then launch the crewed ship who take most of the fuel from the tanker leaving it just enough to land. no extra launches needed 

I like this idea, it is basically a propellant depot, I guess. I suspect refueling operations will be very tricky, given the need for an inertial kick to get the fluids flowing, there are a lot of things that could go catastrophically wrong. Even if the MarsShip is only refueled once,  I don't see any reason to risk having crew aboard during that operation anytime soon.

47 minutes ago, sevenperforce said:

You forget that the vehicles which land on Mars need Earth return capability. Earth return is also highest pucker-factor, so they want to use the Earth EDL as the baseline and have the most experience with them. Which means they need to design in Mars EDL. 

Just because the ship will return to Earth doesn't mean the crew needs to land with it. I didn't forget, I guess I am assuming that Mars and Earth are just different EDL problems, but you could be right that Mars is just an easier version of Earth (assuming you were going to land on rockets in either case).

12 minutes ago, wumpus said:

then Starship should be able to land on its own (it might have to dump propellant before landing).

Dumping hundreds of tons of props during an emergency separation/ landing seems like it would be hard to do. It seems like you would be descending in a cloud of combustible material, right before your landing burn.

[tater]

10 minutes ago, wumpus said:

Wouldn't a LES for Starship just need enough thrust for Starship to detach and leave Super Heavy Booster (or whatever they call stage 1)?  Enough high-boost short-duration SRBs should do the trick, and then Starship should be able to land on its own (it might have to dump propellant before landing).

The problem is that SS itself could have an issue. Given the nature of how it lands, for EDL everything works, or they die (I suppose there are multiple engines, so engine failure has redundancy at least).

An ejectable crew pod seems like a reasonable compromise for crew vehicle testing since they have so much mass margin.

[RealKerbal3x]

8 minutes ago, wumpus said:

Wouldn't a LES for Starship just need enough thrust for Starship to detach and leave Super Heavy Booster (or whatever they call stage 1)?  Enough high-boost short-duration SRBs should do the trick, and then Starship should be able to land on its own (it might have to dump propellant before landing).

Possibly, though if you have several abort SRBs then you run into some potential problems:

1) If anything goes wrong with Starship itself during its ascent burn then you have no way to escape.

2) You have to lug several large abort motors around, which are essentially dead weight after you reach orbit. You could detach them when they're no longer needed, but that goes against the philosophy of rapid reusability.

3) Any abort mode depends on the landing sequence working, which isn't good if you've already had a failure that requires an abort. As @tater already said, EDL with SS requires everything to work.

At least for the first crewed flights, a detachable crew cabin is best. Given that it already would have heatshield tiles on one side, it could probably enter biconically and land via parachutes.

 

[Technical Ben]

1 hour ago, magnemoe said:

They could launch one tanker, then launch more tankers to fill up that tanker, then launch the crewed ship who take most of the fuel from the tanker leaving it just enough to land. no extra launches needed 
This could also be practical if launching expensive stuff like an space telescope to L2 as you just do one docking operation. 
You also avoid the problem of getting stuck in orbit if something interrupt launches like problems on launchpad or you need to inspect the tankers for some issue. 

Now you could technicality abort on Mars, assuming you have both header tanks in the nose and just the top 1 or 2 floors of crew module being ejected you will have pretty decent dV.
Yes this require that you land at an base who has an rover who can dock to the escape module somehow also that the escape module is stable without atmosphere. 
 

That sounds quite interesting, for quicker turnarounds of crewed launch/transit. Have an orbital fuel dump/depot, but due to boil off, only refuel it last min to rendevouz with the crew ship.

[wumpus]

14 minutes ago, tater said:

The problem is that SS itself could have an issue. Given the nature of how it lands, for EDL everything works, or they die (I suppose there are multiple engines, so engine failure has redundancy at least).

An ejectable crew pod seems like a reasonable compromise for crew vehicle testing since they have so much mass margin.

If you have an ejectable crew pod, I'd make sure that it is fully capable of re-entry and use that to return from orbit.  By the time Starship fails a landing attempt, it probably requires more delta-v than an ejectable pod would have.  The idea of using SRBs to detach Starship would be for routine launches with many passengers (and obviously only works for booster failures).

For a Starship failure, I *think* that the engines to orbit and the landing engines are different (and almost nothing is frozen in Starship's design).  They still will require quite a few common parts, but perhaps they can either be reliable or redundant (preferably both) to use for both orbit and abort.

6 minutes ago, RealKerbal3x said:

Possibly, though if you have several abort SRBs then you run into some potential problems:

1) If anything goes wrong with Starship itself during its ascent burn then you have no way to escape.

2) You have to lug several large abort motors around, which are essentially dead weight after you reach orbit. You could detach them when they're no longer needed, but that goes against the philosophy of rapid reusability.

3) Any abort mode depends on the landing sequence working, which isn't good if you've already had a failure that requires an abort. As @tater already said, EDL with SS requires everything to work.

At least for the first crewed flights, a detachable crew cabin is best. Given that it already would have heatshield tiles on one side, it could probably enter biconically and land via parachutes.

 

Since the abort motors should only fire for a few seconds, they shouldn't be all that heavy.  After that you can light the raptors (presumably you'd light both, cutting down on the thrust you'd need on the abort motors).

I've already said I don't like the idea of landing on Starship at all (at least until *all* the bugs are worked out), and would far prefer to land via Dragon for the first >>100 landings (Falcon 9 boosters failed to land on launches 81 and 83 after a string of successes).

[Technical Ben]

Doesn't SS already have enough thrust for most aborts it's self?

[tater]

4 minutes ago, wumpus said:

If you have an ejectable crew pod, I'd make sure that it is fully capable of re-entry and use that to return from orbit.  By the time Starship fails a landing attempt, it probably requires more delta-v than an ejectable pod would have.  The idea of using SRBs to detach Starship would be for routine launches with many passengers (and obviously only works for booster failures).

Needless complexity, IMO.

SS will survive even a bad reentry long enough to let the escape pod eject the crew, IMO.

[RealKerbal3x]

Just now, Technical Ben said:

Doesn't SS already have enough thrust for most aborts it's self?

It's got plenty of thrust, but that's because it has to push 3600t of propellant and up to 150t of payload. Fully fuelled, it doesn't have the TWR to do a safe abort, even if you fire all of the Raptors at once.

[SOXBLOX]

Pretty sure it has just enough thrust with all Raptors to barely ascend once fully fueled.