Starship Artemis Contract & Lunar Starship

[thespacecow]

We've seen so little testing of possible thruster systems that I think they must be planning on using SuperDracos for Option A.  Landing mass should be somewhere around 300t with the ascent prop.  I've been figuring the Raptor shutdown height as 100m, with -10m/s residual velocity.¹  That yields about 710kN of thrust, which is 10 SDs, arranged in 5 pods, 70º apart.  Shouldn't be too hard to keep them out of the way of the drogues.

The potential fly in the ointment is they also need SDs for initial ascent.  If they fix the backflow bug, there's no layout problem.  But if they don't fix the bug, and they need 8 separate SDs to get the HLS high enough to start the Raptors without FOD risk, now you're looking at 9 pods, 40º apart.

That would be getting a little crowded.

There's no way they would use SD for that, Elon would cancel HLS first before putting hypergolic on Starship...

(These are small engines, not easy to spot at McGregor, but I'm pretty sure NSF video showed some candidates. Plus NASA HLS update also mentioned them developing hot gas thrusters)

[rsdavis9]

1) Approach and stationkeeping preliminary to hookup. Ship and depot are parallel and roughly aligned for hookup close enough for the arms to reach. Solution: Very slow maneuvering with no avoidable jerks. Stop and allow slosh to settle as necessary. Apply settling thrust on both ships as necessary. Given the potential mass disparity between ships, some combination of deep throttling and clustering of small maneuvering engines would be called for.

This is the most fraught.  Ultimately, prox ops and docking is a control problem.  You can make the control loop as long-duration as you want, but it does no good if you can't predict the next state with some bounded amount of error.  That's hard to do.

Even extremely low-speed collisions of Starship parts that aren't designed for colliding (i.e., the probes and drogues) is very, very bad.  Getting the Ships close enough while avoiding the collisions is very difficult.

2) Hookup of arms and drawing ships together to make the QD connection. Solution: slow and gentle. Give time for settling thrust and propellant viscosity to settle things down.

I think this is pretty easy.  Whatever slosh force you have to absorb during retraction is no greater than the force you applied to retract in the first place.

3) Propellant transfer and maneuvering for thermal control. Solution: Until we have a sense of expected thermal mitigation maneuvers we have no specifics on the forces that will slosh the propellant. In a general way we can say gentle without jerks is good. Another general point is that it will be best to introduce propellant at the bottom of the tank so that the mass and viscosity of the proceeding propellant acts as a damper on violent movement. Periodically stopping the transfer to give some settling time may he necessary. It may also be beneficial for the propellant inlet tube to have two or more valved outlets at the tank bottom so that if a swirl (for example) builds up the incoming fluids can be switched to counter it.

Until it's worked out it will be a slow and tedious process, probably slowest for the first tanker transfer and faster as the propellant volume increases.

It can't be so slow and tedious that the depot accumulation cycle is too long.

I'm starting to think that propellant containment may be a better way to go.  A sump tank at the base of the Starship can be very light:  it only has to withstand hydrostatic forces during launch (i.e. the ullage pressure inside the sump is the same outside the sump), and the lateral forces are completely determined by the forces applied while maneuvering.

And it's a cheesy dewar.  You've still got some ullage gas to conduct/convect heat from the outer walls, but you can reduce that pressure as much as you want with very little boil-off, because the prop in the dewar stays cold.

The fly in the ointment is sizing the tanks for all of the needed applications.  I don't know how to do that.

Hover slam for docking. It is only a change in the direction of the acceleration vector that causes the slosh. Do it fast and precisely.

[meekGee]

1) Approach and stationkeeping preliminary to hookup. Ship and depot are parallel and roughly aligned for hookup close enough for the arms to reach. Solution: Very slow maneuvering with no avoidable jerks. Stop and allow slosh to settle as necessary. Apply settling thrust on both ships as necessary. Given the potential mass disparity between ships, some combination of deep throttling and clustering of small maneuvering engines would be called for.

This is the most fraught.  Ultimately, prox ops and docking is a control problem.  You can make the control loop as long-duration as you want, but it does no good if you can't predict the next state with some bounded amount of error.  That's hard to do.

Even extremely low-speed collisions of Starship parts that aren't designed for colliding (i.e., the probes and drogues) is very, very bad.  Getting the Ships close enough while avoiding the collisions is very difficult.

2) Hookup of arms and drawing ships together to make the QD connection. Solution: slow and gentle. Give time for settling thrust and propellant viscosity to settle things down.

I think this is pretty easy.  Whatever slosh force you have to absorb during retraction is no greater than the force you applied to retract in the first place.

3) Propellant transfer and maneuvering for thermal control. Solution: Until we have a sense of expected thermal mitigation maneuvers we have no specifics on the forces that will slosh the propellant. In a general way we can say gentle without jerks is good. Another general point is that it will be best to introduce propellant at the bottom of the tank so that the mass and viscosity of the proceeding propellant acts as a damper on violent movement. Periodically stopping the transfer to give some settling time may he necessary. It may also be beneficial for the propellant inlet tube to have two or more valved outlets at the tank bottom so that if a swirl (for example) builds up the incoming fluids can be switched to counter it.

Until it's worked out it will be a slow and tedious process, probably slowest for the first tanker transfer and faster as the propellant volume increases.

It can't be so slow and tedious that the depot accumulation cycle is too long.

I'm starting to think that propellant containment may be a better way to go.  A sump tank at the base of the Starship can be very light:  it only has to withstand hydrostatic forces during launch (i.e. the ullage pressure inside the sump is the same outside the sump), and the lateral forces are completely determined by the forces applied while maneuvering.

And it's a cheesy dewar.  You've still got some ullage gas to conduct/convect heat from the outer walls, but you can reduce that pressure as much as you want with very little boil-off, because the prop in the dewar stays cold.

The fly in the ointment is sizing the tanks for all of the needed applications.  I don't know how to do that.

Hover slam for docking. It is only a change in the direction of the acceleration vector that causes the slosh. Do it fast and precisely.

  can't even if you wanted to..b the end state is zero g, there'd be rebound.

Not a simple problem.

[OTV Booster]

1) Approach and stationkeeping preliminary to hookup. Ship and depot are parallel and roughly aligned for hookup close enough for the arms to reach. Solution: Very slow maneuvering with no avoidable jerks. Stop and allow slosh to settle as necessary. Apply settling thrust on both ships as necessary. Given the potential mass disparity between ships, some combination of deep throttling and clustering of small maneuvering engines would be called for.

This is the most fraught.  Ultimately, prox ops and docking is a control problem.  You can make the control loop as long-duration as you want, but it does no good if you can't predict the next state with some bounded amount of error.  That's hard to do.

Even extremely low-speed collisions of Starship parts that aren't designed for colliding (i.e., the probes and drogues) is very, very bad.  Getting the Ships close enough while avoiding the collisions is very difficult.

2) Hookup of arms and drawing ships together to make the QD connection. Solution: slow and gentle. Give time for settling thrust and propellant viscosity to settle things down.

I think this is pretty easy.  Whatever slosh force you have to absorb during retraction is no greater than the force you applied to retract in the first place.

3) Propellant transfer and maneuvering for thermal control. Solution: Until we have a sense of expected thermal mitigation maneuvers we have no specifics on the forces that will slosh the propellant. In a general way we can say gentle without jerks is good. Another general point is that it will be best to introduce propellant at the bottom of the tank so that the mass and viscosity of the proceeding propellant acts as a damper on violent movement. Periodically stopping the transfer to give some settling time may he necessary. It may also be beneficial for the propellant inlet tube to have two or more valved outlets at the tank bottom so that if a swirl (for example) builds up the incoming fluids can be switched to counter it.

Until it's worked out it will be a slow and tedious process, probably slowest for the first tanker transfer and faster as the propellant volume increases.

It can't be so slow and tedious that the depot accumulation cycle is too long.

I'm starting to think that propellant containment may be a better way to go.  A sump tank at the base of the Starship can be very light:  it only has to withstand hydrostatic forces during launch (i.e. the ullage pressure inside the sump is the same outside the sump), and the lateral forces are completely determined by the forces applied while maneuvering.

And it's a cheesy dewar.  You've still got some ullage gas to conduct/convect heat from the outer walls, but you can reduce that pressure as much as you want with very little boil-off, because the prop in the dewar stays cold.

The fly in the ointment is sizing the tanks for all of the needed applications.  I don't know how to do that.

All true that.


Slosh comes in two flavors. Chaotic splashing and back and forth non splashing swells. Chaotic absolutely has to be damped down before the ships are near enough to touch. With settling thrust and gentle lateral thrust applied the swells will damp down in their own over time. Unlike chaotic splashing the swells will be periodic and easy (a relative term) to predict and counter.


Still, as you say, time is the enemy and a tank within a tank may well be the way to go. This is probably one of those things where modeling will be inadequate and hands on will show the way.

[TheRadicalModerate]

We've seen so little testing of possible thruster systems that I think they must be planning on using SuperDracos for Option A.  Landing mass should be somewhere around 300t with the ascent prop.  I've been figuring the Raptor shutdown height as 100m, with -10m/s residual velocity.¹  That yields about 710kN of thrust, which is 10 SDs, arranged in 5 pods, 70º apart.  Shouldn't be too hard to keep them out of the way of the drogues.

The potential fly in the ointment is they also need SDs for initial ascent.  If they fix the backflow bug, there's no layout problem.  But if they don't fix the bug, and they need 8 separate SDs to get the HLS high enough to start the Raptors without FOD risk, now you're looking at 9 pods, 40º apart.

That would be getting a little crowded.

There's no way they would use SD for that, Elon would cancel HLS first before putting hypergolic on Starship...

(These are small engines, not easy to spot at McGregor, but I'm pretty sure NSF video showed some candidates. Plus NASA HLS update also mentioned them developing hot gas thrusters)

It's not forever, and HLS Starships are going to start out being expendable.  Handling MMH/NTO a couple of times a year on an expendable isn't going to be a major impediment to ops.

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

I would expect to see a pump-fed thruster in the 50-100kN range show up for a whole bunch of different applications on Starship:  low-delta-v maneuvers like deorbit and TCMs, coarse attitude control, and landing for both lunar and Mars applications.  It probably has to be pump-fed because the pressurant demands for a pressure-fed system are simply too large for burn times exceeding one minute.  I think they've put this off simply because they can live without it for a while.

The big problem with SD is the backflow bug.  I guess there's an argument to be made that you need thrusters for landing, but not for ascent, but I suspect the FOD risk is just as great on ascent.  But I also think it's fairly easy to qualify a fix for the backflow bug; they just haven't needed to implement it for the D2.

[TheRadicalModerate]

Slosh comes in two flavors. Chaotic splashing and back and forth non splashing swells. Chaotic absolutely has to be damped down before the ships are near enough to touch. With settling thrust and gentle lateral thrust applied the swells will damp down in their own over time. Unlike chaotic splashing the swells will be periodic and easy (a relative term) to predict and counter.

A truly excellent banner!  It can be mounted on the wall opposite the one that says, "There's never enough time to do it right, but always enough time to do it over."

The worst case slosh is one where most of the prop flies straight across the tank into the far wall--but with a quasi-chaotic impact.  That's likely what you get with any braking maneuver:  the forward motion pins the prop against one wall, and the braking detaches it sorta-kinda cleanly, in the opposite direction.

[Paul451]

"There's never enough time to do it right, but always enough time to do it over."

(Or the military version, "Slow is smooth, smooth is fast.")

Off-topic aside: I feel the intent of that meme (rushing doesn't actually save time) only applies to things where you know, in detail, what you have to do. A familiar task. When applied to something you haven't done before, it still applies... but with the opposite intent. It's faster to do it multiple times until you get it right, than it is to get it right on the first attempt. Significantly faster. (And the gap increases with the complexity of the task, because it applies to each sub-task.)

[thespacecow]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

[rsdavis9]

Hover slam for docking. It is only a change in the direction of the acceleration vector that causes the slosh. Do it fast and precisely.

  can't even if you wanted to..b the end state is zero g, there'd be rebound.

Not a simple problem.

How much rebound is there from a liquid which is nearly incompressible? I would think not a lot. Also a hover slam can be done where the lowest g force is at the end of the slam(if you have complete throttle control). I would think they are not going to do mating operations with raptors and will be using gaseous thrusters with complete throttle control.

So yes the term hover slam is probably the wrong word. It isn't slamming at the end. It is more like a deceleration that is smooth in change in both magnitude and direction of the acceleration vector. Also the direction of the vector being the most important to not change.

[meekGee]

Hover slam for docking. It is only a change in the direction of the acceleration vector that causes the slosh. Do it fast and precisely.

  can't even if you wanted to..b the end state is zero g, there'd be rebound.

Not a simple problem.

How much rebound is there from a liquid which is nearly incompressible? I would think not a lot. Also a hover slam can be done where the lowest g force is at the end of the slam(if you have complete throttle control). I would think they are not going to do mating operations with raptors and will be using gaseous thrusters with complete throttle control.

So yes the term hover slam is probably the wrong word. It isn't slamming at the end. It is more like a deceleration that is smooth in change in both magnitude and direction of the acceleration vector. Also the direction of the vector being the most important to not change.

Plenty I think, since the tank wall is elastic (in stretch, but especially in shape, as the fluid hammer is not symmetrical) and then there's pressure waves.

A lot of Starship "why don't they just XXXX" questions are answered by the fact that while it looks all shiny and rigid, it's very flimsy and elastic and you can't treat any of its geometry as a given.

[OTV Booster]

May I suggest a frame of reference that makes slosh a little bit more understandable?

Assume an unobtainably stable condition with unmoving fuel in the nose of an idealized tank. Apply small momentary thrust in the normal thrust direction. The fuel WILL NOT MOVE. The ship will move around the unmoving fuel. The physics is the same for moving fuel and moving ship but in my mind following the actual energy flow helps clarify the later non idealized behavior.

Lets follow this a bit and see where it leads.  The part of bolus of fuel in contact with the tank walls will shear off leaving a thin unmoving film and in the process act as a slight brake on the ships movement. As the bolus contacts the first beginning inward curve of the rearward dome it will pick up an inward vector that increases as the dome curve steepens. Ignoring the snail tracks left on the walls, this inward movement is the first fuel movement relative to an outside frame of reference.

Because this new vector is applied around the circumference, the outward radial energy transfer to the tank cancels out and only the 'down' component has (literal) impact on the tank.

A lot more follows and it gets chaotic fairly fast unless everything is happening so slow that viscosity is the dominant force. Not gonna happen but viscosity will be there and it has an impact that may be worth examining.

   Big whorls have little whorls
   Which feed on their velocity,
   And little whorls have lesser whorls
   And so on to viscosity.
         Lewis Fry Richardson


Chaotic movement will warm the propellants. How much, I do not know.

[TheRadicalModerate]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

The last test I heard of was about 3 years ago.  Has there been something since then?

I'm almost certain they have something in the works.  But if it were a key element for Option A/B, we'd have seen a lot more of it by now.

[rsdavis9]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

The last test I heard of was about 3 years ago.  Has there been something since then?

I'm almost certain they have something in the works.  But if it were a key element for Option A/B, we'd have seen a lot more of it by now.

Maybe it fits into "it just works". COPV's with high pressure O2 and CH4. Valves to thrusters with igniters.
Simple?

[TheRadicalModerate]

Maybe it fits into "it just works". COPV's with high pressure O2 and CH4. Valves to thrusters with igniters.
Simple?

That doesn't sound like the test summary you'd send to the NASA mission assurance guys.

[OTV Booster]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

The last test I heard of was about 3 years ago.  Has there been something since then?

I'm almost certain they have something in the works.  But if it were a key element for Option A/B, we'd have seen a lot more of it by now.

Not really. There's no reason for landing engines to be at BC until shortly before they're ready to test them. And unlike Raptors, they can be crated and are small enough to hit the loading dock without raising a stir.


IMO, the next big thing will be making orbit and V3 needs at least one good suborbital first. Then they need get on the stick on propellant transfer. Transfer will need maneuvering engines that might or might not share heritage with lunar landing engines. I hope they do.


I also suspect that transfer maneuvers will need a throttle range beyond the ability of any one engine, leading to clustering. And engine clustering is an SX trade mark. Rethink the HLS with each landing engine replaced by a cluster of 5-8 (a totally arm wavy number) small engines.


There's every reason to expect something more than cold gas thrusters for transfer tests. Maybe that'll be our window on the landing engines.

[OTV Booster]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

The last test I heard of was about 3 years ago.  Has there been something since then?

I'm almost certain they have something in the works.  But if it were a key element for Option A/B, we'd have seen a lot more of it by now.

Maybe it fits into "it just works". COPV's with high pressure O2 and CH4. Valves to thrusters with igniters.
Simple?

A gas/gas pressure fed engine question.


Are they simple enough that scaling up or down is easily calculable?  Not necessarily dead on but close enough that it's relatively trivial to get them tweaked in?

[TheRadicalModerate]

Not really. There's no reason for landing engines to be at BC until shortly before they're ready to test them. And unlike Raptors, they can be crated and are small enough to hit the loading dock without raising a stir.

Not BC, McGregor.  Nobody's running a test campaign there, either.  It's not a big deal adding the thrusters late in the program, but it's a big deal not to be testing them yet.

Rethink the HLS with each landing engine replaced by a cluster of 5-8 (a totally arm wavy number) small engines.

That's what I've been assuming.  12 engines suspiciously close to SD's thrust will work fine, and allow one engine out.  However, in the artwork, HLS appears to have four pods of 3 each.  That's a bit un-SD-like.

There's every reason to expect something more than cold gas thrusters for transfer tests. Maybe that'll be our window on the landing engines.

That's a fair point, but settling thrusters are gonna be <1kN.  Those are the kind of things you could test in a chamber, rather than on a test stand.

[thespacecow]

Remember, we're talking about NASA-grade crew certification here.  If they have a brand-new, mission-critical, crew-critical thruster, which will have to burn for about 45s on descent and a few seconds on ascent, we'd have seen a lot more activity than we've seen.  It's not like SpaceX is secretive about stuff like this.  If we don't have evidence that it exists, it's likely because it doesn't exist--yet.

That's my point, we do have evidence. Like I said, NSF video caught something may be this, and NASA also mentioned it in their HLS update. There's an old thread about this thruster, I think some of the evidence were posted there.

The last test I heard of was about 3 years ago.  Has there been something since then?

I'm almost certain they have something in the works.  But if it were a key element for Option A/B, we'd have seen a lot more of it by now.

It's true that we haven't heard about it for a while, but if they're replacing it with SD we'd expect more SD test at McGregor, no? I don't think we have that either.

I could see Elon trying to get rid of landing thruster all together, but convincing NASA to go along with it is a big ask.

[TheRadicalModerate]

It's true that we haven't heard about it for a while, but if they're replacing it with SD we'd expect more SD test at McGregor, no? I don't think we have that either.

I could see Elon trying to get rid of landing thruster all together, but convincing NASA to go along with it is a big ask.

SD are continuously qualified for D2.  I wouldn't think qualifying them for HLS would require the kind of test campaign we'd see for a completely new thruster.  I'm sure it'll require some qualification, but it would be something fairly modest, which could happen later in the program.

One argument against this thesis:  the artwork we see shows pods of three thrusters, not two.  That would require some more extensive requalification, even for SDs.  But it could just be an artwork thing.

[OTV Booster]

It's true that we haven't heard about it for a while, but if they're replacing it with SD we'd expect more SD test at McGregor, no? I don't think we have that either.

I could see Elon trying to get rid of landing thruster all together, but convincing NASA to go along with it is a big ask.

SD are continuously qualified for D2.  I wouldn't think qualifying them for HLS would require the kind of test campaign we'd see for a completely new thruster.  I'm sure it'll require some qualification, but it would be something fairly modest, which could happen later in the program.

One argument against this thesis:  the artwork we see shows pods of three thrusters, not two.  That would require some more extensive requalification, even for SDs.  But it could just be an artwork thing.

Here's the Musk design process as I conceive it.


There is an apparent need for landing engines. Musk does some BOE calcs and comes up with a size and engine count that makes preliminary sense. Musk says to pencil in the engines, at that size and count, pending further refinement.


As design conflicts arise, things like docking cone placement (which is also penciled in), both sides of the conflict get more attention and become less notional. The render artists may not be tightly bound into this cycle because EVERYTHING is notional to some extent up to the point where NASA or Musk insists the design be frozen.