The fascinating problem of the starship landing legs

[Corvus Corax]

Hey everyone. New here. Been enjoying the YouTube livestreams since the beginning. So fun.

Ever since the prototype landing legs had their go with sn5 and sn6, I've been absolutely fascinated by the problem space of how to optimize the landing system:
Maximizing footprint...
Minimizing mass...
Maximizing robustness...
How to compromise optimally between these three factors?

As many of us probably do, I have opinions...

I don't know exactly how the attachment system works, so I guess I'm about to find out, but I threw 6 renders into this post that demonstrate 2 iterations of what I think would be the best options.

Both work under the assumption that the aft flap wings should contain two of the legs, and that the flaps should be reopened just prior to touchdown (when aerodynamic forces have been reduced to negligible levels, the engines have fired up, and the flare has been completed). My reasoning for this: The flaps have a wide footprint, and are very strong compared to the aft skirt. Both use a windward and leeward leg that extend just prior to touchdown in a manner similar to F9's legs do. The windward leg is heatshielded by a shroud that protects its entire length. The flap legs retract into the flaps so that they are protected by the already-extant heatshields. 

The engines are rotated radially by 30 degrees to make space in the skirt; a strong flanged mounting system for the leeward and windward leg deployment pistons would then have room to be installed.

I'd love to talk shop about it with anyone who's interested 

[_MECO]

Hey everyone. New here. Been enjoying the YouTube livestreams since the beginning. So fun.

Ever since the prototype landing legs had their go with sn5 and sn6, I've been absolutely fascinated by the problem space of how to optimize the landing system:
Maximizing footprint...
Minimizing mass...
Maximizing robustness...
How to compromise optimally between these three factors?

As many of us probably do, I have opinions...

I don't know exactly how the attachment system works, so I guess I'm about to find out, but I threw 6 renders into this post that demonstrate 2 iterations of what I think would be the best options.

Both work under the assumption that the aft flap wings should contain two of the legs, and that the flaps should be reopened just prior to touchdown (when aerodynamic forces have been reduced to negligible levels, the engines have fired up, and the flare has been completed). My reasoning for this: The flaps have a wide footprint, and are very strong compared to the aft skirt. Both use a windward and leeward leg that extend just prior to touchdown in a manner similar to F9's legs do. The windward leg is heatshielded by a shroud that protects its entire length. The flap legs retract into the flaps so that they are protected by the already-extant heatshields. 

The engines are rotated radially by 30 degrees to make space in the skirt; a strong flanged mounting system for the leeward and windward leg deployment pistons would then have room to be installed.

I'd love to talk shop about it with anyone who's interested 

I like this, but having the flap legs pivot down from the inside of the flap rather than the outside would give a much larger footprint overall. And if it were up to me I would just put two oversized Falcon 9 booster legs on the leeward side.

You wouldn't need any windward legs if you just fold the flap legs forward, giving you a respectable four-point footprint.

[Corvus Corax]

I like this, but having the flap legs pivot down from the inside of the flap rather than the outside would give a much larger footprint overall. And if it were up to me I would just put two oversized Falcon 9 booster legs on the leeward side.

You wouldn't need any windward legs if you just fold the flap legs forward, giving you a respectable four-point footprint.

I tried folding the wings and using them as windward feet, but the footprint is just WAY too small in the Y-axis; the wings are too short to extend beyond the skirt...they're closer to the axis of the vacuum bell. The ship would fall over backwards :/

Otherwise, is this what you're talking about:

[Rocket Science]

Your renderings are great, thank you! A concern I expressed a couple of years back is putting the landing loads though the control surfaces however... Keep them coming!

[brlavalamp]

Another problem with legs in the flaps is that the flaps are not designed to take  very much load from the edged or bottom, so I am not sure that you save much structural mass by adding legs there.  The skirt is plenty strong and already extends closer to the ground than the flaps, so it is better to mount the legs there.

If it were up to me, I would keep the current vertical leg design, but beef up the nose lateral thrusters to help maintainable verticality during leg deploy and leveling, which would reduce the need for a large leg footprint.

[50_Caliber]

From seeing Musk's tweet's about this, it sounds like the leg design is very much up in the air. I wonder if they might alter the base with a slight bulge so it can house a larger leg. We've seen a modified base on Starship with the tanker design, there's also versions that have the extruded housing on the side. They might even choose to go with the super-stubby legs and then have a different version with much larger legs. Why carry around the extra weight of larger landing legs if you don't need them? The super-stubbies could be used on prepared landing pads with only the lunar and mars SS with the big deluxe legs.

[adrianwyard]

Very nice renders.

We know that using the body flaps also as legs was the initial and preferred design, but they've apparently set it aside. So we can guess it's challenging, presumably for the reasons that 'Rocket Science' mentions, i.e. designing an actuated hinge and flap that can handle the roll-direction loads as well as support the weight of the vehicle/landing in the up/down direction is complex and/or heavy. So any new design will need to bring something new to the table.

Some ideas at random:

* The hinge doesn't have to support the vehicle in all positions, so perhaps you could save weight by having additional bracing lock into place only when they're moved into landing position.

* The flaps don't need to be 'wing-thin' - a thick flap with the same areas is just as effective as a thin one. So you can definitely fit leg mechanisms in there. (Thickness costs a little drag on the way up.) Crazy idea: they could be made sufficiently thick to allow a Falcon 9 style leg located in the root to extend *within* the flap once it's at the right position. (Note: In this case the leg's loads would not go through the hinge at all.)

* A less crazy idea is to just bring back the legs from the 2016 ITS design (below) and put the body flaps outboard of that. You lose a little effectiveness as they are closer to the center of mass, but on the other hand we know they are concerned about expanded plume heating, so perhaps it's an OK tradeoff.

* The body flap roots might not need to be at 180 degrees, so you could move them windward if that helps the leg span. (Their effectiveness decreases as the total wingspan/area lowers, so there is a trade-off.)

[Corvus Corax]

Some ideas at random:

* The hinge doesn't have to support the vehicle in all positions, so perhaps you could save weight by having additional bracing lock into place only when they're moved into landing position.

* The flaps don't need to be 'wing-thin' - a thick flap with the same areas is just as effective as a thin one. So you can definitely fit leg mechanisms in there. (Thickness costs a little drag on the way up.) Crazy idea: they could be made sufficiently thick to allow a Falcon 9 style leg located in the root to extend *within* the flap once it's at the right position. (Note: In this case the leg's loads would not go through the hinge at all.)

* A less crazy idea is to just bring back the legs from the 2016 ITS design (below) and put the body flaps outboard of that. You lose a little effectiveness as they are closer to the center of mass, but on the other hand we know they are concerned about expanded plume heating, so perhaps it's an OK tradeoff.

* The body flap roots might not need to be at 180 degrees, so you could move them windward if that helps the leg span. (Their effectiveness decreases as the total wingspan/area lowers, so there is a trade-off.)

Thanks for the kind words. The renders are kinda crude compared to some of the starship graphics floating around out there, but I like working on the technical aspects way more

I always suspected that they ditched the tripod fin (tintin) design because that extra fin is just dead weight (and a lot of it, too)

I think my favorite suggestions from your list is thicker wings and locking braces... I'm not so concerned about vertical loads on the wing and its bearings, but rather the folding forces that might try to bend the legs out from under the wing. Imagine the leg is an aileron tabbed onto the back of the wing: any torque on landing would try to deflect that "aileron"...that is the weakness of this design. 
Note that F9 legs are forked at the hinge to provide much needed lateral stability.
I think that a more wedgie wing would allow the root of the leg to fork in a similar way to provide that much needed lateral stability.

Adding wedge / extra thickness to the wings would add too much dead weight, so I think a good compromise is to just add blisters at the point of triangulation and to make that triangulated structure into a locking hinge like this...

[Lars-J]

Legs on the fin are not a good idea, IMO. For several reasons:
 - Adds unnecessary weight and complexity to fins
 - Design does not work for lunar Starship.
 - More than one leg design to qualify AND make adjustable to handle non-flat landing spots.
 - If fins are malfunctioning, so are your legs, this is not good

And IF you used fin legs, the leg feet should be at the very edge, to give the most benefit. (like the Tintin design)

[Nomadd]

 The legs will use the same structure as the holdown points to the booster.

[Corvus Corax]

The legs will use the same structure as the holdown points to the booster.

Wait is this verified? That would narrow the speculative domain quite a bit and give us a much better idea of what direction they're heading in.

[Nomadd]

The legs will use the same structure as the holdown points to the booster.

Wait is this verified? That would narrow the speculative domain quite a bit and give us a much better idea of what direction they're heading in.

Yep.

[wes_wilson]

I'd like to throw out an idea I don't believe I've seen dissected yet.  I'm not a good artist but I've attached a sketch trying to show what it would look like. 

Using the hard points and staying within the base of the starship cylinder, what if instead of extending cylinders as legs they extended arches as legs.    They could have 2 half circle arches that together form a circle of 30ft diameter which would follow the base of the skirt. These two arches would be hinged to the starship at their apex using the hardpoints existing for super heavy attachment.  When released they would fold straight down with the end points of the arches ending up outside the skirt circumference.  The end points of the arches act as four feet outside the skirt diameter giving a base with a diameter of 42.4ft.  The Starship sits on the apex of the two arches and their four feet. 

Upsides are that it gives a wider base than the diameter of the ship while keeping all the leg structure out of the way of engines and control surfaces.  The arches provide some energy absorption and take the forces of landing back to the mount points used for super heavy.  Gravity does a lot of the work of deploying them.  The apex would be 15ft high placing the skirt and engines 15ft above ground which is healthy clearance.

Lastly, while you get 4 "feet" with two arches.  You could also do two sets of arches for four arches total; offsetting the two pairs of arches by 90 degrees.  The feet would all fold out to the same places giving the same 4 "feet" areas but each "feet" area would now have the ends of two arches.  So four arches would mean you could lose any 1 arch without loss of any feet; or any 2 opposite arches without loss of any feet.  Losing two adjacent arches would cost a foot.

Weight?  No idea, probably a lot?


 

[wes_wilson]

I'll reply to my own post    My diagram assumes the arches extend down vertically and basically just lock in place.  You can get an even wider stance with the feet in exchange for a lower apex if the arches are extended past the vertical position.  Basically, within a certain range, the spread of the feet becomes adjustable if needed.

 

[Twark_Main]

Hey everyone. New here. Been enjoying the YouTube livestreams since the beginning. So fun.

Welcome! It's a heck of a party.

I know you meant the beginning of the YouTube channel, but YSK that NASASpaceFlight has been around since 2004, while the YouTube channel only started last year (March 2019).

[KBK]

The leg attached to the body but the forcing rods or forcing hydraulics/gearing within the skirt.

Ie, pressurizing/geared/motorized rods in the skirt, between the engines..under the skirt protection covers (engine debris reflectors).

this way, you get your sleek sides, and your balanced load, tucked in tight to the body...

But your forcing rods are motorized via heavy gear ratios, located under the skirt. The strengthening or leg re-enforcing system to handle the vertical load is folded into the leg proper. And attached to the stable and solid skirt and possibly the thrust puck assembly, via a stabilizing ring, to distribute the unwanted extra angular/twisting/offset forces, as the legs touch and settle.

So, one single leg design, but iterated four/six times, with offset/unequal force compensation via KISS mechanicals.

That area, the skirt puck area... is highly reinforced so a bit of extra build quality/rigidity in that area is there to stabilize the landing gear's force distribution, and be part of the landing gear actuation and retraction system, and be part of the debris cover system.

Every mm of leg extension helps, so it may be a better way, all while keeping mass down, keeping simplicity up, keeping the system simple and of a multi-problem solving nature. That multiple things are taken care of at the same time.

Imagine an electrically geared bar, with shock force compensation built in (shock absorbers). This bar, between the engines just a bit. this acts as a gear advantaged offset thruster for the leg which is on the side of the skirt. With maximum gear leverage capacity. and with maximum leverage advantage when the load on the leg is being settled, where the built in shock absorber aspect comes into play. Ie, the longest travel range of the system, almost horizontal in position while the leg itself, moves very little, in the up and down direction.

It can even swing into place and lock, from underneath/inside the skirt (the rod coming through the sidewall, but the hole is covered when the leg is folded), in the landing sequence...as the gimballing of the engine is in the past tense. This..If the system is required to be shorter, to clear the engine assemblies.

Also, the maximum gear advantage can come into play with some possibilities to handle some misalignment when the legs are to be retracted after a given take off. And, the maximum gear advantage can be used to make the rocket more vertical after landing, via said highly geared motors being brought into play. Settling during/after landing can be a problem -this can compensate. ie, to facilitate in the offset load of when using the crane to get objects in and out of the cargo bay.

Simple, robust, takes advantage of structure that already exists and adds little mass in fixing the problem at hand.

The big problem... is that it has to, in some fashion, be near the bell assemblies or the tubopump/chamber assemblies.

Almost like a F9 leg but turned upside down. Aerodynamics, well ..placement would have to be even handed/mirrored on the hot side and since the heat creation would be in the skirt area, not on the tank system or tank/wall join areas , then it might be possible to compensate for that without too much trouble (comparatively speaking).

I'm writing this before my first coffee, so bear with me.

And, if you are really really good in your design implementation, then the lower skirt outer ring becomes a thrust puck vibration control point (via the leg shock absorber system), with additional torsional stress relief for the thust puck, and an overall thrust puck re-reinforcement point when under launch loads..where you are gaining maximum mechanical advantage. ie trying to compress a metal circle (think archways), and the circle is located at the best mechanical advantage location..the outer edge of the bottom of the rocket proper.

Since the rocket mass is much lower on landing, relatively speaking, the overhead is there for dealing with the potentials in offset stressing during the landing proper.

This will all add mass, but one one is careful, not too much. On the good side of the ledger.. it will add margin, in safety and reliability, re the idea of reuse of a fragile package. All things to be looked at, as the package has multiple parts... where some can be added, or modified, and some can be thrown away.

[Corvus Corax]

I don't see how they are going to land on the moon with any of the solutions I've seen anywhere. The surface area of the feet is always minuscule and would create unacceptable point loads on the surface. The craft would sink down to the engine bells, wouldn't it?

The only thing that seems remotely feasible is the 2016 ITS legs, but those look insanely heavy to me. I suppose, since they seem to be designing a lunar-specific design, they could easily get a ship into orbit with much bigger feet like ITS, and then resupply/retank that for its trip...But it's a hell of a redesign.

The only other option is to prepare a hardened landing pad, but they'd have to get quite a bit of equipment up to the landing sight for that to work, so...how? 

[wannamoonbase]

I don't see how they are going to land on the moon with any of the solutions I've seen anywhere. The surface area of the feet is always minuscule and would create unacceptable point loads on the surface. The craft would sink down to the engine bells, wouldn't it?

The only thing that seems remotely feasible is the 2016 ITS legs, but those look insanely heavy to me. I suppose, since they seem to be designing a lunar-specific design, they could easily get a ship into orbit with much bigger feet like ITS, and then resupply/retank that for its trip...But it's a hell of a redesign.

The only other option is to prepare a hardened landing pad, but they'd have to get quite a bit of equipment up to the landing sight for that to work, so...how? 

One of the lessons of Apollo was that the lunar surface is much stronger and denser then many thought.  It's had millions of years to get compacted.

Even if the landed mass was 360 tons (hot swag of vehicle, payload and some fuel) that would be 60 tons per leg and 10 tons in 1/6th moon gravity.  That is very very manageable.   

The actual touchdown on the moon is maybe one of the easier problems, outside of being level.  Way less force than an Earth landing.

[Corvus Corax]

It's neat to imagine the legs adjusting their length on approach as it surveys the terrain. Mechanically, too. Reminds me of the International Docking Standard mechanism.

[CorvusCorax]

Hey everyone. New here. Been enjoying the YouTube livestreams since the beginning. So fun.

Ever since the prototype landing legs had their go with sn5 and sn6, I've been absolutely fascinated by the problem space of how to optimize the landing system:
Maximizing footprint...
Minimizing mass...
Maximizing robustness...
How to compromise optimally between these three factors?

As many of us probably do, I have opinions...

I don't know exactly how the attachment system works, so I guess I'm about to find out, but I threw 6 renders into this post that demonstrate 2 iterations of what I think would be the best options.

Both work under the assumption that the aft flap wings should contain two of the legs, and that the flaps should be reopened just prior to touchdown (when aerodynamic forces have been reduced to negligible levels, the engines have fired up, and the flare has been completed). My reasoning for this: The flaps have a wide footprint, and are very strong compared to the aft skirt. Both use a windward and leeward leg that extend just prior to touchdown in a manner similar to F9's legs do. The windward leg is heatshielded by a shroud that protects its entire length. The flap legs retract into the flaps so that they are protected by the already-extant heatshields. 

The engines are rotated radially by 30 degrees to make space in the skirt; a strong flanged mounting system for the leeward and windward leg deployment pistons would then have room to be installed.

I'd love to talk shop about it with anyone who's interested 

Neat, a fellow raven   welcome to NasaSpaceFlight

Something similar was posted on twitter a few days ago, and commented on by Elon, I think your variant has more merit though

https://twitter.com/ErcXspace/status/1334237562823725056

    Here is my weird Starship Leg Design, 2 deployable inside the skirt and 2 inside the flaps. All self levelling. (Really rushed the modelling part on this one, please excuse me😅) pic.twitter.com/Tkuip3neFw
    — Erc X (@ErcXspace) December 2, 2020

https://twitter.com/elonmusk/status/1334321901783511041

    Not bad
    — Elon Musk (@elonmusk) December 3, 2020