Homegrown Legs

[sevenperforce]

Ever get tired of stock's shoddy landing legs that don't work as advertised?

Solution: build your own.

Easier said than done. It is easy enough to build a set of legs, but not ones that start folded and then deploy at landing.

But I managed!

Here it is on the pad.

Planning on using an expendable upper stage and sending a relaysat to KTO, but bringing the first stage back with RTLS, SpaceX-style.

Detail view of the landing legs on the pad:

The legs start folded, held against the side by paired docking ports. I could have done it with decouplers, but I wanted them to be refoldable, at least in theory. The leg itself has a crossbar which is suspended between paired structures coming out of the rocket body to form dual hinges.

On the pad, the rocket is held up by double-docking-port clamps, which are themselves held to the pad with launch clamps which will not be used. When the rocket launches, the lower docking ports release by action group. These same docking ports become the locking mechanism for the landing legs later on.

I went with five engines because I wanted to do it pure-stock; otherwise I would have tweakscaled down the Vectors and done it with nine engines, just for kicks.

Ignition:

And clamp release!

Very thrusty off the pad, but that's what five vectors will do.

Throttling down around Max-Q, mostly to help loft my trajectory a bit. The higher and slower I stage, the better shot I have at getting the boostback burn just right.

I turned off fuel flow for the bottom tank, so I don't have to worry about how much fuel I am reserving.

Down to fumes, so I cut the outer four engines and throttle the single engine back up in preparation for staging:

Staging!

Switch view to orbit to point more downrange, then back to the booster for the boostback burn.

This is a rapid sequence of events. Set SAS to Orbit-Radial-Out to begin the RCS flip, activated the airbrakes, mashed X to cut throttle, and turned the lower tank fuel flow back on.

Single engine back on and burning hard. Switched the navball back to surface so I can see when my surface prograde marker noses back over toward the west:

Money shot of the boostback burn, with detail of the lower landing leg locking clamps:

The prograde marker is now over on the west side of the navball, so this should be enough for a boostback. Cutting the engines and heading back to the second stage.

The second stage has just cleared the atmosphere, so I'm blowing the fairing and nosing down.

Burning halfway between Radial-In and Prograde in order to push to orbit as quickly as possible. Throttling down and following the prograde marker would be more efficient, 'tis true, but it's time-consuming and I don't have much time.

You can see the first stage is still climbing:

Almost to orbit...

Made it!

Throttled down and switching back to my first stage.

Quickly set surface orientation and retrograde. The airbrakes form a nice little landing crosshair. I should come down just off the coast.

I won't need an entry burn.

Here come the gees!

Punching the engine to full throttle just after the accelerometer shows terminal velocity. 

In pad testing, I used separatrons to force the legs down to snap into place. I thought I might have to do the same, but it turns out that the upward thrust of the landing engine results in enough differential force to pop the legs out, once they are released. I set the docking port release to the Gear action group, just to make it more authentic.

Just about to throttle down...

The legs cannot be released at a high airspeed, or the airflow holds them up no matter how hard I thrust. Which works well, because I need to drop the gear pretty close to the ground anyway.

Gear dropped!

At this point, all four separate landing legs are viewed by the game engine as debris. So they are falling separately from the stage.

However, the hinge keeps them stuck to the stage, and the gees on the stage itself from the engine lifts it up, causing them to rotate down:

Oddly, there's always one leg that rotates slower than the others, but it's never the same one. I guess that makes sense, because they are all separate parts.

Punching Z again to make sure I get a good inertial lock on the legs.

As soon as the ports touch, it is viewed as a docking event and the game switches views and turns off my engine. So this was a lot of constant Z-mashing.

Then I quickly switched SAS to Radial-Out and throttled down to avoid air-launching.

I think I kept my TWR above 1:1 the entire way down.

A bit of a bounce, Bulgariasat-Style:

But I recovered!

There she is, standing tall and not at all wobbly.

Detail view of the leg locking assembly. 

And done!

[Kerbalstar]

Wow. Amazing.

[Geschosskopf]

Ingenious.

[qzgy]

Neat. Hinge improvements could probably be made, but it looks like it works fine.

[Alpha 360]

Brilliant. Just Brilliant. I wish I could recover my stages with the same success. Is there a craft file............? I need this in my life.

[sevenperforce]

15 hours ago, qzgy said:

Neat. Hinge improvements could probably be made, but it looks like it works fine.

I played around with different hinge designs, specifically to try and get it down to just four hinge assemblies, each holding the corner of a leg (like SpaceX's), but was unsuccessful. Hinges are not my strong suit. Got a better idea?

14 hours ago, Alpha 360 said:

Brilliant. Just Brilliant. I wish I could recover my stages with the same success. Is there a craft file............? I need this in my life.

To be fair, my boostback burns are nowhere near as precise as I would like them to be. I had to fly it a half-dozen times to get the boostback right. Landing was a good bit easier, though; I quicksaved once I had a good EDL trajectory, but I didn't end up needing to quickload. I've had a lot of practice with suicide burns.

Here's the craft file, though it'll only be up for a couple of months: https://files.fm/u/xepzwm6e

[qzgy]

1 minute ago, sevenperforce said:

Hinges are not my strong suit. Got a better idea?

I am no expert in good hinge designs, but what I can say is that a design that is pretty common is to hold an RCS ball within a cage of thermometers. A stronger and sometimes smoother hinge/bearing is an RCS ball within more RCS balls. Maybe that might work for you?

Perhaps it might be a good idea to ping @EpicSpaceTroll139 for more help/suggestions?

Oh, and a good mod for this kind of thing is Collide-o-scope. Has been very helpful for me in figuring out the dark arts of stock hinges.

[sevenperforce]

Just for fun, here's the rest of the mission.

Back to the upper stage. Our (convenient) loiter is complete, so let's kick on the engine for the transfer burn.

And SECO-2 in a good orbit!

Payload release.

The first time I flew this mission, I was an idiot and actually forgot to put a probe core on the payload. At least this time I remembered. Also remembered to get those solar arrays out before my probe core dies.

Now to deorbit the upper stage.

A little radial burn changes the argument of periapsis and lowers it to a good disposal orbit.

I totally guesstimated the amount of propellant I'd need for circularization, but I figure three donuts should be plenty. The Ant is efficient enough after all.

Setting up the circularization node for the relaysat to see when I need to start burning.

Throttleup to begin slow circularization.

Of course this is probably far higher acceleration than what real-life GEO comsats do.

Getting there...

And there we go!

Tiny inclination correction burn.

Setting thrust limiter way down to refine my orbit.

Pretty soon...

And that's a wrap!

The payload was just for kicks, of course; the primary mission was designing and deploying fancy landing legs that would work the way I wanted. With a crane and a custom leg-lifting rover, I believe I would be able to refold the legs and refly, though of course the design would need the upper decoupler replaced with a docking port in order to add another upper stage and payload.

But as these things go, it seemed to work pretty well! I will have to remember this design if I ever want to develop a really robust reusable launch system. I've sometimes thought about playing a hard mode career game, and having a reusable launch system with virtually unbreakable legs would enable me to launch payload after payload without ever having to build/recover first stages.

[EpicSpaceTroll139]

Nice work! The leg system looks rather robust. I look forward to seeing more of this kind of thing! 

I like that the legs deploy without assistance. I wasn't able to get all legs to lock reliably doing that with my "Failcan 1."

Thus I used regular landing legs as actuators. Perhaps if I changed my descent profile...

[Geonovast]

Ugh, not even noon and I'm out of likes.  Gonna be a rough day.

Consider this post a like, I guess!

[SuperHammy3]

That's a genius idea. I normally just slap a few girders on as permanently extended landing legs. Gets the job done, and can absorb impacts quite well.

[sevenperforce]

On 1/24/2018 at 11:49 AM, EpicSpaceTroll139 said:

Nice work! The leg system looks rather robust. I look forward to seeing more of this kind of thing! 

I like that the legs deploy without assistance. I wasn't able to get all legs to lock reliably doing that with my "Failcan 1."

Thus I used regular landing legs as actuators. Perhaps if I changed my descent profile...

Hmm, I like the stock landing legs as actuators. That didn't occur to me. Did you use wing parts for the landing legs structures? I believe the landing legs need an autostrutted girder core for maximum impact resistance.

Can I get a closeup of the hinge mechanism? Mine is very robust but also rather bulky.

If I rebuild the leg mechanism for a 3.5m booster, I may try putting a Spark on the tip of the legs along with an Oscar-B (perhaps as part of the hinge mechanism) and setting up the action groups such that I can automatically refold the legs for reflight. Then all I would need would be a crane to drop in a replacement upper stage+payload, lift it up, fold the legs, and drive it back over to the custom launchpad.

 

[EpicSpaceTroll139]

Here's a close-up.

For each leg there are two "cages" composed of thermometers which are attached to the core, inside which are contained 2 linear RCS port. Each one effectively acts as a ball-in-socket joint, so the two together keep the leg on the proper axis. Note that the cages could probably be built using fewer thermometers than shown in the picture, I just haven't tested that.

Anyways, visually, the legs appear to be made from wings, but it is the small, retractable (but disabled in this regard), autostruts do wheel gear at the tip that give them their strength and impact tolerance. Thus this leg design does not require heavy I-beams.

Edited January 29, 2018 by EpicSpaceTroll139

[sevenperforce]

15 hours ago, EpicSpaceTroll139 said:

Here's a close-up.

For each leg there are two "cages" composed of thermometers which are attached to the core, inside which are contained 2 linear RCS port. Each one effectively acts as a ball-in-socket joint, so the two together keep the leg on the proper axis. Note that the cages could probably be built using fewer thermometers than shown in the picture, I just haven't tested that.

Anyways, visually, the wings appear to be made from wings, but it is the small, retractable (but disabled in this regard) wheel gear at the tip that give them their strength and impact tolerance. Thus this leg design does not require heavy I-beams.

Loving it!