Landing leg appreciation thread

[FinalFan]

Like many of you, I've experienced serious issues with landing legs having excessive bouncing.  But it's important to remember the promise of landing legs.  I landed a probe on Minmus lately on a slope and this happened: 

I came down vertically and the legs automatically evened me out.  It was beautiful.  Hopefully this is what we can have right from the beginning in KSP2, and new players can experience this without bitter memories of landing leg shenanigans. 

[The Aziz]

"excessive" is an understatement

 

Edited October 14, 2019 by The Aziz

[AHHans]

3 hours ago, The Aziz said:

"excessive" is an understatement

I prefer "unphysical". But then, I studied physics, so it might not sound as damning to others as it does to me.

[klgraham1013]

I'm not sure if Star Theory is tagging their posts with #BlameUnity yet.

[Superfluous J]

1 minute ago, klgraham1013 said:

I'm not sure if Star Theory is tagging their posts with #BlameUnity yet.

But it's good to know we already are.

[AVaughan]

Personally I consider that a classic an example of what happens if any of the legs contract past their maximum allowable travel, and the physics engine then decides to fix that by moving the leg back to its maximum allowed travel, and in doing so applies a large force to the rest of the rocket. 

Far too often KSP/Unity doesn't seems to get the stiffness of the springs and the damping of the shock absorbers right.  Look at the video frame by frame.  The landing leg seems to offer almost no resistance to compression until it is almost at the very end of its allowed travel.  That is a massive problem since letting the landing leg exceed its maximum allowed travel will result in the physics engine over-correcting and potentially destroying the lander. 

Let's look at it frame by frame.

Touch-down appears to be at about 35.27 seconds with the lander descending at 0.5 m/s.

At 36.02 seconds the lander is still descending at 0.5 m/s.

At 36.03 seconds the lander is now descending at 0.4 m/s.

At 37.52 seconds the landing leg is almost completely compressed, yet the lander is still descending at 0.4 m/s.  (One potential work-around for the player is to use the engine to kill the vertical descent immediately after touchdown, so hopefully the landing leg never gets this compressed.  Tapping F to temporarily disable SAS, so that the craft can rotate, then turn it back on when the craft is parallel to the surface, might also help as then all 4 landing legs should be in contact, and hopefully all 4 can slow the descent before one bottoms out).

At 37.53 seconds the lander is now descending at 0.3 m/s.

At 38.05 seconds the lander is still descending at 0.3 m/s.

At 38.07 seconds the lander is now descending at 0.2 m/s.

At 38.13 seconds still descending at 0.2 m/s, but the landing leg looks like it is about to bottom out.

One frame later (38.15) the lander actually rotates downslope, as though the leg has bottomed out.  Descent rate now 0.1 m/s.

Next frame (38.17) nothing appears to happen.  (I'm guessing a video frame without a physics tick).

38.18 further rotation downslope, descent rate now 0 m/s.  SAS reports that is is unable to control the rotation.  Landing leg is starting to extend.

38.20 looks like another frame without a physics tick.

38.22 massive rotation, and the lander is now ascending at 9.8 m/s.  (Note that to achieve that change in velocity over 2 frames (0.04 seconds), requires acceleration of 25 gees).  Lander leg is now almost fully extended.  So roughly 3 seconds to compress on the way down, and 0.04 seconds to extend on the way up.  Massive overreaction from the physic engine, presumably because the landing leg exceeded it's allowable travel. 

Personally I think KSP needs to set landing legs stiffer by default to prevent them bottoming out, and then set the extension damper high enough that the legs don't extend fast enough to cause the craft to bounce when players do land a little hard.  (Isn't that the point of damping in a suspension system?) 

Also for landing on really low gravity moons, the player really need to control their descent rate better.  Contact at 0.5m/s and then using a tiny blip of the engine to kill the descent rate (and possible tapping F to toggle off SAS for a few seconds to allow the craft to rotate to match the slope) might have resulted in a successful landing.  (Otherwise just ignore the landing legs and land on the engine bells).

[KerbMav]

So, wasnt there a mod that made the stock landing legs behave?

[manek22]

That's why I always change the Spring/Damper settings on the legs (advanced tweakables?).  Usually set the spring to 1 or 1.1, and the damper to 1.3 or 1.4
But be careful, they like to change during mission, so before the decent, double check them.

[magnemoe]

Have switched to static grinders on most landers. 
One fun thing with landing legs is that if you set dampening to 0 and spring to max the lander will jump higher and higher on each bounce. 
Yes they are generating free energy, its makes as much sense as the other landing leg mechanics in KSP. 

The above lander on slope reminds me is discharging all the force at once once fully cocked, however I suspect it give of more energy than is put into it. 

More fun in that I also had bugs with hinges. 

This is an standard bug here. has happened on multiple of this tugs even with a bit variations in design 

[klesh]

Would be great if Star Theory would be on here explaining to us how they've fixed this and it won't be a problem in KSP2.

[RealKerbal3x]

1 minute ago, klesh said:

Would be great if Star Theory would be on here explaining to us how they've fixed this and it won't be a problem in KSP2.

Maybe instead of using the stock Unity wheel/landing leg physics system (like KSP1 does) they’ve created their own, more stable system?

[klesh]

4 minutes ago, RealKerbal3x said:

Maybe instead of using the stock Unity wheel/landing leg physics system (like KSP1 does) they’ve created their own, more stable system?

Would be nice to know, wouldn't it?

[RealKerbal3x]

2 minutes ago, klesh said:

Would be nice to know, wouldn't it?

Indeed.