The Landing Legs & Gears Is Killing Me!

[linuxgurugamer]

1 hour ago, I_Killed_Jeb said:

That's pure conjecture on your part, you realize that right?

Based on experience.

[Cubivore]

We've already established that mods don't have any effect, like do I need to make a video of my bone-stock spaceplane sitting at a standstill on Minmus, bouncing around so badly that taking back off is virtually impossible without quickload abuse? I don't understand why people are being incredulous about this being an issue in the game itself, entirely separate from mods, when Squad themselves have both tried to address it in the past and are attempting to address it again in the next update. I just want the damn game to work correctly, and I sure as hell don't appreciate people implying that this problem is somehow even remotely my fault when I'm playing the game the same damn way I've played it since twenty freakin' eleven.

[FinalFan]

2 minutes ago, Cubivore said:

We've already established that mods don't have any effect, like do I need to make a video of my bone-stock spaceplane sitting at a standstill on Minmus, bouncing around so badly that taking back off is virtually impossible without quickload abuse? I don't understand why people are being incredulous about this being an issue in the game itself, entirely separate from mods, when Squad themselves have both tried to address it in the past and are attempting to address it again in the next update. I just want the damn game to work correctly, and I sure as hell don't appreciate people implying that this problem is somehow even remotely my fault when I'm playing the game the same damn way I've played it since twenty freakin' eleven.

And if someone else built the same spaceplane and sent it to the same spot on Minmus would the same thing happen?  Inquiring minds want to know! 

[Rocket In My Pocket]

14 hours ago, Cubivore said:

We've already established that mods don't have any effect, like do I need to make a video of my bone-stock spaceplane sitting at a standstill on Minmus, bouncing around so badly that taking back off is virtually impossible without quickload abuse? I don't understand why people are being incredulous about this being an issue in the game itself, entirely separate from mods, when Squad themselves have both tried to address it in the past and are attempting to address it again in the next update. I just want the damn game to work correctly, and I sure as hell don't appreciate people implying that this problem is somehow even remotely my fault when I'm playing the game the same damn way I've played it since twenty freakin' eleven.

Bugs can be affected by all sorts of other things, from drivers, to hardware, all the way to playstyle. They don't always affect everyone universally.

No one is being incredulous about the existence of the bug, we're just reporting that some of us haven't experienced it; which is valuable information as it points to some other contributing factors possibly being the cause of the bug. My attempts to replicate the bug on my end weren't to "disprove" anyone, they were so I could hopefully actually see said bug "in the flesh" so I could understand it myself.

I'd encourage you to go back and re-read the thread, I don't think anyone was trying to be anything but helpful.

[Tyko]

16 hours ago, Cubivore said:

We've already established that mods don't have any effect, like do I need to make a video of my bone-stock spaceplane sitting at a standstill on Minmus, bouncing around so badly that taking back off is virtually impossible without quickload abuse? I don't understand why people are being incredulous about this being an issue in the game itself, entirely separate from mods, when Squad themselves have both tried to address it in the past and are attempting to address it again in the next update. I just want the damn game to work correctly, and I sure as hell don't appreciate people implying that this problem is somehow even remotely my fault when I'm playing the game the same damn way I've played it since twenty freakin' eleven.

@Rocket In My Pocket specifically asked for people to send him stock craft that are experiencing the issue. I suggest you take him up on it. Make a stock spaceplane and send him the craft file along with steps to reproduce. If he wants he can then try it himself.

[Tyko]

Has anyone tested 1.5 yet? 

[jkf1203]

Did they remove spring and damper settings on landing legs? I'm not able to access any and thought maybe they put in some type of advance setting to re-enable them but I can't find anything. The default setting is pretty bad. Bouncing all over Minmus like a rubber SuperBall.

[OHara]

7 hours ago, Tyko said:

tested 1.5 yet? 

Briefly.  Generally things seem better.  

Bouncing is damped on the light-aircraft wheels (the example craft for bug 19427 now behave well) and landing legs no longer jump on docking/undocking (19343), but fully-compressed struts still kick back violently on the Exploring Gilly scenario (no bug-report really describes that effect, so I'll add one soon).

<speculation> Possibly things were adjusted so the craft in bug-reports work better, but others might bounce worse.  </speculation>  Also, there is an aerodynamics bug with body lift in version 1.5.0, so keep your old installations intact for a while.

There were several distinct symptoms, several related bugs on the tracker now marked 'ready to test', and four relevant lines in the bug-fixes part of change-log.    
I am pleasantly surprised, particularly by Squad now taking the time to update the public bug-tracker.

@jkf1203, There is a new button 'spring/damper: auto' that you can de-select to see the former sliders.

Edited October 17, 2018 by OHara

[jkf1203]

16 hours ago, OHara said:

 

@jkf1203, There is a new button 'spring/damper: auto' that you can de-select to see the former sliders.

@OHara Hmm. Thanks for the info. I might need to do a full re-install as I'm not seeing that on any of my landing legs.

[RealKerbal3x]

10 hours ago, jkf1203 said:

@OHara Hmm. Thanks for the info. I might need to do a full re-install as I'm not seeing that on any of my landing legs.

It might be an advanced tweakable. Make sure that's turned on through the main game settings.

[Tyko]

I've landed my first small lander (<350kg) in 1.5.1.

I'm still seeing the never ending bouncing behavior when I set Spring/Damper to Auto. If I manually adjust the settings so it's not bouncing it still vibrates. The lander never ends up sitting still, it behaves almost as if it's sitting on a vibrating surface - it kinds of dances around and spins in a slow circle.

Are others still experiencing problems too?

Edited November 7, 2018 by Tyko

[OHara]

Landing gear still bounce.  I marked the relevant bugs as "not fixed" (because the given examples still bounce) but noted that now we have the ability to over-ride outside of the VAB which makes the bug much less annoying.

Even with the new {auto/override} button set to override, the spring strength is still automatically adjusted.  Put a full ore tank on legs, and notice that their compression returns to the same level after we dump all the ore, or hack gravity to a different value.  The slider we get to use seems to be an adjustment on top of lower-level automatic adjustments.
I remain convinced that automatic adjustment of the spring stiffness is driving the eternal smooth bouncing.  Other problems might be causing the quick jitter.

@Shadowmage also provides an automatic adjustment of ride-height in KSPWeel, and  I can see that there is a desire to have (working) automatic adjustment

On 10/5/2018 at 5:25 AM, Blasty McBlastblast said:

On Eve this time, again smooshed suspension (stock game, standard wheel settings)

These designs were tested thoroughly on Kerbin before launch and worked just fine, but act weirdly once delivered to a new planet.

but I don't think we want that in KSP, because KSP is a craft-building game.  My rockets that work fine on Kerbin don't work very well on Eve, and I don't think we want auto-compensation of thrust.

I suppose the force is F = k x + d v for a spring constant k [N/m] and compression x of the spring, and damping d [N/(m/s) for relative velocity v of the ends of the strut.  An automatic adjuster would want to set k = mg/L  for some desired compression L that sets the ride-height.  How to guess m×g when each leg supports some portion m of the mass of the craft and could be used with different values g of local gravity ?   KSP seems to take the average force supported by the leg to be m×g, but that force is higher than mg at the bottom of a bounce so the spring pushes a little to hard on the way up, driving an oscillation.

An automatic adjuster might want to set d = 2 sqrt( k m) to give critical damping, but now it is even harder to guess m.  KSP seems to take m from the average force on the spring, underestimating m in low-gravity situations and thus underestimating the ideal damping.

I think KSP players can handle landing gear with constant spring constants and dampers---say k = 400 kN/m and d = 200 kN/(m/s) for the LT-2. Having the craft sit higher on the gear when landed on Minmus would seem normal.  

Alternatively, a nonlinear spring-force (also an option in KSPWheel)  F = k x x / L  where L is the design compression, would give each part a wider range of reasonable usefulness, because there is (in well-designed legs) at least 2L travel to the struts, so on a high-gravity world with full tanks they could take 4× the design load would not bottom them out.

Edited December 16, 2018 by OHara

[Shadowmage]

On 12/15/2018 at 6:50 PM, OHara said:

@Shadowmage also provides an automatic adjustment of ride-height in KSPWeel, and  I can see that there is a desire to have (working) automatic adjustment

Indeed.

The problem with not having automatic adjustments is that the physics systems don't cope well with setups that are outside of a narrow range of inputs (either KSPWheel or Unity wheel-colliders).  Notably, they fail terribly when spring force = ~0 or less than needed to suspend the load (bottoming out), and also have simulation instabilities when spring force is 'higher than needed'.  Damper ratio has more issues as well at extremely high or low values.  These are not 'real' problems that would be experienced on a 'real' vehicle setup with those parameters, but are purely an artifact of the timestep based simulation.

The problem with having automatic adjustments is one of craft design -- you cannot possibly know in advance how a player will design a craft, or what its intended use is.  It might have 27 wheels on it, but only 5 will be in contact with the ground at any point; so how would one, purely analytically, know this craft layout and the correct spring/damper values to give to each wheel?  You really can't.

The problem with user specified values is that most people are really not engineers.  Give them some basic inputs like N/m  (the K in the spring equation) and they'll have no idea what to do with it.  I tried this route on KSPWheel for awhile, and close to nobody could figure out how to configure the wheels properly for their vehicles.  Lots of complaints of 'wheels aren't working right' -- when it was simply users not configuring them properly for their use.

 

Manually specified parameters work.  If specified properly.  Nobody (almost nodody...) wants to though.  So now we are where we are today -- with bodgey, hacky, bug-ridden implementations of spring/damper 'auto-tune'  (the KSPWheel implementation is just as physically incorrect as the stock one;  stock one is just using the Unity Wheel colliders that are buggier).

[Drethon]

 

On 9/28/2018 at 6:57 AM, Sharpy said:

@steve_vSquad has interest in fixing them properly; they've been trying since the "upgrade". Thing is this is Unity/PhysX problem and all Squad really can do is tweaking the parameters. It's still vastly better than it was in 2.1/2.2 but still, I use landing legs for landing, not for staying in place. The moment the craft comes to a stop, legs go up. Hacks of using cargo bays for landing legs are still fully valid.

 

 

 

They've probably tried this but sounds like a situation for having physics gear while landing, but once touched down transition to a non physics gear that doesn't flex.

[I_Killed_Jeb]

40 minutes ago, Drethon said:

 

They've probably tried this but sounds like a situation for having physics gear while landing, but once touched down transition to a non physics gear that doesn't flex.

This is something I've been thinking about for a while. For landing gears (I won't at all venture into the scary world of wheels), is it prohibitive to simply swap out the suspension-equipped legs for rigid, statically locked parts when you click on an option to "lock"? Like, it can even be a completely different part that just happens to look the same, that just gets substituted in instantly.

[FinalFan]

3 hours ago, I_Killed_Jeb said:

This is something I've been thinking about for a while. For landing gears (I won't at all venture into the scary world of wheels), is it prohibitive to simply swap out the suspension-equipped legs for rigid, statically locked parts when you click on an option to "lock"? Like, it can even be a completely different part that just happens to look the same, that just gets substituted in instantly.

I suspect that part substitution would end up being one of those ideas that sounds simple but is actually very problematic.  Locking, though, I have hope could be done.  (Toggle the travel settings instead of changing the part.)

[soulsource]

Hmmm, in 1.5 it seems that there's some crazy "lock orientation once movement gets small enough" code in place. I just had a lander on Minmus, that had SAS turned off, and trim at zero, and still it ended up tilted (with 2 out of 4 legs above the ground). Small torques (reaction wheels) could only make it oscillate around that orientation. Once a stronger torque (RCS) was applied, it shifted, but just to find a new "equilibrium" orientation, that wasn't realistic either...
Either that, or the collider of the Poodle is now a meter longer than the visual model...

[Bej Kerman]

Ugh, I just don't like how there are only 3 types of landing legs that we've already had for years and only had slight alterations. I landed a booster with 30m/s of delta-V to spare but, of all times, THIS is when the landing legs would decide that they'd bounce and destroy the booster. I like KSP but sometimes it's evident that the programming and engine suck more than a perfect vacuum. It frustrates me that Squad still has no evidence that they have ever tried/are trying to fix the problem, nor have ever talked about implementing larger F9/FH style rigid booster legs. Ik this is a slightly old forum but I really needed to vent somewhere appropriate. 

[Cubivore]

Landing legs are still absolutely terrible. Landing gear is still absolutely terrible. I really don't understand, this used to not be a problem at all. It's totally killing my will to play. This should be #1 on the priority list for 1.7.

[Chel]

Are the landing gear really that bad? I can land my craft quite easily, maybe it's your landing approach and actual craft that may be the problem

[Cubivore]

20 minutes ago, The_Cat_In_Space said:

Are the landing gear really that bad? I can land my craft quite easily, maybe it's your landing approach and actual craft that may be the problem

https://www.twitch.tv/t0zzle/clip/AgitatedFitAntBudBlast

Nothing about how you build your craft or how you approach your landings should ever make this a possibility. Ever. Period.

Fix the landing legs.

And being accusatory in suggesting that the problem is actually me is both unhelpful and flat-out insulting, but I guess I shouldn't be surprised since people have done that repeatedly in this thread.

Edited March 25, 2019 by Cubivore

[GRS]

It's just getting more realistic, this also means that you need to be careful in Landings, and get rid of Part Clipping in your routines, especially in weak parts.

[steve_v]

On 3/26/2019 at 11:48 AM, The_Cat_In_Space said:

Are the landing gear really that bad? I can land my craft quite easily

It varies with craft geometry, gear placement, tweakable settings and craft mass/body gravity vs. gear count. There are probably more.
No matter which way you dice it though, nothing should ever lead to endless bouncing.
Perpetual-motion is a physics simulation fail, there's no other explanation.

 

3 hours ago, GRS said:

It's just getting more realistic

Are you serious? That was sarcasm right?

Realistic wheels do not bounce endlessly without energy input or launch craft violently into the air at the slightest provocation.
The perpetual motion machines are the result of the janky new physics code, not added realism, wheels have been broken since 1.1.

Are we even playing the same game?

 

3 hours ago, GRS said:

get rid of Part Clipping

Part clipping has not been a problem for years, same-craft collisions were disabled.

3 hours ago, GRS said:

especially in weak parts

While it has nothing to do with clipping, craft flexibility does aggravate the suspension jank, hence the recent autostrut band-aid. Autostruts were created solely as an attempt to fix the horribly broken suspension system.

 

On 3/26/2019 at 12:02 PM, Cubivore said:

Nothing about how you build your craft or how you approach your landings should ever make this a possibility. Ever. Period.

Indeed. I would think anyone with a reasonable grasp of common-sense physics would see that from the video you linked.

@GRS registered here after Squad borked wheels/landing gear, so I can only surmise that he thinks this janky behaviour is intentional.

Edited March 27, 2019 by steve_v

[GRS]

I came with 0.23 first, then came in 1.0.4 and 1.0.5 before 1.5, difference ??? Landing Legs broke more easily, and more bouncy, not sure how are those landing gears, i'm always careful in Landing and consciously avoid Part Clipping, they Krakened my Trajectory (as what happenned in my Jool 3 mission), and i don't like Complaining...

For Gilly, putting Landing Legs don't seem to be a good idea, you can drain some RCS to stand or something, or use your design to float a bit.

Edited March 27, 2019 by GRS

[steve_v]

16 hours ago, GRS said:

I came with 0.23 first, then came in 1.0.4 and 1.0.5 before 1.5, difference ???

I can't make heads or tails of that sentence, or the rest of your post for that matter.