Potential fix for 1.1 wheel collision problem! (didn't work)

[GoSlash27]

All,

 I suspect that the wheel collision problem (craft skittering/ flipping/ wheels exploding/ etc) is due to a transcription error in the .cfg files or coding.

 I swapped the suspension distance and position values in the .cfg file for the fixed gear and the problem disappeared.

I would like others to try this fix for various wheels and see if it also solves their problem.

 Be sure to keep accurate notes of the original suspension values so you don't get hosed up. Please report your results here (good or bad) so we can take it to Squad for a hot fix. Or... you know... not (depending on how it goes)

Thanks!

-Slashy

 

Edited April 23, 2016 by GoSlash27

[Jarin]

Been having fun with aircraft-on-ice, but would certainly not mind having traction again. Will test this when I get home.

[ghpstage]

Sounds good.

I'll give this a go on my collection of heavier fixed landing gear aircraft and get back to you.

Edited April 21, 2016 by ghpstage

[Guest]

  On 4/21/2016 at 10:40 PM, GoSlash27 said:

I swapped the suspension distance and position values in the .cfg file for the fixed gear and the problem disappeared.

Expand  

What's probably happening is that the increased suspension distance is reducing the downforce stress on the wheel, lessening the chance of explosion.  targetPosition probably doesn't have to change.  Try maybe just upping the suspension travel (suspensionDistance)?

E: Also, jittering and skittering is likely happening because the suspension has so little distance for travel...

Edited April 21, 2016 by regex

[GoSlash27]

  On 4/21/2016 at 11:06 PM, regex said:

What's probably happening is that the increased suspension distance is reducing the downforce stress on the wheel, lessening the chance of explosion.  targetPosition probably doesn't have to change.  Try maybe just upping the suspension travel (suspensionDistance)?

Expand  

I'm just getting started myself, but so far my results indicate that targetPosition cannot be greater than suspension distance. Still experimenting...

[Guest]

@GoSlash27, suspensionDistance is assigned to Distance and targetPosition gets assigned to Anchor in Vehicle Physics Pro.  Here's the apropos info:

  Quote

Distance (m)

Length of the suspension travel.

Anchor (%)

Relative position in the suspension travel (compression) where the wheel is located at design time. This setting can be understood easily by adjusting it in the Editor and watching the gizmo in the Scene view.

Expand  

Since Anchor is a percentage, set targetPosition appropriately (that is, between 0.0 and 1.0 inclusive).

Edited April 21, 2016 by regex

[GoSlash27]

  On 4/21/2016 at 11:16 PM, regex said:

@GoSlash27, suspensionDistance is assigned to Distance and targetPosition gets assigned to Anchor in Vehicle Physics Pro.  Here's the apropos info:

Since Anchor is a percentage, set targetPosition appropriately.

Expand  

Swapping these values didn't fix the steerable nose gear for me, but it worked for the fixed gear.

 This is good info. Do you have the source? I'd like to look at the other values

Best,
-Slashy

[Guest]

  On 4/21/2016 at 11:22 PM, GoSlash27 said:

This is good info. Do you have the source? I'd like to look at the other values

Expand  

Ask.  I have to correlate the two codebases (or dig into KSP's in general) since they use different names.  Here is the link where I found the info, though.

http://vehiclephysics.com/components/wheel-collider/

Edited April 21, 2016 by regex

[ghpstage]

Seems to have them a little more survivable, but I just experienced some ice skating for the first time and the trick doesn't work on the nose gear so i'm guessing the swapping the values dose that coincidentally?

Edited April 21, 2016 by ghpstage

[GoSlash27]

Cool, thanks!

[Guest]

  On 4/21/2016 at 11:28 PM, GoSlash27 said:

Cool, thanks!

Expand  

Actually it's going to take some time tonight, I'm still looking at pre-release code and a lot of that has changed.  However, if no one else answers your questions first I'll be updated in about four hours.

[GoSlash27]

*update*

The only gear I can reproduce the failure mode with is the fixed gear. All the rest seem to be fine.

 

[WinkAllKerb'']

remind me that's more or less where i get stuck long long ago with unity and landing gear. it used to not really allow any single variation in the hierarchy due to the amount of dynamic element within a single entity.

thks for sharing the infos and thks to every contributor and tester.

(Minor stuff to isolate doing some really weird results thing series it seem.)

 

Edited April 22, 2016 by WinkAllKerb''

[Mad Rocket Scientist]

Does anyone have a video of the wheel collision problem?  I can't tell whether I have it or not.

[cubinator]

  On 4/22/2016 at 1:14 AM, Mad Rocket Scientist said:

Does anyone have a video of the wheel collision problem?  I can't tell whether I have it or not.

Expand  

If your wheels do pretty much what you expect them to, you're not having the problem. If they're jumping all over the place, not getting any traction, spinning around everywhere, exploding when they shouldn't, or any combination of the above, you will know you're experiencing the wheel jitters.

[bewing]

Sorry slashy, but I don't think your theory is right.

@Renegrade and @sgt_flyer have an alternate theory that a bounce in the wheels gets transmitted as a force into the craft's joints. The joints flex (with no damping) and then return all that force to the wheels, which then bounce again.

I have a plane that jitterbugs severely -- see the panther.craft file on this bug report: http://bugs.kerbalspaceprogram.com/issues/9288

-- It supports their theory, because if the wheels are mounted on the wings, the plane jitterbugs. If the wheels are mounted to the fuselage instead, there is no jitterbugging.

I just made your suggested mod (.9 suspensionDistance, .1 targetPostion) to the LY1.

I removed all the fuel from panther (the ly1 can't handle the full weight), and replaced the rear LY10s with LY1s.

After reloading KSP and launching that plane, it jitterbugs for a bit, and then starts spinning in a medium-speed circle. 

 

[GoSlash27]

bewing,

 Thanks for testing it out. I'll accept "no" for an answer

Are you experiencing jitterbugging with anything except the fixed gear? I've seen bug reports for other gear, but can't reproduce them myself.

Best,
-Slashy

[bewing]

  On 4/22/2016 at 3:08 AM, GoSlash27 said:

 

Are you experiencing jitterbugging with anything except the fixed gear? I've seen bug reports for other gear, but can't reproduce them myself.

Expand  

Precisely what do you mean by "fixed gear"? The LY10s that jitterbug intensely for me are steerable (see that panther.craft). sgt_flyer faked up an MK3 design with the heavy landing gear out on the tips of the wings, and that jitterbugged, too.

 

[A_name]

I haven't experienced these problems, how do I reproduce it?

[sgt_flyer]

from my own tests, every landing gear has a suspension (even the LY-01 fixed gear !) - but the 'disable suspension' toggle seem only to have a cosmetic effect on most landing gears or landing legs (visually block the suspension in the current position - however the .craft still has the exact same physical 'rebound' like if the suspension was active regardless of the toggle) . the button seems to only have a physically noticeable effect on rover wheels suspension (and yup, suspension is important for a good grip !).

for the jittering / suspension danse, here's my MK3 test plane (if you fill up the fuel tank, it'll worsen the effect even more, up to linkage failure between the wings and the body) - the 4-wheel heavy landing gear is affected here. - you can see how much the wings flexes

https://www.dropbox.com/s/n7j4u8bvcjowaia/test%20jitter.craft?dl=0

interestingly, roverwheels don't trigger jittering in the same configuration (with XL3's in place of the heavy landing gear on the tip of the wings) :

https://www.dropbox.com/s/nn7rpq94yd9vuss/test%20no%20%20jitter.craft?dl=0

so a good question, what's so different between rover wheels and landing gears suspension ?

from my tests, the main difference between rover wheels and affected landing gears is the damper ratio : the LY-60, ly-99, and ly-10 are the only ones that have a 2.0 damper ratio, while other wheels have a 1.0 damper ratio.

putting a 1.0 damper ratio on the affected landing gear negated the jitterbug for me on my MK3 plane :

i tested the same modification on the small landing gear, and it also fixed @bewing's panther jitterbugging.

as those landing gears were not prone to jitterbugging when centermounted, it's more a combination of wing flexing & overreacting suspension it seems

so the problem seems to not be the same for all landing gears (nice way to complicate the problem even more ! )

 

i also tested how the 'disable' button affected landing gear / legs / roverwheels

here's a copy of my tests and observations on those :

  Reveal hidden contents

variant of the testing system :

i'm putting a x4 radial symmetry of wheels / gear / legs around a fuel tank, and i use radial rocket engines to vary the weight of the .craft on the suspensions - observation with or without suspension toggle. the following pictures are concerning the relevant problems, with my observations for each part :

 

 

Landing Gears :

disabling suspension all landing gear doesn't seem to have a physical effect (only cosmetic on some landing gears - there's still a similar visible damping and rebound physical effect)

LY-01 fixed landing gear : disabling suspension has no effect. the wheel also slightly clips into the runway at all times.

LY-05 steerable landing gear : disabling suspension works (conserves the compression it had when disabled). extended suspension visually shows a gap between the suspension and the wheel. the wheel also slightly clips into the runway. (clips more when the suspension is disabled extended) small jitters observed under 20% wheel stress, with or without suspension. edit : the visible gap might be the result of the swapped values described by @GoSlash27 edit2 : just tested the swap, it doesn't seem to fix the suspension gap.

LY-10 small landing gear : disabling suspension has no effect. the wheel also slightly clips into the runway at all times.

LY-35 medium landing gear : disabling suspension has no effect.

LY-60 large landing gear : disabling suspension works (conserves the compression it had when disabled). the wheel clips through the runway when the suspension is disabled extended. when reactivating the suspension after the wheel clipped the ground, the landing gear instantly compress back (no variation in total height, no physics reaction)

LY-99 extra large landing gear : disabling suspension works (conserves the compression it had when disabled).the wheel naturally hover slightly above the runway. the wheel clips through the runway when the suspension is disabled extended. when reactivating the suspension after the wheel clipped the ground, the landing gear instantly compress back (no variation in total height, no physics reaction)

---

Rover wheels :

Rovemax model S2 : disabling suspension works (conserves the compression it had when disabled). the wheel clips through the runway when the suspension is disabled extended.

Rovemax model M1 : disabling suspension works (conserves the compression it had when disabled). the wheel clips through the runway when the suspension is disabled extended.

TR-2L Ruggedized Vehicular wheel : disabling suspension works (conserves the compression it had when disabled). the wheel clips through the runway when the suspension is disabled extended. visual gap between the suspension and the wheel when the suspension is extended. edit : the visible gap might be the result of the swapped values described by @GoSlash27

Rovemax model X3 : disabling suspension works (conserves the compression it had when disabled).the wheel naturally slightly clips inside the runway. the wheel clips more through the runway when the suspension is disabled extended.

---

Landing legs :

LT-05 Micro Landing Strut : disabling suspension works. the leg instantly toggle between current and almost maximum extension. when disabled extended, the leg clips more into the runway when landing. total height unchanged.

LT-1 Landing Struts : no discernable variation in total height change. the leg constantly clips through the runway (toggle or not), impossible to determine the change in height.

LT-2 Landing Strut : no discernable variation in total height change. the leg constantly clips through the runway (toggle or not), impossible to determine the change in height.

in the end, wheels are currently a bugfest anyway

 

Edited April 22, 2016 by sgt_flyer

[GoSlash27]

  On 4/22/2016 at 8:16 AM, sgt_flyer said:

from my tests, the main difference between rover wheels and affected landing gears is the damper ratio : the LY-60, ly-99, and ly-10 are the only ones that have a 2.0 damper ratio, while other wheels have a 1.0 damper ratio.

putting a 1.0 damper ratio on the affected landing gear negated the jitterbug for me on my MK3 plane :

Expand  

Hmm... I'll have to give that a shot.

I've been playing with the distance, offset, and target position on the LY-01. I've stopped the jitterbugging for light loads, but not heavy ones.

Best,
-Slashy

[WinkAllKerb'']

  On 4/22/2016 at 8:16 AM, sgt_flyer said:

 

so the problem seems to not be the same for all landing gears (nice way to complicate the problem even more ! )

 

Expand  

regarding that perticular point, i always wondered if the 3Dmodelling software, choices made there with the various 3d and object layers , then the import of the 3d object to unity have an impact or not.

As Slashy said some kind of "transcritption" stuff as well on that "3dsoftware import to unity level" and then the hierarchy stuff once into unity. (notice that thoose observations of mine are not really up to date at all but still)

(may be something interesting could be someone experienced try doing the same 3d object using various 3dsoftware, then import to unity, then eventually apply a few variation here and there within the hierarchy, and/or to check from wich 3d software each wheels is imported from and if it as an impact or not)

Edited April 22, 2016 by WinkAllKerb''

[brusura]

Changing the suspension travel on medium landing gear from 0.2 to 0.8 helped a lot during take off and landing, but the ski problem remains as before, as I gently touch the A or D the plan spin violently

[bewing]

I find that changing the "friction control" to override (and sometimes reducing the friction from 1 to .6) can help a lot with spinning problems -- especially on the rear wheels.

[sgt_flyer]

  On 4/22/2016 at 1:22 PM, bewing said:

I find that changing the "friction control" to override (and sometimes reducing the friction from 1 to .6) can help a lot with spinning problems -- especially on the rear wheels.

Expand  

a slight toe in helps a lot too with rolling stability