Chains keep moving like worms - How can I stop them?

tiagoandriottitiagoandriotti Posts: 39
edited December 1969 in Carrara Discussion

Hi

I did a simple chain, an object duplicated 7 times, changed keyframe to physics and the Physical proprieties to metal (effects tab) in all of them.

After some tests I changed the bounce to 0 but the chain that should fall and stop moving in 1 second keeps moving until 2,5 seconds, like worms lol.

Is there any trick to stop these objects moving?

Comments

  • 3DAGE3DAGE Posts: 3,311
    edited December 1969

    HI TiagoAndriotti :)

    what version of Carrara,. and what physics engine are you using.?

    Generally there are a couple of things which can effect the physics,
    One is the Mass, Bounce and Friction settings on the chain you made,
    and the other is the same settings on the objects it's colliding with.

    both things effect the physics, even if the object it's colliding with isn't a "physics" object.

    Think of a bouncing ball on a rubber mat, and a concrete floor,. one will absorb more of the force than the other.

    Hope it helps :)

  • evilproducerevilproducer Posts: 9,050
    edited December 1969

    Physics is a simulation, but it's not a scientifically exact simulation. It's more of an artists conception of a simulation, since that what Carrara is in essence, an art and animation program.


    You can convert your physics simulation to keyframes by going to the motion menu and changing it back to keyframes. Carrara will ask how accurate you want the conversion. From there, you can either delete the extra keyframes after 1 second or compress them to fit in that space.


    You could also play around with friction, gravity, etc. etc. Personally, I think converting the motion to keyframes is the easiest.

  • tiagoandriottitiagoandriotti Posts: 39
    edited January 2013

    Thanks 3DAGE :), you're right, the floor is not "physics", and the rubber ball with the concrete floor is a good reference, I will make some tests with other configuration.

    And thanks evilproducer about the animation conversion trick, it is a great idea. Meanwhile I'm learning the physics behavior I can make this :)

    I'm using Carrara 8 x64 windows

    Post edited by tiagoandriotti on
  • 3DAGE3DAGE Posts: 3,311
    edited December 1969

    No Problem :)

    I'm glad to help.


    A couple of other things you can do to limit movement,

    Add a "Damping force"

    Add a "Direction force", pointing down, and switch off the Scene Gravity,. .. the direction force will act like the gravity, but,.. the force value can be key-framed ,.. so you can have no force. or a strong force to stop the object bouncing.

    :)

  • JoeMamma2000JoeMamma2000 Posts: 2,615
    edited December 1969

    I think the advice given so far might be missing the real culprit.

    What the OP is describing is continuing motion when the object SHOULD be coming to rest.

    Mass, Bounce, and Friction are merely material parameters. They define what type of material you are using. And while that obviously affects the object motion during the simulation (falling on concrete gives a different motion than falling on rubber), it shouldn’t have much to do with the problem of objects moving after they should be coming to rest (ie, “worm” motion). You should be able to (and can, BTW) decide you want a metal chain falling on a concrete floor, set the Mass, Bounce, and Friction appropriately, and have the simulation behave.

    And yes, the Bullet simulator isn’t physically accurate, but I don’t think we’re talking about physical inaccuracy, but rather an issue with the simulation not behaving correctly. And I don’t think that has anything to do with conversion between physics and keyframes. Or Friction and Gravity settings. Nor do I think you need a Damping Force or Direction Force.

    Instead, I think the key issue here is some things that haven’t been mentioned, and those deal with the simulation accuracy. If you go into the Scene settings, under Physics, you’ll see three critical settings that define how accurate the simulation will be. Simulation Accuracy, Geometric Fidelity, and Collision Distance. All three of those have a huge effect on how well the simulation will behave.

    Collision Distance is critical because it defines how much distance the simulation should allow between objects. If you start out with two objects spaced 1 inch apart, but set the Collision Distance at 2 inches, the simulation might freak and give some “squirmy” results. You need to make sure your collision distance is small enough. BTW, I think it is in % of a scene size unit.

    But more importantly, in this case I think you’ve probably got the default settings for Sim Accuracy and Geometric Fidelity, which are probably the culprit for the “wormy squirmy” motion. Geometric Fidelity defines, I believe, how closely the sim models the object surface relative to its actual surface (eg, does it simulate a torus as a simple cube?). And Simulation Accuracy, which I think might be the real cuprit here, defines just that…how accurate the simulation is, and how much of a trade-off it makes versus speed.

    So I would crank up the Simulation Accuracy setting, and maybe also the Geometric Fidelity (not sure if that will matter in your case), and make sure your collision distance is low enough. I usually start with something like 10% for that.

    I just did a chain simulation I to show how you can have a nice, smooth, falling chain simulation that “damps” as it should and comes nicely to rest, as long as you use good values for Simulation Accuracy and the others. I'll post the link here in a bit...

  • JoeMamma2000JoeMamma2000 Posts: 2,615
    edited December 1969
  • tiagoandriottitiagoandriotti Posts: 39
    edited December 1969

    Wow JoeMamma2000, that's a nice animation.

    I'm a beginner in animation and physics, for now I use it just to achieve some natural look in my composition. After making some "easy" tests I decided to make a natural looking heavy chain, first in the ground and after that and animated version.

    I will save your considerations and try to apply it.

    Thanks a lot

  • DartanbeckDartanbeck Posts: 21,543
    edited December 1969

    Awesome sim, Joe
Sign In or Register to comment.