6Dof linked physic object hang down a little bit

Troubleshooting physics
1

Hi

Sorry bothering you again. I now also try to figure out to make a servo like object. To be able to do complicated moves I tried out the SixDofJoint. The first was to limit all to zero, just for testing, but it seems that this joint is not 100% fixed then the attached block just hangs a little bit lose. Just a bit but visible.

And here is what I did:

SixDofJoint dof = new SixDofJoint(hubPhy1, hubPhy2, new Vector3f(0.5f,0,0), new Vector3f(-0.5f,0,0), false);
dof.setAngularLowerLimit(Vector3f.ZERO);
dof.setAngularUpperLimit(Vector3f.ZERO);
dof.setLinearLowerLimit(Vector3f.ZERO);
dof.setLinearUpperLimit(Vector3f.ZERO);
dof.setCollisionBetweenLinkedBodys(false);
bulletAppState.getPhysicsSpace().add(dof);

Any suggestions?
Christian

2

Hm for exact simulations physic engines are nearly unusable, i suggest to implement this yourself with two cubes,

one rigidbody in normal mode
and one as kinematic mode wich is moved by the servo.

3

Hmmm not exactly what I want as this is some sort of pluggable robot :slight_smile:
Just wonder why this hangs down. I thought I did this with C++ bullet and it was not hanging. Btw the collision is set to false but set it to true as well do not change a thing.

4

It seems that the stiffness of the joints (all, no mather if I use hinge or 6dof) is not very big and can not realy be influenced (at least I did not found a way so far). The picture explains:

EDIT: I played around with mass and size. A bigger shape makes everything slower, a bigger mass as well makes it slower (it as well fall down slower), but it has a higher stiffness.
If I have a small shape (1,1,1) and mass 1 everything is faster but as well less stiff. If I use a too huge mass and a small shape the simulation goes wild.
So I’m not that sure what I shall do in the end :slight_smile: I do more experiment. Seems to be very tricky.