Using Angular velocity to controll joints

Troubleshooting physics
1

I have a lamp, with 3 joints and 4 links. All the links have mass assigned. I am using (node.setAngularvelocity()) directlyin the desired direction to move a link around the joint axes.

But I actually want to use torque? I cant move the links when I say node.addTorque(vector(1,2,3)). 


  • Is it ok to use Angular velocities directly on a NODE?
  • Why cant I move the same Node, by applying torque?



    Finally, oncy my joints reach the desired pose, the A_velocity is still getting applied, which is causing the lamp to jitter a bit…
  • What would be the best way to stop it jitter, once reach the desired angle? Since Gravity is acting , if I stop applying Avel, they collapse… or even if I scale it down :slight_smile:
2

I found that appleing torque directly to an object works, but you need to manually unrest (Node.unrest()) to actually get it to move.



(Irrisor is this a bug or expected behavior?)

3

If results are ok, using velocities is ok :slight_smile: - you might get problems with constraint resolving, though (e.g. if more joints are collisions influence your lamp)

Using torque should work. You need to use pretty high values (thousand, probably) as with forces.

What about using powered joints instead? That's usually the best solution.



As for stopping the jitter and to prevent your lamp from slowly collapsing, you can apply an additional "fixed" joint (without any joint axes) while your lamp should not move.

4

No, applying force should unrest a node automatically.

5

Not force, torque.

6

dito (if it does not it's a bug)

7

You might want to take a look then,

I can post a test case but I already have tested it quite thoroughly.  I can call .setForce() and balls move, if I call .setTorque() I must call .unrest() or the rotation never happens. :slight_smile:

8

you were right - should be fixed now, thanks