ode activation

[MENOZ]

hi, ODE solver don't works as rbd solver when dealing with activation.

It has strange behaviuors

I attach a file with a basic scene where the objects are activated one per frame.

with RBD solver all works, with ode solver no.

a bug?

activationODE_fail.hipnc

[Emanuele Berti]

Hi,

what do you mean with 'strange behaviour'?

Because I've seen your scene and I'm not sure about what you are after.

Try to reduce the 'Error reduction' in the ODE solver to something like 1e-007 and set the object type to capsule in the ODE primitive tab...

it should do the job.

Even with these changes you won't have the same behaviour.

My 2c,

Emanuele

[hopbin9]

Yea, I don't think this is an activation issue.

The second sphere is activated to close to the first (based upon the settings) and is popping into the others causing a chain reaction.

If you change $OBJID*2.3 to $OBJID*2.4 to position the spheres, then it works fine.

Edited November 25, 2010 by hopbin9

[MENOZ]

Yea, I don't think this is an activation issue.

The second sphere is activated to close to the first (based upon the settings) and is popping into the others causing a chain reaction.

If you change $OBJID*2.3 to $OBJID*2.4 to position the spheres, then it works fine.

@hopbin9: I see, I think that changing the error reduction is better because it's more general and i'm not forced to change the object position.

@Emanuele: this solved part of the problem. Now that objects are activated and falls down correctily, bounces are still incorrect: the first ball bounces, others don't.

I find that checking "solve per object" on the ode solver makes correct bouncing balls, but if they are one on top of the other they seems to use an old position to calculate bounces.

The simulation goes right if checking the box "create active object" on rbdObject.

In the hip file there are sticky notes that shows various scenarios.. maybe this helps

thank you!

ode_scenarios.hipnc