Page 1 of 1

Momentum RBD: Create and orientation force

Posted: 10 Nov 2014, 11:48
by Pancho
I tried to blend the current euler, quaternion and rotation values with my goal orientation. None of it really worked. It never reaches the goal orientation and if something changes, than with strange flipping issues.

I just want to create a force which tries to establish the goal orientation as much as possible. I can SET this orientation easily, but I didn't succeed in making it a force like the velocity. GoalPos - currentPos > Normalize > linear interpolate with current velocity. Perfect! But for the orientation it seems to be more difficult. Can anybody shed some light?

Thanks in advance
Thomas

Re: Momentum RBD: Create and orientation force

Posted: 14 Nov 2014, 20:41
by Pancho
I haven't found a solution yet. If I mess around with the orientation the RB intercollisions gets messed up and objects intersect. Setting the orientation seems to "override" the internal orientation of Momentum, resulting in wrong intercollision.

The angvelocity doesn't seem to be responsible for the orientation since it's axis is independent from the orientation. Although I believe the rotation axis should be aligned with orientation axis in order to have a smooth transition between rotation movement and the correct end position.

Did anybody ever needed to align the RBs with the surface of a collision/goal object? Any help to create this force is highly appreciated!

Still hope that Exocortex will shed some light into this....

Re: Momentum RBD: Create and orientation force

Posted: 17 Nov 2014, 13:43
by Pancho
O.k., to illustrate the problem. Here are the three results.

To me it seems that orientation doesn't get update internaly. The "Set orientation" works, but it doesn't get recognized in the RB calculations. It looks just like an override. As you can see I tried to increment the rotation with my wanted axis. If the Weight is set to 0 then nothing happens, set to .5 it overrides the internal orientation by 50%, set to 1 I finally get my orientation as wanted. But then the RBs just intersect. THey still believe they aren't rotated. It works like a post-simulation operator.

In the next post is my ICE tree. I can't believe that there is no way to align the objects with my surface with a force. Anybody an idea?

Re: Momentum RBD: Create and orientation force

Posted: 17 Nov 2014, 13:43
by Pancho
Pretty simple ICE tree.

Re: Momentum RBD: Create and orientation force

Posted: 17 Nov 2014, 13:48
by Pancho
P.S.: Is there anybody left at Exocortex who still takes care of the SI support? I didn't get a reply from them within one week and the google user group is pretty much abandoned. In the past Ben always came up with great solutions within a few hours.

Re: Momentum RBD: Create and orientation force

Posted: 17 Nov 2014, 23:25
by wireframex
Just try to connect him with skype

Re: Momentum RBD: Create and orientation force

Posted: 18 Nov 2014, 00:51
by Tekano
pancho, I did not try to work out current rotation to angular momentum as was unsure of format. it seems its -180 - 180 per axis..? anyway I just set the orientation. seems to work fine. the half submerged thing that you had could be the centres.. did you move them to bottom of instance objects. or the substeps. put that up also if you are getting penetration.


Re: Momentum RBD: Create and orientation force

Posted: 18 Nov 2014, 01:25
by FXDude
Lol cool =) thanks Tekano!

Re: Momentum RBD: Create and orientation force

Posted: 18 Nov 2014, 03:08
by Tekano
aaand here is a cheeky way to orient along the local Y using just angular velocity.! the trick is to give a small Euler vector. because you dont want it to spin right? Anyway there is a very specific formula to do what you are asking and I could not bebothered to do that so here is an 'approx' method later
angular_velocity_hack.JPG

Re: Momentum RBD: Create and orientation force

Posted: 18 Nov 2014, 12:19
by Pancho
Thanks so much!!!

Yesterday evening just before supper my head started to work again. I understood angVelocity the whole time as a dynamic force, taking care of the "wild" rotation of the instances. But it is exactly this angVelocity which can be used to change the orientation.

I hope the image sheds some light.

Thanks for the suggestions and help!