Shatter scripts, constraints, and dynamica questions

[curtainsNZ]

Hi all, just wondering if I can get some help on a few issues I've run into with the Dynamica plugin for Maya.

I'm basically trying to create a building demolition shot - so I'm using a shatter script to fragment my geometry, then use the Dynamica physics plugin to simulate these pieces as rigidbodies. Small pieces etc will then be created using instanced geometry to particles.

The problem I've run into is getting unshattered geometry to turn into the smaller fragments during a simulation. Has anyone come up with a solution to make shattered geometry move with the unshattered geometry before it's mass is keyed to make it fall apart? I'm guessing it involves some kind of constraints system but I haven't had much luck yet.

So far I've tried parent constraining the shards to the unshattered geometry but the entire simulation seems to move, as illustrated here:
http://vimeo.com/10888170

Has anyone come across a solution to this yet?

Cheers for any help!

[majestik666]

Looks like the pivot is wrong ...

Not sure how Dynamica handles it

[curtainsNZ]

Exactly -like the entire simulation's pivots are moving. Does anyone know if is is actually possible to move dynamic geometry using keys/constraints...?

[sparkprime]

Looks OK to me. What is the problem exactly? Why is the camera at a funny angle, and I assume there's an invisible wall there too.

What I was planning to do was to fracture by removing the rigid body and creating two new rigid bodies using a subset of the compound parts in each. You also have to move the CG and adjust inertia but this can be taken into account too. It would also be possible to use completely separate compounds for each fraction of the original mesh, in order to not constrain the design original compound (e.g. using a whole box in your example instead of lots of pieces in a compound). I don't see the need for constraints anywhere.

[curtainsNZ]

Hmm -I suppose that video doesn't show the situation too clearly.
The problem is that there is a ground plane (just a rigidbody plane with 0 mass) there already - but it is level. This shot was taken from a side orthographic view to best illustrate the odd angle of the plane that the shards seem to collide with.
Here's another shot that might illustrate the problem a bit more succinctly:
http://vimeo.com/10948635

Regarding your solution - so it's possible to generate new rigidbodies during a simulation? That would change things entirely...

[sparkprime]

Yes, in fact it is sometimes necessary to take the body out of the simulation and put it back in again, in order to reset some caches. Creating a new rigid body is just a case of allocating storage for it, and placing it into the simulation between internal time steps.