Bullet: strange sim when using rbdfracturedobject.

[anupamd]

I'm getting strange results when using bullets RBDFracturedObject dop. I have attached the HIP file.

You can see how when the cluster of spheres fall to the ground, they all tend to be drawn to the average center of the cluster of objects. Which leads me to believe that bullet is calculating an invalid center of mass for each object? Or using the same center of mass for every object.

using copyObject dop is NOT an option, trying to get this to work with voroniFracture geometry.. every piece will be unique.

Is there a way to get around this without having to go into the code?

- Anu

bullet_test3.hipnc

[fsimerey]

I'm getting strange results when using bullets RBDFracturedObject dop. I have attached the HIP file.

You can see how when the cluster of spheres fall to the ground, they all tend to be drawn to the average center of the cluster of objects. Which leads me to believe that bullet is calculating an invalid center of mass for each object? Or using the same center of mass for every object.

using copyObject dop is NOT an option, trying to get this to work with voroniFracture geometry.. every piece will be unique.

Is there a way to get around this without having to go into the code?

- Anu

Same problems for me. Another problem with bullet, I have stranges calculations when i use 'Object Transform' on any RBD objects.

Any idea ?

Thanks for reply.

Join a file very explicit with a switch (houdini rbd solver/bullet solver).

Bullet_fracture.hipnc

[StephenW]

I'm having exactly the same issues.

As far as the object transform stuff goes, I found a crappy workaround which was to uncheck 'Use object transform' and use ch expressions in the position/rotation editboxes. For some reason it's happy with those.

Haven't found any way around the fractured object issue though. Looking at vimeo, clearly some have managed to get it working properly. This one in particular is great: http://vimeo.com/6638366

[sidandmj]

this is because the solver creates all geos in a fracture object with a center of 0,0,0 not the actual centroid of each fractured chunk

so to fix the problem, you must first center all your chunks in sops but store their original position as an attribute, then use a position dop to retrieve the attribute and reposition the pieces. this will then transform the centroids as well and result in a happy sim.

or we need to find a way to rewrite the plugin such that it computes the centers the way an rbd fracture dop does.

I'm having exactly the same issues.

As far as the object transform stuff goes, I found a crappy workaround which was to uncheck 'Use object transform' and use ch expressions in the position/rotation editboxes. For some reason it's happy with those.

Haven't found any way around the fractured object issue though. Looking at vimeo, clearly some have managed to get it working properly. This one in particular is great: http://vimeo.com/6638366

[StephenW]

this is because the solver creates all geos in a fracture object with a center of 0,0,0 not the actual centroid of each fractured chunk

so to fix the problem, you must first center all your chunks in sops but store their original position as an attribute, then use a position dop to retrieve the attribute and reposition the pieces. this will then transform the centroids as well and result in a happy sim.

Thanks for the tip. In case it's useful to anyone else, find attached a quick test that drops a fractured ball and recomputes COG with a quick and dirty points from volume average. It was done in H11 so if you need to load it in 10, just replace the voronoi stuff with a shatter node. I tried putting the dopexport in the foreach so I didn't need to make 10 of them but that didn't go well. It's a bit budget but works...

There's a lot of penetration going on with the Bullet collisions here. I've tried using rydalch's version (http://gitorious.org/~crydalch/bullet-physics-solver/rydalch-bullet-physics-solver) to get substeps and improve this, but it seems to crash out a lot.

Time to look at the code...

test1.hip