With the internal glue set to -1 my rbd glue object still breaks apart ! before even hitting the floor when used with the blend solver set up. I cannot see any reason for this to happen any ideas ?

r

To follow up here is a scene where the RDB glue solver is set to -1 , if you set the initial state of the rbd locally it all works as expected. I really have no idea why blending from key frame to rbd should cause so many problems. ie effecting the internal glue value.When I even reference Craigs Gnomon dvd just to double check , the result is not as expected.

Any ideas whats going wrong ?

Attached 2 examples

circusmonkey
With the internal glue set to -1 my rbd glue object still breaks apart ! before even hitting the floor when used with the blend solver set up. I cannot see any reason for this to happen any ideas ?

r

Hey circusmonkey,

I think I see what you are trying to do. I think the problems is how you are blending between the two states. One state uses $FF < 25 and the other is what is not $FF < 25. I think the latter state is the problem. What you need is a state that is perpetual that will blend with a state that is defined by $FF < 25. I changed it to $FF > 1 instead of not $FF < 25. I hope that this makes sense. Here is the updated file.

Cheers

I think I see what you are trying to do. I think the problems is how you are blending between the two states.

I could be wrong, but I think to really fix this is even trickier than that, unfortunately. I believe what's happening is related to the way the RBDGlueObject works internally. Basically it creates an empty object with the name of the top level RBDGlueObject, then creates all the pieces with a RBDFracturedObject and glues them to that top-level object. The object that the pieces are glued to is stored in the “glueobject” property on the Position data for each piece when the RBDGlueObject is created.

With the BlendSolver setup here, however, the ObjPosition DOP is creating a KeyframedPosition data from the object's animation, then for the first 24 frames the BlendSolver is copying that into the Position data on the pieces, promptly overwriting the “glueobject” parameter on all the pieces with nothing. So if you look in the Details view at the “glueobject” property for the pieces with your setup, it's never set to anything, though it should be set to “box_object2”. So the pieces are not actually glued to anything as far as DOPs is concerned, and as soon as the RBDSolver takes over at frame 25 and they each have angular velocity, they fly apart.

Fixing this takes a few steps. See the attached file.

1) Your group parameter on your BlendSolver is set to “piece*”, but you need the Position data in the empty box_object2 updated as well, so change that to “*”.
2) You need to get the KeyframedPosition set properly for the empty box_object2, but the ObjPosition DOP depends on the object having Geometry subdata, which box_object2 doesn't, so you can't use it on that object without getting an error. Instead I've just Fetch'ed the data from one of the pieces and Apply to the box_object2.
3) Add an RBD State DOP with a group setting of “piece*” that resets the glueobject to “box_object2”, but do that for the data named KeyframedPosition, so that value will get blended in in the first 24 frames and be set from then on. This is where you can re-set the value for the glue strength as well.

Thanks for the gold reply.!
I am beginning to think that there's quite a few fiddles to be made , just to get some functionality out of the 3 rbd obj nodes when coupled with the object RBD obj position DOP and the active DOP Thanks for clearing up the confusion !

Cheers r