Rigid Body interpenetration where there shouldn't be any

[tjcoolid]

I have an animated model of a injection molding machine. At one point in the animation cycle, the machine molds open, and an ejection mechanism pushes out 24 plastic parts (bottle caps) that then fall out onto a converyor belt below the molds. I'm using dynamics for the caps to fall due to gravity and then have them carried away on the conveyor belt. Everything aside from the parts falling and being carried away is keyframed. I got everything working properly through correct use of "passive key" and "active key," but now I want to add a second set of bottle caps as the machine reaches the ejection point in the second cycle. I have a second set of bottle caps that are invisible and completely out of view until the appropriate frame, and then I wanted to set an "active key" at the right moment so that the second set of caps would fall as expected from the appropriate position. I ran my animation back to frame zero to play through, and now the first set of caps falling is all fouled up, and I'm seeing an error message indicating that there are interpenetration issues with rigid bodies of the first set and rigid bodies of the second set. How could that be? I've tried several different ways of making sure the second set of bottle caps is nowhere near the first set when the first set falls, yet there still seems to be some sort of conflict. I feel like I must be violating some cardinal rule of rigid body dynamics, but I don't know where to start troubleshooting at this point. Can anyone please help?

[tjcoolid]

I've tried several different attempts to set this up differently to see if I can achieve the results I'm looking for, and I may be making progress. I decided to remove "rigid bodies" from all of the caps (both the first and second set) and create a single active rigid body for both sets instead of having two separate rigid bodies. I then was very sure to set positional keyframes in addition to active/passive keyframes for the second set of caps so that they would not interfere with the first set when they become active and drop due to gravity. This seems to have been successful, but now I've run into a different problem. When the second set snaps into position, becomes visible, and becomes active, they do not drop due to gravity. I've made sure that they are linked to my gravity field using the Dynamic Relationships Editor. The caps start to eventually spin, but they remain stuck in their x,y,z position. I'm getting an error message that two of the caps from this second set are interpenetrating, but again... I don't see how that's possible.

I'm starting to suspect that this has something to do with where (and when?) an object is when I create the rigid body and that it may not be as straightforward as "the rigid body is where the object is." Could it possibly be like moving a group vs. moving an object within a group? A group can remain at 0,0,0 while an object within a group can be anywhere. Is there a parallel when working with a mesh and its corresponding rigid body?