Spurious horizontal movement with mesh walls

Hi,
Sorry to flood the bug report forum. Here's a synopsis of a bug that i recently encountered that causes spurious motion of particles that interact with mesh walls. The simulation (which can be recreated using the attached files) involves just one particle falling under gravity. The particles falls down vertically on a bottom wall. I tried three different ways to model the bottom and here is a summary of what I observe in each case:

1. Bottom wall modeled using the "primitive" option with "fix wall/gran/...": This case works as it should. The particle bounces a bit and settles down.

2. Bottom wall discretized using a coarse two-element mesh (use file mesh2.stl): As soon as the particle hits the bottom mesh it gains a spurious velocity in the x and z direction and starts to move diagonally. The forces on the particle are negligible (~1E-15), however the velocity in the x and z direction is around 1 mm/s and stays constant after the first few thousand iterations. This particular behavior is processor independent.

3. Bottom wall discretized using a fine mesh (use file mesh1.stl): The particle hits the bottom mesh it gains a spurious velocity in the x and z direction and starts to move diagonally. However unlike the previous case it oscillates back and forth as if bounded by some invisible wall. This behavior is processor dependent. I see it if I run it on 4 processors but not if I run it on a single processor.

Thanks,
Tapan