I think after the last changes in FixWallGran::post_force_eval_contact the line
vectorAdd3D(x_[cdata.i],cdata.delta,contactPoint);
Should be
vectorSubtract3D(x_[cdata.i],cdata.delta,contactPoint);
because the sign of delta has changed in the past.
One can check this with the script attached and having a look in post/fwc*.dump, the expected contact points are at z=0 and z=0.009 (wall bottom and top) but they are at the opposite side of the particle which touches the plane . (particles are expected to wiggle between bottom and top because cor=1 and gravity is off)
edit: added 2nd testscript with diagonal wall (in_V.txt) and two images to show the difference/problem.
Attachment | Size |
---|---|
in_pp.txt | 2.63 KB |
lid.stl_.txt | 309 bytes |
in_v.txt | 2.53 KB |
v.stl_.txt | 649 bytes |
correct.png | 30.98 KB |
wrong.png | 31.09 KB |
ckloss | Wed, 01/14/2015 - 19:10
Hi Christian,
Hi Christian,
correct! I'll fix it for the next release
Many thanks
Christoph