Hello,
Missil, I am using solid to solid contact. I tried soft method and pinball and the results are similar.
Greg, Because the solids are cylinders with different diameters; if they don't slide freely when having radial deformation there will still be forces in the contact plane.
I checked the nodal forces for nodes of one of the cylinders and I notice out of 1700 nodes almost all of them only have x forces (axis x perpendicular to contact palne so all forces should be in x direction) but some node also have x and y forces. It's like they get stuck to the surface of the other solid and can't move or something like that. I tried the analysis with and without initial penetration elimination, no difference though.