Good to know, I assumed it’s just a fault of Mortar contact’s general issues with convergence because surface to surface worked as expected.
Standard approach in Abaqus, but may indeed cause issues in CalculiX. However, based on this post, I suspected that Mortar may not work well with rigid bodies so I also tested without it, but there wasn’t much difference.
The indenter was originally deformable. I only made it rigid later to quickly reduce the problem size (as mentioned above - it’s commonly done in Abaqus).
Interesting. I’ve heard PaStiX might be even better for contacts, but I usually avoid it and rely on Pardiso as the most reliable solver (and one of the fastest).
I’ll look for other, simpler examples. My experience with Mortar contact so far is that when I select it, the analysis that used surface to surface contact usually stops converging right from the start, but maybe this type of contact just needs to be used even more carefully in CalculiX (apart from the bug you’ve noticed).