A good approach to making sure the contact definitions are working well is to solve the model with a coarse mesh. When that is converging nicely, then refine the mesh in the areas you need finer spatial resolution. That way you don't spend a lot of time waiting on a model that will not converge to learn that sad fact.
Have you requested 3 or more Newton-Raphson Residual plots under the Solution Information folder? Those are very helpful for convergence diagnostic purposes.