Convergence Problem???
I would recommend using more than 3 adaptive passes as a default. For HFSS, I usually set deltaS to 0.005 and maximum number of passes to 10 or 12. Check the convergence before viewing the results. It is a good habit to get into.
Also remember that converging on input impedance is no guarantee that the radiation boundaries (and therefore gain, radiation efficiency, etc.) have been adequately sampled. This is a common mistake when analyzing radiation problems.