Well, the graphical part of that is embarrassing, and yes that is what was happening there. However, the continuity constrains are still violated significantly and do not converge after rerunning the optimization with the same setup as in PR4. Is that expected?