With more dodgy hacks making row and column coupling identical we progress. Next will be working around linearPDE wanting dirac points