solverdynamics_simple seems to almost work if integrated only for 23 h. but the expected switching behaviour still isn't seen.