Various controllers contain discrete-time filters and control gains that should be scaled for the time step dt but are not. This means that the performance of the controller depends unreasonably on the time step chosen for simulation. Once fixed, controller responses should look similar for "small enough" time steps.
I plan to address this once the "Long run" Hercules simulator is merged, and certainly needs to happen before a version 1.0 of WHOC is released.