Split of cooling time cfl

[pgrete]

The adaptive and Townsend integrators do not require a limitation on the timestep a priori. Nevertheless, from a physical point of view it may be desired to limit the timestep anyway so that the thermal dynamics and other dynamics don't decouple. While this is currently supported/tied to the adaptive rk12, and rk45 cooling integrators, we should separate that option and properly document it.