I tried to break it by setting the parameters for the fitting to unrealistic values and by making Dmax for the slow bigger than for fast. But although it took a while for the fit to compute, it did not break.