Hi Luis,
The appropriate change here is not to decrease the Max Step Size setting but to decrease the Relative Tolerance setting.
After changing the Relative Tolerance setting to 1e-4 from 1e-3 the simulation proceeds as expected. One indicator this might be an accuracy issue, and not something more fundamentally wrong with the model, is that the change in current at the "discontinuity" is quite small relative to the ripple .
You should set the Max Step Size back to the original setting of 1e-3 as having an artificially low Max Step Size simply slows down your simulation in the majority of cases.
Regards,
Bryan