Dear Warson,
Obviously a RF gap that's too steep isn't possible, and a 3D study may be needed to check, but experience shows that a reasonable ramp poses no problem. And indeed, RF experience in this field is also necessary. Attached is an example of a ramp calculation carried out by GenDTL including frequency shifts directly in the cell design, which should minimise the expected losses in the post-coupler.
Regards,
Didier