Dynomotion KFlop?
Can pretty much program it to work however you want, and it'll work with Mach 3.

If you're running drivers that accept differential inputs, I may still have enough parts around to create a step/dir differential driver board (or I could just do the SMD work, and you can deal with the through hole parts). Only drawback is with just a bare KFlop, interfacing to the low voltage signals can be a bit of pain (you can always add a Konnect to get 24V I/O).

Only other options I can think of are Po-Keys, and the CNC-Drive controllers. PMDX also do some, but I think they're Mach 4 only.