Ordering 7i76/7i77/whatever breakout board might be useful, but I do have working electronics on a (home brew) PCB that accepts a DB25. I might want to continue that route since I already use home-brew stepper drives around dsPIC33F DSP's and I do want to continue on the stepper-with-linear-scale-feedback path. My earlier experiments on the BF20 were quite promising, but that machine has too much mechanical backlash. The sound of a hunting 'servo' certainly makes you feel nervous.
I have this idea to run the steppers in velocity mode using FOC / vector drive as 'regular' high torque/low RPM motors. Voltage mode drive would make it easy to detect (and act upon) stall, would not suffer from the 90-degree phase shift a current-mode drive encounters when it runs out of bus voltage, and it doesn't produce the typical 'stepper noise'. And because a stepper with stall detect is fully synchronous I do not need a rotary encoder on the shaft and it makes life for the servo loop very easy since actual RPM is extremely close to commanded RPM. Final position feedback through 1 micron linear glass scales. Sounds like a great setup for machines that do not need more than ~200W of mechanical power to drive the screws. Stepper motor+drive+scale costs less than a single servo motor+drive.
a gantry style machine needs at least 4 axes worth of linear encoders, and possibly shaft encoders which totals to 8. Initially 4 step/dir interfaces are needed; and later on PDM (or a SPI protocol) to the drives to command velocity might be more suitable.
Whether a card supports the amount of encoderrs/stepgens or not depends on the firmware. In
this[.url] document there is not much information about the 5i25/6i25.