To see what would happen, I ran some gcode from early #CDCNC testing. But faster. I think the CD mill did this at 350-400 mm/min iirc. For this first go with "1551" axis I set feed at 600 mm/min.
That was going pretty well until it died at 2.1 mm depth in the profile slot (of 2.3 mm to pocket bottom and 2.8 mm to complete the profile). No surprise. The slot profile was a mean test. #CDCNC never got anywhere near that far with the slot. For that project it wasn't long before I reCAMmed this job as 1) the pocket, then 2) another pocket surrounding the outside profile of the part, wider than a slot, with the outside face sloped so there was only one vertical wall to deal with. This just chewed thru 20 laps of the profile, with the spindle slowing down in the slot then regaining rpm when back in the pocket. I think that counts as success for the axes to feed hard enough to bog down the spindle motor.
This unfortunately highlighted a problem that I didn't have with the horizontal configuration of the CD mill: chips laying on top of the part and getting re-re-recut. I tried knocking them away with a toothbrush, but still ended up with acrylic froth crust on some surfaces. I figure that contributed to making the slot hard to cut. Not sure what to do about that. I was kinda liking the small footprint potential of the vertical configuration, and don't relish the idea of compressed air blowing chips around a tabletop.
Next up: is Z hysteresis small enough to run paths with varying Z vs a succession of descending waterline paths?