What about Parasolid, STEP, or IGES? I've never had much luck getting openScad files into SW.
I'm not sure about using a floating ball. At least I found that as the angle of the joint increased that lateral forces could easily separate it. That's why I ended up pouring ball seats even though the plain ring and ball is plenty enough to hold things normally (in my regular delta bot). During prints I would hear occasional chattering as the magnets would separate and then click back in place under load, the seats prevent them from moving laterally so the magnets can keep applying their full force. With a ball magnet and full cup seat I would expect you'll get more even force than a countersunk ring, but it's still going to want to slip. With two floating seats one of them will almost always be taking shear forces and it'll hurt even more as the seats pull away from the poles of the ball. If you fix the ball to one of the ends you can use a much smaller rigid mount and a larger seat on the floating side to preserve your range of motion.
I'm not sure about using a floating ball. At least I found that as the angle of the joint increased that lateral forces could easily separate it. That's why I ended up pouring ball seats even though the plain ring and ball is plenty enough to hold things normally (in my regular delta bot). During prints I would hear occasional chattering as the magnets would separate and then click back in place under load, the seats prevent them from moving laterally so the magnets can keep applying their full force. With a ball magnet and full cup seat I would expect you'll get more even force than a countersunk ring, but it's still going to want to slip. With two floating seats one of them will almost always be taking shear forces and it'll hurt even more as the seats pull away from the poles of the ball. If you fix the ball to one of the ends you can use a much smaller rigid mount and a larger seat on the floating side to preserve your range of motion.