Nema17 motors driving 1204 ballscrews ...that ok?

[Jonathan]

The efficiency of a ballscrew is typically over 90%, compared to maybe 30-40% for a threaded rod which relies on sliding friction, so you would expect a big increase in feedrate. The torque required to accelerate the ballscrew would be greater due to the increased diameter, however since the pitch is greater the required angular acceleration is actually less, so overall you'll be much better off with the ballscrew.
The main reason for ballscrews having higher efficiency is they rely on rolling friction, not static friction, so you may want to consider trying the method I demonstrated here, where the screw nut is replaced with a bearing at an angle resting on the threads. That converts the drive to mostly rolling friction, so the efficiency is greater, plus it's also easy to get very low backlash. It's a very cheap way of upgrading, but I'd still stick with ballscrews if you can fit them in due to the accuracy (among other things).

Perhaps by using timing belts and pulleys to drive the screws you could fit in bigger motors? You could also use a favorable ratio to get higher resolution for milling PCBs. Incidentally, what feedrate and spindle speed do you currently use?

£66+15 for that spindle seems a lot to me when compared to the cost of the small brushless motors and controllers from hobbyking. Granted they're not that quiet, but I'm sure you'd find ways round that. The substantially higher RPM should make up for that too, since you could cut the PCBs faster.