I'm back for a second round. This time I changed tacks - I have turned the $270 CNC paper cutter into a 2.1D mill for mil-etching PCBs - instead of the original drag-knife method. See video here
http://www.youtube.com/watch?v=rDRN1_i0vJU and some updated info here
http://research.microsoft.com/en-us/projects/papercutterpcbs/I haven't calibrated the setup yet like ipm, position tolerance and repeatability and minimum isolation track width & repeatability but that will come. The original method I was trying to perfect was to drag a carbide knife through copper cladding to affect electrical isolation but the swarf remaining was a problem.
Mike
[Non-text portions of this message have been removed]