I have been struggling in getting milling PCBs just right. I just couldn't get the depths right and the isolations between the traces looked ragged and looked like they were milled.
The main problem, I think was getting a truly level board. Even milling a flat pocket does not give a flat surface as there are always imperfections in the flattest of PCBs.
I ran across some software called Autoleveller which modifies your gcode to add probing gcode. When the code rns on your machine, it probes every 5 millimeters (or the distance you set). It then adjust you z axis to allow for board imperfections.
I used Diptrace to design the board. The then exported a gerber file of the bottom traces and a N/C drill file for the holes. I imported these files into Flatcam which produces the raw g-code. I then loaded the raw g-code into Autoleveller and produced the leveled code. That code was loaded into LinuxCNC and run on m machine.
Here is the outcome. I manually cut the outline myself and think I cut it too close. I really should have allowed the code to cut the board. But it is only a prototype after all.
I am pretty impressed with how the traces and isolation turned out. I don't think I could have done better with the toner transfer method.
The main problem, I think was getting a truly level board. Even milling a flat pocket does not give a flat surface as there are always imperfections in the flattest of PCBs.
I ran across some software called Autoleveller which modifies your gcode to add probing gcode. When the code rns on your machine, it probes every 5 millimeters (or the distance you set). It then adjust you z axis to allow for board imperfections.
I used Diptrace to design the board. The then exported a gerber file of the bottom traces and a N/C drill file for the holes. I imported these files into Flatcam which produces the raw g-code. I then loaded the raw g-code into Autoleveller and produced the leveled code. That code was loaded into LinuxCNC and run on m machine.
Here is the outcome. I manually cut the outline myself and think I cut it too close. I really should have allowed the code to cut the board. But it is only a prototype after all.
I am pretty impressed with how the traces and isolation turned out. I don't think I could have done better with the toner transfer method.