The way I got around this with my CNC was to clamp a piece of scrap wood and then route a shallow hollow so it's totally flat using a 1/8' end mill. A difference of a tenth of a mm can be enough to ruin your traces. If you've got a CNC that supports auto levelling then you're in luck! Otherwise, you've got a bit of fun and games.

You'll probably cry a handful of times when you're using FlatCAM, so you might want an intermission or spread your work over a couple of days so you don't get too worn out. It'll crash a lot so keep saving and don't feel bad when you start crying - that's what the tissues are for. Next, grab a box of tissues and work out how FlatCAM works. Make sure to have a conductive path from the center of the board to the rest of the blank copper PCB, because we'll need it to level the drill bits. So I designed the PCBs in Eagle (Although KiCad would work just as well), exported the board profile + traces Gerber files and drill Excellon file to FlatCAM V8.991. This needs to have a wire soldered to one corner Unfortunately, I couldn't for the life of me find any 2 sided boards that aren't FR4, so you're limited to that epoxy board material). PCB (Not FR4, the glass dust is horrible for you and will screw with your CNC. Also, the larger the board the more problems you have with levellingĬNC router with that electronic collision detection thing Small IC packages will also be hard, as the drill bit tends to tear the copper to an extent, so thin traces get ripped up. Before I start: tl dr - it'll take a lot of time and effort, you'll probably only be able to do single sided boards, and I absolutely couldn't get any SMD packages under 0805 to work reliably. So I've done this before, figured I'd write up my experiences with trying this.

Cimatron import files