SMD circuits

From Open Source Ecology
Jump to: navigation, search

From Leo x:

My experience with a $7000 CNC router indicates that milling high performance PCBs will be hard. And, as you say, very slow. If we want double sided through hole PCBs with .1" pitch components with teens of pins, no problem at all. But as soon as we want multilayer with vias and 0.5mm pitch SMD QFPs let alone BGAs (all required to make the most of FPGAs) then that's a totally different matter. Boards for these components are right at the edge of what is possible with any router: 0.12mm track spacing and 0.1mm tracks. The smallest mills I can find are 0.1mm, and that doesn't leave much room for error.

Multilayer boards means: manufacturing multiple 2 layer boards to close tolerance (by whatever means), bonding them together with close position tolerance, drilling vias (again close tolerance), plating the vias.

SMD further requires: masking the pads (close tolerance), lacquering the boards, cleaning them, tinning/silvering/hot-air-solder-levelling/whatever the pads, making a paste stencil, applying the paste, pick n place, and then the thermally controlled soldering cycle.

None of this is impossible or even _that_ hard in isolation, but it adds up. It's a lot more than three stages :-). I'd like us to start with reasonable expectations and useful short-medium term goals, and somehow build up to the complete type 3 system in due course.

In this case I'll leave the conventional PCB work to others for the moment. I'm going to pursue complete circuit printing + pick and place for the RepRap.