For prototyping, I've noticed stencils are quite low-cost nowadays, which is fantastic.
For those unfamiliar with the process, they are used to apply solder paste to the correct areas on a PCB prior to component placement and reflow soldering.
The basic principle can be seen in this screenshot from a sparkfun stencilling video:
In that video, the PCB is secured in the XY plane using scrap PCBs that are temporarily taped to the desk, and then a laser-cut mylar template (it has laser-cut apertures which correspond to the pads on the PCB where solder paste is needed) is held in place (aligned by eye) and then paste is applied using a squeegee.
So, that's the fundamentals of the process, but in practice it may need some refinement for certain use-cases.
I'd like to use low-cost stencils (actually steel sheet stencils are low-cost nowadays, the Sparkfun video is six years old) but for perhaps half a dozen boards or more; in other words not production volumes, but "a few to a couple of dozen prototypes" volumes, and then discarding the stencil. Furthermore, nowadays tiny packages are very common for hobbyists to use too, so clamping of the stencil by hand isn't feasible. Perhaps alignment even needs to be inspected under a magnifier/microscope unless there are registration holes or pegs that can be used to obtain some accuracy without having to align by eye each time. How much accuracy is needed? I think an error around 0.1mm may be a suitable limit, because there are components that have 0.5mm or 0.4mm pitch and an offset of 0.1mm could mean that 25% less solder paste is on the pad.
There are many other parts to the process that could be refined too, but for now I wanted to focus on the the alignment/clamping topic.
The commercial solution is a stencil printing machine; for low volume prototyping there are manual versions of such machines. The way they work is that the stencil has a metal frame around it, and this is clamped/fitted into the machine on a hinge. The frame is lowered to do the pasting, then the frame is raised on the hinge to take out the pasted PCB and insert the next PCB.
However, framed stencils cost more. It is possible to self-frame, but this requires a stencil with lots of space around it. The squeegee may be 100mm wide or more, but hobby PCBs are often smaller than 100mm. A low-cost stencil may only be 20mm wider all around.
Adding a custom frame to a low-cost stencil is therefore hard when the squeegee itself could be of a similar size to the stencil.
I tried to think of a way to solve this issue. The diagram here shows the current thought process.
Basically the PCB would fit on a raised platform that is attached to a base. The raised platform could be swapped out for different PCB projects or drilled with more peg positions or have straight edges for aligning the PCB into a repeatable XY plane position. The height of the raised platform would be fixed for a certain PCB thickness but could be modified with some shim if needed. The raised platform could be screwed to the base from the underside. The base and raised platform would probably be constructed from aluminium sheet.
The stencil sheet would be clamped on one edge between two blocks shown in blue. This could be achieved using screws or winged nuts. The precise alignment would come from loosening the screws, moving the sheet manually and then tightening it. If the stencil has 5mm holes then the clamping could be done with 3 or 4mm screws as an example, to provide for adjustment.
The underside of the lower clamp block would have a couple of pegs that would fit into the base. These would allow for repeatable accuracy with insertion and removal of the clamp from the base. After a board has been pasted, the stencil and base would be carefully lifted and put to one side while the PCB is removed and a new PCB is inserted. Then the stencil/clamp would be placed back onto the base, and the alignment should not have changed.
Although commercial screen printing machines use a hinged action for lifting the stencil, I'm wondering for small boards if this is needed or not; perhaps it is just as easy to lift up the stencil from the unclamped end slightly with a ruler for example, and then lift off the clamp from the base. I don't mind if I need to go carefully because I'm not doing production volumes; I can afford to take some extra minutes of care per board.
I already have some ideas about how I will go about this which I'll jot down as I experiment (and eventually write it up in a blog post so that it can be replicated), and I'm curious if others were to do this, what tips/suggestions would you have?
I'm still researching some materials and methods for assembly. For peg alignment I'd clamp the materials together and drill right through, so that the base and clamp are all drilled identically, unless I can find off-the-shelf pegs and sockets that I could incorporate into the design.
I might use bits of optical breadboards etc to save having to drill too many holes precisely.
Message was edited by: shabaz (Corrected the diagram to indicate the scrap PCBs to prevent damage to the stencil on edges)