So, I've been asked to make a GAL-like PCB that can be 'manually programmed' without need of soldering for an educational setting.
Meaning, I have signals coming in ('horizontally') that can be linked up to a wired AND ('vertically') by means of some connection, those ANDs then being linked up by an OR to go into the input of a Flipflop.
Thing is, the requirements are that I'd need to have at LEAST 5 inputs (inverted and non-inverted), 5 Flipflops, and all the flipflop outputs returned normal and inverted, with 5-10 AND-paths. Even if we assume the minimum of 5 on each, that would require
5(ff) * 5(and-paths per ff) * ( 10(input) + 10(ff-output) ) = 500 connection points per board
If I use any connector whatsoever - even just pinheaders, that makes it hard to produce and expensive, which both is not ideal.
Ideally, the connectors would be 'implemented' only on the PCB, with the connection bit (the 'jumper' if you so will) being the only thing that actually costs money per unit.
I've been thinking about solutions like having small screws (like M2 or something?) with nuts to connect the 'grid', on a PCB that is not throughplated, but that would be a pain for the users to assemble and disassemble, and would be fairly expensive at around 10-20 cents a connection, checking supplier prices for M2 screws.
Another thing I've been thinking about were those little earring needles with the retention clip - they can be bought cheaply from that one marketplace with an "express" attached to it, and would be easily affixable to the board. Problem would be the connection, as it is not spring loaded, however you could affix some conductive rubber, as is used in for example, controller buttons to it, to provide a 'natural spring' to it, holding it together.
I've also thought about just putting the pads next to each other, like in a controller button, then electrically connecting them with one of those conductive pen inks or electrically conductive stickers - that might work, but the cleanup would be a problem.
This would be a series of maybe 100-200pcs, maybe repeated every 2 years, so not extremely high volume.
Any other bright ideas how to implement an ultra-low cost connector that is implemented fully by just the connector that closes the connection, that's not one-way?