We're determined to come up with something that's cheap(er) and easier to make.
So the first thing to do was simplify: lose the leds and concentrate on making a large input matrix. Suddenly we were back to an earlier idea - copper tape and cardboard!
The basic idea is to create a membrane style matrix - criss-crossing lengths of copper tape and separating them with a "mesh" between the two layers (a layer with lots of larger holes cut into it). Then by applying pressure onto one surface, the two bands of copper tape touch and create a crude pushbutton. A description of this approach can be found here (http://forum.arduino.cc/index.php/topic,13091.0.html)
Unlike a push button matrix, we don't have a great amount of force pressing onto the contacts, only the weight of the playing piece, to activate the button. What we needed was a cheap way to pull the playing piece down onto the contacts - magnets of course! If we placed a ferric-based (steel/iron) board under our playing surface and embedded a small magnet in the bottom of each playing piece, there would be enough "pull" on the piece to drag it down, making the pushbutton contact. But A4 sheets of steel are not easy to come by, even in the BuildBrighton hackspace.
Now not everyone appreciates that the change in UK coinage in about 1992 meant one important thing - changing from a cupro-nickel to a copper-plated steel for one penny pieces means that any penny coin made after 1992 is attracted to a magnet. So we set about making a matrix of penny pieces:
The penny pieces sit slightly recessed from the top surface, so that when we glue another layer of copper strips onto this, there is a slight gap between the copper strip on the top layer and the piece piece below.
Note the slight gap between the top of the penny and the laser-cut board surrounding it - this will give us our deflection, allowing the top layer to flex, creating the button press, but - just as importantly - causing the top layer to return (restoring the air gap between the top and the penny) when released.
(of course, the keen-eyed will have noticed that we've drawn this top layer incorrectly - the copper traces on the top should be running the other way, at 90 degrees to the traces on the bottom layer)
With the top layer in place, we did manage to create a working keyboard matrix. But the amount of deflection was too great, even for a small, strong magnet. The buttons could be activated by pressing them with a finger, but placing a playing piece over the buttons yielded nothing!
So with a quick revision to our design, we got rid of all the pennies and used a sheet of cardboard as a separator, where the 2mm laser-cut mdf had been before. To make the base magnetically attractive, we simply stood the whole assembly on a small piece of galvanised steel, just to prove the concept worked:
This time, the results were much better!
Everywhere that we placed the magnet (when over a hole in the separator membrane layer) resulted in a positive connection. When the magnet was removed, the button returned to it's normally-open position.
We're going to play about with different membrane thicknesses; obviously, a thinner separator will result in a more responsive board to play on and will also mean we can use a slightly less powerful magnet. It may even be possible to swap the magnetic parts over - using an A4 magnetic sheet as the base layer, and penny coins or washers in the bases of the playing pieces. We'll have to wait until our next eBay order is delivered in a few days time......