The basic process was:

• Create a base out of plywood.
• Create a bunch of pads, out of lucite, carpet pads, hardened paper board, steel sheets and/or copper wire.
• Wire it all up, screw it all down.
• Try it out and realize it isn't quite good enough, dismantle it, and rebuild, several times...

This has been done throughout the later part of September 2002. I keep putting off version 3, although I had hoped to have it complete in early October 2003, I didn't work on it until May 2004. Mark 4 and 5 were worked on around October 2004.

In case you can't tell, click on any of the images for a higher-res version.

Early on, I intended to cover each of these with aluminum flashing, wrapped around the edges, and screwed in place.

Here is another image of the base, with it all screwed in place. The image is a little blurry....

Later on, I found that my table saw worked pretty well, but long after the worst of it was done. The worst that I am refering to, was not this, it's later...

This is an early picture where I removed the control circuitry. As is mentioned elsewhere on the web, the little black tabs you can barely make out on the top of the circuitboard are conductive, but not soldierable..., you have to attach wires elsewhere. More on this later.

• nice soft plastic blue bottom
• a layer of thin plastic with wire traces on the surface
• a very thin layer of foam, with holes evenly spaced across the surface
• a layer of thin plastic with wire traces on the bottom
• a piece of carpet protector(!)
• nice soft plastic with silk-screened images on it.

In the first picture, you can see the a potion of the top layer plastic, and the carpet protector. I suspect it is there to add weight to reduce slipping and to make it seem more solid. The pile of green stuff nearby is the ribbon that ran around the edge of the soft mat.

The second picture (blurry, again, sorry!) is a picture of one of the inner layers with the circuit traces on it. You can clearly (heh) see the places where the buttons go.

Of course, each square is again 11" square.

In the center of the pad is a brown square of paperboard (or something like that...), the same material you typically find on the back of do-it-yourself shelves (although this is more dense).

That brown square is part of switch. There are two of these squares, one set above the other, with copper wire wrapped around them. Wires run from those squares to the controller circuitry, shown earlier. The top square is kept off the bottom one by resting on top of the carpet padding.

Perhaps an image will help.

In the left side of the image, you will notice a piece of the original soft mat. I decided to use the images off the old pad to improve the look on my new one. More images further down.

I was probably daydreaming, thinking about what it will look like....

The silver disks in the image are two sets of washers. The larger washers had too large a hole, which would let the drywall screw heads slip through. The smaller washer on top holds the larger one on.

At this point, the wiring is run off the board, but isn't acutally hooked to anything. I wanted to do some testing before committing myself to this layout.

Looks pretty good, doesn't it!?

I hooked it up, tried it out, and it sorta worked. There were several problems, though.

The first was that the pads were squishy, they sunk rather than clicked. Not a good feel. The second was that the solder joints sucked. The solder between the copper wire wrapper around the paper board and the wires that ran off the board, pictured above in red and white were horrible, and tended to break. (I thought it was working okay when I tried it, but when I dismantled it some were clearly broken.)

So, after a little trial and error, I went back to the drawing board. I had been discussing what I was doing with several friends, and came up with a better alternative...

Shortly a third piece of steel would be placed above the two plates. When pressure is applied the top plate touches both bottom plates and conducts electricty between them.

This is a much better view of the two lower plates.

Under the corners is a layer of paper board and perhaps a thin additional spacer. The top steel sheet, two layers of lucite and the graphic are layered above that and resting on the corners.

At this point I think I was still cutting holes but not yet wireing anything.

Here is an ever so slightly closer version.

Wires were a little bit of a problem, trying to run them along the edges of the pads... But I managed. Unfortunately, one of the pads was raised up a bit higher and never quite sat right. It was the first pad I assembled. It is the down arrow (which is the one on the right hand side in the picture).

Another closeup of one of the pads, the up button, in fact.

Another shot of the whole pad with slightly different lighting.

This is a picture of my recent new connection design. I've run wires up from the bottom of the board, where one will connect to the top metal plate, and the other will connect to the bottom. I didn't want to recut all the bottom plates, so I mostly just wired them together (to form a single virtual plate).

The two bent metal clips are connectors that connect to the screws.

The left clip slides between the metal plate and the paperboard, and wraps around underneath the paperboard. It has some slip, which should be okay, as long as it is tight to the screw.

The right clip slides between the plastic image and the top metal plate. It too has some slip, and I think is at more risk of slipping. I also had to worry about bumping against the lower metal plate, but in several cases I taped the bottom plate with electrical tape.

I stappled the wires down... and on like the last staple pierced one of the wires... fuck. Fortunately it conducts just fine!

The pad tends to work better still, than all the others, but the left pad is weak, sometimes doesn't respond, sometimes bounces... It must be the metal tabs aren't making very good contact. So, on to Mark 4!

This worked very very badly... and it didn't take long before I moved on to Mark 5.

The absolute best thing about this design is that it is much easier to take a square apart and put it back together. I can take five minutes, remove two of the corners, rearrange stuff, and put it back together, and know that it is all positioned correctly.

As you can see from the springs sitting in the middle, I was going to use springs to keep the lucite/metal arrows away from the electrodes. I ended up using some metal strip springs from those do-it-yourself metal frames. I don't have any pictures of them, but they are about 4 inches long, and are bowed (sorta like a bell curve).

I used two strip springs per arrow. I wedged one end of the strip springs under the paperboard, to keep it from sliding around. I ran into a problem where the strip springs would move around and accidentally come into contact with the electrodes, shorting the arrow out...

That isn't to say there aren't problems...

One problem was that the pad tended to bow, so I put a square of paperboard in the middle of each pad, roughly the size of the square made by the 4 springs in the previous picture (about 3 inches square). This helped, although I should replace it with something slightly larger, and perhaps slightly thicker.

The one remaining problem is the noise. It is loud. My next design may be to replace the electrodes with microswitches. I've purchased a bunch, but it will require a thicker board, and so will probably need to be a new mat, built from scratch.