I’ve been toying with the idea of designing / building a multi-touch table for a while; and recently an application idea came to mind that has re-kindled the dream of building an FTIR touch surface with an LCD behind it. I’m going to explain the physical build of the box, and included are some images of a 3D model so you can see the overview.. I’ll also post the design elements seperately to further illustrate how the pieces are supposed to fit… This isn’t a completely finished design, but the area I am concerned about is shown.
To give some background on why I have chosen this design before you take a proper look at it; I have a large glass coffee table in my lounge, it’s got a very elegant metal structure beneath it with lots of free space; I tried to model the frame holding it up, but couldn’t get the curves right; needless to say, I’d like to turn this table into a multi-touch coffee table. I thought about using the original glass, but I can’t see it working without me cutting a hole in the middle of the glass and I’d rather save the piece so if I ever wish to turn it back into a normal coffee table, I don’t have to hunt for another large piece of bevelled toughened glass (Plus, I would probably smash the piece trying to cut out a 22” diagonal rectangle from the middle...)
My main concern was that the touchable surface will be in the center of a much larger piece of acrylic, and I’d obviously like the two to be as flush as possible. As such I have designed a central box to house the 22” LCD matrix, two Microsoft NX-6000 Cameras I read about on here with a 71 degree viewing angle which will allow me to cover the full touch surface with a box just over 8 inches deep (This obviously can be adjusted if I find more suitable imaging hardware, I’m looking into a few different multi camera setups to provide a capture resolution close, if not greater than that of the currently chosen LCD panel). Lighting, again, I’m being experimental by adding boards to the box internals, as high as they can be without obscuring the cameras view; in the new trenches i will add the illumination sources and give them reflective sides, above the height of the partitions, I intend to coat with a light absorbant material; most likely a matte black paint (blackboard paint potentially), I am also looking into adding fresnels to the top of these channels to disperse the light out over the whole panel as evently as possible. I’d love some input on this concept, is it workable? has anything like this been tried before? if so, what were the results? (Obviously each lighting area will have it’s own ventilation. Back to the top again; The frame housing the IR LED’s is 1/8” lower than the height of the Acrylic touch surface. The larger piece of acrylic will be cut down and have a lip to allow the two pieces of acrylic to sit flush along side each other. I am wondering what sort of effect this will have on the FTIR concept; I’m sure there will be some IR leakage, but will it be significant enough to screw up touch recognition?
I would obviously like to bond the two pieces of acrylic together; possibly permanently; using some form of plastic cement or other process to give a completely flush coffee table surface, with a touchable inner screen. I am thinking about some form of IR reflector to place between the two pieces with the tiniest ammount of Acrylic left at the surface purely to bond together and polish smooth. Can anyone shed any light on whether this is even worth trying?
So here’s the top sheet, as you can see, the inner window has a lip of around half an inch which will sit over the frame for the touch window:
Here’s the framed acrylic; perhaps a few too many LED holes, but at this stage I just wanted some holes in there:
Here’s a top down view of the main housing, the layer on top is shown as a piece of glass to represent the LCD matrix, further refinements need to be made to finalise this setup; also please note that the diagonals you can see inside from the two points were me working out the viewing areas for the two cameras, although it doesn’t appear so; the camera fields actually overlap by approximately 20 pixels, sketchup just isn’t accurate enough:
And now the finished item; note it is missing the frame to support the whole structure, but I’ve got that item sat in front of me here, I might take a picture with my cam if anybody is interested :
You can also see the ventilation holes for the lighting channels; these will obviously be fanned and baffled in the final design to improve airflow and reduce light leakage respectively
Let me know what you think.