Oh, Hi! I didn't see you come in...

So yeah, about my MT table...

As you can guess by my long absence on this blog, I dropped that project a while ago, mostly a question of time and motivation. But I did not log my final progress so here it is :

* I found a bandpass filter for my PS3-eye, namely 2 layers of dark photo negatives. It's easy to cut, it doesn't cost anything and it works quite well.

Apart from that, I was working on a small prototype of my table, a 28cm x 18cm screen. The goal was to check the feasibility of the real deal, and to verify if my schematics for the frame (which contains all the electronics) were correct. The results :

* It took me forever, but I finished the circuitry. 65 LEDs, 8 resistors and 2 switches. The idea was :

• 6 rows of 10 IR LEDs (plus a resistor) in the frame, for the FTIR effect,
• 4 IR LEDs on the back of the screen (one in each corner), directed toward the camera, to allow automatic calibration. The camera would see the 4 corners and a program would deduce where the screen is in the video feed to compute the coordinates of the touch points. Plus a small switch to turn that off.
• a final green LED to indicate the table is turned on, and a general switch.

20 LEDs in place in a part of the frame

Problems were :

• Soldering LEDs one to another is not that easy. I think it requires 4 hands, or some tools that I did not have. Anyway it took me forever. Furthermore, the risks of burning / breaking a LED are high.
• A string of 10 LEDs welded together is much more fragile that I expected. I would often put them in place, then for some reason have to remove them and break them in the process.
• In a string of LEDs, if one stops working they all stop. After that you're forced to test each diode one by one, find the one that does not work anymore, unweld it and weld its replacement all of this without breaking anything else.

My conclusion is : do not try to do the whole electronics yourself, unless you're really patient. I had only 60 diodes for the prototype, but I was thinking of around 400 LEDs for the final project. It would have been a nightmare. That's basically the reason why I abandoned the project. So LED strips are definitely something to consider if you want to build a FTIR or a DSI table.

* I finally finished the electronics, put the glass in and closed the frame, and tested it with the camera. It worked! But I didn't seal correctly the frame and I had some IR leakage, which created some kind of LED-LP effect. When I opened it up to correct that, half the LEDs stopped working. I did not have the courage (nor the time) to find out why, so I postponed the project to when I would be motivated again. And that never happened.

* Oh and I also bought and tested a teflon coating for the screen, to replace the compliant surface (an idea I got from here).

However it never really dried and left an oily substance on the surface, not what I wanted.


What I learned from the project :

• The small prototype was a good idea, it didn't cost much and allowed me to consider a a few problems I did not initially think of (mostly about volume optimization inside the frame)
• If I was ever to start again, I would not contemplate the idea of soldering hundreds of components together ; I would go for commercial solutions (I'm talking about LED strips here, not the whole table), even if they are more expensive.
• 'twas fun

So now that I am a bit more motivated to write for this blog, and that I don't have that interesting project to talk about, I'll try to write about User Interfaces, and Usability (my new interests), mostly in multitouch interfaces. And whenever I'll see something of interest about MT tables or other tech-related subjects, I'll notify you.

A+, comme on dit en France.