Hi all. I’ve been surfing around the web, your forums, and youtube for a few days now trying to get some information on general multitouch tech. I’d like to build such a table fairly cheaply with the bits I have.

Currently, Im looking at an FTIR, LCD method. I’ve tried to build a cheap and cheerful mockup of the ultimate design here:

This seemed promising but I’ve since hit troubles. I have added a wide-angle lens to the camera (a butchered QuickCam) and added a piece of floppy-disk as a filter (I believe it was Peau Productions where I saw this). Touching the acylic results in nothing… no blobs at all. I suspect what was actually happening (in my video) was some kind of DI effect.

Anyways, after reading up on all this I have a few questions. I’d like to take apart my Dell 17” panel and use this as a basis for a table. Ultimately, the end goal is to have the table recognise objects, but for now, touches are more than fine. I’d like to know what the best method is?

I’ve seen this link http://nuigroup.com/forums/viewthread/4309/ but I’m not sure how this can actually work. With FTIR, most people seem to use a compliant surface so this must be some special kind of Acrylic. I notice people talk about Endlighten and I’ve seen it used in various projects but Im guessing this is for DI only projects right?

So, how should I proceed with an LCD panel being the major factor? I intend to get a PS3 Eyetoy, add the floppy disk filter and use a set of Osram LEDs OSRAM at RS.

Finally then, some assumptions that I’m not sure about and would love your feedback:

1) Some kind of compliant surface using a form of silicone (I’d rather like to avoid this! )
2) 1 LED per Inch on all 4 sides
3) Reasonably tall box to take into account the field of view of the eyetoy
4) Floppy disk filter
5) Backlight at the bottom of the box
6) Mirror foil inside the box walls to reflect all light
7) 8mm standard Perspex on the top
8) 3mm standard perspex as a support.

Its getting the right materials and tolerances to get a good FTIR effect that im stumbling with.

Greetings,
I can only see three IR diodes - If they have an illumination angle of 30 degrees (That’s 15 degrees ether side of the centre line of the LED) then only the diode that is closest to the camera will have any effect.  (See attached 1LED.gif) The three beams will not merge until close to the center of the box. The IR light will travel to the far edge and will escape. If you move the camera more to the left in the picture, you might get better results.
As you are currently playing with 3 LEDS move one of the outer diodes to the top edge so that its beam crosses the camera at 90 degrees to the center diode and if you are able move the third so that it crosses the camera at 45 degrees to the other two (Say from the bottom). The idea is to flood the area above the camera with lots of IR. You could just rotate them so the center line crosses the center of the camera lens, but I think you will find a better result with the IR source coming from completely different directions. (With three LED’s the ideal would be at 120 degrees to each other)

When the light leaves the diode and enters the Perspex some will escape because of the angle but eventually because of the difference between the two mediums (Perspex and Air) it will reflect internally and bounce along the length of the Perspex.. If we added another layer of Perspex then the IR light would pass into that medium because they both have the same refractive index. If we made the Perspex have a different surface plane then the angle would reflect the IR beam at the angle of incidence. (See Refractionxx.gif s) This is the effect we are trying to recreate - by introducing a secondary medium so the angle of reflection changes - the light will reflect at an angle so the camera will pick up the reflected light.

The problem arises if there is still a layer of air between the touching surfaces - the light will continue on its original path. The touch surface requires that there is NO air gap to work correctly. The frequency of the IR light is absolutely minute and the grooves in our finger prints looks like the Grand Canyon in comparison, so it is only the peaks of our finger prints that are actually reflecting any IR light. You demonstrated this very nicely when you used Blue Tack. So- the more IR light we can get to reflect off the tips of the ridges of our fingers - the better this will work.

Colin

Hi,

nice testing setup and observation. I just curious.

I am not understand well 7 + 8 items, you have 8mm perspex shine up from the sides by LED, and then you have 3mm as support on the top of it. Is that right?  if it is true, would IR will be weaken see by the cam?

Is it your LED is already 120 degree angle? what kind?

Almost the end of the video, I can’t hear well, what do you put on the top to target very bright blobs.

Thanks

Hey thanks guys. Good tips allround.

I have tried using 14 LEDs on all four sides of the same perspex but still no luck. In fact, if anything it was worse.

Using the floppy disk filter I can’t see any effects of FTIR with just perspex and a finger. If what you are telling me is correct (and I certainly believe it is) then how is the guy in the other thread achieving blobs with no compliant surface? I need to know so I might bug him so more because I feel that creating a compliant surface is the most tricky part of any touch setup and I haven’t seen any touch setups using an LCD panel and a compliant surface.

Yes, the distance needs to be within the wavelength of the infrared to make FTIR work. I suspect the effect in my video is some kind of reflection of IR (it probably catches some off the top of the LED).

Im using bog standard Perspex. Some people talk about Endlighten perspex but im guessing this works with methods that are not FTIR.

Hi,

What wavelength are your IR LEDS? 850nm is the the best wavelength to use based on the sensitivity of the majority of cameras.

Are you using acrylic sheet or is it polycarb or hard polystyrene? Only acrylic sheet will work for an FTIR effect as most other plastics will actually partialy block IR.

Have you removed the IR filter in your camera, this is somthing most people strugle to get their heads around.  The camera will have a small square of dichroic glass that blocks Some near and most far IR.  This filter will be sat somewhere in the lens assembly somtimes in the lens other times it is sat on the CCD. THE IR FILTER MUST BE REMOVED!!!

You will now replace this filter with another filter that does the reverse. Blocks the visable spectrum (wavelengths) so that all the camera can see is IR.

Now the floppy disk is alright but it doesnt block very well.  You are must better of buying a bandpass filter or BP for short.  These come in wavelength specific flavours ie 850nm.  Only letting the IR light for a specific wavelength through.

Also you could install more LEDS perhaps double them.