After crawling the forums for a while, I’ve found several different topics that have diverted to Capacitive talk. Since I believe this may be a viable alternative, started new topic. You can find the other discussions here:

http://nuigroup.com/forums/viewthread/415/
http://nuigroup.com/forums/viewthread/451/P45/

---------------------------------------------------------

As I posted in these two other topics, here is a purposed design of the iPhone’s capacitance system:

http://img225.imageshack.us/img225/6287/iphone9c3a840ctb8.jpg

---------------------------------------------------------

My thought is:

---------------------
LCD Display Layer
---------------------
LCD Backlight
---------------------
Capacitive Layer
---------------------

Unfortunately, I do not know what the minimum effective distance of a DiY capacitive layer might be (haven’t built one yet). The above would probably be the cheapest since the optical properties of the capacitive layer do not need to be taken into account. However, if the minimum distance from finger to capacitive makes it impossible, then the project becomes infinitely more complicated. I’m not sure why HowStuffWorks thinks that the capacitive layer is between the user and the LCD, it may or may-not be so.

I’m hoping to bring together several experts in on this in the next few days/weeks (whenever I find the time). I am a student at the University of North Texas (US) and our electrical engineering department is always looking for new projects.

Keep in mind that the reason most people in the group are using IR and cameras is due to cost, size, and DIY. Capacitive sensing is really only practical up to a certain size, and requires a lot more knowledge and work. Keep in mind that recognizing objects (like fiducials) also is not possible with a capacitive sensing technique. Since many of us are designing large scale screens, many times greater than the size of an iphone or lemur, it makes sense to use IR + cameras. It would cost an incredible amount (if it’s even possible) to make a capacitive sensing screen that is the size of an average table.

For those that are making small devices, this could be useful if a cost efficient DIY is possible. We can always use more thoughts on alternatives, so thanks for starting the topic.

The idea to make capacitive sensing is to add multitouch features to LCD-screen, not projected. Of course in projected display it is very efficient to use cameras: The space needed to camera is already there.

My point is to make flat - really flat screen which position and angle can be adjusted. Also when talking of graphical design etc. it is important to have good image quality. Now when I am playing with this 15” screen the main goal is after all to make 30” LCD multitouch screen with some other features as well:
-some sort of plug and play: The screen is only needed to plug and install the driver. That’s it.
-the screen can be used with hands and maybe also with pen (like wacom)

I have watched several videos the last few days on capacitive sensing. It has been successfully used in several large scale application (too tired to find the links now, will post later). I agree that it would be very difficult, however my hopes are to create a multi-touch device approx. 22” diagonally and 3"-4" thick. To this end, IR+Camera+Proj are overly large. I do not want a big table, I would like a large laptop/wall-display.

I am still experimenting with how to create a capacitive screen that is both powerful enough to sit behind an LCD (not between it’s layers), and sensitive enough to enable a more efficient finger locale definition than IR+Camera can provide. Additionally, capacitive technology, although expensive, can have separate hardware for locale definition and therefore should be less inclined to be CPU intensive.

Of course, I am a newb, and your opinion matters greatly. I’m still looking into the technology and the feasibility of doing this. The above is simply my understanding of the advantages of using this more complicated system.

EDIT:

@Kamelisko

That was might thought as well (as seen in other thread). The problem is that you can increase the voltage across the wires (increase the electro/magnetic field around them) but this also causes some more problems with self-inductance and cross-capacitance with other wires that you do not want to be capacitors. This is how I envision the capacitive system working:

- Vertical lines of wires, sheet of plastic, horizontal lines of wires.
- Each wire has a sinusoidal electric current passing though it (all the same, aka in-phase to cause as little cross-interference as possible).
- A control system to measure the capacitance between pairs of wires.
- The control system can measure a change in capacitance between wires in the horizontal plane. Same/Another control system to measure the same in the vertical plane.
- By finding the greatest change in both of these planes, the system can interpret it into x,y coordinates.

We need to do some testing of how powerful we can make this (for max range) without causing problems. I do not have the setup at the moment, I doubt it will be anytime REALLY soon, but until then I recommend we do more research on how this system has been applied in other techs.

Prolly should have posted this in the new thread, but whatever.

Ok. I got a few new ideas for this thing.

Maybe it’s possible to do capasitive “camera”.
In this link (which I posted earlier in other thread) seems to be used 10mm glass. So maybe the 5mm to 8mm gap is not too much for the sensing of touch. okay… my idea:

Would it work if you take some aluminuim foil and make small holes to it let’s say gap as 8mm. Something like the attached picture.

Then to do every of the holes as individual capasitive sensor described in the previous link.

The actual problem is - like before - the circuit to analyze the touches… Let there be someone who can help us with this…

Tell me what you think.

Great idea! I found this after following some links on the page you gave.

Unfortunately the download doesn’t work, but the idea is to use silver pens and copper tape to make a PCB substitute. Would be a cheap way to test some of our ideas, just time consuming. However, one can use the silver/copper prototype all the way through to the finished product, making any changes neccessary along the way, and once the computer is completed, finalize it with a real PCB.

Also, not sure if the article will work the way it needs to for our design. The system proposed in your link depends on having fixed positions for buttons, however on the computer, the buttons needed could be anywhere. The electronics to decifer that would be completely different. We need to have some kind of triangulation system, whereas this just has gate logic for whatever button is pushed, not How close is it to these three different sensing pads? and that could cause problems.

But as far as the circular parallel plate capacitors, I think they are perfect. However, I believe the depth on most LCDs is going to be in the range of 15-20mm, and that is a large gap to jump compared to 10mm. We may have to slim the LCDs acrylic backlighting sheet down to just a few mm and see how it affects the overall brightness.

I am still trying to get a meeting discuss with come faculty in the Electrical Engineering and Computer Science departments here at UNT to discuss how to setup the circuits for our system. I hope they will help us get a jumpstart on what components are necessary to do the comparisons we need.

My idea for the surface is similar, however I think it will enable us to get higher density and provide a better medium for any triangulation technique.

The examples are using sensors from Cypress.  Qprox also makes capicitance sensors that are easy to integrate with microcontrollers (I say this based on experience)

However, these sensors work great when you have only a few buttons like a keypad, or a slider or wheel (like ipod).  The challenge with sensor based technology is how can you increase the resolution AND speed for multi-touch? You’ll prolly need to interface the sensors with a fast and powerful processor like ARM or Intel or something (??).  Someone posted a video-clip of a multitouch laptop implementation from microsoft on the forums.... if you search for and try to determine what sensors they used, it might give direction to your efforts.

in other developments...... mistibushi unveiled their research in embedding tiny light sensors directly with LCDs.... so they had a pixel for RGB and a pixel that was actually an input.... sorta like those projects that use LED’s as touch sensors.... while their engineering is beyond DIY, its not inconcievable to imagine in the near future being able to buy an LCD screen that has 3 output pixels for color, and one input pixel for multitouch!

@nima

Since you have experience with using Qprox’s capacitive sensors, you may be the most experienced person in this topic.

How difficult would it be to program a microcontroller to triangulate the touch based upon the different buttons? My hope is to have an array of small capacitive buttons packed closely enough to give a fairly high resolution output. By using a microcontroller to measure the change in capacitance across all of the capacitors, the change in just a few can be measured, find the mean of the change, and output that as an x,y coord. (Ref. prev. posted pic)

Perhaps there is a more simple way as well.

- Wire the capacitive buttons in parallel with each other in three different directions (horiz, declining, inclining) [Red=(+), Blue=(-)].
- Place gate interrupts between each capacitor on each of these lines [Black].
- Microcontroller A to measure capacitance of each line [Green].
- Microcontroller B to time the gates for MC_A to measure [Orange].

Use MC_B to time the gates so that in one cycle, all of the capacitors are parallel along the horizontal. Use MC_A to measure the capacitance of each horizontal line. If one or more lines have different capacitance than others, that is one Y coordinate. Next cycle, all of the capacitors are parallel in the decline. Same procedure. Next cycle, all of the capacitors are parallel in the incline. Same procedure. One should be able to figure out the location of the touch, and with fewer calculations that comparing the capacitance of each button to each other.

I found interesting material of capasitive touch screens. They claim that with these things it would be possible to do capasitive multi-touch.

CapSense:
http://www.cypress.com/capsense/
go to Applications -> Touch screens

Here is a link to seminar (or webinar as they name it) which presents some info of making touch screens. I haven’t watched it yet. Needs registration.
http://www.techonline.com/learning/livewebinar/1100001

Looked at their presentation, quite nice but they can only support “semi-multi touch” i.e one finger + the “presence” of another finger…
Does anybody know if Apple used projected capacitance technology? Does that mean that they can only sense one fingers X,Y and the presence of another finger?

Apple probably doesn’t use semi-multitouch, since you can scale an image by pinching, which requires detecting the motion of a second finger, not just its presence.