multiple touch switch problem

Thread Starter

stimulus

Joined Jan 14, 2007
4
Hi, my first post here hope someone can help me out..

The circuit is simple: a touch connection current being amplified by a transistor to drive an LED.

I would like to have multiple independant touch controlled LED circuits (resistive not capacitive touch switches) Ie. two touch plates per LED, where the amount of pressure placed on the plates varies the brightness of the LED.

When there are two of these, such as in "Touch LED Circuit 1.jpg", the problem becomes that the two LED's can't be operated independantly of each other. (Because of the body being a common point).

Can remedy this by using two voltage sources, as in "Touch LED Circuit 2.jpg", but this seems like overkill, as I eventually wish to have about 10 of these LEDs.

So is it possible to use one voltage source and somehow make those currents independant when more than one touch switch is closed?
 

Attachments

beenthere

Joined Apr 20, 2004
15,819
Hi,

You have conflicting goals in your project. The circuits you present are basically flip-flops, where touching the appropriate contact will change the state, and therefore the LED that is on. These are on-off in operation. And, yes, a change to the basic circuit can let you operate only one led per circuit.

Witshing to add a brightness control in the form of a pressure sensor ghanges the nature of the circuit. I guess you intend that one of the two plates per LED will do the switching, while the other will vary the intensity.

The intensity input will have to be maintained over time for it to be useful. If the operator is not to have to maintain pressure on the control, then some means has to be cme up with to hold the input. It might be possible to do this using analog circuitry, but I suspect it would be difficult. A partially digital solution would work, but again at the cost of a certain ampunt of elaboration to the circuit.

I can immediately think of placing the voltage fron the force sensor to an LM3914. The ic normally drives an LED bargraph, but has a number of comparators internally. This would let it give an output that indicated the pressure on the sensor as a function of the voltage from it. The output from the 3914 could be used to latch on one of ten current sources to the LED, giving you ten levels of illumination.
 

Thread Starter

stimulus

Joined Jan 14, 2007
4
Thanks for reply. I should have given a better description, sorry!
I'm a musician developing a performance interface for dancers.
By pressure sensitivity I really meant contact surface area. So more contact -> less resistance
This also means that two dancers may touch each other to bridge a connection and vary resistance by how much contact they make.
I'm using large metal pads at the moment, but will be looking into other conductive material in the near future.
Also the metal pads are connected via fairly long wire to the circuit, meaning capacitance has been a problem. I seemed to have remedied this by placing 0.1uF caps between the transistor base and V+ and between base and V-

The LEDs are mounted inside small tubes facing LDRs, which are connected parallel to potentiometers of a MIDI interface (unfortunately dont have the time before the show to fully develop my own PIC-based interface, so have to make do with hacking a commercial one)

In my full circuit I'm also using the current driving the LEDs to further drive a transistor that turns an IR LED on with minimal touch contact. This IR LED is also mounted in a small tube facing an IR transistor which I use to close a MIDI note-on switch.

So with each pair of contact pads, touching them yields a Note-On with minimal touch, and further contact increases LED brightness/MIDI controller data.

I have a working prototype for one pair of pads using 9v DC, however to use more than one pair at the moment I have to use a seperate battery per pair.

Main problem is as follows with one voltage source:

skin contact is modelled as in my circuit diagram below, so a resistor between each contact point with a common centre point.

V+ is connected to Pad 1a and Pad 2a
Pad 1b is connected to base of T1 transistor
Pad 2b is connected to base of T2

Touching just Pad 1a and Pad 1b yields a certain resistance between V+ and T1 base. Now when Pad 2a and/or Pad 2b are touched, the current running between V+ and T1 base is also affected because of the common centre point in the skin resistor network.

With a seperate voltage source for each pad circuit, current isn't shared in this way.

How does one create this same independence with just one voltage source?

I could easily use a battery source for each pad, but I would rather use a 9vDC power supply to give a steady voltage.

hope this wasnt too long winded,any help is greatly appreciated!
 

Attachments

beenthere

Joined Apr 20, 2004
15,819
Hi,

Sorry to be dense, but I can't tell if this stuff is being worn or being danced on. Can you post more of your circuitry? I can't quite grasp the full picture.
 

Thread Starter

stimulus

Joined Jan 14, 2007
4
Sure, here is a schematic of my working prototype.. for some reason using the 2.2v LED indicator inversely helps the IR LED switch on in a cleaner manner.

the pads will be arranged in any number of ways, and using variety of metalic/conductive objects/material..

for testing purposes I'm standing on pads 1a / 2a, and touching pads 1b and 2b with my hands.

hope this clears things up
 

Attachments

beenthere

Joined Apr 20, 2004
15,819
Hi,

This is where I hope you're not too invested in the setup you have with the miltiple contacts to bridge to do the switching.

Check these people out - http://www.qprox.com/ - their chips may make your life easier. There's probably a way to get rid of the wires, too.
 

Thread Starter

stimulus

Joined Jan 14, 2007
4
thanks for you suggestion.. I'm keeping qprox in mind, although i really enjoy the human-human contact aspect. Do you think its possible to design a circuit based on qprox where an entire person could become a switch?
They say on their website that any small metal object can create a field.
How would one get rid of the wires?

Have reworked the circuit quite a bit. The basic interface circuit is still the same, ie.

V+ -> electrode1 -human(s)- electrode2 <- NPN transistor base

with long leads to the pads the transistor needs about 0.1uF capacitor between base and ground to keep it from being permanently on due to wire capacitance.

Eventually this signal switches on an optoisolator.
Then using cmos logic, the signal is processed to get rid of bouncing and the trail effect of the 0.1uF capacitor charging again after a note is triggered:

1. Input edge triggers 40ms Negative Pulse [4528 as retriggerable monostable]
2. Input AND Negative Pulse [4093 schmidt trigger nand]

0000011001010001111111110000 Input
1111100110000110011111111111 Negative Pulse (00)
0000000000000000011111110000 Output

3. Output to 4016 switch, closing switch on midi keyboard.

Works very well, only get the odd very short glitch note, too short to trigger the 4528 and thus be filtered out, but is easily filtered out with software.
 

Chartman

Joined Jan 28, 2007
3
Qprox works on proximity, not touch..... A simple touch switch can be made using a 2N7000 with a 10M pull up on the gate.The two touch pads are ground and the gate.Or you can invert it. 10M to gnd and touch between gate and vcc.
 
Top