How can I sense the touch of a hand through a thick surface? I'm working on a device that has to be completely enclosed in a 1-2" thick wooden casing (or 1-3mm thick metal) due to design requirements, so no chance of pushbuttons.
I thought we could use capacitance touch sensing, but I don't know what will work.
Any ideas?
If you really need to detect touch through a metal surface, you could possibly do it using 4 insulated Piezoelectric pickups. There is a technique for detecting touch that cross references the frequency detected by 4 equidistant piezo pickups. It's a technique called SAW or surface acoustic wave. It's normally used for touch screens that are 3 dimensional or have surfaces that are prone to damage or deformation. It can detect location of a touch point, velocity of a touch (It can distinguish a tap from a press, to emulate a mouse click for instance) and it can detect a dragging or moving point.
If you wanted you could have a very simplified version to implement a few basic buttons or tap sensors. Instead of doing all the annoying frequency calculations, you could just position a few piezoelectric pickups where you want the actual buttons, then use a threshold detection to determine which has been pressed (the one with the largest signal peak). They are quite affective as they don't pickup any airborne vibration, so you don't have to worry about any loud noises setting it off, although dropping it on the floor would probably do it.
I've used the old Qprox QT110 (well, it's not really touch sense, but near proxy) to detect touch through a wood board 4 cm thick. I had to carefully tune Cx and Rs by trials, but the final result was good enough.
If the metal is non-magnetic, you can use reed relays inside and small magnet outside to operate controls.
(my answer migrated from electronics-exchange)
Through 2" thick wood: Also tough, but not impossible. I assume you mean distinguishing between a finger/hand touching the box, and a finger/hand placed 1mm away from the box.
If you can place one or more metal plates on the outside of the box — not providing a total covering of the entire box! — then there are a number of possibilities using electromagnetic fields. I can think of a few (no guarantee that either would work!). Put your thinking cap on, and imagine how the touch of a person's skin can alter an electromagnetic circuit network.
The biggest thing that can make a difference is whether you can bring a conductor out from the inside of a box, through the wall, and connected to earth ground, or to a metal plate on the outside of the box.
1) The resistance of a person's skin can close an electric circuit -- Have two semicircular plates, separated by a small gap. The impedance between the plates is a very large resistance in parallel with a small capacitance. If you bridge the gap between the plates with your skin, you drastically reduce that resistance. If you can somehow measure that resistance, you're done:
If you can connect wires through the box walls to the plates, that's easy, just use a voltage divider and a source of DC or AC voltage that is small (don't hurt the person).
If you can't connect wires through the box walls to the plates, you might still be able to distinguish touch vs. no touch, but it'll be tougher. Maybe have another pair of plates, this time on the inside of the box, on the other side of the box wall from the first pair of plates, but with a much larger gap. Measure the capacitance between the inner plates. There may be a large enough change in the capacitance between the two inner plates with / without a person bridging the gap on the outside, that you could distinguish. There are a number of techniques/chips that can do capacitive sensing. Cypress has CapSense, Atmel has QTouch, ST Micro has S-Touch, etc.
2) A person's body is a reasonably good antenna and will pick up noise. (Look at an oscilloscope probe when you touch the tip with your finger.) If you have a plate on the outside of the box, and you touch it, you transfer that noise to the plate's electric potential. Think of this case as a person "shaking" electric potential up and down. If you can sense this shaking relative to a reference potential, you've got a sensor. The problem here is that finding a reference potential is touch. You really need a connection to earth ground, since a box floating in the air doesn't have access to a fixed reference voltage. You might be able to sense the differential voltage between two plates... dunno though.
Just my two cents.
It should be possible to pick up the minimal bending of the metal with a strain gauge.
I've seen them register the minuscule bending of a cm think steel bar pushed lightly with a finger. Though I don't know what quality of strain gauge and amplifier you need to achieve that. In any case it will be rather expensive (~10$ up).
Detecting the change in frequency of an RF oscillator might work through wood, due to the change in capacitance caused by the proximity of a hand to the oscillator tank circuit. It won't work with a metal case, of course.
As Leon mentioned the RF oscillator will work. I would use a heterodyne oscillator which consists of a fixed frequency oscillator, a variable frequency oscillator and a non-linear mixer (such as a diode). Your body acts as a small capacitor to change the frequency of the variable oscillator. It is easier to measure the change in frequency of the mixed frequency.
This is the way a Theremin works. The Theremin is sensitive at distances of a foot or two and the antenna will work inside a wood case.
The Q-touch sensors may work too. Can you mill a feature in the wood so that the contact was a closer than the thickness of the wood? Q-touch sells springs made for this purpose. One end of the spring attaches to the PCB and the other end rests just below the top of the surface.
There's a relatively new material with the fancy name of "Quantum Tunnelling Composite"
http://en.wikipedia.org/wiki/Quantum_tunnelling_composite
It acts as resistor by default, but when pressure is applied, it transitions to a conductor. You can buy them in the form of even more strangely named "QTC Pills".
They're basically solid(y) state touch buttons.
(my answer migrated from electronics-exchange)
Through a 2" thick conductor? No feasible way, unless you have another electrode that the user is attached to or standing on. (in which case you could use an electrometer-type approach; that's kind of how those touch-activated lamps work)
Think about it:
People are essentially non-magnetic so that rules out a solution that uses magnetic fields/circuits.
If you're talking electric fields, a 2" thick conductor is millions (probably billions) of times more conductive than the human body. Any electric field you try to generate from inside the box, whether DC or AC, will be greatly attenuated by the time it gets outside, and the incremental effect of a human body touching or not touching it is insignificant. Capacitive effects are small (low picofarads) and if you could get the frequency high enough to measure it, you lose because of skin effect: a metal box attenuates AC fields more as you go higher in frequency.
If you try to sense environmental fields originating outside the box, (e.g. treat the box as an antenna, whose reception characteristics change with the addition of a human body touching it) you also lose because of skin effect attenuation.
Aside from electromagnetic fields, you have thermal: possible if you touch a box long enough to sense heating it up. Or radioactivity, or light, or....
