I've been playing with capsense again after seeing a Paul Stoffgren Youtube featuring capsense.
One aspect is touching bare metal that has a 1K or higher resistor between it and the input pin.
Is that enough?
I have a nice little setup on breadboard with one pin charging my cap sense jumper wire through 1M ohm while another pin digital reads it through 1K ohm "protection", and it works very fast, really well but the bare end of the jumper has to be touched.
If less than a vicious spark won't get through then hey, maybe some precaution could be taken otherwise it's pick different resistors and try to get better range. With 4.7M ohm I can tell if a finger is near the coated wire but there is no jump on touch, for instance. And there's always using a bigger piece of metal. It's just that trying to tune those can be a real lot of work where touching a bare wire gets very positive readings FAST.
If this 1K R is okay the net step is to use 1 pin to charge a whole port's worth of sensors and see if I can read them fast enough to get not touched from touched for the whole set. I might be able to make a "slider" with lots of wires, or not.
The Arduino can have 1mA of current pushed into an input pin. So with 1k resistor, the Arduino is protected for 1V, so from -1V to +6V.
You can use perhaps 10k, but that is still not a lot. Perhaps clamping diodes before the 1k resistor would help.
That tells me I need to work it out so bare metal is not touched.
That 1K resistor affects the circuit enough as it is.
The larger the conductive surface used to sense with, the more effect a finger will have. The "fun" part is telling close from touch when you don't actually touch the surface.
With 4.7M between the pin doing the charging and a jumper wire I can pick up a finger near the insulated wire but no jump when I touch the insulation. It's not a touch sensor, it's a 'close' sensor. But if I squeeze the wire and hold a bit, then I get a useful raise which is not what I want. So it looks like I have to go from wires to strips or something.
Note on the video that the outer part or those surfaces have tape over them, and they work well. I want something thicker with a printed insert under it.
One thing though. Knowing charge time is quicker untouched, a touch routine only has to time out a fraction more than that. It doesn't have to run the 2000-some a touch on the tape can take, only enough more than a finger being close will get, maybe half that.
All for 'buttons' with no moving parts that should take ages to wear out.
Well I can spend 5 to 10ms debouncing so maybe I should quit aiming for analog-read sensing times.
