Hey everyone,
I've continued to work on the Keyglove project, but I've run into a weird signal problem that I don't understand. I'm hoping someone here can shed some light on it for me.
Basically, for more rapid prototyping, I got a 2"x2" proto board from SparkFun and soldered two rows of 18 spring terminals on it, back to back with female header extensions in between them, like so:
My latest revision of the glove is using 28 AWG enameled magnet wire for sensor leads; these are attached at one end to each sensor on the glove, then the other ends are inserted into the spring terminals in the picture. It makes changing things around much simpler than when I was soldering male header pins onto every sensor lead for every revision and test--especially because I have labeled the spring terminal blocks.
However, I'm seeing all kinds of weird touch readings whenever I plug in the proto board. This actually happened before I plugged in any sensor lead wires as well, so I don't think it's the sensor wires (which are currently all bound tightly together for cleanliness--I wondered if this might cause interference or crosstalk or something weird).
As a quick review, the controller code detects touches by rapidly changing one pin in the combination to OUTPUT mode, bringing it LOW, then testing the voltage on the other pin. If it's also low, then it records the touch. Lastly, it sets it back to INPUT mode, brings it HIGH, and moves on to the next combination. All of this happens for all 60 combinations about 500 times per second.
The controller is supposed to register no touches when none of the sensors are touching (duh). However, when I plug in the proto board, it starts to register random ones at random times. Some of them seem to be pretty static (e.g. there are always five or six particular combinations that always show as touching), but others are random. The random effect seems to be worse when I am testing legitimate touches; if I touch one set together, it registers a dozen more with some randomness as long as I hold the first two together.
On top of this, it appears that if the Arduino board is disconnected and the proto board is not plugged in for a while, and then I plug the Arduino into the PC and the proto board into the Arduino, it measures touches correctly for a little while (maybe one minute if I'm lucky) before going back to the weird random readings--almost as though some interfering charge took time to build up while it was connected and in use. The proto connector board has no capacitors, and the glove doesn't either. It is only fabric and wire. I would expect flaws in my build process to be the same no matter how I'm connecting the sensor lead wires to the controller board, so I don't think it's that. If I connect the wires directly into the digital I/O header pin sockets on the Arduino board, the problem goes away. I was hoping to avoid this though, just because it takes more time and the wires don't stay put very well.
Does anyone have any ideas or recommendations? I'd really appreciate it.