Hey all,
I'm working on my first microcontroller project, with the AT90USBKEY demonstration board from Atmel. I want to build a USB keyboard emulator whose buttons are four touch sensitive panels. The user will be using their feet and jumping on these panels (playing Dance Dance Revolution). My skills lie firmly in electronics, not woodworking or materials skills, so the simplest option (Building mechanical switches for the panels) is out for me. Anyway, then I wouldn't get to play with a microcontroller!
At the moment, I have one of my IO pins hooked up through a couple of metres of ribbon cable to a 20cm*20cm single-sided blank fibreglass board, which is my prototype touch sensor. My microcontroller switches that pin to an input and enables the pull up resistor. This slowly charges the pad up. I keep sampling the (digital) input pin until it reaches the HIGH level. Based on how long the pad took to charge, I can distinguish between a hand on the traces of the pad and nothing touching the pad. When I'm done, I switch the pin to an output with a low logic level, which discharges the pin.
However, the timing difference between these two states is not very large. I probably don't have any safety margin to correct for a dirty sensor, for instance. If my touch pad acted like a variable capacitor to ground, I thought I could add a resistor between the microcontroller and the pad to increase the time taken to charge. I added a 4K7 resistor (After the ribbon cable), and the pad did take a little longer to charge, but the difference between the "Hand on" and "Hand off" state was smaller, if anything. I tried a 10M resistor, and now putting a hand on the sensor makes no difference to the timing. So obviously my "Pad as a capacitor" idea isn't quite right. What am I missing?
Cheers, Nicholas Sherlock