An LED hooked up to the comparator output on the daughter board acts perfectly well-behaved, but when I write an ISR to blink the Arduino LED when the interrupt switches, it just goes crazy. The threshold voltage on the comparator is set to 2.6v and the output of the comparator swings between 4v and 200mv when the comparator trips. Again, it works when you measure it.when the dark wedges are between the LED and photodiode, the photodiode voltage is about 200mv, and when the clear wedges are letting the light beam hit the photodiode, the voltage is like 4v. For some reason the arduino pullup would only pull it up to 2v. I have them hooked up to an LM311P comparator, with the emitter grounded and the collector running to +5v with a 1k pullup resistor and to the Arduino. Right now I have a IR LED and a photodiode on opposite sides of a litho film 'pie slice' encoder wheel.

I would have suspected bounce when the encoders switch, and the long run of fairly high-impedence down the cord from the arduino to the encoder, that's why I added a comparator to the encoders to 'drive' the cord to the arduino. It must be some form of bounce or interference from the PWM, the motors, the cords, or something. It always works when you measure it statically.on will be 5v and off will be <1v.but when I try to hook it up to the arduino pin, the interrupt trips as if it's counting diode noise or just static. I have tried commercial reflective IR sensor/led packages bounced off a white card attached to the motor shaft, I have tried a reed switch and a magnet that spins around once per revolution, and I have tried multiple IR break-beam approaches. I've literally been working for weeks trying to get an encoder to work. The motor is hooked up to the arduino through a monophonic audio cord and 3.5mm jack/plug as a connector and the encoder is hooked up using a separate stereo audio cord and plug as a connector. I'm not sure if it's electrical interference with the PWM or what.

When tested with a voltmeter, the encoders always seem to work, however, when attempting to actually use them hooked up to the interrupt pin, they fail epicly. Everything is just completely noisy and impossible to debounce. I want to encode its rotation, so I have tried a string of different encoding schemes and nothing works. I have a gearhead motor about 2 feet from my arduino, that I'm driving with PWM. I have this application that is driving me bonkers.

Watch me bounce and touch the ceiling