Yet another Arduino clock thread!

First step towards fixing the light sensitivity problem. I've procured some IR filter material - this is ~3mm thick apparently black plastic which appears opaque to visible light but has no effect on the operation of the opto-interrupter where the peak emitter wavelength is 900nm. I've cut a small rectangle of this and tacked it to the sensor sides of the optos with a drop of superglue. Started another run yesterday afternoon, early indications are quite good but another cloudy morning here (1st June too...) so no sunshine. Next step if more is needed is to dissect the sensors as a couple of people have suggested, and mount the two halves further apart with collimator tunnels as well as a filter. If that fails then Plan C is modified electronics for the optos with AC coupling to reject ambient light signals. Watch this space.

Update. The filter material greatly reduces the effect though doesn't eliminate it entirely. Since the clock will eventually be mounted in a case with a glass door, I have decided to stick with the filter and the current sensors and use an infra-red absorbing film on the glass to filter out the wavelength of ambient light that the opto device uses. I have found a 3M product that looks ideal and should be easy to apply. So the filter on the sensor only passes the IR, and ambient IR will be greatly reduced by the film. Should work well enough I hope, we will see.