A work in progress
Some time ago, I came up with a low-friction bearing using LegoŽ train magnets. The shaft of this anemometer (an instrument that measures wind speed) floats on the magnetic field between two magnets with like poles turned together.

The shaft is held upright in two places, and balanced with a large wheel. The result is a shaft that turns smoothly and freely with little friction, and needs only a small initial force to begin turning.

I attached three panels to the top to act as vanes, and they will catch even the smallest breeze and begin to turn. The challenge I face now is finding a way to measure the speed.

My first thought was to use the optical sensor from the RIS. I built a small housing that fit around the edge of the flywheel without touching it. I then made a black mark at a single point on the wheel. My idea was that the optical sensor could measure rotations by sensing the passing mark, and that I might be able to convert that information into windspeed.

In practice, the sensor was unable to "see" the mark at high speeds. Some more experimentation on the width of the mark might give better results.

I put the project away for awhile, until I recently got a LegoŽ Speed Computer, which has a rotation sensor that will convert rotations into mph. This seemed the perfect solution at first, and I rebuilt the anemometer using the speed sensor as one of the upright supports.

However I found the speed sensor greatly increased the friction on the shaft, and the initial force needed to start the vanes turning went up as a result. It would be possible to calibrate the sensor to account for the friction, but my estimate is that it takes a wind of over 10mph to set the vanes turning, so the use would be limited.

I am interested in refining this instrument, and welcome any ideas other users might have.