AVR-Chat

>
>So you see, the biggest problem in accuracy is to positionning exactly the
>sensors.
>Far distance between the sensors, better accuracy.

Matching the optos will also be a problem. They aren't known for speed 
either, though that's solveable.
Breaking tensioned wires might be a lot easier and more repeatable.
IIRC the earliest chronographs used that..

Of course you could always grab a microwave burglar alarm and process the 
doppler ouput.
They are designed for <1MPH but that's all in the processing electronics, 
just take the homodyne mixer output directly.
I used to know the formula for doppler shift offhand, the typical alarm 
runs around 10.5 GHz