I built one a couple of years ago -
I found the schematic I think on the large format forum. My EE friends helped pick out the components. I think it's generally accurate at slower speeds, but I don't have a way to independently test the device. At faster speeds the signal becomes stretched out and it's hard to pick specific on an off points. Lately I made a small slit mask that I put on the front of the box to try and limit extraneous signal and produce more abrupt on and off signals. It may have helped some, but is also let me measure the speed of a Zorki shutter at the beginning and end of the shutter travel so I could set the tension on the drums.