First, I got a sturdy bicycle wheel from a local used-bicycle shop. The gentleman gave me two wheels for $10, and even threw in some nuts and washers. I already had some 3/8 inch plywood in the garage, so I decided to use that for my supports. I ripped down some 2x4's into 1.5 x 1.5 inch boards with my circular saw to use for structural components. Anything resembling that dimension should work, and use a table-saw for ripping if you own one! I also bolted two small pieces of plywood together to act as a support for the camera arm. A few hours later I have the initial components in place:
More views from the first day: from the top, from the side, and on the floor.
The basic idea is that I will take a picture once for every spoke on the wheel. That will give me regular intervals, and spokes are easy to detect because they have a small spoke-head on the outside of the rim. So, I got a small switch with a roller on the end from Fry's. It cost about $3.00.. As the switch rolls around the rim, it will be triggered when it rolls over a spoke-head. The switch will be connected to my parallel port, and I will write a simple program to read the parallel port and take appropriate action. Using my friend's electronics kit and a resistor he had, we wired the switch into place. We are using parallel port pin 0 to power the circuit, and we are reading pin 7. Here is the initial switch setup:
Another view of the switch electronics.
The wheel was not perfectly round, although it was close. The switch had to be perfectly positioned to allow the 1/16 inch tall spoke-heads to trigger it, so I had to mount the switch in such a way that it would conform to the wheel as it turned. My solution was to create a small sled out of a piece of thin sheet metal I got from Home Depot (about $6.00). This sled slid along the rim, and kept the switch from getting too close. I also mounted the whole sled on a thin arm of sheet-metal so it could move in and out. Here is how the switch and sled ride olong the rim:
Another view of the sled, which should have been rotated!
Sheet metal is a very poor spring, so I soon had to add a small rubber bungee cord to the contraption, which you can see in future photos.
About this time, my Makita cordless drill's battery died in the Arizona heat (it was about 100 degrees in the garage!). I had to cool off the charger and battery in the freezer before they would charge up again, so I had to get Midieval with an antique Brace and Bit.
Finally, ready to take the thing back inside. I had already tested the code to make sure I could see the switch activity, and I had figured out how to use the 'webcam' program to grab a picture off of a borrowed IBM webcam. The webcam is extremely shitty with regards to picture quality (176x144 or so resolution), but it works good enough to see that the basic idea might work. One problem is that the USB cord is so short that I can't do a full revolution the way my computer is set up now. I also had to put up a background made of card-board so that the web-cam would produce something resembling an image. I placed a piece of pottery my brother made onto the center support for a test object. Here's the webcam's view:
With lots of caveats, the Cyclecam seems to work. The sensor is finding all of the spokes, and the webcam is generally taking OK pictures when told to. It's unfortunate I don't have the space or cable length to do a full rotation, but I'll fix that one way or another soon. See pictures from this initial setup