Here is a clock that I built for a class I took once. It is built mostly out of stuff that I had laying around. It turned out to be pretty accurate. In my tests, I was able to tune it to an accuracy of one of two seconds lost or gained over ten minutes. Not bad for a bunch of cobbled together hardware! It is based around a mechanism that I saw once that regulated a hit and miss steam engine. Basically what is happening is I have a rotating part that has two arms that can be thrown outward by centripetal force. Electricity comes in through a wire that connects to the brass strip at the top o f the clock. The electricity flows through the brass strip to the steel vertical beam of the governor. It then goes into a thick brass wire which acts as a hinge for the swing arms. From there it travels down some wires to some more brass wire to a steel washer. The washer rests on a brass cylinder which has a brass wire that brushes against it. From the brush, the electricity goes through a wire to the motor and then back into the battery. Ok, now that I have described how the electricity flows, let me explain how it works... If I let the motor have continuous power, it would spin the swing arms so fast that they would become horizontal. Yes, I tried it, and it was rather scary because I didn't have the arms quite balanced at that point, and the whole governor shook loose and flew across the room. Needless to say, I made sure that they were ballanced before I did that again... Anyway, my goal was not to have the motor run freely, but to regulate its speed. Here is the key of that regulation. When the motor turns on, it speeds up the rotation of the governor. When the governor spins faster, the arms start moving outward. This pulls the washer upward and breaks the flow of electricity through the system. Of course this allows the motor to slow down to the point where the washer comes down enough to make contact again. At that point it speeds up again. This governor does this many times a second which results in a pretty even speed overall. I did notice that if I timed short intervals (like less than a couple of seconds) it might be pretty inaccurate since over the short duration it might speed up a little or slow down a little. If I timed long intervals, it all seemed to even out and be quite accurate. The gearing of the clock is from old toys and tape recorders. There is not much to say about that part as it is pretty self explanatory. This was a pretty fun project. In the beginning, I had planned to make it have some rotating metal cylinders that had paint on them to allow electricity to pass or not pass. There would have been some brushes that could slide over them at the press of a button so that some coo coos could have been activated. Unfortunately, that was wishful thinking since my other classes needed a lot of time and I couldn't do as much as I had hope to. Nevertheless, I am happy with how it turned out.