Posted by Jonathan (68.14.212.239) on September 24, 2003 at 03:37:33:
In case any of you have read my previous posts and care to know more:
I have made a small model of legos, cardboard, and synthetic clay and it does not work. However, I'm sure this is due to high friction and light weights. I have done the math simulating much of it's movement on paper, and continue to be convinced of it's potential. I plan on making a larger one (the first is 5 or 6 inches tall) and will tell you all what I find.
I believe that the key is in Newton's laws of angular motion. Generally one must move weights far or close to the axle to get a net inbalance, but I have convinced myself that this is futile. The point of moving the weights in or out radially is to change the radius their force acts on. I think that there are other ways, just one being enbodied in the device I'm working with, to get a torque differential that doesn't require moving the weights away from their normal, circular path. Cowlishaw's device, esp. when simplified(digitized) the way I have done, acts like a left/right asymectrical differential. My calculations show that not only will one weight 'overpower' the other (two weight version is the only type I've simulated), but that the falling weight falls slower than the ascending weight ascends. However, this device doesn't fit with all eye witness accounts of JB's wheel. If we assume that some version of the device I'm working on is the same as JB's, then my only explaination for the (4lb?) cylinderical weights, (iron?) springs, and falling noises (8/rev) of his wheel is that they are all just unnecessary complications, probably for confusing onlookers. Then again, maybe he's just smarter me and all those were necessary, and radially moving weights isn't a futile meathod for PM.