Originally Posted by Red91
This is why I love you guys, I posted this at m.net too and they're like OMG THAT'S SO DANGEROUS SOMEONE COULD GET HURT! Post it here and Joe points out ways to improve the slicing accuracy.
Well, you have a couple of different force vectors at work here, and while you'd need to hire a competent physicist to really model the behavior of this system properly, a couple of things are immediately evident.
The motion of the car body, as it accelerates both longitudinally and on the axis if its suspension, is being coupled into the sword in a very elastic manner. This, along with the energy absorbed by the blade when it contacts the fruit, causes two separate oscillations (the flexing of the blade of the sword itself, and the deformation of the duct-tape bond) which interact and cause the blade to move in a highly unpredictable manner.
Rigidly coupling the sword handle to the chassis would simplify the problem greatly, however the exact position of the blade (on the time axis) would still be fairly nondeterministic.
I propose instead a coupling which permits the sword some degree of axial movement, and which incorporates both a large mass (to increase the inertia of the sword as a whole) as well as a damping mechanism tuned to combat the oscillation of the blade. In other words, use a steadicam-like mechanism to keep it stable on the Z-axis, and allow the entire
sword to deflect when it absorbs energy from impact with a pineapple, with the aim of minimizing snap-effects in the relatively elastic blade.
If you can get that thing stable enough, you could slice grapes with it.