William "Chops" Westfield mac.com> writes: > I never managed to get a kelvin generator to work, usually blaming it on 'losses exceed generation." But Someone I know made one and it worked fine. Some tricks I remember: - It was Winter, and the air so dry my nose was trying to bleed. Probably 20-25% RH. Everything was pulling sparks, me touching the doorknob etc. - The flow must come in separate drops and it takes ages to adjust the flow on both spouts. - Using some sort of metal mesh or screen on the cups for the drops to go through was helpful. It matters a lot if this screen is on top or bottom of the cup. I think it needs to be on top. I don't know if the drops are to touch or not touch the wire mesh. - The voltage built up very slowly at the beginning even if one column was charged with a rubbed stick first. - Rumor has it that when the device is well charged, the cups start deflecting the drops in various ways. That of course prevents more charging= ..=20 This device could move the leaves of a jar electrometer apart ~2cm and pull a rare 1cm or so spark after 10 minutes of tinkering and seemingly doing nothing operation. Not very impressive. Using the object area formula for capacitance you can figure out the charge 'step', i.e. Cfreestanding_sphere =3D 4*pi*epsilon*epsilon_0*r^2 for the dr= ops and similar for the remainder of the contraption. Assuming some efficiency eta in charge separation and transfer you can figure out an upper bound in charge/time and thus delta_v/time in the device. I think you can already se= e it will take a while to do anything interesting, the capacitance ratio between drop and machine being what it is. I am sure that people who built huge (HUGE) van de Graaff generators for physics in the 1940s-1960s had these calculations pat down, and they must exist somewhere. There was at least one using a 'patched belt' with conductive areas on an insulating belt. The closest parallel to Kelvin's dropper generator is the Wimshurt machine, I think. (and with that, I think that the drops need to not touch the mesh?= ) -- Peter --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .