Torus?

[weirdwilbur]

I think, if air is a fluid, that we could smooth it's travel by altering the shape of the current design to that of a torus with a bell on the top side and the baffle escape slot on a plate that laterally bisects the interior.
http://en.wikipedia.org/wiki/Torus
  Fabrication might be done using the inner tube for a 12 inch tire. After inflation, it might be covered with plaster to make a mold?and that's as far as I've been able to get. This effort would give me the wall curvature, but what medium would I use in the mold? I assume some sort of heavy plastic.
  My proper understanding of the finished product is very weak here. It might be that the  sculpture would be mostly cut away in order to sculpt it nicely into the bell outlet, side inlet and the flat baffle board which would also act as the top of the catch barrel.
  Am I on the right track? It's late. I may need a break from thinking about this for a while. Why can't we buy heavy plastic donuts off the shelf?

[weirdwilbur]

OK that was dumb. The baffle board would NOT bisect the torus laterally but rather be sculpted across the bottom to preserve the curve, of course.

[alan m]

im not sure i understand what you are asking properly . are you trying to shape the top of the seperater into a bell shape , if so no, the pipe needs to extend down into the chamber by 1/2 the diameter of inlet pipe abover the baffel

[weirdwilbur]

Thank you Alan! See? I'm real vague on the dynamics. The hard numbers help a lot. Am I correct in thinking that the size of the inlet, at a given cubic feet per minute, would determine the optimum outside diameter of the cyclone?

[alan m]

its not truly a cyclone but yes to a degree . there arnt any rules persay . usually people use a smaller diameter for smaller sized pipes and larger dia for larger pipes. my 6" version will sit on top of a 45 gallon (eu gallons i think) , i think the diameter is around 21-22 inchs. my 4" is around 14- 16 " , any my 50mm one is around 12".  people usually let the drum the have or can find easily dictate the diameter.

there probably is a sweet spot where all those data lines up best but i am not sure where it is or how to find out without building a few dozen seperaters