Andrew,
You have done a great job of writing this up, and it is evident that you have done a LOT of experimenting along the way to validate and refine the design. Thanks for sharing this.
I am curious about "the BEST solution for the bottom [which] is a radial that is a conical shape with a very slight axial inflow." I am not sure if I am picturing correctly what you mean. I am envisioning something like the following cross-section, except at a much shallower angle than my crude ascii-graphics suggest (note that this will need a fixed font to show correctly):
|\ /|
| \ / |
| \ / |
| | | |
---- -----
What I am attempting to show is a cross-section of a short "plug" that goes in the end of the tube. This plug would have a smaller-diameter hole at the bottom to let the fluid in, but the "inner walls" of this plug would slope up at a shallow angle. Thus, when the fluid enters the bottom, it would be forced to flow up the shallow slope by centrifugal action, and in the process would suck in more fluid. Or at least that is my theory. :)
Is this what you have in mind for "the best solution"? If not, any chance of a drawing to show what you have in mind?
Thanks again for an excellent write-up!
--- In Homebrew_PCBs@yahoogroups.com, "cunningfellow" <andrewm1973@...> wrote:
>
> > Robin Wrote:
> >
> > I doubt if there is an absolute need for the
> > vanes at the bottom of each rotating tube
> > <SNIP>
>
> Early versions didn't have vanes.
>
> They had startup reliability problems.
>
> The vanes fixed the startup issues.
>
> The BEST solution for the bottom is
> a radial that is a conical shape with
> a very slight axial inflow.
>
> I knew people would complain about
> needing a lathe. I didn't want to
> make an impeller that needed a 5
> axis CNC machine to make :D
>
> (The toy wheel was something to make
> it easy for the average person to do)
>
> > <SNIP>
> > Likewise, I doubt that the three vertical
> > strips you glued inside the tube would be
> > necessary
> > <SNIP>
>
> Without the strips you do get banding.
>
> I think it was Alien Relics here that tried
> to make his tube with a spiral of holes
> and found his banding was the exact distance
> of the spirals.
>
> RoelCNC made his tubes with a bottom bearing
> and it does not have such banding. (even
> without the dams/strips)
>
> This indicates to me the out of round
> condition is causing the issue.
>
>
> > <SNIP>
> > I didn't see any lower bearing. Do you
> > have the end of the rotating tubes rubbing
> > against the lower plastic framework?
>
> No lower bearing. Thats why care is needed
> in the building to keep it round.
>
> > <SNIP>
> > I think the problem of evenly distributing
> > the etchant on the PCB might be solved to
> > some degree by cycling the speed of
> > rotation.
>
> I have done this speed modulation and it did
> even out the banding.
>
> I have also used a gear head motor to slowy
> cycle the PCB holder up and down.
>
> Neither is now neccesary now I have the
> banding issues sorted at the source.
>
>
> > <SNIP>
> > went to my
> > brother's place for lunch. Lo and behold
> > he had some of the same plastic containers
> > in his pantry! The were made by Icon:
> >
> > http://www.iconplastics.com.au
> >
> > near Melbourne. He recalls he bought them
> > at Bunnings.
>
> Yes the one I used was "icon" brand from
> bunnies (where lower prices where just at
> the begining)
>