J. Phil Thien's Projects

Nice!

that's a much cleaner build than what I cobbled together:
  http://www.jpthien.com/smf/index.php?topic=1085.0

I think Phil was referring to something like this:

http://www.jpthien.com/smf/index.php?topic=992.0

Andy

Continental 3200CL 32 Gallon Clear Huskee Trash Can

  http://www.webstaurantstore.com/continental-3200cl-32-gallon-clear-huskee-trash-can/6902632CL.html?utm_source=Google&utm_medium=cpc&utm_campaign=GoogleShopping&gclid=Cj0KEQiA1NWnBRDchObfnYrbo78BEiQA-2jqBZzOkDUiVqd1-qWmeu9OcRuNqecreK_zhB1y-FZrV4waAm758P8HAQ

or

http://www.amazon.com/Continental-Huskee-Gallon-Clear-Trash/dp/B004FNW71S

might try a more squarish piece like this one:
  http://www.jpthien.com/smf/index.php?topic=1119.0

you can use laminate (don't use contact cement though.  you may need to move it around it position it.) or some other flexible and sturdy material.  Still attach it with some combination of grooves, sealant, and screws.

Andy

Mike,

That last picture is very similar to mine.  Except your round-to-rectangle transition is more graceful.


BernardNaish: I definitely agree to get the current CFD model to where we understand it before we start making changes like eccentric shapes, chamber heights, and inlet angles.

I hope we didn't run off superptrucker........

Andy

Build thread:

http://www.jpthien.com/smf/index.php?topic=1085.msg5880#msg5880

I too would like to see how these variations affected the computer model,


Andy

I can confirm that a curved inlet does force the particles to the outside of the separator.

I haven't tired to photograph it, but the intersection of the particle flows looks like a "merging traffic" sign and no apparent turbulence where the flows meet.


Andy

I had a chance to test it more today and the bypass is due to overloading.  What I thought was bypass was dust still in the bag filter and the ring from before making the separator.  I took it off and let the impeller discharge to open air and nothing came out visually unless I overloaded the duct.  I did the typical pour a pile of dust on the driveway and suck it up while watching the outlet.

The dust stays suspended along the walls for several revolutions, I'm guessing that's because of the increased velocities from the spiral.  It does eventually drop into the can.

So, I'll be using this since it's working well.

I might try a version that is simpler to construct like the industrial versions bill70j posted awhile back.  It has 180deg of arc at one radius and 180deg of arc at a larger radius. 

If you want to try an Archimedes, might look at how Bill Pentz made a spiral template:
  http://billpentz.com/woodworking/Cyclone/blower.cfm#spiral_configuration


 

Dan did something similar here:
  http://www.jpthien.com/smf/index.php?topic=992.0

Never mind.

A $5 3in puller from the auto parts store worked just fine.  Had to use a 1/2in deep well socket to extend the threaded rod though. 

Dan,

What did you use to remove the impeller so you could rotate the housing 180 deg?

Looking at the impeller, I'd guess something like this:
  http://www.harborfreight.com/automotive-pulley-puller-66868.html

but the reviews about the tips breaking concerned me.

thx,
andy

The spiral or scroll design is not new.  It was presented a while back:
  http://www.jpthien.com/smf/index.php?topic=492

Extending the spiral around on itself seemed to also address Phil's idea of a curved inlet.

Part of the fun of doing this is seeing if things work or not.

I'd been thinking about a spiral top hat for a while.  At some point a while back, Phil mentioned having a "lead-in" section of the duct that curves to put the material on the outside of the duct before it enters the chamber.

I combined those ideas.

The spiral is an Archimedes spiral.  In this design, every 90deg, the radius increases by 1/2in.  My brother had been playing with Ruby script plugins for Sketchup to make spirals; so, I fiddled with the settings (initial radius and growth rate) to get a spiral I liked.  The spiral is for 450deg.  The last 90deg (or first 90 deg depending on how you look at it), represent the lead-in curve section of the duct.  That kept the width of the rectangular duct cross section consistent. 

I sized the top hat for an 18in diameter 10gal tub.  I wanted a tub less than 20gal so if I can get a bag in the tub (without it sucking up into the top hat), the weight of the waste would be in the <20lb range.  Sawdust density: http://www.engineeringtoolbox.com/density-materials-d_1652.html

I'm getting some by pass but I think it's because the outlet is too close the the shortest radius.  If I were to do this again, I'd make the initial radius larger by about 1 to 1-1/2in (this would be radius from the center of the outlet duct to the wall at the closest point).

As you can tell from the pictures, my plexi bending skills leave a lot to be desired.  I used metal duct tape to seal both sides of the wall when ever I had a break....

Cutting the spiral with a jig saw was the hardest part for me.  Once I had one spiral mounting guide for the plexi, I screwed it to another spiral mounting guide and used a flush trim bit to even out any differences.  (I wish I had access to a CNC for this.)

I followed the general top hat assembly techniques of several others here.


The top view picture of the top hat shows the path of the plexi following the spiral.


I think the initial 90deg curve in the duct is a really good idea based on watching the material fly around the outer wall of the top hat.  If you build a spiral, make it a little larger that what I made.  Using a spiral may be more trouble than it's worth unless you have access to a CNC (or better jig saw skills).

I've attached a Sketchup drawing and the Ruby script plugin for Sketchup along with a few pictures.

Hope this helps somebody.
Andy

assuming you use some kind of adhesive in the groove for the polycarb like Goop or maybe epoxy, you can use a flush trim router bit to make the slot flush with the outer wall.

It was explained to me that it's related to noise reduction:

Quote
.... [It] attaches to the inside only.  It doesn't attach 2 pieces of acrylic, in theory it's supposed to reduce noise by introducing a "2nd gore point" that's out-of-phase with the main gore point...  in practice, I think the jury may still be out on that, though.