Large volume laminar flow machine

[karimwassef]

So now to the unsolved problems... There are two issues I haven't resolved yet.

The first is buoyancy of the acrylic box... It's full of cavities. Above, Below, and even in the diverter structure. It's also got a ton of acrylic, but that's not the heaviest stuff. If it's glued watertight, it might float. I'll need to do the math.

If it's not, then water will seep in... that water would rarely go anywhere until maintenance... then it becomes a 300lb mass of water and acrylic that's going to explode if I try to remove it... maybe I'll fill it with foam... anything is better than air (even though foam is mostly air). The foam would at least stop the 300lbs of water from getting in.

....

The second is more difficult (I think).

The overflow glass and tank sides are essentially the ductwork now - so the acrylic can be removed for maintenance (and in case of a massive fish invasion)... but the top still needs to be sealed.

I need smooth glass sides (no lips or edges) so I can slide the side assemblies in (the back duct is simpler).

Now, how do I seal an acrylic top on a glass tank? I can use a gasket material and 0.5" thick glass/acrylic should be ok? But then, where the pressure? Do I drill holes into some glass ledges outside the tank and then expand the acrylic sections to cover them? Use plastic screws to button it all down like a skimmer assembly?

Would that work?

 

[karimwassef]

By the way, this has elements of a mechanical assembly, pumps, ducts.

All of it is now programmable with an alternating DC motor outside and controlled flow DC propeller pumps inside.

It's fully sealed and below water level to reduce waves.

Its parts are removable for maintenance.

The diverter should work to create a uniform, directional and reversible flow. Laminar? I don't know, but by the time it gets to the middle of the tank, I think so.

It will suck and blow at the same rate - no standing waves.

It is a risk for anything under 0.5" in size, but I have that now with my open flow tank anyway.

I can build it with conventional tools - no 3D printing...

So I think this works for me (once we solve the last couple of gaps) and I can get back to building the demo.

The only thing is doesn't have is PVC! Oh wait, it does! a 10" one too, but not used as a pipe

Thanks for all your feedback and ideas. You guys and gals rock. Almost there....

 

[salty joe]

Nice design! It looks a lot more doable than a bunch of vertical squirrel cages.

The entrance to the acrylic baffle looks like a choke point-wouldn't you want that as open as possible?

I'm sure I'm missing something, but I don't see the value of the ramps, the water only has one place to go.

 

[karimwassef]

Ramps=Reduced resistance and minimizing settling/dead zones.

My side diverter takes the 10" duct down to 1.5" at the entrance. That's a lot of back pressure and it is a pinch point... But this is redirecting the entire flow into the structure at 90 degrees in only 2" of available space... Unless I grow the design, I don't see how to open it up.

It's still 10" x 23" in the back going to 1.5" x 18", so it's not like 1.5" PVC pipe... But still almost 1/10x the aperture.

I'll take some snapshots with dimensions and put the seal locks in place to get that clear.

 

[karimwassef]

Oh I see... Just before the turn, I bring it down to 1". Every step in the green structure is 0.5" by the way.

Here's the issue... I can make it smoother and more open, but then the flow will push forward more that change direction by 90 degrees. That means one half of the diverter (the one more towards the front) would get more flow.

In fact, my first version was much smoother in terms of flow and openings, the problems is that I need to create local pressure change to change the direction of flow. So I made my turns more severe to give equal opportunity to the forward as the backward turn.

 

[karimwassef]

Do you think the diverter structure will work? I'm using more logic than engineering here.

 

[karimwassef]

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsv0kpaf5d.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsv0kpaf5d.jpg" border="0" alt=" photo 0_zpsv0kpaf5d.jpg"/></a>

Having hard corners makes the forward and back flows share better.

 

[karimwassef]

Here's another way to look at the openings. If the flow from a point before and after a split is nearly vertical, then the distribution will be more even...

Looking at that, I got the first and last ones right, but the ones in the middle (yellow) need to be more angled.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsp8ngrmgr.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsp8ngrmgr.jpg" border="0" alt=" photo 0_zpsp8ngrmgr.jpg"/></a>

If I add the orange sections, it'll work better, but add more resistance again.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zps29gmhail.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zps29gmhail.jpg" border="0" alt=" photo 0_zps29gmhail.jpg"/></a>

If it's possible to make the little red triangle sections and attach them, that may help. I don't know if my local vendor can make beveled edges like this, much less strips of just the bevel.

 

[karimwassef]

As this gets more complicated, I'm wondering if I can't make a LEGO mold (negative) and then pour a dense foam/concrete or other material to make the actual structure?

I don't think anything is as good (surface resistance, rigidity, etc...) as smooth plastic though...

 

[karimwassef]

Here are three alternative diffuser designs

<a href="http://s1062.photobucket.com/user/karimwassef/media/2_zps20ott5je.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/2_zps20ott5je.jpg" border="0" alt=" photo 2_zps20ott5je.jpg"/></a>

The first constricts the turns (narrower), but then adds corner rounds. I think these might be available off the shelf (the bowed out version anyway).

The second just constricts the corners.

The third uses the fact that is a fractal concept and extends it to 1/4" acrylic sheets too. The combination is clearly more difficult to build and 1/4" sheets don't glue down as well (for me, anyway), but this would be the best structure - if only I could make it.

 

[Michael Hoaster]

Wow Karim, this is a very ambitious project! You live for this stuff, don't you?

I've read through all of it, and I won't pretend to understand much of it, but I think it's way cool! It's guys like you that push the technical boundaries of the hobby.

Good luck! I look forward to seeing the final execution of it!

 

[karimwassef]

Thanks.

It's a lot of learning from the group, really. I know what I want to get to, but doing it without the people who have been giving me help, advice or just questioning my assumptions... Couldn't do it.

I think we're close.

I do get excited about doing something good that's never been done before and maybe help others get there too.

This project was just a need after snorkeling in Cancun and then coming back to my inadequate tank's flow... And that's with dual massive surges and 4 propellers in the tank, beyond the sump circulation.

 

[salty joe]

Maybe it'd be worthwhile to get a bunch of 1/2" acrylic pieces and use something temporary to hold them together-like maybe wax from a toilet wax ring. Then maybe a test run using food coloring to track the flow. You could also use wax to round or knock off corners.
Assuming similar results from a partially full open top baffle and a closed baffle under a bit of pressure.

Just curious, what kind of timing-control do you plan for the pumps? How about the switching motor?

 

[karimwassef]

I'm going to build the diverter prototype with LEGO blocks. Should be here Monday.

I'll just set it up in a wide flat plastic storage box with a powerhead.

The problem with Legos is that they trap air and float, but I might just assemble it underwater. They're also at a 0.31" pitch instead of 0.25". That means the whole diverter grows by 24%. Since it's a replicating fractal, the outlets are in the order of 2 (2, 4, 8, 16, 32, 64...) so I can't just drop a few outlets if the building blocks get more coarse unless I want to go down by a factor of 2. Between 1/2" and 1/4", I went from 32 outlets to 64. If I ever figure out a good way to do it, I'd go 1/8" and get to 128!! Legos also have small gaps between them - in theory. But I don't remember that as a kid... We'll see. I've been studying Legos a lot the original additive manufacturing technology.

Since it's a pressurized sandwich in a plastic box, I think it'll hold.

For the corners, I think the plastic quarter round may work, or just use epoxy putty shapes as needed (for important corners). Maybe superglue to secure the putty to the Lego surfaces.

 

[karimwassef]

For controls, I'll use my Apex dimming control outputs. I use them for my actuated surge controls now, but that's moving to an arduino board that takes input from the Apex through the power outlets.

The motor is a challenge, but that's most likely a quadrature encoded stepper motor connected to the arduino with input commands from the Apex. Basically, the brains is the Apex (turn left, turn right, high flow, low flow) and the local muscle nerve-connect is the arduino, and the muscle are the pumps and motor.

 

[karimwassef]

For the final build, I was thinking of a slotted base to line up the pre-assembled vertical pieces, but that's tough. I'm going with small 1/2" strips cut to length glued down on the base and top... Then the vertical elements will sit in them.

 

[salty joe]

Here's another way to look at the openings. If the flow from a point before and after a split is nearly vertical, then the distribution will be more even...

Looking at that, I got the first and last ones right, but the ones in the middle (yellow) need to be more angled.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsp8ngrmgr.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsp8ngrmgr.jpg" border="0" alt=" photo 0_zpsp8ngrmgr.jpg"/></a>

If I add the orange sections, it'll work better, but add more resistance again.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zps29gmhail.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zps29gmhail.jpg" border="0" alt=" photo 0_zps29gmhail.jpg"/></a>

If it's possible to make the little red triangle sections and attach them, that may help. I don't know if my local vendor can make beveled edges like this, much less strips of just the bevel.

Since the water wants to move to the right, I wonder if a small horizontal block positioned above the arrow heads in the first drawing would do the trick? That way you are only choking off what actually needs to be choked.

 

[herring_fish]

Wow! You have moved a lot in the last few days. I'll chime in with just a little tuning here.

Could you calculate the opening sizes and modify the design accordingly? I quickly tried to communicate what I was talking about but a lot more time would give a better result. The change is in the top graphic and the one underneath is there for comparison.

I noticed that the initial pinch point looks like it is the size of only two of the final outlets. That is 2 to 32 or 1 to 16. Using CAD, you could adjust the length of the vertical dividers and eliminate one or some of the horizontal spacers to get a better total flow. You would only improve it a little, perhaps up to 4, 6 or 8 to 16 would be all that you could get but anything would help.

Again, I did this quickly (CAD and a calculator does a better job). I moved some walls up and opened the openings where possible. A few of the changes might me backward but I hope that you get the idea.

It looks like mine is about 7 to 16 in the first opening. Being rough, the height prior to that appears to be only 5 wide so ignore that. I don't know if the second level would go up or down. Then you see the layer of horizontal plastic that is removed. Below that I guess you can't change much because of the thickness constrains of the plastic but the higher vertical ones are just length changes.

 

[karimwassef]

It's been iterative. My latest goes to 64 final slots with 1/4". I'll take a picture with dimensions.

I also extended the length of the exits to improve laminarization, but that also constricted the sections before.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zpsg94gpwce.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zpsg94gpwce.jpg" border="0" alt=" photo 0_zpsg94gpwce.jpg"/></a>

 

[karimwassef]

Here are the dimensions.

I used minor offsets (1/4" to 1/2") to create a slight bias towards the reversed side flow.

<a href="http://s1062.photobucket.com/user/karimwassef/media/0_zps2u2yej5q.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/0_zps2u2yej5q.jpg" border="0" alt=" photo 0_zps2u2yej5q.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpslstfcxzq.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpslstfcxzq.jpg" border="0" alt=" photo 1_zpslstfcxzq.jpg"/></a>

right side closeup
<a href="http://s1062.photobucket.com/user/karimwassef/media/2_zpsdu9mnark.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/2_zpsdu9mnark.jpg" border="0" alt=" photo 2_zpsdu9mnark.jpg"/></a>

left side closeup
<a href="http://s1062.photobucket.com/user/karimwassef/media/3_zps0yktg6xx.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/3_zps0yktg6xx.jpg" border="0" alt=" photo 3_zps0yktg6xx.jpg"/></a>