Structured Packing Quest

[MooseMan]

Very clean, precise and inspiring work Big Swede, you clearly know your way around a tool shop, and have the brain space to conceptualise and bring a project to reality!

[shadylane]

Brings up a question, how open should the mesh be? If it's too fine, it may as well be a solid sheet, as I think vapor would have a tough time moving THROUGH the mesh holes as opposed to being simply channeled along the tube form. Or maybe that would be better? I think some vapor migration through the mesh is the answer, as this would increase effective surface area hugely. But not TOO much? We are basically making thousands of angled tubes for vapor/reflux mingling. I have no clue.

If this works as-is, I'll probably skip external gearing, make some mesh form, and try producing a 2" biscuit form.

I haven't a clue either but your logic sounds reasonable.
Just a guess for a 2" incher, maybe try 20 Mesh SS window screen. It's cheap and readily available.

[Roo]

The base machine is done. Rolls are 7" long. All I need to test is to slap a handle on it. Hopefully tomorrow I'll give it a try. I've got two test rolls of SS mesh. The price seems decent for a respectable size roll. I actually weighed the rolls to try and get a hack on cost-per-biscuit once we're at that stage.

Looking good. do you have a sense of what it might cost for 6" column packing? I have been looking at it purely from a speed of run point of view but packing seems pretty expensive, even if you put a premium on your time.

I imagine the cost of materials would only be a small portion of overall cost as creating the biscuit from the formed mesh seems like it might be time consuming.

Regarding mesh size, I have no idea of actual optimal sizing. I guess all you can do it try and hopefully pick a size that sits somewhere in a viable range. How many biscuits of a particular mesh do you think you need to do a test evaluation run?

I'd happily contribute to the cost of mesh for some experiments, but I suspect that it is the time and effort to make the biscuits and conduct the test runs that will be the biggest issue.

Would results from testing with smaller diameter columns scale to larger? What is the minimum diameter column you are thinking the biscuits will be used for?

Do you know of anyone on the forums that has a setup to run controlled(ish) evaluations?

[BigSwede]

TY for the kind words. And it works! It handled various mesh sizes/thicknesses well. And I am a dumb-ass, an angled strip DOES travel. I'm not sure why I was so convinced it wouldn't. No worries, the 7" rollers can handle foot-long strips at any angle.

I haven't a clue either but your logic sounds reasonable.
Just a guess for a 2" incher, maybe try 20 Mesh SS window screen. It's cheap and readily available.

I agree. If you zoom in on the very first picture of the "Sulzer" packing, the mesh looks about 20 or 30. This stiffer wire also holds shape better.

Struct25.jpg

This test was I believe 30 mesh. Stacked and held nice.

Struct26.jpg

Looking good. do you have a sense of what it might cost for 6" column packing? I have been looking at it purely from a speed of run point of view but packing seems pretty expensive, even if you put a premium on your time.

I imagine the cost of materials would only be a small portion of overall cost as creating the biscuit from the formed mesh seems like it might be time consuming.

The 16" x 48" roll of mesh from Amazon was $12. It feels like a lot of mesh for not much money, because once formed it feels like mostly air. I'll know a lot more once I make a biscuit. I can weigh the biscuit and compare that to the mass of the original roll. $$/biscuit, that sort of thing.

After months on the old SPP machine, all the s--t that went into making it work, this feels 1,000% easier... IF it works. Cheaper too, SS MIG wire was very pricey.

[BigSwede]

I'd happily contribute to the cost of mesh for some experiments, but I suspect that it is the time and effort to make the biscuits and conduct the test runs that will be the biggest issue.

Would results from testing with smaller diameter columns scale to larger? What is the minimum diameter column you are thinking the biscuits will be used for?

Do you know of anyone on the forums that has a setup to run controlled(ish) evaluations?

I appreciate the offer. The $$ so far has been trivial. $26 for 2 rolls of mesh, and metal I had laying about.

My gut says a 3" or 4" column needs a deeper mesh form. This MAY be able to migrate to a 3" as pictured, but this is all very new. I do have a possible test driver who knows a lot more than me about reflux and has the data to compare. But that's a ways off yet.

I had some gear cutters ordered that arrived today. These would make a KILLER roll set for beefier mesh, with a nice rounded channel form rather than a pointy peak.

Struct27.jpg

Save that for the future. I had one last thought on all this before biscuit making... there is a tool used to repair screen windows, it looks like a dull pizza cutter. Maybe stack and space a bunch of thick fender washers, big washer, small spacer washer, big washer, etc, maybe 4 or so. Mount on a handle so again, it rolls like a pizza cutter.

In a piece of hardwood, rout a series of channels. Lay the screen over the wood, press the pizza cutter tool into the mesh to form channels. You can roll back and forth to press the channels in. Then, move the mesh using the last formed mesh line into the final wood female channel so as to maintain even spacing. Make more channels.

[30xs]

In my mind the mesh is basically giving the reflux a place to cling. The open area between the crimps are going to be the main up flow area.

[Roo]

That is great progress, it looks really encouraging.

[Stonecutter]

This is some innovative stuff. Really cool BigSwede, looking forward to the field tests.
The pursuit of a fine drop continues!!

[BigSwede]

Quick update: The machine makes mesh. Here is samples of 20, 30, 40 mesh, and I'm leaning towards the latter.

Struct30.jpg

Looking at the few pictures of small structured mesh biscuits, especially the very first one in this thread, shows a peak to peak count across the diameter of about 10. So, for a 2" diameter biscuit, we'd want a pitch distance of 0.2", 5 peaks per inch. The machine does that. But I'm not happy with the form, which are shallow, pointy peaks rather than smooth hemispheres. Without a good form, I'm thinking we may as well just roll up mesh into a swiss roll and stuff it in there.

I do have gear cutters, which make a MUCH better form. Did a mini set out of delrin plastic, and checked out the mesh it produced.

Struct31.jpg

To the left is the new form, to the right is the current triangular stuff. So I like the shape of the new gear-form mesh. The problem is pitch. This cutter set produces 7 peaks per inch... too fine. I'm going to recut the rollers shooting for 5 peaks/inch, and a deeper sinusoidal form.

I really want to nail the mesh before cranking out a dozen biscuits, and later wish I had started with the right mesh shape. The good part is I can drop new rollers right into the machine, so it's not starting over, just new rollers.

One nice thing, I was able to spot weld the SS mesh with ease for the biscuit bands:

Struct29.jpg

It's coming along, slowly. One of those things where planning and prep is pretty important.

[Roo]

Quick update: The machine makes mesh. Here is samples of 20, 30, 40 mesh, and I'm leaning towards the latter.

Awesome progress, you really are giving it a nudge. Pretty sweet that it takes a spot weld easily

Also I wonder if you could take the formed mesh samples and conduct an experiment to help get an understanding how the different designs behave.

Can you measure the biscuit (or a section of mesh) dry and then submerge in water to see how much liquid it retains. I have no idea what the target performance is, but I think it might be worth trying to find some kind of simple test to characterise the mesh behaviour, then when you get results from a few trial runs you may have the some way of correlating and using that to inform the next iteration.

It is my understanding that you want the mesh fine enough that when the refluxing liquid drops onto it it doesn't just drain through but it is coarse/open enough that it maximise the liquid/vapour interaction without massive holdup. Raining water onto the mesh in the expected orientation alone may not be enough without something modelling the vapour moving up.

An alternative thought... make 1 biscuit in each size and install into a glass section underneath an upper packed section (scrubbers or whatever) and operate with straight water to see the hydraulic behaviour and get closeup slomo video (I think most phones would do a good job nowadays). Swap the biscuit out for a different one and repeat.

Using water hopefully would reduce costs/waste and still be illustrative.

What diameter column are you targeting? I thought this better for larger columns, or is it because of the cost limiting it to the big boys with big columns?

[BigSwede]

Roo, the target now at least is 2" and 3". I think a good wavy shape will work for both, but going larger diameter? Are many hobbyists doing reflux with a 4" column? My hope is that this stuff will allow good performance with a shorter column.

I like your concepts for testing. One of the caveats mentioned earlier is that reflux from a condenser needs to be well-distributed, and I haven't put any thought into that yet. Let's say you have a 3" column and your condenser puts out a relatively confined stream smack in the middle of the diameter. If this stuff works as I think it will, only the middle of the biscuit would be active as the reflux stream zig-zags it's way downward, and vapor upwards will maybe bypass a lot of the reflux and simply hug the periphery.

So a simple hydraulic test for any sort of falling reflux distribution would be easy, like you say. Sight glass, RC above that, and dump in water at the top, see what happens. At least that would give an idea about vapor/reflux mingling. Different mesh sizes might show very different behaviors.

The true test of course will be HETP performance vs. known hobbyist random packings. If three feet of 2" dia. scrubbies gives 90% ABV at 2KW, what does this stuff do?

I'll say this at least, so far the costs for materials will be pretty low. Not scoria low, but well below SPP. And the construction is a lot more interesting than clipping 10,000 SPP nuggets while watching TV.

[Roo]

Not sure if this is going to work, but there is a video showing some manufacturing of structured packing from an Alibaba vendor. Interesting if nothing else https://www.alibaba.com/product-detail/ ... 30397.html

[Windy City]

I would love to get this in 4"

[Yummyrum]

Using water hopefully would reduce costs/waste and still be illustrative.

Understand your thinking Roo , but water has a high surface tension compared to Ethanol .
So a drop of water placed on mesh would likely sit there in a globule, whereas , a drop of Ethanol will fall through the mesh .

[Roo]

Using water hopefully would reduce costs/waste and still be illustrative.

Understand your thinking Roo , but water has a high surface tension compared to Ethanol .
So a drop of water placed on mesh would likely sit there in a globule, whereas , a drop of Ethanol will fall through the mesh .

Fair point (it did occur to me), while I was thinking about not wasting good ethanol... equally not wasting time is probably as important.

As I understand it, the way this type of packing works is dependent on the wetting behaviour of the packing. And as you pointed out, that behaviour would be different with water. It might be possible to infer one behaviour from the other, but that would be yet another variable I guess.

I have been recently looking at different options for larger diameter columns where SPP would be challenging... While Dixon rings (setting aside cost) look interesting (also made from a mesh), I have read there is a requirement to "pre-wet" them by performing a controlled flooding of the column at the start of operation. I think (not certain) it is because they are made from a fine mesh and if not flooded at the start trap air and don't fully wet, thereby reducing efficiency.

[Windy City]

I have always used Shady's Sugar Shine for sac runs and testing.
It is inexpensive to make and consistent.
For testing I run low wines at 40% and then with no cuts I dilute again to 40% and run again with different parameter.
When I was trying to dial in my pumped side column I think I ran the same run 4 or 5 times

[BigSwede]

Not sure if this is going to work, but there is a video showing some manufacturing of structured packing from an Alibaba vendor. Interesting if nothing else https://www.alibaba.com/product-detail/ ... 30397.html

OK those machines are pretty cool. Uses a stamp mill/die combo, and some of the mesh has holes punched. Without being a chemical engineer and knowledgeable on this stuff, I wonder of the holes are for lateral vapor flow? Maybe the mesh (if too fine) limits vapor flow, and the holes are needed.

I would love to get this in 4"

Beefier teeth cutters are set up and the new rollers are in progress. Barring catastrophe, it'll be biscuits very soon.

[Windy City]

I reached out to the distributor listed in that video trying to get more info on sizing.
She told me each biscuit is 100mm/4" tall she also recommended a liquid distributor.
I have read this in several places that a distributor makes a big difference.
This looks a lot trickier to build and I just may buy these from them.

image.png

[tommysb]

What is the idea with that Windy? Does it just (re)direct down-flowing liquid, or does it act as a substrate/packing for reflux? Looks very interesting!

[Windy City]

What is the idea with that Windy? Does it just (re)direct down-flowing liquid, or does it act as a substrate/packing for reflux? Looks very interesting!

It evenly distributes the reflux from the reflux condenser onto the packing.
I have read a couple of studies that said it made a big difference.

[LWTCS]

What is the idea with that Windy? Does it just (re)direct down-flowing liquid, or does it act as a substrate/packing for reflux? Looks very interesting!

It evenly distributes the reflux from the reflux condenser onto the packing.
I have read a couple of studies that said it made a big difference.

100% what windy said. Prevents channeling.

[BALDR]

What is the idea with that Windy? Does it just (re)direct down-flowing liquid, or does it act as a substrate/packing for reflux? Looks very interesting!

It evenly distributes the reflux from the reflux condenser onto the packing.
I have read a couple of studies that said it made a big difference.

I was told the same thing from the manufacturing my dixon rings. I showed them a picture of a distributor I stole from Salt Must Flow. They wrote back that it would work. But thiers would work better. Their price $268 US. I found slicing the pie into 6 pieces worked best,plus rounding the edge towards the bottom helps.

Centering Plate.jpg

[Salt Must Flow]

^ ^ ^

I mainly use that "distributor" for centering reflux through the underlying sight glass mainly so reflux doesn't run down against the glass. It works great! If you bend the petals to point where you want the reflux distributed, it will do it.

[BigSwede]

I reached out to the distributor listed in that video trying to get more info on sizing.
She told me each biscuit is 100mm/4" tall she also recommended a liquid distributor.
I have read this in several places that a distributor makes a big difference.
This looks a lot trickier to build and I just may buy these from them.
image.png

I'm digging this. At first I thought "That's some really weird packing" and then realized it is a reflux distributor, which is important for function.

The good news is you really only need one (maybe 2? half way down the column might be a thing). I think I can replicate that (or something close) with spot welded / TIG SS sheet.

If you enlarge it and stare long enough, I think you can get a feel for how it works, the troughs that guide the liquid, etc. Complex little bugger though. If sold in a size one could drop right in, it'd be worth simply purchasing.

[BigSwede]

Salt's device prompted me also to make one along those lines. This started as a short tube section brazed to a ring. Holes get drilled around the periphery, then slotted with a hacksaw. This gives petals to bend and mess with. I haven't tried it yet.

Struct32.jpg

[Salt Must Flow]

BigSwede, either method should work to distribute reflux enough. I'd bet it's not really important anyway. When you look at glass columns with zero reflux distributors, centering collars, etc... they work just fine and "channeling" doesn't appear to be a real thing with hobby diameter columns.

[shadylane]

I'm thinking a shotgun dephleg would distribute the reflux good enough.

[BigSwede]

BigSwede, either method should work to distribute reflux enough. I'd bet it's not really important anyway. When you look at glass columns with zero reflux distributors, centering collars, etc... they work just fine and "channeling" doesn't appear to be a real thing with hobby diameter columns.

Yeah, I think that is the key. Some of these gigantic columns for industry are many meters wide. Splashing alone at the top of a hobby column should move reflux over most of the column area, to begin its work.

Still, fiddly things appeal to me.

[Roo]

I have been looking at larger diameter packed columns for neutral production not for volume of liquor but for manageable run times that fit in with life. Maybe it's a fool's errand, but time will tell.

One of the big things that comes up all the time when looking at column diameters above 4" is the distribution of reflux at the top of the packed column. As many have noted, the problem with poor reflux distribution can significantly degrade the efficiency of the packing. I have heard that doubling HETP could be a real possibility. In a height challenged domestic shed that is a killer, and that isn't even accounting for the cost of the actual packing.

One of the liquid distribution designs I came across and looks like something that could be DIYed is a "Gravity Liquid Distributor". One can be seen here https://www.machengineering.com/tower-i ... tributors/

gravity_liquid_distributor.png

They would sit below a dephlegmator, and work by the vapour rushing up through the chimneys and the refluxing liquid falling down and captured in the bath. The liquid forms a shallow pool and there is a positive pressure generated by the vapour rushing up towards the condenser. With the right number of right size holes the return liquid is squirted down in a controlled pattern onto the lower packing.

It is hard to imagine something like this working (or being needed) for a 2" column, but I could see it being needed for a 6" column.

It would be easy to make, start with a blank bubble plate disc that fits into the triclamp joint, drill larger holes for the chimneys and the pattern of smaller liquid distribution holes. then solder in the chimneys from short copper lengths.

Increasing the chimney length allows for higher turn down ratio, I have read up to 1:2 ratio (running at half the design power).

What got me interested in this is it improves packed column extraction performance without increasing the packed height. Obviously the distributor increases it, but nowhere as much as doubling the packed height.

[Roo]

The gravity distributor I posted above is obviously for a much larger column than even the largest home distiller might use.

But one thing I am not certain of is the intended vapour path only through the chimneys? Or is the raised sides of the "bath" smaller than the column ID, leaving an outer annulus for the vapour as well, keeping the falling reflux away from the walls preventing channelling?

I stumbled on a app note on the design, can't find it now, but if anyone's interested I can try and find it again and post a link.