J. Phil Thien's Projects

Possibly the most beautiful top hat renderings ever!

The design represents the current thinking for a 1 1/2 H top hat. The recent thread on CFD raised some interesting ideas, however, and Phil's follow-up questions/thoughts pointed the way to what may prove to be a better design.

The first idea is that there is a benefit to moving the inlet away from the top of the top hat.  CFD modeling showed that particles were slowed when the inlet was tight to the top, and that separation improved when the inlet was lowered slightly from the top.  Phil speculated that it may be best to have the inlet closer to, but not tight to, the baffle rather than at the top.  Since we want the particles to end up at the bottom, why not introduce them closer to the bottom?  Air is more agile and will find its way out. 

The second idea is that the outlet may perform better if it is flush with the top of the top hat.  CFD modeling (with a high inlet) showed there was less pressure drop with a flush outlet.  The modeling held airflow constant, but in our systems, less pressure drop means higher airflow rates and higher velocity, so separation should improve.  Would lowering the inlet diminish the benefit of a flush outlet? I doubt it, but CFD could answer that question.  (hint, hint)

Combining these two ideas could be the foundation for the next generation of top hats.

Aaaaargh, thanks Schreck, too many variables to play with. I desperately need to build one of these things right now. I will put the inlet and outlet wherever the consensus says, if there is such a consensus. Without knowing much about the theory, I would have thought the entry at the top would be most effective as the velocity of air entering the chamber would keep the particles against the wall. If there were 'dead' air above the inlet, would this not cause more turbulence with more chance of particles being held in the chamber longer than necessary?

I think I may keep the inlet at the top for now but it will be very easy to play around with the outlet position by simply pulling or pushing the outlet pipe lower or higher in the chamber.

Rendering CAD models does make it look good but the real test will be if the actual product looks anything like the rendered model :) It shouldn't be too far out as the CAD model is used to generate the toolpaths for the CNC mill.

Mike

"I would have thought the entry at the top would be most effective as the velocity of air entering the chamber would keep the particles against the wall. If there were 'dead' air above the inlet, would this not cause more turbulence with more chance of particles being held in the chamber longer than necessary? "

The CFD model showed particles were slowed down by the top when the inlet was tight to the top.  Slower particles are more likely to be pulled into the outlet airstream.  Air is probably better (more slippery) than a hard surface.  Less friction means higher velocity. 

I don't think there is any "dead" air in a top hat - it is all spinning around.  The slowest spinning air is that which is up against a surface.  At the side walls and the baffle, this helps dust to drop. 

I've never tested in 2H unit with the inlet shifted LOW, towards the baffle.  I've made a couple of them now and each one of them has had the inlet HIGH, towards the top.  The primary reason for the 2H unit with the inlet ABOVE the outlet pipe was to prevent bypass.

In at least a couple of ways the simulations don't seem to closely follow my own observations.  I do not know why.

Personally, though, I would hold the inlet towards the ceiling of the separator, rather than down low.

Mike, I agree with Phil. The computer models did not fit the observations that have been measured. I think it unwise to go along with any conclusion drawn from them. In the long run it may be that the model will show the direction we could look but it is not yet consistent with other work yet. I hope we get some more results from the model.

Hope your build goes well. Look forward to hearing your results particularly the fluffy outcome. Regards, B

Thanks for all the input guys. I will continue with the project, as is, pending any revelations within the next few days. In the end analysis, if something does come of the CFD analysis, it is not beyond the whit of man to re-design and implement any changes. Judging from all the units already made, and documented here, anything even half resembling the best practice will be far better than what I have at the moment.

I will attempt to keep this thread alive with my progress. The plugs are already drawn up for the curved inlet and I am working on the toolpaths. With any luck and a prevailing wind, I may be able to start machining the plugs this weekend.

All the best,

Mike

Quote from: phil (admin) on May 08, 2014, 12:26:50 PM
I've never tested in 2H unit with the inlet shifted LOW, towards the baffle.  I've made a couple of them now and each one of them has had the inlet HIGH, towards the top.  The primary reason for the 2H unit with the inlet ABOVE the outlet pipe was to prevent bypass.

In at least a couple of ways the simulations don't seem to closely follow my own observations.  I do not know why.

Personally, though, I would hold the inlet towards the ceiling of the separator, rather than down low.

Phil,

Any thoughts about the drop slot with a 2H unit?  It would seem the start and stop positions have less significance with the taller unit.  In fact, there was an earlier question about a continuous slot that seems like a good question for 2H units.  Of course that brings construction issues! Since the slot would most likely have to be bridged at two or three locations.

My first drop can had a round tangential inlet pipe and a "baffle wall" that extended down from the top plate. This worked very well but when I switched over from a shop vac to a DC the scrubbing became very bad. I think the effective drop slot width was too wide. I installed a Thien baffle and a rectangular inlet with a 1 1/2 inch width and I was back to or better than I was before. I left the "baffle wall" in place only cutting it down in height to match the outlet pipe projection that was the  usual 1/2D.

It would be interesting to try this floating baffle but what of the cheese grater slicing effect (I think Phil called it that)? Such a baffle plate could be suspended by four thin steel rods from the top plate much as retired2 used to stiffen his baffle.

What if the outlet pipe had a thick wall with the difference between its outside diameter and the wall of the chamber made equal to the width of the inlet rectangle? My build with a "wall baffle" plus the Thien baffle is very similar.

Quote from: retired2 on May 09, 2014, 06:12:05 AM

Quote from: phil (admin) on May 08, 2014, 12:26:50 PM
I've never tested in 2H unit with the inlet shifted LOW, towards the baffle.  I've made a couple of them now and each one of them has had the inlet HIGH, towards the top.  The primary reason for the 2H unit with the inlet ABOVE the outlet pipe was to prevent bypass.

In at least a couple of ways the simulations don't seem to closely follow my own observations.  I do not know why.

Personally, though, I would hold the inlet towards the ceiling of the separator, rather than down low.

Phil,

Any thoughts about the drop slot with a 2H unit?  It would seem the start and stop positions have less significance with the taller unit.  In fact, there was an earlier question about a continuous slot that seems like a good question for 2H units.  Of course that brings construction issues! Since the slot would most likely have to be bridged at two or three locations.

A continuous drop slots turns into a problem as the bottom drum fills.  When sawdust approaches the level of the baffle, a continuous slot really increases scrubbing something awful.

For a 2H unit I don't think the positioning of the baffle relative to the inlet is less important, just harder.  The incoming airstream is going to bounce of the walls and ceiling of the separator and then want to dig down into the drum and if the drum is full, it is going to resuspend debris.  Introducing a handful or two of bright chalk dust into a new separator should give one a good idea of where to position the baffle.  Just don't do the chalk with the filters installed as it might blind them.

Phil, my proposed design is a 1.5H unit and I can see your point about the positioning of the baffle. From the inlet, in my design, there is a small shelf that gets narrower the further it goes round the wall. This continues for approximately ninety degrees. This might suggest that I move the baffle round ninety degrees as the shelf will keep most, probably not all, particles high until the shelf runs out. Does this sound reasonably logical?

Mike

QuoteIntroducing a handful or two of bright chalk dust into a new separator should give one a good idea of where to position the baffle.

Maybe you could elaborate on the positioning? Assuming that the chalk makes a spiral pattern along the wall, is it correct to say that the solid portion of the baffle (i.e., no slot) should begin where the chalk spiral intersects with the baffle? That is, should there be a 120 degree landing zone ("runway") on the baffle for the air path to take before it hits the slot.

In my mind, the alternative is that the spiraling air path would shoot straight through the baffle slot into the waste container, which would create a lot of turbulence, and stir up  collected dust. Of course, my mental image of what's going on has often been in error ???, so I appreciate any clarification.

jdon, you're thinking about it correctly.  If the chalk demonstrates a fan or splayed pattern, you want to align the baffle such that the debris below the baffle is protected from a sudden widening airstream above the baffle.

Another question, if I may. I read somewhere on the forum that the outlet should be 10% larger than the inlet. When I was drawing up the 'curved inlet' I assumed this meant the inlet and outlet of the inlet itself and now I am questioning my belief and am wondering whether it was the outlet, from the top of the separator that should be 10% bigger. My curved inlet, from circular to rectangular, does have the rectangular section with 10% more area that the circular end.

I am about to start machining the plugs and thought I had better check first.

Thanks, Mike

retired2 suggests that the cross sectional area of the inlet rectangle should be 10% greater than the cross sectional area of the pipe from which it morphs. This to reduce losses caused by turbulence in rectangular compared with circular ducts. So you first assumption was correct. It is not that the outlet of the chamber should be 10% larger than the inlet to the chamber.

Bernard, thanks for the verification.

I've just made a couple of small modifications to the inlet so I can make better use of the materials I have to hand for making the plugs. Getting closer to some actual construction and the 2mm polycarbonate sheet should be arriving later this week. All the toolpaths are now complete and, barring any unscheduled shopping trips, I can start machining tomorrow.