Inlet shape - round or rectangular?

[retired2]

Phil,  I'm getting close to building my first top-hat separator, and was wondering if any of your testing included variations of the inlet shape?  Most everyone goes to great pains to insert a round inlet pipe tangentially and then shape the end to match the inside diameter of the separator.  This makes perfect sense since the separation process is dependent upon spinning the dust laden air stream around the inside perimeter of the separator.

Since the objective is to get the waste against the outside wall, would it make more sense to use a retangular inlet that is taller than wide.  Also, key to this concept would be an "eccentric" transition piece that would go from round to rectangular, keeping the area roughly constant.  For example, my piping is 5" diameter so I was thinking of transitioning from 5" round to 5"x4" rectangular, or even 5.5"x 3.5".  My thinking is the transition piece to rectangular moves the air stream closer to the outside wall before it enters the separator body. 

An eccentric transition piece might be more difficult to fabricate, but it seems if a rectangular inlet is better, then the transition is optimal if it is eccentric rather than concentric.

Comments?   

[phil (admin)]

I've been thinking about this quite a bit lately.

When dust travels down a pipe on the way to the separator, it hugs the wall and spins in a helical fashion.  It hugs the wall because the air is moving slower at the wall than in the center.  In spins in a helical fashion due to random currents that setup a rotation that, once started, prevails (kind of like water that spins in a circle as it flows down your drain).

So imagine dust travelling down a round pipe and then transitioning to the rectangular section that is the cyclone's inlet.  If we're lucky, the rotation within the round pipe will be such that it will hit the part of the rectangular inlet closest to the outside wall of the separator.  That would be optimum, as we want the dust to hug the outside wall of the separator.

But of course, the currents within the pipe are random and we cannot predict a positive outcome in this regard.

One thing we could do, however, is have a slight bend in the round pipe before it gets to the rectangular inlet.  If we could curve the round pipe over the three feet or so before it hits the rectangular portion of the inlet, so the pipe basically becomes an extension of the wall of the separator, we might expect that we'd have exactly what we want.  That is, the dust would stop rotating (if we have enough of a curve) while still hugging the wall of the pipe and then hug the outside wall of the rectangular transition and then hug the outside wall of the separator.

How much we'd have to curve is the question.  I imagine that, over three feet of pipe, we'd be looking for the middle (18") to be displaced about 8" or so (just a guess) from the ends.

Does any of that make any sense?

I know conventional wisdom is that the pipe feeding a cyclone should be straight for several feet before entering the cyclone's inlet.

I'm all about being unconventional.  I'm pretty sure a slight bend to the pipe would be superior.

Finally, it is the case that the rotating air inside the separator will want to drive the incoming dust against the wall.  But there is some turbulence at that point.

I'm splitting hairs a little bit.  The difference in separation performance between rectangular inlet and round inlet is probably so small as to be difficult to measure.  And the rectangular would really benefit from the slightly bent incoming pipe.  But would probably outperform the round inlet in either regard.

I'm going to stop now, I'm really rambling...

[retired2]

Phil,  Amazingly, my original question was going to be a two part question, but I decided not to complicate things till you answered the first part.  Well, you answered both parts. 

The second question was of course, what effect a turn ahead of the entry point might have?  But your response showed some deeper insight that I had.  I wasn't aware of the helical behavior of the air stream, but I was thinking that a turn results in the heavier particals being thrown to the outside wall.  Maybe that is an over-simplification and it really doesn't work like that.

Whatever the behavior of the dirt laden air in a turning pipe, if it is known and predictable, then surely some turns are beneficial and some are counter productive.  I would guess that most of us have some kind of change in direction just prior to entering our dust collectors or separators. 

Chuck,  I think an implementation like you sketched might produce too much turbulance and therefore be counter productive.  I was thinking of an external transition piece that bolts to the mouth of the separator.  I would not make any changese to the typical top hat design other than the inlet port would be rectangular in shape rather than round.  The transition would be smooth like the one on top my thickness planer and it's length would be determined by the fabricator. 

[galerdude]

On my top hat configuration the inlet is approximately square and I'm pleased with it for two reasons; one, it was way easier to fabricate and two, it appears to work really well.

Thanks,
Gale

[retired2]

Chuck, your concept might well be the next generation separator, the "Thein2"!!  It incorporates a rectangular inlet and Phil's idea of a turn.  Of course, I'm not sure we know what the turn should look like, i.e. what radii, what direction, and how far from the inlet?

When I get a little closer to actually cutting wood, I may take a drive over to my sheet metal fabricator and see how much a transition piece would cost. 

[Vodkaman]

Fascinating discussion.

We bend over backwards to eliminate turbulence, rightly so, but sometimes vortices can be used to our advantage. Consider this for discussion purposes:



Maybe a block is a bit too much, a bump may be better. If you force or compress the flow inwards, after the bump, it will spring back. Positioned correctly, its inertia will force the incomming flow towards the outside of the chamber.



Dave

[retired2]

The "vortex trip" might work, then again it might act like a snow fence.  If you live in the part of the country that uses these fences you know they cause the snow to drop out and drift on the downstream side of the fence.

As Phil points out, there are a number of ideas that might provide improvements in performance, but they are likely to be small and very difficult to measure or quantify.  For sure the improvements cannot be calculated with mathmatically formulas, so that means the ideas must be tested the hard way - build it!

[Vodkaman]

Thanks Chuck.

Retired2 - I do agree. Quantifying results is the problem. I don't think a bank of pressure gauges is going to help either. I don't think anything can beat measuring dust in and dust out, repeated several times for consistency. I just wish I had the funds, as this is the kind of stuff that I love doing, ideas, building, prototyping and testing is my full time hobby. I am an amateur inventor, which means I don't get paid.

Even though we cannot prove any of the ideas, I still think that we should continue to collect them. Occasionally someone will suggest and idea that maybe seems insignificant, but will jump off the page. Also ideas generate ideas. So I say keep 'em comming.

Dave

[retired2]

Thanks Chuck.

Retired2 - I do agree. Quantifying results is the problem. I don't think a bank of pressure gauges is going to help either. I don't think anything can beat measuring dust in and dust out, repeated several times for consistency. I just wish I had the funds, as this is the kind of stuff that I love doing, ideas, building, prototyping and testing is my full time hobby. I am an amateur inventor, which means I don't get paid.

Even though we cannot prove any of the ideas, I still think that we should continue to collect them. Occasionally someone will suggest and idea that maybe seems insignificant, but will jump off the page. Also ideas generate ideas. So I say keep 'em comming.

Dave

Vodkaman, your comments make me chuckle because they put you among a group of people in this world, me included, who have a compulsive need to understand how everything works whether we truly need to know or not, and we are always on a quest to make them work better whether they need to or not!

I don't know how many times in my life a new aquaintence asks me "Are you an engineer?"  Most of the time it is asked in a way that is not necessarily intended to be a compliment.  It is more like "If you are an engineer that would explain your nutty obsessive complusive behavior and your need to fix things that ain't broke!"

[Vodkaman]

 :D me in a nut shell. Proud to be an engineer.

Dave

[tvman44]

retired2, I am kind of like you I spend more time figuring out how things work than is necessary but love doing that.  I am surely no engineer.  Just always made my living fixing things.   8)