J. Phil Thien's Projects

Thanks for the useful picture and PDF plan. It looks as if your rectangular inlet pipe extends right down to the baffle plate and I wonder how far is the end of the outlet pipe from the baffle? Very tidy build did you use bendy MDF for the side wall?

Have you tried having a longer section of rectangular duct between the transition and the top hat? Perhaps 12" to 24" long.  It should ensure that the flow in is less turbulent. I would like to know if this reduced turbulence improved separation or otherwise. Perhaps it works well enough for you not to need to do any more experiments but I am hoping you are as inquisitive as I am.   Regards from Wessex.

That is a dead link on my system!

I think there mus be something strange happening with your 1 1/2 HP DC. It should be perfectly capable of collecting from any of your machines! What size pipe or ducting are you using? Are there any leaks in your system? Have you got your blast gates the right way round- that is with the locking screw on the side away from the blower so that it helps to seal the flow? Is the filter on the DC clean?

Definitely not worth doing what you propose but very worth while to correct any problems above.

Good luck with the checks and let us all know how you get on.

Alan, I think it is working. Now we need to know the distance between the water levels with the manometer connected to your collecting drum. That is underneath the Thien baffle!!!!!

Regards,      Bernard

The link above is to a manometer designed for use in measuring air velocity not pressure. It is perfectly valid for that application but not for measurement of the pressure above and below the Thien baffle. Hence the manometer arms must be kept vertical to measure these pressures and NOT angled as it must be for air velocity measurement. I do not think we can measure the air velocity in a DC system using this sort of manometer because the air will be too turbulent but I would love to be proved wrong. If someone can try it and tell us all if it will work then we might gain the use of a valuable tool.

Alan, I look forward to seeing the pics. Please add a sketch of exactly where you took the measurements.

Regards,   Bernard

A "U"loop of plastic with each vertical arm 24" long should do it. I doubt if any DC will pull more than 24" of water vacuum. If it does make both limbs longer.

Fill with water colored with a little red dye to half way up and mark this point as zero. Remember that it is the difference between the levels in the two arms that we need hence 1" difference in this level will equal 1" of water vacuum. It is much easier to mark up a scale from zero every half inch. Then each division of the scale will equal one inch because one side will go up the same distance as the other goes down.

The left hand loop is traditionally connected to the container being measured and the right hand limb left open to the atmosphere. We are measuring inches of water pressure compared with atmospheric pressure.

Looking forward to hearing the result. Thanks for trying.   Bernard

Alan, You are correct regarding the size of the orifice determining the vacuum achieved. Phils' original work found that the 1.125" slot was optimal both for maximal separation and for ensuring nothing was trapped in the slot. On occasion he has also suggested that the slot be increased to 1.25" wide when shavings were being snagged. I feel that a row of holes would clog with shavings however we could have one Thien separator for these large shavings followed by another designed more for getting the fine dust out of the air. I suspect the circulating air flow "sees" the narrow slot as a single orifice. It would be worth seeing if the quarter inch right angle presented by the hardboard usually used  is optimal or would it work better if it were beveled? I suspect it is best left as a right angle but would love to know.

I am hoping that a member with a Thien and a manometer will try measuring the pressure across the baffle!

Thanks for your reply.

How the Thien Baffle Works.

We do not yet know how the Thien baffle removes wood chips and dust from an air flow. We do have considerable evidence that it does work from the posts in this excellent blog together with some information, some data and a lot of very innovative ideas. We might also look towards the evidence available from the use of cyclones both those used by wood workers and those developed for the removal of household dust such as the Dyson vacuum cleaner.

I have written here a description of how I suggest it might work. I have drawn extensively from evidence presented in this blog and from the science of fluid dynamics that I have gleaned from Wikipedia and elsewhere, interpreted in the light of my own knowledge and experience gained in my former work as a technologist. I have used some wording from these sources and I make no claim to them.

Nothing I state here has any validity until it has been shown experimentally to be correct. I do hope, however, that my contribution stimulates the experimental work that I think needs to be done. Man is inquisitive as well as essentially practical.

The Thien Baffle Top Hat.

I am considering here only the Thien baffle when made up in the "Top Hat" design. It may or may not be equally applicable to other ways of using this baffle. I have chosen this because I consider the Top Hat to be a superior method of construction and use. I have built two crude cyclonic drop can separators for use with shop vacuum cleaners that work extremely well without a baffle but stopped working when connected to a high volume dust collector. Hence my thinking here is more concerned with HVDC than HPLV shop vacuum cleaner based systems.
Bernoulli's principle states that for any flow, an increase in the speed of the fluid occurs simultaneously with a decrease in pressure and vice versa. Swiss scientist Daniel Bernoulli published this principle in his book Hydrodynamica in 1738
Bernoulli's principle relates to the principle of conservation of energy. This states that, in a steady flow, the sum of all forms of mechanical energy in a fluid along a stream is the same at all points on that stream. Of course this is theoretical and assumes no losses due to friction from the stream container.
If a mixture of air and dust is flowing horizontally in a section of pipe or tube, particles are subject only to their speed and their weight, if the speed increases it can only be because the fluid in that section has moved from a region of higher pressure to a region of lower pressure AND if its speed decreases it can only be because it has moved from a region of lower pressure to a region of higher pressure. Consequently, within a fluid flowing horizontally, the highest speed occurs where the pressure is lowest, and the lowest speed occurs where the pressure is highest.
The gases in air have very little weight compared to the wood chips or even the tiny grains of extremely fine dust. As our flow enters the circular chamber of the separator the wood debris will be thrown against the outside wall because they are heavier.
These heavier parts of our flow require more energy than the air does to flow at the same rate and as the energy input is essentially fixed they will not move as fast as the air so will fall downwards and this will be increased as these particles will lose some of their energy because of the friction against the chamber wall against which they have been thrown because of their greater weight. The rough nature of the wood debris is not as slippery as the air so the effect of this friction will be greater on them than on the air.

A curved air stream will develop a pressure gradient perpendicular to the direction of flow, with the lower pressure on the inside of the curve. Bernoulli's principle predicts that the decrease in pressure is associated with an increase in speed. So that as the air passes around the separator it slows down towards the outside wall of the chamber. This means the chips and dust will tend to fall even faster than they would in a laminar flow.

Inlet Pipe Design.

Air travelling along a round pipe will tend to rotate particularly if it is made by welding or folding a length of steel sheet into a pipe with spiral walls. It will also occur due to the spin of the earth and if a rotary blower is used. A rotating flow is highly complex consisting of many smaller turbulent flows. When introduced to a cyclone separator such disturbed flows reduce the efficiency of the separation of the dust from the air. By changing a round pipe into a rectangular section tube, by means of a transition adaptor, the rotation and its associated turbulence can be halted and if the rectangular section is long enough a steady laminar flow achieved. This is presumably why ventilation ducts in buildings and fume hood trunking in laboratories are rectangular rather than round in section. Recently some dust collectors have begun to use a rectangular duct, between the blower and the simplistic system that directs the flow either to the filter or the debris collecting bag, for the same reason.

Bernoulli's principle as applied to the Thien Top Hat separator above assumes that the airflow is horizontal and axially laminar hence for it to work well the flow into the chamber needs to be as axial and laminar as possible. This means it must be as free of rotation and its associated turbulence as we can get it. Hence, in my opinion the inlet pipe needs to be rectangular in section.

We know that the optimal width of the drop slot for the Thien separator baffle is 1.125 inches. If the inside width of the rectangular inlet duct is also 1.125 inches then it will contain the highest possible concentration of wood debris immediately above the slot. Bernoulli effects as described above should then have the greatest possible opportunity of forcing all of the debris into the collection container under the baffle. Clearly the cross sectional area of the rectangular section needs to be the same, or slightly more, than that of the round pipe from which it flows.

Water Pump Effect.

Our wood burning stoves work well because a draught of air is drawn up the chimney stack by air currents passing horizontally across the top of the chimney liner. This can be improved by adding a rain deflector over the top of the stack because it creates a more laminar air flow. In other words we have a tube held at right angle to a laminar air flow.

Consider the humble Bunsen burner. Here we have a vertical flow of flammable gas passing up a burner tube past a hole in the side through which air is drawn because of the Venturi effect which is intimately associated with the Bernouilli principle. The water pump, another piece of common laboratory equipment, works the same way except in this case the fluid flow is water from the main water supply passing across the top of a tube at right angles to it that creates a considerable vacuum in that tube drawing in gas or fluids.

Similarly if we present a laminar flow of air and wood debris passing across the top of a tube held at right angles to that flow then a vacuum will be created in that tube and some of the mixture, that closest to the pipe will be drawn in. I suggest that the drop slot in a Thien baffle is just such a "tube" albeit rectangular and being much longer in one direction than the other. If the mixture closest to this tube is richer in wood debris then more of this will be drawn below the baffle than air.

I suspect that if we can accurately measure the pressure above and below the baffle then we would find that we have a vacuum in the collection container. This may explain why quite a number of contributors have noticed their tin collection can collapsing under the pressure of the atmosphere pressing against their outside walls. I suspect that this collapse occurs less when the system is dead ended as little or no air is passing the drop slot. The implication may be that the collection chamber IS critical after all in that it must be rigid enough to be able to withstand this pressure to achieve the greatest separation effect.

Other Possibly Relevant Effects.
Bernoulli's principle can also be derived directly from Newton's 2nd law. If a small volume of fluid is flowing horizontally from a region of high pressure to a region of low pressure, then there is more pressure behind than in front. This gives a net force on the volume, accelerating it along the stream. I do not yet see how this affects the Thien baffle but I suspect it may and I will think on.
Conclusion.
I have been trying to stop myself thinking about and carrying out experimental work on the Thien baffle because I thought I had moved on from my previous existence and would rather smooth wood these days. However I cannot resist it and when I have completed some of the jobs on my list, a daughter needs her shoe storage cabinet after all is said and done, then I intend to examine rectangular inputs; outlet pipe EXTERNAL diameter; height of outlet pipe and the depth of the baffle under the end of the outlet tube. I can really only afford to do this for one diameter of Top Hat, probably about 20" and as my Clubs' system uses 4" diameter spiral steel tube and a 1 ½ KW blower I am also stuck with these fixed factors. It would be great if someone else could repeat these tests with different separator, blowers and/or tubing diameters.

I have posted a proposed standardised testing method elsewhere on this blog and will confirm this and the results obtained on this blog if I may do so.

I need to build a Thien top hat separator so there is no reason why I should not include a large removable flange in the top plate so I can swap in different outlet pipe configurations. Building it so that I can have the baffle height adjustable is somewhat more difficult but I might be able to do it. I will scratch my head for a bit.

Building different widths of top hat is further than I want to go so I will leave that to other people.

I agree that feeding from different machines in a standard way that mimics typical workshop operations would be a very interesting test. However to keep it a bit simpler I will make up a large quantity of different wastes and mix them thoroughly to create a standard mix that I can keep by me for testing. I will then feed them to the system in a standard way. I will weigh how much material is fed in and how much is contained in the Thien collection can and the difference will be a reasonably accurate measure of efficiency.

Much as the past draws me to want to measure pressures and air flows I am not going to do that as I want to stick primarily to hand woodworking.

Happy wood smoothing.

Thanks for your response alanm.

Ref: My point A. I am suggesting a 24" long rectangular tube with a section of 1 1/8th" by 5 5/8" inside dimension giving a cross sectional area slightly greater than that of our systems 4" tube.  The transition between the round pipe and the rectangular tube would need to be as long as possible and at least twice the diameter of the round tube. This should get us a turbulent free input system as we can achieve.

Ref: My point C. I am suggesting that the gap between the outer wall and the outlet tube be reduced to 1 1/8th" regardless of the outside diameter of the outlet pipe. Hence the outlet pipes wall would need to be increased by setting it into a cylindrical block of wood to make a "thicker" wall. The purpose of this would be to maintain the greatest possible turbulent free air flow circulating around the outer walls while still having a separator wide enough to have a reasonably good air flow. With a bit of luck the air bleeding over into the bell mouth of the outlet pipe will tend to push the dust and shavings down towards the baffle and through its slot.

I am stuck with a 4" metal collection system as our Club has just installed it to collect from all our wood shredding machines. As the fan inlet to our new HVDC is 5" I connected a simple drop can with a tangential inlet pipe between it and the 4" collection system. This had worked well when we were just using some 2" hose and a shop vacuum cleaner (LVHP) but not with the new 4" metal system and the much greater volume of air moved. Some chips now get scrubbed out of it and the fine filter we have fitted gets quite a lot of dust on it after a short time. Hence we need a cyclone or a Thien separator. I would rather not use a cyclone because of the air movement loss cause by the cyclone and by the long pipes needed to get up and down from the top. Hence my prolonged study (several sleepless nights) of the very interesting posts on this site.

I am trying to tie down some details so I can make one and at the same time hopefully add something to what is known about making one so I can share it with you all.

I am also trying not to revert to my former life as a research technologist as I really only want to work wood now. My experience suggests that a series of experiments with a standardised collection source and standardised material being moved needs to be done to test various elements of the Thien separator. This would allow us to tease out the critical point which is how much of the fine dust is caught by the separator and how is this affected by various configurations. Phil has done some of this I know and if I still worked in a university it would be easy to get this done as a student project.

Disposing of the waste collected needs to be done with great care as tipping dangerous fine dust into a plastic trash bag or venting it outside is not just irresponsible it is potentially dangerous to the operator and to others. Damping this down with water is not an option as there would be a danger of it igniting spontaneously. Ours gets given to the chickens that one of our member's keeps who are kept happy playing with it and who mix it with their droppings making it good for composting while the birds keep cleaner.

Regards Bernard (sometimes known as prof amongst other less positive names though I have never been one.)

Completely fascinated by the interesting ways that Phills' important discovery has been modified as wood workers attempt to probe the science behind it in a very practical way.

1. We are creating a circulating doughnut of air but mismatching a round inlet pipe to a rectangular cross section. That may be why this pipe needs to hug the top. Hence there will be quite a lot of turbulence probably reducing efficiency. If the inlet tube is also made into a rectangle and extended far enough back, to where it morphs into the round collection pipes, to produce a minimally turbulent linear flow we may improve the removal of fine dust.

2. We are relying on the outlet pipe to form the inner layer of our doughnut and know that the height of its end above the baffle is a key factor. I suspect that setting the outlet pipe into a larger diameter cylinder (of wood?) would help reduce turbulence and increase airspeed that we know would probably improve performance.

3. If a fluid moves across the orifice of a tube held at right angles to the fluid flow, then a negative pressure is created inside the tube.  (water pump effect dating to Roman times.) The drop slot is just a wide narrow tube and the vacuum created is what draws our dust and chips into our can.

I suspect that I may be able to get very good dust and chip extraction from this airflow by:

A. Taking my 4" inlet pipe and connect it to a 24 " long rectangular inlet pipe of slightly larger surface area then I will get good untroubled airflow into my separator.

B. Making the inlet section width 1 1/8" ID wide to match the drop slot.

C. If my 5" outlet pipe is set inside a cylinder of wood so that the gap between the two walls is 1 1/8th "

D. Set my outlet tube plus thick collar so its end is > 2 1/2" from the baffle plate.

E. Arrange for the bottom end of the outlet tube to line up with the bottom of the rectangular section input pipe.

F. If I form the end of the outlet tube assembly into an aerodynamic bell mouth.

G. Make the inside diameter quite small, say 12", to get a high circulation velocity and hence a greater vacuum.

G. Listen to the feedback I hope I get from this post and make my Top Hat as smooth and airtight as I can.

If I have stolen your ideas please forgive. Looking forward to your usual honest feedback.