Thread: Fibreglass Cyclone Project

Originally Posted by Muchacho

As promised, I knocked up a simple manometer to my fibreglass cyclone system, with a pressure tapping just upstream of the fan inlet and took a number of readings in the following configurations.

1) Inlet sealed, fan only at zero flow 6.5"

2) Fan, cyclone, 2.5m pvc and 3m flex, fibreglass fitting and thicknesser 3.4"

3) Fan, cyclone, 2.5m pvc and 3m flex 2.4"

4) Fan, cyclone, 2.5m pvc (no flex), 1.25"

The above setups also include the 400 x 600 louvred outlet through the shed wall.

The first reading was not really any surprise, as this seems about the standard 1000 cfm as per BobL's previously published chart.
I'd be interested if anyone has a pressure vs flow chart for a 2HP fan. Would be interesting to see if the 3.4" reading is anywhere near the suposed minimum 800 CFM at the machine.
The final reading has blown me away. According to Bill Pentz, a cyclone built to his design with neutral vane and inlet ramp should be about 2.25".
Have to say this has made my day as it appears that I may have met my objective of making an efficient system to compensate for the less than ideal fan.
Now, what would this thing do with a 3HP fan ???

To make sense of these readings we need to know exactly what is open and closed.
e.g. for the first reading, is the other side of the impeller fully open, i.e. is disconnected from the cyclone?

Static pressure readings only make sense if there is no flow so all the intakes should be closed and only the other side of the impeller or cyclone should be open.
As soon as there is flow bernoulli kicks in and the results can be meaningless.

Originally Posted by Muchacho

Now, what would this thing do with a 3HP fan ???

Well if it was in my shed and anything else was running at the time it would probably pop the circuit-breaker .

Doug

Originally Posted by John Samuel

Bob,
The table got messed up when I pasted it. I have reproduced it below as a picture.

OK got it.

210 mm of WC at 50Hz does not sound right for such a big system. Thats about 8.5" of WC which is the same as generic 2HP DC generates and will move a max of about 1250 cfm through a 6" pipe. This says to me you have leaks. Almost certainly part of these will be all the blast gates.
The only way to be sure of getting a true reading is to disconnect all the ducting and blank off the cyclone inlet.

The data taken for the machines is with the blast gates open and air flowing through that machine.

These results do not mean anything as once there is flow the bernoulli effect will kick in and mess up the reading.
All the measurements need to be done with zero flow. So block the duct off at the machine and close the relevant blast gates and that will tell you what pressure you have available at that machine. Then you can use the attached chart to determine the max flow. Bear in mind that this is the max flow. Even a naked pipe end imposes a restriction on flow so it will be less that the max. A machine even with 6" inlet/outlets will impose more restriction etc. The only opening that imposes near zero restriction to flow is a bell mouth hood.



The effect of increasing the pressure is not linear. Increasing the pressure by a factor of 10 increases the flow by a factor of 3, doubling the pressure increases flow by about 40% so a 40% increase in pressure only increases the flow by ~25%.

Originally Posted by doug3030

Well if it was in my shed and anything else was running at the time it would probably pop the circuit-breaker .

Doug

Don't get too excited - lets wait and see what happens when we are sure about the measurements. His max pressure is only 6.5" - that is significantly less than a generic 2HP DC.

OK,

I disconnected the ductwork, except for the foot or so of flexy which is attached to the cyclone inlet. To that I connected about 18 inches of 6 inch pipe with the end sealed off. Into that piece of pipe I put the PVC manometer tube, and sealed it in with silicone.

The data did not change. 297mm and 210mm of water at 60 Hz and 50 Hz respectively. If there are any leaks there, I can't find them. So, these data suggest the CV blast gates are pretty darned good because I got the same readings by just tapping into the ductwork.

Bob's chart is truncated at the top, so I could not read the exact outcome in CFM, but it seems enough to do a good job.

The measurements with air flowing through the machine tell me nothing about static pressure, but do help me understand how much air flows through the machine, and might warn me about a restricted airflow. The lower the inches of water, the higher the air flow is. The drill press shows 145mm of water, only a little more than an open pipe much closer to the cyclone, but it is little more than an open pipe itself. The table saw shows 155mm of water, indicating a high air flow, but that is to be expected because it has a 150mm and a 100mm line. The linisher, drop saw and drum sander are quite close to each other. Because the drop saw port is simple and open, it seems to draw a bit more air. So, these measurements all agree with my expectations, and indicate all these machines are getting a reasonable air flow. If one of them was showing a significantly higher measurement in inches of water than the others, it would likely be an indication of restricted airflow.

Originally Posted by BobL

Don't get too excited - lets wait and see what happens when we are sure about the measurements. His max pressure is only 6.5" - that is significantly less than a generic 2HP DC.

I was responding to Muchacho's suggestion to run his cyclone on a 3hp system if you look back. I really dont think I can run two 3hp appliances at once.

OK I think we are in a position to perhaps leave the graph and use an equation () which will allow the flow to be calculated at any pressure.

For 4" ducting the relevant equation is
Flow (in CFM) = EXP(0.5491*(LN(Press in inches of WC))+4.80052)
If you uses Excel you can use the equation " =EXP(0.5491*(LN(X))+4.80052)" , where X is the Press in inches of WC

So at 10" of WC the flow is EXP(0.5491*(LN(10))+4.80052) = 430 CFM

For 6" Ducting the equation is
Flow (in CFM) = EXP(0.5358*(LN(X))+5.9365)

So at 10" of WC the flow rate is 1300 cfm and at 12" it is 1433 cfm.

Please note this is the MAXIMUM flow rate assuming a short, non-tortuous path and nothing like a machine in the way.

John, the fact that you get the same pressures whether you include the ducting or not is indeed a good sign of no leaks.
While I agree that if the dynamic measurements (measurements done with air flow) were to be high that would indicate that there was a blockage and while they might correlate with expectations using these pressures in the above equations to attempt to quantitatively determine flow will not be accurate (have a look at what BP says about using a simple manometer to measure flow rates). A simple manometer is indeed simple and does not measure accurately or linearly across a range and in practice they will need calibrating. Even calibrated units like Maghellenics are +/- 10%. Simple manometers cannot even be used reliably for comparative measurements in a dynamic situation because the pressures generated in various flows are sensitive to where in the air stream the probe is, self generated turbulence and the orientation and shape of the test probe. Assuming you have left the pressure probe in the same place and have not moved it around between measurements your relative measurements are probably OK but I would not want others to think this can be a reliable method. Static (no flow) pressure measurements can really only be used for leak detection and to determine a maximum flow. To accurately measure the flow a pitot tube should be used to measure the pressure differential inside the air stream and this needs to be done systematically across the duct diameter to include wall boundary effects.

For folks reading this and think - wow my 2HP DC generates 8.5" so I'm doing pretty well - this is only one part of the picture.
What matters is now is what flow the impeller can maintain as restrictions like long ducting and throttles machines kick in.
A smaller impeller may in fact be able to generate a decent pressure but as soon as restrictions kick it cannot overcome these and the flow drops away - this is why a complete fan curve is needed to really understand what is going on.

Bob's chart doesn't indicate which pressure, total, static or dynamic, this would be good to clarify.

Concerning Bernoulli at a guess most of us would have our machines on the same level as our DC so I reckon that part would be next to nothing, the velocity (dynamic) part is indeed important and is used in conjunction with static pressure when using a pitot tube to indicate total pressure, this then gives us flow via some calcs, (or chart) which leaves me a bit confussed when Bob says static pressure measurements are meaningless, just thinking about this, it would make most sense if indeed the chart pressure values were total pressure and this would then make a purely static pressure test value and then reading off the chart a bit erroneous, this might be what Bob is meaning



Pete

Originally Posted by pjt

Bob's chart doesn't indicate which pressure, total, static or dynamic, this would be good to clarify.

It's for static pressure and smooth ducting with zero length, no twists or junctions etc.
It's the maximum flow one could expect using a short length of smooth straight pipe - I think that chart is specifically for smooth ally pipe - but that should not matter as smooth ally and smooth PVC have similar resistance. I think it also assumes a blunt end openings to the pipe - blunt ends generate a small amount of resistance.

Concerning Bernoulli at a guess most of us would have our machines on the same level as our DC so I reckon that part would be next to nothing.

The Bernoulli effect I'm referring to is the change in pressure inside a tube due to the flow of fluid past the tube opening.

To measure the static pressure with no flow the sensing tube can be placed in any orientation as shown below and you will get the same reading.

As soon as the air moves you will get a different reading dependent on a bunch of factors relating to the shape and position of the tube opening.
The distance into the tube will be one such factor, as the air moves faster in the middle of the duct the Bernoulli effect will be greatest in the middle, but even having it flush against the wall the Bernoulli effect will still be significant


The tube may indeed be cut at right angles to the long dimension but even a whisker of difference as shown in exaggerated form at A and B will make a big difference in the readings.
Even a fleck of dust caught on the tip of the sensor will be enough to make a difference.
This is also why pitot tube are made small, with a specific rounded shaped ends, and oriented into the flow in exactly the same way every time (preferably using a hard mechanical stop) and MUST be calibrated to give accurate results.

If a pressure sensing tip is installed and left in place the relative results obtained may provide qualitative (ie this is more/less than this) but at best to +/-10%. However the actual pressure reading obtained still cannot be used with the static pressure chart to determine an actual flow rate.

, the velocity (dynamic) part is indeed important and is used in conjunction with static pressure when using a pitot tube to indicate total pressure, this then gives us flow via some calcs, which leaves me a bit confussed when Bob says static pressure measurements are meaningless, just thinking about this, it would make most sense if indeed the chart pressure values were total pressure and this would then make a purely static pressure test value and then reading off the chart a bit erroneous, this might be what Bob is meaning

A static pressure test (ie no flow) will provide an accurate assessment of the maximum flow of a system (and is useful for determining leaks) using the chart.

I sometimes refer to the flow rate chart when I measure actual flow rates of straight lengths of specific diameter pipes or openings using an anemometer (which measures air speed directly) - usually my results agree with the chart but not always. My readings are usually less than what the chart shows, probably because of measurement difficulties due to turbulence and in measuring the wall effect.

By total pressure I assume you mean that read by a pitot tube?
This is the difference in pressure between the pressure of the air directly into the air stream (sometimes called velocity pressure) and the pressure at right angles to the air flow.
Measuring these pressure accurately runs into all the problems I refer to above.
Producing a chart of these pressures for general use is of no value because unless you have a calibrated pitot tube they cannot be measured.
In practice most pressure based air speed measurement systems like Pitot tubes are calibrated to convert whatever pressure they read directly into an air speed and do not go through any absolute pressure algorithm step.
EG the maghallenic pressure/air speed uses something like a FPM = Constant * SQRT(Pressure reading/air density)

Even if you could caibrate you own pressure sensor tip it is still subject to the vagaries of the above problems when the sensor is moved around.

And finally the total pressure reading still does not take into account the duct wall/boundary effect. Bill Pentz does a bit of hand waving about this and says something like deduct 10% for this effect but this is tad sloppy in my book especially for vacuum cleaners and higher velocity flows.

Bob,

Thanks for the formula. I plugged it into a spreadsheet and got (=EXP(0.5358*(LN(12))+5.9365)) = 1434 CFM

Can you enlighten me about the figures 0.5385 and 5.9365? Are they a function of the pipe size?

I'm aware of the limitations of a simple manometer with air flowing. Still, the fact that the pressure readings when running a machine group closely together whilst showing the deltas I would expect given the number of ports etc is gratifying. As you suggest, the numbers themselves mean nothing; it is the shape and pattern in the data that interests me.

I still might have a rush of blood and get a Pitot tube. It would be nice to get some meaningful dynamic CFM data.

It seems to me you nailed it when you said, "What matters is now is what flow the impeller can maintain as restrictions like long ducting and throttles machines kick in." Methinks this is one area where the CV impeller comes to the fore. When I see those off cuts being whizzed into the ductwork at the table saw and drop saw, I'm pretty sure the dynamic air flow is pretty good.

Used BP's Static Calculator spreadsheet to work out losses over my longest run (linisher) at 1,000 CFM and got 8.3 inches. The calculated velocity was 5093 ft/min. For the drop saw the pressure loss was 7.2 inches and the velocity was identical at 5093 ft/min. So, for the linisher I need a static pressure of 8.3 inches of water to maintain 1,000 CFM, and I have nearly 12 inches of WC. By lifting the CFM required to 1250 CFM I got 11.6 inches of WC, so 1250 CFM is the estimate of the flow through the linisher. Using the same method, I got a flow of 1400 CFM through the drop saw. According to the formula you provided my max CFM is only marginally more than this.

Originally Posted by John Samuel

Bob,

Thanks for the formula. I plugged it into a spreadsheet and got (=EXP(0.5358*(LN(12))+5.9365)) = 1434 CFM

Yep that sounds right for 12" of WC and a 6" pipe.

Can you enlighten me about the figures 0.5385 and 5.9365? Are they a function of the pipe size?

The figures are empirical values extracted from the pressure/flow graph and yes they are pipe size (and also wall smoothness dependent, which in this case is for smooth ally so they don't really apply to something like )flex

I'm aware of the limitations of a simple manometer with air flowing. Still, the fact that the pressure readings when running a machine group closely together whilst showing the deltas I would expect given the number of ports etc is gratifying. As you suggest, the numbers themselves mean nothing; it is the shape and pattern in the data that interests me.

Yep I agree.

I still might have a rush of blood and get a Pitot tube. It would be nice to get some meaningful dynamic CFM data.

slippery but interesting slope

It seems to me you nailed it when you said, "What matters is now is what flow the impeller can maintain as restrictions like long ducting and throttles machines kick in." Methinks this is one area where the CV impeller comes to the fore. When I see those off cuts being whizzed into the ductwork at the table saw and drop saw, I'm pretty sure the dynamic air flow is pretty good.

I like it!

Used BP's Static Calculator spreadsheet to work out losses over my longest run (linisher) at 1,000 CFM and got 8.3 inches. The calculated velocity was 5093 ft/min. For the drop saw the pressure loss was 7.2 inches and the velocity was identical at 5093 ft/min. So, for the linisher I need a static pressure of 8.3 inches of water to maintain 1,000 CFM, and I have nearly 12 inches of WC. By lifting the CFM required to 1250 CFM I got 11.6 inches of WC, so 1250 CFM is the estimate of the flow through the linisher. Using the same method, I got a flow of 1400 CFM through the drop saw. According to the formula you provided my max CFM is only marginally more than this.

The problem with static Calc is estimating the pressure losses at the machine ports since these cannot be calculated. The pressure you measured for the linisher was 170 mm or 6.7" which at 60Hz will give a flow 0f 1048 CFM - so my guess is the real flow might be somewhere between the 1250 and 1050 - whatever it is, is it good.

A good test of any pressure/flow data is if the flow data from two branches add up to be less than the max flow.
eg for your TS you have 155 mm or 6.1" WC - using the formulas that I provided at 60 Hz you would have 328 CFM at the blade guard and 997 CFM for a total of 1326 CFM which is within 10% of the max flow rate of 1434 cfm so I would rate that pretty good. When I see calcs that add up to be less than the max flow then I am more likely to believe this - it's when the sum of 2 branches add up be more than max flow that means there is something wrong.

It would be interesting to know where the tip of your manometer is in the air stream and the shape of the tip.

Originally Posted by BobL

It would be interesting to know where the tip of your manometer is in the air stream and the shape of the tip.

The tip of the manometer was cut square with a Stanley blade and inserted at 90 degrees to the duct so that the tip protruded into the duct either not at all or, more likely, only one or two mm into the duct. I was trying hard to avoid adding turbulence when taking dynamic measurements.

When designing the ductwork, I had only one 90 degree bend in the main line, which is why I used 6X15 degree bends ... to reduce pressure loss to a minimum. To avoid putting a 90 degree bend down to the linisher I used two 45 degree bends with 3 or 4 feet of straight pipe between them. I don't know about air, but I recalled a class from my teens where we were taught that such a set-up creates less turbulence in water, so I thought it was worth a shot. I am very happy with the ductwork, and am confident that given the limitations of my shop is is basically a good design it terms of limiting pressure loss.

Ah yes cheers Bob, It all becomes clear, a tube inserted into the pipe with all gates closed (no flow) will effectively measure the pressure (Pascal's law) but as soon as there is flow (a gate open) Pascal goes out the window and Bernoulli throws in a bit of velocity pressure which could either drive down or suck up the water level simply because of a slight angle into or away from the direction of flow, so for these simple manometer tests best bet is insert tube just thru duct wall at 90° with a clean cut on end of tube and any flow values we obtain from the chart use for comparison within our own system if we make mods or changes, but to make a comparison with another system based on flow values obtained from the chart best be taken with a grain of salt.



Pete

Originally Posted by John Samuel

The tip of the manometer was cut square with a Stanley blade and inserted at 90 degrees to the duct so that the tip protruded into the duct either not at all or, more likely, only one or two mm into the duct. I was trying hard to avoid adding turbulence when taking dynamic measurements.

Humm . . . . . . it's difficult to know what this setup measures.

My guess is that it measures residual impeller pressure plus a bernouili effect, even though the manometer tip is near the wall, the fact that is perpendicular to the flow means that there will still be a significant bernoulli effect from the air rushing past the tip.

Unfortunately it also won't avoid turbulence as the flows are well over the few thousand FPM even close to the wall.

Despite what I said in my previous post a clue that the pressure values can't be used with the equations are the results for all blast gates open results - if you plug those pressures into the static pressure equations you will get less than the flow for any machine which is not right. Some (probably most) of this will be bernoulli pressure and some will be residual back pressure from the impeller interacting with the ducting.

Muchacos idea that the static pressure will be lowest when the flow is greatest is correct. Unfortunately the system is nether static, and the pressure being measured using these simple manometers inside moving air flows is not the static pressure but something else.

I went through all of this when I first started playing with a Pitot tube and it got very complicated so I gave up especially as I just got a hold of a nice hot wire anemometer.