Just don't forget that apart from some basic training in HVAC systems I'm just a few pages of the book ahead of (some of) you guys - it's been so long since engineering school that i can't remember very much.
The 'constant' factor in fan performance is perhaps that if you reduce it to the maths you can say that the fan HP is proportional to the CFM x pressure drop. (x = multiplied by) Using the word 'constant' in the mathematical sense of its just being a fixed number - the proportional sign in a maths equation can always be replaced by an equal sign and a constant. So you can say that:
fan HP = a fixed number x (CFM x pressure drop)
This tells us quite a lot, in that e.g. it shows that if you want to run a small bore system at very high air speeds (which creates a large pressure drop) so that you still get a decent CFM that it takes quite a lot of extra HP to overcome the extra pressure drop - even though you still will have only the same CFM available at the machine. It's also saying that the reverse is true - that if you want to maximise the CFM for a given fan HP you need to minimise the pressure drop, which means maximising the duct size. (with the proviso that we can't drop below 4,000fpm and still expect reliable chip and dust transportation - which is more or less the design basis of all of our dust collection systems)
The catch is that the formula isn't all that useful in the everyday use sense for designing air systems. That's because the constant is a composite of several different numbers (fan efficiency, motor efficiency etc, and the pressure drop is sometimes expressed in terms of factors which relate to the system/ducting) which are not easily available without doing a lot of testing, and which unless you're very sharp with a pencil are easily got wrong. There's a more complete version of the formula is here http://tiny.cc/mqyh8 - plus some links and a good discussion of how the various factors influence system performance.
The result of this is that engineers tend for convenience to use the sort of tabled information pointed to in the posts above (fan curves, estimated system pressure drop based on ft run of ducting) which numbers have long since been pretty much the standard throughout the ventilation industry.
When for example you go into a 3,450 rpm fan curve or fan table and extract CFM and HP at the total system pressure drop you are expecting you are accessing what is probably a test derived number which by definition includes all of the above. Which means actually that you are using numbers based on e.g. a presumed typical motor efficiency.
Your motor may not be quite the same. There may be small differences too if you are using an industrial makers fan tables to predict the likely performance of the actual fan you have. Similar issues arise too when estimating the pressure drop for your ducting from tables.
Which is why these methods are close but not 100% accurate. My dust system for example is drawing a bit under 4HP with one gate open, whereas similar layouts in the US seem to draw something a little over 4HP. The total pressure drop is probably a little lower than most (short duct runs, few bends, 160mm ducting), so it's probably not down to the ducting being restrictive. It's very likely instead that the low amps/HP draw is actually the result of having used a high efficiency Euro ABB motor which (guessing) maybe draws maybe 6 - 8% less amps per HP output compared to an older design like the Leeson. (high efficiency motors are now required by EU legislation)
Anyway.. It's all interesting, but in the end (unless we work for NASA) we're all in the same boat of practically having to rely on rule of thumb methods like these - and a fair amount of seat of the pants judgement. It tends to work out fine, because unless you are really skimping these numbers tend to have some cushion built in anyway.
PS On adjusting system resistance so that your fan motor draws full load amps in normal use David. It can be done, and maximises the CFM delivered from a given fan. The potential problem is as you know that it could if care is not taken create a situation where there is an increased risk of overloading the motor by opening too many blast gates or whatever.
So unless you're very aware of the issues there's an argument for sizing the fan and motor so that it delivers the CFM you need somewhere safely below full load amps to cover against unexpected eventualities. Ditto in the case of the fan, if you can run down the curve a bit in normal use it'll do better if faced with an increase in pressure drop when e.g. hooked up to a restrictive machine or whatever.
Last edited by ian maybury; 02-19-2012 at 10:05 AM.
I was wondering if I could convince Rod Sheridan (or others) to take a Dylos reading in a carpeted area of their home with a normal amount of foot traffic. I am very curious about the relativity and meaningfulness of the various Dylos readings in the shop environment. I am also curious about the cost of running my 5hp cyclone for 10-15 minutes just as an aircleaner....I assume it to be reasonably high given current electrical costs in my area. I may just do that and post the findings sometime.
Yes Ian, care must be taken when adjusting fan speed. The fan itself can only go so fast before exceeding the capabilities it was designed for. My backward inclined aluminum fan can't speed up as much as a heavy steel fan so I limit myself to the 63.5 top end. It's pretty easy to stay within the amp limits of the motor when the vfd is displaying real time amp draw. A few experiments give you the info you need to keep everything within limits. My best purchase was $100 for a fan type anemometer though. While not totally accurate it provides the real data you need at each port or gate to know what you are really getting. I doubt anyone can tell by feel the difference between 3500 and 5000 fpm. The dust collector companies should develop one to be sold with their systems. They would make some money, although they may disappoint the buyers of their small systems. Dave
CFM, SP, HP all vary depending on the system and fan performance match-up. The only thing said to be constant is that the fan is a "constant-volume device". I think of the blades on the fan wheel like a shovel. Each shovel scoop will pick up a certain amount of volume. If you want to increase the volume, you pick up more scoops, i.e. turn the fan faster. If you are shoveling sand or styrofoam, you pick up the volume, but they weigh different amounts. This concept contributes to the HP requirement of the fan in terms of air density. (elevation, temperature, moisture, SP)
However, the fan is not always a "constant-volume" device. Look at a fan curve and you will see that the CFM changes with SP for the same fan speed. On many fan curves, a HP curve is also presented. The HP also changes with the CFM and SP according to the equation Ian presented. If you maintained constant HP and efficiencies, the CFM and SP would offset evenly (I think Oneida has a system that does something similar by changing the fan speed to maximize HP or amp draw). However, since fan efficiency (and motor efficiency) change throuhgout the fan performance range, CFM and SP are defined by the performace curve.
The system will also have a curve. If you measure CFM and SP in your system, you can define multiple points on the system curve by picking another CFM and calculating the resulting SP. The resulting SP will be the ratio of the volume change, then squared, and multiplied by the original SP. If you graph this on the fan curve, the operating point will be where these two curves intersect. However, the system curve changes if you open or close gates, modify duct runs, etc. When the system curve changes, you have a new operating point on the fan curve. The fan is pulling a different CFM, different SP, and different HP.
Basically, there is no "constant", but there are relationships and inherrant truths in how the DC systems work. When setting up a system, it is called a "balance" for this reason. Everything in the system works together and depends on the other components. It may not be as important to know the numbers (however sometimes it is) for a home shop, but if you know how the hoods, duct system, and fan performance interact with each other, it will give you an idea of how your performance will be affected by changes you make to different parts of the system.
Because a typical fan is not a positive displacement device, it also is not a constant volume device. To be a positive displacement device, the impeller would need to seal to the blower enclosure, something near impossible to do. Describing the mechanics of a fan or pump is beyond my expertise, but basically the impeller creates a vortex that pulls everything along with it, much like a tornado. The closer the impeller is to the housing, the more efficient it is, and the smaller particle it is able to pass without jamming. That is likely one reason cyclones are a bit more efficient, in that the large particles have already been removed so the fan-to-housing clearance can be smaller. In blower-first units, the clearances need to be large enough to pass offcuts without jamming. OK, now that I have said that, hopefully someone that actually designs fans or pumps will jump in and correct me, as I only used to specify pumps in my past life.
I can't answer on how impeller clearance influences fan performance Ole as I've no experience, but it's an interesting question. The impellers we use are highly inefficient, 40% efficiency is a number that comes to mind. They tolerate all sorts of horribleness though.
If we could bump the efficiency without running into other issues there's on the face of it a lot of room for improvement.
Bill P reported doing some testing of a airfoil bladed impeller. It seems they are the next rational step up, and a lot more efficient, but can run into problems with stalling and buffeting when the blast gates are shut and the air speed drops down.
Do you have any relevant experience Michael?
Ian, the type of fans used on our collectors are typically radial blade type (they look like a paddle wheel). They have a lower efficiency but can handle a high dust loading. A backward incline fan would probably be the next step. It is typically used where you have low dust loading but would probably not be recommended for single stage collectors. It would probably be fine for applications after a good cyclone. The backward inclide fan wheel has blades in the shape of an airfoil design but uses curved plates to get the shape. An airfoil wheel has hollow airfoil blades made from thin gauge metal. A true airfoil wheel is typically used for relatively clean air applications. This is usually due to the chance for particulate buildup in the fan and errosion of the thin material on the blades. It is likely that in a home shop, the collector would not see enough use to errode the blades.
Radial wheels are the least efficient, but also the least expensive. I imagine manufacturers of our collectors are trying to hit a price and performance point. However, they are spot-on to use the radial type for single stage collectors and most of the cyclones you see. A good backward incline fan can be 80%+ efficient, and I would put a typical radial blade fan in the 50% range. The efficiency will depend on housing clearances, inlet cone clearance (if one is present), and where you are on the fan curve. There is a portion of the fan curve for each type of fan where it is most efficient.
Ole, a centrifugal fan is working primarily by centrifugal force. The blades on the wheel pick up a volume of air and throw it toward the outlet of the fan. For a given air density and system, the fan will behave as a constant volume device. If the air density were to change, the fan will still try to pull the same volume. For example, if you have a system that normally runs at 600F and you size the fan motor for the operating conditions, you will not be able to start the fan at 70F at full speed and full load. You need twice the horsepower at 70F because the fan is trying to pull the same volume of heavier air. If you slow the RPM, the CFM goes down linearly, and vice versa.
Where this gets confusing, is that the heavier air has more resistance in the ductwork, and therefore changes the system curve. Lets say you have enough motor to start the fan at 70F. You will have a steeper system curve hitting the fan curve at a different spot (CFM and SP). If you measure the flow, you have a lower flow, and higher SP. The fan is trying to deliver a constanct volume, but the system has more losses with the heavier air. Very similar to centrifugal pumps, and the curves also look similar, just different units. CFM vs. GPM, and "wg vs. Ft.Hd. It takes more horsepower to pump the same volume of hydraulic oil than it does water if you have the same system, same flow, and same head requirements.
A positive displacement blower is normally used in high pressure applications (inches of mercury). They have very flat fan curves and can generate the maximum static pressure over the entire flow range if you have enough horsepower connected. The efficiency is very similar over the entire range also.
Great write up Ian.....
am I correct in thinking that DC systems are similar to the liquid flow systems, where you get your actual volumetric flow rate from the intersection of the pump curve and the system curve?
Would the term head loss equate to WG?
Does any one out there know how much loss per foot one might expect in the typical diameters (4", 5" & 6") of typical flex hose (assumed to be stretched straight)?
The little geek inside of me wants to tap my piping with a pitot tube and see if I can measure the flow rate....
I think I can still write out Bernouli's equation....
The loss depends on the duct velocity, Here is the loss per 100 ft at ~4000 FPM
4" - 350 CFM - 11.1"wg
5" - 550 CFM - 8.4"wg
6" - 800 CFM - 6.8"wg
The losses in 100 ft of metal duct (smooth wall) for the same flows are 7.1"wg, 5.5"wg, and 4.5"wg for 4", 5", and 6" duct, respectively.
I'm a geek too, thought about the pitot tube as well for some flow readings and SP readings. I think I will do it before I upgrade to 6" system duct.
Not sure it was pointed my way Matt, but yes the principle seems to be the same as far as the pump/fan and system curves go. Measurement of pressure drop also - it's just that the units (ft head and inches water gauge) are different.
The convention is to treat air at pressures below 25in WG as incompressible, meaning that the maths of flow in ducts tends to be fairly straight forward, and the same as water in pipes.
What I'm not too sure about is just how precisely this holds true - it's too long since i did fluids.
If the incompressibility rule held 100% then closing a blast gate should either stall the fan, or bust a duct. It doesn't, because in the end fans do slip when their pressure capability is exceeded. The curve of a truly volumetric fan would likewise just be a horizontal line.
The rule of thumb likewise says that a very short restriction (like a venturi) will restrict the flow in a system in more or less the same way as if the full length of the ducting was that size. That doesn't seem to be entirely the case - for example David gets quite a decent step up in CFM by running a larger header where the velocity required for transportation is lower.
Against that the fact of running a larger duct for part of a run will reduce the frictional losses in that part of the duct too, so it may be that this rather than any significant degree of compressibility is the source of the increased flow.
tee hee. You're going well for a guy that didn't want to get sucked into the technicalities of dust system outside of work Michael. They're pretty addictive....
Nice explanation Michael.
An example of a more efficient blade design can be found in home central vacuum systems. Mine is a three stage Ametek type. (curved blades, with enclosed impellers - set in stages).
Although there is some science to designing these things, there is still a remarkable amount of empirical work as well (I was working with Carrier a long time ago on some ultra high efficiency energy systems, and they have a room in the R&D lab there with every blade/vane design you can possibly imagine - hung up on the walls for future reference).
Our dust collectors arent 'designed' - they are fabricated based on simplest manufacturing method.....
Yes, I would agree that our hobby dust collectors are definitely built based on a manufacturing method and equipment selection that leads to a "price point". If the cone of a cyclone can be shorter to prevent cutting another sheet, then that's the design, efficiency may be sacrificed. Not necessarily blaming the manufactures, consumers are as much to blame as anyone, myself included. However, on hobby cyclones, we usually have an after filter, so low efficiency is not the end of the world. Just means your after filter gets plugged quicker. It would be great if you could blow it all outside, but that has its own pluses/minuses.
Yes, very addictive...
Whatever the reason Michael it's great to see some more engineering based input about...
For more charts on cfm and pipe size and the loss at various velocities go to www.airhand.com. Particularly informative as to the increase in resistance that occurs as velocity increases . Dave
03-06-2012, 11:22 PM
Those numbers don't correspond with what I have heard before. For example Bill Pentz says PVC actually produces less static pressure drop (which makes sense to me as it is just about as smooth as it gets). I definitely don't think 4" metal duct is equivalent to 6" PVC as your numbers seem to indicate.