PDA

View Full Version : Accuracy of anemometer vs pitot tube to measure CFM



Mathieu Beauregard
03-14-2008, 1:00 PM
I have a small cheap LaCrosse anemometer that I've been using to measure the impact of different configurations in my dust collection system. I use it to determine if the changes I make to the ducting/cyclone are good (higher velocity) or bad (lower velocity).

here it is: http://www.lacrossetechnology.com/ea3010/index.php

Here's how I measure it:

1- readings right at the end of the pipe are way too high and inaccurate so I've drilled a slot in a 10 feet long straight piece of pipe. I insert the anemometer in there and hope it doesn't slip out of my hand...

2- I take the reading in mph, convert it to feet per minute and apply the following formula: CFM = FPM * (Pi*r2/144)

Using this formula I get numbers between 600CFM at 750CFM at my table saw (depending on duct, filter and cyclone). I have a home made cyclone with a 12" impeller, 2HP motor and 6 inch ducts.

In terms of absolute CFM numbers, I'm curious as to how off the measurements with the anemometer are compared to a pitot tube setup (a la Bill Pentz).

Did anyone here experiment with both methods (anemometer and pitot tube) ?

Here's my hacked-up cyclone setup:
http://pages.videotron.com/guizzmo/cyclone/

Tom Veatch
03-14-2008, 2:47 PM
...In terms of absolute CFM numbers, I'm curious as to how off the measurements with the anemometer are compared to a pitot tube setup (a la Bill Pentz)....

It won't matter whether you're using an anemometer or a pitot tube as long as you are only taking a single measurement of the velocity. Under those circumstances - single measurement - the pitot tube would disturb the flow less and, assuming equivalent calibration of both systems, have the potential for a more accurate point measurement.

Due to viscosity effects, the velocity distrubution across the duct will vary from zero in the boundary layer at the duct wall to a maximum velocity at the duct centerline. To get a true CFM reading, you need to take multiple reading across the duct area and integrate those values. A pitot rake with multiple adjacent tubes (looks a little like tines on a garden rake, hence the name) to take simultaneous reading at multiple points can be used for such a task.

As far as a single reading, the pitot tube is probably the simpler system of the two, (pitot or anemometer) and would have less opportunity for introducing measurement error as long as it were properly mounted and calibrated. But since the pitot tube actually measures dynamic pressure, the calculations to convert to airspeed will introduce error unless the air density and barometric pressure at the time of the measurement is well known.

Personally, I used the same LaCross anemometer to get approximate, comparative values of the flow velocity at the centerline of my ducts under different conditions and called it good. A properly calibrated pitot rake in a rigorous test environment could yield more precise volumetric flow rates, but that level of precision is far beyond what is needed to set up a functioning dust collection system. Comparative readings, taken consistenly at the same point in the duct cross section with either pitot or anemometer are accurate enough for this purpose.

Mathieu Beauregard
03-14-2008, 3:08 PM
Thanks for the great info Tom.

I did notice variations in velocity from the center to the side of the duct. I initially tried to use the average velocity feature of the Lacrosse anemometer, but I got inconsistent results. I now use a single reading always in the same spot.

It helped me optimize my setup by trying different duct configurations.

chet jamio
11-19-2008, 9:07 AM
I'm new to woodworking and am doing research for my own dust collection. A search turned up this old thread which is releavant to what I"m doing. I also plan on using a handheld anemometer to determine cfm flow. Rather than Tom's description of using a pitot rake, I'm planning on using some fluid dynamic equations.

101439

The chart above shows a correlation between centerline speed [fpm] and volumetric flow rate [cfm]. The left is 4" duct and the right is 6" duct. What was interesting to me is that the ratio is basically 0.83 over the entire range of pipe sizes and speeds that we need. This is true for 2"-15" ductwork for the 2000-6000 fpm range. I added a cfm column for the math adverse.

I'm purchasing an anemometer now and I may also rig up a pitot tube. As Tom mentioned, it's difficult to measure dynamic or stagnation pressures withour error. The tube has to be exactly parallel or perpindicular to the airstream, depending on which you're measureing.

Jeffrey Makiel
11-19-2008, 9:58 AM
I use to be an HVAC engineer before I became a fat and old desk jockey. Tom's post was very thorough and insightful.

However, 600 cfm would be in the right ball park based upon the system you describe. Keep in mind that your system's flow will vary widely based upon which branch is being used, the hood arrangement on the tool, and the cleanliness of the filter. Trying to get an accurate snap shot may not offer much.

In a less scientific approach, the wealth of testimony from posters in this forum has yielded the opinion that a 2hp, 12" wheel single stage collector with 6" duct in a double size garage shop seems to flow enough air. Often the discussion shifts to filter quality and cleaning maintenance as the primary considerations.

-Jeff :)

David Keller NC
11-19-2008, 9:59 AM
Chet - Remember that you will need to correct the kinematic viscosity for the particular temperature when you measure it.

What I might suggest to those without an engineering background is to ignore the absolute CFM you calculate with a velo. meter, and just use the meter to assess the relative drops in velocity (assuming the same diameter of pipe) in the system you put together. You can then compare that to the manufacturer's stated CFM to get a rough idea of whether your piping system is efficient and your blast gates are not leaking. This is arbitrary, of course, but I'd think getting something like at least 40% of the velocity at the machine port compared to the inlet to the DC is reasonable.

chet jamio
11-19-2008, 10:19 AM
In a less scientific approach, the wealth of testimony from posters in this forum has yielded the opinion that a 2hp, 12" wheel single stage collector with 6" duct in a double size garage shop seems to flow enough air

The problem is that different people desire different results. When they get those results, they claim the design "flows enough air", but that doesn't mean much. There are people who want 400 cfm chip collectors, while others want 800 cfm dust collectors. I'm not saying which is right, but I am trying to help people determine if they're meeting thier specific goals, whether it's 400 or 800 cfm.


Chet - Remember that you will need to correct the kinematic viscosity for the particular temperature when you measure it.

In the range of 40-100 degrees F, the ratio stays at 83%. That was the point of my post. In the woodworker's range of duct size and flow rates, the 83% value always works.


What I might suggest to those without an engineering background is to ignore the absolute CFM you calculate with a velo. meter, and just use the meter to assess the relative drops in velocity

That is the opposite of my intent. I was trying to publish a simple chart that anyone could use to convert velocity readings to cfm values. This would help those of us with single stage units who can't trust the manufactuer's claims. This way, we can determine if we're getting the cfm that we desire.

Ken Garlock
11-19-2008, 10:56 AM
If the vacuum will suck the chrome off a trailer hitch, it is good enough.:D

You gents can save your calculator batteries by going the Bill Pentz website (http://billpentz.com/woodworking/cyclone/index.cfm)and read about his research.

Bill White
11-19-2008, 1:45 PM
I have a slack tube manometer that ya can get cheap. PM me.
Bill

David Keller NC
11-19-2008, 4:12 PM
"In the range of 40-100 degrees F, the ratio stays at 83%. That was the point of my post. In the woodworker's range of duct size and flow rates, the 83% value always works."


Well, if you've done the calcs, I'll take your word for it. Getting out my grad school fluid dynamics book is not my idea of a fun evening. ;-)


But, there are other assumptions here, one of which is that an idealized laminar radial flow profile is the case. The presence of pipe bends, junctions, blast gates, the use of corrogated pipe, and most importantly, the presence of dust and chips in the fluid stream is likely to yield substantial deviations to the laminar flow assumption, so a chart would be perhaps a decent measure of relative differences between pipe sizes, but iffy as far as the validity of the absolute numbers.


Besides, why would this matter? I'm assuming that someone that's setting up a dust collection system has only a few choices of horsepower/single vs. multiple stage/amperage draw, and most of the available collectors from the various manufacturers within a given class (for example, single stage 1-1/2 hp) are going to be darn close. What I'd think someone doing this would be considerably more interested in is how well their blast gates seal, minimizing pipe run length, and avoiding unnecessary bends in the network. Under those circumstances, I'd think using a velocity meter to measure relative differences instead of trying to calculate absolute cfm would be considerably simpler.

chet jamio
11-19-2008, 6:04 PM
But, there are other assumptions here, one of which is that an idealized laminar radial flow profile is the case
Laminar flow only applies for Re < 2300 and turbulant flow begins at Re > 4000. For a DC, Re is above 100,000. Woodworkers are no where near laminar flow. I have made no assumptions about laminar flow.

Regarding bends, junctions, and blast gates, you are correct. I didn't reprint the Bill Pentz data which is "have at least 8.5 pipe diameters to let the entering air stabilize..." For completeness, that should be restated in this discussion. His value of 8.5 is based on the ASHRAE test procedure for HVAC systems, and I haven't yet calculated what this value should be for woodworkers.


the presence of dust and chips in the fluid stream is likely to yield substantial deviations

I don't plan on doing any testing with the DC connected to a saw which is doing any cutting. I don't believe anyone would try this.


Besides, why would this matter?

For me, I want to know if my 10 year old Grizzly with an aftermarket oversize bag and my ductwork can provide the flow I want. The only way to do this is to measure the absolute flow rate. Yes, I can use relative measurements to optimize my system, but what if I'm only flowing 200 cfm with 10 feet of 6" smooth walled ductwork? I can optimize all day and only reach 300 cfm. This absolute cfm measurement is the final answer on whether your system is meeting your requirements.

Ben Davis
11-19-2008, 9:21 PM
I'm not trying to sound trite here, but I have to ask 2 questions.

(1) If you're already getting reasonable readings, what exactly are you trying to accomplish? Are you trying to tweek your DC design? Do you want a permanent reading point? I'm confused at to what you are actually trying to do.

(2) I think the pitot tube idea is brilliantly simple, e.g. airplanes. However, I'm worried that we're forgetting that this is a dust collection system and there will be all kinds of debris impacting the open tip of the tube. Maybe what you really want is a venturi set-up.

Just some thoughts.

Myk Rian
11-19-2008, 10:55 PM
With 35 years experience working with automotive paint ovens and building management systems, I can say that trying to get a perfect balance in a system is a lesson in frustration. Imagine standing on a rooftop in the winter, balancing the airflow between 2 40" ducts running from paint spray booths, to an incinerator system. Not a lot of fun.
We would take 10 readings of each duct in a vertical pattern, then 10 in each horizontally using a pitot tube and an inclined manometer graduated at .01". Do the math, make damper adjustments, and do it all over again.

To simplify balancing a DC system, a single pitot tube mounted in a straight section of the duct, with a distance of at least 8 times the diameter before and after the pitot tube will give a "usable" reading. The tube won't get clogged with debris since it is pointing downstream.

David Keller NC
11-20-2008, 9:43 AM
"For me, I want to know if my 10 year old Grizzly with an aftermarket oversize bag and my ductwork can provide the flow I want. The only way to do this is to measure the absolute flow rate. Yes, I can use relative measurements to optimize my system, but what if I'm only flowing 200 cfm with 10 feet of 6" smooth walled ductwork? I can optimize all day and only reach 300 cfm. This absolute cfm measurement is the final answer on whether your system is meeting your requirements. "

Ahh - that makes more sense. I was looking at it from the standpoint of a new installation. In my case, there really isn't much room for choices, so absolute cfm measurements don't really help. In other words, there's insufficient room for a 2-stage system, and 1-1/2 hp is the limit that my wiring will stand, so I just went with the max I could get.

Interesting about the Reynold's number. Guess I should've done the calcs - plug (or turbulent) flow is generally easier from a computational standpoint. I'd have guessed that it would've been much closer to laminar flow.