DC system flow results

[Ole Anderson]

Now that I have installed my Dwyer air pressure meters I did some measurements and, if I am interpreting the results correctly, it seems I have plenty of airflow at my tools. The biggest issue I ran up against is determining which published fan curve seems to fit my system. With all gates closed I was reading 8.3" wc SP at the inlet to the cyclone. Curves for the Oneida 2 hp Super Dust Gorilla as published in the 2006 American Woodworker mag indicate that the manufacturer claims about 11" of SP at zero flow while the mag measured about 11.5" SP at zero flow. If you go to the Oneida website and click on "fan curve" on the "Product Detail" tab of the 2 hp SDG, that fan curve shows about 10.6" of SP at zero flow. So, when I took the most recent of these 3 curves, it appears that with 8.3" of measured SP, I was getting a huge leak in the system of about 750 cfm. I knew that could not be the case as virtually all seams are either taped or siliconed, and only the blast gates appears to have minor leakage upon close inspection with the system under maximum static pressure. With all gates closed I was reading about 0.3" of pressure on the filter gauge indicating some leakage but certainly not 750 cfm. At that point I suspected that the main SP gauge was giving an incorrect reading so I fabricated a manometer using 1/4" plastic tubing. Using the same port at the inlet, the manometer was tracking the Dwyer gauge within 0.1" SP. So next I began to suspect the impeller itself or the fan curves. I opened the "Technical Specs" tab on the Oneida 2 hp SDG page and saw a "Spec Sheet" pdf at the bottom. I opened that and found yet another fan curve. Meanwhile, I disconnected the last 7" flex pipe from the rest of the duct system to eliminate virtually all possibilities of leakage. With the end of the flex covered in plastic I read 8.75" SP and 0" SP on the filter. With the plastic removed and the system running wide open, except for losses after the pressure port, I read 2.6" SP with 3.4" of pressure on the impeller. Those two numbers correspond almost exactly with the fan curve in the pdf. While disappointed that the maximum SP was 2" to 3" less than other published figures, I at least felt confident that I had a curve that was within the ballpark. I noted that the motor was drawing 10.5 amps at full load vs a published FLA of 11.5. I would be a bit happier if they had loaded up the motor with a more aggressive impeller.

So, IF pressure losses across the filter are proportional (which I suspect they are not), my leakage is 0.3"/3.4" x 1200 cfm or 105, say 100 cfm. Next I measured the SP with various gates open. First with both the 3" top and 5" bottom TS gates open I am seeing 5.5" SP (2.1" on a fairly clean filter) which is about 940 cfm less leakage of 100 leaves 840 cfm for the TS. Next the 6" line to the miter saw had 5.8" SP (1.6" filter) which from the curve is about 890 cfm less 100 cfm leakage leaves 790 for the MS. All four drops to the other tools measured almost exactly the same at 7.3" SP (1.1" filter) which is about 630 cfm after deducting leakage. Not bad for a 5" drop choked to a 4" hose. If my assumptions are reasonable, it shows that a 2 hp cyclone can indeed function adequately to serve a properly sized duct system. Doing it again, I would opt for the 3 hp SDG as its dead headed pressure is 12.6" vs 8.9" for the 2 hp and it will provide an available static pressure of 9.8" at the sweet spot of 800 cfm vs 6.8" at 800 cfm for the 2 hp SDG.

[David Kumm]

Ole, that is really interesting. Oneida fan and system curves have always baffled me to your results are helpful. I'm most surprised at the 630 cfm through a 4" port or a velocity of over 7000fpm. That is more than I would expect the 2 hp system could pull through that size opening given it's SP rating. Oneida impellers are quite efficient at low pressure so 800 cfm through a 6" line seems right at 4000+ fpm. I would have guessed no more than 5-6000 fpm through the 4" port so I learned something. Dave

[Alan Bienlein]

Curious as to the size of your impeller?

I draw 10.5 amps with all gates closed and 18.5 amps with three 6" gates open. Surprisingly it's the same reading if I disconnect the duct work at the inlet of the cyclone. I have yet to try it without filters. I'm running a 5 hp motor and 16" impeller.

[John Lanciani]

Hi Ole,

I posted this a while ago but it didn't generate much dialog so I'll try again here. Maybe it will add some more data points to your observations:

I decided to do some further testing of my dust collector to get a better handle on the relation between static pressure, airflow, and motor power output. My day job is for an electric utility and about 50% of my time is spent dealing with power quality issues so all of the test equipment needed was close at hand.

My dust collector is an Oneida V3000, and it is fitted with a US made Leeson 3hp motor. The pertinent specs off of the motor are as follows;

Full load amps – 13
Full load power factor - .91
Efficiency – 81.5%
Output – 3hp / 2.25 kW
Capacitor start / capacitor run

I used the following test equipment; UEI EM151 digital manometer, Fluke 41B power harmonics analyzer, Fluke 33 true RMS clamp on ammeter, Fluke 189 digital multimeter.

The CFM numbers are extracted from an Oneida published fan curve for the V3000. SP measurements were taken from a pitot tube at the cyclone inlet.

Horsepower was calculated using this formula; voltage (238) x current x pf x eff. / 746

The results were as follows;

SP ______CFM___ Amps___ PF___ HP
12.4 (max)_<100____7.3 ____.71 ___1.34
12 _______200_____7.5____ .73____1.42
11 _______400_____8.7 ____.81____1.87
10 _______525_____9.6 ____.83____2.07
9 ________650____10.1 ____.86____2.26
8 ________775____10.3____.86____ 2.3
7 ________900____10.6____.86____2.37
6 _______1000____10.9____.87____2.47
5 _______1100____11.1____.89____2.57
4 _______1200____11.3____.89____2.61
3.7(min) __1250____11.4____.89____2.63 (1.97kW)

One other interesting piece of information, the locked rotor / inrush peak current was 159 amps.

I’m not really sure that this information has too much value other than to illustrate the relationship between SP, CFM, and HP for a typical small shop dust collector. It did reassure me that my system as installed is working as advertised. There is a little bit of headroom on the motor so it should live a long and happy life

Edited to add- system ductwork is 6" PVC and I manipulated blast gate openings to get the SP steps in even increments. 3.7" was the lowest I could get with everything wide open.

[Michael W. Clark]

Ole,
Great information on the performance and I would recommend you keep these initial SP readings somewhere for reference. This will be valuable information in the future. I'm not sure how you know the CFM values since the curves don't match? Did you take a flow reading?

Either way, knowing the SP at the cyclone inlet (system static pressure) for various branches open will allow you calculate the flow difference as these readings change. The system SP is going to vary with the square of the change in flow. So, if your TS scenario were to read 5.0"wg SP instead of 5.5"wg, then the flow would be about 95% of what it is now (896 CFM if the CFM values are correct). When you notice the drop in system SP, then you would look at the filter DP to see if it has increased to cause the reduced flow.

Because of the relationship between flow and SP, a small change in flow is easily detected with a SP measurement. Of course, if you make duct changes or modifications to the system, then you have to do another baseline reading.

Nice job!

Mike

[Michael W. Clark]

The CFM numbers are extracted from an Oneida published fan curve for the V3000.

It did reassure me that my system as installed is working as advertised.

Hi John,
I may be missing something here, but I am not sure I follow you. If you pull the CFMs from the fan curve, the equipment will, by definition of the fan curve from which you are pulling data, perform as advertised. Your data proves that the amp draw varies as you add resistance to the system, which is to be expected. The more resistance, the less flow, the lower the amp draw because the motor is doing less work. I am not sure how this proves that the fan is delivering the CFM advertised unless you measured the flow with the pitot tube.

Mike

[David Kumm]

Michael makes a great point. I would think the next step would be to measure the velocity or cfm at each port and correlate those with your sp to create your own curve. Since the Oneida one doesn't seem to correlate your real world flow might not match the chart. Dave

[Ole Anderson]

Let me add that my big assumption is that the fan curve from the pdf is reasonably correct in the sweet spot. It is good at each end of the curve, ie. wide open and totally shut, but I don't have a data point in the middle to verify the curve where it is most important, that is in the 500-1000 cfm rate of flow portion of the curve. However just having my flow and SP hit on the ends of one of the Oneida curves within a half an inch of static pressure seems a major triumph and gives some credence to the Oneida curve that I posted (middle figure IMG_1986.jpg, an edited camera shot of my laptop screen). I posted the third figure just to show the difference in the fan curves presented within the same Oneida document. As per David's last post, I suppose my next step is to get an anemometer to actually measure the velocity and therefore the rate of flow at the end of the various ports. Anyone have a line on a reasonably priced anemometer?

Allen, frankly I don't know the diameter of the impeller, I am wishing I would have measured it prior to installation. As well as reading the info on the motor tag as to FLA, SF, etc. Very difficult to access now.

Another note: while looking up the price of my 5 year old Amprobe ACD10-pro clamp on meter (for a response on an RV forum) I found out it has been recalled. FedEx picked it up today and returned it to Fluke, where they are supposed to return a newer version at no cost to me. Cool, as the newer ones have even more functions.

[Michael W. Clark]

Since you already have the Magnehelics, the cheapest solution may be a pitot tube and some hoses. I think you can get a "mini" from Dwyer for around $50. However, you would want a Magnehelic in the 0-2" range. At 4,000 FPM, you will see a VP of around 1.0"wg. Take flow readings at the cyclone inlet so they correspond with your SP readings. You can take them at the hood as well, but the only thing it will tell you is the leakage. Unless you have some major leakage, there will be more error in the flow measurement than the leakage. Flow readings are +/-5% accuracy and are usally only guaranteed within +/-10%. The straighter the duct run the better with about 2/3 of the straight section upstream and 1/3 downstream of your test location.

Flow readings are the only way to truly compare your performance with the Oneida curves. You have to see if the flow and SP match the same point on the curve. I'm not sure how the Oneida curves are given (I guess I could look), but make sure the SP you are comparing is the SP on the curve. Some hobby DC manufacturer's use cyclone inlet SP and an assumed filter DP. Most fan manufacturer's use FSP, FTP, or ESP. These are not the same as inlet SP.

I have a friend that still has access to the airflow test equipment. I may get some readings on my Jet DC 1100 as is and again after I put a baffle in it and convert over to 6" duct.

Good luck,
Mike

Mike

[Jim Neeley]

Great information guys, thanks!

I'll add in my $0.02 here.. I work as an Instrumentation / Controls system applications engineer to support my woodworking hobby and, when metering flow, an important accuracy consideration is the "number of pipe diameters" of straight pipe upstream and downstream of the measuring point to ensure a fully-developed turbulent flow.

Dwyer recommends minimums if 8.5 diameters upstream and 1.5 diameters downstream of the sample point for measuring particulate-laden air flow. Applying this is just multiplying the pipe diameter (i.e. 6") by the upstream value (8.5) meaning they recommend at least 51" of straight pipe upstream of the pitot tube and 7" x 1.5" = 9" downstream.

For Dwyer's PDF confirming this and information on taking the readings and performing the calculations, download http://www.dwyer-inst.com/PDF_files/160S_iom.pdf. Dwyer also show you the method you should use to collect the samples and how to perform the calculation.

As I said, just my $0.02... YMMV.. :-)

Jim in Alaska

[John Lanciani]

Hi John,
I may be missing something here, but I am not sure I follow you. If you pull the CFMs from the fan curve, the equipment will, by definition of the fan curve from which you are pulling data, perform as advertised. Your data proves that the amp draw varies as you add resistance to the system, which is to be expected. The more resistance, the less flow, the lower the amp draw because the motor is doing less work. I am not sure how this proves that the fan is delivering the CFM advertised unless you measured the flow with the pitot tube.

Mike

Hi Mike,

Assuming that the Oneida curve for my unit is accurate (I do believe that it is close), all my numbers really show is that my ductwork is the limitation in my system since I can not get the SP below 3.7" at the cyclone inlet, and that Oneida did a fairly good job at matching the motor and impeller to the 6" inlet on the cyclone. Oneida also offers the exact same unit with a 2 hp motor which would clearly be undersized for the application.

I don't currently have the equipment to measure airflow but that would be the next step. Is there a relatively cheap and easy way?

[Rick Moyer]

No help here. I just wanted to say that after reading thru half your post my head was spinning faster than your impeller!:D

[Jim Neeley]

I don't currently have the equipment to measure airflow but that would be the next step. Is there a relatively cheap and easy way?

John,

Like everything else in life, "relatively cheap" is a relative term :) but :Yes. Check out the link I posted for the method.

[John Lanciani]

Jim,

What about just using an anemometer at the inlet and calculating flow based on airspeed and opening size? Do you think that would be reasonably close? The manometer I have is only a single port so it looks like I'm out of luck there.

[David Kumm]

You can find hand held fan type anemometers for under $100 to measure velocity at the machine ports and convert to cfm. There is some inaccuracy due to the differences in speed around the entire opening but you can find the high low and average velocity and be pretty accurate. CFM at the machine is the relevant number anyway. Dave

[Michael W. Clark]

Hi all,
I know Jim's link went to an S-type Pitot tube, it will work, but requires a much larger hole and also has a measurement factor associated with it. It is often used for environmental stack testing and very large ducts.

This is the option I would recommend http://www.dwyer-inst.com/PDF_files/122d.pdf. The 166/167-12 is very handy and will accommodate up to a 12" duct. You only need a 3/16" hole in the duct for insertion. The model with the shorter tip (1" I think) would probably be the easiest to use on smaller duct.

[Jim Neeley]

Jim,

What about just using an anemometer at the inlet and calculating flow based on airspeed and opening size? Do you think that would be reasonably close? The manometer I have is only a single port so it looks like I'm out of luck there.

Theoretically, yes. The difficulty is that, unlike blown air sucked air comes from all directions at once so air speed falls off at roughly 4 times pi times the distance squared. I wouldn't hazard how close it would be but I wouldn't wager much on it being really close.

As for my PDF link, although it's designed for tubes down to 4" diameter, as Michael says, a smaller one would be better. When I posted the PDF it wasn't to provide a product recommendation; just to show the methods.

Jim in Alaska

[Ole Anderson]

Well, I got an early Christmas present from my son, the one that got me a great price on all my ductwork. It was an anemometer. Not real high end, but it does the job. I just tested it on my 4" router table dust inlet. The diameter at the end is 4-1/8" so that gives me an open area of 0.0926 sq ft. The velocity measured was 6160 ft/min, so the flow rate was 570 cfm. (the meter is reading 616 with a x10 that you can barely read in the pic) Not bad for a 2 hp dust collector (Oneida Super Dust Gorilla). Here is to those that say a 2 hp just doesn't cut the mustard. Hmm, I wonder what the 3 hp would do? And this was with about 5 feet of 4" slinky hose, a 4" blast gate connected to about 10 feet of 5" duct with an increaser and one 90 and a 45 elbow before hitting the 7" main. While a pitot tube or a hot wire anemometer inserted in a small hole in the duct would be more accurate, I am quite happy with the results. As stated earlier, measuring the velocity on the suction end of a hose is problematic as air is coming from all directions. My rough assumption is that the slight loss of open pipe end area from the anemometer pickup balances the directional issue. The only other place I could easily use the device is on my bottom 5" pickup to the TS, but there I have a bellmouth so the directional flow issue is even more pronounced.

[David Kumm]

Ole, I've used that model and it matches with the UEI I PMed you about. You will find that when you move it around the opening it will vary about 500-1000 fpm but that is all. Check it at the blast gate before the flex to see what the flex takes off. My guess is that the velocity on a 5" might be close and then the cfm gets close to maxed out. I've found that before and after the blast gates changes things too. Blast gates to cost some flow. Your 7" main helps as well. Would be good info to post what you read at 5" and 6" ports to see where the CFM maxes out. Dave

[Rick Potter]

I would be curious to see how the ratings vary with the filter on and with it off, like I have. I've always been curious about how much, if any difference there is. Perhaps, next time you clean the filter, you could check it out for those of us whose heads are spinning with all those numbers.

Enquiring minds, with direct vents, want to know.

Rick Potter

[Michael W. Clark]

Ole, that's great. That is a pretty good velocity through the 4" duct. Also, you should have an even higher CFM when connected to the router table (I see you are using a quick connect fitting.) The losses for the open end duct you are measuring are higher than if this duct is stabbed into the side of the cabinet or connected to a hood with some type of taper. For example, if your duct velocity is 4000 FPM, the hood loss for an open end duct is about 1.93"wg, the loss for a port on the side of cabinet is 1.5"wg. You could check this by looking at your inlet SP reading with the hose connected or disconnected. You should also see a slight increase in fan motor amps as you would be pulling more CFM.

Basically, the design of the hood (or lack of hood) affect the system pressure losses, which affects where you are operating on the fan curve. We do this experiment with different hoods at the ventilation conferences and the results are usually quite noticable. This is another benefit to measuring the airflow in the duct with everything connected, but the anemometer can be accurate for inlet air velocities. If you were to measure the duct CFM with a pitot tube in the same configuration as above, you would probably see very little variation. Remember, most airflow readings are considered +/-10% accuracy unless you have a permanant flowmeter and associated insturmentation (and these are usually verified with a Pitot tube.)

Mike