musingsheader image
0 Helpful?

Simple Methods for Measuring Air Flow

Homemade flow hoods, the credit card test, the garbage bag test, and calibrated laundry baskets

Posted on Jan 23 2015 by Martin Holladay

UPDATED on April 15, 2016.

To commission a ventilation system or a forced-air heating system, or to troubleshoot problems with these systems, it’s essential to be able to measure the rate of air flow through registers and grilles. Most home performance contractors measure air flow with a flow hood. Flow hoods vary in accuracy, but they all share one attribute: they are expensive (generally $1,600 to $3,200).

If you want to measure air flow, but you can’t afford a flow hood, you may be interested in using one of the inexpensive approaches to air flow measurement described in this article.

Are flow hoods accurate?

Iain Walker, Craig Wray, and Max Sherman, three scientists at the Lawrence Berkeley National Laboratory (LBNL), examined the accuracy of commercially available flow hoods in a 2002 paper, “Accuracy of Flow Hoods in Residential Applications.” Although most of the flow hoods they looked at provided fairly accurate measurements in commercial buildings with large HVAC(Heating, ventilation, and air conditioning). Collectively, the mechanical systems that heat, ventilate, and cool a building. grilles, they were often inaccurate when used in homes. “Potential errors are about 20% to 30% of measured flow, due to poor calibrations, sensitivity to grille flow non-uniformities, and flow changes from added flow resistance,” the researchers wrote.

When it comes to measuring air flows, the researchers noted that the level of accuracy required to identify room-to-room residential pressure imbalances is +/-25%, and for ensuring room load and comfort requirements it is +/-20%. For such purposes, a flow hood with a potential error of 20% to 30% isn’t ideal.

Robert deKieffer, an energy consultant from Boulder, Colorado, has also looked into the problem of flow-hood accuracy. “The Shortridge flow hood I used wasn’t anywhere close to accurate,” deKieffer told me. “They charged $75 to calibrate it, and I still got a bad number after it came back. I thought, ‘Wow — I spent a couple thousand bucks for this tool and it doesn’t work. And I spent hundreds of bucks to fix it and it still doesn’t work.’”

I recently contacted the authors of the 2002 paper for updated information on flow hood accuracy. Iain Walker noted, "Most of the old flow hoods still have the same performance — they haven't been updated or improved. What has changed is that there are now several powered flow hoods on the market from Retrotec and The Energy Conservatory in the U.S. and several European manufacturers, and [these powered flow hoods] are good at measuring residential HVAC air flows."

According to Chris Stratton from LBNL, "There are differences between measuring inlet (return) and outlet (supply) flows. We found that most flow hoods can measure inlet (return) flows with a level of accuracy that is acceptable for most applications (within 10%). For outlet (supply) flows, only the powered hoods were this accurate (with one possible exception)."

Powered flow hoods are manufactured by The Energy Conservator (which makes the FlowBlaster), Retrotec (which sells a European flow hood, the Flow Finder Mk 2), and by Observator Instruments (which makes the Diff Automatic).

For more information on powered flow hood performance, see two 2012 reports from LBNL researchers: Laboratory Evaluation of Airflow Meter Devices and Field Evaluation of Airflow Meter Devices and System Flow Verification.

A homemade flow hood

Howard Leonard, a trainer of HVAC service technicians in Phoenix, Arizona, built himself a homemade flow hood using a cardboard box and a hand-held digital rotating-vane anemometer (sometimes called a velometer). Leonard made his flow hood out of a U-Haul moving box (see Image #1 at the top of the page). He formed the cardboard into a funnel-shaped tube, one end of which has inside dimensions of 12 inches by 12 inches. Since the opening is one square foot, the reading from an anemometer (which measures wind speed in feet per minute) equals the air flow in cfm. For the most accurate result, the area of the opening should be a little larger than 144 square inches; it should actually measure 144 square inches plus the area of the anemometer that is being inserted into the air flow.

Hand-held anemometers have been dropping in price. Leonard uses a Kestrel 1000 anemometer from Nielsen-Kellerman that is widely available for $74. “I like the Kestrel 1000 because you don’t need two hands to use it,” says Leonard. “It’s shock-resistant, it’s water-resistant, and it’s relatively inexpensive. It’s also more sensitive at lower velocities than some of the other affordable velometers.”

To use the flow hood, the cardboard tube is placed over a supply register with the 12"x12" dimension facing away from the register. The anemometer is set to averaging mode, and is moved in a traverse across the opening of the hood. This method is more accurate than taking a single reading in the center of the opening.

To keep one’s hand out of the air flow, the anemometer can be taped to the end of a slender stick. Although some people may question the accuracy of Leonard’s measurements — in fact, Iain Walker reports that the accuracy of vane anemometers is "not so good" — Leonard maintains that his homemade flow hood is accurate enough to diagnose typical residential duct problems, and is infinitely better than no flow hood at all. “I am not looking for perfection,” says Leonard. “I’m just looking for tools for the regular Joe out there. My game is ‘how can we do our job without being millionaires?’”

Gary Nelson, president of the Energy Conservatory in Minneapolis, Minnesota, advises builders of homemade flow hoods that the longer the cardboard tube (up to a point), the more accurate the measurement, since a longer tube gives the air flow a chance to become less turbulent. Nelson suggests using a tube that is about 3 or 4 feet long.

A credit card, a cardboard box, and a piece of tape

For years, home inspectors have been testing bath exhaust fans using the toilet-paper test: at a minimum, an operating exhaust fan should be able to hold a sheet of toilet paper tight to the grille. One of the main virtues of this method is that the necessary test equipment is readily available in most bathrooms.

Building scientist Terry Brennan of Westmoreland, New York, has invented a method for testing exhaust fan flow rates that is better than the toilet-paper test. The test requires a cardboard box large enough to cover the fan grille. A rectangular hole is cut in a side of the box that will be vertical when the box is placed with the open end covering the grille (see Image #2, below). The hole should be slightly smaller than a credit card, with the long axis oriented vertically. The credit card is installed inside the hole with tape along the top of the credit card, so that it forms a small swinging door.

With the fan operating, the box is positioned over the grille. The fan should depressurize the box, causing the credit card to swing inward. Using a gauge (for example, a pencil), you can measure the amount of swing — that is, distance between the cardboard opening and the bottom of the hinged credit card. The amount of swing should be at least 1 1/2 inch. Brennan performed calibrating tests to correlate credit-card swing distance with air flow. One and a half inch of swing corresponds to 25 cfm; 2 inches corresponds to 35 cfm; and 2 1/2 inches corresponds to 48 cfm.

Brennan notes that the accuracy of this test method decreases with higher air flow rates, since the resistance introduced by the cardboard box becomes more of a factor at higher air flows.

For more of Terry Brennan's tips for measuring air flow rates, see Is Your Ventilation System Working?

The garbage bag test

Don Fugler, formerly a senior researcher at the Canada Mortgage and Housing Corp. (CMHC), developed the garbage bag test to measure air flow rate through HVAC registers, bathroom exhaust fans, and clothes dryer exhausts.

The test rig is constructed by taping the mouth of a garbage bag to a bent coat hanger or a shaped piece of cardboard to keep it open (see Image #3, below). To conduct the test on a supply register, crush the bag flat, place it over the register or exhaust hood, and count how many seconds it takes for the bag to inflate (see Image #4, below). Then consult the tables (Image #5) to translate seconds to cfm.

To conduct the test on a return air grille or the suction port of an exhaust appliance, you need to hold an inflated bag against the exhaust grille and count how many seconds it takes the bag to deflate. Note that the lower the air flow rate, the more difficult the rate is to measure.

While Fugler’s garbage bag test doesn’t give a precise measurement, it provides useful data.

A calibrated laundry basket

If you own a digital manometer, you can use an ordinary plastic laundry basket to measure the air flow of a residential HVAC register or grille. This technique was developed by the ever-resourceful LBNL scientists, Iain Walker and Craig Wray. Walker and Wray contend that measurements made with a calibrated laundry basket are often more accurate than those made with a $3,000 flow hood.

Walker and Wray described the laundry-basket technique in an August 2003 paper, “Evaluation of Flow Capture Techniques for Measuring HVAC Grille Airflows.”

Walker and Wray note that a plastic laundry basket can be used to introduce a known level of resistance to the air flow leaving a grille or register. The laundry basket also provides a volume within which the turbulent air flow can settle down enough to be measured. It’s important to choose a basket with holes of a uniform size.

The typical basket needs several modifications to minimize “flow nonuniformity.” Most laundry baskets have too many holes; Walker and Wray solve the problem by covering most of a basket’s holes with something that should never be used on ducts: gray duct tape (see Image #6, below). “The optimum place to leave the open holes is in the middle of the basket, about halfway up,” says Walker. “You are aiming to leave about 30 square inches of open holes.”

With the basket placed over a grille, a digital manometer is used to compare pressure readings inside and outside the basket (see Image #7, below). The inside reading is taken from the bottom of the laundry basket — that is, the area of the basket farthest from the grille. To help average the air pressure in the region being measured, Walker recommends installing a coil of soaker hose (a canvas hose used to irrigate garden plants) around the inside perimeter of the basket’s base. The two ends of the circular hose connect to a plastic tee, while the third end of the tee is connected to a length of ¼-inch clear plastic tubing that leads outside the laundry basket and connects to the manometer.

The basket is further modified by installing a diffuserIn a forced-air heating/cooling system, the diffuser is a register or grille attached to ducting through which heated or air conditioned air is delivered to the living space. In a tubular skylight or an electric light fixture, the diffuser is a cover plate through which scattered light is delivered. screen across its mouth (see Image #8, below). “With some registers — registers with highly directly flows — there is a problem caused by all of the air flow being directed to one side of the basket,” said Walker. “So we put a diffuser screen near the open side of the basket to break the flow pattern. It’s possible to use either a coarse bug screen or landscaping fabric. The easiest way to install it is just to duct-tape it in, or, if you want to get really sophisticated, to install it with double-sided tape and Velcro.”

The final step is to attach weatherstripping around the rim of the basket, where it will be pressed against the floor, wall, or ceiling.

Different laundry baskets will of course introduce different levels of air flow resistance, so each laundry basket must be calibrated — that is, the manometer readings obtained using the basket must be compared to air flow measurements obtained with equipment known to be accurate. “We did the calibration to account for the discharge coefficient so no one else has to,” says Walker. For tinkerers without access to calibration equipment, Walker suggests measuring the area of one of the open holes in the side of the laundry basket and multiplying by the number of holes. (Although this method results in less accurate measurements than those obtained using a basket calibrated with a very accurate flow hood, the inaccuracy introduced is only on the order of 10%.)

Once the area of the open holes is known, the following formula can be used to convert the measured pressure difference (ΔP) to the air flow of a grille:

Air flow in cfm = Open hole area in square inches x 1.25 x √ (ΔP in Pascals) x (1 + 0.055 ΔP)

Walker explains, “The odd term at the end of the formula, 1 + 0.055 ΔP, is what accounts for the insertion loss of the flow hood. That is, the added resistance of the flow hood reduces the flow out of the register being measured, and this factor corrects for the flow reduction. The correction factor assumes that the pressure differences being measured are in the range of about 5 to 10 Pascals. If your measurements go much above 10 Pascals, then the correction factor isn’t good enough anymore. At that point the basket should be modified by opening up more holes.”

The accuracy of a laundry basket will vary, depending on the types of air flows being measured and the accuracy of calibration. Using a laundry basket in their lab, Walker and Wray achieved accuracies from 3% to 10% for some types of air flow measurements. However, Walker is well aware of the limitations of the technique. “Both the laundry basket and the garbage bags are good for anything you would want to do except to measure duct leakage,” notes Walker. “If you try to use these tools to add up the air flow of all the registers in a house, and to compare that number with the air-handler flow to figure out the difference, your answer will only be good to an accuracy of plus or minus 15%. But that is the range you want to know about to determine duct leakage. So you cannot measure duct leakage with these techniques.”

Robert deKieffer, taking the laundry-basket idea to heart, has introduced a few technical refinements. “The problem with laundry baskets is that they are big, bulky, and hard to use,” says deKieffer. “Lately I’ve been using a wire-mesh wastebasket that I bought at Target. It’s smaller than a laundry basket. I taped the bottom and top with clear tape, which looks better than duct tape. The good thing about the mesh wastebasket is it helps you look like a professional. It doesn’t look like a laundry basket, and it doesn’t look like it came out of a dumpster.” DeKieffer's wastebasket is shown in Image #9.

DeKieffer calibrated his wastebasket by comparing his manometer readings to those obtained with a high-end flow hood. Once he had completed his calibration, he prepared a simple lookup table to convert pressure differences in Pascals to cfm, and taped the table to the bottom of the wastebasket. “The reason I like to use the trash-can technique is speed,” says deKieffer. “I can get an answer in 10 to 15 seconds and move on. It’s so much simpler than using a flow hood.”

When deKieffer has demonstrated his calibrated wastebasket to HVAC contractors, the response has been enthusiastic. Since manometer prices are dropping, the technique is accessible to more and more contractors. “I use a small Testo manometer,” says deKiefer, referring to his Testo 506-2. “It’s smaller than a pack of cigarettes and it has a magnet on the back. It cost me about $200, and it’s accurate to plus or minus one Pascal.”

Don’t abandon your garbage bag yet

Although deKieffer and Walker have performed trail-blazing work with laundry baskets, they both note that for many applications, nothing beats the good old-fashioned garbage-bag test. Using a garbage bag is certainly simpler than using a laundry basket, since no manometer is required. Walker’s lab measurements show that the accuracy of the garbage-bag method can be improved by building a plywood frame for the bag opening. This frame helps the bag maintain its shape, and provides a flat surface to better seal the perimeter of the opening. To use the bag, Walker recommends inserting a flat piece of cardboard between the plywood frame and the HVAC grille, and then pulling the cardboard away briskly at the start of the test. “The sound of the wood frame contacting the surfaces around the grille gives a consistent audible stimulus to begin timing,” wrote Walker and Wray in their paper.

Garbage bags provide a useful level of accuracy. “The garbage bag technique is in the same range of accuracy as the calibrated laundry basket,” says Walker. “The reason I like inflatable garbage bags is they are visibly more appealing. The homeowner can see the bag fill up with air, and can see if it fills up really fast or really slow. There is another advantage: if you use a flexible frame — basically a coat hanger — then you can put it on almost any grille in the house. But if the grille is in a tight spot, you can’t use a laundry basket.”

Instead of a plywood frame or coat-hanger wire, DeKieffer uses a 12-inch round galvanized starting collar — a ductwork fitting — for the mouths of his garbage bags (see Image #10, below). “We attach the garbage bags to the starting collar with a rubber band,” says deKieffer. “It’s a lot easier than trying to tape a garbage bag to coat-hanger wire.”

Author's note: Portions of this article originally appeared in Energy Design Update.

Martin Holladay’s previous blog: “What is Comfort?”

Click here to follow Martin Holladay on Twitter.

Tags: , , ,

Image Credits:

  1. Image #1: Air Conditioning, Heating & Refrigeration News
  2. Image #2: New York State Energy Research and Development Authority
  3. Images #3 and 4: Canada Housing and Mortgage Corporation
  4. Images #6, 7, and 8: Iain Walker, LBNL
  5. Images #9 and 10: Robert deKieffer

Jan 23, 2015 10:46 AM ET

How about ductless heads?
by Aaron Birkland

Is there good and accurate way measure air flow through a ductless head, for the purpose of diagnostics (say, to compute the heat output)? The ACEEE Pacific Northwest study used a small vane anemometer in some fashion, but no mention of an apparatus to constrain geometry to a known quantity (like the cross-sectional area of a flow hood), and I could not find any specific comments on the accuracy or %error of their measurements. The geometry of ductless heads (with long, narrow "supplies" and "returns" spaced relatively closely together) would seem to constrain the kinds of techniques one could use.

Jan 23, 2015 10:53 AM ET

Response to Aaron Birkland
by Martin Holladay

I don't know of any techniques for measuring air flow through a ductless minsiplit head, although I can imagine ways to make a custom tube or chute out of cardboard, plywood, or sheet metal (with appropriate gaskets) to funnel air flow to a measuring device.

There aren't many reasons for performing this type of measurement -- other than the one you mention, computing heat output, which is chiefly of interest to researchers -- since there is no need to balance the air flow through a ductless minisplit with airflow through any other appliances in a house.

Jan 23, 2015 11:26 AM ET

by Hobbit _

I tried building a flow hood; it was pretty much a lose because of
the turbulence/nonlinearities still present at the exit end. I got
better results by doing an even scan of a whole register opening
with the anemometer in averaging mode, and measuring the opening.
MPH times 88 gives FPM, over the aperture sqf gives CFM. While a
moving scan might not have given the precise *average* airflow, it
got plenty close enough for the redneck math I was doing.


Jan 23, 2015 12:16 PM ET

Testing and tuning ductwork
by Jim Baerg

Terrific topic, one I've struggled with. Here are a few comments;
First, Retrotec is selling the European Flow Finder MK2, not manufacturing it.
Secondly, Corbett Lunsford has a terrific video demonstrating 8 different testing devices:
Finally, just getting pressure readings of ductwork with a good manometer is very useful; static pressure, house pressure changes when the furnace turns on, pressure differences across bedroom doors, etc.

Jan 23, 2015 12:33 PM ET

Edited Jan 23, 2015 12:38 PM ET.

Response to Jim Baerg
by Martin Holladay

I have corrected the reference to the Flow Finder MK2 that is available from Retrotec; thanks for informing me of my error.

Thanks also for the link to the great video from Corbett Lunsford. It was interesting to see that the result of his garbage-bag test was fairly consistent with his other test methods -- a good indication that the garbage-bag method is accurate enough for many uses.

Jan 23, 2015 3:30 PM ET

Measuring Mini Split Airflow (Response to Aaron Birkland)
by Kohta Ueno

Aaron and Martin--the usual methods I have seen other do on measuring mini split airflow is some type of a capture bag or chute on the return side--NREL (who runs Building America) published a monitoring protocol for this work:

I believe that some of my colleagues at Steven Winter also did some work along these lines, but I'm drawing a blank on any written documentation--it might have come up over beers at a conference.

Jan 23, 2015 3:43 PM ET

by Aaron Birkland

Fascinating. Did you observe a significant difference between sampling the velocity at a single arbitrary point, vs sweeping the anemometer along the entire length of the outlet whilst in averaging mode?

With regard to a flow hood, I would also worry about any static pressure it introduces influencing the result, as they are designed to operate in free space, without any significant static pressure at all.

Jan 23, 2015 4:09 PM ET

Response to Kohta Ueno
by Martin Holladay

Thanks for the link. When faced with the daunting task of developing a new testing protocol, I'm glad to see that the researchers used the usual materials -- garbage bags and duct tape.


Ductless minisplit air flow testing 1.jpg Ductless minisplit air flow testing 2.jpg

Jan 23, 2015 4:11 PM ET

NREL Protocol
by Aaron Birkland


That was quite informative, especially the sections that describe the apparatus used to measure flow during initial calibration. It would seem that the NREL method is sensitive to anything that would change the relationship between fan RPM and airflow (such as a dirty filter), but the protocol addresses that.

The method that Hobbit_ ended up with seems much easier for ad-hoc measurements by homeowners, but it's nice to know that there's an accepted protocol for long-term monitoring.

Jan 24, 2015 5:07 AM ET

to Aaron ..
by Hobbit _

At any given supply register, for example, I'd find significant
variance as the duct boot behind it certainly wasn't delivering
laminar flow. Thus an average over the whole thing.

There's the issue of the grille, of course, but for narrow slats
I ignored it and didn't have opportunity to think about compensating
for wide ones like in the GreenDream bathroom-fan video. I figure
if the anemometer *spans* a couple of runs of grille it'll get
part higher airflow and part far less, probably averaging to about
the same as if the grille wasn't there. Holding a little farther
away might let the streams re-combine, too. One could always remove
the grille to be sure. There's also the question of just how far
off the edge to scan -- the anemometer doesn't have a zero radius,
so I can't stop at an exact place. Hanging maybe a third of its
radius off the edge and then going back seemed to suffice. Again,
just a hack for ballpark figures, not hard science...


Jan 25, 2015 6:50 PM ET

Great Article
by Kye Ford

Great article Martin you seem to have read my mind as I have been trying to conceive of a homemade contraption to balance mechanical/ERV's.

I have been using Hobbit's approach with an anemometer but I have been thinking about cutting a plastic 1 gallon milk jug in half and then slitting some rubber tubbing to create a gasket to make a small homemade flow hood.

I am not familiar with air disturbance but I figured for the 4, 5, or 6" round ducts I use I should be able to get a fairly accurate measurement.

This article brings out the inner MacGuyver in all of us.

Jan 25, 2015 11:49 PM ET

Edited Jan 25, 2015 11:51 PM ET.

I'm quite pleased with our Alnor 6200
by Curt Kinder

We do quite a lot of airflow testing in connection with residential duct systems. Nearly all registers we encounter and specify are designed to flow 10 - 200 CFM especially with new construction homes coming in at 1500 - 2000 SF per ton. Those flows are too little for the big commercial flow hoods intended for 2x2 commercial diffusers flowing up to 2000 CFM (That and the $3k price for one of the big boys is a major deterrent)

Before investing $1400 I diddled around with cardboard and anemometers along the lines suggested in this blog, but they are slow, unsatisfactory, and, quite frankly, too amateurish for professional / production use.

While I support the innovations described here for DIYers it's hard to imagine a pro willing to spend kilobucks on blower door rigs, duct blasters, and IR cameras piddling around with propellers and pizza boxes...just sayin'.

Jan 26, 2015 9:51 AM ET

Response to Curt Kinder
by Martin Holladay

You wrote, "It's hard to imagine a pro willing to spend kilobucks on blower door rigs, Duct Blasters, and IR cameras piddling around with propellers and pizza boxes."

True enough. But the equipment you list would cost about $9,000 or $10,000 to purchase -- possibly more. That type of equipment is certainly required if you are a home performance contractor.

But not everyone who reads GBA is a home performance contractor. Some of us are owner/builders without much money in the bank -- but with a burning interest in building science and diagnostics.

Feb 20, 2015 5:10 AM ET

Evaporative cooling indirect/direct
by James Rinkevich

I have a climate cooling problem I need to cool about 50000 BTU/hr peak with an IDEC such as an Ultracool/Mastercool II 7112 with IM-70 precooler in conditions of 0.4% occurrance of 112F DB and 70F WB. I think the current ducting 18" diameter and up duct venting (4 units) is undersized. How can these be used to determine the needed fixes?

Feb 20, 2015 7:57 AM ET

Response to James Rikevich
by Martin Holladay

I think that you have posted your question on the wrong page. Your question belongs on GBA's question and answer page. Here is the link:

If any GBA readers want to tackle James's question, we'll need some decoding first.
DB= dry bulb temperature
WB = wet bulb temperature
IDEC is a stumper, but probably means "indirect/direct evaporative cooler."

Register for a free account and join the conversation

Get a free account and join the conversation!
Become a GBA PRO!