Airtight ductwork is one ofthe keys to a quality heatingand cooling system. It helpsto make a home more

healthful, more comfortable, and moreenergy efficient. By reducing the loadon the equipment, it also reduces wearand tear.

There is only one way to make surethat the ductwork has been thor-oughly sealed: Test it. A simple pres-sure test can measure the airtightnessof the air distribution system and helpsyou hold your installers to a high stan-dard of quality.

Of course, airtight ductwork is onlyone of the keys to quality hvac. AManual J load analysis and Manual Dduct design, proper refrigerant charge,and correct airflow over the evapora-tive coil are all just as vital. But poorlysealed ductwork can negate good workin all those other areas: You can prop-erly size the unit, lay out the distribu-tion system correctly, tune thecompressor perfectly, and set the fan

just right, but if leaky ductwork blowsmost of your conditioned air into theattic, or pulls most of the system’sreturn air from the crawlspace, you’llhave a system that costs too much torun and doesn’t do its job.

Pressure testing is not a new conceptin home construction. Plumbers have

their work pressure tested on every job,and the test forces them to get it right— meaning no leaks, period. But whenplumbing leaks, there is an immediateconsequence: Things get wet. Air leak-age, on the other hand, is invisible andseldom causes immediate trouble.Instead, homeowners gradually become

APRIL JLC 2003

A pressure test

with telltale fog

will keep your

duct installers

on their toes

by Michael Uniacke

DuctworkPRESSURE-TESTING

Ductwork

aware of comfort problems, high utilitybills, and decreasing indoor air qualityover time. The test is the only way toidentify the defects in advance.

In ten years of testing ductworkbefore drywall is hung, I’ve seen someexcellent ductwork installations andsome atrocious ones. I’ve learned thatif you want consistently good results,you have to test consistently. Alltradespeople do better work when theyknow they face inspections. As onelocal mechanical contractor’s lead manput it to me, “There are three kinds ofair distribution systems: regular,sealed, and sealed and tested.” If youtest your installer’s work, it will get bet-ter. And realistically, until you starttesting, and showing your installersthe results, they won’t even know thattheir system leaks. They certainlywon’t know where it leaks. The testmakes them face reality, but it alsohelps them learn.

The duct airtightness testing systemI use is called Minneapolis DuctBlaster, from the Energy Conservatory

(612/827-1117, www.energyconservatory.com). It consists of a calibrated vari-able-speed fan, a fan speed control, anda digital pressure-reading gauge called adigital manometer (see Figure 1). Thetest itself is pretty simple: We attach theDuct Blaster fan to the air distributionsystem at a return grille, a supplyplenum, or the blower compartment onthe air handling unit. We temporarilyseal off all the registers and grilles. Thenwe turn on the Duct Blaster fan andapply pressure. The Duct Blaster systemmeasures the airflow needed to create atest pressure of 25 pascals (a 0.10-inchwater column) in the duct system. Thisairflow rate is our duct leakage measure-ment. We compare the duct leakagereading with a recognized standard, andwe give the system either a pass or a fail.The whole process takes less than twohours (mostly for setup and takedown)and typically costs around $220.

Over the years, my company hasheaded off a lot of potential problems bytesting ductwork before the drywall ishung. When the drywall is up, we can

APRIL JLC 2003

Figure 1. The Minneapolis Duct Blaster is acalibrated variable-speed fan that’s tiedinto the heating or cooling air distributionsystem at a supply plenum, return grille, orair handler compartment (right). The digitalpressure gauge (below) calculates air leak-age rates from pressure and fan speed data.

Stev

e W

righ

t

still gain access to the ducts, but theproblems are harder to locate and assess.

Why Airtight Ductwork MattersGood duct sealing makes a major

contribution to the healthfulness,safety, comfort, and efficiency of a newhome. Sealing the ducts is even moreimportant than sealing the buildingenvelope, because when the air han-dler is running, the pressures in the airdistribution system are much greaterthan in the building or the outside air.Pressure differences are what driveinfiltration; a hole or crack in thebuilding envelope is not an air leakagepoint unless it sees a pressure. Ductsalways operate under pressure, so ahole in the ductwork is always a leak.

To clarify the importance of pres-sure, let’s look at some numbers. Wind,stack effect, exhaust fans, and the likegenerate air pressures across the build-ing shell that range from 0.5 to 10 pas-cals. On average, a house is usually inthe low end of this range. (A pascal is avery small metric unit of pressure.There are 25 pascals in 1/10 inch ofwater column. If you were to put astraw into a glass of water and suck thewater up the straw 1 inch, you’d createan inch of water column pressure, or250 pascals.)

Pressures created in air distributionsystems when the air handler is run-ning range from 10 to 125 pascals, tensof times greater than the ordinary pres-sures acting on the house envelope.And these pressures, unlike the wind,are continuous when the air handler isrunning. That is why house infiltra-tion rates can double or triple whenthe equipment is active.

Temperatures. Remember, too, thatleaks in supply ductwork involve con-ditioned air, not room air. The air thatescapes from supply ductwork hasbeen heated to 140°F or cooled to58°F, and it is not getting to the roomit was meant for. If most of the duct-work is in the attic or crawlspace(which is often the case), all thisheated or chilled air is now leakingstraight to the outdoors. The resulting

energy loss is much greater than theloss that occurs when 70°F room airleaks out around a window, for exam-ple. That’s why simple duct tighteningmay cut heating and cooling costs by15% to 30% in many homes.

Duct leaks also typically affect com-fort, because the system doesn’t deliverthe intended amount of conditionedair to satisfy the design load for aroom. Rooms at the end of long ductruns suffer the most: The farther awayfrom the fan a room is, the greater thelikelihood that air will find its way outof a leak instead of going where it’smeant to.

On the return side of the air distrib-ution system, leaks pull ambient air oreven hot attic air into the system, fur-ther compromising efficiency. Leaksalso threaten indoor air quality: Airfrom a crawlspace may carry pesti-cides, moisture, radon, and moldspores, all of which can get suckedinto the return leak, then sentthrough the supply side to everyroom in the house.

Energy penalty. How much do theseleaks affect energy consumption? A1999 summary report of 19 separatestudies from around the country(www.aceee.org/pubs/a992.htm) sug-gests that the average annual energy

savings potential in a typical housefrom sealing the ductwork is around17%. These studies include both heat-ing and cooling climates. Anywhereductwork can leak outside of thehouse, there is the potential for a largeenergy penalty. In one study,researchers at the Florida Solar EnergyCenter found that sealing ductwork inexisting homes cut cooling bills byabout a third.

Locating the LeaksWhere are these leaks happening?

Basically, leaks are possible at any jointor seam in the system, from the airhandlers supplied by manufacturers,right on through to the sheet-metalsupply boots your installer brings tothe job. If you look at all the joints,you’ll find the leaks.

There are leaks on sheet-metal trunkducts every 48 inches, where the trunksections connect. Every collar attachedto the trunk, where flex duct branchestake off, represents a possible leak. Sodoes every adjustable elbow. We alsofind leaks where flex duct connects tosupply boots, in the boots themselves,and around the plaster grounds. I’moften surprised by how leaky some ofthe manufacturers’ mechanical unitsare. I’ve tested units that leaked 40 to45 cfm right out of the box.

I used to commonly find flex ductsthat either had never been connectedor had been accidentally disconnectedby another trade. During the early tomid 1990s, most framed return cavitieswere major leakage sites. Disconnectedducts and framed cavities still tend tobe the biggest culprits, but practiceshave improved significantly.

Stud and joist bays used as returnsare still common in my part of thecountry and often present a lot ofproblems. All framed cavities shouldbe lined with OSB, plywood, ductboard, drywall, or sheet metal andthen sealed with mastic at every joint.Hvac contractors sometimes blame theframers for not lining the cavities, butwe don’t accept that excuse — if itleaks, you fail.

APRIL JLC 2003

Researchers at the

Florida Solar Energy

Center found that

sealing ductwork in

existing homes cut

cooling bills by

about a third

The most frustrating leaks are wherethe hvac installer tried to seal the systemand failed — where someone did thesealing but didn’t get it right. For exam-ple, it hasn’t yet dawned on everyonethat each joint and seam have to besealed on four sides, not just three. If acrew hasn’t had the benefit of sometraining, and has never seen a systemtested and fogged, it’s a good bet the sys-tem won’t pass. Installers just can’t pic-ture the consequences of an averageduct sealing job.

Running the TestA Duct Blaster test on a single air dis-

tribution system takes about 45 minutesto 11/2 hours to perform. Most of thattime is spent sealing off registers andgrilles and setting up the equipment.Once a system is prepared, it takes onlyabout 2 minutes to perform the test.

The first step is to attach the DuctBlaster fan to the air distribution system(Figure 2). We can attach it to theblower compartment on the air handleror to a return grille. If the air handler isnot set when I arrive to test, I connectthe Duct Blaster directly to the trunklines in the mechanical room where the

APRIL JLC 2003

Figure 2. The author tapes a plas-tic transition fitting to a piece ofcardboard cut to fit the openingin the air distribution system(above), then places the assem-bly in the opening (above right)and connects the fan to the open-ing with a length of flex duct(right). It’s best to test the systemwith the air handler installed,because the air handlers areoften a leakage point. If the airhandler has not been set, theauthor hooks up the Duct Blasterto the trunk line where the airhandler will go. The equipmentcan also be attached to a returngrille (below).

air handler will eventually be set. Butwe prefer to test with the air handlerinstalled, because of the leakage we’veseen in even brand-new units.

The second step is to seal off the airdistribution registers with a poly prod-uct called Duct Mask (Figure 3). Itcomes on a roll (also from the EnergyConservatory), so I simply run a beltthrough the roll and wear it aroundmy waist. Duct Mask is perforatedevery 4 inches and has adhesive onone side. It makes sealing up the sys-tem such a simple task that I considerit essential. But I now have somemechanical contractors who seal allthe supply registers that are installedin floors with sheet metal and mastic.This keeps construction debris andrain out of the ductwork and also pre-pares the system for testing.

The last piece of equipment to hookup is the digital manometer, which issimple to set up and operate. The digi-tal manometer I use both measures thepressure in the duct system anddirectly displays air flow through theDuct Blaster fan in cubic feet perminute, which is convenient and cutsdown on errors.

I can often tell how leaky a ductsystem is simply by turning on theDuct Blaster and seeing how quicklyor slowly the duct reference pressuresrespond. A leaky system will requirethat I ramp up my Duct Blaster fanspeed to overcome the numerousleaks in the system. In fact, in somehomes the ducts are so leaky that Ican’t pressurize the air distributionsystem at all. In a system with areturn that leaky, it’s very possiblethat more air is being drawn into thesystem from the crawlspace or atticthan is being pulled from the house.

Once the air distribution system ispressurized to 25 pascals, I read the airflow through the Duct Blaster fan frommy digital manometer.

It’s important to recognize that theDuct Blaster measures the air leakage ata test pressure of 25 pascals, not theactual duct leakage when the system isrunning. The actual duct leakageunder normal use depends on wherethe leaks are located and what pres-sures they see. The closer the duct leaksare to the air handler fan, the higherthe pressures. The typical leaks whererefrigerant lines enter the coil, for

example, will see much higher pres-sures than the leaks around the supplyboots at the end of branch runs, andthus will leak more air.

The tests are performed at a uniformpressure of 25 pascals because that rep-resents a typical average operatingpressure in residential systems andgives us a quick way to compare themeasured leakage rate to acceptedstandards.

The Fog MachineThe shortcoming of a digital mano-

meter readout is that tradespeoplecan’t picture the leakage; in theirminds, that number is not associatedwith anything. Also, we still don’tknow just where the leaks are. Byintroducing theatrical fog into the sys-tem through the Duct Blaster, we canmake most of the leaks visible (Figure 4,next page).

The fog pouring out of duct leaks is areal eye opener. No one argues with thispart of the test. Even after ten years ofusing a fog machine out in the field, I’mshocked sometimes that so much fogcan leak out so fast. It is truly telling.

In my experience, installers really

APRIL JLC 2003

Figure 3. The author carries a roll of adhesive poly Duct Mask on his belt (left) and uses the material to seal registers and grilles (right).

appreciate being on hand when theirwork is tested and fogged. It’s often thefirst time they’ve seen their worktested, and the fog drives home theimportance of paying attention.

Sealing the system. Fortunately, ductleakage problems are simple to elimi-nate during new construction if youknow where to look for leaks and howto seal them. The cost varies depend-ing on the system’s size and complex-ity, but most systems can be sealed forsomewhere between $150 and $600.Research has shown that the cost isrecovered within one to five years;after that, the savings go to the home-owner’s bottom line. And of course,the greatest benefit is a more healthfuland comfortable home.

The products of choice are water-based mastic, accompanied by a fiber-glass mesh on larger holes. The mastichas the consistency of mashed pota-toes. It is easily spread over joints inthe ductwork with an inexpensivepaintbrush or one’s hand. On cracksand gaps wider than 1/4 inch, a 2-inchfiberglass mesh tape is placed in a bed

APRIL JLC 2003

Figure 4. When a systemexceeds the allowable airleakage standard, the authoruses a theatrical fog generatorto introduce visible vapor intothe Duct Blaster fan intake.The fog makes it easy to findthe leak locations.

Stre

amlin

e A

/C

Stre

amlin

e A

/C

Streamline A

/C

of mastic to reinforce the seal. Theadvantage of mastic over duct tape isthat it provides a long-term durableseal. The water-based mastic cleans upeasily with water.

Duct Tightness StandardsThe goal of duct sealing is to ensure

that the system is sufficiently airtight.That does not mean submarine airtightor hermetically sealed. We aim not toeliminate leakage entirely (althoughsome contractors come close) but toreduce the leakage to an acceptablethreshold. I use one of two standardsfor duct system airtightness, and eachgenerates a “not to exceed” duct leak-age number.

One standard is 3% of the livablesquare footage. For example, a 2,000-square-foot home should have nomore than 60 cfm of leakage in the airdistribution system. The other com-mon standard is based on the size ofthe air-conditioning unit, which youcan read off the unit’s label (or theinstaller can tell you). It’s commonpractice to assume 400 cfm per ton, soa 4-ton unit would be rated at 1,600cfm; the air leakage should not exceed5% of the total airflow capacity, or inthis example, 80 cfm.

These levels are realistic on the job;conscientious workers can easily sat-isfy them. In my experience, if a crewcan’t meet these criteria routinely,either they don’t have enough trainingand experience, or they aren’t tryinghard enough.

Who Should Test?There are a growing number of third-

party testing companies like mine. A listof testing contractors by state is avail-able on the Energy Conservatory web-site (www.energyconservatory.com).

I believe that hvac companies shouldconduct in-house testing of air distrib-ution systems. If your hvac contractoralready owns a Duct Blaster and otherdiagnostic tools, that tells you that thecompany truly wants to do qualitywork. Companies that don’t own aDuct Blaster have to ask someone else

to test their work, as well as show themtheir mistakes.

I’ve worked with architects whomake the builder pay for the first testand the hvac contractor pay for thesecond test if he fails the first one. If asystem fails, as a part of my fee, I’llspend up to an hour on the job help-ing the crew find the leaks with myDuct Blaster and fog machine.

Reaping the BenefitOn most jobs, there is a tug of war

between quality and profit, and profitroutinely wins. This means ductwork isoften not sealed, or it’s sealed but nottested. When customers have a comfort,air quality, or utility bill problem, it addsto their distress to learn that no onebothered to implement a simple quality-control step. They begin to wonderwhat other corners the builder cut.

I always advise my builder and hvaccustomers to at least offer sealed duct-work and testing. If you offer it andthe customer turns you down, there’sa record, in case of later complaints,that the customer chose low priceover quality.

For builders who do test, we file atest report, which becomes a recordfor the house. I encourage my buildercustomers to use this report as evi-dence that they are truly committedto quality behind the drywall. If youseal and test ductwork and the con-tractor across the street doesn’t, youhave a better product, and you canuse our documentation to build yourreputation.

Satisfied customers also provide amarketing advantage. A customer whomoves into a new home that is morehealthful, comfortable, and energyefficient is more likely to tell friendsthat you are someone who builds ahome that works.

Michael Uniacke runs AdvancedInsulation, Inc., an Arizona company thatsupplies insulation, energy auditing, diag-nostics, testing, and consulting services.

APRIL JLC 2003

Sources of Supply forDuct-Sealing Mastic

AM Conservation Group, Inc. 908/852-6464www.amconservationgroup.com

Biddle Company800/878-7876www.trustsm.com

Carlisle Coatings & Waterproofing, Inc. 800/527-7092 www.hardcast.com

Carrier Aeroseal, LLC800/772-6459 www.aeroseal.com(aerosol duct sealing franchiser)

Ductmate Industries800/245-3188 www.ductmate.com

Duro Dyne Corporation800/899-3876 www.durodyne.com

Foster Products Corporation800/231-9541 www.fosterproducts.com

McGill AirSeal Corporation800/624-5535 www.unitedmcgill.com

Mon-Eco Industries, Inc.800/899-6326

RCD Corporation800/854-7494 www.rcdmastics.com

Rectorseal Corporation800/231-3345 www.rectorseal.com

Shelter Supply800/762-8399 www.sheltersupply.com