Duct Leakage Outside the Envelope
Why this matters
A leaky duct inside the house is mostly a balance problem: the air leaks into a room it was not meant for, but it stays in the house. A leaky duct in the attic, crawlspace or garage is a different animal. It throws away conditioned air the customer paid for, it changes the pressure of the whole house every time the blower runs, and on the return side it imports whatever that unconditioned space holds: attic heat and insulation dust, crawlspace moisture and soil gas, garage exhaust and fuel vapor. A shop that only thinks of duct leakage as lost capacity misses half of what it does, and often misses the half that explains the complaint.
Where the envelope line falls
The envelope, or pressure boundary, is the layer that separates indoors from outdoors, and the question for every duct is which side of it the duct sits on. A vented attic above the ceiling insulation is outside. A vented crawlspace is usually outside. An attached garage is outside, and it is a fire separation as well. A sealed, insulated crawlspace or an unvented attic insulated at the roof deck may be inside, and the classification of those two spaces belongs to The Attic and Crawlspace, Inside or Outside the House. Once you know which side, the leak's effect follows from which side of the blower the leak is on.
Supply leaks and return leaks push the house in opposite directions
ATTIC SECTION, blower running
roof deck
----------------------------------------
ATTIC (outside the envelope)
SUPPLY duct -> leak blows house
air OUT into attic
RETURN duct <- leak sucks attic
air IN to system
======= ceiling + insulation =========
HOUSE
supply leak > return leak: house (-)
return leak > supply leak: house (+)
A supply leak to outside takes air the blower drew from the house and dumps it outside the envelope. The return keeps pulling its full share from the house, but less comes back through the registers, so the house loses air and goes negative. Replacement air comes in through every crack, chimney and recessed can.
A return leak to outside pulls outside-the-envelope air straight into the system. The supply side delivers its full share into the house, but the return pulls less from the house, because part of its intake came from the attic or garage. The house gains air and goes positive, pushing house air out through the shell. And the imported air goes through the coil and into every room.
Most systems have both kinds. What sets the house's pressure is the difference between them, which is why a house can have a lot of duct leakage and a small net pressure, and still be importing attic air into every bedroom.
Worked example, one system, two leak patterns
The system: a 3-ton air handler moving 1,200 CFM, ducts and air handler in a vented attic. Say the attic sits at 120 F on a summer afternoon and the house at 75 F. The figures for the leaks are illustrative, of a size duct testers commonly find in an unsealed attic system.
Pattern A, supply-dominant: 120 CFM supply leak, 40 CFM return leak.
- Air delivered to the house through registers: 1,200 minus 120 is 1,080 CFM.
- Air pulled from the house by the returns: 1,200 minus the 40 drawn from the attic is 1,160 CFM.
- Net: the house loses 1,160 and receives 1,080, so 80 CFM more leaves than arrives. The house runs negative with the blower on, and that 80 CFM is replaced by infiltration from outdoors and from the attic through ceiling leaks.
Pattern B, return-dominant: 40 CFM supply leak, 150 CFM return leak.
- Delivered to the house: 1,200 minus 40 is 1,160 CFM.
- Pulled from the house: 1,200 minus 150 is 1,050 CFM.
- Net: the house receives 1,160 and loses 1,050, so 110 CFM more arrives than leaves. The house runs positive and pushes house air out through the shell.
What the return leak costs in capacity. In Pattern B, 150 CFM of 120 F attic air mixes with 1,050 CFM of 75 F house air before the coil. The mixed air entering the coil is (1,050 times 75 plus 150 times 120) divided by 1,200, which is (78,750 plus 18,000) divided by 1,200, about 80.6 F, so the coil sees air nearly 6 degrees warmer than the room. The sensible heat that leak drags in is about 1.08 times the CFM times the temperature difference, the standard-air factor ACCA and ASHRAE residential practice use, which assumes sea-level air: 1.08 times 150 times 45 is about 7,300 Btu/h. Against a nominal 36,000 Btu/h system that is about 20 percent of the nameplate gone before any latent load from humid attic air is counted. Check it from the other side: 7,300 divided by 12,000 Btu/h per ton is about 0.6 ton of a 3-ton unit, which is the same 20 percent.
What each pattern looks like to the customer. Pattern A shows up as drafts, dust at baseboards and recessed cans, and a house that never quite feels dehumidified, because the negative house is pulling humid outdoor air in through the walls. If there is a natural-draft water heater or a fireplace inside the house, the negative pressure works against its venting. Pattern B shows up as a system that runs long and cannot hold setpoint in the afternoon, a musty or hot smell at registers, and in winter, house moisture pushed out into the cold walls and ceiling. Same duct system, opposite pressure, opposite symptoms.
What changes the answer. Move the leaky return from the attic to the garage and the capacity math is similar but the imported air now carries car exhaust, solvent and fuel vapor. Move it to a vented crawlspace and it imports moisture and soil gas, radon included. Bring the whole duct system inside the envelope with a sealed attic or a conditioned crawl and the same leaks stop moving house pressure, because air leaking into a conditioned space never left the house. That is why duct location matters as much as duct tightness.
The garage and the combustion zone, handled separately
Return leaks in a garage are a safety finding, not a comfort note. An air handler in an attached garage, or return duct running through one, can pull a running car's exhaust into the house. If you find a return opening, a gap in the return plenum or an unsealed cabinet panel in a garage, treat it as a carbon monoxide route: tell the homeowner in writing the same day, and if a CO alarm in the house has ever sounded with the car running, route it to the HVAC trade urgently. See related: The Attached Garage as a Pressure Neighbor.
Return leaks near a natural-draft appliance depressurize the space around it. An air handler in a basement or closet with a draft-hood water heater beside it, and an open filter slot, is a backdraft setup. Do not test it by firing the appliance; route it to a qualified gas technician. See related: The Combustion Appliance Zone and the Door That Closes It.
How to see it without a duct tester
A full leakage-to-outside number takes a duct pressurization test run with the house held at the same pressure by a blower door, and that belongs to the HVAC or energy trade. See related: Duct Leakage Test Numbers Explained for Other Trades. But several field cues point at leakage to outside without any test:
- Dirty insulation at duct seams and around the air handler, darkened where air has been filtering through it for years.
- A warm or cold attic duct joint you can feel by hand on a hot or cold day, a sign the seam is passing air.
- A house that changes pressure when the blower starts. A manometer reading the house with respect to outdoors, taken blower off and then blower on, that moves by a pascal or more is telling you the supply and return leaks to outside do not balance. A swing toward negative means supply-dominant, toward positive means return-dominant. See related: How to Read House Pressure With a Manometer.
- A closed-up room that pressurizes or depressurizes more than its door should explain.
- Disconnected or crushed flex visible from the attic hatch.
Locating the specific leak is a method choice among a duct tester, smoke and a pressure pan, covered in Duct Blaster vs Smoke vs Pressure Pan Leak Location Method Decision Matrix. The repair methods belong to Duct Sealing Methods Reference. And where you are in an attic doing any of this, the attic hazards are your own: walk only on framing or boards laid across it, watch for nails through the deck, treat old loose fill of unknown type as possible vermiculite and do not disturb it, and get out to cool off on a hot afternoon before heat sickness sets in rather than after.
The failure modes
Sealing only the supply. A crew that seals supply joints because they are easy to reach and leaves the return leaks alone moves the house from Pattern A to Pattern B: the depressurization goes away and now the system imports attic air through the return with nothing balancing it. Seal both, and check the house pressure blower on and off afterward.
Counting the leak only as lost air. A tech who writes "10 percent duct leakage, recommend sealing" has described Pattern A's capacity loss and nothing about where the house is now pulling its air from. The customer's actual complaint, dust or humidity or a smell, is answered by the pressure direction, not the percentage.
Fixing the duct and leaving the cabinet. The air handler cabinet, its filter slot and the plenum connections are part of the duct system and are often the leakiest part, because they get opened at every service.
How to know the read was right
You read it right if you can say, for the system in front of you, which side of the envelope each duct and the air handler sit on, whether the house moves negative or positive when the blower starts, what space the return side could be drawing from, and whether a combustion appliance or a garage is involved. If the pressure swing points one way and your leak findings point the other, keep looking: the biggest leak is usually somewhere you have not opened yet.
References
- See related: Duct Sealing Methods Reference; Duct Blaster vs Smoke vs Pressure Pan Leak Location Method Decision Matrix
- See related: The House as One Pressure System; The Attic and Crawlspace, Inside or Outside the House
- See related: How to Find Duct Leaks Outside the Envelope; The Attached Garage as a Pressure Neighbor
- ACCA and ASHRAE residential practice for the 1.08 sensible heat factor at standard air conditions