Failed Airtightness Test on a Cold-Climate House at Finishing Stage
New two-storey house in a cold climate. Walls are 38 × 140 mm studs at 400 mm o.c., OSB sheathing, 25 mm exterior XPS, 6-mil polyethylene air and vapour barrier on the warm side, lapped and sealed, with 12.7 mm board. The air-barrier target in the specification is 2.5 ACH50. The blower-door test on a nearly finished house returns 3.6 ACH50. A smoke pencil shows movement at four places: the top plate where the ceiling poly laps the wall poly between two joists; an exterior-wall receptacle box where the electrician cut the poly and left the box proud of the stud face; a bath-fan housing screwed through the poly with no gasket; and the corner where the tub drain and supply come up through the bottom plate. Three rooms are already boarded, taped and painted.
1. How do you decide which leaks to fix first, and how do you confirm a leak is real rather than a thermal artefact?
Read the worked answer
Rank by consequence, not by how easy the hole is to see. In a cold climate, an air path that lets warm moist interior air into the wall or ceiling is worse than its size suggests, because the moisture condenses on the first cold surface it reaches and stays there. That puts the top-plate lap and the bath-fan housing at the top of the list: both are near the top of the stack, both are subject to stack effect, and both feed air into a rim or ceiling cavity where it can freeze and thaw undetected. The tub penetration is next. The receptacle is the smallest hole, but it is on an exterior wall and it interrupts the barrier at a known location, so it is quick to fix and should not be skipped.
Confirming rather than guessing matters. A smoke pencil in a depressurised house shows air movement, but a plume can also come from a convective loop or from a cold surface chilling the air. I would use the smoke together with infrared under a maintained pressure difference, and tape off one suspect at a time and re-test with a second blower-door run or a zonal pressure test. That gives a number for each repair, which is what the specification target of 2.5 ACH50 is measured against. Repairs are then prioritised by measured contribution and by moisture risk, and the final test is redone after the repairs and before any of the three finished rooms are closed up again.
2. The exterior-wall receptacle has been cut through the poly. What is an acceptable repair, and what is not?
Read the worked answer
The objective is a continuous air and vapour barrier that is sealed to the box, not merely a plastic patch stapled near it. Acceptable: cut the poly back to sound material, fit a box that is listed for air-tightness on an exterior wall, or a gasketed vapour box with a flanged face, and seal the poly to the flange with a compatible acoustic sealant or the sealant the membrane manufacturer lists, tooled into the corner. The box then sits flush enough for the device plate to cover the seal. Alternatively, a listed membrane boot or gasket system made for the purpose. Either way the joint is continuous and flexible, because the wall will move.
Not acceptable: a smear of caulk around a box that is already proud of the stud, with the cut edges of poly left open behind the plate; a patch lapped over the hole and stapled only at its edges, which leaves a gap where the cable enters; expanding foam used as the primary seal; or taping the poly directly to a warm, dusty or damp box face, where the tape will let go within a season. The cable entry itself needs sealing, not just the perimeter.
I would also check every other exterior-wall box in the house. If the electrician cut this one, the same detail will have been repeated, and it is cheaper to deal with them all before the painters return than to discover them at the next blower-door test.
3. The three finished rooms include a leak location. Do you remove finished board to repair the barrier, and how do you justify the cost?
Read the worked answer
Not automatically, and not by demolishing whole walls. The question is whether the barrier can be made continuous at that point with a local repair, and what the consequence of leaving it is.
At the top plate, the ceiling poly and wall poly can often be reached from the attic or by lifting insulation; the lap can be made good and sealed from above without touching the ceiling. At the tub corner, an access panel or a small cut behind the tub or from the floor below may serve. Where the only access is through a finished wall, a clean cut around the location - roughly a 300 mm square, cut to joists or studs so it can be patched - is far cheaper than removing a sheet, and it can be sealed and patched to a level that the painter can finish. What I would not accept is reaching in through the device box and smearing sealant blind, because neither of us can then say the barrier is continuous.
Justification for the owner: the target is a performance specification, and 3.6 ACH50 against 2.5 is a measured failure that has to be remedied before the air-barrier acceptance; concealed condensation in a cold-climate wall is a rot and mould problem, not a comfort problem, and remediation after occupancy costs many times the patch. I would quantify each repair by a re-test, fix the high-contribution leaks first, and document the before-and-after numbers for the consultant rather than arguing the point in the abstract.