Airtightness, Infiltration and Moisture

A blower door measures at a pressure the weather never produces

Natural pressure differences across a building envelope are a few pascals and they vary constantly with wind and temperature. Measuring leakage at those pressures would be measuring noise, so the standard test imposes an artificial and repeatable difference — fifty pascals — and reports the flow required to hold it.

That gives a number which is excellent for comparison and for verifying workmanship, and which is NOT the building's air change rate. A house at five air changes per hour at fifty pascals is not exchanging five volumes an hour in ordinary weather; it is exchanging a small fraction of that, and the fraction depends on the climate, the building's height and how sheltered it is.

The conventional conversion — dividing the fifty-pascal figure by a number around twenty — is a CORRELATION fitted to measured houses, not a physical relationship. It carries real scatter, and it is applied to a building type it may not describe. Using it to produce a natural infiltration rate is defensible; presenting the result as a measurement is not, and every page here that performs the conversion states the assumption it used.

The same test can be divided two ways. Air changes at fifty pascals divide the flow by the volume inside; air permeability, q50, divides it by the area of the envelope — every wall, the roof and the ground floor — and is the figure UK regulations limit. A large building holds so much volume for each square metre of skin that it looks far tighter in air changes than in q50, and US whole-building tests report at 75 Pa (0.3 in. w.g.) rather than 50, a figure that converts only with the exponent of the building's own leaks.

The exponent says what kind of leaks you have

The power law's exponent is not a fitting nuisance; it carries information. A value near one half indicates flow through sharp, well-defined openings — holes, gaps, unsealed penetrations. A value near one indicates laminar flow along long narrow paths, such as the interface between two overlapping sheets.

Real buildings sit between, and where they sit hints at where the work is. A building with a low exponent has discrete holes and will respond well to targeted sealing; one with a high exponent is leaking diffusely and needs a continuous barrier rather than a tube of sealant.

It also matters for the conversion in the previous section. Because the exponent is not one, leakage does not scale linearly with pressure, and two buildings with identical fifty-pascal results can behave differently at the four pascals that the weather actually delivers.

Height beats area: the stack effect decides which leaks matter

Warm air inside a building is less dense than cold air outside, so it rises and pushes out at the top while cold air is drawn in at the bottom. Somewhere in between is the NEUTRAL PRESSURE PLANE where the difference is zero.

The consequence is that leaks are not equally important. A gap at the ceiling plane and a gap at the base of the building are each working under the full available pressure, while an identical gap at mid-height is working under almost none. The driving pressure grows with the height of the building and with the temperature difference, which is why stack effect dominates in tall buildings and in cold weather and is almost absent in a single-storey house in mild conditions.

In practice this makes attic sealing the highest-yield work available. Penetrations through the ceiling plane — top plates, recessed lights, flue chases, plumbing stacks, the attic hatch — sit at the top of the stack and are usually hidden under insulation, so they are simultaneously the leakiest and the least visible part of the envelope. Weatherstripping a window is visible, satisfying and worth a small fraction as much.

Wind adds a separate pattern: positive pressure on the windward face, negative on the leeward and on the roof. It reinforces stack effect on one side of a building and opposes it on the other, which is why infiltration is not steady and why a single measurement on a windy day is not comparable to one on a still day.

The driving pressure is outdoor air density × gravity × the height from the neutral plane × the temperature difference ÷ the absolute indoor temperature. It is small — a 12 m (39 ft) shed with 20 K (36 °F) across its walls sees about five pascals at floor level — and it is enough: through an open door it drives air at roughly two metres a second (400 ft/min), which is why a warehouse's heating is decided at its dock doors.

Air leakage carries far more moisture than diffusion does

This is the fact that reorganised building science, and it is worth stating flatly: the quantity of water vapour carried into a construction by AIR MOVING through a hole is orders of magnitude greater than the quantity diffusing through the same area of material. The comparison usually quoted is between a small deliberate gap and an entire wall's worth of diffusion, and the gap wins by a factor in the hundreds.

So the assembly that stays dry is the one with a continuous AIR barrier, and the vapour retarder — which is what most people think of as the moisture control layer — is the secondary measure. A perfectly specified vapour barrier with an unsealed penetration through it is not a partly effective vapour barrier; it is an air leak with a plastic sheet next to it.

The direction matters too. In a heating climate, interior air leaking outward through the assembly deposits its moisture when it reaches a surface below the dew point, which is typically the cold side of the insulation. In a cooling climate the flow reverses and so does the condensing plane, which is why an assembly designed for one climate can fail in the other.

This is also why sealing the attic plane fixes problems that look like ventilation problems. Frost on roof sheathing, damp insulation and ice dams are usually warm indoor air arriving where it should not, and adding vents to remove it treats the symptom while the air barrier treats the cause.

Build tight, ventilate right

A leaky building is ventilating itself, badly. The rate is uncontrolled, it is highest exactly when the weather makes it most expensive, it is lowest in the mild still conditions when pollutants accumulate, and it delivers air to wherever the leaks are rather than to where people are.

Removing that leakage removes a service, and the service has to be replaced deliberately. A building sealed without mechanical ventilation will have higher humidity, higher carbon dioxide, higher concentrations of everything emitted indoors, and a real risk of combustion appliances being starved or backdrafted. Every airtightness target in every standard is paired with a ventilation requirement for this reason, and the pairing is not optional.

The targets themselves are also volume-normalised, which produces a quirk worth knowing: because the figure divides by enclosed volume, a large house reaches a given air-change target with proportionally more leakage than a small one. Standards that want to compare envelopes rather than buildings therefore also quote leakage per unit of envelope AREA, which is the fairer measure of how well something was built.

Dehumidification: the rating condition is not your basement

A dehumidifier's capacity is quoted in litres or pints per day at a RATING CONDITION — a stated temperature and relative humidity, warmer and more humid than most of the spaces the machine ends up in. Its output in a cool basement is substantially lower, because both the moisture available in the air and the refrigeration cycle's performance fall with temperature.

Sizing therefore means sizing against the conditions the machine will actually see, and the load is the sum of what enters — infiltration carrying outdoor moisture, ground moisture through slabs and walls, occupant activity — rather than the volume of the room. A machine sized on floor area alone is sized on the wrong quantity, which is why the same room can need very different capacities depending on whether its walls are wet.

Pull-down is a separate question again, and the reason it takes far longer than the air volume suggests is that most of the water is not in the air. It is absorbed in timber, plaster, furnishings and stored goods, and it leaves those materials slowly and at a rate set by their own properties. A room's air can be dried in an hour; the room takes days, and the humidity climbs back overnight until the materials have given up their share.

What has to be measured

Airtightness is one of the few building properties with a cheap, decisive test. A blower door gives the whole-building figure; running it while walking the building with a thermal camera or a smoke pencil shows WHERE, which no calculation can; and a duct pressurisation test separates envelope leakage from duct leakage, which otherwise gets attributed to the wrong place.

For rates rather than leakage, tracer gas measures the actual air change in the building as it stands, and continuous carbon dioxide monitoring is a cheap proxy for whether occupied spaces are getting enough air. Neither can be inferred from a fifty-pascal number.

The calculators here convert between the quantities, estimate a target, and size the materials for sealing work. They are planning tools, and the pages say so — the number that matters is the one the test produces after the work is done, which is why airtightness targets in standards are verified rather than declared.