Air tightness test: how a blower door test works

The written version

Airtightness: where the air gets in

These labs are a visualisation aid to help you picture how a house works. They are not technical advice: the models and figures are simplified, so please don't rely on them for design, specification or buying decisions. Ask a qualified designer, engineer or installer about your own home.

A house loses heat through its walls and roof, and through every gap that lets warm air out and cold air in. The air test puts a new house under pressure with a fan in the doorway and measures how much air gets in through the gaps. In England every new home has to pass one. This lab runs the test on a model house and then looks closely at the ten kinds of junction and service hole where the air gets in, as built and sealed.

How an air tightness test works

A calibrated fan in the back door pulls air out of the house until it is 50 Pa below the air outside. Air rushes back in through every gap, and the fan has to move as much air as the gaps let in: 1,345 m³/h in this typical new build. Divided by the area of the envelope (264 m²), that is an air permeability of 5.1 m³/(h·m²) at 50 Pa, inside the Part L limit of 8. At 50 Pa all the gaps together pass as much air as one hole of about 670 cm², a little more than a sheet of A4 paper. The three biggest are the window reveal (180 m³/h, 13% of the total), the floor edge (160 m³/h) and the joist ends (130 m³/h). Seal the five biggest and the result falls to 2.7; seal all ten and it reaches 1.7. What is left is the background: small gaps in joints and laps all over the house, which only careful work everywhere brings down.

The air test in seven experiments

  1. Pull the house down to 50 Pa

    The tester shuts every outside door and window, fits a red airtight panel in the back door and switches on the fan set into it. The fan pulls air out until the house is 50 Pa below the air outside: the water in the manometer stands 5.1 mm higher on the house side, and every wall feels the push of a 20 mph wind at once. Air rushes back in through every gap. Once the pressure holds steady, the fan is moving out exactly as much air as the gaps let in: 1,345 m³/h in this house.

  2. Divide by the envelope

    A big house leaks more than a small one built just as well, so the flow is shared over the envelope the air has to get through: the ground floor, the walls up to the top ceiling and that ceiling, 264 m² here. 1,345 m³/h ÷ 264 m² = 5.1 m³/(h·m²) at 50 Pa. That is the air permeability, q50, and Part L limits it to 8. Divided by the 265 m³ of air inside, the same flow is 5.1 air changes an hour, n50, the figure Passivhaus limits to 0.6.

  3. Draw the airtight line

    Architects draw the airtight layer round a section of the house in red, and a good design lets the pen go all the way round without lifting. In this house the line is the parge coat on the inside face of the blocks, the membrane over the ground-floor slab and the membrane above the top ceiling. Here the pen lifts 13 times at ten kinds of gap: junctions where two layers were never joined, and holes made for pipes, cables and lights. Part L asks for drawings that show where the line runs, all the way round, and what each part of it is made of.

  4. Behind the plasterboard

    Look closer at a socket on an outside wall. The plasterboard is stuck to the blocks with dabs of plaster, which leaves a gap behind it, and nobody parged the blocks. Air comes in from the cavity through the joints between the blocks, runs along the gap behind the board and out through the socket box, and the smoke pencil shows it blowing into the room. One box passes about 3.6 m³/h at 50 Pa and this house has 14. Approved Document L says plasterboard linings are not an air barrier. A parge coat on the blocks before the board goes on seals the whole wall at once, sockets and all.

  5. Tape the junctions

    Junctions leak where two layers meet and nobody joined them. Under this window the foam has cracked, and air runs from the cavity, under the frame and out beneath the window board. Airtightness tape stuck to the frame and lapped onto the parged reveal joins the window to the wall's airtight layer, and the jet stops: the windows' share falls from 180 to 11 m³/h. The floor edge and the joist ends are the same kind of problem, and these three junctions are 35% of this house's leakage.

  6. Holes for pipes and lights

    Every service that crosses the line needs its own seal. The soil pipe goes up through the bathroom ceiling into the cold loft through a hole far bigger than the pipe, and 110 m³/h of loft air comes down round it. A rubber collar slid down the pipe and taped to the ceiling membrane closes it. Better still is a hole that never gets made: with a service void of battens under the membrane, downlights and cables sit below it and never pierce it. Part L asks for holes to be core drilled, no bigger than they need to be, and sealed with grommets or collars.

  7. Seal it all, then let fresh air in

    Seal all ten and the result falls from 5.1 to 1.7 m³/(h·m²). Turn the fan off and the gaps still let air in on their own, pushed by the wind and by warm air rising: about 67 m³/h on an average winter day in the typical house, roughly 1,120 kWh of heat a year. Sealed, that falls to about 22 m³/h and 370 kWh. Part F asks for at least 117 m³/h of fresh air for this 108 m² house, or more with four or more bedrooms (133 m³/h for four), so a home this tight needs ventilation on purpose. Part F calls a home highly airtight if it is designed below 5 or tests below 3. Its guidance on background vents with occasional extract fans does not cover those homes, so it points them to continuous mechanical extract or MVHR, or to expert advice. The MVHR lab shows how heat recovery keeps the warmth.

Ten places air gets into a new house

Each detail is one of its kind in a house like this one. The flows are at 50 Pa in the typical build; a sealed detail keeps 6% of its flow.

1. Window reveal

Under a first-floor window, front wall. The frame was foamed in and the foam has cracked. Air runs from the cavity, under the frame and out beneath the window board.

Airtightness tape: Airtightness tape stuck to the frame all the way round and lapped onto the parged reveal before the board goes on. One window passes about 23 m³/h as built, and the house has 8 windows (180 m³/h in all).

2. Floor edge

Where the ground floor meets the outside walls, all round. The floor membrane stops at the wall and the parge stops above the floor. Air comes up the gap between the screed and the blocks and out under the skirting.

Membrane lapped and taped: The floor membrane turned up the wall and taped to the parge coat, which is carried right down to the slab. One metre of floor edge passes about 5.3 m³/h as built, and the house has 30 metres of floor edge (160 m³/h in all).

3. Joist ends

Where the first-floor joists meet the front and back walls. Each joist end sits in a pocket in the blocks with gaps round it, and nobody parged inside the floor zone. Air gets into the floor void and out into the rooms above and below.

Joist hangers: Joist hangers, so the blockwork runs on unbroken, and the parge coat carried up through the floor zone. One joist end passes about 2.8 m³/h as built, and the house has 46 joist ends (130 m³/h in all).

4. Soil pipe

Where the soil pipe leaves the bathroom for the loft. The hole cut for the pipe is far bigger than the pipe, and the membrane round it is torn. Loft air comes straight down round the pipe.

Pipe collar: A rubber pipe collar slid down the pipe and taped to the ceiling membrane. One soil pipe passes about 110 m³/h as built.

5. Loft hatch

In the landing ceiling, under the cold loft. A plain board resting on a timber stop, with no seal. Air comes down all round its edge.

Airtight hatch: An insulated airtight hatch with a compression seal and catches that pull it tight, its frame taped to the membrane. One loft hatch passes about 100 m³/h as built.

6. Downlights

In the first-floor ceilings, under the cold loft. Each light sits in a hole cut through the plasterboard and the membrane. Loft air comes down through it like a small chimney.

Service void: A service void: battens under the membrane drop the plasterboard 50 mm, so a slim light sits below an unbroken membrane. One downlight passes about 7.5 m³/h as built, and the house has 12 downlights (90 m³/h in all).

7. Pipes and cables

Where the water main and cable ducts come up through the kitchen floor. The membrane was cut roughly round each duct and the duct ends were left open. Air from under the slab comes up round them and through them.

Grommets: A top-hat grommet taped to the floor membrane round each duct, and the duct ends sealed round the pipe and the cable. One pipe or duct passes about 12 m³/h as built, and the house has 5 pipes and ducts (60 m³/h in all).

8. Socket

Sockets and switches on the outside walls. The blocks were never parged, so air comes through the joints between them, runs along the gap behind the plasterboard and out through the socket box.

Parge coat: A parge coat over the blockwork before the plasterboard goes on, which seals the whole wall behind every socket at once. One socket or switch box passes about 3.6 m³/h as built, and the house has 14 socket and switch boxes (50 m³/h in all).

9. Fan duct

Where the bathroom fan's duct goes out through the back wall. The hole round the duct is bigger than the duct and loosely foamed. Air comes in round the duct and behind the plasterboard.

Duct collar: A rubber duct collar slid along the duct and taped to the parge coat. One fan duct passes about 20 m³/h as built, and the house has 2 fan ducts (40 m³/h in all).

10. Letterbox

In the front door. A plain flap that never quite closes. Air comes straight through the slot.

Sealed letterbox: A draught-sealed letterbox on the inside, with brush seals and a second flap. One letterbox passes about 35 m³/h as built.

Air tightness rules in England: Part L and Part F

A limit for every new home

Approved Document L (2021 edition) sets a maximum air permeability of 8 m³/(h·m²) at 50 Pa. Since it took effect in June 2022, every new home has to be tested; before that, builders could test a sample of each house type on a site.

The limit is a backstop

The energy calculation uses the tested figure, and the notional home it is compared with assumes 5, so a leaky result can fail on energy while passing the limit. Good self-builds aim for 3 or less.

Two ways to test

The blower door test at 50 Pa, or a low-pressure pulse test, which measures at 4 Pa with a limit of 1.57. Converted with the same flow law as this lab, 8 at 50 Pa is about 1.55 at 4 Pa.

The method and the tester

Tests follow CIBSE TM23 (2022), which replaced the ATTMA standards as the approved method. It asks for the zero-flow pressure before and after, a range of readings that includes 50 Pa, and a good fit through them. Building control can accept a certificate from a tester registered with ATTMA or Elmhurst, and the results, fails included, go to building control within seven days of the final test.

How it has to be built

Approved Document L lists what makes a home airtight: drawings that show where the air barrier runs; services grouped and sealed with grommets or collars; holes core drilled and no bigger than they need to be; joist hangers considered instead of building joists into the inner leaf; plaster, a parge coat or a liquid membrane on dense blocks, because plasterboard linings are not an air barrier; taped laps on membranes and sheathing; window and door frames taped to their openings; and airtight loft hatches.

Photos of the details

The completion report for building control (the BREL report) comes with photographs of the build, including the airtightness details at the floor, walls, roof and openings.

Ventilation follows

Approved Document F calls a home highly airtight when it is designed below 5 or tests below 3 m³/(h·m²). Its guidance on background vents with occasional extract fans covers only less airtight homes, and it points highly airtight ones to continuous mechanical extract or MVHR, or to expert advice.

From 24 March 2027

The 2026 edition of Approved Document L, published in March 2026, takes effect on 24 March 2027 (a new home whose building notice, initial notice or full plans application goes in before then can still be built to the 2021 rules if its foundations and ground floor structure are complete before 24 March 2028). It keeps the same limits and a test for every new home, and asks for the test kit to have been calibrated at least every 24 months. In the SAP calculation its notional home assumes 4 m³/(h·m²) and continuous mechanical extract ventilation, a step down from the 5 it assumes today.

Elsewhere in the UK

Wales has its own Approved Document L, with the same limit of 8 and a test for every new home. Scotland (Section 6 of its domestic Technical Handbook) and Northern Ireland (Technical Booklet F1) set their own standards, so check the figures where you are building.

When to test during the build

Every fix in this lab goes on before something covers the joint: plasterboard over the parge coat and the window tape, screed over the floor membrane, and ceiling boards under the loft membrane and its pipe collars. The test that goes to building control is done on completion, when all of that is hidden. An early test, once the airtight layer is complete and the windows and outside doors are in, puts the house under pressure while every gap can still be found with a smoke pencil and sealed. The NHBC Foundation's guide to building airtight dwellings suggests a second test once the services are in, then the final one.

What happens if a house fails its air test

Approved Document L asks for the airtightness to be improved and the house tested again until it meets the limit of 8 m³/(h·m²) and the energy targets worked out with the measured result. The steps are:

  1. Report the result

    The tester's results go to building control, fails included.

  2. Find the leaks

    With the fan holding the house under pressure, the tester or the builder goes round the airtight line with a smoke pencil, or a thermal camera in winter, and marks each place the air comes in.

  3. Seal the biggest leaks first

    Tape, collars, grommets and parge go on where the line is broken. Sealing the five biggest leaks in this lab takes the typical house from 5.1 to 2.7.

  4. Test again

    The house is retested until it passes, and the as-built energy calculation uses the result that passed.

What air leakage costs in heating

Air that leaks in has to be heated. With the fan off, wind and warm air rising push about 67 m³/h of outside air through the gaps of the typical house on an average winter day, 0.25 of its air every hour. At 5°C outside that takes about 330 W of heating, or roughly 1,120 kWh over a year: about £100 at an example 9p per kWh of heat. Built like a draughty older house, the same home would lose about 2,340 kWh; built to Passivhaus (0.6 air changes an hour at 50 Pa, a q50 of about 0.60 here), about 130. A tight house still needs fresh air: Part F asks for at least 32 l/s (117 m³/h) for this 108 m² house, or more with four or more bedrooms (37 l/s for four), which a Passivhaus home's gaps, at about 8 m³/h, come nowhere near.

The test kit, the airtight layer and the leaks

Blower door test kit

Blower door

An adjustable aluminium frame with a red airtight fabric panel, fitted in the back door for the test, with the fan set into a hole in the panel. Every other outside door and window stays shut while it runs and the inside doors stay open, so the whole house is one space.

One fan is enough for a house: it only has to move as much air as the gaps let through.

Calibrated fan

A variable-speed fan whose flow has been measured in a lab, so the pressure at its inlet tells the gauge how much air is going through. Speed it up and the house is pulled further below the pressure outside.

For a very tight house the tester fits a ring over the inlet so a small flow still reads accurately. Part L asks for the kit to have been calibrated by a UKAS-accredited lab or its maker, and the 2026 edition at least every 24 months.

Digital gauge

A two-channel pressure gauge. One channel reads the house against outside, through a hose run outdoors; the other reads the pressure at the fan's inlet, which gives the flow.

Before the fan starts and again at the end it reads the zero-flow pressure: what the wind and the warm house do on their own. ISO 9972, which the UK method builds on, does not accept a test while that is over 5 Pa.

Water manometer

A U-shaped glass tube of water with one end piped into the house and the other open to outside. Pull the house down and the water climbs on the house side: at 50 Pa it stands 5.1 mm higher. That is how small the test pressure is.

The scale on this one is drawn a hundred times life size so you can see it. Testers use digital gauges; a water tube is how pressure was first measured, and why it is still quoted in millimetres of water.

Pressure hoses

Thin tubes carry the pressures to the gauge and the manometer. The outdoor one is laid away from the fan's jet and out of the wind, or gusts make the reading jump.

Test software

Steps the fan through a range of pressures, logs a reading at each, fits the curve Q = C·ΔPⁿ through them and reads off the flow at 50 Pa.

Here it takes 8 readings from 25 to 60 Pa. CIBSE TM23 asks for a range of at least 30 Pa that includes 50 Pa, a fit with r² of at least 0.98, and an exponent n between 0.5, for clean round holes, and 1, for long thin cracks.

The tester

Part L asks for an air test on every new home, carried out to the method in CIBSE TM23 by a tester registered for that kind of building. Building control can accept a certificate from a tester registered with ATTMA or Elmhurst.

Before the test the tester closes trickle vents, tapes over extract fan and ventilation grilles, fills the traps under sinks and baths, and opens every inside door. Anything else taped up is recorded on the certificate as a departure from the method. The results go to building control, fails included, within seven days of the final test.

Windsock and thermometer

Wind, and warm air rising inside a cold house, push air through the gaps even with the fan off. They also nudge the gauge before the test starts.

On a windy day the zero-flow pressure jumps about and the readings scatter, so testers wait for a calmer one.

The airtight envelope

The envelope

The whole house, with the slice on the bench marked through it. The airtight envelope is the box the air has to get through: the ground floor, the outside walls up to the top ceiling, and that ceiling. Here it is 264 m² and holds 265 m³ of air.

Part L shares the fan's flow over the envelope's area. Passivhaus divides it by the volume inside instead.

Airtight line

The red line round the section: the parge coat on the inside face of the blocks, the membrane over the ground-floor slab and the membrane above the top ceiling. It only works if it is continuous, so every junction where two of them meet is planned and joined, and every pipe or cable through it is sealed.

Part L asks for drawings that show where the air barrier runs, all the way round, and what each part of it is made of. If you can trace it without lifting your pen, the design is airtight on paper.

Cavity wall

From outside in: brick, a cavity filled with mineral wool, dense concrete blocks, a parge coat, and plasterboard stuck on with dabs of plaster. The blocks and their mortar joints let air through on their own, so the parge coat on their inside face is what makes the wall airtight.

Part L asks for the inner leaf's joints to be fully filled, and for plaster, a parge coat or a liquid membrane on dense blocks. In a timber-frame house the same job is done by OSB sheathing with taped joints, or a membrane.

Parge coat

A thin coat of sand and cement, or a ready-mixed airtight plaster, trowelled over the whole inside face of the blockwork before the plasterboard goes on. It fills the joints and the open pores of the blocks in one go.

It has to go everywhere the line runs, including behind the bath, inside the floor zone and down to the slab, where nobody will ever see it. Wet plaster straight onto the blocks does the same job.

Plasterboard on dabs

Dry-lining: plasterboard stuck to the wall with blobs of plaster adhesive. It is quick and flat, but it leaves a gap behind the board that joins every socket, switch and skirting on the wall together.

Approved Document L: internal plasterboard linings are not appropriate as an air barrier. On unparged blocks the gap behind the board carries air from one leak to the next.

Ground floor

Hardcore, a concrete slab, a membrane over it, insulation, a screed and the floor finish. The membrane is the floor's airtight layer, so the floor leaks at its edges and wherever pipes come up through it.

The hardcore under the slab is open to outside air, so wherever a pipe or the floor edge breaks through the slab and its membrane, there is a path straight in.

Airtight ceiling

Under the cold loft the airtight layer is a membrane fixed to the underside of the ceiling joists, its laps taped, with the plasterboard below it. Every light, pipe and hatch in that ceiling is a hole in it.

Air that leaks up into a cold loft carries warm, moist house air with it, and the moisture can condense on cold roof timbers.

Cold loft

The loft is outside the heated house: the insulation lies on the ceiling and the roof space is ventilated at the eaves. During the test it is outside air, so loft air comes down through any gap in the ceiling.

Service void

A zone of battens on the room side of the airtight layer, with the plasterboard fixed to them. Cables, lights and pipes run in it, so they never need to cut through the line.

Each cable or light kept in the void is one less hole to seal.

Finding leaks: smoke pencil and cutaway details

Specimen stand

One junction of the house cut open at a large scale, layer by layer, with the airtight line drawn on it in red. Choose a gap on the desk or in the panel to bring it up, and seal it to see the fix go on.

Gaps are drawn wider than life so you can see the air go through. The leaks round a window frame are cracks in the foam or sealant that fills the fitting gap, often well under a millimetre wide.

Smoke pencil

A puffer that lets out a thin stream of harmless smoke. With the house under pressure the tester holds it by a joint: if the smoke is blown into the room, air is coming in there.

A thermal camera does the same job in winter, showing cold streaks where outside air comes in.

Leaks at junctions

Window reveals

The gap between each window frame and the wall round it. Foam fills it but cracks as the frame moves, and air finds its way round the frame and in under the window board. One window here passes about 22.5 m³/h at 50 Pa, as much as a clean slot 0.26 mm wide all round the frame, and the house has 8.

Fix: tape every frame to the airtight layer all the way round before the plasterboard goes on. Part L asks for window and door frames to be taped to the openings round them.

Floor edge

Where the wall's airtight layer meets the floor's. Unless the two are joined, air comes up between the screed and the blocks and out behind the skirting, all 30 m round the ground floor. Each metre passes about 5.3 m³/h at 50 Pa, as much as a clean slot about 0.27 mm wide. A real crack is a little wider, because air rubbing along its sides slows it down.

Fix: turn the floor membrane up the wall and tape it to the parge coat, carried right down to the slab, before the screed and skirting hide the joint.

Joist ends

The first-floor joists rest on the front and back walls. Where their ends are built into the blockwork, each one leaves gaps in the inner leaf, and once air is in the floor void it comes out at the edges of the rooms above and below. This house has 46 joist ends.

Fix: Part L suggests joist hangers, so the joists never pass through the inner leaf, and the parge coat has to carry on through the floor zone. Proprietary seals for built-in joist ends also exist.

Loft hatch

A lid in the airtight ceiling that people need to open. A plain board resting in a frame leaks all round its edge, straight into the cold loft.

Fix: an insulated, airtight hatch with a compressible seal and catches that pull it tight. Part L names loft hatches specifically. A loose push-up lid is pressed shut when the test pulls the house down, but on cold days warm air rising in the house pushes up on it, so it needs the catches.

Leaks round pipes, cables and lights

Soil and vent pipe

The soil pipe rises through the bathroom ceiling into the cold loft and out through the roof. The hole cut for it is usually far bigger than the pipe, and the gap opens straight into the loft.

Fix: a collar or grommet taped to the ceiling membrane. Part L asks for holes to be core drilled, no bigger than they need to be, and sealed with grommets or collars.

Pipes and cables

The water main and the incoming cables come up through the kitchen floor, and other pipes and cables go out through the walls. Each one crosses the airtight layer.

Fix: group services so they cross the airtight layer in as few places as possible, and seal each one with a grommet or collar taped to it. Part L asks for exactly this.

Downlights

Recessed lights cut holes in the ceiling under the loft. Each one is a small chimney into the cold roof space, about 7.5 m³/h at 50 Pa, and this house has 12.

Fix: a service void below the ceiling membrane so the lights never pierce it, or airtight fittings or covers taped to the membrane.

Sockets on outside walls

A socket box on an outside wall sits in the gap behind the plasterboard. If the blocks behind were never parged, air comes through them, along the gap and out round the box: about 3.6 m³/h for each of the 14 boxes here.

Fix: parge the blockwork before dry-lining, which seals behind every box at once, or use airtight boxes where the wall cannot be parged.

Extract fan duct

The bathroom extract duct goes out through the back wall in a hole bigger than the duct. For the test the tester tapes over the fan's grille, so what leaks is the gap round the duct.

Fix: a duct collar or grommet taped to the airtight layer. In everyday use, a fan with a backdraught shutter stops air coming back through the duct itself.

Letterbox

A flap in the front door. A plain flap lifts in the wind and lets air straight through.

Fix: a letterbox with brush and flap seals on the inside, or a post box on the wall outside and no hole in the door.

The model is a two-storey house of 108 m² with masonry cavity walls and a cold loft: 265 m³ inside an airtight envelope of 264 m². Air flows through the gaps as Q = C·ΔP^0.65. Each named leak is sized by an illustrative share of a typical test, and a sealed one keeps 6% of its flow. Infiltration with the fan off uses the divide-by-20 rule of thumb, scaled for wind and temperature, and the yearly figures assume 2,100 degree days and an example 9p per kWh of heat. The details are drawn with their gaps wider than life. The numbers are the right size rather than exact for any one home.

Questions people ask

What is a good air tightness test result for a new build?

A good result for a new build in England sits well under the Part L limit of 8 m³/(h·m²) at 50 Pa, and careful self-builds aim for 3 or less. The energy calculation compares the tested figure with a notional home of 5, or 4 under the 2026 rules from 24 March 2027, so a result near 8 can pass the limit and fail on energy. In this lab a draughty older house tests at 10.7, a typical new build 5.1, a careful self-build 2.3 and Passivhaus about 0.6.

Do all new homes need an air tightness test?

Every new home in England has needed its own air tightness test since Part L's 2021 edition took effect in June 2022, and the 2026 edition keeps the rule. Testing a sample of each house type on a site ended then. The test follows CIBSE TM23, and building control can accept a certificate from a tester registered with ATTMA or Elmhurst. Regulation 43, which sets the duty, applies to new buildings, so Part L does not ask for a test on an extension or a conversion.

How does a blower door test work?

A blower door test fits a calibrated fan in an outside door and pulls air out until the house is 50 Pa below the air outside, like a 20 mph wind pushing on every wall. Air rushes back in through every gap, and the flow the fan needs to hold that pressure is the leakage: 1,345 m³/h in this lab's typical new build. Divided by the 264 m² envelope, that gives the result, 5.1 m³/(h·m²). A real test, stepping through several pressures, takes most of an hour with setting up.

What happens if a new build fails its air test?

If a new home fails its air test, the leaks have to be found and sealed and the house tested again until it passes. It has to meet both the limit of 8 m³/(h·m²) and the energy targets worked out with the measured figure, and every result, fails included, goes to building control. Leaks are found with the house held under pressure, using a smoke pencil or, in winter, a thermal camera. In this lab, sealing the five biggest leaks takes the typical house from 5.1 to 2.7.

Where do new houses leak air the most?

New houses leak most where two parts of the airtight layer meet and where pipes, cables and lights pass through it. In this lab's typical new build the window reveals pass the most air, 180 m³/h at 50 Pa or 13% of the total, then the floor edge and the joist ends, and the three junctions together carry 35%. Next come the soil pipe, loft hatch, downlights, service ducts, sockets on outside walls, fan duct and letterbox. The NHBC Foundation's guide to building airtight dwellings lists all but the letterbox.

When should you do an air test on a self-build?

Early as well as at the end: the air test that goes to building control is done when the house is finished, but an earlier test finds leaks while they can still be reached. The NHBC Foundation's guide to building airtight dwellings suggests a first test once the airtight layer is complete and the windows and outside doors are in, a second once the services are installed, and the final test on completion. Before plasterboard, screed and boxing-in go on, a leak can still be taped, collared or parged.

← All labs

Planning your own build?

Livedin maps self-build plots across England, with the register data, architects and planning context for each area. It is free to browse.