Draught-Proofing an Old House, Opening by Opening

Sorting the doors into three piles

Doors divide into three groups and the sheet is worth ruling into three columns before the round starts. The first pile takes a seal in the rebate and nothing else: the leaf is sound, it closes on its stops evenly, and the gap is consistent enough that one profile will cover it. The second pile needs joinery before any seal is bought, because a leaf that has dropped on its hinges, bowed away from the frame at the lock side, or been painted until the rebate is full will defeat every profile on the rack. The third pile needs a threshold rather than a strip, because the gap under it is not a tolerance, it is a clearance over a carpet somebody laid in 1994.

Measure the moving joint as the full perimeter of the opening, all four sides. Air sealing is not casing: casing skips the bottom of a door because the bottom has no architrave, and air does not care about that. But the bottom is a different product — a sweep, a brush strip or a drop-down seal, sized by the width of the leaf and bought as an item — so it comes off the roll figure and goes on its own line. Count the same way at every opening and the two lines never get confused at the counter.

Take gap widths while you are there, at three heights on each jamb and at both ends of the head. Record the largest and the smallest, not an average, because the profile you buy has to seal the widest point without binding at the narrowest. A folded card, a coin, a feeler set, anything with a known thickness works; what does not work is holding the door and saying it feels about right, since a seal is asked to bridge that gap for the life of the door with no further adjustment.

Then group the openings by size before touching the arithmetic. Six identical bedroom doors are one calculation multiplied out; a front door, a back door and a pair of French doors are three separate ones. Seal comes in whole rolls with a waste allowance built in, so a run of small openings does not consume small amounts of anything — it consumes a whole roll per profile per gap band, and the leftover ends are what you go back for.

1. Walk the ground floor first with the extract fans running, since that is where the largest and most reachable leaks in an old house sit.
2. At each door, close it on a strip of paper at four points and note where the paper pulls out with no resistance.
3. Check the leaf against the frame with a straightedge before writing any product against it.
4. Write perimeter, leaf width, largest gap and smallest gap on one line per opening.
5. Mark which openings are fire doors or escape routes; those take intumescent and smoke seals to their own specification, not a draught profile off a rack.
6. Total the perimeters by size group, not by room, because the merchant sells against the group and not against the room.

The sashes, and what a film can honestly do

Single glazing loses heat two ways that feel identical from an armchair. Air comes through the joints, and the glass runs cold enough to chill the air against it, which then falls down the face of the window and crosses the floor. The second is not a draught in any measurable sense and no amount of weatherstrip touches it. It is the one that makes a sealed room with a big window still feel unpleasant to sit in.

Shrink film answers the second problem by adding a still-air layer in front of the glass, and it answers it well enough on a genuinely cold single-glazed sash that people notice within a day. It goes on the room side, taped to the frame or the bead rather than the glass, because the trapped layer is the whole mechanism and a film stuck to the pane traps nothing. Tape to a clean dry surface, tension it with a hairdryer until the wrinkles pull out, and test the tape on a hidden inch first — on old distemper, flaking gloss or bare timber it will take the finish with it when it comes off in March.

Measure to the tape line, not the glass. The film has to cover the opening plus the band of frame it sticks to, and rolls come in fixed widths that do not tile neatly onto a bay of odd Georgian sashes, so the waste is real and worth carrying in the number rather than discovering it two windows from the end. Roll dimensions come off the product's own packaging: there is no standard governing shrink film coverage, and the sizes differ between insulating kits, static cling and solar control films that solve an entirely different problem.

Then be honest about the limits, because they are what the guide is for. A filmed sash cannot be opened until spring, so no window serving as an emergency escape opening — governed by IRC Section R310 in North America and by Approved Document B in England — gets filmed, and neither does one whose trickle vent would end up behind the sheet. On a modern sealed unit the gain is small enough to be inside the noise. And if the room is genuinely damp, moving the cold surface from the glass to the film moves the condensation with it rather than removing it. Where the window is worth keeping and the improvement has to last, Historic England's advice note on secondary glazing describes the durable version of the same idea, and it is joinery rather than a kit.

The joints that never move

The fixed joints are where an old house usually loses more air than the moving ones, and they are invisible because the trim covers them. A frame was shimmed into a rough opening, packed with whatever came to hand, and the architrave went on over a void that runs the full height of both jambs and along the head. Behind skirting there is often a continuous gap onto a floor void, and under a window it is that gap, not the sash, that puts cold air across your ankles. Loft hatches, boxed pipework, the plaster around a soil stack and the hole a cable was pulled through all belong on the same list.

Sealant selection is a specification job even at this scale. ASTM C920 classifies elastomeric sealants by movement capability and by the substrates they are tested against, and ASTM C1193 is the guide to using them: joint width to depth ratio, backer rod, and the rule that a joint bonded on three sides tears itself apart the first time it moves. On the outside of an old building the substrate is the problem — friable brick, soft lime render and chalky paint give nothing to stick to, so rake the joint back to something sound, and where the reveal is genuinely powdery accept that a compressible backing and a bead is a maintenance item, not a permanent seal.

Quantify these as bead length and convert to tubes, and buy more than the count says. Every tube has an end that hardens in the nozzle overnight, a gun-load worth of waste at each changeover, and the runs that get found once the architrave is off are always longer than the ones measured with it on.

The appliance that had been breathing through those gaps

There is one way this work kills people and it is worth reading before the first roll goes on. An open-flued gas fire, a back boiler, a solid fuel stove, an open fire or a gas range takes its combustion air from the room it sits in, and in a house that leaks through every door and sash, that air arrived through the gaps you are about to close. Seal the ground floor properly, then run a powerful kitchen extract, and the path of least resistance for replacement air can become the flue, pulling combustion products back down into the room.

What governs it depends on where you are. In North America the indoor air method in NFPA 54 and the International Fuel Gas Code sizes combustion air openings by appliance input, requiring two openings — one near the ceiling, one near the floor — each with the stated minimum free area; the outdoor air methods use different factors and an unconfined space may need no dedicated opening at all. In England, Approved Document J covers combustion appliances and BS 5440-2 sets the ventilation provision for gas appliances up to 70 kW net. Neither is a calculation to settle by yourself on somebody's stove.

So write every open-flued appliance inside the sealed volume onto the survey sheet before any product is bought, keep the permanent air vents that serve them, and have a spillage test done by a registered engineer after the work rather than before it. Carbon monoxide alarms in the affected rooms are a five-minute purchase on the same trip. A sealed room with a working alarm is a good outcome; a sealed room with a bricked-up air brick and a cheerful new draught excluder is the other one.

Proving it, and what a tighter house does next

Repeat the depressurised round with the same fans running, the same doors open and the same smoke source, ideally in similar weather. That comparison is worth more than any single reading, because it uses the house as its own reference and finds what you missed rather than scoring what you did. The openings that still pull will mostly be the ones where a profile was chosen against a remembered gap instead of a measured one.

The measured version puts a number on it. A blower door reports flow at fifty pascals, and that flow taken over an hour, divided by the volume of the house, gives air changes per hour at the same pressure, which is the currency every standard and target is written in. Watch the unit the fan reports in before dividing anything: a reading in cubic metres per hour divides straight into a volume in cubic metres, while one in cubic feet per minute is a per-minute figure and has to be put on an hourly footing first. The direction of travel matters more than the absolute figure in an old building: new-build thresholds under the adopted energy code, such as the air leakage provisions at Section R402.4 of the IECC, are written for construction that was detailed for airtightness from the drawings up, and no amount of strip round a Victorian sash is going to reach them.

Tightening the envelope also removes ventilation nobody had specified. Uncontrolled infiltration is a bad way to ventilate a house — it happens most on the coldest, windiest days and least in still humid weather, which is exactly backwards — but it was doing something, and the moisture from cooking, washing and breathing left through it. ASHRAE Standard 62.2 and Approved Document F both make the same point from opposite sides of the Atlantic: the answer to a tighter house is deliberate ventilation, in the wet rooms first, not a return of the gaps.

Watch the first winter for where that moisture goes instead. It condenses on the coldest surface in the room, which in an old house is a window reveal, an uninsulated lintel, an external corner or the wall behind a wardrobe on the north side. Mould appearing on a reveal a month after draught-proofing is not a coincidence and not a reason to undo the work; it is the humidity that used to leave through the sash telling you the extract fan runs for four minutes and the bathroom door is shut too soon.

Keep the trickle vents open. They are the controlled version of what you removed, they are the cheapest ventilation in the building, and taping them shut on a cold night is the single most common way a good draught-proofing job turns into a damp complaint by February.