Window Installation: the Joint Nobody Sees
A high-class window is worth nothing if the joint around it is not designed. The declared performance belongs to the frame tested in a laboratory, not to the hole in the wall: it is the joint that decides how much of it survives on site.
Overview
TAV. 00The window is the only envelope component that is bought with a certified performance and then installed by somebody else, often in a hurry and almost always without a drawing. Hence the gap everyone knows and nobody measures: excellent windows badly fitted, whistling in the wind and growing mould on the reveal. One principle holds it all together, and it takes a minute to learn: the joint must be tighter inside than outside. Inside you stop the vapour pushing from the warm room towards the cold; outside you let out whatever got in anyway, and stop only liquid water. Swapping the two tapes — by far the most common mistake — turns the joint into a trap: vapour enters, cannot leave, condenses in the filling and rots it from within. The second principle is that fixing is mechanical and foam is not a fixing. Foam fills and insulates, it does not carry: the weight of the window and the wind load must be discharged into the masonry through anchors or straps, at a regular spacing and closer at the corners, where the frame works hardest. The sub-frame is not a luxury: it is what separates shell from finish. Fitted while the wall is still open, it gives a square reference, allows building around it without touching the window and — thirty years later — lets the window be replaced without demolishing the reveal. Qualified installation is covered by standards, whose logic is the design of the joint and the qualification of whoever executes it; but the substance stays a craft matter: a square opening, a continuous bearing plane, two tapes the right way round.
The site score
TAV. HThe detail of the craftthe two diagonals
The sub-frame and the bearing plane
Checking the opening, fitting the sub-frame square and plumb, mechanically anchoring it to the masonry and forming a continuous bearing plane at the sill.
«Measure both diagonals. If they differ, it is not the window you will adjust later: now is when it gets fixed.»site notebook — editorial synthesis
click a beat · arrow keys ← → walk the site · dashed pauses are the chemistry at work
The principle, not the product
Why tighter inside than outside
It is the only thing to remember, and on its own it explains nearly every pathology of the window node. Warm room air holds far more vapour than outside air, and that vapour pushes towards the cold. If the joint is open on the inside, vapour enters the filling; if it is closed on the outside, it can never leave. It condenses inside the joint, where there is neither ventilation nor sun, and the filling stays wet for years. Close it on the inside and let it breathe on the outside, and the vapour does not get in at all, while the little that slips through defects finds its way out. It is the same logic as external insulation and a ventilated roof: stop vapour on the warm side, let it dry on the cold side.
Continuous external sealing all round the perimeter is the elegant version of the same mistake: it looks more careful and does more harm.
What skipping it costs
The sub-frame: a part you pay for twice if it is missing
The sub-frame looks like an added cost and is the first thing cut when the quote is squeezed. In fact it decouples two trades that otherwise obstruct each other: with it in place the mason closes the opening, plasters and finishes without waiting for the window, and the installer arrives at an already verified square reference. Without it, the window must be fitted before the finishes and then sits on site for weeks, badly protected, at risk of knocks and of dirt in the hardware rebates. The real bill, though, arrives at end of life: replacing a window built directly into the masonry means demolishing the reveal, and therefore redoing plaster, sill and paintwork throughout the room. With a sub-frame, the old frame comes off and the new one goes on the same day.
It applies to refurbishment too: if the old sub-frame is sound and square, keeping it and fitting over it often beats demolition.
How it ages (and what betrays it)
TAV. PQuestions from the site
TAV. QIs a blower door test really needed on a refurbishment?
It is not compulsory everywhere, but it is the only way to know whether the work is worth what it cost. The point is not the final figure: run with the envelope closed and the finishes still open, the test turns an invisible defect into a located one that can be fixed in half an hour. Run after plastering, it merely certifies a problem you can no longer reach. On a partial refurbishment, where new windows meet old irregular masonry, it is even more useful than on new build: that is where joints hold the surprises.
Can I fit the window directly into the masonry opening?
You can, it has always been done, and it works until the window has to be replaced. There are two drawbacks. The first is a site one: the window arrives before the finishes and stays exposed for weeks to knocks, plaster dust and mortar splashes, which never come out of the hardware rebates. The second is at end of life: without a sub-frame, replacement means demolishing the reveal, and therefore plaster, sill and repainting the room. If the budget forces the choice, at least protect the frame with film and agree in writing who answers for damage until handover.
Is polyurethane foam enough as a fixing?
No, and it is not a matter of caution: foam simply has neither the strength nor the long-term stability to transfer loads. The window is heavy, the sash shifts the centre of gravity as it opens and closes, and wind pushes and pulls on the whole surface: these actions must be discharged into the masonry through anchors or straps, sized and set at regular spacing. Foam does two other things, both important: it fills the joint with insulation and damps vibration. Asking it to carry load means discovering two years later that the window has dropped and no longer shuts.
Should the joint also be silicone-sealed on the outside?
Not with a continuous bead all round the perimeter, which is the most elegant and most damaging mistake at the window node. The outside must be watertight against driving rain but open to vapour, because it is the side from which the joint dries: seal it hermetically and water entering through any defect — and sooner or later it will — has no way out and stays trapped in the filling. External sealing is used selectively, at singular points and over the drip, while the real barrier is the breathable tape. Continuous silicone, moreover, cannot be inspected: it hides the defect instead of signalling it.