Insulating Sash Windows: Every Option, Ranked by What It Does

Most people insulate a sash window in the wrong order. They get quotes for double glazing at four figures a window, decide it is too much, and do nothing for another three winters, while the actual problem is a 4mm gap running the full height of the meeting rails that could have been closed in an afternoon.

The useful thing about sash windows is that somebody has measured all of this properly. Research carried out for Historic Scotland at Glasgow Caledonian University put a real timber sash and case window into an environmental chamber and tested it bare, then with each upgrade in turn, then with the upgrades combined. The numbers below are mostly from that work, and they are more useful than any sales brochure because they compare everything against the same window.

Two different problems that feel the same

A cold sash window is losing heat two ways, and they need different fixes.

Air leakage is cold air physically moving through the gaps: between the sash and the pulley stile, along the meeting rails, under the bottom rail, and around a shrunken parting bead. This is what you feel as a draught on your ankles. It is the cheapest thing to fix and it is the thing that makes a room feel cold at a given thermostat setting.

Conduction and radiation are heat passing through the glass and timber themselves. You do not feel this as a draught. You feel it as a cold surface a metre away from you, and it is why sitting near a single-glazed bay in January is unpleasant even in a sealed room.

The split matters. In the Glasgow Caledonian tests the glazed area was 55% of the whole window, but roughly 72% of the heat was going through the glass. So the glass is where conduction losses concentrate, and the joints are where the draughts are. Fix the wrong one and you will be disappointed.

What each measure actually achieved

These are reductions in heat loss through the glazing compared with the bare single-glazed window, with the resulting centre-of-glazing U-values. The bare window measured 5.4 W/m²K at the centre of the glass; measured as a whole window in a guarded hot box, it came out at 4.4 W/m²K.

Draught-proofing sits outside that table because it addresses the other problem. In the same programme, draught-proofing the window cut air leakage through it by 86%. The whole-window U-value barely moved, from 4.5 to 4.2, which is within the measurement uncertainty. That is the point: draught-proofing does almost nothing to the U-value and an enormous amount to how the room feels.

Start with the gaps, because they are cheap and they are most of the discomfort

A properly draught-proofed sash has brush pile seals routed into the timber, not stuck on the surface. The staff bead and parting bead are removed, a groove is machined into them, and a nylon pile carrier is fitted so the pile bears lightly on the sash as it passes. The meeting rails get their own seal, usually a compression or blade seal set into the underside of the upper meeting rail. The bottom rail gets a seal at the cill. The whole thing has to be set so the sash still runs freely, which is the skill in it: too much pile and the window binds, too little and you have paid for nothing.

Sizing matters more than most people realise. Pile heights are usually specified between about 4mm and 7mm to suit the gap, and if a fitter uses one size everywhere you will get some windows that jam and some that still whistle. Our article on brush seals and their pile heights goes into the routing depths.

Two practical notes on cost. Draught-proofing a single sash window in London commonly runs somewhere in the £250 to £450 range, and larger or more complicated windows more, though access and condition move it about considerably. And draught stripping is a qualifying energy-saving material, so it currently carries 0% VAT in Great Britain until 31 March 2027, after which the rate is due to return to 5%. Secondary glazing and double glazing are not on that list and are standard-rated. On a whole-house job that difference is real money.

The things you already have: shutters, blinds and curtains

If your house still has its original panelled shutters folded back into the boxing, you own the second most effective measure on the list and it costs nothing to use. Closed shutters cut heat loss by 51% in testing. They only work at night, which is exactly when you need them and exactly when the temperature difference across the glass is greatest.

The catch is that most London shutters have been painted shut, or their knuckle hinges have seized, or the boxing has been filled in by a previous owner. Freeing them up is joinery work rather than insulation work, and it is usually worth doing before anything else is bought.

Insulating the shutter panels raised the figure to 60%. In the test that was done with a 9mm aerogel blanket (thermal conductivity around 0.013 W/mK) let into the panels behind a 6mm plywood facing, covering 55% of the shutter area. The researchers estimated that a shutter designed from scratch to maximise the insulated area could reach 80%. That is a genuinely interesting option for a room with sound original shutters, and it is completely reversible.

Curtains are the weakest measure on the list at 14%, which surprises people. The reason is that a curtain hanging on a pole above the reveal lets warm room air fall down behind it, run across the cold glass and spill out at the bottom. A curtain fitted tight to the reveal with a pelmet closing the top performs far better than the same fabric on a pole. Honeycomb blinds did well at 36% because the cells trap still air and they sit close to the glass. Our comparison of blinds that fit sash windows without damaging the joinery covers the fixings.

Secondary glazing: the best single measure by the numbers

At 63% and a U-value of 1.7, low-e secondary glazing outperformed slim double glazing in the same tests. Combined with closed shutters it reached 75% and a U-value of 1.1, which is better than most modern replacement windows.

It works because you are adding a second sealed air layer with a low-emissivity coating and, critically, a second airtight line. It also does something no glazing upgrade can: it cuts noise properly, because the wide air gap between the original glass and the secondary pane damps sound transmission in a way a 6mm cavity never will.

The trade-offs are honest ones. You have a second frame inside the reveal, and you have to open it to open the window. Slim aluminium systems with a mill or bronze anodised finish disappear surprisingly well against a white reveal, but they are visible. If the window is in a listed building this is usually the option a conservation officer will prefer, because it is entirely reversible and touches no historic fabric; our article on secondary glazing in listed buildings explains why.

Changing the glass

Two routes here, and both mean taking the glass out of your existing sashes.

Slim double glazed units are typically 4mm glass, a 4mm to 6.5mm gas-filled cavity and a 4mm low-e inner pane, giving 12mm to 14.5mm overall. In the Glasgow Caledonian tests slim units gave 55% and a U-value of 1.9. They roughly triple the weight of the glass, so the sash weights, cords and joints all need attention at the same time, and the edge seal is shallower than a standard unit which shortens the working life. The detail on unit depths and sightlines is worth reading before you commit.

Vacuum glazing is the newer option: two panes with the air evacuated between them, around 6.7mm to 8.3mm overall, with quoted U-values near 0.7 W/m²K. That is triple-glazing performance in something thin enough to go into a rebate that would refuse a slim unit. It costs more per square metre and the range of suppliers is narrower, but for a fine-barred Georgian sash it is sometimes the only glazing upgrade that physically fits.

Both are permanent changes to a historic window. Neither is a controlled fitting under the Building Regulations: Approved Document L is explicit that a controlled fitting means the entire window including the frame, and that replacing glazing, or a window in its existing frame, does not have to meet the energy efficiency requirements. Replace whole windows and the 1.4 W/m²K standard applies instead.

Do not seal the house shut

This is the part that gets skipped. A Victorian house was built leaky, and the moisture your household produces has been leaving through those leaks for a century. Close them all and the moisture stays, condenses on the coldest surface it can find and grows mould there.

The Building Regulations handle this at the point of window replacement. Approved Document F requires replacement windows to carry background ventilators where the originals had them, and where they did not, sets minimum equivalent areas of 8,000mm² for habitable rooms and kitchens and 4,000mm² for bathrooms unless the ventilation is being provided another way. It also has a whole section on making sure ventilation is not reduced when energy efficiency measures are carried out in an existing home.

Draught-proofing existing sashes does not trigger those specific requirements, but the physics does not care about the rulebook. If you tighten up every window in a flat, make sure something else is bringing air in: an extract fan that actually runs on in the bathroom, a working kitchen extract, or a small permanent vent somewhere sensible. The connection between airtightness and condensation on the inside of the glass is direct and immediate.

The order that makes sense

1. Repair first. A window that does not close properly cannot be sealed. Get the cords, the beads and any soft timber sorted before you spend on insulation.
2. Draught-proof. Biggest change in comfort per pound, currently zero-rated for VAT, and it does not close off any later option.
3. Free up the shutters if you have them, or fit a honeycomb blind or a properly boxed-in curtain if you do not. Nearly free, and worth more than most people assume.
4. Sort the ventilation. Extract fans working, vents clear, before you go any further.
5. Secondary glazing if the rooms are still cold or noisy, and particularly if you are anywhere near a bus route.
6. Reglaze only if the appearance of secondary glazing is unacceptable to you, or the window is being fully rebuilt anyway and the glass is coming out regardless.

Steps one to four will get most London sashes from genuinely uncomfortable to unremarkable, for a fraction of the cost of glazing work. If you want that done as a single visit, our sash window draught-proofing service covers seals, meeting rails and a balance overhaul together, and the wider sash window repair service works across every London borough. If you would rather try it yourself first, the DIY draught-proofing route and its failure points is a fair place to start.