0800 690 6513









Chris Wood, Bill Bordass, and Paul Baker

October 2009





This Executive Summary should be read in conjunction with the research report prepared for English Heritage by Dr Paul Baker, at Glasgow Caledonian University. It summarises the report’s conclusions, and sets the timber sash window research into a broader context.



English Heritage supports national and global efforts to reduce energy consumption and greenhouse gas emissions, in which existing buildings are being asked to play an ever-increasing role. A few years ago large-scale reconstruction was widely advocated, but this is now seen to be impractical and the emphasis has moved towards refurbishment, with a presumption of window replacement. English Heritage is questioning the extent to which this is necessary or desirable, because:

• traditional windows can be very durable: many original Georgian and Victorian windows are still in place, whereas modern windows tend to be designed to have very much shorter lives (typically 20 years);

• current calculation methods may be pessimistic about the performance of traditional windows and the opportunities for improvement; and

• window replacement can easily destroy the character of a traditional building, as has been widely demonstrated over the past 30 to 40 years in nearly every part of the UK.

Consequently, English Heritage has funded research at Glasgow Caledonian University [GCU], carried out in conjunction with Historic Scotland, who commissioned similar work on a different timber window. The English Heritage work also included tests on condensation.

The results presented here are part of an onging programme of research into the thermal performance of traditional building materials and components. Testing of a steel and cast-iron windows with leaded lights is currently underway.



English Heritage provided GCU with a traditional 2 x 2 traditional double-hung vertical sliding sash window in poor condition. The window, which measured 1.77 x 1.16 m overall, was installed between two environmental chambers, one of which (“the cold room”) was maintained at 2°C to simulate outdoor winter conditions, whilst the other (“the hot room”) was held at 22°C to represent room conditions. Heat-flow tests were undertaken with the window as-found, after repair, and again after draughtproofing. The effects of adding curtains, blinds, shutters and secondary glazing were also evaluated. Formation of condensation was also monitored with and without secondary glazing, and with the window both open and shut.



Heat loss from a room through a window during the heating season is complex, with three main mechanisms:

• By convection and conduction, from the warm room air to the colder surfaces of the glass and the frame.

• By the colder surface of the window absorbing infra-red radiation from the room.

• By uncontrolled air leakage, which can either bring in cold air from the exterior or take warm air out from the interior; often called air infiltration, this can occur even when the window is closed.


From the first two columns of Table 1, it can be seen that relatively simple methods of

insulation can substantially reduce heat loss through the glass of single-glazed windows. In addition, the increased internal surface temperature of curtains, blinds, shutters and secondary glazing (shown in the third column) will limit downdraughts and reduce radiation losses, which may make the room feel more comfortable at a given temperature. The effects of the frame were estimated by relating the measured U-values for the glass of the single glazed window to those from sophisticated hot-box tests of a similar traditional window at the National Physical Laboratory. This then allowed the two-dimensional heat conduction model FRAME to be calibrated and used with the empirical values to estimate the overall conduction heat loss of the window assembly. The results are summarised in the fourth and fifth columns of Table 1, which show that estimated heat losses by conduction and radiation

through the window as a whole can be reduced by:

• 40 to 50%, simply by closing curtains or lowering plain blinds;

• 50 to 60%, by using shutters or insulating blinds with reflective surfaces facing outdoors;

• over 60%, by using secondary glazing with a low-emissivity coating; and

• even more where curtains, blinds or shutters are used alongside secondary glazing.


The results show that a traditional window with low emissivity (“low-e”) secondary glazing is perfectly capable of meeting the regulatory requirements for new buildings during the day, and can do still better at night, when the blinds and shutters are closed. Further savings could be made, for example, if the secondary glazing used insulating frames (made for example of timber), or if it were to incorporate double glazing, as is widespread practice in northern Europe.



Heat loss by air infiltration was inferred by setting up air-pressure differences across the window assembly and determining the relationship between air flow and pressure. At a 50 Pa pressure, the leakage rate through the window as-received was 183 m/hour. Repairs cut leakage to 120 m/h, and in-situ draughtproofing (which still left the window operable) to 26 m/h with the window closed. With secondary glazing installed as well, leakage fell to 8 m/h.



None of the results for conduction heat loss shown in Table 1 are significantly affected by the measures to reduce air leakage discussed in Section 5. To get the best performance it is important to tackle both aspects. Indeed, we estimate that if the window tested had been installed in a typical house, air leakage would have been responsible for:

• over 60% of the overall heat loss through the window in its as-received state;

• about half its overall heat loss after the joinery was repaired but the window was not

draughtproofed; and

• less than 20% of the heat loss if the window were fully draughtproofed but not insulated.

By combining repair with draughtproofed secondary glazing, total heat loss could be reduced to one-quarter of that of the window in its original state; and by even more at night with shutters, curtains or blinds in place. Thus it is certainly not essential to replace existing windows to obtain levels of improvement in thermal performance that make traditional timber sash windows comparable with standard modern windows




The original window was so leaky that air infiltration would have met most wintertime ventilation requirements for occupied rooms, even when it was closed. The refurbished window would likewise have allowed sufficient ventilation, but draughtproofing reduced the measured air infiltration rate to about 60% of that achieved by a standard trickle ventilator (such as those often fitted to modern windows).

At this stage, one needs to reconcile heat retention with ventilation and moisture removal. In some rooms, it may be possible to leave some windows tightly closed, because sufficient air can be provided elsewhere (e.g. via other windows, ventilators, or mechanical systems). In other cases – especially where there is only one window per room – the window will need to provide ventilation as well. With secondary glazing, the simplest approach is to open the inner window by a small amount, and rely on infiltration through the original window; if necessary this original window can be opened slightly too. The risk is that this might cause condensation problems, so to study the possible outcomes GCU undertook some tests of condensation with

secondary glazing.



While occasional condensation is acceptable, persistent condensation is unslightly and can lead to decay of materials and mould growth. The condensation tests involved collecting and weighing the amount of water deposited on the inside of the original window over a 5-hour period, with the warm room at 22°C with a relative humidity of 60%, and the cold room at 2°C as usual.

The results for condensation on the inside of the original window were as below. No

condensation was found on the secondary glazing itself. When the secondary glazing was closed and reasonably well sealed (as in the tests), it protected the primary window from condensation. With the secondary glazing slightly opened, the

presence and amount of condensation on the original window depended on the direction of air movement:


• if the outside air flowed inwards to the room, there was no problem; however

• if there was little or no flow, condensation on the original window increased; and

• if the flow was outwards and the main window was also open, condensation increased further.

The results demonstrate that there will be a risk of condensation in cold weather if the indoor air is humid and the secondary glazing is left open for ventilation, unless the flow of air can be controlled to be always, or at least nearly always, inwards from the exterior to the interior. These risks will be highest in rooms where there are high rates of moisture generation, or low rates of ventilation.



The timber sash window tests at Glasgow Caledonian University suggest that:

• There are major opportunities for improving the thermal performance of existing windows by relatively simple methods, including traditional curtains, blinds and shutters.

• There is a good potential for improvement from draughtproofing, with air infiltration

through the repaired and draughtproofed window being somewhat less than through a

standard trickle ventilator.

• There is potential for further improvement where secondary glazing with a low-emissivity coating is used as well. This gives good performance in the daytime, and better still at night when curtains, blinds and shutters can be closed.

However, when designing secondary glazing to avoid heat losses, it is important to ensure that ventilation is sufficient, and that the risk of condensation is minimised. The box below provides some simple guidelines for this.


As discussed above, draughtproof secondary glazing with a clean low-emissivity hard

coating facing the outside offers major reductions in heat loss through existing windows by controlling conduction, radiation and air leakage. There are potentially further opportunities to reduce conduction losses if double or vacuum glazed secondary units can be accommodated, and by improving the thermal insulation of the frames. However, secondary glazing can be relatively airtight, so other means of ventilation may need to be considered. Condensation on the primary system may arise if the secondary system is opened for ventilation in cold weather and/or where rooms are relatively humid, (typically owing to high rates of indoor moisture production and/or low ventilation rates), and the air leakage is outwards


 These condensation risks will be minimised where the secondary glazing is either:

• able to be kept closed in cold weather, because there are alternative means of


• located where the normal direction of air flow is from outside to inside, e.g. on the

windward side of a building, on the lower floors, or where a designed natural or mechanical extraction system helps to ensure inward airflow

• fitted with devices which avoid reverse airflow in adverse circumstances, e.g. a one-way trickle ventilator or a small fan incorporated in the window; or

• where the primary and secondary window assemblies incorporate some alternative

means of ventilating between the exterior and the room interior, but bypassing the

cavity between the primary and secondary glazing (e.g. a bypass trickle ventilator on the secondary glazed unit).



This report summarises the results of research on the thermal performance of traditional windows and methods of reducing heat loss carried out by the Centre for Research on Indoor Climate & Health, Glasgow Caledonian University [GCU] on behalf of English Heritage.



In 2004, the UK’s carbon dioxide emissions stood at 559 million tonnes per year, with 27% of this attributed to the energy used in people’s homes. It has been estimated that approximately a third of the CO2 emissions from the average home could be saved by adopting simple energy saving measures, but achieving further reductions in carbon emissions from UK households to meet the UK Government’s 80% target is a major challenge. Some hold the opinion that traditionally-constructed buildings are energy inefficient and should be replaced with new build rather than refurbished. However, whilst the operational carbon emissions of new buildings are often lower than traditional buildings, the latter already embody carbon and energy is required to demolish them and dispose of the resulting waste, and produce and transport new building materials. Existing buildings also have cultural and societal value. The

question is how can we maintain our architectural heritage whilst improving the existing housing stock in response to climate change and the urgent need to reduce CO2 emissions?

The options for upgrading the thermal performance are particularly limited for pre-1919 dwellings with solid wall constructions. Traditional single-glazed windows are considered as perhaps the easiest option for replacement with modern double glazing. Traditional windows are often considered to be draughty, prone to condensation and hard to maintain. On the other hand, with good care and maintenance traditional windows may outlast modern replacements and may be considered as a sustainable resource. The heat lost through a single glazed window is about twice that through a double glazed window meeting the current UK building regulations and standards (for example, a reference U-value of 2.0 W/m2 K is given for SAP calculations [1]). Whilst secondary glazing may be effective as an option to preserve existing

traditional windows, there is little information on the performance of more traditional (and cheaper) methods of reducing heat loss, such as shutters, blinds and curtains.

The work presented here quantifies the effectiveness of relatively simple measures to improve the thermal performance of traditional windows by draught-proofing, and using blinds, curtains, shutters and secondary glazing.

A traditional Victorian sash window with single glazing was provided for testing. The window was mounted in an insulated panel between the two independently controlled rooms of an environmental chamber at GCU. Under a 20°C temperature gradient, the conductive heat flow through the glazing was measured using heat-flux sensors for the glazing only and with the various improvement options. The reduction in conductive heat loss and U-values were estimated. The improvements in the air-tightness were assessed after refurbishing the joinery, and again after draught-proofing. Condensation tests were also carried out on the window

before and after installation of secondary glazing, since the assessment of condensation risk should be considered as part of the overall evaluation of the improvements made to the window.



English Heritage provided an old two-by-two panel sash window for testing. The condition of the window was poor, with a decayed sill and the frame out of true, which resulted in a particularly bad gap between the top of the upper sash and the case. The paintwork and putty had also deteriorated, and one pane was cracked.

The outer frame dimensions were approximately 1770 mm high by 1160 mm wide, and 130 mm deep. The exposed sash dimensions were 1557 mm high by 923 mm wide. Each of the four panes was approximately 71.5 mm high by 42.5 mm wide, giving a glazed area of 1.22 m2, compared to the total window area of 2.05 m2

(i.e. 59% glazing).Top:


After the first series of tests on the window in the as-received condition, the basic joinery was refurbished by skilled joiners from Historic Scotland [HS], and additional improvements were carried out by GCU as follows:

• Frame squared up (HS)

• Part of sill and outer section offrame replaced (HS)

• Broken pane & putty replaced (HS)

• Gaps filled with plastic wood (GCU)

• Sash boxes and top of window sealed off with plywood and sealant (GCU)

• Coat of white primer applied (GCU)

The window was subsequently draught-proofed using a proprietary system, which

comprises a flexible sealant applied to the frame after careful preparation. One surface

is coated with a detergent solution to prevent the sealant from sticking; the sealant is then applied to the adjacent untreated surface. The sealant is allowed to cure and the

detergent removed, and the sashes can then be opened as usual.


The test window was installed in a 300 mm-thick insulated panel mounted between the two rooms of the GCU Environmental Chamber, with the window frame set flush with the cold face of the panel as recommended by BS EN ISO 12567-1:2000 [2]. Sealant was used around the joints between the window and the insulated panel in order to seal all gaps and hold the window firmly in position.

The Environmental Chamber is designed to test the performance of building materials & components under the range of climate conditions experienced in the UK. The chamber consists of two walk-in rooms: an “exterior” room which can be used to simulate outdoor weather, and an “interior” room to simulate typical indoor environmental conditions. The exterior room also has the facilities to simulate driving rain and solar radiation (using infra-red lamps) on a wall surface. Both rooms can be pressurised. The aperture formed between the rooms can accommodate a wall up to 3m wide by 2.4 m high. By moving the interior room different wall thicknesses can be constructed. The two rooms can be controlled within the temperature and humidity ranges shown below. The temperature and humidity in both rooms, and the driving rainfall and infra-red lamps, are fully controllable from either built-in controllers or a PC.


The effect of the various options on the conductive heat loss through the glazing was estimated as follows: A U-value was calculated from the average heat flux meter reading and surface temperature difference between the outer glazing surface and the room facing surface of each option with a correction for the standardised internal and external surface resistances and the thermal resistance of the heat flux meter, using the following equation: where Tand Tare respectively the internal and the external surface temperatures, and Q is the heat flux. The term 0.17 is the sum of the standard internal and external surface resistances. The term 6.25 × 10 is the correction for the heat-flux meter. This approach is justified because the boundary conditions in both rooms of the chamber are unknown and would require extensive calibration outside the scope of this investigation. However, steps were taken to reduce excessive air movement in both rooms by baffling of the air conditioning system. Without baffling, it was observed that the heat flux increased, and calculating the glazing U-value from the heat flux divided by the air temperature difference gave unreasonably high results.

The reduction in conductive heat loss was calculated from the improvement in U-value due to the option compared to the value for single glazing only.

The temperature of the surface of the curtain, shutter or blind facing the interior (warm) room is also reported, for comparison with the glazing temperature of the window without the option. This gives an indication of the improved comfort that should be experienced by lower radiative losses to a better-insulated window.


The test results are shown in Table 3, and compared in the figures on the next page.

The estimated uncertainty in the U-value measurements is 0.3 W/m2K. In the conductive heat loss measurements it is about 6%, due largely to temperature stratification down the window. For example, during the testing of the modern roller blind the average temperature of the

inside surface of the top pane was 11.4°C, compared with 9.5°C for the bottom pane (near to the heat-flux meter locations). This stratification was confirmed by a thermographic survey of some of the options.