DOUBLE GLAZING DEBUNKED, PART THREE

Last week’s post ended with the assertion that ‘when it comes to insulation, the best window is no window at all.’ I didn’t mean to suggest that we should forego windows entirely in our houses - of course, habitable rooms need windows, and they are a regulatory requirement. But this is a false binary. What I meant is that windows can and should be consciously sized so as to achieve the best trade-off between light and heat. I put this in italics because the question ‘How much of the wall needs to be window?’ should always be asked, but more often than not isn’t; and the calculation itself is almost never made. Instead, ‘go big and go double glazed’ is the default mantra. Here I would like to demonstrate, via a calculated example, the effect window size can have on the overall rate of heat transfer of the wall it sits in, and the room it serves.

For our example, imagine a room that’s 3m square (room area 9m2) with a 2.4m ceiling. Only one wall is an external wall (with a total area of 3m x 2.4m = 7.2m²), and it contains one window. Assume that the outside temperature is 5°C and the indoor temperature 25°C, for a difference of 20°C.

Next let’s establish a single-paned and double-paned window option for the room. I looked on the WERS website and chose a manufacturer (Capral) at random, then took the worst-performing of each of their aluminium framed fixed single and double-glazed windows: 6mm clear single glazed with a U-value of 6.3, and 6mm clear/12mm air gap/6mm clear double glazed with a U-value of 3.4. Remember that the R-value is the reciprocal of the U-value, so to obtain the R-value simply divide the U-value into 1. So for the single glazed window, 1/6.3 = R0.16; for the double glazed window, 1/3.4 = R0.29.

Given these values, the single glazed window is transferring heat at the rate of 20°C/R0.16 = 125W/ m²; the double glazed window, 20°C/R0.29 = 69W/m².

Let’s also assume that our windows are 1.5 metre wide by 1.5 metres high, i.e. 2.25m² in area. So the total heat transfer of the single glazed window is 125W/m² x 2.25m² = 281.25W, and that of the double glazed window is 69W/ m² 2.25m² = 155.25W.

What area would we need to reduce the single-glazed window to in order to reduce its total heat transfer to that of the double glazed window, i.e. 155.25W? The answer is obtained by 155.25W/125W/m² = 1.24m², for example a window roughly 0.9m x 1.4m. For a 9m² room, this window clears the minimum natural lighting required by the Building Code of Australia, being 10% of the room area, or in this case 0.9m².

In our example using the windows given, it can be seen that reducing a window’s size by around 45% has the same effect as double glazing it. A shortcut way of calculating this equivalence is to simply take the difference between the two U values (6.3 - 3.4 = 2.9) and dividing the single glazed U-value (3.4) into this (2.9/3.4 x 100 = around 45%).

Note that this example hasn’t taken into account the effect of the increase in area of the wall that accompanies the reduction in window size, because for any reasonably-well insulated wall, the effect is negligible in comparison to the effect of the change in window area. But the calculation is worth doing anyway, if only to demonstrate just how terrible the insulative performance of even double glazed windows are when compared to even a moderately insulated wall!

For a 2.25m² double glazed window, there is 7.2m² - 2.25m² = 4.95m² of wall area. Assume a wall with an R value of 4.0, which transfers heat at a rate of 20°C/R4.0 = 5W/ m². The total heat transfer of the wall is 5W/ m² x 4.95m² = 24.75m². Add to this the 155.25W total heat transfer of the window, and we obtain a figure of 180W for the wall and double glazed window together. For the single-glazed example, we have 7.2m² - 1.24m² = 5.96m² of wall area, for 5W/ m² x 5.96m² = 29.80m². Add to this the 155.25W total heat transfer of the window, and we obtain a figure of 185.05W for the wall and glazed window together.

In conclusion, I hope that this and the previous two posts in this series have been persuasive in making the case that double-glazing shouldn’t necessarily be an automatic choice, and that its advantages should be weighed against other considerations such as cost, lifespan, and a more realistic appraisal of the need for natural light; also, I hope I have demonstrated that single-glazing is by no means obsolete but is very much still a viable option in many, and perhaps even most, cases.

DOUBLE GLAZING DEBUNKED, PART TWO

In last week’s post, I made the case that insulated glazing units (double glazed windows being their most common form) are neither green nor even particularly effective.

In next week’s post, I hope to back up these claims with a concrete example; first, however, a short digression is required here into how the insulative properties of building materials and elements are measured.

The basic measure of a material’s ability to transfer heat from one side of itself to the other is called its thermal conductivity, defined as the rate of heat flow through one unit thickness of a material subject to a temperature gradient. The unit of thermal conductivity is W/ m⋅K, watts per metres kelvin, or W/ m⋅°C, watts per metres Celsius. For example, the thermal conductivity of concrete is given as around 1.30 W/ m⋅°C. From this basic figure, the heat transfer coefficient of a particular material for any particular thickness can be calculated by dividing the thickness of the material (in metres) by its thermal conductivity, then multiplying this figure by the temperature differential across the material. In practical terms, this means that a 0.2m thick solid concrete wall with a temperature gradient of 20°C (e.g. the temperature on one side of the wall is 10°C and the temperature on the other is 30°C) transfers heat from one side of itself to the other at the rate of (0.2m / 1.30W/ m⋅°C) x 20°C = 3.08 W/ m2⋅°C.

Since most building elements today are not monolithic but composites of cladding, timber, insulation, plasterboard, and so on, the insulative performance of a of a complete building assembly like a wall, floor, or roof is determined by adding together the individual heat transfer coefficients of each material, plus coefficients of surface thermal resistance at the external and internal air boundaries; this figure represents the thermal resistivity of the assembly, also known as the R-value (°C⋅m2/W). The overall heat transfer coefficient, or U-value (W/m2°⋅C), is simply the reciprocal of the R-value, i.e. it can be obtained by dividing the R-value into 1, just as the R-value can be obtained by dividing the U-value into 1. Thus the higher the R value, the better the insulative properties of the element; the lower the U-value, the better the insulative properties of the element.

The R-value tells you how many watts (joules per second) of heat you can expect to transfer across one square metre of a given building element for any given temperature difference across the element. For example, say you have a simple one-room cubic building, without openings, whose walls, floor and roof all have an R-value of 4.0. The outside temperature is 5°C and the inside temperature is 25°C. That means you have . Rearranging the equation 20°C⋅m2/W = R4.0 to 20°C⋅m2/R4.0 = W gives us a value of 20/4 = 5W per m2. Meaning we are losing 5 watts of heat from the inside to the outside for each square metre of wall/floor/roof. Suppose the building is 5m long by 5m wide by 3m high, giving a total surface area of 110m2. 5w/m2 x 110m2 = 550 Watts, meaning that to maintain the 25°C temperature in the room you would need to run a 550w heater.

Whereas the insulative ability of solid building elements such as walls and floors is usually indicated by an R-value, that of windows, in contrast, is given by a U-value. The reason R-value is not used for windows is that while R-values for well-insulated walls might be as high as 8 or more, the typical window, at least historically, has an R-value of less than one, and these numbers are unwieldy for use in calculations. In any case, that different values are needed for measuring the insulative performance walls and windows should serve to remind us of the fact that even the best double glazed window is a poorer insulator than a minimally insulated stud wall. In other words, when it comes to insulation, the best window is no window at all. A bold statement, perhaps, but one I will support with a calculated example next week.

DOUBLE GLAZING DEBUNKED, PART ONE

Insulated glazing unit (IGU) is the industry name for any glazing product that consists of two or more panes of glass separated by a metal or polymer spacer, with the whole assembly forming a thin sealed chamber that contains an insulating layer of air or other gas (typically argon). Insulated glazing was first patented as far back as the 1860s, and IGUs have been commercially available since the 1940s. Though triple, quadruple and even sextuple glazing is available for use in colder climates, double glazing is by far the most common type of IGU seen in Australia, where it has steadily gained market share to the point that it is now arguably seen as the standard choice (at least outside the tropics) in new houses, particularly since achieving first a five-star, then a six-star energy efficiency rating became mandatory in most states in the 2000s. IGU’s themselves are not mandated in the building code, but they are one of the easiest ways to ‘tick the boxes’ in the formal and largely meaningless exercises known as thermal energy assessments (which is a whole other subject in itself). Indeed, double glazing has become somewhat emblematic of ‘green’, ‘eco’ or ‘sustainable’ architecture - feel-good, nebulous and largely sham concepts that generally indicate the uncritical application of energy-intensive, high-tech solutions to perceived ‘problems’ in building design and construction.

But does double glazing work? Well, that depends what you mean by ‘work’. IGUs perform as advertised out of the box, but will they work for the lifespan of your house? Almost certainly not. Lifespans (and warrantees) given for IGUs range from around 10 to 25 years; the failure mode is almost always the failure of the seal, and an IGU is only an IGU as long as the seal retains its integrity. If you look closely at the strip of metal or plastic separating the panes of glass in an IGU, you will see two rows of tiny holes. Under these holes is a layer of desiccant. Once the seal fails, moist air enters the gap, the desiccant eventually becomes saturated, and all you have at that point is two expensive and very closely spaced single-glazed windows prone to internal condensation. If being ‘green’ is your concern, bear in mind that the whole IGU must now be replaced, with all the additional embodied energy that implies.

A sectioned timber-framed IGU showing the desiccant layer (white) under a perforated metal strip

Older, low-tech alternative to IGUs exist that provide much of the insulative benefits of IGUs without the limited lifespan. One very old solution is the use of external storm shutters, but these have the disadvantage of not being able to be used during the day. A more modern solution, common in cold climates from the early 20th century until the advent of IGUs, is just to use two single-glazed openable units in a single frame, separated by ten centimetres or so. While the large gap does mean that there will be some convection of air which will reduce the insulative performance, it also allows the internal faces of the panes to be easily cleaned, and the fact that the cavity is not sealed means that there is no seal to fail - the inevitable fate of all IGU’s in the end.

But perhaps the most fundamental ‘solution’ to this ‘problem’ of heat transfer across windows doesn’t require the application of technology at all, either high or low. Rather it simply requires a change of attitude, which is perhaps why it is almost never mentioned. It requires us to go right back to basics and challenge one of the assumptions that underlies the adoption and perceived necessity of double glazing in the first place: the idea that larger windows are always better and more desirable than smaller.

In next week’s post, we will demonstrate how this ‘no tech’ approach works, by first reviewing the physics of heat transfer and looking at how the insulative properties of materials and building elements are measured and calculated, and then applying this knowledge via a practical example to highlight the influence of window size on heat loss from a room or building.

LOCAL HEATING

Heating the entire volume of a house or room with a fan-forced convection device such as a split-system air conditioner is a very recent luxury. Before gas and electricity, heating was far more ‘local’ to the body, and was usually achieved with a radiant heat source, be that an open fire, stove, or brazier. Then as now, conductive heating was also employed, and at the most local level possible: by using the heat of the body itself to warm the layer of air trapped between it and clothing or blankets.

In the unsealed and uninsulated traditional Japanese house, there were three main ‘stations’ of heat that the inhabitants used to keep warm throughout the day and night: the kotatsu, the bath (heat by conduction), and bed.

The kotatsu is an excellent example of the kind of evolved emergence and holistic integration of parts that is so often found in vernacular ‘design’. It is a low table with a top that sits loose on the frame; between the frame and top is sandwiched a padded futon (here meaning a blanket or quilt rather than ‘mattress’) which drapes down on each side to the floor and is placed over the laps of those sitting at the table, so enveloping their legs in the heated space created between the floor and the futon.

A modern Japanese kotatsu

In the modern version, the heat source is a small electric space heater attached to the underside of the frame. In the traditional version, the hori-gotatsu or ‘sunken’ kotatsu (presumably evolved from the irori, the hearth sunk into the floor of Japanese ‘living rooms’ in farmhouses and elsewhere), there is a pit sunk into the floor that contains a small charcoal brazier and is covered by a grate flush with the floor to protect the legs. In some cases, there is a pit for the legs roughly the size of the table itself and the depth of the lower legs, so users can sit as if in a chair rather than cross-legged; the brazier is contained in a smaller pit within this pit.

Extended family gathered around a farmhouse irori.

The modern kotatsu (top) and the more traditional hori-gotatsu (bottom).

The key to the effectiveness of the kotatsu is in the clothing of those using it: traditional Japanese clothing such as the kimono are open at the bottom, allowing the heat from the kotatsu to rise up into the space between the clothing and the body; the clothing can also be drawn closed or open at the neck to prevent or allow the heated air from escaping as necessary. The kotatsu also forms the locus of the social activity the Japanese call kazoku-danran: sitting together in a family ‘circle’ to eat, talk, play games, and so on. So the kotatsu can be seen as part of a system, a highly satisfying vernacular solution that integrates not only the function of heating with the furniture and the architecture, but also with the clothing, and even with the manner of social interaction.

A birds-eye view of kazoku-danran around the kotatsu

Similar solutions can be found in the west, though perhaps not so sophisticated as the kotatsu. The high-backed, winged armchair, for example, achieved its form for functional reasons in the days before central heating. When faced towards an open fire, the cupping shape of the chair collects the radiated heat; the high back and wings block cold draughts to the head, and the the arms allow a blanket to be more securely draped over the legs.

DESIGN CONDESCENSION

From time to time I come across articles on interior design blogs or in other places where the writer traces the development of a particular aspect of architectural or interior design through its history. In these articles, there is often a faint undercurrent of condescension or superiority, as if to say, ‘haha look at those silly premoderns, luckily we moderns know better.’ This attitude is driven by an underlying assumption of inevitable and endless progress, be it social, material or technological, that confers redundancy on everything that came before the present.

A good example of this is kitchen design. The author will sketch out the history of kitchens, comparing the separated and poky little lean-to kitchens of the nineteenth century unfavourably to the modern ‘open plan’ that is ubiquitous today, and imply bafflement that anybody would have chosen to do it any way other than we do. As an aside, it is stating the obvious to point out that between the two ends of this kitchen design spectrum there are all kinds of in-between ‘semi-open’ design possibilities that allow the best of both worlds, but for whatever reason these possibilities are rarely explored; nor in any case are the eminently rational motives behind the design decisions buried in these old and ‘primitive’ kitchens.

Before electricity and even gas, all cooking was done with wood or coal, and the risk of fire was very real. By separating the kitchen off the back of the house, the risk of a kitchen fire taking out the entire house was reduced, particularly in the case of a brick house where the lean-to kitchen was effectively fire-separated from the main dwelling. Cooking fires also generate a lot of heat, which isn’t necessarily wanted in the rest of the house, especially in an Australian summer.

No electricity also means no mechanical extraction fans, so a separate kitchen was the only way of preventing smoke, soot, oil, cooking smells, and water vapour from permeating the walls and furnishings of living areas.

These are only some of the ‘technical’ reasons for kitchens being the way they were; there are also social factors that I won’t go into here. The point is that the design decisions of past buildings shouldn’t be dismissed as historical or superannuated, but rather taken seriously and even learnt from.

Design, like evolution, has no telos; design features, like the features of biological organisms, simply represent the fittest or best responses to the prevailing conditions of the environment in which they exist. If, as I believe, we are leaving our historically anomalous environment of extreme energy and resource abundance, and re-entering an environment of energy and resource scarcity that is almost beyond living memory in the first world, then we will also witness a reversal of the design ‘progress’ seen by techno-progressives as irreversible, and the re-emergence of many of the design elements, and much of the design wisdom, contained in old kitchens and other spaces.

CEILING HEIGHTS

If you’ve lived all your life in newer buildings, you’re probably familiar with the sense of expansiveness and ease you feel on entering a Victorian or Edwardian house, then noting how high the ceilings are compared to those in your own home. What happened?

Regulation of ceiling heights in Australia goes all the way back to 1810, when, under the Governorship of Lachlan Macquarie, an order was issued to the effect that “no Dwelling-house is to be less than nine Feet high” (this figure probably refers to the ‘pitching height’ of the rafters, which is de facto roughly the ceiling height). Presumably the order was felt necessary because builders and developers were trying to skimp on material costs by building low, and nine feet (2.7m) was settled on as the minimum required to provide amenity to occupants. In the Australian climate, tall rooms have the advantage of being cooler in summer, because warm air will pool near the ceiling, leaving cooler air near the inhabited zone at floor level- the difference can be 5° C or more. In the short mild winters, high ceilings presented less of a disadvantage than they do today, because heating then was radiant- open fireplaces heat surfaces and bodies directly, rather than heating the air of the entire space, as is the case with modern air conditioning systems. Taller rooms also allow for taller windows, allowing light to penetrate more deeply into rooms.

As the prosperity of the colonies grew, so did ceiling heights. In particular, a fall in material costs in the 1860s saw ceiling heights of twelve or even fourteen feet (3.6 or 4.2m) becoming relatively common in the homes of the affluent.

The 20th century saw ceiling heights swing back in the other direction. Following World War 2 in particular, austerity conditions and materials shortages put pressure on building regulations to reflect new economic realities, and the minimum ceiling height for habitable rooms was reduced from 9 to 8 feet (2.7 to 2.4m), where it remains today. This represents a reduction of just over 10% in required wall materials. Taller ceilings may also require taller cornices, skirting boards, doors and windows if they are to remain in proportion. Ceiling lights need to be more powerful or more numerous the further they are from the floor. In two storey houses, increasing the floor-to-floor height means more space and material required for the stairs. So dropping the ceiling can mean substantial savings, and for the ‘marginal’ prospective buyer whose ability to afford a house is borderline, the difference might mean being able to scrape together the deposit for a mortgage on an off-the-plan volume-built house.

While ceilings have dropped over time, the size of the average Australian house has more than doubled since 1950.  One way of interpreting this is that we've sacrificed vertical space for horizontal, and not because families have grown (they’ve shrunk), but to accommodate all our extra stuff. Vertical space is seen as not as useful for this purpose; its value is more intangible, more difficult to articulate, and harder to defend against the material advantages of ‘building out.’