Rooflight U-values explained, and how to read the data sheet

A U-value is a rate of heat loss. It says how many watts pass through one square metre of a construction when the air on one side is one degree warmer than the air on the other.

The unit is W/m²K, read as watts per square metre per kelvin. One kelvin is the same size as one degree Celsius, so a difference of 20 degrees between a heated room and a cold night is 20 kelvin. A rooflight with a U-value of 1.4 loses 1.4 watts through every square metre for each of those degrees.

The whole idea fits in one line of arithmetic. Heat lost equals the U-value, multiplied by the area, multiplied by the temperature difference. Take a rooflight of 2 square metres with a U-value of 1.4, on a night when it is 20 degrees warmer inside than out. That is 1.4 multiplied by 2 multiplied by 20, which comes to 56 watts. Swap in a unit at 2.8 and the same rooflight loses 112 watts. The glass has not changed size. The room simply pays twice as much to keep its temperature.

Two points follow from this. First, lower is always warmer, because the number is a loss and not a score. Second, the area counts as much as the U-value. A large lantern at a good figure can lose more heat in total than a small flat rooflight at a mediocre one. The figure describes each square metre, and your roof has however many square metres you cut into it.

The number on the data sheet bundles three separate processes together: conduction, convection and radiation. Knowing which one a component tackles makes the rest of the sheet easier to read.

Conduction is heat passing through solid material: the glass itself, the spacer between the panes, the aluminium or timber of the frame. Convection is heat carried by moving air, and it happens inside the sealed gap between panes as well as on the room side of the glass, where air cools against the surface, sinks and is replaced. Radiation is heat leaving as infrared, straight across the gap and out through the glass, with no air needed at all.

Each part of a modern unit deals with one of these. A low emissivity coating cuts radiation across the cavity. Argon, which is denser than air, slows the convection and conduction inside the gap. A warm edge spacer reduces the conduction along the perimeter of the glass. A thermally broken frame interrupts the conduction through the metal. We take each in its own page, and the overview of how heat escapes through a rooflight puts them in order of importance.

A data sheet may quote any of three figures, and they describe different things. The glass has one, the frame has one, and the finished unit has one of its own.

The centre-pane value, often written Ug, covers the glass alone, measured well away from the edges. It is the best figure a double or triple unit will ever produce, because it leaves out the spacer, the sealant and the frame. The frame value, Uf, covers the frame section by itself, and in an aluminium product it is often the weakest figure on the sheet. The whole-unit value, Uw for a window or roof window, combines glass, edge and frame in proportion to their areas, using the actual dimensions of the product.

Only the whole-unit figure describes what you are buying. The other two are ingredients. Our answer on the difference between a whole-unit and a centre-pane U-value goes into the calculation, and the table below sets the three side by side.

A quote that gives a single U-value with no label deserves a question. Ask which of these it is, and for which size of unit.

A data sheet is a marketing document that happens to contain physics. The figure printed in large type is usually the most favourable one the manufacturer can honestly defend, and there are four common ways it drifts away from what ends up on your roof.

The first is the centre-pane habit already described. A triple glazed pane quoted at its centre value looks transformational, yet the same pane inside a wide aluminium frame with a standard spacer will have a whole-unit figure noticeably higher. The edge and frame are a larger share of a small unit than of a large one, so the gap between Ug and Uw is widest on the small rooflights that go into bathrooms and landings.

The second is the size of the test unit. Whole-unit values are calculated for a particular set of dimensions, and a different size gives a different answer. A sheet that quotes a figure for a large unit may be optimistic for the smaller one you ordered, because the frame takes up more of a small opening.

The third is the angle. Heat rises, and the air inside a sealed unit circulates differently when the glass is tilted. A pane that is vertical loses heat more slowly than the same pane set flat or at a shallow pitch, since the convection currents in the cavity are stronger when the glass faces the sky. A U-value measured for a vertical window cannot be carried across to a rooflight. Check that the figure is stated for the slope the unit will actually sit at, and be wary of any sheet that borrows a wall-window number.

The fourth is the gas. Argon fill improves the figure, and the improvement is real when the unit is new. Over the years a small amount of gas can leave the cavity through the seals. A sheet normally quotes the figure for a full fill, so it describes the day the unit is made. That is a fair basis for comparison, but it is worth knowing that the number is a starting point.

The number is the output. The inputs are choices you can see on a quote, and each one moves the result by a known amount in a known direction.

Start with the glass build. Single glazing passes heat freely. A basic sealed double unit from the era of clear glass and a plain aluminium spacer is far better, and a modern double unit with a coating and gas fill better again. A third pane adds another sealed gap and lowers the centre-pane figure further, at a cost in weight and in light. Our page on triple glazed rooflights and when the third pane earns its place weighs that trade.

Then the coating. A low-E coating is a microscopically thin metal oxide layer on one glass surface. It lets most visible light through and bounces long-wave heat back into the room. It does more for the U-value than any other single addition to ordinary double glazing.

The spacer is next. The metal strip that holds the two panes apart is a path for heat round the edge of the glass, and in an older unit it is a cold line that shows up as condensation along the bottom of the pane. A warm edge spacer is made of a less conductive material, which lifts the edge temperature and helps the whole-unit figure more than you would expect from such a thin component.

Last comes the frame, where much of the difference between products now hides. An aluminium section is light, strong and slim, and metal conducts heat very well, so an unbroken section undoes part of the glass's good work. A thermally broken frame splits the metal into inner and outer halves, joined by a low-conductivity barrier. Timber and PVC frames start from a better position, and each has its own limits of size.

Most sheets follow the same order, and once you know where the information sits you can check a product in a couple of minutes.

Begin with the heading figure and ask which U-value it is. Then look for the glass description, which should read like a recipe: the thickness of each pane, the width of the gap, the gas, the coating and its position. A line such as a laminated inner pane, a wide argon cavity and a warm edge spacer tells you there is a real build behind the number. A line that says only "energy saving glass" tells you nothing.

Next, find the test size. Many sheets give Uw for one reference dimension, and a few give a small table of values across a size range. If your rooflight is much smaller than the reference, expect the real figure to sit above the printed one. Then read the footnotes, where the conditions live: the slope, the gas fill, the method of calculation, and whether the glass is an option or a standard.

The calculation method is worth a glance. In the UK, window and rooflight values are normally worked out to the European standards for thermal performance of windows, doors and shutters, using measured or calculated data for glass, spacer and frame. You do not need to follow the method. You do need to see that it was followed, because a figure with no stated method is an opinion.

1. Identify the figure: Ug, Uf or Uw.
2. Confirm the size and the slope it applies to.
3. Read the glass build, spacer and frame type beside it.
4. Check that the sheet is for the unit on your quote, not a neighbouring product.
5. Note whether any good result depends on an option you have not chosen.

Building Regulations do not ask for the best U-value available. They set a limit, and a product has to be at or below it for the situation it is used in.

In England the thermal rules sit in Approved Document L, which gives maximum U-values for new elements and for replacement glazing. The limit depends on what you are doing: building an extension, replacing an old unit in an existing house, or creating a new dwelling. The numbers differ between windows, roof windows and rooflights, and they change when the Approved Document is revised, so any figure written down in a guide goes out of date. We keep the current picture on our page about Part L and what it asks of rooflights, and the specific question of whether a swap counts is answered at whether a replacement rooflight has to meet current U-value rules.

Two practical points are worth having in mind. Replacing an existing unit is usually a controlled fitting, so the new one is expected to meet the current requirement rather than copy the old one. And where the work is notifiable, we handle the Building Control notification, so the paperwork does not land on you. A written U-value on the quote, stated as a whole-unit figure, is what makes that conversation simple.

It helps to treat the regulation as a floor. A limiting value protects you from a poor product, but a room that will be used every day, such as a loft bedroom, sometimes justifies a better one. That is a decision about comfort and running costs, and it belongs on the quote as a choice with reasons.

Improving a U-value always reduces heat loss, and the useful question is by how much, in the context of a whole house.

Go back to the arithmetic. Heat loss is proportional to the U-value, so halving the figure halves the loss through that area. That sounds dramatic, but rooflights are a small fraction of a roof, and a roof is one of several surfaces that lose heat. A loft room with two small rooflights gains only a little from moving from a good unit to an excellent one, while a large lantern over an open-plan kitchen has far more to gain, simply because it has more glass.

The bigger gain is often comfort, and the U-value predicts it. A cold pane cools the air next to it, which drops and is felt as a draught on the shoulders and ankles. It also makes you feel colder because your body radiates heat to cold surfaces. A unit with a better whole-unit figure has a warmer inner surface, so the room is more even, and condensation is less likely to form on the glass and frame. The same effect is what you notice as a cold skylight on a winter morning.

Orientation and the season also complicate the picture. A south-facing rooflight can collect useful sun through the colder months, which offsets some of its loss, while a north-facing one only loses. The U-value is constant, but the benefit of the glazing in total is not, and you should weigh both together rather than treating a single number as the whole answer.

It is tempting to treat a low U-value as a sign of a generally better rooflight. It is not. The figure describes heat and nothing else, and two products with the same U-value can sound and look quite different.

Sound travels differently from heat. Quiet in a rooflight comes from the glazing build, in particular a laminated inner pane, the thickness of each layer, the gap, the seals and the upstand. A heavier laminated pane can sit in a unit with an excellent U-value and an excellent sound character, but one does not guarantee the other. Extra glass thickness may move the thermal figure a little, and the coating and gas may do nothing for the sound. Our page on whether acoustic laminated glass changes a rooflight's U-value explains how the two interact.

In Crawley, where aircraft, road and rail are part of what a loft bedroom hears, we write both requirements into one specification. The thermal figure goes on the quote as a whole-unit value, and the glazing build for quiet goes into the quiet spec that comes with every quote. The two are chosen together, so a warmer room does not end up noisier or the other way round.

Light works the same way. Extra panes and heavy coatings reduce the share of daylight that reaches the room, so a very low U-value can come at a price in brightness. The transmittance figure sits next to the U-value on the same sheet, and the two should be read as a pair.

When two quotes arrive with different U-values, the lower figure is not automatically the better product. It is better only if the figures mean the same thing.

Put the two sheets side by side and check four things. Are both figures whole-unit values? Were both calculated for a comparable size and the same slope? Does each list the glass build, spacer and frame? Is any good result tied to an upgrade that is priced separately? A low figure that turns out to be a centre-pane value for a larger unit than yours is not a saving, it is a mismatch.

If the answers differ, ask each supplier for the whole-unit value of the exact size on the quote. A good supplier can give it without fuss. If you are choosing between a few builds for a particular room, we will set the figures out in one table, with the glass, the frame and the regulatory limit beside them, as part of the written quote. You can start that conversation through the quote form, or over WhatsApp at +44 7880 284133.

These are the points that come up most once the basics are clear, and each has a short answer here with a longer one linked where it exists.

Is a U-value of 1.0 twice as good as 2.0?

For heat loss through that square metre, yes. Half the U-value means half the heat loss at the same temperature difference. Whether that is noticeable in the house depends on the area of glass and how much of the total loss it represents.

Can I rely on the centre-pane figure if the frame is slim?

A slim frame helps, since less of the unit is frame. It does not remove the edge and frame losses. The whole-unit figure already reflects the slimness of the frame, which is why it is the one to use.

Does the U-value get worse over time?

The physics of the glass stays the same, but the performance of a sealed unit can decline if the seals fail, because gas escapes and moisture gets into the cavity. Visible misting is the sign. A unit in that state is one to replace, and our page on misted and failed rooflight replacement explains the process.

Do I need the lowest U-value on the market?

Not necessarily. You need a value that meets the regulation for your project, suits the room and sits with the noise and light choices you have made. Beyond that, each extra step costs more and returns less, and the area of glass often matters more than the last decimal place.

The parent page on energy efficient rooflights in Crawley brings the glass, frame and fitting decisions together, and is the place to go next if you are planning a specification.