Sizing a Rooflight for the Room Below

Where oversizing comes from

The instinct is understandable. A dark room feels like a shortage of daylight, and a shortage suggests the remedy is more of what is missing. Brochure photography reinforces it, because a photograph of a bright room does not carry any of the consequences of the glass that made it bright.

Three of those consequences show up later. The first is heat. Glass in a roof faces the sky rather than the horizon, so it collects sun for a longer part of the day and at a steeper angle than a wall window does, and a room can go from dark to genuinely unpleasant on a summer afternoon. The second is glare and contrast. A large aperture in one part of a ceiling creates a pool of brightness and leaves the rest of the room looking darker by comparison, which is the opposite of the even light people imagined. The third is heat lost the other way, since glass performs less well than an insulated roof build-up.

None of that argues for a small unit. It argues that size is a balance rather than a maximum, and that the right answer is the one where daylight, comfort and the structure all land in an acceptable place at once. Our Didcot page goes further into overheating in small rooms, and our Summertown page into what happens at the large end.

Three constraints set the size

By the time a survey is finished, the range of sensible sizes has usually narrowed a great deal, and it has narrowed for three separate reasons that have nothing to do with taste.

Each of these can be the binding one. On a trussed roof the structure will usually decide before anything else does. On a small room with a south-facing slope the comfort question decides. On a visible slope of an older building the elevation decides, sometimes conclusively.

What matters is that all three are checked rather than the most convenient one. A unit sized purely on daylight can be structurally awkward and visually wrong. A unit sized purely on what drops between rafters can be too much glass for the room it serves. The size that works is the one surviving all three tests, and finding it is a short exercise once somebody has been on the roof and in the room.

What the structure will allow

The roof structure is the constraint people meet first, because it is the one with an immediate cost attached.

A pitched roof is a series of rafters at regular spacing. A unit narrow enough to drop into a gap between two of them, with the frame carried on trimmers at head and foot, is a modest piece of work and the covering is the main event. A unit wider than that gap means cutting a rafter, and once a rafter is cut the load it was carrying has to go somewhere. That is done with trimming members transferring the load to the rafters either side, and it is a designed alteration rather than a detail on a drawing.

Trussed roofs are stricter. A trussed rafter is an engineered component working as a whole, and its members are not there to be cut and reconnected on site. In practice a trussed roof either takes a unit fitting between trusses or it takes a designed alteration prepared by a structural engineer.

On a flat roof the equivalent question is the joists, their span and their direction. Cutting across joists on a wide opening changes how the deck behaves, and where a kerb sits relative to falls and outlets is part of the same question.

The useful point is that the step from a unit fitting the existing structure to one that does not is not gradual. It is a jump, and knowing where that jump sits on your particular roof is often the single most valuable output of a survey.

What the room actually asks

The second constraint is the room, and it is asked about far less often than it should be.

Start with what the room is for. A landing or a hall needs enough daylight to feel navigable and pleasant, and very little more, because nobody sits there. A kitchen wants working light across a surface. A bedroom wants daylight in the morning and a way of stopping it in summer. A study with a screen in it has a glare problem that a generous aperture makes worse rather than better. These briefs are genuinely different, and they lead to different sizes on the same roof.

Then ask what daylight the room already receives. A rooflight added to a room with a decent existing window is doing a supplementary job, and it can be modest. A rooflight in a room with no other opening at all is doing the whole job and reasonably wants more.

Then ask where the sun goes. A slope facing towards the midday and afternoon sun collects considerably more heat than one facing away from it, and the same aperture will behave very differently on those two slopes. On a warm orientation, the answer is often a smaller area of better glazing with solar control, rather than a larger area with a blind added afterwards to undo the decision.

Finally, ask about the volume of the room. A small room with a low ceiling has less air to absorb heat and less surface to spread light across, so both the benefit and the penalty of a given aperture arrive faster than in a large one.

Proportion against the roof itself

The third constraint is what the roof can carry as an object, and it is judged against the roof rather than against the room.

A roof slope has a rhythm to it: the line of the eaves, the ridge, the courses of slate or tile, the position of any chimney or dormer already there. A unit sitting comfortably in that rhythm reads as part of the roof. A unit filling most of the available slope reads as a roof that has been replaced with glass, which may be exactly what was wanted on a modern rear extension and is rarely what was wanted on the main roof of an older house.

This is why a unit sized entirely from inside can look wrong from outside. The room has no idea what the elevation looks like. Somebody has to stand back from the building, on the pavement or across the garden, and judge the aperture against the slope it is going into.

On a constrained or designated roof, this stops being a judgement and becomes something settled with somebody else. Where a building is listed, sits within a conservation area, or is subject to a local direction affecting what can be altered, the acceptable size and position on a visible slope may be determined by the authority rather than by the client. Where that applies, establish what is acceptable first and design the room’s daylight around it, because the reverse order produces drawings that have to be redone. We install rooflights rather than determine consents, and we say so plainly rather than guessing on anybody’s behalf.

A pair against a single

One of the most useful moves available is to stop treating the aperture as a single object.

Where the structure allows it, a pair of smaller units set apart frequently serves a room better than one large unit of similar overall area. The reason is how light behaves. A single aperture makes one bright area with fall-off around it, and the eye adapts to that brightness, so the corners read as gloomy. Two apertures separated along the roof throw light into two parts of the room, the overlap fills the middle, and the result is more even. Even light is what people actually mean when they say a room feels bright.

There is a structural argument too. Two units narrow enough to sit between existing rafters may avoid cutting anything at all, where one wide unit covering the same span would need a designed trimmed opening. The saving there can be substantial, and it is one of the few places in this job where the better result is also the cheaper one.

The heat argument runs the same way. Two smaller apertures on different parts of a slope, or on different slopes, spread solar gain through the day rather than concentrating it, and they give more options for putting glass where the sun is weakest.

It is not a universal answer. A single unit reads more calmly in a small room, and on a flat roof a single larger piece of glass or a lantern can be the right architectural move. But the pair is worth putting on the table every time, because it is the option people do not think to ask for.

Position matters as much as area

Where a unit sits changes the room at least as much as how big it is, and it is the cheaper variable to adjust.

Light falls where the aperture points. On a pitched roof the unit is angled, so daylight arrives at a slant and lands on the wall and floor opposite the slope rather than directly beneath the glass. Move the unit along or up the slope and the bright area moves with it. That means a unit can be placed to light the part of the room that is actually dark, or the surface somebody works at, rather than the middle of the ceiling.

Height on the slope matters as well. A unit set higher up throws light further into the room, which is what you want where a room extends back from the slope. A unit set lower gives a view out for somebody standing, and puts more light near the outside wall.

Position also decides how the unit is used. A rooflight over a stairwell will never be reached by hand. One over a worktop is at a sensible height for a handle. That feeds straight back into whether the unit needs to be fixed, manual or powered.

Position also interacts with everything already discussed. Where the rafters run, where a purlin sits, and where the unit will be visible from all constrain the choice. A well-placed modest unit routinely outperforms a badly placed large one, which is not what anybody expects to hear.

Size follows the survey

The practical conclusion is a sequence rather than a number.

A size chosen from a catalogue before anybody has been on the roof is a guess with a code attached to it. It may be structurally awkward, it may be wrong for the orientation, and it may be a size the elevation cannot carry. Worse, it anchors the conversation, because once a client has a code in their head every alternative feels like a downgrade rather than a better answer.

The order that works runs the other way. Establish what the room is for and what daylight it already gets. Look at the structure above the ceiling and find where the openings can go without redesigning the roof. Look at the slope from outside and see what it can carry. Establish, on a designated or constrained building, whether anybody else has a say. Only then does a size, a count of units and a position come out of it, and by that point it is a conclusion rather than a preference.

We work that way because the alternative produces changes late, and changes late in a roof job are expensive. On our installations the survey settles the specification, the specification produces the quotation, and the Building Control notification is handled as part of the work. If you take one thing from this page, take the habit of asking what the roof and the room will allow before asking what is available.