Should I replace all my rooflights if one has failed?

A sealed rooflight fails at the edge, where a ring of sealant holds gas between the panes and keeps moisture out. That seal is a small, separate component on every single unit, bonded during manufacture and never shared between rooflights, even when several were ordered together and fitted on the same day. One unit's seal giving way says nothing about the bond quality of the seal three metres along the same roof.

What does travel between units fitted at the same time is age and specification, not condition. If four rooflights went in together during one extension job, they are all the same number of years into the same kind of seal, holding the same gas fill, on roughly the same pitch. That shared starting point is a reasonable prompt to look at the others. It is not proof that they are about to fail, because two identical units on the same roof can still age at different speeds once the compass direction, the shade from a chimney or a neighbouring roof, and the standing water at a low corner all start to differ from one opening to the next.

The strongest case for treating a set of rooflights together is age plus exposure, not age alone. A row of units on a south-facing rear extension, all fitted the same week, will have gone through an identical run of hot afternoons and cold nights, so if one has clouded over, the neighbouring panes are worth checking on the same visit rather than waiting for each to announce itself in turn. A unit tucked in the shade of a chimney, or facing north on the same roof, has had an easier life even if it came from the identical order, and treating it as due for replacement purely because a sibling unit failed is guesswork rather than a finding.

Peterborough's new-town townships give a clear version of this. Bretton, Orton and Werrington have large numbers of homes from the same building phases, often with more than one roof window or rooflight added at a similar point in the house's life, sometimes as part of a single loft conversion. A failed unit in one of those homes is a fair reason to have the rest of that job's units looked at together, since they share both an age and, usually, a pitch and orientation. It is a weaker reason to assume every rooflight in the house needs replacing if one was added years apart from the others, on a different roof face entirely.

Where several rooflights sit in a single line of sight, such as a run of three flat units over a kitchen extension or two roof windows either side of a landing, a mismatch between an old frame and a new one is not just cosmetic. Frame profiles have generally slimmed down over the years, so a current unit fitted next to an older box frame can leave one pane reading noticeably smaller than its neighbour from the same seat at the table or the same spot on the landing, even though both openings are the same size in the roof.

This is worth thinking through before ordering, not after. If the units sit apart, in different rooms with no shared sightline, a mismatch in frame width is not something anyone in the house will actually see day to day, and replacing one now with the rest to follow later causes no problem. If they sit together in the same eyeline, matching frame width and bar position across the whole run is usually the better outcome, and that argues for planning the set as one job even if only one pane has clouded over so far.

• Rooflights in separate rooms: a mismatch is invisible day to day, so timing each replacement on its own condition works fine.
• Rooflights in one continuous run or one shared view: a newer, slimmer frame next to an older, bulkier one is visible from the seat that matters, so matching profiles across the set is worth planning for.
• A room with a single rooflight: the question does not apply, since there is nothing else in that view to match.

The practical case for doing more than one unit in a single visit is about the roof being open, not about the units themselves. Access, whether a scaffold tower, a cherry picker or roof ladders and crawl boards, has to go up before any unit can be reached, and that access serves every rooflight within its reach for as long as it stays in place. Reusing it for two or three units in one visit is more efficient than putting it up again for each pane in turn as it fails over the following years, and the household goes through the disruption of an open roof once rather than several times.

That case is strongest when the units are close together on the same roof face, so the same access reaches all of them, and weaker when they sit on opposite sides of the house or on entirely separate roofs, where a shared visit gains little. It is also independent of whether every unit actually needs a new glass unit yet. A survey can confirm a neighbouring pane still has years of seal life left, in which case leaving it in place and revisiting the access argument when it does eventually fail is a reasonable call.

Not every rooflight in a set needs the same scope of work even when several are replaced on the same visit. A unit whose frame is sound can usually take a new sealed unit into the existing frame, while a neighbouring unit with a corroded or distorted frame needs the whole rooflight replaced. Replacing the Glass Unit or the Whole Rooflight and When a Failed Frame Means Full Replacement both go through how that split is decided on a survey.

The two scopes can also carry different notification requirements, since a glass-only swap into an existing frame is treated differently to installing a new frame under current Building Regulations. Do I need Building Control to replace a misted glass unit? sets out when that notification applies. We handle the Building Control notification for the units that need it, whatever mix of scopes a set turns out to need.

The reliable way to answer this for a specific house is a survey of every unit in question, not a guess from the ground. A surveyor can read the condition of each seal, note the frame material and its state, check the pitch and orientation of each opening, and look for the manufacturer's own date reference on the spacer bar where one is visible. That gives an age and a condition for every rooflight being considered, rather than treating "one has failed" as a stand-in for the state of the rest.

From there, the choice usually comes down to the three factors already set out: whether the units share an age and exposure, whether they sit in one shared view, and whether the access argument makes a combined visit worthwhile. Every quote, whether for one unit or a set, comes with a sightline sketch showing the view from where you actually sit or lie in each room, so you can see how a part-replacement or a full set will look before any roof is opened. What Affects the Cost of a Rooflight sets out the factors that shape the cost either way, and our parent guide to Misted and Failed Rooflight Replacement in Peterborough covers the full range of what gets replaced across VELUX roof windows, flat rooflights, lanterns and conservation units.

Does one misted unit mean the others will fail soon too?

Not necessarily. It means the others are the same age, which is worth checking, but exposure differs from one opening to the next even on the same roof, so a neighbouring unit facing away from the sun or sitting in shade can have years of seal life left in it.

Can I replace just one rooflight and leave the rest until they fail?

Yes, where the units are not in a shared line of sight and access is not a shared cost. Where several sit in one continuous run over the same room, matching frame width across the set is usually worth planning for even if only one has clouded over so far.

Will a new rooflight look different from the old ones next to it?

It can. Frame profiles fitted years apart are rarely identical, so a current slim frame next to an older box frame can leave one pane looking smaller than its neighbour from the same seat, even where the two openings in the roof are the same size.

Is there a cost advantage to replacing several rooflights in one visit?

The main saving is around access rather than the units themselves. Putting up a scaffold tower or roof ladders once for several rooflights on the same roof face is more efficient than repeating that access for each pane as it fails in turn, though a survey may still confirm that some units have plenty of seal life left.