Guide · Skylight & Rooflight Installation & Replacement Chelmsford
Telling a failed unit from a failed installation
Water has appeared somewhere near a rooflight, and underneath every other question sits this one: has the glass failed, or has the roof around it failed? A glazing unit can be finished while the detail holding it is in good order. The detail can have opened up while the glass is still doing everything asked of it. From a sofa on a wet Tuesday in February the two look identical.
Telling them apart decides the work. A failed unit is an order, a day and a new frame set into an opening that is already sound. A failed installation is carpentry, covering and flashing, and dropping a new unit into a detail that has already let go simply buys a year. This page sets out the evidence a homeowner can gather without going onto the roof, and what each piece of it points at. It sits under skylight replacement, where the wider process is described.
Two faults that look identical from the room below
A sealed unit fails inside itself. The cavity between the panes takes on moisture, and no fitter caused it and no detailing will cure it. It is a component reaching the end of a working life, in the same way a boiler does.
An installation fails at the perimeter. Water crosses a lap, runs behind an upstand, or condenses in a cold construction because the insulation stops short of the frame. The glass is innocent. What both have in common is that the first thing anybody notices is dampness somewhere near a rooflight, which is why the diagnosis has to start with position rather than with the product.
Where the moisture is sitting, and what that alone tells you
Before any theory, establish exactly which surface is wet. It narrows the field faster than anything else and it can be done with a torch and a dry cloth.
| Where the water is | Most likely cause | What it points to |
|---|---|---|
| Between the two panes of glass | Sealed unit at the end of its life | New unit, opening untouched |
| On the room face of the glass | Surface condensation from room air | Ventilation and glazing performance |
| On the frame face, running down | Condensation on a cold frame, or a gasket | Frame generation, or the seal |
| On the plaster reveal only | Cold perimeter, or water crossing the detail | Insulation continuity, or flashing |
| On the ceiling away from the opening | Water travelling before it drops | Entry point somewhere uphill |
Haze inside the cavity, which is never an installation fault
Moisture visible between the panes is the one unambiguous reading on the whole list. The cavity is a closed system. Nothing that happens on the roof puts water into it, and nothing done to the roof afterwards takes water out.
It shows first as a faint bloom in one corner on a cold morning, then as a milky film across the lower third, then as tide marks etched into the inner face of the outer pane where mineral deposits have dried in. That progression is covered in detail on misted sealed units and why they go. For diagnosis, the only thing that matters is that a hazed cavity means a new unit and says nothing whatever about the quality of the original fitting.
Water on the room face of the glass, which is usually the air in the room
Condensation on the inside surface is warm moist air meeting the coldest thing in the room. In a kitchen with a hob running or a bathroom after a shower, the coldest thing in the room is very often a piece of glass lying in a roof plane radiating to a clear winter sky.
It is physics rather than a fault, and the tell is that it appears on cold still nights with no rain, clears through the morning, and follows the household routine rather than the weather forecast. A better unit raises the inner pane temperature and reduces how often it happens. Moving the moisture out of the room is what actually settles it.
Marks on the reveal, and the two routes water takes to get there
A wet mark on the plaster lining is the single most misread symptom in the trade, because two entirely different mechanisms produce almost the same stain.
The first is water crossing the perimeter from outside, running down the inner face of the kerb or the trimmer and appearing where the plaster meets the frame. The second is interstitial condensation, where the insulation stops short of the opening, the timber around the reveal runs cold, and moisture from inside the house forms within the construction. The first is worse after rain. The second is worse after a cold week with the heating on and the windows shut. Same stain, different fix.
Note the date and the weather every time a mark appears, because that record diagnoses more reliably than any single inspection.
Timing against the weather, the most useful evidence available
Rain is a test rig that runs itself, and most people already have weeks of data without realising they collected it.
Dampness that appears during rain or within two hours of it starting is water getting in. Dampness that appears on clear cold nights with no rain at all is condensation. Dampness that shows up at the end of a cold spell, as the roof warms and the void dries, is moisture that was already inside the construction. Those three patterns cover the great majority of cases, and they can be separated by nothing more than a note on a calendar.
Wind direction, and the fault that shows from only one quarter
Steady rain falling vertically rarely finds anything. Rain driven sideways at thirty miles an hour finds everything, which is why the prevailing south westerly across the open ground between Chelmsford, Great Dunmow and Braintree does the diagnostic work.
If the mark only ever appears when the wind is from one direction, the weakness is on that face. Wind driven rain climbs a vertical surface, so it will push up an upstand that would be perfectly adequate in still conditions, and it will get under a flashing lap that sheds gravity flow without complaint. A fault with a wind direction attached to it is a perimeter fault, every time.
Sun after rain, and the mark that arrives late
There is a pattern that confuses almost everybody. It rains overnight, nothing happens, the sun comes out at eleven the next morning and a mark appears on the reveal by lunchtime.
That is water already held in the insulation or the deck being driven inwards by solar gain. Warm the outer layers of a wet flat roof and the vapour pressure pushes moisture towards the cool side, which is the room. It means the water got in some time ago and the construction is holding it. On a warm deck flat roof it also means the insulation under the covering is likely to be saturated over an area rather than at a point, which changes the scope of the work considerably.
A frame that has moved, and a sash that no longer pulls down
Rooflight frames are mechanically fixed through a flange into the kerb or the trimmers. Where the timber at the base of an upstand has taken water and gone soft, those fixings lose their hold, and the frame lifts a fraction under wind uplift and drops back.
Signs are specific. A hairline of daylight at one corner of the internal lining. A dull knock in a gale. A gasket that has taken a permanent set, so a strip of paper closed in the seal pulls out with no resistance where it should grip. On an opening unit, worn hinges let the sash sit proud on one side, and the pane then only seals across half its length. All of that is installation and hardware, not glass.
Weep holes and drainage channels that have closed up
Rooflight frames are designed to get wet inside. Water that gets past the outer gasket collects in a channel within the frame and drains out through weep slots along the bottom edge, and this is normal engineering rather than a shortcoming.
Block those slots and the channel fills. Once it fills above the internal upstand of the frame section, water tips over the edge and runs into the room, at which point it presents exactly like a failed seal. Leaves, moss, pollen from the limes and pine needles all do it, and on a flat rooflight sitting in a corner where debris collects it can happen within two autumns. Clearing the slots is the first thing to check on a unit that only ever weeps in heavy rain.
Flashing laps, and the side that gives it away
On a pitched slope, water is shed by overlap rather than by seal. Each component has to sit over the one below it in the direction of flow, and a single reversed lap is enough.
Three failures turn up repeatedly across Essex slopes. A head apron that is too short for the tile depth, so wind driven rain gets behind it on a deep concrete interlocking profile. Side soakers dressed under the tile where they should sit over, common where a kit was substituted for the correct one. And a cement fillet used at a junction where a dressed lead detail belongs, which cracks within a few winters as the timber moves. Each of those produces a mark on one specific side of the reveal.
Water arriving from further up the roof
A rooflight is often a witness rather than a culprit. Water entering at a ridge, a chimney, a valley or a slipped tile runs along the underlay or across the deck and drops at the first obstruction it meets, and the first obstruction is very frequently the timber trimmed around the opening.
The tell is a mark that is not centred on the rooflight. If the stain sits at one end of the reveal and continues along the ceiling line away from the glass, the entry point is uphill. On a flat roof, water can travel several metres between layers before it appears. Cutting the correct diagnosis short here is expensive, because a new unit into a sound opening will not change a thing.
An upstand that is too low for the roof it sits on
The height from the finished roof surface to the top of the kerb is the number that decides whether standing water, drifted snow or wind driven rain can reach the frame. The industry figure is one hundred and fifty millimetres, and it is not arbitrary.
Older flat rooflights across Essex sit on shallower kerbs, sometimes barely fifty millimetres, because that was ordinary practice when the extension went up. On a roof that drains cleanly it can survive for decades. On a roof that ponds, or where a parapet lets snow drift against the unit, it is the fault. Whether the existing kerb stays or is rebuilt is set out on reusing an existing kerb and when not to.
Testing in order, dry checks before any water
Diagnosis has a sequence, and running it out of order produces a false result that costs money.
- Dry inspection first. Torch, dry cloth, and a note of which surfaces are wet and which are merely cold.
- A moisture meter at the reveal, the ceiling and, where accessible, the deck, to distinguish a surface film from a saturated construction.
- Controlled wetting last, starting at the lowest point of the detail and working upwards in stages, ten minutes at each, with somebody inside watching. Wetting the whole area at once tells you only that it leaks, which you already knew.
Which findings mean a new unit, which mean new detailing, and which mean both
Sorting the evidence gives three outcomes and they carry very different programmes.
Moisture in the cavity, a chalked or crazed frame, or an obsolete unit that will not meet an escape opening means a new unit, and the opening is usually left alone. A wind driven mark on one face, a blocked drainage channel, a short apron or a low kerb means the detail is rebuilt, which is carpentry, covering and flashing around a frame that may itself be reusable. Soft timber at the base of the kerb, or wet insulation over an area, means both, because the new unit needs something sound to fix into and there is no sensible order other than doing the two together.
What to record before anyone visits, and what a survey adds
Four things gathered in advance shorten the whole process. A photograph of the wet area with something in frame for scale. The date and weather each time it has appeared, going back as far as memory allows. A photograph of the unit from inside showing the frame and reveal together. And a photograph of the data plate on the sash or frame edge if there is one, which gives manufacturer, model and size in a single shot.
The survey then adds what photographs cannot: the state of the timber at the base of the kerb tested with a bradawl rather than judged by eye, the fall of the roof around the opening, whether the covering has life left in it, and how the insulation meets the frame. From there the answer is either a unit, a detail, or both, and the fixed price follows the diagnosis rather than the guess.
Keep reading
More on skylight & rooflight installation & replacement chelmsford
Sizing a rooflight for the room
Read the guide →Roof window or rooflight, and the difference
Read the guide →Flat roof against pitched roof rooflights
Read the guide →What happens on installation day
Read the guide →Cleaning and maintaining a rooflight
Read the guide →All about skylight & rooflight installation & replacement chelmsford →
Get a quote
Tell us about your roof
We start with the building, the covering and the slope. Then we tell you what will suit it, and what it will cost, as a fixed written number.
- Surveyed before it is priced
- 10-year workmanship guarantee
- Building Control notification handled
- New installations and replacements