Leadwork Around a Chimney: Soakers, Aprons and the Back Gutter

The Stain Is a Metre Downhill and the Stack Is Dry

A ring on a bedroom ceiling roughly a metre below the chimney breast, brown at the edge and dry in the middle, that darkens after driven rain from one particular direction and never after a still wet day. Nine times out of ten somebody has already been up, found nothing obviously wrong, run a bead of sealant round the visible edge of the lead and gone away. It stops for a season. Then a gale comes in off the same quarter and the ring gets a second edge.

Water arriving in a roof space at a stack rarely lands where it leaves. It runs down the inside face of the masonry, along the top of a ceiling joist, and drops at the first thing that interrupts it, which might be a cable clip two rafter bays away. So the stain is evidence that something at the stack has failed, and evidence of nothing else at all. Where it dropped tells you where the joist was, not where the hole is.

The stack is also the one place on a shedding roof where four different problems all present identically. Three of its four faces above the roof line do three different jobs, and the fourth face is a dam standing across the fall of the roof. This guide works from the symptom outward: rule out what is not the flashing, then get the four pieces of leadwork right, in the right code, at the right girth, and order them as pieces rather than as metres.

Rule Out the Other Three Before Anything Comes Off

Leadwork gets blamed for water that never touched it. A stack standing above a roof is exposed on four faces to weather nothing else on the building sees, and it is usually the oldest masonry on the house. Saturated brickwork passes water straight through to the inside face; open perpends do it faster, and the stain that follows behaves exactly like a failed apron.

Look at the flaunching first, because it is the cheapest thing on the roof and the most reliably cracked. A sand-and-cement fillet holding the pots takes the full thermal cycle of a dark masonry mass in the sun, and once it has parted from the pot the crack runs down into the head of the stack and fills it every time it rains. From below it looks like a hairline. From a ladder it is a funnel.

The second suspect is the flue itself. A stack whose fireplace was blocked up and whose pot was capped without a vent breathes nothing, and the moisture in it condenses out on the coldest surface it can find, which is the masonry above the ceiling line. Hygroscopic salts from decades of combustion pull it further. Salt staining is fluffy and white at the edges and comes back when it is painted over, and no amount of new lead will touch it.

The third is the covering next to the stack rather than the stack itself. A slipped slate, a cracked tile or a nail-sick fixing within a course or two of the leadwork puts water into the same place, and someone standing on the ridge looking down at a tidy line of lead will not see it. Lift the courses either side of the soakers before condemning them.

Only when those three are eliminated is it worth pulling metal off. Test with water, not with opinion, and test in the order the roof drains — lowest first, one zone at a time, with somebody in the loft with a torch and a phone. Hosing the whole stack at once tells you that the stack leaks, which you knew.

1. Get someone into the roof space with a torch, directly under the stack, and let their eyes adjust before anything is turned on.
2. Wet the covering a metre below the stack first, for ten minutes, and wait another ten. Nothing yet means the covering below is sound.
3. Move to the front apron and its two corners next, and give it the same ten minutes on, ten minutes waiting.
4. Then each raking side in turn, working up the slope, so a wet result names one detail rather than four.
5. Then the back gutter, which is where the water has been standing all along in most cases, and which usually shows within a minute or two.
6. Only then wet the masonry above the roof line, and give it half an hour, because saturation is slow and it is the answer that gets missed.
7. Photograph what ran, from inside, before anyone starts stripping. It is the only record of the diagnosis that survives the repair.

Soakers With a Cover Over Them, or One Piece Dressed to the Tile

The choice on the raking sides is settled by the covering, not by preference, and getting it wrong is the commonest reason a stack that was reflashed last year is wet again this year. A soaker is a small rectangle of lead with an upstand, laid with a course and hidden by the next one, which is only possible where the units are flat and thin enough for something to sit between them without lifting the course above. A stepped flashing dressed straight onto slates with no soakers underneath looks identical from the ground and relies entirely on the dressed edge staying down, which it does until the first gale.

Where the covering has a profile, the argument reverses. A pantile or a Roman tile has a pan and a crown, and there is no flat soaker on earth that follows both. Metal dressed over the crowns alone leaves the pan wide open beneath it, so the cover flashing itself has to be cut and dressed down into each pan, which is slow, skilled work and takes noticeably more girth than the flat detail. Several tile makers supply their own soakers or a proprietary side abutment system for exactly this reason, and where the roof is under a manufacturer's warranty that system is usually the only detail the warranty recognises.

The Code on the End of the Roll

Rolled lead sheet is specified by code, and the code is a thickness with a weight and a colour attached to it. BS EN 12588 is the material standard; the colour marking on the roll end is how it is identified on site, and it is worth checking rather than trusting the delivery note, because Code 4 and Code 5 look identical in a stack of offcuts and behave very differently on a back gutter. The weights below are simply the thickness multiplied by the density of lead, 11.34 g/cm3, so any of them can be checked with a micrometer if a roll turns up unmarked.

Buying up a code where the piece is exposed, wide or walked on is cheap insurance; buying down is how a back gutter ends up creased and split at the fillet within a few winters. The practical constraint on the other side is weight and handling: a 6 m roll of Code 5 cut 600 mm wide is 3.6 square metres and around 91 kg, which is a mechanical lift or two people on a scaffold, never one person on a ladder. Cut it to piece length at ground level.

The second constraint is thermal movement, and it is the reason the manuals cap piece length rather than trusting the fixings. Lead expands roughly 29 parts per million per degree, so a 3 m length going through a 40-degree swing between a frosty dawn and a dark slate roof in July moves about 3.5 mm, and it will move at whichever end is least well held. That is what tears a nail hole into a slot and what creases a long apron into a ripple. The Lead Sheet Training Academy manual gives maximum piece lengths for each application — commonly 1.5 m for flashings — along with lap and fixing centres; take the figures from the current edition rather than from memory, because they have been revised.

Three material notes that cost money when they are missed. Lead in direct contact with fresh cement mortar, and lead in contact with oak or western red cedar, needs protecting with a bituminous coating, because both attack it. New lead washed by rain produces a white lead carbonate that runs down and stains everything below it, permanently on light-coloured tiles and render, which is what patination oil applied on the day of installation prevents. And lead is a health hazard to the people cutting and dressing it: OSHA 29 CFR 1926.62 governs lead in construction in the United States, and the Control of Lead at Work Regulations 2002 in the United Kingdom, which between them mean washing before eating, no dry grinding, and controls on any hot work.

How Far the Apron Runs Out, and How Far It Goes In

The apron is where most of the argument on site happens, because the figure that matters is invisible once the job is done. It runs out over the covering below far enough that wind-driven water crossing the lap still has somewhere to go before it reaches the top edge of the metal, and the conventional minimum in the lead manuals is 150 mm measured up the slope from the face of the stack. Shorten it to tuck neatly behind a course line and the detail is now relying on the covering below it, which was never designed to be watertight on its own.

The other end of the same piece is the upstand and the chase. The upstand runs up the face of the stack, the top of it is turned 25 mm into a joint raked out to receive it, and it is held there by lead wedges at not more than about 450 mm centres before the joint is pointed. Raking out is a bed joint for a straight run and a stepped set of bed joints for a raking side; cutting a diagonal chase across the brickwork is quicker and is why so many stacks have a crack running corner to corner ten years later. Both ends of an apron are turned 50 mm around the corners of the stack, which is the detail that stops water walking round the end of it and is the first thing to check on a stack that leaks only in wind.

Girth follows from that, and girth is what actually gets ordered. Twenty-five millimetres into the joint, 75 mm up the face and 150 mm out over the covering is 250 mm before anything is dressed, and dressing over anything other than a flat slate consumes more — a bold profile can take another 50 to 75 mm to follow the pan. Order the strip cut to a girth that has the dressing allowance in it. Lead is supplied cut to width from the roll at no waste, and it is the one material on the job where ordering the exact strip width is easier than trimming it on a scaffold.

The Face That Stands Across the Fall

Everything above the stack arrives at its upslope face, and unlike a wall abutment it has nowhere to go but sideways. That is why the back detail is a gutter rather than a flashing: a sole with a fall across it, a tilting fillet so the lead is not asked to turn a hard right angle, an upstand against the stack, and a turn-up at the back carried under the covering above. Water leaves it at both ends, over the top of the side soakers, which is why the back gutter goes on after them and laps over them — the whole stack is worked from the bottom of the slope upward, apron first, soakers up the two sides, gutter last.

Above a certain width that arrangement stops coping and the roof needs a framed saddle instead — a small ridged structure behind the stack that splits the flow into two and sends it round. In the United States the residential code makes this explicit for masonry chimneys, requiring a cricket where the stack is more than 30 in (762 mm) wide measured parallel to the ridge, subject as always to the edition adopted locally. British practice reaches the same place from the other direction, treating a lead back gutter as the standard detail at any width and adding a timber saddle where the width and the snow load make a flat-soled gutter a bad bet. Either way, the geometry is the same: half the up-slope-facing width, times the pitch of the roof, gives the ridge height the saddle has to reach.

This is the face that fills with debris, and the failure is almost never the metal. A back gutter on a roof with a tree over it collects leaf litter, holds it against the stack, and then holds water in it through the winter. Ice builds behind the dam, backs up over the 150 mm turn-up that was perfectly adequate for running water, and the stain appears inside. Where a stack sits on a north slope under trees, that gutter needs to be on somebody's list twice a year, and it is worth saying so in writing at handover rather than after the second call-out.

What Is Left When the Scaffold Goes

Oil the new lead on the day it goes on. It takes minutes, and the alternative is a white carbonate wash down whatever sits under the stack, permanent on light render or a pale tile. While the scaffold is still up, run water down the roof above the stack — not at the flashings, at the covering upslope of them — and watch the back gutter discharge at both ends. That is the test that proves the detail rather than the sealant.

Then write down what nobody will be able to see again: the code used on each piece, the girths, the piece lengths and where the laps fall. It is folded into the joints and buried under the courses, and the next person at this stack will otherwise be guessing at it from a ladder. A photograph of each face before the covering closed over it, with a tape in the frame against the apron cover and the gutter sole, is worth more to that person than any invoice.