Replacing a Failed Lintel Over a Window

The Wall Has Five Ways of Cracking There and They Do Not All End in a New Lintel

A crack at an opening corner is a stress concentration finding the weakest line available to it, and the opening corner is the weakest line in almost any elevation. That means the corner tells you where the wall gave, not why it was pushed. Before pricing a lintel, rule out the four other things that make the same mark.

A corroding steel lintel is the easiest to confirm and the most commonly missed, because it usually opens a horizontal joint rather than a diagonal one. Mild steel expands to several times its own thickness as it rusts, so a galvanised pressed-steel lintel that has lost its coating jacks the course immediately above it upward. Look for a bed joint over the head that has opened along its whole length while the joints either side of it are tight, brown staining bleeding out of the perpends, and a soffit that has begun to laminate. Where the lintel is a plain angle behind a soldier course, the soldiers tip forward before anything cracks. BS EN 845-2, the product standard for prefabricated masonry lintels, ties a declared load to a declared durability class for exactly this reason: the capacity in the table belongs to a section that still has its section.

Progressive foundation movement is the one that must be excluded rather than assumed away. Its signature is that it is not local: the same rotation shows in more than one opening, cracks are wider at the top where the wall is rotating outward at high level, doors bind on one leaf, and it moves with the seasons. BRE Digest 251, Assessment of Damage in Low-Rise Buildings, with Particular Reference to Progressive Foundation Movement, is the document that sorts crack damage into categories by what the repair actually involves, and the boundary worth knowing is the one between damage that can be dealt with by repointing and damage that needs masonry cut out and pieced in. The Institution of Structural Engineers' Subsidence of Low-Rise Buildings is where the diagnosis logic sits. Neither is a substitute for a monitoring record: tell-tales or fixed studs, read on dates, across enough of a year to catch a clay soil's seasonal cycle.

The remaining two are movement the wall was never given room for. Clay brickwork expands irreversibly after firing and keeps doing so for years, so an unjointed run of new brick pushes at its returns and finds an opening corner to relieve itself at. Concrete blockwork does the opposite and shrinks, which is why the spacing convention for its control joints exists at all. On a blockwork or rendered-block elevation, count the joints the run should have had before you condemn anything structural, because a crack from an opening corner on an unbroken forty-metre panel is shrinkage doing what shrinkage does.

What Sits Over the Head, and How Much of It the Lintel Owns

Open up a window head on a post-war cavity wall and you may find one combined lintel carrying both leaves, or two separate members that have nothing to do with each other, or an outer-leaf lintel over a timber inner head. Which of those it is decides whether this is one operation or two, because a cavity wall with separate lintels has to be propped leaf by leaf. The outer leaf carries its own weight and nothing else; the inner leaf carries the floor. Treating them as one member because they share an opening is how a prop ends up under the wrong leaf.

The load the permanent lintel takes is not the whole panel of masonry above it. Sound brickwork bridges over an opening, and manufacturers' safe working load tables are written around that: a load-distribution case in which the lintel carries a triangle of masonry over the span plus whatever floor or roof load lands inside that triangle, provided there is enough height of masonry above the opening for the triangle to form and no second opening cutting into it. BS 5977-1, Lintels: Method for Assessment of Load, is where that convention in the UK tables descends from; it is no longer a current standard, so read the load cases printed in the manufacturer's own table rather than a remembered version of it. BS EN 1996-1-1 (Eurocode 6) and TMS 402/602, Building Code Requirements and Specification for Masonry Structures, both treat arching as something the geometry has to earn, not a default.

That is the permanent case. It does not apply to your props, and section by section this page keeps the two apart. Where the wall above is shallow, where another opening sits inside the triangle, or where the mortar is soft enough to rub out with a thumb, the arch is not there to be relied on and the honest assumption is the full rectangle up to whatever actually stops the load coming down: a slab spanning elsewhere, a ring beam, or the top of the wall. To turn that area into a weight you need the masonry's weight per unit of wall face. Block manufacturers publish it directly. For clay brickwork, multiply the leaf thickness by the bulk density: a 215 mm solid wall at around 2000 kg/m3 comes out near 430 kg/m2, and BS EN 1991-1-1 (Eurocode 1), Actions on Structures: Densities, Self-Weight and Imposed Loads for Buildings, is where characteristic densities for masonry are tabulated.

A Safe Working Load Is Not a Moment, and the Span Is Squared

Two different documents answer the capacity question and they answer it in different currencies. A prefabricated steel or concrete lintel arrives with a declared safe working load in kilonewtons per metre or in total kilonewtons, published against a span and a stated load-distribution case, and that figure is the end of the argument for that product used that way. Catnic and IG Lintels both print the load cases alongside the numbers, and using a table figure under a different distribution — a point load from a beam landing mid-span, for instance — voids it silently. A reinforced concrete or reinforced masonry lintel, by contrast, gives you a moment capacity from its section and its reinforcement, and a moment has to be converted before it can be compared with anything.

The conversion is the simply-supported case: maximum moment under a uniform load is wL squared over eight, so the allowable uniform load is eight times the allowable moment divided by the span squared. The square is the part people underestimate on site. A member that was comfortable over a 900 mm cottage window has 36 per cent of that capacity over a 1.5 m one, which is why a bay that was originally three narrow lights and has since been opened out into two wide ones so often turns out to have a head that never had any reserve. It is also why the replacement rarely wants to be like-for-like without checking: the original may have been under-designed from the day it went in, and it has simply taken forty years for the arch above it to give up.

Bending is not the only check and this page does not pretend otherwise. Shear at the ends, deflection under the same load, and the bearing stress where the reaction crosses into the jamb are three separate limits, and on short heavily-loaded spans it is shear or deflection that governs rather than bending. Deflection is the one that matters visually: a lintel that satisfies strength and sags past the limit the finish can take will crack the plaster above it again within a year, and the owner will reasonably conclude the repair failed.

Getting the Weight Off the Old One

Propping a wall over an opening is temporary works, with everything that word carries: a designer, a brief, a check, and a record. BS 5975, Code of Practice for Temporary Works Procedures and the Permissible Stress Design of Falsework, sets out the procedure, and the Construction (Design and Management) Regulations 2015 put the duty to plan the temporary condition on the people organising the job rather than on the person holding the prop. On a single window head in a two-storey house that reads as heavy-handed until the first time a strongback goes in under a bed joint that turns out to be a lime joint with no bond across it.

There are two schemes and they are not interchangeable. Needles are timbers or steel sections passed right through the wall above the opening and propped both sides, and they hold the wall through its full thickness. Masonry support heads clamped to props — Acrow's Strongboy and its equivalents — sit in a raked-out bed joint from one side only, which is quicker, disturbs less, and carries far less: the attachment's published safe working load is a fraction of the prop's own, and it is the attachment that governs, not the prop. Their literature is the only place that number lives. On a cavity wall each leaf needs its own support unless somebody has demonstrated the ties can carry the outer leaf on the inner, which on a wall old enough to have a failed lintel they generally cannot.

1. Establish what is over the head before anything is drilled: probe the soffit, take out one perpend over each bearing, and find out whether there is one member or two and what the inner leaf is doing.
2. Take the temporary load as the full rectangle of masonry above the opening, not the triangle. Arching is a property of a wall that is intact and undisturbed, and you are about to cut into the exact zone the arch springs from.
3. Add whatever floor bears on the leaf being supported. Joists running parallel still often have a trimmer landing on the wall, and a ceiling below counts as well as the floor above.
4. Check what the props stand on before choosing sole plates. A suspended timber floor mid-span is not a slab, and the required bearing area comes from a pressure the temporary works designer sets, not from a habit.
5. Rake the joints for the supports out of a bed joint with continuous bond, keeping clear of the perpends, and never at the same course on both sides of a leaf.
6. Load the props by hand and stop at snug. Over-jacking lifts the wall, opens the joints above and drops it again when the prop relaxes, which is a crack you built.
7. Leave the props in until the new member has taken its bearing, been packed solid to the underside of the masonry, and the packing mortar has had the time its mix wants.

Cutting Pockets That Will Take the New Bearing

With the load transferred, the old member comes out and the bearings are the next thing to fail. Almost every lintel found rotted, rusted or cracked has damp masonry at one or both ends, and cutting a pocket into that damp masonry gives the new member a bearing on the same material that destroyed the old one. Cut back to sound, well-bonded units. Where that means going deeper than the intended bearing, the pocket is rebuilt in dense units and a strong mix and given time to gain strength before the member goes on it, and where it means going deeper than half a brick into a jamb pier, the pier itself is now the question rather than the lintel.

How much bearing is enough is a question with several right answers depending on what you are installing, and it is the one figure on this job that people habitually shave to make a member fit an existing pocket. It never fits by shaving the bearing. A prefabricated lintel's declared capacity is stated against a minimum end bearing, and installing it on less does not derate it gracefully — it takes the published figure out of scope entirely. The International Residential Code's anchored masonry veneer provisions carry a minimum end bearing for the noncombustible lintel that supports veneer over an opening, and TMS 402/602 governs bearing stress where the reaction crosses into the masonry, which is a separate check from the length.

Two details close the pocket out. The bearing is packed solid to the underside of the masonry above with a dry-ish mortar or slate pinning, because a member with a 10 mm gap over it is a member the wall reaches only after it has moved 10 mm. And on an external cavity wall, the tray goes back. BS 8215, Code of Practice for Design and Installation of Damp-Proof Courses in Masonry Construction, covers the cavity tray at a lintel; combined lintels that are formed to act as their own tray still need stop ends and weeps through the outer leaf, and the tray maker's detail and the warranty provider's standard set the spacing. A perfectly sized lintel with no tray over it produces damp at the head within two winters, and the owner will call that the same failure again.

Toothing the Courses Back In

The rebuild is small and it is where the repair becomes visible or invisible. Courses come out over the head to get the old member free, more come out along the crack itself if it has to be cut and pieced rather than repointed, and the two pockets need units. Take the count off the face area with the joint width you are actually working to rather than off a rate, because a repair to a wall with 6 mm joints and a wall with 12 mm joints do not want the same number of units per square metre, and add enough waste to allow for the ones that break coming out of the pallet and the ones that turn out not to match once they are wet.

Matching is a longer conversation than quantity. Reclaimed units from the same period usually beat new specials on a facing elevation, and where the wall is nineteenth-century soft red the new mortar should be softer than the units and softer than the surrounding mortar, not harder. A repair mix stronger than the brick moves the failure from the joint into the face of the unit, and spalled faces cannot be repointed out. Dampen absorbent units before laying: a dry brick with a high initial rate of absorption pulls water out of the mix faster than the cement can use it and leaves a joint with no bond at the interface, which is a defect that shows up as a hairline crack around every perpend two summers later.

Filling the Crack Last, and Only Once It Has Stopped

The crack is the last thing to touch and the first thing the owner wants done. Filling it before the cause is corrected buys a repair that reopens on the same line, and reopening on the same line is what convinces people that structural repairs do not work. Once the new member is in, bearing, packed and cured, the wall has its load path back and the crack is a fabric defect rather than a symptom. Where the diagnosis pointed away from the lintel to foundation movement, the crack does not get filled at all until the monitoring record says the movement has stopped, which is a different timescale entirely.

Rake out rather than skim over. A surface fill across an open joint has no thickness to bond with and pops out inside a couple of freeze cycles; the working rule is to rake to a depth of at least two to two and a half times the joint width, back to sound mortar, using a hand chisel or a raking wheel rather than a disc that widens the joint and burns the arrises. Rake the perpends as well as the beds. Brush and dampen before repointing. On any pre-1919 wall the replacement mortar should be lime-based and matched by sieve to what is coming out, because a cement repoint in a lime wall traps water in the units and moves the decay from the joint into the brick face.

Take the quantity off the actual joint run rather than by area, because a stepped crack repair is a long thin measurement and estimating it as a square metre rate gets it wrong in both directions. Count the beds and the perpends along the crack, add the joints disturbed around the pockets and the rebuilt courses, and work in the raked depth you have committed to rather than the depth you hope to reach. The volume is small enough that a bagged repair mortar is the sensible buy, and small enough that running out mid-elevation means a visible colour change where the second batch starts.