Building a Free-Standing Garden Wall That Stays Up

Gravity Is Nearly All the Prestress There Is

Treat the wall as a plain vertical cantilever, a metre of run at a time, and give the brickwork 20 kN/m3 — the order of figure BS EN 1991-1-1 Eurocode 1 Part 1-1 tabulates for clay masonry in its densities annex. A strip of thickness t and height h weighs gamma t h per metre of run. On the point of rotating about its leeward face its restoring moment is that weight acting at t/2, and the overturning moment from a uniform wind pressure q is q h squared over two. Set the two equal and almost everything cancels, leaving one expression worth carrying around: the pressure that tips an unbonded wall is gamma t squared over h. Thickness squared on the top, height on the bottom, nothing else in it.

Put the numbers in for the wall people actually ask for. A half-brick skin, 102.5 mm, standing 1.8 m comes out at 117 N/m2. That is the whole of its gravity resistance, and 117 pascals is the stagnation pressure of a wind of about 14 m/s — enough to move the top of a young birch, not enough to be worth a phone call. Take the same wall to one brick, 215 mm, and the figure quadruples to 514 N/m2, because thickness is squared and height never was. Neither is a design wind pressure anywhere in Britain, and the second is not comfortably clear of one.

Yet half-brick walls stand for decades in sheltered courtyards, so gravity plainly is not all of it. What holds them is the flexural bond of the bed joints — the tensile strength of the mortar-to-unit interface, which BS EN 1996-1-1 Eurocode 6 tabulates in two directions and modifies through its National Annex. Bending a wall about a horizontal axis cracks it along a bed joint, and that is the weak direction: the smaller of the two tabulated values, very sensitive to unit suction and to how well the bed was filled, and effectively gone if the joint was laid in frost or disturbed after it began to stiffen. Designers routinely refuse to rely on it where a collapse would put masonry on a footpath, which is why the published tables are as conservative as they look.

The practical consequence is short. If the wall can be thick, make it thick, because the return on thickness is quadratic and the return on everything else is not. If it cannot, break the span with piers so the panel bends the strong way. If neither gets there, it wants reinforcement or a designer, in that order. The actual height limit is a real calculation with a real wind pressure in it — BS EN 1991-1-4 Eurocode 1 Part 1-4 in the European system, ASCE/SEI 7 Chapter 29 in the American, where a solid free-standing wall is an 'other structure' with force coefficients that follow its proportions. No calculator on this page performs that check; the ones below tell you what the wall weighs, consumes and costs.

The Trench Is Sized by Rotation, Not by Bearing

Work an example and the point makes itself. A one-brick wall, 215 mm, standing 1.8 m above the top of its footing, on a strip 600 mm wide by 400 mm deep in concrete at 24 kN/m3. The wall weighs 7.74 kN per metre of run and the footing another 5.76, so 13.5 kN presses down on every metre of strip. Push it with an illustrative 0.6 kN/m2 of wind: 1.08 kN per metre at mid-height, 1.3 m above the underside of the footing, an overturning moment of 1.404 kNm per metre. Divide by the vertical load and the resultant sits 104 mm off the centre of a 600 mm base, whose middle third only reaches 100 mm. The heel is in tension. Widen the strip to 750 mm and it gains another 1.44 kN of its own weight, the eccentricity drops to 94 mm, and the middle third reaches 125 mm.

Now look at what that widening did to the bearing pressure. At 750 mm wide with a 94 mm eccentricity the toe carries about 35 kPa and the heel about 5. Thirty-five kilopascals is nothing; almost any ground that would hold a fence post carries it without noticing. The width of a garden wall footing is therefore not a bearing calculation at all — it is a rotation calculation, and the soil's strength barely enters. That is the opposite of the intuition most people bring from slab and pad work, and it is why widening the trench by 150 mm is the standard cure when a wall turns out a course or two taller than the sketch said.

Depth is a separate question with a separate answer, and the wall does not answer it. The ground does: below local frost penetration in any climate that has one, and in shrinkable clay below the depth at which seasonal moisture movement stops mattering — which nearby trees push a long way further down. NHBC Standards Chapter 4.2 Building near trees is the UK reference, and it will call for a footing far deeper than a garden wall needs for strength. Where a mature tree sits within influencing distance of the line, the honest options are to move the wall, found it deep, or accept that it will crack and put movement joints where you would rather the cracks appeared.

Two things get missed in the trench itself. A footing stepped down a slope steps in whole courses of the masonry above it rather than in convenient increments of concrete, so the brick sets the step height and it has to be set out before anything is dug. And the concrete is a structural element, not a levelling bed: it wants a specification and a trench bottom that has not been rained into a slurry and poured over anyway.

Gate Piers Live a Different Life From the Rest of the Wall

Everything above assumes the load arrives as wind, spread over an area, a few times a winter. A gate pier is loaded by a person, at a point, several times a day, for years, and the load has a horizontal component the rest of the wall never sees: a hung gate hangs its weight off two hinges on one face, which is a permanent twisting moment about the pier's vertical axis, and a gate allowed to slam adds impact to it. That is what pulls gate piers out of plumb and opens the joint between pier and panel — the classic garden wall defect, and one that shows up within a couple of years rather than after a gale.

So a gate pier is a separate design from the wall it interrupts. It wants to be bigger in plan than the intermediate piers, it wants a pad rather than a share of the strip, and past a light pedestrian gate it wants reinforcement carried from that pad up into the pier, with the hinge fixings resolved into it rather than plugged into a brick face. A 0.9 by 0.9 by 0.45 m pad is a reasonable opening sketch for a pair of domestic pedestrian gate piers; a vehicle gate is somebody else's calculation, not a scaled-up version of this one. Set the centres off the gate leaf plus its hinge and latch clearances before the trench line is marked, because an opening 30 mm short is corrected in the ironmongery and one 30 mm long is corrected in the brickwork.

Clay Grows, Concrete Shrinks, and the Joints Are Not Interchangeable

A long masonry run will move, and the direction depends on what it is made of. Fired clay leaves the kiln bone dry and takes up moisture from the air for years afterwards, expanding irreversibly as it does — most of it early, all of it in the wrong direction if the wall is restrained. Concrete units do the opposite and shrink as they cure and dry. So a clay wall wants expansion joints sized to let it grow, and a blockwork wall wants control joints placed to decide where its shrinkage cracks appear. They are not the same detail, not on the same spacing, and copying one interval onto the other material produces cracking in a wall that was jointed all along.

A garden wall makes this worse than a building does. It is unrestrained at the top, exposed on both faces, often built dead straight for twenty or thirty metres without a return to absorb anything, and it swings through a far wider temperature range than a wall inside a heated building. UK practice puts clay brickwork movement joints at intervals of the order of ten to twelve metres, with the joint width derived from the panel length rather than fixed; PD 6697 carries the recommendations, and the brick maker's literature carries the movement figures for the specific unit, which is where the number should come from. Piers and gate openings are natural places to put one, since the wall is already interrupted there.

For the blockwork case the calculator below applies a widely used simplification — spacing at the lesser of twice the wall height or 7.6 m — to give a joint count for a straight run. Read it as a budgeting figure. It knows the height and the length and nothing else: not the units' documented shrinkage, not whether the run has openings or steps, and above all not whether the wall is clay, where the whole logic runs backwards. NCMA TEK 10-2 Control Joints for Concrete Masonry Walls is the guidance it simplifies, and the layout belongs to whoever designed the wall.

Why the Order Comes Out at Twice the Guess

A fourteen-metre wall 1.8 m high has 25.2 square metres of face, and standard 215 by 65 units on 10 mm joints give a nominal module of 225 by 75 — about 59 bricks per square metre, or a shade under 1,500 for the run. That is the number everyone quotes, and it is the number for a wall you would not build, because a 102.5 mm skin at that height is the top row of the table above. What actually goes up is one brick thick, two leaves bonded together, so the real figure is close to 3,000 before a single pier, cut or breakage. Piers add their own volume, and every one is a course-by-course bond into the panel with cuts in it.

Mortar behaves the same way and surprises people harder, because the bed is now a full 215 mm deep rather than a leaf's worth. Something around a cubic metre of mixed mortar for that wall is the right order of magnitude, and a cubic metre of mortar is a little over two tonnes of material to get on site, mix and lift — which changes what the job needs in the way of a mixer, a labourer and somewhere to stand a sand heap. Cement is the one item on the list with a shelf-life problem: what has sat through a wet fortnight under a torn sheet is not the cement the specification assumed.

Waste runs higher here than on a house wall, for reasons specific to the job. Every pier is cuts. Every movement joint is a stopped end. Copings break in handling more readily than bricks do. Ends and returns consume units without consuming face area. And the client is looking at both sides of everything, so the tolerance for a badly matched top-up delivery is nil — which is the argument for ordering the whole quantity in one batch.

1. Take the face area off the elevation, then double the brick count for a one-brick wall before adding anything else.
2. Add the piers as separate volumes at their real plan size and full height, not as a percentage on the end.
3. Price coping by the run plus every end and return, with a breakage allowance of its own.
4. Total the concrete across the strip and every pier pad before booking a delivery.
5. Order the facing units in one batch, and keep the geometric quantity and the waste allowance on separate lines.