Case Study: 2210mm of Ceiling Inside a 2500mm Limit

Most of our case studies are about what a garden room is. This one is about a number. The customer came to us wanting as much interior headroom as they could possibly get, and they were willing to spend what it took to get it. They also refused to trade away build quality or insulation to do it. Those two goals pull in opposite directions, which is what made this project worth writing about.

Permitted development allows an outbuilding up to 2.5m tall measured from ground level. That is the whole budget. Everything has to fit inside it: the floor structure, the ceiling, the roof joists, the fall on the roof and the roof covering. On a small room this can still allow for reasonable space. On a bigger one it's a real problem, because a longer span needs deeper joists, and every millimetre of joist depth comes straight off the ceiling height.

This build is 7.3m by 4.8m externally, sized to keep the internal floor area just below the 30m² threshold at which building regulations approval kicks in. It is, in other words, about as large as a garden room gets before it stops being a garden room - and therefore about the worst possible case for headroom.

We finished with a 2210mm ceiling, while also keeping it extremely well insulated. Here is how we achieved that.

Diagram of the 2500mm height budget showing a 2210mm interior, the roof build-up above it, the floor structure below ground, and the four decisions that recovered 275mm — The height budget. Four decisions recovered roughly 275mm between them - and not one of them was taken out of the insulation.

First, We Went Downwards

The 2.5m limit is measured from ground level. It says nothing about what you do below ground level. So the cheapest millimetres available are the ones you dig for.

We designed and commissioned an excavation pit covering the entire footprint of the building, around 300mm deep. The bottom half was filled with MOT Type 3 aggregate, laid over a geotextile membrane and deliberately left uncompacted. That last detail matters and it is the opposite of what you would normally do with a sub-base. This aggregate is not there to carry load - the ground screws do that. It is there to hold open a void that soil cannot fill, so rainwater has somewhere to go and it doesn't flood the floor frame which is underground as well. Type 3 is graded with fewer fines than Type 1, so it holds more empty space between the chips, and compacting it would destroy the very thing it was laid for. It leaves roughly 3m³ of drainage void under the building.

Around the perimeter we set concrete lintels as retaining walls. These carry nothing structural either. Their job is simply to stop the surrounding soil from slumping back into the pit over the years and quietly filling in the space we had just paid to dig. They were finished with patio paving laid all around the pit.

Excavation cross-section showing the 300mm pit, the Type 3 MOT drainage layer over a geotextile membrane, and the concrete lintel retaining perimeter — The groundwork drawing we issued for the excavation. Levels are taken from the front ground level; the retaining perimeter has to come out dead level for everything above it to work.

The completed excavation pit filled with MOT Type 3 aggregate, with concrete lintels forming a retaining perimeter and the existing patio in the foreground — The pit, finished and ready for the ground screws. Unremarkable to look at, and the single largest height gain in the whole project.

The ground screws then went into the pit with their base plates 125mm below ground level. In a normal build those plates sit around 25mm above it. That 150mm swing is the biggest single win in the project. The groundwork is involving and messy, but it if you want to achieve maximum headroom - it is worth.

A Floor Frame With Nowhere to Go

Sinking the screws only helps if the floor frame is shallow enough to sit in the space we made. Our standard floor frame is 150mm deep. For this build we designed one at 100mm.

You cannot simply make a joist shallower and hope. A timber joist gets its stiffness overwhelmingly from its depth, so removing 50mm removes a great deal more strength than it sounds like. We bought that strength back with width instead, using 4x3 C24 timber tripled up and bound together into single structural member where the loads demanded it. More timber, more labour, more cost - and the same load carried at two thirds of the depth. The cavity still takes the full 100mm of PIR, exactly as it would have at 150mm.

Comparison of a standard 150mm floor frame sitting above ground on screw plates against the 100mm tripled-beam frame sunk into the pit, finishing level with the patio — A standard frame puts your floor a step above the patio. This one lands flush with it, and the whole floor structure sits below the line the 2.5m limit is measured from.

With the plywood subfloor and the laminate on top, the finished interior floor came out level with the external patio. That was the point of the exercise - but it also gave the customer something they had not asked for. There is no step at the door.

Low angle view along the patio showing the paving running level with the interior floor through the open sliding doors — The patio and the interior floor are the same level. Everything that made this possible is buried out of sight.

Trading Spacing for Depth in the Roof

The roof joists span a little over 4.6m, which is a long way for a garden room. The standard answer for this span is 2x9" joists at 225mm deep, and that is the standard answer for a good reason: depth is by far the cheapest way to span a distance in timber.

Depth was the one thing we did not have. So we ran the same trade as the floor, in the other direction: 3x7" joists at 175mm deep, set at 400mm centres instead of the usual spacing. Thicker sections, more of them, closer together. It is a more expensive roof by some margin, and it hands back 50mm of ceiling.

Comparison of standard 225mm deep roof joists at 600mm centres against 175mm deep joists at 400mm centres spanning the same 4.6m — Same span, same load, 50mm less depth. The bill goes up and the ceiling comes down - which was exactly the trade the customer asked us to make.

Two Slopes Instead of One

A flat roof is never actually flat. It has to fall, or water sits on it. That fall is built up in firring strips on top of the joists, and the build-up eats height at the high end. Across 4.8m of depth, even a gentle 1:80 fall has to climb a surprising amount by the time it reaches the far side.

So we split it. Rather than falling continuously from one edge to the other, the roof runs up to a high line set 1.5m back from the front and running left to right, then falls away in both directions. The pitch itself never changes - it stays at 1:80, and it is still very much a flat roof. But neither run has to climb as far as a single 4.8m run would, which takes another 25mm or so off the thickest part of the build-up.

Section through the roof depth comparing a single fall across the full 4.8m against two falls from a high line set 1.5m back from the front — Two shorter runs instead of one long one. Same fall, same flat roof, lower peak. The falls are drawn exaggerated - at 1:80 the real rise is a few tens of millimetres.

This means that some small amount of rainfall is actually flowing towards the front of the building, but we accepted that as a reasonable trade-off. We also added a small extra layer of fibreglass to divert the water away from the front door.

Roof plan seen from above, showing the ridge set 1.5m back from the front and a wedge of extra fibreglass on the front slope that pushes run-off out to either side of the doorway — The roof from above. The front 1.5m drains forwards, so a wedge of extra fibreglass splits that run-off and walks it out to either side before it reaches the doorway.

The Flitch Beam Over the Doors

The best detail in this build is one you cannot see at all. The customer wanted a wide opening at the front: a sliding door with side lights, 3.4m across in total, and as tall as we could make it.

To span an opening that wide you would normally drop in a rolled steel joist. It works, it is well understood, and on this building it would have been a serious mistake. Steel conducts heat hundreds of times better than timber. An uninsulated steel beam laid across an insulated envelope is a thermal bridge in the most literal sense - a cold strip running the full width of the wall, right above the largest opening in the building. Warm, humid indoor air finds it, condenses on it, and mould follows. On a build where the entire point was uncompromised insulation, that was not a trade we were willing to make.

The alternative is an all-timber beam, but at 3.4m that means going to around 225mm deep - and 225mm of header depth comes straight out of the door height we were trying to protect.

So we used a composite flitch beam: a 10mm steel plate sandwiched between two timber members and bolted through. The steel does the structural work. The timber carries the steel and, more importantly, keeps every face of it on the warm side of the insulation line. The result spans 3.4m on a header just 150mm deep, with no thermal bridge anywhere in it.

Three options for spanning the 3.4m door opening: a steel RSJ which bridges heat, a 225mm timber beam which costs headroom, and the 150mm flitch beam used in this build — Steel is strong but cold. Timber is warm but deep. A flitch beam is the detail that refuses to choose.

The Result

Excavating for the screws found 150mm. The floor frame found 50mm. The roof joists found another 50mm. The two-slope roof found 25mm. Together that is roughly 275mm of ceiling that a conventionally built room of this size would simply not have had.

The finished interior ceiling height is 2210mm, inside a building that measures under 2500mm to the top of its roof. We went back with a tape measure to check, because a claim like that is worth checking.

Tape measure run from the finished floor to the ceiling inside the completed room — Floor to ceiling, inside. The slope in the ceiling line by the door is the front run of the two-slope roof, showing itself.

Tape measure run up the outside of the building to the top of the roof coping, reading under 2500mm — And the same building from the outside, measured to the top of the coping. This is the number that has to stay under 2500mm.

A member of the Precision Rooms team standing inside the finished room reaching up to touch the ceiling — The least scientific measurement we took, and the one that actually tells you what 2210mm feels like to stand in.

None of It Came Out of the Insulation

This is the part that matters most, because it would have been very easy to find those millimetres the lazy way. Thinner insulation is the first thing to go when height gets tight. It went nowhere on this build:

Roof: 150mm rigid PIR between the joists in a ventilated cold roof build-up - 0.16 W/m²K
Walls: 100mm rigid PIR in the frame cavity, plus a further 50mm of rigid PIR across the outside of the frame to break the thermal bridging through the studs - 0.16 W/m²K
Floor: 100mm rigid PIR in the frame cavity - 0.25 W/m²K

The 50mm layer over the outside of the wall frame is worth dwelling on, because it is the reason the walls perform as well as the roof despite holding less insulation. Timber studs conduct heat far better than the PIR sitting between them, so in an ordinary wall every stud is a small thermal bridge repeating every few hundred millimetres. Wrapping the whole frame in a continuous layer of rigid insulation cuts every one of those bridges at once. It also, incidentally, is why the as-built building came out 100mm wider and deeper than the groundwork drawings called for.

The cold roof deserves a note too, since a ventilated roof needs a clear air path above the insulation to work. The firrings that create the two falls sit on top of the joists and lift the deck clear of the PIR, leaving a void that varies between roughly 45mm and 70mm along the run. With a single ventilated run of 4.6m, that is comfortable. The firrings ended up doing two jobs at once: shaping the roof and ventilating it.

Was It Worth It?

Honestly, this was an expensive way to build a garden room. Excavating a pit, tripling up floor timbers, over-speccing the roof joists, fabricating a flitch beam - every one of those decisions costs real money, and none of them shows up in a photograph. A customer who wanted a straightforward room would not spend any of it, and we would not suggest they should.

But this customer knew exactly what they were buying, and what they bought was a room that does not feel like an outbuilding. It feels like a room in a house. That difference is 275mm, and 275mm turned out to be worth a great deal to them.