U-value of a suspended timber floor insulated with wood fibre or sheep’s wool

Wood-fibre, sheep’s wool or hemp batts between the joists of a suspended timber floor, with the membranes that make them work.

Insulating a suspended timber floor with 100mm of wood-fibre batts (STEICOflex 036) between the joists, held in a windtight vapour-open membrane with an airtight layer under the boards, takes a typical floor (30m of exposed perimeter around 50m², P/A 0.6) from about 0.81 W/m²K to 0.29; 125mm reaches the 0.25 Approved Document L asks of a renovated floor, and 150mm gives 0.22. Sheep’s wool and hemp batts (λ 0.039) do about 0.01 worse at the same depth. One measured case is worth knowing: in a UCL study of a London terrace, 100mm of wood fibre took the floor from a measured 1.04 to 0.36 — a 65% cut, a little more than the standard model’s 0.57 to 0.23 — and the authors put what was left above the model’s 0.23 mainly down to the fit: batts that didn’t fill the space, chamfered edges and gaps, so fit it tightly.

Open this build-up in U-Monkey — pre-filled, free

Every layer editable · live U-value · Part L & Future Homes Standard targets · free PDF report

The build-up (outside → inside)

Batts at their declared λ: STEICOflex 036 0.036; Thermafleece CosyWool 0.039 and UltraWool 0.035; IndiNature IndiTherm 0.039; STEICO universal 0.048 under the joists. Joists at BR 443’s 10.8% for floors. U-Monkey’s default floor: 30m of exposed perimeter around 50m² (P/A 0.6), 225mm void, 1,500mm² of vents per metre, sand or gravel (λ 2.0). The UCL rows use the published inputs of Pelsmakers & Elwell (2017) — perimeter 9.1m, area 27.7m², clay (λ 1.5), 220mm walls, the floor 0.17m above ground, void walls U 1.7, 2,400mm² of vents per metre, 19mm boards — as U-Monkey’s validation case for that floor does. The membranes carry no thermal resistance and are not entered as layers. The pre-filled link opens as an upgraded floor.

Typical U-values for this construction

SpecificationU-value (W/m²K)Notes
100mm wood-fibre batts (STEICOflex 036)≈ 0.29As pre-filled
125mm≈ 0.25Meets the 0.25 for a renovated floor (0.249)
150mm≈ 0.22
175mm≈ 0.20
200mm≈ 0.18
100mm sheep’s wool (Thermafleece CosyWool, λ 0.039)≈ 0.30Thermafleece’s own table: 0.31
150mm sheep’s wool (CosyWool)≈ 0.23Thermafleece: 0.24
150mm sheep’s wool (Thermafleece UltraWool, λ 0.035)≈ 0.22Thermafleece: 0.22
140mm hemp batts (IndiTherm, λ 0.039)≈ 0.24
100mm wood fibre + 22mm wood-fibre board under the joists≈ 0.25The board covers the joists (0.252)
150mm wood fibre + 22mm wood-fibre board under the joists≈ 0.20
The UCL field-study floor (its published inputs), uninsulated≈ 0.57The paper’s model: 0.57; measured on site: 1.04
The UCL floor, 100mm wood fibre at the paper’s λ 0.038, joists counted≈ 0.26
The UCL floor, 100mm wood fibre, joists left out (as the paper modelled it)≈ 0.23The paper’s model: 0.23; measured: 0.36

Figures computed to BS EN ISO 13370 with BR 443 conventions, exactly as the pre-filled calculator works them out, with any pass or fail judged on the unrounded U-value. Indicative for the generic build-up and the footprint and depth shown: your products, thicknesses, perimeter, area and ground will move the number, and that's what the calculator is for.

What changes the number

Depth and λ come first: each 25mm of wood fibre between the joists is worth about 0.02–0.04 on a typical floor, and sheep’s wool or hemp at λ 0.039 does about 0.01 worse than wood fibre at 0.036 for the same depth; a 22mm wood-fibre board under the joists covers the timber and is worth about as much as another 25mm between them. The floor’s shape and its void then move it, as on any suspended floor — the perimeter, the area and the ventilation openings all go into BS EN ISO 13370. The membranes do the rest, and they are where the measured case teaches most: in Pelsmakers and Elwell’s UCL study (Energy and Buildings, 2017) of a pre-1919 London terrace, 100mm of wood fibre held in a tightly stretched, stapled breather membrane cut the floor’s measured U-value from 1.04 to 0.36 — 65%, a little more than the 60% the BS EN ISO 13370 model predicted (0.57 to 0.23, the joists left out). The real floor lost more heat than the model said before and after; the authors suggest the uninsulated difference may come from the model’s inputs, the wall–floor junction it leaves out or its steady-state assumptions, and what was left after insulating mainly to the fit — ill-fitting batts, chamfered edges and gaps — hence their conclusion that “good installation quality is important to maximise heat-flow reductions.” U-Monkey reproduces the model on the paper’s published inputs (0.57; 0.23 without the joists, 0.26 with them). Historic England’s guidance on suspended floors favours natural fibres in older buildings, “allowing both air and moisture vapour to pass through slowly”, says the most reliable installations include some form of restraint below the insulation, and notes that a membrane with low vapour resistance below the joists “can usefully enhance the effectiveness of the insulation by reducing air movement through the floor”; it quotes 0.22, the 2010 figure for a new floor; a renovated floor needed 0.25 then and still does (Approved Document L, Table 4.3, with a lesser standard where floor levels prevent it, note 8), and a new floor now needs 0.18. Keep the void ventilated — Thermafleece asks for at least 150mm of ventilated space below the insulation — and every air brick clear.

Next: check the junctions

A U-value covers the flat area of an element — but heat also escapes where elements meet, and at modern U-values those junctions can be 20–30% of a dwelling's fabric heat loss. Now you know your U-values, model the junctions with the free ΨMonkey psi value calculator (no subscription; PDF reports £25 +VAT a junction):

Or read thermal bridging explained to see how junction Ψ-values feed your SAP calculation.

Frequently asked questions

What is the U-value of a suspended timber floor insulated with wood fibre or sheep’s wool?

Insulating a suspended timber floor with 100mm of wood-fibre batts (STEICOflex 036) between the joists, held in a windtight vapour-open membrane with an airtight layer under the boards, takes a typical floor (30m of exposed perimeter around 50m², P/A 0.6) from about 0.81 W/m²K to 0.29; 125mm reaches the 0.25 Approved Document L asks of a renovated floor, and 150mm gives 0.22. Sheep’s wool and hemp batts (λ 0.039) do about 0.01 worse at the same depth. One measured case is worth knowing: in a UCL study of a London terrace, 100mm of wood fibre took the floor from a measured 1.04 to 0.36 — a 65% cut, a little more than the standard model’s 0.57 to 0.23 — and the authors put what was left above the model’s 0.23 mainly down to the fit: batts that didn’t fill the space, chamfered edges and gaps, so fit it tightly.

Does underfloor insulation work as well as the calculations say?

In proportion, yes; in absolute terms, only as well as it is fitted. In a UCL field study of a London terrace, 100mm of wood-fibre insulation between the joists cut the measured heat loss through a suspended timber floor by 65% — from 1.04 to 0.36 W/m²K — slightly more than the standard model’s 60% (0.57 to 0.23). But the real floor lost more heat than the model said both before and after, and the authors tied what remained after insulating mainly to the fit: batts that didn’t fill the space between the joists, chamfered edges and gaps. U-Monkey gives 0.26 for the same insulated floor with its joists counted. Fit the batts tightly to the joists and the boards, restrain them with a stretched membrane below, and seal an airtight layer under the boards.

Is this calculator free?

Yes — building the element, reading the U-value, the compliance targets (including the Future Homes Standard), the condensation-risk screening check and the thermal-mass figures (decrement factor, time lag, κ) are all free in your browser, and the Building Control-ready PDF report is free too. No sign-up.

Does it follow the official calculation method?

Yes — the combined method of BS EN ISO 6946:2017 with BR 443 conventions, including timber bridging, mortar joints in blockwork (and in solid brickwork where BR 443 §4.2 requires them) and air-gap corrections, BS EN ISO 13370 for ground floors, suspended floors and basements, and ISO 13788 for the condensation screening check.

Other common build-ups

Floors

Walls

Roofs

Different build-up? See the full floor U-value calculator, or start from scratch in the main U-value calculator.

Need the wider compliance picture? U-values feed straight into SAP calculations — or let our New Build Compliance Pack handle the lot. Background reading: what U-values actually are, and why two calculators can give different answers for the same wall.

Your build-up won't be exactly this one.

It never is. Open this one pre-filled, then swap the insulation, change the thicknesses and watch the U-value move — free, in your browser, with a Building Control-ready PDF at the end.

Open this build-up pre-filled or call 01202 623236