A window datasheet says U-value 0.85 W/m²K. A competing datasheet says 0.80. The specification demands 0.80, so the second window wins — except the two numbers came from different calculation standards, different reference sizes, and different spacer assumptions, and the "losing" window may in fact be the better thermal product. Whole-window U-values are outputs of a defined calculation, and reading them without knowing the calculation is how fenestration procurement goes wrong.
This is the calculation, as EN ISO 10077-1 defines it.
The three-zone formula
EN ISO 10077-1 splits a window into three thermal zones and weights each by how much of the window it occupies:
Uw = (Ag · Ug + Af · Uf + lg · Ψg) / (Ag + Af)
- **Ag, Ug** — glazing area (m²) and glazing U-value (W/m²K, from EN 673 or measurement). This is the number glass suppliers quote: 1.1 for standard double low-E, 0.5–0.7 for good triple glazing. - **Af, Uf** — frame area and frame U-value. The frame's own thermal transmittance comes from a 2-D heat-flow simulation per EN ISO 10077-2 or hot-box measurement. - **lg, Ψg** — the visible glass perimeter (m) and the linear thermal bridge coefficient (W/m·K) of the glass edge, where the spacer bar couples the warm and cold panes.
Nothing else enters. No air-leakage term, no solar term — EN ISO 10077-1 is a pure conduction calculation at steady state. Solar gain is a separate number (the g-value or SHGC, [treated here](/resources/blog/window-u-value-vs-shgc-climate)), and airtightness is tested under a separate standard.
A worked example on the reference window
The standard reference size for a single-sash window under the EN system is **1.23 m × 1.48 m** (it also anchors Passive House component certification). Take a triple-glazed unit in a frame with a 100 mm face width:
- Total area Aw = 1.82 m² - Glazing Ag = 1.03 × 1.28 = 1.32 m² (72%) - Frame Af = 0.50 m² (28%) - Glass-edge perimeter lg = 4.62 m - Glazing Ug = 0.70 W/m²K, frame Uf = 1.0 W/m²K, warm-edge spacer Ψg = 0.04 W/m·K
Uw = (1.32 × 0.70 + 0.50 × 1.0 + 4.62 × 0.04) / 1.82 = (0.92 + 0.50 + 0.18) / 1.82 = **0.88 W/m²K**.
Three readings of that arithmetic are worth pausing on.
The glass flatters, the frame decides. The glazing contributes 0.92 of the 1.61 W/K total — but it earned that at Ug 0.70. The frame, at barely a quarter of the area, contributes almost as much heat loss per square metre of window as the far larger glass area, because its U-value is worse. Swap the Uf 1.0 frame for a thermally-broken aluminum frame at Uf 1.4 and the same glass delivers Uw ≈ 0.99 — the window just lost its sub-0.9 rating without the glazing changing at all.
The spacer is a tenth of the result. 4.62 m of glass edge at Ψg 0.04 adds 0.10 W/m²K to Uw. Published warm-edge Ψ values run roughly 0.03–0.05 W/m·K against roughly 0.08 for a conventional aluminum box spacer (the German flat-glass association BF publishes representative tables). Spacer choice alone can move a whole-window U-value by about 0.1 — the difference between passing and failing a 0.80 specification.
Size is not a detail. Because frames are the weak zone, a small window (higher frame fraction) computes worse than a large one with identical construction. That is why the reference size exists — and why comparing a Uw quoted on a 2.4 m sliding door against one quoted on the 1.23 × 1.48 m reference is not a comparison.
Why European and North American numbers do not match
North America rates fenestration under NFRC 100, and the two systems disagree by design:
- **Boundary conditions.** EN/ISO calculates at 0°C outside / 20°C inside; NFRC at −18°C / 21°C. Colder conditions change gas-fill convection and radiation exchange, so the same physical window scores differently. - **Method.** NFRC solves the whole product numerically; the EN system combines analytically-derived component values (Ug, Uf, Ψg). - **What gets compared.** NFRC rates the whole product at fixed model sizes under one condition set; the EN system characterizes components separately, which is what lets a frame system carry its Uf across many window builds.
The practical consequence: an NFRC U-factor of 0.17 Btu/h·ft²·°F is not simply "0.97 W/m²K" — the unit conversion is right, the boundary conditions are not. Cross-continental procurement needs the rating system named next to every number. (It also produced a real design divergence: European IGUs settled near 16 mm cavities, North American near 12–13 mm, each optimal under its own rating conditions.)
What this means for frame material
Run the formula in reverse: to hit Uw ≤ 0.80 with realistic triple glazing (Ug 0.60–0.70) on the reference window, the frame must deliver roughly Uf ≤ 1.0–1.3 with a warm-edge spacer. Thermally-broken aluminum reaches Uf 2.5–4.0 in common systems — the formula simply does not close. Insulating frame materials — pultruded GFRP at a bulk conductivity around 0.3 W/m·K, timber, uPVC — are what make the arithmetic work, which is why [Passive House certified windows](/resources/blog/passive-house-window-u-value-requirements) are dominated by them. The material-level comparison is on the [FRP vs aluminum windows page](/technology/frp-vs-aluminum-windows); F1's pultruded GFRP-PU [fenestration systems](/products/fenestration-systems) hold PHI component certificate 2491wi03 with certified whole-window performance to Uw 0.78.
To run the EN ISO 10077-1 arithmetic on your own frame, glazing, and spacer combination — including the pass/fail against cold-climate program targets — use the free [whole-window U-value calculator](/technology/u-value-calculator). It exposes every term of the formula above, so you can see which zone is costing you the rating.
