ALPES Aluminum Windows and Doors Energy Saving: A K Value Is a Rate, Not a Saving, and Three of the Four Inputs Are Yours

"Energy saving" is not a property of a window

Ask ten suppliers whether their aluminium windows save energy and ten will say yes. The word is doing three different jobs in that sentence, and it gets away with it because all three sound like the same claim. One of them is a rate of heat flow. One is how much air gets past the closed sash. One is how much sun is let in. They have different units, different test rigs, different owners, and in a cooling season the third one can point in the opposite direction to the first two.

Here is the sentence we would rather have printed under our own name: a window does not save energy. A window, in a particular opening, in a particular climate, in front of a particular heating or cooling plant, changes one term of a calculation by an amount that depends on three things the factory does not have.

We publish a K value on three of our series. We are a factory with every motive to make that number sound like an outcome. This article is about what it is, what it is not, and why you are better off taking the method than the number.

Three paths for heat, and only one of them fits on a catalogue line

Path	What drives it	The number that describes it	Who actually owns it	Does it appear in a window catalogue
Conduction through glass and frame	Temperature difference between inside and out	Whole-window Uw or K value, W/(m²K)	Split between the glass supplier and the system calculation	Sometimes, usually as a series value
Air leaking through the closed joint	Pressure difference, and the quality of a seal you cannot see	Air permeability or infiltration, m³/h, or per area, or per joint length	Half the product test, half the installer's junction with the wall	Almost never in our part of the industry
Sun coming in through the glass	Irradiance, orientation, shading, season	Solar factor g, or SHGC, dimensionless	The glass build, which is a separate purchase order	Very rarely, and almost never series-wide
Air you let in on purpose	Occupants, doors, ventilation strategy	Air change rate, not a product number at all	The building, its users and its mechanical design	Never, and correctly so

Only the first row has a number that a catalogue can print on one line. That is why it wins every argument, and that is why the other three decide whether the first one mattered.

The one row in the product standard that has no class

EN 14351-1 covers windows and external pedestrian doorsets, and for most performances it gives you a ladder to specify against. Thermal transmittance is the conspicuous exception.

Characteristic	Determined by	What the standard offers	What that means when you write a specification
Resistance to wind load	EN 12211 and EN 12210	Classes 1 to 5, then E	You can write "class 3 minimum" and enforce it
Watertightness	EN 12208	1A to 9A, then E	A band, so a claim can be checked against a boundary
Air permeability	EN 12207	Classes 1 to 4	A band, and the one most often left out of an enquiry
Thermal transmittance	EN ISO 10077 or EN ISO 12567	No classification at all. A declared value in W/(m²K), or NPD	There is no band to catch a number that has quietly moved between specimens

That asymmetry is the whole problem in one table. For wind, water and air you can demand a minimum class, and a class has an edge that a claim can be tested against. For thermal transmittance there is no class, only a value, and a value can be carried from one specimen to another without anybody noticing, because there is no band to catch it.

Two things follow. First, the only defensible question is whose specimen, under which standard, at what size and with which glass build. Second, NPD — no performance determined — is a legal declaration under the same standard, and sometimes the more honest one. A declaration that refuses to state a thermal value is not hiding something worse than a declaration carrying a number whose specimen nobody can name.

And there is one line in the same CE guidance that matters a great deal for a sliding door manufacturer: door height windows, lift-and-slide doors and folding-sliding doors are classed as windows for the purpose of testing and classification. Our lift-and-slide doors are therefore asked the same question as our casement windows, in a frame whose depth, travel and threshold share almost nothing with a casement's. Being classed together for testing does not make them comparable for specification.

The arithmetic nobody runs: air against glass, and why the climate cancels out

Two formulae, both public, both elementary, both found in every heat loss guide ever published.

Conduction: Q = U × A × ΔT, in watts, where U is W/(m²K), A is area in square metres and ΔT is the temperature difference.

Infiltration: Q = 0.33 × L × ΔT, in watts, where L is the air leakage in cubic metres per hour. The 0.33 is the volumetric heat capacity of air in Wh/(m³K), which is air density of about 1.2 kg per cubic metre times a specific heat of about 1005 J per kilogram per kelvin, divided by 3600. Several public calculators and engineering guides publish exactly this constant and this formula.

Now put numbers in. This is our own arithmetic and every input is stated so you can ｒｅｐｌａｃｅ it. Take the window reference size used across European schemes, 1.23 by 1.48 m, which is 1.82 square metres. Take our published AW90 system K value of 1.3 — and read the caveat immediately: that is a system value for a series, not a whole-window value for your opening.

• Conduction coefficient = 1.3 × 1.82 = 2.37 W/K
• Setting 0.33 × L equal to 2.37 gives L = 7.2 cubic metres per hour
• Across 1.82 square metres that is about 3.9 m³ per hour per square metre

So roughly seven cubic metres of air an hour carries as much heat as the conduction through the whole window. Where does 3.9 sit on the published ladder? A published classification table puts the top EN 12207 class at 3 m³/(h·m²) at a test pressure of 100 Pa. The crossover point and the top of the class ladder are the same order of magnitude.

That is not an accusation against any product, including ours. Test pressure is a strong wind, not an annual average, and nobody lives at 100 Pa. It is a statement about scale: the air term is not a rounding error next to the conduction term. It is the same size.

Now hold the K value fixed at 1.3 and change nothing but the condition of the seals. A Chinese technical article tabulates infiltration per metre of opening joint by window type and condition. We have used its ranges, and we assume about 5.4 m of opening joint for a single sash in that reference size — our assumption, stated so that you can ｒｅｐｌａｃｅ it.

Window type and seal condition (published range)	Leakage per metre of joint, m³/(h·m)	Total for about 5.4 m of joint, m³/h	Air coefficient, W/K	Against the conduction term of 2.37 W/K
Thermal-break casement, gaskets intact, multipoint locking	0.3 to 0.8	1.6 to 4.3	0.5 to 1.4	About one fifth to three fifths of it
Thermal-break casement or uPVC, gaskets aged	1.0 to 2.0	5.4 to 10.8	1.8 to 3.6	About three quarters to one and a half times it
Sliding aluminium, pile intact	1.5 to 2.5	8.1 to 13.5	2.7 to 4.5	About 1.1 to 1.9 times it
Sliding aluminium, pile failed or absent	2.5 to 4.5	13.5 to 24.3	4.5 to 8.0	About 1.9 to 3.4 times it
Old steel or single timber, distorted frame, worn hardware	3.0 to 5.0	16.2 to 27.0	5.3 to 8.9	About 2.3 to 3.8 times it

The K value is 1.3 in every one of those rows. Same window, same wall, same weather: the air term moves by a factor of roughly nine while the printed number does not move at all. If you have ever specified a window on its U value alone and then sat next to it in January and felt cold, this table is the reason.

And now the part that is genuinely uncomfortable for a manufacturer to write. The largest leakage path on a real building is frequently not the sash at all. A published technical article on air infiltration sets out five paths and notes that the frame-to-wall joint is often the largest on site, is not measured by EN 12207 at all, and depends on the subframe, the tapes, the sealant and the workmanship. The same article says it in one sentence worth repeating: a window can legitimately hold a class 4 certificate and still be the leakiest element in the wall once installed, because the test measures the product and not the junction between the product and the building.

About half of the energy story is in the wall, and the wall is not ours. We would rather tell you that than sell you a lower K value you cannot feel.

Two denominators, one window, and a pressure you cannot carry across an ocean

EN 12207 reports air permeability against two different references. One is related to the total area of the unit, in m³/(h·m²). The other is related to the length of the opening joint, in m³/(h·m). Both are reported. They are not interchangeable, and a specification that names only one of them has not finished its sentence.

There is even disagreement in what gets published. One German classification table gives class 4 as 3 m³/(h·m²) with a joint-length companion of 0.75 m³/(h·m); another published summary gives the joint-length figure as 1. We are not going to adjudicate between them, and we are not going to print either as though it were settled. The useful point is structural: there are two columns, and the second one is the one that tracks the seals.

Then comes the pressure, and this is where honest reports stop being comparable. A published summary of BS EN 12207:2016 gives the conversion Q = Q100 × (p / 100 Pa) raised to the power of two-thirds. Run it once, at the North American test pressure of 75 Pa, and the factor is about 0.83 — our own arithmetic on a published formula.

So the identical unit, tested honestly under two regimes, reports a number about five sixths as large in one as in the other, and neither laboratory made a mistake. Airflow through a narrow joint is turbulent rather than laminar, which is why the exponent is two-thirds and not one. Two honest reports, two different numbers, one window. Which is why the only safe wording in an enquiry is: give me the measured permeability, at a stated test pressure, in both references, on a specimen that matches my configuration.

From a rate to a saving: four inputs, and we hold less than one of them

A K value is a rate. A saving is a quantity. The bridge between them is the degree-day method, which is public, old and easy to check:

Annual heat loss in kWh = U × A × HDD × 24 / 1000, and the energy you actually buy = that figure divided by the efficiency of your plant.

Four inputs. U, the value under discussion. A, your opening. HDD, your local heating degree days at a base temperature somebody has to choose. And plant efficiency, which turns delivered heat into gas, electricity or oil.

We can offer you something less than one of the four. Our published K values are system values for a series. They are not whole-window values for your opening, and they are not on the same boundary conditions as an EN ISO number or an NFRC number. We have written about those three systems elsewhere; this article starts where that one stopped.

The method is still worth carrying, because it collapses to one constant you can hold in your head. Take a difference of 1.0 W/(m²K) between what you have and what you are buying, one square metre of window, and a heating season of 2,000 degree days:

• Delivered heat: 1.0 × 1 × 2000 × 24 / 1000 = 48 kWh per square metre per year
• At 90 percent boiler efficiency: 48 / 0.9 = about 53 kWh of fuel
• With a heat pump, the same 48 kWh of delivered heat is divided by a completely different number, and that single substitution can move the answer further than the window does

That is the sentence we would like you to leave with: the saving is a product of four numbers, and the one everybody negotiates is frequently the smallest of them.

Input	Where it comes from	Your value
Delta U: existing window minus new window, W/(m²K)	Your survey of what is there now, and a calculated value for what you are buying	____
Total glazed area, m²	Your window schedule, not your floor area	____
Heating degree days, and the base temperature used	Your national meteorological or energy agency, or your energy consultant	____
Seasonal efficiency of the heating plant	Your MEP engineer, not your window supplier	____
Delivered heat saved, kWh per year	Delta U × area × degree days × 24 / 1000	____
Energy purchased, kWh or cubic metres per year	Delivered heat divided by efficiency	____

The caveat that works against us: the better the window, the worse this method fits

Degree days are counted against a base temperature, and the base is not a law of nature. A published university course on energy in buildings is blunt about where the usual base came from: an assumed balance point of 15.5 degrees Celsius was appropriate for estimating the heating consumption of buildings from the 1950s and 1960s, and it is still in use today.

What that base really represents is the building's balance point — the outside temperature at which internal gains from people, appliances and lights, plus solar gain, stop being enough to hold the set point. Insulate the building and the balance point falls. In the worked figure published by that same course, an uninsulated house balances at just under 15 degrees Celsius and the insulated version of the same house balances at about 12.

So here is the uncomfortable consequence for anyone selling high performance windows, including us: every improvement you make moves the base, and the standard base does not move with it. The degree-day method, applied with a fixed conventional base, will flatter the saving from the first improvement and will overstate the saving from the last one. The better the envelope gets, the more the conventional method overstates what the next window upgrade is worth. We would rather say that in public than let a number we calculated become a number you promise.

The published degree-day totals do not even agree inside one country. One public calculator gives the UK standard year as 2,463 degree days to a 15.5 degree base, with London at about 2,033 and northern Scotland above 2,900. Another public source gives London at about 2,100, Newcastle at about 2,400 and Plymouth at about 1,900. The two versions of London are close. They are not the same. And that figure is the base of every number downstream of it.

Aluminium, and the gap a polyamide strip has to carry

If you want to know why "aluminium" and "energy saving" were once treated as opposites, one table answers it. These are thermal conductivities in W/(m·K) taken from a Chalmers University thesis table on energy efficient windows, a peer-reviewed sensitivity study of window frame components, and industry technical pages.

Material	Thermal conductivity, W/(m·K)	What to notice
Aluminium extrusion alloy	160 in two academic tables; 200 to 218 in pages quoting the 6063 alloy	The spread is real, and it is about which alloy and temper the source meant
Steel	50	A 36 steel entry door sits here, roughly a third of aluminium's figure
Soda-lime glass	1.00	Two hundred times worse than aluminium, and it is the material you look through
Polyamide PA66 with 25 percent glass fibre	0.25 to 0.30	The thermal break material we specify
EPDM	0.25	The gasket, not the structure
Rigid PVC	0.17	The material aluminium is usually compared against, and lost to
Softwood	0.13, with 0.12 to 0.18 commonly quoted	An entry door solid wood leaf sits in this row
Polyurethane foam	0.03	Better than still air, because the air cannot move
Still air	0.024 to 0.025	Only while it stays still, which is the whole trick
Argon	0.0168	Lower than air, and only one of three mechanisms in a cavity
Krypton	0.009	Lower again, and priced and used accordingly

Two divisions anybody can check, and both are our arithmetic: aluminium against polyamide is roughly 500 to 700 to one; aluminium against softwood is roughly 1,200 to 1,600 to one. On conductivity alone an aluminium frame should lose every comparison it enters. It does not always lose, and that is the interesting part.

A thermal break does not remove the aluminium. It cannot: the polyamide strip is also the structural connection holding the inner and outer halves of the profile together, and it carries shear and thermal movement for the life of the window. What the break does is force the heat to travel around. Our C116 publishes a 34 to 44 millimetre multi-chamber polyamide strip. Read that as a detour length, not an insulation thickness. Thirty-four millimetres of PA66 is not thirty-four millimetres of wall insulation, and two systems with the same strip width can perform differently because the chamber layout and the isothermal line are not the same.

The multi-chamber part is not about air being intrinsically good either. Air at 0.024 is only good while it is stationary. The published reasoning in the Chalmers thesis is explicit: convection is avoided by making the cavity small enough that it does not start, and that is precisely what has been done to the cavities in window frames. A multi-chamber profile is a set of cavities sized to stop air from moving, not a set of air gaps. Which is why a wider cavity is not automatically a better cavity, and why the cavity in your glass unit is a separate decision from the strip in your frame.

On the glass side: one of our sliding series publishes a nominal 27 millimetre insulating cavity with inert gas fill. Dividing the two published gas conductivities, argon at 0.0168 against air at about 0.0249 at 10 degrees Celsius, gives about 0.67. The gas conduction term drops by roughly a third. That is real, and it is also one of three mechanisms in that cavity — the other two being radiation across it, which is what a low-emissivity coating addresses, and conduction along the spacer at its edge. Anyone who sells you the gas and stays quiet about the coating and the spacer has told you one third of it.

The label that folds three decisions into one letter, and points the wrong way on your south elevation

The British Fenestration Rating Council's Window Energy Rating does something no other number in this article does. It takes all three — the whole-window Uw, the solar factor g, and an air leakage term L — and folds them into a single band running from A++ downwards. As a comparison tool for a homeowner choosing between two products, that is a genuine service.

It is also where people stop thinking. A published CPD module on sustainable aluminium window systems says the important thing plainly: there is no direct correlation between the U-value and the energy rating, and it is possible to achieve an A rating with U-values ranging from 1.0 to 1.7 W/(m²K).

The same source then says the thing that matters most on a real project. An A-rated window is designed to maximise the energy it gains from the sun. Put large ones in a bright south-facing room and you may find that this makes the room uncomfortably warm in summer. It may therefore be better to specify windows and doors with a low U-value, which would be good at conserving heat in winter, rather than a high window energy rating. Conversely, on the north side of the house, a specification that maximises solar gain may produce the most comfortable conditions.

A door is not a window, and the reference size proves it

Here is a piece of arithmetic that uses nothing but two published reference sizes, and it is the reason a door value and a window value should never be put in the same column.

European schemes commonly use a window reference of 1.23 by 1.48 m. Approved Document L, in the edition published online by the Welsh Government, gives standard door sizes of 1.23 by 2.18 m for doors up to 3.6 square metres, and 2.00 by 2.18 m above that.

Quantity	Window reference, 1.23 x 1.48 m	Door reference, 1.23 x 2.18 m	What moved
Area	1.82 m²	2.68 m²	About 47 percent larger
Outer perimeter, which is where the frame and the seals live	5.42 m	6.82 m	About 26 percent longer
Perimeter per unit area	About 3.0 m per m²	About 2.5 m per m²	About one sixth less frame per unit of glass

The glass grows faster than the frame does. Because a whole-unit U value is an area-weighted average of the glass, the frame and the edge, the same profile family built as a door rather than as a window will tend to report a better number — not because the door is a better product, but because the fraction changed. Anyone who lines a door value up against a window value and calls it a ranking has compared two different fractions.

The same document contains a rule worth copying into every enquiry you send: when a single U-value is calculated for a product range of doors, the configuration chosen for the calculation should be the worst performing in that range. If a supplier cannot tell you which configuration was calculated, you do not have a value. You have a family name.

Then there is the thing no U value can capture. A door is the only part of the building envelope that is opened on purpose, dozens of times a day, by people who are not thinking about energy. A 42 inch entry door on an exposed site exchanges the air of the room every time it swings. The perimeter seal, the closer, the door sweep, and the presence or absence of a lobby are doing work that the leaf's U value cannot do and that no U value measurement includes.

This is also why the threshold is an energy decision twice over. Our sliding doors publish a low track at 8 mm and a high track at 43 mm, and one series offers a 45 mm high-ｄｒｏｐ bottom track with an optional 85 mm water barrier. Those are normally discussed as drainage and accessibility choices. They are also air choices, and they are made exactly once, at order time, in a factory several weeks of shipping away from your site.

What we publish, and what you will still have to ask for

Item	What we publish	Why there is a gap	What to do about it
Thermal transmittance	K = 1.3 for AW90, 1.6 for C120, 1.7 for C116	These are system values for a series, not whole-window values for your opening, and not on NFRC boundary conditions	Request a calculated Uw for your exact size, mullion layout and glass build, with the standard named
C113, TS130 and every sliding door series	No K value published	A sliding sash must retain a clearance in order to move, so its sealing is contact-based rather than compression-based. Not publishing is not the same as being worse	Ask for the calculation for your configuration, and compare sliding against sliding
Solar factor or SHGC	Nothing published series-wide	It is a property of the glass build, and the glass is a separate purchase order on a separate supply chain	Specify the glass first, then ask for g on that build
Air permeability class	Nothing published	No class can be transferred from a laboratory specimen to an installed junction, and we would rather not imply that it can	Ask for the measured value at a stated pressure in both references, and treat the frame-to-wall joint as separate scope
Warm-edge spacer and edge psi value	Not published	It belongs to the sealed glass unit's supply chain, and it appears in no frame catalogue	Ask the glass supplier; this is the one that decides whether you get condensation complaints
Insulation strip	PA66GF25, multi-chamber, 34 to 44 mm on C116	Width is a detour length, not a U value, and it says nothing about chamber geometry	Ask for the frame Uf, not the strip width
Glass cavity and gas fill	Nominal 27 mm cavity with inert gas on one sliding series	The gas changes one of three mechanisms in the cavity, and only by about a third of that one	Ask for the whole Ug with the coating and the spacer named alongside the fill
Salt spray and shear results	240 hours and 5,000 N, belonging to AW90	They measure coating corrosion resistance and the composite bond between strip and aluminium. Neither is a thermal, air or water rating, and neither transfers to another series	Use them for what they are, and do not let them stand in for a performance class
Your degree days and base temperature	We do not have them and cannot estimate them	They belong to your site and to your building's balance point	Get them from your energy consultant, in writing, with the base temperature stated
Your existing windows	We have never seen them	A saving is a difference, and we do not have the subtrahend	Survey or model what is there now before you compare anything
Plant efficiency and orientation	Neither is ours	The same delivered heat becomes very different purchased energy; solar gain is directional and can outweigh everything above it	Your MEP engineer and your architect own these two, and they are not small terms

Twelve lines to paste into your next enquiry

Field	Fill it in like this	What happens if you leave it blank
Opening size in millimetres, per unit	1210 x 1470, not "standard"	Any U value you receive belongs to somebody else's opening
Configuration	Which panels are fixed, which open, and in which direction	The reference configuration is usually the worst in the range, and you will not know which
Mullion positions, in millimetres	From the left jamb, to the centreline of each mullion	Frame area changes, so the area weighting changes
Glass build, written out	Thickness, cavity, thickness, coating, spacer type, gas fill	Most of the thermal performance lives in this line and it will be assumed
Whole-window U value required, with the standard named	Maximum 1.4 W/(m²K) to EN ISO 10077, for example	Three standard families give three different numbers for the same window
Solar factor required, with the standard named	A target g or SHGC, and whether it differs by elevation	One glass build gets applied to four elevations that need different things
Air permeability	Measured value, test pressure, and which reference: area or joint length	You get a class that is a range, from a pressure nobody stated
Frame Uf, asked separately from Uw	A number, with the strip width and chamber description	You end up comparing strip widths, which is not a performance comparison
Spacer type and edge psi value	Warm edge or aluminium, and the psi value in W/(m·K)	The frame performs and the edge of the glass still fogs or runs
Threshold height required, in millimetres	The number, plus any accessibility constraint that caps it	It is decided at order time and cannot be changed on site
Your design degree days and base temperature	A number, plus the base it was counted to	Every saving figure downstream is unanchored
Existing windows, in one line	Type, glass build, frame material, condition	Nobody can tell you what you will save, only what the new one costs to run

Ten questions, and the answer that sounds helpful but is not

The question	The answer that sounds like yes	Why it carries no information
Are your windows energy saving?	Yes, all of them are	No standard, no specimen, no size, no climate. Every supplier on earth can say it
What is the K value?	1.3	Whose specimen, what size, what glass, and system or whole window are all unanswered
Is it A rated?	It would be, if we tested it	A rating is issued to a configuration by a scheme, never granted to a profile by a factory
Do you have a U value?	Yes, 1.4	1.4 under whose boundary conditions. Three families of standard, three different answers
Are they airtight?	Very airtight	Airtightness is a measured number at a stated pressure, in one of two references
Does the thermal break make it as good as PVC?	Our break is the best available	Conductivity is not transmittance. Ask for Uf on a named frame section
Will it save me money?	You will notice it on your bills	No baseline, no degree days, no plant efficiency, no orientation
Is the argon worth it?	Argon is much better than air	It moves one of three mechanisms, by about a third of that one
Same profile as the door, so same value?	Yes, same system	The reference size is different and the frame-to-glass fraction moved
Can I use the catalogue number in my submission?	Of course	A submission needs a value for your opening. A series name is not that

Questions we get asked

What is the one number I should ask for first?

A calculated whole-window U value for your actual opening, your actual mullion layout and your actual glass build, with the standard named and the specimen described. Not a series value, and not a value carried over from a size you are not buying. Everything else in this article is downstream of that one.

Why do the sliding doors have no K value?

Because a sliding sash has to keep a clearance in order to move at all, so its sealing is contact-based rather than compression-based. That is a statement about the mechanism, not a claim about our numbers. Asking us for a calculation on your configuration is the right move; asking us to rank a sliding door against a casement window on one number is not.

Is a lower K value always the right target?

No. In a cooling-dominated climate the solar factor frequently dominates, which is why four elevations with one glass build is usually the wrong answer. The published CPD guidance on aluminium systems makes the same point from the other direction: on a bright south elevation, a window chosen to maximise solar gain can make the room worse, and a low U-value may serve you better than a high energy rating.

Do the 240 hour salt spray and 5,000 N shear results mean anything here?

They mean exactly what they say and nothing more. They belong to AW90, they measure coating corrosion resistance and the shear strength of the bond between the polyamide strip and the aluminium, and they do not transfer to any other series. Neither is an air, water, thermal or structural performance rating, and we will not let them stand in for one.

Will better windows let me install a smaller heating plant?

Possibly, and the degree-day calculation will tell you by how much if you feed it honest inputs. But remember the balance point: as the envelope improves, the building needs heating at a lower outside temperature, so the standard base temperature progressively overstates what further improvement is worth. This is the rare case where the honest answer is smaller than the marketing answer.

Who actually calculates the building's energy use?

Your energy consultant or your MEP engineer. What a factory can give you is evidence about a specimen: profile sections, chamber geometry, strip material and width, glass build, test reports and calculations on a defined configuration. What a factory cannot give you is a number for your building, because a building number needs your weather, your orientation, your plant and your baseline. We supply the first kind and decline the second. You can see the sort of projects we have supplied on our villa projects in Saudi Arabia, Dubai and Australia.

The short version

• "Energy saving" is three different things wearing one word: conduction, air leakage, and sun. Only the first appears in a catalogue.
• Thermal transmittance is the one characteristic in EN 14351-1 with no class ladder. There is only a number, so there is no band to catch a number that has moved.
• Air and conduction are both linear in temperature difference, so their ratio does not depend on your climate. About seven cubic metres an hour carries the same heat as conduction through a 1.8 square metre window at K 1.3.
• Hold the K value fixed and change only the seals, and the air term moves by roughly a factor of nine. The printed number never moves.
• EN 12207 has two denominators and one test pressure. The same unit tested at 75 Pa reads about five sixths of its 100 Pa value, honestly.
• A K value is a rate. Turning it into a saving takes four inputs, and a factory holds less than one of them.
• As a rule of thumb from the degree-day method: one W/(m²K) of improvement is about 48 kWh per square metre per year of delivered heat in a 2,000 degree-day season, before plant efficiency.
• The better your envelope gets, the more the conventional degree-day base overstates the next improvement. That cuts against us and we would rather you heard it from us.
• A thermal break lengthens the path, it does not remove the aluminium. A strip width is a detour length, not an insulation thickness.
• A window energy rating folds U, g and leakage into one letter, and the same letter is opposite advice on two elevations. It is a comparison tool, not a project answer.
• A door is not a window: the reference area grows by about 47 percent while the perimeter grows by about 26 percent, so the frame-to-glass fraction moves and the value moves with it.
• Ask for a value on your opening. If you cannot get one, you do not have a specification, you have a family name.