The Earth Ad­vant­age En­er­gy Per­form­ance Score Pi­lot Pro­ject for the En­er­gy Trust of Ore­gon published in 2009 found that the most commonly used energy models were wrong as much as 96.6% of the time, regularly overstating energy savings.

Testing and Rating Windows

Testing & Rating Windows

Your windows occupy an environment filled with air convection. Room air convection moves heat to your windows. Air convection between glass panes helps pass the heat through your windows to the outside air, and wind whisks heat away from your windows into the great beyond to add your tiny contribution to the problem of global warming. This is not, however, the environment in which windows are tested and rated for thermal performance.

Heat moves through your windows in three ways: radiation, conduction, and convection. All of these processes operate all the time and at the same time.

Of the three, convection is the most important, followed closely by conduction. Radiation in net heating climates is a distant third.

Unfortunately, because windows are mostly glass, there is not much that can be done to combat any of these thermal transfer processes.

Glass is a good conductor and a poor insulator, and since glass makes up 90% of most windows, keeping heat from transferring through your windows is ultimately a losing battle. Heat flow can be slowed, but not stopped.

The very best "super-insulated" windows barely reach R-9 (using three panes, a super-insulated frame, low-E coatings, and Argon or Krypton fill gas) compared to a minimum R-19 recommended for your walls and R-48+ in your ceilings. And, these are windows that start at about $4,500+ per window. A more typical thermal window tops out at R-2 to R-3 for around $1,200.00.

So, to heat, even a well-insulated window is nothing more than a gaping hole in your house's insulation. (But, there are exceptions. See: Windows as Heat Sources?, elsewhere on this page.)


Radiation is the movement of heat in electromagnetic waves. It is how the sun's heat gets to Earth through the near-vacuum of space.

Thermal Conductivity of Common Materials

Conductivity is a measure of the speed or rate at which heat moves through a material.

This is more complex than it sounds since conductivity is affected by factors such as temperature and the atomic-level composition of the material.

Ordinary carbon, for example, is not a particularly good heat conductor, but Graphene, a slightly different physical form (or allotrope) of carbon, is the most conductive material so far discovered.

A material being tested is held at a specified uniform temperature to make results as universal as possible. Typically the lab measures the amount of heat passing through one square foot of the material one inch thick in one hour.

The result of the test is a value called a Thermal Transfer Coefficient (often abbreviated W/mk and informally called a K-value) which reflects the general conductivity of the material.

If heat travels very slowly, very little will pass through the material in one hour. If quickly, more heat will get through. The higher the number, therefore, the more conductive the material.

R-value and K-value are related. R-val­ue is a measure of how much time is required for heat to move through a material from one side to the other. Obviously, it takes less time for heat to travel through a highly conductive material, but it also takes less time to travel through a thinner material than a thicker material.

Consider K-value to be the speed of a car and R-value the time required to travel a specific distance.

Traveling at a constant 30 mph (K-Value), a car would travel 1 mile in 2 minutes (R-Value). At 10 mph, it would take 6 minutes. K-value is unaffected by distance. The speed or rate at which heat travels through a material is always the same no matter how far heat travels.

R-value is affected by distance. It increases as the heat travels farther. It is also affected by speed. It takes heat a considerable amount of time to pass through one inch of a low-conductivity material like aerogel (R-10), but almost no time at all to get through aluminum (R-0.61), a highly conductive material through which heat travels very quickly.

Material K-Value
Aerogel (silica)0.001
Air (gas) 0.004
Aluminum 44.028
Argon (gas) 0.003
Fiberglass insulation 0.008
Glass (window)0.169
Graphene (The most conductive material)600.000
Gypsum drywall0.030
Hardwood (oak, maple)0.028
Krypton (gas) 0.001
Mineral wool (insulation)0.007
Plaster (wood lath)0.049
Softwoods (fir, pine)0.021
Vinyl (PVC) 0.034

Anything warm, even air, radiates some heat. If it is very warm, it radiates a lot of heat. Put your hand in front of a radiant heater, and you can feel the radiated heat. Put your hand in front of a warm wall and you won't feel any heat. Your wall is, in fact, radiating heat, just not enough for you to feel.

While radiation operates all the time, summer and winter. During the winter, however, it is not a prime mover of heat out of your house.

The only things radiating in your house are your walls, furnishings, and the room air — not much radiation — so there is not much radiation going out of your windows (unless you happen to have a radiator right under a window, then there will be more). But, there is some, and window engineers combat this form of heat loss using low-E (for "low-emissivity") coatings.

Low-E coatings are metal or metal oxides thinly deposited on the surface of the window glass. They reflect thermal radiation but allow light waves to pass through, so you can still see out.

Properly used, they reflect back most of the long waves of heat radiation that originate inside your house but do not reflect as much of the short UV waves that come from the sun.

This effect allows sun rays to come into the room, warming it up but does not allow long-wave radiation originating in your house to escape, helping retain heat in your house.

In summer it is a different story. Radiation from the sun is a prime source of heat gain in the summer. In net cooling climates, like most of the American South, a different configuration of low-E coatings is used to keep heat-producing UV radiation out of your house but still allow light to come in.

Unfortunately, low-E coatings are not tuneable. They cannot be designed to let heat in the winter but keep it out in the summer.

In climates that have both a hot summer and a cold winter, you have to pick one of the other. In Nebraska, we usually opt for allowing radiation in to help combat winter cold, although as global warming gets worse, we may revisit that decision.


Conduction is the movement of heat on a microscopic level from molecule to molecule inside a material.

Heat one end of a steel bar with a propane torch and soon the other end is too hot to touch. Heat moved through the metal by conduction until it reached the other end of the bar.

To reduce heat loss by conduction, materials that do not conduct heat very well are used. Most insulation, for example, is made of low-conductivity materials.

Almost Science Fiction

Aptly nicknamed "frozen smoke," aerogel is a solid material so lightweight that it is just slightly heavier than air. Invented in 1931 by chemical engineer Samual Kistler to win a bar bet, it is, despite its name, not a gel,

but a dry solid that feels more or less like Styrofoam®.

Aerogel is almost all air: millions of tiny cells filled with air. This makes aerogel an excellent insulator. Aerogel produced in a laboratory setting can have an R-value of up to R-30 per inch. — 10 times greater than fiberglass insulation. Aerogel now in commercial production has a much lower R-value, about R-10 per inch. — still higher than any other form of insulation other than a pure vacuum (R-45/in).

It would be a great insulator in windows if it were transparent. It's not. It's translucent but not transparent. And, while silicon aerogel is used in premium skylights, it is not yet suitable where clear viewing is needed.

It's also very expensive. Its production is slow and costly, although recent advances in production methods will probably bring the cost down soon.

There is little question that eventually researchers will come up with a transparent aerogel — they're pretty close now. In fact, a company called Aspen Aerogel has already produced a modified aerogel that in thickness of 1/2" or less is as transparent as ordinary window glass, but it's still in the largely experimental stage.

Sandwich a thin slice of transparent aerogel between two pieces of glass, and you have super-insulated window glazing. A mere 1/4th inch of aerogel could potentially yield a windowpane with an R-value of 12.5, which is approaching the R-value of a 4" wall.

This could make windows net heat sources in winter. Most house windows would take in more heat from solar radiation than they would lose. Large window walls eventually could be a building's primary source of heat.

Initially, aerogel windows will be very expensive, but, like early LCD television screens, the price will come down fairly quickly as new production methods come online and competition intensifies.

But, don't look for it in your neighborhood window store. It won't be there for a few years yet.

Unfortunately for window engineers, the primary material in a window is glass, and glass is not a low-conductive material. If it were, then the simple solution to insulating windows would be to install thicker glass.

But, while glass is not as good a conductor as most metals, it's good enough to move heat right along.

There is not much that can be done about it until science develops a less-conductive glazing material. Such materials are in the experimental stage, but none is quite ready for widespread commercial use (See "Aerogel: Almost Science Fiction", this page).

Heat also moves through window framing materials by conduction. This is where today's efforts to reduce conductivity are concentrated, simply because it is an area in which something can be done.

Most of today's window frames are very insulative, but this has little effect on the overall conductivity of the window since non-glass parts make up no more than 20% of a typical window.


Convection is the movement of molecules within fluids (i.e. liquids and gases).

Convection does not take place in solids because solids do not flow, nor in a vacuum, because there are no molecules to move around. But, in gases, like air, molecules move freely and with that movement transfer heat and cold.

Most of the literature on controlling the movement of heat into and out of your house concentrates on air moving through gaps in your building envelope. But convection is actually more complex. It is not just one, but a series of four processes.

Air Infiltration & Exfiltration

Air moving out of your house can be a major cause of heat loss, and not just through your windows.

Warm air escapes from your house and cold air gets into your house through tiny cracks and gaps in walls and roof, and through gaps in and alongside your windows.

A lot of homeowners are motivated to replace their old windows because they are "drafty". The proper cure for drafts is weatherization, not window replacement.

Weatherization is simple and costs little — certainly much less than replacing your windows, which is akin to buying a new car because the old one has a flat tire.

Room Air Convection

To get through your windows heat must first be moved to your windows.

Most heat moves to your windows piggy-backed on moving air.

Cool window glass attracts air convection currents like ants to honey. Convection currents do their best to move all the heat in the room to your window glass.

Some heat is lost through the walls and ceilings, of course, but, according to Department of Energy estimates, [1] in a well-insulated house as much as 25% of your total heat loss is through windows.

A 2000 sq/foot house has about 3440 square feet of walls and ceilings and about 200 square feet of windows. Windows comprise just 6% of the building envelope.

If six percent of your building envelope loses 25% of your heat, you can see how weak the thermal protection of a window truly is. Heat views your windows are just large holes in your wall's insulation. And, if the windows are the easiest way out of your house, that's the path heat will take.

Your room air molecules give up heat to the glass in your windows by bumping up against the glass molecules. Then the heat moves, molecule by molecule, by conduction, through the glass where it is again in contact with air. The journey does not take very long. Window glass is only about 1/8" to 3/16" thick.

How to slow heat transfer through glass is one of the perplexing problems confronting window engineers. It is not possible to alter the composition of glass making it a better insulator without affecting transparency, the property of glass that makes it useful in windows.

All of the current solutions invo;ve using two (or more) panes of glass with some sort of insulating material sandwiched between them. The usual sandwich filler at the moment is air. It's not a particularly good insulator, but better than glass. A dual pane window/air sandwich has an R-value of about 2.2.

Inter-Pane Air Convection

The air trapped between two panes of glass is often described as "dead air", but it is anything but dead. It is, in fact, a very lively convection current that draws heat from the inside pane of glass and conveys it to the outside pane.

The air touching the warm inside pane picks up some heat and starts to rise — warm air, as you know, rises. It soon reaches the top of the window sash where it cannot rise any further.

Eventually, however, pressure builds and it gets crowded against the colder outside pane. It gives up some heat.

Now colder and heavier, it falls, and soon there is a constant convection current moving inside the glass – down the outside glass and up the inside glass – conveying heat from the warm inside glass to the cold outside glass.

The process is, unfortunately, very efficient and ruthlessly unstoppable.

But, while convection cannot be stopped, it can be slowed. So, this is an area where window engineers try to retard heat flow.

One solution that has been tried is to eliminate most of the air between the panes, creating a partial vacuum. We know that vacuum is an excellent insulator — in fact, one of the best insulators, rated at R-45 per inch. Neither convection nor conduction works in a vacuum.

The problem is that so far no one has been able to create a vacuum in a glazing system that will hold up over time.

The extremes of climate and the brutal environment in which windows live always defeat the vacuum seal in the end, usually in fairly short order.

Every year there is a new technique with the promise of solving the vacuum problem. And, every year it turns out that it also does not work in the long term.

Fill Gases

So, the solution most often adopted by window manufacturers at present is to replace the air between the glass panes with a heavier gas such as Argon or Krypton (which has nothing to do with Superman® — that's Kryptonite, not Krypton).

Heavy gases are more and flow more slowly. They are also less conductive. Heavy gases can greatly slow down convection currents and reduce conduction, slowing heat transfer by as much as 50%, achieving R-values of up to R-4.

Unfortunately, however, the gases are not permanent. They leak out. Not necessarily because of seal failure (although seal failure is a fairly common problem), but because the gas molecules simply work their way through the seal material, while air molecules, in turn, work their way in. Once gone, the gases cannot be renewed. They're gone for good.

Seals have certainly improved since gas-filled IGUs were introduced over 40 years ago. But, they still permit gas exchange. Window companies say they have it under control, and that leakage is down to as little as 1% per year. [2]

That's what they say, but how confident are they in what they say?

Their confidence can be judged by the fact that not one single manufacturer guarantees its fill gas against leaking. Not a single one.

Fill Gases Are Guaranteed to Leak Out

No manufacturer guaranties fill gases will not leak. They know full well it will leak out over time. This warranty language from Milgard Windows — which offers one of the best warranties in the business — is typical:

"For Milgard Products with argon or krypton gas-filled insulating glass, Milgard injects the gas at the time of manufacture. The gradual dissipation of the gas may occur naturally over time and is not a defect. Other than gas loss due to seal failures, this warranty does not cover the gradual dissipation of inert gas or the amount of inert gas remaining in the Milgard Products at any time after manufacture."

The only guarantee is that they will not leak out suddenly, indicating seal failure.

The same laboratory study that showed a leakage rate of "as little as" 1% per year also showed a possible leakage rate of "up to 18%" per year.

At that rate, you would lose half of your fill gas in just 32 months. So that extra $5,000 you spent for a house full of Krypton-filled windows bought you nothing more than a temporary boost in your windows' energy performance.

Atmospheric Convection

Eventually, fill gas or no fill gas, heat will reach the outer glass pane. Conduction takes over again to move it through the glass, and now (unless this is a triple pane window, (in which case you need to go back a few paragraphs and start over) the heat is outside.

The heat jumps from the window glass to the outside air through conduction, bumping molecules, until the heat reaches the outside of the glass where it heats up the outdside air molecules pressed against the glass.

The heated air molecules don't just stayed put, pressed against the glass. , If they did, heat transfer would be minimal. But, they don't.

The culprit is atmospheric convection. They are immediately whisked away by atmospheric convection currents, or, what we usually call the "wind", and new cold air molecules move in to take their place.

Wind is constant. It may be just a mild breeze, or it may be a tornado. Even if you can't feel it on those hot, muggy summer days, it's still moving, if only a little. In winter, these air currents keep moving cold air molecules against your warmer window glass where they keep drawing heat out of your window.

It's a never-ending process, or, more accurately, it will end someday, only after the sun explodes or the Earth just runs out of air.

Heat Movement and Window Testing

The complexity of heat movement through windows makes accurate models of the process very difficult. And, the inaccuracy of the models is reflected in the testing process.

The "Hotbox" Window Test Environment

Does this look even remotely like your house?

In this “hotbox” testing environment approved by the National Fenestration Research Council, air convection is almost eliminated.

The heat source is a blazing radiant heater positioned just behind the window rather than warm room air and minor radiation from walls, ceilings, and furnishings.

This very artificial testing environment is a far cry from the real-world environment of the windows in your home.

Windows are tested in a very artificial environment that does not even remotely resemble your home — the place where the windows will actually be used.

For testing in a laboratory, a window is placed between a hot plate and a cold plate inside a tightly sealed, environmentally controlled chamber, colloquially known as a "hotbox".

One side of a window is heated. The window's thermal performance is then estimated based on how long it takes for the temperature on the non-heated side of the window to equal the temperature on the heated side.

After a few calculations on the computer, the result is then stated as the U-value of the window.

This method of testing heat transfer is well-established and reliable when it comes to testing simple materials.

For more complicated assemblies like windows, it is less reliable because it eliminates much of the complexity that exists in the real world.

In testing windows, the two most important simplifications are

The Elimination of Convection

While it may seem illogical to eliminate a process known to contribute substantially to heat loss through windows, there is a reason. Convection is much too hard to control, harder yet to standardize in a test environment. It's even very hard to model.

What is a U-Value?

Windows are not rated by R-values like every other insulation product in the U.S.


Windows use an obscure measure called a U-value or U-factor, based on a testing and rating protocol established by the Na­tion­al Fen­es­tra­tion Re­search Coun­cil, a window-in­dus­try-spon­sored association.

U-value is not a measure of how well a material insulates. It is, in fact, the opposite. It is a measure of how well a material transfers heat.

A material with a high U-value permits a lot of heat flow. A steel block, for example, has a relatively high U-26.2. A material with a low U-value transmits little heat. A block of Sty­ro­foam® has a U-value of just U-0.15.

Originally all insulation materials were rated using U-values. Then insulation makers realized that showing resistance to heat flow rather than heat flow itself would better help the public understand insulation effectiveness. So they created the R-value rating in use today in North America. R-value is nothing more than the inverse or reciprocal of a U-value rating.

U-value rating of a window is more understandable once it is translated into the corresponding R-value.

The calculation is easy, just divide 1 by the U-value. For example, a double-pane thermal window typically scores about U-0.45, which converts to an R-value of 1÷0.45 or R-2.2.

How much insulation is this? Your house walls without any insulation at all are about R-3.2, and with insulation are R-11 to R-19. So, it's not much insulation.

Why do window companies use U-Value rather than the more familiar R-Value?

Simple, U-values are less well understood and cannot be directly compared to R-Values, so they can be made to sound impressive.

One young and enthusiastic window salesman told us that his company's windows were rated U-0.45 which meant that they "allowed only 4.5% of the heat to pass through." We explained to him what the U-value actually meant before we gave him the bum's rush.

Thermal window U-values when translated to R-values, are anything but dazzling. As far as heat is concerned, a window with U-0.45 or R-2.2 is just a large thermal hole in your otherwise well-insulated R-19 wall.

For more information on R-values and U-values, see Will the Real R-Value Please Stand Up?.

Climate scientists have been trying for decades to develop an adequate model with only marginal success even when using supercomputers to do the calculations.

A successful testing process must be relatively simple, inexpensive, and easily replicated not only from window to window but from lab to lab.

Air convection is far too complex. It is almost impossible to duplicate air convection currents reliably from one test environment to another. So the solution adopted by the testing protocol is to eliminate convection from the tests as much as possible.

Single Heat Source

Another simplification used in laboratory testing is the heat source.

In your home, the heat source is complex and multi-faceted: Most of the heat lost through your windows comes from warm room air. But, your inside walls, floor, ceiling, and furniture contribute a little radiant heat (not much, but some) and a radiator or hot air register inside the room could contribute quite a bit.

In the laboratory, the sole and only heat source is a bank of radiant heaters aimed at the window.

Radiant heat can be precisely controlled and greatly simplifies measurement, so it is the standard for laboratory testing. But, it over-produces the one heat transfer process, radiation, which in the real world of window heat loss, is the least important process. In the lab world, however, it is the most important.

Windows Designed for the Laboratory, Not for Your House

As a consequence the the compromises made to enable laboratory testing, test results for windows are inaccurate and very skewed.

Lab tests have concluded that about 70% of the heat loss through your windows is by radiation. Conduction and convection account for only 30% or so. This is true, however, only in the lab environment in which almost all air has been eliminated and radiant heaters are aimed directly at the inside of the windows. How close is this to the actual environment of your house?

The real-world effect of this elimination, however, is that the testing model is an unreliable predictor of how windows will behave in the real world full of air and convection currents.

The Earth Ad­vant­age En­er­gy Per­form­ance Score Pi­lot Pro­ject for the En­er­gy Trust of Ore­gon published in 2009 found that the most commonly used energy models were wrong as much as 96.6% of the time, regularly overstating energy savings.

These inaccuracies produce unfortunate results.

The most detrimental is that they affects how windows are designed and built.

They are made to test well, not work well. Window manufacturers tend to build windows that score high in the test environment but do not necessarily perform well in the real world.

For example, most window manufacturers place a lot of emphasis on their low-E coatings to block radiation. Yet, outside the laboratory setting, radiation plays only a small role in heat loss.

Field studies have shown that low-E coatings have very minimal effect in winter and just a modest effect in keeping our houses cool in summer.

But, in lab tests, windows with better low-E coatings score well because the sole heat source is a blazing radiant heater, so window companies emphasize low-E coatings rather than working on measures to reduce convection and conduction, which are the process by which most of your heat is lost in the real world.

So, a low U-value rating for a window does not tell you how well the window will perform in your house, UNLESS, your house is a hotbox chamber with most of the air eliminated and a blasting radiant heater located just behind the window.

If your house is more conventionally arranged, say with furniture, carpets, and other typical house-type stuff, heated by forced air or a room radiator, the U-value tells you almost nothing useful.

A test of window performance in a normal house environment has not been done.

Window manufacturers are not the least bit interested in showing that their windows do not perform in the real world as well as advertised, and no government agency seems to have been aroused enough to do a comprehensive formal field study, although this would seem to be right in the Department of Energy's bailiwick.

From the limited field studies that have been done, however, we know that actual thermal window performance is well below that predicted by U-value ratings.

There is plenty of evidence that properly restored old wood windows with storms perform at least as well as new thermal windows, and in the long run, as seals start to leak and the fills and coatings that temporarily boost new window thermal performance start to degrade, restored old windows may perform better.

Yet, in lab tests, old windows generally perform poorly.

Restoring Old Windows

There are substantial differences in the design and manufacturing of modern windows compared to traditional wood windows.

The important difference is that modern windows are designed and made for energy efficiency above all other considerations. Old wood windows were designed for longevity, endurance, durability, and ease of repair and restoration.

The oldest still working window in North America dates from the 1600s. Old windows can be restored and restored again to preserve and extend their longevity while improving energy performance.

How this is done is the subject of the next section, Restoring Your Old Wood Windows.

1. U.S. Department of Energy (2005).

2. The study most frequently quoted to support the claim that gas leakage is at less than 1% per year is Plavecsky, J. "Leaking Out the Facts", Door and Window Maker, Sept-Nov, 2000. However, that's not precisely what the study concluded. Plavecsky found that the amount of leaking was determined by quality control in the factory, the method of inserting the gas, the type of seal used, and the care with which the seal was installed. He tested seals in a laboratory environment using a process of "accelerated aging....said to represent five years of normal field exposure". Leaking was found to be as high as 18% and as low as 0.9% per year over a simulated five-year period. Quite a range and there is no way for a consumer to judge whether a replacement window is at the high or low end of this leakage rate.

Rev. 08/17/23