GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Guest Blogs

Comparing North American Window Frames to European Frames

Thin North American window frames with a high U-factor sometimes perform about the same as European window frames with a lower U-factor

Image 1 of 3
This European window has a frame with a low U-factor. While impressively constructed, a disadvantage of this window design is that the frame is so wide that it limits the amount of solar gain that the window can admit. Windows with narrower frames have a higher percentage of glazing (by area) than windows with bulky frames.
Image Credit: Image #1: Makrowin
This European window has a frame with a low U-factor. While impressively constructed, a disadvantage of this window design is that the frame is so wide that it limits the amount of solar gain that the window can admit. Windows with narrower frames have a higher percentage of glazing (by area) than windows with bulky frames.
Image Credit: Image #1: Makrowin
This illustration shows the three types of operable window frames that were entered into PHPP for the exercise described in the article.
Image Credit: Images #2 and #3: Stephen Thwaites
This illustration shows the three types of fixed window frames that were entered into PHPP for the exercise described in the article.

[Editor’s note: The author of this article, Stephen Thwaites, is a window manufacturer. His company, Thermotech Fiberglass Fenestration, is located in Ottawa, Ontario.]

Presumptive European Superiority Syndrome applies not just to a window’s glass but its frame as well. After all, many European Passivhaus window frames have lots of insulation in them — so they must be better than North American frames, which are at best less well insulated, but more likely not insulated at all — right?

I may be biased, but I’m not so sure. Let’s take a closer look.

You need to know more than the U-factor of the window

The Passivhaus system for rating windows emphasizes insulating ability over solar gains. But the Passivhaus approach — choosing windows based on a lower U-factor, as long as solar heat gain coefficient (SHGC) of the glass is greater than 0.5 — does not always result in the lowest energy bills.

Recognizing the flaw, the Passivhaus Institut introduced a letter rating scheme that rewards slimmer frames. (For an explanation, see page 3 of this document.) Slimmer frames allow greater solar gain than bulkier frames, which reduces heating loads, making it easier to achieve the Passivhaus heating load target. While the letter rating scheme does help, it still doesn’t always lead to the lowest energy bills.

Typically, European Passivhaus window frames, even those with an “A” rating, are bulkier than North America’s most energy-efficient windows.

Window frames do not contribute solar gains

A window frame, no matter how well insulated, will never be as energy-efficient as the triple glazing that it surrounds. That’s because, unlike the glass which contributes solar gain during the heating season, the frame cannot gain energy; it can only lose energy. So, while windows with bulky frames typically have frames with a lower U-factor than windows with a slimmer frame, they also always contribute less solar gain than windows with slimmer frames.

Some common European Passivhaus window frames are almost 5 inches (125 mm) wide, but most are more like 4 inches (100 mm) wide. Even A-rated European frames frames are still typically 3 1/4 inches (85 mm) wide. And they are often that wide for both operable and fixed windows.

[Editor’s note: In this article, the word “width” is used to describe one of the measurements of a window frame, as shown in the illustration at left. Readers should be aware that some writers use the word “height” to describe the same dimension.]

Narrower frames let in more sunlight

North American outswing windows usually have frames that are about 2 3/4 inches (70 mm) wide. This is an advantage over European Passivhaus windows. On smaller windows it can be a significant advantage. For a 24″x35″ window, a 4-inch-wide frame leaves you with a window that is only 50% glass. A slimmer 2 3/4-inch-wide frame results in a window that is 63% glass for the same size of window.

More significantly, North American fixed frames are about 2 inches (50 mm) wide – providing noticeably more glass area than their more bulkily framed European counterparts. For a 48″ x 48″ fixed window, a European “A” rated window with a 3 1/4 inch (85 mm) wide frame is 77% glass. The same window with a 2-inch (50 mm) wide frame is 84% glass.

Narrower frames can lose less heat

When comparing heat loss through window frames (unlike when making many comparative heat loss calculations), area is a variable, not a constant. Despite insulating more poorly, a narrower frame may lose less heat.

For example, to have the same heat loss as a 4-inch-wide R-10 frame, a 2-inch-wide frame would only need to be R-5. So narrower frames have several interesting advantages.

Narrower frames are not the whole story

To be very clear, just as there is more to energy-efficient cars than tire pressure, there is more to energy-efficient windows than frame width.

It certainly possible to have a long and tortured discussion about the about the many and often conflicting factors that affect window energy performance. But once again, the best way to assess “better” is not to argue over more beer, but to run PHPP.

Modeling a house with PHPP

So I ran PHPP. I compared the performance of windows with different frames at the same house in Lancaster, New Hampshire, that I used when I ran glazing comparisons. (More information about the house can be found here.)

I took the PHPP spreadsheet for the house and kept everything constant except for the characteristics of the window frames.

I compared three frames that I thought might be representative: two European Passivhaus frames and a North American insulated fiberglass frame (see Images #2 and #3, below). The first European frame is based on a common PVC frame geometry. It is a time-tested design, updated with multiple insulating chambers. The second European frame is based on a newer, more advanced PHI-listed “A”-rated (slimmer) frame. Not as widely available, this ‘masked’ sash approach, while tough to screen, is designed to be energy-efficient. The North American insulated fiberglass frame is based on a Passivhaus version of a 20-year-old insulated fiberglass window. While the operable frame is “stock,” the fixed frame is modified to accommodate a 1 7/8 inch thick (48 mm) IGU.

The results are in the table below.

There are lots of variables

Keep in mind that your mileage may vary. Your mileage will especially vary, sometimes wildly, depending on building design, climate, proportion of fixed windows, and of course a specific window’s characteristics.

The Problem With Software

In a comment responding to my glass-focused guest blog, Jin Kazama quite correctly wrote, “The problem with using software is the involved labor and complexity to get accurate results, and the number of iterations required to measure possibilities. Most regular folks do not possess the required knowledge and the interest for the simulations, nor the budget to have someone else do it.”I think he was responding to my repeated advice to use PHPP to compare windows. I thought that my response to his comment was best placed here because the rate at which I advise the use of PHPP is even higher in this blog than the previous one.For Canadians and others in similar heating climates, there is a simple/crude tool to rank residential windows by their effect on heating bills. It’s the ER (or the Energy Rating). The ER (a value displayed on window labels) is calculated by estimating heating season solar gains through a window, and then subtracting the estimated heating season losses. With few exceptions, simulations like PHPP or HOT2000 show that houses with higher ER windows will have lower heating bills. Like the similar systems in the UK and Denmark, the ER rating is intended for houses in heating climates. Most Canadian firms report the ER for their windows. If you’d like to calculate the ER for yourself, or recalculate it with CEN data, the formula is as follows:ER = (Gains – Losses ) + 40 = (72.2* 0.8 * SHGCwindow – 21.9 * Uwindow – 0.54 AL) + 40 where:72.2 is the average Canadian heating season solar gain on a vertical surface (average means average of the four orientations, and averaged over a 5064 hour / 211 day heating season for an average location in Canada (based on where people live, not geographical geometry) in W/m^2.0.8 is a reduction factor for shading.SHGCwindow is the SHGCglass, reduced by the light-blocking effect of the frame (some grade school math may be required)21.9 is the average heating season indoor/outdoor temperature difference for an average Canadian location in degrees Kelvin/Celsius.0.54 is an air leakage constant to convert test results into an annualized loss.AL is the tested air leakage in m^3/h/m^2 of window.40 is the marketer’s fudge factor (MFF) so even double-glazed windows can have a positive ER. For new hinged or fixed windows you can easily simplify the equation by making the air leakage term zero. You’ll lose maybe 0.5 ER of accuracy, which is an error of about 1-2% for most triple-glazed windows. (For reference, the average “weather coefficients” correspond fairly closely to the weather in Ottawa, Ontario.)People too horrified at the simplifications of the ER to touch it with a 10-foot (3.3 m) pole might consider the ERS, or Energy Rating – Specific. Specific means a specific orientation in a specific city. There is off-the-shelf ERS data for about a dozen Canadian cities. As an example: for a south-facing window in Montreal, Quebec:ERSsouth = (112 * 0.8 * SHGCwindow – 22.7 * Uwindow – 0.54 AL) + 40Other Montreal orientations use other values for the solar gain; E/W = 60 W/m^2; N = 34 W/m^2.If you were so inclined you could compare your weather to Montreal’s and create your own version of ERS to rank window choices for your location. If you were so inclined.For more accuracy, without the input effort of PHPP or HOT2000, there’s always LBNL’s RESFEN or NRCan’s HOT2XP.So there are ways to compare windows that do not involve complex software.

The best bet to compare windows is to run PHPP for your building.

PHPP predicts that this house, as built, will use 4.5 kBTU/ft^2/yr (14.2 kWh/m^2/yr). Note that the common European PVC frame’s bulk would have increased the heating requirements of the house by 9%. This increase is enough to put the house over the 4.75 kBTU/ft^2/yr (15.0 kWh/m^2/yr) qualification threshold for the Passivhaus standard.

For this house, the Passivhaus version of the North American insulated fiberglass frames produce a 14% lower specific heat demand than a common European PVC window.

The results also show the advantage of slimmer frames. Although the “A”-rated PHI-listed window frame insulates much better than the Passivhaus version of the insulated fiberglass frames, its extra bulk results in a nearly identical heating bill.

All three frames that we looked at in the example would be good choices from an energy point of view. According to PHPP, the spread between them for the Lancaster house is only about 450 kWh/year. For some people this may be the difference between qualifying for the Passivhaus standard and not qualifying for the Passivhaus standard. For others it is inconsequential.

Again, results like this are project-specific. Your mileage may vary; but in this case, a North American offering held its own when compared with windows that most people would assume were much more energy-efficient choices.

Again, the best bet to compare windows is to run PHPP for your building.

While European windows are terrific, until you’ve run PHPP for your building, there should be no automatic presumption of their thermal superiority.

Stephen Thwaites is a professional engineer and the technical director of Thermotech Fiberglass Fenestration in Ottawa, Ontario.


  1. AntonioO | | #1

    Are the NA windows actual or "theoretical"? In other words, is there a window company I could call up and spec the NA window you describe? Also, can you say more about the 48 mm thick IGUs you simulated--spacing and glass thicknesses and types, for example? Does Thermotech have such an offering? Thanks in advance.

  2. STEPHEN SHEEHY | | #2

    cost comparison
    I understand your point, but comparing a European UPVC window, a European clad wood window and a North American fiberglass window without giving any cost data doesn't help someone thinking about buying windows. Is there a NA UPVC window with specs similar to the Euro one? Can I get the same heat gain (assuming I want heat gain) by just buying a bigger Euro window? Will I pay more? Less?

    In case anyone's interested, I just got delivery of Intus UPVC windows for about $40 per square foot. One problem for those of us who are consumers, not building professionals, is that we must rely on builders, architects or friends to recommend products. We really can't all go through the process Stephen has outlined.

    I won't notice if my windows are better or worse than any other option. But if I can sit by the window when it's bitter cold and not need a hat, I'll be happy.

  3. Stephen Thwaites | | #3

    Availability? & 48mm glass
    Yes, the frames are real. We accomodate the 48mm glazing in the fixed frame with a pre-painted roomside wood glass stop. So that could be done by anyone.
    Another approach for those seeking a NA window with Euro-like performance would be using a regular thickness IGU and filling with Krypton. The Uframe would be a bit higher, but the Uglass would be a bit lower.

    In this blog, only the frame changed, the glass properties were the same for all 3 frame options. CEN procedures calculate Uframe without any glass in the glazing cavity. To calculate Uframe the glazing cavity is filled with an insulated block. The wider the glazing cavity, the greater the contribution of the insulated block, the lower the Uframe....

  4. Stephen Thwaites | | #4

    Costs et al
    I wanted to focus on physical principles, as discussion of them always seems to be in short supply. You are correct, price is an important part of any decision. Pricing is to some extent job specific. It can and does vary based on options like color, glass, auxilary profiles, order size and currency exchange rates to name a few factors. These factors make it hard to have a simple discussion about costs.

    I believe there are several Euro-like NA fabricated PVC windows. There is at least one on the PHIUS window list.

    Yes, if you want to match the solar gain of a slim framed window with a bulkier framed window, making the bulkier framed window appropriately larger will do the trick. The fixed window example in the blog was for a 48x48 window. With a 2" wide/high frame the glass area is 13.4 ft^2. A bulkier 3 1/4" wide/high framed window with the same glass area is 50x50. Since the 50x50 is at 17.7 ft^2 is about 10% bigger than the 48x48 window, it's probably 5-10% more expensive than the 48x48 fixed window. Also it's a minor point, but the larger window will result in slightly smaller wall area, slightly reducing the overall inslualting value of the envelope.

    Hope that helps

  5. Smagneron | | #5

    PH Comfort Criteria

    Great series of blogs on a much needed discussion topic. I like that Stephen Sheehy pointed out that increasing the size of your windows could be a strategy to overcome the wider Euro frame issue. But as you astutely pointed out, the wall’s overall effective thermal resistance would be reduced. A balancing act that would be performed in the PHPP! Another advantage to the wider Euro frames is that, if installed as an inbetweenie, it has the potential to be over insulated from the exterior, reducing the Psi-install thermal bridging affect of the frame. Something that is a little more difficult to achieve with slimmer NA window frames.

    The overall energy balance as a result of a window analysis in PHPP is fair enough. Slimmer frames with high SHGC will likely come out on top for Space Heating Demand (although will likely result in increased Overheating). However, there is the issue of comfort, which Passivhaus prides itself on. Radiation asymmetry, or mean temperature radiation should also be considered when choosing a window. That’s to say, a higher insulating window frame, the higher its interior surface temperature will be. The warmer this surface is, the more comfortable a person would be next to it on a cold winter’s day.
    The PH comfort criteria states that interior surface temperatures should not be more than 4.2C (7.2F) lower than the indoor temperature. The equation used to assess this is as follows:

    Interior Surface Temp = Room Temp – U-value of assembly * Interior surface film resistance * (Room Temp – Outdoor Temp)

    Let’s assume that we’re in Ottawa, where the PHPP design temp is around -20C (-4F), the Room Temp is 20C (68F) and the Interior surface film resistance is 0.13 m2K/W (R-0.23).

    Using your three examples, for the operable options, the results would be as follows:

    Classic Euro PVC 1.10 W/m2K (0.19 BTU/hr.ft2F) – 14.38C (57.9F)
    Newer European Clad Wood 0.71 W/m2K (0.12 BTU/hr.ft2F) – 16.37C (61.5F)
    NA Insulated Fiberglass 1.32 W/m2K (0.23 BTU/hr.ft2F) - 13.26C (55.9F)

    The only frame that passes the PH Comfort Criteria is the Newer European Clad Wood window frame. In fact, a window frame would have to have a Uf – 0.78 W/m2K (0.13 BTU/hr.ft2F) to pass the PH Comfort Criteria in Ottawa.

    This is great in theory, but I’m wondering if anyone has real life examples of discomfort next to windows due to frames that would not meet the PH Comfort Criteria in their location???

  6. AntonioO | | #6

    What's with all the Stephens?
    Just kidding.

    Seriously, thanks Stephen Magneron for providing some kind of comfort metric. In lieu of the impossible idea of having someone set up a display where consumer selectable windows are placed in front of a freezer to see what it feels like to sit in front of them on a cold day, this is probably the best I can expect to get. Just a question or two: How does one know the surface film [heat] resistance for a material? Also, intuitively, I would think that number would itself be at least somewhat temperature dependent. Would my intuition be incorrect or incorrect? Thanks in advance.

    BTW, I've been trying to determine if any of these numbers pertaining to energy performance are ever measured or if they are all simulations. At the very least are the simulations verified by a real world measurement from time to time?

    Also, not to be a nit picker but a temp delta of 7.2 degrees F is equivalent to a delta of 4.0 degrees C, not 4.2. I might try putting a thermocouple on the frame of my current lousy windows to see how much the temp differs from the room air temp. I bet it's much bigger than 4 degrees C. If anyone out there has great windows installed in their homes (or works for a window manufacturer), please do the same and let us know what you find (especially if you have the kind of simulations that would allow you to compare your measurement to your simulated value).

  7. AntonioO | | #7

    reply to Stephen (Thwaites that is) post 3
    Is there a reason why it is that when I go through the IGU offerings posted on the websites of NA window companies, I see no mention of a 48mm IGU. (For the record I've found the same for some European companies.) A couple of your NA competitors boast how they have frames that were built for the thicker, high performance glazing units--not modified, blah, blah, blah. However, when you look at the specs of what they're selling, it's basically the same that everyone else is--tri-pane units in the mid to upper 30s millimeters thick. Must I call them all up and ask for an IGU to my specs?

    Also, as for the argon to krypton change, one then has to ask how does the cost and performance of 18 mm of argon compare to the cost and performance of 12 mm of krypton for a window specified the same in every other way.

  8. jackofalltrades777 | | #8

    Why No Full Disclosure?
    Why is there no disclosure at the BEGINNING that the author of this article works for Thermotech Fiberglass out of Ottawa, Ontario? I am not talking about a two sentence blip at the bottom of the article that once you read the entire article you realize that it was written by someone working for a window company and the entire article had a bias to it.

    That's like reading an article claiming Chevy cars are superior but hiding the fact until the end of the article that the author of article works for GM.

    Also, it would be nice to have a disclosure that Thermotech had an extermely large amount of window seal failures that came from a production run that lasted over 5 years. So thousands of windows went out that experienced seal failures.

  9. GBA Editor
    Martin Holladay | | #9

    Response to Peter L
    Stephen Thwaites is fairly well known, and no attempt to deceive was intended on the part of GBA.

    That said, you make a fair point. I have added an "Editor's note" at the beginning of Stephen's article to address your concern.

  10. Bronwyn Barry | | #10

    Selective Comparison
    Thanks for another great post, Stephen T. However, I have to support Stephen M.’s comment, highlighting the window performance criteria that have been omitted (or perhaps ‘fudged’ a bit?) Without including them this is either a selective comparison or you might be saving the full comparison for a future post (?)

    Stephen M. covers both the comfort criterion and the installation psi-values that are important for overall performance. He misses an additional component to the comfort criteria described by PHI as the ‘hygiene requirement,’ which I interpret as a durability item. It’s an insurance that condensation will not occur at the glass edge on the interior of the frame, resulting in moisture damage either to the frame itself or in the case of PVC or fiberglass frames, to the sill and framing beneath where that condensation inevitably migrates.

    While this hygiene requirement is not a performance number per se, it would disqualify the thinner North American frames Stephen T. uses in his comparison. Not all North American frames though. There are actually three North American frames that have met all PH requirements and also achieve the highest efficiency class rating. Unfortunately only one is currently in production. It is the Synergist out of California, listed here: (For those needing metric translation, the Synergist’s IP frame u-value is 0.148 Btu/[hr.ft2.F] and width from base of frame to edge of glass is 3.8”.) The others are the Northwin and the CaliPassiv, both searchable on the PHI database. (Full disclosure: I designed the CaliPassiv profile.)

    As you repeatedly advise, the PHPP will reveal the full performance implications of specific window choices based on all their performance metrics and must include (a realistic) psi-install. I'm noting here that you’ve selectively chosen to share the performance comparison for the fixed window profiles only. The psi-install values you’ve used for your North American frames are optimistic at best. These could only be similar to the Run #2 frame psi-install values if you totally over-insulated the exterior frame. You can’t do that for the operable units, so you’re implying that all windows in this project are fixed. Interesting… If you were being more realistic, you should compare the operable frame units and include a realistic psi-install value, not these numbers which conveniently skew these results to support your argument.

    I’ve calculated that a beneficial psi-install can improve your Heating Demand number by 22%. Being able to take advantage of over-insulating a thicker frame and setting it into the middle of the wall insulation can significantly boost performance. For these reasons, frame u-value and IGU performance cannot be isolated and compared selectively. The most convincing argument against your North American inferiority will be a PHI certificate. It costs around $3K and does not require destructive testing or shipping of samples to Europe. Given these posts, I'm really surprised Thermotech has not qualified already…. Seriously, quit the whining and the selective data comparison and just get it done.

  11. GBA Editor
    Martin Holladay | | #11

    Response to Bronwyn Barry
    Two things struck me about your comments.

    First, I learned something: that the Germans who developed window standards for the Passivhaus Institut describe the need to control interior condensation as a "hygiene requirement." This is a particularly German way to describe condensation, don't you think? (When my family traveled by VW bus to Germany in 1963, when I was 8 years old, we were stopped for inspection at the border crossing from France. The German border guard scowled at the dust on the side of our vehicle. He reprimanded my father, telling him in no uncertain terms that "Auf Deutschland ist alles sauber.")

    Second, I notice that you assume that the desired pinnacle, the Mount Olympus, to which window manufacturers aspire is a certificate from the Passivhaus Institut. You may be right -- but it is an unproven assumption.

  12. Stephen Thwaites | | #12

    Over insulated Inbetweenie – Practical Considerations

    Stephen M wrote in reply #5
    “Another advantage to the wider Euro frames is that, if installed as an inbetweenie, it has the potential to be over insulated from the exterior, reducing the PSI-install thermal bridging affect of the frame. “
    I'm glad this got mentioned.
    My 'argue over more beer' answer is that it is tough enough already to get 'outie' windows installed properly with sloped subsills and unsealed bottom flanges. In my view properly installing an inbetweenie is 'a bridge too far' for all but the very very best installers. Additionally, and perhaps more importantly, the added cost of a more complex installation needs to be balanced by clients against the energy savings.....

  13. Stephen Thwaites | | #13

    Over Insulated Inbetweenie – Energy Considerations

    Stephen M wrote in reply #5
    “Another advantage to the wider Euro frames is that, if installed as an inbetweenie, it has the potential to be over insulated from the exterior, reducing the PSI-install thermal bridging affect of the frame. “

    Its true that masking the outside face of an inswing window reduces heat loss throught the frame. Based on PHI numbers for the Newer Euro Clad Window, the masking can add effectively reduce Uframe by up to 18%. This 'up to' needs to be reduced for at least 3 factors;

    1. Sills are not masked, because they need to drain water. For a square window this reduces the masking effect by 25%. (so 18% effective reduction in Uframe becomes 13.5%) For a window twice as tall as it is wide, the masking effect reduction is lower - 17% (so 18% effective reduction in Uframe becomes 15%)

    2. The thinner the wall cladding, the greater the porportion of the window's exterior face that can be covered with insulation, the greater the effect of masking. The dwgs on PHI's site seem to uniformly show acrylic stucco over foam. If you are cladding with other materials like wood , or cement board, or worst of all stone, the 'up to' will be reduced accordingly.

    3. Mullions are not masked so the extent of their presence will reduce the effect of any exterior overinsulation. (But they do eliminate the PSIinstall, so one mulled window in a larger RO does better in PHPP than the sum of its parts in two or more RO's)

    But this is still kind of an 'argue over more beer' answer.
    And maybe, Stephen M., you can help with some better quality answers. Being less familiar with PHPP, I did not attempt to run this level of detail through the Lancaster model. An experienced PHPPer, like Stephen M., can routinely handle these types of comparisons. Hopefully he has some time to throw some 'your mileage will vary' numbers at us.

  14. Stephen Thwaites | | #14

    PH Comfort Criteria and more
    Stephen M,
    I have the exactly same question about the PH Comfort Criteria as it applies to window frames.
    I have a hard time imagining a scenario where say a solid wood window frame would cause discomfort, if its glass and surrounding wall were of PH quality. Even if the non-compliant frame was thermally broken aluminum, it would be seem to be such a small contributor to either roomside natural convection or roomside radiation effects that it is hard to imagine it affecting comfort.
    But these are just semi-educated musings. If anyone has any references to research on the effect of relatively smaller area cooler surfaces to overall comfort, I'd love to read them.

    If I understand your comments correctly, then the Comfort/Hygiene Criteria is perhaps a condensation requirement. The challenge I would have with this approach to identifying condensation risk, is that condensation is due to a cold spot or spots. Uanything is an area average that hides the effect of cold spots.
    As an extreme example, consider the images below. The frame on the right would have a Uframe significantly better than the frame on the left. However, both frames would have the same potentional for condensation.

    So I agree that that the slimmer Fglass windows in the Blog will not meet cold climate PH comfort requirements. For some people this will be a 'deal breaker'. My point is that not all NA windows are necessarily less energy efficient, than Euro windows. and in fact, some in some cases, are more energy efficient.

  15. Stephen Thwaites | | #15

    Replies to Antonio's questions (#6 &#7)

    Film coefficients are set out in various standards, which in turn are based on peer reviewed research which is based on hyper accurate research grade experimetal equipment.

    My view, based the what I remember of round robin testing (test labs sent samples and report results, without knowing the other lab's resuts), simulations are as good or better than testing because of the potential for measurement error.

    Admittledly, this is not the view of NFRC. On the other hand, PHI, PHIUS and NRCan all accept results from accreditted simulators.

    The lower the rate of heat loss the less confidence I have in “production” laboratory test results.

    You are right, i've made it seem like wider 48mm IGU are more available than they are. Many smaller companies will be more receptive to customization at this time of year. So it never hurts to ask. Certainly the greater the perceived demand there is for thicker IGU, the sooner the market will accommodate.
    As an alternative to 18mm of argon, 12mm of krypton, while more expensive, does provide for a lower Uglass when calculated to the CEN stnd. – about 10% lower.

    Hope that helps

  16. Stephen Thwaites | | #16

    PSI (sigh) Confusion
    I've resubmitted the table to Martin to identify PSI as PSIspacer. I think Bronwyn has perhaps identified it as being PSIinstall'n. Upon further review it should've been more clear.

    All 3 cases in this Table used the same Uglass and PSIinstall'n from the final PHPP spreadsheet for the house. I did this not to present real windows, but to isolate components. I thought comparing NA frames with NA glass to Euro Frames and Euro glass might not highlight the independent effects of glass and frame.
    Bronwyn is correct real world comparisons for PH houses would also include installation effects. And certainly the NA fglass operable frames would not fare as well as the newer Euro Clad frame on that score. On the other hand, I think the NA fixed windows slimness might offset this difference. And in heating climates the higher gain NA glass can be very helpful in lowering energy consumption to a greater extent than Euro glass – especially on modestly glazed buildings.

    But this is my experience, which is better than nothing, but not better than running PHPP for each building.

    [Editor's note: See Stephen's revised table below.]


  17. jackofalltrades777 | | #17

    I Still Disagree with the Delta-T (and it rhymes!)
    I disagree that "0" is the better choice for the northern US. For the NFRC to use a delta-T of 70F, it takes into account a 0F outside temp and a 70F inside temp. I would like to see where in the Northern US does it average 0F? Nowhere but the Arctic Circle does it average 0F. Nowhere outside of northern Alaska where you might actually see a 70°F average delta-T during the heating season.

    The European delta-T is more realistic and the windows are designed with wider gaps between the window panes. A wider gap as found in European windows would see better energy benefits. NFRC calculations favor narrower spacing.

    Then you have the spacer technology. Most US manufacturers are still using metal (aluminum or stainless steel) for the spacers, while Euro manufacturers have warm edge spacers (Swiss design) which don't introduce metal into their spacers.

    As mentioned, pay the $5k and get the ThermoTech windows tested and PH certified if you really believe that they perform better than European windows. Talk is cheap. Run the third party tests and post the results.

    Also, while you are at it. Get someone in sales and marketing to clean up the BBB rating of ThermoTech because the "F" rating they currently have doesn't instill confidence in their product.

  18. Stephen Thwaites | | #18

    Third Party Data?

    at the risk of a being scolded by PHIers,

    now includes data from at least 2 FGlass window makers

  19. jinmtvt | | #19

    Peter L
    Average temperature is not what you are looking for since delta T increases with difference.

    At -20c there is much more energy loss than at 0c.
    We need windows that performs when it counts,
    not when it's 10c outside.

    Could we plase keep this BLOD CLEAN AND NOT DISCUSS Thermotech rating
    nor the fact that they do not sell PH rated windows.
    Does not have anything to do with current discussions.

    The BBB rating of his company does not mean he cannot discuss window performance here.

    Peter do you work for Fibertec? :p

  20. user-1081937 | | #20

    Little Bang Theory
    I have to admit I tend to glaze over (no pun intended) when reading about window component specifications and testing protocols, this is very nuanced territory and not for the timid like myself. If I get it vaguely right the argument presented is- yes our window frames are not as efficient as European frames but don't worry we make it up with more glass thus more solar heat gain.

    If this is correct then if I am designing in PHPP then to find my magic balance I simply have to up my window size, say 10% to achieve the same or a little more glass with my EU unit. This was clearly stated by Stephan S and I think is the entire point. I am not designing my building around my windows, I design my windows around my building. Even a uber passive solar Passive House is maybe 20% glazing on the south side so I have plenty of room to play with. Yes my EU frame is slightly larger but because the U factor is so much better I am betting it hardly nudges my kbtu/ft2.
    The larger the window the less effect and visa versa, a smaller window the larger the effect depending on just how lower the U factor is in the frame. As a designer I can handle this no problem and dollars to donuts (not sure what the means really) I bet the EU unit typically comes out on top.

    Then in the real world I have to account for the shading factor so the edge glazing typically has the least solar gain, especially the top. If I need to nudge my PHPP then I over insulate my frame which is painless with EIFS and only a little trouble with wood. Maybe its worth it, maybe not but I have the option. I know there is an extraordinary amount of physics for what is the most complex part of a high performance project- the window -which will continue to elude me but it important to see this in its application in a building and not isolated.

    Reading Martin's remark "Second, I notice that you assume that the desired pinnacle, the Mount Olympus, to which window manufacturers aspire is a certificate from the Passivhaus Institut. You may be right -- but it is an unproven assumption." The Far Side immediately came into my head.

  21. AntonioO | | #21

    Response to Stephen T (#15) ...Caution, caution
    Stephen T,

    Speaking as an experimental physicist, I would just say that a "presumptive inferiority" of measurement testing as opposed to simulations testing may be wrong headed. A well regarded theoretical physicist once told me that any theory or calculation is still only a mathematical exercise unless it conforms to or is confirmed by a real physical measurement or observation. Certainly we would hope that the simulation tools you are using to produce the numbers you are reporting have been so confirmed. The great value of a simulation in my opinion is that you can easily simulate many, many, many times the number of situations that you could ever hope to measure. You also state: "The lower the rate of heat loss the less confidence I have in “production” laboratory test results." One little warning: trends in measurement difficulty often correlate well with trends in simulation difficulty, meaning one would not be ill-advised to have less confidence in the simulations as well. The good thing about experiments, from my experience, is that ONE way (not the only way) you know when you get to the difficult to measure point is that different measurements produce different results--letting the observer know that the measurement confidence has decreased. Often times, another physical principal has crept over the boundary of your assumptions. On the other hand, everybody using a particular simulation tool or different tools based on the same prescriptive model will always get the same number for the same parameters and might assume from this that confidence has not decreased.

    Edit: typo fixed.

  22. Smagneron | | #22

    Replies to Stephen Thwaites' comments #12 and #13
    Stephen Thwaites wrote in comment #12

    “My 'argue over more beer' answer is that it is tough enough already to get 'outie' windows installed properly with sloped subsills and unsealed bottom flanges. In my view properly installing an inbetweenie is 'a bridge too far' for all but the very very best installers. Additionally, and perhaps more importantly, the added cost of a more complex installation needs to be balanced by clients against the energy savings.....”

    You are correct that inbetweenie windows require more labour during the design and installation stages as well as more material, which all contribute to increased costs. However, I hope that this additional cost is part of the learning curve, and that over time the detail of inbetweenie windows will be more familiar to designers and installers so that it’s almost at no additional cost.

    Another consideration with inbetweenie window installations is the increased shading of the window due to the reveal compared to an outie installation. This would negatively affect the Space Heating Demand, but may be compensated by the better Psi-install potential (not to mention the non-energy benefits like comfort and hygiene as Bronwyn pointed out in comment #10).

    Stephen Thwaites wrote in comment #13

    “Its true that masking the outside face of an inswing window reduces heat loss throught the frame. Based on PHI numbers for the Newer Euro Clad Window, the masking can add effectively reduce Uframe by up to 18%. This 'up to' needs to be reduced for at least 3 factors; 1) The Sill, 2) System for over-insulating frame, and 3) Mullions”

    1) The PHPP recognizes that the sill will be more difficult to detail a better Psi-install coefficient than the head and jambs. It is required to analyze all four details as a result.

    2) You are correct that the choice in wall cladding will impact how creative you need to be to over-insulate. But that’s not to say that the same material must be used for the reveal/return that is used on the walls. A site installed aluminum ‘extended brick mould’ could be used to cover the insulation over the frames, even when brick is installed.

    3) The treatment of mullions will depend on your definition of mullions. If the mullions are purely decorative and are installed on one lite of glass, then only its affect on shading would be taken into consideration as a percentage of the area of glass it is installed on. If the mullions are as a result of ganging windows together within one rough opening, then the PHPP requires that you treat each window as individual windows. Each window then needs indication as to which sides of the window are directly against the rough opening. It is only these sides that benefit from calculated Psi-installation values. The frames in the field of the rough opening remain as is.

    I have also seen PHI certificates for thick framed Tilt and Turn windows that can be installed as inswing or outswing. Infact, the outswing windows in some of the installation details outperform the inswing window. This has the added benefit of being able to push the inbetweenie window out closer to the exterior for more solar gains.

  23. AntonioO | | #23

    "Dollars to doughnuts"?
    I think the phrase was more meaningful when you could buy a dozen doughnuts for a dollar, meaning you would be offering a 12:1 payout for a winning bet against yourself. Not quite that old, but I do remember a doughnut shop from my childhood where you could get a dozen doughnuts for a buck-fifty. Now dollars to doughnuts could be one to one odds depending on where you're buying.

  24. Bronwyn Barry | | #24

    Driving in fourth gear
    Martin – that’s a fun travel story. Germans can generally be relied upon to behave, well, like Germans. As a general rule, we love their engineering here in NA, but maybe have cultural differences in our appreciation for ‘hygiene’ (and maybe rigorous window calculations.)

    Stephen T. – I agree that the installation of in-betweenie windows is more complicated and requires architects and builders to flash properly. (Windows were installed in this manner for hundreds of years before the nailing flange came along in the 60’s and the knowledge of how to properly flash was lost to both architects and builders. Pity. Our industry is slowly re-learning this skill. Architects are now hiring another consultant – the Water-proofing Consultant – because either they like to give away work to skirt liability or they don’t like to make any money :-) I hope that our industry has re-learned flashing basics by now and includes site inspections to ensure this critical assembly is installed properly.

    As Stephen M. mentions, the ‘halfie’ install does become easier and simpler as our experience and comfort with it grows. (There still aren’t the great install z-clips available here that I’ve seen in Austria & Germany, but they’re coming.) The ‘halfie’ install also has additional benefits other than performance: it provides an easy way to replace windows in the future without having to damage the exterior siding or stucco around the window. That’s a durability and resilience benefit that considers the long-view. (Historical preservationists, take note.)

    To your comment on identifying condensation risk: this is not measured by using an overall u-value, but is determined at a single point on the sill section where the glazing meets the frame. That spot is always the coldest surface on the whole window and the most common region that interior surface condensation occurs, so it’s a logical place to measure for this compliance requirement. A detail of this test can be found on slide #12 of this presentation: Interestingly, LBNL’s THERM software was updated last year to allow simulators to test for condensation risk. There is talk that the NFRC will include this as an optional test for their certification.

    Psi-spacer: thanks for clarifying your inclusion of psi-spacer rather than psi-install. However, my earlier comment still stands. This is a highly selective comparison and is totally insufficient evidence to make a claim that NA window performance can be comparable to some EU windows. (I do appreciate that you include the caveat: your mileage may vary!) Leaving out the psi-installation info is like comparing three different cars by only driving them all in forth gear because two of the cars don’t have a fifth gear. It’s a totally skewed comparison and penalizes the car with a fifth gear. You have admitted to “Being less familiar with PHPP,” and that you “did not attempt to run this level of detail through the Lancaster model.” I’m therefore inclined to forgive this very smelly comparison as a PHPP beginner’s error. While I too would dearly love to support NA window manufacturers and REALLY DO WANT THEM TO BUILD HIGH PERFORMANCE WINDOWS, the fact-o-matter is that EU windows are kicking our butts performance-wise, and many now compete easily on cost too. Customers are figuring this out and the window import business is doing very nicely.

    I note that your windows are made in Sauk Rapids, MN. That’s ICC climate zone 6. Even according to the confusing PHIUS data you linked to, only the fixed, ‘net shade’ window qualifies for your local region. Rather than spend dozens of hours writing these ‘we’re not inferior’ articles and driving the PHPP in fourth gear to generate fishy performance numbers, how about simply improving the design and performance of your windows?

  25. jackofalltrades777 | | #25

    Bronwyn Barry Sums It Up Best...
    "Rather than spend dozens of hours writing these ‘we’re not inferior’ articles and driving the PHPP in fourth gear to generate fishy performance numbers, how about simply improving the design and performance of your windows?"

  26. Stephen Thwaites | | #26

    Caution, Caution response to Antonio (#20)
    Given your background i am probably not the right person to answer your measure vs. theoretical question.
    While i believe the answer is that there is lots of detailed peer reviewed research supporting WINDOW & THERM, life long researchers are in a better position to answer your questions.
    The 2 names that i can think of are Steve Selkowitz at LBNL in Berkley and Hakim Elmahdy at NRC in Ottawa.

    The comments about 'production laboratory results' refer to the difference between NFRC accreditted test labs and the research labs like LBNL or NRC that produce original peer reviewed research.

    Hope that helps.

  27. Stephen Thwaites | | #27

    PHI Compliance
    The point of my blogs was to show that a presumption that Euro windows are the most energy efficient choice is not always a correct presumption.
    I am making no claims about PHI compliance of NA windows.

    Slim framed fiberglass windows, whether made by my firm, or at least 5-6 others in NA and at least 2 others in Scandanavia, while not complying with all of the PH window criteria, can result in buildings that use less energy than if the same building used an offical PHI window.

    My understanding is that an official PH building does not require PHI listed windows.

    If a project team decides that it wants official PHI listed windows its building, then so be it.

  28. AntonioO | | #28

    Reply to Stephen (26)
    Thanks for the references, Stephen.

  29. Bronwyn Barry | | #29

    Really sorry
    I can't agree with your conclusion, Stephen T. What you’re trying to show is what you (and others) sell right now works just fine. And that may be the case for this particular project if you don’t include comfort, condensation risk and an additional boost in performance from over-insulating the frame. As I said earlier, I'd love to see more locally produced, high performance window options on our market. Thermotech could help lead the way.

    I honestly wouldn't care if they weren't PHI certified either. However, that is where I've seen the most innovative window solutions, specifically in the realm of frame design. PHI sets a ‘must be better than U-window 0.14 or R-7’ high bar for certification. Hundreds of companies have now hurdled that bar, and keep raising it every year. Check out this R-9.4 window certified with a higher delta T for arctic climates: The link includes installation details showing how the header and jamb MUST be over-insulated because of the design, and this is possible for all finish material options, not just EIFS as you assumed previously.

    While the US has generated some of the most innovative glazing options, particularly in the techno-geeky realm of electro- and photo-chromic glass, windows still require good frames. Those frames should ABSOLUTELY NOT cause discomfort or interior condensation. As I elaborated earlier, the 'fifth gear' of windows is the installation detail. Not many builders and architects here are using that gear – yet. But they will be. It’s a relatively cheap performance improvement and good bang for the buck since windows have to be purchased and installed anyway.

    I’ve perhaps belabored my point here, but I’m a natural optimist and wouldn’t be where I am if I wasn’t doggedly persistent. This impressive country was the first to put a man on the moon. Almost all of us now carry ‘smart’ phones – mine says ‘Designed by Apple in California’ - that can perform some insanely complex functions. (Their utility is debatable!) Our military can unfortunately now kill a person on the other side of the planet, using a drone. I've yet to see or hear a really good explanation for WHY, given all the information and technology available to us, a better performing window frame is too much to ask for?

  30. GBA Editor
    Martin Holladay | | #30

    Response to Bronwyn Barry
    You wrote, "And that may be the case for this particular project if you don’t include comfort..."

    What nonsense. Are you really alleging, Bronwyn, that homeowners who install triple-glazed Thermotech windows will have comfort complaints due to the difference in between the U-factor of Thermotech frames compared to the U-factor of some European frames?

    If that's what you are claiming, it's balderdash.

  31. Bronwyn Barry | | #31

    Martin - Stephen set up this whole comparison to compare a typical NA window frame with 'many European Passivhaus window frames.' He set the comparison benchmark, not me. All I'm pointing out is that there are additional criteria that he is conveniently omitting in the comparison.

    If you want to compete, at least have the balderdash to compete on a level playing field and quit trying to move the goalposts to favor your home team. In most countries, including this one, that's called cheating.

  32. GBA Editor
    Martin Holladay | | #32

    Response to Bronwyn Barry
    You're avoiding the question.

    You live in California, Bronwyn. Are triple-glazed Thermotech windows uncomfortable (due to the U-factors of their frames) in your climate? If so, how cold does it get when you notice this comfort problem?

    I live in Vermont, and last winter the temperature dropped to -25 degrees F. I was never uncomfortable reading a book beside my triple-glazed Thermotech windows. I'm glad you are worried on my behalf, Brownyn, but get serious. This is far beyond a typical princess-and-the-pea type tantrum. This is utter nonsense.

  33. Bronwyn Barry | | #33

    The 'comfort' criterion I was referencing is the PHI one that Stephen Magneron calculated in comment #5 above. He said, "The only frame that passes the PH Comfort Criteria is the Newer European Clad Wood window frame. In fact, a window frame would have to have a Uf – 0.78 W/m2K (0.13 BTU/hr.ft2F) to pass the PH Comfort Criteria in Ottawa." I incorrectly assumed you would get that, but since you didn't, there's your answer.

    Whether or not the measured effects of this criterion make a significant difference to comfort is a topic for much further investigation. The PHI comfort criterion is not one that is determined by where I live, or where you live, or where Stephen lives.

    If you want to include a perceived comfort measurement as another "selective criterion" in this comparison, that's your prerogative. I don't believe that is what Stephen T intended, or would support. As he stated in comment #16, "I did this not to present real windows, but to isolate components." It's a laudable exercise and he achieves his goal admirably. I agreed above that looking through that single lens, it works.

    Stephen T is a very rational guy. I'm pretty sure he would concede that this particular topic requires more than one lens. Martin, it appears you're not willing to concede anything, but want to add 'perceived comfort' as a new criteria. Perhaps it's time for a Pretty Good Window manifesto? :)-

  34. GBA Editor
    Martin Holladay | | #34

    Response to Bronwyn Barry
    Thanks for explaining what you meant. I now understand what you intended to say.

    However, I don't think I can be faulted for thinking you meant what you wrote, instead of what you intended to write. When you wrote "comfort," I thought you were talking about comfort, not "compliance with the Passivhaus Institut's comfort criterion."

    My misunderstanding has nothing to do with my willingness to concede anything; I simply misunderstood a confusingly written sentence. Thanks for clarifying what you meant.

  35. AntonioO | | #35

    "calculated comfort" vs. "perceived comfort"
    Granting that with odd terms like "hygiene"--a component of the "comfort criterion"--having been defined and that Bronwyn stated clearly that the "hygiene requirement is not a performance number per se" it is still hard for me to conclude that "calculated comfort" is not supposed to in some way give an indication of "perceived comfort." This is especially so given how the comfort criterion was defined initially by Stephen M. Intuitively, it would seem that a window with a higher interior surface temperature (in the heating season) would be characterized as more comfortable by perception. Of course, perception rides along with the perceiver. That being said, given that Martin has perceived his Thermotech windows to be comfortable in pretty low outdoor temperature conditions, I wonder if he might share what he perceives to be a comfortable indoor air temperature. In other words, Martin, what do you set your thermostat to in winter if you have one?

  36. GBA Editor
    Martin Holladay | | #36

    Response to Antonio Oliver
    Windows are often responsible for comfort complaints in winter. Sometimes, people complain that the windows leak air, and this is possible. More often, a discomfort perceived as a draft is really radiational cooling of one's skin. This is especially common when the outdoor temperature drops below 0 degrees F. The warm skin of a person sitting near a window radiates heat to the cool glazing. This chill is sometimes perceived as a draft.

    If you upgrade from double glazing to triple glazing with one or two low-e coatings, these complaints usually disappear. The reason is that the temperature of the innermost pane of glazing is much higher on a triple-glazed window than a double-glazed window.

    I have never heard of a comfort complaint related to the U-factor of the frame. The temperature of the interior glazing is always the significant factor in these complaints. After all, the frame of the window represents a relatively small percentage of the window's area.

    I don't have a thermostat at my house. I imagine that my indoor temperatures range from 68 to 72 degrees F, depending on my mood and the output of the wood stove.

  37. AntonioO | | #37


    Thanks. Your point about heat radiating from the person toward a colder window glazing is a good one. With the rate of radiation being driven by the temperature difference between the two objects, getting that window temp up is key. But if we're talking about high performance glazing with a lower conductance than the window frame, why shouldn't I expect the frame then to be colder than the glazing itself? Or is there a good enough point reached by most windows (or a Pretty Good window) such that any radiation effect would be negligible and the risk of condensation directly on the frame is likewise small? And, no, there is no need to repeat your point about the frame representing a small fraction of the windows area. Thanks again.

    P.S. You might be surprised how many people still have single pane windows in decades old frames that actually do leak air.

  38. GBA Editor
    Martin Holladay | | #38

    Response to Antonio Oliver
    The U-factors of the North American frames that Stephen Thwaites discusses in this article range from 0.15 to 0.23.

    When I surveyed the window specifications of a wide range of North American manufacturers of triple-glazed windows for a 2010 GBA article, the whole-window U-factors (that is, the NFRC U-factors) of the surveyed windows, including both frames and glazing, ranged from 0.13 to 0.29.

    In other words, the frames that Stephen is discussing have U-factors that are in the same range as the whole-window U-factors. The frames do not represent a "weak link," and are therefore unlikely to have a surface temperature that is lower than that of the triple glazing.

    This reinforces my belief that it is hard to imagine that someone sitting next to one of these windows would feel comfortable next to the glazing, but would somehow feel a chill because of the window frames.

  39. Stephen Thwaites | | #39

    Over Insulated Frames - General

    Bronwyn is right about the level of innovation in PHI windows. The PHI list of certified windows is by far the largest cluster of innovative window designs that i have ever seen. Over insulating is one of the many innovations to be found there.

    Inswing windows lend themselves to exterior overinsulation. Overinsulation is easiest with thin claddings like EIFS. You can overinsulate with thicker claddings, it's just that with thicker claddings, the insulation layer is porportionately thinner. And as Stephen M. suggested on buildings with thicker cladding the sides of the window opening could be clad with something thinner.

  40. Stephen Thwaites | | #40

    Over Insulated Frames – Energy Savings

    A key question is how much does this technique save? With some coaching (thanks Stephen M.) I did some additional work with the PHPP model for the Lancaster House.

    Based on the PHI certificate for the Newer Clad Wood Window the difference in PSIinstall'n between the masked and unmasked conditions is 0.007 Btu/(hr ft F) for wood framed walls. This degree of overinsulation is based on a thin cladding, like EIFS, over foam overinsulation.
    The Lancaster house, built 4 years ago, was modelled in an older version of PHPP, so I had to manually adjust the overinsulation effect on PSIinstall'n for the sill. (To drain water sills are not overinsulated). The windows are, on average, 1.5 times taller than they are wide. Based on this I reduced the overinsulation effect by 1/5th.

    The result is that the over insulating 15 Newer Clad Wood Windows saves about 0.1 Btu/ ft^2 or 0.4 kWh/m^2. For a house that just mets the PH stnd., this measure would save another 2.5%. By most people's standards it is probably not a lot of energy. For the Lancaster house this is about 65 kWh/yr.

  41. Stephen Thwaites | | #41

    Over Insulated Frames – Horse Race

    For the Newer Clad Wood Window the savings of overinsulating is 0.1 Btu/ ft^2 or 0.4 kWh/m^2. This is enough that the (frame+overinsulation) combination of the Newer Clad Wood Window, is marginally more energy efficient than the NA Fglass Window.

    In this case marginally means about 16 kWh/yr. I might be wrong , but I think most PHPP users, would call this a tie.

    Keep in mind this is all based on maximally overinsulating. Using thicker claddings will reduce the overinsulation effect.

    Also keep in mind that the window that forms that basis for the Newer Clad Wood Window is not your average 4” wide framed Euro Window. To my knowledge, this 'A' rated window is not yet available in NA. If it does arrive here, it will need an insect screen for most markets. Accomodating a screen that can be removed from the roomside will likely increase the frame width, or limit the once the installation effect of over insulating the frame is considered.

    So, in my view, its still fair to say that while Euro windows are certainly very impressive, there should be no automatic presumption of their thermal superiority.

    And once again, running PHPP or HOT2000 is the best way to optimise your building.

  42. Bronwyn Barry | | #42

    Stephen, thanks for taking the time to look into the psi-install numbers on this sample project. It looks like your windows work well on this project. It would be great to conduct a more extensive study into this issue. Perhaps we can collaborate on that in future?

    Many thanks again for a great post.


Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |