GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Presumptive European Superiority Syndrome

Triple-glazed windows from Europe don’t always outperform triple-glazed windows from our side of the Atlantic Ocean

This Passivhaus residence is located in Lancaster, New Hampshire. Several iterations of the PHPP modeling software were used to evaluate different glazing options for this superinsulated house.
Image Credit: Garland Mill Timberframes

UPDATED on October 22, 2014 with an Addendum.

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

Most of the world, especially the green building community, assumes that “European” implies “more energy-efficient.” When it comes to windows, this automatic presumption of superior energy efficiency is both so common and so misplaced that it deserves a name: the Presumptive European Superiority Syndrome.

Different standards

Higher R-values (lower U-factors) for European glass are for the most part a product of different test standards rather than a refined European mastery of physics.

European window standards (CEN) are based on an outdoor temperature of 32°F (0°C), while North American standards (NFRC) are based on an outdoor temperature of 0°F (-18°C). At the warmer exterior temperature of 0°C, there is a thermal benefit to a wider to 11/16-inch (18 mm) pane spacing (that is, to a wider gap between the multiple panes in an insulating glazing unit, or IGU). This benefit does not exist at the colder North American design temperature of 0°F.

In other words, at 32°F, glass lites optimally spaced for 32°F insulate better than the same glass lites optimally spaced for 0°F. That’s why three lites of Europe-optimized glass has a European Uglass that is lower lower than a North American Uglass. (“Uglass” is the center-of-glass U-factor, as opposed to the whole-window U-factor shown on NFRC labels.)

For example, consider a high-solar-gain triple with two low-e coatings as shown in the drawing below:

Depending on the inter-pane spacing and the calculation procedure, this glass unit can have very different U-factors.

The table below compares the U-factors of optimally spaced glass for each of the NFRC and the CEN Standards.

This triple-glazed unit, when optimally spaced for each standard and evaluated according to each standard, has very different values for Uglass. In this case the CEN Uglass is 19% lower than the North American Uglass. (0.61 vs 0.76 W/m^2K or R-9.3 vs R-7.5).

So European windows have a significant marketing advantage. Their Uglass is calculated differently, so they claim a significantly better insulating value for their glass. It’s sort of like runners comparing times for the 100, without specifying yards or meters — only worse.

More bluntly, stating the Uglass value of a European window to North Americans without identifying it as a European Uglass -value is misleading. For that example tabulated above, it is very misleading.

For more information on the implications of the above table, see the Addendum at the end of this article.

Different Low-e 1

So European glass looks better because of a more favorable standard. Some people (usually Passivhaus enthusiasts) will say that European windows have a lower Uglass not only because of differences in standards, but also because “they have better glass than we do.”

I may be biased, but I’m not so sure. Take a closer look.

Because the Passivhaus program is based on European (CEN) standards, my analysis is based on calculating Uglass to CEN conditions.

The best insulating glass on the Passivhaus Institut’s Component Database for Glass is Guardian Europe’s “ClimaGuard Premium2.” It is listed as having a metric U of 0.49 when filled with krypton and having a G/SHGC of 0.53.

In imperial units that’s U-0.086 or R-11.6. That’s a mind-bogglingly good insulating value. WINDOW7.1 produces the same results for Climaguard Premium2 under CEN conditions with a glass unit constructed as shown below. So, not surprisingly, the Passivhaus Institut’s listing is a legitimate result.

It is true that no single North American low-e coating can produce exactly this combination of SHGC and U-factor. European low-e coatings are different. European low-e coatings have SHGCs and U-factors partway between those of North American high-solar-gain low-e coatings and low-solar-gain low-e coatings.

Logically, then, the best way to create a European-like unit with North American glass is to create a hybrid unit as shown in the diagram below. The North American hybrid unit has one high-solar-gain low-e and one low-solar-gain low-e. This hybrid unit comes awfully close to Guardian Europe’s ClimaGuard2. According to WINDOW 7.1, using Guardian Glass, such a hybrid unit has a G/SHGC of 0.52 and a metric U of 0.53. In imperial units that’s U-0.090 or R-11.1.

This Guardian version of a North American hybrid unit insulates slightly less well (about R-0.5 lower), and allows just slightly less solar gain (the SHGC is about 0.01 lower) than Guardian Europe’s Climaguard Premium. We could argue all day about how close is close, but the only way to evaluate the difference is not to argue over more beer, but to run PHPP.

So I ran PHPP. I compared not only a North American hybrid unit based on Guardian North America’s glass, but also three other North American manufacturable (Cardinal, PPG & AGC) hybrid units. Using a PHPP model for a house in Lancaster, New Hampshire, I compared them all to Guardian Europe’s Climaguard Premium2 glass.

Built by Garland Mill Timberframes in 2010, the Lancaster, New Hampshire, house is a highly glazed house in a heating climate. Its south-facing glass area is equivalent to about 17% of the floor area. This is high, but most of our custom house customers, whether building a Passivhaus or not, end up with houses with large south-facing glass areas. So, by my logic, it’s not an unreasonable example.

I used the PHPP spreadsheet for this house to compare the effect of several different glass options. The table below compares four North American hybrid glass options to the most energy-efficient European glass option. The table compares their respective G/SHGC and Uglass characteristics, along with the PHPP-generated specific heat demand.

All of the above options — some of today’s most crazily energy-efficient, most optimally spaced IGUs, with rather expensive krypton gas fills — are more energy-efficient than what was used in the 2010 New Hampshire house. That’s why all of the above glass options produce much lower specific heat demands than the 4.45 kBTU/ft^2/yr (14.2 kWh/m^2/yr) in the as-built PHPP model.

Running the PHPP model with the four North-American-made glass units produced specific space heat demands ranging from 3.3 to 3.5 kWh/m^2/yr.

So the North American hybrid glass results in 10%-20% higher projected heating bills for this cold-climate Passivhaus than the best European glass. For this house, the difference is about 160 to 320 kWh/yr. As a worst-case scenario, with $0.15/kWh electric-resistance heating, this 10% difference amounts to about $25 to $50 per year.

I suspect that to most people the differences aren’t much. But the bragging rights do go to the European glass. Your mileage may vary, but my guess is that most heating climate Passivhaus buildings would probably get similar results. So that’s it then: the European glass is better, right?

Not so quick, Poindexter.

Different Low-e 2

There is at least one other worthwhile option to consider. It’s an option that takes advantage of North America’s sunnier climate. This option uses North American high-solar-gain low-e coatings for both of its low-e surfaces. Combined with a low-iron outer lite, let’s call it a super-gain glass unit. An example of this unit is shown below.

For these low-e coatings, the G/SHGC is higher than the highest rated PHI-listed unit, which is good. On the other hand, the Uglass is also higher, which is not so good. Again, the best way to assess “better” is not to argue over more beer, but to run PHPP.

The table below shows G/SHGC and Uglass for four North American high-solar-gain low-e glass options. More importantly, it also shows the respective specific heat demands when input into the Lancaster, New Hampshire PHPP model.

Again, and as always, your mileage may vary. However, for this heating climate example, the North American high-solar-gain low-e results in about a 20% lower space heating demand than “the best” European low-e.

Because this house so energy-efficient, this 20% spread at Passivhaus efficiency levels does not result in a large dollar savings; it is nevertheless important to note that there is likely no incremental cost to achieve this savings. (That is, there is no incremental cost to go from “North American hybrid” glazing to “North Americana super-gain” glazing.)

The key point is that despite not insulating as well as the European low-e glazings, North American high-solar-gain low-e glazings take advantage of the fact that, for the most part, North American heating climates are sunnier than European heating climates.

So not only can North American glass match the performance of European glass, but in heating climates it can also surpass it.

I do think it’s fair to say, then, that when it comes to energy-efficient glass, that there should be no automatic presumption of European superiority.


Some people have wondered how two windows with high-solar-gain triple-glazed IGUs — one with an IGU with optimal spacing according to the NFRC standard, and one with an IGU with optimal spacing according to CEN — would fare when tested to the standard for which they were not optimized. The super-short answer is that would they do worse, but one would do worse than the other.

The CEN-optimized unit loses all of its “advantage” when simulated with NFRC conditions. The NFRC-optimized unit, on the other hand, is almost imperceptably less insulating when simulated to the CEN conditions. An expanded table that shows all 4 results is shown below. The added information is in italics.

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


  1. user-2069108 | | #1

    Cool analysis
    This is what I have been looking for ever since I learned that European Ug's were measured differently than in North American.
    So European sealed glass units have spacing of 18mm, which are optimized for their climate conditions (or is it optimized for testing?). And this wider spacing (vs 12.5 mm in NA) is not advantageous domestically, therefore, these European windows would not really perform as well as advertised here?
    In the first table, we see this comparison, sort of, but I think we have to read more into it. The U calculations are for the same glass configuration, but with different spacing. What isn't clear, however, is whether the U calculations are also based on the different testing (Outside temperature of 0 degrees C versus -18 degrees C domestically). If the numbers are, in fact, calculated as I indicated, what is the Ug of the same glass configuration using NA testing and the European spacing?

  2. AntonioO | | #2

    First, thanks to Steve Young for asking my first question. It's not clear to me why a lower U from bigger spacing is not an advantage, period, irrespective of where the window is installed.

    Second, are the large relative energy savings (up to 20%) only applicable because the heating load is already so low for such an energy efficient house? In other words are the windows a relative weak point in the structure? Certainly, someone living in a house like mine--double wythe brick and cinder block (yep, real "cinder") walls that have no insulation, and laughable attic insulation--should not expect even the best windows to save a huge amount on energy, right? I use my house as an extreme example here, but in general isn't that the case?

  3. Guilbault | | #3

    Really appreciate the effort to actually model the differences
    Antonio Oliver, regarding your first question as to why bigger spacing between glazing surfaces is "not an advantage period": it has to due with the formation of convective gas currents between the surfaces. At high enough differential temperatures between surfaces in multi pane glass, convective loops will form in the insulating gas between the surfaces. This worsens performance.

    Changing the spacing changes the conditions under which convection can be caused in the insulating gas. A smaller spacing can discourage convection, but does so at the cost of reducing the amount of insulation provided by the gas at temperatures where convection is not an issue. Optimum spacing is determined by the type of gas and the design temperature of the window.

    The question you need to ask is "where do you need the performance?" At moderate temperatures, or at more extreme temperatures?

    In North America many manufacturers have chosen spacings with an intent to mitigate losses from convection at more extreme temperature differentials, which necessitates a sacrifice in insulating performance at more moderate temperature differentials which do not cause convection of the insulating gas.

  4. spaulson | | #4

    Good info. , but
    Thanks for taking the time to put this information together. These are definitely differences that need to be taken into account and understood. I think there are a few things being left out that are important to consider as well.

    What happens to the cooling load and winter overheating by going to the higher SHGC glass package?

    From what I have seen some of the better performing low e coatings from the US have some low VT numbers. What is your experience?

    Do we want to be using 0 degrees F for our rating number? It seems as though this is a on the low side for most of the US.

    Just some thoughts that it would be good to hear your opinion on.

  5. jinmtvt | | #5

    superb technical article ..
    Very interesting and informative Stephen .
    I always welcome deep technical information and even if i did not understand all of the small details on first reading, the base info was very easily understood with the accompanied PHPP exemples !

    About european products ...
    What about the total window U value from PH certified units compared to our/your offerings ?
    There are major differences in framing designs and materials and also how the glazings are installed and often recessed in the frames.

    Also some typical german type installations benefit product performance as announced by installing the frame within a recessed insulation cavity , limiting its exposure .

    I have yet to experience a NA window that was assembled and designed with the same quality as good german windows though .. some manufacturing lines videos on youtube of euro factories
    show very advance production systems .

    Merci Jonathan, for explaining why the spacing matters , good simple explanation!

  6. jinmtvt | | #6

    Scott Paulson : cooling load
    Scott Paulson : cooling load has nothing to do with glazing ... provide shade
    architectural features

    Stephen , i imagine that the colder the design temp is the better the us smaller spacing benefits ?
    Is it a on/off thing at a certain temp or does it have a gradient ?

    Does windows7.1 / PHPP take into account the glass spacing up to your modeled energy experiment ?

    Lastly, the results favor high SG because of a pretty large south fenestration %,
    would you have an available model with much lower south % windows to play with and test how much difference between the high SG glazings and the hybrid/euro type.
    And perhaps a test with a much cheaper glazing alternative ( regular triple/double , argon and low-e ) if it is not too time extensive for you ( i do not know the implication of running such tests in PHPP labor-wise )

  7. spaulson | | #7

    Reply to Jin
    You are correct that there should be proper shading. I don't see much if any in the above pic. Shading will help in the summer, but won't likely help in winter overheating or potentially overheating in the swing seasons.

    This just isn't as cut and dry as the data may appear. Very good and valuable data to understand the complexity in the decision, but we need to make sure we are taking everything into account.

  8. jinmtvt | | #8

    Scott :
    Yes something i've been pointing out in way too many recent projects
    ( lack of proper architectural shades on at least south side windows )

    This subject is soo complexe going back up to mass effect on thermal regualtion etc....
    but this report presents interesting information nonetheless to most of us

  9. user-626934 | | #9

    0C and 0F
    The vast majority of people in the U.S. live in places where the average temperature for the 3 coldest months is 32F(0C) and higher, therefore the CEN modeling protocols are more appropriate in most situations. Even in Canada, you have to go to a place as cold as Edmonton to get close (but not quite) to 0F average for Dec/Jan/Feb. Ottawa, home of Thermotech, splits the difference at about 15-16F.

    North American (and European) manufacturers optimize their products within the rules of the "league" they play in...NFRC in North American, CEN in Europe. The question I've always had is this: why do the NFRC rules stipulate an outdoor temperature that is sub-optimal for 90%+ of all North American locations?

  10. Stephen Thwaites | | #10

    reply to Steve Young (#1)
    The first Table shows the results from WINDOW 7.1 for 2 different conditions; the NFRC Winter Design (0F) & the CEN conditions (32F). The optimal glass spacing for each condition is different. The first Table shows the Uglass for each condition's optimal spacing.
    If the roles are switched and 18mm spaced glass is simulated under NFRC conditions then its Uglass is higher than if simulated under CEN conditions.
    Similarly if the 1/2" spaced glass is simulated under CEN conditions instead of NFRC conditions then its Uglass is higher --- but not nearly as much.
    An earlier draft had this info in the Table, but thinking it weakened my point about differences between the two standards i took it out.
    I will work on a way to re-incorporate that info.

    hope that helps

  11. AntonioO | | #11

    What physically changes, Stephen Thwaites?
    Stephen, you say the following in your article:
    At the warmer exterior temperature of 0°C, there is a thermal a wider gap between the multiple panes in an insulated glazing unit, or IGU
    Although, you didn't say so, Jon Guilbaught says this is an effect of convective heat loss being mitigated by a smaller spacing. Admittedly, I don't quite get that. Can you say a bit more about the physical properties behind your statement?

    On the absolute temperature scale of Kelvins, a change from -18C to 0C is only a change from 255K to 273K--not that big. For a noble gas like Krypton or Argon both those temperatures are well above where a lot of interesting stuff happens in this column of the periodic table. The kinetic energy (being proportional to temperature) would only change by seven percent for such a temperature change. Correspondingly, a RMS velocity change of the gas molecules of about 3.5% would occur. What am I missing?

    One other question, if I'm considering a European window and see a U value for glass (not the whole window unit) is that a calculated number or a measured number?

  12. Stephen Thwaites | | #12

    replies to Antonio Oliver (#2 & #11)
    Differences in Uglass are independant of other component U values. But your intuition is correct that PH quality windows in an otherwise uniinsulated building won't materially change overall energy consumption.

    As far as the why questions goes - why a different temperature difference produces a different Uglass goes, others know more than me. Hopefully Greg Smith will wade in here.

    I'm not sure, but suspect that Euro Uglass numbers are simulated values. Given the accuracy of simulations, and the potential for measurement errors i would trust a simulated Uglass over a measured Uglass.

    Hope that helps

  13. Stephen Thwaites | | #13

    reply to Scott Paulson
    re: overheating
    For the Lancaster NH house, PHPP says with the Euro (or N.American Hybrid glass) the frequency of overheating is 9%.
    The N.American High Solar Gain Glass, according to PHPP, would have a frequency of overheating of 13%.
    These figures are from one PHPP model, your mileage will vary.
    Perhaps someone with more experience with PHPP could comment on these figures.
    In my experience, some winter overheating is appreciated as novel in cold climates. On the other hand no one wants to live in an occassional oven. And the definition of "some overheating" and the definition of "an occassional oven" can vary from person to person.
    re: VT of 'better performing' low-e
    Some of the lowest SHGC glazings do have a lower VT. In general they don't help a heating climate building meet the PH energy targets because they allow so little solar gain. For that reason none of this group of glazings was considered in my piece.

  14. Stephen Thwaites | | #14

    reply to Jin Kazama (#5)
    Re: Euro Windows
    This blog is about glass. You are correct that there is more to window energy efficiency than just glass. Without spilling the beans on a possible future blog; where North American products have an advantage is the slimness of the frame. North American fixed frames in particular are often alot smaller than their bulkier Euro counterparts. The glazed area of a window, in a heating climate can be a net energy gain. Even the most highly insulating frame, is a net energy loss. Sometimes a larger glass area more than compensates for a less well insulating frame.
    The best way to assess this effect to run PHPP or other relatively detailed software for your building.

  15. Stephen Thwaites | | #15

    reply to Jin Kazama (#6)
    Re: Spacing and Temperatures
    I would imagine that over the temperature ranges that a building would experience, that the bigger the temperature difference, the smaller the optimal space between panes. I have no idea if this is a linear relationship, but i imagine someone, somewhere has studied this.
    I think the relatively narrow range of spacing and temperatures discussed in this piece are likely well modelled by WINDOW 7.1.
    Re: Results, Trends
    Because PHPP results trend the same way as the ER (Energy Rating is a CDN single number rating that is the sum of the gains and losses) i believe the trend of the PHPP results.
    A less over glazed building would have a lower absolute difference in energy consumption between high gain and lower gain glazings, but the high solar gain glazings would almost certainly yield a lower energy bill. Again run PHPP or similar software for yourself, for your building in your location, then consider against your costs..

  16. kevin_in_denver | | #16

    The Most Important Window Spec is the Price
    For anyone who is obsessing about window specs, please re-read:


    In an average house in an average climate, windows that have a 3x lower U value only save $100-$200 per year.

  17. Bronwyn Barry | | #17

    Poindexter's inferiority complex...
    Nice post on a topic that has vexed the North American PH community for many years. The explanation of how these ISO (used by the Passive House Institute - PHI) and NFRC calculation protocols differ is important to understand. However, the conclusion that NA glass gets the short end of this stick is too simplistic. As with most window-related data, the answer is complicated and more accurately stated as 'it depends.' (I conducted a similar study on this issue a number of years ago using a one project, one product sample size and reached the same inaccurate conclusion, but also managed to figure out the protocol anomalies. My graphic representation of the differences can be viewed here:

    A really comprehensive study of this issue has been done though. A coalition of smart Canadian companies had the foresight to fund a proper study by RDH Engineering. A ppt was presented at Passive House North 2013 and can be viewed here: The conclusion reached by RDH was that once all the additional calculation metrics between the two testing protocols are included, there is no correlation between them. No easy 'conversion' factor can be used and the additional variables of size, frame and spacer are complex enough that only u-values derived from the same testing protocol methodology can be reasonably compared. (It includes a great illustration of the spacer width issue nicely explained above by Jonathan Guilbault.)

    After my own rudimentary study into this issue I did manage to connect both the powers-that-be at the NFRC and at the Passive House Institute to see if there was a path to help remedy this situation for PHPP users in North America who want to source local windows. Not surprisingly, the brief discussions elicited variations of 'our protocol is better' murmurs from both sides. To its credit, PHIUS made an attempt to 'harmonize' the two protocols with their window 'data certification program' and generated a hybrid between the two protocols. It lands somewhere 'mid-pond.' (Unfortunately neither original parties currently recognize the PHIUS data, which can now only be used for PHIUS+ projects.)

    Fortunately any number of local simulators have the ability to generate window performance data using either NFRC or ISO protocols. (Thanks to free US tax-payer-funded software called THERM.) PHI has hinted at possibly implementing their own verification service, but has preferred to encourage performance and design improvements by retaining their high bar 'must be better than U-w: 0.8W/m2K' certification standard. Having recently designed a window to meet this standard, I can confirm that it's a challenge, but not that hard. For this reason, I'm surprised more North American window companies have not stepped up to the plate. Poland, China and many other much smaller nations have managed it easily. :)

    So, Poindexter, I recommend you revel in your fabulous 'poutine and politesse' and not feed any inferiority complex. This is no apples to apples comparison. As they say in GoT, 'Winter is Coming.' Time to get busy and design a window that perfectly meets the Canadian cold climate using the advantages of the great glass we do have available here, but with better frames and wider spacers. I'd love to see them.


    PS. All due respect to Stephen Thwaites. The only similarity I ascribe to his fantasy character 'Poindexter,' whom I address here, and Stephen is that they are both Canadian.

  18. jackofalltrades777 | | #18

    Check the BBB
    The author of this article works for Thermotech Fiberglass out of Ottawa, Ontario. Maybe he would like to explain why the BBB has given his company an "F" Rating?

  19. jackofalltrades777 | | #19

    Flawed Article
    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 that is just one factor why they are designed and perform better. This then plays into how the window panes are gaped between the spacer. 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.

  20. jinmtvt | | #20

    Peter L and John
    I would rather have a high performing product when it is required : c'est a dire when it is deep cold !

    -20C is double energy loss than 0c .. so average temp doesn't mean anything because we have a set point for Delta T which is around 20c .

    Last winter my night time temperatures where under -15c for something like 4 months ,
    and more often than not, lower than -20c . This is the time when u need everything to perform at its best ( and where the HeatPumps are at their lowest COP so cost is doubled )

    Peter : their business is not BBB acredited, so they can't resolve complaints etc on BBB ... i will not vouch for them, but i have yet to see another owner of a window manufacturing company have as much interest and take as much time to discuss with everyone than Stephen.
    The more sales you make, the more possible problems you get.

    Bronwyn Barry: wow super stuff here , thank you ( from a canadian :p )
    Some more reading stuff here .

    Kevin Dickson: even if it is worth only 100$ per year in economics, we still need to discuss and understand it nah ?? But thanks for dragging us back down to earth here :)

    Stephen : about frames

    i do not understand what you mean here ... we are discussing about ~ r10 glazings
    but yet your own window frames are about 3" thick insulated cavity
    which should end up being from R12 to R14 depending on insulation
    so why would a smaller frame to glazing benefit our american windows ??

    If you wish to compare a shitty PVC frame with only a few compartments and no added insulation
    to a high end R11 glazing well ok .. i guess that smaller frame is good then

    But how hard is it really to design an R12 frame ?

    ( between i still do not understand why your details on your website show an empty area all around the glazing that could be at least filled with backer rods or other insulation , whereas most European windows are completely sealing IGU units in their frames )

  21. GBA Editor
    Martin Holladay | | #21

    Thin frame advantage
    The advantage of thin frames isn't about the U-factor. It's about the SHGC.

    The thinner the frame, the higher proportion of the window area is glazing -- and the glazing is a lot better at admitting solar heat gain than the frame.

  22. Peter Hastings | | #22

    Genuine Superiority of European Triple Glazing
    The real superiority of European triple- and double-glazed high-performance windows lies in their market share. There is simply a higher proportion of good windows in Europe. Market penetration in the US and Canada is very small. Fortunately this doens't matter much since better or worse window performance is such a small part of the overall thermal envelope. Nonetheless, it is the relatively large market in Europe for high-performance that leads to "manufacturing lines videos on youtube of euro factories show very advance production systems". The production systems are advanced because the market is large and secure. This is not due to a highly-educated, green-savvy clientele - the average European home-owner's knowledge of windows doesn't extend much beyond knowing how to open and close them and what products are best for cleaning them. So what is the difference ? Building codes. Until NA gets national building codes which are stringent and enforced (no more "I don't need to build to code in my county") there will be no mass market for high-performance windows and no mass production. Of course this will be resisted fiercely in the US on the grounds that it is central regulation undermining state's rights.

  23. Stephen Thwaites | | #23

    0C and 0F
    I suspect the origins of N. Americans using of 0F to calculate Uglass is related sizing heating equipment.
    Like many who have thoughtfully responded i'm not sure N.Americans are well served by this.
    By far, a much bigger problem than undersizing heating equipment is reducing the carbon emissions from annual energy consumption.
    Accurately calculating annual energy consumption has more to do with average annual heating season temperatures than heating design temperatures. (using this logic, Europeans should also probably calculate Uglass at a warmer temperture)
    Changing the temperature at which N.America calculates Uglass is an almost inconceivably difficult task. Manufacturers have a big voice. An incremental rachet to E.Star generates tantrum tornados that are frustratingly completely and utterly oblivious to our carbon problem.
    Even if NFRC changed the temperature at which it calculated Uglass, not all glass makers would want to change to a wider air space. Wider air spaces are more prone to winter time thermal stresses, which can result in broken glass. Winter design temperatures are colder in many parts of N.America than in Europe - so the risk of breakage for wider spaces is higher here. There are ways to mitigate this risk, but all add cost.
    However, despite these concerns, i still think that 0C (32F) is the better point at which to calculate Uglass - just not sure how to change the standard.

  24. GBA Editor
    Martin Holladay | | #24

    Response to Steve Young (Comment #1)
    Q. "What is the Ug of the same glass configuration using NA testing and the European spacing?"

    A. See the new Addendum to the article, created by Stephen Thwaites this morning.

  25. Stephen Thwaites | | #25

    inferiority complex ? who me?
    My motivation for writing this piece was not to judge, or correlate the two (NFRC and CEN) approaches. As you point out lots of smart people have decided they can't be correlated. Neither was my motivation to even expose my inferiority complex, although it may yet do so....
    My motviation was two-fold;
    First, i wanted to point out a European supplier, using a Euro Uglass should identify that Uglass as being different from and not comparable to the Uglass from a domestic supplier. To do otherwise is misleading.
    Second, it was to point out that if a window person* in a heating climate wanted to 'play in the PH sandbox' they could do so with wider spaces and N. American glass. There is no inherent reason why Euro glass is better for the N.American PH market. In fact in some cases there is a reasonable argument for the opposite - which i tried to make. (* by window person i mean either a factory based manufacturer and/or a site based builder of fixed windows)
    P.S don't eat poutine; its real bad for your choloesterol.
    On the other hand, being polite is good for your soul.

  26. AntonioO | | #26

    Indoor temp
    Another question. This may have been stated but perhaps I missed it. But when the U values are calculated (measured) by the two ratings agencies, what indoor temperatures do they use for their calculation (measurement)?

  27. AntonioO | | #27

    CEN indoor test conditions, anyone?
    Found NFRC conditions on their website:

    Tin = interior ambient temperature of 21.0º C (69.8º F)
    Tout = exterior ambient temperature of -18.0º C (-0.4º F)

    still looking for CEN indoor conditon. Maybe it's the same as NFRC

  28. Stephen Thwaites | | #28

    CEN Indoor Conditions for Uglass

    From WINDOW 7.1
    T indoor = 20C, or about 68F

  29. [email protected] | | #29

    those are two great


    those are two great links, thanks!!

  30. user-2069108 | | #30

    Is the PHPP smart?
    Well, if all these window specifications are calculated for ratings, why doesn't the PHPP simply allow the raw data to be plunked in and let it do the calculations to suit the PHPP model - all adjusted for the climate that is experienced?

    Stephen and Bronwyn - thank you for your diligence on this issue. I see that Canada is going to demand that Europeans have to provide NFRC numbers. Hopefully, the Americans will do the same. It is a real PITA that we cannot easily compare products. (I guess there aren't enough European windows being sold in the US for the local manufacturers to cry foul - the only way anything happens down here.)

  31. [email protected] | | #31


    Thank you for posting that was a great explanation, one of the best descriptions of the differences that I have read. Much better than anything that I have written addressing the same subject.


    Whether dealing with walls or window IGU’s (leaving radiant heat gain/loss in an IGU aside), it’s about controlling air movement within the space. Convection currents are bad whether they are inside your wall or inside your IGU. And while a wall can be filled with all sorts of different materials to limit air movement; not so much an IGU. And also like a wall, a deeper IGU in a window is potentially better - provided that you can control (within reason) convection currents.

    As an aside, the correct term is INSULATING Glass Unit, rather than INSULATED Glass Unit.

    When one pane of an IGU is warm and the other is cold, the air (gas) between the lites of the IGU is going to move. It’s going to form a convective loop and there is no practical way to stop it - at best you can limit it.

    The combination of warmer air rising and cooler air sinking results in a convective loop within the IGU. It's actually a rather efficient heat transfer device. While convection isn't the actual heat-loss mechanism, that would be conduction, convection is certainly feeding the conductive loss.

    Replacing air with a gas fill is one way to help limit convection currents. In that role argon is good, krypton is better, xenon is more so, and as far as I am aware no one has yet to try radon.

    Another way to control convection is to match the airspace width to the temperature difference of the inner and outer lites. The wider the delta, the narrower the airspace needed to limit convection; conversely the narrower the delta the wider the airspace can be.

    Almost seems to be a bit of a “chicken and egg” paradox involved there.

  32. Bronwyn Barry | | #32

    RSVP's, PHPP and info on other PPPs.
    Greg - glad you liked the links. There are loads more on that same Slideshare site.

    Steve - the issue is not so much whether the PHPP is smart, but more about how it's structured to accept inputs. The NFRC outputs don't correspond with the required inputs for the PHPP. I had to learn THERM myself in order to extract some of the required info needed in my own early comparative studies... It would be great if it were that simple.

    All - I am encouraged by the general response to this post and Stephens' support for lobbying the NFRC to modify their temperature delta. That would be huge. (The LBNL window group is the place to apply pressure. They are smart scientists and have done the calculations for this already.) However, from my assessment, it appears that there are larger vested interests maintaining the status quo.

    The US window market consists of a handful of very large players (rather than the diversified mid- and small-sized producers that supply the European market.) These heavyweights have big production facilities set up to make the same frames they've always made (which may have the 'advantage' of being narrow, but can't hold the thicker IGU's.) Moving to a wider IGU - especially triple pane - requires a thicker frame, stronger hardware and likely some re-tooling at the production facility. Change costs money. Since there's very little incentive to improve performance beyond the minimum to meet code in the more progressive states, they'd rather keep throwing money at sales and marketing, than improve product performance. Until the NFRC sets a 'high bar' standard - like PHI does - manufacturers will continue to certify anything with glass in it, no matter how poorly it performs.

    The NFRC and Big Window Co's are unfortunately not the only hurdles to overcome. While I was developing the window profile that did achieve PHI certification, I learned that our local IGU producers are reluctant to use the newer, better performing spacers. (I specified the Super Spacer because it is US-made.) Because they don't have experience with these spacers, they don't want to take the liability for seal failures, despite the success of these same spacers on the EU market. (US air molecules behave differently, and our climate is ‘special’ goes the theory.)

    To compound this challenge, I found there were very few available options for mid-range SHGC glass packages with good VT ratings. Even the multi-national glass companies with outlets on both continents sell glass with completely different coatings and performance specs on the US market, than what they sell on the EU market. The federal tax credits provided for low-SHGC glass for commercial buildings in the lower states bears much responsibility for why North American glass production is skewed towards certain glass availability. (Clearly someone decided that it was easier to use glass to block free heat gain, than it was to teach architects and builders to design decent shading for buildings.)

    When I did find a combo that had the performance and visual specs I wanted, and we went to order it, we were told they only make that in Texas, in 6mm panes for commercial projects, and we needed a huge order to get it. Deeply frustrating!!!

    There are however, valid reasons to be cautious. Wider 3-pane IGU's actually do behave differently than narrower ones, and very differently than 2-pane IGU's. Which layer the metal low-e coatings are placed on is also important. The EU has figured all this out over the past twenty years and has come up with various combinations that work for their climate and performance requirements. We'll have to do the same here and be prepared for a few panes of cracked glass and a few seal failures.

    The bottom line is that change is hard and whining about it doesn't make it easier. Getting on with it and figuring all these 'kinks' out is actually really interesting. The window I designed has not yet made it to production for these and a few other reasons. That doesn't mean I'm giving up. I hope Stephen and all the other wonderful members of the North American window manufacturing community don’t give up either, but keep sharing great info with a market that is still learning how (and why) to demand better windows.

  33. GBA Editor
    Martin Holladay | | #33

    Response to Bronwyn Barry
    Thanks for the great post -- full of lots of interesting information.

  34. rdhvicenergy | | #34

    Further Technical Information & Discussion on NFRC vs ISO
    Hi - for those interested a few years ago we completed an in-depth well peer reviewed study of NFRC vs ISO standards as a spin-off to work we did in Canada on the ER Study. There is even more to the picture than is nicely summarized in this article. The full reports for those interested in digging deeper are available for download here:

    A short presentation from Brittany Hanam presented at the 2013 Passive House North Conference is also linked for those not wanting to read the reports:

    Key conclusion is that window and SHGC/g-values U-values determined from both standards will be different and neither is necessarily wrong. From a glazing perspective European IGUs are optimized for wider argon filled gaps, while North American IGUs are optimized for narrower argon filled gaps between the lites of glass. This is due to different calculation methodologies and outdoor air temperatures. There are lots of other differences including solar heat gain calculations as well. The study highlighted the need for climate specific window design and considering outdoor temperatures when determining, something that can be done using either standard.

    I would suggest we a community try to adopt some sort of U-NFRC and U-ISO (subscript) terminology when discussing windows at online venues like this. This will help avoid confusion as U-NFRC does not equal U-ISO. Modeling in THERM and other programs can calculate either, though with NFRC, some laboratory validation is also required as part of NFRC100/200.

  35. jinmtvt | | #35

    incidence angles
    For n00bs such as meself, took some time to find the right info,
    and since we are discussing glazings actively here.

    Here is a snap of a study paper i found :

    Sensitivity of Fenestration Solar Gain to Source Spec~rum and Angle of Incidence
    W. Ross McCluney, Ph.D.
    Member ASHRAE

    Was pointed out by many more experienced members in the past , the relation to SHG and sun angle on the glazing.

  36. Stephen Thwaites | | #36

    BBB Rating
    Until about 10 years ago, i believed all of the IGU durability claims of the firm that makes our preferred spacer. Tragically, it turns out, that these claims only apply to certain sealants. So we have about 5 years of a very large number of extremely premature seal failures.
    The problem very quickly emptied our financial and our emotional larders. It's been a struggle ever since. The fire at our pultruder and subsequent 8 month shutdown didn't lighten the load. Neither did the recession.
    Having said all that, there is a certain nobility in the struggle and the hope of better days ahead.
    I expected this piece to bring comment on our burdens, but I thought it needed to be written.
    Nobody was really clarifying why Euro windows seemed so much more energy efficient to GBAers. I felt compelled both to clarify and to explain why N. American low-e's were not inferior.
    I also wanted to show the utility of PHPP or equal software in selecting windows – that window rating systems alone don't always lead to the window that results in the lowest energy bills.
    At the end of the day atmospheric carbon is the problem. In my view, the more we know about ways to reduce emissions through lower energy consumption the better.

  37. Stephen Thwaites | | #37

    Although my piece was is about glass, spacers aren't exactly off topic, so I though I should add some context/spin to Peter L's comments on spacers (back in reply #18).
    About 25 years ago nearly all spacers in North America (and Europe) were aluminum. Today, I believe aluminum spacers are less than 10% of the North American market. I would suggest this is in large part to SuperSpacer, a foam rubber spacer that was introduced in the late 80's. When introduced, it was ignored by all the large companies, because there was no high speed equipment for it. In the early years it was really only adoptable by small, less automated shops, typically smaller window firms that made their own IGU.
    While its adoption was, and arguably still is, low, it forced the rest of the industry to face the energy and condensation issues associated with alum'n spacers and developed alternative solutions.
    Today the large firms use, for the most part, stainless steel spacers, which are warmer than aluminum and yield respectable Uedge and PSI values.
    Of course, a no metal spacer like Swissspacer or SuperSpacer yields even lower Uedge and PSI values.
    It is, however, a little unfair to infer that there are no no-metal spacer options in N.America. They are not easy to find, but they've been here for 20 years.

  38. Stephen Thwaites | | #38

    High R-value Frames
    Jin keeps asking about insulating values of frames, (mostly recently #20) and while this is a bit off topic from glazing – I think i can make a brief hopefully helpful comment.
    It is possible to concieve both of a R-10 foam block frame and an R-10 glazing unit. However, when combined they probably won't end up being an R-10 window because of bypass or 'scoot around' losses.
    A more illustrative example might be installing an R-20 VIP that would be about 1/2” thick in an opening in a R-20 strawbale wall, which itself might be about 10” thick. The overall assembly would not be R-20. This is because there are lower R-value heat loss paths, bypass paths or 'scoot around' paths near the edge of the VIP. Paths that would not involve the entire thickness of the strawbale wall.
    Both testing and simulations like THERM capture these 2-D heat transfer effects. These 2-D heat transfer effects are why frame R-values are lower than their otherwise 'theoretical' value.
    Hope that helps

  39. jinmtvt | | #39

    Stephen :
    Who are you pointing out in TO area ??
    I heard that Fibertec now makes their own IGU ??

    In a frame like the one you use on your fixed units,
    the IGU is help up in a "void" section of the frame where insulation is at its minimum.
    I've seen several details of european windows where the IGU is sitting much lower inside the frame, with insulation close to is "hiding" the spacer section in the frame insulation.

    So my guess is that there is less loss of R value with that type of design.

    Martin : true, had not thought about the ratio of glazing VS frame for SHG ...
    This is still climated/location related though.

    I would like to ask if i may,
    when calculating SHG, can we use the mean daily insulation energy directly or is there better data out there wich considers the effects of angles and spectral distribution etc ??

    The SHG factor on a glazing, how exactly is it calclated at NFRC ?
    Is the sun angle factored in ? Can we use the number directly in rough calculations ?

    I love discussing windows tech info, keep this going Stephen,
    very nice info.

  40. [email protected] | | #40

    spacer trivia
    Thanks again, Bronwyn.

    I would agree with Stephen that less than 10% of North American companies still use aluminum spacers. Personally I wouldn't be surprised if it's not much more than 5%.

    The primary spacer system used in North America is probably PPG's Intercept spacer.

    It has been around a long time and it is more or less reasonably reliable (in large part based on the quality of the IG fabricator), but it isn't quite state-of-the-art.

    While Intercept is tin-plated steel, PPG has an upgraded version called Supercept which is stainless steel. Using stainless has improved the spacer's energy performance numbers and reportedly has also improved durability/reliabilty.

    Edgetech, the folks who brought Superspacer to market, really did a lot to introduce the discussion of spacers when discussing window systems. They also brought one heck of an advertising campaign to accompany their spacer system.

    It's quite common to see mention of Superspacer when discussing glass and spacer options, but what most people don't realize (I suspect) is that there are actually five variations of Superspacer available and they don't offer the same energy performance numbers between the different versions.

    Most spacer systems available today are installed using dual-seal technology - one seal air/moisture and one structural. Superspacer is unique in that their structural seal is internal and their air/moisture seal is exterior.

    Superspacer uses an acrylic seal for structural and "allows" a variety of different materials to be used for the air/moisture seal. In my opinion most of the options for air/moisture seal should provide excellent results, but a couple of them are questionable for long-term reliability performance.

    Also available in the "non-metallic" spacer arena are Duraseal and Duralite.

    Duraseal is a direct descendent of "Swiggle" For those who are not familiar with Swiggle...I will say that there was probably a good reason for the development of a follow-up version.

    Duraseal has a descendent of it's own called Duralite.

    Although they used to be separate companies and competitors, both Superspacer and the Dura-spacers are now owned by Quanex , so former rivals are now family.

    Prior to having the same ownership both Duralite and Superspacer claimed to have the warmest edge in the industry. Now that they are no longer competing how does Quanex advertise which spacer has the warmest edge? An interesting problem for the marketing folks.

    Cardinal is by far the largest provider of stainless spacers, with (as advertised) 500,000,000 currently under warranty. Guardian has recently announced that they will be entering the IG business with an IGU using stainless spacer of their own.

    Cardinal also claims the lowest seal failure rate in the industry at .2% over 20 years for their XL spacer. Cardinal recently introduced the Endur stainless steel spacer as an improvement over the XL in energy performance with the same long term reliability as the XL.

    So how do the spacers compare per warm edge performance?

    The best (warmest) superspacer variant and duralite are pretty close at the top spot.

    Then a couple more superspacer versions, endur stainless, the remaining superspacer versions, duraseal, XL, supercept, swiggle, intercept, aluminum.

    While the rankings, as I listed them, are based on real numbers obtained thru testing, it wasn't my research or testing, so I am not claiming that they are gospel, but I do believe that they are accurate.

    Also keep in mind that the edge temperature differences between the various spacers can be very slight. For example the difference in edge temp between the top performing superspacer and the endur is less than a degree and when you are discussing triples the differences between the different "warm-edge" spacers become pretty much insignificant to both edge and overall performance.

  41. jinmtvt | | #41

    Greg :
    Any data/info on what kind of effect different recent spacers have on heat loss ?

    What happens to the spacer performance when it is hidden further down in frame insulation
    as many german windows are designed for ?

  42. Stephen Thwaites | | #42

    Sourcing glass, more frame info et al, reply to Jin (#40)
    Sourcing glass
    The Toronto 'glass scene' i mentioned is the wholesale market. It`s tough for individuals to buy from them. If you are still interested in more info contact me through our website.

    Voids in frames can be filled. This improves window R-values for triples. Sometimes voids need to be left open to allow room for hardware or for water to drain. The voids you described, if filled, can, albeit very slightly, reduce building energy consumption. It has been done to help buildings that need decimal place improvements to hit PH targets.
    Recessed Site Lines
    In most windows the top of the spacer lines up with the site line of the frame. Unsurprisingly, there is a thermal benefit to recessing the top of the spacer below the siteline. The more it's recessed the warmer the edge of glass. However to reduce thermal stresses, few window makers are recessing the spacer much more than 6mm (1/4").
    Calculating SHG
    I believe NFRC calculates SHGC at normal incidence. It is easy enough to check, all their standards are on their site.
    There are stnd formulae for converting horizontal radiation data into vertical surface data for any given hour (or minute). Given the wide range of simple to complex softwware, it may be easier to use an existing pkg than research and write your own code/spredsheet. For something simple try NRCan's HOT2XP, an express version of the time tested HOT2000 software.

    Hope that helps

  43. Jason Peacock | | #43

    North American vs European Windows
    I like European Windows.

  44. Stephen Thwaites | | #44

    Sourcing Glass
    Bronwyn's latest post (#33) mentions the challenges in sourcing glass.
    Perhaps one way to find more energy efficient glass suppliers is to contact warm edge spacer suppliers for names of glass firms in your area.
    Sourcing glass is basically a simple demand problem. If there is enough demand the glass fabricators will build triples, with whatever low-e coatings & spacers you want. But they have to see the demand to invest their time and their money. In general, and unsurprisingly, smaller firms will be more flexible than larger firms.
    At some point, if there is enough demand, multiple firms in a market will step-up and pricing become more competitive. This has only recently happened in the Toronto area. Hopefully it is also happening in other areas....

  45. jinmtvt | | #45

    Look at how the "spacer
    Look at how the "spacer section " of the IGU is tucked in the frame and is insulated at least somewhat
    on the euro vs the NA window ...

    i have never compared this directly in THERM, but it must have some impact on the weakest point of the glazings .

    Most high end windows from Deutsch manufacturers have very close fit tolerances on the contour of the glazing unit to the frame, whereas most of the NA windows " including the fiberglass ones "
    are designed to be assembled with a free space all around, and is often uninsulated also.

    It exposes the edges and bottom of the spacer to the exterior temp at a higher degree,
    and i see no advantages to this method . ( at least fill the empty space with backer rods or something that blocks air )

    First time i realize the glass is much thicker on the euro glazing,
    is this the norm ? Must be hell of a lot heavier and tougher.

    The size difference of the frame is almost bogging though.
    This is not the best euro designs i've seen though,
    some are shorter and feature more insulation within the frame empty spaces .

  46. Stephen Thwaites | | #46

    Section Comparison Comments

    There is no question Euro style windows whether made in Europe or N.America are well made. There are lots of reasons to like them. In my view, some of the reasons listed in Jason's photo (reply # 45 ) are not as accurate as they could be. Here's my take on them;

    Costs, due to the effects of among other factors; options, shipping, currency exchange rates, perceived risk and general competitive pressures, pricing fluctuates.

    Not sure of the basis for this comparison, but would love more info. WINDOW 7.1 assumes 90% argon when it calculates Uglass for either NFRC or CEN temperature scenarios.

    Glass Thickness
    I am not aware of any N.American Casement that uses less than 3mm thick glass. In Canada 3mm is the legal minimum for all windows. It is used for glass areas up to 15ft^2 without issue. I believe in the US, some smaller Double Hung windows might use 2.2 or 2.5 mm glass.

    Air Leakage
    The key to good air leakage numbers is a center seal, which is present in both cross sections (as it is in most variants of hinged windows). Dig a little and you'll find that Casements on the PHIUS window list have air leakage rates of 0.01-0.02 cfm/ft^2. T&T windows on the same list have air leakage rates of 0.05- 0.01 cfm/ft^2. So both types of windows are tight.
    I suspect the 0.08 figure for the N. American window is based on our info. If it is, the units for the 0.08 are m^3/h/m of crack length. Converting 0.08m^3/h/m of crack length into more conventional units, the air leakage becomes a respectable 0.02 cfm/ft^2.

    Thermal Breaks
    A good way to quantitatively assess the thermal quality of a frame is with Uf. Based on PHIUS data, the respective Uf's in Jason's image are 1.3 W/m^2K and 1.1 W/m^2K. By this measure the Euro frame insulates better that the N.American frame – about 15% better.
    In my view, a better way to quantitatively asses the thermal quality of a frame is to multiply Uf by the frame height to get a conductance or heat loss per unit length of frame. For this example the respective conductances would be 0.09 W/m/K and 0.13 W/mK. By this measure the N. American frame loses fewer Watts than the Euro frame – about 30% fewer.
    So, interestingly, despite insulating about 15% worse, the it loses about 30% less heat.

  47. jinmtvt | | #47

    Stephen :
    The heat loss related to frame length comment ... nice thinking :)


    we are comparing a high performance unit from NA to a mid performance unit from EURO .
    Most good windows from Germany are in the .8 to 1 wm2k .

    But in any way , this is splitting hair in 2 .

    Stephen since you are in the window business,
    would it be possible for your to give us an approximate pricing difference
    from using lesser performing glazings up to the Climaguard/180/400 high end ??

    I mean, < 0.6wm2k is imressive figures, but just how much need we to pay for going up to those high end units ??

    Maybe a list of available high performance glazings with their SHGC and wm2k could be a nice addition to this thread ... hard to understand where we are coming from for uninitiated such as myself.

    Lastly, when using glazing which is better performing that most frames,
    narrower frames such as the ones used in NA must have an even higher advantage,
    as pointed out by Martin about differing glazing total area.

    I still do not understand though how the NA window depicted in JAson's picture,
    does not "shortcut" the glazing through the spacer since it is basically rested against exterior window stop without any insulation.

  48. AntonioO | | #48

    Stephen: frame length?
    Please define frame length. And what do you take as the frame lengths in the two cases pictured?

    Also, is U an intrinsic property or quantity? For example, when we talk about resistance to current flow there is resistivity and resistance. Resistivity in this case is the intrinsic quantity and for a given cross sectional area (perpindicual to the direction current flow) of material the resistance of the material will increase as the length along the direction of current flow increases. Therefore resistance is an extrinsic property (not intrinsic) as the number measured depends on the amount of material resisting current flow. Moreover, a resistance measurement already has accounted for the dimensions of the material. I had always thought that U was an extrinsic property that already took into account the dimension through which heat was flowing. Is my thinking flawed? By multiplying by a length dimension you seem to be implying that U is an intrinsic property similar to resistivity in current flow and unlike resistance in current flow. Please explain.

  49. jinmtvt | | #49

    the euro frame being "thicker" or wider , will have more frame surface for the same total size of window

    If your frame has higher performance than your glazing then its all good,
    but in the case of high perf glazing, the frame is a loss in total performance.

    the wm2k is directly proportional fo the surface erea of the calculated element

  50. AntonioO | | #50

    So when you say frame thickness do you mean the same dimension that Stephen refers to as frame length? It's not clear to me that this is the case. I would think of thickness as the dimension from the outside of the window to the inside. And that is the direction of heat flow. But perhaps you refer to thickness as the distance from the edge of the spacer to the outer most edge of the window frame. That dimension is important in determining the fractional cross sectional area of a window opening which is frame as opposed to IGU. If that was the point, it at least now makes some sense to me. But I would still have to ask what dimensions were used for each window. The only dimension I see in the photo is glass thickness. I can't say what any of the other dimensions are.

    I also think it's not necessarily correct to say that for "high performing glazing" that the "frame is a loss in total performance." . It's correct to say that the frame will add no heat gain value, but it's relative performance compared to the IGU would depend on how the frame is constructed. I would think it rather easy to build a higher performance frame by filling those open voids with insulation. Perhaps someone will explain why this may be much more challenging than it seems.

  51. GBA Editor
    Martin Holladay | | #51

    Thickness, width, and length
    If this drawing depicts the cross-section of a window frame -- and frankly, I don't know if it does -- I would say that the frame is 60 units thick and 70 units wide. I would reserve the word "length" for the distance from the sill to the head, or from one jamb to the other.


  52. Bronwyn Barry | | #52

    CAUTION: High SHGC glass may lead to overheating.
    I've just re-read this post and see that its conclusion may upset the balance point of your PH design. While it's important to consider Heating Demand when designing a PH project, a PH is all about designing a COMFORTABLE building, so cooling and overheating must also be considered.

    Stephen writes:

    "The key point is that despite not insulating as well as the European low-e glazings, North American high-solar-gain low-e glazings take advantage of the fact that, for the most part, North American heating climates are sunnier than European heating climates.
    So not only can North American glass match the performance of European glass, but in heating climates it can also surpass it."

    While this statement is undeniably true, it doesn't take into account the negative side of using high SHGC glass. (I too was excited about our great North American insolation until I started to look at how this affected cooling loads and overheating, especially here in California.)

    There is a very good reason that typical IGU's in the EU are not much higher than 0.5. (It’s not because they don’t know how to build them.) Experienced users have found that the windows simply need to be 'net energy positive' on their overall balance, and not be used as substitute heaters. (The glass should also not be used in lieu of shading devices, but that's a whole other topic.) Even in the EU, which does not have as much solar access as we do in North America, overheating became an issue in the early projects and careful shading was emphasized. Designers learned not to be greedy with either their glass area or their SHGC. We would do well to heed this lesson.

    What I've found in our projects here in California is that it's ok to have our window energy balance be slightly tilted to the negative and keep our SHGC in the 35-45% range. We still need to carefully shade our glass, especially for the shoulder season, when we have our hottest months here in the Bay Area. Managing internal gains in PH projects appears to be much more of a challenge than optimizing solar gain, and in fact I'm finding we must carefully control our solar gain in order to not have significant overheating issues. (We don't like to have more than 1% overheating in our projects.) As Stephen correctly says, “your mileage may vary.”

    Finding ways to flush excess heat has been critical to the comfort (the ultimate success) of our PH projects. I give full credit to my business partner Allen Gilliland here, for finding ways to incorporate night ventilation cooling into our projects. The glass, frames and SHGC are all very important, but as with all PH projects, we have to step back and look at the bigger picture. We must each find the balance point for our local climate. Be very cautious if you do select high SHGC that it does not tip you into having to deal with higher cooling load or overheating issues.

  53. jinmtvt | | #53

    Antonio and Martin :
    I also thought he meant the cross sectional since discussing heat loss.
    It is obvious from the pictures that the "width" is less on the NA window,
    thus cross sectional is lower for the same "length" of framing or total window size
    not sure how he came up with this numbers though ... we let him explain i guess :)

    Bronwyn Barry: nice input, but isn't this very location related problems ?

  54. Bronwyn Barry | | #54

    More solar availability in North America...
    means more 'free' heat but also more opportunity to overheat and then need to pay for cooling. (Just because we can do something, does not mean it's always a good idea.)

    Restraint, in the case of SHGC, may be the wiser option? (I write that as a question because I haven't modeled this for a variety of climates, so am simply speculating here and urging caution. Guaranteed there are exceptions.) My point is that the 'big picture' should always be considered and is the joy and strength of the PHPP and whole point of Passive House. Miss that and you've missed everything.

  55. Stephen Thwaites | | #55

    Terminology, Conductance
    If anyone can find an official naming scheme for the various frame and window dimensions i'd sure like to see it. In the mean time, I suggest we start with the terms shown in the diagram below. Hopefully that helps clarifies the terms I used when desscribing Conductance (W/m/K)

    If conductance is a touch confusing ignore it and just compare total heat loss. Pick a window size, calculate the frame area, pick a temperature difference and calculate the heat loss through the frame
    Heat Loss = U* A * temperature difference

    Conductance = U* h

    So for the N.American frame
    Conductance = 1.3 W/m^2K * 0.068m
    = 0.9 W/mK
    and for the Euro frame:
    Conductance = 1.1 W/m^2K * 0.120m
    = 1.3 W/mK

    hope that helps

  56. Airithol | | #56

    Conductance / greed

    Thanks for starting this useful thread.

    I can see from your example (post 56) that the NA frame has a lower conductance value.

    For my own spreadsheet, I have been using UAΔT to calculate heat loss, and as you predict, frames with lower insulating values (higher Uframe), but smaller area are lower in overall heat loss (of frame).

    Conductance, (Uframe * Height) seems like a comparison metric (similar to 0.6ACH50?) to me. Is there a direct use for this value, an alternate way to calculate heat loss? Can you extend your example a little?

    In my own example (in feet, forgive me...) I have a window frame with Uframe 0.1670 and a height of 2.75".

    This gives me 0.1670 * (2.75/12) = 0.038 conductance. What then?


    Thanks for your insightful posts (and links). Your comments on not being too greedy are spot on. I am in the middle of the exact balancing act you describe and the temptation to push for higher solar gain is difficult to resist.


  57. jinmtvt | | #57

    i hate to point to it once more ..
    but most of the overheating in many climates is due to missing shades
    Look at the Lancaster example from the picture in this thread..

    Where is the south shading ???
    this is basic error that should never happen in a PH level building.

    And there are existing solutions for south-west/west shading that are relativly inexpensive .

    Some climates might be more difficult to tune for, but if the proper shadings are not in place, you will have to choose the wrong and expensive compromises right at the begining of the project.

    Stephen: is frame height usually considered at the highest point ?
    As per the last detail you posted, the interior side of the frame is shorter than the exterior
    leaving more glazing to air which side is taken into account when going for calculations ?
    it's probably insignificant difference in result, but being here would be fun to know if there is a standard :)

  58. Stephen Thwaites | | #58

    Conductance, cont'd
    The best way to compare window performance is to run PHPP or similar software. This is especially true when comparing different approaches, like Euro inswing and NA outswing. Window Rating schemes, do not neccessarily, lead to window decisions that produce the lowest energy bills.

    Conductance is an illustrative tool to show how optimising Uf, does not neccessarily lead to the lowest energy bill. It avoids the messiness of calculating frame areas. But it is only one tool. The best tool is to run a whole building energy simulation.

    A building is a system. Optimising all sub-systems may not lead to an optimal system.
    Run PHPP or similar software and to find your optimal for your building in your climate.

    Jin: frame height is taken to the highest point. The dwg Martin provided and i nomenclatured is missing both glass and a roomside glass stop. The roomside glass stop usually tops out at the same height as the exterior glazing leg of the frame.

  59. jinmtvt | | #59

    thanks for confirming the measurement standard.

    Problem with using softwares is the involved labor and complexity to get accurate results,
    and the iteration required to measure possibilities.
    Most regular folks do not posess the required knowledge and the interest for the simulations,
    nor the budget to have one do it.

    So then we must use generalize assumption from other's projects .

    It is easy to do as a designer/entrepreneur of a building project..and mostly required.
    Much less for the regular builders that live with owner's limitations.

    Slowly crawling back to my conception that to move building efficiency forward in mass,
    regulations and entrepreneur need to step in ..not indifidual home owner,

    thanks again Stephen for pointing out once more that Uf is not the only value to take in account when looking for hole window performance.
    Although i believe that Uf/Uw includes this to a degree for a fixed size ?

  60. AntonioO | | #60

    Diagram helps
    Someone once said a picture is worth a thousand words. I must say that a diagram is pretty valuable as well. After seeing your diagram and nomenclature I now understand your comparison. Your units initially confused me and made me believe you had confused thermal conductance (extrinsic) with thermal conductivity(intrinsic). A two dimensional cross section to explain a three dimensional problem sometimes leads to such confusion. I assume also that your numbers for window dimensions are probably based not on the photo posted by another commenter but on the two windows you were originally comparing in your blog post. Anyway thanks again, Stephen.

    This has been a highly illuminating and opinion changing conversation. Before I had assumed that there was no reason to consider "overly expensive" European windows when I had a much better value proposition right here in North America, especially Canada (and including Thermotech). I now question that belief. Perhaps the Europeans offer a worthwhile value as well. I have more homework to do.

    If Martin is still reading. I would ask that he consider inviting other manufacturers in North American and from across the pond to respond to this blog and provide the pros and cons of their respective products.

  61. jinmtvt | | #61

    Antonio :
    would you please explain your change of believes ??
    I believe that Stephen was trying to point the contrary to what you just stated ...
    So what makes you now think that EURO windows should be given a thought ?

  62. GBA Editor
    Martin Holladay | | #62

    Response to Antonio Oliver
    You wrote, "If Martin is still reading, I would ask that he consider inviting other manufacturers in North American and from across the pond to respond to this blog and provide the pros and cons of their respective products."

    Anyone is free to post a comment on this blog, including employees of window manufacturers. All we ask is that these representatives don't use our web site to post blatant advertisements for their products.

    Stephen Thwaites has done a great job explaining the technical issues that underpin any performance comparisons between North American and European windows. He has also refrained from promoting his window company. (Thanks, Stephen.) If any other window experts, including employees of window manufacturers, want to submit a guest blog that discusses the issues raised by Stephen, they are welcome to submit their proposed blog to me by e-mail: martin [at] greenbuildingadvisor [dot] com.

  63. AntonioO | | #63

    Response to Jin

    Several key points came out that I have to give more consideration:

    1. Perhaps Euro windows are not necessarily so much more expensive than those by the North American manufacturers anymore as pointed out by other commenters. I'll have to check that out.

    2. Solar heat gain is not as important to me as how well the product as a whole insulates. It's basically a site and orientation issue. Solar heat gain as a benefit is literally a one sided issue as I see it.

    3. Going back to my initial comment, all else being equal (which I know isn't the case), give me the larger inter-pane spacing any day of the week. With a noble gas as the insulation I still haven't been satisfied as to why the IGU with 12.5mm spacing will perform the same but the IGU with 18mm spacing will perform worse because of convection when the temperature outside drops from 273 to 255 Kelvins. Admittedly, it's been some time since my life revolved around doing physics problems so I'm a bit rusty, but that one just still leaves me fogged.

    4. I have to go back and pay attention to air leakage. Not saying one is better or not but I have to go back and look at the numbers. I don't think I've given air leakage the thought it deserves, probably because it appears not every company reports numbers for this.

    5. Are all the Euro manufacturers giving up an easy opportunity to do a better job insulating their frames as in the photo posted? Jin, you seemed to suggest that this was a sub par Euro frame pictured.

    6. Is the glass in Europe really 50% thicker? That's a potential plus to a person like me whose property is adjacent to a road that's busy certain times of day. Maybe thicker glass lets in less noise.

    Having said all that, a good NA window is not out of the question for me. But it looks like the Euro windows have some properties that I value and should consider more before ruling them out.

  64. AntonioO | | #64

    Final thought..
    On the convection issue, has the same analysis been done but using argon instead of krypton as the insulating gas? That could possibly tell us a lot about the convectional flow and what drives the results seen.

  65. jinmtvt | | #65

    Antonio :
    i have no knowledge whatsoever of convection physics , but it would be interesting to learn more about the phenomenon in IGU .

    Just have a look at the PHI database :
    a few catergories contain fixed and openable windows

    Believe it or not, there are probably a few hundreds more window designs taht aren't on this site because they fail to meet the standard performance by only a margin,
    but are still very noteworthy.

    As you can see some window designs are pretty simple but adavanced in performance.

    i like very much this kind of design :

    As for the infiltration rates .... maybe Stephen could step in to explain how those are tested and confirmed from NA manufacturers.

  66. Stephen Thwaites | | #66

    The Rabbit Hole of Window Air Leakage
    Most new fixed or hinge windows are quite air tight.
    Most would have test results of 0.05 cfm/ft^2 of window area or less.
    For the record, this air leakage is measured at 75 Pa.

    By my very quick calculation;
    Even at 0.05 cfm/ft^2 of window area;
    For a house or other low rise building in Ottawa, the air leakage heat would translate into an a heat loss of 2 kWh/ft^2/yr (ft^2 refers to window area).

    For the Lancaster NH house mentioned in the column (window area was about 25% of floor area) heat lost to window air leakage might be 3% of the annual heating bill.

    Hope that helps

  67. amattock | | #67

    International Window Study
    This study on International Window Standards by RDH may bring clarity to those with questions about the different in standards between EU and NA:

  68. AntonioO | | #68

    Thanks, Andrew Mattock
    That was a great read, Andrew. One thing that was not clear to me in chapter 4 was whether the study was speaking of real world performance or simply the ability to achieve a lower U value in a calculation when the study suggests that NA windows would not perform as well as a similar product with a larger gap spacing. Do the experts know how well the calculated U values conform to real world performance? Though I trust the value of a good simulation, it seems that sometimes good experimental measurements could sort a lot of these differences out. But since I was trained as an experimentalist, I may be biased. That being said, I'd be interested to know the physical models behind the curtains of these calculation tools.

    [Edit: By physical models in that last sentence, I mean physical equations if it's that kind of model.]

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |