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

Community and Q&A

High-Performance Windows in Low R-Value Walls

washxhouse | Posted in Energy Efficiency and Durability on

I just came across a 2010 article Martin wrote for this site entitled ‘Do Europeans Really Make the Best Windows?’, in which he concludes as an almost tossed-off aside:

“The high cost of triple-glazed windows is hard to justify unless you’re building a superinsulated house in a cold climate. But once your wall specs reach the R-40 level, triple-glazed windows start to make sense.”

… This surprised me and made me question what I think I understand about building science. I’m tempted to argue from a marginal case to understand the boundaries of the logic here but it’s probably more constructive to look at this in the case of my home.

I’m in a 1987 contemporary in SE Pennsylvania (4A/5A border) with ~R19 walls, about 2700 sq ft in 2 stories above ground with an unfinished basement. I have about 550 sq ft of windows, which are about 1/2 operable casement and 1/2 fixed, all original (low-performing, leaky, cold) dual-pane with a wood interior aluminum exterior.

When I look at the heat loss composition of my manual J, I see windows at 28,500 BTU/h heat loss as the largest contributor by a factor of 3 to my total number of 71,500 BTU/h.

Granted my walls at R19 are contributing 11,000 BTU/h but I can’t do a ton about that, but according to Martin’s claim above, since my walls aren’t high-performing, it’s not really worth the cost of going with high-performing windows.

In my case, given that windows present such an enormous contribution to my heat loads in what is a heating-dominated climate, I can see for sure that there’s a benefit to be had by improving performance there, and I doubt Martin meant to imply that there is no point in getting at least middling-performing windows with R19 walls.

But am I wrong in thinking that because my place has an unusually high amount of glazing that an investment in high-performing windows would pay off, regardless of the performance of the walls?

Is it perhaps a matter of degree that’s important here, and perhaps Martin meant to suggest that I’d be better served investing in ‘good enough’ windows at the upper middle end of the performance range than paying perhaps double that cost to get what is potentially a marginal benefit from the highest-performing stuff on the market?

I could certainly see the cost-benefit skew becoming unfavorable at the highest-performing end in windows just as it does in almost every other category of product, making the highest-end stuff only reasonable for certain niche cases.

I guess what I’m feeling is – that was a very brief, very general statement with a ton to unpack that upended what I thought was a pretty clear motivation to get the best-performing windows I can afford. Now I’m not so sure. For example, should I not bother with triple glazing at all?

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. STEPHEN SHEEHY | | #1

    I guess there are two issues.
    First, should you replace the windows? You probably won't save enough money on heating costs to justify the enormous expense of replacing all those windows.
    But you'll be a lot more comfortable. New windows won't be drafty. Your house will be much quieter. You won't be wiping up puddles of condensation.
    So if you decide to replace, whether it makes sense to opt for really good windows, say U .20 or better, depends on the cost difference from Energy Star qualified, which I think is U .30. If the difference is 10% I'd do it. 40%, probably not.

    1. washxhouse | | #2

      I think this is a helpful starting point for putting together a framework for making these kind of decisions.

      Taking this further, maybe we could say once you've committed to replacing the windows for comfort and energy's sake (rather than purely for an energy-related savings ROI), you're best off doing so at some minimum U-value (and likely a max air infiltration value) that will result in some appreciable improved comfort and energy performance.

      After that, you're in cake-icing territory, where you're now talking about spending potentially a lot more to get a little more performance, but maybe not appreciably more comfort. If the cost increase for high performance is significant over the 'pretty good' option, the performance improvement would also need to be quite significant to justify it on the energy ROI side.

      Given that in my case I'm running a (new, efficient) heat pump, I'm not likely to realize really significant energy ROI, so it comes down to comfort for the most part, which boils it down to: what is the good enough U-factor to make this place feel less drafty in the winter, and make it comfortable to sit in the rooms with tons of windows? Is it 0.3? Lower?

      Given the level of investment and disruption involved, it would be a bummer to undershoot the level of performance necessary to meet that threshold.

      1. Andrew_C | | #3

        "...framework for making decisions..." Old article by Martin:

        Replacing windows has always been near the top of the pyramid. Cost of solar, various energy sources, efficiency of ASHPs change, but fundamentally windows are still going to be near the top of the pyramid. I'd only use triple pane if I had a noise issue (busy road, etc). Although, if you've got the money and have done the other stuff, there are benefits to really nice windows, as Stephen points out.

        1. washxhouse | | #6

          This is helpful. Without having seen this I'm glad to say I've been working my way up the pyramid but it doesn't feel brilliant to think about going high-endy at the top 😅

  2. walta100 | | #4

    Washx if you want to make the economic case for replacement windows it seems to me you have a very tuff road in front of you.

    The only real way to make your case is to build computer models of houses and run the numbers.

    When I built my house I did just that and could not find a ROI for the price difference between the new cost on the two windows. The numbers become impossibly stacked against your premise when you start talking about replacing serviceable existing windows.

    I do agree with Stephen that the extra cost can more easily justified as a creature comfort.

    Another comfort issue is the fact that the surface temp of the window glass set the winter humidity level for your home. The dew point of the air in the home will never exceed the surface temp of the glass.

    If you want to build the computer model check out BEopt software it is free to use but has a learning curve. You can change the weather interest & inflation rates and the cost per sqf of different windows.

    I am not sure I want to contemplate the dystopian world of hyper inflated fuel costs it would take to give replacement windows a ROI.


    1. washxhouse | | #5

      That does simplify things, and leaves the focus on 'if you're going to replace windows for comfort, what's the good-enough performance level that's going to tick the box?'

      IIRC I read in some NFRC literature somewhere that ~60 deg F interior surface temp is the point at which it becomes uncomfortable to be near windows ... is there some math I can do that says 'for your window to stay warmer than 60 X% of the time, you need a U-factor of Y'?

      1. pjpfeiff | | #7

        One pretty simple and useful tool is at

        1. washxhouse | | #8

          Dang that's a cool tool. Thanks Paul.

        2. AC200 | | #9

          Interesting tool. It states the calculations are in the absence of perimeter heat. I would take that to be no forced air heat duct registers at the window locations where they are almost always placed to combat draft and radiant discomfort.

          If this is the case, then the tool calculates the worst case scenario and perimeter heat would improve things.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |