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

Window Installation Tips for a Deep Energy Retrofit

Use ‘Dudley boxes’ for a hassle-free way to install new windows in walls that are being retrofitted with exterior rigid foam

We built a plywood “Dudley box” for each rough opening. The Dudley box includes an integral plywood flange, and the flange is the same thickness as the vertical furring strips that will later be installed over the rigid foam. One key aspect of the Dudley box concept: the box is sized to match the window, not the rough opening.
Image Credit: All photos by Joel Schuman
View Gallery 17 images
We built a plywood “Dudley box” for each rough opening. The Dudley box includes an integral plywood flange, and the flange is the same thickness as the vertical furring strips that will later be installed over the rigid foam. One key aspect of the Dudley box concept: the box is sized to match the window, not the rough opening.
Image Credit: All photos by Joel Schuman
The rigid foam is installed right over the window rough openings. You don't want to cut window openings in the rigid foam until all of the foam has been installed. The openings in the foam are established with a laser level, to make sure that windows line up on the exterior of the home. The rough window sill is flashed with peel-and-stick tape before the Dudley box is installed. Daniel Dudley assembles a Dudley box. The inside of the Dudley box should be about 3/8 inch wider and taller than the outside of the window frame. The 8-inch-wide flanges are made from 3/4-inch plywood. Cross section of a Dudley box Caulk the back of the flange before you install the Dudley box. Daniel Dudley has finished assembling one of the first Dudley boxes. Daniel Dudley (left) and Woogs O'Keefe (right) are installing a Dudley box. To secure the Dudley box to the house, attach the box's flange to the rough opening studs with long nails or screws that extend through the rigid foam. There is no need to attach the sides of the Dudley box to the rough opening jambs. A Dudley box has been installed in one of the three second-floor window openings. Woogs O'Keefe installs a window. Once the Dudley box has been secured in the rough opening, installing the window is a snap.
Image Credit: All photos by Joel Schuman
The sixth window has been installed. The aluminum window flanges have been flashed with peel-and-stick flashing. The edges of the peel-and-stick flashing — where the flashing meets the rigid foam — have been reinforced with Dow Weathermate tape. The interior trim can be assembled in the same way as a plywood Dudley box. In other words, you can make interior Dudley boxes, too.

In May 2011 we began a deep energy retrofit of our old, cold, drafty house in Saugerties, New York. Because the house was poorly and cheaply built in the 1840s (apparently from scraps and salvage), we were leery of opening up the walls from the outside, lest we find that the clapboard siding and incomplete sheathing were all that had kept the house from collapsing.

Because we were happy with the interior after 25 years of piecemeal renovation, we couldn’t open up the walls from the inside, either. Instead, we opted to wrap the house with two staggered 2-inch-thick layers of rigid polyiso insulation, carefully taping the seams to form a water drainage plane, screw furring strips through the insulation to the studs, and finishing with fiber-cement siding.

We replaced all the doors and windows. For windows we selected triple-pane, argon-filled, low-e, fiberglass-framed units from Inline Fiberglass. We ordered the windows with factory-applied aluminum nailing flanges (which Inline refers to as “brickmold” flanges). These were not replacement windows that replace the sash in an existing frame. They are the same complete units used in new construction. The old window frames weren’t in good enough shape for us to even consider “replacement” glazing, nor were the window-to-wall connections sound enough to work up as part of a tight shell.

Historic renovation specialists tackle their first deep-energy retrofit

Our carpenters, Dan Dudley and Bob “Woogs” O’Keefe, had previously specialized in historic renovations. If it hadn’t been for the economic downturn and the slowdown in their high-end work, I doubt they would have considered tackling our modest project. Luckily for us, they took the job and brought us not only their broad experience but their creativity, insight, and common sense.

That none of us had ever done this before didn’t faze anyone. We didn’t see this as a pioneering effort and we expected to find well-documented procedures and techniques each step of the way.

Where can we find step-by-step instructions?

But when it came time to install the windows, we were shocked to find little usable information on how to actually do it. What little we found was mostly idealized cross-section drawings of a completed retrofit, with energy-nerd analysis of resulting benefits. All the building science and how-to articles seemed assume replacement of not-so-old windows in relatively recent construction, with standard framing in good condition.

We found no discussion of how to replace old, deteriorated windows in old, shoddy, irregular construction after re-skinning the exterior with thick rigid foam. That is, nothing of use to the carpenters working on our house.

So we figured out our own simple, repeatable, method that gracefully handled the challenges we found in installing thick foam and new windows in old construction. These challenges included:

  • Irregular walls which have racked, sagged, twisted, bowed, and gone lumpy from unknown causes (we resisted the temptation to explore any more than necessary).
  • Walls of varying thickness due to additions and modifications.
  • Old windows not uniform in size, construction or installation details.
  • Existing rough openings apparently framed without use of squares, levels, rulers, tapes, or plans. Even windows which once lined up were now cocked, skewed, racked and misaligned.

Typical extension boxes take a lot of shimming

Instructions from the window manufacturers for new windows assume that the rough openings (ROs) are close to perfect. You order new windows’ net or heel size 1/2 inch smaller than the RO and shim to fit, screwing through the window frame to the RO studs.

Typical retrofit detail drawings for rigid exterior foam insulation add an “extension box” built in place by nailing strips to the jamb studs. Then you shim the window to the extension box and screw through the side of the window frame to the extension box (or all the way through to the studs).

That couldn’t work for us, as some of our existing ROs had four different side lengths. Many were cocked more than 1/2 inch, and some were out of line with the neighboring ROs by more than an inch. Dan and Woogs recoiled from the idea of so much shimming and reframing, not to mention a stick-at-a-time extension box for each window, and we decided there had to be a better way. In addition, we were concerned about the lack of a positive seal between extension box and the studs, and between the casing, rigid foam, and extension box.

Cut loose from the rough opening

Our key insight was that if we built an extension box that had enough strength and integrity to stay square and planar during the installation process, we could ignore the RO and not try to correct it.

To do this, we built extension boxes sized to the window, not the RO, with a wide, integral plywood flange. The plywood nailing flange is built from 3/4-inch plywood, so it has the same thickness as our furring strips.

We positioned the windows on the wall as if the wall were perfect. We screwed the box through the flange to the building face, into the edge of the RO studs, instead of through the side of the extension box to the side face of the RO studs.

In addition to being much simpler and faster, this technique provides a blast-proof air seal and a deluge-proof connection to the rigid foam drainage plane.

Step-by-step instructions

Measure the new windows. There’s a trade-off here. The more clearance between the new window and the old RO, the easier it will be to correct for alignment and position problems. But smaller windows mean less glass, less light, less solar gain and less view.

We settled on no less than 1 inch of clearance, all around, but if I had it to do over I’d make it 1 1/2 inch to make it easier to apply spray foam. So net or heel size should be 2 to 3 inches smaller than the RO. If you are replacing a bunch of windows that are the same size, a smaller heel size will make it easier to line up the windows later.

Although our old windows were double-hungs, we chose casements for our new windows. Sure, casement windows are substantially more energy-efficient, but the advantage that is relevant to Dudley boxes is that casement windows let in more light and offer a better view. Since using Dudley boxes can reduce heel size — especially if replacing a row of misaligned windows — switching to casements recovers or even increases the glazing area. (We also prefer the modern, European look of a plain casement mounted close to the outer wall plane.)

Make the Dudley boxes. (Although the box design was very much a three-way collaboration, I think Dan contributed the most and deserves to be immortalized by having these named after him.) The sides of the box are made of 1/2-inch CDX plywood. The face flanges are 3/4-inch CDX plywood (to match the 3/4-inch furring strips on the rigid foam insulation).

The internal dimensions of the box are crucial. The box should be between 1/4 inch and 3/8 inch larger than the heel size of the window. Tighter isn’t better. Ideally, the gap will be wide enough that you can run the tip of a spray foam applicator along all four edges after installation and completely fill space between the window frame and extension box.

If the extension box is too tight, you risk voids in the spray foam, compromising the thermal barrier and air seal. The depth of the extension box is determined by wall thickness. You want the inside edge of the box to be 1 to 2 inches shy of the inside wall plane. Don’t try to make it flush to the inside wall; the extension box isn’t directly tied to the interior casing.

The width of the face flange should be around 8 inches. If this seems too wide, keep in mind that in addition to securing the window to the wall, it will be part of the drainage plane and will be the nailer for siding.

We made the Dudley boxes on site with a table saw and nail gun. We nailed and glued the sides together, roughly squared them up by diagonals, and then applied the flanges, again with glue and nails. The 1/2-inch side pieces are nailed to the 3/4-inch flanges. (Fortunately, it doesn’t really matter if some errant nails stick out a bit.)

Make sure the flange pieces are cut square, and tight joints between flange elements will keep the box square. Be generous with the waterproof yellow glue, as a good seal of the extension box joints could pay off in reduced air infiltration if your spray foam seals aren’t perfect.

Install the rigid foam insulation. We found it easiest to install the windows as we installed the foam. Sure, the ladders and scaffolding were already in place, but more importantly we applied full sheets of foam over the window openings. Since we were living in the house during the renovation, we didn’t want to darken whole rooms and then come back days later to do the windows. So we did each wall in two passes: downstairs, then upstairs.

First, we took out the old windows. Then Dan and Woogs inspected each RO and replaced any rotten framing. (There was more than we expected).

Then we applied both layers of foam, staggering the joints where possible, taping all the seams (double tape on horizontals), and securing the foam to studs with furring strips and 6-inch or 8-inch HeadLok screws.

Our usual practice was not to cut the window openings before applying the foam. Sometimes, however, a partial sheet of foam would suggest covering part of a window opening, so we weren’t absolute about this. Again, it’s important not to cut the foam back to the RO at this point.

Lay out the new window positions on the foam. Here’s the crucial innovation of this technique. First, we transferred the vertical and horizontal center of the existing RO from inside to outside by sticking an awl thru the foam from the inside and then transferring the distance to top and bottom to the outside. Where walls had multiple windows in a row, we did all the windows at once.

Using a laser level, we established the top line of the new windows by determining the lowest top point of the existing ROs and going down 1/4 inch to 1/2 inch. That’s the top cut line through the foam.

We measured down from that top line exactly the outside height of the Dudley box. That’s the bottom cut line. Next we established the vertical center line of each window and measured out from it exactly half the outside width of the Dudley box. Those are the vertical cut lines.

When that’s done, you have the cut lines for the Dudley boxes marked on the foam. It’s time to step back and eyeball those black rectangles. We say, “If it looks right, it is right.”

Fit the Dudley box. Cut through the rigid foam on the lines. (We used a 6-inch drywall taping knife sharpened on one side edge.) Each opening should be exactly the size of its Dudley box.

Test fit it. It should be a press fit, and it might take some effort. Don’t push it all the way in yet; just make sure it starts okay. Then remove it and apply three parallel beads of caulk on the foam around the opening. If you apply one fat bead it won’t squeeze flat, but three smaller beads will squeeze easier and merge together. (This caulk is just one element of the air barrier connecting the windows to the wall. It is not the primary drainage plane water barrier.)

Secure the Dudley box to the studs. From the inside, measure from each edge of the Dudley box to the nearest stud, header, or sill plate. Transfer the measurements to the outside face of the flange. Secure the Dudley box to the framing with sturdy screws thru the flange and foam.

Do not screw through the sides of the Dudley box to the RO framing. Even if you shimmed perfectly, the box would be likely to distort. It’s important not to stress and deform the box.

Don’t worry about strength; the screws through the flanges are plenty strong enough to take the weight of the window, and you’ll be adding even more support later. As with the furring strips, the long screws to the studs put the foam under compression, with the result that the Dudley box is strongly anchored to the wall.

Mount the window. Drill the aluminum window flanges (the brickmold flanges) with countersunk holes spaced according to the manufacturer’s specification. (This is usually 4 inches from corners, 8 to 12 inches apart along edges). We did this on the ground before hoisting the windows to the scaffold.

Securing the window takes two people (at least, depending on size): one inside and one outside. Lay beads of caulk on the inside of the aluminum flanges. Again, this caulk is part of the air barrier, but it does not come into play as a water barrier unless the primary drainage plane has failed.

Position the window in the Dudley box, with one person inside and the other outside. The inside person puts shims between the bottom edge of the window frame and the Dudley box. The Dudley box won’t deform because it is evenly supported by the foam where the window bears. If the box has not racked or twisted, a few shims should give an even gap all the way around the window.

Check the window for plumb. If the opening in the foam was cut plumb and square, the window should sit perfectly in the box. If the box was sized accurately, the gap between box and frame should be even on all sides.

Put one screw at each corner of the aluminum window flange, into the plywood flange of the Dudley box. Check the window for smooth operation and troubleshoot or adjust as necessary. (I think we had to slightly adjust 2 out of 28 windows.) Screw the rest of the aluminum window flange to the flange of the Dudley box.

Disregard the manufacturer’s instructions to screw through the window frame to the studs (or extension box). The factory-applied aluminum flange is plenty strong enough secure the window to the wall, and you’ll be adding even more support shortly.

Flash the window. The idea is to continue the drainage plane formed by the foam’s outer foil layers over the plywood flange and over the aluminum window flange. We used 6-inch self-adhesive butyl tape in the following sequence:

  • One piece at the bottom over the aluminum window flange to the edge of the plywood flange, vertical over the aluminum brickmold;
  • Two pieces over the sides: first over the aluminum window flange, then wrapping the edge of the plywood flange to the foam;
  • Two pieces at the top, first over the aluminum window flange, and then wrapping the edge of the plywood flange to the foam.

We paid a lot of attention to corners. You can find many fine instructional flashing videos on YouTube, as well as a series of helpful videos here on the GBA site.

Foam the window. On the inside, spray low-expansion door-and-window foam (the yellow stuff, not pink or orange) in the gap between window and Dudley box. Once it has set, pull the shims and fill that gap with foam to eliminate the thermal break.

Try not to let too much foam boil up past the window frame; it’s a pain to trim later. Seal the junction between the Dudley box and the rigid foam and the studs with spray foam. You don’t need to completely fill the gap to the studs, just make it air tight where the box meets foam.

Note that all this spray foam provides more than ample support for the window and plenty of mechanical grip as well. We didn’t test to destruction, but my sense is that these windows won’t come out unless the whole wall disintegrates. We did test for air infiltration with a blower door and theatrical smoke; there wasn’t any.

Apply interior trim. You will find that the Dudley boxes greatly simplify the interior trim process. If you were rigorous about keeping the gap even between box and window frame, you can lay jamb material flat against the Dudley box and get just the right reveal with minimal shimming (none, usually).

Because the walls are now thicker by 4+ inches, we found MDF to be the best material for painted jambs on windows with a painted casing.

We converted all of the downstairs windows from extension-jamb-and-casing to drywall returns. With Dudley boxes, there’s almost no shimming. You just lay 1/2-inch drywall flat on the Dudley box side, and the even 1/4-inch gap gives a perfect 1/4 inch reveal on the window frame. Smiling carpenters!

Upstairs, we aimed to disturb the existing walls around the new windows as little as possible. Dan and Woogs realized they could make complementary inside Dudley boxes — in other words, Dudley boxes made of window trim — and completely avoid the stick-by-stick window trimming process. To make this work, you have to be organized and accurate with tape and ruler. You measure from the window frame to the wall plane to get the box depth on each side (just as for extension jambs). We decided the flange (casing) width room by room to minimize wall patching. The “inside Dudley box” should slip inside the outer Dudley box with a satisfying “Whoosh!,” and a few nails are all that’s required to secure it. Our inside walls are pretty lumpy in places, but surprisingly little caulking was needed behind the casing edges. In some rooms I didn’t even have to touch up the walls — I only painted the inside Dudley boxes.

A faster, better method

Dudley boxes go together much more quickly than built-in-place extension boxes, so you’ll save a lot of time. If you have multiple windows the same size, as we did, you can cut the parts all at once on site and assemble as needed. Or you could make them in the shop and speed things up even more with a biscuit joiner or your rotary clamp machine.

Dudley boxes are easier to install. Once you have the window locations accurately marked on the foam, the rest is easy — no fussing with shims, levels, etc., and much less chance of warping or twisting, because the boxes hold their shape.

Finally, you get a better performing wall, because all connections are accessible with spray foam, and they are backed up with caulk. Try it on your next retrofit. If you don’t agree it’s a better way, I’ll be surprised.

Joel Schuman is a semi-retired furniture designer and software developer living in Saugerties, New York.


  1. dante03 | | #1

    Looks like a great system
    Thanks for the info on how you accomplished this. Seems like a great system that accomplishes all of the goals of your project.

  2. dante03 | | #2

    Investigative work
    Thinking about this a little more (and in regards to any deep energy retrofit), I'm curious what kind of investigative work you did prior to installing the exterior foam - specifically to avoid causing moisture issues inside the wall down the road? If you didn't open up the walls on the interior, how do you know that you are not creating further issues down the road with existing vapor barriers and previous remodeling that has happened prior to this renovation? Assuming that the previous work happened at different times, did you verify the existing insulation and vapor barrier in each assumed area of the previous remodels?

  3. dankolbert | | #3

    Great piece, minor quibble
    What a great story - congratulations on both the project and the creativity of you and the carpenters. Looks like a lot of fun.

    We have been putting the first piece of tape behind the bottom flange, to provide a pathway for any bulk or condensate water that might find its way in. Then the sides and head go over the flanges (and bottom layer). We try to minimize nails in the bottom flange as well.

  4. Joel Schuman | | #4

    Re: Investigative Work
    Moisture management was on my radar at design time. We didn't open up the interior walls THIS TIME, but I know what is in the walls. When we did our first not-quite-a-gut on-the-cheap reno in 1990 we opened up each wall segment from the inside at one point or another. It was shocking! Most studs were scabbed, some were fished together out of tiny scraps less than 2 feet long - I swear to Lstiburek! Surprisingly, for an 1840s house, there was insulation: stacked bricks - I swear to Lstiburek! Seems sailing ships bound for Saugerties used low-fire brick as ballast. No good for building, so they dumped 'em on the shore and they were free for the hauling. I couldn't find an R-value for loose stacked brick, so we removed what was easily accessible and replaced it with loose fiberglass batt. But most of the brick is still there. The older part of the house (1840s) has decrepit clapboards directly on the studs - no sheathing - I swear to Lstiburek! We feared the 1940s brick-look asphalt shingle siding under the aluminum siding we peeled off was all that was keeping those parts standing, so we left it on. The new part (1880s) has 1" sheathing, but the house has been modified many times, and it has little integrity.
    I don't think the term "vapor barrier" was in play in the 1880s. And although we could see tar paper in places, mostly we could see through the exterior walls to the out-of-doors. Did I mention our house used to be drafty? I proceeded as if there were no vapor barrier in place and conservatively rated the walls at R-1.
    We used the two overlapping, staggered layers of TuffR Commercial rigid polyiso foam as both the water barrier (drainage plane) and air barrier. No house wrap. The keys to making that work are having skilled and conscientious installers who are committed to perfect joint taping and a good plan for connecting the windows to the walls (i.e. Dudley boxes).
    We didn't do any fancy modeling. No WUFI. We read and discussed Martin Holladay's "Are Dew-Point Calculations Really Necessary?" on GBA and everything we could get our hands on from BSC, factored in the difficulty and labor cost of doing a really good job with house wrap, and decided it was more cost effective to do a really good job with two layers of 2" foam.
    Our blower door tests were good, and one year later the house proved to be still too dry in winter (<30%RH without humidification), so we're not concerned about condensation. Even if there's occasional condensation in some overlooked section of wall, given the lack of vapor barriers, the house is probably drying to the inside pretty quickly.
    Did I answer your question?

  5. Joel Schuman | | #5

    Minor Quibble
    Thanks - project was way too much fun!
    The flashing tape manufacturer's tech rep was on site when we started, so we followed his suggestions. But I think your sequence might be better. Does the extra 1/16" under bottom flange cause any distortion?
    I would not skimp on the bottom nail/screw spacing on the aluminum flange unless I got green light from manufacturer or engineer. Might be okay, as the yellow foam provides a huge amount of support and adhesion. But hard to get to later if not.

  6. jinmtvt | | #6

    Nice write up.
    One question,

    Nice write up.

    One question, why not screw the inside of the "dooooodeley" box to the RO ??
    once it is screwed from the exterior, it shouldn't move much, also easy to play with screws if it does
    using screws with under head cutters )
    Problem here is that all of the weight of the windows are on the exterior screws which extend the thickness of the insulation as leverage. ( not so bad with 2" ..but what about 4 and 6" ?? )
    And only 4 small screws in the interior plywood flange to the RO would replace proababley 20 long screws in term of strength because of he leverage.

    Ur system is good for smaller windows, but windows in the 6ftX6ft + size with triple glazings ...not too sure.

    I've installed very similar windows to yours in the past , with the same type of alum nailing flange.
    Makes for quick isntallation and super easy flashing, but very restrictive on window placement within the wall.
    I'd have to consider plus and minuses next time i purchase and isntall simialar windows, to go with or without the brickmould nailing flange...

    Also, i believe there would be additional energy gains if one would find a way to install
    windows in a similar situation, without the box thermal bridging
    ( let's say box would stop mid frame of window and rest would be insulation similar to how many EU windows are isntalled )

    Any word on energy efficiency and comfort in your neue/old haus ?

  7. jinmtvt | | #7

    Joel: fastening the top and
    Joel: fastening the top and bottom flang doesn't change much, but is usually required by the window manuf for correct installation/warranty ... but really not necessary with this type of window
    even more if the frame is spray foamed form inside to the box aftewards.

    On small windows ( 4X5 or less ) one could probably really get away with spray foam and 2 screws through frame sides to the RO ( 2 per side )

  8. Joel Schuman | | #8

    One question
    That's more than one question, so I'll answer your last one first. Before retrofit: We would stay out partying all night in order not to come home to a cold drafty house. My wife wore knee high sheepskin UGGs as house slippers and a full length expedition weight bathrobe. Our dog hid under the covers when we got up in the morning. After retrofit: We never leave and our friends come here to party because we have the warmest and coziest house. We run around nearly naked. The dog naps on the window sill because the internal glass temp is 71F, same as the walls. After one year, we have a little data on energy savings, although last winter was not typical. Keep in mind we also installed geothermal and solar DHW. We are paying about 25% of what we used to for energy. Better than we hoped.
    Now the other Qs:
    1. Reasons not to screw thru the DUDLEY (please, he reads this blog and he's very sensitive for a carpenter:) boxes to the RO jambs. There's a variable size gap between the box and the RO, and it would take a dab hand with the impact driver not to overtighten and distort the whole box (remember it's a nailed and glued integral unit). If the side is distorted, it creates problems mounting the window and more problems later trimming. The screw or nail is an obstacle to smooth application of the yellow spray foam. Ideally, you run the nozzle right down the edge of the box in one quick full-throttle pass and see the back of the rigid foam disappear.
    2. I don't think you are right about where the weight bears, and I don't think there is any "leverage". In the plane of the window frame, the weight of the Dudley box bears directly on the 4" rigid foam (two 2" layers). The yellow Dow foam under the window frame has substantial compressive strength, easily enough to support the window without taking into account the (possibly greater) tensile contribution of the foam along the sides and top. I believe that the foam is more than enough to bear the weight of our heavy triple-glazed windows. The screws thru the flanges do NOT support the window weight directly. Their purpose is to place the foam between the Dudley box flange and the RO framing under serious compression to create a tension-compression structure that locks the box to the wall. I'd love to see this modeled by an engineer. I suspect the glass would break or the wall would fail before the box or window frame blew out.
    3. Our largest is 62" x 54". About as much as three guys can get up onto a scaffold without a lift. I don't see why a larger Dudley box wouldn't work just as well, although larger sizes might require temporary stiffeners.
    4. Only direct thermal bridging on our installation is thru the aluminum brick mold at the windows' extreme outboard edge.
    5. You're right, fastening thru the window frame is what the manufacturer wants for new construction. Our dealer acknowledged that it was not optimal for our purposes and is honoring the warranty. It's an important point you are making here: if you plan to color even a little outside the lines, check with the dealer/manufacturer first.

  9. dante03 | | #9

    Investigative Work
    Joel - yep answered my question. That is one area of question that I have had when I look into doing deep retrofits. Making sure all the details are covered prior to making serious changes to a house can lead some big liability. This being your house and knowing those details in advance helped I'm sure. Thanks again for a great article.


  10. jinmtvt | | #10

    Joel: i didn't think about
    Joel: i didn't think about the foam underneath or missed that in the text ...
    the peel stick also will help alot keeping it up against the upper foam board
    the only thing i could see is that it would sag somewhat with time ( many years probably )
    and end up creating some "gap" between the top of the box ad the foam board on top of it
    but then again ..maybe not ( and again the tape helps it all )

    I am not sure i understand about the thermal bridging ..
    don't you have a plywood going all the way from inside to the alum nailing fin ?
    probably almost meaningless because of its ok R value and extremely low area
    still would like to see solutions that use only insulation at some point
    ( are you installing insulation inside around the window frame also ? )

    25% of how much ?
    had you calculated some payback term ?

    anyhow, very nice input on window and exterior insulation !! congrats on your new thermal enveloppe!

  11. Joel Schuman | | #11

    Re: Thermal Break and Payback
    Jin - I'm just a rookie, but look at the cross section. The only thermal bridge I see is where the aluminum nailing flange meets the edge of the window frame itself, which as you say has little cumulative area relative to the wall and window. Every other connection goes thru foam.
    Last year our total for gas/electric and firewood was 25% of the previous 5-years' average gas/electric, firewood, and OIL. Mild winter, sure, but we do use our woodstove as entertainment (cheaper to watch than cable). Also, we're retired so we can do laundry and take showers when the sun is heating most of the water. As for payback, the original estimate from the energy guys we worked with was 15 years total, including federal tax breaks and state incentive programs and pegging energy prices to inflation. But now it's looking like it could be much sooner - does 8 years sound crazy?

  12. jinmtvt | | #12

    8 years would be fabulous!!
    8 years would be fabulous!! factored in probable energy cost increase yet ? :)

    What is your total energy costs per year ?

    Maybe i am getting your drawing wrong, so i painted it up a bit.
    Blue is cold outside
    green = thermal bridges i see ( not very important ones, it is wood after all.. not steel ...and i've put the screws in just for fun ..but its more to tease than anything else :p )

    yellow = where you used spray foam ??

    The plywood boxes touches from exterior ( under the aluminium fin , wich is same as direct exterior ) toward the interior directly in your drawing.
    Again it is not an important problem, because it is wood ..not concrete aluminum or steel ...
    and almost all of the exterior insulated buildings i've seen lately, use some some of wood boxes around windows ( no real choice here ). Only wanted to point it out with you
    ( that is if i do get the drawing..if not i apologize )

    have a nice cozy week end! ( cold here..we just passed a -20c day -30c night week .. )

  13. GBA Editor
    Martin Holladay | | #13

    About payback
    You wrote, "As for payback ... - does 8 years sound crazy?"

    Well, frankly -- yes.

    I'm just guessing, but here are some back-of-the-envelope calculations:

    You used to spend $2,000 a year on heating fuel.

    Now you're spending only 25% of what you used to spend -- so you are saving $1,500 a year.

    In 8 years, you've saved $12,000. That's great. But there is no way you bought 28 triple-glazed Inline windows, a PV system, a solar hot water system, and paid two skilled carpenters to put two layers of 2-inch polyiso and new siding on the outside of your house, for $12,000.

  14. Joel Schuman | | #14

    About Payback
    Martin brings up the key question of what to include in the cost column. We did more than just the energy work. The carpenters also put on a two-story addition, fixed structural defects, etc. And we did cosmetic changes, like a new roof, that were not strictly necessary but have some effect on energy use. So if I include all the costs that Martin mentions and anything else that has even tangential bearing on energy use, the payback period is more like 30 years.

    But as a non-accountant, the whole calculation of payback seems dodgy because you have to make assumptions about where energy prices will be at each point in the payback period. Likewise tax laws, if you have a nig carry-forward tax credit, as we do.

    We certainly did not consider payback when we made the decision to do the retrofit. We were happy about the anticipated lower expenses and improved cash flow, but we would have done it anyway. This is our home - we did it for comfort, pleasure, and because we believe reducing personal energy use is an ethical, moral, and social choice as well as a financial one.

  15. homedesign | | #15

    Outie Windows Rock
    Maybe if I lived in climate like Fairbanks Alaska... I would consider a "middie" window.
    Otherwise... I think Outies are the way to go.
    Try to imagine building an Innie Skylight :--)

    Joel, I really liked your article and your system...
    I think windows are the extra tricky part when using outside insulation stratagems.
    I think your method is far superior to the BSC method shown in this video
    and the alternate method that Lstiburek's favorite crew (Synergy) used on Hobbit's Deep Energy Retrofit

    anyway...Kudos to you and your crew

  16. GBA Editor
    Martin Holladay | | #16

    Response to Joel Schuman
    You wrote, "This is our home - we did it for comfort, pleasure, and because we believe reducing personal energy use is an ethical, moral, and social choice as well as a financial one."

    Those are all excellent reasons -- and it looks like you and your crew did a great job. I'm jealous; the finished product must be a pleasure to come home to.

    I agree that payback isn't everything. However, when the topic comes up, it's important to be accurate in our math.

    Thanks very much for your blog. I think that Dudley boxes are a great method of installing retrofit windows.

  17. nvman | | #17

    Shimming the window boxes
    You mentioned a lot of problems with the walls yet you never said anything about have to shim the outside of the window boxes so that all 4 corners of the nailing flange are in the same vertical plane. Were the walls not as bad as you thought?

  18. Joel Schuman | | #18

    Re: Shimming
    When the carpenters removed the old aluminum siding and we saw just how bad the walls were my knees sagged and I got all in a sweat about how/whether the new windows would go in - whether they would be all snaggly and not work. Because you can't shim the Dudley boxes without shimming the furring strips. Otherwise the siding would look all swoopy at best, and at worst wouldn't bend enough to accommodate the windows. Shimming furring strips would have been a lot of extra (unbudgeted) work.
    But the two 2" layers of rigid foam - and it is pretty darn rigid - strapped tight to the studs had the near-miraculous effect of smoothing out the lumpy walls almost completely. We learned how to tweak and adjust the plane of the wall by varying the tension of the HeadLok screws that connected the furring strips to the studs. Given the flimsy construction of the walls, I think the exterior foam may have actually straightened out the framing. I kept checking the inside wall surfaces for cracks or gaps, but saw none.
    Note that I said ALMOST completely. Tthere was still measurable (and in many cases easily observable) irregularity over a whole wall. But locally - around each window - it wasn't enough that we needed to shim the Dudley box flanges - they were all well within caulking distance.
    The building had racked and swayed quite a bit over the years, so some walls were not even close to vertical, and the front wall in particular had "melted" badly at the bottom (almost 2" from vertical in the first 8 feet). We didn't try to correct that. The window installation instructions stress keeping the window vertical, but we put them in the walls we had, even when they were 3-4 degrees off vertical, and crossed our fingers. The windows operated perfectly and are still doing so after 16 months. I think keeping the window frame planar is more important than anything, maybe everything, else.

    PS More benefits of the exterior foam I had not grokked fully:
    1. The house is quiet. We lives three doors down from a 5-way intersection on a state highway at the bottom of a steep grade running across an old noisy bridge (concrete-filled steel deck). Not quite as bad as living under an airport approach, but almost. Now we hear nothing. Really. When our kids visit from the city it's too quiet for them to sleep.
    2. Vibration. The house used to sway in a strong wind or when heavy trucks rolled across the bridge. Strapping the rigid foam to the studs kind of trussed the whole thing together. Now it feels like the walls are made of stone or concrete.
    3. The walls are thick enough that you can sit in the window wells (9"- 11" sills). That opens all kinds of possibilities.

  19. esp71 | | #19

    Those old Bricks
    Brick in-fill in wall cavities was very commmon in New England (at least) in the 17th, 18th and even the 19th century. It's a technique called "nogging", and was intended as draft stopping. They were usually pointed though, not just dry stacked. As a preservation carpenter who works for a New England non profit that runs house museums I've seen this numerous times. Most recently we did work on a 1695 house that had clay used for setting the bricks. I have also seen lath and plaster applied to the back side of the sheating in an early 19th century house for the same purpose. Those folks were trying to deal with the same issues we are. When every BTU was obtained by swinging an axe, they did all they could with what they had to keep the heat in.

  20. ecdunn | | #20

    FInished pics
    Can't wait to see the photos of the finished project. Also, a better drawing on how the trim and siding work.

  21. user-1136306 | | #21

    Dudley boxes for OLD, non-flanged windows
    I really like the Dudley box idea.. I am renovating a 1771 Saltbox where the lath and plaster were attached directly to 1 1/4" think oak plank sheathing..and windows were nailed into the oak planks ONLY through their exterior casings...I am planning to add polyiso on the outside of the oak planks, then furring strips and new clapboard. Since I had to gut most of the walls, I am able to add new 2x4 interior walls for wiring and new plaster as well as for additional fiberglass insulation.

    BUT, I've seen nothing published on foam and window installation where you intend to RE-USE the original OLD windows..These windows are in good condition with thick, old growth lumber for frames and casings...As such I am stripping paint and re-using them to retain the old look...Exterior storms are still there as well, so I plan to weatherstrip sashes and the old storms to retain originality. What is the best way to install these 4" window frames into Dudley boxes through the polyiso foam sheets? The windows have no flanges, and at this point I am considering adding my own using bent aluminum coil attached to the exterior sides of the frames...
    Can someone advise on this? Because these window frames are thick and sturdy, I could screw the frames from the inside into the sides of the Dudley boxes and RO studs, but that creates shimming and twisting issues as described already by Joel in his posting...Also the aluminum flanges assist in air and drainage sealing which is otherwise difficult without flanges...But the flexibility and lack of rigidity of the coil may be a problem...any better ideas?

    Unless I am misreading, I also thought that Joel indicated that the 3/4" "flanges" of the Dudley boxes were positioned on top of and nailed to the edge of the 1/2" plywood - yet the pictures show that the 1/2" plywood is nailed into the 3/4" edge of the flanges.. am I mistaken? and the butt joints of the side flanges into the top/bottom flanges are only fastened together with glue and no mechanical fasteners like biscuits or pocket screws? What holds them in alignment until the glue dries in jobsite built situation?

    Lastly, is 4" of polyiso required if I can use 2" polyiso outside with R-13 fiberglass inside? Cost is a big factor here..

    Wonderful article with great attention to details.. Thanks SO much for the guidance...

  22. Joel Schuman | | #22

    Re: Dudley Boxes for Old Windows
    Ooops! You're right! I went over all the photos, and the 1/2" sides overlap and are nailed to the 3/4" flanges. The butt joints of the flanges are toe-nailed together (Dan and Woogs are fearless with a big nail gun) at the outside edge, which keeps them lined up while the glue dries. We considered biscuits, but since it would have added work we decided to see if glue-and-nails was enough. It was. Each box also had a temporary diagonal brace to keep it square while the glue dries and while the box is installed.

    I don't think the second layer of 2" polyiso is needed if you can put R-13 fiberglass in the studs.

    Why can't you nail the windows back into the Dudley box flanges thru the exterior casings, as was done originally?

  23. GBA Editor
    Martin Holladay | | #23

    Correction noted
    I haved edited the text of the article to clarify that the 1/2-inch plywood sides were nailed into the 3/4-inch plywood flanges.

  24. ohioplayer | | #24

    Architectural Intergrity
    I am all for deep energy retrofits and even plan to do one on my 100 year old Dutch colonial. That said, I was saddened to see what looked like a lovely brick home with double hung windows remodeled with lap siding and casement windows.

  25. Joel Schuman | | #25

    Re: Architectural Integrity
    That isn't real brick - it's asphalt shingle siding doing a pretty convincing imitation of brick. As for the switch from double hung to casement, we're fans of casements. Their energy performance is undeniably superior, they present a much better view (we have great views), and to our eyes they have a more pleasing appearance, both inside and out.

  26. user-1122346 | | #26

    Using off-site production of sheathing and cut-through windows a
    Your post, “Window Installation Tips for a Deep Energy Retrofit,” is a terrific and welcomed discussion of the ‘wacky world of DER,’ and I very much appreciate it and all the interesting comments that your post engendered. The issue of bucks for windows in an exterior foam re-sheathing retrofit of an existing building is a perennial one, and the Dudley Box approach makes a whole lot of sense.

    Another way to improve on this process might be to use one-sided structural insulated panels (SIPs), such as Nailbase, as a starting point, instead of adding up layers of foam. I did note, in the photographs accompanying your post, that you did use, in some instances, DOW’s structural insulated sheathing—SIS—but that still requires using furring strips for siding attachment. The use of SIP-based material provides for OSB sheathing, and a lot more options for siding and trim nailing, not to mention even better anti-racking and other structural bracing improvements. To my way of thinking, however, the most potential for cost savings in using SIP materials is from the precision off-site manufacturing of large sheathing elements, through the use of computer-aided manufacturing software and production tools. The “one-sided” SIPS, where the OSB is the exterior face, attached to 4-6 inches of rigid foam, which generically I call insulated sheathing panels (ISPs), can be produced with window and door cut-outs, as well as dimensional lumber attached in the cut-outs to create window and door extension jambs or “bucks.” Fewer, bigger, integrated pieces arrive on site for installation, reducing labor steps and on-site waste, among some other potential cost-reducing benefits.

    That’s the concept behind RetroSheath (, but there is another requirement to be satisfied before this approach can move from interesting concept to field application. This requirement is to effectively and accurately capture the three-dimensions of exterior walls and their elements such as windows and doors. While this could be done by hand (tapes and laser range finders, and constant re-checking!), the time needed is prohibitive to capture the level and resolution of measurements required to drive SIP CAM processes at the needed tolerances (generally thought to be no more than plus/minus one-eighth inch). There are also some related questions still in need of answering, including whether computer-aided design and drafting (CADD) software can use this data effectively, or whether these programs need additional development.

    I’ve recently submitted a research proposal to NYSERDA that seeks to address the practicality of using digital technologies for measurement capture, and, specifically, I’m seeking to assess whether digital graphical processing has reached the point where it offers affordable, sufficient, and valuable new ways of taking the measure of existing structures, by looking at current photogrammetry and laser scanning services and their tools and procedures. Currently, there are companies that do this sort of thing, but at prices that are supported by high-end projects, such as historical architectural preservations. Given how cheap high-resolution digital cameras are, how powerful even mid-range PC and laptops are in graphical processing, and the many new and shockingly capable 2D-to-3D programs there are out there (including Google’s free Sketch Up), all trades that need accurate and detailed measurements (contractors in general, many manufacturing businesses, etc.) may be at an important inflection of how traditional things can get done in new ways, vastly cheaper and better. Perhaps the promise of digital is finally coming to construction.

    Posts like Joel Schuman’s, and the ongoing discussions around DER in the pages of GreenBuildingAdvisor and other trade publications, as well as conference presentations at the key conferences such as NESEA’s upcoming BuildingEnergy 13, all move us forward. Reducing the cost of DER remains a critical requirement, and Dudley boxes are one more point along this way.

  27. user-977322 | | #27

    Great minds think alike?
    Reading this post was a confirmatory experience. It’s nice to see innovation is alive and well. This past year, we went through the same DER process on our circa-1890 Victorian-era home. Unlike Joel’s home, however, ours was apparently very well built and framed with native white oak. As a result most of the RO’s turned out to be virtually perfect, even after 120+ years, which allowed us to use just the flange part of the “box” over the polyiso for nailing the windows, and to size the windows to the RO’s.

    That said, several old window openings in a slightly newer part of the house turned out to have been monkeyed with by the three stooges at some time in the distant past, and were in awful condition, the RO’s consisting of various fragments of 1x, 2x, and other less readily identifiable substances that collectively resulted in polygonal to rhombic RO’s. Ergo, our creative contractor constructed boxes virtually identical to your Dudley boxes to overcome the challenges, and also fastened them in identical fashion, i.e., only through the flange to avoid distortion. Worked like a charm, extremely strong, easy to foam and trim. Necessity is the mother of invention, so I imagine other experienced contractors have probably thought of soemthing like this as well. You've done a great service by providing such a well written explanation.

    As for architecturally correct windows, you do not have to compromise (much) on energy savings and other features if you want to stay with double hungs. Fourteen of our 22 windows are DH’s, and we went with Inline Fiberglass’s Eternity line of triple-glazed DH’s, which were released late in 2011. These windows live up to their phenomenal performance specs, as reported by one of Martin’s earlier posts. Particularly air infiltration: you can stand in front of them in the face of a 50 mph wind on a sub-zero day and not feel the slightest draft, nor observe any effect on a smoke pencil. Their details also go really well with an older home, especially with the stainable oak veneer option. The only downside that we can see, and it’s a minor one, is their stiff operation, but that isn’t terribly surprising considering how robust their seals are. I would unconditionally recommend them to anyone concerned about maintaining historical integrity without sacrificing performance.

  28. marleyjune | | #28

    This is one of the best articles ever on GBA
    With this detailed explanation and photos, I'm finally confident that I've found the way to do new exterior insulation and window replacement the right way. The comments and responses add much to the information, as well. Thanks.

  29. Mike Steffen | | #29

    Quibbles revisited, perhaps major
    I appreciate the innovative approach to window installation in retrofits presented here. However I want to harken back to the comment by Dan Kolbert early on. He noted that the installation method shows the flashing tape applied over the flange at the sill of the window. That small detail implies that this window installation method is not concerned with managing water penetration through the window system itself. That may be OK for certain applications but not for others, many others I would argue. It is clear that many window products leak. Some leak early during their service life (some before the owner/user evens takes occupancy!)' others leak later during their service life. Managing the water that penetrates through windows is one key element of window to wall detailing. In the details exhibited here, I do not see that this type of water penetration is managed. Am I missing something? The way I see it, if windows leak where this method has been used, the plywood box and the framing materials will be at risk of moisture damage.

    BTW for clarification I am not talking about window leakage at the joint between the window and the adjacent wall but rather leakage through the window unit itself - for example a wood frame window with metal cladding that may exhibit leakage through the joinery at the sill to jamb corners at the bottom of the window.

  30. GBA Editor
    Martin Holladay | | #30

    Response to Mike Steffen
    Even if you prefer a different method of flashing windows, there's no reason you can't use a Dudley box. All you have to do is install a sill pan or peel-and-stick flashing on the Dudley box sill if you want to do it differently. Then remember to leave the bottom flange of the window untaped.

    Of course, you would need to adjust the height of the Dudley box to be sure that the left enough room for the sill flashing details.

  31. dalesizemo | | #31

    furring strips
    Great article but I'm left wondering how about the finishing details of the furring strips that presumably were installed in the vicinity of and in between the windows.
    I'm installing two half inch staggered layers of polyiso foil faced insulation on the exterior of an osb sheathed double wall new construction project in Georgia. I was planning to install plywood boxes in the rough window openings joining the two walls and extending out to the thickness of the foam, but my framers didn't always follow the RO dimensions I gave them which allowed for the 3/4 plywood boxes. I was planning to install Marvin Integrity fiberglass windows so the nailing flange would be behind the furring strips utilizing a drip cap taped to the foam to allow any moisture that gets through the Nishiha staggered shingle siding to drain. After reading this article I realized going head and bringing out the window nailing flange to the furring strip level will make installing the Nishiha staggered shingles easier (leaving more window protrusion), but I have a couple of questions on how to flash above the windows and maintain the flow through ventilation that helps dry out moisture getting behind the siding.
    To keep the ventilation flowing from the bottom of the wall to the attic, should I leave space between the vertical 1 x 2 furring strips and the box frame of the window all the way around so air can flow through? If so, how much space?
    Should I leave the bottom flange of the window units, which range from 3' x 3' up to 6' x 6' fixed glass, unattached with the recommended 12 inch piece of wrb taped to that bottom flange to allow water to drain (will the top and side flange attachments be enough)?
    Since i am using 1 x 2 furring strips, and my RO are relatively close, can't I use just a 3-4" wide 3/4" plywood box faceframe rather than 8 inches?

  32. Joel_Schuman | | #32

    Re: furring strips
    You have it right. We left a 1" gap between the furring strips and the top and bottom flanges for air circulation between rain screen (siding) and drainage plane (foil foam facing).
    And you certainly can use narrower flanges if your ROs and siding attachment details permit. Our ROs and adjacent framing were wild, really wild.
    We screwed our bottom alum nailing flanges to the box flanges. Dunno if it was necessary. We flashed under and over, and given the rubbery nature of the flashing tape I doubt the screws would leak even if the flashing over them failed.

  33. dalesizemo | | #33

    drip cap
    Thanks! I'm using a product called Protecto Super Stick Building Tape and it is somewhat thinner than the butyl type adhesive window tapes, so I'm nervous about the bottom screws for that reason, but could get some of the thicker butyl tape for the sill pan area.
    What type of siding did you go with and what kind of drip caps did you use above the windows once you finished taping them?
    Thanks again

  34. Joel_Schuman | | #34

    Re: drip cap
    We sided with James Hardie plank w/ 6" reveal smooth surface, factory painted.
    Our carpenters had intended to fabricate drip cap from flat alum for the windows. But when we installed the first window and saw how well the factory nail flange shed water (mfr calls it "brick mold" but protrudes more and has a slope on the top), we decided we didn't need it. I hose-tested after flashing and before siding and saw no water coming in anywhere. These windows are widely used in commercial/industrial bldgs with no drip cap and similar drainage plane connections.

  35. Cabbydog | | #35

    Entry Doors
    You did not mention how you treated any doors. How would one handle entry doors and sliding glass doors in this situation. I will be adding 2 layers of 2 inch rigid foam to the exterior of my Tahoe cabin.
    I would like suggestions for handling entry door and sliding glass doors. Thank you.

  36. GBA Editor
    Martin Holladay | | #36

    Response to Whitney Johnson
    I don't know how Joel did it. But if I were you, I would build some door bucks out of 2x lumber, not plywood, to use as rough openings for your doors.

    You would probably want to rip the 2x lumber into 4" or 4 3/4" widths, and secure the lumber for the rough openings to the exterior side of the existing wall sheathing with long screws.

  37. Cabbydog | | #37

    Entry Doors
    Thank you Martin for your reply.
    Can you tell me what the negatives are to, instead of furring out the doors to the exterior, creating boxes but leaving the doors in the plane of existing framing and thus, having them recessed 5" from the exterior edge of the siding?
    Thanks in advance.

  38. GBA Editor
    Martin Holladay | | #38

    Response to Whitney Johnson
    Q. "Can you tell me what the negatives are...?"

    A. With your proposed method, you have to come up with waterproof sill (threshold) details to prevent moisture entry under the door.

  39. Cabbydog | | #39

    Entry Doors
    2X4 Frame construction.
    The main entry door is mounted in a frame attached to the studs with the threshold mounted to the subfloor, The sliding doors are mounted to studs, header and subfloor. The siding is 1-2" proud of the doors. In my mind, a typical installation.The doors open on to a redwood deck which attaches to the exterior of the structure between the subfloor and sill plate. If one leaves the doors where they are and carry the exterior out 5", how does that change moisture entry from under door?
    What if one or two iron grate(s) were placed in front of each door in lieu of a 2x6 deck board?

    Hanging the doors to the exterior of the stud frame just doesn't seem as strong and secure as having them within the stud frame.
    What am I missing?
    Serious questions.

    And just to be clear, the box I propose is more a Staple than a Box as it would not have a threshold, just sides and head.


  40. GBA Editor
    Martin Holladay | | #40

    Response to Whitney Johnson
    You're right. If all of the rigid foam is exterior to the framing and rim joists, there shouldn't be any water-entry concerns with a recessed door. I was thinking of recessed doors in double-stud walls (for some reason), and my concerns don't apply in your case.

  41. Pvo88 | | #41

    Very helpful article. I’ve been looking for similar articles on how to install standard doors and french or sliding doors (with exterior foam insulation) but haven’t come across much. Can you point me to any resources or is the practice the same as with the windows?

    From a functional standpoint (door swing) it seems installing the doors as “innies” makes the most sense and the windows as “outies” but then the house wrap plane become complicated. Any advice here?

    1. GBA Editor
      Martin Holladay | | #42

      Start with these three articles:

      "All About Doors"

      "Sill Pans For Exterior Doors"

      "How to Flash an Exterior Door"

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |