This is the second part of a blog series by architect Elden Lindamood about the design and construction of his own home. The first installment was called A Low-Energy House for Northern Minnesota.

Progress has continued on the new house, and a few things have come up along the way to warrant minor changes in the foundation insulation details and window installation. It is refreshing to work with a builder who asks questions and works with me to alter things and arrive at alternate solutions that seem more practical and buildable to him without losing my performance or aesthetic intent. Too many times I have arrived on project sites to find a builder who simply “did it the way he’s always done it” despite the bid set that specified otherwise.

I’d like to share some of the details that my builder and I altered from the original design. I’ll also share a few experiences with the nailbase insulation panels in case a reader is considering that wall construction method.

Should the base be crushed stone or sand?

One of the first conversations I had with my builder, Steve Johnson, that resulted in an alternate detail involved the insulation surrounding the frost protected shallow foundation (see Images #2 and #3, below). The original and revised details seem very similar at first blush, but there are a number of differences which were all carefully considered for constructability and performance.

Right out of the gate, the excavator encouraged me to use sand for the base rather than the crushed gravel that I had specified. My original intent was to take advantage of crushed stone’s lesser tendency for capillary action on a site with a high water table to help ensure that the slab, the slab insulation, and horizontal wing insulation would stay dry. In talking it through with the builder and excavator, however, I agreed since my site is dead flat for hundreds of feet in any direction, and since there was no way to run a drain tile to daylight, that using sand (which would be cheaper and easier to work) was an acceptable solution in this instance.

Had it been a full basement below the water table, I wouldn’t have relented. Steve also suggested that he could run the sub-slab vapor barrier / soil gas barrier continuously down, under, and up the form for the thickened slab edge, helping to protect it from capillary moisture (see Image #4, below). I agreed that was a good plan, and we were on our way.

Foundation insulation changes

As the drawings show, I switched from XPS to Type IX EPS for the under-slab insulation. This was primarily a matter of cost and availability, with a very small energy penalty (and a slight global warming potential improvement). The builder installed the EPS in two 4-inch layers, with the vapor barrier placed between the two layers.

We then discussed my detail showing a single continuous sheet of insulation from beneath the footing, extending out to form the horizontal wing insulation. I thought I was being helpful by detailing it so Steve wouldn’t have to cut any more foam than necessary, but because of how he wanted to set his wood foundation forms, that was going to be difficult. I was hesitant to let that one go because I was concerned about a joint or gap at the lower, outer slab edge. However, Steve wanted to install the vertical exterior foam post-pour, so the vertical insulation could then go on prior to the wing insulation, creating a lapped joint that I was happy with.

This detail also allowed Steve to slope the exterior wing insulation away from the house without the need for the extra tapered insulation that I envisioned placing on top of it. Additionally, Steve asked if the 2’0″ dimension that I had initially indicated from the top of slab to bottom of thickened edge was code-required. I looked into it and determined that the 2’0″ dimension was required to the bottom of foam, not the bottom of the concrete. This allowed Steve to cut his forms and vertical insulation to 24 inches instead of 27 inches.

All of this discussion lead to a good foundation detail that fit the builder’s preferred process, met my performance goals, and saved some time and money. This is how I wish all projects would go — but I digress.

Insect worries

One of the things about a frost-protected shallow foundation that has always made me uneasy is the potential for insect intrusion into the foam. Northern Minnesota is not a termite zone, but will it be 100 years from now?

I searched fruitlessly for information on how deep in the soil termites and carpenter ants will burrow. Regardless, I wanted to protect my foam, and also block any possible path from the foundation insulation into the wall insulation.

I initially explored wrapping all of the wing insulation and the vertical face of the foundation foam with stainless-steel screen cloth, but that was prohibitively expensive, especially for something I wasn’t sure was necessary. I eventually settled on covering the top of the wing insulation with Ice& Water Shield, and extending that up the face of the foundation. That would later be lapped with more peel-and-stick flashing to connect the Zip sheathing to the face of the foundation.

This created a nice barrier against burrowing insects (and varmints) from the exterior, and broke the potential path between the foundation insulation and the exterior wall foam with a peel-and-stick flashing. I also couldn’t find any information about whether ants or termites would chew through Ice & Water Shield, but I assumed they would not. Subsequent layers of metal flashing would further reduce that risk.

Frost heaving pushes up the flashing

Steve placed a prefinished metal flashing over the top of the peel-and-stick flashing to protect and finish the foundation (see Image #5, below). This was then taped, and we were ready for the nailbase panels — with one small hitch.

Since we are building in the winter, in northern Minnesota, despite the fact that the weather has been mercifully mild so far, winter construction brought its first complication. We had a significant amount of rain before I backfilled the trench against the metal foundation flashing. The soil that I backfilled with was sandy in many places, and muddy topsoil in others. The soil was saturated.

A few weeks after that was done and Steve was getting ready to place the first nailbase panel, he noted that the wet soil had grabbed the metal flashing as it froze, and heaved it up about 1/4 inch. This meant that the joint between the bottom of the nailbase panel and the top of the foundation insulation wouldn’t be as tight as I hoped.

Since there wasn’t much that could be done, shy of waiting until next May, we decided to place a strip of expanding foam tape at the rear of the joint, and a liberal bead of sealant at the front of the joint (see Image #6). This should serve to keep insects and vermin out of the joint.

I’ll definitely watch the joint over the next couple of years as things settle in to be sure it doesn’t open up. If it does, I’ll have to caulk it. Steve is inclined to think it will never be a problem.

Sealing the nail heads

With the foundation complete, the above-grade walls were framed and sheathed with Zip System OSB. The joints were taped — the taped sheathing will be my air barrier — and the sheathing was also taped back into the framed window openings. Since this is for air control only, and the windows will be “outies” at the face of the nailbase panels, the openings didn’t need to be flashed at the Zip sheathing, but simply sealed.

Even though Huber says you don’t need to tape or seal the nail holes in the panel fields, my partner Catherine and I spent a few hours taping little squares of Zip tape over each nail hole. It may not make any difference, but since I am going for a ridiculously tight envelope, I thought it was worth the minimal effort (see Image #1 at the top of the page).

The nailbase panels cupped

A previous project taught us a few things about using nailbase panels for our high-R walls.

The first issue is that the nailbase panels were cupped. Having an OSB skin on only one side makes them less stable than a SIP, so they all seem to cup with the center of the OSB skin moving outward about 1/4 inch to 1/2 inch.

This isn’t problematic in and of itself, but it does make it critical that the attachment screws get a good purchase in order to pull the center of the panel tight to the wall. The panel manufacturer, Extreme Panel, usually provides 8 1/2 inch self-tapping screws with their 8-inch nailbase panels. These screws are long enough (normally) to grab the structural sheathing beyond.

Note that the manufacturer only requires you to screw into the sheathing, which is why the 8 1/2 inch screws are what they usually provide. Because of the cupping issue, we ordered 9 1/2 inch wood-tipped screws instead, hoping to get a better bite in the sheathing and pull the panels tight. (I should mention that the nailbase manufacturer was very helpful in ordering the non-standard screws, and has been a pleasure to work with).

Even with these more aggressive screws, Steve said he thinks it is best to hit a stud so that he doesn’t just strip out the hole trying to pull the panel center tight. He added that if he were to do it again, a 10-inch screw might be preferred.

The nailbase skin is smaller than the nailbase insulation

We also stipulated that the nailbase panels have an OSB skin measuring 47 3/4 inches x 95 3/4 inches and an EPS panel measuring 48 inches x 96 inches. This was done on the recommendation of a builder; based on experience on a previous job, he thought it would aid in getting the joints between panels really tight.

It turns out that although the manufacturer was very accommodating at providing the panels we requested, the OSB skins aren’t all glued perfectly centered, or straight, on the foam panels. Thus Steve learned that to ensure accuracy, he needed to measure his cuts off the foam, not from the OSB edge.

Since this is winter construction, the panel adhesive, which you are to apply liberally between each panel, is cold and a bit stiff. That means that the panel joints don’t seem as tight to me as they might be if the sealant was more easily squished. Again, there isn’t much we can do about that except wait until May, so we’ll make do.

Fortunately, because of panel screw alignment with the studs, the panel joints don’t often fall on a stud in the cavity wall and the potential for thermal bridging at a panel joint is somewhat minimized. It does make me reconsider the merits of multiple layers of foam with staggered seams, though.

Steve also commented that he really liked the large washers provided with the screws, and that it was much easier to install the panels in a flat plane than it is to install furring strips directly over thick foam.

The nailbase corners come together in a mitered joint

Steve opted to miter the outside corners of the nailbase panels. He cut them short so that he could fill the gap with canned spray foam.

The manufacturer suggested a butt joint with a piece of extra OSB glued on, but Steve liked the miter joint idea better. He did the second corner with a 43-degree cut so the gap gets wider toward the outer edge, allowing for easier filling with the canned foam (see image #7, below). He is using a track saw with a chainsaw-type attachment that allows him to make his miters in the thick material.

Other than making the exact kind of mess you’d expect from cutting EPS with a chainsaw, the cuts are straight and the cutting method seems to work great. His non-mitered cuts are executed with a Festool track saw, cutting 3 inches deep on each side and then finishing off the unreachable center of the cut with a handsaw. Using multiple layers of thinner foam would make cutting easier because you could use a table saw or circular saw, but I think the time lost with the nailbase panels is made up in the speed of installation. I’ll ask Steve to share further impressions on the process when the job is finished.

Also note that I opted to extend the trusses out over the nailbase panels, as opposed to running the nailbase panels up to the roof deck (see Image #1 at the top of the page). I did this because at the time it seemed easier than notching the nailbase panels around the truss tails or holding them below the soffits. Also, the nailbase panels are expensive, so reducing the area covered by the panels and filling the area with the attic cellulose seemed more prudent. You really could do it either way, but this is the direction I chose.

Planning for an attic catwalk

I had the truss manufacturer add an additional single horizontal member to each attic truss at 28 inches above the bottom of the trusses. This allowed Steve to install a catwalk using plywood salvaged from the concrete forms. The catwalk runs the length of the attic (see Image #8, below). This added a couple of hundred dollars to the cost of the truss package, but if I ever have to do any work in the attic, I think I will be happy I don’t have to wade through 24 inches of cellulose.

Steve suggested some good installation details for the ceiling air barrier. He installed the polyethylene before the interior partitions were framed, so the membrane is continuous across the top plates of the partitions. He instructed the plumbing and electrical trades that he wants as few holes as possible in that membrane.

He also sealed the ceiling air barrier to the lower of the two top plates at the exterior walls (see Image #9). This makes the air barrier continuous through that plate to the exterior Zip sheathing. (The Zip sheathing was also sealed to that plate when the sheathing was installed.)

When penetrations are made in that plate, the penetrations can be sealed more easily from the interior side of the penetration, eliminating the possibility of air infiltration between the plates, which I have seen when conduction blower door tests at other jobs (see Image #10).

In my next blog, I hope to share some photos of the window installation and to talk about the changes we made to those details. In the meantime, Steve is trying to get the envelope closed in just as the Minnesota winter is getting really harsh.

Elden Lindamood is an architect with Wagner Zaun Architecture in Duluth, Minnesota.