We met with our energy/passive haus consultant, our structural engineer, and our insulation expert today.
We first met Marc, our energy consultant, this winter up at his place south of Lebanon, New Hampshire. We found him because he's an expert in Passive House, which is a building technique we'd become interested in because of its energy efficiency and the level of comfort it would provide. It's turned out to be a good choice, because although we are not going for full Passive House certification, Marc is a very general expert, and mostly uses Passive House techniques rather than necessarily going for full Passive House certification.
This flexibility is good for us, because it turns out that there are a number of architectural features in the house that aren't really compatible with Passive House--in order to both have those features and have Passive House, we would have to go to some very expensive, high-tech materials. Conversely, we could compromise on the aesthetics. We don't really want to do either, and Marc has been helping us to navigate that.
Going in to the conversation today, we were planning to build a 14" thick double wall, with 6" studs on the outside and 4" studs on the inside, and some combination of fiber and foam inside the cavity. We were planning to use open web trusses for the second floor cantilever (see floor plans), as well as in the roof.
Our structural engineer came to the meeting with a wall section, which was pretty good, but had some thermal bridging issues (that's why we have an energy consultant) and insufficient insulation in the slab. He'd done the roof the way we expected, but switched to 2x10's in the floor because of some ambiguities in the architectural drawings. He was the one who's proposed going to 6" studs on the outside wall--we'd originally been planning to use 2x4's.
At the end of the conversation, we'd kept his 2x10's, and his 2x6's, but at Marc's suggestion switched from open web trusses in the roof to I-joists. Instead of building a second inside wall, we're going to add somewhere between 4" and 6" of foam to the outside. Why?
The roof I-joists work better than open web trusses because of the way they're constructed. An open web truss is essentually a structure built up of dimensional lumber, with plates to connect the lumber and form a lattice that joins the top and bottom members, which are continuous along the length of the truss. So there are a lot of irregular air gaps inside the truss, which is handy for running ventilation pipes--that's why we wanted it in the second floor.
Unfortunately, these same gaps make the labor of insulating the roof substantially greater. In addition, there's no convenient place to nail a blocker that would allow us to vent the roof. Consequently, if we want to vent the roof (and we do), we would have to add additional 2x3's along the length of the each truss on either side to provide nailing for the vent--a fairly substantial amount of work, and additional expense. In comparison, the top member of the I-joist is in the perfect place to act as an offset to which we can nail the blocking for the vent.
Unless we hear something really surprising, we're assuming at this point that we're going with I-joists in the roof, and not web trusses. Really surprising would be that the I-joists cost thousands of dollars more than the trusses. We really don't expect that to be the case.
Another concern we had is that a 2' I-joist, which is the thickness we need to get a clear span (no supporting wall in the middle), is pretty big. It's going to stick out a long way on the north side, so we're essentially going to have a 2' box sticking 7' out on that side. We spent a bunch of time thinking of ways to make the box smaller, and then I asked Camilo, our architect, what he thought of a box that size, and he said he didn't think we needed to be concerned about it. That simplified things quite a bit on that end.
Additional complexities on the roof--the south overhang is probably going to be constructed as a box that we nail to the rim joist on that side, rather than an extension of the I-joist. This is because a 2' box sticking out 1' on that side really would look odd. We'll do the same thing on the sides. As I type this I'm already starting to think that we would be better off keeping the box, but that's something to debate with the engineers later. The reason is that since the roof is at about a 30º angle, the overhang if we kept the 2' box on the south side would probably be in the ballpark. And the side box has to match up against the south box and the north box, ideally, or else it's going to look a bit strange. But it could work either way, I think--just a question of aesthetics.
Why did we switch to single-wall, exterior insulation? Mainly because I've been really sweating it on the question of moisture in the wall. I really don't want a wall that rots after ten years. I don't want moisture accumulation in the wall. One of the ways moisture accumulates in the wall is that the wall is colder than the dew point. The way this happens is that insulation runs through the wall. On the inside, it's, say, 68 degrees--well above the dew point. On the outside, it's below zero--well below the dew point. Somewhere inside the wall, then, vapor is going to condense into water droplets, which are then going to make the wall wet.
So what if we put R-25 on the outside of the wall in the form of 4" of poly-iso-cyanurate closed-cell insulation? What happens is that the dew point is now inside that poly-iso. Poly-iso is closed-cell, so very vapor-resistant. As a consequence, we don't expect condensation at all. But if there is any condensation, it's going to be inside the poly-iso, outside of the structure and the sheathing, where it can't cause any harm.
My main resistance to this kind of construction is that I've never done it, and hence don't know how. Marc explained how, and managed to convince me that I could do it. Going this route has a lot of advantages. For one thing, I need to build only one set of outer walls. For another, I can build the walls now, and insulate on the outside later. Given the lateness of the season, this is a very good thing.
As for the 2x10's, we decided to stick with them because they're easier to work with, available off the shelf, and we decided that everything we wanted to put inside the trusses, we can just do in soffits, because nearly all the wall space where the soffits would go is either in the kitchen or in utility rooms; in the kitchen, it's space that would otherwise be virtually unusable, above the cabinets. So there's a bit of soffit in the downstairs guest bedroom, maybe, but we may not even need that. So we get a lot less complexity in the floor design, in exchange for not that much complexity in the ventilation design. I'm inclined to leave the ducting exposed in the utility rooms, but I expect we'll cover it in the kitchen.
Oh, as far as the insulation goes, the main takeaway was that we can do it with dense-pack cellulose, both in the roof and the walls. This is good because cellulose is cheap compared to some of the stuff we were looking at. We'll probably get about R-70 in the roof, and either R-40 or R-50 in the walls, depending on what Marc's models tell us to do. R-50 would require two inches more of poly-iso.
Hm. We have bearing walls in the middle of the house, so that requires a thickened slab under the bearing walls. We went over the insulation detail for that--we're looking at about 6" of foam under the slab, and where the bearing wall runs we'll have to dig down an additional 6" to install more foam, with an overlap and a hole to account for the thicker slab at that position.
We're tentatively going with a window design for the big windows in the front where we have a 4x6" stud between each window. Need to follow up with Camilo on whether or not that will look okay. This prevents us from having to levitate the corner of the building, and we probably need structures that strong to hold the windows anyway, because they are large and contain a lot of glass--on the order of 50 pounds, which has to swing out without bowing the thing it's attached to.