Rob Myers is building a timber-frame house in Ontario, Canada, at a site on the Bonnechere River an hour and a half west of Ottawa. The first installment of his blog series was A Timber-Frame House for a Cold Climate — Part 1.
When assembling the REMOTE shell for a timber-frame building, the first word that comes to mind is “precision” — especially for the framing. As my friends will attest, this job was a perfect fit for me, since I happen to believe that if you are going to all the trouble of measuring and cutting a piece of wood then you may as well measure and cut it to (exactly) the correct length. (You can tell that I don’t do this for a living!)
There are two good reasons to build precisely. The first is that the wall and roof have to be spaced out from the timber frame by 5/8 inch and 7/8 inch respectively to allow the drywall and tongue-and-groove ceiling to be installed by slipping them between the frame and the shell. This allows completion of the outer shell (along with the air/water control layer) while still having full access to the inside wall and ceiling bays for plumbing and electrical work.
There are no worries about re-sealing any penetrations, since all work occurs inside the control layers. The interior finish surfaces can then be relatively easily applied without fitting around posts and braces.
The second reason for precision is that applying the rigid foam and strapping is a challenge in itself without having to guess at the location of every stud or rafter. It is much easier if the framing members are exactly where they are supposed to be and the spacing is maintained over the length of the member by using blocking.
Framing the wall and roof
The wall was constructed using 2x4s at 16 inches o.c. Since the wall is non-structural, the stud spacing could have been 24 inches. But because of the design, thermal bridging is not an issue and I feel that drywall just looks better with studs spaced at 16 inches.
The depth of the wall (3 1/2 inches) was selected so that when the insulation is added to the inside bay, the proper balance of inner and outer insulation is maintained (as calculated for my climate). The top plate of the wall was tied to the timber frame using Timberlok screws and 5/8 inch plywood spacers.
In the same manner, the roof was constructed of 2x6s at 16 inches o.c. (The maximum rafter span from purlin to purlin was less than 4 feet). There were no rafter tails or rake overhangs; everything terminates at the edge of the wall.
The framing was then sheathed with 1/2-inch plywood which tied the walls to the rafters, forming a continuous shell. In my opinion, plywood is preferable to OSB. I also chose plywood because the house would be going through a Canadian winter without a finished roof, and plywood is the material that I thought would best handle the abuse.
As the plywood was installed, I sealed each seam with Siga Wigluv tape. There were no issues with the Wigluv tape adhering to the plywood in the cold — it was near freezing most days — but I always waited for any frost to dissipate before working. Since the tape can be applied to the roof/wall intersection, the air control layer is continuous from foundation to peak.
Innies, outies and in-betweenies
The 2×8 box extensions for the windows and doors were added during the wall framing. I decided to go with in-betweenie windows since this gave a reasonably balanced look from both the inside and outside and resulted in slightly improved thermal performance.
As I mentioned, one of the advantages of self-building is that labor is cheap, so I also opted for a somewhat complex design without worrying too much about the consequences. I chose flanged windows because I felt that they would be easier to install and air seal. The windows were triple-glazed fiberglass and I ordered orientation-specific glazing: solar-block glazing for the west-facing windows, and glazing with a high solar heat gain coefficient (SHGC) for all the rest.
The house design follows passive solar principles, but the best view (and hence the largest window) is to the west, and trees shade one end of the house to the south, so the site is not ideal for passive solar. However, I settled for somewhat compromised performance and kept the view and the trees.
Since the window flange would be in the middle of the wall, the depth of the rough framing box was set so that it also terminated in the middle of the wall. This allowed me to later apply insulation to the outside of the flange to provide a thermal break right up to the fiberglass frame itself. (I used two strips of EPS to make it easier to remove the frame if it was ever necessary).
Before assembly of the window boxes, I planed a slope with a small shoulder onto the rough sill so that it would drain to the outside. The rough sill thus has a physical barrier to water intrusion (a small step up). This was easy to do and I think saved a lot of fiddling during the window install.
Air and water control layers
At this point the timber frame is completely enclosed with a plywood box (except for the windows and doors) and the air control layer is complete. I like to ensure that materials are compatible, so I standardized on one manufacturer for the roof and wall water control layers. I applied Tyvek DrainWrap on the wall sheathing and DuPont Roofliner synthetic underlayment on the roof sheathing. The Drainwrap provides a better drainage plane between the EPS and the plywood, although after the house is completed I can’t imagine a circumstance where this would actually be needed. The Roofliner is tough and can be exposed to sunlight for six months. It worked well as an exposed roof surface over the winter.
As mentioned before, I try to build systems that have redundancy. With taped seams, the DrainWrap and Roofliner resulted in a second complete air control layer as well as a water control layer.
Both materials were applied shingle-style, from sill to peak. The materials were fastened with cap staples, and then all seams were taped using Tyvek tape. I taped over the cap staples in some areas, but this was a mistake: the tape tended to lift when the weather got severe, especially on the roof.
The detailing around the window boxes was pretty much standard. I used DuPont FlexWrap and StraightFlash with proper overlapping technique, tying the window flashing into the DrainWrap. The sill flashing was wrapped completely around the box so that it ended at the inside edge of the window. The windows were then installed using silicon caulk, and the flange and head were taped using StraightFlash. I used cap staples to ensure there was a drainage channel at the sill. The house went through a very long and cold Canadian winter in this unfinished state, with only minor leaking through a few of the cap staples.
Installing rigid foam on the walls
I would have to say that the thought of cutting and applying multiple layers of rigid foam, followed by driving very long screws through the strapping and foam and into a 1 1/2-inch-wide framing member, filled me with some trepidation. I was not exactly looking forward to the task, to say the least.
Metal sill flashing was applied to the perimeter and sealed to the base of the wall sheathing using pest-block foam. The flashing offers physical protection to the top edge of the slab foam and the bottom edge of the wall foam, and it ensures that any water is diverted past the edge of the slab.
The first job was to get the rigid foam. I tried to use recycled foam, but couldn’t find a source in Canada that could provide the quantity necessary. Trucking recycled foam from the U.S. wiped out any cost advantage (not to mention any environmental advantage).
In the end, I ordered 4×8 sheets of 3-inch EPS from a local manufacturer. Because I needed a large quantity, the price was almost the same as used material. I also bought the foam a year early, thinking that most of the shrinkage would occur before I used it. The sheets did shrink about 1/4 inch, but I couldn’t find any actual data on rates of shrinkage so I’m not sure whether this was a useful strategy.
I worked out the cuts and the sequencing for application using a 3D drawing program. The foam was cut on the ground and placed without further adjustment (an added advantage of precision building).
I built a very inexpensive hot-wire foam cutter table and a few hot-wire angle cutters to enable accurate cutting of the foam without foam dust. (I discovered that I have a severe sensitivity to the dust and I now try to avoid it at all cost).
The table has a 2×4 frame with an OSB top. (Originally, it was a concrete form.) The arm has to be reasonably stiff. I built mine with a 2×4 column and 1×4 strapping. The arm is 28 inches long; this allows me to cut a 4’x8′ foam sheet into any width. The cutter is nichrome wire from . One end of the nichrome wire attaches through a hole in the table to a screw connector, while the other end wraps around an eye bolt.
The thinner strip of wood on the top of the gadget provides tension on the wire when it is heated (attach the bottom of the wire, press down on the strip, and attach the top of the wire).
The power is provided by a 50-watt 12-volt halogen transformer. I used a Variac to fine-tune the temperature of the wire, but my guess is that a decent dimmer would work just as well.
The fence has a single screw at one end and pivots to adjust the width of the cut. When I make an adjustment, I just screw the other end of the fence to the tabletop. Not pretty, but it works!
The second cutter is for flush cutting the foam on the corners after it is installed.
The third cutter is for cutting a bevel on an edge (two scraps of wood with the wire screwed across the end so that it cuts the desired angle when moving along the edge of the foam). All just plug in to the same 12-volt power supply. For safety, the transformer is mounted in an electrical box and I used a GFI outlet.
The cutters allowed me to rip the 4×8 sheets while maintaining a perfectly square edge as well as cutting angles to match the slope of the flashing at the sill and the valleys.
They allowed for better quality work in less time; and aside from providing ease of installation, they almost entirely eliminated the use of spray foam to fill gaps. Spray foam is messy, expensive, and requires more work while up on the roof. (I do as much work as possible on the ground — I am not as young as I used to be!)
The cutters also allowed for some details that I would not otherwise have attempted. For example, I wanted to insulate the flanges of my “in-betweenie” windows but still allow for removal of a window if that ever became necessary. (I’m trying to be kind to future generations). I was able to custom-cut foam “flange covers” that friction fit around each window; these flange covers provide continuous insulation to the edge of the frame.
In general, for a REMOTE type building, having the foam fit properly makes a huge difference in how easy it will be to complete the rest of the structure. Roof sheets were cut on the ground and then simply assembled, with almost no spray foam. The rigid foam for the roof valley was cut with a bevel for a tight fit, and as a bonus this also resulted in offset seams. (However, I did have to first hand cut the full sheet using a fine-toothed saw).
For me, the small amount of work required to make the hot-wire cutters more than paid off. But for other builders, the usefulness of these tools will depend on how much foam the project requires.
Securing the rigid foam to the walls
The first 3-inch layer of rigid foam was fastened to the wall using one or two 3 1/2 inch screws and Wind-Lock washers. The second layer was fastened using longer Headlok screws, again with a Wind-Lock washer.
The screws for the first layer are embedded by the next layer, so I didn’t bother removing the first-layer screws as I went along. However, the second-layer screws were removed as I added strapping, since they were thermal bridges and served no useful purpose after the strapping was installed. The holes were filled with a shot of spray foam.
Although not technically necessary, I covered the rigid foam on the walls with a layer of Tyvek housewrap. This primarily protects the foam during the rest of the build. But it also offers a first line of defense against water intrusion. (It’s cheap insurance).
The 1×4 strapping on the walls provides ample bearing on the foam, so there is virtually no compression unless the screw happens to be at an edge (such as near a window). In general, the screw will actually strip out of the wood before there is a lot of visible compression (which is a good reason not to try to countersink the head by overdriving).
The strapping was installed 16 inches o.c. using a 9 1/2-inch Headlok screw that penetrates through the foam and into a stud. The screws are driven in a slight upward direction (about 5 degrees off horizontal) to help prevent settling of the strapping when the siding is applied.
Each piece of strapping was marked from the drawing for stud location and then pre-drilled using a simple jig to set the angle. The hole is also countersunk so that the screw head does not interfere with the installation of the siding. (Countersinking the screw by over-driving it will cause wavy strapping or a stripped screw hole).
If a screw needs to be reset, the original hole was filled with spray foam. Note that different screw patterns will be necessary for different outer wall weights and foam thicknesses.
The corners presented an interesting problem, since there is nothing behind the rigid foam to screw to. Typically, two 10-inch-wide pieces of 3/4-inch plywood are used to form the corners, and these are held in place by screws near the edge driven into the corner studs of the wall. I worked out a slightly different method using 2x4s sunk into the foam and screwed through the diagonal into the corner studs. The 2x4s float on a channel in the foam, so the foam corners had to be precise.
This method allows for firm attachment of the corner and each piece of strapping can then run continuous to the corner, which I feel improves the strength of the assembly.
Rigid foam on the roof
The rigid foam on the roof presented different challenges than the rigid foam on the walls. The total foam thickness is greater (9 inches vs. 6 inches) and the strapping consists of 2x4s on the flat, not 1x4s. In addition, in my design the eaves have to be joined to the strapping and installed at the same time. So the whole process is a little more complex.
The rigid foam itself was installed in pretty much the same way as for the walls, except there are three layers rather than two. The fit at the eave and rake edge of the roof has to be reasonably accurate so that excess foam doesn’t interfere with the installation of the eave and rake overhangs.
Because the bottom of the eave also rests on foam, I was worried about sagging or the roof becoming uneven as it was assembled. I decided to use continuous 2×4 strapping from the ridge to the soffit with the eave pre-built on to the end of the strapping. I pre-cut and assembled each eave extension and then joined it to the strapping. As an added bonus, the eave assembly then keyed to the edge of the wall, so it actually made assembly a little easier.
As with the wall strapping, each screw was predrilled and countersunk using a small jig to ensure that the angle was consistent and the hole was straight. Predrilling the strapping was especially important for the roof, because the strapping and foam are both thicker. If you simply try to drive the screw without a pilot hole, then the slightest variation in grain direction will move the screw off axis and it will miss the rafter. I also inserted all screws before lifting the piece to the roof — it is much easier to work on the ground whenever possible rather than hanging off the roof.
This is a cold roof (ventilated) design, so the strapping had to run from top to bottom in order to create a ventilation channel. This meant that the strapping follows the inner rafters which in turn meant that warp and twist of the strapping created major problems when I tried to hit a rafter with the screw. A small twist of just 3 degrees would mean that the screw would miss the rafter (assuming everything else is perfect!). To minimize the problem, I rejected any strapping that was obviously twisted or bent. As it was applied, each piece of strapping was screwed first at the top and then the bottom to flatten it. I paid attention and adjusted for any twist when driving the second screw.
If there was a warp in the strapping, I snapped a line from center to center before installing the piece. After a screw was installed in each end, the center was pulled into the correct position, as judged by comparing the screw hole to the snapped line. The rest of the screws were then driven, and usually these lined up perfectly. Only about 5% of the screws missed on the first attempt.
If a screw did miss, it was fairly obvious. Trying to adjust and re-drive a screw without removal is difficult. (The screw tends to bend and follow the channel in the foam). It is easiest to just remove the screw, re-drill in a new location, and reinstall. The original hole is then filled with a shot of spray foam.
To ventilate the valleys, I used lattice-type strapping to allow air movement through to the ridge.
The addition of the rake extensions was pretty much standard building practice, but I did make use of continuous lattice type strapping to strengthen the overhang.
One other detail that may be of interest was the installation of an air-sealed insulated metal chimney support for the wood stove (available from ). A typical chimney collar and flashing arrangement is not even remotely airtight, and there is thermal bridging due to the clearance necessary for the insulated pipe component.
Finishing the outer roof
The rest of the roof was straightforward. A layer of 5/8-inch tongue-and-groove plywood was installed. Henry Blueskin rubberized membrane was used on the eaves, valleys, and rakes. This was followed by a layer of roofing underlayment. (Any roofing underlayment approved for metal roofing would be fine here, but I used the DuPont Roofliner since I had lots on hand). I then installed a hidden fastener metal roof.
I have been working on the inside of the house this winter without any of the inner insulation installed, and it is still easy to keep warm with a medium-size wood stove. I’m looking forward to seeing how the house performs when it’s finished.
Rob Myers’s previous blog was A Timber-Frame House for a Cold Climate — Part 2.
Rob Myers has worked as an analytical chemist, high-tech manager, and purveyor of fiery foods, all of which served to support a lifelong love of woodworking and building (not to mention a wicked tool habit). He is currently on a (possibly permanent) sabbatical from any real job while he builds an off-grid timber-frame home near Eganville, Ontario.