Friday, April 28, 2017

Construction - First Cathedral Ceiling

The first cathedral ceiling presented an opportunity to apply the concepts detailed in the two posts on vapor and air barriers.  And it is interesting how much the final design of the ceilings deviates so much from the original design that I so naively and confidently detailed in a prior post before fully understanding vapor and air barriers.

I am deliberately inserting this post between the two posts on vapor/air barriers in order to reference it while discussing air barriers in the second of the two posts.

Original Design
My original design called for a 3 1/2" tall "mini-attic" between a fabric stapled to the tops of 2 x 12 rafters (with which to confine the blown-in rice hull insulation) and the roof sheathing. After more insight into moisture control and air infiltration/exfiltration in wall and ceiling assemblies, and, after a meeting with the consultant who will certify our project for energy efficiency, a different design for the mini-attic emerged. Also, instead of using 2 x 12s, I opted for trusses but only after thoroughly parsing I-joists as well. 

Perhaps the most succinct article on the subject of cathedral ceilings that I have seen is How to Build An Insulated Cathedral Ceiling on the Green Building Advisor site.  It clearly informed me that my original design would have been a disaster, that air sealing the ceiling/roof assembly is much more important than ventilation between the insulation and the sheathing, but that a dedicated air space sandwiched between a double layer of sheathing as described below for our project would be an advanced design worth the additional time and expense.

Roof Trusses
Two-by-twelves seemed a bad choice for three reasons.  First, 11 1/2" of insulation would yield an R-factor of only 35 when our target was at least 45.  Second, the thick 2 x 12s would allow considerably more energy-robbing thermal bridging than either trusses or I-joists. (Ever notice how easy it is to identify cathedral ceilings vs. conventional attics by the snow melt pattern on the roof?  Striations appear over cathedral ceilings because melting is faster over the 2-bys than over the insulation between them, whereas melting over conventional roofs is more uniform.) The third reason for avoiding long 2 x 12s is that their length and girth make it more likely that they come from old growth trees while I-joists and trusses have certifiably sustainable sourcing.  

I-joists 16" tall would have been a little cheaper than 16" trusses but would have required considerable job-site customization.  They would have
Roof trusses resting on truss walls (click on the picture to enlarge)
had to have been plumb-cut and reinforced on both ends and, since bird-mouths cannot be cut into the lower chords, the top plates of the walls would have had to have been fitted with wedge-shaped support boards.  And, while the trusses could be customized at the factory to fit flat on both 15" wide walls, the I-joists would have rested on the inside top plate on one wall and the outside top plate of the other. These considerations made the minor 
up-charge for the trusses a good trade-off.

The one downside to the decision, though, is that, while the trusses are significantly less thermal bridging than 2 x 12s,  I-joists would have been even better.  In hindsight, I would have used trusses 18" tall instead of 16".  The difference in cost would have been manageable and the extra two inches would have boosted the R-value by at least 7 points which would presumably off-set the loss from thermal bridging.

Mini-Attic Design
The air barrier for a conventional attic must be done at the level of the drywall as discussed in the recent post on vapor/air barriers.  In our case, however, the barrier will move to the level of the tops of the trusses due to our choice of tongue and groove pine ceilings that will be more permeable than drywall.  Instead of mesh on top of 2 x 12s as originally envisioned, I installed 1/2" plywood sheathing as the first of two layers of sheathing. The first layer will serve as the "floor" for the mini-attic; the second layer will double as its "ceiling" and as decking for the roof.  Since vapor passing through a wall or ceiling largely depends upon moving air, air sealing the floor of the mini-attic as described below will virtually eliminate vapor penetration through the pine ceilings.

At the time of this writing, I had covered the plywood with 6 mil sheet plastic anchored with batten boards to protect it for a few months until the mini-attic could be completed in conjunction with roofing the rest of the house.  And, for whatever it is worth, the first attempt to protect the plywood was a failure.  I conscientiously anchored the plastic with batten boards fastened with nails.  However, it took only a short time for wind blowing across the surface and coming up through the spaces between the sheets of plywood to heave the plastic enough to work the nails loose from the relatively thin (1/2") plywood.  The battens either blew off the roof or clung loosely to the plastic.  In either case, the nails protruding from them ripped holes in the plastic to the extent that I had to recover the roof with new plastic after taping the seams between the plywood sheets and filling the nail holes (last photo below).  This time I screwed the batten boards to place.  The moral is "use screws"; do not depend on nailed battens and don't even think that staples alone will work.

Just before the final roofing goes on, I will use construction screws to fasten 2 x 4s on edge on top of and fastened to the roof trusses through the first sheathing. I will then nail the second layer of sheathing to them.  The result will be a 3 1/2" ventilation space -- mini-attic -- that communicates with the outdoor air via continuous soffit vents in the north eave and a continuous ridge vent between the south edge of the roof and the overhangs for the second story windows.  

According to Joe Lstiburek at Building Science Corp., plywood for the first layer of sheathing is a better choice than OSB because it will allow water vapor to pass through it should vapor escape the living space, negotiate the less-than-impermeable wood ceiling and rise through the insulation. By contrast, the impermeability of OSB would block vapor which then could harbor mold, rot the sheathing, if not the trusses, and degrade the R-value of the insulation. OSB for the second layer of sheathing is acceptable however because any vapor from below will be vented from the mini-attic through the soffit vents and does not have to find its way through the second layer of sheathing.

The code calls for 1" minimum ventilation space between the insulation and the sheathing of a conventional cathedral ceiling.  Lstiburek suggests at least 2" for the air space while questioning the efficacy of any air space directly in contact with the insulation.  Our mini-attic will not only provide 3 1/2" instead of an inch or two but will also have sheathing separating the air space from the insulation.

The roof will overhang the walls 24".  I plan to extend the edgewise 2 x 4s that carry the second sheathing outward as support for the overhangs.  As discussed below, the 2 x 4s will not complicate air-sealing as would rafter tails extending from the roof trusses.

Sheathing the Short Truss Wall
The trusses are plumb cut flush with the short wall, i.e., there
 are no rafter tails extending from the trusses to interfere with
sealing the junction between the wall and the roof with a
continuous run of flashing tape
For the same reason I used plywood instead of OSB under the mini-attic, I used it for sheathing the short truss wall (and plan to use it for all of the exterior walls). It is important to note that, by plumb-cutting the roof trusses and leaving off the rafter tails, all of the wall sheathing could be abutted against the roof sheathing in order to simplify air sealing at the junction between the two. If rafter tails had been present, the sheathing would have had to have been cut and fitted around them -- a tedious job with a less-than-ideal outcome when it comes to air-sealing.  I was able to use a continuous run of flashing tape to seal the junction between roof and wall whereas, with rafter tails, tape, caulk and spray foam on the
 inside would also have been necessary for the inevitable gaps between the  tails and the wall sheathing.  

Air Sealing the Roof Sheathing
Blocking between trusses to stiffen the junction of the roof
sheathing and the wall sheathing and to facilitate caulking it
 from the interior, in addition to having taped the junction on
 the exterior (click on photo to see detail)
The clips used between sheets of plywoodsheathing are spacers to allow for expansion without buckling.  However, the space also would allow air infiltration and exfiltration that would be totally unacceptable.  Using caulk in the cracks would be counter-productive eventually since it loses its flexibility with age.  Spray foam would be rigid from the git-go.  So thank god for flashing tape. I used it not only to close the gaps left by the clips but also where the sheets of plywood met over the trusses. The nice thing is the tape will remain flexible indefinitely.

After the front wall and the rake walls for the second story have been sheathed withplywood, the junction between them and the roof sheathing will be handled in the same manner as the junction between the short wall and the roof. Then, considering that (1) the roof-wall junction is sealed with tape on all four sides, (2) the cracks between sheathing panels of both the roof and the walls are sealed with tape and (3) proper air-sealing is done around the windows when they are installed, the second story would theoretically be ready for a blower door test well in advance of the drywall stage.