Friday, January 29, 2016

Design - Roof and Cathedral Ceilings


The design of cathedral ceilings is an interesting, complicated and confusing issue and one that has caused much waffling on my part.  Apparently, the majority of cathedral ceilings extant today have been constructed with the top-to-bottom combination of roofing, sheathing, rafter/insulation and drywall.  And the outcome has been that most have failed, are failing or are destined to fail because of moisture condensation at the top of the insulation. So what should do we do different?

Red Flags
The entire roof for the living quarters of our house will be shed type with cathedral ceilings (no attic).  My research on cathedral ceilings has led me to conclude that the sheathing should not be applied directly to the tops of the rafters.  The emerging consensus is that 6 mil moisture barrier on the bottom side of the rafters is no answer (actually should be avoided in our climate) but that there should be ventilated space between the rafters and the sheathing -- a "mini-attic" if you will. 

And there are alternatives to a mini-attic .  A good one, if you can afford it, is to fill the entire space between sheathing and drywall with spray foam insulation, as opposed to using compacted fiberglass or compacted cellulose.  Spray foam completely blocks air infiltration and the moisture that it carries.  Another approach that apparently works as well or nearly as well is to hold conventional insulation back from the top few inches under the sheathing and fill the space with spray foam insulation.  This would seem to be an economical alternative to all spray.  Most of the other solutions that crop up online are mostly iterations on the mini-attic approach.

We plan to use rice hulls for insulation for both the walls and the ceilings,  And, since I have had no information on rice hulls as an air barrier, I plan to equate them to dense-packed cellulose and dense-packed fiberglass.  This means that, at the 15" thickness of our walls and ceilings, the hulls will probably stop most of the moisture-bearing air infiltration that penetrates beyond the air-tight drywall detailing

For more on the thinking that went into the final design for the roof, check out an earlier post:  Timeline - Design Evolution - Energy Efficient Roof.

Solving the Moisture Problem
In order to create a ventilated "mini-attic" above the insulation and avoid the moisture problem, I plan to fasten salvaged 2 x 4s on top of  the 2 x 12s rafters then nail the sheathing to the the 2 x 4s.  When insulation is added later, it will be held level with the tops of the rafters to create a 3 1/2"  space between them and the sheathing. Ventilation will occur when the "mini-attic" allows convective air movement between vents in the soffet at the lower end of the shed roof and vents at the upper end located next to the wall between rafters.

The ceiling that will support the rice hulls will be wood (see below).  It will have to be installed a little at a time and insulation blown in.  The question becomes, how do I keep from filling the 3 1/2" mini-attic with insulation instead of holding it back level with the tops of the 2 x 12s?  The best answer so far seems to be stapling some sort of strong fabric, such as fiberglass screening or weed barrier, to the tops of the 2 x 12s before adding the 2 x 4s.  It would not impede air and moisture transfer through the ceiling but would be strong enough to control the rice hulls.  

The disadvantage of using fabric is that the 2 x4s would have to be installed from above while balancing on the rafters rather than from below from a scaffold or ladder.  Or the fabric would have to be installed one rafter at a time as 2 x 4s were fastened.  This choice could be avoided altogether by using sheet goods, such as plywood or OSB board, over the the rafters instead of fabric.  The moisture accumulating at the top of the insulation would be pulled through the sheet goods by the air movement in the mini-attic but not quite as fast as through fabric.  The sheet good approach would have one other perk that interests me.  Air-sealing tape could be used over the junctions between sheets for an easy way to eliminate air-infiltration through the ceiling or, failing that, caulk could be used from below. If I did this in conjunction with our plan for something similar with the exterior walls, the entire envelope would be sealed.

If it weren't for having to maintain the space for the mini-attic, 16" tall I-joists could be installed to give an R-48+.  But, again, how do we maintain the attic space?  If I decide in favor of I-joists, it would have to be with a 12" height with the addition of screening or sheeting on top then 2 x 4s on edge.
Structural screws are available in a
variety of lengths

Fastening the 2 x 4s on edge to provide space for the mini-attic is structurally feasible and DIY-friendly due to the advent of structural screws and impact drivers.  Where long lag screws would have been used in pre-drilled holes in similar situations in the past, self-threading, "star"-driven construction screws are used.  Not only are they faster to use but are also considerably stronger at much smaller diameters than lag screws.  (The International code now even allows construction screw connections between rafters and the top plates in lieu of rafter ties.)

Increasing the Thermal Efficiency of the Roof
We can increase the thermal efficiency of the roof in two ways:  (1) top-to-bottom design that minimizes heat loss in winter then (2), adding foil-backed sheathing that also reduces heat gain in summer.

With rice hull insulation, at R-3+ per inch, flush with the tops of 2 x 12 rafters, the R-value for the roof would be 36+ which is 6 over the recommended minimum for our climate zone. However, buying good quality 2 x 12s in the lengths necessary to span the open areas of the house may be impossible or impossibly expensive.  Our original budget assumed that we would be using salvaged lumber* and I had not comparison-shopped dimension lumber vs. I-joists at the time of this writing.

In the unlikely event that another salvage opportunity comes up or 2 x 12s of sufficient length can be bought at prices that are considerably cheaper than I-joists, I would go ahead and use 2 x 12s. In which case, I am toying with the idea of adding edgewise 2 x 4s to the bottoms of them before attaching the tongue and groove pine ceiling.  The additional 3.5" of depth for the cathedral ceiling would increase the R-factor at least by 10, giving us a total R-factor approaching 50. (Parenthetically, the pine ceiling is both an aesthetic choice and a structural one -- to support to the rice hull insulation which is a mite heavier than fiberglass or cellulose.)

While salvaged 2 x 12s would have been about as green as it gets, engineered I-joists are greener than new dimension lumber because they minimize thermal bridging, are available in heights taller than dimension lumber to accomodate more insulation, are available in very long lengths and, best of all, are made from sustainable plantation trees. Moreover, they are manufactured to exacting standards. But their use requires special knowledge that experience with dimension lumber does not automatically impart -- a challenge I will have to meet if we use them.

While the "mini-attic" approach should solve the moisture problem, it does nothing to prevent the conduction of heat in and out through the rafters, i.e., thermal bridging.
I-joists 12" tall would be the ideal way to hold thermal bridging to a minimum because
Man-made I-joists

their vertical components are so skinny compared to dimension lumber. But, in case the I-joists are too expensive, I have been thinking about ways to avoid thermal bridging through conventional 2 x 12s. One solution would be to sandwich insulating shims between the bottoms of the rafters and the 2 x 4s installed below them. The shims could be cut from extruded polystyrene insulation boards and glued to place temporarily until the 2 x 4s could be fastened with long construction screws. When it comes time to do the final comparison shopping, I would not be surprised if the combination of dimension lumber, insulation and the pricey construction screws might make I-joists a reasonable choice after all.  More on this at the time of construction.

Actually, for our passive solar sun-drenched home, our worry is as much about heat gain through the roof in summer as heat loss in winter.  So I plan to pay a little more for OSB sheathing having a foil backing that will help to deflect the sun's radiant heat before it can raise the temperature in the mini-attic and challenge the R-factor of the rafter/insulation complex further down.

Steel Roof
Our roof will be a highly reflective light colored steel roof which, compared to asphalt shingles that are made from petroleum, lasts longer, has a recyclable end-life and is cheaper upfront.  I will underlay it with 30#felt then, instead of self-adhering bitumen-type material for the eave edges to prevent damage from ice dams, I will substitute unused roll roofing that was a Craigslist find at nominal cost. Unfortunately, our budget dictates exposed fasteners for the metal panels, as shown in the photo, instead of the more desirable concealed fasteners. For a more complete discussion of our choice of steel roofing, check out a prior post on roof cladding.
______________________
*  Originally, the rafters were to have been 2 x 12s 22 feet long salvaged from an old implement shed at about half the cost of new ones.  Until the "rafter fiasco" (discussed in another post on Craigslist shopping), all of my experience with Craigslist has been nothing but positive. However, the outcome that I have to live with is that all of the rafters will probably have to be store-bought.

______________________

Update:  Winter 2021
This only one of several posts have dealt with the evolution and final design for the roof.  To fast-forward, the one we built was supported by roof trusses and double-sheathed so as to create a dedicated 3 1/2" mini-attic space between layers and provide a space above the ceiling that was tightly-packed with insulation that was either 16" or 18" thick (+/- R-60).

2 comments:

  1. This is such a great resource that you are providing and you give it away for free. I love seeing blog that understand the value of providing a quality resource for free.
    Certified Precision Roofing

    ReplyDelete

As a do-it-selfer-in-training, I welcome your comments.