Thermal Bridging on Cathedral Ceiling
Climate zone 7B. 58lb snow load. No AC in most residential structures in this area.
I have a section of a proposed build that is 20×40 and the entire ceiling is cathedral with an 8/12 pitch, metal roof, and a couple of valley connections to the other rectangles of the floor plan. There will be heavy timber trusses in half and rafters and ridge beam in the other half. These are structural and also important architecturally. On top of that are concealed purlins (running perpendicular to the rafters and slope of the roof). These are planned to be 2x8s 24” oc. We use 5/8 OSB For roof decking in this area. I was planning to use a synthetic roof felt with a high perm rating like GAF Deck Armor. Then, thanks to learning things from this site, we’ll do the cross hatch 1×4 grid 24” oc and then a metal roof. We’ll be the first in the area (that I know of) not to put the metal right on top of Grace Triflex or something like it.
Seems like the best idea for insulation is flash and batt. We’ll use 5-6” of closed cell foam (CO banned the HFC blowing agent a year or two ago so that eases my conscience). I keep trying to figure out how to use repurposed roofing polyiso to counter the thermal bridging in the purlins, but this is somewhat problematic because I’ll need to add strapping outboard of the polyiso so the sheetrock screws don’t pop, thereby coming down the timber further with insulation, strapping, and sheetrock. Also, this makes hitting the ratio of closed cell foam more difficult.
what would seem to be easier is a sheet of EPS or GPS between the purlins and OSB. I’d go with the higher PSI rating. I like it because you put the EPS at the coldest section of the roof where it will perform the best, the foam is the air barrier, we’ll be sure to observe the 61% ratio of foam to fluff that Joe calls out.
the main question/problem is does this have a meaningful impact on the roof decking and purlin diaphragm? I know the conventional wisdom for rigid foam between studs and wall sheathing is it will reduce the shear strength. Though, it seems like Huber with Zip R has figured it works even with 2.5” of iso between the stud and sheathing.
building inspector is easy going and legitimately might not even see it so that’s not my issue. I don’t have an engineer for the house so it’s not an easy ask, e.g. I’m coming to GBA.
in general, does this sound like it will work and if so, how thick of foam board should I use? I assume for structure the thinner the better and otherwise the thicker the better. Plus, I don’t want to go too thick because I need to get the strapping for the vented roof attached through 5/8 OSB and foam to stud. That’s why I thought high PSI GPS and maybe 1” or 1.5”?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
To start with SPF is something you design out of new builds. In most cases there are much cheaper and more efficient ways to air seal and insulate.
Complicated roofs are hard to vent and hard to make energy efficient, so a bit of design up front can save a lot of build dollars and operating cost down the road. This doesn't mean you have to eliminate the architectural features such as your trusses and beams, just looking at different way to combine them in an efficient manner.
With trusses and purlins, the simplest way is to set rafters on top of the purlins running along the roof slope. Insulate both with batts but use batts for the rafters that are 1.5" to 2" less than the full depth. This will give you a vent gap above the insulation for a pretty simple vented roof. Since the rafters and purlins only cross over a small area, this eliminates most of the thermal bridging from the assembly. A vented roof also helps with snow melt in snow country which might not be a big issue at your snow load.
The rest of the roof, use I-joists instead of dimensional lumber. The thin OSB web of these eliminates most of the thermal bridging of the rafters. The bottom of the top flange of the I-joist makes for a convenient 1.5" spacer for a vent gap spacer. Parallel chord scissor trusses supported mid span by the beam is also a good option.
The one spot where you have to watch for thermal bridging is large beams cutting through the roof. A much better design is to set the beam under the rafters instead of flush, this way the rafter insulation can be carried over the whole beam.
The challenge with any of this type of timber/truss structure is air barrier continuity. Without careful up front design, it is easy to end up with a bunch of joints that relay only on caulk to air seal. With timbering, it is easy to have couple hundred feet of joints between the ceiling and the timbers. A couple hundred linear feet times even a very small 1/32" gap makes for a very large hole. Ceiling air leaks are usually the biggest source of energy loss in heating climate and also the main culprit behind ice dams.
The best is to set a sheet good over the timber truss/beam that overhang the timbers before the rest of the roof is installed. A wide piece of flashing tape is a simple way to do this, you can then seal the ceiling air barrier (ie 6 mil poly) to this tape. Same idea for the wall air barrier to ceiling joint.
Thanks for the info, but this doesn’t address my primary question about rigid EPS/GPS foam in between sheathing and purlins (could be rafters).
Also, in the roof assembly you describe above are the purlins concealed? Right now my plan is a concealed (by Sheetrock) purlin.
More on your described assembly- are you saying “sheet good” being osb before the purlins and the purpose being something uninterrupted to put a ploy layer on?
I won’t do any assembly that requires poly. It’s not used here and the few roof and wall problems that have occurred have either had SIPs or poly. Interestingly, the standard here for cathedral ceilings is 8-12” of open cell foam and unvented roof with triflex or something similar and typically metal roof(<25% shingle roof).
The short answer is you can't put rigid between rafters/purlins and sheathing. Besides code issues, I don't see how you can walk on such a roof to install the sheathing and there is no easy way to get an air barrier right under the foam. Without a proper air barrier, the assembly can have the same issues as a leaky SIP roof.
The purlins would be concealed. Having though about it a bit more, I would skip the purlins/rafters, hang a beam off your timber truss tip and set I-joists/truss rafters above the beam. Pretty much the same construction as the rest. Much simpler build with minimal thermal bridging.
You are correct, it is to keep the air barrier continuous across any timbering/beams. Sheet good could be OSB/CDX or wide flashing tape.
I doubt that a roof or wall would fail from poly in zone 7, would bet money the issue was elsewhere. SIP is a completely different animal and problem.
There is absolutely nothing wrong with poly above the ceiling in a vented roof assembly (cathedral or attic doesn't matter). Same for a stick framed wall without exterior rigid. It is also code in my area and in most cold climates. A smart vapor retarder is generally a better option though in almost all cases.
About the only assembly where poly can be a problem is a hybrid insulated roof. There you want a class II or III vapor retarder (ie painted drywall).
An unvented open cell SPF roof in your climate is asking for trouble, not something I would ever consider installing even in my much warmer zone 5 climate. A roof with issues is a very expensive fix, not an area where it is worth to take any chances.