Spray foam vaulted ceiling with thermal break
I’m trying to build a pretty good house in the Midwest climate zone 5A. My house has some vaulted ceilings which need to be unvented because of the somewhat complex roofline. I’m also trying to use techniques my builder is familiar with (ie using spray foam instead of foam on top of roof). To get to a thermally-broken roof, my plan is to use 2×12 rafters with a 2″ strip (not sheet) of EPS/Polyiso tacked to the bottom, to give a total of 13 1/4″ of insulation space. The drywall will then be attached to this. To get to R60, I will use 5″ spray foam, and then fill the remaining 8.25 inches with either fiberglass or cellulose. Does this plan seem make sense given my contractor constraint or do you see some flaws? Thanks in advance. -Dale
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Dale,
You're talking about the flash-and-batt approach. In Climate Zone 5, a flash-and-batt roof needs at least 41% of the total R-value of the roof assembly to come from the spray foam layer.
You didn't tell us whether you were thinking of using closed-cell spray foam or open-cell spray foam. I hope you are thinking of using closed-cell spray foam -- that's the only way that this assembly will work.
You have about R-32 of closed-cell spray foam, and about R-30 of fiberglass, for a total R-value of about R-62, with 51% of the total R-value coming from the spray foam layer. That will work.
Martin,
Yes, I'm planning to use a closed cell spray foam. Do you think the strips of foam creating the R8 thermal break will sufficiently break the thermal bridging?
There's not really such a thing as "sufficiently break the thermal bridging", unless you have some sort of specific goal in mind. Less bridging is better, and having those foam strips significantly reduces it, relative to the standard assembly with no strips. It might even be enough to prevent prominent rafter-shaped snow melt patterns on your roof, though someone with more experience with this particular setup would know better.
To do a decent job of eliminating thermal briding you need about 1/2 to 1/3 of the R value of the cavity insulation to be the foam strip. So, so with a R60 roof, you need somewhere between R15 to R20 foam strip.
Putting much less than that will not make enough of a difference in the whole assembly R value to be worth the effort.
EDIT:
You can get a similar performance roof by using TJI joist for the roof and skipping the foam strip. The thin OSB web does a decent job of reducing the thermal bridging.
I'm not sure you are making the right comparisons there Akos. While it is certainly preferable to have a higher fraction of the R value in the foam strip, I think the real question is if adding the 2" strip significantly increases the performance of the assembly. It most certainly does: since it provides more room for insulation, the whole assembly R value goes up a fair amount (from ~R46 to ~R54), and even the R value of the "stud" portion of the wall nearly doubles (wood studs are a bit over R1/inch, so the 2" of foam brings that portion of the assembly from ~R12 to ~R23). Since the added cost and labor is relatively small, it seems like a worthwhile addition.
I agree that TJIs can be a good solution as well though. My quick calculation (I just cut the framing factor by four) puts a 13.25" TJI assembly at about equal performance as the 2x12s with foam strip, so the common 14" TJI is a good solution.
>"To do a decent job of eliminating thermal briding you need about 1/2 to 1/3 of the R value of the cavity insulation to be the foam strip."
Aedi has it right- there is no such thing as "eliminating" thermal bridging (short of making a monolithic structure out of a continuous insulating materials. ) The best that can be done is a thermal break, reducing the heat transfer through the thermal bridge.
Thermally breaking a more conductive material (such as a wood rafter) becomes very noticeable when the thermally breaking insulation approaches or exceeds the R-value of the conductive material. At 11.25" a 2x12 has an R value of about R13.5. At 2" foil faced polyiso comes in at about R12-R13, which cuts the heat transfer through the thermally bridging rafter by nearly half, noticeable/measurable thermal break.
It's pretty difficult to install ceiling gypsum long-screwing/nailing through 2" of foam though. Using 1.5" polyiso and a half-inch CDX nailer strip or 1x2 furring Bonfiglioli style is a bit easier, albeit somewhat less of a thermal break:
https://www.finehomebuilding.com/membership/pdf/9750/021250059.pdf
Are strips of CDX or 1x2s as described below strong enough to hold up in a cathedral roof application? I'm considering this exact application, but in 10 years don't want the inside of the roof to come crashing down!
"It's pretty difficult to install ceiling gypsum long-screwing/nailing through 2" of foam though. Using 1.5" polyiso and a half-inch CDX nailer strip or 1x2 furring Bonfiglioli style is a bit easier, albeit somewhat less of a thermal break:
https://www.finehomebuilding.com/membership/pdf/9750/021250059.pdf "
1x4s are commonly used as ceiling strapping to attach drywall to. You could put this same arrangement of 1x4s up over your foam. I wouldn't trust 1/2" plywood here though -- I think you'll get too much deflection. 1x2s tend to be pretty crummy wood usually, and will probably split pretty easily. Strips of 3/4" plywood might be an option though, similar to what is commonly used to create a rainscreen gap on the exterior.
Bill
Thanks Bill. with 1 3/4 wide LVLs and rounding up to two inches for 3/4 plywood strips does seem like it would be stronger along with 1x4s. Thanks.
I am with the above posters on using TJI's for rafters to add height for insulation and the reduced thermal bridging through the plywood web. They also work nicely for creating a dedicate airspace for vented cathedral roofs. It is very easy to attach a rigid material to the underside of the top flange to create the airspace. Another plus is full insulation over the exterior walls. Have used this system many times for additions and winter snow lays uniformly on the roof, no heat loss related melting.
In the Northeast it is pretty standard to run 1 by 3 strapping perpendicular to the roof framing. If you did this after you put up your thermal break, you can screw the strapping to the framing and then easily install your GWB to the strapping.
Peter
One thing that is nagging at me from this setup is whether the thermal break in the studs is in the right location.
Say if we were looking at steel studs with spray foam in between, no amount of foam strip against the steel stud would give you back the full R value of the spray foam. Essentially, the thick layer of spray foam between the studs would be mostly wasted.
Would it not be also the case here but to a lesser extent with 2X dimensional lumber and spray foam? Essentially for the thermal break to be effective, should the foam strip be installed beside the high R value /inch spray foam?
Thank you to all for the thoughtful and thorough responses. -Dale