Flat Roof Insulation Over Rafters
I am building an addition with a low slope roof (1/4″ per foot) that will also serve as a 2nd floor walkout deck. My plan is to build an unvented roof assembly using 2″ of high compressive strength (60 psi) XPS insulation (R10) directly on top of the rafters/joists and then install 3/4″ tongue and groove sheathing over that with no sheathing in between the rafters and insulation. I like the idea of the thermal break between the rafters/joists and the sheathing. The 2 x 10 cavity below will then be one layer of foil faced Polyiso and one layer of R30 Rockwool for a total assembly R-Value of 53. Any thoughts? Anyone with the math skills to calculate if 60 psi is sufficient compressive strength to ensure the insulation will not crush? Is there any downside I am not considering? We are just north of Philadelphia on the border of Climate Zones 4 and 5. I have added a sketch to illustrate the plan. My main question is–Is 60 PSI strong enough not to crush without solid plywood underneath?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Lets assume zone 5 for now. That means you need 40% of the assembly R value as exterior rigid. With R46 bellow that means R30 above. That feels like a bit too much rigid, I would reduce to code min here, in most cases for a roof with exterior insulation you can use U factor based compliance which would be R38 assembly.
This could be a single layer of R23 batt bellow and R15 rigid above. It is not a problem to not fully fill the rafter bay as long as the batts are pushed up against the roof, they can be held in place with insulation wires.
The rigid can't go directly over the rafters, it needs to go above the roof sheathing. This is not an issue with rooftop decks, you can set the decks onto 2x on flat to spread the load or onto adjustable deck block. Deck block as the easiest if you have the height as you can level the deck without having to deal with tapered floor joists. Regular 20 PSI rigid is fine.
My question is, Why can't the insulation go directly over the rafters? We are doing that on the walls putting R6 directly to the studs and then sheathing over that. If we can get the compressive strength in the insulation so that it won't crush on top of the rafters, it seems like a win win because we have a really good thermal break between the framing and the sheathing and we don't have to spend double for two layers of sheathing.
There is a big difference in something like Zip R which is engineered to be installed that way VS something DIY. You would have to get an engineer to sign off on it for the building department to accept it.
Walls are also not roofs, loads and air infiltration issues are very different.
Your sketch shows some cut and cobble as well, this needs to go. Cut and cobble is never air tight enough to prevent air leaks, once moisture makes it past the rigid, it has no way to dry out.
You also have to watch for the air barrier under the rigid. This needs to be solid to avoid any moisture issues down the road. You can google about SIP ridge rot to see what the problem can be if not addressed properly.
The simplest, lowest cost and least likely to cause issues is to go with rigid above the roof sheathing.
The design load for your deck is probably something like 40 psf live load + 10 psf dead load = 50 psf total load. At 16" o.c., that's 66.7 lbs per linear foot along each joist, or 5.6 lbs per linear inch. With 1.5" joists, that's 3.7 psi.
Foam compression ratings are based on 10% deflection so if you have 2" foam, at the rated load it will crush 0.2". That's more than you would want on a deck but there is no such thing as zero deformation; I would say that 0.02" (about 1/64") would be acceptable. Deformation isn't perfectly linear but pretty close so if you want to increase the bearing capacity by 10X then the compressive rating should be 10X. In other words, instead of designing for 3.7 psi, design for 37.0 psi. Round that up to 40 psi. 60 psi is overkill, in my opinion, but it should work.
There is another value that engineers use for foam under continuous load; for creep, they use a safety factor of 3X. But your deck will not be under continuous load, other than the 10 psf dead load, which is really more like 5 psf and not worth worrying about.
Others have mentioned your insulation ratio problem, which could lead to condensation. I'll also mention that XPS is the worst insulation you can use if you care about the environment; the blowing agents are 1400 times worse than CO2 in terms of climate impact, and CO2 is not benign, it's just the currency we use for measuring climate impact. XPS' insulating value drops to just over R-4/in over time. GPX would be slightly better.
When I have to use rigid foam I use either polyiso, EPS or GPS, all of which are available in high densities but it might take time to get them. In this case, polyiso would be the safest because it's impermeable to moisture. For a safe (40%) ratio, you would need at least R-20 in the foam layer and no more than R-30 in the mineral wool layer. Because the concern is long-term moisture damage and not just code compliance, I use the long-term, worst-case R-values; R-5/in for polyiso and GPS, and R-4/in for XPS, GPX, EPS and mineral wool.
Structurally, according to the IRC, wall sheathing needs to be in direct contact with the framing for resistance to lateral forces, but there is nothing in the IRC that says roof sheathing has to be in contact with the rafters. If you don't have another source for racking control, you might want to put down diagonal metal straps or something else to prevent lateral movement.
There may be a shear concern as well with having foam under the sheathing. I'm not certain but I'd definitely want to research it. Also, there may be code implications.
Best of luck with your project.