Big low-sloped mono-pitched roof: vented or unvented?
peter_r
| Posted in Energy Efficiency and Durability on
I’m building a new home outside of Vancouver (similar climate to Seattle or Portland) with a big mono-sloped roof at a shallow 2.5:12 pitch (see image below) and how the roof will be vented is keeping me up at night! The roof will be framed with 2x12s and there will not be an attic. The roof covering will be standing-seam metal sitting on plywood protected with a peel-and-stick membrane. The roof faces mostly south.
Vented option:
I strap the roof with 2x3s on the flat and at 45 degrees to create ventilation channels that go from lower to higher elevation (for stack effect), and nail my plywood sheathing on top. From below, I attach rigid but vapor-permeable insulation (e.g., rockwool or wood-fiber) to protect from wind washing, and then fill the rest of the 2×12 cavity with rockwool batts. I install continuous soffit vents around the entire roof perimeter.
Unvented option:
In this option I nail the plywood sheating directly to the 2×12 roof joists. I then spray 5″ or so of closed-cell spray foam, and then fill the rest with rock wool batts.
In both cases, I’m considering plywooding the bottom of the joists too, to create a reliable air barrier, and using 2×4 strapping on edge to create a “service cavity” so that I can install lightning, etc., without risk to the air barrier.
QUESTION:
The vented option will be more work and money and also has the disadvantage (aesthetically) of making my roof appear thicker than I want). But there’s something I really like about the fact that it can breathe to the outside as well as to the inside, although I’m not sure how well the ventilation will work because the roof is quite flat. Is this worth the trouble?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
Peter,
I don't think in this climate that any vented cathedral roof works below 3/12. I don't think diagonal strapping yields any appreciable difference, and I don't think a roof slope at or near 3/12 with that small an air-space, and that long a span, would effectively vent. So my vote is for unvented.
Just a note on roof thickness and aesthetics: Within reason, changes in the depth of the structure shouldn't affect the appearance of the roof, as the overhangs can be framed independently and be any thickness you want. Even applying foam above the roof deck can be dealt with so as to leave thin overhangs. This link shows how a local architect does it: http://www.osburnclarke.com/active-pas
Peter,
I agree with Malcolm. Here is a link to an article with more information: "Insulating Low-Slope Residential Roofs."
It seems to me that you may have made decisions on this house (including decisions on your 2x12 rafter framing) before you had an insulation plan. Insulation decisions should come early on in the design process. More information here: "Plan Ahead for Insulation."
The roof looks more than just a shed roof, I think all those complications plus the low slope make any type of venting impossible.
Going with TJI instead of 2x lumber reduces the thermal bridging of the roof, increasing the overall effective R value. Here they are cheaper than dimensional lumber for long spans, so it is an upgrade that actually pays twice.
Thanks for the replies, Malcolm, Martin and Akos. Malcolm, that's a beautiful home you linked to!
My wife just watched a video on spray foam and is now very much against it, and I worry adding 4" of rigid foam on top of the sheathing will either make my roof thicker visually than I want, or complicate the framing to hide the added thickness. My current preference is a solution with insulation between the ceiling joists, and then a (possibly-insulated) service cavity of cross-strapped 2x4s. This article talks about using rockwool in an unvented roof (rather than spray foam) and without any foam above the roof sheathing, provided the air barrier between the home and roof is very tight:
https://foursevenfive.com/blog/insulating-unvented-roof-assemblies-foam-free/
Isn't the argument here that what's key in a unvented roof assembly like this (no exterior foam) is to eliminate air movement from the conditioned space onto the cold condensing surface on the back side of the sheathing? And spray foam is just one way of accomplishing that?
Now, if an approach like this works without any ventilation, then what would be the harm of additional soffit ventilation and a 1" air space, even if the ventilation paths are crappy? Some air movement against the back side of the roof sheathing (or at least easier vapour diffusion) is better than none, no?
I understand why venting is bad if your sheathing has exterior foam on top, but if it doesn't then it seems venting is either pointless or helpful. I don't see how it can be harmful, all other variables held constant.
"Isn't the argument here that what's key in a unvented roof assembly like this (no exterior foam) is to eliminate air movement from the conditioned space onto the cold condensing surface on the back side of the sheathing? And spray foam is just one way of accomplishing that?"
I guess the question is: what are the other legal ways? Every roof here, vented or unvented, is required to incorporate a warm-side vapour-barrier and an effective air-barrier. The code doesn't say if you do a good job of those you can disregard the other requirements for the assemblies.
475 says their modelling shows these unvented roofs will work. Neither the IRC or our code agrees they are safe.
And that's your first hurdle. Both vented and unvented roof assemblies have distinct code requirements in the BCBC. You can't say I'm building an unvented assembly with a vent channel. It's an either or thing.
Thanks Malcolm. I see the issue now: the roof assembly 475 is discussing isn't code compliant. I missed that part.
Deleted
Peter,
I think people worry a bit too much about the extra thickness. From the ground level looking at a roof 15' to 20' in the air one would be hard pressed to tell how thick it really is.
I'm not sure what the BC code requires for R value, but ours (OBC) is R31. The exterior to interior insulation ratio is 40%, so that means R12.5 above and R19 bellow. You can reach that by putting 2x6 batts between the rafters pushed up against the sheathing and 2" of polyiso on top of the roof.
Sections of my roof have 2.5" of exterior of insulation, one would be hard pressed to tell from the street where there is exterior insulation and where it is just the 2x8 rafters.
Ignoring code, here is a partial list of things that effect sheathing rot risk:
Wetting air flow
Ventilation/exterior drying
Sheathing temperature (ext foam ratio)
Sheathing material
Interior humidity
Vapor retarder performance
You can made up for a lack in some of these (eg a 2.5:12 pitch vs 3:12) by improving some of the others. But it probably won't fit prescriptive rules/recommendations. A WUFI expert might be able to sort it out and come up with something that meets your requirements (including code) and provides lower risk than many prescriptive approaches.
You can add a vent channel (even a poor one) to a roof that meets unvented specs and IMO it will decrease risk.
What a tremendous amount of finicky work! All those parts needing to be installed perfectly in order to create a long lasting roof.
Why not give Insulspan in Delta BC a call and get a SIP roof? One long panel with structure, insulation and air/vapor barrier in one. A crane could put it on in one day. Multiple thicknesses and sizes available.
https://www.insulspan.com/
Scott,
You have touched on what I think is a very important point about these assemblies. They seem to ignore that houses are a) Built by fallible humans, and b) Get modified over time by both their occupants, and all the other things that affect the longevity of a house as time passes. Something that needs to stay perfect to work is asking for trouble.
Malcolm, I've always wondered why someone who wants to build a simple, unvented roof wouldn't just use a SIP panel. Why bother with rafters and sheathing and insulation, hiring people to install all those separate pieces and then try to close it all up tight so that it functions as an unvented roof assembly when you can get the same benefits in one product - a SIP panel.
Now, some people wonder about "how do I install lighting or wiring in a SIP panel ceiling - it's a big sandwich made out of OSB sheathing with foam in the middle? You don't, you add a shallow false ceiling to create room for the wiring and electrical boxes. Then there's no penetrations into your ceiling structure at all.
By the time you calculate out all the materials and labor needed to assemble your stick framed roof you could have the SIP roof installed. And it goes in so much faster.
Scott,
I think people are still gun-shy because of the re-occurring failures of SIPs roofs over the years. If a period of time goes by during which we see that the problems are past, I think they will enjoy better acceptance.
Peter,
How about a structural ribbed metal roof, like 1 1/2" B-deck, on top of your roof sheeting?
E.g. https://www.newmill.com/products/steel-deck/roof-deck/b-deck/standard-b-deck.html?optin=false
Large ribs up, small ribs down, vented at the lower and upper edges of the roof. About 2/3 of the roof area is ventilation channels. This has solid support, unlike the 2x3 wood strapping, and it has high beam and compression strength. Have the air barrier on the ceiling, and let the roof assembly vent upwards.
I've done this on 2/12 slope roofs, in climate zone 6 with snow loads, and the thermal buoyancy of the air does cause air to flow up the slope of the roof. (Air flow can be felt, with a moistened hand, at the high-side eaves of the roof, and it can be seen in vertical mirage in the air, when viewed parallel to the high-side eaves.)
Partially, the reason I tried this is that I am in a California Wildland (fire) Urban Interface Zone, so our WUI Code compliant prescriptive construction is un-vented roof assemblies (spray foam and fluffy batts). And, also, standing seam roofs leak, when water stands on them, as when ice dams water at the cold roof eaves. And there are no drainage channels between the flat roofing metal and the roof underlayments (a problem like perched water behind masonry veneers). So the low pitch, un-vented, standing seam roofs, in the Sierra Nevada mountains where I am, oftentimes have long-term leakage problems.
So, the air-channels of the B-deck, over a vapor permeable fire-resistant water-barrier roofing underlayment, over a permeable gypsum fire resistant roofing board, over an insulated roof structure, with an interior air barrier, provides adequate drainage and ventilation. And it is a very fire resistant roof buildup. And the deep ribbed steel decking is a strong structure.
Also, the vented ribbed roof is, more-or-less, a cold roof. So, at low slopes, without any snow-stops, or snow holders, the B-deck metal holds snow. Low-slope standing seam roofs, in this same area, are mini-glaciers of creeping snow, or are occasional, dangerous, avalanches of sliding/shedding snow. The expensive clamped-on-the-seams snow holders, on standing seam roofs, look very expensive, when you can see them, and they hold snow... until they don't hold the snow... and they depart the roof, with the snow. Not much of anything damages 20 gauge 0r 18g B-deck.
A plus is that the B-decking is insolation reflectively zinc coated. If someone wants decorative heat absorbing coloration, it can be painted colors... And (...I've never done this), if someone really wanted wood flitches or dead-dinosaur-tar-saturated-paper-stuff on the top of their roof, wood-product sheeting and the other stuff could be fastened on top of the ribbed metal...