Roof construction for a low-slope roof?
Hello,
Our house is being built near the boundary of zone 4 and 5 in Illinois with a 1:12 pitch metal roof. I’ve done a lot of research on low-slope roofs because our builder does not have a lot of experience with flat roofs. After reading a couple of Martin’s articles and his advice on a previous question, we’re thinking we would like to create an unvented roof assembly (see sketch). We’re thinking of filling the rafters with enough dense-packed cellulose to get the total roof assembly to R50. I also included our elevations. Since this is all new to us, my wife and I would greatly appreciate a little help with a few questions we have.
Does this diagram look like a good approach?
Are the rafters built using roof trusses?
Any details that I should pass along to our builder about the roof trusses?
Should the ceiling be flat or should it follow the roofline? We’re open to either option.
Thanks in advance!
Kevin
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Kevin,
First, check with your roofer that the type of metal roofing that you intend to install can be used on a low-slope roof. Some types of metal roofing can't be installed on a 1:12 roof.
Q. "Does this diagram look like a good approach?"
A. Yes. You'll need 8 or 9 inches of cellulose to get R-30.
Q. "Are the rafters built using roof trusses?"
A. You could use 2x10s or trusses -- either one -- as long as the span isn't too long. If you don't know how to size rafters or trusses, talk to an engineer.
Q. "Any details that I should pass along to our builder about the roof trusses?"
A. Make sure that they can bear the roof load and the snow load; have them sized by an engineer if your builder has any doubts.
Q. "Should the ceiling be flat or should it follow the roofline?"
A. Either way will work. If you don't want your ceiling to be at the bottom of the rafters, you'll need to install some type of air barrier at the bottom of the rafters. This air barrier will have two functions: it needs to contain the cellulose, and it needs to stop air flow.
Thank you Martin! Exactly the info we were looking for. If our ceiling follows the roofline, the air barrier would be the drywall, which would also contain the cellulose, correct? Seems like this would be a cost savings rather than creating both an air barrier and the flat drywall below.
Kevin,
Yes -- drywall is an excellent air barrier.
One final question that just popped into my head - would recessed lights, ventilation fans, or other fixtures attached to the drywall spoil the drywall air barrier?
A couple of comments...
4" of XPS won't perform at R20 for the full 50+ year lifecycle of most metal roofing. As it loses it's blowing agent over a handful of decades it's performance starts out higher than R20, but eventually drops to about R17. The R5/inch labeling is based on something like a 20 year average performance.
R17 might be sufficient for dew point control, but it might not be.
If instead you went with 5" of EPS, it's R value will be pretty much the same on day 25,000 as on day-1, since it's much lighter blowing agent dissipates quickly, and it's labeled-R is it's fully-depleted R.
There are air-tight gasketed insulation contact rated recessed light fixtures, but they need to be inspected and installed carefully for air-tightness. There are also surface mount LED fixtures that can be mounted on standard electric boxes that present a far smaller and far shallower penetration into the insulation layer, which may be a better alternative. Any penetration of the ceiling gypsum needs to be detailed for air tightness.
Kevin,
Q. "One final question that just popped into my head - would recessed lights, ventilation fans, or other fixtures attached to the drywall spoil the drywall air barrier?"
A. Absolutely. You want to minimize all penetrations and electrical boxes in an airtight ceiling, and all penetrations need to be very carefully air sealed. If you care about energy performance, you won't have any recessed can lights in your ceiling.
Here are links to two articles with more information:
Questions and Answers About Air Barriers
Ban the Can
Martin,
After discussions with our builder and supplier, we are now considering a 60 mil EPDM roof instead of standing seam metal. From an aesthetic standpoint, the middle portion of the roof will not be seen because it slopes towards woods to the west. With the 1:12 pitch, I don't think you'll be able to see much of the outer portions of the roof either.
Would the EPDM roof perform as well (or better) as standing seam metal?
Looking at the cross section that I included in the original post, would the cross section for EPDM be the same (substituting EPDM for the metal)?
Any other considerations I should know or discuss with our builder before deciding between EPDM and standing seam metal?
Thank you for all your help with these important decisions for us!
On low slope roofs in the snow zone EPDM will be work better than standing seam, since it can't leak even under high winds, whereas metal roofs will. With limited slope to drain well it can take forever to be rid of any leakage moisture. EPDM won't last as long as metal, but there's a reason it's the most common roofing for very low slope roofs.
Take the long term R value degradtion of XPS seriously if you're only putting up 4" with R30 on the interior. If you're on the zone 5 side of the line 5" would be better (or 5" of EPS). If the cavity-R is over R30, the long term R of the above deck insulation needs to be proportionally higher. In zone 5 make the exterior R at least 40% of the total, in zone 4 it's 30% MINIMUM.
On low slope roofs in the snow zone EPDM will be work better than standing seam, since it can't leak even under high winds, whereas metal roofs will. With limited slope to drain well it can take forever to be rid of any leakage moisture. EPDM won't last as long as metal, but there's a reason it's the most common roofing for very low slope roofs.
Take the long term R value degradtion of XPS seriously if you're only putting up 4" with R30 on the interior. If you're on the zone 5 side of the line 5" would be better (or 5" of EPS). If the cavity-R is over R30, the long term R of the above deck insulation needs to be proportionally higher. In zone 5 make the exterior R at least 40% of the total, in zone 4 it's 30% MINIMUM.
EPDM is great, and you can get it in white, which will reduce the cooling load.
A cheaper alternative that I usually use is spray polyurethane:
http://www.trsfoam.com/sprayed-polyurethane-foam/spf-myth-busting
The upside is that a well maintained SPF roof can last indefinitely. (It gets recoated every ten years or so). The downside is that you should inspect it carefully every year.
Kevin,
As an architect I'd be inclined to take another approach and increase the slope on the roofs. By going to 3 in 12 I'd bet both the interior spaces and and exterior elevations would benefit - and you could use conventional roofing details.
Edit to address your other questions:
As none of your windows appear to be above 8 ft or so you could use mono trusses for all the roofs without much compromise. If you decide to slope the interior ceilings you might want to think about some higher ones to bring in more light. With sloped ceilings you still have the choice of solid lumber, I joists or parallel chord trusses for the structure.
My builder and roofer are suggesting a vented assembly as shown below. They have not done an unvented roof with the layered foam above the sheathing and seem more comfortable doing the following approach and say it would save a lot of $ in time and materials. The trusses were designed such that there would be 2' at the narrow ends of the roof. Our builder would like to blow about 16" of cellulose and leave the area above the cellulose vented. He recommended no venting in the center of the roof due to more chances for leaks. What are your thoughts on this approach? Any concerns? Thanks for all your help!
Kevin,
It might be time to find a different roofer. If you can locate a roofer familiar with commercial construction, you'll be in better hands -- because installing rigid foam above the roof sheathing is a standard method of insulating low-slope commercial roofs.
The approach shown in your sketch is associated with failures, because there isn't enough of a difference in elevation between the "inlet" vents and the "outlet" vents. I call this type of ventilation "faith-based venting." Air rarely follows the "smart arrows" that some designers draw to indicate where the air is supposed to enter, and where it is supposed to leave.
For more information on good ways to insulate this type of roof, see Insulating Low-Slope Residential Roofs.
Martin,
A commercial general contractor put me in contact with a roofer familiar with low slope roofs. This roofer and his engineer have proposed using fully adhered 90 mil EPDM with 1/2 inch mechanically fasten densdeck between the EPDM and plywood. After showing them the previous 2 sketches (vented and unvented), they suggested an unvented assembly. However, rather than laying foam above the sheathing, their suggestion was to only insulate from underneath the decking using a blown fiberglass with some sort of adhesive that makes it bond to the decking or a spray foam. Will this approach work or should I insist on foam insulation between the EPDM membrane and the sheathing? Also, is the 90 mil thickness overkill? Most of the applications I have read about use 60 mil or 75 mil, but we do want a roof built to last. Thanks for helping educate me on this important issue for us.
Kevin,
I stand by my advice. For an unvented flat roof, you want at least some -- or ideally, all -- of the insulation to be above the roof sheathing. That way the roof sheathing stays warm and dry.
Martin,
I have gone back and forth with the roofer several times. After discussing the importance of foam above the sheathing, our roofer seems reluctant to do 2 layers of XPS foam above the sheathing and is now talking about a vented assembly again. The roof trusses and sheathing are on and we need to move forward on this issue. So I guess I'm looking at 2 options now:
Option 1. A vented assembly like shown in this sketch. After reading your article on insulating low-sloped roofs, I also discussed adding center ventilation to the roof. The roofer recommended adding 24" vent in the center of the roof like shown in this link:
https://www.roofvents.com/flat-rooftop-vent.html
Would the combination of the soffit ventilation and a center vent like shown in previous link provide adequate ventilation if the ceiling was air-tight?
Option 2 (that builder seems reluctant to do). Unvented assembly with 2 layers of 2" XPS between the EPDM/densdeck and sheathing. Also spray enough insulation directly to the bottom of the roof deck to get at least R49.
I should add that the ductwork will be in the floor trusses between the first and second floor.
Thank you!
Kevin,
Between the two options, I would choose Option 2. Remember that this option requires you to insulate the attic walls.
If you want to improve Option 1, skip the 24-inch vent (which will be prone to leaking) and install a real doghouse vent as described in this article. (Whatever you do, avoid the temptation to use a powered attic ventilator -- in other words, a fan.)
There is also Option 3: Skip the rigid foam above the roof sheathing, and install R-49 of spray foam under the rafters. I don't recommend this approach, because it is expensive, but it would work.
This may be a dumb question, but do you have a picture, link, or details about how to construct a doghouse vent? My attempts to Google returned a few pictures of actual doghouses with a vent but no details about ventilation for a home.
Kevin,
Here are some images. The first image is the one that appeared in my Fine Homebuilding article about insulating this type of roof (Insulating a Low-Slope Shed Roof).
.
Martin,
Another roofer has proposed using 4.4" of polyiso insulation above the sheathing for an unvented assembly. This builder said that XPS and EPS aren't compatible with the EPDM adhesive. Quote from builder - "Like pouring gas in a styrofoam cup, it melts it away. They are typically used in ballasted or mechanically attached systems."
The issue I have with polyiso is that I have read that the effective R value in cold weather is lower, and polyiso should only be used in warmer climates.
The roofer also proposed using closed cell foam on the underside of the decking. The two insulation contractors I spoke with also recommended closed cell over open cell for this application. Their concern with open-cell is that it would absorb moisture like a sponge and get moldy. However, I have read in your articles that open-cell foam should be used in this situation because closed-cell foam does not allow the roof sheathing to dry to the interior.
What are your thoughts on the polyiso? What are your thoughts on closed cell vs open cell foam on the underside of the decking? Should I be concerned about the open cell foam absorbing moisture and becoming moldy like this particular roofer and the 2 insulation contractors mentioned?
Thank you.
Kevin,
Q. "What are your thoughts on the polyiso?"
A. Here is a link to my latest article on the polyiso question: Cold-Weather Performance of Polyisocyanurate. I tend to agree with John Straube, who told me, "One option [to the polyiso dilemma] is to stick with polyiso and just make it thicker. If we do that, let’s call polyiso R-5 per inch."
Q. "What are your thoughts on closed-cell vs. open-cell foam on the underside of the decking?"
A. Either type of spray foam can work, as long as you are sure that the roof sheathing is very dry on the day that the spray foam is installed. It's true that open-cell spray foam has been associated with some wet sheathing problems; to learn more about the circumstances surrounding these failures, see Open-Cell Spray Foam and Damp Roof Sheathing.
In your case you have to weigh the advantages of open-cell spray foam (it's more environmentally friendly, and it will allow your sheathing to dry to the interior) with uncertainties concerning reports of damp sheathing problems. I don't think there is any risk of damp sheathing with open-cell spray foam if you have an adequate thickness of rigid foam above the roof sheathing, because the rigid foam will keep the sheathing above the dew point during the winter. Warm sheathing is dry sheathing.
Our roofer has suggested 2 different approaches:
Option 1.
90 mil EPDM - 1/2" cover board - 2" polyiso - 7.5" EPS. This option would also require a 10" knee wall built around perimeter of roof. Total cost about $53,000. Also said to spray foam vertical walls inside the rafter area for additional cost of about 5000. His proposal also included a 7 1/4" fascia - curious how this would work since the roof assembly will be thicker than 7 1/4".
Option 2
90 mil EPDM - 1/2" cover board - 2 layers of 2.2" polyiso. This option would require 3 2x4's nailed around perimeter. Also need to add 6" closed cell spray foam on underside. Proposal also includes 7 1/4" fascia. Total cost about $68,000.
Would you guys recommend option 1 or option 2? We have about 3800 SF of roof.
I should quit my editing job and go back to being a roofer... I could make some real money then.
Kevin,
If those are your only two options, I would choose Option 1, which provides more rigid foam above the roof sheathing, and has a lower total cost.
What Martin said- option 1 is the better approach.
The 6" of closed cell foam of option 2 is both more expensive, and thermally bridged by rafters, undercutting it's performance. In option 1, even if the2" polyiso is derated to R4.5/inch you would have a continuous R40+. which would beat code max for compliance on a U-factor basis (rather than on center-cavity R value).
53k or 68K? Yeow!
Edit: Maybe I should expand on that a bit. We aren't party to how this project evolved, but it appears to be a good illustration of the problems that can occur when a house is designed without taking into account the implications of the decisions being made. Specifying a low slope roof, without any apparent benefits, which ends up leaving the owner with two expensive and unpalatable construction choices, just makes no sense.
i come back to my earlier comment: It would take very minor changes to end up with a conventionally built roof, which would also make the elevations more aesthetically pleasing. The best way out of this situation is to avoid it altogether.
This is my 3rd comment of the morning ... great site. Earlier this year we had a building designed, with a flat roof, 30x50 ... open web trusses a bit oversized for load limits, ply, then ice and water, then foam, the ply and epdm... I enjoyed the view from up there and the lines were good to look at etc... I put up some plastic over the Ice and water because a storm was coming.
We had warm sunny days, on third day went to take plastic off, still water up there on top, then reconsidered last winter, then promptly ordered another set of trusses with an 10 pitch, bigger overhangs and put up a metal roof. No flat roofs for me any more.