GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Insulating a low-slope EPDM roof from the ceiling below

hobbitrock | Posted in General Questions on

Hi. I recently had an Epdm roof installed on my 140 year old house. It’s low slope and meets with a gable roof on the other part of the house. The Gable is standing seam. We had a tapered ISO board system installed as the underlayment to help with drainage and make the edge details easier to deal with. I reframed the roof and sheathed it myself before the roofing contractors started so I could save money. At the time the ceilings underneath the roof were all closed so I installed 4 inch rockwool batts right on top of the plaster/lath of the ceilings below. I was trying to maintain a 2 inch air space for ventilation and even though 4 inches or rock wool doesn’t give me much R value it was better than nothing which is what it had been before.

I then created a continuous soffit vent with cor-a-vent and made sure there was at least 2 inches of space between the old skip sheathing of the gable so that it could all share the same vent space. The roofers installed a ridge vent on the metal gable roof.

Long story short. I have since torn open the ceilings of 2 of the rooms under the EPDM roof meaning all the insulation I did above those rooms needs to be redone. I started researching again recently because of the cold winter we’re having and realized that because I have ISO board on top of the roof sheathing, the roof assembly should probably be an unvented “warm roof” situation.
If so that’s fine for the two rooms I have open. The other room will just have to stay the way it is or have cellulose blown in from the attic side providing I can seal of the rafter bays on the soffit side.

My question is…..is my thinking correct on this? Should I ignore the fact that I built a soffit vent and just fill the cavity between the drywall and the underside of the roof sheathing? And if so is rock wool a good product to use considering I can really pack it in there much better than fiberglass. Note- the space is very small….too small to access from gable attic in any way. 6 inches at the eaves maybe 11 or 12 where it meets the gable
I have started insulating the bathroom with this design in mind by cutting 2″ XPS and installing it tight to the underside of the roof sheathing, spraying the voids and gaps with canned spray foam with the intention of then filling the remaining cavity with Roxul? I really don’t want to rent a spray foam kit or hire a contractor for these relatively small spaces. Budget will not allow.
I apologize if it seems like this question is something that has been answered here before. I guess because of the age of my house and the limited access it seems like a unique dilemna. The performance and durability is more important to me than energy efficiency in this case simply because tearing off and re roofing in a few years due to rot from condensation seems like a worse thing to do than having insufficient r value. Of course a warm roof design that actually worked would bring my R value up significantly…..probably close to recommended levels.
I also thought about blowing cellulose in the bays that I don’t have access to. I could do this from the attic if I seal off the bays on the soffit side. This would be the dreaded dense pack approach but I’m wondering….since I have ISO board on the roof roof is the dense pack method really so bad? Thanks.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.

Replies

  1. GBA Editor
    Martin Holladay | | #1

    Joseph,
    You are correct that it makes no sense to introduce a stream of cold outdoor air between two layers of insulation -- the mineral wool insulation below and the polyisocyanurate insulation above. If you ever want to create a vented roof assembly, remember that the ventilation channel has to go above the uppermost layer of insulation.

    There are a few rules of thumb that you have to follow when you are insulating a low-slope roof or trying to create an unvented roof assembly. Here is a link to an article that will explain what you need to know: Insulating Low-Slope Residential Roofs.

  2. hobbitrock | | #2

    Thanks so much for the quick response. When the roof was going on I was unaware that the iso board was anything more than a substrate. I now realize that I totally wasted my time building the soffit vents. It would seem that the presence of iso board would make the roof lean heavily towards a warm roof design by default. I just don't see iso board mentioned very often in the discussions about such things. I also am finding that installing the xps on the underside is harder than I thought it would be due to many framing members sistered together in the bays. That's why I'm considering blown in cellolose. That and the fact that I have limited access to portions of the roof. I will re read your article this evening after work. Again, thanks for the reply.

  3. Expert Member
    Dana Dorsett | | #3

    How thick is the polyiso at the thinnest edge, and what is your location?

    Both are critical in determining exactly how much fiber insulation can be installed below the roof deck without risking excessive winter moisture in the roof deck. Depending on the answer pounding it full of cellulose may be mistake, or it could be the best thing you could possibly do.

    Installing XPS from the under side may be a mistake too depending on the thickness, since XPS has a fairly low vapor permeance and could potentially create a moisture trap.

  4. hobbitrock | | #4

    Hi Dana. I'm in New York State. Iso board is 1/2 " at the edges I believe. The rafter bays are 6 inches at the eave and the depth increases from there as it rises slightly to meet the gable. I am using 2" thick XPS glued to the underside of 5/8" zip system sheathing. The glue is foam board adhesive. My idea with the XPS was to get a solid layer of closed cell foam without having to spray it. I am spraying the voids at the edges of XPS with great stuff.

  5. GBA Editor
    Martin Holladay | | #5

    Joseph,
    You are located in Climate Zone 5 or 6, depending on what part of New York State you're in.

    It sounds like you want to create an unvented roof assembly that is insulated with a combination of rigid foam and mineral wool.

    Building codes require that the foam layer of your roof assembly (the polyiso above and the XPS below) have a minimum R-value of R-20 (in Zone 5) or R-25 (in Zone 6).

    We can assume that the 1/2 inch of polyiso above your roof sheathing has a (nominal) R-value of R-3. That means that the foam that you plan to install directly under your roof sheathing needs to have a minimum R-value of R-17 (in Zone 5) or R-22 (in Zone 6). Then you could make up the rest of the needed insulation (up to the R-49 requirement) with mineral wool batts.

    What I'm trying to tell you is that your plan to install 2 inches of XPS (R-10) under the roof sheathing with the cut-and-cobble method isn't sufficient. You need a thicker foam layer to stay out of trouble. Moreover, cut-and-cobble insulation is somewhat risky.

    To learn more about minimum R-values for foam insulation in this type of roof assembly, see How to Build an Insulated Cathedral Ceiling.

    To learn more about the cut-and-cobble method of insulating, see Cut-and-Cobble Insulation.

  6. hobbitrock | | #6

    Hi Martin. Thanks again for the replies. I am in Climate Zone 5. Truth be told, I originally wanted to have a vented roof. But I got scared off of that idea from various sources saying that if I had only 2 inches of air space then it would be inadequate. Plus the realization that I've got insulation on top of the roof deck. Even if it's thin at the edges.
    This project was a re-roof on an old house. Nothing about the profile of the design was changed. Because it's only 6 inches deep at the eaves where there's only 1/2" of iso board above, I'll never be able to meet code. I can only do the best I can do. It's actually sort of a fluke that I'm even able to access the portions of the underside of the roof that I've got access to. I had not planned on ripping those ceilings out at all.
    So......cut-and -cobble. That's funny. Yeah, as soon as I started doing it I began to regret it. It's a lot of work. But it is installed in a portion of the open ceiling and ripping it out sounds like a nightmare.
    Sounds like my best option is a combination of three different methods-
    1) go with a 2 part spray foam on the underside of the xps that I've already installed creating the "peanut brittle" configuration, then fill with roxul.
    2) Use the same 2 part spray foam on the portion of the roof that I have access to that doesn't have xps installed, then fill with roxul which by the way would be dense packed from below by hand.
    3) blow in cellulose into the bays where the ceilings are intact. Even though the dense packed flat roof is warned against, I've got iso board above the deck which is one of the ways it can work. Right?

  7. Expert Member
    Dana Dorsett | | #7

    In zone 5 the exterior R needs to be at least 40% of the total center-cavity R on a roof to be able to do it without an interior vapor retarder. A half inch of iso (R3) isn't going to cut it against 5.5"-6" fiber (R18-R24). It may be fine at the thicker end though.

    Blown cellulose in the cavities would share the moisture burden with the roof deck, which adds some resilience. But you would also want to use a "smart" vapor retarder like Intello Plus or Certainteed MemBrain on the interior side, between the cellulose & ceiling gypsum, detailed as an air barrier.

    Relying on cut'n'cobble is indeed risky over the long term, and is no substitute for sprayed ccSPF under a roof deck, since the cut foam isn't exactly bonded to the roof deck- even a small air leak can create a big problem. The additional performance of the higher-R cavity fill with cut'n'cobble or closed cell polyurethane is remarkably low, since more than half the heat transfer would still be happening through the rafters. You'd be better off filling the rafters with cellulose or rock wool, and installing continuous 1" XPS on the interior side of the rafters, since you'd have about 1-perm of drying capacity through the foam (which is nearly a class-II vapor retarder at 1"), the foam would be thermally breaking the rafters, and the foam would be in a location where it's easy to get a reliable air seal. The cellulose-fill + smart vapor retarder solution is somewhat lower thermal performance, but offers far more moisture resilience than the 1" continuous XPS.

    If you installed the MemBrain at the rafters it would be fine to put as much as 2" of Type-II EPS (R8.4) between the vapor retarder and ceiling gypsum as a thermal break. OR, you could install strips of 1.5-2" polyiso on the rafter edges and use deeper cellulose with the vapor retarder between the fiber & polyiso strips and the gypsum.

    Unless the rafter spacing and depth in this antique is magically identical to current framing standards, batt insulation is nearly impossible to get right. Blown insulation is really the right solution, 19 times out of 20. If you can't find a cellulose contractor who can do it within your budget, this is DIY-able with a 1-stage box-store rental blower, using landscaping fabric stapled to the rafter edges ~2" o.c. You won't be able to hit 3.5lbs density, but you should be able to hit 3, which is good enough if you're also installing an interior side smart vapor retarder, which should limit the moisture-cycling of the cellulose sufficiently to prevent settling over the long term. (3lbs density is over 2x the density of open-blown attics.)

  8. hobbitrock | | #8

    Thanks Dana. Yeah- Cut and Cobble is out. I knew it as soon as I started that it couldn't possibly work for my situation. Access is simply not good enough. I'm going to get an estimate from a spray foam company next week but honestly I think I'm leaning towards going back to a vented design.

    I was checking in my attic of the adjoining gable roof to see how the access would be for blown in cellulose and realized that the air way is not as bad as I had originally thought. I know my soffit vent is adequate and if I'm willing to lose a bit of ceiling height which I'd have to do to get the r value then I could have a pretty open vented channel. Not 6 inches but still pretty breezy.

    Question if I go that route- you had mentioned installing rigid foam on the underside of the framing.
    If taped properly this would go a long way toward air sealing the ceiling plane- correct? Which is effective in keeping the warm air out of the roof assembly. Gotta be a good thing right?. If I did this though should I expect problems because I've got an assembly with two vapor barriers. The foam between the drywall and the framing being one, and the ice and water shield/iso board/EPDM above the decking being the other.

    By the way, I don't mean to use the terms vapor barrier and air barrier interchangeably. I just assume that the rigid foam board has both characteristics to some extent. What I'd really like to do is simply leave the venting channel open, insulation below that, then sheetrock. No vapor barrier other than the latex paint. At least that way if I had a moisture problem I'd be aware of it sooner. Or if the rigid foam is somewhat forgiving ( you mentioned it having 1 perm of drying capacity) then that would be ideal. I guess I don't really understand how that works but I don't like the idea of a roof assembly with rubber on top and some sort of waterproof material below.

  9. GBA Editor
    Martin Holladay | | #9

    Joseph,
    Re-read my article on insulating low-slope roofs. If you want a vented roof assembly, you really need a minimum of an 8 inch gap between the top of the insulation and the underside of the roof sheathing. You also need to install a cupola or doghouse in the middle of the roof with vents on all four sides.

    Low-slope roofs don't get the airflow through ventilation channels that steep-slope roofs get, so the rules are different.

  10. Dana1 | | #10

    Low slope roofs are nearly impossible to vent adequately.

    At 2" XPS runs about 0.6-0.8 perms which is only a class-II vapor retarder. The vapor retardency of 2lb spray polyethylene is comparable. That's fully an order of magnitude higher vapor permeance than vapor barriers such as 6 mil polyethylene, and offers sufficient seasonal drying capacity toward the interior.

    Rigid foam is not an air barrier unless detailed as such, which gets harder to do when you cut it into a gazillion pieces.

  11. hobbitrock | | #11

    Got it. Thanks for both replies. Back to the rigid foam as a thermal break that you suggested Dana....This seems to be useful in multiple scenarios and it wouldn't need to be thick foam if the R value was already above the rafters. 1/2" or 1" between the drywall and framing seems like it would work. One would not need to cut it into many pieces though....it would be installed in full sheets with seams staggered in relation to the drywall. In that sense air sealing at this layer would be attainable as long as seams were taped etc. Correct? I'm just wondering which type of foam one would use. XPS? Like I said, I don't want the equivalent of plastic sheeting.

    Or......am I just better off installing airtight drywall directly to the framing and not worrying about a thermal break.

  12. Dana1 | | #12

    If you aren't going to reduce the vapor retardency of the interior side to at least Class-II levels it will be clear code violation, and susceptible to problems. You really DO want the smart vapor retarder, since it's easier to detail as an air barrier, and it offers faster drying rates most of the year, along with low class-II moisture loading rates in winter.

    To hit class-II vapor retardency with full sheets of 1.5lbs density XPS, it needs to be at least 1.5" thick. At 1" it's close, but not quite there. Read the specs of the particular product- most XPS is between 1.2-1.8 perms @ 1", and would be about 2/3 of that at 1.5". The class-II boundary is 1 perm. While 1.2 perms wouldn't be too risky, 1.8 perms is probably more vapor open than ideal. (Standard latex paint is 3-5 perms.)

    If you go with just air-tight drywall & dense-pack you can paint the drywall with "vapor barrier latex" as the primer coat, which should get you close. The IRC is silent about unvented roof assemblies with class-II vapor retarders, but demands higher exterior R to be able to use only standard latex paint (or other class-III vapor retarders) as the interior side vapor retarder.

  13. hobbitrock | | #13

    Thanks for that. I will look into the Intello products. Just for clarification, when I used the term latex paint, I was being too general. The plan would be to use something along the lines of Ben Moore Super Spec for the primer coat.

  14. Expert Member
    Dana Dorsett | | #14

    In Re: Benjamin Moore Super Spec: "Perm rating of 1.7 when tested under ASTM D1653"

    That's still a Class-III vapor retarder, but on the vapor-tight end of Class III.

    You can buy Intello Plus via internet or phone from this place:

    http://foursevenfive.com/product/intello-plus/

    I'm not sure where else to get it N. America.

    Certainteed MemBrain is less expensive, but sometimes hard to find on local shelves. The midwestern retailer Menards carries it- you can order online and have it shipped:

    https://www.menards.com/main/paint/drop-cloths-plastic-sheeting/poly-film/8-x-100-membrain-smart-vapor-retarder-air-barrier-film/p-1902435-c-8188.htm

    Both are pretty good products.

Log in or create an account to post an answer.

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |