GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Rigid foam under rafters!

nj_homeowner | Posted in General Questions on

Hi everyone –

I have Martin’s Musings book, and have also read through many wonderful articles on this site.  Still, I am not sure whether our plan for insulating the attic in our new house is a good one.

The house is in northern NJ, climate zone 4.  Our attic will be finished, conditioned space, with an insulated cathedral ceiling under a vented, 12/12 pitched roof.  Rafters will be 12″ thick.  We will have a dormer and a couple of skylights.  There will be 3 foot knee walls on each side, but we are planning to frame them after the entire roof is insulated, ridge to soffit.  Code says insulation should be about R49.

Can we insulate successfully with batting in the rafters and rigid foam underneath?

What type of rigid foam would be best?

Should we do two or more layers and tape the seams?

Will this be a good, durable, efficient roof system?

I have studied the articles and guidance here, but can’t find much information about this approach.  We want to avoid spray foam, and it’s too expensive for us, anyway.  Putting rigid on top of the roof, as you show in various drawings here, sounds great but my contractor, architect, and everyone else I speak with says doing so is both unfamiliar and cost prohibitive.

We would like our house to be efficient, comfortable, and reliable but also affordable and buildable by our contractor.

Thank you in advance,


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.


  1. GBA Editor
    Martin Holladay | | #1

    If you have dormers, you need an unvented insulated roof assembly. The dormers prevent a vented approach.

    Q. "Can we insulate successfully with batting in the rafters and rigid foam underneath?"

    A. No. The two possible approaches involve either:

    (1) an adequately thick layer of rigid foam insulation installed on the exterior side of the roof sheathing (with or without supplemental fluffy insulation between the rafters); or

    (2) an adequately thick layer of closed-cell spray polyurethane foam installed directly against the underside of the roof sheathing (with or without supplemental fluffy insulation on the interior side of the cured spray foam).

    More information here: "How to Build an Insulated Cathedral Ceiling."

  2. Expert Member
    Akos | | #2


    I've installed 2.5" polyiso bellow a cathedral ceiling. Its doable but a bit fussy, any ceiling penetrations are a real pain, I would stay away from those. I put the vapour/air barrier between the foam and the rafters (this was taped and caulked), didn't do anything with the foam seams. I put the drywall directly over the polyiso with 4.5" long drywall screws (available from commercial drywall places). Make sure to mark rafter locations otherwise they will be hard to hit with 4.5" screws.

    Depending on how the dormer if framed, venting dormers is doable in most locations, except the bottom of the roof section bellow where the dormer starts. Sprayfoam there is your friend.

  3. Expert Member
    Michael Maines | | #3

    Brian, will your dormer be a gable (aka doghouse) style, or a shed style? You can still vent a shed dormer, if you plan for it.

    A house I designed with a simple gable roof is just about done, with vented rafters filled with cellulose, and 2" of polyiso insulation below the rafters. I try to avoid using foam because it has higher embodied energy per R-value (meaning it contributes more to global warming) than other insulation, but in this particular case it made the most sense. The builder said it was easy enough to install.

  4. GBA Editor
    Martin Holladay | | #4

    If you take Akos's advice, remember that his or her suggested approach will only work in rafter bays that can be vented, top and bottom.

  5. nj_homeowner | | #5

    Thank you everyone for the thoughts.

    Our one dormer is a shed dormer. Also the house is shaped like an L, so there are actually two gabled roofs that meet, with a hip and a valley. Those features plus the skylights are the fanciness.

    Martin - I have carefully read, and re-read, what I think are all of the relevant articles here, and indeed I did not realize that your view is so unequivocal.

    I'm having trouble understanding why our plan cannot be adapted to work well.

    There are apparently smart people here, including Akos and Maines above, who seem to be saying our plan in fine.

    Moreover, our architect, in practice for 30+ years, says our roof is not particularly challenging, and he has no concerns about doing it as proposed. He did say he would do some extra treatments, such as drilling holes in certain critical locations, to enhance performance.

    Our builder also says he has no concerns about doing a vented assembly as described. He's been building for 19 years.

    So I'm having trouble reconciling your conclusion with those of the professionals on our project, and those of other folks we are hearing from. You feel they are mistaken? How will our roof fail if constructed as proposed?

    Can you please share any further explanation, or hopefully guidance about how to successfully build a roof as we are contemplating?

    Are you suggesting in your last comment that perhaps our plan will be fine if we do special things for the rafter bays that are interrupted?

    Akos - What penetrations are you referring to? What kind of vapor/air barrier do you recommend? Our plan does sound similar to yours, so I'm glad to hear it worked for you.

    Thank you again,


    1. Expert Member
      Akos | | #7


      Here we are required to have a 6 mil plastic vapor barrier by code. Since it was for my home, I just went through and taped and caulked all lapped joints and staples to seal it up. Airtight drywall or taped foam seams work, whichever your contractor prefers.

      I have exposed collar ties and a couple of electrical boxes for lights that also had to go through this assembly. I tried my best to seal those up, but they still leak more then I like.

      Also watch for venting the bays bellow and above your skylight.

      Good luck,

      P.S. Martin is correct. The best performing and robust roof design is one with exterior foam.

    2. Expert Member
      Michael Maines | | #8

      Brian, your additional details change things. A traditional approach at valleys, hips and other odd spots is to drill holes to allow a little airflow, but the building code requires at least 1" of clear ventilation space from eave to ridge, as Martin noted, and building scientists like Dr. Joe Lstiburek recommend more. (I typically spec 1 1/2" because it's available in a product from Accuvent.) Buildings can be forgiving of less-than-perfect details, but once you move beyond a simple gable roof, it gets progressively more challenging to do things right. Few builders revisit old projects, and it often takes 10 years or more for problems to show up, so the fact that a builder has done something a certain way for a long time is never a good argument if it goes against industry-accepted best practice.

    3. GBA Editor
      Martin Holladay | | #11

      The problem with roof valleys is that it is quite challenging to vent the rafter bays that "die" in the valleys. A similar problem happens with hips and the rafter bays that die under skylights and dormers.

      The only robust solution, in my experience, is to "double strap" the roof with 2x4s. After the framing is complete, but before the roof sheathing is installed, the builders install two layers of flatways 2x4s above the rafters. These 2x4s are installed either 24 inches on center or 16 inches on center. The first layer is installed 90 degrees to the rafters -- this layer of strapping allows air to move laterally at the tops of "dead" rafter bays. The second layer is installed 90 degrees to the first layer -- that is, above each rafter. Then the sheathing is installed.

      Finally, ventilation baffles are installed in each rafter bay, from below.

      Note that an engineer needs to approve this approach, because the sheathing is no longer fastened directly to the rafters or trusses.

      1. nj_homeowner | | #15

        Thank you for this information, Martin. I am interested and will suggest to the architect.

        Do you have a sense for the added cost of building this way, and getting approval from an engineer? Would you imagine that it's much less than doing rigid on top would be for us?

        It would be so great to have a robust approach that is affordable and acceptable to our architect and builder!!

        Thank you,


        1. GBA Editor
          Martin Holladay | | #16

          Costs are intensely local, so I'm not going to speculate. You need to talk to your contractor for an estimate.

          In many parts of the country, the least expensive solution is to buy some reclaimed (used) rigid foam on Craigslist, and install it above the roof sheathing. For more information on this approach, see "How to Install Rigid Foam On Top of Roof Sheathing."

      2. nj_homeowner | | #20

        Hi Martin -

        Do you know how the double-strapping affects the skylight installation or final appearance of the skylights?

        Thank you,


        1. GBA Editor
          Martin Holladay | | #21

          If you double strap your roof, and then install plywood or OSB roof sheathing about the double strapping, your skylight installation will proceed normally. The flashing kit deals with the upper layers -- the sheathing, the roofing underlayment, and the roofing -- and not anything under the sheating.

          The appearance of the skylight will not be affected, but the skylight well will be 3 inches longer than it would be without the strapping.

          1. nj_homeowner | | #22

            Thank you. Just asked my builder, and he is concerned that by inserting the strapping between the roof rafters and sheathing, we will lose a lot of strength... Do you think that is a valid concern?

          2. GBA Editor
            Martin Holladay | | #23

            There's no doubt in my mind that you can double-strap a roof, and that the roof will be adequately "strong" (depending on what you mean by "strong"). But you have to plan the fasteners carefully, and you should heed the advice I provided in Comment #11: "An engineer needs to approve this approach, because the sheathing is no longer fastened directly to the rafters or trusses." So consult an engineer.

  6. Jon R | | #6

    It's not clear to me that you can't ventilate any rafter bay, top and bottom, using shingle vents.

    1. Expert Member
      Michael Maines | | #9

      Jon, what do you mean by shingle vents? I only know that as a ridge vent product made by Airvent. Airvent also makes the EdgeVent, which allows intake at any portion of the roof ( that what you're thinking of?

      1. Jon R | | #10

        Yes, like the latter (which is called shingle vent by other manufacturers). Or DCi SmartVent. Maybe Cobra IntakePro.

        Also see DCi ValleyVent, which allows air to flow between stud bays (evidently without structural effect).

  7. Expert Member
    Peter Engle | | #12

    We built a kitchen addition 25 years ago with 8" fiberglass in vented cathedral ceiling cavities, full-length 1" baffles above the FG and 1" foil-faced polyiso on the interior, taped at the seams. There are two skylights in the roof, but no other interruptions. This is in coastal NJ, also CZ4, but significantly warmer than inland. We notched the tops of the rafters above and below the skylights to allow for some lateral ventilation. I now know that this probably does very little to help with moisture removal.

    We added another room over part of the kitchen a few years ago and that gave us opportunity to strip and inspect the whole roof. Everything was fine - no discoloration, no rot, no signs of condensation except at the 4 recessed lights that seemed like a good idea at the time. The house runs pretty dry in winter (RH 20%-30%) and the kitchen is kept somewhat cool (about 65 degrees). So for us, it has worked.

    For contrast, I was in a 15 year old house last week that had 12" I-joist rafters, 10" of fiberglass in the cavities. No baffles, but natural airspace above the batts that ranges from 1/2"-2". No VB on the interior, only the kraft facing, and no interior insulation. The house has hip roofs that are vented at the soffits, and with a ridge vent installed on the hips. The ridge vent was installed 5 years ago when the house was reroofed because of rotted sheathing. The sheathing is once again soaking wet and rotted, despite the vents on the hips and there is now structural damage to the flanges of the I-joists. It seems that the added ventilation is just pulling more moisture from the house into the cavities because of lack of attention to air sealing details between the house and kneespaces.

    So Brian, my answer would be that your approach could work if you use foil-faced foam and you are scrupulous about taping the seams and other important air-sealing details. Don't have ANY penetrations in the foam layer. If you keep the moisture out of the cavities, there isn't any moisture to condense on the sheathing.

    However (here is the big caveat): This approach doesn't meet code. It will be subject to failure if you mess up even a little bit on the installation. It is not a forgiving roof system and it may be unreliable. It is very difficult to know whether or not it is working properly without invasive inspection. If it fails in 10 years, nobody is going to help you out with repair costs. Done properly, your roof structure should last forever, but there are countless examples of improperly constructed cathedral ceilings that haven't made it past 10 years - many far less.

    If you think the risk is worth whatever benefits you think you are gaining and your architect is willing to sign off on it, go right ahead. But the cost of doing it "right" is not much higher, and the risk is far lower.

    1. nj_homeowner | | #13

      Thank you very much for this insight, Peter. We are trying to be frugal, but I absolutely want to get this right, and we are willing to pay some more to get it right.

      The biggest problem I am having is that everyone is totally unfamiliar with the rigid-on-top configuration -- I don't know that I can even get a price on it.

      Do you know someone in NJ who is experienced with this kind of roof insulation, who may be interested in giving us a proposal?

      Doing the job with spray foam will cost perhaps $10k more than our current plan, which seems like a lot, and I really want to avoid spray foam in any case.

      Also, how does this not meet code?


      1. GBA Editor
        Martin Holladay | | #14

        According to the building code, you can't install fluffy insulation in an unvented rafter bay unless either (a) you have installed a sufficiently thick layer of rigid foam above the roof sheathing, or (b) you have installed a sufficiently thick layer spray foam against the underside of the roof sheathing.

        This is explained in my article, "How to Build an Insulated Cathedral Ceiling."

        You might want to re-read Comment #11, which is directly below Comment #8. In that comment, I described a way to vent your type of roof (one that includes valleys, dormers, and skylights). If you can figure out a way to vent every rafter bay, you can install the fluffy insulation between the rafters.

    2. Jon R | | #17

      > If you keep the moisture out of the cavities, there isn't any moisture to condense on the sheathing.

      +1 on this. While code lists 1" gap and 300:1 venting, the reality is that the amount of drying needed is entirely dependent on the amount of wetting (primary due to air leakage).

  8. Expert Member
    Dana Dorsett | | #18

    >"Rafters will be 12″ thick."

    Not very likely- a milled 2x12 is 1.5" thick and 11.25" deep.

    >"Code says insulation should be about R49."

    Code also says compliance can be achieved by a U-factor of U0.026 or less.

    U0.026 is R38.5 "whole-assembly R", accounting for the thermal bridging of the framing, the R values of the roof decking, roofing, interior finish ceiling, interior & exterior air films, air films inside of any unfilled framing cavity, etc. In most houses that can be achieved with an empty rafter bay with interior side gypsum board and 6" of 2lb roofing polyiso (R34-ish) , a standard structural roof deck, and a nailer deck atop the polyiso onto which to mount a standard asphalt shingle layup.

    At 6" it'll take 7.5" pancake head timbers screws to install the nailer deck and a ~1x8 facia board at the roof edge, which some contractors (and some architects) might find objectionable.

    An easier to build path to compliance on an R-value basis would be to split the R between the above-deck foam and install some rock wool in the rafter bays. When going that route one has to be mindful that there be sufficient exterior R for dew point control at the roof deck. In the US climate zone 5 parts of northern NJ that would be R20 minimum on the exterior for up to R30 in the rafter bays. That could be done cheaply with 4" of reclaimed roofing polyiso (or 5" of reclaimed EPS) up top and R30 rock wool batts in 2x8 rafters ( or deeper). In the zone 4 locations it can be done with 3" of reclaimed polyiso (or 4" of reclaimed EPS) and a low density R38 fiberglass batt compressed to 9.25" in a 2x10 rafter bay.

    With 3"=4" foam the edges can even be finished with a 3" drip-edge rather than a facia board, but even if it's a facia board it's not as prominent as a 1x8.

    Sources for used roofing foam aren't hard to find in your area:

    1. Deleted | | #19


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |