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

Recommendations for insulating existing space over garage with a cathedral ceiling

Andrew K | Posted in Energy Efficiency and Durability on

I live in Massachusetts (Zone 5A) and have some space over the garage that I would like to finish as an office. The space has a sloped roof with 2×8 rafters, with soffit vents and a ridge vent (but no existing insulation). I’ll be adding some 5 foot kneewalls and a ceiling, leaving about 4-5 feet of cathedral ceiling between the two.

What are my options for insulating the space? I have looked at closed cell insulation but quotes have come in at about $5-6K to insulate the entire roof deck (no venting). So I’m looking for cheaper alternatives.

I am thinking this (outside in):

Roof deck | 1″ vent ” | Fiberglass Batt | 2″ Rigid foam insulation | Drywall

Where should the vapor barrier go? Should I use faced batts or something else (either between the fiberglass and rigid foam or on top of rigid foam?)

Which rigid foam would would I use? XPS or polyiso?


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

    Here are my comments:

    1. A 2-inch-deep ventilation channel is better than a 1-inch-deep ventilation channel.

    2. Dense-packed cellulose insulation will perform better than fiberglass batts.

    3. Polyiso is more environmentally friendly than XPS.

    4. If you install 1x3 or 1x4 strapping (furring strips) on the underside of the polyiso, you'll have an easier time installing the drywall. And if you choose foil-faced polyiso, the air space will provide additional R-value to the roof assembly.

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

  2. Expert Member
    Dana Dorsett | | #2

    What he said. Only to add:

    Code requires a minimum of 1.5" for vent channels, but on lower angle roofs even that isn't enough.

    Standard rock wool batts or high density "cathedral ceiling" batts can work if installed to near perfection. Dense-packing cellulose in a vented assembly is damned-near impossible to achieve without adding something both rigid & vapor permeable to maintain the the vent channel without it collapsing due to the pressure while dense packing.

    A flash-inch of closed cell foam against the roof deck is sufficiently vapor retardent to protect the roof deck, which would allow you to safely dense-pack the rest, and go un-vented. This does not meet the letter of code, but there is ample evidence that it works. Presenting this document to the inspectors would likely allow you a waiver:

    If you go unvented you can't use foil facers on any interior-side foam. Fiberglass faced roofing iso would work. EPS would be fine too. XPS has a heavy greenhouse footprint (heavier per inch than closed cell foam) but that too would cut it. If XPS, 2" is the absolute max, or it becomes too vapor retardent.

  3. Andrew K | | #3

    Thanks guys.

    If I go with a R23 5.5 inch rock wool batt, would I need to put baffles behind it (between roof deck and batt) or are the batts rigid enough not to fill the vent space?

  4. GBA Editor
    Martin Holladay | | #4

    Rock wool insulation is air-permeable, so you need an air barrier on the top side.

    The air barrier separates the warm air held between the insulation fibers from the cold air passing through your ventilation gap. Your ventilation baffles need to be installed in an airtight manner, because these baffles are your top-side air barrier.

  5. Andrew K | | #5

    Ah, ok.

    Could I build these out of some strips of wood and 1" XPS from the inside? I'm guessing that the accuvent baffles I can buy at local store won't cut it?

  6. GBA Editor
    Martin Holladay | | #6

    Your options are listed in the article I linked to. Many builders are satisfied with AccuVent baffles. It's also quite possible to make your own baffles, using plywood, OSB, cardboard, or rigid foam. Your suggested material -- 1-inch XPS -- probably has a high enough R-value and a high enough permeance to keep you out of trouble.

  7. Andrew K | | #7

    Thanks for your clarifications and patience.

    If I was to go with R-38 closed cell instead (no venting), would I still have to worry about ice dams in the winter?

  8. GBA Editor
    Martin Holladay | | #8

    If the snow on your roof is very deep, it is always possible to have some ice dams, especially if your roof is unvented. However, those types of ice dams only occur in high-altitude locations or locations near ski areas, as long as your roof has no air leaks and the R-value of the roof insulation is adequate.

    In Masschusetts, I think you would be fine with spray foam insulation totaling R-38, even with an unvented assembly. Of course, R-38 of spray foam is expensive.

    For more information on ice dams, see Prevent Ice Dams With Air Sealing and Insulation.

  9. Andrew K | | #9

    Thanks again.

    One final question. I read in one of your articles the different ways of insulating the cathedral ceiling / kneewall (one way is to do sloped ceiling only vs. part of floor / kneewall / sloped ceiling).

    I will be insulating the entire sloped ceiling up to the small attic portion. Is there any harm or benefit to also insulating the kneewalls with unfaced insulation?

  10. GBA Editor
    Martin Holladay | | #10

    Your question is somewhat unclear. If you are insulating "the entire sloped ceiling," I hope that means that you will be extending the insulation all the way to the floor (near the eaves). If you do it that way, the cramped attic behind the kneewall will be within the home's conditioned space.

    If that's what you are doing, there is no need to insulate the kneewall. But if you want to insulate the kneewall, there is no harm in doing so.

  11. Peter L | | #11

    One way to get a better cathedral ceiling design is to use SIPs. They provide excellent air sealing (as long as you pay attention to the panel joints) and they provide the structural and insulation, all in one panel. They also eliminate any thermal bridging, unlike standard cathedral ceilings with bridge via the joists. One can get R-50 panels and if you want to go higher on the R-Value, you can always use furring hat channels/Z-channels on the interior and then install rigid foam board within the furring strips and then drywall. That should get you to around R-60 +/-. Preferably Steel SIPs are better in rainy/wet climates than OSB SIPs. One doesn't have to create a cold roof with Steel SIPs but it is advisable to do so with the OSB SIPs to prevent "SIP rot." Also an EPS foam core is easier on the planet than XPS foam.

  12. Expert Member
    Dana Dorsett | | #12

    Martin wrote back in response #4:

    "Rock wool insulation is air-permeable, so you need an air barrier on the top side."

    That''s only sort-of-true. It's quite air permeable from a moisture transport point of view, but (unlike low density fiberglass) rock wool batting is sufficiently air-retardent to not lose performance to convection, and will pretty much perform at it's specified R value as long as there is an air barrier on at least ONE side of the assembly. The same is true for snugly fit high-density "cathedral ceiling" fiberglass. While it doesn't hurt to have a top-side air-barrier, it's less critical here than in other assemblies. Where it's simple to add a top side air barrier, go for it, but don't sweat it where it's difficult/awkward/impossible. If it's wide-open, using strips or chunks of 2" foam as spacers to the roof deck and 1-3" EPS rough-cut to fit, sealed at the edges with can-foam works pretty well. EPS is preferred, since it's 3-5x as vapor-permeable as XPS. Since it''s a cut'n'cobble job, there's no point to using virgin-stock either. There are multiple sources of reclaimed and factory seconds foam in MA (search your local craigslist materials section for "rigid insulation", or call up the Insulation Depot in Framingham. Using reclaimed goods the R/square foot is often cheaper than batts.

    A full cavity fill of closed cell foam would create a moisture trap for the roof deck, between fairly impermeable roofing, and barely permeable foam. It's not only expensive, it's a potential problem.

  13. Andrew K | | #13

    Thanks again for the info.

    Somewhat related question about insulation in the floor in this room. The room is above a garage (not heated), floor has 2x8 joists which currently have existing R-19 insulation (faced, with paper on side of room).

    I was thinking of having it dense packed with cellulose instead, would I need to keep a vapor barrier?

  14. GBA Editor
    Martin Holladay | | #14

    Q. "Would I need to keep a vapor barrier?"

    A. First of all, plywood subflooring is already a vapor retarder. Second, airtightness matters more than the presence or absence of an interior vapor retarder. For more information, see:

    Do I Need a Vapor Retarder?

    Vapor Retarders and Vapor Barriers

  15. Dirk Gently | | #15

    Martin, I realize a 2" vent is better than a 1" but, Why does the Accuvent Catherdral vents only allow for a 1" space. Accuvent is referenced in your fine link to building a cathedral ceiling but seems like the emphasis should a 2" air gap custom built chutes from eps/xps in this article.

    Note here: Regular accuvents allow for 1.5" of air gap. The cathedral ceiling version allows for only 1".
    Perhaps a solution here would be to modify the regular accuvents to be used in cathedral ceiling application? If this is true...then could the regular accuvents handle 1.8 lb -3.5.b dense pack (should that be an option for Andrew)

  16. Dirk Gently | | #16

    Andrew, In regards to your insulating the knee walls and cathedral ceiling. I have done that on my existing house. In 04 I remodeled upper level of my cape and try as I might to seal every crack with obsessively fit batts, caulking, and ridgid foam.....I was not 100% happy with results. In 2010 I had the underside of the rafters sprayed with open cell to turn those tiny attic areas into conditioned space. I was unsure if I needed to tear out the knee wall insulation so left it in place.
    I can tell you there is a temperature difference of several degrees between the occupied portions rooms and the attic areas. I have no idea if that is a problem or not....but in my mind it helps cut on my already cheap heating costs. No idea if that helps but it is a real world scenario anyhow.

  17. GBA Editor
    Martin Holladay | | #17

    Designers of buildings and designers of building materials compromise all the time. The deeper a ventilation channel, the better it works. However, there is a limited depth to rafters -- and when you increase the depth of a ventilation channel, you usually end up decreasing the depth of the cathedral ceiling insulation. So designers compromise.

    An Accuvent designed for cathedral ceilings provides a 1-inch-deep air space. That means that the ventilation channel won't be as effective as a 2-inch-deep channel -- but you'll have a little bit more room for insulation.

    If you are building your own site-built baffles, you can choose any depth you want (as long as your local building inspector approves of your approach).

    Usually, rafters aren't deep enough for adequate insulation, no matter what depth you choose for your ventilation channel, so you need to scab on some type of framing with plywood gussets or strapping to make the rafters deeper. As long as you are doing that anyway, you might as well plan for a 2-inch-deep ventilation channel.

  18. Dirk Gently | | #18

    I am doing nearly the same exact project of mostly 2x8 rafters and dormers above a (to be heated) garage. rather than start a new Q and A is it ok to piggy back on this one? If not I will try to delete this post.
    I am having the same difficulty choosing a system as most on this site. I got a permit before code was updated in zone 5 in MA so I am only held to R-30.....but would like to do much better.
    I have 2 low slope shed dormers on the roof. 1 dormer is 2x10 so I could go with 2" gap. other is 2x8 so to get R-value was planning on 1" gap. Unfortunately these ceilings were built to match existing home and have a low head room inside. There is about 900 square feet of ceiling area to insulate with roof piches ranging from 3/12 to 12/12
    My first choice was to do 1" gap with dense pack Cellulose net and blow. 2" felt faced Polyiso (reclaimed) attached to rafters detailed to be an air barrier and to reduce thermal bridging. held in place with 1x3 strapping. This would go from soffit to ridge.
    In this scenerio is there a problem with the screws thru the foam causing air leaks? Also does anyone know what type of tape will adhere best to felt faced reclaimed Polyiso?

    I got a price for 7" open cell SPF which I thought reasonable but my wife did not. Also mother nature did not like this approach as well. 2" rigid insulation would be also added to underside of rafters.

    The third option is go with unvented cut n cobble system of reclaimed insulation. Not sure which type of insulation to use here EPS or Polyiso. There would still be 2" of rigid foam attached to underside of rafters here......again not sure if polyiso or EPS is better (Perm concerns)

    Now I have the added confusion of Spider or Optima dense pack instead of cellulose. I have concerns that if it is not packed to at least 1.8 Lb density then it will not have air ceiling properties.

    Any help in rigid board type is much appreciated. Again my concern is getting the Perm rating correct.

  19. GBA Editor
    Martin Holladay | | #19

    Have you read the article below? It may answer many of your questions.
    How to Build an Insulated Cathedral Ceiling.

    Q. "In this scenerio is there a problem with the screws thru the foam causing air leaks?"

    A. No.

    Q. "Does anyone know what type of tape will adhere best to felt faced reclaimed polyiso?"

    A. I would advise you to contact one of the customer service representatives at Small Planet Workshop or 475 Building Materials. These two distributors sell high quality European tapes.

    Q. "Any help in rigid board type is much appreciated."

    A. Most green builders consider polyisocyanurate to be the most environmentally friendly type of rigid foam. Of course, recycled foam is better for the environment than new foam.

  20. Expert Member
    Dana Dorsett | | #20

    You can't really tape felt-facers- air seal the seams with fiber reinforced duct-mastic and hope for the best. Reclaimed roofing foam will likely have less overall shrinkage & dimension change issues than virgin-stock. If a little bit of micro air-leak ends up developing over time, that's what the vent space is for.

  21. Matt Zahorik | | #21

    Martin & Dana,

    This is a followup to #2 and #4

    I too have a partial cathedral ceiling to insulate, with an existing ridge vented roof, and I'm struggling on what material to use to air seal on top of the batts.

    In my case, with 2x10 rafters, the layers are:

    plywood roof plane (soffit and ridge vents)
    nominally 2", minimally 1.5" vent space
    some sort of air barrier here...
    7.25" rock wool batts (R-30)
    two layers 2" foil faced polyiso, 4" total (+R-25), staggered seams, taped, caulked/foamed.
    potential strapping

    I'm in SE New Hampshire near the Maine border, so zone 5, not that far from zone 6.

    Martin said:

    "Many builders are satisfied with AccuVent baffles."

    Do you have a concern in this setup that the AccuVent represents a vapor barrier, thus having a vapor barrier above and below the batts, giving the batts nowhere to dry should moisture be driven in? Would something like Tyvek on top of all or part of the batts but below the venting space act as an air-barrier but allow vapor permeability?


    - Matt

  22. GBA Editor
    Martin Holladay | | #22

    If you don't like AccuVents because they aren't vapor-permeable, you can make your own site-built ventilation baffles. Many builders have done it. You can use plywood, fiberboard, or cardboard if you want (all of which, in my opinion, are better than Tyvek for this purpose).

    I still think AccuVents can work fine, and here's why: moisture problems in cathedral ceiling bays are almost always due to ceiling air leaks, not vapor diffusion. If you can make your ceiling airtight (or close to airtight), you're not going to have moisture problems.

    The small amount of moisture that enters cathedral ceiling bays via vapor diffusion is absorbed by the rafters. The rafters distribute the moisture, which evaporates from the tops of the rafters into the ventilation gap at the top of the rafters.

  23. Matt Zahorik | | #23


    Thanks - that's very helpful.

    My thinking, which was poorly stated and thought all the way through, was if the assembly got wet - "if moisture was driven in" - not necessarily from vapor diffusion. I like to build in some resiliency against bulk water intrusion, such as a roof leak that went unnoticed, rain driven into the ridge vents, etc. so that such failures don't snowball into a major repair of the assembly.

    Upon actually thinking it all the way through, if you properly lap the accuvent so water can flow down and out, and you're 100% careful about sealing all accuvent edges and seams against air (and therefore water), the assembly should withstand bulk water short of full failure of the roof. If the caulk/foam loses elasticity and starts to separate after many years, the assembly drying will be sub-optimal but if the water is infrequent it still should dry through the method you describe.

    - Matt

  24. Dirk Gently | | #24

    Thanks for asking the accuvent question Matt.
    Please post if u decide to use them. I too struggle with the same decision.

  25. Expert Member
    Dana Dorsett | | #25

    The vapor permeance of framing wood species is about 1 perm @ 1", similar to 1" XPS or 1" closed cell polyurethane.

    But with the Accuvent your drying area is much reduced- the framing is only ~15% of the surface area in a roof assembly. I'm not totally convinced the extremely low permeance of Accuvent won't be a problem if you have an actual air leak into the insulation from the interior side. It's not a simple thing to model, but it could be tested. With more permeable baffles the assembly becomes more resilient.

  26. DrFlayNuby DrFlayNuby | | #26


  27. Deshawnlek Deshawnlek | | #27


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |