GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Q&A Spotlight

How to Insulate a Cathedral Ceiling with Mineral Wool

If the rafter bays of a sloped ceiling are packed with insulation, is venting the roof necessary?

One approach to venting the roof:

Installing sheets of rigid foam insulation slightly below the tops of the rafters and sealing the edges with foam is one way of providing an air channel for ventilation.
Image Credit: M S

John Roy is building a house in southeastern Massachusetts, and at least part of it will have a cathedral ceiling. He’s thinking of insulating the ceiling with dense-packed rock wool.

The president of a local insulation company tells him there’s no need to install air chutes in the rafter bays before the insulation is blown in because the insulation does not absorb water. The local building inspector is prepared to go along with the recommendation providing soffit vents are installed.

Is Roy getting good advice? That’s what he’s asking in this Q&A post, the subject of this week’s Q&A Spotlight.

First, will that be enough insulation?

Whether unvented cathedral ceilings should be dense-packed at all is a question that’s generated plenty of discussion in the past. Leaving that aside for the moment, a key question for Roy is whether mineral wool packed into a 2×10 rafter bay will provide enough thermal insulation for his locale, and the answer is no, it won’t.

As GBA senior editor Martin Holladay points out, a 2×10 is really 9¼ in. deep. Using the values published by the National Association of Home Builders Research Center, mineral wool at a density of 2½ pounds per cubic foot would have an R-value of about 34 if it completely filled the rafter bay.

The 2006 International Residential Code requires ceilings in Roy’s area (Climate Zone 5) to be insulated to at least R-38, Holladay says.

If part of each rafter bay is dedicated to an air chute, reducing the amount of insulation that can be added, the numbers get worse. This means that whether Roy adds air chutes or not, he’ll need to find a way to add more insulation if he wants to meet IRC requirements.

Debating the merits of ventilation

Despite assurances from the insulation contractor, building codes won’t allow the kind of roof assembly Roy is contemplating.

“According to my understanding of the International building codes, you can’t install an air-permeable insulation like mineral wool in a cathedral ceiling unless you first install ventilation chutes on the underside of the roof sheathing,” Holladay says. “Styrofoam ProperVents are too flimsy to withstand the pressure of dense-packed insulation, so you’ll probably want to install plastic AccuVent or site-built ventilation chutes.

“It doesn’t make any sense to install soffit vents if you don’t have ventilation chutes between the top of the insulation and the bottom of the roof sheathing,” he adds. “If it were my house, I would install ventilation chutes (well sealed to prevent air leakage). I would also install 2 inches of foil-faced polyiso under the rafters, held in place with 1×4 strapping, and I would blow the mineral wool through holes in the polyiso.”

A poster named M S directs Roy to an article by building scientist Joseph Lstiburek of the Building Science Corp., who says flatly: “As fantastic as the dense-pack approach is for walls, it is a pretty dumb approach for unvented cathedral ceilings and flat roofs. And, no matter how hard some of us work to try to stop people from going ahead and doing it, they persist on doing it. First, it is a building code violation. Second, it violates the physics.”

It would seem Roy should abandon ship on whatever plans he had for dense-packing an unvented cathedral ceiling. So what’s the best way of adding ventilation?

Making airtight ventilation chutes

The most effective ventilation chutes will be site-built, Holladay says. “Usually, the builder starts by tacking up 1″x1″ sticks in the upper corners of the rafter bays,” he writes. “Then rigid material (Celotex, thin plywood, or rigid foam) is installed against the sticks, with caulk or spray foam to make the installation airtight.”

While providing an excellent air seal, these chutes are not the fastest to install, Holladay says.

M S says chutes might be made from sheets of polyisocyanurate foam insulation, sealed at the seams with spray foam. The foam would simultaneously serve as baffle and insulation. By dense-packing the remainder of the rafter bays with cellulose, at R-4 per inch, the roof would be rated at R-39.

“Now, as Martin suggests, overlay the face of your studs with rigid insulation, either polyiso or XPS,” M S says. “Putting up another 2 in. of polyiso gets you close to R-52.”

What about a vapor-permeable chute?

If the ventilation chute carries away any excess moisture from the cavity insulation, Thomas Jefferson says, it would seem “wise” to make the baffle with a vapor-permeable material.

“Foil-faced polyiso does add R-value but it’s not permeable,” Jefferson says. “I’d suggest hardboard (Masonite) or another fiber panel to just hold back the dense pack without impeding water vapor. The dense-pack will be only ~R-31, so you’ll need to furr out below the joists or add polyiso below, as Martin suggested.”

Holladay disputed Jefferson’s opinion that “this vent chute serves partly to carry out any excess moisture from the cavity insulation.” According to Holladay, “No one wants interior moisture migrating through the insulation to the ventilation cavity — certainly not deliberately,” he says. The idea is to prevent any moisture from getting into the ventilation cavity and to provide a mechanism that helps keep the roof sheathing dry.

“The purpose of the ventilation channel is to lower the humidity of the roof sheathing — in case some sloppy carpenter did a bad job of caulking the ventilation chutes in place, or in case some electrician mistakenly installs wiring or a can light in your ceiling,” Holladay says.

Isn’t the purpose of a ventilation chute to prevent the accumulation of moisture? asks Jefferson. “Moisture can come from a few different sources, and one could be below the chute,” he says. “Why design a vent chute to intentionally exclude water vapor that just might find itself in the cavity insulation? Not that it’s intentional for the insulation to be damp and need drying to the exterior, but the opportunity of drying reduces risk.”

Not really, Holladay says: “The purpose of the ventilation channel is as a last-ditch method of keeping your sheathing dry in case your assembly has flaws. But you don’t want to build in any paths that allow or encourage vapor to flow toward your cold sheathing.”

Spray-in foam works, too

Given the difficulties of building in an effective ventilation chute, Roy asks, what about the possibility of using closed-cell foam directly on the underside of the roof sheathing?

Closed-cell foam provides an air barrier, but to control condensation it must be applied to a minimum R-value, depending on what climate zone you’re building in, Holladay says. That ranges from R-5 in zones 1-3 to R-35 in zone 8.

Our expert’s opinion

Peter Yost, GBA’s technical director, had this to say:

Roof assemblies, like wall assemblies, should be designed and insulated to dry in at least one direction. We tend to focus on how building assemblies get wet from interior moisture during the cold winter months, but more importantly, we have to focus on how they dry from any of a variety of moisture sources, not just wintertime interior relative humidity.

Roof assemblies are like wall assemblies that never get stood up straight. So, while they do see more bulk water, they can be designed and insulated without back-venting the roof sheathing (“hot roofs”), just as walls are. There are two critical elements, however: selecting materials for ALL layers that collectively create directional drying AND a continuous and contiguous (to the insulation) air barrier in the unvented roof assembly.

On the other hand, back-venting the roof sheathing is a more forgiving approach, if the soffit-to-ridge vent chutes are continuous and well-sealed with respect to the roof cavity insulation.

You can “decouple” the roof cladding and roof sheathing (in terms of wetting and drying) from the rest of the cathedral roof assembly with either a top-side air space or a vent chute underneath the roof sheathing. Each presents construction detail challenges, and costs. And frankly, the choice of roof cladding material and your climate can strongly suggest just how important some sort of venting is to the performance of the roof assembly. For examples, see this article from Building Science Corp. as well as this one.

For more information on vapor profiles and drying, take a look at this blog.


  1. TJ Elder | | #1

    The opportunity of drying does not necessarily promote wetting
    Where I differ with Martin here is that he seems to describe some intentionality to this water vapor that I suggest might find itself in the dense pack cavity. That is, the usage of a vapor permeable material for the vent baffle would somehow encourage water to soak into this cavity insulation, knowing it could escape into the vent chute. I don't think the permeability of the baffle would attract moisture into the cavity, and my argument is that it's possible there could be moisture in there due to air leakage, excessive indoor humidity, etc.

    Let's recall also that the joists are 2x10 and the cavity will lose an inch or so due to the vent chute, so the insulation depth isn't sufficient without dropping the ceiling. If there were polyiso installed below the joists (as suggested) this would create a vapor barrier, and it would make sense to allow the joist cavity to dissipate moisture toward the exterior.

  2. User avater
    Armando Cobo | | #2

    Based on my experience...
    “Roof assemblies are wall assemblies on an angle”, I can’t remember who said that first. Having said that, what I’ve done for many years is a material permeability and dew-point analysis on that roof assembly. In most cases in CZ4 –CZ5, having 9 ¼” space to insulate, I will use dense packed cellulose, BIB or open cell to R32 under the sheathing and ADA sheetrock so the assembly dries to the inside; and a good moisture barrier and 1”-2” of taped rigid foam over the sheathing.
    Again, you are going to need do a good job on the moisture management details and moisture barriers above and below the sheathing on any kind of roof assembly.

  3. Wayne Hutson | | #3

    insulate cathedral ceiling
    very interesting discussion. I invite any critiques of the building practice I use . Even though I am just a builder in Canada(near toronto, ontario) I will gladly offer how we would commonly handle building John Roy's roof .

    We apply 1X3(16" OC) strapping perpendicular on the outside of the rafters and the sheathing/ water shield/shingles is then attached to the strapping. This differs slightly from how Holladay suggested of tacking up 1"X1" in the upper rafter bays. we find the strapping exterior to the rafter bay quicker and obviously allows for a bit more insulation.

    We press 2" rigid foam between the rafters and up tight against the strapping which allows a 3/4" ventilation shute. We next
    spray a flash 1.5" of closed cell rigid foam(sprayed from inside) to seal the rigid insulation.
    Next a 2X6 mineral wool batt insulation. the 2X10 rafter bay is now full. Under the rafters would be another 1-4" of rigid foam for a thermal break but not a class 1 vapor retarder. No poly is applied in this detail . Next is strapping (1X3) to hold the rigid foam and create an air space and then final inside ceiling finish eg drywall.

  4. TJ Elder | | #4

    Foam Sandwich

    You use this assembly on a regular basis? It doesn't seem very user friendly or cost effective. A couple things I'd say are not ideal as you've described. First, the spray foam has an irregular surface that makes it hard to effectively install batts. Second, it's generally not a good idea to sandwich an open cell or fiber material between two closed cell materials. Third, the plywood should be in direct contact with the joists if there is a structural need for the diaphragm. Maybe where you are that's not a concern.

    I'd suggest it would be more cost effective to drop 2x3s below the joists using plywood gussets, basically a horizontal Larsen truss, then dense pack the ceiling. There are a few projects in the Green Homes section using that approach. That could also work in combination with your spray foam method.

  5. User avater GBA Editor
    Martin Holladay | | #5

    Response to Wayne Hutson
    You wrote, "We apply 1X3 (16" OC) strapping perpendicular on the outside of the rafters."

    Did you really mean "perpendicular"? Or did you perhaps mean parallel?

    If the 1x3s are perpendicular to the rafters, then your vent channels are from rake to rake instead of eave to ridge.

  6. Aj Builder, Upstate NY Zone 6a | | #6

    Wayne is describing nailing
    Wayne is describing nailing the 1x3 on top of the joists, parallel with the joists in length but in a T. So one piece of lumber becomes the 3/4" stop for the foam sheeting that is installed to make the vent area.

    He is killing two birds with one stone and is maintaining the diaphragm assembly Mr. President.

    Interesting assembly. I love hearing about actual assemblies alternatives other builders are trying.

    Wayne, tell us about your wall and foundation assemblies.

  7. TJ Elder | | #7

    AJ, you're right.
    That makes more sense. Running the 1x above and parallel to the joists would also allow venting above shear blocking, without needing to V-cut the blocking.

  8. Wayne Hutson | | #8

    Correction- Thanks Martin
    My apologies. The strapping that is perpendicular is holding up the rigid foam on the underside of the rafters . On the top side of the rafters the strapping is running parallel with the rafters forming a T like aj stated.
    The vent channels run eave to ridge.

    Thanks for the comments Thomas. I may have incorrectly assumed the unfaced expanded polystyrene or extruded polystyrene would still allow the cavity to dry since they are not considered class 1 vapour retarders.


  9. David Ely | | #9

    Nice system Wayne
    I wonder if the strapping could be stopped an inch or two from the valleys and hips so the decking cantileavers that little bit and allows more complete ventillation. Those locations are ordinarily the bane of vented roof systems.

    There might be a structural consideration unless the strapping is close enough to allow the decking to be nailed through it to the hip or valley rafter. Quite often valleys are doubled to carry the snow load which might provide that extra width for nailing. Has anybody done anything similar?

  10. Aj Builder, Upstate NY Zone 6a | | #10

    The huge problem with doubled
    The huge problem with doubled valleys as they are spec'ed now is the heat loss going thru all that low R material making valleys very prone to snow melt water intrusion.

    We need to put out plan sheets showing some kind of foam insulation used in the build up of such valleys. I do like the idea of outsulation in regard to a home with a large area of valleys as maybe the only solution at this time.

    Any ideas?

  11. Radu P | | #11

    Simmilar problem I encounter
    This is a 60 yro house in New England with cathedral ceilings and 5.5" rafters. Of course not enough insulation. inside the rafters is 60yro batt insulation that probably shrinked to 3". We are going to redo the roof this summer and we're thinking of pulling the plywood out and install closed cell spray foam insulation between the rafters from top. If we fill the cavity that will give us R36. The 100 point question is: do we need to keep a ventilation shute in between the sprayed foam and the plywood or not?

  12. User avater GBA Editor
    Martin Holladay | | #12

    Response to Radu
    It's basically impossible to spray closed-cell foam in such as way as to make to foam completely fill the rafter bays. (Open-cell foam can be oversprayed and then trimmed, but closed-cell foam is very hard to trim, so the installers don't fill cavities all the way.)

    You will end up with an air space between the top of the cured foam and your new plywood sheathing, whether you want it or not. For all intents and purposes, this airspace will act like a ventilation channel.

  13. Radu P | | #13

    Simmilar problem I encounter
    Thank you Martin. So I can do closed cell foam at least 4.5" in between the rafters and I have the ventilation present. That gives me R30 and if compared with R12 that I have now it is not bad.
    If I add a 2x3 shim on top of the existing rafters I can do 7" closed cell foam. and still get 1" of air left.
    I have to see what my energy savings numbers look like.
    I would like to talk with a structural engineer to see if my 2x3 shim add-on if glued and nailed in place is going to increase the strenght of the roof/ building. We got a lot of snow this winter and I was unconfortable with the 2x6 rafters...

  14. Thomas Gray | | #14

    existing cathedral
    Have read all of the above comments and still not clear of best approach- we have an existing ranch, 1960, 2 x 6 rafters on 4:12 pitch. Roof is in good shape, client intersted in fixing from below. I suggested remove drywall, add 2 x 3 to bottom of rafters, spray foam to existing deck or dens pack w/ cellulose after beadboard ceiling goes up. There is no ridge vent. Soffits are deep so we can get continuous vent there but pinch point above wall plate will be a problem. I have heard others justify dense pack of cathedral by likening to SIP panels which have been used for years w/ no vent chutes. TGray in Pittsburgh PA

  15. User avater GBA Editor
    Martin Holladay | | #15

    Response to Thomas Gray
    You write that you aren't clear about the best approach. That may be because there are many approaches and all of them can work.

    If you choose to dense-pack the rafter bays with cellulose, you MUST include a ventilation channel between the top of your insulation and the roof sheathing. That is a code requirement.

    If you have a beadboard ceiling -- I assume you are talking about tongue-and-groove boards -- you won't have an air barrier unless you first install gypsum wallboard. A board ceiling can create a disaster, so don't forget your air barrier.

    More information on insulating sloped roofs can be found here: Creating a Conditioned Attic.

  16. Walter Strobl | | #16

    Using 2 X 6 Rafters
    I find your discussions very interesting, since I have been pondering a similar question. I also have a 2x6 ceiling over an existing garage that has always been unfinished. I have always intended to finish it, but you know how it is other things get in the way. How can you achieve the insulation value necessary? Furring out the 2x6 causes another problem for me in that the entrance to the room is challenged by ceiling height. I cannot move the entrance since that would involve tearing out an existing bathroom. I may have room for just adding a 1 or 2 inches of rigid insulation under the rafters, but I would have to see. I always intended to just screw the rigid insulation to the rafters and then the drywall through the insulation into the rafters using very long drywall screws to reduce the thermal bridging. The house was built in the 90's and passed all of the codes then. But how do I meet the codes of today and pass the Ontario Building code? I need to provide a gap under the roof sheathing as well which further reduces my R value. So what is the best approach? Can I use get away with less than an R-38?

    Martin you mentioned that when using closed cell insulation that you would end up with an air space. Would this airspace not create a place for moisture to accumulate since it would end up being between the drywall and the insulation? Or would it dry to the inside since the closed cell is a moisture barrier even though it is not ventilated?

  17. User avater GBA Editor
    Martin Holladay | | #17

    Response to Walter Strobl
    If you don't have enough depth to provide code-minimum insulation, your best bet is to install closed-cell spray polyurethane foam, since that has the highest R-value per inch. You can also insulate your roof by installing rigid insulation on top of your roof sheathing. Of course that means you also have to install new roofing.

    If you install closed-cell spray foam in a way that leaves an air space between the cured foam and the interior drywall, you won't have a moisture problem. What moisture are you worried about? Exterior moisture won't diffuse through the closed cell foam, and interior moisture will never see a cold surface where it can condense. The air in the space behind the drywall will be at interior conditions.

  18. DKlemp | | #18

    This thread is only 8 years old, so here goes - can someone explain to me why strapping above the rafters to form the "tee" (Wayne, AJ) and then placing foam insulation inside the rafter bays is any different than just placing the insulation directly above the rafters (which is supposedly a no-no)? You would need purlins to stabilize the rafters, etc., but purely from a moisture-transport standpoint, what's the difference? Is it that moisture inside the roof cavity would travel through the wood rafter, around the insulation baffle and out of the roof assembly - that seems a little far fetched, but i cant see why one is allowed and one is not when they seem functionally the same. Lots of good information, but the more i read, the more unclear this seems.

    1. Rick Evans | | #19


      You can absolutely put the insulation above the rafters (and on top of your roof sheathing/decking). So long as the exterior foam is thick enough, it will "warm up" the roof decking so that interior moisture can no longer condense on the underside of the sheathing.

      If you put the foam on the inside using the "Cut and Cobble" method, or the "Tee method" you referenced above, then it works but the science is different. Here, you are not warming up the sheathing, it will remain cold. But, the strapping adds a ventilation layer between the top of the rigid foam in the cavities and the bottom of the roof decking. Therefore, when the sheathing gets wet (from interior moisture finding its way around the rigid foam) it has has an air space to dry.

      In the former scenario, the sheathing will likely never accumulate moisture. In the latter scenario, you are assuming that the cold sheathing will ultimately accumulate some moisture but you are giving it a chance to dry with the air space.

      In both instances, the rigid insulation needs to be thick enough for your climate zone.

      Hope this helps!!

      1. Rick Evans | | #20

        Just want to add that both of these assemblies must be a pain to build, especially in a colder climate (Zone 5 and up).

        Raised Heel Trusses are the way to go! :-)

        1. Malcolm Taylor | | #21


          Which plays a big part in why a lot of the high performance assemblies aren't going to be picked up by production builders. I spent over a decade d0ing spec. housing here in the wet PNW. The builds typically took around four months from excavation to putting them on the market. There was no time to let the project sit until good weather allowed the installation of exterior foam on a roof deck.

  19. Seth Guikema | | #22

    I’m also picking up on this older thread with a related question. We will be getting a new roof (metal) and plan to retrofit the portion of the house that has a cathedral ceiling. It is early ‘90s, 2x10 rafters, with R30 fiberglass with no air barrier other than the fiberglass facing. It is “vented” but very poorly. Our thought is to seal the soffit vents, pull the sheathing and vent baffles, put down new sheathing and then put 6 inches of polyiso on top, meting the 41% exterior R value rule for ceilings for our zone 5 location. One top of the polyiso is a layer sheathing, and then the metal roof. Two questions:
    1) Do I need a vapor barrier membrane on top of the sheeting just below the metal roof?
    2) Is the kraft paper on the fiberglass enough of a vapor barrier to worry about in terms of the ability of the assembly to dry to the interior? With the polyiso exterior to the rafters and current sheathing, we’re dependent on drying to the interior with this design.

    1. User avater
      Jon R | | #23

      Data I've seen indicates that variable Class II/III vapor retarders (like Kraft facing) improve the performance where code levels of exterior foam make it not required. Because contrary to popular belief, the sheathing still goes below the dew point, condensation/sorption occurs and some moisture accumulates. A Class I VR would be a problem (and isn't allowed by code).

      Be conservative (as you plan to be) - de-rate the polyiso and don't build right to the *minimum* R value limit here. Use a vapor permeable, fully adhered underlayment under the metal roof to reduce the moisture trap for the upper sheathing.

      IMO, it would also be a good idea to, before any of the upgrades, use a blower door to improve the air sealing of the ceiling drywall. Interior side air sealing has advantages over the exterior side air sealing that you will also do.

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |