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

SPF Cathedral Ceiling – Ventilation compromise?

user-882055 | Posted in Green Building Techniques on

I am planning to insulate my Victorian 2.5 story house’s attic. I’ve already torn down the plaster, and am gearing up to spray on the polyurethane. The problem I’m having is that

a) The rafters are only 4″ deep (real 2x4s).
b) I’d rather not furr them out any more than 2″. Any more will reduce the headroom unacceptably.
c) I’d like to get at least R30. The only real hope I have of reaching this (within the limits above) is 5″ of closed cell SPF, or 6″ of XPS with a poly vapour barrier.

I’ve researched the concerns and benefits of “hot roof” (aka un-ventilated) cathedral ceiling assemblies (where the foam is sprayed directly to the underside of the roof sheeting) on this site and others and I’ve found the following:

i) A 2″ or thicker layer of SPF is much better at eliminating condensation than simply ventilation
ii) SPF Hot Roofs might be more prone to ice dams than ventilated attics, but usually this is more than compensated by the improved insulation performance compared to fiberglass with an unreliable vapour barrier
iii) The Ontario Building Code (and I think the Residential Building Code in the US) now allows unventilated Hot Roofs.
iv) SPF Hot roofs might not dry as quickly or show symptoms after a failure of the shingles / roof membrane. As a result, a leak may go undetected until the roof is severely damaged.

The first three arguments are reassuring, but the last one raises enough doubt that I’m wondering if I shouldn’t try something new. Here’s my thought: “if the only problem with SPF Hot Roofs is that you can’t detect a leak, why can’t we design a channel just for that purpose?”

To that end, here’s my proposed roof assembly, from top down:

– Existing Shingles
– Existing Decking (with existing vents)
– 1/2″ thick Corrugated Plastic Roof Sheets between rafters. Possibly sprayed with water-soluble paint
– 3-3.5″ of spray closed-cell polyurethane
– 2″ of wallboard over the top of the rafters
– drywall

The idea is that I don’t need the ventilation channel to remove much heat or vapor, so it doesn’t have to be very thick. It’s just there as a “backup roof”, to carry any leaks down to the soffit where they will be detectable. Furthermore, I could paint the exterior side of the currogated sheeting with water-soluble paint, so that any leak will stain the soffit noticeably.

The only disadvantage I can think of is this: If, (in 15 years or so when the current roof wears out) I want to replace the shingles – I can’t improve the insulation by adding above-deck foam (the ventilation channel would negate that benefit, unless I could seal it up).

Is this a great idea or a terrible one? I have to decide within a week or so!

my research sources are here:

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

    Your idea will work. Basically you are installing closed-cell foam with a ventilation channel between the top of the foam and the roof sheathing. Other installers have done it before, although no one I know have has suggested using corrugated plastic roofing as a ventilation chute. (It's a good, rugged choice -- but fussy to install, I imagine.)

  2. terrell wong | | #2

    Depending on your location this is acceptable without drainage thingy. Some studies shows this will reduce the life of your shingles, others say the opposite. Not helpful. But as long as there is not an additional vapour barrier on the roof and you are not in an extreme location you should be fine. I would suggest saving your pennies so that when your existing roof is due for replacement you can add a metal one which self ventilates. This also gives you the oportunity to add polyiso above the roof at a later date if you are so inclined. Terrell Wong, Stones Throw Design Inc.

  3. user-882055 | | #3

    Do you think it's worth the "fussiness"? It would be much simpler to have the insulator foam right to the decking, and it would give me the option later of insulating under the shingles. I'm just worried a bout leak detection & drying. Is that unfounded?

  4. Riversong | | #4


    You don't indicate whether your roof has existing continuous soffit and ridge vents and uninterrupted ventilation pathways (no hips, valleys or dormers).

    If so, I would not suggest compromising the many benefits of roof ventilation, including reducing summer radiant heat gain.

    Your reason number 2 is meaningless, since increased insulation performance does nothing to reduce the potential damage from ice dam backup and leakage, and comparing a high perfomance option to the worst performing option ignores everything in between.

    The most important function of roof ventilation is not to be able to detect a roofing leak but to dry out the roof sheathing and framing in the event of a leak. No hot roof can offer this benefit, and it's one that can make the difference between a durable structure and one that will have a substantially increased potential for catastrophic failure.

    Here's a much simpler and more effective option. Install 3" of foil-faced polyisocyanurate insulation board between the rafters, leaving 1" air space below the roof deck for ventilation (anything less is inadequate), and seal the joints and edges with canned foam. Install an additional 2" of polyiso below the rafters, also taping joints and sealing edges and penetrations to serve as the air barrier. Then drywall with long screws or, better yet, apply horizontal strapping (you can reduce the subrafter foam to 1½" if necessary) and then drywall.

    This will give you better than R-30 with an excellent air/vapor barrier, a good vent and drainage channel, and a radiant barrier against summertime gain.

  5. user-882055 | | #5

    Robert: The house has vented soffit panels and individual roof vents (three pairs per side) positioned half way up (below the knee wall) and near the ridge. I have no idea if that's adequate - for instance the soffit panels are fastened to a great big sill plank, which is probably blocking most of the air flow. There is one dormer over the stairs, otherwise the roof is a continuous 12/12 gable.

    I thought about something like your rigid board plan, but a) I can't find a supplier for polyiso in Canada. The reno warehouses only carry polystyrene. And b) I didn't want to have to cut each board to fit old uneven rafters and collar ties.

    The ice dam issue seems to be moot. The thinking I read was that once you have a good solid R30, *and* a functioning air flow barrier, the heat loss through the roof should not be enough to melt snow in all but the warmest of days with the deepest of snow coverings. Most of the heat loss with fiberglass / cellulose attics comes from air exchange, which SPF effectively stops. Plus, for ventilation to prevent ice dams, it would have to be guaranteed to remove most of the heat that leaks through the insulation. That might be a tall order when we've already assumed a deep snow pack which might be covering the vents.

    I've pretty much abandoned the idea that, with spray foam, you need good airflow under the roof deck (that seems to be the growing consensus). But a way to identify leaks would still be useful. (The best method would be to have the vapour and heat barrier above the structure, but my roof is in too good a condition to tear up)

  6. Riversong | | #6


    You're correct in guessing that your existing roof vents are inadequate. Nothing short of continuous soffit vents at 9 per linear foot of net free vent area and continuous ridge vents with 18 and exterior wind baffles will perform as required. Your double row of roof vents will short circuit the circulation path.

    But you're incorrect about the parameters for ice dam formation. In a cold climate, an R-38 cathedralized roof with any more than 9" of fresh snow will maintain melting conditions at the roofing plane.

    Soffit vents are, of course, protected from snow and ridge vents rarely get covered because of wind effects. In any case, snow is somewhat air-permeable until it becomes glazed with ice.

    But, if you are not planning to improve your roof venting, then you may as well eliminate it and take your chances with what arguably has become the growing convention, if not consensus about hot roofs as a reasonable alternative to a properly vented roof.

    However, almost everyone in the building science community, while acknowledging that hot roofs can work if perfectly done, continue to recommend vented roofs in the north country.

    Bill Rose, ASHRAE, Illinois Building Research Council:
    Airtight ceilings are a more reliable way to ensure a dry attic than venting, but in practice most houses fall into a middle ground where venting balances moisture input.

    Don Fugler, research director, Canada Mortgage and Housing Corporation
    Houses with ceilings tight enough to meet Canada’s strict R-2000 standard “probably could get by without roof venting.” But most Canadian houses, even new ones, don’t have such perfect ceilings.

    Ned Nisson, author, The Superinsulated Home Book, editor, Energy Design Update:
    "I hesitate recommending [unvented roofs] to clients unless I’m absolutely assured of impeccable quality control. In my opinion, roof ventilation is cheap insurance against expensive callback problems. Why gamble?"

    Wayne Tobiasson, research engineer U.S. Army Corps of Engineers, Cold Regions Research Center:
    “Because of the monumental problem of ice damming, there is no question in my mind that the ventilated roof is an order of magnitude better in cold regions.”

    Anton TenWolde & William B. Rose, members, ASHRAE:
    We recommend venting of attics in cold and mixed climates. However, if there are strong reasons why effective attic vents are undesirable, unvented attics can perform well in cold and mixed climates if measures are taken to control indoor humidity, to minimize heat sources in the attic, and to minimize air leakage into the attic from below, or vice versa. The necessity and effectiveness of vents in cathedral ceilings in cold and mixed climates is still a contested issue. Unvented cathedral ceilings can perform satisfactorily in cold and mixed climates if the cavity is properly insulated, measures are taken to control indoor humidity and minimize air leakage into the roof cavity, and a vapor retarder is installed in the ceiling.

    Paul Fisette, director of Building Materials Technology and Management at the University of Massachusetts, Amherst:
    "There are many ways to treat the symptoms [of ice damming], but proper air sealing, insulation, and attic venting are the best ways to eliminate the problem."

    Joe Lstiburek, Building Science Corporation:
    "Vented attic/roof designs have the advantage of a long, proven historical track-record. However, they work best with airtight ceiling/attic interfaces and where ductwork and air handlers are not located within attic spaces. The increase in the use of complex roof shapes and cathedral ceilings has resulted in problems with vented roofs."

    "In extreme snow regions it is necessary to add a vented air space between the roof cladding (shingles) and the rigid insulation to avoid ice damming. The vented air space is needed to flush heat away trapped by the insulating value of relatively thick snow."

    Journal of ASTM International, Volume 6, Issue 4 (April 2009)
    Peter E. Nelson P.E., Senior Principal,Simpson Gumpertz and Heger, Inc., Jason S. Der Ananian P.E., Senior Staff I-Building Technology,Simpson Gumpertz and Heger, Inc.:

    All of the unvented roof assemblies are intolerant of incidental water leakage and the moisture-sensitive layers (i.e., sheathing and gypsum wallboard (for open-cell polyurethane insulation)) exceed the threshold for decay. In hot, humid climates, the most durable roof assemblies are the vented, open-cell polyurethane systems with shorter drying time of the interior gypsum wallboard when compared to the unvented roof assembly; both the sheathing and gypsum wallboard dry out within 2-1/2 months. In an unvented assembly, the drying time for sheathing is similar but the drying time for gypsum wallboard increases to 6.5 months on average. Alternatively, unvented permeable shingled roofs are a viable option in hot, humid climates, although they are slightly less durable. The least tolerant roof assemblies in either climate are the unvented closed-cell polyurethane roof assembly due to trapped moisture and slow drying of the roof sheathing (up to 12 months in Miami, FL and 27 months in Boston, MA).

  7. user-882055 | | #7

    That's a fantastic compilation Robert. Thanks for being so thorough. I've read the Ananian et al report, but not the others. Definitely the "perfectness" of the moisture barrier seems to be critical, and most of your experts are prudently hedging their bets.

    A new option for avoiding ice dams: As I understand them, ice dams will only form when the meltwater refreezes over the uninsulated, clod soffit portion of the roof. If I carry the hot roof right down to the fascia and back across the soffit, the roof should be an even temperature all the way down - ie, no place for water to re-freeze. Thoughts?

  8. Riversong | | #8


    That's an innovative solution, but for one problem. If you have gutters on your roof (and gutters are the first line of defense against exterior moisture problems), then they will be cold spots that will collect water, begin to freeze and potentially allow water to back up under the roofing.

    Renovation is always a challenge involving the balancing of pros and cons. But for new construction, I would never design or build a house without both rain gutters and a well-ventilated roof. Belts and suspenders.

  9. TFMdMgCRGe | | #9

    I'm in a similar situation--I have an attic space that is underinsulated and does not allow sufficient space for ventilation and insulation, so I am researching doing an unvented roof.

    Regarding the potential of ice dams forming in the gutters, does anyone have experience with heated gutters? Not trying to start anything--just trying to explore all the options besides bulldozing the house.

  10. Riversong | | #10


    Heating gutters or roof eaves with electrical heat tape is an expensive option and often fails to work because it melts snow or ice only immediately around the heat tape and that meltwater then refreezes somewhere else.

  11. user-882055 | | #11

    OK, here's the post-mortem:

    I took Robert's list of expert opinions seriously and decided I would create a proper ventilation channel before spraying the foam. I went with the cheapest channels I could find - the thin styrofoam ones with the 1.5" deep cones for separation. At the soffit and the roof, I nailed up some furring strips and painstakingly (emphasis on the pain) cut 1/2" rigid foam to block off air cavities. The whole exercise took me three weeks worth of evenings and probably $400 worth of material.

    The spray foam guys came today. They were livid. Turns out the stupid little grooves and cones in that commercial ventilation channel make it really hard to eliminate voids in the spray. They finished the job, but it took them much longer than they had estimated. So here's a warning to anyone following my path: Use the (more expensive) sturdy plastic ventilation channels if you're going to spray foam.

    For the foam producers / retailers: why don't you get together and come up with a ventilation system that is ideally suited to spray foam? I'm thinking little plastic spacers followed by some stiffish plastic that comes on a roll. This shouldn't have been so hard!

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |