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
Musings of an Energy Nerd

Insulating Low-Slope Residential Roofs

Unless you’re careful, your low-slope roof can end up with damp sheathing

Image 1 of 2
Rigid foam insulation can be installed on top of the roof deck. Most low-slope commercial roofs are insulated above the sheathing, not below the sheathing. [Photo credit: studio-tm.com/construction blog — Project: splyce.ca]
Image Credit: Images #1 and #2: Studio-tm.com/construction blog — Project: splyce.ca
Rigid foam insulation can be installed on top of the roof deck. Most low-slope commercial roofs are insulated above the sheathing, not below the sheathing. [Photo credit: studio-tm.com/construction blog — Project: splyce.ca]
Image Credit: Images #1 and #2: Studio-tm.com/construction blog — Project: splyce.ca
The most common types of rigid foam insulation installed under low-slope roofing are EPS (shown in the photo) and polyisocyanurate. Both types of foam are available in tapered configurations to create a slope on an existing flat roof. [Photo credit: studio-tm.com/construction blog — Project: splyce.ca]

UPDATED on April 8, 2016

There are lots of ways to insulate a low-slope roof, and most of them are wrong. In older buildings, the usual method is to install fiberglass batts or cellulose on top of the leaky ceiling, with a gap of a few inches (or sometimes a few feet) between the top of the insulation and the roof sheathing. In some cases, but not all, there is an attempt to vent the air space above the insulation to the exterior.

It’s rare for anyone to inspect the roof sheathing — unless, of course, the boards gets spongy enough to be noticed when the building is re-roofed. If there were any way you could squeeze into the tiny attic under the flat roof, however, you would probably see evidence of mold or rot.

Defining our terms

What’s a low-slope roof? It’s a roof that is flat or almost flat. This type of roof is common in urban areas (for example, on triple-deckers in Boston and row houses in Philadelphia), as well as in the Southwest. Some of these roofs have parapets — perhaps on just one side of the roof, or perhaps on three or four — while others have no parapets at all.

These roofs are either framed with deep roof trusses, or are framed with roof rafters that are separate from the lower ceiling joists (creating a cramped attic between the flat roof and the ceiling). In some of these buildings, the attic is high enough to allow a person to climb into the attic through a hatch; in others, the attic is too cramped for human access.

Venting the attic under a low-slope roof is possible but difficult

So what’s wrong with insulating a flat roof the traditional way? Nothing, really — as long as the job is done correctly: that is, with an airtight ceiling and adequate attic ventilation.

The problem is that most of these roofs aren’t built correctly. The ceilings leak air, and the attic ventilation is inadequate. There’s just enough ventilation to pull warm, moist interior air through ceiling cracks; once the moist air is in the tiny attic, the moisture accumulates in the cold roof sheathing. The result is rot and mold.

Here’s some advice from Joe Lstiburek, a principal at the Building Science Corporation: “If you have an airtight ceiling, and you have an air gap of at least 6 inches between the top of the insulation and the roof deck, and if you have perimeter air coming in at vents at the soffit or fascia above the insulation, and if you also have ventilation openings near the center of the roof through some kind of cupola or doghouse — not just a whirlybird turbine vent — there is nothing wrong with your roof assembly,” Lstiburek told me recently. “You can build a 2 foot by 2 foot doghouse that sticks up a few feet, and put in some rectangular vents. If the ceiling is airtight, then the makeup air comes from the outside. That’s the least expensive way to do things.”

Lstiburek continued, “The problem with this type of roof is that it is rarely executed correctly. Usually, architects don’t want to provide any ventilation around the perimeter. Or the architect won’t provide a deep enough truss to get enough insulation. If you just have a few whirlybird vents and a leaky ceiling, the whirlybirds will suck moisture-laden air out of the building and the roof will rot.”

Bruce Harley, the technical director for residential energy services at the Conservation Services Group, shares Lstiburek’s contempt for turbine vents. “I dislike turbine vents,” Harley told me. “I’d prefer a big mushroom vent or two over a turbine vent.”

The right way to vent a low-slope roof

If you want to build a low-slope roof that is insulated with fluffy insulation, here are the details you need to include:

  • Specify very deep roof trusses. The trusses should be deep enough for 12 to 16 inches of insulation (depending on your climate), plus room for an air gap of at least 6 inches between the top of the insulation and the roof sheathing. Even better: frame the roof separately from the ceiling, so that there is an attic that is deep enough for human access.
  • Provide vents at the perimeter of the shallow attic. These can be soffit vents, fascia vents, or wall-mounted vents, as long as the vents allow exterior air to connect with the air gap between the top of the insulation and the underside of the roof sheathing.
  • Provide one or more vented cupolas (“doghouses”) in the center of the roof. Most building codes require 1 square foot of net free ventilation area for every 300 square feet of attic floor area; half the ventilation area should be located at the perimeter of the building, and half of the area should be located at the cupolas near the center of the roof.
  • Perform air sealing work at the ceiling before the insulation is installed. Pay close attention to electrical penetrations, plumbing vent penetrations, the top plates of partition walls, and access hatches. The ceiling should be airtight as you can make it.

Bruce Harley emphasizes the importance of air sealing. He said, “Besides the standard bypasses — the partition walls and plumbing penetrations — remember that these older masonry buildings often have furring strips at the perimeter walls, and the cavities created by the furring strips may reach into the attic and need to be air sealed.”

If you forget to vent the attic, lots of things can go wrong

Builders in Arizona often use open-web trusses to frame low-slope residential roofs. Some of these builders cut corners: they omit the air space above the insulation and don’t bother to install any ventilation openings. They just jam the fiberglass batts up against the underside of the roof sheathing, wire the insulation in place, and cross their fingers.

Oops. About seven years ago, this insulation method was implicated in a cluster of wet-roof failures in Arizona. The first signs of problems were drywall cracks at the intersections between ceilings and partition walls — classic signs of truss uplift. (Truss uplift occurs when the top chord of a roof truss experiences different humidity conditions from the bottom chord; the humidity difference causes the trusses to deform.)

Uncertain of the cause of the drywall cracks, one of the builders called in William Rose, the well-known building scientist from the University of Illinois, to investigate. Rose discovered that the homes had wet roof sheathing — due in part to the type of roofing installed on the affected homes (white membrane roofing). “In December, January, and February, the fiberglass was wringing wet,” Rose told me. “In this climate, radiant effects become really important. There is nothing standing in the way of the roof radiating out to space. You have a whole lot of heat loss from the roof surface, day and night. With this white roofing, 80 percent of the heat that hits the roof is reflected. The sun can’t keep up with the heat losses to the sky. What you’ve created is a sky-powered cooling coil, and the fiberglass insulation is like a dirty condensate pan. The roof sheathing gets so cold that it is sucking wetness out of dry air.”

John Tooley, a senior building science consultant at Advanced Energy Corporation in Raleigh,

North Carolina, was also called in to help investigate the case. “At one roof I investigated — it was a flat-top roof assembly with a hot tar membrane roof coated with an off-white elastomeric coating — we pulled the roof off to take a look,” said Tooley. “The roof deck was totally saturated, and there was mold growth all over the bottom of the sheathing. The moisture content was greater than 30 percent. The fiberglass insulation was totally wet. This was in a house that was less than a year old.”

Tooley told me that this type of failure was common. “If you busted open roofs all over the Southwest, you’d find that the lower the pitch of the roof, the more you would see the evidence of moisture,” he said. “All of these roofs get wetting and drying cycles. If the wetting cycle is long enough, mold will grow and the insulation will get wet. I think if you cut roofs open, you will often find out that they are wet.”

The recommended solution to the problem of these wet Arizona roofs was to add a layer of rigid foam insulation above the roof sheathing.

What if you don’t want to depend on roof venting?

Let’s face it — it’s hard to vent a flat roof. That’s why most commercial low-slope roofs, including the roof on your local WalMart, are unvented.

In many ways, it’s easier to build an unvented low-slope roof than a vented low-slope roof. If you go this route, there are several possible ways to proceed:

  • You can install a thick layer of rigid foam insulation (6 inches or more) above the roof sheathing.
  • You can install a more moderate layer of rigid foam insulation (2 to 4 inches) above the roof sheathing, supplemented by a layer of air-permeable insulation below (and in direct contact with) the roof sheathing.
  • You can install a layer of closed-cell spray polyurethane foam roofing on top of the roof sheathing, supplemented by layer of air-permeable insulation under the roof sheathing. (For more information on spray-foam roofing, see Spraying Polyurethane Foam Over an Existing Roof and Roofing With Foam.)
  • You can install a thick layer of closed-cell spray polyurethane foam on the underside of the roof sheathing.
  • You can install a more moderate thickness of closed-cell spray polyurethane foam on the underside of the roof sheathing, supplemented by a layer of air-permeable insulation below that.

Of course, the total R-value of your roof insulation must at least meet minimum code requirements. Moreover, if you install a combination of foam insulation and air-permeable insulation, you need to be sure that the foam insulation is thick enough to keep the roof sheathing (or the lower surface of the foam insulation) above the dew point during the winter. The minimum R-values for the rigid foam insulation needed for this type of roof assembly are shown in the table below.

These roof assemblies dry inward

The insulation methods described above — those used for unvented low-slope roofs — are similar to the methods used to create an unvented cathedral ceiling. To read about the methods in greater detail, see How to Build an Insulated Cathedral Ceiling.

While vented roof assemblies are designed to dry to the exterior, unvented roof assemblies are designed to dry to the interior. That’s why an unvented roof assembly should never include interior polyethylene. (If a building inspector insists that you install some type of interior “vapor barrier,” you can always install a smart vapor retarder like MemBrain to satisfy your inspector.)

For more information on roof assemblies with exterior rigid foam, see these two articles:

  • How to Install Rigid Foam On Top of Roof Sheathing
  • Combining Exterior Rigid Foam With Fluffy Insulation
  • A (somewhat) controversial approach

    What if you need to insulate an existing low-slope roof with attic access on only one side of the building? This type of attic might be 3 feet high on the high side, but might taper down to only 6 or 8 inches on the low side.

    Doing it the right way probably requires some ceiling demolition and a spray-foam contractor (if the work is performed from the interior), or else requires new roofing (if the work is performed from the exterior). Either approach is expensive, so some contractors have figured out a way to insulate this kind of roof without demolishing the ceiling and without installing new roofing.

    Bill Hulstrunk is the technical manager at National Fiber, a manufacturer of cellulose insulation. When I interviewed Hulstrunk recently, he described a technique to insulate tapered attics with limited access. “With that type of attic, we crawl in and do as much air sealing as possible on the side with good access. Then there is a point where the attic gets too confined and you can’t crawl in there to do any air sealing. So we’ll dense-pack the side of the attic with limited access, and then we’ll blow in loose-fill cellulose on the side of the attic where there was enough access for air sealing work. It is always a good idea to have some vents on the side walls, above the top of the insulation, to provide some connection between the air above the loose-fill insulation and the outside. If we have done a good job with the air sealing, we have reduced the amount of moisture that will get up there. But in case there is some moisture that gets through, it’s good to have some way to allow the moisture to be able to make its way to the exterior.”

    Hultrunk’s approach receives qualified endorsement from Bruce Harley. In his book, Insulate & Weatherize, Harley writes, “Even an excellent dense-pack job can allow some air movement. In an unvented cathedral ceiling or flat roof, this can deposit moisture at the roof deck, especially in a home with high humidity. There are two basic strategies to avoid increasing the risk of condensation and potential damage to the roof deck: using foam insulation to control condensing temperatures, and ensuring an opening from the unvented cavities into a larger, vented space. The first approach — using continuous foam insulation — is the only proven, code-approved method for an unvented roof. … The second approach is to provide a partial venting path for the closed dense-pack area. … If one end of the dense-packed area is open to a vented attic space (preferably the top), any wetting effects appear to be balanced by drying toward the vented space. This approach can also be used under low-slope roofs (for example, a row house or shed dormer), where access near the low side is impossible. Experience has shown that up to one-third of the total attic area can be dense-packed without venting, provided that the remaining attic space is vented normally. Note that this method does not conform to standard code requirements but has been accepted by many local building officials. And I would consider this approach much more risky in climate zones 6 to 8 [than in warmer zones].”

    When I interviewed Harley about using dense-packed cellulose in low-slope roofs, he was cautious. “I am not comfortable trying to dense-pack an entire attic cavity, especially where it gets deeper in some areas, and therefore harder to dense-pack well,” Harley told me. “The risks are just too high. There are examples of dramatic and expensive failures. There are lots of questions: how effectively did you really dense-pack over areas where there may be air leaking? Are there small flaws in the cellulose? What is the moisture load in the house? We don’t really have control over that.”

    Harley still endorses limited use of the dense-packing method, as long as a list of conditions is met: “For a common rowhouse in Chicago or Philadelphia, in climate zone 5, in a building with effective code-compliant venting of the attic space, we have seen pretty good results from an approach that includes dense-packing the lowest part of the the attic,” Harley said. “But never more than one-third of the total attic area.”

    Repairing a problematic roof

    What if you are called in to repair problems in an existing building with a low-slope roof that shows signs of moisture?

    “We fix the problem roofs — the ones that get moldy — one of two ways,” Joe Lstiburek told me. “When the roof sheathing gets moldy, people freak out. The usual way we repair them is from the inside. We take out the gypsum ceiling and the insulation, and we spray 2 or 3 inches of closed-cell spray foam on the inside of the roof sheathing and the inside of the short walls. We encapsulate the mold. Then we repair the ceiling and we blow the space full of cellulose. If the space is too deep to fill with cellulose, we sometimes blow low-density spray foam over the high-density spray foam, because low-density foam is cheaper than high-density.”

    The second way to fix this type of roof is from the exterior. “We’ll install two 2-inch layers of polyiso on top of the sheathing, then a layer of OSB, and we’ll screw it all down and install the roofing. From the inside, we’ll install fiberglass batts up against the roof sheathing, held in place with metal pins. That’s a foolproof method.”

    A quirk in the code

    Lstiburek also explained a code quirk that sometimes results in superinsulated ceilings. “In California, the code for multifamily construction requires that you need to install sprinklers if that ceiling space is not completely filled with insulation,” he said. “Sometimes we are called into a project with 2-foot-deep parallel chord trusses. If they install 14 inches of insulation, and a ventilated space above, the system works perfectly fine. But to save the cost of the sprinklers, they have been building unvented ceilings, putting R-20 of rigid foam on top of the roof sheathing and R-50 of fluffy insulation in the roof trusses. It turns out that R-70 insulation is cheaper than sprinklers.”

    Martin Holladay’s previous blog: “Are Affordable Ground-Source Heat Pumps On the Horizon?”

    Click here to follow Martin Holladay on Twitter.

125 Comments

  1. Bill Rose | | #1

    Some observations that were not publishable
    Good job.

    Whenever you do research you always have the recorded findings and then some side observations — never enough for publication, but they get filed away for later (or not). We had two pertinent observations in this Arizona study, without enough support to draw publishable conclusions:

    1) Venting didn't work. It may have had a slight effect of warming the underside of the roof, but with low-slope it's all wind and no buoyancy, and at night the air is quite still. Plus, they were adding vents without adding an air space. Not helpful and not publishable.

    2) Cellulose insulation was installed in one single case. The cellulose was at the bottom of the assembly rather than at the top, of course. That assembly had no problem at all and stayed really dry. I think we underestimate the impact of even seasonal moisture storage in cellulose.

    Bill Rose

  2. Jin Kazama | | #2

    I sure hope that nobody is
    I sure hope that nobody is still building "vented flat roofs" nowadays.
    steel deck + vapor membrane and exterior insualtion is the only way to do it.

  3. Jin Kazama | | #3

    Martin: just a quick thought,
    Martin: just a quick thought, the same "dew point" rule/calculation applies to roofing assembly then walls ?
    If so, it would mean that some quantity of interior insulation could be possible even if within the final vapor membrane ?

    I do like the steel roof deck used a a reflective radiance material though ...would loose this property if packed up with some insulating materials ??

  4. Skylar Swinford | | #4

    Dimensional Stability
    Martin,

    Thanks for the great article. I noticed in the photos that the roofers/insulators are using sheets of EPS installed in staggered layers that are considerably smaller than an average 4'x8 sheet, but you didn't touch on this detail in your article. The Manual of Low-Slope Roof Systems, has a great section explaining the importance of using smaller insulation boards to improve dimensional stability. For example, the National Roofing Contractors Association (NRCA) recommends using rigid insulation boards that are 2'x4' to reduce stress on roof membranes. Lstiburek does a great job explaining why rigid foam insulation expands and contracts in his "correction" to Foam Shrinks, and Other Lessons. The Manual of Low-Slope Roof Systems also has valuable information on inverted (IRMA) or protected membrane roof (PMR) systems that readers in more extreme climates may find interesting. Plus the manual contains more great info on the pitfalls of ventilating low-slope roofing and the authors do a great job hammering home the importance of preventing ponding with adequate slope and drainage.

    So it looks like the roofers/insulators in the photos got the smaller rigid foam boards right; however, there doesn't appear to be a structural air-barrier at the roof sheathing. Perhaps they are relying on SPF or SIGA tape below...any thoughts?

    Readers may appreciate the article: Breaking With The Code for more observations regarding the typical ineffectiveness of ventilating low-slope roof assemblies.

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

    Response to Jin Kazama
    Jin,
    Q. "Does the same dew point rule/calculation apply to roofing assemblies as walls?"

    A. No. All of this is explained in my article, Calculating the Minimum Thickness of Rigid Foam Sheathing. While most of the article talks about walls, there is a section ("Is there a similar chart for unvented cathedral ceilings?") that talks about roofs.

    The reason that the recommendations differ for roofs is that rafters are usually deeper than studs. It's common for rafters to have more insulation than studs; that makes the sheathing colder, which means that the rigid foam has to be thicker.

  6. User avater GBA Editor
    Martin Holladay | | #6

    Response to Skylar Swinford
    Skylar,
    Thanks for the added information, and for the useful links.

    You're right that the roof in the photo would probably have benefited from an air barrier between the roof sheathing and the insulation (or at least taped plywood seams).

    For more information on dimensional instability in roofing foam, see my July 2000 article on the topic, "Shrinking Insulation Boards Plague Roofers." The article was published in the "Notebook" section of JLC; scroll down to the end of the section to find the article.

  7. User avater
    Dana Dorsett | | #7

    Expanding on Martin's response to Jin Kazama
    Not only are the insulation depths bigger, in a roof assembly there is an extremely low-permeance exterior finish surface (the roofing material itself) that is often wet with dew/rain/snow anyway. Literally ALL drying of the assembly is toward the interior in an unvented roof.

    This can be true of wall assemblies with foil-faced foam on the exterior too. The IRC prescribed minimums are just that, MINIMUMs, not optimal, not necessarily the best. Personally I'd normally shoot for more exterior R, and a more vapor-permeable exterior R. (EPS is about permeable at R15 as XPS is at R8) When there are foil facers in the mix I prefer to go 25-50% more than the IRC prescriptive min- margin counts.

    But the moisture buffering of cellulose also counts. At IRC prescribed minimums there's a real difference in the seasonal moisture cycling of the sheathing between mineral/glass wools and cellulose (any density) for the fiber layer of the stackup. Bill Rose's second observation in the first comment is not a fluke.

    FWIW: A few years ago I dense packed a portion of the underside of 2x10 rafters 12" o.c of a north side flat roof shaded by the taller portion of the building. It has tapered EPS on top of the roof deck under a membrane that has since been replaced. Other parts of that roof had ~R10 rock wool batts between the rafters. Some of the roof decking on the portion with the batts became punky and had to be replaced during the re-roofing, but so far the dense-packed portion remains solid. This is in a building with low occupancy rates and few interior moisture sources, and though the foam/fiber ratio wouldn't cut it from a code point of view in a residential building, I'm not too worried- it seems to be doing the job.

  8. Jin Kazama | | #8

    Seriously, how much more
    Seriously, how much more labor is is to make a vented roof and install cellulose/batts on ceiling than
    it would be to simply use more rigid insulation on top.
    With the current EPDM and other membrane prices and technology ...

    I am considering a balasted ( pebles ) epdm on top of a "to be determined" R value of rigid insulation
    that will be fastened to the steel deck through a peel stick membrane that will serve as inside vapor barrier.

    I don't see how one could do simpler than that labor wise.

    wood sheathing has no place in a roof, always ends up leaking somehwere and then it needs to be torned out again whereas a leak on a steel deck can usually be fixed without destroying the roof.

  9. Dan Kolbert | | #9

    Bill's Sample of One
    I realize it would be dangerous to draw much of a conclusion from the one cellulose-insulated roof, but it certainly is interesting and begs for more investigation. Bill H. is chuckling to himself somewhere in the background.

  10. Kevin Dickson, MSME | | #10

    A 12" SIP can solve all the moisture issues and support itself
    An EPS structural insulated panel for the roof/ceiling system can really make it easy. Because of all the issues it solves, the extra cost is warranted. If the engineering is done right, the SIP will save a lot of money on roof joists and beams because of its inherent stiffness.

    A SIP is also the most elegant way to eliminate thermal bridging at the roof/wall connection.

    At R-48, however, it's a little less than optimum. Is anyone making a 12" polyurethane SIP? Or a 16" EPS SIP?

    Cover it with white EPDM, and you should be good for 50-70 years, even in hail country.

  11. User avater GBA Editor
    Martin Holladay | | #11

    Response to Kevin Dickson
    Kevin,
    As I wrote in the article, "The insulation methods [for low-slope roofs] are similar to those used to build an unvented cathedral ceiling." And in my article on cathedral ceilings, I wrote, "Another possibility, of course, is to build your roof with structural insulated panels (SIPs)."

    SIPs are rarely used for commercial low-slope roofs, because they are more expensive than the more common approach (installing rigid foam on top of the roof sheathing). The other problem with SIPs is the tricky matter of sealing all of those seams.

    Considering the higher cost of SIPs and the seam-sealing problem, most designers will probably stick with more conventional approaches to insulating low-slope roofs.

  12. Rich Backus, GMB CPHC | | #12

    Low-Slope Venting: "All Wind, No Buoyancy"
    Bill, Martin:
    If we define 'low-slope' as 'flat, or nearly so', do we state a pitch < 1 in 12 ?
    And to avoid a nerd-fest on that definition, the question is this: At what pitch would we expect buoyancy / air density differential to create ventilation air movement worth considering as a mechanism for drying ? (ignoring for discussion wind pressure differentials as a factor, and assuming Joe's 6" air space as entirely inviolate/ extant, except for dust and pollen particles, and the stray African Swallow carrying coconuts.)

  13. User avater GBA Editor
    Martin Holladay | | #13

    Response to Rich Backus
    Rich,
    Your question may be of interest to researchers, but probably not to builders. For builders, such hair-splitting falls into the category of "misplaced accuracy."

    If you are worried that your roof won't get enough air flow for proper venting, the solution is simple: build an unvented roof.

    For a builder, a low-slope roof is like pornography for a Supreme Court justice: you know it when you see it.

  14. John Walker | | #14

    Air-cooled insulation
    Martin:

    Isn't free ventilation pulling perimeter air up and through "dog-houses" reducing the insulation value of the fiberglass/cellulose insulation layer? Moving air is the enemy of effective insulation...

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

    Response to John Walker
    John,
    Q. "Isn't free ventilation pulling perimeter air up and through "dog-houses" reducing the insulation value of the fiberglass/cellulose insulation layer?"

    A. The degree to which wind-washing degrades the performance of air-permeable insulation depends on many factors. Ideally, the incoming ventilation air is introduced above the top of the insulation layer; if this is the case, then a ventilated low-slope roof is no different (at least in theory) from a ventilated attic.

    In a classic ventilated attic, wind-washing does slightly degrade the performance of the insulation near the soffit vents -- especially if the insulation is fiberglass, and especially if there is no wind-washing dam. If a wind-washing dam has been installed, or if the builder has chosen cellulose insulation instead of fiberglass, the thermal degradation caused by wind washing is much less.

    In general, there are three ways to minimize the negative effects of wind-washing:
    Choose cellulose, not fiberglass
    Install wind-washing dams near soffits
    To make up for possible wind-washing effects, make the insulation a little bit deeper than you otherwise would.

  16. Mitchell Daniels | | #16

    Open vs closed-cell foam on underside of a low-slope roof
    Good morning.

    Planning an addition in Montreal, Canada, with a low-slope roof (Soprema modified bitumen roofing on plywood or OSB on engineered joists).

    Was planning on spraying about 6 inches of closed-cell foam on the bottom of the plywood deck (as air and vapor barrier).

    I am being told that 12 inches of open-cell foam on the bottom of the plywood deck (as air barrier) with a paint vapor barrier may be a better option.

    The logic, as I understand, is that if there is a leak in the roof and closed cel foam was used, the owner will have no idea of the problem. The moisture will be trapped above the foam and the plywood runs the risk of rotting.

    However if open-cell foam was used with a paint vapor barrier, if a leak occurs, there is a much better chance that the owner will see a stained ceiling and take action before rot sets in

    Any thoughts/comments? Thanks in advance!

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

    Response to Marc Daniels
    Marc,
    I strongly urge you to choose the closed-cell spray foam.

    In your climate, open-cell spray foam (which is vapor-permeable) allows interior moisture to reach the roof sheathing by diffusion. Over the course of the winter, the sheathing can get wet enough to begin rotting.

    One solution that was recommended in the past (even I recommended it until I learned why it is bad advice) is to spray vapor-retarder paint on the underside of the cured spray foam. The only problem with this solution is that it doesn't work.

    Engineers at the Building Science Corp. conducted tests that showed that vapor-retarder paint only works when you spray it on gypsum drywall (a smooth surface). When you spray the paint on an uneven porous surface like cured spray foam, it is worthless as a vapor retarder. The interior vapor goes right through the paint.

    One possible solution (if you really want to use open-cell spray foam) is to cover the cured foam with a layer of gypsum drywall, and then to spray the drywall with vapor retarder paint.

    Better yet, just choose closed-cell spray foam.

    More information here: Creating a Conditioned Attic.

  18. Mitchell Daniels | | #18

    Response to Martin Holladay

    Thank-you Martin.

    I was not clear in my description. As for the open cell foam option, the plan was to fix drywall to the bottom of the engineered joists and then apply a vapour barrier paint to the drywall. The idea being that if there is a roof leak, the moisture would travel through the plywood roof deck and through the open cell foam and down to the drywall. The owner would then be aware of a leak and would able to fix before rot sets in.

    I would prefer to use closed cell foam. The concern in this case is that if there is a roof leak, water would be trapped in the plywood deck by the closed cell foam. The owner would have no idea of this until the plywood deck would rot out.

    Further thoughts?

  19. User avater GBA Editor
    Martin Holladay | | #19

    Response to Marc Daniels
    Marc,
    The question you raise is a thorny one. Many people have proposed the same theory: that roof leaks will show up faster if you use open-cell foam rather than closed-cell foam.

    Maybe. However, I have my doubts that the situation is as simple as the theory proposes.

    I used to work as a roofer, and I know from experiece how hard it can be to trace a roof leak by looking at ceiling stains.

    Some roof leaks show up as ceiling stains or ceiling drips immediately. Those are the easy cases.

    Far more common are the roof leaks that start slow and aren't noticed for a year. This happens all the time, even with conventional building materials like wood framing and cellulose insulation. When the ceiling stain finally appears, it can be 20 feet sideways from the leak. Water moves, and water does tricky things. Many building materials are absorbent, complicating the situation.

    It's even possible that closed-cell foam, being more waterproof, is more likely to protect vulnerable building materials from damage than open-cell foam.

    The fact is, there are too many variables to make a clear-cut ruling on this issue. The bottom line: no matter what type of roofing, sheathing, or insulation materials you use, you want to repair any roof leaks as fast as possible. Sometimes fast repairs are possible because the owners are paying attention, and sometimes fast repairs are possible because of luck.

    Other times, even when you think you're paying attention, a hidden roof leak can do considerable damage before anyone notices.

    Finally, if you are really worried about this issue, there is a simple solution: design a building with a ventilated, unconditioned attic, and go up in the attic with a flashlight regularly to inspect the roof sheathing.

  20. Mitchell Daniels | | #20

    Response to Martin Holladay
    Thanks again - very much appreciated!

  21. Confused Homeowner | | #21

    Insulating a cool, self-drying, vented, low-slope roof
    I own a mid-century modern home in the Seattle area with a 1/12 pitch roof over vaulted ceilings separated by 2x6 joists with R13 batt insulation, venting at the soffit and ridge, and a condensation issue that started after the modified bitumen roof was replaced last June with a white TPO membrane. I've been reading all of Martin's articles and I'm still not satisfied that I have the best solution. It seems that the root of my condensation problem is that we've lost sight of the self-drying properties of a torch down or modified bitumen roof assembly because we're so focused on cool roofs using membranes with lifetime material warranties. If you read some of these articles below, you'll see that the heat absorbing properties of these older roof systems are part of the design. You may also notice the precautions for colder climates in the U.S. Department of Energy Guidelines fo Selecting Cool Roofs.
    http://web.ornl.gov/sci/buildings/2012/1995%20B6%20papers/090_Desjarlais.pdf
    http://www.epdmroofs.org/attachments/2012-jan_coolroofscausecondensation_dregger_wr01123.pdf
    http://www1.eere.energy.gov/femp/pdfs/coolroofguide.pdf
    Unfortunately, I put complete trust in my roofer and now I have a leak free lifetime roof that condensates inside and doesn't dry out very quickly. In my search for a solution, I've found that most, if not all, of the technical documentation for TPO membrane are focused on commercial roof assemblies and they require an "adequate" insulation layer above the roof deck. It seems to make sense that some insulation above the roof deck would be appropriate when replacing a modified bitumen roof with a cool membrane roof to make up for the difference in solar reflectance and thermal emittance properties. However, residential roofers tend to dismiss the need for insulation above the roof deck as a waste of R value. What do you think? If I re-roof again this summer, should I go back to modified bitumen or should I add a layer of insulation above the roof deck and use a darker color of TPO or PVC? I've also considered replacing the batt insulation, either with a dense packed blow-in-blanket system followed by 3" of rigid insulation above the roof deck or maintaining the vented system with 4" of closed cell spray foam above the ceiling, or 4" of rigid insulation sealed in above the ceiling. These other options are a lot more expensive and are still not foolproof. I've already ruled out the option of removing the ceiling to fix the problem from underneath because it's too disruptive.

  22. User avater GBA Editor
    Martin Holladay | | #22

    Response to Confused Homeowner
    C.H.,
    The problem you are experiencing is described in my article, under the subhead, "If you forget to vent the attic, lots of things can go wrong."

    In those paragraphs, I describe how white membrane roofing stays colder than dark roofing, encouraging condensation. Building scientist William Rose describes this type of roofing as a "sky-powered cooling coil," while "the fiberglass insulation is like a dirty condensate pan. The roof sheathing gets so cold that it is sucking wetness out of dry air."

    Soffit-to-ridge venting doesn't work on a low-slope roof, as my article explains. For all intents and purposes, your roof assembly is unvented. The solution, as you correctly realize, is to install rigid insulation above your roof sheathing.

    Since your roofer recommended a defective roof assembly, your roofer should fix the problem at no charge to you.

  23. Confused Homeowner | | #23

    Insulating a cool, self-drying, vented, low-slope roof
    Since you said, "for all intents and purposes, your roof assembly is unvented", would you recommend converting to an unvented roof assembly? I've observed that dripping and other signs of condensation damage are concentrated under roof bays that have inadequate venting. This leads me to believe that my venting does work, but it needs to be improved with additional vents above and below my chimney and skylight. I'm also afraid of what might happen in a completely unvented roof assembly if I fail to completely seal off air leaks from the ceiling.

    If I stick with venting when I re-roof, then should I keep the R13 fiberglass batts if they seem to be in decent shape? Or should I replace the batts with closed cell spray foam or rigid polyiso that has been cut to fit? Both approaches would provide the benefit of a higher R value and the disadvantage of making any future changes, such as electrical wiring, more difficult. Spray foam would create a better moisture barrier, but it's more expensive and I think it would likely involve use of harmful hydroflourocarbon (HFC) blowing agents.

  24. User avater GBA Editor
    Martin Holladay | | #24

    Response to Confused Homeowner
    C.H.,
    Q. "Would you recommend converting to an unvented roof assembly?"

    A. Your R-13 batts do not provide anything close to minimum code requirements for roof insulation. In your climate zone (Climate Zone 4C), the 2009 IRC calls for a minimum of R-38 insulation. The easiest way to provide the insulation you need without disturbing your interior finishes is to install rigid foam insulation on top of your roof sheathing and to seal your vent openings.

    Q. "If I stick with venting when I re-roof, then should I keep the R-13 fiberglass batts if they seem to be in decent shape? Or should I replace the batts with closed cell spray foam or rigid polyiso that has been cut to fit?"

    A. You don't want to install polyiso between your rafters. (Polyiso is best installed in a continuous layer, not cut into narrow strips as you propose.) Since R-13 is insufficient, you should choose a method of insulation that allows you to achieve R-38 or more.

    When it comes to venting this type of roof assembly, I stand by the advice given in the article. The only way you can make venting work on this type of roof is if you can maintain a gap of at least 6 inches between the top of your insulation and the underside of your roof sheathing, and if you can cut a 2 ft. by 2 ft. hole in the roof sheathing near the center of your roof, so that you can install a doghouse above the roof to act as a vent outlet.

  25. Confused Homeowner | | #25

    Batts, Dense Pack, and Exhaust Fans
    Thanks to your advice, my solution is becoming much clearer. I understand that I'm not required to bring my mid-century home up to 2009 IRC standards, but achieving R-38 or higher does seem like a worthy goal. One concern about sealing the vent openings is that the 3.5 inch thick R-13 batts have probably been in place for over 45 years and are likely not achieving the stated R value. They also leave two inches of available space in the 2x6 joist bays. Should I replace the R-13 batts with 5.5 inch thick R-21 fiberglass batts or with a dense packed R-24 blow-in-blanket system (BIBS)? Also, since all this insulation will make my house tighter, I'm planning to replace my contractor grade bathroom exhaust fans with Panasonic WhisperFit 80 cfm fans to help control indoor humidity levels. Since these fans are 5.5 inches high, I'm concerned about how much insulation they would displace below the roof deck. Would it make sense to build dropped soffits above my shower and bathtub for exhaust fans so that only the ducts will need to penetrate the ceiling?

  26. User avater GBA Editor
    Martin Holladay | | #26

    Response to Confused Homeowner
    C.H.,
    Q. "One concern about sealing the vent openings is that the 3.5 inch thick R-13 batts have probably been in place for over 45 years and are likely not achieving the stated R value."

    A. Obviously, sealing up the vent openings would not be a solution if that were the only action you performed. Clearly, you need to add R-value to the existing assembly.

    Q. "Should I replace the R-13 batts with 5.5 inch thick R-21 fiberglass batts or with a dense packed R-24 blow-in-blanket system (BIBS)?"

    A. Either choice would be an improvement over what you have, but neither choice would meet minimum code requirements. Only you can decide whether you can afford to do better.

    Q. "Since these fans are 5.5 inches high, I'm concerned about how much insulation they would displace below the roof deck. Would it make sense to build dropped soffits above my shower and bathtub for exhaust fans so that only the ducts will need to penetrate the ceiling?"

    A. No one should ever install a bathroom exhaust fan in the insulated bays of a cathedral ceiling. You have two choices: you can build a soffit under the existing ceiling, or you can install a wall-mounted exhaust fan (assuming that your bathroom has an exterior wall). The fan should be installed as high in the wall as possible.

  27. Robert Vandermolen | | #27

    2/12 roof--to vent or not to vent
    I am building a ranch style home in the Chicago area (zone 5) with a low slope roof (2/12). We are installing a white TPO or PVC roof. My builder recommends going with a non vented roof using closed cell insulation to seal air leaks and meet energy codes. My architect recommends going with a vented roof using blown in cellulose to meet energy codes. I was leaning toward the vented roof until the architect drew in this linear cupola that ruined the lines of the house. I was hoping to have my house featured in Fine Home Building and not Mother Earth News. But I also don't want to be featured as an example of what not to do.

    How much rigid foam do I need to install on top of the roof to prevent the condensation effect that was describe when using the white membrane roof?

    Can I go with a standard low profile ridge vent and get the needed draw on a 2/12 roof?

    Or should I go with the closed cell and pray the installer seals it all off and lays it on thick?

  28. User avater GBA Editor
    Martin Holladay | | #28

    Response to Robert Vandermolen
    Robert,
    I stand by the recommendations given in the article. If you want a vented roof, you need the doghouse. If you don't want the doghouse, build an unvented roof.

    Rigid foam above the roof sheathing is a much better approach than spray foam under the roof sheathing (since the rigid foam addresses thermal bridging through the rafters). This is a standard approach for commercial construction, so talk to a roofer who handles commercial roofs.

    If you want to put all of your insulation above the roof sheathing, you can just consult the code books and do the math; I imagine that you will need at least R-38 in your climate zone. If you want to combine rigid foam above the roof sheathing with fluffy insulation below the roof sheathing, you will need to install at least R-20 of rigid foam above the roof sheathing in your climate zone.

  29. Confused Homeowner | | #29

    Weighing options to achieve similar R-Value
    I'm almost comfortable with the idea of replacing the R-13 batts in my 2x6 roof bays with dense packed insulation, either cellulose or fiberglass, plus adding at least three inches of polyiso on top. This article by Joseph Lstiburek both confirmed this approach as well as scared the heck out of me.
    http://www.buildingscience.com/documents/insights/bsi-043-dont-be-dense
    Here are some of the options I've considered for increasing the total R-value of my roof assembly from least expensive to most expensive...
    1. Keep R-13 batts and add 4" of polyiso above deck
    2. Keep R-13 batts, fill cavity with dense packed fiberglass or cellulose , add 3" of polyiso above deck
    3. Keep R-13 batts, add 2" fiberglass batts, add 3" of polyiso above deck
    4. Remove R-13 batts, fill cavity with dense pack, add 3" of polyiso above deck
    5. Remove R-13 batts, add 1" of closed cell spray foam above ceiling, fill cavity with dense pack, add 3" polyiso above deck
    Would you rule out any of these options? Which option, if any, would you choose?

  30. User avater GBA Editor
    Martin Holladay | | #30

    Response to Confused Homeowner
    Confused,
    As long as you follow the advice in this article -- especially concerning the minimum ratio of foam insulation to fluffy insulation -- any of your proposed options will work.

    The minimum R-value requirement for your layer of rigid foam insulation or spray foam insulation depends on your climate zone, so I can't evaluate your proposed foam thickness choices without more information.

    Your choice will also depend on whether you are willing to replace your roofing.

    In terms of performance, the best option would be option #4 -- as long as 3 inches of polyiso meets the minimum R-value requirements for this approach in your climate zone.

  31. Jerry Chwang | | #31

    No parapet wall on one side - how does insulation stay on roof?
    This might seem like a silly question, but how do insulation boards above the sheathing stay on the roofs that don't have parapet walls all the way around? Will the ballast be sufficient?

    I have a low slope roof where we've left the low side without a parapet wall, so it acts as one big scupper. Just thought about the XPS on top and wondering how I keep it from 'sliding' off? If I go with a green roof as ballast, this gets even trickier as my dirt will wash off for sure?

  32. User avater GBA Editor
    Martin Holladay | | #32

    Response to Jerry Chwang
    Jerry,
    Rigid insulation is attached to the sheathing with cap nails or cap screws (available at any roofing supply house). The perimeter of the roof is usually trimmed with flashing. If these concepts are new to you, you should hire a roofing contractor.

  33. Ron Stanley | | #33

    Best option for a newly constructed low-sloped roof
    The more I read this and other sources on the subject of insulating, the more certain I am that it can be very easy to get it wrong. I have a 100 year old craftsman house in Tacoma (Marine 4C) that is in need of major updates (siding, windows, insulating, and roofing). I am about to start my remodel-redesign which includes changing the architectural style from craftsman to modern. The first phase will include complete removal of the 8/12 gable roof structure. In its place I will be constructing a shed style 1/12 roof structure with gray TPO roofing. The new attic space will have a 60” tall wall on the high side (North) and a 34” tall wall on the low side (South) leaving plenty of space to access all areas. I-joist will span between the two walls creating a 9 ½” cavity below the ¾” OSB deck. The current ceiling has R21 High Density glasswool batts between the 2 x 6 joist and it has ½” drywall below. There are numerous penetrations in the walls and ceiling, (plumbing, can lights, electrical, bathroom fan) none of which were carefully sealed by the previous owner.
    Along with replacing the roof structure, I will also be removing and replacing all of the siding and many of the windows. I will be adding 3” of exterior rigid insulation and an open rain screen. The exterior insulation will run up the walls to the roof overhangs.
    The existing attic is unconditioned and vented and shows no signs of moisture issues. Given my location and planned redesigned roof structure, what would be the least problematic approach to take after doing my best to seal all of the wall and ceiling penetrations in the attic:
    1. Add more insulation to the ceiling and keep the attic space unconditioned and vented. If so is there any benefit or harm in adding a layer of rigid insulation between the OSB deck and TPO roofing?
    2. Add 6” of rigid foam between the OSB deck and the TPO roofing making the attic space part of the conditioned volume by removing the existing batts and eliminating venting.
    3. I’ve got it all wrong and should instead do this…
    One additional note, I am trying to do all of this while my family continues to live in the house so it is important to have very limited disruption to the interior space.

  34. User avater GBA Editor
    Martin Holladay | | #34

    Response to Ron Stanley
    Ron,
    I stand by the advice given in my article; the details of your house don't change my advice.

    Considering the penetrations you list, and the fact that your roof lacks a doghouse, it makes the most sense to follow the usual practices of commercial roofing and to install all of the insulation above the roof sheathing. If 6 inches of rigid foam is enough to get you to your minimum code requirements for R-value, then that's the way to go.

    Don't forget to seal your attic vents and insulate your attic walls.

  35. Ron Stanley | | #35

    Response To Martin Holladay
    Martin thanks for the response and this article. I have read and reread this article and all of the suggested linked articles. I have been finding it difficult to determine which advice applies for my application (low sloped roof over an attic space). I did not know that a 1/12 sloped roof would not vent soffit to soffit and would require “dog houses” mid span. I shared the details of my house because the multiple references to attic size “tiny attic” and “cramped attic spaces” made me wonder if my larger attic volume changed the equation in any way. I trust your advice and will go with the unvented more commercial style roof assembly.
    If I understand the information in this article correctly (You can install a more moderate layer of rigid foam insulation (2 to 4 inches) above the roof sheathing, supplemented by a layer of vapor-permeable insulation below the roof sheathing) along with the information regarding minimum rigid insulation for unvented roofs (Table R806.5 specifies the minimum R-value for the foam installed on top of the sheathing -- R-10 for Climate Zone 4C) than can I go with 4” of rigid polyiso – R value of 23.6 according to the new LTTR on top of the deck and then move my existing R21 glasswool batts up to the bottom side of the deck for a total of R44.6? I already have enough 4” thick polyiso to cover the roof deck and it would be great if I did not need to add another 2-4 inches on top for both cost and aesthetic reasons.

    In the future, would there be a problem with adding more vapor permeable insulation to the underside of the deck? It seems some people suggest having a 2/3 outside to 1/3 inside insulation ratio. I believe this is to keep from isolating the warm side of the deck and turning it into a cold surface that could allow condensation to occur.

  36. User avater GBA Editor
    Martin Holladay | | #36

    Response to Ron Stanley
    Ron,
    If you install R-23 of polyiso above the roof sheathing in your climate zone, your roof will exceed the minimum requirements for above-sheathing foam -- giving you a big margin to work with if you decide to add thicker fluffy insulation between your rafters. I think you will be fine.

    If you really want to perform the dew-point calculations to show how far you can go with your plan, here is a link to an article that tells you what you need to know:
    Are Dew-Point Calculations Really Necessary?

  37. Ron Stanley | | #37

    moving forward with the unvented roof assembly
    Martin,
    The roof deck is built (3/4 inch t&g osb above 9.5 inch I-joist), the 4" polyiso is on site, and I'm looking for a few last words of advice. There was a mention in the comments and your article of placing an air barrier directly on top of the roof deck and below the rigid insulation. This is also described in, "Complex Three Dimensional Airflow Networks" By Joseph Lstiburek but not much detail about the airflow barrier is given. Also, many articles speak to using multiple layers of rigid insulation, offset, to eliminate "pathways". My plan, maybe not a good one, had been to put the insulation directly on top of the osb deck, put a layer of densdeck above the insulation, and fully adhere the tpo to that. I could change the assembly to - osb deck, air barrier, 4" polyiso, additional 1" polyiso that is faced prepared for fully adhered systems, and then tpo. My questions are: Is the air barrier recommended and if so, what type of product do you suggest? How important is it to have multiple layers of offset insulation if an air barrier is used?

    I am also having trouble finding any information on the best way to build up the outer perimeter of the roof deck to form a nailing base for fascia, gutters, and flashing. Do I just build up using 2x4's or lumber to the height of the rigid insulation?

  38. Confused Homeowner | | #38

    Mounting Solar on Low Slope Insulated Roof
    I'm hoping to install a solar array after re-roofing my low-slope roof but I'm a little worried about heat loss from the mounting brackets that would be bolted into the roof deck. I will have an R-24 blow-in-blanket-system in the 2x6 joist bays below the plywood deck and 3 inches of poly iso above the roof deck covered by a PVC membrane. How concerned should I be about thermal bridging from the mounting brackets? I suppose my alternative would be to use a ballasted system for mounting solar panels. A ballasted system would not penetrate the membrane and poly iso, but it would likely be a lot heavier since it uses bricks to hold the rack system in place.

  39. Graham Swett | | #39

    Low slope unvented assembly
    Working on a townhome project. GC is onboard with doing unvented attic space below the low-slope trussed roof assembly. Climate zone 5B. Looking at using ccspf on underside of roof deck w/ white EPDM topside. Appears to be a viable assembly according to: http://www.buildingscience.com/documents/digests/bsd-102-understanding-attic-ventilation
    but I am concerned that there is no drying potential and any roof leak will probably go unnoticed.......thoughts?.......thanks.

  40. User avater GBA Editor
    Martin Holladay | | #40

    Response to Graham Swett
    Graham,
    Your chosen approach is one of the standard approaches for insulating an unvented low-slope roof assembly, and it will work.

    Most commercial roofers prefer to install rigid foam above the roof sheathing, because that approach costs less, and still allows the roof sheathing to be inspected from the interior when necessary. But you can do it your way if you want.

    You wrote, "I am concerned that there is no drying potential and any roof leak will probably go unnoticed." There are lots of things to worry about in life -- and if you're really worried about that issue, you should have built a steep roof over a ventilated attic.

  41. Graham Swett | | #41

    Response to MH
    Thanks Martin........as architects we are tasked with having to balance the demands of the client, the market, the planning and development dept, the building codes and a host of other factors that do not always align to allow us to take the path of least resistance. All we can do is approach the final design with intelligence so that the final built environment does not become a drain on the end user and a black-eye and potential law suit for the design/development team........thanks for contributing to the brain trust.........

  42. Sascha Zerbin | | #42

    best way to insulate a converted attic with a flat roof
    Hello Martin,
    I am in the middle of converting my attic into living space when I started researching about possible ways to insulate the ceiling/roof and I was reading your article with great interest! Thank you for bringing some more light to this challenging topic!

    Despite the many good suggestions, I am still not 100% sure about the best way to tackle my particular problem. Here are a few facts:

    The house is located in St Louis, climate zone 4,
    built in 1880, with original sheating, the flat roof is Modified Bitumen with a white coating) the sloped mansard roof to side is black slate. The slate is new and the flat roof is good for another 5-7 years. The ceiling joists are 2 x 10 in size and I am planning to attach the ceiling drywall directly to save as much ceiling height as possible.

    The 2 attached pictures might give you a better idea...

    In order to get the R-30 for the ceiling, I could use fiber glass batts or rolls, but that would negate any form of ventilation between insulation and sheating. From what I understand the best option might be the insulation with rigid foam on top of the sheating....but there is still some years left on the flat roof, I am planning to put the insulation on top of the sheating while putting on a new flat roof with new sheating...

    My questions are: What is the best and most cost effective, interim (5-7 years) insulation until I will put the new roof on?

    Is it better to coat the flat roof with a darker coating to avoid the "cooling coil" effect?

    Should I insulate the walls or between the sloped rafters or both?

    And what type of vapor barrier is best to use in my particular case?

    I also thought about creating a space of 3 inches between insulation and sheating by using 2x4 as spacers and putting a layer of rigid foam boards2-3 inches, sealing all the gaps with closed cell foam and fill the space underneath with fiberglass, than the vapor barrier and the drywall. But I might not reach R-30..

    Thanks for any suggestions!

  43. User avater GBA Editor
    Martin Holladay | | #43

    Reponse to Sascha Zerbin
    Sascha,
    In Climate Zone 4, the 2009 International Residential Code calls for a minimum of R-38 ceiling insulation, not R-30.

    I suggest that you insulate along the roofline rather than attempting to insulate the kneewalls.

    Since you will be creating an unvented insulated roof assembly, you can't use fiberglass or a similar air-permeable insulation. Your only two choices are spray foam insulation or a combination of spray foam insulation and air-permeable insulation.

    For more information on this type of insulation challenge, see How to Build an Insulated Cathedral Ceiling.

  44. Sascha Zerbin | | #44

    thanks for getting back so quickly!
    Hello Martin,

    the 09 IRC is amended in St Louis when it comes to insulation...

    Would it be possible to use rigid closed cell boards combined with closed cell spray foam instead of just spray foam? This way I could reduce costs by 80%!!

    And which class of vapor barrier/retarder would you recommend?

    regards

  45. User avater GBA Editor
    Martin Holladay | | #45

    Response to Sascha Zerbin
    Sascha,
    The method that you are proposing is called the "cut-and-cobble" approach. It is not recommended for unvented cathedral ceilings because of the risk of moisture problems and rot.

    To read more about the method, and to read warnings about its use for cathedral ceilings, see Cut-and-Cobble Insulation.

    If you decide to install spray foam insulation, you only need a vapor retarder if you choose open-cell spray foam. (Closed-cell spray foam is already a vapor retarder.) The usual vapor retarder for installations of open-cell spray foam is a layer of vapor-retarder paint installed on drywall. (The drywall is necessary for fire protection.)

    Don't attempt to install the paint on the cured spray foam -- that method won't work. It has to be installed on drywall.

  46. Gregg Berkholtz | | #46

    I'm getting a little worried here...
    While researching what you've called the cut-and-cobble method, I stumbled across this blog posting. After reading the article and then every comment, I'm a bit worried about our roof installation.

    We recently purchased a low-sloped (1 in 12) house in Zone 4C (Portland Oregon, Multnomah County). This 1950's house had a leaking torch-down roof...so at our roofer's advice we steered towards TPO as a replacement surface. After tearing off existing roof to the decking, what they installed was...listed in order from decking up: 2" PolyIso, FR-10 underlayment, 60Mil GenFlex TPO.

    Since we had opened the house's interior (total gut; for new electrical & plumbing, and replacing wall-insulation), our roofer then advised us to install new high-density batts in-between 2x6 joists; leaving 1" of airspace between batts and roof deck underside. We then needed to open the soffits a minimum 1", and use a 1"-2" circular saw to open up as much blocking as possible, as close to the top plates as we can get.

    From what I'm reading here, it appears our installation is likely to encounter moisture/condensation issues... Am I understandings things correctly?

  47. User avater GBA Editor
    Martin Holladay | | #47

    Response to Gregg Berkholtz
    Gregg,
    If you have rigid foam above your roof sheathing, you definitely don't want a ventilated air gap underneath your roof sheathing. So the first order of business is to close off the soffit vents in an airtight manner.

    The second issue concerns the decision to combine a layer of rigid foam above the roof sheathing with air-permeable insulation under the roof sheathing. This approach can work, but according to code requirements, the air-permeable insulation has to be in direct contact with the roof sheathing. Here's how the code reads: "In addition to the air-permeable insulation installed directly below the structural sheathing, rigid board or sheet insulation shall be installed directly above the structural roof sheathing as specified in Table R806.5 for condensation control."

    Your rigid foam sheathing is fortunately thick enough for your climate zone, but the gap between the fiberglass and the roof sheathing is a code violation (as well as bad practice). You should bring this fact to the attention of your contractor, who is responsible for following the building code.

  48. Gregg Berkholtz | | #48

    Response to Martin Holladay
    Thank you greatly for the information. The idea of an unvented roof is completely foreign to me - so your feedback greatly helps us better understand things.

    Since we have enough rigid foam sheathing above the roof deck (2"), it sounds like our best course of action is to focus on the void boxed between our rafters and roof sheathing/drywall.

    In 1/3rd of the house, we've already drywalled, so it's damage control: it appears our most reasonable path is to drill & blow-in a dense cellulose to fill the gap between fiberglass bats and roof sheathing, and then seal openings once complete.

    For the rest of the house, we had originally planned on a 1-2" airspace between batts and roof sheathing. Since that part of our project is just beginning (two roofs to deal with - one is on 2x6 & the other on 2x10 rafters), it appears we have additional options.

    Modeling after your article's recommendations:

    1) Install 5 1/2" high-density batts in the 2x6 space; filling the cavity. Then covering the rafters with plastic just before putting drywall up (this was also suggested by an insulation installer that visited today).
    2) Cut-and-cobble with spray foam to to fill the gaps between boards and rafters.
    3) Fill cavity with dense pack (e.g. a bib system)...this was the preferred recommendation of today's insulation installer.

    I get the impression that cut-and-cobble would actually leave an airspace with 2x6 rafters (we have access to 3" PolyIso boards at $10/sheet). We understand its extremely labor intensive, but at this point its cost and damage-control for us.

    There's another space where we have 2x10 rafters, still under a 1/12 flat roof. Today's insulation installer also suggested using 8 1/4" high-density batts (R30), plastic, drywall; and leaving venting/soffits open. With what I'm learning these past few days, I'm not sure how I feel about that recommendation.

    Our goal is to not tear off the roof surface, again...if at all possible.

  49. User avater GBA Editor
    Martin Holladay | | #49

    Response to Gregg Berkholtz
    Gregg,
    You're getting closer to understanding a good roof assembly -- you are almost there. There is one remaining problem: you definitely don't want interior polyethylene with this type of roof assembly. An interior vapor barrier is a no-no.

    While vented roof assemblies are designed to dry to the exterior, unvented roof assemblies are designed to dry to the interior. Polyethylene would interfere with the assembly's ability to dry to the interior.

    The theory behind wall and roof assemblies with exterior rigid foam is explained in this article: Calculating the Minimum Thickness of Rigid Foam Sheathing.

    On the underside of your roof assembly, it's always a good idea to pay attention to airtightness. You want your drywall layer to be airtight, but not vapor-tight. To make sure that the assembly is airtight, don't install any recessed can lights, and make sure that you seal carefully at all wiring penetrations and plumbing vent penetrations.

    There are lots of articles on the GBA site that delve deeper into these issues, including:

    Questions and Answers About Air Barriers

    Vapor Retarders and Vapor Barriers

    Forget Vapor Diffusion — Stop the Air Leaks!

    Do I Need a Vapor Retarder?

  50. Gregg Berkholtz | | #50

    Response to Martin Holladay
    I thought an interior vapor barrier might be a bad option, especially with a TPO roof membrane; where would any moisture escape? Although the plastic sheeting was recommended by yesterday's installation installer, so thank you very much for reinforcing my thoughts on this.

    In addition, my understanding of the bib system they recommended also sounds like it'd be a problem; their design would use tyvek to hold blow-in cellulose. What's strange is that our installer said he "cleared it with his technical/legal department" while we chatted (he was exchanging photos and TXT messages). They sounded so convincing...

    For our situation, to get something better than the ~R12 from existing 2" PolyIso above the roof deck, it sounds like packing the space with un-faced high-density insulation might just be our least-cost option, as long as we're careful to seal all top-plate/blocking edges (and any interior or top/bottom-plate holes) with expanding spray foam.

    I'll dig into your links this afternoon - the information is much appreciated!

    [Editor's note: To read the answer to this comment, and to read subsequent comments, advance to page 2 by clicking on the number 2 below.]

  51. User avater GBA Editor
    Martin Holladay | | #51

    Response to Gregg Berkholtz
    Gregg,
    Your most recent comments show a new misunderstanding. Unlike polyethylene, Tyvek is not a vapor barrier. Tyvek is vapor-permeable, and won't cause any problems if you install it on the interior side of your roof assembly.

    However, Tyvek isn't a good choice for holding cellulose in place when dense-packing cellulose between framing members -- because Tyvek is an air barrier. When an insulation contractor uses a blowing machine to pack insulation between your rafters, the blowing machine pushes the insulation with a huge amount of air, and that air has to go somewhere. That's why experienced cellulose contractors use an air-permeable membrane (for example, InsulWeb) as netting to hold in the cellulose. For more information on this topic, see How to Install Cellulose Insulation.

    I'm beginning to wonder whether you have selected an experienced contractor to help you, or whether you are getting advice from someone who is in over his head.

  52. Gregg Berkholtz | | #52

    Response to Martin Holladay
    You're right, of course, about the Tyvek - I don't know what I was thinking when typing that part out. After all, it's why we wrapped the exterior walls with that stuff...heck, my wallet is made of Tyvek; sweat still gets to the paper on hot days.

    Your point about trapped-air during the blowing process is an excellent one. In hindsight, it's an obvious question; where is all that air supposed to go during installation - although I didn't put two&two together until your feedback. Thank you for taking the time to reframe that. This is now a screening criteria as we sort through contractor proposals.

    As for finding a qualified insulation specialist, it seems we could summarize desirable professional qualifications (at least for my needs/goals), as someone whom has training/experience specific to "unvented low-sloped 1 over 12 TPO roof".

    It's too bad you're not in the Portland Oregon area - I'd be arranging an appointment with you ASAP.

  53. Gregg Berkholtz | | #53

    Rigid PolyISO board is a vapor barrier?
    Gosh, it's difficult to find a qualified local installer...

    Just met with another installer, he insisted that 3.25" of R-MAX R21 rigid board installed via cut&cobble (e.g. with all edges sealed by expanding foam) would be our best option to insulate rafter space under the roof deck. He was confident we'd have sufficient vapor control with just the ~2" gap between roof deck underside and cut&cobble rigid board. This advice, specifically engineering a gap between roof deck underside and cut & cobble rigid board, appears to violate code (IRC R806.5). This consult was from a local contractor insulation supply house (fairly large company; 6 regional locations...apparently made it past legal...) - bid was for materials only...

    Tonight, I came across your article discussing how rigid board acts as a vapor barrier: https://www.greenbuildingadvisor.com/community/forum/general-questions/21111/using-polyisocyanurate-insulation-exterior-home

    This leads me to conclude that rigid board's vapor barrier properties effectively disqualifies cut&cobble for a TPO coated low-slope roof assembly, especially when there's 2" of rigid board already above the deck sheathing.

    Looks like our best strategy is to totally scratch the cut&cobble method, and proceed with packing a 5.5" rafter space with air-permeable/unfaced insulation just prior to drywalling. Our main options seem to be:
    1) 5.5" high-density R21 batts.
    2) 6.5" R22 batts (e.g. to counter insulation settling effects?).
    3) dense-pack blow-in insulation.

    As for the rest of your points, there's a clear need to be mindful of:
    1) Never use a "vapor barrier" paint on unvented roof/ceiling assemblies.
    2) Use closed-cell expanding spray foam before batt installations to ensure air-tightness at all wall and top-plate penetrations into the ceiling (e.g. electrical wiring & plumbing vents). Also ensure air-tightness around any ceiling penetrations (such as code-required single light/junction box per room...maybe a switched outlet is a better option...).

    As for meeting insulation code, R33 (2" Poly + R21 batts) certainly exceeds the requirements for a 1950's structure remodel, so I believe we're doing great there.

  54. Adrian Nicolici | | #54

    1963 Low Slope Roof & Silver Coat
    Hello Mr. Holladay, thank you for answering questions and keeping this thread active. We recently bought a 1963 home in the Pacific Northwest that has a low slope roof. It currently has a torch down roof that appears to be fiberglass based (you can easily see the fibers throughout the rolled-out sheets). My roofer recently installed a vent for our range hood fan and while he was up there I asked him about the condition of the roof. He said it was decent but that it was getting old and would need replacement in a couple of years. As a way of extending its life, possibly for another 5 years or so, he suggested silver coating the roof to keep the temperature and UV down as well as to seal any current small cracks/etc.

    Previous to this, the inspector we used during the purchase of the house noted after peeking into what areas of the tiny attics were accessible (parts of the house have no attic due to vaulted ceilings) that the condition of the trusses and sheathing were good for a house that age and he noted no excess moisture or any mold. I have since verified this above the kitchen after installing some can lights and additional insulation (only 2" rock wool was originally installed) in the small attic there as well as in the adjacent family room w/ a vaulted ceiling. The roof is vented throughout - a continuous saw-fit vent runs down the length of the low ends of the roof and this connects to each of the cavities in the vaulted ceiling areas as well as to the small attics that exist. I'm not entirely certain about the vaulted areas but in the attics there are a few mushroom vents near the top of the roof. I am not sure how air escapes the vaulted ceiling areas as there don't seem to be vents in those areas... (There may be a cavity common to those areas at the peak of the roof and above the glue-lam beam that runs down the spine of the house on the inside and this may connect to the attic spaces...)

    My question has to do with the silver coat and your mention of cold roofs causing moisture buildup. If I apply a silver coat to this roof to extend its life, I am worried that its temperature would drop dramatically, causing condensation from this point forward in time, leading to mold and structural damage due to moisture buildup. Would you agree with this concern or do you think the venting is probably sufficient?

    Thank you.

  55. User avater GBA Editor
    Martin Holladay | | #55

    Response to Adrian Nicolici
    Adrian,
    I don't have a clear understanding of your ceiling insulation.

    Since you tell me that you have a cramped but vented attic, I assume that the insulation is against the ceiling, with air above the insulation, and that there is no insulation against the roof sheathing. Is that correct?

    You wrote that the house originally had 2 inches of rockwool -- again, I'm assuming that the insulation was installed on top of the ceiling -- and that you have installed "additional insulation." What type of insulation did you add? Where was it installed?

    It sounds like you have an imperfectly vented low-slope roof with fibrous insulation above the ceiling. If you install a reflective coating to make your sheathing colder, the sheathing could certainly begin to rot.

    Rather than investing in a coating to lower the temperature of your roofing and your roof sheathing, you would be better off sealing your roof vents and installing rigid insulation above the roof sheathing, followed by new roofing.

  56. Adrian Nicolici | | #56

    1963 Low Slope Roof & Silver Coat
    > no insulation against the roof sheathing. Is that correct?

    Correct.

    >again, I'm assuming that the insulation was installed on top of the ceiling -- and that you have installed "additional insulation." What type of insulation did you add? Where was it installed?

    Correct - on the ceiling. Additional insulation was more rock wool on top of the old, still on the ceiling. but only in areas with an attic.

    >If you install a reflective coating to make your sheathing colder, the sheathing could certainly begin to rot.

    Yes, I think you are right. I will stay away from this option.

    >Rather than investing in a coating to lower the temperature of your roofing and your roof sheathing, you would be better off sealing your roof vents and installing rigid insulation above the roof sheathing, followed by new roofing.

    This does seem to be the frequent fix from what I've read and its on my list when the roof will eventually need to be replaced. For now, to extend the life of my current roof, I think I will simply coat with something like Gaco silicone coating which comes in dark colors and will not reflect the heat away from the roof but still seal nicely and protect against UV. A cheap short term fix:

    http://www.gacoretail.com/gacoroof.html

    How much insulation would I need to install on top of the existing roof after sealing the vents in order to keep the previously vented spaces from getting too cold and condensing moisture? I would like to minimize the affect on the appearance of the house... Does regular latex paint inhibit moisture transfer back into the house? Is sealing the ceiling (no pun intended) very important after such a modification?

    Thanks.

  57. Mike S | | #57

    Spray foam Roof in Philadelphia
    Martin,

    I am getting bids for spray foam for an unvented low slope roof in Philadelphia (zone 4). Most of the SPF contractors I speak with are pushing 10 inches of open-cell to reach R38 in the roof assembly. I was leaning towards doing around 2.5 in of closed-cell for an R15 and doing R30 in fiberglass batts on top.

    Which assembly would you prefer using? Is there a benefit to one or the other?

  58. User avater GBA Editor
    Martin Holladay | | #58

    Response to Mike S
    Mike,
    I'm not quite sure why you are proposing "2.5 in of closed-cell for an R15 and doing R30 in fiberglass batts on top." I hope that was a typo. When you combine closed-cell spray foam and fiberglass batts, the fiberglass batts go on the bottom, not the top -- and the two types of insulation have to be in direct contact.

    Assuming that you get your spray foam/fiberglass sandwich in the right order, either of the two approaches you listed will work. However, the best approach always includes rigid foam above the roof sheathing -- since that approach addresses thermal bridging through the rafters.

  59. Liam Knute | | #59

    for the unemployed, indebted homeowner with a flat roof...
    ... We have a choice between zero insulation (the status quo) or something, albeit imperfect. My mid-century modern will be torn down whenever I sell - maybe in 10 years? Meanwhile, I had to gut the interior due to a leaky roof and very bad mold problem. With the drywall and ceiling off, now is the time to add some insulation. It's near Everett, WA.

    Its a 1:16 low-slope roof (nearly flat), with newish black torchdown on top, a base sheet, and shiplap roof decking. The rafters (joists, trusses) are only 2x8s (so only 7.25" vertical gap). Furring strips run perpendicular below the rafters.

    I cant afford exterior insulation and a new roof. My credit cards are maxed out. I'm doing all the work myself.

    I'd welcome thoughts on my ceiling "plan", from top down, below the decking:

    3.75" air gap below roof decking (best I can do).
    Attic foil, permeable, between rafters, draped loosely, resting on...
    3.5"Roxul R15 ComfortBatts (atop furring strips).
    3/4" air gap due to furring strips
    2" RMax polyiso rigid foam , foilbacked, taped seams, R14 (screwed into furring strips)
    1/2" gypsum (screwed through foam into rafters w 4" screws)

    It has a 25x30' main living-dining-kitchen area, 3' soffits. No ceiling vents between gypsum and roof, only grill-vents in soffits. (I can't afford new holes in roof and added risk of leaks.)

    Any suggestions? Things to watch out for? Better solutions for the same money? Thank you!

  60. User avater GBA Editor
    Martin Holladay | | #60

    Response to Liam Knute
    Liam,
    Some comments:

    1. Working from the underside, it's going to be hard to drape the "attic foil" on top of the Roxul batts.

    2. I assume that the product you describe as "attic foil" is a radiant barrier that has been perforated with tiny holes in an attempt to address the problem that it is a wrong-side vapor barrier. I don't think that these perforated foil products are as vapor-permeable as the manufacturers pretend. If you want a vapor-permeable air barrier at this location, install a plastic housewrap like Tyvek.

    3. Although you plan to include an air gap between the Roxul and the roof sheathing, you don't mention whether there will be any ventilation openings to connect this air gap with the outdoors.

    4. The air gap between the bottom of the Roxul batts and the rigid foam is potentially problematic, because it allows air movement. It would be better if you could remove the existing furring strips so that the rigid foam can be directly under the Roxul batts. Then you can install the furring strips underneath the rigid foam, where they belong.

  61. Liam Knute | | #61

    Thanks for your reply. Very much appreciated!
    Many thanks for your reply!

    I'm wondering if it's even worth doing the rigid-foam at all (for now). Maybe I should just do only the 3.5" R14 Roxul for now. Not up to code (R30), of course, and not as energy-efficient, but up until now (for 50 years) there was essentially no insulation in the ceiling at all (just 3/4" acoustic ceiling tiles), so R14 would be better than before. And the house will be a tear-down for sure, whenever I sell it. Not sure if it's worth the time and money to do the rigid foam, I'll have to think about that.

    (I thought of using 5.5" Roxul, for higher R-value, leaving only a 1.75" air ventilation gap on top -- but even with a house that isn't too air-tight that seemed to go against what I've read.)

    Correct on #2, the 'attic foil' is a perforated radiant barrier. The goal would be to keep out some of the summer heat coming down from the black roof. (Not a big deal, but I think I could unroll it loosely on top of the furring strips, then stuff the Roxul batts up under it. Tyvek would work for vapor, but no radiant heat effect -- but I'm not sure the foil would help much, anyway.) Previously, the original vapor barrier was a paper/foil, laid with the foil side down!, nailed to the rafter bottoms, above the furring strips. (The wrong way, I think: if anything, it brought down heat into the house.)

    On #3, the only 'formal' external ventilation opening to the outdoors currently is a few (3? I haven't counted) 3"x8" rectangular vent openings in the soffits. I could add more of these? Or some in the fascia boards? (I'm trying to avoid poking new holes in the roof if I can.) The soffits and house in general are not particularly air-tight (board-and-batten exterior, ship-lap sheath/siding, no Tyvek yet, etc.), although I am slowly working with Great Stuff and caulking to seal up major cracks, but there's a long way to go (I'm not even going to try and seal the board-and-batten, so that's major 'ventilation' -- and heat-loss -- right there.)

    For the walls, given that the drywall is already removed, my intention given time and money constraints (correct me if I'm misguided!) is to staple Tyvek around each of the studs and to the interior side of the shiplap sheathing (which is covered on the exterior by cedar board-and-batten siding outside), before putting 3.5" Roxul in the walls. A nuisance, but I don't have the money or manpower right now to remove the cedar siding and put the Tyvek outside the sheathing (where it belongs).

    Just to enlighten me, if I do go with both Roxul and foam for the ceiling, what's the concern in point #4 about the 3/4 air gap (furring strips) between Roxul and rigid-foam? I'm guessing it's that with (relatively) open air movement, the insulation value is cut way down (unless it were airtight which it won't be)? That's one reason I'm thinking maybe don't bother with the rigid-foam for now (save the money until I can do an external/second-roof solution, maybe in 5 years).

    I wasn't sure if the Roxul would stay in place in the ceiling rafters, without the furring strips to hold it up, but I just tried and it does seem to work! Friction trumps gravity. :-) (I tried to attach a pic that shows a Roxul batt in place above the furring strips -- it seems to hold in place even if it is pushed well up above them.) Good to know.

    Again, thank you very much for your insights. (I know there's a lot of higher-end and maybe more interesting projects out there. Making-do is where I'm at, and I greatly appreciate your thoughts on this old house.)

  62. User avater GBA Editor
    Martin Holladay | | #62

    Response to Liam Knute
    Liam,
    Q. "The original vapor barrier was a paper/foil, laid with the foil side down!, nailed to the rafter bottoms, above the furring strips. (The wrong way, I think)."

    A. You're wrong. If a builder includes an air space with a foil radiant barrier facing the air space, then the radiant barrier will raise the R-value of the air space, regardless of which side of the air space the foil is installed on.

    Q. "The only formal external ventilation opening to the outdoors currently is a few (3? I haven't counted) 3"x8" rectangular vent openings in the soffits. I could add more of these."

    A. Your plan does not comply with the recommendations in my article. These openings will not provide the ventilation you seek. If you proceed with your plan, your roof sheathing boards are at risk of moisture accumulation and possible rot.

    Q. "What's the concern in point #4 about the 3/4 air gap (furring strips) between Roxul and rigid-foam? I'm guessing it's that with (relatively) open air movement, the insulation value is cut way down (unless it were airtight which it won't be)?"

    A. Yes, that's what I am worried about. You are setting yourself up for convective air currents, or for the creation of air pathways that allow interior air to find cracks in your wall assembly or roof assembly.

  63. Liam Knute | | #63

    Thanks again! (Update: turns out chimney chase has 2 vents!)
    Sorry for the slow reply. I mainly want to reiterate my thanks for your reply (and the broader service you provide to so many people, on so many topics).

    In case this is of assistance to anyone who stumbles across this thread:

    On closer inspection, it turns out my roof *does* have something of a doghouse/cupola vent, embedded in the side of the 6' wide chimney chase (above the roof)! Never having looked at them closely, I'd assumed they served some chimney function. But in fact, as far as I can tell, they are there to vent out the dead air space between the roof and ceiling. The vents open down into the area where those joists rest on the edge of the chimney (as revealed by shining a light at night).

    This 1958 house was ahead of its time in some ways (I think built off of Popular Mechanics or similar designs) -- e.g., radiant heat in copper pipes embedded in concrete floor, still functional today (recently pressure-tested).

    Thanks for clarifying the foil radiant barrier.

    I wonder if the ship-lap roof deck dries out faster than does a modern plywood/OSB deck?

    The black torchdown (on a base sheet) also allows some vapor to pass through, unlike modern PVC membranes.

    Due to time and financial constraints, for now I'm forced to make a compromise and omit the rigid foam. (N.B. this is not recommended, I know.) So, only 3.5" R15 Roxul* for now, and hoping the 3.75" air gap will vent enough moisture until I can afford to do it right (new roof, rigid foam on top). Not the best, but perhaps better than zero insulation -- provided there is adequate ventilation in the joist area.

    (*Canadian 3.5" Roxul is rated R14, in the USA the same dimension Roxul is rated R15. Not sure if there are any differences in product, or if the difference is in the rating systems.)

    Thanks again, and best wishes for the holidays.

  64. User avater GBA Editor
    Martin Holladay | | #64

    Response to Liam Knute
    Liam,
    Q. "I wonder if the shiplap roof deck dries out faster than does a modern plywood/OSB deck?"

    A. Yes, it probably does. It will certainly be better able to withstand a few wetting/drying cycles than OSB.

  65. Insulating Dilemma | | #65

    I know it's unconventional and not ideal...
    My house is a mid-century modern that seems to have done well for the previous 50+ years, but due to the severe winter we are having here - just North of Boston maybe 3 miles from the coast (so Zone 5 or 4 Marine?) we encountered an ice dam that backed up to a leak for the first time (during our tenure here).
    Obviously I don't want to encounter this problem again, but it got me thinking not only about the cause, but how to reduce our very high heating bills.
    Last year we put R-30 fiberglass rolls over anything existing in the area of the attic we can easily access...basically the middle of the house.
    The 3/4 of the house which is not cathedral ceiling has a rise of I'd say 1 foot to every 4 feet (1/4?) but 10 feet from either end there is a structural beam which pretty much blocks off those last 10 feet from normal size human access, making it pretty much impossible to get in there to insulate.
    Based on the icicle pattern outside around the time the leak occurred I suspect that the supplemental heater we had running in one of the cooler end rooms contributed to the melting causing the ice dam (it was above said cool room). So we stopped running it for the time being.
    The attic is unconditioned with hatch access and contains A/C unit and ducts. The original existing insulation is/was insufficient and typical of the time - fiberglass not extending beyond the tops of the joists. There is a ridge vent and soffit vents (although Springtime will now merit checking these out to see if the amount that appears to be there is actually there and adequate). There is also a powered vent at the top that can be switched on/off. Roof is newer with a single layer of asphalt
    shingles.
    I would blow cellulose into those areas, but am very concerned about accidentally shooting it right over the soffit vents and blocking them since access is so bad.
    Then I considered sliding polystyrene boards back there as best we could (thinking something is better than nothing) but saw a post somewhere indicating this could be a bad idea because of the air gap between the 2 insulations that would allow condensation to occur on the bottom of the poly board. They suggested, though, that plywood would be acceptable since it is more absorbent than the poly. So, would using, say, luan to slide more fiberglass into place back there work?
    I am trying to think outside the box a little due to the constraints we face, but do want to do right by the house. I don't want to cause a problem trying to fix one and I am not looking for a major project - interior (ripping out ceilings, etc) or exterior (decent existing roof). But maybe there is something to be done when the time does come for a new roof? Given that this was not really in the budget at all, and I'm not looking for a major project, what, if anything, can you suggest about how to proceed with this area?

  66. User avater GBA Editor
    Martin Holladay | | #66

    Response to Insulating Dilemma
    I.D.,
    When it comes to ice dam problems, halfway measures don't work very well. If this attic has ductwork and an air handler, and if enough snow is melting to cause leaks, it's time to seal your vents and to install thick insulation above your roof sheathing. It won't be cheap, but that's what you have to do.

    During the roofing work, have your roofing contractor remove 4 or 5 sheets of plywood or OSB so you can enter the attic from above. Have an air sealing contractor waiting in the wings. You want to seal all of your air leaks; then you want to re-install the roof sheathing, put down about 6 inches of rigid foam, a second layer of roof sheathing, and new roofing.

  67. Insulating Dilemma | | #67

    Halfway measures
    Martin,
    Thank you very much for your quick response. I will absolutely take that under advisement. I believe that having had this experience we are now better prepared (alert) to avoid the ice dams in the future. It will be a matter of keeping the snow off the edge area and watching the total amount of snow up there. We had it collect to quite a depth - I'd say maybe 3 feet in one area and no less than 2 every where. The insulating properties of the snow itself probably contributed to the melting in the area above the heater. Very unusual weather making this scenario 'possible.' It is in trying to reduce the heating costs that I am even considering what might work for "halfway measures."

  68. Michael Quinn | | #68

    low slope roof with insufficient tapered polyiso
    Martin- I'm in St. Louis , zone 4a. The builder of my historic replica home with a very low slope roof put tapered polyiso rigid foam under a tpo membrane - sloping from 6.75" thick in the front of the house - to a half inch thick at the rear of the house. So the rear half of the 25' x48' foot deep house has less than the r-15 insulation required by code - above the sheathing. Thus allowing a cold sheathing on that half - and I assume - eventual condensation. The roof sheathing is plywood and I have 16" deep truss bays under the sheathing. The builder wants to put fiberglass bats between the truss bays to get to r 38 total. I now have insulated hvac ducts running through the truss bays - and hvac metal trunk lines running thru a dropped soffit in three rooms - which would still beopen to the truss bay "attic" above. My architect says The code reads that I need to put 2" of closed cell spray foam on thebottom of the sheathing and then i can put the bats as well. I have read dozens of pages of articles and posts and at a loss as to the best way forward. Ripping off the newly installed tpo roof to add additional polyiso on top would cost thousands acording to the roofer. So, Should I create the foam sandwich with sheathing between polyiso above and closed cell foam underneath - for the rear half of the house -- where the polyiso is not thick enough? Or should I use open cell underneath? also should i add additional fiberglass batts between the trusses - and should i worry about covering the can lights to prevent moisture from rising into that space? and whats the best way to cover them? (or maybe I should pay more to replace them with flush mount lights.)

  69. D Dorsett | | #69

    Several ways to skin that cat.
    If the roof is not shaded and not pitched northward it may do just fine as-is, since the solar heating increases the average temp of the roof deck. It's usually north facing and shaded roofs that end up with problems.

    As little as 1" of closed cell foam is sufficient protection for the roof deck in the thin spots, since it is roughly 1- perm vapor permeneance, a minimal class-II vapor retarder, and provides a non-wicking first condensing surface. At 2" you'd be at about 0.5 perms, which is still sufficient drying capacity, and even less wintertime moisture uptake.

    The IRC 2012 chapter 8 prescriptive for R15 on the exterior of the fiber insulation based on an R49 total-R for zone 4A, not R38. It's the RATIO of exterior R to the that determines the average temp at the roof deck, which is what determines how much moisture will accumulate. If your total R is over R49 you have to increase the foam-R, but conversely, if it's less, you can safely go with less. So the prescriptive level is really only 30% of the total R. At R38 it means that you would only need R11.4 at the foam/fiber boundary for dew-point control with R26.6 of fiber. So with R6 polyiso and 1" of ccSPF (R6+) on the underside of roof deck you would have adequate dew point control for R28 batts under the roof deck in that section.

    A bigger issue is the air-tightness between the batts and the interior space. IRC chapter 8 prescriptives assume that there is moderately air-tight class-III vapor retarder between the air permeable fiber and the conditioned space, which is going to be tough to do can-lights and duct penetrations using gypsum board. It's worth installing a broad-sheet air tight membrane type smart vapor retarder such as Intello or MemBrain tight to the batts, above the ducts etc. Any penetrations of the membrane need to be carefully air sealed, and if the can-lights penetrate it use only air-tight fixtures. If conditioned space air (or duct leakage) convects through the fiber insulation it can move orders of magnitude more moisture to the roof deck than vapor diffusion through latex paint. If it's impossible to install the smart vapor retarder on the roof deck side of the ducts in an air-tight manner, it may be worth using open cell foam rather than fiberglass, even in locations where there is adequate foam above the roof deck.

  70. Michael Quinn | | #70

    response to D Dorsett #70
    Thank you so much for responding. I need to make decision asap on this (by this weekend!).
    The roof is not shaded at all and is pitched ever so slightly at 1/8" per foot - to the west.
    So if i wanted to go R49 total R, then i would need to increase the ccsf to R-14.7 which would be 2 inches thick? And could then put R34 of fiberglass batts between the trusses for R 49 total (since the polyiso at a half inch gives me virtually nothing on top - for the rear third of the house)

    Re: your 2nd paragraph/concern: The insulated hvac flex ducts run through the open webs of those trusses -so I don't see how I could possibly get anything solid across ABOVE them. Though I could attach something across the bottom of the trusses and soffits-- before the drywall ceiling gets put on under that. But I guess that wouldn't help? So you are suggesting I use open cell sf under the whole front half of the roof as well - in stead of Fiber - to reach my total R value of R49? And also attach the open cell layer to the closed cell layer in the back half of the roof too? Is there no other acceptable option? i was really trying to minimize the use of foam if possible - though I know I need to use some at least in the rear half. Thank you - I very much appreciate your input.

  71. User avater GBA Editor
    Martin Holladay | | #71

    Response to Michael Quinn
    Michael,
    Before I answer some of your questions, I want to provide general advice to GBA readers.

    Here's that advice: Don't be like Michael, and find yourself figuring out your insulation details at the last minute, after you have chosen your roof framing members, framed your roof, installed ductwork, specified and installed your above-sheathing rigid foam, and installed your roofing. At that point, you have painted yourself into a corner, and it's far too late of a time to scratch your head and try to figure out how to insulate your roof assembly. All insulation details need to be determined at the planning stage, before your foundation hole is dug.

    Michael,
    You have somewhat of a mess. Your ducts are in the wrong place. You forgot to come up with a plan to install an air barrier on the underside of your insulation. Your above-sheathing insulation layer is too thin.

    If your ducts are in the way, you may need to temporarily remove your ducts so that you have full access to the underside of your roof sheathing. Then you can call in the spray foam contractor to install enough spray foam to keep your roof sheathing out of the danger zone. Then you need to come up with a plan to install an air barrier, and to leave enough room between the air barrier and the cured spray foam for enough fluffy insulation to reach R-49.

    Once that's all done, you can re-install your ducts -- if they fit.

  72. Michael Quinn | | #72

    Response to Martin #72
    Yes I agree. i should have never trusted my builder and architect without digging into all of this further myself - months ago. My education has grown by volumes since reading articles on this site and others - only too late. (I would have never gone with the low slope roof if I knew what I know now.) How is it that virtually no one seems to understand these dynamics (I've talked to numerous builders, architects, and roofers locally)? Nevertheless, the duct work was placed above the ceiling on the 2nd floor after I decided to put a separate zoned furnace and a/c system on the 2nd floor (from advice from articles on this website) instead of the builder's proposal to put in one large HVAC system with long duct runs, electronic baffles and zones - in the basement - for all floors. So that "smart" decision, in turn, has totally complicated my roof insulation situation. Lesson learned the hard way. But where else could I have put the ducts for the 2nd floor zone- this was the only option I was given.

    In any case - the spray foam can be applied above the flex ducts without issue. I have 8" above those in the 16 " truss bays. What I don't understand is the air barrier. So you are saying that all of my ducts - supply and return air - should have been put in a soffit below my 16" high attic, by lowering the ceiling (which we already did in the 2 bathrooms and laundry/furnace room)? (and If i lowered the ceiling in other rooms I coud achieve this.) And then foam should have been applied first - and then drywall should have been attached to the bottom of those trusses to completely seal that attic space. And then the lights and hvac ducts go below in the soffit?

    At this point I am just looking at how to make the best of a bad situation.

    Alternatively I guess I could have the roofer install doghouses on the top of the roof and side vents in the walls above the fluffy insulation - to create a vented attic - but then I lose the insulatiing value of the polyiso on top and I still have the ducts to deal with?

  73. User avater GBA Editor
    Martin Holladay | | #73

    Response to Michael Quinn
    Michael,
    Only you can decide what to do. I certainly recommend that anyone who is building a house with ducts should:
    (a) Make sure that the ducts are installed inside the home's conditioned space; and
    (b) Plan where the ducts will go at the design stage.

    Moving ducts at this stage can certainly be tricky.

    It's hard to know the best way to install an air barrier under your fluffy insulation layer without making a site visit and looking at your roof assembly.

  74. Michael Quinn | | #74

    response to #74
    Thank you. Conditioning this attic space would not be an appropriate option in this case - via the use of foam under the sheathing, and a "loose" ceiling? I think I may be confusing the techniques.

  75. User avater
    Dana Dorsett | | #75

    This is tough without the visual...
    If the ducts are surrounded by fiber insulation a duct leak could insert a disasterous amount of moisture behind the air barrier & vapor retarder over the course of a winter. If encapsulated in open cell foam they really can't.

    If you don't have an air barrier fairly tight to the interior side of the fiberglass it's too risky. Installing the air-barrier on the interior side of the 16" trusses it leaves a huge bypass/convection channel behind the air barrier, potentially thwarting it's function if it ever develops an air leak.

    Filling the trusses full of fiber it would limit the convection rates due to the air-retardency of the fiber, but you'd be at R55-R60 below the roof deck, which would require proportionally more R above the roof deck.

    Have you painted yourself into a tight enough corner yet?

    With 2" of cc foam against the roof deck you have sufficient vapor retardency to fully protect the roof deck, then it becomes a matter of limiting the amount of moisture in the fiber. If you can move the ducts outside the trusses you can fill the rest up with cellulose blown in netting, and apply MemBrain detailed as an air barrier on the interior side of the trusses. It wouldn't meet the letter of code, but it would be pretty safe. The cellulose would redistribute and buffer any moisture that got by the MemBrain, but can release it more quickly than it takes it on. When the roof deck warms up in spring since the MemBrain becomes vapor open as the cellulose released it's adsorbed moisture into the cavity air on warmer hours of the day, raising the relative humidity to well over 50% for those hours. The high cavity-air humidity opens up the "smart" vapor retarder. When the cavity air moisture is re-adsorbed overnight in early spring the vapor retarder becomes vapor tight again taking on very little extra moisture from the conditioned space air. It dries nearly an order of magnitude faster than it takes on moisture under those conditions.

    This can work with air-tight can lights sealed tightly to the MemBrain, but isn't a great idea if the ducts are in the insulation layer. There are some pretty low-profile LED surface-mountable fixtures that might be easier & better than sticking with the recessed light approach.

    If the ducts can't move 2" of closed cell foam, against the entire roof deck, followed by 8-9" of open cell foam (installed in two lifts) gets you to code min in a less-risky fashion than using an air-permeable fiber. The open cell foam would load up with some moisture where it meets the closed cell foam over a winter, but if you keep it under 35% RH in the winter it won't load up too much. If you wanted to guarantee that it doesn't load up much, you could flip the stackup, and install the open cell directly against the roof deck with 2" of closed cell applied to the interior of the open cell, but the open cell foam surface won't be very smooth, and it's hard to get the thickness of the closed cell right if going that route- it's not really a great solution.

    There's no easy way out, but don't pull the trigger on any of it until you've really thought this out and priced it out. None of the options are cheap, and a rushed decision at this point could cost quite a lot to rectify later. Finding somewhere else for the ducts would be great if you can, since it simplifies the rest, making it more likely that it'll be done well.

  76. Michael Quinn | | #76

    Response to 76
    Thank you again Dana. I will definitely look at moving the insulated 6" flex supply ducts out of the truss bays but the uninsulated (so far) metal trunk and return ducts are in a soffit under/attached to the bottom of the trusses in 3 small rooms. Those would be much harder to move. What if I forgo the fiber bats altogether and just use the cc and oc foam to reach my r value across the whole roof (more in the back where the polyiso is lowest). ...And as u suggest in your first paragraph - encapsulate the ducts in the foam as well. I could also replace the can lights with flush lights.

  77. Michael Quinn | | #77

    Response to 76
    Ps I will post a pic

  78. User avater GBA Editor
    Martin Holladay | | #78

    Just use spray foam?
    Michael,
    Considering the way you have painted yourself into a corner with this roof assembly, just going to 100% spray foam may be the best solution.

  79. Michael Quinn | | #79

    Response to 79
    Agreed. Thank you both.

  80. Brian Burtch | | #80

    Builder Installed Insulation
    Martin,

    I wanted to get you to weigh in the situation we have in a new house where things are already closed up. We have a low sloped roof built with tapered trusses that vary in depth from probably about 14" to maybe 24" at the high end. The roof assembly is unvented. There is 2" of rigid insulation installed over the entire roof surface and the builder then installed 12" of r-38 batt insulation at the ceiling level (not the underside of the roof decking). It also appears that the side walls of the attic are not insulated above the 12" batt insulation line. The ceiling is penetrated by some can lights and a skylight. We are trying to determine if this condition is satisfactory for our Indiana climate and if it will cause any problems down the line. If not, we are trying to determine the best, most minimal approach to remedying the situation. Any input would be much appreciated.

  81. User avater GBA Editor
    Martin Holladay | | #81

    Response to Brian Burtch
    Brian,
    Your builder made a number of errors. It's up to your builder to fix these errors, at the builder's expense.

    If the builder wants to use a combination of rigid foam insulation above the roof sheathing and air-permeable insulation underneath the roof sheathing as part of an unvented approach, then the air-permeable insulation must be in direct contact with the roof sheathing. You can't leave a big air gap between the two insulation layers. This is a code requirement; according to the 2012 IRC, Section R806.5, "Air-impermeable and air-permeable insulation. The air-impermeable insulation shall be applied in direct contact with the underside of the structural roof sheathing as specified in Table R806.5 for condensation control."

    That was error #1.

    Error #2 was your builder's choice to install only 2 inches of rigid foam insulation above the roof sheathing. That is insufficient to keep the roof sheathing above the dew point in your climate zone. In Climate Zone 5, the building code requires this type of foam insulation to have a minimum R-value of R-20 (about 5 or 6 inches of EPS, 4 inches of XPS, or 3.5 inches of polyiso). This is a code requirement, so your builder's use of 2 inches of rigid foam is a code violation.

    Error #3 was the failure to insulate the attic walls.

    Errors #4 and #5 probably weren't code violations -- and in fact I'm speculating about them. But from a building science perspective (not a code perspective), the recessed can lights were a big mistake. (They are a cause of air leaks.) And I'm speculating that the skylight chase has insufficient wall insulation. (That's a guess.)

    The situation can be remedied from above or below. If you want to remedy the situation from above, the roofing can be removed and an adequate amount of rigid foam can be installed above the roof sheathing. That would be R-49 of rigid foam (about 12.5 or 14 inches of EPS, 10 inches of XPS, or about 8.5 inches of polyiso).

    If you want to remedy the situation from below, you'll need to demolish the ceiling and throw away the batts. Then install R-39 of spray foam against the underside of the roof sheathing.

    In either case -- whether you remedy the errors from above or below -- you'll need to insulate the attic walls.

    For more information on this topic, see How to Install Rigid Foam On Top of Roof Sheathing.

  82. Brian Burtch | | #82

    Builder Installed Insulation
    Martin,

    I was pretty sure this was the case, but just wanted to confirm...sounds like we have some major fixing to do. Now, I have another question. If we were to do another house in the future, it would be sufficient to do R-20 foam above the roof deck with the remaining r-value to be batt insulation fastened to the underside of the decking, correct? Thank you again for the help.

  83. User avater GBA Editor
    Martin Holladay | | #83

    Response to Brian Burtch
    Brian,
    Q. "If we were to do another house in the future, it would be sufficient to do R-20 foam above the roof deck with the remaining R-value to be batt insulation fastened to the underside of the decking, correct?"

    A. Yes. I would also recommend that you include a good air barrier directly below the fiberglass insulation -- something like gypsum drywall, Tyvek, or MemBrain.

  84. Brian Burtch | | #84

    Builder Installed Insulation
    The challenge with the air barrier is the ceiling would likely be constructed with tapered trusses to achieve the slope needed for the flat roof drainage. I'm not sure we would be able to get a continuous air barrier directly below the fiberglass insulation because of these truss members interfering.

  85. User avater GBA Editor
    Martin Holladay | | #85

    Response to Brian Burtch
    Brian,
    You wrote, "I'm not sure we would be able to get a continuous air barrier directly below the fiberglass insulation."

    All the more reason for choosing a better insulation system! Fiberglass batts have a lot of disadvantages.

    Ideally, the insulation details are determined at the planning stage, before any construction begins. It's important to specify trusses that are compatible with the type of insulation one prefers -- or, alternatively, to choose an insulation type that works well with the type of trusses one prefers.

  86. Laramie Hartmann | | #86

    Look before your leap
    I recently purchased a mid-century modern home with a leaking flat roof in Jacksonville, FL. My GC replaced it with a two layer product (Elastoflex SAV Base Sheet and Polyglass Modified Roofing). We discussed installing iso board insulation on top of the deck to insulate and add slope, but alas, this didn't happen, and I didn't realize he had pulled it from the proposal until after installation. Shame on me. The "attic" is vented around the entire exterior with soffit openings, but there is no doghouse to increase ventilation. Instead, the middle of the roof has a small hut atop that contains two air handlers and ductwork. I plan on air-sealing this hut and creating a conditioned space to reduce duct work sweating and increase general energy efficiency in the HVAC systems, but I'm unsure what to do with the remaining roof.

    Do I just live with poor insulation and keep the vented roof? I've pulled ceiling drywall down in several sections, and the original batt insulation if compressed, dirty, and covered with debris that the roofers dropped between gaps in the roofing deck. Is it worth cutting holes across the ceiling in order to blow in insulation? I'm afraid I may have no real options other than suffer the consequences of poor planning.

    Thanks.

  87. User avater GBA Editor
    Martin Holladay | | #87

    Response to Laramie Hartmann
    Laramie,
    All of the different approaches to insulating a low-slope roof are described in my article.

    Apparently, you have a poorly insulated roof. If you want to beef up the insulation layer, you need to follow the advice given in the article.

    If you want a vented roof assembly, you need (a) to make sure that there is room in your attic to install a thick layer of new insulation, plus a minimum of 6 inches of space between the top of the insulation and the underside of the roof sheathing, and (b) you will have to install a cupola or doghouse as described in the article.

    If you can't do that because of lack of attic space, you will need to seal the attic vents and install insulation above your existing roofing. That would be painful, of course, because your roofing is new.

  88. Steve Wilson | | #88

    Low slope design with SIPS
    Martin,
    I'm designing a home with a low slope .5/12 or 1/12 unvented roof with exposed timber or gluelam beams inside, and possibly a SIPS roof in snow country, Heber, Utah at about 6,000 ft. elevation. The roof material will likely be a white Sika Sarnafil membrane.
    Do I need any level of Polyiso above the SIP panel and below the Sarnafil roofing, or what would you suggest in this application?
    Thanks,
    Steve

  89. User avater GBA Editor
    Martin Holladay | | #89

    Response to Steve Wilson
    Steve,
    I don't have enough experience to answer your question definitively. If this were my roof, however, I would want to make sure that (a) the R-value of the SIPs at least meets minimum code requirements (R-49), and (b) that the SIP seams are carefully sealed on the interior with tape, as well as with spray foam, and (c) that there is a ventilation channel above the SIPs.

    I would consult with the SIP manufacturer to discuss the design of this roof.

    For more information on installing rigid foam above SIPs, see How to Make a SIP Roof Better.

  90. Steve Wilson | | #90

    ventilation channel above the SIPs.
    Martin,

    Thanks for your quick response. On your comment 'C' above, are you suggesting a 'cold roof' system, where furring strips are placed on top of the sips OSB and then another layer of OSB on the furring strips with the roofing material on top of that, thereofore leaving a channel where air can move freely?

    I will also get all the info from the SIPs manufacture I can as per your suggestion.

    Thanks again,
    Steve

  91. User avater GBA Editor
    Martin Holladay | | #91

    Response to Steve Wilson
    Steve,
    Yes, I am suggesting a cold roof. I've heard too many reports of SIP failures -- SIPs with rotten OSB at the seams, especially near the roof ridge -- to advocate installing a SIP roof in a cold climate without ventilation channels above the SIPs. The ventilation channels provide a margin of safety, reducing the chance of OSB rot.

  92. Trisha Megan | | #92

    Mushroom Vents
    Hi Martin. I hope I can pick your brain on this one...We have a row house in Chicago. The drawing indicate (it seems it was built as drawn) that we have a 1/4" per foot modified bitumen roof. We don't have soffits or ridges (just parapets on three sides and a gutter along the back). The roof structure is min 16" wood roof trusses at the low gutter end and they get bigger at the front. We have R49 Green Fiber blown in cellulose insulation in the trusses. There is no vapor barrier (as the insulation suggested) and there is two layers of 5/8" drywall at the ceiling. We have mushroom vents at 1 sq. ft. of net free area for every 300 sq. ft. of attic floor space. My questions are: 1) I've read online that we need 1sf vent for 150 sf of floor space but there is an exception if you vent ridges and soffits and have a vapor barrier it can be 1:300. It doesn't really address flat roofs. Do you typically see the 1:300 on flat roofs like the drawings indicate? and would adding more vents be better? 2) The parapets are so low (typically 6") that we can't add rigid and keep the min parapet height required by code and to make things worse I don't think zoning would allow us to add a doghouse if it was taller than the parapets because the building already had to get a zoning height exception. 3) If all we have are the mushroom vents, is there any requirements about where they should be located on a flat roof? The drawings didn't show their locations (only the net vent area required). Thanks in advance!

  93. Tim Miller | | #93

    Low Slope Venting
    In Michigan, the issue our firm has run into on commercial projects (R-2, type 5A multi-family) is in regard to section 1203.2 "Attic Spaces" of the 2012 IBC. Multiple jurisdictions have indicated that even though we are utilizing 4" or more of polyiso rigid insulation on the exterior side of the roof sheathing (low slope roof) the 24" open web/ parallel chord roof trusses form an attic space and they have been requiring us to vent the low slope application. What w have been doing is using intake and exhaust vents similar to the Air Vent Pop Vent and Aura Vent to provide the required intake and exhaust. It our professional opinion that when providing the insulation on the outside of the roof sheathing this is not required, however the AHJ seems to differ in opinion. What are your thoughts?

  94. User avater GBA Editor
    Martin Holladay | | #94

    Response to Tim Miller
    Tim,
    4 inches of polyiso has an R-value of R-20 to R-24, amounts that will work in southern Michigan -- but only if you install an additional interior layer of about R-29 of insulation on the interior side of the roof sheathing. You didn't tell us whether you are doing that. The interior insulation needs to be in direct contact with the underside of the roof sheathing.

    This type of roof is an unvented roof assembly. Obviously, it makes no sense to introduce cold exterior air on the interior side of your insulation layer -- so your local building department is wrong.

    You may also be wrong, however, if you have (a) forgotten to install any interior insulation, or (b) installed the insulation on the attic floor, with an air space between the top of the interior insulation and the roof sheathing.

    Note that the attic walls also need insulation.

  95. Kenneth O'Brien | | #95

    ducts in unvented roof
    I'm working on a 4 story wood frame building with low slope roof trusses. We are putting all the insulation on the plywood roof deck (rigid foam insulation). We are using an air-tight drywall approach for the ceiling air barrier on the underside of the trusses. I can't seem to find any info on any advantages/disadvantages to putting the ducts in the truss space vs. in soffits under the drywall ceiling. Have there been any studies into possible losses, inefficiencies in placing the ductwork in the truss space? I can't locate anything, except some vague cathedral ceiling graphic which seems to indicate about a 2% efficiency loss, but I don't know where that number came from. Thank you.

  96. Phil Mason | | #96

    Re-Roof insulation upgrade advice
    My late '60's flat roof, 2 story home is getting to the point of needing roof maintenance. It currently has an EDPM roof with no added external insulation as far as I know. There is no attic. As far as I can tell there is fiberglass insulation between the ceiling joists which I would guess are 2x8. The attic is "vented". This consists of some slots under the soffit. I do not know of anything else.

    I have 3 plastic bubble skylights. They are great for light but not much else. I would like to upgrade those as part of any repairs. Some fixed Velux ones seem a good option.

    During the humid part of Indiana summer (zone 4/5) it "rains" in the bathroom on the second floor and in the middle of the house. Obviously reaching dew point. We live in a woods and do not have to use AC much. Have not seen other issues.

    First off, I would seal off the soffit vents. Then I am looking at 2 options.
    1. adding 4 inches of Polyiso outside and going back with some kind of rubber roof
    2. Using SPF to get both insulation and re-roof.

    I appreciate any relevant questions or considerations I need to take into account. I figure now is the time to start shopping. Thanks in advance for your expertise.

  97. User avater
    Dana Dorsett | | #97

    The polyiso solution is greener.
    3lb spray polyurethane is blown with a high global warming potential HFC245fa (about 1000x CO2). Polyiso is blown with pentane (about 7x CO2).

    At 4" the polyiso would be in the R22-ish range. With R20+ above the roof deck it would be fine to fill the 2x8 bays full-up with blown fiber insulation and close off the venting, so long as there isn't a true vapor barrier on the interior side, such as 6 mil polyethylene.

  98. User avater GBA Editor
    Martin Holladay | | #98

    Response to Phil Mason (Comment #97)
    Phil,
    Either approach will work from a building science perspective. If you choose to install spray polyurethane foam, you are correct: the spray foam, if installed above the roof sheathing, can act as both insulation and roofing. (Of course, your spray foam contractor has to choose a roofing foam for this application, and the cured foam has to be protected from sunlight with a liquid-applied coating or a layer of small stones).

    Dana Dorsett is correct that, from an environmental and "green building" perspective, the polyiso option is preferable to the spray foam option.

  99. User avater GBA Editor
    Martin Holladay | | #99

    Response to Kenneth O'Brien (Comment #96)
    Kenneth,
    Q. "I can't seem to find any info on any advantages/disadvantages to putting the ducts in the truss space vs. in soffits under the drywall ceiling."

    A. If all of your insulation is above the roof sheathing, and if you have a roofing membrane above the insulation, then you already have an airtight insulated roof assembly. It's perfectly OK to put the ducts in the "attic" (truss space) if you want, since these ducts are inside the home's thermal envelope. You don't have to build interior soffits.

    Note that your drywall ceiling doesn't even have to be airtight. After all, your roof assembly is already airtight.

    Remember to insulate the "walls" of your attic -- in other words, the "rim joist area" of your roof trusses.

  100. Quinn Sievewright | | #100

    How far should the rigid foam extend past the wall elevation
    Hello,
    Our planned low slope (1:12) roof will have 2' overhangs on three sides and a 6' overhang on the front elevation. For ease of install we'd probably cover the entire 2' overhangs with the rigid foam, however is there any merit/need in covering the larger 6' overhang with foam as well?
    Or could one bring the insulation beyond the vertical elevation of the underlying exterior wall insulation (in our case R6 Roxol over sheathing) and level the roof with with 2x timber to the level of the adjacent rigid foam.
    This would essentially result in a hollow unvented and uninsulated space sandwiched between the soffit (Cedar T&G) and top roof sheathing (ply). The top roof sheathing will likely be covered in torch on or EPDM.
    Thanks very much for your thoughts.

    [Editor's note: To read the answer to this comment, and to read subsequent comments, advance to page 3 by clicking on the number 3 below.]

  101. User avater GBA Editor
    Martin Holladay | | #101

    Response to Quinn Sievewright
    Quinn,
    Q. "Could one bring the insulation beyond the vertical elevation of the underlying exterior wall insulation (in our case R-6 Roxul over sheathing) and level the roof with with 2x timber to the level of the adjacent rigid foam?"

    A. Yes, you could do that.

    -- Martin Holladay

  102. Quinn Sievewright | | #102

    Thanks
    Thanks Martin.

  103. Jon R | | #103

    Unvented question: I get that
    Unvented question: I get that adding the right amount of rigid foam over moisture susceptible sheathing keeps the sheathing warm and so minimizes or eliminates condensation at the sheathing (caused primarily by exfiltrating air). But then I see references to then adding a second layer of moisture susceptible sheathing immediately over the foam, with no ability to dry to the exterior (and very little to the interior). Doesn't this revert to the same issue (eg, cold, wet, rotting OSB)?

  104. User avater GBA Editor
    Martin Holladay | | #104

    Response to Jon R
    Jon,
    The upper layer of sheathing is not as vulnerable to accumulating moisture for a simple reason: it's almost impossible for warm, humid interior air to come in contact with this upper layer of sheathing.

    The lower layer of sheathing, on the other hand, is vulnerable, because warm humid interior air can contact it in the winter.

    That said, some builders (especially builders in climates where ice dams are frequent) like to install a ventilation channel directly below the top layer of sheathing. This approach certainly works: it helps keep the sheathing cold during the winter (reducing the chance of ice dams) and it allows the upper layer of sheathing to dry to the exterior if it ever gets damp.

    The usual method for creating such a ventilation channel is to install 2x4s on the flat, 16 inches o.c. or 24 inches o.c., laid from the soffit to the ridge (typically, with each 2x4 above a rafter). The 2x4s are installed above the rigid foam, below the top layer of roof sheathing. This creates a ventilation channel that is 1.5 inches high.

    -- Martin Holladay

  105. Jon R | | #105

    Thanks. So the assumption
    Thanks. So the assumption is that there is no exfiltrating air (which would bring interior air into contact with the cold upper OSB) and no fluttering roof membrane (pumping interior air in and out). Sounds reasonable (but perhaps best avoided anyway) in the case of a fully adhered membrane - sounds questionable in all other cases.

  106. User avater GBA Editor
    Martin Holladay | | #106

    Response to Jon R
    Jon R,
    If you follow the recommendations provided on the GBA site, you'll have an air barrier under the rigid foam, and you'll probably tape the seams of the rigid foam (or install the rigid foam in two layers with staggered seams).

    So, unless you are an unusually sloppy installer, fluttering roofing membranes aren't going to pull indoor air through the foam to contact the upper layer of roof sheathing. Nor will there be any exfiltrating air.

    Again and again, GBA articles emphasize: the key to a long-lasting insulated roof assembly is an airtight ceiling.

    -- Martin Holladay

  107. Michelle K | | #107

    We want to try the "somewhat" controversial approach
    We have a 1960s split level, just outside of Philadelphia. One of the attic spaces is 34’ long x 18’ wide. It has a low slope roof where the tallest part is 3.5 feet sloping down to about 4 or 5 inches. A good 8 feet is not accessible. It is ventilated: a ridge vent running the length, 2 gable vents and “fascia vents”—i.e. where the roof meets the fascia board there is a gap in some of the bays. We are in the process of air sealing the accessible parts. The roof was replaced 3 years ago. At the time, we did not have the benefit of the knowledge we have gained here at GBA, so there is no rigid foam on the roof sheathing or any of the other measures that would have been optimal. So, we are in a position to only do as good as we can. I read this article and we are hoping to try the "(somewhat) controversial approach"—to add dense-pack cellulose for the 8 feet that is not accessible, and then add loose-fill cellulose up to R49. My questions: Do we need to add ventilation baffles on the side that we dense-pack or is it okay if the cellulose is touching the roof sheathing? Do we need to worry about blocking the “fascia vents” with this method? The 2 gable vents will be mostly covered by insulation so I was thinking of boxing them out with insulation board to keep some of the ventilation and to minimize wind washing —will that work? Are there any tips or tricks or products for air sealing top plates and other intrusions in super tight spaces without going from the outside, or from the space below? I was thinking that a Dow FrothPak could work but I doubt that we could do a thorough job from 8 feet away. Thank you in advance.

  108. User avater GBA Editor
    Martin Holladay | | #108

    Response to Michelle K
    Michelle,
    Q. "Do we need to add ventilation baffles on the side that we dense-pack or is it okay if the cellulose is touching the roof sheathing?"

    A. The side that is dense-packed all the way up to the roof sheathing can't possible benefit from ventilation openings. If there are any ventilation openings on that side of your roof, they will be blocked -- either deliberately, by carefully air sealing them before the insulation truck arrives, or accidentally, once the attic is dense-packed.

    On the other side of the attic, as stated in my article, ventilation openings above the top of the insulation -- as many openings as possible -- are a good idea.

  109. user-7067800 | | #109

    A (somewhat) controversial approach
    Martin,
    We own an apartment building with a flat roof in Chicago (climate zone 5). We have a roof tear off planned this summer. We will also have the attic space insulated with R-49 cellulose. The insulation contractor plans to do exactly what you describe in your section "A (somewhat) controversial approach".
    We have a cool roof (we painted it ourselves), and we will replace it with a cool roof. In 17 years we've had 2 or 3 incidents of condensation resulting in leaks into the top floor.
    To prevent condensation in the future, we are considering adding insulation to the roof layers (currently on the table is 4 inches for R-23). We would then not vent the attic space. Does this plan sound reasonable? If not, how can we adjust it to make it reasonable? It's a lot of extra money, but if there's a good chance it would reduce condensation, we may bite the bullet. TIA

  110. User avater GBA Editor
    Martin Holladay | | #110

    Response to Judico
    Judico,
    Adding rigid foam above the roof sheathing will certainly move your roof assembly in the direction of safety.

    In Climate Zone 5, at least 41% of the total R-value of the roof assembly needs to come from the rigid foam layer for a safe installation. If the total R-value of the roof assembly is R-49, you would want R-20 rigid foam and R-29 of cellulose. That means you would have to remove some cellulose.

    If you leave the R-49 cellulose, you would need a lot of rigid foam to make the assembly work. You would need R-34 of rigid foam. That would give you a total roof R-value of R-83, with 41% of the total R-value coming from the rigid foam (since 34 is 41% of 83).

  111. Jun Piesse | | #111

    Thickness of top ply sheathing and XPS to rock wool seam
    Hello GBA,

    Thank you for this article. It and all the comments have been very useful.

    I'm building a warm deck over a living space. It is 3 feet deep and 30 feet wide running above the first floor on the southern edge of the house.

    -It has 1 inch ply sheathing over the joists with taped seams.
    -I plan to follow this with XPS, at a thickness suitable to my climate zone, with taped seams. (It was made without using the harmful blowing agents outlined on this website)
    - 3/8 ply sheathing screwed to the joists through the XPS
    - Glued down PCV single ply membrane with welded seams.

    I have two questions..

    Would 3/8 inch ply be suitable for light foot traffic- or should I use a thicker tongue and groove product.

    The XPS will overlap rock wool board used on the exterior of the walls. I can't tape this seam, however they will be compressed together by the (fascia) board screwed to the beams and top layer of deck sheathing nailed down to the fascia. Would this be okay? Or should XPS be used for the last bit of insulation at the junction where the wall meets the deck?

  112. User avater GBA Editor
    Martin Holladay | | #112

    Response to User-6877107
    User 687etc.,
    First of all, can you tell us your name?

    Membrane roofing is not a walking surface. If you want a walking surface, you'll need to install sleepers above the roofing, followed by some type of deck. You'll find articles on this type of deck with a little Googling.

    I think that 3/8 inch plywood is too thin, but you should contact the manufacturer of the roofing membrane to determine their requirements for a cover board.

    I don't understand how the XPS on your roof will "overlap" the mineral wool on your walls. I think you'll need to post a sketch to clarify your question.

  113. Jun Piesse | | #113

    Martin, Thank you kindly for
    Martin, Thank you kindly for your reply.

    I'm Paul. Sorry for leaving out some important information.

    The PVC is a 2mm single ply membrane that is specified for light foot traffic .

    The deck itself is essentially a warm roof over a the first floor. The second floor roof cantilevers over it, protecting it from most of the rain.

    So the manufacturer recommends a 9mm thick board made from a mix mineral fiber and volcanic sand. It functions much like ply, but doesn't burn as easily. Here, in Japan, some houses are sheathed with it. Locally made ply here is much cheaper and readily available. Would 2 layers of 9mm ply staggered protect the xps from compressing under foot traffic?

    By overlap, I mean where the XPS from the roof meets the supporting wall. It has exterior rock wool board. The vapour barrier above the roof sheathing is attached to the vapour control layer . On this website it seems that usually exterior insulating foam is taped at the seam where the walls meet the roof but taping XPS to rock wool board isn't possible. Would this cause any problems?

    I have attached an simple sketch of the roof. It doesn't include the smaller details such as drip edges.

    .

  114. User avater GBA Editor
    Martin Holladay | | #114

    Response to Paul (Comment #114)
    Paul,
    If your house is in Japan, I don't think I can be of any help to you. Building materials available in Japan may be totally different from materials available in the U.S.

    Concerning the coverboard that is required under your membrane roofing, you should follow the installation instructions provided by the roofing manufacturer.

    I don't see any particular problem with the way that your rigid foam overlaps the mineral wool.

  115. Jun Piesse | | #115

    Thank you Martin,
    I

    Thank you Martin,

    I understand your point about the building materials- many of the materials are different and some are the same- regardless of this, I very much enjoy reading and learning from information posted here.

  116. Rick Ramirez | | #116

    Venting roof without Cupola and relatively small depth
    Martin,

    Kind of a nuanced situation here. We have a family home in the Florida Keys (Zone 1) that had damage with the latest hurricane and one section of roof was redone in the old tar and gravel method. I know we should have insulated the roof deck but did not. The plaster below was completely destroyed and we are going to either use drywall or tongue-in-groove for the ceiling. The roof structure is deck boards above 3x6 beams. There is a 2x material used as facia all around this roof and there are not currently vents.

    The nuanced aspect is that this is a vacation home where the house is largely unconditioned - i would say 95% of the time there is no AC being used. The beams are not water damaged in any way from the roof leak nor are there signs of moisture issues (mold, etc). I'm wondering if it is worth it to add insulation. We could use mineral wool (typical 2x4 size) that would fill around half of the cavity and then add venting in each bay but I'm wondering if there is enough air flow across a flat space because cutting a cupola doesn't seem to make sense in a brand new flat roof. We haven't had problems with moisture damage as you can see but it does get extremely hot there and they are the bedrooms, so we are looking to help the space without causing damage.

    I know this is not an ideal scenario, but most remodeling scenarios are not. Any help in direction would be appreciated. -Rick

  117. User avater GBA Editor
    Martin Holladay | | #117

    Response to User-7034460
    User 703etc.,
    First of all, can you tell us your name?

    Insulation in Florida ceilings is a code requirement. Of course, many older homes aren't in compliance with building codes. But if you post a question on GBA, asking, "Should I install insulation in my uninsulated roof," I'm going to answer "yes." It's the right thing to do, for you as well as the next owner of the house.

    If you can't afford to insulate your roof assembly, and you choose to crank up the air conditioner when necessary instead of installing insulation, that's your choice (as long, of course, as your home is in compliance with local regulations).

    If you want to insulate this type of roof assembly -- which is too shallow to vent -- from the interior, your only choice is to install closed-cell spray foam on the underside of the roof sheathing.

  118. Rick Ramirez | | #118

    RE: Venting roof without Cupola and relatively small depth Read
    Hi Martin,

    My name is Rick. I updated my profile so hopefully that gets put on the posts.

    Your answer regarding insulation is precisely why I am here. I would like to make the house more comfortable without causing moisture issues. I simply was supplying pictures showing that we have had zero moisture damage with the existing unvented roof assembly even with a major roof leak. Also, I think you may have misinterpreted what I was saying regarding AC use. The fact of the matter is that efficiency and economy aside, the AC is simply not effective in the summer even if you crank it.

    As an aside, we just got finished reroofing an older cathedral ceiling on a different property where the roof assembly was not vented, where they used a vapor barrier, and where the roof deck was completely rotted because of this. I consulted your article here (https://www.greenbuildingadvisor.com/blogs/dept/musings/how-build-insulated-cathedral-ceiling) and used site-built baffles out of XPS. Anyways, this has made me concerned about not merely insulating a structure, but insulating it correctly.

    I will look into closed cell foam installers or look into DIY kits I have come across in the past. However, are there any other remedies that would work? What about using ridgid foam such as XPS (something like this: https://www.menards.com/main/building-materials/insulation/insulation-panels/owens-corning-reg-foamular-reg-extruded-polystyrene-insulation-4-x-4-x-8-r-20/p-1444450493765.htm)? I understand that R-30 is code but perhaps this could be something installed more easily - a 4" sheet with a 2" sheet would create R-30 (see https://dcpd6wotaa0mb.cloudfront.net/mdms/dms/EIS/43522/43522-FOAMULAR-250-XPS-Insulation-Product-Data-Sheet.pdf).

    Thank you for helping and providing a forum to discuss issues like this.

  119. User avater GBA Editor
    Martin Holladay | | #119

    Response to Rick Ramirez
    Rick,
    The articles on GBA are all consistent: If you want to insulate this type of roof from the interior, your only choice is closed-cell spray foam. Once you have at least a portion of the R-value installed as closed-cell spray foam, you can complete the job using the flash-and-batt method if you want.

    If you want to use rigid foam, the rigid foam must be installed on the exterior side of the roof sheathing, not the interior. That means you'll need a second layer of roof sheathing and new roofing.

    In case you missed the articles that explain this -- even though the article where you posted this comment explain these facts -- I'll post links to more articles so that you understand the consistency of GBA advice on this topic.

    How to Build an Insulated Cathedral Ceiling

    How to Install Rigid Foam On Top of Roof Sheathing

    Flash-and-Batt Insulation

  120. Jeff D | | #120

    Hi,
    I have a question regarding adding roof insulation to an older modular home in northern Canada, zone 8. This house currently has two roofs. The first roof is the original roof that came with the house. It has a very small slope and made of tin. There is currently batt insulation of unknown R value (put probably 10) between the ceiling and this first roof. The house underwent some remodelling at some point to give it a gable roof. I recently found out that this gable roof has no insulation and has small number of small soffit vents and nothing else. There is enough room between the two roofs to enter this space by cutting a small hole in the gable roof.
    Condensation builds up on the underside of the first roof. I want to add insulation to increase the R-value of the roof (to reduce heating costs) and also eliminate the condensation. Please see the drawing for an elevation schematic of my roofs.
    1. Where should the insulation be placed?
    2. Should I be looking to improve the ventilation of the second roof or consider the entire assembly as an unvented roof?
    Thanks

  121. User avater GBA Editor
    Martin Holladay | | #121

    Jeff,
    Since this is a low-slope roof, it is best treated as an unvented assembly, for the reasons provided in the article on this page.

    The best approach would be to cover the top side of the old roofing with closed-cell spray polyurethane foam -- as thick a layer as you can afford. If you take this advice, make sure to seal up any soffit vents with closed-cell spray foam as well.

  122. Jeff D | | #122

    Thanks Martin for your response. Can I cut and cobble the top side of the old roof (including the soffits) instead?
    If I understand the article well, I would avoid putting poly on the ceiling if I do some ceiling renovations to make sure the ceiling dries inwards, correct?
    Thanks again. It's very interesting to read these articles.

  123. User avater GBA Editor
    Martin Holladay | | #123

    Jeff,
    Your metal roofing -- it's probably steel roofing, not tin -- can be covered with rigid foam insulation if you want. Your use of the term "cut-and-cobble" surprised me -- that term is usually reserved for jobs where the insulation is installed between framing members. Then I realized that there may be framing members installed on top of the metal roofing. Is that the case?

    If you go that route, you need to make sure that the rigid foam is protected from squirrels and other rodents. Seal up any vents to keep out the critters.

    1. Jeff D | | #124

      You are correct Martin. The new roof was build on top of the old roof with 2x4 framing. I assume this was done to make the house look more "stick-built".
      XPS would work in this case because I want the roof insulation to be as air impermeable as possible, correct? It shouldn't be a problem to keep the rodents out since they currently have no way in. I discovered the roof construction not too long ago when I was removing the wood stove chimney after having installed a pellet stove.

  124. User avater GBA Editor
    Martin Holladay | | #125

    Jeff,
    Any of the three most common types of rigid foam -- EPS, polyiso, or XPS -- will be air-impermeable. Green builders try to avoid the use of XPS, however, because it is manufactured with a blowing agent that has a high global warming potential. For more information, see "Choosing Rigid Foam."

Log in or create an account to post a comment.

Related

Community

Recent Questions and Replies

  • |
  • |
  • |
  • |