Replies

1. 
   Dana1 | May 04, 2015 09:51pm | #1

   The pictures don't look anything like a complete fill with foam. It looks (being generous) closer to an ~8-10" average depth which BTW does't not meet code minimum in MA, with some spots as thin as 6".

   Even a full 11.25" (the nominal depth of a milled 2x12 rafter bay) of open cell foam would not meet the code minimum R49, and it's unlikely that your framing fraction is low enough to make it on a U-factor basis. To meet code on a U-factor basis it has to measure out with a whole-assembly average of U0.026 or less, which is essentially uniform R38 without the thermal bridging of the rafters. At the 8-10" average depth you'd be lucky if it performed at R38 average even center-cavity, let alone after calculating the thermal bridging.

   No matter what the above-deck layup is, unless it has about R20-R24 of rigid foam, the roof deck is at some risk of wintertime moisture accumulation. While vapor barrier latex on air-tight wallboard would work on paper, the long term air tightness is not guaranteed. A better solution would be to first bring the foam contractor back and fully fill the cavities, trimming it flush with the rafter edges. Then, install a layer of MemBrain detailed as an air barrier on the rafters, lapping & taping any seams between adjacent sheets ONLY at the rafters, and cauking the edges to the framing elsewhere. Then install 1.5" of UNfaced EPS (and not XPS) as a continuous layer on the underside of the rafters, and long-screw the ceiling gypsum through the foam to the rafters. That would bring you up to a center-cavity R of about R45-R46 or so, still shy of a code-min R49, but the R6 thermal break over the rafter edges brings the assembly reliably below U0.026.

   At 1.5" Type-II EPS has a vapor permeance of about 1.8-2 perms, which might be sufficiently protective of the roof deck on it's own, but in combination with the MemBrain the wintertime moisture accumulation at the roof deck will be extremely slow, but the drying rate of the roof deck will be only limited by the 2-perm EPS. If you paint the ceiling gypsum, use a standard latex paint, not vapor barrier paint. Don't use vapor barrier paints on the ceilings below, or on any of the partition walls, etc. That way even the most modest amount of air exchange between the attic & the rest of the house will keep the humidity levels in the attic space tracking that of the rest of the house.

2. 
   djhaus18 | May 05, 2015 12:17am | #2

   Since it appears this installation won't meet the minimum code, is that something that the town building inspector should flag?

   Also do you think there are other viable options to address this issue? I want to be prepared for what the builder and/or foam installer might suggest as correct or incorrect solutions. For example, would adding closed cell foam over the open cell to fill the gaps and rafter bays completely to a full 11.25 depth be another viable option? Perhaps in the areas behind the walls the rafter bays could be foamed beyond the 11.25 depth since no gypsum would be attached to the rafters

   Thanks again,
   -David

3. GBA Editor
   Martin Holladay | May 05, 2015 04:57am | #3

   David,
   The spray foam installer made at least two mistakes, and they are serious ones.

   1. As Dana explained, the installation does not meet minimum code requirements for total R-value.

   2. The kneewalls were built, and the ducts and HVAC equipment were installed, before the drywall and required vapor retarder were installed. This violates the installation requirements of most brands of open-cell spray foam, including Icynene. Installing the required drywall (or, better yet, rigid foam) now will be awkward, and may require the kneewalls and ductwork to be temporarily removed and later reinstalled.

4. 
   Dana1 | May 05, 2015 07:47am | #4

   As stated previously, a full fill of 11.25" of o.c. foam still leaves you about R10 shy of code min. You could go 3" deeper beyond the rafter depth and hit that (you'd make it in only 2" on a U-factor basis, if you could guarantee a full 2" of depth over the rafter edges) but it would leave you with only the most awkward methods of attaching the ceiling gypsum.

   If you complete the full fill with no sinks or voids above the plane of the rafter edges, insert a smart vapor retarder detailed as an air barrier, and add R6 semi-permeable rigid foam (the 1.5" of unfaced EPS), it makes code min on U-factor, yields a reasonably flat ceiling and doesn't require ridiculously long fasteners for the gypsum.

5. 
   djhaus18 | May 06, 2015 01:12am | #5

   I spoke with the foam installer. He informed me that because my town had adopted the 'Stretch Code' that the attic R-value required is only R-38. This doesn't quite seem right to me. The foam they used is R-3.9. He claimed the foam is 10" in or more in each bay and therefore the R value was met. He also indicated that gypsum with wallboard would cover the entire roof deck even behind the partition walls and that the gpysum would be primed and painted with a vapor retarding sealer/paint.

   I am considering having the builder increase the foam to completely fill each bay to get the full11.25 depth which would be about R-43. Would this combined with the gypsum and vapor retarding paint be enough to address potential moisture issues?

   Thanks,
   -David

   -dave

6. GBA Editor
   Martin Holladay | May 06, 2015 04:39am | #6

   David,
   Here is information on the Stretch Code: "The stretch code appendix offers a streamlined and cost-effective route to achieving approximately 20% better energy efficiency in new residential and commercial buildings than is required by the base energy code. This is largely achieved by moving to a performance-based code, where developers are required to design buildings so as to reduce energy use by a given percentage below base code, rather than being required to install specific efficiency measures."

   Here's a question for the insulation contractor: "Since you are skimping on attic insulation, other features of the house must be built to a very high standard to make up for the low R-values in the attic. What are these extraordinary other features?"

7. 
   djhaus18 | May 06, 2015 08:23am | #7

   Thanks Martin. I will speak to the builder about what else is being done. I guess my lingering concern still remains about moisture in the attic. Do we think the open cell foam at R-38 coupled with all of the rafters covered with gypsum and vapor retarding paint will be enough to prevent moisture on the roof deck?

8. GBA Editor
   Martin Holladay | May 06, 2015 09:03am | #8

   David,
   If the open-cell spray foam is upgraded to R-38, and if the contractor is really able to install drywall on the underside of the rafters -- a feat that will be difficult behind the kneewalls, where the ducts are -- and if vapor-retarder paint is applied to the drywall, you won't have any moisture problems.

9. 
   Dana1 | May 06, 2015 09:07am | #9

   For a vapor retarder to protect the roof deck it has to be completely air-tight. That's damned hard to achieve with an air-tight gypsum approach given where you're starting, and if the rafter bays are not completely filled you have large convection channels to amplify any air leakage. Air transported moisture is a HUGE risk here, even if you installed a broad sheet membrane vapor retarder like Intello or MemBrain.

   Even if they have the heat-loss math to prove that they're 20% better than code for the whole house even at a sub-code R38 in attic, from a long term moisture resilience point of view it's worth filling the rafter bays completely, planing the o.c. foam flush with the plane of the rafter edges, and installing a smart membrane vapor retarder rather than paint. But as long as the air gap between the gypsum & foam is truly miniscule- as tight as possible, and not an easy convection channel, and you religiously caulk the gypsum to the framing you'll probably be OK with a vapor retardent paint approach.

10. GBA Editor
    Martin Holladay | May 06, 2015 09:18am | #10

    David,
    I disagree with Dana concerning the need for the drywall to be airtight. Unless the spray foam installers were extraordinarily sloppy, 8 to 10 inches of open-cell spray foam constitutes a perfectly adequate air barrier.

11. Expert Member
    Dana Dorsett | May 06, 2015 04:17pm | #11

    The 8-10" of foam is great air barrier against infiltration, but not for stopping convection behind the drywall from drawing in air through leaks in the drywall layer. If the drywall layer is to be an effective vapor retarder, it has to be air tight:

    At 8-10" of ocSPF in a 2x12 rafter bay it means you have 2-3" of convection channel between the drywall & foam, a channel that is subject to convection due to the temperature difference at the face of the foam relative to the room air. A 2-3" gap is convection channel bigger than is required for a vented roof(!), but in this case it's on the "wrong" side of the insulation. If they were only going for 8-10" of foam rather than a full fill it arguably should have been a vented roof, with the 2-3" channel on the roof deck side, not the gypsum side.

    If the gypsum is not air tight it hardly matters what the vapor retardency of the paint is, since the leaking air can get behind it, and WILL due to the buoyancy forces of the denser colder air next to the foam falling to the bottom of the channel leaking into the conditioned space on the other side of the gypsum, drawing air into the channel higher up.

    The vapor retardency of ocSPF itself is usually over 5 perms @ 10", rarely under 2- the same range as standard latex paints and not sufficiently protective on it's own. So even fairly slow air leakage into the channel from the conditioned space side adds up to a real moisture burden potential. If the gypsum is air tight and painted with 0.5 perm paint, the potential is extremely low.

    If the average depth of the channel shrinks to 1/16" or less it's not exactly an air barrier, but it's somewhat air-retardent. The rate of convection from those buoyancy forces can't be very high, and the amount of moisture getting by the vapor retarder is that much smaller, even with some leakage at the margins.

    Given that 11-12" of ocSPF without interior vapor retarders has been demonstrated to deliver higher than safe sheathing moisture content even on walls with vented cladding that can dry toward the exterior, the problem is necessarily worse in roof decks with ~0.1 perm felt & shingle layups, which have inherently NO capacity to dry toward the exterior, and are subject to radiational nightime cooling, with lower average wintertime temps than wall sheathing:

    https://www.greenbuildingadvisor.com/articles/dept/building-science/cold-sheathing-double-wall-construction-risk

    http://www.buildingscience.com/documents/bareports/resolveuid/03990e0f18594c57b6643490d62bea0e

12. 
    djhaus18 | May 06, 2015 10:02pm | #12

    Dana,

    If the rafter bays are completely filled with foam to remove the convection channel does that reduce most of your concern for moisture? I don't think I'm going to arrive at a perfect solution with the installer I'm just trying to reduce as much risk as possible given the foam is already installed.

    -dave

13. GBA Editor
    Martin Holladay | May 07, 2015 07:35am | #13

    David and Dana,
    First of all, we all agree that it would be best to call back the spray foam contractor to finish the insulation job -- meaning to add enough open-cell spray foam to fill the rafter bays, and then to trim the insulation flush. This is the standard method for installing open-cell spray foam, which is easy to trim (unlike closed-cell spray foam, which is hard to trim).

    However, I'd still like to resolve the issue that Dana raised -- the claim that there will be convection currents between the drywall and the cured spray foam that will undermine the performance of the assembly.

    Dana is wrong, for several reasons.

    First of all, it is simply untrue, as Dana asserts, that "If the drywall layer is to be an effective vapor retarder, it has to be air tight." A vapor retarder is designed to address vapor diffusion, not air leakage. Measuring the rate of vapor diffusion through a material has nothing to do with air leakage.

    That said, there are many reasons why you want an air barrier. Fortunately, the installed open-cell spray foam is a very good air barrier.

    Second: Let's imagine that drywall was never installed. (This is not a good idea, because you want to address vapor diffusion -- but I'm introducing a thought experiment.) On the interior side of the cured spray foam, the air is at interior conditions -- let's say it is at 70 degrees F. The spray foam is an air barrier.

    Are there convection currents? Who cares?

    Will the air in the attic degrade the performance of the insulation layer? Not really -- even if someone installs a ceiling fan up there to move the air around. The air in the attic is indoor air; the air on the exterior side of the roofing is outdoor air; and you have a good air barrier and a pretty good (not quite code compliant) layer of insulation separating the two. Sure, you'll get heat transfer through the roof assembly -- and that heat transfer rate is a function of the insulation's R-value. No air leakage, though -- so convection doesn't matter.

    Now, let's add the drywall and some vapor-retarder paint. The air between the back of the drywall and the spray foam -- not much of a layer of air, by the way, especially if the insulation contractor came back to finish the job -- isn't moving much, because there aren't any air leaks to the exterior. It's basically still air.

    Even if the air moved a little, it's interior air, and it isn't flowing from the interior to the exterior. You have a good air barrier.

14. 
    Dana1 | May 09, 2015 08:13am | #14

    If room air gets by the vapor retarder, the vapor pressure across the vapor retarder is nil, and the vapor retardency pf the vapor retarder doesn't matter. A vapor barrier is only useful if it's sufficiently air tight that vapor diffusion is the primary mechanism of moisture transfer.

    With the gypsum leaking air the vapor retardency of the FOAM then matters. When the vapor pressure at the interior surface of the foam it tracking that of the conditioned space, 11" of open cell foam is not sufficiently low permeance to protect the roof deck. It's not even reliable protection for the sheathing of a wall assembly with vented cladding, where the sheathing can dry toward the exterior, as demonstrated by the in-situ monitoring of Carter Scott's double studwall with 12" of open cell foam:

    http://www.buildingscience.com/documents/bareports/ba-1501-monitor-double-stud-moisture-conditions-northeast

    And in a roof assembly there is the aggravated circumstances of no drying toward the exterior, and colder sheathing temperatures.

    So sure, the foam is a great air barrier, but it's not a good enough vapor retarder, thus the vapor retarder layer has to be sufficiently air-tight to make it's vapor retardency relevant.