Insulation value of snow
We are building a house in Truckee, CA. Elevation is approx 6500 ft. Our plan is to use R38 unfaced fibreglass insulation under the structural roof deck, 2″ rigid foam insulation over the structural roof deck, then with 1″ vent channels from the eaves to the roof vent, and then another layer of plywood sheathing to attach the asphalt shingles. My question is when estimating the condensing temperatures, prior discussions that I have read seem to ignore the insulating factor of snow. The typical conditions in that area are to accumulate a substantial depth of snow (1-3ft) on the roof that remains throughout the winter.
My sense of the use of the layers is that the first layer of sheathing will dry to the inside and the second will dry to the vent channels. Additionally, the vent channels will keep the roof deck cold enough to reduce the incidence of ice dams.
Do you have any insight on this combination of materials and the final insulation and condensation result when snow is added to the equation?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Burt,
Depending on how fluffy the snow is, its R-value varies R-0.5 per inch to as much as R-2 per inch. The latter figure is unlikely.
You're near Reno, Nevada. I know that Reno is in Climate Zone 5, so I'm going to assume that Truckee is, too.
If I understand correctly, you are insulating your roof assembly with a combination of rigid foam and fiberglass batts. To make this type of roof assembly work, the minimum R-value of the rigid foam layer in Climate Zone 5 is R-20. Your 2 inches of rigid foam is too thin.
Here's what's going to happen: your roof sheathing will be too cold in the winter to keep your roof sheathing warm enough, so your roof sheathing is at risk of moisture accumulation, mold, and rot. Snow is unlikely to save this roof.
Here is a link to an article with more information: How to Build an Insulated Cathedral Ceiling.
If that is the case, why would the same not happen with a conventional cathedral ceiling roofing system which is R38 below, with a ventilation space under the sheathing to keep the roof sheathing cold enough to not create ice dams? I have read the reports that talk about the R20 rigid on top, but most assume that you are not providing ventilation space under the top layer of sheathing. They all seem to end with a footnote that says "when in snow country, an air space underneath is advisable to prevent roof dams." Wouldn't that same air space solve the moisture accumulation problem?
Burt,
Correct me if I'm wrong, but it sounds as if you plan to install a ventilation channel above the rigid foam. That type of ventilation channel is an excellent way to reduce the chance that you'll have an ice dam, but it doesn't help remove moisture from your roof assembly the way a code-mandated ventilation channel under your roof sheathing does.
The code does not allow you to install fiberglass batts under your roof sheathing, unless you either (a) include a ventilation channel between the underside of your roof sheathing and the top of your fiberglass batt -- and I don't think you have that -- or (b) you include an adequate thickness of rigid foam above the roof sheathing -- in your climate zone, that means a layer of rigid foam rated at R-20 or more.
Q. "Why would the same [moisture accumulation] not happen with a conventional cathedral ceiling roofing system which is R-38 below, with a ventilation space under the sheathing?"
A. The reason that you don't get moisture accumulation in a cathedral ceiling with a ventilation space under the sheathing is that the air flow helps carry away moisture and keep the sheathing dry.
Truckee is significantly cooler than Reno due to the higher elevation. With a mean climate of nearly 8000 HDD, Truckee is definitely a zone 6B climate, which means literally HALF the R-value would have to be above the roof deck for dew point control on an unvented assembly.
http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?ca9043
The IRC prescriptive in TABLE R806.5 presumes R49 for the total center-cavity R. For higher R assemblies, the exterior foam needs to grow proportionally:
http://publicecodes.cyberregs.com/icod/irc/2012/icod_irc_2012_8_sec006.htm
For vented roofs, the 1" minimum (2" is better in snow country) air channel has to be UNDER the roof deck to protect the roof deck from wintertime moisture.
If you haven't specified the rafter depths yet, going with 2x8 /R30 and 6" of rigid polyiso above the roof deck gets you there. At 6" the foam-over is still pretty manageable. The timber screws get really long when you're doing say, R75 all above the roof deck.
If it's easier or cheaper, a 3.5" polyiso nailbase panel (polyiso with half inch plywood or OSB facer laminated on one side) on top of 3" fiber faced roofing polyiso is pretty straightforward, but it's also possible to use 2 layers of 3" or 3 layers of 2" with an OSB layer through-screwed to the structural roof deck as well.
Thanks for your comments, Dana.
Burt,
So, if you are in Climate Zone 6, building codes require that any rigid foam installed above the roof sheathing have a minimum R-value of R-25, not R-20 as I recommended earlier.
If you end up installing more than R-49 worth of insulation, the required minimum thickness of the rigid foam goes up. For more information, see Combining Exterior Rigid Foam With Fluffy Insulation.
If I understand the answers correctly, the condensation problem is in the lower (structural) sheathing. Considering that, it becomes a ratio problem, lower versus upper insulation. It appears you are saying the solution is to reduce the lower to R30 from R38 and then get the upper value up to 20 or 25. One solution we have looked at is 2.5in of polyiso for a value of R16.7 and 1 inch of EPS with a value of 4.35 to get to a total of 20 on top. 3in on the first layer gets us R20.3 on that layer. It is all getting pretty expensive to put it together that way. It seems the easy way out is just to use the truss bays to get a lot more insulation below, allow for some vent space and not do the foam on top at all. FYI, nailbase panels are really expensive, about double the cost of the foam plus the plywood.
In a 6B climate (that's you, Truckee) you need to raise the outer R to equal or exceed the R value of the cavity fill. R20 exterior works with R30 cavity fill in Reno, but not where you live. R25 exterior doesn't work with R30 cavity fill either. You need at LEAST R30 on the exterior of the structural roof deck to be moisture safe with R30 in the cavities.
If you're going for just code min, derate polyiso to R5 per inch for performance at temperature in that stackup & climate, and you're good with only 5," and use R23 rock wool in 2x6 rafters. That still works from a dew point perspective even with derating- you'd be at R48 center cavity performance. An it would easily pass muster from a code-R review since the labeled R would be R27.5 on the exterior (R5.5 /inch is typ labeing for roofing iso) making the center cavity labeled R value R50.5. It would still outperform most R50 installations of cellulose on the attic floor, due to the better thermal break over the rafters.
The material cost of nailbase polyliso is indeed up there compared to foam + plywood, but the labor costs are higher with the latter. In high labor cost areas it can be a wash or even cheaper. I've personally never been involved with a project where it penciled out favorably to go with nailbase panels, but I've been told by reliable sources that it sometimes works out that way. Clearly YMMV.
To save money on rigid foam, figure out how much you need, add 20-25% to that, and see if you can't get a truckload of reclaimed roofing polyiso drop shipped to your site. Nationwide Foam has staging depots in several locations in the lower 48 and the freight might not be all that bad if there's one near you: http://nationwidefoam.com/
I have multiple vendors of reclaimed and factory seconds rigid foam in my neighborhood (including the above), not sure if there are others near you, but Nationwide is true to their name- they'll ship almost anywhere (for a price.) The cost of decent quality used foam is typically 25-35% virgin stock pricing, it adds up fast on bigger projects with thicker stackups.
Burt, yes, insulating the attic floor is usually much cheaper than insulating the attic roof. If that's an option for you, I would choose that. And it also allows you to go up to R60 or R75 with very little additional cost.
If you insulate the roof, I do think that the snow helps you. The code doesn't give you any credit for it, and you count count on it--I hear that there have been recent years with less than normal snowfall in your area, and that might happen again. Given that you have to at least meet the code requirement, all the snow does for you is make that code-minimum design less risky than it would be otherwise. That's a nice confidence booster if you go that way, but it's still more expensive and not better than insulating the attic floor.
We have scissors trusses. The end bays are probably not big enough to get the values we want and it would be difficult to get it blown in properly into a decreasing space. Even if the R value was acceptable, I am trying to solve the thermal bridging issue, too. It sounds like R19 fibreglass batts or the R23 rock wool below and R25 above will give the right ratio. (when all else fails, reread the code, thx)
Thanks for all the input to help me get some clarity to how to assemble this properly. When I checked back with my insulation vendor, his reply was: "most people just do R38 and that should be enough". That wasn't my intention, I am trying to do better than minimums, just not break the bank doing it. All info has been appreciated.
Burt (and Charlie) - I'm not sure that the snow helps at all . The idea of the vent space above the insulation is to keep that last sheathing layer at or near the outside temp. You have insulation layer(snow), cold air space(vent), then the insulated roof deck. That cold air vent space disallows help from the snow, it would seem. It does help with ice dams,though.
IRC 2012 and 2015 code minimum for attics in Truckee's climate is R49, unless they can demonstrate it still meets a whole house performance criterion. I'm not quite sure how folks are getting away with only R38 under Title 24 rules, but I'm not very well versed in the nuances of CA Title 24 2013, but it often differs from IRC prescriptives. This may be one of those places where it differs.
[edited to add] I just looked it up:
--------------------------
SECTION 150.0 – MANDATORY FEATURES AND DEVICES
Any newly constructed low-rise residential building shall meet the requirements of this Section
(a) Ceiling and Rafter Roof Insulation.
The opaque portions of ceilings separating conditioned spaces from unconditioned spaces or ambient air shall meet the requirements of Item 1 or 2 below:
1. Ceilings and rafter roofs shall be insulated between wood-framing members with insulation resulting in an installed thermal resistance of R-30 or greater for the insulation alone. Attic access doors shall have permanently attached insulation using adhesive or mechanical fasteners. The attic access shall be gasketed to prevent air leakage.
EXCEPTION to Section 150.0(a)1: Insulation of rafter roofs in an alteration shall be insulated between wood-framing members with insulation resulting in an installed thermal resistance of R-19 or greater.
2. The weighted average U-factor shall not exceed 0.031 that would result from installing R-30 insulation between wood-framing members.
(b) Loose-fill Insulation.
When loose-fill insulation is installed, the minimum installed weight per square foot shall conform with the insulation manufacturer's installed design weight per square foot at the manufacturer's labeled R-value.
----------------------------
Looks like R30 is code-min state-wide, independent of DOE or CA climate zone.
Kevin,
You are right that those vent channels make the temperature at the surface of the insulation close to the temperature of the outside air, regardless of the insulation provided by the snow. I forgot about them when I wrote my answer below. Oops. Thanks for pointing that out.
But I think the snow does very slightly help still--on the coldest winter nights, the sky is usually clear, and roofs can get colder than the outside air through radiation to the sky. That's why the code prescribes a bigger percentage of the insulation exterior to the sheathing for a roof than for a wall. So if you could count on snow, you could probably use the wall prescription instead of the roof prescription. Note the big if there.