14" truss joists with a standard fill of cellulose, or would it need to be dense pack? No insulation on the deck just plywood, 30# paper and shingles,Ice dam on the eves and valley's, 5/8 drywall on the underside and latex paint. Would it be better to use a low perm primer on the ceiling? Will it meet 2009 IEEC? My understanding is that the State of IL energy code is based on 2009 IEEC.

Posted By ilgeo on 01 Nov 2011 10:39 AM
14" truss joists with a standard fill of cellulose, or would it need to be dense pack? No insulation on the deck just plywood, 30# paper and shingles,Ice dam on the eves and valley's, 5/8 drywall on the underside and latex paint. Would it be better to use a low perm primer on the ceiling? Will it meet 2009 IEEC? My understanding is that the State of IL energy code is based on 2009 IEEC.

Without exterior foam that would be a code violation per IRC 2009, which calls out for a minimum of R20 of air-impermeable insulation above the fiber in Zone 5.  Unlike walls, roofs have essentially ZERO capacity to dry toward the exterior (especially when you have 6" of snow & ice up there), so it takes a larger fraction of the total R to limit the wintertime accumulation of moisture in the roof deck than if it were a wall assembly.  With a more complicated vented roof deck assembly with a vent gap between the structural roof deck and a separate nailer deck for the shingles you can get away with a lot though- it's the moral-equivalent of a rainscreen gap in a wall assembly.

There are some vendors (notably National Fiber) who swear that in a 2x12 cavity dense packing to 4.0lbs density is sufficiently air retardent and moisture buffering that you can skip the exterior foam and interior vapor retarders (no need for vapor-barrier paint) in US zone 5,  but I haven't seen independent confirmation.  Assuming they're correct, , note that more typical dense-packing densities are 3.0-3.5lbs per cubic foot, so getting to 4 lbs is going to require a diligent professional who tests the density as they go along and remediate where cores show densities below 4.0lbs.  This bit o' bloggery is a timely disussion of this very question.

An alternative solution would be to apply 1-2" of closed cell (~1-1.5perms @ 1") spray polyurethane foam on the roof deck and dense-pack at normal dense-pack densities below it.  This too is a code violation per IRC2009, since it doesn't meet the minimum R value on the foam layer, but in fact it works, and works well- see table 3 on page 11 .

This last document simulates stackups with 1.8lb density JM Spider rather than 3lb cellulose as the fiber in the spray foam + fiber stackups, but the air retardency of 3.5lbs cellulose is comparable to (or even better than) 1.8lb spider, and the cellulose will buffer winter moisture drives better. 

This approach works because the condensing surface is the polyurethane foam, which won't wick moisture toward the roof deck even during condensing temps, and by using a fairly air-retardent fiber the total quantity of moisture is limited. That combined with the modestly-permeably 0.5-1.5 perms of the foam layer, the roof deck can still dry toward the interior.  With Spider some of the fiber would actually get wet during condensing conditions, but with cellulose the hollow & wicking properties of cellulose fibers allows it to absorb quite a significant amount moisture internally without losing R value or actually feeling wet- it buffers the moisture.  At 2" of ccSPF you'd be good to go even in northern MN, but a 1" flash-foam should be good enough for N-IL unless the roofing on the north facing pitch is a light-colored metal.  (At about a buck a square foot per inch of thickness the difference in cost between 1" vs. 2" is more than pocket change.)

Without the foam, the cellulose fibers in the wood of the roof deck compete with those in the insulation when the roof deck is below the dew point, and more moisture ends up in the roof deck,  but only about half or less of what  it would soak up it were dense-packed fiberglass with no wicking or buffering capacity. It would be interesting to see if painting the wood with semi-permeable acrylic paint to reduce the wicking at interior surface to the roof deck would improve the performance of dense-packed cellulose alone.  It might, but I wouldn't try selling that theory to the code folks without better evidence.  The WUFI simulations on the ccSPF + dense-packed fiberglass is convincing enough to me (and possibly your local building department.)

I could use ceramic paint on the underside of the roof deck and everything would be just fine. no?

Posted By ilgeo on 01 Nov 2011 02:36 PM
I could use ceramic paint on the underside of the roof deck and everything would be just fine. no?

Not necessarily- it depends on it's vapor permeability and wicking characteristics, the density & thickness of the cellulose for buffering the wintertime moisture.  (If it were simple to analyze it would have been done already.)

Dana sorry I was joking about the ceramic paint. If I put 2" cc on the underside of the deck than I can met thermal and moisture requirements with 3# dense pack cellulose?

Ceramic nanobead paints may be worthless as insulation, but they're pretty interesting as protective coatings. Whether they'd be useful here is anybody's guess.

According to the Building Science guy's WUFI simulations you'd be in great shape with 2" of cc and 3.5lbs cellulose, but 3.0lbs density might be cutting it close. You might want to print out that document in case you need to make the case for a variance. At 12" of depth you may need do go more than 3.0lbs density with the cellulose but with R12 of cc foam on the exterior it won't undergo nearly as much seasonal moisture cycling than if it's cold edge was the roof deck itself. It's the total seasonal moisture cycling that causes cellulose to lose loft, and by keeping it warmer/drier it would not be necessary to take it all the way up to 4.0lbs (the National Fiber recommended density in 2x12 assemblies when there is no exterior foam.) Center-cavity you'd be over R50 with that stackup. The whole-assembly R depends on the properties & dimensions of the webbing of the truss-joist rafters, but I'll bet it's well over code for Zone 5 (= R38 between 16" o.c. joists.) You'd be good-to-go with that stackup even in northern MN or the MI-UP with everything but a but a heat-rejecting metal "cool roof" for the exterior finish.

Dana
What about many Passivhaus projects that use thick unvented cellulose roof assemblies? There are a series of articles showing tight homes featuring 14-16" thick, non vented roof assemblies without exterior insulation or spray foams. These seem to be getting passed by local code officials and claim so because they meet the typical insulated value of a roof assemebly not classified under the lesser r; cathedral ceilings. What are your thoughts? They seem to be torn over that Green Building Advisor. Some who have used the roof system without issue continue to do so, while others still hold onto the code claims and suggest the vent space below the sheathing. I am trying to finalize my own future roof design. The plan is to use 2x10 rafters with plywood strips cut at 15" and hang 2x4 from the 2x10 to create a much deeper insulated space and create a thermal break over my entire roof, resulting in around an r-55-60. The roof would be ZIP sheathing as well as the walls, and no overhang. The roof and wall sheathing would be taped to create a completely air tight shell. overhangs would be added later.

Izerarc,
Passivhaus is were I got the idea for this roof assy. It was an article that someone on gbt had posted a link to a year or so ago. I think the house was in or near Urbana Il

In US zones 3 & 4 you can get away with a lot with an unvented high-R cathedralized-roof assembly that won't necessarily work on the cold-edge of zone 5, even if it doesn't meet the letter of the IRC for air-retardency.

If you can point me to a specific PassiveHouse design that's an all-cellulose cathedralized roof, I'd be able to comment on it. Where do I find the aforementioned " ...series of articles...?

An air-tight ZIP roof deck should perform nicely until you fill it full of nail holes and one of the nails leaks. If you used 2x furring (through-screwed to the rafters) and an exterior and laid down an OSB nailer deck on the furring it would be a vented assembly with far fewer code & moisture issues to be concerned about. Even when the roof leaked though the nailer deck the cellulose would stay dry. The nailer deck is an expense, but compared to a couple inches of ccSPF it may still work out favorably, and is a rock-solid way to keep the ZIP dry enough in zone 5 without interior vapor retarders (maybe even zone 6.)

http://www.passivehouse.us/passiveHouse/PHIUSProjectsSmithHouse.html#1

Here is a prime example of why I should not operate from memory...16" tji vented deck...

Yep they vented it.

IIRC they used 1.8lb Optima as the fiber in that house, EPS for the foam bits. 

Optima (Spider too) is half the weight and somewhat higher-R than dense packed cellulose but 1.5x the price/R (get quotes), and doesn't buffer moisture.

I'm starting to think the best bang for my buck is to go to 16" tji with dense pack cellulose, or drop the tji and go to scissor truss, vented attic with r52 of loose cellulose. I'm sure the later is more cost effective, but looses the effect on the interior design.

Dana
This is not the passivehaus article I was thinking of (not a ph at all) but it is where I originally got the concept.
I think I have the pdf of the ph saved on my home computer...will have to see if I can find the source I got it from.

http://www.kaplanthompson.com/_images/publications/09.06-jlc.pdf

The other thought (more costly) is to use 14-16" TJI and use the upper flange as the vent channel, and rip 1/2" plywood in 2' stripes and caulk and nail to the underside of the upper flange to create a 1.5" vent channel. This should till allow me to have a pretty air tight roof assembly that is vented.
In our area, cellulose has had a big increase in price, and many contractors are pricing blown fiberglass the same as cellulose, especially since they prefer the ease of blowing fiberglass anyway (less bulk to transport, less mess and dust, etc).

Posted By lzerarc on 03 Nov 2011 01:10 PM
Dana
This is not the passivehaus article I was thinking of (not a ph at all) but it is where I originally got the concept.
I think I have the pdf of the ph saved on my home computer...will have to see if I can find the source I got it from.

http://www.kaplanthompson.com/_images/publications/09.06-jlc.pdf

The other thought (more costly) is to use 14-16" TJI and use the upper flange as the vent channel, and rip 1/2" plywood in 2' stripes and caulk and nail to the underside of the upper flange to create a 1.5" vent channel. This should till allow me to have a pretty air tight roof assembly that is vented.
In our area, cellulose has had a big increase in price, and many contractors are pricing blown fiberglass the same as cellulose, especially since they prefer the ease of blowing fiberglass anyway (less bulk to transport, less mess and dust, etc).

That roof & wall assemblies in the Kaplan Thompson pdf uses an interior 3-ply reinforced polyethylene membrane made by ParPac.  I couldn't find an ASTM E 96 perm spec on it, but I suspect it's pretty low. It would block winter moisture drives from reaching the roof deck, but it also means the assembly MUST dry toward the exterior, through a metal roof, no less.  They didn't detail how the applied the metal roofing, but if there is no vent space between the roof deck and roofing  and a sufficiently high-perm top treatment to the ZIP/OSB there's no way for moisture to leave the assembly.  They mis-applied a mesh-type siding rainscreen material (John Obdyke Home Slicker) as a metal roof underlayment but it's nowhere near thick/deep enough to vent a roof. In figure 2 it looks like they used copious quantities of Grace Ice & Water Barrier in some areas, which is very low-perm.

I'd be a bit concerned about the long-term prospects for this roof.  They even noted in the article that the Huber ZIP was swelling up a bit after wetting. Perfection is a lot to ask of a roof, and at some point it WILL get wet, and it can't dry very quickly to the exterior. The cellulose will soak it up for quite awhile, but you'll never see it on the interior with that membrane in place, and the ZIP will likely be history by the time it's discovered & rectified.  Mounting the roofing on cross ventilated purlins (purlins mounted on furring parallel to the rafters) the assembly would have both good drainage and top side ventilation of the sheathing, making it much more resilient to the water that will find it's way behind the metal roofing.

The walls also are also a drying-to-exterior-only stackup with interior ParPac vapor barrier. They didn't specify the siding type or the thickness of the Home Slicker used but even the 6mm version would enhance the wall's drying capacity quite a bit. (A 10mm rainscreen would be required by code if it were in Canada, but this is in Maine.) I'm much less concerned about the walls hanging in there forever.

Dana
do you think the parpac is the biggest issue? I do not see where they are using the drainage mat under the metal roof. I have read a few other similar publications that do not have any sort of interior vapor barrier. Here is another article where they do not appear to use any sort of a interior barrier.
http://www.ihtmv.org/PDF/HPHomes.pdf

Stack effect means that you have moist warm air pushing to find a crack and you don't want to trap it within the roof. For venting in little space, you can use AtticWrap.

http://www2.dupont.com/Tyvek_Weatherization/en_US/products/residential/resi_atticwrap.html?src=tyvek_atticwrap_g_Broad

Posted By lzerarc on 05 Nov 2011 12:16 AM
Dana
do you think the parpac is the biggest issue? I do not see where they are using the drainage mat under the metal roof. I have read a few other similar publications that do not have any sort of interior vapor barrier. Here is another article where they do not appear to use any sort of a interior barrier.
http://www.ihtmv.org/PDF/HPHomes.pdf

That looks like a code violation, and not a great way to insulate a roof.  They may be counting on the relatively mild winter climate on Martha's Vineyard (of the SE coast of MA, NE of eastern Long Island NY)  for fewer condensing hours at the roof deck, and a the massive moisture buffering capacity of R49 dense-packed cellulose to keep the roof deck sufficiently dry.  (It's quite mild relative to nearby Boston- it's the warmest possible edge of zone 5, or the cooler edge of zone 4.)  Figure 3 specifies vapor retardent paint on the walls, and that probably applies to the roof assembly as well, but I can't advocate that methodology.   It may actually work in that location in a WUFI simulation, but I wouldn't  do it. YMMV.

The dense-pack in the walls doesn't come all the way to the gypsum either, leaving a potential thermal & hygric bypass gap between the vapor barrier paint and the insulation, but in a double studwall cellulose fill that's probably not a deal-breaker in that climate.

Using doubled up foil faced iso on the interior of the foundation wall doesn't thrill me either, since it demands perfection of both the foundation-sill seal, the waterproofing on the sub-grade concrete, and the foil facer that is in contact with the sub-grade foundation.  Iso is hygroscopic (unlike polyurethane or polystyrene foams)- it will wick and retain moisture, however slowly.  Any ground moisture finding it's way into the concrete puts the layer of iso and the foundation sill at risk at any points of imperfection.  As drawn the exposed lower edge of the iso is in direct contact with the footing without benefit of a capillary break or vapor barrier the way the slab is, which is a moisture bypass of the exterior foil allowing moisture to creep into both layers of iso over time, with no way out.

In extremely well drained sandy soils well above the water table this may work out OK, but ICF or using and interior of  2.5" of unfaced EPS  + unfaced batts/blown fiberglass in a studwall to achieve the same whole-wall R-value would allow the concrete to dry toward the interior, a much more resiliant approach that doesn't demand long term perfection.

No house is perfect, but on a good-better-best practices scale this one is barely on the "better" side of "good" from a long term resiliancy point of view, IMHO.

Everybody's a critic, eh?  

jonr:  Can you really dense-pack against Tyvek without bowing it out to contact the roof deck? (I've never seen it done.)  If not initially, over time & pressure I'd think polyolefin would stretch under the tension- it's not nearly as structural as even half-inch gypsum, which has real issues with 24" on-center framing + dense pack  Even if it doesn't contact the roof deck between rafters/trusses, it would be susceptible to nail damage from the roofing.  Leaving at least 3" of space between the deck and the Tyvek would probably be OK, but I doubt 2" would be enough. 

OTOH they package cellulose for shipment at 2x+ that density in heavy polyethylene, so maybe it works.

You are assuming applied from below? I would apply stabilized cellulose from above with whatever ceiling strength and cellulose density required to support it and keep it from moving.

I think what Dana is saying (and I agree with) is if you use the Tyvek product, it is a flexible wrap spanning from joist to joist. The pressure from the dense packed insulation will press the Tyvek up against the underside of the roof sheathing.