The 475 Blog has a new post about unvented dense pack cellulose cathedral ceiling/flat roof systems using Intello membrane on the inside to allow the assembly to dry to the inside.  I am thinking about this and my assembly would look like:

Shingles
Roof felt/ice guard
OSB (taped seams)
Dense pack Cellulose
Intello
drywall/paint

Any other thoughts on this?  The other usual options would be spray foam all or at least a few inches of the cavity or rigid foam over the exterior surface of the OSB.

Thanks!
Scott

With the Intello and perfect air sealing it can work in most climate zones.

With no thermal break over the rafters the net U-factor of the assembly isn't going to meet IRC 2012 code min for zones 4 or higher with 2x12 rafters, since it takes more than 13" of cellulose to hit R49, and a 2x12 only gives you 11.25".  In zones 3 or lower you can meet code with 2x12 rafters and cavity-fill only though.

Foam above the roof deck outperforms foam between the rafters, since it thermally breaks the rafters.  The ratio of foam-R to total R required for dew point control at the roof deck is different in different climate zones.  The prescriptive values in Chapter 8 presume a total R that is only code-min.  If you go higher than code for total R, the foam-R has to be increased proportionally. eg:

For US zone 5 they prescribe R20 on the exterior, which is about 40% of the R49 code min for zone 5.  If you fill a 2x12 rafter bay with cellulose you'd have about R42 in the rafter bay, and you'd need R28 above the roof deck for dew point control.   If you only installed R8 (say, 2" of EPS) above the roof deck it would meet code for total R value, but you would still need a smart vapor retarder such as Intello or MemBrain on the interior to control moisture at the roof deck.  Byt with R28 or more above the roof deck you could skip the Intello in a 2x12/cellulose + exterior foam stackup.

How deep are the rafters, and where are you located?

Cincinnati, Ohio. I'll ask the designer about the rafter depth, we have not broken ground yet or else I would measure them.

Cincinatti is on the warm edge of US climate zone 5A, smack on the edge of zone 4A.   IRC 2012 code-min is R49, not sure which rev. of the IRC  OH code is based on, but R49 isn't insane.

With 2x10 rafters you'd get about R34 out of dense-pack, and if you went with R20 above the roof deck you'd be fine, despite being on the zone 5 edge of the line.  In zone 4 you'd only need 30% of the total R on the exterior for dew point control.  Either 5" or 5.5" EPS nailbase panels above the structural roof deck would get you there.  If you used polyiso nailbase panels you need to derate a bit from the labeled R, but a 4"-4.5" panel would do it.

Alternatively, you could go with 16" TJI rafters w/cellulose and Intello or MemBrain behind the ceiling gypsum, if you promise to get religious about air-sealing the vapor retarder and avoid punching a lot of holes through it for lights, etc.

Dana,

That seems to be the issue-- this is doable but it really has to be perfect. Thanks for the recommendations!

I don't see where the code allows the use of Intello plus exterior side foam thinner than prescribed. Or requires adjustment for increased rigid foam if more air permeable insulation is used (although it makes sense).

Intello or MemBrain, cellulose, then the prescribed amount of rigid foam would be a good, conservative way to go. Or add a 1/4+ " vent space (edit: above the sheathing).

Posted By jonr on 15 Jul 2015 04:37 PM
I don't see where the code allows the use of Intello plus exterior side foam thinner than prescribed. Or requires adjustment for increased rigid foam if more air permeable insulation is used (although it makes sense).

Intello or MemBrain, cellulose, then the prescribed amount of rigid foam would be a good, conservative way to go. Or add a 1/4+ " vent space.

Code requires a class-II vapor retarder when the vent area of the attic doesn't meet code minimum cross sectional area.  Intello (and MemBrain) is a Class-II vapor retarder when the proximate air is dry, and becomes more vapor-open when the RH is high, which lets it dry quickly.  You can safely fudge this one as long as you don't poke it full of holes. 

And with the air retardency and hygric buffering capacity of dense-packed cellulose it will tolerate minor air leaks, just not big ones.  The cellulose will redistribute the moisture loads, and a smart vapor retarder will let it dry fairly quickly.

A vent space of only 1/4" doesn't meet code- it takes an inch, but for low-angle roofs you'll need a lot more than an inch.

A flash-inch of closed cell spray polyurethane foam applied to the underside of the roof deck alone is sufficient vapor retarder to protect the roof deck, and as long as you have the smart vapor retarder to limit the moisture burden in the cellulose it'll be fine.  See this document.

> A vent space of only 1/4" doesn't meet code

1/4" is right out of the code (for wood shingles and note, still in the R806.5 unvented section). I'd say one needs to stick to the requirements of 5.1-5.3 for such designs. Ie, when no vent space is required, a 1/4" vent more than meets R806.5 code - it doesn't shift to an undersized vented attic.

Posted By jonr on 15 Jul 2015 10:21 PM
> A vent space of only 1/4" doesn't meet code

1/4" is right out of the code (for wood shingles and note, still in the R806.5 unvented section). I'd say one needs to stick to the requirements of 5.1-5.3 for such designs. Ie, when no vent space is required, a 1/4" vent more than meets R806.5 code - it doesn't shift to an undersized vented attic.

The 1/4" in R806.5 is the required space between WOOD shingles and the underlayment ABOVE the roof deck, not a vent space below the roof deck:

"3. Where wood shingles or shakes are used, a minimum ¹/₄-inch (6 mm) vented air space separates the shingles or shakes and the roofing underlayment above the structural sheathing."

The under-roof clearance requirements are in R806.3:

"A minimum of a 1-inch (25 mm) space shall be provided between the insulation and the roof sheathing and at the location of the vent. "

And what I'm suggesting is adding a non-required vent space between the shingles and the underlayment, ABOVE the roof deck. This adds resilience and doesn't turn an otherwise code compliant R806.5 unvented design into one that violates code. R806.3 doesn't apply to this case.

See here for discussion of such a "vented over-roof" on an unvented design (edit: for definition, since there seems to be confusion over what the stack looks like).

The entire discussion of the over-roof in that document consists of two sentences in the discussion of unvented :



Not exactly relevant to this discussion, since this is not a high snow load area.  And 1/4" gap above sheathing and ice-shield membrane is not a drying path, even if vented to the outdoors, since it's impermeable to water vapor. Even if you traded the membrane for #30 felt it would need more than 1/4" of vented air gap to provided any appreciable drying, and you can't even create that gap without adding another layer of nailer-deck, so you might as well give it 1.5" ( with 2x furring). That's a lot more expensive than Intello or Membrain, and not as effective, since the ~1 perm structural sheathing is exposed to the interior moisture drives, requiring the moisture to move through it, rather than having that moisture hit the drying channel first.

I'm glad we now agree that one can add an optional 1/4+" vent gap (above the rafters) and remain code complaint and it will add to moisture removal. Assuming ice-shield below the gap is a straw man and when a stack is designed to need no vent gap at all, I suggest that 1/4" is appreciably more than that.

Intello/MemBrain is beneficial - but it doesn't exempt one from 5.1-5.3. As mentioned, "1/4+" - sure, more will remove more moisture with little cost increase. Whether even more optional insurance is worth it - it depends and there is always the "where do you stop" question.

If there is good research on above the rafters roof vent channel size vs moisture performance, I'm interested.


Less than perfect air sealing (lights) and a lack of reserve resiliency sufficient to remove this moisture did this roof in.