Venting Cathedral Ceiling/Roof Assemblies

[Riversong]

I came across a recently published article in the ASTM Journal, "Compact Asphalt Shingle Roof Systems: Should They be Vented?". http://www.astm.org/DIGITAL_LIBRARY/.../JAI102057.htm

The abstract concludes:

Our study showed better durability of vented roof assemblies with permeable insulation in cold climates due to redundancies that can tolerate incidental moisture and provide visual indicators of roof leakage; roof sheathing typically dries in 1-1/2 to 2 months. All of the unvented roof assemblies are intolerant of incidental water leakage and the moisture-sensitive layers (i.e., sheathing and gypsum wallboard (for open-cell polyurethane insulation)) exceed the threshold for decay. In hot, humid climates, the most durable roof assemblies are the vented, open-cell polyurethane systems with shorter drying time of the interior gypsum wallboard when compared to the unvented roof assembly; both the sheathing and gypsum wallboard dry out within 2-1/2 months. In an unvented assembly, the drying time for sheathing is similar but the drying time for gypsum wallboard increases to 6.5 months on average. Alternatively, unvented permeable shingled roofs are a viable option in hot, humid climates, although they are slightly less durable. The least tolerant roof assemblies in either climate are the unvented closed-cell polyurethane roof assembly due to trapped moisture and slow drying of the roof sheathing (up to 12 months in Miami, FL and 27 months in Boston, MA).

Has anyone seen the full article? It certainly supports my long-standing belief in the importance of ventilating a roof, the value of permeable insulation, and the importance of allowing all thermal envelope assemblies to be able to breathe for drying the inevitable leak.

Responses?

[ThingOfBeauty]

Re: Venting Cathedral Ceiling/Roof Assemblies

That looks like a good read, hopefully we can find a copy! good post.

I was just discussing this with a bunch of DC area builders. Making a long story short, in our 4000-5000 HDD climate, unvented cavities seem to perform OK. There are a lot of unvented, cold-side-vapor-barrier roofs in existing DC row houses.
A couple of them said they figure normal indoor humidity levels (granted, in normally rather leaky houses so rather low) will dry a roof from the inside if you don't impede it too much. If it leaks probably all bets are off, but even vented cavities don't withstand leaks of almost any size, right?
In Vermont or similar cold climates, I would guess an unvented roof would be more vulnerable.
Without more info on the assumptions and method used in the ASTM study, it's hard to know exactly what this paper even says. But personally I'm comfortable saying a vented roof is less risky than an unvented one.
Regarding always using permeable assemblies, that seems to make sense except if you put a low-permeable foam insulation where the dew point happens inside the foam layer; or using an impermeable sheet on a water-tolerant surface like a masonry wall, with insulation on the other side? Seems like those could work as well.

[Dancing Dan]

Re: Venting Cathedral Ceiling/Roof Assemblies

I'll bite. It's the topic of next month's building science discussion group

[Dancing Dan]

Re: Venting Cathedral Ceiling/Roof Assemblies

Started reading - one interesting thing is they don't include cellulose in the comparisons, only open and closed cell SPF and FG batts. And the abstract said the best performing system was open cell w/ vents.

I guess I'm not sure how a vented assembly gives you better visual on leaks - if your soffet is dripping?

[ThingOfBeauty]

Re: Venting Cathedral Ceiling/Roof Assemblies

A little bird whispered in my ear that it's a theoretical, WUFI-based study. So, sadly not a real-world evaluation.
The simulation they ran included a one-time "leak" where they simulated water getting into the sheathing and watched how long it simulated to dry out. They assume shingles at .5 perms while acknowledging there is no tested or accepted value for shingle permeability. (sounds about right but still...)
Basically, to me it amounts to "unvented cavities don't dry as well if you get them wet once".
They do say they regularly see roofs get wet once and dry out...that seems like news to me...in my experience they leak either never or frequently.

That's my take for now.

[David Meiland]

Re: Venting Cathedral Ceiling/Roof Assemblies

I'd say 98% of the cathedral ceiling/roof assemblies around here are vented and have FG batt insulation. My experience is that the wind blows right thru them, including wind-washing thru the insulation, leaking air thru can light penetrations, and all the rest. They definitely dry well, if they ever get wet. IMO part of what's needed is a baffle system that actually separates the insulation from the sheathing in terms of air leakage. The cardboard or foam baffles that are used do not do this. I have seen one or two cases where someone installed wood strips to the sides of the rafters and then foamed in rips of XPS as baffles, filling the rest with FG. That probably works pretty well, and it could even be done to convey a roof leak right out the soffit vent.

This whole area is one that needs improvement. I'm not sure I want to go to SPF, at least as it's currently used.

[Martin Holladay]

Re: Venting Cathedral Ceiling/Roof Assemblies

David,
I agree with you. My problem with venting cathedral ceilings is that there is usually no air barrier between the insulation and the vent channel. That's nuts.

Of course there are ways to do it better. The best is to insulate on top of the roof sheathing with rigid foam, and then create vent channels above the foam. Install another layer of plywood or OSB, and then your roofing. This assembly is also available as a vented nailbase panel from several manufacturers.

[Riversong]

Re: Venting Cathedral Ceiling/Roof Assemblies

I got a copy of the full ASTM Journal article and below are the conclusions of the WUFI hygrothermal modeling study which examined 24 different roof assemblies for each of the two climates: Boston MA and Miami FL. The study was run for a 10-year period, using hourly local weather data, and examining only the last 8 years - after the construction moisture content reached equilibrium. They also introduced a single event leak (to simulate a wind-driven rain storm) of an average 0.20 lb water/ ft²h for a total duration of eight hours in the third summer of each simulation to examine the effects of incidental leakage.

Each roof assembly included: asphalt shingles; either felt or self-adhering membrane; either OSB or plywood; 9¼" of either fiberglass, open-cell spray foam or closed-cell spray foam; either poly VB (with fiberglass), or variable VB (with open-cell foam), or no VB (with closed-cell foam); and gypsum wallboard.

The following conclusions contradict the conventional wisdom that closed-cell spray foam does not require venting and protects from water damage and rot. This is one of only two studies I'm aware of that considers the inevitable leakage that will occur at one of more times during the life of a house (the other was a CNRC MEWS study of plywood vs foam sheathed walls, which also demonstrated dangerously poor drying potential with foam).

Conclusions

Our comparative moisture analyses show that important design considerations, such as providing ventilation and selection of insulation or water-resistive barrier, can significantly influence the robustness of moisture sensitive layers in roof assemblies due to incidental leakage in various climates. These design choices ultimately affect the durability of the roof system.

Based on our analysis, we present the following guidelines for determination of roof performance, taking ventilation and its effect on the drying due to incidental leakage as a major factor:

 Vented roof systems with permeable insulation in cold climates are durable because they include redundancies that can tolerate incidental moisture and provide visual indicators of roof leakage.

 The least tolerant roof assembly in either climate is the unvented closed-cell polyurethane insulation roof assembly with SRAM applied over the sheathing. This roof assembly creates a vapor trap and is slow to dry although the SRAM is supposed to prevent leakage from wetting the sheathing. Additionally, the closed-cell polyurethane foam will not allow leakage water to filter through and can promote deterioration of the wood roof structure with no visible indication of a roof leak.

 In cold heating climates, vented roof assemblies clearly outperform unvented assemblies with respect to drying potential. Multiple types of vented assemblies in this climate are considered durable, including permeable insulations such as glass-fiber batt and open-cell polyurethane insulation.

 In hot humid climates, the most durable roof assembly is the vented open-cell polyurethane with either felt or SRAM applied over the sheathing due to decreased drying time of the interior gypsum wallboard when compared to the unvented roof assembly. However the addition of venting in a hot, humid climate further complicates the location and construction of the air barriers and vapor retarders. Functionally, the foam needs to be applied to a second layer of sheathing, which also serves as a redundant air barrier. A layer of blocking would form a vented cavity between two layers of sheathing. The first layer of sheathing could also be changed to glass-fiber faced gypsum sheathing, to improve the fire resistance of the assembly.

 Unvented permeable shingled roofs are also a viable consideration in hot humid climates, although it would be slightly less durable. The unvented roofs would be less expensive and simpler to construct that the vented roofs. These unvented roofs result in increased drying time of the inboard gypsum wallboard. The insulation must remain permeable to avoid a vapor trap.

Given the trend towards unvented compact steep-slope roof systems with various types of spray-applied foam insulation, designers should consider performing transient hygrothermal modeling to analyze the potential for condensation and accumulation within the roof assembly and to understand the limitations of the system as a result of incidental roof leakage. Transient hygrothermal models can provide designers with a relatively inexpensive method of performing a durability analysis using alternate building components in a variety of climates.

[Dancing Dan]

Re: Venting Cathedral Ceiling/Roof Assemblies

I wish they had done dense-pack cellulose, vented and un-vented.

[ThingOfBeauty]

Re: Venting Cathedral Ceiling/Roof Assemblies

Well, we already knew that a loose assembly with air blowing through it will dry better than a tight one. So that's not exactly news.
Since this is a simulation, it can't show how much water soaks into open-cell foam (a LOT in my two experiences) which will severely impact the time of drying, or that a fiberglass-insulated ceiling's energy performance is a joke.
You know, foam was not used around here until the last few years but we still had plenty of roof leak damage. There are three possible rates of leakage: small amounts that will dry out; large amounts that go right through and warn everyone; and middle amounts that cause progressive damage to the sheathing but don't show for a while. I'm not sure this study does much except to say that making an assembly that doesn't dry out as fast means there will be certain cases where one of the rare leaks that would dry out won't dry anymore.
WUFI doesn't accurately model water flowing through a 3-dimensional assembly, just the theoretical drying potential through simulated models of the layers in the assembly. I'm afraid we'll have to wait and see how things work in the real world to see how this issue really turns out.

[Riversong]

Re: Venting Cathedral Ceiling/Roof Assemblies

WUFI doesn't accurately model water flowing through a 3-dimensional assembly, just the theoretical drying potential through simulated models of the layers in the assembly. I'm afraid we'll have to wait and see how things work in the real world to see how this issue really turns out.

I believe that WUFI is broadly accepted as the most sophisticated moisture modeling software and is routinely tested against actual scenarios for accuracy.

We could, as I've been suggesting, use the precautionary principle and not create future potential problems or we could wait, as you suggest, and see what future problems are created by our present shortsightedness.

My prediction: in 10 to 20 years, we're going to discover all kinds of rot and mold problems in unvented cathedral ceilings insulated with spray foam and covered with bituthane, as well as in wall systems with outsulation.

[David Meiland]

Re: Venting Cathedral Ceiling/Roof Assemblies

insulate on top of the roof sheathing with rigid foam, and then create vent channels above the foam. Install another layer of plywood or OSB, and then your roofing. This assembly is also available as a vented nailbase panel from several manufacturers.

Martin, that would work (foam above the roof deck) but you're going to have a tall fascia board. My preference might be to build as usual... rafters with sheathing... then insulate with spray foam between the rafters... then felt the top of the deck... THEN add furring strips along the rafter tops, another layer of plywood, another layer of felt, and then roofing. If the roof leaked it would be contained above the insulated structural assembly. I'm sure there are problems with this, but if the goal is to insulate our ceiling/roof assemblies better I think we're going to be going to less permeable insulations and we'll need to really keep the water out of them. My guess is that the average, well-constructed and well-detailed conventional cathedral ceiling/roof does get some water in it during high wind/rain events, and that it simply dries out. With foam or other materials it might not. That concern has stopped me from using SPF so far.

[ThingOfBeauty]

Re: Venting Cathedral Ceiling/Roof Assemblies

I believe that WUFI is broadly accepted as the most sophisticated moisture modeling software and is routinely tested against actual scenarios for accuracy.

WUFI is a really neat tool, no doubt about it. Try it out, play around with it, it's a very educational experience. Free too.

WUFI does have shortcomings though. In particular the three-dimensional behavior of a water leak in a non-vertical assembly can't realistically be modeled by a 2-D algorithm. The joists aren't even in the model for one.
Overall I would expect the inaccuracies to make your point even more strongly. I particularly believe that open-cell foam's benefit in the drying phase is oversold, because I've seen it holding lots of water. So, I would expect either kind of foam to perform poorly in the imaginary, small-irregular-leak-that-could-dry-in-a-leaky-cavity scenario.
Building Science suggested that anyone with a SPF roof should have it IR camera'd annually to see if there are hidden water leaks! wonder how that calcs out on the price/performance scale.

Overall I agree that a vented roof has a lot of advantages and much less risk. But, this study doesn't seem to me to be a slam-dunk, in fact due to its shortcomings I would say it's no more useful than conversation, and much less useful than actual field observation.

[mike donalds]

This is widely using in most cathedral: baffle ceiling This will help eliminate the stray reflection of the stage. They can also be used for sidewalls, rear walls or suspended on the ceiling. Larger ceiling baffle system, such as our 600x1200x50mm or 1200x1200x50mm sizes are ideal for church applications. Baffle ceiling will effectively absorb sound in cautious applications. ThBaffle ceiling will effectively absorb sound in cautious applications. The ceiling baffles insulation board is suspended from the ceiling structure, both sides can absorb the sound.insulation board is suspended from the ceiling structure, both sides can absorb the sound.