Unvented roof assembly remodel — technical question
swimslowly
| Posted in Energy Efficiency and Durability on
[Editor’s note: At the request of the GBA reader who posted this question, the question has been deleted.]
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
R. Bouchard,
Physics is unrelenting, isn't it?
There is always a way to support insulation if you need to do so. The denim insulation needs to be installed directly against the roof sheathing; that is both a code requirement and the recommendation of building scientists. It would also be a good idea to have an air barrier under the denim insulation (and if you install tongue-and-groove ceiling boards, that air barrier will be missing).
One possible solution to your dilemma is to install 1"x1" ledgers at the proper height, on the sides of each rafter, in each rafter bay. These ledgers would be installed 8 inches below the roof sheathing. Once the denim batts are installed, you could install plywood under the batts to properly support them. The plywood could be fastened to the ledgers and diagonally to the rafters. If the plywood is installed with attention to airtightness, it becomes your interior air barrier.
Concerning your second dilemma: there is always spray foam.
Your questions highlight a point I have made repeatedly: that it's always a good idea to nail down your insulation details before construction begins, rather than scratching your head halfway through the project.
Martin Holladay,
Thank you very much for the response and suggestions. The 1"x1" ledgers with plywood under the batts is a good solution to both of the problems with the first roof assembly. We will use that method and pay attention to airtightness. I have done my best with planning, but unfortunately there is minimal attic access and the previous owners took the plans for the house - so some things have to be changed/modified as we go along as different problems are discovered.
With regard to the second roof assembly I'm hoping we can adequately air seal the ceiling and vent the underside of the roof according to the suggestions in "Lstiburek’s Rules for Venting Roofs", but unfortunately we will not know until the roof is removed if there are any obstructions for venting as there were on the first roof assembly. From what I know now, the only area that could be problematic is a section of the roof that peaks with clerestory windows instead of a normal ridge (see attached photo). Would it be sufficient to have a ventilation path notched out of the rafters to vent that section of the roof to the nearest ridge? With the assembly in the Lstiburek video - is the top side of the insulation exposed to the 2" insulation wind baffle? Nothing covers the top of the insulation in this assembly? It seems we could use the 1"x1" ledgers, plywood for airtightness, 9" of insulation and have a 2" space for ventilation within each of the 2X12" rafter bays (of course there would be soffit to ridge vents installed for ventilation). In this assembly could there be a combination of 1" foam on top of the plywood combined with the denim insulation or does the cavity need to be filled with only vapor permeable insulation?
Personally I feel the risks of spray foam do not outweigh the benefits - as long as other design options are available. Thank you in advance for your time and advice,
Ryland
At 3" the polyiso isn't sufficiently derated for your actual climate extremes to work- you'll need at least 4" if you're using R30 fiber on the interior side. At 3" you're really only looking at ~R16 performance at outdoor temperatures that matter from a dew point control point of view, when you need it to perform at R20 or better to have any margin.
Cotton provides a substantial amount of hygric buffering, and putting it tight to the roof deck would be more protective than leaving a gap. If you installed Martin's 1x ledgers for making a plywood or or OSB support there, make it a compression fit, and caulk or can-foam the OSB to the sides of the rafters to make it a true air barrier. At 1/2" the OSB/ply would be a 1-2 perm vapor retarder, which would also be protective.
Roofs with hips & valleys are lousy candidates for soffit to ridge venting if any portion of it is going to be a cathedralized/unvented ceiling or roof, and attempts to vent them can cause as many issues as it fixes in a zone 5 location.
Hi Dana,
Thank you for the additional information, I've read that one layer of XPS on top of the Polyiso could be a good compromise and improve the winter performance of the Polyiso rather than adding another inch of foam?
There are essentially two separate roof assemblies on the roof that are not thermally connected. The livingroom has a cathedral ceiling which has its own roof, and the rest of the house has its own roof assembly (see attached drawing). Since the two areas are not thermally connected it seemed okay to combine the unvented assembly with the vented assembly - am I mistaken? In the EPA booklet on moisture control it says that in an unvented assembly the foam should be on the interior side of the high-perm insulation (denim) when you combine vapor permeable insulation with impermeable insulation. Based on this information I had planned on including two layers of foam underneath the denim insulation before the ventilation channel. In this assembly is there anything between the denim and the ventilation channel?
Thanks in advance for your advice. I know an on-site building scientist/engineer would be ideal, but we are in a rural location and I have been unable to find anyone knowledgable to assist with the technical details of the project.
Ryland
Ryland,
You wrote, "In the EPA booklet on moisture control it says that in an unvented assembly the foam should be on the interior side of the high-perm insulation (denim) when you combine vapor permeable insulation with impermeable insulation."
I am not familiar with this booklet, but its advice is the opposite of code requirements for unvented cathedral roofs, and is also the opposite of recommendations from consultants at the Building Science Corporation.
You want the air-impermeable insulation (the foam layer) to be on the exterior side, and the air-permeable insulation (the fiberglass, cellulose, or denim) to be on the interior side. This approach limits condensation risks.
Hi Martin,
My apologies for the typo - I meant to write in a vented assembly. So it should read: "In the EPA booklet on moisture control it says that in a vented assembly the foam should be on the interior side of the high-perm insulation (denim) when you combine vapor permeable insulation with impermeable insulation. Based on this information I had planned on including two layers of foam underneath the denim insulation before the ventilation channel. In this assembly is there anything between the denim and the ventilation channel?" The booklet I referred to is here: http://epa.gov/iaq/pdfs/moisture-control.pdf, and the section I was referencing is on page 56.
My apologies for the confusion and thank you in advance for your help,
Ryland
Ryland,
Your use of the phrase "before the ventilation channel" is confusing.
If you want to include a ventilation channel, the ventilation channel has to be on the exterior side of all insulation layers. The usual location of the ventilation channel is between to top of the uppermost insulation layer and the underside of the roof sheathing.
It's also possible (but more unusual) to install the ventilation channel above the roof sheathing. If you choose this approach, most types of roofing require the installation of a second layer of roof sheathing above the ventilation channel.
Hi Martin,
My apologies for the confusing terminology. In a vented assembly, if the uppermost insulation layer is denim does there need to be anything between the denim and the ventilation channel?
In an assembly with ventilation channel above the sheathing is a 1.5" ventilation channel sufficient or is the minimum also 2" for this type of assembly? Thank you,
Ryland
If you're going to be putting a layer polystyrene foam on the exterior of the polyiso to better manage the average performance, EPS has a lower environmental footprint solution than XPS. It also has more stable long term performance.
XPS is blown with HFC134a (global warming potential ~1400x CO2), whereas EPS is blown with pentane (7x CO2), and whereas almost all of the pentane is gone by the time sheet-EPS hit's the distributor's yard, the HFC134a bleeds out over the first few decades and the R-value of XPS falls along with that loss. At 50 years the XPS has barely higher performance than EPS of the same density, but it's done 200x the climate damage.
Like XPS, the R-value of EPS rises with falling temperatures. When the average temp through a outer layer of Type-II EPS is 25F (as it will be, for much of January, at the cold edge of zone 5 ) it's performance will be about R4.7/inch, which will outperform polyiso at the same thickness. At an average foam-layer temp of 40F (the way it might be in March) EPS will only run R4.5/inch, and underperform polyiso- the polyiso performance slews pretty quickly with rising temps in that range. But it's the performance of the 8-10 coldest weeks that matter most from a dew point control point of view.
Ryland,
For answers to your questions, and lots of other information, you may want to read this article: How to Build an Insulated Cathedral Ceiling.
Q. "If the uppermost insulation layer is denim does there need to be anything between the denim and the ventilation channel?"
A. In most locations, air-permeable insulation like denim needs an air barrier on all six sides. In this location, the performance of the insulation will be degraded by wind-washing unless you include an air barrier above the batts. You can either use a site-built baffle or a commercial product like AccuVent. There is more information on this topic in the article I linked to.
Q. "In an assembly with ventilation channel above the sheathing is a 1.5" ventilation channel sufficient or is the minimum also 2" for this type of assembly?"
A. A 1 1/2 inch high channel is sufficient.
Hi Martin,
Thank you so much for your patience and all the information. Over the last few months I have read the article you linked to several times (which is extremely informative), but this is a challenging project in several ways so I've been doing my best to verify all the details. Is plywood an acceptable air barrier for a site-built baffle, or does the material need to be more (or less) vapor permeable? Are there other materials that would be suitable for a site-built baffle? Thanks again,
Ryland
Hi Dana,
Thank you for your in-depth explanation of the environmental impacts of both EPS and XPS insulation. I was having trouble ordering an adequate supply of the Polyiso insulation from our local building supply store so thought 3 layers of 1.5" XPS might be a simpler (and quicker) solution. However, after reading your post I decided to be patient and just go with the 4" of Polyiso on the unvented assembly even if it delays progress on the roof. Any delays we incur seem minimal compared to the environmental impacts of several hundred sheets of XPS insulation.
Ryland
Ryland,
Q. "Is plywood an acceptable air barrier for a site-built baffle?"
A. The answer can be found in the article I linked to (How to Build an Insulated Cathedral Ceiling), where I wrote:
"Some builders ... make their own site-built baffles. ... A vent space can be created by installing 1 inch by 1 inch 'sticks' in the upper corners of each rafter bay, followed by stiff cardboard, thin plywood, OSB, fiberboard sheathing, or panels of rigid foam insulation. (If you use rigid foam for your baffles, it probably makes more sense to choose thin EPS or XPS rather than foil-faced polyisocyanurate, to allow a bit of outward drying, however slow, by diffusion. A thin layer of EPS or XPS is somewhat vapor-permeable, while foil facing is a vapor barrier.)"
Ryland, one small point, don't know if you are a long time builder or have an expert very experienced crew working on your project. But, goof up your build even slightly and you will be spending lots of extra money for a poor insulation job to even setting up your roof for moisture failures here and there, etc.
Such a big project to me should not be done by a crew that has never done it without at least pulling someone who has years of experience to guide and check your work. Or go visit a site being done right... etc.
Your roof is complex, your solutions for insulating are too. Hope you don't reverse an idea accidentally.
Foam done right works, done wrong, well we end up replacing them up this way. We always build vented rigid foam roofs and unvented spray foam roofs.
Hi AJ,
I interviewed many contractors for the job from the area and there was literally no one who was qualified to oversee the work. One insulation contractor, who apparently does many homes in this area, recommended 1/2"-1" of spray foam to air seal and then partially fill the rest of the cavity with blown cellulose (essentially flash & batt). When I told him the 1/2"-1" of insulation was not sufficient to stop interior condensation in this climate he seemed rather offended and claimed he's done this many times (and on his own house) and has not had any problems. This was generally the attitude of every contractor I met in the area - and people from Chicago are not willing to travel 2 hours out of the city to work on our house. So, unfortunately, my only option has been to become as educated about everything as possible and work with our very experienced carpenter who is willing (unlike many people) to take the time to do everything right.
I've attached the semi-final drawings of the roof assemblies - if you have the time and see anything incorrect please let me know. Thank you for your concern and assistance,
Ryland
Hi Martin,
Thank you for the quote from the mentioned article, I must have been focused on the section about unvented assemblies when I read the article the last few times. Thank you again for your time and assistance, it's greatly appreciated. I've attached rough diagrams of the proposed assemblies in a previous message - if you have extra time and notice any errors I would appreciate any feedback you might have.
Ryland
I do vented rigid foam like you propose.
I do NOT do unvented rigid foam.
I do NOT do cut and cobble.
Good luck with your build Ryland.
aj
In the vented assembly of your last posts, if the interior-side polyiso is covering the rafter edges (not cut'n'cobbled) it will perform at it's full rated R, and there should be no moisture issues, as long as you have sufficient roof pitch and vent gap. Code demands only 1" of ventilation gap minimum but 1.5" or even 2" would be better. With foil faced polyiso it means you're be locked into a vented assembly for the duration- you can't seal it up and add exterior foam later without it becoming a moisture trap.
On the unvented assembly drawing having 4" between the batt and the gypsum will undercut performance. Some sort of air-barrier needs to be snug to the fiber to limit convection moisture transport adequately.