The example of the failed roof with the polyiso on the interior.
GBA Editor
| Posted in Energy Efficiency and Durability on
It appears this sandwich had fiberglass in the roof cavity?
I was thinking of using 1″ poly iso on a cathederal ceiling to give a thermal break to the framing. Spraying poly iso 4″ closed cell 2# foam in the rafter bay. Then adding the additional inch of sheet foil faced on the interior. With the closed cell we have never had a moisture problem occur. Our location is north eastern Pennsylvania. The roof system has ice and water shield on it. Our feeling is that the closed cell foam doesn’t let any air leaks occur, keeping moisture out of the interior to exterior sandwich.
Feed back please.
Thanks,
Joe Billingham
Billingham Built
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
Joe,
If you can avoid air leaks, your proposed system will probably work -- although it would certainly be better to omit the Ice & Water Shield.
Problems usually arise because of air leaks at the ridge, which is often imperfectly sealed (or which develops air leaks due to truss or rafter shrinkage). If air leaks out the ridge, indoor air can easily be drawn into the ceiling assembly around cracks at partition top plates or (I hope you don't have any of these) recessed can lights. Then you get condensation and rot.
So -- your proposed assembly will work if everything is perfect. If something isn't perfect, it's a little risky.
Unfortunately, hygro-thermal dynamics don't follow "feelings" but rather the laws of physics. And we must never forget that Murphy's Laws also apply to all human endeavor.
The roof assembly you describe would work perfectly if the framing was also plastic. Burying wood between impermeable layers is never a wise approach. Burying manufactured wood members in such a "sandwich" is an even more foolish approach.
Would still like to know if it was a fiberglass filled cavity.
The sprayed in place foam provides a seal to all members. Lights would be within the sealed cavity to reduce hot spots and ice daming.
We have used spray foam in roofs for 30 years but have never tried to give a thermal break on the ceiling rafters. We have used the thermal break on walls with the spray foam in the stud cavity. We're hoping increase our thermal envlope values in the roof
Joe,
If you are asking about the photo on this page:
https://www.greenbuildingadvisor.com/blogs/dept/building-science/vapor-profiles-help-predict-whether-wall-can-dry
The answer is, the rafter bays were filled with cellulose.
There are two ways to tell:
1. If you click on the picture and then enlarge it (by clicking the plus symbol in the magnifying glass in the lower right-hand corner), you can clearly see the clumps of cellulose.
2. Peter Yost explained more about the roof assembly, including the fact that cellulose was used, in his post on the same page. The relevant post is dated Aug. 11, 2010, 6:54 AM EDT.
Joe, of all your foamed jobs have you been able to go back and visually inspect how the product has performed over the years? How quickly do you usually cover up the insulation in your projects? The reason I bring this up is- I had the roof deck in my own house spray foamed so I could finish-out the attic and then the project was put on hold. When I went back to install the drywall (about a year later) and the foam had pulled away from several of the rafters. The gaps could have easily gone unnoticed if it weren't for my paranoia and "under the magnifying glass" inspection, but several locations were close to a 1/16" gap along the entire rafter. Fortunately I had the chance to foam fill all of those gaps, but it could have added up to quite a bit of leakage gaps. I can't say for sure that it was the foam curing/shrinking or the rafters drying and shrinking, or a combination of the two.
The foam insulation does perform very well, but we can't assume any installation will be 100% perfect, especially in a cathedral ceiling situation where the foam isn't continuous. I would be concerned with creating the 'sandwich' in any case. At least let it dry in one direction. I chose to dry both directions due to my mixed-humid climate and so far so good.
Hunter,
Your observations present the missing link in the spray-foam-prevents-air-movement gospel. Yes foam can shrink, both during curing and over time (all plastics lose their plasticizers as they age). Yes wood does shrink (and swell and move), both initially and cyclically.
And yes, even "perfect" air/vapor barriers (like the foil-faced polyiso in the picture Martin linked to), are never perfect at installation and become less perfect over time.
All buildings leak - both air and water. That's the assumption upon which all hygro-thermal decisions must be based. The black mold and rot in the picture is high at the hip and ridge and in the upper part of the hip rafter, so it's not from a roof leak but from air-borne moisture rising up into the "perfectly sealed" roof assembly.
In addition to being a designer/builder for more than 30 years, and an instructor in hygro-thermal engineering and sustainable design, I was also a town emergency manager. There is a risk/consequence matrix (http://www.riskagenda.com/images/bite6big.gif) that we use to plan our emergency response and mitigation strategies. Highest priority is given in planning for high probabliity/high consequence events, but consideration is also give to low probability/high consequence events.
If you think it's improbable that air or moisture could leak through a material, but the consequences of such a leak would be disastrous, then plan for failure. To do anything less is irresponsible.
Hello Joe- Brooke Rush is doing a rebuild of a leaky (air-wise) cathedral roof assembly right now in our area. He is incorporating closed cell spray foam and XPS in the system, with a huge amount of attention to detail. I do believe he has documented this project to share/present and he has a couple of lessons learned applicable to our area. contact me at Shelly's Perkasie or call Brooke and tell him I sent you.
This roof detail example makes me wonder how a a designer should have designed that cathedral ceiling so it shouldn't fail, in the immediate future... and in the long term. What if the roof had been specified with tar paper and asphalt shingles instead of modified bituminous membrane and shingles? The roof would have been able to dry to the exterior and probably no problem?
But in twenty years when the home-owner needs to replace the shingles, to ensure that the roof can't ever possibly leak, he might splurge and opt for modified bituminous membrane as an underlay. Would the roofing contractor know the building science to advise the customer that the self adhesive membrane would be a bad idea. The roofing contractor would have to know that there was rigid poly-iso board under the drywall... and know his building science to offer that advice.
That's not very likely!
So how do you design a roof detail that is resilient enough to overcome ignorance?
Andrew
Vent it.
Andrew,
Asphalt shingles without Ice and Water Shield don't provide enough drying to prevent this problem.
The best solution would have been to insulate the roof above the roof sheathing with rigid foam.
Not true.
"Asphalt shingle roof systems are vapor permeable because joints between individual shingles allow vapor to pass through." - American Society of Home Inspectors
#15 felt has a perm ranging from 6 (dry) to 31 (wet) - BSC Building Materials Property Table.
A shingled roof over #15 felt underlayment allows drying to the outside. Hi-tech polymeric underlayments and bituminous self-adhering membranes limit or prevent outward drying.
According to a 2009 report in the ASTM Journal, "Compact Asphalt Shingle Roof Systems: Should They be Vented?":
 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 (self-repairing adhesive membrane) 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
 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.
 Unvented permeable shingled roofs are also a viable consideration in hot humid climates, although it would be slightly less durable. These unvented roofs result in increased drying time of the inboard gypsum wallboard. The insulation must remain permeable to avoid a vapor trap.
There is never one "best" solution, and this suggestion would result in the same moisture trapping as the Ice and Water Shield, the only difference being that if any water did accumulate in the sheathing it would be warmer and more vulnerable to mold and rot.
I've been using the foam for a least 30 years. In roof conditions we tipically foamed directly to the roof sheathing. In northeastern PA using 3" of 2# closed cell we have never had any issues. The roofs last longer. Never have any ice daming. We have never experienced "pull away" from the framing.
My questions arose because we recently read how 30% of a framed surface is in contact with the exterior. In hoping to cut this down in the roof of an addition we're planning we were thinking of using the 1" polyiso on the interior ceiling under the sheet rock. Then we saw this article that brought up all the questions.
I will contact Brooke Rush. I know he has used the foam for a long time.
Joe Billingham
The addition of a thermal break such as the Polyiso board over the studs on the inside is a good idea it will increase the systems r-value and performance. As long as you use enough closed cell foam to keep the condensation point within the closed cell foam at large delta-t's the system will be fine. Using closed cell cell foam along with an air permeable insulation such as cellulose can result in moisture problems if the system is not engineered correctly. Hence the rotted roof in the picture, the polyiso board is a vapor barrier and so is the ice and water shield. The moisture would not have been able to get through the closed cell foam if it had been installed to cause this problem. You can have shrinkage of any closed cell foam and this is most likely caused by installer error. Closed cell foam is very sensitive to how it is installed in order to obtain the physical properties the manufacturers state the material must be manufactured on site properly. Always ask your installer for their certification. If they don't have a certification to spray foam you shouldn't use them. Keep the ice and watershield if you want but the roof is better off with felt paper. Too many layers of plastic with the sandwhich approach to make it durable. Use the polyiso board on the inside for a thermal break just don't tape all the joints, you already have an air barrier with the closed cell foam. If your worried about the continuity of the air barrier ask for a Q/A program to check the installers work. Even the best installer can miss a spot.
Steve Shaw
One of the primary advantages of a layer of foil-faced polyiso interior to the rafters is that it can create a near-perfect air barrier with taped seams and foamed edges (assuming to can lights or other unsealed penetrations).
It's foolish to do this installation without sealing the joints, since it's not going to allow any appreciable inward drying with unsealed joints. But the effect of this thermal break will be to lower the temperature of the framing which, with only 3" of spray foam under the sheathing, may have more than 50% of the side surfaces exposed to potential air-borne moisture with the risk of condensation and negligible drying potential.
Spray foam may, if properly mixed and installed at the correct temperature, be an air barrier material, but it doesn't necessarily create an air barrier assembly when it's interrupted every 16" with framing. An air barrier requires continuity on all six sides of a structure.
Martin, I recall from another post that you have used or advocated for polyiso sheets on the inside of a cathedral ceiling. What then is the remainder of the assembly? Cellulose between rafters, followed by sheathing and roofing? My understanding is that the least permeable layer should be on the outside for cold climates, so that the inner face of the poly-iso will be fairly warm if moist air gets in the cavity, and thus won't condense. It depends on how much of vapor barrier the poly-iso is, and if there is another VB in the assembly...
Chris,
Here's a good cathedral ceiling assembly:
-- A ventilated channel under the roof sheathing. The best ventilation chutes are site-built, using rigid material attached to battens in the corners of the rafter bays.
-- Dense-packed cellulose.
-- Foil-faced rigid polyiso -- ideally thick enough to increase the R-value of the assembly to R-50 or R-60.
-- 1x3 or 1x4 strapping, screwed through the polyiso to the rafters.
-- Drywall.
What does the strapping accomplish over the polyiso besides and airspace?
Chris,
The polyiso and strapping go up first, before the dense-packed cellulose. (The cellulose installer makes holes in the polyiso for his hose, and later patches the holes.) The strapping does three things:
-- it holds the polyiso in place so that the polyiso doesn't pop off when the cellulose is blown in the cavities;
-- if the polyiso has a foil facing, the air space adds R-value to the ceiling assembly;
-- the strapping makes it much easier for the drywall installers to hang rock.
Martin,
I was going to spray 4" of closed cell foam. Have never had any issue in 30+ years of using it. I am going for the additional thermal framing break. For that, on the interior I wanted to add the 1" foil faced to the bottom of the rafters with the strapping for the sheetrock.
I've read nothing in all the comments or my experience that has shown this might be a problem
Joe, then you either didn't read the study results I posted or are choosing to ignore them because of a bias for non-breatheable materials such as bituthene and closed-cell foam or are foolishly assuming that you can do a perfect job that remains perfect for the life of the house with zero possibility for leakage or moisture accumulation.
"The least tolerant roof assembly in either climate is the unvented closed-cell polyurethane insulation roof assembly with SRAM (self-repairing adhesive membrane) 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"
I've had no issue in 30 years and we are improbing the sealing way beyond what we did 30 years ago. The exteriors are foam sheathed, taped and sealed with channeled air screen under the siding. I want to still improve the framing transfer in the ceilings. With the sealing of the 1" polyiso seams as you suggest, an additional furred surface air barrier on the interior I see this as a way to improve the thermal faming break in the roof.
Theres lots of discussion in the middle, but anted to say that the roof proposed by the OP will work fine, without issue. Except that the "thermal break" layer is pretty much worthless. You will have two thermal layers seperated by an airspace that will by difficult to keep airtight. Now factor in that the sides of the rafters are unisulated in this airspace. Seems like a lot of work for almost no benefit.
The foam against the roof is hard to screw up, and somewhat idiot proof.
Now, just wanted to point out to all that the picture in the referenced article head is NOT a cathedralized roof. It is in fact a small, cramped, open attic space ABOVE the cellulose insulation layer. I have not seen this roof assembly fail if it is cathedralized (densepacked between drywall and roof sheathing). I realize it is supposed to, but they haven't yet that I know of. And I live under one right now.
The roof above created a condensing surface with the airspace above. You can see the line where the cellulose stopped and framing (ceiling joists or collar ties) under the roof rafters.
I believe the cellulose has miracle curative powers sometimes because it buffers and wicks moisture out of many places where I should've seen a problem, but didn't.
-Rob
As to the side of the rafter--- It is insulated. The last rafter sits on the top plate. The 3" - 4" of foam covers the inside of the rafter, insulating it and sealing cracks. The exterior is in line with the exterior wall framing and is covered with 1" of foil faced polyiso.
The no vent above the cellulos from eve to ridge would be my guess of the problem sourse.
Joe - Agreed at the edge of the roof, I was speaking about the middle of the roof. The roof rafter will have 3 or 4" of foam, and then be exposed for 4 or 5 inches down to the thermal break layer.
-Rob
Now I understand. So my objective is to get some thermal break. The interior 1" polyiso applied to the bottom of the rafters is doiing that. It doesn't solve it 100%. Where it makes sense to put it on the exterior of walls, It's more labor and material to add on the exterior of the roof. I think under is the best. Remember for years we did nothing. Someone will come up with a product to do this. Why couldn't 1" polyiso be fabricated to the underside of a plywood panel? Glue it in place and tape the seams.
Joe
The picture being discussed in this thread is this cathedralized roof with interior thermal break and exterior impermeable membrane, a disaster waiting to happen with either fibrous or foam cavity insulation (it would have been worse if the sheathing had been OSB):
https://www.greenbuildingadvisor.com/sites/default/files/P6070001.JPG
Robert,
Yes, I completely agree. OSB would have been mush.
Joe,
Your "30+ years of no problems" is acceptable only if you have gone back and opened up these roof assemblies to prove there aren't any problems.
Have you kept in contact with the majority of the homeowners over the last 30+ years?
Have you noted the when the home changed ownership, and explained you were the contractor who installed the foam insulation to the new homeowners?
I doubt it.
Robert, The roof is not cathedralized., or at least not the part with the rot.
Perhaps it is a clipped or vaulted ceiling below, but the top is clearly an open attic.
The bottom apparently is cathedralized, but it looks fine otherwise.
The "cathedralization" appears to end at the romex penetration.
Without knowing what bypasses are in the ceiling below, venting may not have mattered.
-Rob
Thanks for providing the hi-res link. If you zoom in, you can see it's a stepped ceiling framed under the rafters.
I don't know why they even tore the bottom off. We usually repaired these from the top down. Keep ripping off sheets until you find something solid!
-Rob
An attic is an accessible space between the highest ceiling joists and the roof framing. By code, an attic must have a minimum of 30 SF of area and a minimum of 30" of headroom.
It's not at all clear from the picture whether that space, which seems to have been filled with cellulose, is an attic or simply a thickened part of the cathedralized roof, which I suspect. If it was full of insulation, then it certainly wasn't an attic.
Hi Robert,
I wasn't too bunched up with the code definition, more defining the characteristics. An open attic has condensing surfaces all over it. The densepacked portion would operate differently.
Because I used to own a cellulose insulation business doing new homes, I'm betting it was not filled to the ridge, that would be a lot of bags! But we are *really* at the point of speculation now.
-Rob
I am hoping someone on this thread can help me. I have talked to numerous local contractors, architects, & engineers but I am getting so many different answers we can't make a decision on it. We live in Central PA. We have a home built in 1969 with all cathedral ceilings. The ceilings are non-vented and we have 3" x 10" Hemlock beams 4' on center with 4'x8' stressed-skin panels on top of them of plywood, foil-faced fiberglass (both sides of fiberglass appear to be foil faced), then the roof deck. We have shingles on top. It is R-19 at best and the foil seems to be crumbling in most places we have been able to access. We would like to add insulation. We are down to 2 options...
1) Add 3" Polyiso Nailboard insulation to the roof deck and re-roof the house. Our main concern with this is adding the extra weight to the 3"x10" beams. Although they do not appear to be the least bit stressed now & it is not that much weight, they are spanning a good distance in 2 locations (24' length of beam)
2) Fill between the beams on the interior with a blown in fiberglass adding R-41 and then install a new ceiling to the bottom of the beams (we do not need to keep the beams aesthetics). This should not be adding as much weight to the roof system as it is from below...we think....
Any thoughts or info are greatly appreciated!
Lara
First I would like to say this is the most information I have been able to find on any type insulating that is not using a standard batts or spray foam. Thanks. Now to my question. I have a cape cod style third floor which I am currently finishing. The rafters are 2x10 . I have used 1 x 1 wood strips which I have fastened against the rafters butted up against the underside of the roof deck to allow for ventilation of the roof. I have then cut and fit 2 layers of polyiso board, which equals seven inches in thickness, inside the rafter cavities with the wood keeping the foam board 1" away from roof deck. When fitting the polyiso board, I left a 1/2" gap between either rafter and the edge of the foam board which I then filled with a closed cell foam to seal the cut edges of the polyiso board and to hold into place. My seams between the two layers are staggered more than six inches. This will leave approximately a 1/2" gap between the drywall and the face of the second piece of polyiso. Any comments on this would be appreciated.
Aaron, You did the work first and are asking how to do it second? Interesting methodology. No soup for you. Also, best to ask a new question sometimes. As to your answer, sealed up foam should do the job with the vent or without the vent. You may have gone to overkill. One caveat though, any error in your work could lead to disaster. High moisture by running humidifiers, watering a hundred indoor plants, having an indoor swimming pool or fish tank than holds ten thousand gallons... etc.
For a huge file of info, join the site at the pro level....
Cheers, it's time for an afternoon libation here, IPA of course.
As far as doing the work first and asking after, I have been researching this for over three months as well as contacting Dow about the polyisoboard and trying to find the dew point depth. I happened to come across this site and decided to post the question. I am halfway done with the installation. The proposed installation is on the roof only, the walls are standard batt. My focus was to try and eliminate as much of the exterior heat from the roof transferring through to the living space.
Aaron,
It sounds like your question is: "how thick should my rigid foam be to keep the roof sheathing above the dew point during the winter?"
The answer depends on your climate. For a full explanation, see How to Build an Insulated Cathedral Ceiling.