Re-roof insulation upgrades for a cathedral ceiling with polyethylene
Hello GBA Community,
I have a re-roof coming up. What are my options for improving the insulation in my cathedral ceiling that has an interior poly vapor barrier? I’d like to do this work without tearing up the internal finish.
My house is an 1850s carriage house in the greater Boston area (zone 5). Around year 2000 it was converted into a condo. All the interior details are modern (drywall, Romex, copper pipes, etc.), but the structure is old. The roof is gable with a cupola; it also has a dormer on one side and two skylights on the other. I believe the rafters are 2×8, and I suspect the roof deck is boards. I know there are fiberglass batts between the rafters, but I don’t know the R value. IR imagery shows that most of the bays are reasonably insulated, but there are a couple cold spots. Thermal bridging is also evident. There are 4 large exposed beams acting as collar ties. I also strongly suspect that I have a poly vapor barrier under the drywall (I know it’s present in the walls).
I’d like to add some continuous insulation when we redo the roof. Ideally I’d add enough reclaimed XPS to come up to the code minimum R-20 above the roof deck, but I am worried about forming a vapor sandwich.
Q1. Is there a safe way to add exterior insulation to this assembly?
Q2. Does the permeance of XPS or EPS foam allow sufficient vapor diffusion to not cause problems assuming the drywall installation is average quality.
Q3. If foam is unacceptable in this assembly, is there a mineral wool product that I can use on a steep roof instead?
Q4. If foam is acceptable, would it be advisable to add a peel and stick vapor permeable air barrier such as Delta-Vent SA under the foam? I am thinking I’d air seal the roof deck with a vapor open membrane, put the foam on top, add 2x4s on the flat and additional sheathing over that. Would this be a dangerous assembly?
This community has been a great resource.
Thank you very much
Pavel
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Pavel,
Q. "Is there a safe way to add exterior insulation to this assembly?"
A. Assuming that the roof shows no signs of moisture problems now, and assuming that you do not discover any signs of damp or rotten sheathing when you remove the existing roofing, I think that you can safely plan to install R-20 rigid foam above the roof sheathing without worrying about the polyethylene.
Here is a link to an article that explains more: How to Install Rigid Foam On Top of Roof Sheathing.
Q. "Does the permeance of XPS or EPS foam allow sufficient vapor diffusion to not cause problems assuming the drywall installation is average quality?"
A. No. The safety of this assembly comes from the airtightness of the air barrier you need to install above the existing roof sheathing, along with the R-value of the rigid foam. R-20 rigid foam isn't very permeable, so the actual vapor permeance value of the rigid foam is irrelevant.
Q. "If foam is unacceptable in this assembly, is there a mineral wool product that I can use on a steep roof instead?"
A. You could install one of the denser available mineral wool products if you want, followed by another layer of roof sheathing and new roofing. This type of roof assembly works, but it is rarely done in the U.S. -- so you can expect push-back from contractors and roofers.
Q. "If foam is acceptable, would it be advisable to add a peel-and-stick vapor permeable air barrier such as Delta-Vent SA under the foam?"
A. A peel-and-stick air barrier is fine, but paying extra for a vapor-permeable product is a waste of money. You won't get any outward drying from this assembly.
Q. "I am thinking I'd air seal the roof deck with a vapor open membrane, put the foam on top, add 2x4s on the flat and additional sheathing over that. Would this be a dangerous assembly?"
A. The added vent space above the rigid foam is fine, but as I said you still don't need the peel-and-stick air barrier to be vapor-permeable.
Pavel. Consider using reclaimed foam (from Insulation Depot or similar providers). It is much less expensive than new material and more environmentally friendly.
Independently of what the labeling says, don't count on XPS to perform any better than R4.2/inch over the long term, so you'll need at least 5", not 4" to hit R20. This is particularly important when using reclaimed XPS.
In US climate zone 5 long as at least 40% of the total R is in the exterior foam (properly derated for age & type) the risk is pretty low, but 50% would be even better. But you need to find out the total depth of the cavity to estmate the R-value of the cavity fill. If they're ~7" deep and filled with mid-density fiberglass you'll be fine with R20, but if it's deeper you'll need more foam-R. The prescriptive R20 is predicated on an R49 (code min) total. And while you may have an class-I vapor barrier on the interior, you don't have a drying path. The only way to guarantee that the roof deck stays sufficiently dry over the long haul is to keep it sufficiently warm, which means a minimum of 40% of the R value has to be above the roof deck.
Thank you all for your responses.
Martin, I don't know about any existing water damage, but the reason I'm thinking about the roof now is because there is a small water leak where one of the collar tie beams intercepts the drywall. The leak only happens when it rains, so I don't think it's condensation related. The current roof has a ridge vent, soffit vents and foam ventilation baffles above the fiberglass. So I am not sure how this assembly will do if I add foam and close off the vents. In other words, I am not sure how airtight the drywall is currently.
The article you link, and every other article I've read on the topic, says that unvented assemblies need to dry to the inside. There is always some disclaimer that there should be no vapor barrier on the inside. I am a little concerned that although the proposed suggestions "should" work, I might still become a building science experiment. I am mostly interested in using the reroof as an opportunity to improve the thermal envelope, but our house is small, and heating bills are not very high, so if there is any appreciable risk of this becoming an adventure down the road, maybe I should forego the added insulation?
"A. No. The safety of this assembly comes from the airtightness of the air barrier you need to install above the existing roof sheathing" Does that mean that I absolutely have to use a peel-and-stick product on the roof deck? If I understand you correctly, the idea is to limit air movement from the inside with airtight drywall, but also to encapsulate the assembly from above? Once again, I've seen horror stories with peel-and-stick trapping moisture. Am I missing something here that makes this assembly absolutely safe, or is this another case of "it should work"?
Steven, yes, I will try to use reclaimed foam if we go this route. This is helpful for both the wallet and the planet.
Dana, Thank you for your comment on the total R value needed. I will aim for 5" or more if we go this route. I wonder if there is a way to estimate the existing R value using thermal imagery? I can imagine getting the temperature in the rafter bays as compared to the thermal bridges from the rafters themselves. Then presumably I should be able to figure out the insulation value given some reasonable assumptions on wood thermal conductivity. Do folks have experience with this?
Once again, thanks for the advice.
Pavel
Pavel,
It's impossible to remove all risk from the work you are contemplating. This is a judgment call.
Here's my judgment: To address 90% to 95% of cathedral ceiling problems, you need to have an airtight ceiling. If you can feel fairly confident that you have achieved an airtight ceiling, you have done most of the required work to limit condensation problems. The airtightness of your ceiling can be tested with a blower door.
Air leaks are tricky, though. If there is a hidden air leakage path that you haven't anticipated, there is always some risk of condensation problems.
The rigid foam above the roof sheathing goes a long way to addressing the remaining risk, because the rigid foam will keep the sheathing warm. If there aren't any cold surfaces, interior air leaks won't lead to condensation.
If you do a competent job of sealing the air leaks in your ceiling, and you do a competent job of installing the rigid foam (and, of course, sealing your soffit vents and ridge vent), I think the risk of this approach is low.
-- Martin Holladay
Honestly I would tear out the ceiling drywall here. That way you could:
1. Remove the poly
2. Remove the existing fiberglass batts
3. Remove the soon-to-be-unnecessary ventilation channels and plug up the roffit and ridge vents from inside to make sure they're really sealed
4. Find the exact location of the leak you mentioned
5. Perfectly air-seal and re-insulate the assembly from the inside
If your rafters are deep enough and you can get enough R-value between them, you might not even need to add any outside foam in the first place! If you can afford a slight loss of ceiling height, you can actually put an inch or two of foam under the rafters--between them and the new drywall. This gives the roof some continuous insulation but preserves a 100% outside drying path, which you currently have. To keep that outside drying path, you'd want to use a vapor-permeable underlayment and above-sheathing ventilation channels.
But you can always add foam on the outside instead if you want, remembering the ratios of inside insulation to outside insulation.
Thank you all for the helpful and insightful comments.
Since my last post I've learned some additional information. First, there is a chance that there might not be any polyethylene after all. I removed a smoke detector and didn't see any plastic. I don't know if the same holds for the rest of the ceiling, but I am hopeful. The rafters appear to be 2x8 and cavities are filled with Kraft faced fiberglass.
I had a roofer come out today to take measurements for an estimate, and he suggested that we use Hunter Cool-Vent panels (although he's also willing to put down reclaimed XPS and build up the new deck). I looked up this product, and it's a layer of PolyIso attached to wooden spacers with OSB sheathing on top. This seems to be exactly what I need, but I have some questions:
1. Have folks on this forum used this product successfully?
2. Should I be worried about R value reduction in cold temperatures? I am in zone 5 and I would spec R20 for Cool-Vent
3. Are there any benefits to using a built-up XPS + sleepers + sheathing assembly instead?
4. Any idea if I should expect a significant cost difference between the two? It seems that one has a higher material cost, but the other has more labor involved.
Thanks for your time!
Pavel
Pavel,
Q. "Have folks on this forum used Hunter Cool-Vent panels successfully?"
A. Yes. The product is mentioned in my article, How to Build an Insulated Cathedral Ceiling. In that article, I wrote, "It's also possible to purchase nailbase (a type of SIP with OSB on one side instead of two) that includes integrated ventilation channels between the OSB and the rigid foam; one brand of these panels is Cool-Vent from Hunter Panels."
Q. "Should I be worried about R-value reduction in cold temperatures? I am in zone 5 and I would spec R-20 for Cool-Vent."
A. In fact, it's a good idea to assume that the polyiso has an R-value of R-5 per inch. You can always add a layer of continuous rigid foam under the Hunter Cool-Vent panel to improve the R-value of the rigid foam layer.
Q. "Are there any benefits to using a built-up XPS + sleepers + sheathing assembly instead?"
A. Most green builders avoid the use of XPS rigid foam for environmental reasons. For more information, see Choosing Rigid Foam.
Q. "Any idea if I should expect a significant cost difference between the two? It seems that one has a higher material cost, but the other has more labor involved."
A. Only you can perform those calculations. By far the cheapest approach is to install recycled (reclaimed) rigid foam.