We recently placed 3.5" of RMax foil-faced polyiso foam board (air and vapor impermeable) neatly cut and placed between the rafters of our cathedralized ceiling home in the Santa Cruz Mountains of California (hot dry summer, very wet and rainy winters, never much below freezing). Home is roughly 80 years old, built originally as summer cabin (very leaky, etc.). The roof is unvented, medium color asphalt shingles in good condition, with no roof leaks after three years living in the home. Roof deck is tongue-and-groove 1x8 pine, painted from below as a finished ceiling several years ago.

3-days after installing and sealing the perimeter of each sheet with can foam, I felt water drip on me while working at home mid-day (~2pm). I started paying attention, and we got many, many more drips, right below the ridge blocking. This occurs only in mid-afternoon from roughly 1pm to 3pm on hot sunny days, and is most evident on the South-sloped roof surface. This is a handful of drops, not a deluge of water, but still clearly unexpected and not good. 

We opened up two south-facing panels yesterday at 2pm. We found a completely dry and hot roof deck, and completely soaked ridge blocking and structural metal hardware. The ridge was beaded up all over with condensation, and the drops I observed were coming from the Simpson hardware at the ridge. I count myself lucky to have seen this issue prior to sheetrocking the whole home. I've added some photos below to show context, install quality, and observed condensation (note, this shows beaded water on the hardware, but it also covered the entire ridge block, just hard to see when painted white). 

I have a clear hypothesis of what's happening in my mind, but I want to know what others experience has been with this type of assembly and insulation strategy. Is it common in your region to cut and place foam board between rafters in an unvented roof assembly? It meets the letter of the model codes and local CA codes.

Have you ever experienced or heard of this issue, specifically condensation on hot summer days in unvented roofs insulated with polyiso or other foam boards? Or moisture stained sheetrock or rotted ridges in such roofs. (Note: I'm very familiar with moisture issues in these roofs in general, but am curious about this group's experience with this specific strategy).    

I also have a sense of how to fix it, but what would this experienced crew suggest? If removed, what would you recommend replacing this insulation with, prior to sheetrocking? 




You need to be a member of Building Performance Community to add comments!

Join Building Performance Community

Email me when people reply –


  • Hi Brennan,

    Physicist Niels Bohr is credited as saying “An expert is someone who has made all the mistakes which can be made in a very narrow field.”  Sounds like you’ve moved even one step closer to expert status with this “experiment”, congratulations!  

    Many good hypotheses have already been shared by you and others, and I agree that Kohta’s advice is particularly spot on. 

    I did want to inquire about the connection at the exterior walls. Can you provide a photo with a piece of the foam removed?  

    I’m curious if the wall plate to exterior sheathing was meticulously sealed to eliminate the connection of the new roof cavity (formed by the addition of foam) to the wall cavity? If not, I can envision the South facing exterior wall heating up in the afternoon and pumping moisture into the new roof cavity.  

    As has been mentioned, the most saturated/buoyant air ends up at the ridge where it condenses on the North facing, thermally conductive materials that are slightly below the 100+ F dew point.

    I hope you’ll get the opportunity to data log temperatures of the various components and the wall and roof cavity RH, and present your experience at the Dry Climate Forum in February.  

    Because when these things occur in fully code-compliant approaches, especially to exceptionally smart people, we are reminded how complicated energy efficiency is, and the high value of expertise. Thanks for helping us all think, learn and improve.  

    Mike MacFarland

  • We've all seen snow-covered roofs where the snow has melted the lines of the rafters and the bays are still cold.  Wood, per inch, is a poor insulator compared to manufactured materials.  If you could live with an architectural compromise, you could install collar framing to create a narrow flat ceiling along the ridge, which could be substantially insulated.  I do not recall if anyone has suggested a vapor barrier, and I am not talking about 4 mil poly.  I have also heard of folks who stripped the shingles, furred out the rafters and constructed a new roof deck on top of the old.  With soffit and ridge ventilation, the delta t would be reduced.

    • So, here's my thinking on furring down the ridge and insulating a bit from underneath. First, it might keep the ridge above the dew point, but it could very well remain in mold growth territory above 80% RH. Second, even if the ridge was protected, the roof deck itself, once the sun stops shining on it, is going to drop below the dew point pretty rapidly, probably more quickly than the water vapor will be reabsorbed into the wood. That may still be problematic.
      On a separate point, I really have to re-emphasize that I don't think the living space air is the source of the moisture, instead it's being driven out of the heated roof deck into the tiny air space above the foam. I think this would be happening irrespective of the quality of the seal at the interior face. Staggered seams, taped edges, sheetrock; doesn't change the moisture emissions from wood heated to 175F. Furthermore, if it was a moisture from below issue, we would probably see this happening in the bathroom or in bedrooms, but instead it's occurring in the vacant living room with a South roof orientation. Very interesting stuff.
      I think the vapor diffusion vents mentioned by Kohta would help. I do believe that periodic sealing along the back side of the foam along the roof slope would help limit this moisture migration towards the ridge, if that airspace was periodically interrupted. In the end, I've gone the cheater's path, and we've ripped out the foam board and replaced it with closed-cell spray foam applied from underneath. Absent a roof leak from above, we should be in pretty decent shape.
      But let's keep this discussion going for the sake of interesting and useful discussions.
      • I am confused.  What is the source of the moisture in your roof assembly?  Theoretically, it should “boil off” and your roof would stabilize.  If you think it is from rain, fog or dew you do have roofing problems.  Your explanation of the moisture path seems improbable.  The ridge is the thermal weak link, and at the highest elevation, much more so than water producing rooms because they are thermally balanced.  I did a quick calculation that a single human being will produce 264 liters of water vapor per day…Add in your daily activities and guess how much moisture is generated.  I would be interested in recording the outdoor and indoor relative humidities over time, with so many air changes per day, they should be the same, or that your home holds more moisture.

    • Stripping a roof just to add a ventilation space is only an attempt at fixing the symptom. If you are going to strip the roofing, then add at least a few inches of rigid insulation first. With adequate vapor impermeable insulation covering the entire roof, there is no need for ventilation.

      Consider custom SIPs panels for that insulation. They go up quicker, come with roof deck ready for roofing and blocking can be included at perimeter for fascia trim.

      • Yes & No - just remember that anyone in a region with a 50# PSF or greater snow load they should have ventilation as snow itself is a good insulator 

        • Yes & No - We ventilate a roof to allow moisture to escape. In a well insulated roof with a vapor impermeable rigid insulation (warning- not all rigid insulation is a Class I vapor retarder!) vapor escaping is not a concern, but heat loss is. Snow has an R-value somewhere around R-1 per inch, depending on moisture content. With 2 feet of snow on a modern roof (with a minimum of R-49 insulation, including at least several inches of the above noted rigid insulation or SPF), the temperature will decrease across the roof and snow. At a certain temperature difference, dependent on ratio of R-values, the temperature at the interface of the snow and the roof can be at or above the freezing point, resulting in snow melt at the roof.

          The solution for that would be a cold roof, where there is at least a 2" air space at the roof deck, to minimize the heat transfer by conduction. There will still be some heat loss (because we can never completely stop heat flow), but it will be by convection and radiation.

          • The discussion seems to be devolving from the central problem of condensation along the ridge. I would appreciate an endorsement of a narrow ceiling along the top of the room, to isolate and insulate the thermal bridging.  Second, while attempts to air-seal all joints of the ridged insulation are noted, the best solution is to stop moisture at the interior source, below the sheetrock, with an engineered vapor barrier.  Stop moisture and reduce the Delta T.

            My double roof model is very expensive and wasteful.  It has nothing to do with ventilation as it seeks to reduce the roof temperature while adding insulation to the roof system.  As you recall, the owner has about R-21 in his ceiling, so adding rigid insulation on the existing roof deck would be helpful, but keep the airspace.  Let’s return to the problem and help Brennen Less solve his issue with his delightful cabin in the woods.

  • Anybody want to bet that if that closed cell layer had been installed on top of the roof deck

    that we wouldn't be having this discussion?

    Charles Bennett

    Mr Insulation;  San Rafael, Ca

  • The thin application of expanding foam along the perimeter has a much lower R-value than the foam board. There are also small gaps between the foam board and the roof deck as others have pointed out.

    When I install with foam board and expanding foam, I always seal the back side of the foam board as well as the inside. Cutting the foam board a little smaller (1/2" smaller than the cavity at least) helps get the foam to fill those side gaps. Sealing the sides is obviously important, but each piece needs to be sealed periodically across the cavity to prevent any type of convection happening in that tiny space.

    I wish there were a good article on how convection can happen in these small spaces and how it can impact the performance.

This reply was deleted.