GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Guest Blogs

An Insulated Cathedral Ceiling for a European Passivhaus

The rafter bays of the Golcar Passivhaus are insulated with fiberglass batts, and a continuous interior layer of rigid foam reduces thermal bridging

We used 12-inch-deep I-joists as rafters to make sure that we would have plenty of room for insulation. Pieces of rigid insulation were attached to the sides of the I-joists to make the sides flush.
Image Credit: All photos: Golcar Passivhaus
View Gallery 7 images
We used 12-inch-deep I-joists as rafters to make sure that we would have plenty of room for insulation. Pieces of rigid insulation were attached to the sides of the I-joists to make the sides flush.
Image Credit: All photos: Golcar Passivhaus
The house as a structural ridge beam made of steel. The blue Solitex membrane is the exterior air barrier. Above the Solitex are battens and counter-battens. This detail drawing shows the different layers of the roof assembly. Short 2x4 rafter tails were scabbed onto the eaves. The scabbed-on rafters are on the exterior side of the building's thermal envelope. The building has two roof layers. Although skylights provide welcome natural light, they are weak links in the home's thermal envelope.

Editor’s note: The author, Bill Butcher, is a Certified Passive House Consultant who was project leader on the Denby Dale Passivhaus and is currently leading the Golcar Passivhaus project, a 2,700-square foot three-bedroom house in Golcar, West Yorkshire, England. Some of the technical and construction-related terms used in Bill’s essay have been translated from British English to American English.

At the Golcar Passivhaus, we’ve gone for an insulated roof assembly to create extra space in the attic area, as opposed to the uninsulated roof we created for the Denby Dale project. You get the advantage of having the extra space and room, but inevitably an insulated roof assembly is the more expensive option, as it is a more complicated build.

The roof detailing at the Golcar Passivhaus is in fact similar to that at our Stirley Farm EnerPHit project.

Structural stability

We have used steel structural ridge beams to carry the main load of building (see Image #2, below). These steel beams are inside the thermal envelope (so they do not create a “thermal bridge”).

I-joists for rafters

For rafters, we used a slightly smaller I-joists than we used at Stirley. In this case the I-joists are 302 mm (11.9 inches) deep. We have improved our I-joist detailing from the recent Stirley Farm project by placing 25 mm (1 inch) of rigid foam insulation in both sides of the web of the I-joist so that it forms a neat unit with flush sides (see Image #1 at the top of the page).

This approach has three main advantages:

  • It reduces the thermal bridging through the I-joists.
  • It ensures that our mineral wool or fiberglass infill in the rafter bays is consistent and flush as opposed to squashed in the I-joist.
  • It reduces air movement around the insulation and thermal bypass within the web of the I-joist.


The roof needs to be windtight and vapor-open on the outside; on the inside, it needs to be vapor-closed and airtight.

In Stirley we used a wood-fiber vapor-open sarking board for the roof sheathing. In contrast, at Golcar we’ve used Solitex windtightness membrane (an air barrier membrane) because of the need for traditional detailing at the eaves and rakes (see Image #3, below).

To further reduce the thermal bridging of the I-joists we’ve used a continuous 50 mm (2-inch) rigid polyurethane insulation board on the interior of the roof assmbly, sandwiching the Intello membrane (a smart vapor retarder).

Build-up of the roof

The layers of the roof assembly are shown in Image #4, below. From the outside in, here are the layers:

  • Natural slates
  • Battens and counter-battens
  • Solitex membrane, taped
  • Full fill of 325 mm (12.8 inches) fiberglass batts in between the 302 mm (11.9 inch) I-joists, so that the batts compress, reducing the potential for sagging and minimizing thermal bypass problems
  • 25 mm (1 inch) of rigid polyurethane insulation attached to the sides of the I-joist webs
  • Intello smart vapor retarder as an air barrier (taped inside)
  • 50 mm (2 inches) of rigid polyurethane insulation with laminated plasterboard on it

Note: We’re not actually using the Intello to its fullest potential here; in theory it could have been replaced by plastic sheeting. We are not going to be able to utilize its ‘intelligence’ and vapor-release properties because we’ve blocked it with the rigid polyurethane insulation. But we’ve used it here for buildability and robustness reasons and cost.

This is not going to create problems with moisture / condensation as we have a vapor-open membrane on the outside, so moisture won’t be trapped in the roof. As a rule of thumb, if the vapor control layer is within the first third of the insulation there is no danger of interstitial condensation.

Aesthetics of the roof/wall junction

The junction between the roof and the wall outside needed to fit in with the West Yorkshire vernacular. At Denby Dale we had a big overhang for shading which disguised the build-up due to the thickness of the insulation within the roof, which gives the building a different look than the normal West Yorkshire vernacular.

At Golcar we haven’t needed an overhang but we have had to create an extra “false” roof with smaller rafters (see Image #5, below) to cover the thickness of the walls; otherwise it would have created an aesthetic problem.

Two roof levels

We have had to ensure continuity of insulation and airtightness between the two roof heights. To achieve this we’ve had to build timber “framing” so that we have continuity of insulation, as well as airtightness and windtightness (see Image #6, below).


We’ve done a bit of cost/benefit analysis on the roof using PHPP, and we concluded that we did not need such a depth of insulation as we used in the Stirley Farm warm roof. This time, we’re going with 302 mm (11.9 inch) deep rafters as we didn’t need 406 mm (16 inch) I-joists. We didn’t need as good a U-factor as we needed at Stirley.

That is the beauty of using PHPP to plan your project; you can take your time and see what you can get away with. So we’ve got a lesser U-factor in the roof at the Golcar Passivhaus, but it still delivers the Passivhaus requirement of 15 kWh/m2*year.


We are using 3 Fakro rooflights (skylights) which are Passivhaus-certified and quadruple-glazed, with an excellent whole-window U-factor of 0.58 W/m2K (equivalent to U-0.10 in the U.S.). They also came with airtightness and windtightness skirts and insulated flashings, so they’ve really thought about how it would work within a Passivhaus / low energy building (see Image #7, below).

The problem with rooflights in a Passivhaus is that – due to the need for water and wind protection for the roof – you can’t wrap them within the insulation in the same way as windows and doors in a wall. You have to put the frame right into the cold of the outside of the roof. The thermal edge losses (Psi values) for rooflights are around 10 times worse than the wall / window installation detailing at the house.

Rooflights are always going to be a weak point thermally. But since the Golcar house is relatively large, we can get away (thanks to PHPP modelling) with using three rooflights.

Bill Butcher is the director of the Green Building Store in the U.K. Bill is a Certified Passive House Consultant who was project leader on the Denby Dale Passivhaus and is currently leading the Golcar Passivhaus project in West Yorkshire, England.


  1. dvaut | | #1

    Scabbed Rafter tails
    Are there any similar details in the GBA detail library showing the layout of the scabbed rafter tails? It seemed like most of the ones I see are cantilevered off the top of the roof or penetrate the air barrier. These look like you could pre-fab them attached to your air barrier (sheathing) on the ground and then fasten it all at once.

  2. dvaut | | #2

    Scabbed Rafter tails
    Ahh yes, upon further inspection I see that the rafter tail's are not an overhang. This would not apply to say an 18" overhang. My bad.

  3. GBA Editor
    Martin Holladay | | #3

    Response to Dillon Vautrin
    If you are interested in more details showing "applied overhangs" or rafter extensions that are scabbed onto the exterior of the wall sheathing after the exterior sheathing has been air-sealed, you should check out the illustrations that accompany these two GBA articles:

    Airtight Wall and Roof Sheathing

    Getting Insulation Out of Your Walls and Ceilings


  4. jackofalltrades777 | | #4

    Could SIPs have been a better alternative to a cathedral ceiling? It seems with all that was involved here to insulate and seal a cathedral ceiling, couldn't have SIPs provided a simpler solution?

    Have you ever used roof SIPs in a Passive House design or are they not used in Europe?

  5. rjparker | | #5

    Membrane Directly Over Trusses and Fiberglass
    It appears that there is no roof sheathing other than the Solitex membrane, as mentioned in the article and shown in the pictures. While the Solitex is held down by the vertical battens, I wonder how common this type of construction is in the UK or the US?

    "In Stirley we used a wood-fiber vapor-open sarking board for the roof sheathing. In contrast, at Golcar we’ve used Solitex windtightness membrane (an air barrier membrane)..."

    "Natural slates - Battens and counter-battens - Solitex membrane, taped- Full fill of 325 mm (12.8 inches) fiberglass batts"

  6. jinmtvt | | #6

    I've seen many similar setups
    I've seen many similar setups in UK and some in France ( web of course..never been there :P )
    Alot of newer UK homes use some different "foil" type insulation as additional roof insulation,
    which end up right under the water membrane.

    I do not belive this is a popular way of building here in NA.
    Never seen anything as such in Canada neway.

    I would've preferred SIPS + additional vapor/water membranes to this labor intensive method.
    Maybe SIPS on cheap beams that could be cellulose filled on a smart membrane celllule ?

    But anyhow, always instructive to see/read about details of "other" systems !

  7. GBA Editor
    Martin Holladay | | #7

    Roofs without sheathing above the insulation
    It's only a matter of time before Ken Levenson posts a comment here.

    Ken Levenson is one of the partners who owns Four Seven Five in NYC, and his company promotes this type of roof assembly. He has photos of a few projects like this in the U.S.

  8. Peter Hastings | | #8

    Reduced Thermal Bridging
    I'm curious as to how putting the solid insulation into the recesses of the I-joists reduces thermal bridging. The cross-section of OSB in the web is unchanged as is the temperature difference between inside and outside. Surely it is simply insulating the void, as you would do anyway, albeit effectively ?

  9. GBA Editor
    Martin Holladay | | #9

    Response to Peter Hastings
    I agree with your analysis. These pieces of rigid foam on the sides of the rafters don't reduce thermal bridging; they just reduce the chance that the fiberglass batts will be compressed or will leave voids.

  10. Ken Levenson | | #10

    Happy to comment
    Great to see Bill Butcher's post. Passive House as been particularly good at increasing the trans-Atlantic building science conversation and this is a welcome development.

    Yes, we think this sort of assembly is a particularly robust approach to cathedral roof construction. Our typical assembly from inside to outside is: gyp bd finish, 1x wd furring, INTELLO airtight and vapor intelligent control membrane, I-Joists with fibrous/vapor open insulation (dense-pack or batt, cellulose, rockwool or fiberglass), SOLITEX waterproof, vapor open and airtight solid membrane, furring, then finished roof.

    This approach allows for a roof that has very high levels of insulation with robust moisture protection and airtightness. We have CAD drawings of these construction details that can be downloaded from our website:

    While this approach is still far from "common" we do see more applications of it where folks are looking for high-performance.

  11. GBA Editor
    Martin Holladay | | #11

    Response to Ken Levenson
    Thanks for your comments.

    Your photo makes me dizzy. I had to rotate it.


  12. Peter Hastings | | #12

    Racking Stresses
    Are these designs using gypsum board or rigid insulation to provide resistance to racking stresses ? I would expect to see OSB (or similar) in there somewhere. I guess if the ridge beam doesn't move longitudinally relative to the wall heads then it's not a problem but if it tries to then the gables are taking loads which they are poorly equipped to resist. What am I missing ?

  13. user-1152547 | | #13

    Comments from a UK perspective
    Thank you for taking an interest. I don't know if I can answer all the comments adequately but I'll try.

    Dillon. There are several issues we have to address with the eaves detail on a Passivhaus, the performance criteria with continuity of insulation, minimising thermal bridging, airtightness, and minimising themal bypass, and to maintain the ascetic look of Yorkshire, coursed natural stone with no eaves overhang.

    Peter. We did consider SIPS but could not find an appropriate product which gave us the U value we needed economically. Timber I joists are relatively economic, readily available and familiar to us.

    rjp. Not having sarking board in Southern UK is normal. For Passivhaus levels of performance we use modern wind tight but very vapour open membranes, taped and well detailed at the eaves and verge.

    Peter. The rigid polyurethane in the web of the I joists has a lamda value of 0.022 as opposed to the glass fibre between the joists being 0.040 improving the general U value of the roof. Also, cheaper glass fibre quilt is supplied in 570mm widths assuming solid timber rafter/joists. This means we alleviate against thermal bypass, convection air movement through and around insulation - a big problem in the UK, although it is not widely recognised yet.

    Thank you Ken for your comments.

    Peter. The steel ridge bolts into concrete pads built into the 150mm inner leaf concrete blockwork which in turn is tied by low conductivity ties to the outer 100mm stone wall, all acting as a single structure. Our structural engineer has also designed steel 'wind posts' within the inner leaf blockwork as we have relatively large wall panels. All steel work is on the 'warm side' with no penetrations through the thermal envelope.

  14. Peter Hastings | | #14

    Thanks Bill
    Thank you for the answers. Would I be right in thinking that the wind posts tie into the blockwork with steel strapping between some courses of the blockwork ? So vertical and in-plane horizontal forces are dealt wth by the blockwork and out-of-plane horizontal forces are transferred to ground by the posts.
    I currently live in a steading conversion with a very similar mixed steel and concrete structure in which pairs of the 'wind posts' are tied with steel cross-beams to form portal frames.

  15. Peter Hastings | | #15

    Steading Conversion
    A steading conversion is like a barn conversion but the raw material is not a barn but one (or more) of the steading buildings on a farm. In my case, the original buildings were used at various times as a cart shed, stables, greenhouse and mushroom shed.

  16. GBA Editor
    Martin Holladay | | #16

    "Steading" is an unusual word, at least to American ears -- but I looked it up, and learned that the word is chiefly Scottish, and refers to the service buildings on a farm.

    "Steading" comes from the Middle English "stede," whence "homestead."

  17. user-1152547 | | #17

    Barn Conversion
    If anyone is interested here is the link to our blogs on the Stirley Barn conversion, recently awarded EnerPhit certification.

  18. user-1152547 | | #18

    Peter; Basically yes to your questions about the windposts, but I've no idea what a steading conversion is!

  19. mfredericks | | #19

    Sky light window flashing?
    Great to see projects like this and learn about other methods of meeting PH requirements.
    I'm wondering what is happening around the skylight (rooflight) with the flashing membrane wrapping over and around each of the battens/counter battens? This is shown in image #3 and #7. It seems that the flashing could have smoothly transitioned to the Solitex if the battens were simply held back by a couple inches around the window.

  20. user-1152547 | | #20

    Rooflight flashings
    Mark. I agree with you! This was the first time we'd used these particular rooflights which originate from Poland where the roof make up is different - they appear to nail their roofing tiles directly to sarking board. The quadruple glazed rooflights came with pre made air tightness and wind tightness skirts as well as insulated flashings. In my opinion my colleagues on site interpreted the fixing instructions in an over complicated way. We will certainly revisit this detail the next time around!

    Incidently we had our first air pressure test on Friday. We are very pleased to find we've managed 0.26ac/hr @ 50 pascals. I do not know the equivalent American measurement.

  21. GBA Editor
    Martin Holladay | | #21

    Response to Bill Butcher
    The equivalent American measurement is 0.26 ac/hr @ 50 pascals. Congratulations!

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |