Exterior roof Larsen truss (CZ2, with vapor diffusion port)?
CZ2. Single story residential with hip roof (just over 4/12 slope).
Looking to make a conditioned unvented attic, insulated at the rafter line (not the ceiling/attic floor).
CZ2 so a vapor diffusion port makes air permeable insulation at the rafter line possible and to code. In everything I’ve seen that is occurring underneath the sheathing, installed from below. That is very difficult in this house due to rafter depth and access issues.
Need full re-roof anyway. Thinking of adding I-joists to roof deck (a la Larsen truss), extending out over current rafter tails. Then insulating with batts or dense blown cellulose, then zip sheathing taped and air-sealed at eaves, with proper permeable vapor diffusion port at ridges (and down a bit for hip condition). Then metal roof over top.
I can’t think why this wouldn’t work, ending up identical to an interior batt/blow and vapor port attic.
Anyone have thoughts?
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
I would place my air barrier on existing roof with something like adhero. Then you can vent or diffusion port, either way, and you can use just osb and felt paper
Yes, that seems to make sense. Reading up on the Building Science Corp stuff no vapor diffusion ports makes it clear that the air barrier should be underneath the fluffy insulation (to reduce warm humid air getting into the insulation in the first place).
That also massively helps with the sequencing, since the roof is immediately weathertight (well, for up to 5 months, although with high UV would be shorter than that!)
You also might want to trim off the existing rafter tails. This would let you run the roof deck peel and stick down the wall for a solid air barrier.
This only makes sense if you are replacing the tin work on the house though.
Thanks, yes, that does make sense as well. Full re-roof is part of the plan (so all flashing connections, etc).
I was looking at details for TJI and hip roofs (TJI jack rafters connected to the hip rafter using something like a Simpson LSSR connector). Not clear to me whether that's needed for this non-structural over-roof, though.
One issue I see is that the details for a vapor port in a hip roof are a little different. See https://buildingscience.com/documents/guides-and-manuals/gm-2101-guide-building-conditioned-unvented-attics-and-unconditioned Figure 9. That involves round holes, but only extending down two feet along the hip ridges. That detail could work with i-joists, but it does leave a few bays low on the hip with blocked vapor pathways. Seems that's just not worried about?
Figure 11 shows using a shortened hip rafter (doubled to account for shortening) so that there are vapor paths from every framing bay (which feed into the space above the shortened hip rafter. That seems great from a vapor channel perspective, but I can't see that working with i-joists (where the top flange would be hanging out over the shortened hip rafters).
There are a few other details for the vapor ports that seem incompatible with the i-joists (e.g., notched rafters at wall/roof connections).
Almost certainly easier to get someone to frame/sheath with 2x lumber anyway (although the connection to the existing rafters is harder with traditional lumber than i-joists where you can screw downwards through the bottom flange).
You could always use 2x lumber at the hips. Or for those few joist lower down just block out the end of the I joist with 2x to make a solid connection and prevent top flange twisting
if you used 2x for the entire roof you could probably get away with placing a few GA2 angles along existing rafters and fasten your new 2x to that. this would give you both shear transfer and uplift prevention. The one issue is it would bear your new rafter either right next to the current rafter or barely on it. You'd have to see what an engineer says but since it is so close to the other rafter all the forces on the existing roof deck would be shearing forces and not bending moment and so it might be fine.