Triangular Window Opening in Dormer

[RFreund]

A client wants to use triangular windows in a dormer and sent the photos below as an example. The problem is, the ridge isn't direclty supported. Traditional framing rules are that you either need to support the ridge or tie the rafters (if it can't spread, it can't come down). However, I want to at least poke a litter deeper, as we all know, wood structures are fairly redundant. So what is holding this example up:

[ol 1]
[li]As this ridge displaces downward, it will need to push against the roof diaphragm. So the roof diaphragm can resist the thrust of the rafters at the valley portion.[/li]
[li]The wood sheathing. First acting as a cantilever beam (in-plane bending) to help carry the thrust load from the rafters not in the valley back to the roof diaphragm.[/li]
[li]The wood sheathing. Second as a folder plate (out of plane bending of the composite sheathing-rafter member). Essentially the roof wants to act like a folder plate spanning between dormer walls.[/li]
[li]In the framing photo there looks to be a lot of nails in the single tie at the 2nd set of dormer roof rafters. Maybe this is an attempt at some sort of tie at that end.[/li]
[/ol]

So my question are -
[ol 1]
[li]Do any of these factors sell you on putting your name behind this?[/li]
[li]You think it's even possible to model these effects?[/li]
[li]Am I missing something?[/li]
[/ol]

I'm guessing that all the factors above are helping, but ultimately this ridge is going to 'sag' over time and rest on the window. All will likely be fine until you go to replace the window and can't get it out.

 

[phamENG]

There is a folded plate analysis method. Not widely used, but it has been published and at least one member here has used it with success. It's fairly involved. For something small like this, that's about all is go for. With a little more room for real shear walls next to the windows I might get cute with it, but I won't rely on sheathing acting as a shear wall if code says the AR is too high.

From a forensic perspective i think you're right, but if we can't reliably quantify it we can't say it meets code reliably standards.

 

[SwinnyGG]

I'd argue there's a potential path home by skinning that 'bent header' on both sides and treating it like a box beam.

The internal bracing shown in the photos doesn't look to me like it would be sufficiently stiff to guarantee load transfer into the dormer walls. Which themselves are partially cantilevered above the roof diaphragm.. Man the more you think about it the more there is going on.

 

[phamENG]

A plywood box beam might work, but those things are a bit of a pain to design and even more of a bear to build correctly. I only know of one or two contractors in my area that I would trust to even attempt to do it the right way. And only one of them really has a shot at pulling it off.

 

[SwinnyGG]

I would agree it's not an ideal scenario. Just not sure how else you get away from not having a tension tie there. Tough little detail to figure out.

 

[phamENG]

As small as the dormer walls are and as short as the span is, I might be temped to design a fixed base on the studs with some extra blocking along the doubled rafter. Though my preference would be to get the architect to give me some walls from floor to dormer bearing so I can pin them to roof diaphragm and resist thrust through bending in the studs. Would be a tough sell, though.

 

[DoubleStud]

Perhaps if I design the red column to be able to take moment fixed at blue arrow, I would stamp it. You have to make sure the double plate (dashed green) can span and able to take the thrust force from the dormer though. Maybe specify LVL instead and detail connections on both sides.

 

[RFreund]

FWIW -> architect accepted some steel rods as rafter ties.
Might still get back to this at some point.

 

[DoubleStud]

Why not just do bent steel beam and call it good?