Perhaps you could run some rectilinear framing across the skylights, strategically placed for aesthetics, to connect the two diaphragm halves without introducing bracing.

Not sure what you mean by 'rectilinear framing' KootK. Care to elaborate?

Just some struts that cross the skylights perpendicular to the skylight run.

Thereby getting the diaphragm shear capacity from the weak axis bending in the framing?
As it happens, there is a truss member perpendicular to the windows, but as the trusses are spaced at 8' o.c. I'm not sure how much I should be relying on this.

I've actually attached an illustration of the condition here (which I probably should've done to start).

• https://files.engineering.com/getfile.aspx?folder=7ab6254a-f36a-42e1-96b8-f

Quote (atrizzy)

Thereby getting the diaphragm shear capacity from the weak axis bending in the framing?

No. No chance you'll have enough stiffness for that. The idea is to still have two diaphragms, with independent chords, but to interconnect them across the skylight with struts to prevent differential movement. The struts would be assumed to carry only axial load, not participate in diaphragm shear resistance.

Ah yes... as I figured then, 2 diaphragms it is...

Thanks for the input KootK.

Atrizzy:
Why not make that window (clear story window) wall a continuous truss over the primary roof trusses. It extends vert. from the top of the left (top chord) roof plane to the underside of the canti. tip (top chord) from the right roof plane. It has vert. chords at each main roof truss location, and it has “X’ed.” diags. in the 8’ panel spaces. This truss primarily transmits diaphragm loads btwn. the two roof planes. The windows are then four triangular units fitted into each 8’ “X’ed.” diags. panel space, and trimmed out. The diaphragm loads are transmitted at each plane change, left-vert. and vert.-rt. canti. tip with nailing and various hardware at roof plane blocking and the horiz. chords of the special truss. The special clear story truss might be a Vierendeel truss if you can use deep enough horiz. t&b chords and wide enough vert. chords/mullions at 8’, to properly develop the joints. The Arch. might have to learn to live with approx. the same glass area, but a slightly difference window appearance if the wants the roof diaphragm and lateral load system to work properly.

Now that I see the reality of your situation, I feel a little differently about it. I was mostly only thinking of shear perpendicular to the ridge for which I stand by my original comments. However, there's parallel to ridge to consider too. And that will create a modest, cross opening shear demand too. To that end, some options are:

1) Do some bracing as proposed.

2) Consider the two roofs as a separate, three sided buildings.

3) DO consider the framing as being able to transfer a bit of shear, vierendeel style.

#2 and #3 will be greatly helped by this being a long building along the ridge line. Only you know the appetites of your client and how hard this diaphragm will be working but, in many cases, I'd still be tempted to find a way to make a go of it sans bracing because there is a high probability that's what your client would prefer. And, of course, it most definitely is a popularity contest.

KootK, your #2 is where I'm heading with this.

It's just very hard for me to assume any real shear connectivity vierendeel style, aside from maybe a nominal amount to keep the roof from shearing apart when the "two" diaphragms begin moving in opposite directions. Gut feel tells me this is OK providing that each diaphragm carries its own capacity to cantilever from the exterior walls.

I share your concern about not enforcing compatibility between the two diaphragm segments and agree with the recommendations above to provide some type of shear transfer through the clerestory window. A vierendeel truss is probably the least offensive to your architect. You may wish to use this project for some inspiration and to help convince the architect: Link. You will also want to provide chord continuity across the step, which presumably won't be an issue if you have wood sheathed end walls.

Let's be a little less coy and get some information on the table.

1) What are the truss members? Steel? Heavy timber? Muliti-ply pre-eng truss?

2) What's the width and length of the building? 30' x 60' ish?

3) What's the length of the skylight run?

I've got some ideas but it's not worth the sketch time unless I can establish some degree of relevance to your problem. Of course, if you're already happy with where your at with this, by all means, feel free to leave it be.

OK, the facts, though I'm still in the early stages:

1) building is about 30'x100', trusses at 8' o.c.
2) trusses are currently multi-ply 2x6
3) skylight run is the entire length of the building, effectively

Due to the large aspect ratio, I imagine the concern I have about differential diaphragm movement may be a bit overblown...
Maybe it's just a matter of adding up the expected deflections of each individual diaphragm and hoping the total is small?

Quote (atrizzy)

Due to the large aspect ratio, I imagine the concern I have about differential diaphragm movement may be a bit overblown...

Agreed, you could potentially be transferring your little baby shears over a dozen 3-ply 2x6. A couple of neat options:

1) Do the vierendeel thing but with a prefab wood truss. Don't even attempt to get them to design it for real as that would likely be hopeless (I used to be that guy so I can say these things). Just say 2x4 everything and 3x4 plates min at all joints. That would be gobs of capacity.

2) Weird truss frames as shown below.

This feels like the kind or project where a steel vierendeel truss is gonna raise some eyebrows.

Anything steel might make me look... less than competent.

But KootK, your prefab truss idea is great. I think I'll look into that, many thanks.