Rafter Tie Continuity at Flush Beam

[TRAK.Structural]

I have a traditionally stick framed gable roof system (rafters with ceiling joists) and need to put in a semi-flush beam to add storage capacity in the attic space. I think one can reason that the tension in the ceiling joists can be transferred through the sheathing on top but that would require that sheets be continuous over the beam.

1. I think some light gage straps are a prudent requirement on top of the joists and going over the beam to transfer the tension. Has anyone done that before, or a different way?
2. Part of me also wants to tie the bottom sides of the joists together across the beam but this seems more difficult due to the beam depth being greater than the joists. Anyone have any ideas how to do this, maybe a twist strap? Is tying the bottoms together overkill here?

 

[XR250]

I usually either show a 4 ft. wide sheet centered over the beam or use straps. I have run into many situations where the joists separated from the beam or the beam plies have pulled apart.

 

[Aesur]

I wouldn't rely on the sheathing unless you can show it can adequately in tension transfer the force in both nailed connection and tension in sheathing. Most of the time I would either strap the top or use HD's and threaded rod through the beam attached to the sides of the tension member and avoid even trying to prove the sheathing works.

I should also note, the strap or HD option is easier to construct as I assume you need shoring which makes it difficult to keep in place while sheathing and once the shoring is in removed, the sheathing is required to be there.

 

[ChorasDen]

Second the HDU or DTTZ option. I get worried about straps not being pre-tensioned, and thus allowing the walls to begin moving outward before sufficiently engaging the straps after elongation of the steel, and thus, pulling the beam plies apart. This requires holes through the center of the LVL, but that should be achievable by whomever is doing this retrofit.

 

[TRAK.Structural]

XR, Aesur, and Choras - Looking back at my original post its worded like a retrofit, but this will be a new build. Don't think that really changes much though.

The Simpson hardware with through bolts seems like a lot of holes, and if I try to space them out (say every other joist) than I have to prove that the "connected" joists can handle the extra load that would be forced through them.

If I go with straps do you see a need to strap over the top of the beam and the bottom?

 

[ChorasDen]

If you strap under the beam, how do you finish this area, and apply a gyp ceiling? Furr out and use a soffit or false finish? Everything is a guess for us, because we don't know the design conditions, what's the magnitude of the loads per ceiling joists? I would consider what light gauge strap you want to use, and it has to be light gauge, because you're connecting 2x_ material, and use TL/2AE to determine the elongation of the strap, if that's low enough that you are comfortable with the plies holding together and the walls staying plumb, then I would look at just adding a strap directly overtop the joists.

 

[ChorasDen]

The Simpson hardware with through bolts seems like a lot of holes, and if I try to space them out (say every other joist) than I have to prove that the "connected" joists can handle the extra load that would be forced through them.

Not totally, what's the stiffness of the sheathing compared the the straps you intend to design with? Will the straps even engage in tension prior to the sheathing supporting the load? Would a DTTZ be a stiffer connection? Can you rely on the sheathing carrying some of the load using a stiffness approach, and thus, requiring less simpson hardware?

All depends on how deep down the rabbit hole you want to go...

 

[Eng16080]

Lots of options that come to mind:

1. Threaded rod with hold-downs, as noted above. This is probably the best solution for a larger tension force. The connection can be located at the joist centroid, which is probably preferable.

2. I think Simpson makes a sort of double twist strap for this situation. It would run over the top of the beam and attach to the sides of the opposing joists near the top. If you keep the top of the LVL like 1/4” lower than top of joists, this will eliminate any interference with the strap. (This is also a good way to account for the expected shrinkage differential between the joists and LVL, not that it likely matters in an attic.)

3. Not sure this is reasonable, but perhaps there’s a joist connector capable of resisting the tension force. Even a typical joist hanger fastened with screws into the beam probably has some useful capacity.

4. As mentioned above, if coordinated properly and contingent on the loads, relying on just the floor sheathing to resist the tension could be justifiable. (I would probably avoid this option though.)

5. If the floor surface doesn’t matter, you could add a flat strap on top of the sheathing to connect the joists.

6. A more exotic solution (and one which doesn’t directly answer your question) is to use an entirely different load path. Instead of transferring tension between the joists, instead use half the width of the attic floor diaphragm to resist the lateral force from the rafter thrust on each side. The tension force essentially transfers into the sheathing (diaphragm) over some length based on the nail spacing and capacity. From there, just check that each half diaphragm can resist the load. In general, I prefer most of the other options over this, although I think this might show that there’s a valid load path through the floor sheathing if a direct tension connection had been neglected.


FWIW, I wouldn’t worry about connecting the bottom of the joists. If the tension force at the rafters is applied at the centroid of the joists and you use one of the options above with the tension connection at or near the top of the joists (options 2, 4, and 5), this will cause a small moment in the joists, which you can calculate. You’ll probably conclude it’s insignificant.

If you really want to connect the bottom of the joists, maybe you can use a shallower LVL of more plies to match the joist depth?

 

[phamENG]

Straps over or the double twist Eng16080 are my go-to. I don't like using sheathing - in wood construction, I want engineered conditions to 'look' engineered. Otherwise, the framer will do whatever he normally does and the GC won't notice.

 

[XR250]

I run into a lot of houses with a separation in the floor finishes above a crawlspace. Usually what happens if that there is a flush girder located a multiple of 4 ft. from the foundation wall so is conveniently ends up at the joint in the floor sheathing. The joists end up pulling away from the girder or the girder separates at the plies.

 

[TRAK.Structural]

Thanks for all the input here.

Eng - I think found the "double twist" strap; Simpson SA Strap. Looks like its intended for pretty much this exact load path. Depending on load it looks like an H2A could also work.

I think a strap in combination with specifying the sheathing to be centered over the LVL is the belt and suspenders approach I like.

XR - Why would the floor joists be pulling away from the girders? Out of plane load from the walls in the diaphragm hasn't reached a perpendicular/resisting element before it hits the sheathing joints over the girders?

 

[XR250]

XR - Why would the floor joists be pulling away from the girders? Out of plane load from the walls in the diaphragm hasn't reached a perpendicular/resisting element before it hits the sheathing joints over the girders?

Good question. I have never been able to determine that but it seems to happen. Maybe shrinkage?
Have not seen any cracks in the foundation that would indicate building separation.
It def. has changed the way I do things now. When possible, I use dropped beams and run all the joists in the same direction so there is continuity.

 

[BridgeSmith]

If the tension force at the rafters is applied at the centroid of the joists and you use one of the options above with the tension connection at or near the top of the joists (options 2, 4, and 5), this will cause a small moment in the joists, which you can calculate. You’ll probably conclude it’s insignificant.

It's also a negative moment, so whatever moment you do get, is counteracting the moment in the joists due to the floor loading.

That does bring up a question for me, though. When designing the tension strap over the top, would you need to account for tension due to bending of the joists under dead and/or live load?