Timber Connection Detailing and Design

[TRAK.Structural]

This one is for the timber framing experts out there.

See attachments. I am designing a truss replacement at the gable end of a rectangular gazebo structure. The existing bottom chord of the truss is currently sagging quite a lot due to failure of the connection between the king post and the bottom chord. Additionally a lot of the wood is deteriorating so a full re-frame of the truss has been decided on.

The truss spans about 25 feet clear so the forces in the compression chord and the tension chord are significant (5+ kips, and 4+ kips respectively). I think I'd like to use notches/bearing type connections because these loads seem like a lot for screws and connectors, although I am open to other options that don't involve custom plates that may have a long lead time. To be clear, these are not intended to have mortise/tenon elements, just simple cuts. My questions are below:

1. Can PSL members (the bottom chord) be notched as shown?
2. Seems like the applicable limit states for the notched connection are bearing capacity for the wood, and block shear? Is it that simple or am I missing something?
3. I've seen some diagrams showing the block shear plane drawn like my blue dashed line, is there also a block shear (or other) number to check on the top chord as shown with my red dashed line?
4. I want to have some kind of mechanical connection for additional insurance and any uplift cases, SDS screws seem reasonable here (currently shown on the detail) but are through bolts clamping the two members together a better option?
5. Any guidance on the angle from horizontal that the green line should be?

Feel free to offer other connection ideas and/or let me know if I am totally missing the mark on how this should be configured.

 

[Eng16080]

[ol 1]
[li]I would think it is ok to notch a PSL like you show, although you may want to check the manufacturer's literature. If the notch is exposed to moisture, I can see water potentially getting trapped there.[/li]
[li]Seems reasonable. I would check block shear (blue) per NDS Appendix E, using a calculation similar to that for row tear-out, except with only a single shear line.[/li]
[li]For the red line, I can't think of a specific check necessary. An axial force from the top chord would seemingly want to cause a block shear failure similar to the bottom chord, but the shear failure length would presumably be the full top chord length, so I would expect that not to control. One thing to note, you may want to account for an eccentricity of the top chord in your compression check considering that the location of the result force (at the green line) is not through the member centroid.[/li]
[li]I don't understand how the SDS screws are connecting the two members to resist uplift. Per the detail it looks like the top chord is sitting in a notch but the two members aren't actually connected. I also don't understand the clamping comment.[/li]
[li]Perpendicular to the top chord longitudinal axis makes sense to me (seemingly how you have it shown).[/li]
[/ol]

Other ideas:
[ol A]
[li]Normally for a connection like this I will use steel side plates and thru bolts. If the forces are high enough, I might use a steel saddle that the end of the top chord bears against.[/li]
[li]Just brainstorming: something like a concealed flange joist hanger could almost work if you squared off the end of the top chord.[/li]
[/ol]

 

[DoubleStud]

typically when I do timber truss like this, I just design it as beam and king post. Most of the houses I design want very large timber and just using beam and king post works. This way all of the diagonal members are "cosmetic", so the diagonal member connections are not critical.

 

[jerseyshore]

I'm with DS, 95% of the time trusses at gable ends of gazebos, covered decks, etc. can just be beams and king posts. For regular Joe Shmoe framers that do this type of small construction that's what they want to see as well.

Also, notching a PSL can be extremely difficult. Unless it's being cut perpendicular to grain, I've seen saw blades crack trying to cut a PSL notch.

 

[TRAK.Structural]

Eng16080: The SDS screws would be toe'd in on each side of the top chord down into the bottom chord. In the event that the top chord is in tension (from wind uplift) the screws would act in combined tension/shear to prevent separation at the heel connection.

DoubleStud and Jerseyshore: Yes, simple beam and post would make the analysis easier, but for a 25 foot clear span that catches a point load at the center from the main structural ridge of the structure I think that beam would have to be very deep. Even if a 16-18 inch deep member works for strength I don't see how it could be stiff enough to prevent the "truss action" from taking over and sending a lot of thrust to the heel joint.

If using the simple beam approach with point load in the center, when the beam deflects downward, the vertical post above also deflects, which means the peak/ridge of the truss must deflect down. But the peak cannot deflect downward because the top chord members prevent this through transferring thrust at the heel joint. Load always follows stiffness, which in reality is somewhere in between truss and simple beam, but I just don't see this long of a beam attracting a lot of vertical load due to it being much less stiff than the top chords in compression.

I've attached an elevation of the gable end for more clarity.

 

[Eng16080]

But the peak cannot deflect downward because the top chord members prevent this through transferring thrust at the heel joint.

This assumes that you've provided a sufficiently stiff connection between the top and bottom chord. If you provide only a minimal connection at that location, as opposed to one which resists the full thrust force, then I would expect the connection to deform as the bottom chord (beam) deflects downward while the top chords simultaneously push outward. By designing the bottom chord as a beam to resist the vertical force, it's similar to providing a ridge beam. Assuming the beam is properly sized, you are providing a valid load path. The reality, as you mention, is of course some combination of truss action and beam action.

 

[TRAK.Structural]

Eng16080: Agreed

However I'm curious how one can justify a "flexible" connection with common connectors/fasteners for wood construction. If the connection cannot handle the full thrust force I would think it would fail by fasteners shearing or pulling out of the wood, etc; as opposed to just moving/rotating and redistributing load. At this point you would have common rafters at the gable end without a connection that can resist any kind of load at the bottom, which is still undesirable.

 

[Eng16080]

TRAK, I generally agree with your comments above. There is likely some minimal amount of deformation that the connection can handle, but after that, there's the potential for the failure that you mention. Would this failure actually occur in in reality? It's hard to know for sure as there's usually a lot of extra strength inherent in roofs which our simplified analysis methods fail to account for.

FWIW, I usually take the same approach that you are here in terms of designing the top to bottom chord connection to resist the thrust. The other method suggested by DoubleStud is simpler overall, but doesn't take full advantage of the truss geometry. I think that method is perhaps preferable where the roof span is relatively short and where a beam can be chosen which has very little deflection. I can also understand the desire to give a builder a design which is harder to mess up.

 

[DoubleStud]

TRAK, you are right, that may be a bit long. I have also done steel beam that is wrapped to make it look wood.