Parapet wall to truss connection 1

[struct_eeyore]

I've got a condition where I have to terminate flat roof wood trusses into a face of a parapeting (can I say that?) wall.
The bonus here is that the roof slopes, so in order to keep the sheathing flush with the top chords, I'm showing the trusses installed on an incline.
I've been mulling over the connection details and have come up with something like this (below)
Wondering if anyone will see any issue with what I'm proposing here - or any other comments/suggestions.
Thanks in advance.

 

[XR250]

You need to show a nailing schedule for the dbl. 2x12 to transfer the shear or your bolt capacity will be diminished.

 

[RontheRedneck]

My first thought in seeing that detail is to ask what's the other end of the truss look like?

CMU walls are not perfect. They'll lean in or out, or wander in and out. There needs to be some adjustability in the design. Like maybe a center bearing wall where trusses lap.

 

[TRAK.Structural]

Can you use insulation to achieve the roof slope instead of sloping structure?

I don't love mixing CMU with concrete in this way, can you make the tie beam out of CMU?

I'm also wondering about how connecting the top chord and the bottom chord to the wall is going to affect things. Is that going to make this act like a psuedo moment connection?

 

[jerseyshore]

Is that supposed to be a really deep concrete beam?

Our typical detail just has a bond beam at the roof level and a PT ledger with typical face-mount truss hangers (THA's or similar). No mixing and matching concrete and masonry. Vert wall reinf would just be designed for the parapet's cantilever.

 

[struct_eeyore]

I'm curious why you all are against mixing concrete and CMU. This is very commonly done in my area, and I don't know of a single issue with this type of construction that has cropped up over the last ~15 years.

The other end of the truss will have the exact same detail. I disagree on the nailing schedule, as we already have double bolts at 24" o.c. The ledgers can span themselves between the bolts - the double ledger is there only to provide additional bearing so the odd truss doesn't slip.

I agree sloping is best done with insulation - hoping I can convince the arch.

I don't believe there will be any substantial moment developed, as the isn't much horizontal restraint provided by the twist strap.

 

[XR250]

The other end of the truss will have the exact same detail. I disagree on the nailing schedule, as we already have double bolts at 24" o.c. The ledgers can span themselves between the bolts - the double ledger is there only to provide additional bearing so the odd truss doesn't slip.

Because you bolts are effectively cantilevered 1 1/2" off the first ledger as they are not composite. Unless you have taken this into account in you bolt capacity calculation (doubtful).

 

[phamENG]

You're in Florida, right? Concrete tie beams on CMU buildings are quite common down there.

If this is a simple monoslope roof, I disagree that insulation is better. Slope the top of the trusses all day long. It will be way cheaper. Not sure why, but tapered insulation seems to be worth many times its weight in gold. Now if you have some complex, compound slope going on...taper away.

And yeah, if you're using both of those plies in the ledger for bearing, they need to be first fastened together (specify nail size and pattern), and then bolted to the wall. Otherwise you effectively have only one ply working as there's a slip plane between them, and friction isn't going to do much for you once those plies shrink a little.

 

[TRAK.Structural]

I don't know if any adverse affect of having concrete and CMU mixed in this way, but I think it complicates, lengthens the duration, and likely increases the cost of the construction unnecessarily. You can use multiple bond beams or even a deep bond beam to do the same thing as the concrete. Concrete is great, I think your detail works in this regard, but it would never fly in my market.

 

[XR250]

My first thought in seeing that detail is to ask what's the other end of the truss look like?

CMU walls are not perfect. They'll lean in or out, or wander in and out. There needs to be some adjustability in the design. Like maybe a center bearing wall where trusses lap.

I think Ron makes an excellent point as far as adjustability . No way the walls are going to be straight. Have a contingency plan for when they are not.

 

[RontheRedneck]

The other end of the truss will have the exact same detail.

Once again, I want to say that's a really bad idea. That means everything would have to be perfect. It looks good on paper. But doesn't work well in the real world.

 

[XR250]

@Ron,

I once saw 3-piece trusses used in a similar situation. Two halves and a central splice member to give a moment connection between them.
Have you ever seen that?

 

[Tomfh]

my thoughts:

1. The connection between truss and wall appears rigid. Is everything designed accordingly?

2. If it’s a wood truss, how do they make that notched piece?

3. Will your connections foul the truss nail plates?

 

[struct_eeyore]

Regarding the nailing of the plies... I initially agreed with the commentary here, but then started scratching my head. If we were to add in another set of bolts to secure the plies together, it would in theory accomplish the same thing as nailing (while bolting the first ply directly to the wall). So where does the discrepancy creep in when we merge the individual bolt pairs into a single unit? It seems that in this case were actually creating a corbel out of the wood plies cantilevering out of the face of the wall, rather than cantilevering the bolts themselves - what am I missing here?

Regarding the construction of walls... This is commercial construction - I would be surprised to see wall dimensions off by more than 1/2"

Regarding fixity... Admittedly I haven't given this much thought, but I have to think that there is inherent flexibility/give between the anchor and the clearances that would maintain a predominantly pinned connection. I'm very concerned about the flashing separating at this location, so there is no way to avoid a top chord connection to the concrete. The other option of course is to show a purely top chord bearing truss.

 

[Tomfh]

In my view, assuming inherent flexibility sufficient to maintain a pin connection is not compatible with continuous “non separating” waterproofing integrity. A pin connection inherently requires some degree of rotation or slip, which means a flexible waterproofing solution is essential to maintain waterproofing under such movement. This could include a sliding flashing, a flexible bandage, or another system specifically designed to accommodate rotation or slip while preserving waterproofing. If it will leak if it separates, you will have issues. I mention this as I have recently completed three reports on jobs where the waterproofing failed along a line similar to this one.

 

[RontheRedneck]

@Ron,

I once saw 3-piece trusses used in a similar situation. Two halves and a central splice member to give a moment connection between them.
Have you ever seen that?

Yup. On scissor trusses that were very tall and couldn't be done in one piece. And one time on a ~70' truss that was too long to be built on the equipment we had at the time.

But never on a flat truss like the one in the OP.

 

[RontheRedneck]

my thoughts:

1. The connection between truss and wall appears rigid. Is everything designed accordingly?

I doubt there would be enough horizontal deflection in the truss to make any difference. But I'd have to run the truss to get some numbers. And how much is acceptable is subjective.

2. If it’s a wood truss, how do they make that notched piece?

It's just a 2X block plated in place. We do it all the time.

3. Will your connections foul the truss nail plates?

I don't see anything that would concern me.

 

[XR250]

Regarding the nailing of the plies... I initially agreed with the commentary here, but then started scratching my head. If we were to add in another set of bolts to secure the plies together, it would in theory accomplish the same thing as nailing (while bolting the first ply directly to the wall). So where does the discrepancy creep in when we merge the individual bolt pairs into a single unit? It seems that in this case were actually creating a corbel out of the wood plies cantilevering out of the face of the wall, rather than cantilevering the bolts themselves - what am I missing here?

Good questions. Would doubling the number of required bolts address this? I suppose but it makes my brain hurt to figure that out. I would like to hear other's comments. That being said, I think there is a practical limit to the amount of plies in the "corbel" My personal limit is 2

 

[Eng16080]

I don't really like this detail, but I'm not immediately coming up with anything better. For projects I've been involved in, the wall typically terminates below the trusses with the parapet built on top or with studs cantilevering up, sistered to the sides of the trusses. If it was me, and this wasn't existing construction, I would push for that solution instead.

Concerning the bolted 2-ply ledger and shrinkage, if the plies were to shrink, I think it's accurate to say the ledgers would be supported by the cantilevered bolts initially. Under load, though, I can see the bolts deflecting downward, causing the plies to come into contact again (despite the shrinkage gap). At that point, the load would be redistributed and resisted more by a T/C couple in the bolts and through wood bearing.

Could you use a solid ledger instead, like a 3 1/2"x11 1/4" LVL and provide a moisture barrier between the ledger and concrete? The LVL would shrink much less than the PT pieces.

Where you show a 5.25" distance between the rows of bolts, that technically exceeds the 5" limit between fasteners measured perp. to grain per NDS Section 12.5.1.3 (and Figure 12H). Especially with PT lumber, I can see this potentially shrinking and splitting at the bolts. Whether 1/4" really makes much difference, I don't know, but that's at least what the code says.

 

[phamENG]

So where does the discrepancy creep in

For one thing, I don't love bolts between wood plies, and I only use them in flitch beams with steel plates. Otherwise its lots of smaller fasteners - preferably wood screws. But to answer the question...

When looking at two wood plies fastened together to share load or maintain deflection compatibility, the load is entering one ply, being transferred through the bolt into the other ply. The bolt is only as stiff as the next ply, so the load is shared amongst the plies according to their stiffness. Generally, that means equally (since they'd each have the same stiffness), but maybe not in the case of a flitch beam or a repair detail where you're attaching a 2x6 to both sides of a 6x6. Either way, the beam is deflecting, and the bolt is simply ensuring deflection compatibility.

In this case, your stiffest ply isn't a ply at all, but a concrete wall. The goal isn't to maintain deflection compatibility, it's to deliver the the load to the support. The bolt is not moving with the beam deflection, so it's not going to share the load. The bolt is fixed in the wall, so it's going to act more as a cantilever support for the ledger. You have two plies that are not otherwise attached and have a slip plane between them. If you apply your load across the top, the ply closest to the wall will deflect less (less movement in the bolt), and the one further away will deflect more (more movement in the bolt). Is there friction between them? Yes. But how much? In steel connections, we measure the tension applied by the bolt. Can't do that here. The wood crushes before the bolt will pick up much pretension. The wood also shrinks and swells - mostly shrinks, since it'll likely be delivered at something close to 19% and come down to 10-12% - meaning any initial tension will be reduced if not completely lost. So there's no reliable, quantifiable friction there.