Deck joists attached to floor joists 1

[buleeek2]

Hello everyone,

Every once in a while I need to design a deck with joists fastened to the side of floor framing. See attached. I am wondering what your approach would be to come up with shear forces in bolts B1, B2, and B3.
See attached and thank you for your input.

 

[Eng16080]

 

[Eng16080]

I would assume all the uplift is being resisted by bolt B3. I would design B3 to resist the force: P x L1 / L2. The other bolts, B1 and B2, would just be there to provide lateral stability for the deck joist. This approach should be conservative.

In reality, I'm not sure how the loads would be distributed to the bolts. I can speculate that perhaps most of the load would want to go to bolt B1 initially since it would technically be the stiffest load path. Since the lever load to B1 would be very high, I imagine the wood grains would crush at the bolt and the resulting deformation would allow most of the load to be redistributed to bolt B2. Then the same would occur at B2 and most of the load would end up at B3. The end result is maybe a linear distribution with B3 taking 2/3 of the uplift, B2 taking 1/3, and B1 taking close to zero. Again, just speculation.

 

[ANE91]

Isn’t this just a continuous beam? Any number of indeterminate analysis methods will solve for the reactions. Model it once as if the bolts provide rigid support and then again idealizing the bolts as spring supports, to account for wood crushing and/or bolt plasticity. The reality is somewhere between the two, depending on how stiff you model the springs.

You can manually check for which of the dowel limit states from the NDS controls (ZI, ZII, etc.). Select a spring stiffness based on which failure mode happens first (i.e., wood crushing, etc.).

 

[XR250]

I would assume all the uplift is being resisted by bolt B3

This.

If you model it..

1) It will not reflect reality
2) It will likely show crazy high loads in B2 and B1

 

[phamENG]

Also, don't use bolts if you can avoid it. Use a cluster of smaller diameter screws at the back end (where Bolt 3 is), and then a staggered pattern down the length. Wood responds better to lots of small fasteners than to a few big ones, where crushing comes into play a lot faster. There are limits, of course, and too many will just tear the wood to pieces, but putting 6 or so 1/4" self drilling structural screws in lieu of a single 3/4" bolt will have much better performance.

 

[TRAK.Structural]

I would assume all the uplift is being resisted by bolt B3. I would design B3 to resist the force: P x L1 / L2. The other bolts, B1 and B2, would just be there to provide lateral stability for the deck joist. This approach should be conservative.

In reality, I'm not sure how the loads would be distributed to the bolts. I can speculate that perhaps most of the load would want to go to bolt B1 initially since it would technically be the stiffest load path. Since the lever load to B1 would be very high, I imagine the wood grains would crush at the bolt and the resulting deformation would allow most of the load to be redistributed to bolt B2. Then the same would occur at B2 and most of the load would end up at B3. The end result is maybe a linear distribution with B3 taking 2/3 of the uplift, B2 taking 1/3, and B1 taking close to zero. Again, just speculation.

Wood crushing and yielding to get the load where you want it sounds like something I would want to avoid.

Why not just put a couple fasteners right over the double top plate and then fasteners at the end (B3) ?

 

[Eng16080]

Wood crushing and yielding to get the load where you want it sounds like something I would want to avoid.

I think that's making it seem more dramatic than what might actually occur. With that said, I would likely do something more along the lines of what pham suggests with many smaller diameter fasteners (nails or screws) assuming that can work with the loads. Maybe just 16D nails or wood screws staggered at 6" o.c. In this case I would have a group of fasteners around the location of bolt B3 which can take the load there.

Why not just put a couple fasteners right over the double top plate and then fasteners at the end (B3) ?

I would do that and have fasteners between. I'd be a little concerned with buckling if there was nothing from B1 to B3.

 

[jerseyshore]

Also, don't use bolts if you can avoid it. Use a cluster of smaller diameter screws at the back end (where Bolt 3 is), and then a staggered pattern down the length. Wood responds better to lots of small fasteners than to a few big ones, where crushing comes into play a lot faster. There are limits, of course, and too many will just tear the wood to pieces, but putting 6 or so 1/4" self drilling structural screws in lieu of a single 3/4" bolt will have much better performance.

I don't want to actually quantify this, but you have a not-so-insignificant percentage of comments on here recommending the use of lag screws over thru-bolts. Crazy how often people want to use bolts over screws for these wood connections. I guess it's the bigger fastener = better type of mindset. You da real MVP.

 

[buleeek2]

Thanks for all your replies.
Attached is the actual situation. The HSS canopy tube is attached to the TJI floor joist.

1. I agree with small-diameter screws, but due to the steel tube, I think I need to go with bolts.
2. When I model B1 B2 B3 as supports, B1 receives crazy (and unrealistic) uplift, B2 and B3 receives a very minimal forces.
3. I would consider spring supports, but could someone elaborate on what k/ft value each support would have?
4. If someone has any thoughts on how to reinforce the floor framing (tubes at +- 32" on center, joists at 16" o.c.) other than blocking, please be my guest.

Thanks,

 

[phamENG]

The HSS canopy tube is attached to the TJI floor joist.

The what's attached to the what now?

 

[TRAK.Structural]

The what's attached to the what now?

ROFL

Buleeek2 - Connecting to the web of an I-joist like this is not a good idea.

 

[XR250]

I'm lost. The Front View does not show tube steel or an I-joist.

 

[Eng16080]

OP: please post pictures directly like this if you can. You'll probably get better responses as some people won't care to download a PDF.

 

[Eng16080]

I don't have all the information that you do, but I would generally prefer an alternative solution to this. See points below:

1. Is there a reason why you have to use an HSS section? I would prefer a channel instead, which reduces the eccentricity at the connection to the floor joist. There's also less risk of water/moisture damage versus the closed HSS shape which has more potential to trap moisture.
2. I probably wouldn't use an I-joist to connect this to, even with the webs packed out. I would prefer to use LVL joists in this location instead. The LVLs should have minimal shrinkage similar to I-joists.
3. In terms of modeling this in software, I'm not surprised by your results. I would probably model it with only two supports, one at the wall and the other at B3. I think the results you're getting are somewhat in line with what I was thinking in my post at the top.
4. As far as using bolts go, that's usually how I connect wood to steel. In suggesting multiple smaller fasteners above, I was assuming this to be a wood to wood connection.

What's the approximate load magnitude here?

 

[buleeek2]

1. I could switch to C-channel, but 4" depth max.

2. The floor is existing and the joists are 9.5" TJI @ 16"o.c. The original concept was to attach the canopy tubes to the existing parapet wall, but I am worried the parapet will not withstand the load.

The load is +- 250 - 300 lbs at the tip of the cantilever.

Does anyone have a better idea than attaching to the TJI webs ?

Thank you,

 

[XR250]

Looks like your cant. is about 12". 4" C channel oughta work. Why are you using such heavy tube steel?
If you went back 3 ft. with the channel, the load would only be 100 lbs +/- at the fastener.
Seems doable - even with shitty I-joists.

 

[buleeek2]

This is what I came up with. The 4" channels don't work for deflection. I modeled the bolts as spring supports (100 lbs capacity / (1/16" deflection) = 19.2 k/ft). See below, any input will be greatly appreciated. I pasted images instead of attachment, as requested.
Thanks

 

[XR250]

So your original drawing was not even close to being to scale? That would be helpful in the future.
I take back what I said about this being OK with I-joists.
Not sure why you are using so many bolts and modeling that bolted connection ain't gonna reflect reality.

 

[Eng16080]

The 4" channels don't work for deflection

How about double channels? Using two channels with the joist sandwiched between is a much better connection IMO. The HSS along one side creates an eccentric connection which will be trying to twist the joists.

The load is +- 250 - 300 lbs at the tip of the cantilever.

You're looking for feedback on this, so can you explain how you arrived at this load? And why is the load applied at the end of the cantilever and not distributed evenly over its length?