Built-Up Sawn Lumber 2-Span Continuous Beam

[ajk1]

Background:

I am checking a continuous 2-span built-up sawn lumber beam of a cottage that was constructed within the last year (I had nothing to do with its design or construction). The beam that I am checking is the main floor perimeter beam supporting the main floor, and the wood stud wall above which in turn supoorts a loft and the roof.

Preliminary checking indicates that this beam is over-spanned. The spacing of the supports (number of supports were reduced), and the size of the beam were both increased by the contractor, from the original system designed cottage, but no engineer sized the new setup.

Although certain simpifying shortcuts could be made when designing such a beam, I cannot take such shortcuts when checking because I may put the owner to needless expense strengthening the beam, where perhaps no strengthening is required if more accurate checking methodology is used.

Given:

The 2 spans are about 7 feet and 11 feet, and 2 of the plies of the 4 ply beam are butted at about the 2 foot locatiion from the centre support in one span, and about 3 foot from the centre support in the other span.

Questions:

1. Does Woodworks software account for the location of the joints?

2. When checking manually, how should the butted plies be dealt with? For example, is there a "development length" over which the ply picks up its share of load from the adjacent plies of the 4 ply beam? Are the nails generally adequate to make that tansfer of load?

3. Are "clear" spans rather than centre-to-centre spans entered into Woodworks?

4. Is there any worked example of design or checking of a built-up continuous wood beam with some of the plies butted within the spans?

 

[Canuck67]

ajk1,

Woodworks does not account for the location of the joints. However, you can input the location of each splice as a point of interest so that Woodworks will provide you with bending, shear, etc... at the specific locations you would like to check.

At the butted ply locations, I only consider the continuous plies, i.e. no moment transfer of discontinuous plies, shear strength of only the continuous plies, etc...

The latest version of woodworks lets you specify which span you have input.

Also, some residential codes have prescriptive methods for splicing beams that is considered to be code compliant. Usually something like, splices of single plies are allowed within 6" either way of the quarter points of the span

 

[ajk1]

To Canuck67:

The point of interest feature is good, but I already manually calculated the moment and shear at the splice location, and checked whether the remaining 2 of the 4 plies can resist the moment and shear at the splice location (I find that the remaining 2 plies can resist the moment and shear at the splice location).

But I am still puzzled as to whether that is all that needs to be checked. Is there a "development length" like there is for rebar in concrete? How do the 2 spliced plies pick up the load so that at the maximum mid-span moment location all 4 plies are equally participating in resisting the load?

For example, if the splice point is 3 feet from the centre support of a 2-span beam of 11 foot spans, then the distance from the splice point to the point of maximum span moment is approximately 0.6x11-3 = 3.6 feet. The total factored load on the 11 foot span beam is 19,000 pounds, so each of the 4 plies must support about 4750 pound factored load. How can such a large load be transferred from the adjacent plies with only the nails?

 

[OldPaperMaker]

ajk1,
You might want to attach a sketch of the way this thing is framed.

 

[ajk1]

It is a standard built up beam of: 4 vertical plies, each ply a 2"x12" #2 spruce member, 2-span beam loaded on its top, with 2 of the plies spliced at 0.25 of the span from the middle support.

I don't have a scanner here at home. I can try to sketch it up and photograph and then attach the photo.I will send sketch later this morning.

I expect that people who specialize in wood design could answer the question of how to design at, and adjacent to, the splice point, particularly how the load gets transferred from the non-spliced 2x12" to the spliced 2x12's so that at the location of maximum moment all 4 plies carry an equal 25% of the load. Or do they?

Is there a "wood forum" for wood design questions?

 

[jayrod12]

Here Wood Design

 

[KootK]

That would be the NDS forum here. Unfortunately, I think that you're on to something here. I don't see how the butt spiced system works unless the discontinued plies are able to dump their shear reactions locally near the butt joint. That's just statics, right?

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[jayrod12]

However, most of the guys that will respond in that forum are the same ones you will get here. So it may be fruitless to attempt to repost in that forum.

The "proper" way to design that splice would be to determine the shear and moment transfer required across the splice. Then figure out the number of nails required to take the load from the 2 plys that are terminating into the 2 plys that are continuous across the splice. Then figure out the number of nails to put that same load back into the new 2 plys that are going to continue on from the splice.

Re-reading that it is really awkward sounding and I'm too lazy to go back and re-formulate a response so see my rudimentary flow chart as I see it.

4 plys carrying load ===>> design nails to transfer 1/2 load into continuous 2 plys ===>> 2 plys carry load for short length ===>> design nails to tranfer 1/2 load into new additional 2 plys ===>> 4 plys carry the load

That is obviously not an exact description of how to design it. But that is the load path as I see it.

 

[ajk1]

TO kootk and Jayrod:

What each of you describe is precisely my thinking...yes statics must be satisfied. It seems to me that the number of nails to transfer a load of this magnitude would far exceed what is normally provided (maybe that is why they say only splice for light loads, but unfortunately they never suggest a load range guide for what would be a light load. I better go thru the calculations and prove it one way or the other.

 

[racookpe1978]

Use screws. Not a carpenter's bag of #8 drywall screw, but real wood screws or bolts rather than just nails or staples.

 

[KootK]

We've gotta be missing something. I see this system all the time in my area and I don't recall ever seeing a big confluence of fasteners at the ends of the plies. Maybe one just has to accept that the force transfer occurs over a relatively long length. Still, you'd think that there would be rules of some kind.

@Jayrod: nice work with the flow chart. Nailed it!

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[jayrod12]

It's not as many nails as you may think. What possibly can your shear load be in a 2 ply beam, a few kips? A 3" common wire nail into the shittiest wood is good for 150 lbs. and the load transfer doesn't need to be instantaneous at the end of the splice, it probably happens over a few feet meaning if you've got 4 rows of fasteners you can transfer 600 lbs per vertical line. Tighten up the spacing in the splice area to 6" O/C and you can produce shear transfer of over a kip per foot of splice.

Maybe I'm over generalizing but like KootK said it's done all the time with no issues.

I'm also of the mind that by the time you provide enough fasteners to transfer the shear then the moment will be taken care of as at 1/4 span it should be near zero.

 

[ajk1]

Just to refresh your memory, I noted earlier in this string, the beam spans 11 feet max, is 4 plies (not 2), 2 of the 4 plies are spliced at the 3 foot from support location, and the factored load on the 11 foot span beam is 19,000 pounds, so the factored reaction is >> 9500 pounds (due to the continuity, 2 span beam).

Aren't you just transferring the reaction from the partial length ply to the adjacent full length ply, which makes the maximum mid-span moment in the full length ply very nearly equal to what it would be if the partial length ply were not there? That's what it seems to me. I will draw the free body diagram and follow the forces through from ply-to-ply, and span-to-span, and see how it works out. In the end, statics of each ply, looked at as a free-body diagram, must be satisfied.

Yes butted plies are used all the time, but it remains to be seen whether they are equivalent to the sum of all the plies, or only to something less.

 

[OldPaperMaker]

ajk1,
I don't think that I understand how this thing is framed. Do you have (2)-2x12 in the center that are 11' long and (1)-2x12, each side of the center plys that has 2 segments that are 8' & 3' long that are somehow attached to the center plys? If so, then the 2 outer plys won't supply any support to the loads from the floors, roof and walls above.

 

[ajk1]

Please see attached sketch and comment thereon.

 

[OldPaperMaker]

ajki,
I think that the attachment didn't coperate and go into the "attachment" mode

 

[ajk1]

I tried it again. Did it upload this time?

One way to settle the issue, is to run the 4 plies on a grid analysis program, with the load applied along the top of each ply, and have a fictitious short member spanning across the end of the spliced plies and connected to the plies each side (simulating the nails). Then see if the mid-span moment in each ply is the same as if there were only 2 plies or the same as if there were 4 plies.

 

[jayrod12]

For the 7' span having a splice at 2' in from the end I get the shear at the splice as ~1600# if the beam is uniformly loaded (1055 plf). For the 11' span it is ~2700#. Those are total shears and so you can reasonably assume each ply carries 1/4 of that which mean you need to design nails for 400lbs at each location for the 7' span and 700 lbs for the 11' span.

So from each continuous member into each of the two discontinuous members you need like 5 nails. Put 4 rows of nails at 6" and you've got 8 nails in a 1' length like I mentioned earlier that's over a kip per foot of shear transfer.

Am I missing something?

 

[KootK]

While I don't have an answer yet, I have some more observations to share:

1) When I opened AJK's sketch, my instantaneous gut reaction was "that's a two ply, continuous beam".

2) This will sound impossibly lame but my next step was to ask my wife. She's also a structural engineer and something of heavy timber specialist. Her gut reaction was the same as mine. She said that she encounters this a lot in older buildings but never does it for new construction. My wife rarely deals with part 9 stuff though and that's where I've seen most of the newer, real world examples that I mentioned above.

3) Assuming that the maximum shears occur near the butt joints, I think that we can all agree that the beam is a two ply member from a shear resistance standpoint. Of course, it's unlikely that shear would govern.

4) If the four plies all experience identical vertical deflections along the length of the beam, then they would all have the same curvatures and therefore the same moments and shears along the lengths of the beam. This depends on the ability of the nails to do their job convincingly. And, clearly, this compatibility of curvature would be somewhat out of whack locally near the but splice.

5) I feel that the stiffness of any moment connections near the butt splices is just as important as the strength of those moment connections when it comes to maintaining compatible curvatures amongst the four plies. When I was a kid, I tried to construct a portal frame fort from three 2x8's nailed together at the lapped corner joints. It sucked. Based on that limited anecdotal evidence, and the fact that you have to make such a connection twice as you cross each butt joint, I question whether or not moment transfer across the butt joints should be relied upon for enforcing curvature compatibility, and true load sharing, amongst the four plies.

6) Based on the above, I think that the composite beam is clearly better than an analysis of two plies would suggest but also definitely worse than an analysis of four continuous plies would suggest. For new construction, I would only count on two plies. I'd only tinker with four ply voodoo if it was an existing condition and I was a bit desperate. From what I've read, that's exactly where you're at AJK.

7) As a way to evaluate an "in between" solution, perhaps one could abandon the notion of moment continuity across the butt splice while still satisfying shear transfer requirements. You wouldn't have perfect 1/4 load sharing any longer of course. However, the spliced plies could still make a meaningful contribution at the maximum moment locations. Again, I would never go to all this trouble for new construction.





The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[ajk1]

To Kootk and your wife: You are a power couple. I had not thought about this from the point of view of curvature but that is interesting and logical. I am not sure that the nailing can cause all the plies to have the same curvature though. I think we are both at the conclusion that we would design this as a 2-ply beam if it were a new design, and that would be a little conservative. I am not that "desperate" about this since it is not my design and I had nothing to do with the construction, and it appears that the contractor just guessed at things. I think I will most probably recommend additional piers where it does not work as a 2-ply beam. I will further study what you have sent...looks like a lot of very good thoughts there. Thanks for taking the time to think it through and write it all out. Much appreciated.