Shear flow built-up wood beam

[jdgengineer]

I'm looking to see if it is at all possible to deepen an existing ridge beam by adding another beam below and then gluing and screwing them together. I understand this is unusual for wood, and I haven't done it before, but just looking to see if it is feasible by calculation. The existing wood is dry and the new we would propose to be either KD or engineered (LVL) to minimize the potential for shrinkage that could compromise the composite behavior.

The existing beam is 4.75" wide x 13.5" tall and I am looking at adding a flat 4.75" wide x 3.5" tall wood beam to the bottom of the ridge.

To determine the screws between the two pieces I am using shear flow VQ/I.

V = 4.63 kips (worst case shear at member end)
Q = A * y(bar) = 4.75"*3.5"*6.75" = 112.2 in^3
I = 1/12 b * h^3 = 1/12 *4.75 * (13.5+3.5)^3 = 1,945 in^4

VQ/I = 4.63*112.2/1945 = 267 # / in

If I were to put in (2) SDS screws @ 3" oc I would get a force per fastener of 267 #/in * 3 " / 2 screws = 401# / screw. (I don't have proper edge distance, just evaluating relative numbers).

Does this look reasonable or is this just a crazy idea? We would also add glue between the joints. For our purposes, assume the existing beam is shored to remove all existing stresses.

 

[1.392DL]

Yes it's possible. IMHO SDS screws would work pretty well if you were adding a steel plate to the bottom of the beam, but for adding another wood beam I worry about sag. I agree gluing is appropriate, but it's hard to quantify. SDS screw are likely not long enough either, whereas SDWS screws might get you the length you need. However, I think you should look into using fully threaded screws like SDWC screws, or longer ones by other mfrs like ASSY. If you install them at a 45, you engage the screw longitudinally rather than in shear, which makes me feel better about beam sag. Draw a free body diagram of the angles screw and youll see there is a clamping force between the existing and new beam. You can research composite CLT construction and reinforced GLB beams for more information on designing wood with fully threaded screws as well.

 

[XR250]

Shouldn't your Q be 4.75x3.5 x 3.5/2 since you want the shear at the new interface.
i.e. Area x distance from centroid of added member to the new interface.

 

[XR250]

..Also, you may not want to count on glue sticking to an LVL. Some have a wax coating on them.
In your situation, I generally add another substantial member below and don't count on composite action.

 

[RPGs]

XR250, AFAIK his value for y(bar) is correct. It is the distance from the neutral axis of the entire built-up section to the centroid of the area above the horizontal plane being analyzed.

jdgengineer, I have made this design effort a few times and each time I was very hard pressed to get the necessary capacity per fastener using NDS connections. I'm sure there are lots of post-installed built-up members out there that work just fine, even if the shear flow/fastener capacities don't calc out--but it has never given me a warm and fuzzy feeling when I looked into it for clients.

I agree with XR250, it's usually best to add a suitable member below to carry everything, or (since shoring is already on the table) remove and replace the ridge beam with a new member.

 

[XR250]

@ RPGreen

Yup, I woke up at 2 am and was picturing the mechanics of the shear and realized I had fucked that up.

 

[jdgengineer]

Thanks everyone.

It's a bit of a weird condition. It's a single-family residence, that has an existing glulam ridge beam. The ridge beam was designed (I have original calculations) to span across a room in a vaulted condition. The original architect added decorative trusses at 1/3 points (with sizable concentric gusset plates, so they have some accidental structural values) that they now want to remove 30 years later. These trusses were not in the original calculations. However, to create concealed connections between 6x8 rafters and the ridge, they let-in the rafters 1" into the ridge beam. The original engineer didn't account for this in the calculations, so we assumed the ridge beam is 2" narrower to account for the let-ins. The ridge calcs for strength but not for deflection considering this reduced width. Assuming the 2" total reduction in width for the whole section is likely conservative as the bottom 6" of the beam are not impacted, and the reduction is only at discrete locations where rafters are located ~4'-0" oc. However, it seemed like a good compromise to account for stress concentrations factors from the let-ins (which shouldn't be too significant since the "notch" is on the compression face).

To remove the trusses, we wanted to improve the deflection capability of the ridge to minimize the potential for noticeable sag after the trusses are removed. Currently, the ridge meets L/336 live load and L/152 total deflection with the above assumptions. We are considering the trusses to essentially shore the beam (at least at 1/3 points) as they appear substantial enough to help support the ridge. Therefore, our idea was to improve the ridge beam section with the 4x below the ridge and then remove the trusses. The beam would then be painted.

We would typically, add a beefier section below to work on it's own, flitch plates, or replace the existing beam, but given that it is an exposed condition and they are trying not to remove the roofing here, this composite section seemed like a reasonable compromise. As there is arguably no "ceiling" I think one could argue the CBC deflection criteria would be L/180 live and L/120 total. The CRC would want L/240 live, with no requirement for total that I am aware of. So technically, I think we already meet the minimum code requirements for deflection, but we wanted to improve that to minimize the potential for noticeable sag. If the composite section behaves as one deep beam, we would have L/671 live and L/302 total (1/2" dead, 1" total).

I think what we have proposed is conservative, and reasonable and one could likely argue doing nothing. But it's a fairly high value property, so we prefer not to take too many chances.

 

[dhengr]

Jdgengineer:
You could put a few split rings or shear plates, in the faying plane, in the first few feet at each end of the new beam. Then use screws for proper clamping of the glue joint, but at a wider spacing, and count their shear cap’y. too. You do have to unload (shore) the existing beam during this marrying process, to make the new deeper beam act properly, as a true built-up member. Pay some attention to the 3.5" deep notch at each bearing end of the new beam. You could use pre-drilled side plates with timber rivets, both sides, and each end to transfer that max. shear flow.

 

[RPGs]

jdgengineer, thanks for taking the time to write that out. I agree with your approach and hope all goes smoothly for you. Sounds like a cool project. High-end residential will always be my favorite.