Rafter design with intermediate support

[gcs3pe]

A builder here is putting up a 38' deep house is using 2x12 rafters (span = 19') and a non-structural ridge board. The frame will be tied together at the eaves with the appropriate thrust resisting connections. In my eyes, that's all good.

Changes get made in the field and now we've got two bearing walls in the attic, 5' on either side of and parallel to the ridge. The builder plans to rest the rafters on the bearing walls (birdsmouths), changing the primary span to 14' and a second span of 5' to the n/s ridge board. Again, no problem, I'm thinking.

I get an email today - he wants to use 2x10s (or 2x8s if possible) for the rafters. My first thought is 2x10s should be plenty, but then I start thinking about the moment on the cantilevered portions of the rafters at the top. Now I'm sitting here, stumped on how to model this and not sure how to answer.

FYI: snow 30psf, pitch 8/12.

 

[msquared48]

Model it as a continuous beam over a support.

I would also be worried about the framing below the new interior bearing wall(s) that have been added and the capacity of that framing, clear to the foundation. That, too, needs to be verified.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask

 

[4thorns]

Is this close to your situation? See attached.

 

[4thorns]

I agree with Mike. The rafters may not carry the roof load even without being notched with a crow's foot. Also if the walls are an after thought to create a usable attic space maybe to be used for storage then the ceiling ties may not have been designed to carry the extra load. You'll have to add that into the equation as well.

 

[gcs3pe]

4 - your drawing is spot on, and you'll both be happy to know the attic walls have been well thought out and the loads are safe to the foundation.

Model the ridge board as a support, huh? That's simple enough. I thought it would be more complicated than that because they don't meet at 90 degrees, and the whole set up is effectively a truss. Actually, I was concerned the ridge board may act more like a hinge, and my recollection of how to treat a hinge is sketchy.

 

[gcs3pe]

4 - your drawing is spot on, and you'll both be happy to know the attic walls have been well thought out and the loads are safe to the foundation.

Model the ridge board as a support, huh? That's simple enough. I thought it would be more complicated than that because they don't meet at 90 degrees, and the whole set up is effectively a truss. Actually, I was concerned the ridge board may act more like a hinge, and my recollection of how to treat a hinge is cloudy at best.

 

[Ron]

No. The ridge board is not the support. The opposing rafter is the support. You stated the ridge board is non-structural. You have two choices with ridge conditions...either the ridge board is a beam picking up the load of the rafters or the rafters have to be tied with a collar tie to keep the wall from spreading out.

Wait until the shingles are loaded on the roof...you'll get your answer rather quickly!!

I agree with Mike's approach and share his concerns.

 

[gcs3pe]

My bad, communication error - I know the n/s ridge board is not a support. And, as I said, the rafters are tied together at the plates with connections that can handle the thrust.

So, I'll model the opposing rafter as a third support (same diagram as modeling the ridge board as a support, but I understand the need to clarify). 2x10s braced along the bottom are all good.

Of course, now the thrust calcs are off....

Thanks.

 

[4thorns]

If the walls are supported to the foundation isn't the thrust at the exterior walls reduced to the 10' section above them?

 

[4thorns]

Actually now that I see it again Collar ties at the top of the attic walls should effectively eliminate any thrust there as well creating a small triangle bearing directly on the attic walls which are good to go correct?

 

[gcs3pe]

I agree. I suspect the builder will put a ceiling in, but I'll make sure he does. In any case, the large thrust forces from the original model (which warranted bolts and/or straps) have now dropped down to a magnitude that can be handled by nails - so thrust drops off my radar with the new bearing walls.

Thanks.

 

[Lion06]

Here are my thoughts. The reason for the thrust at the lower support is a result of the support at the ridge being a roller that allows vertical movement and only provides translational restraint horiztontally. In a similar fashion, you still get a thrust at the lower support if you have a true ridge beam, and that thrust is a function of the stiffness of the ridge beam. The more it deflects vertically, the more thrust you get at the lower support.

I want to preface this paragraph by saying that I'm referring to 4thorns first sketch, not the second - i.e. the sketch that does not have a collar tie at the top of the intermediate wall. Now, when you introduce this intermediate support, you get a vertical spring support, not a pin. This makes a huge difference. Your model is essentially a sloped beam with three supports. A pin at the base, a vertical spring at the intermediate support, and a vertical roller at the upper support. The stiffness of the verical spring at the intermediate support (assuming the load i transferred out at these joists and the wall doesn't go directly to ground) is not very stiff. I think what you'll find is that it's not very effective, both in reducing the lateral thrust at the base or changing the moment diagram of the rafter).