Rafter horizontal reaction on the exteriro wall 1

[m1208]

I am working on a roof project with 4:12 roof slope. The ridge beam span is 30'. The ceiling is exposed and there are no ceiling joists to tie the roof rafters to the ceiling joists. The windows in the exterior wall have cut into the wall top plate partially. The exterior walls are adequate to resist the vertical reactions from the roof rafters. I am concern about the horizontal reaction from the roof rafter pushing the existing exterior wall out. The owner does not want any trusses in the ceiling area. Any suggestion is appreciated.

 

[SteelPE]

So, you have a ridge beam vs a ridge board? While I realize that deflection of the beam will cause some lateral thrust at the top of the wall, I would think most would ignore this. Am I missing something?

 

[m1208]

Ignoring the ridge beam deflection, does the end reaction of the rafter (sitting on the wall) has any horizontal force? Or is it basically vertical reaction supported by the wall? The reason I am asking is because the weight of the roofing material is perpendicular to the rafter and thus has vertical and horizontal reactions.

 

[BAretired]

If the ridge beam is adequate to carry half the roof load, the rafters do not require a horizontal reaction at the wall to carry gravity load. Rafters should be seated flat on the top plate. If they are seated on a sloping bearing surface, the connection would need to resist shear parallel to the bearing.

BA

 

[XR250]

I agree with the others, however, a 30 ft. span ridge is pretty significant. What size ridge are you using?

 

[KootK]

The reason I am asking is because the weight of the roofing material is perpendicular to the rafter and thus has vertical and horizontal reactions.

It's important to recognize that this is not the case. You may have horizontal reactions but there is no horizontal component to the gravity load. Gravity's good like that.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[msquared48]

As BA stated, the rafters must have flat bearing at the wall and at the ridge beam, assuming there is one, or there will be a lateral kick from the vertical load irrespective of any deflection induced kick.

Another way to avoid this is to install a strap tie over the top of the rafters at the ridge to eliminate the kick. Depends on the roof framing layout though as to whether or not this can be done.

This scenario has been discussed in many other strings in this forum in the past.

Mike McCann, PE, SE (WA)

 

[KootK]

Yeah, we might have to make a white paper on rafters someday. Perpetual confusion.

Rafters should be seated flat on the top plate. If they are seated on a sloping bearing surface, the connection would need to resist shear parallel to the bearing.

This is true.

If the ridge beam is adequate to carry half the roof load, the rafters do not require a horizontal reaction at the wall to carry gravity load.

This is true but could use some expansion in my opinion. Unless you`re near a plan corner, there simply is no possibility of a horizontal reaction at the wall because the wall is pinned at the foundation and pinned to the underside of the rafter.

the rafters must have flat bearing at the wall and at the ridge beam, assuming there is one, or there will be a lateral kick from the vertical load irrespective of any deflection induced kick.

I believe this to be untrue.

1) I believe the reactions delivered by the rafters to the supporting structure to be unrelated to whether or not the bearing surface is flat.

2) Since the supporting wall is pinned top and bottom, there can be no lateral kick at the wall. Only unrestrained lateral movement.

3) Since there can be no lateral kick at the wall connection, there can also be no lateral kick at the ridge beam connection.

I believe the model shown below to be the appropriate way of visualizing the typical rafter condition. I forgot to throw in some load but you`ll get the idea.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[XR250]

@ Kootk,

If the rafters have a sloped bearing at the top of the wall and no connection, they will tend to slide the wall inwards under gravity load. Adding toenails or similar shear connection will prevent this if the slope is not huge. Having a flat bearing removes this shear load. So technically no kick, but the wall can be unstable if the angle of friction is less than the slope. This is all theoretical, of course, as there will generally be a connection there in practice.

 

[KootK]

If the rafters have a sloped bearing at the top of the wall and no connection, they will tend to slide the wall inwards under gravity load. Adding toenails or similar shear connection will prevent this if the slope is not huge. Having a flat bearing removes this shear load. So technically no kick, but the wall can be unstable if the angle of friction is less than the slope.

I agree with this, as it is similar to BA`s point, but I don`t believe that it affects any of my statements. At a sloped wall to rafter connection, one of two things is true:

1) the component of the vertical rafter reaction parallel to the rafter is able to be resisted and the system is stable OR;
2) the component of the vertical rafter reaction parallel to the rafter is NOT able to be resisted and the system is unstable.

In case one, there is no net horizontal load delivered to the top of the wall. In case two, forces are not in balance and it`s meaningless to discuss reactions. I think this is the crux of where we tend to go wrong with this stuff. We confuse a tendency towards lateral movement in a hypothetical unstable condition with a net lateral force in the stable condition.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[Jc67roch]

m1208 I think you need to clarify if you have ridge BOARD or BEAM. The two are not the same or interchangeable and heavily influence reactions at the wall.

See the following link:

http://blog.buildllc.com/2013/08/ri...rds-modern-structural-solutions-for-a-frames/

A ridge BEAM will carry half the rafter vertical load. Because a ridge beam is substantial enough to carry the vertical reactions of the rafters (and the rafter reactions are vertical only - due to gravity as pointed out earlier) there is no kick out at the walls.

However 30 ft is a VERY long ridge beam and based on even modest rafter spans would have to be pretty sizeable (definitely better than a single 2x). I suspect you have ridge BOARD, which is not substantial enough to resist any vertical loads/rafter reactions. This results in a requirement for a collar tie OR ceiling joists. These resist the kick out forces and ensure the walls see only vertical loads from the rafters.

If the owner can live with a section of flat ceiling near the ridge, you can install collar ties (roughly 1/3 the way down the rafter vertical height). These can be drywalled over as part of the ceiling.

OR you can install periodic ties from wall to wall - depending on the span across the room and their dead weight you can use engineered lumber or dimensional lumber, adequately sized to resist their deadweight in bending and the tension kick out force. I place these at some interval along the top of the wall, and clad them to make them look like just intermediate beams (can even make them look like hand hewn logs). The spacing is based on the capacity of the wall top plate and roof diaphragm to resist bending from the kick out forces between them. But generally I find 4 to 8 feet on center works.

Lastly, you could add columns. Determine the capacity of the ridge board for bending under the laterally unrestrained rafters, and place columns under the ridge board at that interval. But this could result in quite a few! You could also provide a NEW ridge BEAM below the ridge board/rafter tops and support this on columns.

As for the owners wishes for no ties or trusses - all you can do is offer him options, but there is no way around the physics.

 

[msquared48]

OK. I understand the diagram to show how the ridge beam deflection will translate into a lateral deflection of the wall. No worries.

However, stand an uncut rafter on the outside edge of a wall top plate and the outside edge of a ridge beam and see if it stays there. NO. It will slide. The steeper the slope, the faster it slides.

Bottom line is that there must be a horizontal component that has to be dealt with. That is why we use flat bearing if we can at the ends of the rafters.

Mike McCann, PE, SE (WA)

 

[oldestguy]

You structural guys are forgetting what actually is out there. For roofs of pretty good slopes, the plywood type sheathing on the roof turns each side into a single unit, rafters integral with the sheathing. If you removed the ridge beam, nothing much would happen. If you took out the center part of the wall, there would be slight sagging there, but no failure. Consider that side roof section as a narrow width beam of significant depth, in the plane of the roof. Not exactly a folded plate, but similar.

XR250. Ever see a rafter slide off a wall, due to no friction or nails holding? How about a side wall tipping in?

 

[msquared48]

We have discussed that too in the past.

It is not forgotten at all, although maybe not discussed here. It does come into limiting the deflection of the ridge beam, the greater the slope, the greater the limitation.

Mike McCann, PE, SE (WA)

 

[KootK]

However, stand an uncut rafter on the outside edge of a wall top plate and the outside edge of a ridge beam and see if it stays there. NO. It will slide. The steeper the slope, the faster it slides.

I agree with the outcome of this mental experiment. However, it is missing a crucial feature. It assumes the absence of BA's shear parallel to rafter force. If you reintroduce that force, as would be required for equilibrium, you will find that the net lateral force on the wall is no force at all.

The sketch below shows my statical analysis of a sloped bearing condition. In essence, it's a convoluted way to say that you can't squeeze a horizontal reaction out of a vertical applied force.

Bottom line is that there must be a horizontal component that has to be dealt with. That is why we use flat bearing if we can at the ends of the rafters.

That horizontal component is dealt with, internally. I think that we use flat bearing seat for two primary reasons:

1) Who the heck wants to bother ripping a tapered top plate.

2) Per the comments by BA & XR250, the flat bearing surface eliminates the need to transfer shear parallel to the bearing surface, at least for gravity loads. Clearly, you'd still need to for wind and seismic.

I suspect that real reason is overwhelmingly #1.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[KootK]

Ignoring the ridge beam deflection, does the end reaction of the rafter (sitting on the wall) has any horizontal force? Or is it basically vertical reaction supported by the wall?

I believe that is jut a vertical reaction supported by the wall.

One thing that you'll have going for you here is the relatively shallow pitch of your roof. Basically, whatever deflection you end up with at the center of your ridge beam, you're walls will kick out about 1/3 of that amount at the same location. So, for a total load ridge beam deflection of L/240, you'll only have about 3/8" kick out.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[KootK]

You structural guys are forgetting what actually is out there.

I agree with Mike, we haven't forgotten squat. Here's how I se it:

1) Yeah, the roof diaphragm does something for the system. Per the document below, we've been talking about this since the early 70's at least. We debate the merits of diaphragm action here regularly and, occasionally, rely on it out of desperation.

2) As structural engineers, our wheelhouse is less "what works" and more "what can we prove works". I'm not about to tell a family living under a KootK stamped roof that they shouldn't worry about having a ridge beam because I haven't seen any examples of them failing on the interweb.

3) Per the document show below, which contains about 100 pages of typeset algebra, proving that folded plate works is really, really, really hard. And getting the designed details properly executed in the field is even harder.

4) One thing that you may not be considering is that these great, deep diaphragm beams require supports at the ends to act as such. Those supports are generally the shear walls at the gable ends, one of which is usually chalk full of windows. If the top plates are spliced together carefully, one can take advantage of symmetry to solve this problem. It deserves special attention however.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.

 

[oldestguy]

Mr. Kootk: Great post. However, in summary, the simple way (easiest) is to go for the belt and suspenders to account for all possibloe unknowns

 

[oldestguy]

Mr. Kootk: Great post. However, in summary, the simple way (easiest) is to go for the "belt and suspenders design" to account for all possible unknowns and is the least cost for designing, least risk. Don't forget to tie it down for when the tornado comes along. Cost for owner is incidental. In my past work the cost for the owner was a prime factor to consider.

 

[KootK]

Mr. Kootk: Great post. However, in summary, the simple way (easiest) is to go for the belt and suspenders to account for all possibloe unknowns

Thanks. I wholly agree if what you mean here is to design and supply a ridge beam and then just count on diaphragm action as some nebulous, bonus capacity. This is what is commonly done.

On the other hand, if what you mean is to actually account for diaphragm action in design then I very much disagree. The design would be anything but simple and efficient. I could design and detail the ridge beam system in an hour. Without the aid of a killer spreadsheet, the folded plate business might take me a day or two. And a good deal more drafting and field review time as well.

to account for all possible unknowns and is the least cost for designing, least risk.

I see risk like this:

1) Design for ridge beam and feel good about having diaphragm action too. Low risk.
2) Omit the ridge beam and design for diaphragm action only. Medium to high risk.

In my past work the cost for the owner was a prime factor to consider.

Cost for the owner is a prime factor in all of my work. I would submit that a properly designed ridge beam system would be much cheaper than a properly designed folded plate system for a few reasons:

1) The ridge beam essentially replaces shoring that might be required for a folded plate system.

2) Folded plate may well necessitate a sheathing upgrade, possibly including blocking. The cost of that could easily dwarf the cost of a ridge beam.

3) The diaphragm connections would likely be much more onerous that would would see with a conventional ridge beam design.

4) With a folded plate design, you`d be asking the builder to do something unconventional. There`s always a premium associated with that.

Sure, if you change nothing but omit the ridge beam, money might be saved. But then that`s not a properly designed folded plate system.


I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.