How to Eliminate Rafter Ties? 4

[PYDC402040]

The conceptual question is - how is the horizontal component accounted for when ridge beams are used to carry the load? Yes, the rafters are designed for bending, and a large component of the gravity load is transmitted to the foundation, but there is still thrust present pushing out the walls. It looks as though rafter ties can be eliminated for engineered buildings with ridge beams, but the code also prohibits walls from resisting lateral loads.


I have a low slop building - 1.5in per foot over 25'-5" span. We'll be using microllam rafters. The architect and client do not want rafter ties, they want to see the slant of the ceiling from inside.

In my case, it seems small at each rafter - just 155 lbs of horizontal force, but the building is nearly 28 ft wide. So this multiplies out to 2175 lbs acting on each wall. Certainly needs to be accounted for.

Do you have to show moment capacity with the ridge beam posts and ends? Even if upsized to double 2x6 and reinforced the bottom connection to resist the moment of 155 lb horizontal at each rafter (over a few sections at a time), isn't this still the "wall resisting the lateral force" which isn't allowed?

I can tell the client they have to have rafter ties, but that also does not seem right as it is such a low pitch roof. Plus, I would like the architect to hire me again (LOL) and don't want to mess with her design too much! The windows will also be full height, and will be custom cut to the slope, so adding a ceiling will change the entire aesthetic, even from the outside. If you're looking at the schematics, the wall on the left has a handful of large openings and is more like a portal than a true wall (this is a sunroom that connects to the interior of a house) if that makes a difference, the largest opening the ridge beam spans is about 10'. So if I'm looking at horizontal capacity of the posts on either side of that opening it would be 387 lb at a 12 ft height.

Thank you for any guidance!

 

[jayrod12]

If there is a ridge beam supporting the rafters, where does the thrust come from? There's no "pushing" of the rafters against each other to create the thrust. Draw your full free body diagram of the situation and if you can't determine where the thrust load would be introduced, then you have your answer.

 

[phamENG]

Any thrust at the walls with a ridge beam is second order. It's a result of the ridge beam deflecting. If you had an infinitely stiff ridge beam, there'd be no outward movement of the rafters. So the trick is to control the ridge beam deflection.

 

[DaveAtkins]

This is a monoslope roof, so it will not produce any horizontal thrust at each end due to gravity loads. Wind loads will produce a horizontal load, but that will be resisted by the roof diaphragm.

DaveAtkins

 

[phamENG]

So it is. Didn't open the attachment.

 

[DaveAtkins]

If you want to prove this to yourself, run it on an analysis program like RISA-3D. One end of the rafter is pinned, while the other end is a roller.

DaveAtkins

 

[phamENG]

One issue with your FBD is that you are assuming both ends are pins. That is only true if the walls are capable of resisting lateral thrust. They are not, so one end should be on a pin and the other on a roller.

Also, you should start your FBD by considering the global conditions. Sum(F[sub]y[/sub])=0 and Sum(F[sub]x[/sub])=0 must be true from the beginning. So if you have no applied forces in the X, where are your X reactions coming from?

 

[Eng16080]

The simple answer here is that most engineers would design a ridge beam and give no additional consideration to any outward thrust from the rafters at the walls.

The better answer is along the lines of what phamENG wrote above, which is that the rafters will push outward as a result of the ridge beam deflecting. This will be most severe at midspan of the ridge beam, where the max. deflection occurs. Determining how far the rafters will spread outward due to the ridge beam deflecting is a simple geometry problem. If my math is correct, the amount of lateral movement, dx, of the the rafter at the wall due to a vertical deflection, dy, at the peak of the rafter, is:

dx = sqrt(x*x + 2*dy*s*x + dy*dy) - x

with x = horizontal rafter span
and s = roof pitch in rise over run

Plugging in the number for your case and assuming a rather severe max deflection of 1" of the ridge beam, it's found that the rafter (at each side of the ridge) will move outward by 0.127". I would consider this insignificant. If it is not insignificant for your project, then simply design a stiffer ridge beam to reduce the deflection.

If you don't like the math above, you can observe the effect quite easily. Simply set a book vertically on your desk. Now, take a pen/pencil/ruler/whatever and set it against the book sloping it roughly the same as your roof pitch. Now move it down the book a little. You'll observe that it barely move outward for such a shallow roof pitch. You could also draw this in cad.

 

[DaveAtkins]

Following up on what phamENG and Eng16080 wrote, once the two walls of the house move outward from each other (~1/16" at each wall if Eng16080's calcs are correct), you no longer have two pins. You have the equivalent of a pin and a roller.

DaveAtkins

 

[PYDC402040]

Thank you guys very much. I'm glad to see that the horizontal component doesn't cause anyone else alarm without rafter ties. I'll move forward without them and reinforce the post to the ridge beam as close to mid-span as possible to help with any lateral pull the rafters might cause. Thank you for the insight that I need a pin at one end and a roller at the other, I did not think of it this way. Thanks again!