Thrust from Rafters on Wall

[asiem]

I have been looking at a one story addition to a house that was built without a structural engineer and is now having problems. The addition is conventional wood light framed construction with 9' 2x6 walls supporting 2x10 rafters at a 4:12 pitch over a total building width of 28'. The rafters are collared with 2x6s about 3' above the eave height (12' above the floor) creating a tray ceiling. My analysis shows the rafters failing and this configuration has, of course, created thrust on the exterior walls which have now begun to bow under dead load alone. Normally under these circumstances I would recommend that a strucutural ridge beam be installed, however, the architect does not want to go this route because of accessibility concerns. Instead, what we are looking at doing is reinforcing the rafters. I can get the rafters to work by sistering each. My quesitons is what to do about the horizontal thrust. A finite-element analysis run pin-roller shows the frame with sisters deflecting 1" in the X-direction vs. 2" without sisters. This seems alot better but keep in mind that the present deflection has occurred under dead load only. I'm worried that if all we do is reinforce the rafters the thrust will be just as bad under full loading conditions as it presently is under dead load only. Any thoughts?

 

[jike]

I am not sure what controls your design.....

Suggestion below is just brainstorming without knowing all the factors.....

Would it be possible to shore up the rafters, remove or loosen up the collar ties, install temporary cables with turnbuckles in there place and reduce the horizontal deflection and then refasten the collar ties?

Would this procedure overstress the rafters?

 

[fbaxley]

Is it acceptable to install a dropped ridge beam? IRC requires that if rafters are not tied to ceiling joists at the top plate, then the ridge formed by the rafters must be supported by a girder designed in accordance with acceptable engineering practice.

 

[asiem]

The rafters as they exist right now are overtsressed in bending due to the tension reaction from the collar tie (since there are no ceiling joists). The collar itself is not overstressed and if I sister the joists they will no longer be overstressed either. I am not so much concerned with how to correct the deflection that has occurred already as I am about finding a rational way to look at the roof frame and wall interaction. My conundrum is that no matter how much I reinforce the frame and how rigid I make it, because of its configuration it will still want to transmit some horizontal force to the wall. The problem that I have is coming up with an appropriate amount force. Taking the horizontal reaction that would be obtained analyzing this roof frame as a "pin-pin" condition seems too conservative to me, but maybe I'm wrong.

To describe the situation in a little more detail. At this point in time all drywall is up. The homeowner has ruled out installing any sort of horizontal tie that would be visible below the current tray ceiling line.

 

[miecz]

Have you modelled the wall into your frame?

 

[jike]

Is the connection between the collar tie and rafter overstressed?

 

[LowLax]

Can you have a slip connection at the top of the wall, to allow for lateral deflection? I know we do this with premanufactured scissor trusses. Simpson makes the connectors.

 

[jike]

How does is the top of the wall braced from wind load (direct pressure and suction)with a slip connection at the roof member?

 

[SacreBleu]

jike,
The slip connection alternates with fixed connection, each truss. Each truss has a fixed connx at one end and a slip connx at other end.

 

[asiem]

Jmiec, after my inital post I tried out the model this way and I think this gives the most accurate representation of what is actually happening. My model is 2-d so I have the roof frame with 2x6 columns (the wall studs) supporting it on each end, with the columns fixed at the base. Under this scenario, I get 1" of lateral deflection at the top of the wall with unreinforced rafters and 0.5" (L/216) deflection once the rafters are sistered.

Jike, you make a good point. The ceiling is in place so I can't see what the connection is right now. However, in my model I'm getting 2k of tension force in the collar. We probably won't be able to do this type of repair because a connection to take this kind of force won't be practical to install on so many rafters.

Low Lax, the thing I don't like about scissor truss clips is that using them takes away restraint from the top of the wall. So when considering wind force on the wall, what holds the wall in place at the top, other than the dead load of the roof?

Thanks for your help guys, I'll think I'll tell the architect to quit being a baby and put in the structural ridge beam.

 

[miecz]

Doesn't the roof deck defeat the purpose?

 

[MikeE55]

Going back to proper support conditions, you should assume a pin on one side and roller on the other, unless at least one of the walls will be constrained from horizontal movement. If you then analyze and determine the walls can withstand the amount of lateral movement you obtain from the pin-roller analysis, you have a good solution. It is incorrect to assume pin-pin because the stress levels on the frame will be unconservative, since the walls cannot support the amount of thrust generated. I have noticed this problem in scissors truss designs. Some truss manufacturer's software assumes pin-pin conditions. I have found this can give stress levels in the truss off by 30% - very unconservative. I agree you should use the ridge beam.

 

[valleyboy]

If you clad each slope in ply sheathing, you could consider each slope as a deep ply-web beam spanning the 28'(8.5m) between shear walls.

Or how about doubling the rafters and installing a structural wall plate to span horizontally between shear walls to deal with the thrust?

VB

 

[bylar]

i see this all the time as a result of the framer or GC (and often Architects.)not understanding the whole principles involeved. I agree with Mike55 regarding the analysis.
I think your only reasonable solution is the ridg beam. If the architect doesn't like it let him be responsible.

 

[miecz]

SacreBleu,

With alternating slip connections, doesn't the roof sheathing keep the trusses in line, thereby preventing the movement the slip joints are meant to provide?