Cross-ridge load path pitched roof

[Martin Gillie]

I'd be interested to hear people's ideas on how cross-ridge wind loads are transferred to shear/gable walls in pitched roofs. Two scenarios

1) 45deg pitch with ridge beam and timber frame walls. I am seeing a substantial lateral load on the ridge beam causing minor axis bending, transferred to the supports at each end. Do people design for this? Is load also transferred through the wall plate>elevation>gable somehow? With a steep pitch, transfer through the sheathing seems dubious to me.

2) As above but with raised collar ties. What stabilized the wall heads?

 

[MotorCity]

The gable wall will act as a shear wall to resist the cross ridge wind load. It is much stiffer than the ridge beam in weak axis bending and will attract the load.

 

[Martin Gillie]

Sure - but my query was how loads get to the shear walls!

 

[Bisbee_Structural]

Agree with MotorCity.

Are you looking at an FE model of the roof? In my experience when trying to make a model of a complicated roof system unexpected load paths come up that are usually ignored in standard design practice (simple hand calcs for wood structures) or are tricky to account for in the model (for example nail slip at a diaphragm will reduce the diaphragm stiffness considerably)

In my experience pitched roof diaphragms are typically designed assuming the roof is flat, and the load path is assumed to be through the sheathing. Be sure to check the deflection using the relevant formulas accounting for issues like nail slip.

I'm not sure what you are asking with the collar ties- as far as I know collar ties are used to resist wind suction at the roof ridge (they prevent the tops of the rafters from lifting/separating at the ridge under the higher wind suction seen at the ridge) and don't have anything to do with transferring wind shear force through the diaphragm.

 

[Martin Gillie]

More just a conceptual description of load transfer. It seems to me that the steeper the pitch, the less the sheathing can be relied upon. At one limit, with a vertical "roof", thin plywood wouldn't be capable of transferring load to a shear wall as it would be acting almost entirely in bending. At the other limit of a flat roof, shear in the sheathing is clearly fine. How steep is still OK?

Probably raised ceiling tie a better term than collar tie - say a little above where a ceiling would be if present.

 

[Bisbee_Structural]

RE: how loads get to the shear walls, in the US I usually see the roof sheathing nailed to the rafter at the gable wall. The top plate of the gable wall is nailed to the bottom of the rafter flush to the outside. The wall sheathing is placed up against the rafter and sheathing panels are nailed to the rafter and the top plate with the prescribed nail spacing. At this point the load is in the gable wall, and you treat it like a normal shear wall. I don't typically worry about the 45 degree roof angle and the sheathing, as the nailing values and panel strength in the AWC codes are based on the smallest values whether taken perpendicular or parallel to the panel direction, and your values will be somewhere in between.

If you don't have it already, I would recommend taking a look at the Wood Frame Construction Manual the AWC publishes in the US. You can download a copy for free from their website.

 

[Martin Gillie]

Makes sense, and for steep pitches too. The rafter, toplate sheathing detail isn't specified here like that IME, although may happen in practice in timber frame. I'll look out the reference, thanks.

 

[Bisbee_Structural]

For a very steep roof, say >45 degrees, I would consider that the wind load perpendicular to the ridge is resolved through the triangulated rafter structure- rafters in tension and compression like a triangular truss, with the proper detailing at the ridge and eaves to transfer the force across the ridge and avoid uplift on the tension side.

Then there would be some kind of horizontal diaphragm (like rafter ties) at the eaves to resolve the horizontal thrust and collect wind load from the walls. This would transfer the wall and horizontal wind load from the roof into your shear walls.

If there's no diaphragm at the eaves, you would need more of a portal frame structure, like one of the laminated wood 3-pin arches you see in 1960s church structures.

If the roof slope is less than 45 degrees I'd tend to see it as a mix of the two, probably leaning more towards diaphragm action, but I'm not aware of any defined cutoff angle between the two approaches.

 

[Bisbee_Structural]

A really steep roof would be like one of those old cool A-frame structures- and as far as I know the rafters on those go all the way down to a stem wall/foundation, avoiding the need for ties or a diaphragm at the eaves.

 

[phamENG]

12:12 (45[sup]o[/sup])or less would be diaphragm action through the sheathing using the flat roof assumption. Anything steeper and I agree that the efficacy of the sheathing to act as a diaphragm probably starts to drop off quickly. For that, looking at it as Bisbee mentioned is a good approach.

You can still consider the sheathing, but I would no longer use the equivalent flat roof assumption. You need to consider the actual vectors and the implications for directions of the reactions. You'd also need to pay closer attention to the gable end shear walls.

In any case, if you're looking at it with FEM software...don't. I've tried. It's a nightmare. Sometimes I'll give it a shot with complex roof cross sections with dormers, attics, and multiple pitches, but it's only a 2d model. Trying to look at it in 3d is overly complex and can give you some weird results.

 

[Bisbee_Structural]

Definitely agree about the FE software- a few years back I modeled a complicated wood roof with hip rafters, valley rafters, etc. that had to be modified as part of a renovation, and because of the complexity of the proposed changes I didn’t feel comfortable just doing hand calcs. It took ages to get it to work properly- fiddling with releases, setting up the sheathing, etc. and still there were lots of things that didn’t make sense.

In the end I had to use the tables in the AWC residential design references to demonstrate that the hip rafters worked anyway, because my model was telling me that hip rafters don’t work which is clearly not the case!