Timber Roof Spread 6

[damo74]

Hi all.

I'm designing a timber cut roof with a lack of triangulation. It typically consists of 2 rafters supported at the feet on block walls. There is a raised collar tie about 1/3 depth from the ridge. There is no ring beam at eaves.

Has anybody any experience in what to check for this case?

 

[DaveAtkins]

I run this situation on the computer (I use RISA 3D). In my experience, the connection between the collar and the rafter is most critical, followed by the bending-axial interaction in the rafter.

DaveAtkins

 

[msquared48]

Right Dave.

And you'd better consider the deflection of the rafters the collar tie is pulling on. Might have to make the rafters larger, maybe double them up with a single member collar tie. Then you get double shear on the bolts too.

Mike McCann
McCann Engineering

 

[damo74]

Thanks guys.

I'm actually more concerned about possible supporting wall spread. What lateral deflection limits do you use for the wall? In the past I've used H/300 as my limit.

Another option I've checked before is to use a pinned support at each of the feet of the rafters and applied the lateral load from the rafters to the wall. This will induce an overturning moment about the bottom of the wall. The resisting moment of the wall is the (weight of the wall + dead load) x lever arm. I've never been able to prove this however.

Any suggested checks on the supporting walls?

 

[ctcray]

SOunds like you need a structural ridge to prevent your rafters from deflecting, thus preventing the spread.

 

[DaveAtkins]

Allowable deflection at the top of the wall depends on wall finishes. H/300 would be fine for masonry--I think you could allow H/180 for less brittle finishes. Remember that if your model has a roller at one support, and a pin at the other, the horizontal movement at the roller can be divided by 2 (half of the movement is taken by each wall).

I disagree with trying to resist a horizontal thrust with a wall, unless it is reinforced concrete or reinforced CMU. Steel cantilever columns might also work. But I can't see a wood framed wall resisting the thrust.

DaveAtkins

 

[damo74]

Thanks guys.

I've actually used DaveAtkins' solution in the past-I was just looking for some comfort in my logic I suppose!

Also DaveAtkins 2nd point I would agree with, as my wall overturning calculations would suggest.

 

[csd72]

In some places they call this an A-frame truss due to the shape.

Rafters cantilever out from the collar ties.

Lateral loads taken by sheathing.

It can all be done pretty easy by hand treating the rafters as a cantilevered continuous beam.

csd

 

[msquared48]

Really, there are three ways to control any lateral wall movement here short of adding a ridge beam:

1. Increase the rafter size; or

2. Increase the E of the rafters; or

3. Lower the collar tie.

Just pin one end of the collar tie, put the other on rollers, and play with the results. The less bending in the rafters, the less lateral movement.

Mike McCann
McCann Engineering

 

[csd72]

or

4. put a load bearing ridge beam in.

 

[msquared48]

As I said csd, short of adding a ridge beam.

Mike McCann
McCann Engineering

 

[csd72]

sorry Mike.

 

[msquared48]

No worries csd.

Another thought, and kind of a corollary to the idea of lowering the collar tie, is to increase the roof slope, leaving the collar tieat the same elevation. The effect is to decrease the tension in the chord tie, and may decrease the rafter deflection. You may be able to reduce the live load with increased roof slope too which would further reduce the tension force. Downside is the rafters are longer which would have the effect of increasing the deflection. The combination though should lower the overall lateral deflection.

Final thought. Even with a ridge beam, you can get lateral deflection of the walls, but to the inside of the structure, not the outside. You compensate for this either by cutting a portion of the rafters flat for flat bearing on the wall top plate, or through the use of sloping rafter connectors as made by Simpson and like companies.

Mike McCann
McCann Engineering

 

[apsix]

Has anyone tried using the roof sheathing as diaphragms to support the roof? I'm just thinking out aloud, never tried it myself.
Obviously a steep roof pitch and appropriate room geometry would help.

 

[damo74]

Another idea I thought of was treating my wall as a laterally supported wall panel supported on 3 sides by the bottom slab support, and 2 buttressing walls.

I've checked this panel with the lateral ud line load from the roof rafters applied to it, spread it over an acceptable width and it works structurally in accordance with normal wall panel design. This should prevent my roof spread.

Thanks guys for the discussion.

 

[HouseBoy]

I'm not clear on how you are doing that.
What is resisting the lateral load at the top of the wall? Are the top plates? What is "normal wall panel design"? I have looked at walls supported on 3 sides but they need to have some bending capacity themselves. Usually, I am looking at a concrete wall fo rthis behavior (like a basement wall). What is it about the stud framed wall that gives it capacity to span horizontally?

 

[damo74]

Houseguy,

Firstly, the wall is supported on 2 sides by buttressing walls and on the bottom by the floor construction.
Secondly, the wall is constructed from masonry, not timber frame and therefore it has some bending moment capacity like a flat plate. The design is in accordance with BS5628.

 

[HouseBoy]

My bad. I overlooked that in your OP.
What is the height and length of the CMU wall? I'd be curious about the horizontal deflection at the top of wall.
Has anyone used the roof sheathing as a diaphram to resist this thrust? It would be great if one could but I'd expect theoretical problems with spaced sheathing (H clips and all) as a diaphram along with other things like a possible lack of chords.

 

[damo74]

The wall height is approx 1200mm high and 4000mm long between butressing walls.

Haven't looked into using the roof sheathing as a diaphragm.

 

[valleyboy]

Hello Damo74

I wondered if you had considered the use of a structural eaves beam to span 4m between the butresses to resist the horizontal thrust? This could probably be achieved in timber over that span, or a small RHS.

Another solution that I have used sucessfully over this sort of span is as suggested by apsix above by incorporating a ply-web beam within the depth of the roof construction to resist the axial loads in the rafters.

With your proposed solution to consider the panel as supported on three sides and subject ot a line load at eaves level, there is often difficulty in justifying the panel to work subject to lateral wind loads without the top restraint - this may not be a criteria for a panel with the geometry you have noted.