A little background:


I'm planning the build of a timber-frame house in the "American" style (long beam) in Normandy, France. I have a local sawmill with CNC machine who are prepared to cut the frame for me (I'd love to do it myself, and may still, but time is a big issue). However I'm battling slightly against the local style of timber framing. Houses here are mostly blockwork with timber roof structures and the local carpentry style is very specific. The carpenter is prepared to push himself a little, but he's already outside his comfort zone and we still haven't met eye-to-eye on the roof. The bents are pretty simple, but do include an assymetric roof which has a height restriction on the underside making a simple tie-beam impossible.

What we are trying to achieve is actually a principle purlin frame since the rafters will be I-joists packed with insulation, but the narrowness (4.5m - 15ft) of the house precludes a centre (or in this case off-centre) post. The carpenters solution is a king post which the rafters meet, the triangle being formed by a pair of rectangular section timbers clamped to either side (Known in french as an assemblage moisé). In a normal building, the king post would be a square timber, the principal rafters and "sandwiching" beams would be rectangular and the king post would be notched so that it was sitting on the paired beams. There are various reasons why this is not an ideal solution for us, but lets assume that it's an aesthetic that doesn't work for me.

Possible solutions.
1) Change the design. This simply isn't an option. A change to this would involve raising the height of the house and would involve a new planning application, not to mention leaving us with more unused space overhead.
2) A half dovetail on each end of the collar tie. The problem with this is supporting the king post. It would allow it to be removed altogether, but then the ridgebeam would have to be joined to the rafter mortise and tenon which is going to be an exciting joint!
3) A pair of collar ties with half dovetails (in rectangular section) replacing the paired beams and metal connectors with traditional joinery. I would however be concerned that the rafters would have to lose a lot of material for that joint.
4) Replace the collar ties with a single beam with a spline pasing through the collar tie and rafter. This spline would then provide the tension resistance that this joint needs. (Obviously this drawing is just to indicate the general idea. Neither the spline nor the peg holes are to scale.) This is currently my preferred idea.
5) Take the king post down to the level of the first floor girt so that the ridgebeam is supported by this girt. This would add a not-insignificant load to this beam and would play havoc with the floor joists (the house layout depends on the joists running perpendicular to the bents and not parallel to them.

So over to you. Any thoughts? Any suggestions? Any questions?

Jon

Hi John,
Have you consulted a timber engineer on this design?
Although the collar joint may prevent the triangle above from spreading, the whole assembly, especially with the eccentric loading on the long rafter, seems to put a great amount of stress both the extended wall posts. Purlins will add even more loading to the rafters. Once that long rafter wants to sag below the collar the problem will intensify.
Option 5 looks like the best solution to me.

The carpenter who would be building the frame (CNC cut) was responsible for the original design. At that point we weren't sure that the roof design was possible. The beams will be 8"x8" (Well 190mmx190mm actually), but I'm not so sure about calculating the deflection on the long span. Dropping that king post down to the first floor beam below will be touch and go. We'll be at 90%+ of it's capacity before snow / wind loads and I don't know if we can source the timber to scale up the first floor beam enough to take the loading. Taking the timber down to the foundations, while appealing, just isn't an option.

Loading on the rafter isn't an issue. The roof structure will be single span I-joists from top plate to ridge beam and I already know that they can take it. My concern is getting the load from the ridge beam down to the posts without something snapping!

I'm going back to the carpenter soon, but I wanted to go with options, and the more I look at this, the more I'm convinced we don't have any.