David, I'm glad you clarified because "The solution has always been to let the rafters rest unattached on top of the plates and purlins, pegged or nailed only to the ridge." does not equal strapping but still the strapping used in high wind resistant construction is not either floating or unattached.

I would also like to add, in the tradition I accustomed to, framing is done with very well seasoned fir timbers -seasoned before any layout and joinery is done- so movement in the final frame is slight, especially compared to framing with green oak.

The reason for this is the very problem Jon is facing, very exposed framing that must continue to look well fitted years into the future, so we use seasoned wood and work with extremely slight tolerances (like, 1/32" in any exposed joinery, and 1/16" in blind joints)

I would also like to state, I don't disagree with Gable by any means, but I should say that his approaches named are just one way to approach things.

In Bernese framing, we mix elements of log building into timber framing including the stacking of deep section timbers at the union of two stories of a structure, and the seating of ridge beams on top of deep section timbers. These things can be done, BUT you have to design for them. Gabel's stated approach is to avoid these things, so that you don't have to go to the trouble of designing for them.

The floating roof is one example of designing for these things.

Strapping as I mention it is, a steel strap nailed to the plate, passed over top of the rafter, then nailed back to the plate, nowhere nailed to the rafter, so that the rafter is free to slide through the strap if it needs. But this isn't really done. The weight of the roof is going to keep it down.

The Storm Lothar several years back took roofs off of one or two houses, this was an Atlantic hurricane that turned inland and struck Central Europe in 1999 or 2000.

This type of construction uses no manner of birdsmouth or other rafter to plate joint. The rafter passes over the plate, allowing us to easily make generous overhangs (like in the Emmental, where overhangs are about 12 feet)

As a side note that may interest some, when framing became more industrialized in Switzerland, a plate seated rafter roof was adopted, mostly in response to the need of a completely open attic space for open hay lofts and driving floors. But they still wished to preserve the generous overhangs responsible for their houses lasting 500+ years. So they inserted lower rafters keyed into the primary rafters, but set out at a slightly shallower angle so that they could kick past the plate. This became a distinctive architectural characteristic of the Canton of Bern from the 18th century on.

Thanks for these comments everyone. As DL alluded to I ended up with a ridge because I was designing for a shallow pitched roof (4:12) to not make an imposing structure where the space under the roof will never be used. I like the idea of the sliding rafters and would probably strap them for peace of mind. I will have plate and ridge of equal depth so it seems that the only shrinkage the will be relevant is that of the tie beam shrinking away from the bottom of the ridge post? This will result in the same scenario of the peak dropping as the ridge post settles onto the smaller tie beam, no?

DL, are you saying that the full rafter just sits on the arris of the plate? no notches? What connection would you recommend at the peak? tongue and fork, or should they just sit on the ridge and not meet each other?

Thank you Gabel for the mathematical piece of it. are you assuming something around 5% shrinkage? Also I didn't follow how you came to "13" of oak" from the example with a 8" plate and 12" ridge.

very appreciative of this forum and your time, thanks.

Uplift should be considered, an open structure and high winds.

Jon,

Sometimes we just sit them on the edge, some times the whole timber is set rotated to match the rafters, sometimes a 5th face is made to match the general slop of the rafters, and sometimes the plate is only reduced at the rafters, so that they slide through a notch, this is the method to use if the attic space is to be insulated.

At the peak, rafters don't need to be strongly joined to each other, unlike a standing rafter roof, the top connection between the rafters here does nothing. So they can be butted and spiked, half lap and peg, or even just set beside each other, which is how it was done when they used log rafters. You peg or spike them to the ridge.

I missed the part of it being an open structure, in this case I would definitely recommend strapping for the rafters.

I got the 13" in my example from 9" of tie beam and 4" deeper ridge than the plates.

Will/do the rafters sit atop the ridge with a birdsmouth or house into the side of the ridge as Roger is alluding to? If the latter, he is correct and only the distance below the rafter bearing should be considered in shrinkage planning.

If I recall correctly, oak shrinks ±5% radially and ±10% tangentially from saturated to oven dry. (I could have that backwards). While in forestry school, I did learn how to calculate estimated shrinkage for different species and different cuts of timber/lumber between two moisture contents, but observation and experience have informed my estimates now Over the years, we have taken several blocks of different species that were green and measured them, marking the measurements on the face indelibly and then checking them over time to observe actual shrinkage. I also like to take a tape measure when visiting past projects to check shrinkage.

I plan for/expect 5-6% shrinkage in oak from saturated moisture content to equilibrium MC.



Apologies DL, but I don't see how unconnected rafter feet could meet modern building codes with regards to uplift. Especially in an open building.

Gabel,
true, which is why I stick to strapping my rafters. Other more complex methods of a sliding joint/fastener could probably be designed too

With the oak I have used I have been very pleased with the stability and relatively low shrinkage. This has been winter cut and quite nice in grade. Also anchor sealed and oiled. There has been lots of small checks but most not more than an 1/8th inch.