My question is which joint is the preferable joint to use in joining rafter feet (joined in pairs with tongue and fork, no ridge beam,no purlins)to a top plate. I notice a lot of designs with rafter to plate joinery utilizing the step lap as opposed to the shouldered birds mouth. Is the step lap a better connection here? I tend to design using the step lap as it allows to possibly have a tail extend for roof overhang, which brings up another question. Whether to adze down the depth of a rafter (say 6" down to 4") allowing 2" for the step and 2" depth for the lap,which allows the tail to be flush with the plate? Or one could technically not reduce the depth and allow the 2" for the step and 2" for the lap, it would still have 2" above the plate,but be 4" in depth at the tail and would eliminate the adzing step. It is also questionable whether to pin this joint on the tail or step portion of the joint.I wonder which of these is the strongest joint and altogether which joint is preferred all things considered?

If you don't make the top of the rafter flush with the top of the plate, then there will be a gap there. The siding usually gets nailed to the plate with the top end of the vertical siding being even with the top of the plate and is notched out around the over hanging rafter.
If you don't have the rafter flush with the top of the plate the vertical siding can extend up to the top of the rafter and butt against the roof decking somehow. This would require some type of fastening up there to keep the ends in line and or hold them somehow, I would think.

If you want to do a shouldered birds mouth, you could add a sprocket to the rafter to gain the roof over hang.
I believe the Historic joinery series shows at least one sprocket....
If not I have a picture of one....

Jim Rogers

In section five of the historic American timber joinery on page 3 is this drawing of a sprocket:



Hope that helps...

Jim Rogers

Hi TB:

Just a quick comment on the diagram above,

I personally have never seen a rafter gained like the one above against the upper plate, now that is not saying that it couldn't be done, but in retropspect it probably would only be for a very small building using sawn material and sawn rafters.

Rafters usually are cut on their feet to the run and rise of the roof, and lay flat on a seating that has been prepared on the surface of the upper plate. On the bottom of the rafter's foot there is a cog that is shaped as the foot is prepared. This cog can extend across the foot or can be located in the centre. The front of the cog is at a 90 degree angle to take the downward thrust of the rafter, while the back slopes on the same angle as the rafter.

The cog sits down in a prepared notch cut into the top of the plate, these notches for all rafters are carefully lined up along the upper plate about 4" (+ -) from the outside upper corner. After the rafter is in position one wooden peg, or 2-6" wrought iron nails driven in from opposing sides hold things in place.

This type of fastening for rafters is one that has been used successfully over many generations, and as long as no rot happens will hold securely under all conditions.

NH

Hi,

When I noticed the above diagram I agreed the connection looks weak in the fact that the birdsmouth on the rafter might sheer right at the corner of the top plate. I think NH is describing a solution but I still do not know what a "cog" is. I have learned to cut a housing on the inside of the top plate to take the bearing away from the birdsmouth. But this can pose problems too if working with green timber. If the seat cut is sitting right on the top of the plate and bearing on a housing shrinkage can in effect negate the housings strength. As the rafter dries in this scenario it would pull the rafter up off of the bearing point in the housing. As crazy as it sounds would it not be better when working with green timber to leave say an 1/8" gap between the seatcut and the top of the plate to account for the shrinkage (as long as the housing is deep enough with enough relish to support?

Then as the timber dries it will sit nicely.

that does sound crazy Mo and I am not sure I quite understand, but it just may a solution. This is exactly the reason I dont like the birdsmouth joint as much as the step lap, that it likes to split the rafter in the "mouth", unless you house the joint with bearing on the plate, but then the shrinkage might put the strain back on the birdsmouth joint. In addition the step lap lends itself more easily to having a tail. Perhaps a lot of people dont use tails much anymore with sips tf? tb

I believe northern hewer refers to a step lap.

A cog is not a step lap, I'll post a drawing of a cog later today.....

Unless I'm mistaken this drawing from page 2 of section 5 of the Historic American joinery book shows a rafter with a cog.

Is the cog on the rafter or the top plate? I think it's the wood remaining on the top plate? I have been using cogs instead of lapped dove tails on the bottom of tie beams in english tying joints. (see the Yellow book). Tim