Okay one more design post for this desk build. I never thought this would take 3 posts to work through just the design phase of this project, but there’s a lot to consider when designing a piece with traditional joinery while taking wood movement into consideration. Certain areas are particularly tricky such how to secured wide flat pieces like the desktop of this desk.
To quickly recap, I left things off talking about a solid wood joinery approach to securing the desktop to the understructure. Having decided on something similar to the inset dovetail keys I shared last week, I decided to go ahead and quickly modify my 3D model to start visualizing that approach.
In addition to the joinery, one change I made was to reduce the intermediate cross pieces from 3 to 2 pieces. I ended up reducing the overall length of the desk from 2 meters (6.56 feet) to 1.8 meters (5.9 feet), and doing so meant that I had plenty of support in the middle of the desk using just 2 cross pieces.
I’ve come across a few different names for the segmented dovetail joinery, including some abbreviated names, but at least one book of mine calls the joint 送り寄せ蟻吸い付き桟 (okuri-yose-ari-sui-tsuki-san). If you do a google search using those characters you’re sure to get some hits. To simplify things here I’m going to call the joint okuri-yose-ari, which is one of the common names for an inset sliding dovetail joint.
One of the reasons a person may want to use okuri-yose-ari is to support a tabletop that is visible on all sides. Typical sliding dovetail cleats require the joint be cut straight through the edge of at least one side of the piece, for the dovetail cleat to slide in place. In those situations I’ve seen people plug the exposed dovetail cut, which certainly works and if done well can look okay. But for those situations which demand leaving the table top edges untouched the okuri-yose-ari is a great way to go.
This desk will initially be placed against a wall, meaning I could get away with sliding dovetail cleats, leaving the exposed cut on the backside of the desktop. But who knows what the future will hold, and if there comes a time when this desk sits more out in the open, I don’t want to have an open dovetail cut visible, and I’m not all that interested in messing around with plugs either. So for this project, that’s another win for the okuri-yose-ari.
Doing a bit more research into these joints, I’ve come across a couple different variations for how to make them. One variation involves cutting a standard long sliding dovetail joint and then cutting a series of notches to essentially break the joint up into smaller chunks. That’s exactly how I modeled the joint in the image below.
Here’s an image of the same variation of joint along with its housing.
On the 5-leg coffee table that I talked about last week, I used inset keys in lieu of a chopped up long dovetail. I really can’t recall how or why I came to the decision to use keys, or why I spaced them so far apart, but it did work out okay for that particular project. I’ve also seen the same method pictured in at least one Japanese joinery book. Keys worked out fine for me, but for the most strength I think it makes a lot of sense to cut the joint out of the one piece, as is picture in the images above.
The video below shows the same segmented variation of the joint being assembled. Here you can more clearly see how the joint goes together, with the female part of the joint clearly visible. The dovetail segments first sit down into the wide spaces that are cut to let the widest portion of the dovetail shape to pass through, and then once down, the whole piece gets tapped into place engaging the male and female dovetail parts.
In one more variation of the joint instead of completely removing the material between the dovetail segments on the male part of the joint, only the sides of the non-dovetail parts of the joint are cut back. I think this method makes a lot of sense for maximizing wood strength by removing the least amount of material.
So once again I went back into Blender to play around with my model a bit and redrew the joinery to illustrate this last variation of the joint.
I really like this joint, and I’m excited to give it a shot.
One final detail that needs to be considered when cutting an okuri-yose-ari joint is what happens when the two parts of the joint are slid together and fully assembled. The female part of the joint needs to be made at least one segment longer than the male part to allow the dovetail segments to be fully inset before sliding the dovetails together. That means that when the joint is installed there will be an open space that needs to be addressed.
In the image below you can see how the person who built this, offset the dovetails on the cleat, leaving a wider open space towards the right end. Offsetting the dovetails that way means that when the joint is assembled the right end of the cleat will effectively cover the open space that remains in the table top.
Here you can see the cleat/leg assemblies installed, with the cleat fully covering the last open space.
In this image from the same blog, but a different project, you can see just the cleats installed before the legs are attached, and on the left the open space is clearly visible. I’m guessing the legs may end up covering that opening.
Previously I mentioned being wary of using multiple sliding dovetail cleats on the desktop understructure, since during assembly all of those joints will have to be carefully slid into place simultaneously. Okuri-yose-ari have the same issue, which is one downside to using these joints, but given all the other benefits discussed so far, I think I’m willing to embrace the challenge. At the moment I’m thinking I’ll use clamps to carefully slide the cross pieces into place, which should offer plenty of control and allow me to work each location little by little.
That’s about it for the design phase of this project. See you next week with some photos from the build phase.
You’ve got this!
I’m eager to see how it turns out. I think you are right that assembling all four cross rails into their housings in the top is going to be a challenge. But maybe not. Do they all have to go in simultaneously? I’m thinking install one cross rail free of the long rails, then the next, until all four are in. Then the long rails tapped on to the assembled cross rails. Then the legs.
One concern I might have, never having tried this, is that the segmented cross rails seem thin, like maybe 15 or 18 mm at the narrow neck of the dovetails? I’d worry about those dovetails breaking off if the top really wants to cup. But I know you previously posted that you have reconfigured them. The pix and video you referenced showed much wider battens. But also maybe a thicker top than you have in mind. It might not be an issue at all, or if it is then make the cross rails thicker/wider in the next go round?
Such interesting joinery, and thanks especially for the Japanese translations and the Japanese script. So useful to me to have both for further study.
Thanks Gary.
I thought a lot about different ways to construct the understructure in a way that would allow me to assemble the parts one piece at a time like you described, but I wasn’t able to find a joinery method that seemed quite right. So the compromise is to preassemble the structure and install it as one piece. I’m using wedged through tenons for joining the long rails to the intermediate cross pieces, as well as the cross pieces on the ends, but those two areas are at 90 degree angles to one another, so there’s really no way to install the long rails after the fact. I plan to take some good photos/video of the installation so hopefully that will all become more clear when I get to posting that.
As for the thickness of the pieces, I’m definitely pushing things towards the thin side, but I’m quite confident that the joinery won’t fail. Overall I think the thickness of the top at 22mm thick makes it “pliable” enough that even if it wants to put up a fight down the road, the cross pieces should be plenty strong to resist the movement without breaking. But it’s definitely a valid concern, and I think about those issues a lot when designing a piece.