Skip to content

Desk No. 3 – Design

Desk no. 3 in oak detail

My wife and I desperately need a desk. Since corona arrived my wife has been largely working from home, and has been using my old workbench as a makeshift desk. The workbench works great for one person, but it’s impossible for the two of us to use together. We’ve patiently made do with this situation, but it feels like its about time to make something that we can both use.

My initial idea for this desk was to make something relatively simple that I could fabricate with leftover material I have on hand. I also wanted the design to be made with demountable joinery for when we inevitably move. After building the heavy duty shelves that I posted about recently, I thought that the desk would be another great opportunity to use kata-sage-ari joints. Those joints really worked out well on the shelving. They are really solid, and I find it incredibly gratifying to build a structure using traditional joinery, held together with wood on wood connections secured with simple wedges. It doesn’t get much more satisfying than that.

So after taking some measurements, and considering the joinery I wanted to implement, my initial concept for this desk came about pretty quickly. I chose to keep the overall design relatively simple, reducing it down to a desktop and legs. No drawers, cabinets or special compartments. Just an old school desk. I designed the desk and rendered all of the images shown here in Blender.

Desk No. 3 initial design

In this initial design the main structure composed of the legs and stretchers are all designed to use the kata-sage-ari joints. Initially I drew the legs square, but quickly made them round. I really love the look of round legs, particularly with rectangular parts like the stretchers piercing through. To my eye it’s somewhat reminiscent of tea house construction, which I’ve always admired.

In the next iteration of the design I tapered the legs to make them feel a bit lighter, and adjusted the size and location of the stretchers a bit.

Desk no. 3 wide nuki

The original size of the stretchers felt too heavy, so I lightened them up on the sides, while keeping the back stretcher wide (in a later iteration, I lightened up the back stretcher too). The other big change came in the form of widening the upper cross piece that joins the legs front to back. Those cross pieces also support the desktop and the desktop understructure, and this is where the structural challenges of this design have popped up. In effort to maintain the demountable nature of this piece, I’m leaning towards using sliding dovetails in a number of places. The downside of sliding dovetails is that you end up having to remove a fair amount of material which can have an negative effect on strength.

This is what the understructure of the desk looks like.

desk no.3 understructure

I’m currently thinking that I would like to join the wide cross pieces to the desktop with long sliding dovetails. I’m also thinking about joining the understructure to the wide cross pieces using sliding dovetail joints. (The long dovetails look a bit weak, but that’s simply because I drew them by eye in Blender. I later confirmed that these joints will be plenty strong after adjusting them to the proper angles.)

Desk no. 3 cross piece joinery

Using sliding dovetails to join the long understructure rails will require taking out a good chunk of material from the cross pieces. So that was one of the reasons I wanted to make the cross pieces wider in order to maintain their overall strength. Doing so also maintains a solid run of lumber for joining the front and back legs. Lastly I curved the ends of the cross pieces to account for the kata-sage-ari leg connections, which need to be adequately offset from the end grain of the legs to prevent the grain from blowing out when locking these joints with the wedges. All of this came together in a pleasing way, with all of the joinery and design considerations feeling purposeful.

One area I’m still not sure about is the other cross pieces that make up the understructure. I absolutely need the long pieces to support the desktop from sagging, but I’m less sure about the three middle cross pieces. The wide cross pieces on the ends, with their long sliding dovetails, will be doing the main work of securing and keeping the desktop flat, but I’m inclined to add the middle 3 cross pieces for additional support. I’d rather overengineer things a bit, especially if I end up building this desk for a customer some day. The problem is how to join these cross pieces to the long support pieces as well as the desktop itself.

One option for joining the cross pieces to the long supports, is to again use sliding dovetails, but that means weakening the long supports. Another option I’m considering, which I think will be less of a burden on the long supports, is to use a pair of wedged through tenons to join the long supports and middle cross pieces. The downside I see to using those joints is that the understructure will become more or less permanently assembled, and I start to move away from my goal of making the whole desk demountable. But I think I’m willing to accept that. I’ll still be using wood joinery after all, and for the sake of the structure I think these joints are going to be a good way to go. Plus they look great.

desk 3.0 understructure wedged tenon joinery
desk 3.0 understructure assembly

Finally there is the issue of how to join the understructure itself to the desktop. Given my joinery choices thus far, assembling the desk will have to happen in a specific order. The assembled understructure will first need to be attached to the wide cross supports on the ends, followed by sliding the desktop onto the long sliding dovetails, and from there the legs can be attached. That leaves just a couple options for joining the understructure to the desktop, and the options I see are either more sliding dovetails or screws. And at the moment I’m leaning towards using screws.

It’s tempting to be a stickler about using only solid wood joinery, but as the design stands the desktop will have to be slid onto the two end cross supports’ sliding dovetails simultaneously, which is already a bit of a challenge. It’s going to take a fair amount of care to slide the desktop onto those joints, working evenly from each side to prevent racking and damage to the assembled understructure. If I added more sliding dovetails to the mix on the middle cross pieces the assembly process would become even more stressful and risky in my opinion. Hence, I’m leaning towards screws. In the end, the sliding dovetails on the end cross pieces will be doing the main work, the screws are just a bit of added assurance.

So that’s where things stand at the moment. I’ll probably make some adjustments here and there, but overall I’m narrowing in on the final design. Despite my initial goal of making this desk quickly and simply, things have gradually gotten more complicated as I delve more into the nitty gritty bits of the design. On top of that, after looking more closely at the leftover stock that I thought I would use to build this desk, I realized that many of the boards are way too warped to use. So things are levelling up again and I’ve decided to build the desk out of Japanese white oak. There’s a stock pile that one of my shopmates recently bought, and the price is right.

I’m pretty excited about this design. I’ve wanted to build a desk for a while now, and building it well, with quality materials, will allow me to have a more professional finished piece to share with customers. And personally using the desk will allow me to see how the design fairs over time. So stay tuned, I’ll be sure to share the construction details as soon as that process starts.

Thanks for reading.

Desk no. 3 narrow nuki
Desk no. 3 narrow corner detail

6 thoughts on “Desk No. 3 – Design”

  1. I appreciate your analysis of strength, joint types and other considerations of your design processes (multiple here). I’m designing a desk with a hinged top for my wife and this blog could not have come at a better time. Thank you. Looking forward to next week’s updates.

  2. I like that design very much. I especially like how the end cross pieces are sculpted and relieved at their ends to accommodate the kata-sage-ari joinery. And the inset rounded legs add some pizzaz.

    For what it’s worth, I think you definitely need the intermediate cross pieces to keep the long rails from twisting and to give you attachment points to keep the top from flexing up or down if it bows in the middle.

    On balancing demountability vs stability vs exposed joinery for those intermediate cross pieces: If you didn’t want exposed joinery and wanted something demountable and not visible, you could, I think, turn those intermediate pieces so they are flat and let them into tops of the the front and back rails with one wide or maybe double dovetails. I will say that I don’t think it is a great solution, but an option. Or leave the intermediate cross pieces vertical and let them into the front an back rails with sliding dovetails that go only 1/3 or so down from the top, as you did on the ends. Then screw the top to those.

    You’ve probably already considered these options and more. I’m just having fun playing along.

    1. Thanks for the comment Gary.

      I like your suggestions for the intermediate cross pieces. I didn’t consider turning those pieces flat, but that is certainly another option, and using dovetails in that situation would also be less taxing on the long front and back rails. In the end though I think I want to keep those intermediate cross pieces vertical to use their strength to hold the top flat, and with that in mind I think you’re suggestion of cutting short sliding dovetails could be another possibility, and would be pretty strong especially if I inset the end of the intermediate pieces into the long rails by an 1/8″ or so to resist racking.

      But I’ve actually grown to really like the look of the through wedged tenons. They aren’t readily demountable, but I think they will be plenty strong and I also realized that sliding dovetail have the potential of being pulled up and out if the desktop were to warp in a weird way…. probably not a huge concern, but something about making the whole understructure more “locked” is starting to appeal to me.

      After writing the post another potential alternative to using screws for fastening the desktop occurred to me, that being sliding dovetail keys. If I went that route, I’d have to also use those same joints in place of the long sliding dovetails on the end cross pieces.

Leave a Reply