I’m finally completed some work that was under a N.D. Agreement, so I can return to regular programming, blog-wise. Look for another post in the Dark Chocolate and Sponge Cake series soon!
Meanwhile, I can talk about a small project I tackled over the last couple of afternoons in my frozen cold shop: a wooden serving tray. Maybe it is hardly worth writing about, given that it took less than 6 hours to put together, but it did allow me to explore some fun joinery that came as a result of considering various options for connecting some sort of feet to the tray top.
I’ve made many sushi geta previously, which are simple wooden tops with transverse battens, held in to the top with full length sliding dovetails, and serving as feet. If you google the term you’ll see what I’m talking about.
The issue with that sort of construction, even with a top as narrow as 6″ (155mm) across, is that seasonal movement in the top across its width will leave the end of the dovetailed battens either protruding or slightly recessed at different times of the year. It’s no big deal, considering what they are used for, but it is nonetheless a drawback to that form of connection. The exposed dovetail joint is also a decorative feature on the geta, a sort of message most people will get, if that’s going to be part of it, design-wise. If you’re not interested in such messages, then it’s a drawback to have the joint exposed.
Also, while a flat-bottomed batten is a relatively simple thing to produce, in order to have a piece that will be less prone to rocking from bottom surfaces that are not co-planar, or are bowed slightly by movement of the top, one might resort to hollowing out a mid-portion of the batten undersurfaces to produce, in effect, a pair of ‘feet’ at either end. This is what I’ve been doing anyhow, and on recent pieces I’ve taken to more deeply machining out a hollow in the bottom surface which leaves much more pronounced 2-footed look.
With a wider top board, even if one were to employ quartersawn material, the dimensional difference between the width of the top (which shrinks and swells) and the length of the battens (which hardly change length at all) would be evident at both the driest and the wettest time of year, so it is an even less ideal solution when you get into wider tops.
And if you’re going to shape a batten’s undersurface to produce 2 contact points widely apart, then you’re well on your way to another form of feet for the tray altogether, one consisting of 4 separate feet instead of 2 battens. That way the feet can float back and forth with the movement of the top board. The batten doesn’t serve much purpose besides that of a foot in a sushi-geta, so though simple, I thought more about having 4 independent feet on the tray.
Along those lines, I considered different ways one could connect those feet to the top. There are plenty of options, and in thinking about placing feet so that they meet the perimeter of the top in some locations, one consideration that came to mind was thinking ahead as far as the chamfering is concerned. It’s an issue with the regular exposed end of the sliding dovetail batten form of construction – it is a joint interface which doesn’t lend itself well to chamfering unless you also partially house the batten sidewalls into the surface.
So, with those considerations in mind, this is the solution I decided to explore, by all means not the only one, but one which I think may have a certain usefulness. Here’s an overview of the connection, assembled:
I’m fortunate to have an off-cut chunk of quartersawn Honduran mahogany kicking around, and that seemed like a good piece with which to make a large serving tray. In preparation for the layout and joinery, I leveled the bottom surface of the board:
Despite a clean appearance, the board had a 3″ (76mm) long slightly scalloped band in the surface, which you can faintly see in the above photo after the first round of passes, and there was no point taking many shavings to get rid of that dip when just 3~4 heavier passes, one of which you see above, did the job. Obtaining a clean, tool mark-free surface is the goal regardless of how many passes one takes, and that was managed somehow.
Then, entirely on the mill, using a sequence of cutters and operations, I produced some double-dovetailed sliding trenches at four points:
A closer look reveals the double dovetail trench:
Then some chisel work followed to clean up the interior abutment:
A small block of tight-grained cocobolo, 4″ (105mm) square and 1.25″ (30mm) thick provided enough for the four feet:
As the run of the grain is aligned between the top and the sliding dovetailed feet, I could have glued without the usual concern for seasonal movement working over time against the connection. But instead I decided to do the connections dry, using just the friction of fit between the parts, and trust that to hold up over time:
This piece will be in use in our home, so I can observe how this dry fit joinery approach plays out over the years. I’m not much concerned though, as I’ve plenty of other pieces with dry-fitted joinery in my living room and they’re doing fine.
Besides, cocobolo, being an oily wood, is tough to glue and my strong preference for that task would be to employ a special epoxy that I have for oily woods. While mixing a bit of glue is no big deal, after a brief mulling over I concluded ‘some other time perhaps’ for that way of doing things. Let’s try an alternative route.
The foot starts its journey, a hand press-in to start:
Then a mallet does the persuasion:
I know there should be final ‘all the way in’ pic inserted here, but I forgot about taking a picture, so…
…and there we have it, four feet fitted:
I applied some wax, which brings out the cocobolo especially:
One can apprehend other uses for this sort of connection, when one sees the parts in the above orientation.
The chamfer follows a pleasing jog around the foot:
A view of the underside – the ends of the feet overlap the ends of their mortises by 1/4″ (6mm):
Before the top and feet were brought together, I did give the top and all edges a finish planing, so once the wax is rubbed out the piece will be done:
That’s our new tray, and has taken up residence inside the dining room cabinet. I look forward to using it.
Thanks for tuning in, and hope to see you next time.
Chris;
Very nice! Love the contrast in wood color! Great design in the double dove tail! Keep it coming. Thanks again!
J.T.
JT,
glad you like it. A lot of woods seem to pair well with mahogany, and rosewood, though difficult to work with, could be thought of as a tad luxurious for this application.Gotta do something with the offcuts though…
Cool, that is an excellent block of wood for the use of separate feet.
It’s always interesting to see the versatility of a well executed dovetail joint.
The piece appears so simple, yet elegant.
Beautifully done. Thanks for showing it.
I debated whether to post about it at all, but your comment, and others show that it was of interest, so I’m pleased about that. giving a sliding dovetail a beveled shoulder allows for certain advantages, so it was worth looking into with this tray, and will be something I use again in some future project I’m sure. Thanks for your comment!
Nice job, I like the chamfer detail. between the board and feet.
Mark,
yes, the entire joint was pretty well arrived at as a result of wanting to have a clean chamfer transition. I’m glad you like it, thanks for the comment!
I like the result even though I want to point out a few assumptions that you’ve made in the lead up to your design. One positive aspect is that you can use short sections of tropical hardwoods for the feet. You didn’t go into an explanation as to how you machined the dovetails or whether you began with a long piece out of which the four feet were separately cut.
You ruled out using transverse battens because, as I understand your thinking, of the problems due to seasonal movement and because the “battens don’t serve much purpose.” The fact that seasonal movement can cause uneven surfaces is how wood is. I think that in today’s industrialized home and work environments, this living aspect of wood can be regarded as a positively tactile feature. But more pointedly, this movement can be made less noticeable by allowing the battens to stand a bit proud from edges of the top plank. I think even more than seasonal movement is that a sushi geta is exposed regularly to moisture on only the top side from serving and washing. This moisture can create crowning, which, I strongly believe, battens can help to counteract. Additionally, battens can allow for increased rigidity when using a thinner top plank.
You made me think about the challenge of keeping a piece that is likely to move and change shape over time sitting level on a tabletop. Battens can be undercut so that one has two points of contact and the other has a wider contact surface centered at the middle of the parallel batten. Admittedly it’s also a compromise with its own different shortcomings.
I look forward to updates as to how this piece performs over the next year or so.
Potomacker, thanks for your detailed comment. Glad you’re still tuning in!
The feet came out of a piece of cocobolo which was 4″ square and 1.25″ thick, as noted in the post, so that answers your question as to which size of piece they came out of.
You also make several assumptions in your commentary about transverse battens. If one makes the batten ends stand out noticeably proud when the piece is made, then the amount they stick out when the top shrinks is even greater and this I do not view as a positive tactile feature. Further, since a portion of what sticks out is the sharp and delicate arris of a dovetail, it is not likely to wear (attractively) well over time. I would worry a little about a person catching a splinter from that portion. I would add that the degree of movement on a wider piece of wood top is going to be a lot more than what is seen on sushi geta, and that was what concerned me in regards to going the route of using battens.
The solution you suggest – making the parts deliberately proud from the get-go – is the same one you see with Greene and Green furniture in relation to breadboard ends on table tops. It is not something you see practiced in Japan or China from what I have seen, though that is not a strike against it otherwise. I employed the strategy on the ‘Square Deal’ tables, using hammerhead corner keys. One consequence of such a method is that it makes the joinery an even stronger visual feature of a piece, and that is not always suitable for a piece’s overall aesthetics.
Sushi geta may well get regularly exposed to water from washing in a restaurant context, however in my home we only make sushi about once per month at the very most, so the amount of washing they are exposed to is quite minor, and in any case it is surface washing, not a soaking. I think, therefore, that most of the seasonal movement with the existing sushi geta I have takes place while they sit dry on a cabinet shelf. And they move plenty I have found, especially noticeable at this dry time of year. Even with sushi geta only 6″ wide, the effects are clear. Maybe if the geta were very regularly exposed to water they might reach some sort of equilibrium where the movement between the parts is more narrow in nature, but that is not what happens in my house.
I am less strongly convinced about the ability of battens to counteract cupping (not crowning, as you state, which is movement relative to the run of the grain) across a board -it depends upon a few things in terms of the sizes of the parts, nature of materials, etc. When wood, especially flatsawn stock, gains moisture and starts to cup, the forces developed are quite high, higher than the modulus of elasticity (MOE) of the batten pieces to resist that load in most cases. Further, some woods have an easier propensity to bend than others, and may not inherently be able to resist movement in that direction.
Wood wedges can be placed in pieces of granite and soaked so as to effect the splitting of rock. That is to say, the loads that arise from wood gaining moisture are quite high. If wood wants to cup, I am doubtful that in most cases the MOE of the battens will do much to resist that – – in fact, if the battens were themselves from less stable material and were tending to crown via moisture gain, their movement may very well accentuate the cupping of the top board.
For the battens to have the greatest chance of constraining cupping, they need to be a significantly taller in section than the top board to which they connect, and, conversely, the top board is therefore kept ideally as thin as possible unless you want giant battens. The aesthetic of the thin board and chunky battens however is not typically at all for sushi geta, in which the top is invariably thicker than the batten sections. So, I think the ability of the thinner battens to counteract the forces of a thicker top wanting to cup is marginal at best, and when I put parts together like that I am not expecting the battens to control the top.
Even with that set aside, the sliding dovetail connection is an arrangement where the end grain abutments in the top’s dovetail trench will resist crushing much better than the side grain surfaces of the batten’s dovetail male portions. Even with a perfectly unyielding batten of some sort the loading induced on the batten dovetail males by the top trying to cup will simply lead to grain crushing on portions of the male dovetails, which, when the top board dries and returns towards original form, will result in a loosened connection as the crushed grain does not spring back fully.
I think the function of battens is also misunderstood by and large – as is the function of a breadboard end on a table top. Many folks concluded that these crosswise pieces stiffen a connecting panel against cupping. I think, when it comes to frame and panel assemblies, they are less about keeping the panel from cupping than they are a mechanical connection which allows the panel itself to serve as a shear resistant diaphragm connected to the door stiles via the batten tenons – the load which it controls is that of the door frame trying to distort under the influence of gravity. That’s why a door made in imitation of that look, but with the battens not trenched into the panel with dovetails but just on the surface, will not resist frame sagging for long, as all the loads would go to the frame corner joints. I think a lot of folks do not understand this function, and instead, if they think about it at all, conclude that the dovetailed battens are there to keep the panel flat. Again, unless the battens are significantly thicker than the panel/top, they will probably not be able to do much to counteract any tendency of the top/panel to cup.
With a table top, similarly, how can a breadboard end stiffen the table top against the loads induced by the boards cupping? Again, this is not something with a hope of functionality unless the breadboard end is quite a bit thicker than the table boards.
With sushi geta of course, the ability of battens to counteract shear loads is completely besides the point. A sushi geta is a simple piece with simple joints and there are not many good joinery options if you want to connect parts together in that configuration. Such arrangements are better confined to utilitarian functions (like sushi geta) I feel as it does not look so good, modern sentiments aside, when the surfaces are significantly proud of one another at certain times of the year. Maybe the layperson would find that offset pleasing somehow (I am skeptical that most would notice), but it is the sort of thing I am less excited about, especially when we are talking about a batten with dovetailed end protruding.
The elegant solution, if one wants to employ dovetailed battens on a top which is going to move, is to make the battens shorter in length than the width of the top, then (optionally) glue/pin the connection at one end (with the batten end portion held back from being flush at the board edge by, say, 6mm or so), and then on the other end of the connection glue in a dovetail plug matching the top’s material, and leave a gap underneath between the plug and the end of the batten to allow the top to move. I’ve done that before on a splayed leg stool, many years ago, and it seems to work well. With this small serving tray, I felt like trying a different approach. No harm no foul?