French Connection 11

A received a question recently concerning the applicability of this three legged bench – did people make these? Well, yes – here’s at least one, a scale model:

This three-legged bench is really a convenient vehicle for the study of placing an inclined saltaire cross between a cross-beam and purlins, along the lines of this study:

In the last post, it appeared I had completed the drawing work on side 2, the second set of saltaire braces, a pair which have their front and rear sides plumb. Well, not so fast there young fella! While working to develop the brace’s joinery details, certain things came to light in regards to the joint where the two pieces cross one another, and that caused a back step in the process.

The first set of braces, on side 1, I had made as square sections, 1.125″ on each side. As one might expect, those braces, as they have faces flush to one another, fit together seamlessly when they are the same size each. When it came to side 2, I followed along with that pattern, making the braces 1.125″ on each dimension. It makes sense to use the same section of material for all the braces, or so I thought. What happens though, when you use the same size section for each brace on side 2, is that the crossing point between them ends up with some goofy little joint interfaces. This is something I hadn’t noticed on the previous drawing go-round (three years back) on this piece.

I’ll show you what I mean, using my prior drawing for the trépied établi:

Now, zooming in on the front of the crossing, and looking at the lower intersection, you can see that they don’t quite meet spot-on, which leaves a slight little barbe (“beard”) indicated by the arrow:

Rotating around the connection, a similar situation occurs at the upper rear intersection, though here the barbe is very slight:

In order to eliminate those barbes, I had to reconsider the dimensions of the sticks. After some thought, I decided to make the plumb measure of the braces equal to the section height of the tripod top beam pieces, at 1.5″:

With this change, which was easily accomplished in a 2D to 3D development, the front face’s barbe disappears:

Not all was well however at the back – there was still a tiny barbe:

This barbe is a compound joinery effect, as the two sticks each have a different slope. The only way to eliminate this tiny barbe, besides ignoring it of course, leaving a less-than-crisp or fudged fit, would be to change one or both sticks in their section thickness. I decided to change one stick, which made it 1.128″ instead of 1.125″:

None of these issues revolving around the fit of these parts as the intersection was apparent from the text, and it was difficult to know exactly what was intended, as the drawing of the tripod bench in perspective does not show the meeting points for this side. So, all i can do is discover the issues and try and find the best solutions I can. I’m grateful for the 3D as otherwise I don’t think I would have been able to discover these issues so readily from the 2D plans. Indeed, I imagine that the second tiny barbe, as shown a couple of pics above, would have appeared in 2D layout as a tolerance error (it’s about pencil line thickness) and I would have just connected the lines – fudged it – and the result would have been a gap in the fit or a not-quite fit.

Anyway, these barbe‘d comments, ahem!, are but a warm-up for bigger fish to fry in this project. By far a bigger concern to me is how I am going to put the bits together with joinery.

Here’s a look at the older sketch of this project, where you can see the prints (outlines) of a post and two braces (one from side 1 and the other from side 2) as they converge at the top surface of the beam:

Clearly, there is too much going on in that space. While a tenon from brace set 1 (to the right) emerges into clear territory on the top surface of the beam, such is not the case for the brace coming from the left side, from pair 2, which largely occludes with the traced outline for the post.

Here’s a view of the same thing from my current drawing, in which I’ve drawn the post’s tenon and the beam mortise already, and you can see where the brace traces emerge on the beam’s upper surface:

I can’t see any way I can put a through tenon in there coming from the brace – and the text indicates no tenon there, however whether the text shows a joint or not, I have a hard time taking the text seriously. After what happened with the 4-legged Mazerolle tréteau project from a couple of years back, where several mortise and tenon connections were indicated which later proved to be impossible to assemble, call me a skeptic. All tenon depictions in these books I take with a large grain of salt. In this instance, I agree though that no conventional through tenon seem possible in that location.

Take a look at this picture from Delataille’s work, which shows a through mortise and tenon connection for the brace:

I’m doubtful that will go together as illustrated, unless some of that new rubberized wood is involved, and that is even presuming that the crossing point of the two braces is formed not from a lap but from three pieces, mortise and tenoned.

Examples of impossible connections abound in Delataille – here are a couple more:

From the maquettes I have seen, these sort of multi-legged saltaire-braced tables invariably have nailed connections. The French aren’t that preoccupied with joinery. In some models you can often spot the odd place where a nail has rusted off and some little portion of a brace has fallen off at some past point in time.


I don’t see any pegs or other joinery telltales.

Here you can see the nailing fairly clearly:

Can’t forget this one, one of my favorite little pieces of insanity:

Every time I see it I laugh and wonder how many hours that took to make!

I guess I would rather avoid using nailed or glued connections on this piece, but am seriously doubting how I might realize that goal. It might be unachievable.

Thanks for the visit, and we’ll see you next time.

10 thoughts on “French Connection 11

  1. Julie,

    thanks very much for the encouragement. While drawing this piece I occasionally feel like scrapping the whole thing, not because of the drawing difficulty, but because I get pessimistic about trying to make such a piece join together cleanly. I'm still not convinced I'll be able to solve it without recourse to metal fasteners, and if that is the case, I won't build it. Just because it can be built doesn't mean it should be built. We'll see what happens – for now I'm walking the path to see where I get to, and will decide from there.


  2. I am very delighted to see you are tackling this most challenging piece. I have looked at the original drawings myself for quite some hours in the past and have always wondered about how any of those tenons could meet properly. Therefore I find it most interesting to see how you are dealing with it, well struggling might be better way of putting it. Not that I have ever considered building such a thing myself at this point.
    I am sure your creativity will supply you with a satisfying solution. I am pretty sure you have thought about it but have you looked into using floating tenons? These would even complicate matters but then they don't have to be angled in the same planes as the sticks.
    If the compagnons used nails to hold it all together the nails can always be substituted with a wooden peg of some sort. Of course I am just guessing here and since my limited skills in 3d drawing I can't make any hard statements. Looking forward to see your progress.

  3. Mathieu,

    good to hear from you. I have thought about using floating tenons and various forms of Japanese joinery, and have also considered driving wooden nails in. The wooden nails would be a last resort, and I'm certain that some floating tenons will be involved here and there.


  4. Kurisu

    This some crazy action – I think you are bending the time, wood, space continuum here and should be careful about entering this black hole.


  5. Hello Chris

    I understand that French people are using nails for fixing the joints. Nice mix:)Russians in some area of Sibiria did not have any joints in their window frames about a century ago. Sticks were simply nailed around glass. Sometimes things are very strange.
    But lets take japanese joints. I ask seriously why they have not used glue there ? Tradition ? Lack of knowledge about glues ? Or in some cases today they are using in japanese traditional joints glue too ?
    Best regards

  6. Priit,

    thanks for your comment and questions. In Japan today much of the traditional furniture is put together with white glue, or 'bondo' as they call it. You don't see glue in traditional timber buildings on any continent. Not that adhesives are not wonderful in certain ways, but the difficulty of clamping such a large assembly, complications which can arise during a glue up with large structural parts, and the reduction in structural flexibility at glued joints are some good reasons glue is not used. Only modern synthetic glues could handle the sorts of loads imposed by the elements on a timber buildings, so glue is not a part of traditional construction. Also, if you want to dismantle a structure for repair or re-use, pegged and wedged joints can be taken apart whereas glued connections, at least with epoxies and so forth, must be sawn out, considerably reducing re-purposing.

    I really like the purity and simplicity of an all wood joinery mechanism.

    Traditional Japanese furniture making used to employ rice-paste- and animal hide/hoof-based glues, however most of it now is just done with white glue, which is not nearly so easy to disassemble later on. I think the advent of those glues had led to a certain decline in the constructional quality, as increasingly glue is relied upon to hold things together instead of a mechanically sound and tight joint, in which the glue merely acts as a reinforcement, and has a weak bond which can be broken with a mallet strike, or application of steam.

    Now, in my case, I try to avoid glue out of both my background in timber construction, and a desire to emulate classical Chinese furniture making practice, which avoided glue as far as possible.

    I don't hate glue, I'm just trying to see how far along I can go without it being part of the equation, and so far, I've managed pretty well without it. Going glue free sets up a design and construction challenge that leads me to strive to make better joinery, to think long and hard about how things can get put together, and even if one day I return to using glue, I think I will have benefited from this journey.

    To me a great virtue in classical Chinese furniture, and in timber carpentry generally, is demountability, which allows for fairly easy repair and modification.


Anything to add?