The Logic of Chair Design

Evan Dunstone, Fallingwater chair. Photos: Lisa McKelvie

Words: Evan Dunstone

Chair design and construction is rightly considered one of the most testing activities in woodwork. Many articles have been written about ergonomics, ‘comfort’ and construction methods. In this article I wish to take a slightly different slant and consider the overall logic of a chair design. I’ve used images of some of the chairs made in my workshop to illustrate a few of the concepts I’m talking about.

Rigid and flexible forms

Timber chair designs range from very rigid to very flexible forms. A chair’s rigidity (or lack of it) depends on the construction methods used (joinery), the materials used (timber type), and the configuration of the frame. The Fallingwater Chair shown above is a good example of a very rigid chair. The original was designed for a client who had some cedar left to her by her grandfather and wanted a set of chairs made in a ‘sort of Arts and Crafts’ style (you will note that the image is of the chair in wenge, not cedar). Unfortunately she only had cedar sections of 100 x 100mm and 125 x 25mm.

Any chairmaker will tell you these are not ideal section sizes. Cedar is not, in my opinion, a great chairmaking timber, as it is soft and relatively weak (it can snap easily in small sections or crush). I know there are many fine examples of cedar chairs, but its nature restricts some design opportunities. Given the client’s brief and the material at hand, I was compelled to make the back legs straight. Straight back legs on a chair are mechanically weak, so the design needed to be rigid and braced.

Why rigid and braced? A flexible chair transfers stress loads along the flexible member to the joinery. The flexing member (no sniggering please) absorbs some of the energy but transfers a considerable percentage of stress through to the joinery. If the timber were jarrah, then it could handle considerable stress at the joint. Given that the timber was cedar, we should avoid placing so much stress on the joinery.

Down load and back load

To stiffen the chair and reduce stress on the main joints, I used lower stretchers. The irony here is that the stretchers are really in compression when they are doing ‘work’. Consider what happens when some big hefty chap flops down on the chair and leans back. There is a down load and a back load (fig.1). The down load doesn’t much matter; it’s the back load that hurts.

Because the back leg is straight, the side-rail to back-leg joint acts as a fulcrum and the bottom for the leg is the pivot point. The stress on the joint is considerable. The stretcher, however, is under compression and relieves the stress significantly. Timber is wonderful in compression, and the stretcher can be really quite thin in cross section (even in cedar). The brace between the stretchers really only helps keep things aligned. Some designers address the inherent weakness of a straight back leg by sloping it deliberately forward at the top. This reduces the movement on the joint, but is difficult to get away with aesthetically.

Four on the floor

With this configuration, you don’t need to worry too much about the front leg joinery while all four legs are on the ground. Yes, the stretcher transfers some force onto the front legs, but the fact that there is weight on the chair actually helps by ‘pinning’ the front legs to the ground. I have never known a chair to fail at the front legs before it has first failed at the back-leg to side-rail joint (assuming that it is soundly made).

If our aforementioned hefty chap rocks back on the chair (fig.2), then the stretchers are still in compression but the back-leg to side-rail joint is positively screaming. There is a good reason the nuns used to smack naughty kids for rocking back on their chairs.

Of course, at this stage we are talking about a well made joint with a sound glue line. If the glue starts to fail, the same forces will slowly pull the tenon out over time. Dowels, of course, will fail in about 10 years if used on this joint, irrespective of glue used. there are three things you can count on—death, taxes and failed dowel joints.

Curved back legs

Now look at what happens if the back leg was curved instead of straight (fig.3). Because the foot of the leg is behind the back-leg to side-rail joint, when a load is applied to the chair, it wants to fall forward. When the sitter leans back, the force is operating in the opposite direction. To some extent, the two forces cancel each other out. In this situation, you can consider not having a stretcher (many of my designs don’t).

Evan Dunstone, Tamar Chair. Photos: Lisa McKelvie

With a curved back leg, you can start playing with a flexible design. The Tamar Chair shown above is my most flexible chair. First let us consider its footprint. The back legs are unusually wide. This is deliberate, because it places the foot of the legs outside the centre line of the down force.

The back legs are ‘falling in’ at the top when under load. If the sitter shifts from side to side, he or she is effectively moving over the top of one leg or the other. The side rail is able to twist (torque) to absorb the forces at play on the leg (fig.4). The light horizontal back supports lock in the relationship of the back legs. Note that they need to be pinned because they are fixed to the leg with shoulderless tenons.

Stiffened side rails

The side rails are stiffened by a horizontal rail that is twin-tenoned and the side rails are fixed to the front rail also with twin tenons. This makes a sort of A-frame when looked down on and is very strong laterally, but able to twist through its plane.

Outrigger legs

The outrigger front legs are the most vulnerable aspect of this design. Outrigger legs are currently very fashionable with designer makers, but tellingly they are not seen on antique chairs. I have used them with very few failures, but I have seen similar outrigger legs on contemporary chairs fail catastrophically. The problem is in the joinery. A ‘normal’ front leg is tenoned from the side rail and the front rail, making a stiff and strong joint. An outrigger leg can only be supported by a single joint from the front rail.

A traditional single tenon is very strong when a down load is placed on it. It also resists a direct side load well. What it doesn’t handle very well is a twisting side load. Guess what? A loaded chair places a twisting force on the front leg which twists the tenon in exactly the wrong way. Not only is the tenon twisted the wrong way, but because there is not a lot of room for the tenon, it is often small.

Then there is the problem of how the side rail intersects with the front rail. If it is too close to the front rail-to-leg in- tersection, the joinery weakens the front rail’s tenon. If it is too far from the front rail-to-leg intersection, the front rail is susceptible to twisting and splitting. I have shown some of the usual solutions to these problems in fig.5.

The solution used on the Tamar Chair is twin shoulderless tenons (shown right) coming off a 30mm thick front rail. The front rail is thick enough to resist twisting, the twin tenons are massively strong and the joint stops the leg from splitting.

Why have I gone to so much trouble to make a flexible chair? It gets back to our hefty chap. He weighs 110kg, drinks too much and tells loud and expansive stories that involve much waving of arms and flamboyant gestures (not that I have many clients like that!). His petite wife weighs 58kg, sits the way the nuns taught her and is long suffering. Fortunately, she doesn’t suffer when sitting down. The Tamar is flexible enough to adjust to the size and (constantly changing) shape of our wine sodden hero, while being perfectly comfortable in its ‘default position’ for our petite princess. The Fallingwater, being a rigid chair, will suit him well enough, but will be less successful for her.

is reduced somewhat. Arms stiffen the back leg of a chair rather dramatically, as they lock the back leg to the front leg in a rather unforgiving way. Fig.6 shows that for the back leg to flex, it has to take the front leg with it and put pressure on the joinery. Traditional green wood lad- der back chairs with arms sometimes have enough whippy flex in their design to give a little, but they are the exception to the rule.
Design to suit
As I always say to my students when I’m teaching chair design, your first consid- eration should be ‘who is the chair for’? If you can imagine your target sitter, then you can design accordingly. There is nothing inherently wrong with a rigid chair or a flexible chair, it just has to be appropriate for the client.

Unfortunately, as soon as you put arms on a chair, the potential for flexibility is reduced somewhat. Arms stiffen the back leg of a chair rather dramatically, as they lock the back leg to the front leg in a rather unforgiving way. Fig.6 shows that for the back leg to flex, it has to take the front leg with it and put pressure on the joinery. Traditional green wood lad- der back chairs with arms sometimes have enough whippy flex in their design to give a little, but they are the exception to the rule.

Design to suit

As I always say to students when I’m teaching chair design, your first consideration should be ‘who is the chair for’? If you can imagine your target sitter, then you can design accordingly. There is nothing inherently wrong with a rigid chair or a flexible chair, it just has to be appropriate for the client.

Evan Dunstone designs and makes furniture in Queanbeyan, NSW. Later in the year Evan will be running a design masterclass. Learn more at dunstonedesign.com.au