Good joinery is key to building a chair that will stand up to the stresses of everyday use and to the test of time. The primary joint in this chair — the joint between the rungs and the legs — is essentially a dowel joint, a round tenon in a round mortise. Dowel joints are notorious for their high failure rate, but, if you have an understanding of the properties of wood and pay attention to those properties during log selection, milling, shaping, mortising, and assembly, it’s possible to make a dowel joint that is strong and long lasting.
There are five aspects of joinery that I work hard to control during the building of this chair:
Straight long grain, particularly in the rung tenons
Grain orientation between mating parts
Moisture content of the mating parts
Precise fit between the mortise and tenon, as well as good surface quality in both parts of the joint
Continuous glue bond over the entire joint
Getting a precise fit can be difficult, but with practice it’s possible to get the size of the mortise and tenon extremely close. The other aspects of making a good joint are not difficult at all — they simply require paying attention at each step of the process.
1. Straight long Grain
Straight long grain within a given part will make the part and the joint as strong as possible so it’s important to pay careful attention to the long grain during the process of building the chair — from choosing a log, milling the log into boards, and milling the boards into part blanks. As many of you know, the straightest possible grain is obtained using a process called riving, or splitting the parts directly from the log. This is particularly important in Windsor chairs where many of the parts are very thin and need to be flexible — in this case riving produces parts with long grain that is perfectly parallel with the edges of the part and with absolutely no grain run-out. These thin riven parts are generally made using woods that split well such as oak, hickory, and ash. In this ladder back chair the parts are heavy enough to be sawn rather than split as long as I pay careful attention to the grain direction during the various milling processes. And since I often build chairs in woods that don’t rive well — cherry, walnut, and maple — it becomes much less wasteful to saw the parts, rather than rive them.
As an example, I always layout parts parallel to the grain, not parallel to an arbitrary sawn edge. This photo shows a walnut board that I plan to use for rungs. The straightest grain generally runs parallel to the bark edge. I begin by laying out a reference edge close to the bark edge and parallel to the grain. Then, working toward the center of the board, I lay out parallel parts from the reference edge. Since all logs are tapers, not cylinders, I usually end up with a wedge shaped piece in the center. Each of these rungs will have very straight grain end-to-end with minimal run-out. There are a few blanks (5th, 6th, and 7th from the right) that have some run-out at the bottom end of the blank. I will probably use these blanks for shorter rungs where I can trim off the part of the blanks that runs out.
See these posts for detailed information on milling for grain direction:
2. Grain orientation and wood movement
Understanding how wood moves along its three primary planes allows me to align the mating parts of a joint so that, to the greatest extent possible, the wood movement in one part matches or compliments the wood movement in the mating part.
Wood is hygroscopic. This means that it absorbs or loses moisture with changes in relative humidity. In a high humidity environment wood will absorb moisture and in a low humidity environment wood will lose moisture. When wood absorbs moisture it swells and when wood loses moisture it shrinks. The amount of swelling and shrinking varies across the three main planes of the wood grain: tangential, radial and longitudinal (long grain).
If you look at a cross section of a piece of wood the tangential plane is roughly parallel to, or tangent to the growth rings. The radial plane is perpendicular to the tangential plane and in a tree runs from the center (pith) of the tree to the outside or “heart to bark.” You can also think of the radial plane as grain that radiates from the center to the outside of the tree. In many woods you cannot see the rays in the radial plane, but they are there. In this piece of white oak it is easy to see the rays. The longitudinal grain or long grain runs up and down the length of the wood. The greatest amount of dimensional change occurs along the tangential plane. As a rule of thumb the dimensional change along the tangential plane is roughly twice that of the radial plane. And along the long grain there is virtually no dimensional change at all.
To demonstrate the relative wood movement in the tangential plane compared to the radial plane you can look at the dimensional change in a cross section view of a round cylinder turned from green or wet wood. After the cylinder dries the cross section changes from a circle to an oval as a result of the wood shrinking more in the tangential plane than it does in the radial plane.
Applying the principles of wood movement when joining two parts is fairly simple, especially with turned parts that can be rotated to the correct grain orientation. The tangential planes in each part align with each other, while the radial plane in one part aligns with the long grain in the mating part. In the illustration below, when joining a single rung to a single leg, the tangential planes (A) align with each other and will shrink and swell in unison and at the same rate (assuming the two parts are the same species). Then the long grain of the leg (B) aligns with the radial plane in the rung (B). Similarly the long grain of the rung (C) aligns with the radial plane in the leg (C). The radial plane has the least amount of wood movement in the end grain of each part so it is the best choice to align with the long grain of the mating part, which has no wood movement.
Not every joint in the chair can be aligned using the ideal grain orientation of the mating parts. Using the rung to leg joint as an example the ideal alignment of the long grain of the rung relative to the end grain of the leg is shown on the left. In the actual chair there are rungs joining with the leg from two sides as shown on the right. In this case I rotate the leg 45° so that each rung, while not ideal, is equally happy (or unhappy, depending on your point of view).
For detailed information on grain orientation and wood movement for every part of the chair see these posts:
3. Moisture Control
Controlling the moisture content of the rung tenon and leg mortise is another simple and effective way to enhance the quality of the rung to leg joint. Since wood shrinks when it loses moisture and swells when it absorbs moisture I can take advantage of that characteristic when shaping, mortising, and assembling parts.
I don’t pay too much attention to the moisture content of the legs other than to get them roughly to equilibrium moisture content — a state where the moisture content is stable and the wood is not absorbing or losing moisture. In my shop this will be in the 8% to 12% range depending on the humidity and time of year.
For the rungs, however, I want to reduce the moisture content of the wood to bone dry or roughly 4% or so. To achieve this I place the rung blanks into the small light bulb kiln shown at left. The temperature inside the kiln is about 125°F to 140°F and usually a week is long enough for the rungs to get to a bone dry state. The parts are too small for me to measure with my moisture meter, but based on everything I’ve read and heard a week is usually long enough. If I am uncertain I can always leave them in the kiln longer. If I wanted to be particularly fussy I could weigh the rungs prior to putting them in the kiln, and then again every day or two. When they stop losing weight they are as dry as they can get.
It is only after the rung blanks are bone dry that I will turn them on the lathe. The entire rung, including the tenon, is turned to final size at this bone dry state and then returned to the kiln until assembly. Just prior to assembly the rungs are removed from the kiln, still bone dry, and assembled. After assembly the tenons will gradually absorb moisture and swell within the mortise, resulting in a very tight joint.
Read more about the light bulb kiln in this post:
4. Precise Fit and surface quality
A precise fit between the rung tenon and the leg mortise is the only part of making a good joint that can be at all difficult. The rung mortise is 5/8″ diameter or 625 thousandths of an inch (.625″). My goal is to turn a tenon that is as close to .625″ as possible. I am generally happy with a range of 5 thousandths of an inch or a tenon diameter between .620″ to .625″.
To achieve such a precise tenon diameter I use several tools to help. From left to right in the photo: a dial caliper that measures in thousandths of an inch; a roughing wrench with an opening that is fine tuned to .650″; a finishing wrench with an opening that is fine tuned to .625″; a 3/8″ wide beading and parting tool; and a small roughing gouge.
Dial calipers are not normally considered woodworking tools, but for this level of precision they are necessary. I use the dial caliper to measure the tenon after turning to confirm that the diameter is within range.
I always turn the tenon after the rung blank has been in the light bulb kiln for at least a week and has dried to a bone dry state.
I have developed a process that makes it fairly easy to get a precise and repeatable tenon diameter as well as excellent surface quality. It begins by rough turning the tenon with the beading and parting tool to a diameter of .650″ using the roughing wrench as a caliper. Next I switch to the gouge and the finishing wrench. I take light passes across the tenon with the gouge and after each pass test the diameter using the finishing wrench as a caliper. When the finishing wrench just barely fits over the tenon it should have a diameter within a range of .620″ to .625″.
Another benefit of this method is that, in addition to a precise diameter, it leaves a very smooth, flat tenon surface which is perfect for a good joint. Here you can see the tenon after each stage of turning. On the left after rough turning to about .650″. And on the right after finish turning to .625″.
Careful attention should also be paid to the surface quality of the mortise. For this reason I use a HSS Lipped Brad Point bit from Lee Valley which leaves a beautiful surface on the mortise walls and drills a very precise diameter mortise. At this writing the 5/8″ bit is about $28 — pricey, but worth every penny. To extend the life of the bit and keep it cutting as perfectly as possible I only use it for chairmaking. Another trick, if you are using a drill press, is to place the bit in the chuck, tighten finger tight, then successively tighten around the chuck so that the bit isn’t pushed off center. This helps to prevent run-out and consequently drills an accurately sized mortise.
For a detailed description of turning a rung, including a 10 minute video, see this post:
5. Continuous Glue Bond
Using modern glues, a precise, tight fitting joint and a good, continuous glue bond are usually mutually exclusive — because of the tight fit most of the glue applied at assembly is sheared off, still leaving a tight fit, but not an ideal glue bond. To remedy this I use hide glue instead of any of the modern glues.
You may have heard that hide glue is reversible — meaning that it will become sticky again and mix with new glue upon exposure to heat and moisture. It is this characteristic of hide glue that allows me to get a continuous glue bond even with very tight joinery.
The method is simple. Prior to assembly I coat the mortise and tenon with a very thin coat of hide glue — this is a called a sizing coat. The sizing coat is allowed to dry. After it has dried, it completely covers the surface of both the mortise and the tenon with a thin coat of glue and has soaked into the wood. Because it has already been absorbed into the wood — think of the glue as growing roots into the wood — it cannot be sheared off no matter how tight the joint. At assembly additional hide glue, which is liquid and heated to about 140°F, is applied to both parts of the joint. This glue reactivates the sizing coat, making it sticky again. Most of the glue applied at assembly get sheared off, but it has done its job by reactivating the sizing coat and providing lubrication for the assembly process. The result is a continuous glue bond.
I will be going into the details of using hide glue in a future post.
The test of time
One of the things I love about building this chair is that inherent in the process is a respect for, and an understanding of, the qualities and characteristics of wood. This way of working wood almost always results in a chair that is well built and beautiful as well. There is nothing particularly difficult in making a joint that will stand the test of time — just careful attention at every stage of building the chair. I have been building chairs using this method for about 15 years and have never had a single joint failure using these methods.
Side Chair Build Series Links:
Previous Post: Turning, Part 2