When we teach people to cut dovetails for the first time, I like to say that everyone on the planet is born with a certain number of sets of bad dovetails in their hands. And the only way to get rid of those bad dovetails is to make them. Eventually, you run out of bad dovetails, and you’re set for life.
The same is true with writing a book. Every time I start writing, there are at least three or four bad chapters that I have to write in order to get to the good ones.
When I say “bad” I don’t mean they consist of dirty Esperanza limericks or are written from the point of view of the wood (“Oh Reginald, plunge your chisel into my lignin and make my mortise walls tremble”). Usually the chapter is fine in the noun and verb department. It simply doesn’t fit in the book. Square-peg-round-hole stuff.
This week I finished up one of these orphan chapters. As I wrote its last few paragraphs, I realized it was going to be recycled. The tone is wrong. Part of the tone problem is that I’ve come up with a better title for the book, and the chapter’s meta devices (there are three points of view) are just wrong.
So here it is. Note that I threw a few placeholder images into the text to break up the grey scrolling. Normally a chapter like this would have 15-20 images.
— Christopher Schwarz
Read other posts from the “Making Book” series here.
How to Build Any Kind of Chair
We had an intern in our workshop who was an odd duck. He was – by the numbers – smarter than anyone I’d met. But when it came to the real world, he had difficulty dressing himself each morning or making sure there wasn’t a piece of scrambled egg on his mouth.
I’ll call him John.
We always insisted our interns build furniture during their summers with us, and the interns sat in our planning meetings each week and were told: Come up with an idea for something to build. You can build it, write a magazine article about the process and end up with a nice piece of furniture and an article with your name at the top. A good deal.
At his first editorial meeting as an intern, John had an idea.
“Have you ever thought about doing an article on how to build a chair?” he asked everyone assembled around the table.
I stifled an immense snort. I asked him: “What sort of chair?”
“Any kind of chair,” he replied.
There was an awkward silence. What John proposed was like writing a bi-fold pamphlet on building a nuclear weapon, or publishing the IKEA directions for ending world hunger. The topic of chairmaking is so huge and complex that writing a 2,000-word magazine article on how to build “any kind of chair” was the most ridiculous proposal I’d ever heard.
And that’s why, more than 20 years later, I can’t stop thinking about it. So, with all apologies to John, I present a magazine article submission titled: “How to Build Any Kind of Chair.”
How to Build Any Kind of Chair
A chair is a platform for your butt that also offers support for your back. It may or may not offer support for your arms. That’s all there is – that’s the end of the definition of a chair.
Though the idea of a chair is simple, the thought of building one intimidates many woodworkers.
Once you understand how chairs work – there’s zero voodoo in it – I think you’ll find making any sort of chair to be straightforward. First you need to understand the numbers that make a good chair. Then the materials that make it durable. And finally, the joinery (if the chair has any – it doesn’t have to).
(Ed: No joinery? Are you not going to explain?)
Before we start talking about this stuff, here’s a trick that will help you remember these lessons. Find the most comfortable wooden chair in your house and fetch a tape measure, sliding bevel and protractor. As we talk about the numbers, the wood and the joinery, measure and examine the chair so you can compare it to your “ideal.” You will learn a lot.
Chair by the Numbers
The seat should be between 15” to 18” from the floor. While 18” is the modern standard, that’s too high for short people. It pinches off the blood supply to the thighs. A height between 16-1/2” and 17” suits a wider range of people without making the chair uncomfortable for the gangly bunch.
(Ed: Need to attribute to “Human Dimension and Interior Space” by Julius Panero and Martin Zelnik)
A chair that’s too low can be difficult to pull yourself out of (which is not always a bad thing). And a low seat can encourage the sitter to kick out her heels. That posture might not feel as stable as when you have your feet flat on the floor. But I think that a too-low seat is better than a too-high seat. So, I always err on side of making the chair a little lower than a little taller.
The size – width and depth – of the seat is important. The seat should be about 16” deep, from front to back. Deeper seats stop the blood flow behind your knees. Shallow seats are not as big a problem as deep seats. I have made chairs with seats that are 13”-14” deep that sit just fine. Shoot for 16”. A little less than 16” is fine; a little more is not.
The width of the seat should be between 20” and 24”, unless the sitter is a pencil or an entire package of magic markers. If you make an armchair too wide, then the armrests will be too far apart to be comfortable. If you make an armchair too skinny, then the sitter might get wedged into the chair. This happens.
The seat should not be parallel to the floor, when measured from front to back. Instead, it should pitch backward a bit to encourage the body to slide backward against the backrest of the chair. If the chair is intended for typing or eating – where an upright posture is preferred – pitch the seat so it falls 3/4” from front to back.
(Ed: Are you sure that seats should never be parallel to the floor? What about stools? Swing-out seats? Might want to qualify your statement.)
If you want the chair for reading or conversation, double the pitch so it’s 1-1/2” from front to back.
If you make the seat parallel to the floor, the sitter might feel like he is being pitched forward. On the other hand, if you make the seat so it slopes 2” (or more) toward the back, you can force the sitter’s back into some unnatural positions, depending on the shape of the chair’s backrest.
I do not expect you to believe that such small changes make a huge difference. So do an experiment. Take a wooden chair in your house and prop up the front feet on 3/4″-thick blocks. Sit in the chair for about 15 minutes. Move the blocks to the back feet and sit for another 15 minutes – if you can stand it.
Oh, while we are putting blocks of wood under the legs, what about the angles of the legs? Aren’t they important? Yes and no. They can be straight up and down and still work, depending on the chair. Think about it this way: The footprint of the chair is what’s important. Think about the sitter and where her feet are and where her back is. If the front legs jut out forward of the seat, her feet might get tangled up in them. Or she might snag a leg as she walks by the chair.
For the back legs, you have to look at the backrest (more on those numbers shortly). If the chair’s backrest leans out a lot past the back of the seat, but the legs do not, then the chair will be tippy. Put another way, the feet of the back legs should be roughly under the shoulder blades of the sitter.
So, the angles of the legs are important. But mostly they control how the chair looks.
(Ed: But can’t you give a range as to what’s acceptable? What’s too much? 20° 25°? And don’t you need to say these angles don’t apply to ladderback chairs?)
The seat of a chair – what we call the saddle – can be scooped out so it’s curvy like the human buttocks or it can be as flat as a board. A little curvature improves comfort. A lot of curvature turns the chair into a Jell-O mold, which is uncomfortable. I aim to scoop out 3/8” to 1/2″ of the seat. But I don’t have a lot of cushion on my pelvis. You might want more or less scooping.
In the end, I err on a shallow saddle as it’s called. You can always add a cushion or sheepskin to a seat to make it more comfortable. A deep saddle cannot be fixed with fluffy accessories.
(Ed: Again, shouldn’t we mention ladderbacks here? The seat is neither flat nor “scooped.” It has some give to it because of the material. Might want to add something about this.)
Now let’s move to the backrest. Every chair has one, otherwise it’s just a stool. The backrest can be as simple as a flat board or a steam-bent slat. Or it can be as complex as a curve of bent and shaped spindles. In general, the backrest is designed to touch the back at one or two places: the lumbar (or lower back) and the shoulders. There is no rule that says which arrangement is best. As a chairmaker, I say that it’s easier to make a chair comfortable when it supports the shoulders. Chairs that support the lumbar only are trickier to make.
The backrest is rarely at 90° to the seat – thank goodness. The backrest should angle back. I think 10°-15° is a good place to start for chairs made for eating or typing. For chairs designed for lounging, try 20° to 25°.
(Ed: Are you saying that ladderback chairs with 90° backs are uncomfortable? You are going to catch flack for that.)
For chairs that support the lumbar, the sitter’s back should encounter the backrest about 9” or 10” above the seat. For chairs that support the shoulders, the sitter should encounter the backrest between 16” and 22” above the seat. How they encounter the backrest is up to you. It can be an array of spindles, a curved piece of wood or even a cushion.
The armrests (if your chair has them) should be about 8” to 10” above the seat. Luckily this height is similar to where the lumbar is, so you can make the backrest and arms as one piece that does two jobs.
Wood for Chairs
It’s easy to become a fundamentalist about how the wood in a chair must be perfect and straight and strong. Not every chairmaker works this way. If a batch of wood is weak for some reason, you can overcome this by using thicker stock. If, on the other hand, you have perfectly rived stock, you can make it pencil thin and the parts will flex but not break.
The goal of all chairmakers is to avoid “short grain,” which is where the grain doesn’t follow the shape of the part. Short grain makes the part prone to splitting. You can avoid short grain lots of ways. The bottom line is to pay attention to the wood and adapt your plans to suit it.
Let’s talk about the different parts of a chair and the kind of wood that’s ideal for each.
For the legs, spindles and stretchers, the grain should be as straight as possible. Ideally, the wood fibers should run dead straight through every piece. That’s why many chairmakers prefer rived wood, where the wood is split along the fibers.
Not everyone can get rived wood with ease. The (not always mentioned) alternative is to saw the wood along the grain to get the straightest grain possible. This produces excellent results. For these straight components, many chairmakers prefer oak, ash or hickory. You can use other species, such as walnut, beech or cherry, but you should consider beefing them up in size.
(Ed: Can you say how much you should beef up the spindles when they are walnut etc.? Maybe 1/8” in diameter? Or is there a test? Maybe striking them with a small sledgehammer?)
For a solid plank seat, it’s best to use a species that doesn’t split easily. Pine and poplar are popular choices in the United States. The elms are excellent seat material. But really, anything will do for a plank seat. You can also make a seat by creating a joined frame and wrapping it in a fibrous material (such as bark or cloth). Or upholstery.
(Ed: Really, pine? White pine? Aren’t you worried about a lawsuit here?)
The arms and other curved components can be sawn out of solid wood or steambent to shape. You also can use a modern cold-bend hardwood (which works a lot like steambent wood without requiring heat).
(Ed: Can we give a source for this “cold-bend” stuff? A search turns up “Pure Timber LLC” in Washington.)
If you saw a curved component out of solid stock, you definitely risk it splitting because of short grain. You can get around this problem by laminating three or four pieces together to reinforce the short grain. Yup, just like plywood. This is an old trick.
A lot of would-be chairmakers get worked up about the sizes of all the components. How thick should each component be so the chair is strong enough to hold a person without breaking? Just like with casework, chairmakers use woods in fairly standard thicknesses.
Legs, stretchers and seats are made from 8/4 stock (which is 2” thick in the rough and finishes off at 1-5/8” to 1-7/8”). Arms are typically made from 5/4 (1-1/4” in the rough that can finish out at 1” to 1-1/8”). Sticks or spindles are usually 3/4” at their thickest and taper to 5/8” at the seat and maybe 1/2″ at the top of a tall chair. However, sticks are usually cut from thicker stock (such as 8/4) so you can straighten the grain and get a strong component.
In the end, almost any wood can be used to make a chair. It might have to be a little thicker than what you see in the woodworking magazines. You might have to glue up a few thicknesses of wood to avoid short grain. And the wood might be so ugly that you have to paint it. But know that even construction lumber can make a good chair. I’ve done it. You just have to be smart about it.
There are three (no four) kinds of chair joints:
- A squared-off mortise-and-tenon joint – the same joint used in casework. The only notable difference is that chairs require better-quality fits than casework. Unlike a cabinet, chairs are tortured by children and won’t hang on the wall in peace.
- Cylindrical joints. The mortise is round, as is the tenon. The fit needs to be snug. And there are tricks to make a wet mortise tighten up on a dry tenon. Truth: Dowels are cylindrical joints, though they are often used poorly.
- Conical joints. These joints are typically used where the legs enter a plank seat. They tighten the more you sit on the seat, but they require special tools to ream the hole to the conical shape.
- Mechanical fasteners. Some chairs – particularly commercial ones – are joined by screws, nails or fancier hardware that provides a mechanical lock. To be honest, if you throw enough wood and metal fasteners at any object you can make it sturdy. But it is unlikely to be graceful. In general, metal fasteners play a limited role in traditional chairmaking. Except to repair old chairs, which is where they often make things worse.
No matter what sort of joint you use in chairmaking, the fit between the components is the key to the joint lasting a long time. Tenons should require effort to enter the mortise. Wedges or pegs should be used whenever possible to reinforce the joint. And use glue. Some purists insist you don’t need it. While that might be true, glue is dirt cheap and doesn’t hurt the strength of the joint. So why not paint some on during assembly?
<ED: Rough transition here. Can you add something that eases readers into the final grafs?)
Of course, none of the above information tells you how to make a chair that is attractive. That is the hard part. And unfortunately, there isn’t a magazine article out there with the headline “How to Make Any Chair Look Good” to help you.
(Ed: Maybe this is a possible article for the June issue?)
So how do you make a good-looking chair? The answer might be right in front of you. If you took my advice above and examined the best chair in your house while you read this article, then you took an important first step. To design nice chairs, you need to look at lot of chairs – old, new, ugly and gorgeous. Find out what separates the fantastic from the flatpack. Measure the angles – your eye can be easily deceived. Make sketches of the chairs you like the most.
But most of all, build chairs. The next one will be better.
Christopher Schwarz is a woodworker and writer in Covington, Ky. He is currently at work on his next chair, which really will be beautiful this time.
When we teach people to cut dovetails for the first time, I like to say that everyone on the planet is born with a certain number of sets of bad dovetails in their hands. And the only way to get rid of those bad dovetails is to make them. Eventually, you run out of bad…Read MoreMaking Book, The Stick Chair BookLost Art PressRead More