by Steven J. Owens (unless otherwise attributed)
I've been researching various knock-down furniture connector approaches for a couple years now. Most woodworking joints, mechanical connector or otherwise, are intended for permanent joints, so this is an odd topic. I'm by no means an expert, which is why I've been taking notes. Here they are.
So far, my favorite is the very historically period keyed mortise & tenon, aka "tusk tenon", but they're also the most work, so I'll explain that at the end, along with brief descriptions of various tools and approaches for the hardest part of the job -- making through mortise, aka holes through the center of wood.
There's also the ever-popular edge half lap, aka interlocking notches, which I'll also discuss below.
However, there are a number of more modern options:
I'm going to try to link to examples of each connector, but since links can go dead, I'll also try to give a short description. Someday I'll take photos of all of these and add them to this article.
I'm not sure if there's a more technical term for them, but bed rail fasteners are a popular option and in fact it looks like a fair number of people in the SCA use them for thrones, and even beds.
Bed rail fasteners are available on sites like rockler or woodcrafters. They're bits of metal that hook together. You embed one in the end of a rail (via a shallow mortise) and another into the side of a post:
There's a version that doesn't require embedding/mortising, but those are significantly less robust than the mortised version.
I experimented a bit with bed rail fasteners but I found the ones that I tried weren't robust enough for uneven ground and lateral force. Maybe other people have had better success with different bed rail fasteners, or simply avoid using them on uneven ground.
I guess if you made one out of really thick, heavy steel, very large, it might do the job, but good luck finding that for sale.
I haven't looked into these in a few years, so it's possible there are better options on the market now. For example, I just googled on it and came across "Kasunen bed rail brackets". These appear to be non-mortise brackets but more substantial than the ones I tried.
A carriage bolt, also called a coach bolt or round head square neck bolt, is basically a long bolt with nut and washer, but the head of the bolt has a weird shape. It has a rounded dome over a four-sided square neck, then the main body of the bolt.
Note: I also recently learned of "elevator bolts", which seem very similar to carriage bolts, except that instead of a domed head, they have a flat head, that ends up being more flush with the surface.
Carriage bolts are designed for the four-sided square neck to sink into the wood. The wood is what holds the head of the carriage bolt still and keeps it from spinning as you tighten the nut.
This makes the carriage bolt smooth and relatively attractive, just a shallow domed button on the surface of the wood. But a few moments thought will also point out that this makes carriage bolts not so great for repeated assembly and dismantling.
The best bet is if you just leave the carriage bolt embedded in the wood and only take off the nut and washer. Otherwise you end up risking chewing up the wood the neck is embedded in, or damaging the threads of the bolt so it becomes a pain to tighten the nut.
Another thing to watch out for in outdoor applications is that the bolt or the nut can rust and then be very hard to assemble/disassemble. You have three options, galvanized, zinc-coated or stainless steel. Galvanized is cheapest but the protection is a chemical coating that can easily get rubbed off. The other pros and cons you can google on yourself.
Another gotcha to watch out for is that some bolts and carriage bolts aren't threaded all the way up to the head. This is fine if your wood is thick enough that your nut never gets to the non-threaded part, but it's something to watch for.
If you still want to go the bolt-together route, steel hanger bolts are interesting. These look like somebody took a thick wood screw and a thick carriage bolt, cut off the heads of both, and welded the cut-off ends together.
Instead of drilling all the way through and shoving a carriage bolt through from one side and adding a washer and nut on the other, the hanger bolt is one half bolt, one half wood screw. You drill a pilot hole and then screw the wood screw half of the hanger bolt into the board.
To actually do the screwing, you use a pair of nuts or wingnuts. You put one on the end of the bolt, tighten it down a half inch or so. Then put on the second nut and tighten that down against the first nut, until the pressure of the two nuts against each other and against the threads keeps them from rotating. Then use those nuts to drive the wood screw half into the board.
If you're using wingnuts, you put the first wingnut on upside down, so the wings don't interfere with tightening the second wingnut against it.
Tee nuts and threaded inserts and barrel nuts are similar takes on the same idea, burying the nut in the wood to hide it. This link has examples of all three:
Tee nuts and threaded inserts require you to drill all the way through the board and then embed the fastener in the board from one side, and screw in the bolt from the other side. Barrel nuts are weirder, you drill two holes at a 90 degree angle, one to embed the barrel nut, the other for the bolt.
A tee nut is similar to a bushing, in that it's a little cylinder that's designed to be hammered into a hole in a board. It serves as an embedded nut.
The tee nut is threaded on the inside, like a nut. One end of the cylinder has a lip with little prongs.
You pre-drill a hole all the way through the board. You insert the tee nut into the pre-drilled hole, then gently hammer it so the little prongs bite into the wood. The prongs keep the tee nut stuck to the wood and keep it from rotating, and you can then screw a bolt into the hole from the other side.
A threaded insert is more of an embedded nut. It's a cylinder with threading on the inside, and reverse threading on the outside.
You drill a hole for the bolt all the way through the board, then on the side that the nut would normally go on, you drill a slightly larger hole, as if for a countersink, and insert the threaded insert into the slightly larger hole.
Usually the outer end of the insert is designed to take a hex wrench, which you use to drive the insert into the wood. The external threading is the reverse of the internal threading, so inserting the bolt from the other side and tightening it doesn't unscrew the threaded insert.
A cross dowel barrel nut is a smooth cylinder with a threaded hole drilled in to its side.
The barrel nut is similar to a threaded insert, but the nut is installed by drilling a hole perpendicular to the bolt hole, similar to a cam lock (see below).
The end of the barrel nut has a slot for a flathead screwdriver, so you can turn the barrel to align it with the bolt hole. The smooth barrel nut slips into the hole. There's no threading or prongs to hold it in the wood, because it's not designed to resist force along its long axis. Once the bolt is tightened in from the side, it will hold the barrel nut in place.
I've seen a lot of cross dowel barrel nuts in furniture I've bought, but generally these seem intended to be assembled and then left assembled.
Also, I've seen comments here and there that seemed to imply that barrel nuts are considered mainly good for perpendicular joints, i.e. T-joints or L joints.
The wikipedia entry for sex bolts gives a whole laundry list of alternative names:"sex bolt, binding posts, Chicago screws, interscrews, barrel bolts, barrel nut, partition screws, door closure bolts, furniture screws, panel fasteners, architectural sex bolts, arch series screws, hinge screws, display fasteners, screw nuts, connector bolts, threaded rivets, grommet nuts, post and screw sets, book screws, and stationary screws."
The most useful term I've found to describe them is "sleeve nuts", because the nut part is indeed a sleeve, a relatively long metal tube that goes all the way through the material.
These are basically a nut and bolt where the nut part is a smooth cylinder almost as long as the bolt. The bolt doesn't go through the nut, instead the other side of the nut is designed to take a hex wrench, which you use to keep the nut from spinning while you tighten the bolt.
I'm not sure about sex bolts, I get the feeling they're more meant to be used with metal parts.
Ikea loves cam locks. A cam lock is kind of like a barrel nut, except that instead of a threaded bolt going into a threaded hole in the barrel, the rotation of the barrel part - the cam, in this case - is what's used to pull the two pieces together. The "bolt" isn't threaded, but the end has a slight mushrooming that the edges of the cam latch onto.
Note, if you google on cam lock you'll see a whole bunch of actual keyed lock mechanisms. Try adding "woodworking fasteners" to the search terms.
Ikea loves them, but have you ever tried to disassemble a piece of Ikea furniture to move them? I'm not sure I'd trust them to do the job.
Also, like barrel nuts, I've seen hints here and there that cam locks are mainly meant for T-joints.
I'm not sure if there are more generic, technical terms for these.
A sash lock is the ubiquitous latch you find on casement windows.
A toggle latch is the ubiquitous latch you find on toolboxes and footlockers.
Technically sash locks and toggle latches aren't meant for furniture purposes, but the principle of a sash lock is very similar to a cam lock, and I've thought they might be very useful for a variety of knockdown furniture purposes, especially in situations where I want to surface mount them, but keep them fairly flattish.
Like a cam lock, most sash locks work via the rotation of a flat cylinder sort of piece. Unlike a cam lock, the sash lock cylinder is partial and as it rotates it extends out of the fitting it's in and through an immobile fitting on the other side.
Toggle latches seem useful for a similar purpose, like keeping ther lid of a box on.
One thing I have not found and haven't figured out the right name or search term for is a toggle latch designed to go around corners. Imagine you have a flat lid, without an inch or two of lip to fasten the upper half of the toggle latch to. It seems like this should exist, and I can think of a few designs where that'd be really handy.
I did find this article about this idea, so apparently it's not that far-fetched:
While I'm at it, for once a pinterest page in the search results is actually useful. Some interesting ideas here:
These come up often, so I'll address them briefly here.
A joint where two boards overlap and are both half cut away is called a half lap. Most often you see this with the flat of the board, the widest part, doing the overlapping. When it's an edge overlap it's called an edge lap half lap, or sometimes shortened to an edge half lap. But note that woodworking is often vague and confusing about terminology, so don't be surprised if people don't call it that.
I'm pretty sure that edge half laps are popular because they're a lot easier to cut than mortises, but they also tend to break easily, and they don't do so well on unlevel ground.
The simplest way to make them less likely to snap off is to simply make the ear, or the bit on the end that's past the notch, much longer.
You can also reinforce them. This can be as simple as proactively driving a few good, long deck screws through, or you can get period and drill a hole and pound some long dowels in, etc. If you go with deck screws, you can countersink them and hide them with a wooden plug made from a bit of dowel.
Or you can go the opposite direction and reinforce with steel. The big box stores sell little plates of steel with a few dozen holes drilled in them. Back when I used them for a project, I was told they were called "mesh plates" (which certainly describes them well) but when I googled on that just now, it didn't turn up what I used.
The classic example of which is the Oseberg viking ship bed, or the Gokstad bed. Most people are far more familiar with this from various trestle tables. I built a bench/couch using this concept:
I have to give credit here to bloodandsawdust.com, where I first learned about tusk tenons.
He has a great little PDF tutorial:
I also have to give credit to two good friends, Angus and Mark, both talented and more knowledgable about woodworking than me. They gave me a lot of good advice and actual help on the projects I've done so far.
For my own purposes, I've used a double set of horizontal key mortises, instead of a single vertical mortise. As the PDF explains, if you're using horizotal key mortises, you need two to prevent the key mortise from becoming an axle that the wood rotates around.
Note: There's also "wedged mortise & tenon" which is a different thing. Wedged tenons are more for permanent use, but people seem to refer to the "tusks" as wedges. I can kinda see why. In fact,when I googled on "wedged mortise & tenon" just now to to find the example link below, the fourth result was in fact a keyed/tusk tenon. So I'm mentioning it here so you don't see a reference to wedged morties & tenon and get confused.
In a wedged mortise & tenon you cut slits into the end of the tenon, insert the tenon into the main mortise, and then pound thin wedges into the slits to expand the end of the tenon:
Moving on, important to note is that in a keyed mortise & tenon, the secondary mortise, called a "key mortise", is slanted on one side, and uses a wedge or "tusk" instead of a dowel pin. That enables you to get a very tight joint by pounding in the wedge.
If you're planning to use this for knockdown furniture, I recommend making the tusks out of a good hardwood, so the thin end doesn't split when you're knocking them loose. Also remember that you want the thin end of the tusk thick enough to pound! Despite that, I find it handy to keep spare tusks for my knockdown furniture.
Also bear in mind that wood will swell over time, so if you're in a wet or humid environment, I recommend bringing along a wooden or polyurethane mallet to knock the tusks out. If you don't have one, then use a scrap of 2x4 and use it sort of like a chisel, i.e. put one end of the scrap against the thin end of the tusk, then hit the scrap with your hammer. In fact, you may find that to be a lot easier in hit accurately than the tusk, even if you do have a wooden or polyurethane mallet.
Cutting the huge mortises for this kind of furniture can be a chore, however. There are a few approaches, which I'll summarize here:
I've only done a few of these, but here are my notes from researching them.
The two most frequent recommendations I heard from woodworkers were using a router, or using the drill-and-chisel method. I've tried the drill-and-chisel and it worked, but was tedious, especially if you're making a piece of furniture with 8 or more huge mortises. I'm planning on trying the router.
Note: Even if you don't want to do the drill-and-chisel method, check the note about drilling out key mortises, at the end of that section.
In general, for almost all approaches, there's an issue with "tear out", meaning that when the tool breaks through from the opposite side, the wood at the surface of the work piece doesn't cut cleanly.
One approach to fixing this is plunge-cutting from both sides. Literally cut to slightly over half-depth from one side, then flip the board over and cut from the other side. However, this means you need to get the cuts lined up exactly right. More on that below.
A more common fix, because it's easier, especially with power tools, is to clamp a thin sacrificial board to the underside. This board holds the surface wood of the work board in place, and the tear out occurs in the sacrificial board instead.
This is pretty much what it sounds like. Use a drill, ideally a drill press, to drill out the bulk of the mortise. The result is a round hole, of course, or several adjacent round holes. Use a typical wood chisel (with beveled sides) to chisel that hole square.
This can work, but for really large, really deep mortises, the chiselling is a fair bit of work. I've only done a few of these:
The approach worked, but they weren't easy and they weren't super clean. Part of it may have been that I didn't know how to properly sharpen my wood chisel. Part of it may have been that I was using cheap pine with fairly uneven grain. Also, I've noticed that in a lot of the videos I've seen online, they tend to orient the mortise so the long side is parallel to the grain.
Note: When I did the twelve 4" x 1.5" mortises, by the time I got finished chiseling out all of those and had to do the the key mortises, I was tired and decided to cheat/be sloppy. I was cutting the tusks/wedges from 1/2" thick boards, so I just used the drill press to drill the key mortise as two overlapping holes, and instead of squaring off the key mortise, just used a belt sander to round off the upper side of the tusks. This worked out fine.
Note: I've seen a few examples where people used the drill press almost like a router, plunging all the way and then sliding the board back and forth. I'm not sure how well this works. I've been told that drill presses aren't designed for lateral stress; that's the major difference between a drill press and a router. Also, this requires being able to move the board. This risks going off kilter with the board, or worse having the drill throw the board at you.
Most woodworkers use a router for simple, small (compared to our needs) mortise & tenons for permanent furniture joinery. Here's a good intro from the Wood Whisperer.
The two big gotchas with a router are a) it will have rounded corners the same size as the bit, and b) most routers can't plunge very far and that makes it of limited usefulness for mortising a 3x3 or a 4x4.
The rounded corners of a routed mortise are basically going to be the size of the router bit. Standard woodworker router mortising is essentially done with a bit the same width as the mortise. You want a mortise 1/2" wide, you use a 1/2" bit, and the ends will be a 1/2" diameter semicircle.
For the big mortises I'm talking about, you'd have to move the router in a circle. I don't have much router experience, but experienced woodworkers seem to dislike that approach. Maybe because it's just slower, or harder to get accurate, and of course they almost never do mortises that big. Here's an example of that from the Wood Whisperer. I've linked it to about halfway through the video but the whole thing is well worth watching:
Note, bits significantly larger than 1/2" seem hard to find (and expensive), and even harder to find in long lengths for the drill-and-plunge trick (see below). Plus they're going to require a lot more physical effort.
Even if you use a 1/4" bit, you're still going to have rounded 1/4" corners. Also note that cutting a large mortise with a 1/4" bit is going to take twice as long as with a 1/2" bit.
An option is to use the bigger bit for the initial work and then clean up with a smaller bit. This means two passes, of course, and time swapping the bits out, or using two routers so you can just put one down and pick up another.
Also note that you can't just plunge all the way in and then route sideways. Generally all sources say you should start shallow and make many passes, increasing the plunge depth maybe 1/4" for each pass. This, of course, takes more time.
You can avoid the limited-plunge-depth problem a couple ways. The "plunge" aspect is actually almost always an add-on case or sleeve that the router bolts into. Mostly they just don't plunge that deep (but see lock mortisers, below).
An option I've considered, and I'm happy to see that the usually pretty competent Samurai Carpenter thinks it's a good idea, is avoid using the plunge mechanism at all. Instead, first drill a hole all the way through with a drill press, then get a very long bit, 2" or 2.5" long, lower it into the pre-drilled hole, then turn the router on:
As I said above, most sources say you should make multiple shallow passes, so I'm not sure what the risks or downsides to this approach are. The Samurai Carpenter is usually pretty clueful... but that may be exactly why somebody like me, with lesser skills, shouldn't try this at home.
Even with drill-and-plunge, you still won't be able to plunge all the way through a 3"x3" or 4"x4" post. You can plunge from both sides, but this requires a lot of accuracy to make the two holes match up cleanly.
Approaches, I'm told, include a) a lot of skill and patience, b) a very good template jig setup, c) a neat trick using two different bits, a top-bearing and bottom-bearing bit.
I don't have a lot of skill, so I'm leaning towards b) or c) :-).
For b), there are a brazillion commercial, super-complicated jig contraptions and homemade template/jig designs.
One useful google search term is "lock mortiser". These gadgets are like jig/clamp plus plunge housing on steriods, just add router. They're neat and maybe a great solution, but just the gadget (not including router, bit, etc) can be $900-$1500. Here's one example, I think that's a Virutex FC16s, around $1500:
Option c), the top/bottom bearing bit approach, is a neat hack.
MM Wood Studio marks both sides and uses a drill press to remove the bulk of the wood (which I think is a good idea), removes a little more bulk with a chisel, then clamps a template to it and routes it clean with a top-bearing bit:
I was going to include an explanation of top-bearing bits versus bottom-bearing bits here, but really you should just go google it and find a good video or two.
Long story short, in the video he makes the first plunge to more than halfway through. Then when he flips it over and does the plunge from the opposite side, he switches to a bottom-bearing bit that then guides along the the already-cleaned-up mortise, so the opposite side matches perfectly:
A pantograph is a framework where you have two pens or whatever, and as you move one, the other moves in correspondence. You can use it for writing or drawing two images simultaneously, or for tracing an image, using a pointer instead of a pen in one of the spots. You can also design it so the traced image is much larger. Some stuff I've seen suggests this was one of the early uses for it.
Pantographs have been popular for use with woodworking routers for decades, but recently this insane - and insanely-skilled - Canadian woodworker, Matthias Wandel, build a really nifty pantrograph router that he calls the Pantorouter:
Wandel sells plans for a modest $10-$20. Needless to say, as awesome as this pantorouter is, you need to be a fairly skilled woodworker to make it.
I wrote him and asked him if there was a set of plans designed to be built with a CNC router. He replied "If you don't have a CNC, the best way to build the pantorouter is with a table saw. If you do have a CNC, the best way to build the pantorouter is with a table saw."
Somebody made (with Wandel's permission) a machined, metal version of the pantorouter and sells it for $1400. It looks awesome, but a little too pricey for me:
Chain mortisers are basically a chainsaw version of a plunge router, or perhaps more accurately "lock mortiser type jig meets chainsaw".
I've heard these have a long history in Japan, and I've read that they're used a lot for timber framing. But I've also read forum comments saying that US timber framers don't like or use them because the mortises they cut are too rough. I don't know who to believe :-).
Here's an example:
And here's another one that goes through the whole timber framing joint process:
In this one the Samurai Carpenter goes over a (small) chain mortiser in more detail:
This is the period approach, and from what I've seen it may be - with sufficient practice, good tight-grained wood, and a properly sharpened chisel - one of the most cost-effective and fastest, cleanest ways to chisel out a mortise.
Most woodworking chisels you'll encounter are not mortising chisels. Regular chisels have beveled side edges. Mortising chisels have flat sides, i.e. a square cross-section, and are generally a lot thicker and made with stronger steel. They're designed for prying out chunks of wood:
Note, one requirement of the traditional approach is that the mortise width must match the mortise chisel width. So if you want to 1.5" mortises, you'll have to find a 1.5" mortising chisel. That's not easy or cheap...
I can't find the video right now, but I'm pretty sure it was one of Paul Sellers' videos on chisel sharpening (he has several) where he mentioned an interesting tidbit; the square edges of the mortising chisel should be sharpened also, because they're designed to cut the wood fibers as you hammer the chisel in. This makes more sense after you watch the technique in the following videos.
I saw a great video of Franz Klaus hand-chopping a mortise in a piece wood up against a plate of glass:
Here's another video, the sound quality isn't as good and there are no/few closeups:
Klaus also has an instructional DVD that's pretty good.
Paul Sellers did a version of this demo with better camera quality, and also does another one with a beveled-edge chisel:
One gotcha with this approach is that in every example I've seen, the long dimension of the mortise is parallel with the grain of the wood. Assuming I get enough practice at both sharpening chisels and mortising, I'm not sure how well it would work cross-grain.
Here's the Pete Follansbee video Paul mentioned:
While we're at it, there's nothing revolutionary in this video, but it has some good closeups and tips on wood chiseling in general:
Here's a video of basic operation:
A hollow chisel mortiser is a lot like a drill press, except that there's a long box of metal around the drill bit, and the edges on the end of the box are sharpened and act as chisels.
Unlike most drill presses, instead of a circular handle, hollow chisel mortisers use a long lever handle, to give you leverage. That's because the chiseling action is pretty much just brute force.
These monsters cost upwards of $12,000 for a used one and require 3-phase power, but they're awesome.
A swing chisel mortiser moves a chisel back and forth laterally, at very high speeds, as it gradually plunges.
These are the machines that they use to route out door lock mortises in a factory that manufactures doors. Here's an example:
I'm still looking for some good information on these two topics. What I suspect from experience is that neither one really matters that much, but I'd like to have more information from a more authoritative or even scientific source.
A tenon is the end of the board with a chunk cut away on both sides. The bit where it widens back out to original dimensions is called the "shoulder". That's the part of the tenoned board that presses against the mortised board (the shoulder presses against the board on either side of the mortise).By the way, the etymology of mortise and tenon are both surprisingly murky, or maybe it's not that surprising, given how old woodworking is.
Tenon comes from the old french tenir or tenere, meaning to hold, which is also the root for "tenet". One source said it might come from the proto-indo-euroepan root "ten-", to "stretch", as in to stretch your mind around a concept.
Mortise comes from french "Old French mortaise (13c.), possibly from Arabic murtazz "fastened," past participle of razza "cut a mortise in." Cf. Spanish mortaja."
A lot of examples have shoulders on all four sides, although obviously the shoulders on the face (the wide side) of the board have to be pretty shallow.
So I've mostly just skipped the shallow shoulder on the face, and just cut the deep shoulders. Nominal 2x6 boards are actually 1 5/8" x 5 5/8", so I end up with the tenon part being 1 5/8" x 3.5", and only two shoulders, each with 1 5/8" x 1 1/8" of area.
I used to be worried about how or whether the mortise and tenon would resist lateral stress, aka "wracking force". I worried that I didn't have shoulders on all four sides. My first mortises, I cut a shallow secondary mortise that the shoulders socketed into a little.
After some experience, observation and thought, I strongly suspect all of that is unnecessary. The two 1 5/8" x 1 1/8" flat shoulders, pressing against the flat of the 2x6 they're going through, provide more than enough resistant to lateral force.
Similarly, I used to be worried about making mortise and tenon fit tightly.
A lot of the woodworking info out there harps on this. But then again, most of that info is about making permanent, glued mortise & tenon joints. Those joints don't have the key mortise & wedge to make sure it's tight. And boy can it be a pain in the ass getting that tenon through that tight mortise - or out, after two weeks in the outdoors humidity.
So, again, I've begun to strongly suspect that the joint doesn't really need - or rather, doesn't absolutely depend on - the reinforcement of having a tight fit between the tenon and mortise. Sure, all things considered a tight fit is probably better than a loose fit, but is it enough better to justify the difficulty it adds to assembling and disassembling?
I still need to do some systematic experimenting to figure out just how loose you can get away with. Or preferably find somebody else's results from doing all that hard work.
Unfortunately, the obvious phrase "loose fit", or "loose tenon" is used to refer to an approach where you put a mortise in both sides of a joint, and then use an entirely separate, third piece as the tenon that goes into both mortises. This makes researching this topic difficult. Another term for this same thing is "floating tenon", so why the hell they had to use up a whole separate useful phrase is beyond me :-).
This page is, again, aimed at permanent M&T joints, but I'm including the link because of the fun acronym CTSBTF, "Cut to size, beat to fit" :-).