precision is optional?

[estimated reading time 7 minutes]

i was going to call this article “cut roughly but carry a big tenon” but i figured that would get me the wrong kind of search optimization… in a break with my usual perfectionism, today i’m going to take a little time to talk about why you shouldn’t worry about being imprecise in your joinery — sometimes. i’ve probably said and written a thousand times this month alone how fundamental precise joinery is to furniture-making. and it is. but there are many things you’re never going to see that make absolutely no difference. so sometimes you’re chasing a disappearing mouse in the dark down a hole that nobody’s ever going to see. and you (and i) need to stop doing that.

i’ll give you an example. a standard mortise-and-tenon joint is brilliant. it’s not my favorite thing to cut but it’s probably the joint i use more than any other in furniture construction. it’s simple, fast and the strongest thing you can do without reinforcing the whole thing with liquid steel or something. do a good mortise-and-tenon and your piece will stand up forever. unless you forgot wood movement. in that case, it’ll crack to pieces and be dead next july. but assuming you’re not oblivious to the joys of moisture (i know you’re thinking it but there’s no need to go there) the m/t will give you heirloom durability. some of the question becomes whether you care how long the piece will last but that’s irrelevant. it’s a solid joint. very traditional in both western and eastern traditions. and it’s … a secret.

hold on. it’s the most popular joint for a thousand years and it’s a secret? well, yes. what does a mortise-and-tenon joint look like from the outside if you haven’t screwed it up? a butt-joint. two pieces of wood coming together end-to-face (or some other configuration but usually end-to-face like in carcass construction). how does anyone who’s not a woodworker know there was serious work behind that joint? that it’s stable and won’t crack? what if someone just glued those boards together that way? it would look exactly the same. mortise-and-tenon joinery — woodworking’s unexpected secret weapon against butt-joint idiocy. i bet you’ve never thought of it that way. but that’s what it is.

of course, we can make it far more obvious by turning it into a through-mortise or adding some pins, maybe even drawbore pins, wedging or tusking the thing on the other side. and there are some great structural reasons to do that and even more aesthetic ones. but when you’re looking for strength it’s rarely necessary to do anything more than the basic, standard joint. it’s what holds my bench together. it’s what holds most of what i build together. i bet the majority of the pieces you’re proud of building (after your teen-angst-pockethole-and-epoxy-rivers phase, that is — not that i had one of those because neither was trending that long ago and my teen building phase actually looked far more like badly-executed timberframing and curved joinery that worked but often took literal days to cut because i loved japanese traditional joints but had to have my own style…) are the same — mortises, tenons and enough glue to flood the yangtze. twice.

why am i talking about this venerable joint in the same breath as inaccuracy, though? there’s nothing more obvious than a badly-fitting mortise-and-tenon. except maybe a badly-fitting through-mortise-and-tenon. and you’re absolutely correct. but let’s look at what has to fit and what doesn’t.

it’s a butt-joint with cheeks and shoulders when you come down to it. two pieces of wood joined together with one of them managing to penetrate the surface of the other at least a little — you can see why people call this the manliest joint, can’t you? there are many ways to cut this joint but here’s the typical one. let’s say you’re joining a 50mm x 100mm board (2x4s for those still counting with finger-sized measurements) to another, end-to-face with the second rotated vertically to be a leg. you take the end of the first board and cut a rabbet 12mm wide, 36mm deep all around the end of the endgrain. now you pick up your favorite chisel and dig a mortise in the other to receive it that’s 36mm deep, 76mm tall and 26mm wide. after a testfit, you slap some glue on the parts and bash them together, stick it in clamps and call it done.

but wait. there’s always cleanup. seriously. if you get a mortise-and-tenon joint to fit right off the saw, you’ve done something unusual. it’s not necessarily a lot but they rarely go together the first time if you’re cutting them by hand. and this is where the obsessive precision takes over. you get out the router-plane and start making sure the mortise is exactly 36mm deep. you start filing the edges. but which edges matter and which can you leave? there’s a three-way division here. there are some cuts that have to be absolutely perfect or you’ll see it in the finished product. there are a few others that have to be pretty close or pva glue won’t hold them properly. and there are some that i intentionally cut with a gap because they don’t matter and there’s no need to even try to get close.

the first set is the shoulders. the line around all four sides of the piece where the tenon begins are vital. they have to be square and straight or you’ll see a gap when you put the boards together. this isn’t negotiable. get out your file and paring chisel and make sure that’s clean. this is where being anal will reward you (a sentence i didn’t think i’d ever write for publication, i assure you).

the next set is the faces and edges of the tenon and the walls of the mortise. these don’t have to be nearly as precise. a little ridge? no problem. a bit of roughness? totally ok. as long as most of those surfaces are making serious contact, that’s plenty of glue surface and you can drive the thing in with a mallet. and i mean that. you want it to be tight enough you have to use the mallet to get it together — or perhaps clamps to push it in place. the wood will compress. you don’t have to file and pare these walls on either side as long as you can force the thing together.

now here’s where things get really controversial. there’s no need to put glue on the bottom of the mortise or the end of the tenon. it doesn’t matter. those pieces aren’t going to give you any significant strength advantage — there’s a reason for this — in fact, there are several reasons. but it doesn’t matter why. people dose this with glue and it’s unnecessary. so does it matter if they match? not in the slightest. here’s my solution. i cut my mortise at 38mm and cut my tenon at 36mm (approximately). why 36? because it’s easy for calculation if i always use the same tenon length. shoulder-to-shoulder measurements matter. tenon-end-to-tenon-end is irrelevant as long as the thing seats in the joint without pushing the sides or legs out.

so you can devote your time and precision to the part someone will see. spend a little time getting the four gluing faces to match close enough to stick together. cut the tenon shorter than the hole it’s going in and leave the floor of the mortise as rough as a glacier-carved riverbed and call it a day.

there are a few other places you can be imprecise with your joinery, too, where you might not have thought of it. this, though, is probably the most obvious.

when you cut through mortise-and-tenons, again, don’t try to make the tenon the right length to be flush with the outside of the case — not even approximately. just cut that thing 5mm longer than it needs to be and sand it flush or to whatever you want the protruding piece to be.

a quick tip on how to do that, by the way. cut a couple of thin spacer blocks to the protrusion thickness (1mm? 2mm?) and stick them between the shoulders and the wall of the case when you clamp the pieces together. sand the tenon flush with the outside, remove the spacers and you have exactly the protrusion you want — the easiest “proud tenons” you’ve ever made. i’m sure i saw my grandfather do this so it’s not a new technique and it’s certainly not mine but i use it whenever i’m building something with sticky-outy details and i’ve noticed people use some seriously-complex measuring approaches to this that i’ve never thought were anything other than massive wastes of time if only they’d had a grandfather to show them what mine did.

this goes for wedged m/t, too. cut too long, sand flush, remove, cut in the slots for wedges and you’re done. by the way, wedged tenons are something i’ve seen done rather badly in one particular aspect — people make the wedges too thin and what they’re actually doing is splitting the wood rather than adding strength to the joint. they often don’t realize this for years and keep teaching others to do it the same way. let’s say your saw kerf is 1.5mm. it might not be (and it’s probably not exactly that) but imagine it is for this example. you cut two wedges in a tenon (no, don’t cut one in the middle — one near each edge is good but one in the middle is just asking for problems — don’t even try). now you make a few wedges and … where’s the end of the wedge going to seat? well, at the bottom of the wedge cut, isn’t it? no. it’s going to dig into the wood because the sides of the wedge are actually too narrow to bottom-out in the slot if you taper the end of the wedge to nothing. what’s a better approach? the thin edge of the wedge should be the thickness of the kerf so it does indeed wedge at the bottom of the slot. the thick part of the wedge (the depth of the wedge from its end) should be exactly the width of the additional space it has to fill in the joint. let’s say you cut your joint an extra 3mm wide at the outside and you have two wedges to go in. that means each should be 1.5mm wider at the thick part of the wedge than the kerf (and the thin part). that 3mm will fill the gap in the joint and give you a perfect after-the-fact dovetail that’ll lock your piece together unless the glue dies.

what about dovetails? those have to be precise, don’t they? well, no. the walls of the tails and pins have to be but when you’re cutting them, there’s no reason you can’t just make the pieces a few millimeters longer in both directions and sand them flush. makes the whole thing so much simpler when you don’t have to worry about three dimensions of accuracy.

anyway, this isn’t a comprehensive guide to all the places you have to be absolutely precise and where there’s something hidden you can completely ignore. but here’s a good way to think about it. if someone’s going to see it, touch it or move something against it, make sure it’s smooth, sanded and finished. if it’s never going to see anything but glue, make sure it’s vaguely accurate and your glue is selected to fill whatever gap you might have (pva won’t fill anything of note so make sure any inaccuracy is in the direction of compression rather than gap like in a mortise). and if it’s non-structural like the end of a tenon in an enclosed mortise or the bottom of its mortise, intentionally cut or route a little short and save yourself the headache of even contemplating whether it fits.

this may be a totally new way to think about joinery for you. it might be how you’ve approached things all along. either way, i hope i’ve either preached to the converted or sung for the choir. thanks for reading!

share on social media...
thank you for reading. your eyes have done me a great honor today.