getting your curve on

[estimated reading time 19 minutes]

today is my mother’s birthday. she was my gateway to the world of wood — and nature in general. she’s an avid gardener and she and my father built the home they’ve now (finally, well into their seventies) retired to, years before i was born. just to give you a sense of how this all began for me, my favorite photograph of my mother is one where she is standing where a year later would be their house but is the middle of a forest, surrounded by stumps, decked-out in bellbottom jeans, kaftan and turban (no, buddhists don’t wear them religiously but it was convenient and she was a flower-child) holding a chainsaw with what looks like a two-meter bar but i’m sure isn’t even close to that — she’s not very large, one might say, so what for me would be a big chainsaw for my mother looks like something built to be used by a giant. anyway, my parents are both academics but they have a respect for nature in general and love for wood and working with wood in particular that inspired me as a child to enter the craft and still motivates me to design and built.

as a result, i have been designing a series of projects to build for them when next i can finally escape the hell of where i have been trapped since the beginning of the pandemic — being immunocompromised has really been a huge problem in my life but never nearly as much as it has the last two years and, having had an anaphylactic reaction to the first dose of the vaccine, i am forever destined to have difficulty traveling because i’ll always be classified as “not fully-vaccinated” — much to my sadness, as i was one of the first people in this country to be given the first dose and i was so happy that day to walk into the clinic and hope things were about to improve.

my parents have simple tastes in furniture (as in most things) and this is probably where i get my extreme minimalism. ok, i’m more extreme about it than they are and i simply refuse to build furniture for myself because … well, i don’t want to own any furniture. but i love building things for them. in the last few weeks, i’ve designed a simple bench and a couple of sitting-stools, as requested, for my mother. i wish i could go to the next step and make them for her today but it’ll have to wait. for now, though, they’ve given me some things to think and talk about. the bench is inspired by japanese traditional timberframing (in style, if not in assembly techniques as it’s actually built with floating dowels — or floating tenons if you’re a domino fan cause those work even better). one of the stools, though, uses kerf-bending or bent-lamination (either will work for the design) to get a pretty-aggressive curve in the top. you could steam-bend it but that is just asking for problems with such tight radii.

when i was a kid, one of the typical phrases used as an alternative to “fuck off” was “get bent” — i never really imagined at the time i’d be taking it quite so literally after i’d had it screamed at me a few hundred times…

there are realistically three ways to bend wood — heat, lamination and partial-segmentation. heat comes in three varieties — boiling, steaming and direct-application. so i guess we have five things to talk about. this could be a little long. then again, they often are.


this technique has always sounded to me like a country of twisted, furry camels — you know, “bent-llama nation”. but i’m weird. and a lecturer in language and culture so what did you really expect? that completely aside, the easiest way to bend wood is to make it so thin it can’t resist then, once it’s bent, glue it together so it can’t go back. there are alternative methods, of course. but when i want to get a piece of wood to take a new shape this is probably the way i will automatically think to use. it requires no specialized equipment, no dangerous temperatures and no waiting time except for the glue to dry. it does, however, require a significant amount of planning and clamps — don’t rush your bent-lam or you’re going to end up with some very sticky firewood-in-training. let’s take a look at the process.

by the way, there are other processes for this. i have tried quite a few of them. i think this one works better. it’s not the simplest or the cheapest. but those are leaving a lot more to chance and i don’t work without a safety-net — i’m not a teen anymore and i don’t freeclimb rock-faces, either. life is a bit safer this way. my time in the shop is, too.

thin slices of wood will bend quite easily. if you’ve ever tried to skin a torsion-box door with a 3mm sheet of skinning material, you’ll know exactly what i mean. or if you’re like me and do a lot of your mockup patterns in 1.5mm laminate-wood panels, this has already become very clear to you. the concept of bent-lamination is extremely simple. if you want a thick piece of wood that turns a corner, you take thin slices and make them all fit that shape, hold them there with glue between them and they’ll become, practically-speaking, a plywood panel in that curve once the glue dries. the only question is how thin you need to make the wood to make it bend in the radius you’re looking for.

a good place to start is that wood thicker than 3mm in most species isn’t going to do much bending. if you want a very subtle curve, you might be able to use 5-6mm-thick pieces for your lamination but i typically don’t start with anything thicker than 3mm. there is a very tiny space for the glue but, realistically-speaking, that doesn’t matter. so if you are aiming for a 30mm-thick panel, you’ll need ten sheets. want a 60mm-thick panel and you’ll need twenty. thankfully, we don’t usually work with panels quite that thick and laminated panels are extremely strong — much like plywood, though their internal structure doesn’t normally have the benefit of the perpendicularly-opposing-grain. so it’s typical to see six-sheet laminations giving 18mm or even four-sheet lams for 12mm — typical plywood thicknesses.

before we get started on the procedure, it’s probably good to remember you’ll see the laminations on the side of whatever you’re making. so either you like the plywood-edge look or you’ll need to cut some banding material. the easiest way to do this is to cut the banding oversized then shape it once it’s been attached to the completed bent segment — you can do this by hand with a spokeshave or chisels or use a handheld trim-router with a pattern-bit. i’ve used both and they’re quick and effective. the bent piece gives you a great reference surface and even completely chiseling the extra away takes very little time in most cases.

there are various schools of thought about bending forms. one is that you need a form that outlines the curve, top and bottom that has a bunch of space between the inside and outside pieces. another is that you need a mostly-solid bottom form and the top can be replaced by a torsion-strap (usually ratcheting). a third is to make the form from sticks that point out to the inflection points of the curve and pull the whole thing tight against those points, allowing the curve to form naturally. yes, these all work. sort-of. but they’re a lot of pain and annoyance to use in-practice and the results are usually questionable. they’re also not very beginner-friendly. so much tension and flying-by-the-seat-of-the-overalls are involved in the process. i’d rather guarantee success — and if this is your first time doing it i suggest you stack the odds in your favor, as the mocking-jay would say (i really miss teaching young-adult literature courses).

what you need for this procedure is a pair of solid bending-forms. you’re aiming for two pieces shaped to the inside and outside of your target curve. i think the easiest way to do this is with plywood (or mdf if you’re not asthmatic like me and can put up with the dust). draw a cross-section of the lamination you’re doing on a big sheet of posterboard. now complete the rectangle it fits in. build the rest of the rectangle, the pieces above and below, from plywood and laminate multiple identical pieces of plywood to get the thickness you need for the entire laminated piece to fit between them without hanging out — you want your form to be exactly the depth of your laminated piece. leave at least 60mm on each end of the thing you’re making to be cut off, though. lamination at the ends is often problematic and you’ll get a little springback if you’re not careful.

once you have those forms made from plywood and glued together, you can figure out how you’re going to clamp it. draw your clamp positions that will hold the two pieces to each other for the maximum amount of force between them. there’s a shortcut on this if you are looking for one and have a good bench, though. and i often take this approach so i’ll describe it first.

if the whole form fits in your bench-vise, you can rejoice because this whole thing is going to be very easy. just clamp it together in your vise and let it dry there. simple.

but it probably won’t unless you’re making something very small or have an unusually-large vise. no problem, though. you have something better than a large vise — a large bench. if you have dog-holes in your bench, setup as if you’re making a full-width planing stop opposing your vise. now you should be able to position the two pieces of your bending-form with one against the moving dogs on your vise and the other against the planing-stop. clamp the thing in the vise with the laminated piece in the middle and walk away — you’re done.

if, however, this is not a possibility for you — for example, if you’re working without a good face/end-vise or, more likely, you’re actually making many laminated parts all at once and only have one or two usable vises, my typical problem with using this approach in practice — there’s an alternative. you can use clamps. a cautionary-tale about clamps. bent-lamination takes a lot of force. a shitload. a fuckton. more than you think. by far. for most of these things, you’re probably going to need a clamp about every 8-10cm the entire length of the curve. so you need to drill large holes on each side of the curve, perpendicular to the tangent of the curve at that intersection-point, so you can get the jaws of the clamp to fit in them. i use f-style clamps, sometimes c-clamps if it’s a small piece. once the piece is in, you can clamp the whole thing tight with pipe-clamps then add all the internal ones through the holes.

here’s another method i’ve used that works — clamping cauls and pipe-clamps. remember how we were going to do this in the vise? well, you can do the same thing with two long pieces of wood and a pair of long pipe-clamps. take the whole thing and lay it on the pipes with the clamps facing up. put a piece of wood on each end — between the clamps’ faces and the bending-form. now take a thin scrap and put it directly in the middle of the clamping form on each end. so you have a clamping-form sandwich, the outside pieces of wood are the bread and above and below the filling is a scrap — maybe 3mm thick and 50mm wide. this will distribute the pressure from the outsides across the entire surface of the form. it’s not necessarily as effective as using a lot of clamps but it’ll get the job done and i’ve never had a piece come apart when done this way.

the actual lamination is the easiest part. i have found it useful to grease the inside of the bending-form. some people use plastic or tape but i think it’s easier to use paste-wax and just skip the packing-tape nightmare — when that stuff sticks to the glue, you can get some weird reactions between the two adhesives and … well, let’s just say i don’t want to repeat that encounter.

with that done, take the all your pieces of 3mm veneer and spread one side with glue — except the final piece. this one will be dry on both sides for obvious reasons. believe me. you’ll forget. i have. it’s not a very calm thing. you end up trying to get all the glue off and it gets very messy, very fast. once you have all your pieces prepared, lay the first in the form, dry-side against the bottom, glue-side facing up. stack them like that until you finish with the final all-dry piece topping off your sandwich. press the top piece in and squeeze the living shit out of it. you’ll get lots of squeezeout. lots. you’ll feel like you made a mistake and your whole world is swimming in pva. that’s ok. if you’ve done it right, this is to be expected. i promise, you can’t squeeze out all the glue. it’ll hold.

clamp it tightly. keep going. you can’t squeeze too hard. honestly if you could park a ship on top of the form you’d probably still be jumping up and down on the deck to add pressure.

leave it for the glue to dry. if you’ve used pva, i’d say 6-8h is plenty but you can leave it as long as you want. if you used structural epoxy (like ecopoxy, totalboat or west-system), which might be a good idea if you’re worried about doing it quickly enough to get all the pieces in, leave it at least 24h before you even think of releasing those clamps. you’ll get springback. don’t worry about it. there are many ways to mitigate that problem but remember you’re edge-banding the whole thing. the edging will hold it in shape once it’s attached.

there are two ways to do that. i call them the pragmatic approach and the commiseration approach. you can call them before and after. cause those are so much easier to spell. the pragmatic approach is to actually glue the edge strips to the piece while it’s still clamped in the form. this way, you won’t get springback at all. the whole thing will be solid. the other benefit is that you don’t really have to wait. once the first step is done, you can simply clean up the squeezeout while it’s still wet and glue the edges on. remember to paste-wax the form or you could really end up with a sticky mess. you want to be able to release the form, remember, not glue the form to the edges. the edges should be trimmed very close to final shape. you can clean up a millimeter or two later but you want it to be very, very tight or it will risk getting adhered to even the best-prepared bending-forms.

the commiseration approach is to wait until the piece comes out the clamps, is smoothed and sanded then attach the edging. this is how i usually do it. i’ve found the “before” version too challenging to make work. i’ve successfully done it but it doesn’t seem worth all that extra effort and physical moving things around. anyway, once the thing is finished, plane or sand the sides to ensure a good bond with the edge, glue it on, clamp it thoroughly and go back to your watching-glue-dry activities (shop naps are fun).

with the whole thing complete, you can trim the lamination to length and scrape or sand the surfaces and you’re all finished. there’s no real complexity to any of this and it’s actually quite fast. it sounds complicated. but all you’re really doing is the equivalent of taking a bunch of sheets of paper and gluing them together to make a strong curved piece of cardboard and i’m sure you’ve all done that at some point as kids — for example, making a curved-wing paper-airplane in school.


this technique is completely different. it relies on math rather than brute-force and the joys of modern glue. it’s actually very simple but it scares a lot of people. the idea isn’t complex, though.

wood is strong even when it’s very thin. but most of its strength is in one direction — resisting torsional bending forces. as it gets thinner, its resistance to direct-bending is dramatically reduced. remember how easy it was to bend 1.5mm or 3mm sheets of veneer to make the bent-lamination we just talked about? well, instead of cutting the wood to make multiple sheets that thick, let’s just use a piece that thick to start. i know what you’re thinking. if you wanted a finished-thickness of 1.5mm, you’d be building a very tiny piece of furniture. but that’s not the point. the wood will actually hold together if that’s all that’s left — in very specific places.

if you’ve ever cut a groove with a tablesaw or tracksaw, you’ll understand the concept. you run multiple kerfs to get the task completed. instead of running them in overlapping passes, though, you space them. this is something that is easier to experience than explain. so i invite you to give it a try.

take a scrap maybe 60-100cm long and cut a single kerf right in the middle with your thickest saw until you’re 1-2mm from completely-through. now cut another next to it, spaced perhaps 5mm away. move another 5mm and do another. keep going until you’ve got eleven kerfs. now move back to the first cut and repeat it in the other direction, cutting ten more at 5mm spacings to a depth of all but the last 1-2mm. now you have twenty-one of them. pick up your board. try to bend it. that’s kerf-bending.

of course, it takes a little calculation (or trial-and-error) to figure out how many kerfs you need, how deep and how far-apart they need to be. but i have good luck for most curves spacing my kerfs 3-5mm with a kerf-size of 3mm on the tablesaw (it looks like a finger-joint) or spacing them about 2mm if i’m cutting them with a kataba (or the manual saw of your choice). you might need a lot of cuts if you’re aiming for a very tight radius on your curve. you’ll need far fewer for a larger radius. don’t assume the wood will compress — just add more kerfs. if you cut to a depth of all-but-2mm, you’ll probably be fine. if you really need it to bend, you can go closer to the edge but the closer you get the easier it will be to snap it when you actually do the bending.

the only other pieces to this are adhesion and banding. once it’s curved how you want, open it back up and add glue in each kerf. then bend it, clamp it in position and let it dry. i highly-recommend using epoxy for this because the gaps aren’t going to be tight enough for the kind of pressure pva requires. it can be done but it’s less likely to stay stable. when that’s dry, clean it up (the edges will be covered in squeezeout unless you forgot to add the glue) and add edging in whatever material you like — probably whatever you used for the bent piece. done. yes, it’s more math. but wow is it ever easy once you get the hang of it.

direct-heat bending

also called hot-pipe or hot-plate bending, the easiest way to bend a piece of wood is to get it hot enough the internal tension relaxes and you can just manipulate it in your hands. this is not an easy process to do safely, though. the simplest approach is to use a copper pipe (no, not galvanized or treated because you’ll be breathing in the off-gassing vapor and that’s a quick way to get yourself hospitalized) and a blowtorch. heat the pipe by sticking the blowtorch in the end and touch the wood on the pipe. i don’t recommend this approach. it’s very dangerous. even with the pipe clamped in a vise, it’s easy for things to come loose and even easier to start a fire — an open flame in a shop full of wood — what can possibly go wrong?

yes. you’ll light your bench on fire and probably yourself. so heat-bending is great. but this is a bad way to do it. if you’re going to heat-bend wood, a much better, safer approach, something i can actually recommend, is to either use a purpose-built wood-bending pipe tool (do a search — i can’t recommend a specific one cause i’ve never used any of the commercial ones) or make one yourself. the way i’ve made them is to take a heating element (electric soldering torch in my case) and put that in a copper pipe, turn it on and … well, it’s still going to be very, very hot. you have to be careful with this and i take no responsibility for what you might end up doing to yourself or your shop. press the wood against the hot pipe and it will bend at that point nearly at a complete 90-degrees if it’s a thin piece of wood.

a couple of caveats. first, like i said, you do this at your own risk. i don’t teach this in class and i have generally avoided even doing it myself because it’s fairly dangerous and prone to burning things — and scorching the wood if you’re not very careful. second, though, it’s not very effective. yes, you can absolutely bend a 12mm-thick board this way without too much difficulty. but if what you’re looking for is anything thicker than that or a denser wood than poplar or spruce you might end up having to heat the wood so much you’re turning it black and toasting the thing before it bends as much as you’re aiming for. this is great for thin things and it involves no glue. where you see this in the wild is instrument-construction. guitars, violins, cellos, etc. they’re often made using this technique but they all have something in common — very thin, light walls. so this isn’t a technique i recommend for typical hobbyist woodworkers — especially not unless you want to risk second-degree burns. the other approaches are far safer and more effective on thicker boards like we tend to use for furniture-construction.


bending wood in hot water is fairly simple but it’s slower than you think. and this is generally only functional for small projects unless you want to buy or build a water-heating appliance. there are better alternatives for those so don’t even contemplate the wisdom of it. it’s a bad idea. always.

but for small things like the sides of boxes that you want curved, this is far easier and faster than using steam or gluing laminations. get your biggest stock-pot — yes, one specifically for this purpose. i recommend a hotplate in your shop and one seriously-large pot — maybe 10-15l, something meant for large families and lots of soup — and fill it most of the way up with water. boil it. when it’s boiling, drop in your wood.

you’re going to need a form much like you had for the lamination process. i already explained how to build a two-part form, though, so i won’t describe it again here. just build a form to fit the piece and be ready to clamp it in place.

i recommend attaching a string to the end of the board or getting a good set of tongs. either way, you need to get the wood out without draining the water. put all the pieces in at once. once you’ve boiled the wood for a while, take out a piece and — very, very carefully — bend it. don’t touch the wood. it’s boiling. stick it in the form and clamp it in place. you’re done. just let it cool — actually, if you want it to hold that shape, it’s probably best to leave it in the form at least 24h, maybe longer. but you can take it out whenever you like if it’s going to be held in that shape by other components, which i expect it will be.

if it doesn’t bend easily enough to go in the form, just drop it back in the water. the longer you boil it, the more it will be like shaping cooked pasta. actually, this is exactly the process of cooking pasta and wood is very much like spaghetti noodles in precisely this way. you can’t really overcook it but you can make it soft enough to be a little difficult to work with. try fifteen or twenty minutes. if that’s not enough, give it another ten or fifteen and keep doing it in those increments until it fits the form. there’s no need to do most for hours. if the piece is big enough to require that kind of boiling time, you’re probably not going to fit it in the pot in the first place.

a word to the wise. students have asked me about doing this using other chemicals as a way to infuse the wood with scent or dye. don’t. most of those things are either flammable or reactive. they’re not usually meant to be heated to boiling. if you get any “smart” ideas about letting the wood soak up finish more easily at that temperature, what you’re actually talking about is building an explosive device in your shop. don’t do that. it’s bad for you. and everyone else. linseed oil will burst into flames. denatured alcohol will start a fire and make your workshop look like a scene from backdraft. water is good. water is happy. wood and water go nicely together. you’ll thank me later. and your eyebrows will still be on your face where they belong.


the only difference between bending with steam and hot-water is speed. steam takes far longer to take effect than heat directly-applied using water. there are many ways to build a steam-box. for long boards, i have generally used large-diameter pvc pipe — you know, like sewage pipe or plastic conduits in a big-enough diameter to fit the boards. you probably want to do all the boards at once because it will take a while for them to get soft enough to bend.

a cautionary note before i describe the procedure. you’re pumping steam into the pipe. it needs a place to come out the other end. a tightly-sealed pipe with steam pouring into it and not coming out has another name — a pressure-cooker. it’s also a bomb. a pipe-bomb. don’t make one of those. make sure you have a nice hole at the other end for the steam to escape so pressure doesn’t build up. if you get this wrong, it’s not my fault. i’m warning you now. you can blow up your shop if you pump steam into a pipe and don’t give it anywhere to go. see why i’m saying this so many times? yeah. i don’t want you to get hurt. be safe. i’m a teacher. i’m obsessed with safety. but i haven’t had anyone rushed to emergency yet and i’d like to keep it that way.

so you get yourself a steam-generator of some sort. you can buy a commercial steam unit (rockler makes one that’s quite good but a bit pricy, though i’ve only used theirs in someone else’s shop so i can’t really give it my stamp of approval) or get a wallpaper-steamer, which is what i’ve generally used. as long as it makes steam, it doesn’t much matter what it is. you pipe the steam into your box — probably a pvc pipe or something made from screwed-together pieces of plywood. oh. don’t use mdf. you’ll soak up the steam and the whole thing will come apart like soggy cardboard. yes, even laminate-coated mdf.

you drilled a nice big hole to let the steam out the other side, right?

anyway, you keep pumping in the steam and letting it go through the box/pipe and heat the wood. let it sit there for some time. it’ll probably take at least an hour. i usually end up having to let it steam for at least three or four if it’s serious wood (i usually use oak or maple, sometimes cherry, which bends a little more easily but is still slow — if you’re bending softwoods it’ll be faster but are you seriously using softwoods for a project involving this much curve? maybe fir but even that’s not very strong and you’re killing a lot of the inherent strength by bending it…). take out a piece of wood and try to bend it in your form. if it doesn’t bend enough, stick it back in and keep cooking. it’s really trial-and-error. there’s no real guideline. even two trees of the same species have different internal tensions and structural strength and different boards in the same tree can vary quite a bit.

much like with the hot-water approach, leave it in the form at least 24h and it’ll hold its shape quite well.


there you have it. lamination, kerfing, heating, boiling and steaming. you, too, can have bendy projects in your woodworking life and these techniques are fun to try (except the direct-heat approach, which you probably shouldn’t).

a few final notes. you can do this with any species of wood but the denser the wood the less easily it will bend. if you want to steam-bend jarrah or live-oak, you’re going to have it in that steam-box for ages. that doesn’t mean it won’t bend. you just have to be very patient. cherry and walnut bend much more quickly because they’re simply much softer and the steam penetrates more easily. if you’re trying to kerf-bend or laminate-bend harder woods, you might need to use thinner strips or leave a thinner outer-strip (at the bottom of each kerf) or, perhaps, make the kerfs closer. like most things, these techniques are very easy to learn using a little bit of trial-and-error.

you will have many people tell you you must use air-dried or green lumber to successfully steam/water-bend. that’s not true. you can (and i have) with kiln-dried stock. will it bend as far for the same amount of time being heated? no. will it bend as far if you heat it longer? absolutely. there’s nothing you can do with air-dried lumber that’s impossible with stuff that’s been in a kiln. just remember the actual structure of the wood has been altered so it is more difficult for it to absorb moisture. it can take two or three times as long for the heat and wetness to have the desired impact so you’ll have to be more patient with it. so if you have access to green lumber or air-dried boards, this might be the best source for your project. but don’t give up because you don’t have that — 80% of the bent stuff i’ve done has been with regular, kiln-dried wood and it works just fine.

there are many ways to have fun in the shop. most of them involve wood. this, though, i suspect will be one of the most fun projects you’ve tried in years. it’s especially good as a project to tackle if you have older kids — teens, i’d suggest, because younger kids aren’t usually all that careful with hot things, if you’re going to do steam or water. but even really young kids love to make stuff and will happily spread glue if you’re doing a bunch of laminations.

a great little bent-lam project to do with children is to make a few small toy cars shaped in the form of a u. make a few bending-forms about 30-40cm — semicircles on the ends of long boards with mating pieces. make 50-60cm strips and glue them in the form the whole thing should take no more than a half hour and be within their attention-span. the next day, go back out with the kids and pop them out of the clamps, quickly sand, drill two axle-holes through them and attach wheels (you’ll have to prepare these in advance, i suspect) and they can be playing with the cars in no time.

wood is fun. sometimes we forget that and just think it’s beautiful or the process rewarding. but it can be whimsical and artistic, too. thanks for taking the time to explore bending with me today. stay flexible, stay fluid, even try out being floppy. enjoy!

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