Digital Wood Joints | Hacker News Reader

Digital Wood Joints

openup.design

141 pointsmontgomery_r68 comments

These are all designed for CNC cutting, and appear to be intended as one-sided jobs worked with the face of the board down. That is generally the simplest way to hold a board on a CNC.

I would caution anyone thinking of doing joinery this way to consider whether it's actually suitable for the application. I was at a coffee shop once where absolutely every chair in the place was starting to fall apart. The reason was that the joinery was all some variation of half-lap, which doesn't constrain the movement of the pieces in all of the directions that matter. Once the glue failed, the chairs started coming apart.

I would also add that the corner joints meant to replace dovetails or a box/beehive joint are unsightly with all the required dogboning and will not improve aesthetically with the addition of glue. I would further point out that there are already quick ways to cut dovetails or box joints with a router quickly and efficiently. It would be hard to convince me that there's a truly useful role for cutting an uglier version of a box joint on a CNC.

Source: am a furniture maker who does some CNC work.

I wonder how much of the simplicity is due to the fact it was created in 2004. I don’t have a good feeling for how quickly things progressed over the last 20 years

I wasn't in the trade in 2004, so some of this is a bit speculative:

It looks like everything on the poster is made to be cut on a 3-axis machine. Stepping up to five-axis is a huge leap in cost at present, and surely was then as well.

Tooling has likely improved in availability and cost since 2004 as well. Automatic tool changers have probably also become more affordable, but just like going to five-axis, you're getting into a whole different class of machine once you start talking about adding that.

Some of the reasons to cut stuff single-sided on the flat are unchanged by any of that though. It still costs you precision (and time) to flip a workpiece over, and you're going to have issues if you don't have good consistency with your material thickness if you need to reference your Z axis to the material surface rather than the bed surface.

Working an end of a long piece remains a pain in the ass for fixturing that involves a hole in your machine bed, and possibly the floor as well. Tenoning a bed rail, for instance, is inconvenient however you do it.

All of my CNC experience is on a three-axis machine, five-axis gets you a lot of flexibility that most places won't have unless CNC work is their primary focus, or at least core to their workflow. I've seen a shop that builds high-end windows with a large five axis machine. I have no concerns about the durability of their products, it's just that most people don't have access to that kind of capability.

Thanks for the informed commentary. I'm surprised that the tools haven't dropped in price faster, but perhaps there is a de minimis based on amount of material / strength / precision engineering required. I got to the site because I have a plan to make a bed frame (poor first project choice I know) and found this whilst looking for joints that might work gluelessly.

In my experience, the cost has significantly dropped, it's just that back then they were even more absurdly ridiculous

If you're making a bed, consider that this is a problem people have worked on for a very long time and there is an entire vertical industry on it.

One reputable company has an entire section in their catalog on it:

https://www.leevalley.com/en-us/shop/hardware/bed-hardware

i'm a hobbyist woodworker with more money than time. i have a pretty basic 3-axis cnc and i thought it would save me time, but it really doesn't. the only thing i actually use it for is cutting out router templates, and even that would be done better with a laser cutter (although a good laser cutter costs a lot more than my cnc).

i could see how a machine big enough for 4x8 sheets with an automatic tool changer, a vacuum table, and all the automatic calibration gizmos might be a time saver for a production shop, but if you're building something that's a one-off or you don't have all the setup automation goodies (which are $$$$$) then setup and programming usually end up taking longer than doing the work the old fashioned way.

for tenon cutting like in the bed rail example you gave, i have a hard time imagining any situation where cnc is going to be more efficient than a domino xl.

4x8 CNCs with a vacuum table really aren't faster. Even the watercooled CNCs I've used are still too slow for joinery. All the furniture shops I've worked in have been dominated by the Domino for most joinery tasks.

I find CNC is a time-saver for one-offs when the work is complex enough that it'd be difficult-to-impossible to do by hand, eg complex curving cuts, engraving/pockets, etc.

I actually saw an unusually straightforward example of this last year - a group of friends and I were making instances of Tyler Gibson's 1-sheet portable bike rack design (it's great, check it out: https://www.thetylergibson.com/building-a-better-portable-bi... )

One group of two-ish people used jigsaws to manually cut the pieces, and I used a Shopbot 4'x8' CNC router. Very roughly, it took about twice as many man-hours to make one by hand, vs by CNC, and the result was less clean. CNC could have done even better, but due to warping of the sheet, it failed to cut all the way through in places, and I had to do a cleanup pass with the jigsaw. And once the upfront cost of generating the toolpaths etc was paid, it would improve again.

Yeah. I want to know how well these joints will last in different scenarios and how the tradeoffs are.

I’ve seen YouTube videos where people test different wood joints to see how much load they can withstand, but I assume the results will be different if you account for things like changes in humidity over time, which causes the wood to expand and contract (which doesn’t happen isotropically).

Fun chair force article: https://www.woodreview.com.au/how-to/the-logic-of-chair-desi...

Fun chair anatomy video by arguably one of the most talented ones currently active.

https://www.youtube.com/watch?v=T2g6kGl66XU

If the glue has failed, there are some serious craftsmanship issues regardless of the joint type.

With the exception of the joints labeled "...with key" these joints are all very remote from the types of joints used in traditional Japaneses temples which do not use glue.

These are mostly western style joints, which are also very beautiful and useful, but generally expected to be assembled with glue.

Great resource!

> If the glue has failed, there are some serious craftsmanship issues regardless of the joint type.

No. You can't simply use whatever joint you want and expect the glue to deal with the (sometimes enormous) forces applied to it.

It really does depend on the joint type. Do you expect a lap joint to hold together without any additional fasteners or glue?

This goes against the conventional wisdom that a properly glued wood junction is stronger than the wood itself, and that under such forces it is the wood that will fail.

The wood would've failed at the same place if you'd carved a whole chair in that shape out of a solid piece of wood. The problem is that the design concentrates forces at the joint (or the "junction") in a way that no material can withstand.

Agree that you "can't expect the glue to deal with the forces"

A good craftsman would not choose a joint that would see such high stresses.

Additionally if the glue is chosen and applied properly, the wood that the glue adheres to should fail long before the glue.

That said, glue is not as simple as it may seem. There are many different types and proper surface prep and application makes a huge difference to ultimate strength.

For example, many people will mix 2 part epoxy until it "looks mixed" which for a clear epoxy happens pretty quickly. In truth, the resulting bond strength is far more closely related to the amount of mechanical energy that has been transferred into the mixture than the visual uniformity.

Lots of ways to go wrong with glue... but a good craftsman should be well aware of these.

Yes, the one thing I was missing from all of this is what is the appropriate application for each joint

As someone who is not a furniture maker and doesn't have access to a CNC machine, I was surprised by how good of a joint I can make with just a doweling jig. It does end up slightly more visible (assuming you aren't painting it), but it creates a very strong joint that is hard to mess up.

As another layperson who always dreamed of making dovetail joints for strength, I was quite surprised at various explorations of joint strength on Youtube that tend to find that mitered box joints with dowels or splines beat most other joints by a wide margin. This is great because they are even easier to make, and can be made to look quite nice if you use a contrasting wood for the dowel/spline.

could you post the link? I'd love to watch :D

https://www.youtube.com/channel/UCJNY2_y5nPoghlAO6FPxHfg

I know a miter joint, a box joint and dowel joint. But what is a mitered box joint with dowels?

How is it more visible?

Dowel joints are great, as far as I know Krenov never used anything else on his cabinet carcasses. I'm not sure why they developed a bad reputation.

Easiest way is to drill through, then cut and sand the dowel flush; leaves a round circle where the grain doesn't match for each dowel.

You are probably thinking of dowels used as loose tenons, where a blind hole is drilled into the mating pieces of wood and the dowel is not visible once the joint is together.

Sometimes people glue a butt or mitre joint, then once the glue has set, drill a through hole for a through dowel from an outside face of the joint.

The blind holes approach is tricky to get perfect alignment on where dowels are being added across a longer length. As a hobbyist I've tried a few cheaper dowelling jigs and had mixed success. This challenge lead to other loose tenon solutions like biscuits or dominos which allow for some side to side misalignment while retaining the ability to keep the visible faces aligned.

The through dowel approach avoids the misalignment problem, but comes with the visibility of the end grain of the dowel on one exposed face. Some people are ok with that.

Dowels are still one of the "strongest" options for end-grain to long-grain joints. Many professional woodworkers now favour dominos simply due to who quick it allows them to work and the additional allowance for some side-to-side alignment.

Excellent points all around. As a fellow woodworker / CNC person, I will say it often frustrates me how people with CNCs demand that the WHOLE process be done on a CNC. 90 degree corner chisels exist and make nice interior corners without needing dogbones. Other shapes can be trimmed/finished by hand. Once the CNC sets up nice reference surfaces and removes all the waste it's actually pretty easy to clean up the insides of dovetails to sharp corners.

Yeah, intuitively I doubt that these will work the way they’re supposed to because well you’re trying to reinvent the joint. It’s true that the tools are different but I think there has been a steep decline in general joint know-how, so being able to innovate joints would require catching back up to where we used to be to decide if it’s worth it. Tbf if anyone could do it it would probably be Germany.

None of these joints is novel in the slightest. Still, is it awesome to have them so well documented. Additionally, the CNC makes them much much easier to use!

Really. Show me one single source other than this page that uses a meander key lengthening joint.

Nice. But the link is to a clickbait site with dark patterns for the privacy policy. The real article is here.[1]

Somebody should run these through a finite-element package to compute their strength. Some of those look fragile.

[1] http://winterdienst.info/50-digital-wood-joints-by-jochen-gr...

Precisely! They look pretty but there's a reason (beyond tools and skills) some of these have never been seen in practice. Would be super interesting to see stress analysis on these.

Charles Hayward [1] has a great book on the topic:

Woodwork Joints https://archive.org/details/woodworkjointski0000hayw_k7x4

Hayward discusses not only the joints, but techniques for cutting them with hand tools.

There are plenty of used copies out there. Hayward has many other useful books.

1: https://en.wikipedia.org/wiki/Charles_H._Hayward

This is phenomenal. I wish I had it to share with a student I worked with last year. Even without this, he came to similar realizations and explored similar territory with significant results.[0] He began by comparing and contrasting traditional Japanese joinery with CNC milled joints. He realized that the critical difference was that Japanese carpentry joints are almost always entirely squared off (due to the flat saw and chisels). Yet the inside corners created with CNC were always rounded. So he let this critical difference be the animating concept of the project.

So if you enjoy this, you might enjoy Luke's work too!

[0]https://issuu.com/lukemurrayarc/docs/portfolio_2023 starting at page 26-27.

Here is what immediately comes to my mind.

    - Figure out a way to parameterize joints
    - Create an automated evaluator that uses finite element analysis
    - Evolve joints with a genetic algorithm

Oooo... that's a great idea! I would love to see the output of that project!

You left out one step/criteria:

- constrain joints so that they can be cut with a tool on a CNC with the material secured on the machine in a convenient fashion in a reasonable period of time

which is something I've been working on.

constrain joints so that they can be cut with a tool on a CNC with the material secured on the machine in a convenient fashion in a reasonable period of time

Could one make that a part of the evaluator/fitness function?

Yes, but it needs to be included as a consideration up-front and not left to afterthought or accident.

Yea, but wood isn’t actually uniform, so a finite element analysis might find a joint that is stronger on paper, but the particular piece of wood it needs to work on makes it weaker in practice.

It might be neat to do it automatically based on a picture of the wood, but that is a much bigger problem.

When working with wood in an automated/industrial environment, there are two options:

- design for the worst-case scenario for a given practice setup (note that wood is graded on how many knots there are and how tight they are expected to be) and work around that, eliminating pieces w/ problematic grain - use engineered lumber where appropriate

This is a marked contrast to some styles of hand-woodworking --- I can still recall walking with a great-uncle in the woods when we came to a walnut tree where the trunk had an odd angle to it (apparently the main trunk had broken off vertically in a storm and a branch had continued the upward growth) --- he immediately took a colorful handkerchief out of his pocket and tied it around the tree, looked carefully at a couple of local landmarks visible through the trees, and then confident of being able to return to the spot, we went to his cabin, fetched a saw and ax and proceeded to cut the (smallish) tree down and sectioning it off, taking it back to his cabin where he covered the ends of the central portion around the angle in melted paraffin.

Almost a decade later during another visit, he announced he had something to show me and got out a rifle which I remembered as having a broken stock --- he had of course restocked it (once it had dried sufficiently) with the tree which had grown at an angle which was perfect to make the stock out of.

Very cool graphic. However emphasizing the juxtaposition between pictures of wood joints on the internet and actual real-life wood joints with the phrase "digital" is a bit perplexing.

I think “digital wood joints” emphasizes that they’re meant to be cut with a CNC, and thus aren’t the traditional wood joints that have been taught for centuries. I don’t think it has anything to do with the fact that this content is available digitally.

Honestly I thought they were trying to make a cute pun by calling finger joints 'digital'

Your interpretation is more likely, though. And explains why so many of the joints are weird and look like they'd leave gaps...

Based only on the pics in the “poster”, it looks like a lot of the joints in the second half have visible gaps. I think I would have expected more emphasis on joints that can be cut with the CNC’s round cutter, but which also come together cleanly.

Also, an over abundance of lengthening / scarf type joints, more than the typical woodworker would use, IMO.

I wish I could find the clip now, but with the advent of practical router bits with pantographs, there was a fashion in the late 1800s to replace dovetail joints with dowel and C type joints. They were really hard to make fast by hand, but easy to do with machines.

They could be made to higher tolerances, and had more surface area, and I assume were more reliable to make/hold draws together.

These are very similar to the idea here.

EDIT: thankyou to mikey_p and WillAdams who pointed out the joint I was looking for was the Knapp Joint: https://www.finewoodworking.com/2018/09/26/history-cove-pin-...

Probably you were thinking of "Knapp joints" (cove and pin) which had a currency of a decade or so (their presence is used to date furniture), but which were replaced by machine cut dovetails.

A CNC router affords a lot of new possibilities, e.g., this box which has an integrated lid:

https://community.carbide3d.com/t/as-funny-as-a-3-dollar-box...

You're thinking of the Knapp joint which were sometimes called dovetails. The pantograph router version is a modern recreation of how the original was made.

This video has some explanation of the new method and why they were originally popular in the first place (mostly limitations of available machines): https://www.youtube.com/watch?v=MlJjsvph3r8&ab_channel=Matth...

Excellent thank you for the video!

I'd like to think we're at a place where CNC machines are cheap and accessible enough that clever tricks or new tools are just around the corner to overcome some of the inherent issues of making basic wood joints on a CNC router.

To me, these look like a stepping stone to something better. They're all intended to be used with a flat-shaped endmill, and all probably have strength and aesthetic issues. One solution is using different router bit shapes in combination with traditional techniques, but this is more labor-intensive.

we're not really close, for two reasons:

1) programming takes a long time, and it only makes sense to take the time to do it if you're making a bunch of copies of something. this is something that could be improved with better software and ux - if cad programs made it easy to just drag and drop joints from a joint library into your model then this would be a different story. a hardware+software solution could also work here, something like a cnc version of https://www.woodpeck.com/multi-router-group.html where the software makes it easy to scale the templates to your work piece.

2) setup takes a long time on the affordable machines. every time you change bits you have to recalibrate. positioning the work piece on the table and clamping/taping it down takes a lot of time. if you have to flip the work piece over then that takes even longer and positioning is even more critical, and programming is more complicated as well. regardless of whether your designs require cutting on one or both sides, you have to program tabs into your design so the router doesn't cut all the way through (or else the piece will move and the router will screw it up), and then you have to go back and cut the pieces out the rest of the way manually and trim off the tabs with a flush trim router bit. the high end production quality machines mitigate a lot of these issues, but now you are talking about a machine that costs at least $100,000 and takes up a whole room.

I've been working on the software end of things:

http://tug.org/TUGboat/tb40-2/tb125adams-3d.pdf

and

https://community.carbide3d.com/t/a-different-sort-of-box/36...

and for more see:

https://willadams.gitbook.io/design-into-3d/programming

Interesting idea, using the strengths of the tooling to expand how joinery can look. Some of these are photos; looks like they made actual wood examples of some of these joints.

What is really missing for me here is complete pieces, utilizing some or many of these new joints throughout. Are they practical? Do they look good as part of a larger whole? Even 3D renders would be a good start, if not an actual physical piece.

Most of these are basically impossible to cut without CNC. Once you’ve embraced CNC as your joint method, why wood is the next question.

If you want to learn wood joints, learn mortise and tenon, dovetails, dowels or floating tenons, and nailed/screwed joints. Some extra love for biscuit miter joints but really that is all the main ones people use.

Offenbach also hosts a relatively large Japan festival (in Germany). Wood joinery, at least in my mind, is a very Japanese craft. I wonder if this is by accident.

There are a couple of Japanese joinery books out there. One that I've read gives the reasoning for all of the joints in the Japanese tradition for joining two boards end to end (thereby creating a longer board, in effect) is that the timber available for building in Japan doesn't come in long, straight lengths.

I can't tell you if that's true or not, but I have a hard time seeing folks putting so much work into developing such joints without a pressing need for them.

There’s a wonderful little carpentry museum in Kobe, highly recommended:

https://www.dougukan.jp/exhibition?lang=en

It focuses a lot on the evolution of precise woodworking tools, like saws and planes. They also had examples of complex joints, made without nails or glue.

Lovely!

Off-topic, but that's a giant fuck you of a cookie popup... it's got 3 options of "high/medium/custom data privacy", a duration of how long I would like to keep those settings (just like those nagware popups of "Try OneDrive/Teams/[whatever fucking bullshit] now!" with no option for "Get off my OS!", only "Remind me later"), a customize toggler that makes it even larger to have more settings. Someone programmed this to frustrate the user with "

And the description for the 3 options is contradictory. Above them it says, "Select a Data Access Level", but selecting "High", the description is "Highest level of privacy"... So what does it do, give the advertisers a high level of access, or give me a high level of privacy?

Yep, such links should be re-written to the original source, in this case:

http://winterdienst.info/50-digital-wood-joints-by-jochen-gr...

Meh, Matthias Wandel's still going to beat you on speed (and quite possibly strength) with his homemade pantorouter or box joint jig.

As much as I like nice joint design, I think a lot of that aesthetic beauty comes from when it's not machinable (or not traditionally anyway, iirc the pantorouter can make some pretty tight dovetails - because of the orientation vs. trying to route them 'normally'). It's not how fancy and curvey can it look, it's how intricate and fine the detail. Chunky dovetails don't look better than box joints in my opinion.

I would argue that it shouldn't be necessary to see joinery at all --- I find full-blind dovetails far more impressive and aesthetically pleasing.

this is an excellent reference - not only for computer-controlled wood processing, but also for digital fabrication. When trying to create pieces made of plastics, you will sometimes need a strong joint, and wood-working is the perfect place to look

the images are great, but need more info about strengths & weaknesses of each type

I've been working on this sort of thing for a while.

For a Japanese spin on this see Tsugite:

http://ma-la.com/Tsugite_UIST20.pdf

which I worked through at:

https://community.carbide3d.com/t/a-study-of-joinery/28492

Traditional joints (box, dovetails, or obscure variations such as Knapp (cove and pin)) require a vertical fixture and 3 setups (at a minimum) --- cut parts to length and machine internal features, mount four board and cut joints in 2 corners, flip boards (with correct orientation) and cut other two corners.

Rabbet joints are simpler --- so simple that they were covered in a video as "The Simple Box":

https://www.youtube.com/watch?v=V93xDM3lXsM

(ob. discl., I work for Carbide 3D)

There have been a number of programs developed for joinery. A current commercial option is:

http://www.g-forcecnc.com/jointcam.html

(but it requires a vertical fixture)

One commercial option became freely available:

https://fabrikisto.com/tailmaker-software/

and ingeniously has an option where a 30 degree V endmill is used, but to cut boards held at a 15 degree angle, affording a 90 degree cut with a great deal of control and flexibility --- this can multiply setups to 9.

A variation I've been experimenting with is full-blind box joints:

https://community.carbide3d.com/t/full-blind-box-joints-in-c...

They're reasonably easily drawn up, though they do have some rather specific tooling requirements (a narrow 90 degree V endmill, a square tool of that or smaller diameter, and to make things easier, a large V endmill)

One test project was so tight that after putting it together for a dry-fit before gluing I was unable to get it apart:

https://cutrocket.com/p/63781eaf9822f/

I've been working on a programming system to make this sort of thing a bit easier:

https://github.com/WillAdams/gcodepreview

and have some sketched out joints which I've not been able to make using existing CAM tools which I hope I'll be able to do using this system (if anyone could recommend books on conic sections, I'd be grateful --- that's where I got bogged down last time).

think the site suffered a hug, not opening for me ;(

I believe it's the same as:

http://winterdienst.info/50-digital-wood-joints-by-jochen-gr...

(which is the original site and which would be better to use/link to)

Reply on news.ycombinator.com