[almog]

> "The strength comes from the tightening/wedging effect when pulling the join apart. If the dovetail is tapered, the join can also tighten when it aligns too – this is highly desirable for gluing, as it means the glue will not be scraped away."

While true, dovetail and other traditional woodworking joints borrow their strength from using the wood grain to support the joint.

I don't know a lot about 3D printing but I'm guessing a lot of the assumptions about woodworking joints strength aren't applicable.

[wongarsu]

Grain direction in woodworking has many similarities to layer lines in 3d printing. That makes techniques from woodworking often more directly transferable than techniques from CNC machined, cast, or injection molded parts.

[almog]

How so? What kind of filament preserves a multiple fibers like structure that resembles wood grain? Sincerely asking since I couldn't find any source for that

[wongarsu]

Each continuous strand of filament is kind of like a wood fiber, in the sense that the strand is much stronger than the connection to the other strands.

Of course 3d printers lay these strands in a 2d plane (and stack layers of these 2d planes), whereas trees grow in more-or-less one dimension. So wood has one strong direction and two weak ones, while 3d prints have two strong directions and one weak direction. And obviously 3d printed strands are layed out by boring and often naive algorithms that lack any of the beauty of naturally grown wood. Then again, for structural applications you also want the most bland and boring wood patterns possible.

[almog]

That much I'm aware of, but as I mentioned in previous comments, with wood strength is high in the direction of the grain but weak across the grain, plenty of academic studies confirm that (and it takes little to verify it empirically).

With 3D printing, despite the analogy, I couldn't find any sources for strength along/across the "grain" so i'm not sure how accurate that analogy really is.

[wongarsu]

With 3d prints this is usually framed as layer adhesion. Within one layer (so in the x-y plane) you are basically "along the grain" since the strands form loops and other 2d shapes within the layer. Between layers (== in z direction) you only have the adhesion between different strands holding the layers together, which is equivalent to going across the grain in wood.

CNC kitchen on youtube does a lot of strength and impact testing, e.g. [1] (results at 9:55). "Across the grain" you have half the strength in PLA (and similar numbers in all filament types except TPU). Or if you prefer manufacturer numbers, [2] is the datasheet for a random PLA filament. It also shows worse numbers in every metric in the Z direction (across the grain).

The difference isn't as severe as in wood, but it's big enough that it is something you have to consider in structural parts

[1] https://youtu.be/dOzVuoBP9gY?t=535

[2] https://polymaker.com/wp-content/uploads/lana-downloads/Poly...

[almog]

Thank you!

[WillAdams]

layer lines.

That said, if the part is small, and one can afford to print w/ 100% infill just put it in a tray of salt and bake to re-melt/re-flow to make a (more) solid part.

https://x3d.com.au/blogs/tips-and-tricks/pla-baking-the-secr...

[almog]

Wood strength is high with the direction of the grain and weak across the grain, is that a property of layer lines too? I just couldn't find any source that claim it to be so.

[WillAdams]

See:

https://xometry.pro/en-eu/articles/3d-printing-strong-parts/

[almog]

Thanks!

[bb88]

Interesting. I didn't know that.

But I do know about 3d printering. :)

Strength in printing is usually a function of the mass used to print with. So more plastic used at the boundary layers will mean stronger joints.

Eg. A PET plastic soda bottle (pepsi, coke) will flex in the body where it's thin, but not on the threads on the top of the bottle. In fact, those are pretty strong, because they need to hold on a cap which keeps the carbonation in.

A good question to ask is if the dovetail joint in plastic as strong as wood? The answer may likely be no, but it may not also need to be. It just needs to be "good enough". And this looks like it might be.

[observationist]

https://github.com/TengerTechnologies/Bricklayers

This came at a perfect time, then - you can cause each adjacent layer to be offset by half the layer height, to maximize the strength.

Annealing parts by baking them at the right temperature for a sufficient length of time is relatively easy to do. Printing at higher temperatures and slower speeds to ensure layers are melted together also results in higher strength parts.