Print-in-Place Chainmail Jewelry - 3D Printing and Designing Decorative Jewelry in Fusion 360

After printing the super cute mini octopus during my Cam Toy project, I got quite interested in moveable parts, such as joints or links, that can be printed in place, which led to this print-in-place chainmail jewelry project.

I really love the idea of jewelry that moves with and/or conforms to the wearer's body. While textiles are great at conforming to the wearer, it can be more difficult to create 3D geometries with them. Macramé is a knot craft that can create 3D geometries that are relatively more rigid. Though I personally am more experienced in making 2D macramé (mostly friendship bracelets), I intended to incorporate some of that into this project.

On the other hand, if you search for 3D printed jewelry online, what comes up is mostly likely 3D printed pendants. Even with headpieces, they tend to be composed of rigid pieces that don't behave like fabric or textiles. However, flexible 3D printed artifacts has been somewhat explored. There are different 3D printed chainmail models on sites such as Thingiverse or Printables. Nervous System's kinematics system creates complex, foldable, and fabric like jewelry that are all printed in place.

Therefore, I started out with the goal to create a jewelry set, including a headpiece, earrings, and a necklace. I wanted the parts to print in place, requiring minimal assembly. Additionally, I would like to create opportunities to incorporate macramé to the jewelry in order to add texture and character.

Due to the amount of time needed to explore existing work, to model the components, and especially to troubleshoot print issues, I decided to focus on the headpiece first.

Modeling: Fusion 360

Slicer: Cura

Printer: Ender 3 Pro

Print material: PLA (more flexible material for the snap fit clips might be even better)

Macramé textile material: nylon(?)/synthetic threads, lighter and glue to secure the threads

I searched for and then test printed available chainmail like fabric model from online 3D printing communities.

As my first step, I wanted

1. to see if printing chainmail is actually doable on my printer
2. to feel out what kind of chainmail design do I like or dislike, how wearable are they, what kind of quirks different types of design have
3. to gauge the time required to print large pieces

Chainmail/fabric design I tested:

• Chainmail - 3D Printable Fabric by flowalistik
• This is a square chainmail, looks and behaves relatively close to actual metal chainmail. It has a simple geometry and can be easily arranged using rectangular pattern in Fusion 360.
• Triangle Fabric by Devin Montes (video)
• This fabric needs to be scaled up for better print quality. My small test print didn't flex and move as smoothly as I wanted. The hinge design on the backside felt a bit stabby.
• However, the design files included modular pieces, eg. a left corner piece, an edge piece, which is quite useful if one were to create rectangular fabric. Certain loops are intentionally left incomplete, so different fabric pieces can be joined together using a 3D pen.
• NASA chainmail by Halling Slimta
• This is the NASA space fabric. I liked how the fabric when laid flat doesn't really have any gaps between the hexagons. When folded backwards, it becomes more rigid and can hold itself up. I liked the circular and spiral design of the back as well.
• Chainmail 3D Printed Fabric that can snap together too by k.makesstuff:
• This one is way thicker than others. It didn't snap together for me unfortunately, possibly due to material or tolerance that needs to be tweaked.
• I didn't like that it really only folds horizontally or vertically. It is not very close to fabric-like in my opinion.
• Tetrahedron chainmail fabric by Luke's 3D
• This chainmail has a very interesting surface texture. When printed with silk PLA or even metal, it looks super shiny and cool. It behaves a lot like fabric and is my favorite one so far.
• The amount of small features on the initial layers makes it rather difficult to print reliably.

Other cool STLs I've found:

• Flat fabric, Hexagonal Fabric by jbergmans
• Escher - A modular 3D Printed Fabric for Fashion construction by Max Samuels
• Penrose Tiles 3D printed fabric by ndv
• Chainmail bracelet - improved by gpvillamil
• Triangular chain necklace by tart2000

Honorable mentions:

• How to 3D print fabric - Step by step by Sara Alvarez
• Wang, Y., Li, L., Hofmann, D. et al. Structured fabrics with tunable mechanical properties. Nature 596, 238–243 (2021). https://doi.org/10.1038/s41586-021-03698-7 (video)
• 3D Printed Dress by Sophy Wong (video)
• Kova by sascha's 3D printed headpieces

I sketched out what my jewelry pieces might look like. The red outlines the areas for 3D printed parts, and the blue lines are parts to be made with macramé.

So the front is a larger chainmail piece, with spikes for aesthetics. It is connected to the back of the head through a printed chain, from which I'd create a faux mohawk with macramé. For the longer hanging pieces on the side of the head, I want to enable swapping of different designs, which would be achieved through snap-fit clips.

In Fusion 360, I started with the snap-fit clips. The clip itself and the notch it snaps onto should have complementing shapes. The tolerance between the clip and the notch also requires tweaking. They need to snap on securely, but still easy to remove and allow smooth movement.

Prototypes #1:

First, I created the notch as a small wedge with divots on its top and bottom. The divots were a bit too close together in my first prototypes and the material broke easily. 1mm worked for my later iteratios. The x-bevel on the side can help the clip open up more easily. For the clips, I created protrusions with the same cone-like geometry that matches the divots. I made four of them, with gaps of +0.2mm, +0.4mm, +0.6mm, and +0.8mm, all of which snapped on very easily.

Prototypes #2:

I added notches onto the clips so that they can snap onto one another to form a longer chain. However, dues to the added materials at the bottom of the clips, the clips became more rigid, especially compared to prototypes #1. None of them bent or opened up sufficiently.

Prototypes #3:

After evaluating previous iterations, I redesigned the clips to allow for more space between the protrusions and the divots. This restored the clips' snap fit function. The +0.8mm gap clip was found to be the best one. The shape of the clips were improved to look a bit more streamlined. For the notches, I tested different fillets. 1.5mm fillet worked best.

Final Design:

I noticed pretty obvious printing artifacts on the clips, due to their thinness. Therefore, the clips were thickened intentionally for the curvier final design. From there I derived an additional ending clip for the bottom of the chain. The notch is redesigned to fit three parallel snap fit chains.

The clips are printed with 0% infill, without support. The notch pieces are printed with 20% infill, without support.

Much inspiration for the headpiece came from the tetrahedron chainmail fabric. I decided to scale up the chainmail in an attempt to improve bed adhesion for the initial layers during the printing process. Instead of small, short tetrahedrons forming surface texture, I envisioned larger spikes lofted from hexagonal links, creating a much more dramatic overall shape, appropriate for a statement headpiece.

Create the hexagon links:

First, I used Circular Pattern to sketch six small circles around the origin, with a distance of 10mm. (Such values should be pre-defined as User Parameters, so that they're easier to adjust later on.) This is going to be the base for the link's posts/stems. I set their radius to be 1mm, resulting in stems with thickness of 2mm. Next, create a construction plane to sketch from, offset from the xy-plane by linkHeight+materialThickness (for mine it's 5mm+1mm), use Circular Pattern again to arrange six small circles but further away from the origin. Loft these six pairs of circles to create link stems. (I had to loft the circles without drawing a hexagon that intersects them, otherwise Fusion refuses to loft cylinders.) Then, on the 1st sketch, draw lines tangent to pairs of circles to create a star shape, and loft the star by materialThickness. On the 2nd sketch, draw concentric hexagons to create a hollow hexagon, and loft it too.

My triangle and circular links were created in a very similar fashion: draw the stem circles and loft; sketch out the base star shape and loft; sketch out the hollow polygon/circle on top and loft.

However, my initial parameters were too small, making the links far too thin. The links were later adjusted by modifying User Parameters or with Extrude.

Create the spikes from the hexagon links:

Using a hexagon link created in the previous step, I filled the top hexagon using Extrude, creating a base. Use a construction plane parallel to the hexagon base, offsetted by spikeHeight (mine is 40mm), to sketch a point for the tip of the spike. Loft the hexagon base to the tip.

In order to create variations, in terms of heights and shapes, I chose to cut from the spike geometry. I used construction planes, offsetted 10mm apart.

The tetrahedrons from the inspo chainmail fabric were completely filled and solid, which contributed to a long print time (~10 hours during my test print). Due to the much larger size of my spike designs, printing entire solid geometries will increase the weight of the pieces significantly, and may cause overhang issues. Therefore, the spikes need to be transformed to shells.

Again, construction planes, using Plane at Angle now. I cut out and removed the "insides" of the spikes, leaving ~1mm thickness as its shell. The tops of the shorter spikes need to be added back. I did that by sketching a hexagon on the top surface and extruding it downwards.

The polygon link models need to be arranged properly, chained together as a single piece, before exporting to print.

Failures:

Frankly, I did not think it through before going in. I drew a cursory sketch of approximately how and where my triangle links should be placed between the hexagon links and went to Fusion. However, there is a reason why the tetrahderon/traingular links were laid "on top of" the hexagon links in the inspo chainmail design. Flipping the geometry (hexagons on top of triangle links) does not work in the same way.

Assembly failure led to a modified circular link design, though at that moment I forgot to consider overhangs from floating circular paths. For a polygon, edges between vertices are straight line segments, which can be suitable for unsupported printing. A circle essentially has infinite vertices and so requires more than the three support points I gave it. I don't think the filament likes to follow curved paths unsupported, either. Unfortunately, I'd sunk a couple of hours going through circular link iterations, before realizing the mistake mid prototype print.

After all that and after re-examining the tetrahedron chainmail, I decided to stick to only hexagons. The smaller hexagon links are taken from the tetrahedron chainmail's STEP file, though in hindsight I could've also just scaled my own links down. They would've been pretty much identical, anyways, moving on.

Figuring out the arrangement:

There were two main factors that affected the process. The overall shape of the headpiece, and the geometry of the links.

Most chainmail STLs I explored show rectangular patterning. However, it was not suitable for mine. I used a combination of circular patterning and mirroring to achieve the result. It is important to ensure that all components are at the same Z value. If not, the slicer might generate bad layers and cause print issues.

Prototyping with only hollow hexagon links took about 4 hours. It's around half the time of a full print. The result already looked really beautiful!

Lastly, I simply added the spike links, and adjusted their positions until they looked great. :D

The full headpiece design took over 8 hours to print. Printed with no infill or support.

For the third and last component of the headpiece, I created a chain design to be printed in place. Each link has a vertical and horizontal loop that interlocks with another link before and/or after it. (Many more fun designs utilizing the same concept can be found here.) I arranged them along a circular path in order to fit the entire chain on the printer. Each link was rotated ~15 degrees. Starting and ending variations were created for assembly with the chainmail parts, though in reality I didn't find them super necessary or helpful. Since the links are so small and delicate, using a clipper to manually open a small gap in the loops works much better.

I added an extra loop on top of the links, as an anchor for potential macramé decorations. Due to time constrains, I wasn't able to actually make the mohawk from my initial sketch, though I do have a couple of ideas to explore, eg. a macramé crown. I'll update on this :)

When printed at the original size of my model, the top loop is pretty fragile. I broke quite a few off when I was removing brims (more on that in Step 9). Therefore, I scaled up its size to 120% in Cura for a second attempt. The result definitely feels stronger.

The final chain is printed with no fill or support.

Due to artifacts on the first print of the centerpiece (details in Step 9), I reprinted it on a raft. It was a really difficult removal and took a lot of effort. In the process, a link broke. Part of the stem under a spike link was missing, so I couldn't just glue it back. I attempted to fix it with a 3D pen from the lab, but drawing with it was much harder than I had expected. The remaining stem was partially melted on accident, which actually helped the repairing. I used nail glue and some tissue paper (as supportive and connective material) to glue the stem back. I took a spare small hexagon link, cut three openings on the edges, and reconnected neighboring links.

To secure the headpiece on one's head, I used square knots to form cords. I made it so that both the front and back attachments are adjustable to fit most wearers. I used staircase knots to finish the adjustable ends.

Here's what it looks like after adding textiles and assembled!

Most if not all chainmail designs have small, delicate details that ask for a lot of precision from the printer. I've encountered more than enough print issues throughout this project. Half of my time was probably spent on troubleshooting print failures, leveling (or more accurately called, tramming) the bed, messing with slicer settings, getting frustrated at the filament, and crying because something went wrong 2 hours into the print.

Poor bed adhesion/de-adhering (given bed is leveled):

Filament de-adhering in the initial layers happened the most often during the project. Yet, it is really deadly to chainmail prints. A piece lifting not only messes up the chain itself, it might also get stuck to neighboring chains, causing issues affecting a much larger area. Due to the small surface area of each chainmail part, the adhesion isn't nearly as strong as larger prints, and any unevenness of the bed shows up more. Additionally, the directional changes of the nozzle movement are quite large and quite often, easily snagging up details.

To counter, there are several things that I've found helpful:

1. SLOW DOWN!: Turn the print speed down to about 65% and even 50% to stop small details from de-adhering. For my chainmail centerpiece (the most difficult print), I set the initial layer print speed to 15mm/s and saw significant improvement.
2. Slow down (part ii): Lowering acceleration and jerk. My Ender's default jerk was 20mm/s, way too high. The print jerk is adjusted to 7mm/s, and initial layer jerk to 4mm/s. I didn't experiment with acceleration much, but theoretically slower acceleration can contribute to higher quality prints.
3. Decrease nozzle gap: More pressure from the nozzle helps stick the print onto the bed better. However, you do need to be careful not to have too much pressure. I've had some pieces got stuck to the nozzle and lifted.
4. Use brims or a raft: The mohawk chain was surprisingly finicky to print. While the main issue was material curling up, using brims or a raft will absolutely help adhere the print to the bed.
5. With brims, it's possible some lines might stick to the print too much, or, not adhere enough. In my mohawk chain's case, even the brims themselves were curling up... not sure what caused it.
6. A raft provides adhesion as well as initial layers' evenness, but the bottom of the print will not be flat and neat. Depending on the air gap between the raft and the print, it'll affect adhesion and removal difficulty. The mohawk chain did great with a raft. However, when I printed the chainmail centerpiece on a raft, it was too difficult to remove. I broke some of the chain in the removal process.
7. Increase the nozzle temperature (filament dependent): With the Overture PLA I was using, its temperature range is labeled as 190 - 220°C. I set the initial layer temperature to be 205 - 210°C and saw better adhesion. However, it does get a bit stringy, but then again, even at 200°C this filament is still pretty stringy.
8. Use glass beds, or add blue/painter's tape to the bed: pretty sure this works, it's just that I don't have either.

Bottom of the print curling (probably filament related):

This is a problem I only saw with the black Overture PLA. You can observe it in the pictures on my snap fit chains. While it might be acceptable for larger prints, curling can cause small chainmail links to quickly de-adhere, and/or propagate problems to higher layers of the print.

I didn't have any curling issue with the Hatchbox PLA filament I was using before. I learned online that curling might be caused by the material cooling too quickly. However, I've already set the bed temperature to 60°C, which is the max bed temp on the filament's labeling. Turning up the bed temp even more didn't seem to help.

As I've mentioned, using brims or a raft can help with this issue, but comes with trade-offs.

I came up with a wack idea to use an iron (for clothes, I have a mini one) to heat and press down on the areas that are curling up. I thought hey maybe it'll help reshape the material and re-stick to the print bed. Nope, didn't work. Might've even caused more curling and de-ahering issues.

Poor Layer adhesion (may or may not be filament?):

During the chainmail centerpiece print (the one without a raft), two of the spikes had significant layer issues towards the top. I think in this case it was also bed unevenness related, but why was it fine in lower layers? In other prints, I also saw different layers gets messed up seemingly randomly and suddenly. Similar instances are observable on the tetrahedron chainmail fabric test, and my hollow hexagon chainmail prototype (on the right).

Not sure what's the underlying issue, and so I don't have a solution for it.