Final: 3D-Printed Sandals

I designed a pair of 3D-printed sandals in Fusion 360 and printed it with TPU 95A in the Bambu printer. The TPU is to be soft, strong, and extremely comfortable as a base for my foot.

I really enjoyed reading about wearable designs/tools for wearable/body-mapped design (e.g. Tactum) and was impressed by the ability of Fusion 360's generative technology to put pieces together in unique ways after creating a generative toolbox with the software, so I wanted to wearable object for my final project with a unique design. I occasionally am averse to wearing sandals in the house since sandals can be uncomfortable at times due to the texture of the sandals. Both my brother and I have size 13 feet (U.S. Men's sizing) for our sneakers/boots, but size 12 for sandals/Crocs/dress shoes/bowling shoes, and even then it's hard to find shoes that are fit well (many are too tight or too loose). Consequently, I wanted to create a pair of sandals that was (a) comfortable and (b) was mapped specifically to the dimensions of my feet. My work in this project is somewhat based off this tutorial, although my mapping and groove design is unique and I printed the sandals in a significantly different way (e.g. tree supports, reduced infill density, gyroid infill) for added prettiness/sturdiness/comfort based on the results of multiple tests. Furthermore, I added my last name to the straps of the shoes via a Boolean operation. I also printed multiple shoes (left and right), while they printed one.

I researched different avenues by which people have 3D printed sandals, and I found that a popular way to do so was by printing with TPU on a Bambu printer, and that TPU would provide a bendable, sturdy, and soft material for the shoe (with the downside being that TPU is stringy). However, there are many different grades of TPU, and the lab had TPU 95A, so I printed a mini-rectangular sole (with rectilinear infill) that I designed myself in Fusion 360 to test bendability/softness and a spring I found with Alejandro on a test print website to test sturdiness/stringiness. I found that the TPU we had was slightly stringy, but it was extremely bendable and sturdy. It did not appear to be soft, but it felt like it had the potential to be, so I decided to stick with it until I had a larger test print.

I began by finding an initial sketch of my feet size and roughly tracing it/building a sandal to it in Fusion 360 just with the basic sketch and extrude tools. My main goal with this was to experiment with how a potential sandal would look like before I tried out any tutorials. I discovered that it was very difficult to create a well-mapped, pretty shoe by just sketching and extruding based on a rough sketch.

I then researched generative slippers, and I came across generative high heels. I fell in love with the abstract, pretty and sturdy nature of the geometry that was created within these shoes, and I was moved to try to generate similar sandals like these.

I explored generative constructions for sandals, and I tried a couple of different rough designs out, but I was not able to generate a sandal design that was fully connected and wearable. I decided that the first step would likely be actually mapping my foot out and creating a non-generative sandal that was wearable/comfortable/fit properly, and then based on that, I could add generative features to it so I don't run into the issue of my sandal not fitting.

Afterward, I had my peer review session with Payton. We discussed my exploratory design, my initial foot sketch, my generative design experiments, and my plans for creating 3D-printed slippers. The main points regarding the non-generative design were to make bottom thick enough to be sturdy and beware of strap generation since it can be funky to print on the sandal or attach it later on. I also had brought up the idea of using hot glue to help remove supports/attach the strap, but she brought up that hot glue isn't too strong and won't be permanent for strap attachment. Furthermore, I also expressed interest in potentially painting the sandals after the course, and she brought up the really important point that I could potentially paint the straps and the top of the base since that would be aesthetic, but beware of friction removing the paint. Regardless, even if I don't end up painting the shoe, I still think the solid color TPU looks pretty. We also discussed potential avenues for generative sandals, and she brought up that Fusion 360 can do great branch-like design, so I was hoping to do something similar if I got the chance to make generative branch-like outlines for my sandals. My full peer review notes can be found here.

I decided to move forward with designing a non-generative shoe since that would give me the best chance of achieving my initial goal of a wearable, comfortable, and well-fitting shoe. If there was time at the end of the course, I would hope to use this design to be the base of a generative design.

I initially added more features (strap, closed all gaps, etc.) to my exploratory sandals and wanted to print it as a test print, but it would have taken 1-2 days to print, and Jennifer mentioned that the length of time needed for that kind of test print wasn't worth it. Furthermore, because my designs were loosely to my foot, I decided to give just the sole as a test print to see how comfortable my sandals could be.

Before we printed the material, we reduced the sparse infill density to 5% and changed the sparse infill pattern from rectilinear to gyroid to allow for more bendability. The end result was an extremely strong yet cushiony sole, and I loved the outcome. It was much softer than the original test sole with the rectilinear infill and 20% density. The print barely fit on the printer.

I then began to construct my mapped sandals. I measured my right foot and found that it was 27 cm long and 9 cm wide. I was slightly confused by this measurement, as that would imply a shoe size of around 9-10 U.S. men's, but the last time I tried to fit into sizes lower than 12, the shoe is really tight. I eventually realized later on this measurement happened because I forgot to stretch my toes out/wear socks (hence resulting in a shorter length), so I had to eventually increase the length of the shoe later on (see Step 14). I took the measurements and a shoe canvas from the tutorial and overlaid them perpendicularly so I could begin designing the shoe.

I designed the sole of the shoe by creating a form outline of faces based on the shoe canvas and duplicating it across the shoe by the width of my foot plus 0.5 cm for extra space. The goal would be that I would eventually use the edges of the faced to map the shoe to my foot. After duplicating, I bridged the two T-spline bodies together with creases and filled any holes remaining to create a solid sole.

I mapped the sole onto my foot by editing the form of each edge until the flat sole aligned with my foot (plus some extra space as a buffer). Afterward, I analyzed my feet, and modified the depth of the sole (by editing the faces) based on the high and low points of my own feet. This overall process was relatively small, yet crucial as it would determine overall shape of my shoe.

I chose to add the strap directly on the design in order to prevent assembly issues later on. I created a form with faces based on the show canvas, duplicated it across the width of my foot + 0.5 cm, and bridged them (with creases off this time). At this point, I had four faces, and I spent a significant amount of time trying to figure out how to map the strap directly onto the sandal (I went back and checked the tutorial, and it did not cover this). Eventually, I realized I could add more points to split edges or add more faces themselves, and because of the complex shape of my foot, I ended up having a lot of edges to map.

The overall mapping process of the design strap took between 1-2 hours as it was difficult to get it exactly right from the top, bottom, left, and right views. I had actually finished it and cut the grooves (Step 12) before realizing that the edges were floating, so I went back in the timeline and spent 30-45 minutes getting the edges exactly right.

I decided that I wanted each of the grooves (which provides the grip) of my shoe to be unique and angled to my liking. This is because many shoes I have had in the past have had terrible grip, so I wanted my shoe to have somewhat varying/less uniform grip to experiment with what that would do to my grip. Consequently, I manually drew tiny triangles across the length of my sole sketch at angles that I thought fit together well. I then extruded them across the bottom of my sole and cut through it to create the grooves.

I combined the strap and the sole into one body and exported it to the Bambu slicer. I printed it with gyroid infill at 5% density, and I chose to add tree supports in order to limit the amount of support cutting I would have to do at the end of the print. TPU supports, from my research, appear difficult to remove, so I was hoping that the branches would make it easier/less targets to remove. Also, printing with tree supports would result in a 5 hr faster print, so I went with that. Furthermore, after consulting with Jennifer, we made the retraction length 2 mm and the retraction speed 20 mm/s for a better/cleaner print and to hopefully avoid some stringiness. The print time was around 15 hours.

Removing the supports was incredibly easy due to my support design as I could just pull the trees off within a couple of minutes, so I did not have to spend much time removing them as the guy in the tutorial did. Furthermore, my shoes were not too stringy, and it was relatively easy to remove the few stringy parts. One side of the shoe (the outer) was more stringy than the other, but that was because the support base was attached to it as I had printed the shoe on that side. I tested my initial shoe print live in front of the class during my presentation slot. The shoe was sturdy and soft, but it could be improved for both comfort and sturdiness. Based on our discussion, I decided to implement the following changes: (a) Increase the infill density from 5% to 12% since the shoe base appeared to be slightly collapsing when I walked, so the base needed to be firmer (b) Edit the strap design to be perfectly attached to the base since the outer right mapping was not correct and a tiny portion was not attached to the base (which would have led to an eventual wear and tear of the strap) (c) Print my new shoes with the inner side as the base so as to hide the stringy side from viewers (d) Add my name to the strap of the design because the strap was a relatively unused design space, and I wanted to create an object using more of the features of Fusion 360 and 3D printing (e) Decrease the front part of the strap's height by 15 mm and the back part of the strap's height by 5 mm since the shoe was coming out as I walked, and I wanted my shoe strap to prevent that (the measurements I obtained from measuring the difference between my feet and the strap and subtracting a couple millimeters from some foot room) (f) Increase the overall length of the shoe by 10 mm since the initial design was too small.

Fixing changes (a) and (c) were relatively quick since I just needed to adjust the Bambu Slicer settings. I realized the reason I had to make change (b) — as I had previously thought my mapping was perfect before I printed the shoe — was because I had missed mapping a line from the Back view. Making the change was relatively easy because I was so familiar with the mapping from earlier. I fixed (e) and (f) by traversing back in the timeline and making the edits in the appropriate form creation. Implementing change (d) — the name — took the longest time because I wanted the name to fit perfectly on my 2 mm thick strap without extruding past it, but also touch the bottom of the strap so that it would be as sturdy as possible. I ended up typing my last name, extruding it, and fitting it over the strap. I then used a Boolean operation to remove the extruded name from anywhere else besides the inside of the strap and the top. When I took my updated design back to the Bambu Slicer, I found that the Slicer had added support generation across the name to ensure its stability. My new print time was now around 17 and a half hours (2 and a half hours increase)

Since my feet have similar shapes, I decided to mirror the right shoe to get the left shoe body to print. I mirrored it by following Autodesk's instructions for mirroring a body, but I realized that while the shoe base and strap mirrored fine, the name was mirrored backwards. After playing repeatedly with the software, I discovered through trial and error that in order to mirror the whole design completely, I would need to traverse back in the timeline and mirror the base and strap separately. Then I would need to duplicate (not mirror) the name and recombine all the bodies into one. Following those steps, I was able to successfully mirror the shoe.

Before I gave my shoes for printing, I wanted to make one last attempt at generatively creating my shoes by playing with the final shoe design that I had created. I separated the name, the strap, and the base, and I assigned them to be "preserve geometry." I then created an obstacle geometry between them, and I gave the design for outcome generation. This was my most successful generative design so far, yet it was also not printable as the mapping/dimensions were now off. I think that in the future, if I was to explore further into generative design for footwear, I would focus on small generative aspects (the outline/form) and create perfectly mapped obstacle geometry and starter parts, but even so, it is likely the shoe would not be a completely well-mapped piece of footwear.

Each shoe print took around 17 hours, but the final sandals came out well. They were much more firmer (yet still comfortable and soft) due to my added infill density, and the sandals were mapped better (no outside strap piece sticking out, the shoe was longer, and the reduced strap height prevented the sandals from falling off). All changes were implemented successfully, and the name was sturdy.

If I had time to do another round of printing, I would increase the width of the sandal by 10 mm and slightly increase the length by 5 mm so the sandal isn't as tight fitting, but these sandals are much more improved compared to my previous print. Removing the main supports from the name was easy as well! There were tiny pieces of supports between letters that could not be removed from the name without me rising damaging the sandals, so I left them there.

Changing print settings can have a significant effect on the print time (e.g. support vs. no support, infill density, etc.) and quality of the print (e.g. gyroid vs rectilinear infill type). I was really surprised by the softness of the TPU with gyroid infill, and I am curious if that was primarily a product of the infill or by nature of printing a large object (since my mini-rectangle sole felt harder). I am also curious as to how well the strap, which is 2 mm thick, will hold up over time after making consistent contact with the upper part of my feet when I walk.

The process of iteratively testing TPU, the base, and the sandals significantly helped me improve the quality of my design. Given more time, I would conduct longer tests on the sandals, including wearing them for multiple days outside and testing how the sandal holds up when my feet move in various ways (e.g. if I slip, how well does the sandal hold) in order to create even stronger sandals. This process is common in product design, and 3D-printed footwear could be an interesting product line to make due to the limited costs to design the footwear.

I would love to explore how different grades of TPU would affect the makeup of the sandals, and if I would need to alter print settings like infill density in order to account for the differences. I would also love to conduct further testing on my shoes to improve their stability and comfort.

I would love to experiment with stain and dye testing on the sandals as well. Adding color to the sandals would make them extremely pretty, but I'm unsure as to how long the colors would last for, so I would love to test different types of dyes/stains and how long they last on different environments.

Converting my sandals to a generative shoe would also be a really beautiful project to take on, especially if I am able to create designs mapped to my feet similar to the generative high heels in Step 4. This will take much more time and a greater understanding of generative design, so I plan to spend time researching this avenue in the future! When I successfully design them, I would love to come back in the Fall, print them, and stain them with pretty colors!