Week 10

Goal

For this week, I wanted to mold a boot. My final project will walk around on many feet, which all need to have traction. Molding is a great way to use materials with high traction and it is a great way to make 100 of something. So, this week was a good opportunity to prototype for my final project.

I decided to do a two-stage mold just to practice, even though my end product was supposed to be soft.

Design

I decided to use Blender to design my boot, because I have never used Blender. The boot I designed is below:

Figure 1

I figured out how to sculpt, but in the interest of time, I ended up keeping the two starting shapes for the design. I uploaded the design here.

I exported the STL file and imported it into Fusion, where I designed a two-part mold. (It quickly became apparent that I had not exported with enough shapes for the cylinder, which I should have addressed earlier.)

Figure 2

Figure 3

The “second” mold is a negative of the boot, and includes walls for registration (instead of bumps) and a hole to help separate the pieces. It is shown in Figures 2-3. It is also accessible at https://a360.co/48CY09l and embedded below (so long as autodesk360.com continues to host it):

Figure 4

Figure 5

The “first” mold is a negative of the second mold. I moved and resized the pour hole and added an air hole, which I had omitted in the second mold originally (in the interest of time, I did not got back and edit the first CAD). I made each side the same size as the provided wax stock. They are shown in Figures 4-5. They are also accessible at https://a360.co/4pnFSrg and https://a360.co/4rG3yso and embedded below (so long as autodesk360.com continues to host them):

Making the First Molds

I made two versions of the first mold: one out of wax on the Genmitsu, and one out of resin on the Formlabs SLA printer.

The wax mold presented significantly more challenges, and I made a few mistakes. I had not used the Genmitsu before, and I found that learning to use it made me understand G-code a lot better. However, it did have a few quirks. Most notable was that its travel in Z was severely limited, but I did not find a way to make Fusion aware of that when generating toolpaths. So, some manual tweaks of the G-code were required. For every missed tweak, the machine would hit a limit switch, and require repeated resets and jogs to get it out of that state.

Another challenge for me were the toolpaths. The coarse run was straightforward, but I did not do a great job with the fine run. I suspect I should have split it up by areas. This is definitely something I hope to get better at in the future.

Figure 6

Most critically, however, I was not able to drill the walls of the larger second mold. When I created the corresponding first mold, the walls created high-aspect-ratio trenches. When the machine got to them, the collar prevented the cuts from completing (Figure 6). I decided not to fix this, since even small walls allowed for registration, but this is something else I will improve the next time I do this. I simply stopped the run and fixed the first occurrence with some tape.

Also, after one of the Z-axis stalls, the machine suddenly moved in the X direction and took a small chunk out of the mold. It’s unclear to me why it did this, but it’s a good reminder to never make assumptions about where the machine is going. I stopped and reset it after which it continued normally.

Figure 7

The final product is shown in Figure 7.

Figure 8

The SLA printing was painless and high-quality (despite using the Fast material). Alan walked me through it on the Formlabs printers in 015. The mistake I made there was printing the entire design, which I had made to include the wax stock to make toolpath planning easier for the milled version. Next time, I will not only not do that, but modify the mold to further reduce material.

Figure 9

The final product is shown in Figure 9. The guck in the picture is from making the second mold - the print itself turned out pristine!

Molding the Second Molds

For each of the two first molds, I made two sets of second molds using two types of silicone rubber: OOMOO 30 and the “food-safe stuff”. For both materials, I used vacuum to remove all bubbles each time I made a new batch.

Figure 10

Figure 11

Figure 12

First, I used OOMOO in the wax molds (while the SLA molds were printing). That worked great!

Figure 13

Figure 14

Figure 15

Then, I mixed the food-safe stuff (Figure 13) by weight and made a set of second molds with that (while simultaneously molding boots in the first set; more on that later). Simultaneously, I molded OOMOO in the SLA print. Both turned out well (Figure 15). The SLA version had the nice walls, as expected.

Finally, I molded the food-safe stuff in the SLA molds. (I also re-molded OOMOO in the wax, because my first set of boots had made a mess, and the first mold was idle.) The food-safe stuff did not cure in the SLA molds. I left it in overnight (much longer than necessary, but when I checked in the morning, it was a gooey mess.

Figure 16

In summary, out of the four combinations, three worked well (or, as well as the original mold allowed), but the food-safe stuff did not cure properly in the Fast SLA. All five attempts are shown in Figure 16:

OOMOO in wax
Food-safe stuff in wax
OOMOO in SLA
Food-safe stuff in SLA
OOMOO in wax again

Molding the Boots

Pipelined with molding the second molds, I was molding boots in the second molds that were already molded. This stage was a lot less successful.

Figure 17

Based on faulty intuition, I started by molding in the opposite material. The food-safe stuff boots in OOMOO cured for the same amount of time as the food-safe stuff in wax (Figure 14), but they did not fully cure (Figure 17).

Figure 18

I then molded the opposite material in each second mold except for the food-safe stuff SLA one. The OOMOO-in-food-safe-stuff boot was almost okay, but also didn’t cure completely. The results are show in Figure 18:

Food-safe stuff in OOMOO in wax
Food-safe stuff in OOMOO in wax
Food-safe stuff in OOMOO in SLA
OOMOO in food-safe stuff in wax

(One of the attempts had some other issues as well.)

Figure 19

After these failed attempts, Kat suggested I do like-in-like and use mold release. I washed two of the OOMOO’s and both food-safe-stuff molds, and lined them with SC Johnson Paste Wax. I then molded OOMOO in the OOMOO ones and food-safe stuff in the food-safe-stuff ones. The food-safe-stuff mold that had been made from the SLA mold turned into a solid block of silicone, but the other three turned out okay (Figure 19).

Final Thoughts

This week was simultaneously my least successful week, and the week in which I learned the most (perhaps those go hand-in-hand). I learned:

how to create 3D toolpaths
how to use a table-top mill
how to use an SLA printer
how to mold

I didn’t get to do any casting, but I imagine a lot of my learning in molding will transfer. I knew going in that a two-stage mold was unnecessary for my final product, but I learned that it’s actually significantly more difficult than a one-stage mold because of cure inhibition.

As is the case with most weeks, there is a lot more I would have tried with more time. This week was especially tight with machine week ramping up in parallel. If I were to redo this experiment, I would go in with a much clearer plan. But, my main goal was to learn how to mold for my final project. The good news is that the first stage mold worked well, so I am confident I will be able to execute successfully. Based on my experience, I plan to use the SLA printer with OOMOO 30.

Sources

Blender file, Boot: https://gitlab.cba.mit.edu/dimitar/kitchen-sink/-/blob/main/boot/boot.blend
CAD, “Second” mold: https://a360.co/48CY09l
CAD, “First” mold, part 1: https://a360.co/4pnFSrg
CAD, “First” mold, part 2: https://a360.co/4rG3yso