Week14 - Molding and Casting
Overview
For Week 14:
- Design a mold around the stock and tooling that you’ll be using
- Mill it and use it to cast parts
I chose is a ruyi-style “auspicious cloud” (祥云) pattern from Inkscape, plus a circular “Fu” motif for a second mold, so later those parttens could be the decorations for my final project.
This page walks through the full pipeline: 2D pattern → CAD → print → silicone → (planned) casting. The silicone story has a clear order: my 祥云 (cloud) mold was where I made a serious mixing mistake; my Fu mold was where I applied what I learned—with Tim’s reminder—and got a cured rubber mold.
Idea and design
For this week’s molding and casting assignment, I wanted to create something that connects with traditional Chinese culture.
The pattern I chose is the “Ruyi cloud” (如意祥云), a classic decorative motif that symbolizes good fortune, harmony, and the fulfillment of wishes. The word “Ruyi” literally means “as you wish,” and this pattern has been widely used in architecture, ceramics, and traditional ornaments throughout Chinese history.
I’ve always been drawn to these kinds of cultural elements, they carry meanings beyond just visual aesthetics. So instead of designing a purely functional mold, I decided to bring this pattern into my project and explore how digital fabrication can reinterpret traditional designs.
Then I thought, for the Ruyi, I could make a rectangle to hold it, and for the round fu, I can make a circle.
2D pattern design (Inkscape)
In Inkscape I drew a 祥云 (ruyi cloud) pattern: loose, looping curves that read well at a distance but still look intricate up close. Working in 2D first was helpful because I could tweak line weight and spacing before worrying about extrusion heights or printability. For fabrication, clean closed paths matter: anything you plan to extrude or boolean in CAD should be watertight on the drawing side, otherwise you end up fixing gaps later in 3D (spoiler: I still had to fix one).
I used Trace Bitmap directly, but the result doesn't looks good. So I chose to draw myself.
I kept iterating the 如意纹路 linework in Inkscape—different scales on the same page, fill/stroke tweaks—until the shapes felt balanced for mold depth, not just for “looking nice on a sticker.”
3D modeling (Onshape)
I moved the artwork into Onshape and built two separate mold positives, each designed as a “tub” the silicone would pour around.
For the rectangular cloud mold, the footprint started as nested rectangles so the outer walls and inner cavity matched the numbers I wanted—easy to check in the sketch before anything was extruded.
The cloud geometry came from imported DXF/SVG from Inkscape, still in sketch form before extrusion—lots of little handles, which is why merging solids matters later.
For the round Fu piece (molding-test-1), I worked from an imported DXF and used sketch tools like mirror to keep the character centered and symmetric—handy when you’re trying to keep the CAD predictable for slicing.
Mold A — circular “Fu” design
| Feature | Size |
|---|---|
| Outer diameter | 100 mm |
| Inner diameter | 80 mm |
| Outer wall height | 40 mm |
| Inner base thickness | 5 mm |
| Fu character relief height | 15 mm |
The Fu character sits on the inner base as raised geometry. I kept the relief height moderate so the cast part would still demold without locking into undercuts.
Mold B — rectangular box with cloud pattern
| Feature | Size |
|---|---|
| Outer rectangle | 100 × 50 mm |
| Inner cavity | 90 × 40 mm |
| Cloud / pattern relief height | 25 mm |
| Inner base thickness | 10 mm |
| Outer wall height | 60 mm |
The taller walls on Mold B gave extra headroom for pouring silicone above the tallest cloud features, which helps avoid underfill if the pour is a bit shallow.
Why these numbers
I wanted molds that would fit comfortably on our printer bed, use a reasonable volume of silicone, and still feel like “real” objects when cast—not tiny test cubes, but not giant buckets either.
3D printing and iteration
I printed the positives in yellow filament: it matched the fortune color and made fine details and surface defects easier to see.
In Bambu Studio, the round Fu job looked straightforward on the build plate—until the first cloud print reminded me that “looking fine in CAD” is not the same as “prints as one solid.”
What went wrong the first time
About 30% through the print, I noticed a gap between the base and the ruyi cloud pattern. That kind of separation is a red flag: even if the print finishes, you get a weak joint, odd shadows in the silicone, and sometimes resin sneaking into seams. The slicer had also generated more supports than felt necessary, which would mean extra cleanup and more places for silicone to grab.
I stopped the print rather than riding it out—wasting some filament is cheaper than wasting silicone and time on a bad master.
Mid-print, the “forest” of auto supports was a visual hint that the geometry wasn’t telling the slicer a simple, confident story yet.
And on the bed, you can actually see daylight under the cloud relief—exactly the kind of non-merged result I didn’t want to immortalize in rubber.
What I changed
Back in CAD, I merged the surfaces so the cloud pattern and base formed one continuous solid—no floating islands, no hairline split. On the reslice, supports dropped to a saner amount.
The slice preview after the fix was night-and-day: same machine profile, but the model finally read as one printable object, with only a thin budget of tree supports.
Result
The second print completed successfully, with only small burrs left for light sanding or careful trimming. For mold masters, I’m learning to accept that last-mile cleanup is normal; what you want is no hidden voids and no false seams that silicone can reproduce forever.
Mold making (silicone)
I made two silicone cases, in this order: first the rectangular 祥云 (cloud) positive, then the circular Fu positive. They are not interchangeable stories—the cloud pour is my failed chemistry lesson; the Fu pour is the one that actually cured.
Safety first
Before opening any chemicals, I put on a mask and gloves. Silicones and their catalysts are not something you want on your skin, and fine aerosols from sprays aren’t great to breathe. Ventilation matters too—working next to an open window, extraction, or at least airflow helps, especially if you’re using mold release in a spray form.
On pour days I also wore a lab coat when working at the bench—mostly to keep spills off my clothes, but it also reinforces the “slow down and be deliberate” mood.
Mold release — why bother?
I applied mold release spray to the 3D-printed positives. Printed plastics—especially with layer lines—can bond more than you expect to cured silicone. Release agent reduces tear-out when you demold and helps preserve fine detail on both the master and the rubber. Skipping it is a classic way to get a ruined first mold.
Case 1 — 祥云 (cloud) mold: failed silicone attempt
This was the first silicone pour I attempted.
I set up the bench with the rectangular cloud positive, a scale, mixing cups, and the silicone—exactly the kind of layout that should make a pour calm and systematic.
I also used mold release on the rectangular master before bringing silicone anywhere near it.
Here is the critical mistake, stated plainly: I poured silicone base without mixing in the curing agent (Part B / catalyst). I thought I was ready; I was not.
Outcome: the material never solidified. Without the right ratio, the silicone doesn’t fully cross-link—you don’t get a flexible, cured mold you can demold. This wasn’t a cosmetic problem or a “wait longer” problem; it was a process failure on my side.
That outcome was embarrassing, but also extremely educational: rubber is not “pour and hope”—it is a two-part system, period.
What the label was trying to tell me (mixing ratio)
After the cloud attempt, the paperwork on the materials finally got my full attention. The jug in our lab is explicit: 100 parts silicone : 2 parts curing agent, by weight.
We also had smaller Part A / Part B bottles on hand—another reminder that “silicone” is never one anonymous liquid; it’s always paired chemistry.
Why ratios matter
Two-part silicones need a specific balance so chains link evenly. Too little catalyst means no reliable cure (what I lived on the cloud mold). Too much catalyst can change work time, hardness, and shrink. This is weigh-first territory, not eyeball territory.
Sunny helped me get oriented on materials and lab practice; Tim is the person who really drove home that I had skipped the catalyst—and what to do differently on the next pour.
Case 2 — circular “Fu” mold: successful silicone attempt
For the Fu positive, I did not repeat the same mental shortcut.
Tim reminded me to actually mix Part A with curing agent the way the datasheet and label specify. I measured by weight on the scale, mixed thoroughly, then poured.
Result: this batch cured. I finally had a solid silicone mold around the Fu master—something I could treat as a real mold-making outcome, not a puddle Lesson Zero.
This is the contrast I want to remember: same lab and the same general materials workflow, but a totally different outcome because the second time I respected both parts of the rubber system.
Casting preparation (AB resin)
My plan for the final positives is two-part AB casting resin poured into the cured silicone negatives. At the time of writing, I discovered my AB resin was the wrong type for this workflow (different resins have different viscosities, cure profiles, and compatibility with the mold material—using the wrong one risks incomplete cure, heat issues, or a part that won’t release cleanly).
So casting is on hold until the correct material arrives. I’m treating that as part of the process: molding and casting depends on supply-chain and spec hygiene, not only hand skills.
Problems and fixes (quick recap)
These rows are intentionally literal: the 祥云 silicone try failed because of chemistry; the Fu silicone try succeeded because I mixed correctly after Tim’s reminder.
| Problem | Fix | Lesson |
|---|---|---|
| Gap between base and cloud at ~30% print | Stopped print; merged surfaces in CAD; resliced | Fix the model, not just the slicer |
| Excess supports | Better geometry + reslice | Cleaner masters = faster finishing |
| 祥云 mold: poured silicone without mixing curing agent | No fix in the cup—the rubber never solidified; treat as scrap and restart protocol on the next mold | Catalyst isn’t optional; read ratio before pouring |
| Fu mold: second pour with correct Part A + catalyst (by weight) | Tim reminded me to mix properly; batch cured | Checklist + scale beats confidence |
| Wrong AB resin on hand | Wait for correct resin; don’t “force” a mismatch | Read product specs before committing material |
Final result (so far)
I don’t have final cast resin parts documented here yet—the resin spec mismatch paused that last hop.
On the silicone side, the scorecard is very specific: the 祥云 rectangular pour never became a solid mold because I skipped the curing agent; the Fu circular pour did cure, after I mixed Part A and catalyst the way Tim insisted I should.
So the “final result” for this week’s molding chapter is really: one failed rubber lesson and one successful cured mold, not two perfect molds.
Common pitfalls (things I’m watching next time)
- Undercuts in the positive that lock the silicone or the cast part.
- Trapped air in deep relief—pour technique and vent paths matter.
- Ignoring draft on vertical walls—demold fights are real.
- Skipping release agent to “save time.”
- Guessing mix ratios instead of weighing.
Reflection
This week felt less like “press print and you’re done” and more like closing loops—except silicone doesn’t forgive skipped steps.
The emotional arc is almost funny in retrospect: I was so focused on looking prepared (bench layout, gloves, pretty pour photos) that I almost missed the actual rule: pourable Part A is not a mold until Part B is in the cup at the right ratio. On the 祥云 mold, reality delivered that feedback in the meanest honest way: a liquid that would not turn into rubber.
Tim’s correction mattered because it happened before I wasted another ambiguous batch—I went into the Fu pour knowing exactly what “correct” looked like on the scale. That second mold curing felt less like luck and more like I had finally joined the same conversation as the datasheet.
I’m looking forward to finishing the story once the right resin shows up: resin casting on top of a mold that actually set.
Status
- Design / CAD: Mold A (Fu) and Mold B (祥云) modeled in Onshape with dimensions as listed above.
- 3D-printed masters: Re-printed successfully after fixing the base–pattern gap; light burrs only.
- Silicone — 祥云 (rectangular cloud): Failed (poured without properly mixing curing agent); silicone never solidified; process reset mentally before the next mold.
- Silicone — Fu (circular): Succeeded after Tim reminded me to mix Part A + catalyst by weight; rubber cured and the mold is usable for the next step.
- Casting: Pending suitable AB resin; wrong material held up this step—not a failure of the mold idea, just a specs mismatch.