Part Selection/Initial Processing

For the midterm project, our task was to create chess pieces through a molding and casting process. John and I teamed up to attempt to recreate a creeper from Minecraft. We downloaded the .stl file from Thingiverse and began to prep the file to 3D print and CNC.

As you can see, the original creeper has a gap between its “legs”. We determined that this geometry would be particularly difficult to CNC, so we slightly modified the file in Solidworks by simply filling in that gap, effectively creating a rectangular base.

3D Printing

To create the chess piece, we had to first create two half positives. We decided to 3D print the front half of the creeper, as we felt that the face would be hard to make using the CNC. We used Meshmixer to cut the creeper in half, and imported the modified stl file into Solidworks, where we added pegs and a wall bordering the design to aid in making the silicone mold.

CAM and CNC Milling

It was then time to CNC the back half of the creeper. Again, we imported the file into Solidworks in order to add the modifications necessary to create the silicone mold. We then proceeded to Fusion 360, following the provided CAM preparation tutorial. We went to Easel and finalized the file to cut on the Carvey. We gathered some wood pieces from the wood shop; however, we found that the pieces we got were too thick, and when we tried to cut, the drill bit collided with the L-clamp before the cut even started. There were no other pieces that were thick enough for our piece, so we decided to glue two pieces of wood together to achieve our desired thickness. This solved the problem, and we were able to CNC the back half positive for the silicone mold.

 

Negative Molding with Silicone

The process for creating the molds was straightforward and went well for our components. As recommended by Dr. Wettergreen, we built up cardboard walls along the perimeter of both of our positives. The “foot” of the creeper was too close to the top of the mold for both positives, so we needed the wall in order to add enough silicone to cover the top. We then roughly measured the volume inside of the walled positives using a bag of small buttons; the buttons were poured into the positive to the desired level, then poured into a measuring cup. The cup was marked with a sharpie so we knew what level to pour the mold material to. Once the volume was measured, the two-part mold material was mixed together in equal parts and poured slowly over the positives.

Both sides came out very well — the side poured into the CNC wood was much rougher than the 3D printed side, but this was a product of the quality of the positive, rather than the quality of the pouring process. However, a glaring issue we discovered once the molds were removed is that our peg alignment was flipped from what we intended. When creating the 3D models, we mirrored the peg and hole alignment so it would aid in mold alignment when clamped together. However, in the actual molds, the corresponding pegs and holes lined up with each other on both halves, defeating the purpose of making them in the first place. While frustrating, we could easily cut the pegs off with a utility knife and use a different alignment method.

Casting

Once the two negative halves were removed from the 3D print and CNC wood positive, the casting process began. The assignment called for eight identical pawns to be created, with at least two colors shown. We used the standard casting material for the class, a quick cure resin with a short pot life (~5 minutes). In order to begin the process, the two mold halves were first secured together. Initially, we chose to secure the halves with wooden skewers piercing through both sides — piercing the mold material posed a challenge, however, so we moved to nails. While this piercing method worked for alignment, it very quickly became apparent that the friction between the mold and the nails made actually sealing the mold halves against each other difficult. We tried to use rubber bands to “clamp” the two sides together, but our first pour ended up in much of the resin leaking out. For the next attempt, we removed the alignment skewers and added pieces of plywood on either side of the mold to prevent deformation of the soft material. We then again put rubber bands around the entire length of the mold. This technique proved to be successful, with no leaks detected. This method was used for all subsequent pawns.

After we felt comfortable with our casting process, we began to add dyes to our resins in order to make our chess pieces more “creeper”-like. We found the provided dyes to be very potent but experimented enough to get a few different shades of green.

 

Post Processing

The post-processing process for the chess pieces was short, as the molds were meant  to produce near-finalized casted pieces. However, the flat “blocky” geometry of the piece allowed for the opportunity to very easily sand the outer edges to smooth out the lines created from the molding process. Additionally, slight variations in the size of the molds led to some unintended obtrusions which could also be sanded down. We began to use the belt sander to sand down the mold lines, but we quickly found that the sanding process discolored the surfaces of the pieces. Faced with the option of having a totally smooth piece that was discolored or an uneven piece that was all the same color, we chose to stop sanding. We felt that the obvious discoloration was more of a detriment to the aesthetics of the piece than the slight protrusion caused by different-sized molds.

Conclusions and Cost

Overall, we felt that the pieces turned out successfully. If doing this project again, we probably could have come up with a more complex piece, or alternatively created our model completely from scratch in Solidworks (the geometry would not have been that difficult to create). Additionally, we would have spent more time on the CNC portion of the project. While our CNC job worked, it could have been smoother if we had more time, and many more revisions could have been done with different bits and CAM settings. Additionally, the wood positive turned out to be larger than the 3D printed half — an issue that could have been addressed given more trial and error time. The 3D printed portion was the most successful part of the project, as both the positive and negative were very high fidelity and showed details very well. In conclusion, we felt our chess pieces turned out very well; we’d gladly play chess with them any time.