For our final project, Anas, Matthew, and I created a reusable mold and casted 8 identical chess pieces from it. It was a great way to utilize several different techniques that we had learned including: CAD/CAM techniques, 3D printing, molding, and casting. Overall, we really enjoyed the process and were happy with our final pieces. This blog post will cover our process, areas of improvement, and a cost analysis. 

Our Process

CAD Model Creation for 3D Printing

First, the process of simplifying the chosen chess piece on Meshmixer was completed without issue/complication. Our spiral/helical geometry did require more than 50% compression of triangular faces but the end preservation of the geometry was good as seen in our Solidworks and later Fusion modelling. 

Moving to Solidworks, we wanted to customize our piece somewhat (as well as make it truly CNC Mill-viable) by adding a wall to cover the hole/channel on the top of our queen piece. This was done by drawing a reference plane based on three points along the top edge of the piece. This allowed for the shape to be manually drawn in by tracing that edge, projecting each relevant triangle edge, sacrificing some smoothness and a lot of time to ensure a solid body would still be present after the extrusion.

A similar process was needed regardless in order to create the back wall for each half of the mold. Again, creation of a reference plane allowed for a simple extrusion. Of note, a small extrusion was also made towards the pieces from its back since there was some slight height discrepancy between points on the back. Evidence of this can be seen on the edge between the piece and back wall highlighted below. 

The channel to prevent air pockets, the pegs/holes needed for the mold halves to hold securely, and the walls were each simple geometries and presented no issues. 

Issues with .stl creation

We had an interesting issue arise in the process of actually 3D printing the mold negatives. On exporting the parts from Solidworks, we believe there was an issue in recognizing the CNC-friendly plane we created. It is unclear whether this is due to the very fine detail in the edges of the sketch or caused by Solidworks attempting to compress the file given the vast number of facets in the original piece geometry. This resulted in the feature only being rendered/recognized close to the outer edge of the piece and was quite apparent on the final pieces. 

We did not stop and return to the CAD file for a number of reasons. First, we initially assumed that this was actually an issue with rendering on 3Dprinteros. It seemed more likely to us that the browser-based software was having visual issues rather than Solidworks exporting poorly. Second and more importantly, we decided that any resulting extra material on the edge could be easily removed, whether on the mold or the final pieces. 

CAM Model For (Potential) CNC Milling

The creation of the CAM toolpath was largely uneventful, thanks to the step-by-step instructions provided. The only major highlights are a large number of likely erroneous collisions/errors detected in the final, parallel pass. While we thought this could be due to minute overhangs created by the helical geometry that could collide with higher segments of the drill bit, but it is more likely that this was due to a learning curve with Fusion, where the physical ordering of CAM processes in the directory would cause issues (running the parallel cut before the adaptive cut would certainly cause conflicts). As exported in our g-code, no collisions were found by observing the simulation.

This process did highlight some interesting differences between Solidworks and Fusion 360. As an example, in an attempt to fix some conflicts/collisions that were present in the initial pass, removing the walls on Fusion was more difficult that it may’ve been on Solidworks. Fusion attempts to be more fool-proof and prevents cuts that could have created under-defined geometries. This may have been due to working on a slightly compressed part (.stl from Solidworks) or the minutiae of working with the created solid body. Overall, both the versatility (CAD and CAM capabilities) and the drawbacks of Fusion were made readily apparent in this project. 

 

Creation of the Mold

After we had our 3D printed mold created, our next step was to create a mold from the 3d print we had created so that we can repeatedly cast more pieces from the negative mold we created. 

 

 

 

 

 

 

The mold came out pretty nicely, and as you can see here we incorporated a channel for the air to pass through so that we could minimize the amount of air bubbles and increase the consistency and the quality of the pieces produced. We found that while the pegs we had created fit very well it would have been better with 4 instead of 2 for added stability when pressed together. Other than that we were very pleased with our mold.

 

Casting Our Pieces

 

The first 4 of the pieces we cast were uncolored and they came out quite nice. We did notice that the pieces we were casting had the same imperfections as the mold had, and they were perfectly transferred to the pieces. As you can see in the picture below, the edges of the queen are frayed from the molding and required touching up during the post production process. We also saw that as the pieces were being cast, the described edge issue worsened as the mold began to break down. In the picture on the right, there was a small crack that appeared in the crown of the piece because of the imperfections that were beginning to appear in our mold, which we will talk about later in the report. 

After these pieces were created, we started adding oil paints into the mixture to give our pieces some color, and we found this to work pretty well when we added in a small amount of paint and mixed it. When it was solid colors we found that it simply lightened up in color, which we liked as it gave it a pastel color look. We also tried experimenting with multiple colors, with two different methods for the last 2 chess pieces. With the very colorful one, we mixed in the colors blue, yellow, and red one after the other and then we gave it a quick stir before the mixture set it, which gave it a far more blended and colorful look. In our last piece, we decided to layer on the pieces with red, blue, and yellow and did not give it a stir before letting it rest and set, which gave it a much more distinct color separation contrary to the previous piece. 

Post-Processing

Once we had created our 8 chess molds, we were ready to post-process them. Since we were really happy with the colors and texture of the chess pieces, we decided not to spray paint or clear coat them in any way. However, we did use a high-grit sandpaper to sand down some of the extra material that formed as a result of the two molds being misaligned slightly. Additionally, we used the sandpaper to sand down the very bottom of our chess pieces as some of our pieces would end up with a raised bottom due to adding too much casting liquid. Sanding down the bottom also helped eliminate some of the glossy edge the casting liquid produced. Once we had sanded down our pieces, we were finished with our chess pieces.

 

Areas of Improvement

Overall, we are really happy with how our chess pieces came out. There were some minor issues such as the fact that the two mold pieces did not always line up perfectly, leading to some extra material on the chess pieces. This could have been solved by adding more pegs and holes to our model so we would have more areas to align the molds. However, we were mostly able to align it ourselves and it was not a huge issue overall. 

Our biggest area of improvement, however, was an issue that stemmed from the original file that we created. Since our .stl file from Thingiverse had a slight dip at the top sphere, we added a flat plane to the top of our chess piece so that it would be physically possible to 3D print and CNC our chess piece. When we 3D printed our two molds, we noticed that there was a printing error on the top of our chess piece since the plane that we added was not properly aligned. As such, when we created our molds, the printing error transferred into the silicone molds and the top of the chess pieces were slightly deformed and not properly created.

Whenever we casted our 3D printed pieces, the top plane around the sphere did not properly form and was choppy and had gaps. Additionally, since the top plane did not properly form, the mold would worsen every time we casted a chess piece with it. We should have verified the plane on the top of our CAD model (or removed it given the only theoretical need for CNC-milling) so that everything would have formed properly.

 

 

 

 

 

Cost Analysis

For this cost analysis, we will look into the cost of: SolidWorks, Fusion360, a Prusa 3D printer, 3D printer filament, silicone rubber molding material, liquid plastic casting material, and the cost of our time.

The student edition for SolidWorks costs $99 per year, while Fusion360 costs $495 per year. A Prusa i3 MK3 3D Printer costs $999 and the 3D printer filament costs around $2.25 according to 3DprinterOS. For the silicone rubber mold, I found the 8 lb. kit that we used for $180 and, with a generous assumption, we can say that we utilized around $15 worth of material from that. For the liquid plastic cast, I found the 15.2 lb. kit that we used for $142.88 and, with a generous assumption, we can say that we utilized around $25 worth of material from that. Finally, we each spent around 8 hours working on this project (including class time) and, assuming an average pay of $12 per hour, this project was worth around $288 of our time.

Overall, the project cost us around $1923.25, but, if we remove software and machine costs, it cost around $330.25.