Inspiration and Model

For this project, we were tasked with creating two chess pieces using techniques for 3D design.

The project involved making casting molds with 3D printing and CNC machining processes, so we knew we were working within the constraints of the CNC machine which can only cut from above. This proved to be helpful when choosing our chess piece from the internet since we had to make sure the geometry was not too complex.

Our team wanted to go with something simple and recognizable, so we chose the Chrysler Building in New York. We agreed it looks aesthetic, the size works as a chess piece, and there is no complex internal geometry that will create an issue. The only red flag we saw was that the building has a lot of flat edges which, at least with our method of casting from two half-molds, might create very visible and messy seams.

 

Processing in Meshmixer and Solidworks

After downloading the STL file, we tried 3d printing it to make sure it looked like it could be molded. We verified this and moved the file into Meshmixer to cut it in half. The shape has kind of a weird diagonal side, so we decided to try doing that side on that CNC. We saved each side and imported them both into Solidworks. Here, we added a base beneath the part to facilitate mold making, along with some other holes and pegs to help with casting. It was a little tricky to work with the part since it was a graphics body, but we got something that looked good. We sent the straight half to 3dprinterOS and the diagonal to Fusion 360.

 

3d Printing

The part was processed in 3dprinterOS easily. It took an hour and a half to print in PLA and looks great. Normally when making a mold of a 3d printed part we would’ve sanded it down to hide the layers, but we wanted it to be visible for this project.

 

Fusion 360 CAM

This was a challenging step for us since neither of us had used Fusion 360 before. After some difficulty, we got the part all set up and oriented. We made the different cuts and added the tool library to create our toolpath. We did a rough cut first that took off most of the material with an ⅛” bit, and then a finishing cut with a 1/16” bit that helped define the shape more. There was some difficulty when we noticed in the simulation where the tool was hitting uncut stock during the finishing cut. We had to play around with the size of the tool boundary to get it to cut correctly, which took a while but eventually worked. There were other problems we discovered when cutting, but I’ll get to that later. With a good-looking simulation, we processed the part in Easel and got ready to cut. 

 

Cutting on Carvey

Jorge cut our stock, and we placed our first try on the Carvey CNC mill. It started pretty well, and successfully cut all of the features, except for the holes in the base. We spent a while in Fusion 360 trying to figure out why this happens, but never found the cause. We weren’t the only team to encounter this issue as well. Next, we paused the cut, swapped the bits, and were about to hit start when we realized we had a problem. The machine wouldn’t re-zero with the new bit, so all of the depths of the cuts would be off. We tried a somewhat creative solution, and just restarted the cut at 200% speed with the 1/16” bit. It spent another half hour passing over everything that had already been cut, and finally started with the finishing pass. However, everything it was cutting was about 0.25” off in the x-direction. All it did was deform our shape, and we were forced to call our first attempt a failure.

We went back to Fusion 360 and made a few changes. The first thing was the discovery that the x dimension for the stock was different between the two different cuts. This meant that on the second one, the CNC thought the part was in a different place entirely, despite having just cut it. We assumed at first that this was just a user error, but it seems like many other teams had the exact same issue and some of them just skipped the finishing cut because of it. I wonder if they had the same error in the stock dimension and if it could have been some kind of default setting. The next thing we changed was including the missing holes in the bounding box for the finishing cut. If the rough cut couldn’t do it, we could just make them with the 1/16” bit. However, it does this by cutting the hole out bit by bit with every pass over it, a method that stressed us out quite a bit and inspired us to turn the feed rate down to 50%. Finally, we realized we could just post process the two cuts separately and load them in the same Easel project. This allowed us to replace the bit after the first cut without needing to start over.

We sent the new Gcode back to Easel and tried again. This time it all went according to plan. It took about an hour and a half to cut it out, and the detail was great, though admittedly worse than the 3d printer. 

Moldmaking

With both of our positives in hand, we could finally make our molds. We made mold boxes using cardboard and hot glue, and we leak checked them with some water. It was a good thing we did that since the box for the 3d printed piece leaked a lot. We mixed parts A and B and poured both molds at the same time. We were careful to pour slowly in the corner and let the material flow over the part, and then we banged both of them on the table to try to get all of the air bubbles out. 

The 3d printed mold turned out really well. There were only a few noticeable air bubbles, and you could see every layer from the print. The mold for the wood part was not quite as nice. There were lots of tiny air bubbles that were noticeable in our final casts. The mold should’ve been shaken and slammed just a bit more to make sure all those bubbles got out. Still, the mold halves fit together, so we went forward with casting. 

 

Casting

We connected the two mold halves together by aligning the holes and pressing the halves together using cardboard and rubber bands. Next, we took some EasyFlow plastic and mixed 10mL of each part since we knew our piece was about 20mL. After pouring the mix slowly into the mold, we waited for 30 minutes and took the piece out. On some iterations, we couldn’t get the two halves to align perfectly, so we tried again and aligned the two halves manually. We conclude that the holes we made to align the molds may not be enough to perfectly align them.

This part of the project was the easiest and definitely the most exciting because we got to see the fruits of our labor.

Cost Analysis

PLA Resin: at a price of $0.05/g. We printed about 7g = $0.35

We used about 0.4 ft^2 of pinewood which costs $2.36, setting the price of wood at $7.87/ft^2.

Liquid silicone – at a price of $25 for 500mL, we used about 100mL per half mold = $10 total

Liquid plastic – at a price of $60/gal, we used about 20mL per chess piece times 10 iterations = $3.18 total

Dye – at a price of $2.98 for 8 oz. We used about 0.1oz = $0.04

Labor – setting a labor cost of $15/hour. We each spent about 16 hours on this project = $480 

 

Total Cost: $495.93

 

Summary

This project did a good job at combining a variety of processes to make one object, so our team learned a lot about the tools available to us in 3D design. The most difficult part of this project was working with the Carvey to make a clean and detailed cut. Our next step would be to research how to make the Carvey machine our part with greater detail.
Ultimately, the biggest thing that we learned from this project was how to use 3D software to make the g-code for our part to send to the CNC machine. This skill will definitely transcend the scope of this class and will help us in our design careers for years to come.