Geometry Selection and CAD Prep
After identifying a rubber duck geometry we liked on thingiverse, we downloaded the .stl and printed a model of our duck. We initially printed it fairly small, and got some feedback to make it bigger.
We adjusted the scale, and then we used MeshMixer to slice the file in half and save the rubber duck half as a .stl file. Once we had the half duck model ready, we imported that into SolidWorks and built the 3D print positive that would be used to make the first mold half. We printed a first go at the positive, and after talking to the TAs and Dr. Wettergreen, realized we had some adjustments to make. First, our vents that we included to prevent air traps while molding were extruded in the wrong direction, and second, we could place them more advantageously based on our geometry and where air gaps may form. We made these adjustments, and also brought in the alignment pegs to make the mold shorter, and then our 3D printed positive geometry was ready.
Since our part is not axisymmetric, we needed a mirror geometry to use for our CNC side positive geometry. We saved a copy of our 3D print side positive and removed the vents (since we only need those on one side). Then we mirrored that geometry and changed the holes to pegs and pegs to holes. To be super sure our mold geometries would align correctly, we used boolean operators in CAD to virtually create the two mold halves. So we basically created a box in CAD and subtracted the geometry of the 3D print side positive from it to get the negative of that geometry, which is the silicone mold. Then we repeated that process for the CNC side. Image of CAD for the CNC side positive and negative below.
Once we had both negative geometries, we assembled them in CAD by aligning the pegs and holes. Then we repeated a similar boolean operation to what we had done before, we created a big box in CAD and subtracted the geometry of the assembled mold halves. You can see below, first the assembly of the mold halves (including a section view), and then the final cast part geometry- a perfectly aligned rubber duck with air vents (and rings showing the clearance between pegs and holes). So we knew the mold positive geometries were ready to be made.
3D Printed Mold Half Creation
Using the 3D printed geometry, we used cardboard and hot glue to add walls onto the print. We also used duct tape on all the edges to prevent any of the silicone from leaking out. We then mixed together the part A and part B of the silicone together for a total of 7 minutes before pouring it into our 3D printed duck geometry. While we were pouring though, we realized near the end that there were a multitude of air bubbles present in the silicone and were extremely worried that our resulting duck mold would be filled with gaps. We found that slowing down our pouring and also hitting the 3d printed part against the table a couple of times would break some of these bubbles.
After waiting a day for our mold to completely set, we removed the silicone mold from the 3d printed geometry, and found our mold to be perfect! The air bubbles we had worried so much about did not produce any gaps. We were now ready to move to our CNC part.
CNC Mold Half Creation
For our CNC part, we imported the 3D file Jordan created into VCarve! We followed the instructions listed in the proficiency section of the CNC assignment. However, when we completed our file setup, the 3D roughing was estimated to take an hour and a half, which definitely seemed too long. After a bit of messing around with the program, we discovered that we had accidentally had our material set to a wood that was not pine, and that we had Z-Level selected as our roughing strategy instead of 3D Raster. Changing this changed our roughing time drastically, taking us down from an hour and a half to twenty minutes. We also heeded the advice both the TAs and Dr. Wettergreen provided, and added a 20% step-over to our 3D Finishing file. We also decided to do only the 3D roughing file, and one finishing file.
However, once we tried saving our two files, we realized that we had accidentally opened the file into VCarve’s free trial version, so the files wouldn’t save individually. Luckily, this wasn’t an issue as we were able to open the created file into VCarve’s paid version on the computer, and save the tool paths separately then. Next time, we will check which program we use carefully to ensure that we do not waste time creating a file on a program that doesn’t function properly.
Next came the most difficult part of CNCing: Setting up the wood! This proved to be such a large challenge, as the wood we were using was slightly too short to be secured well on all sides. In addition, it was difficult to tell when a piece of wood would be secure enough to be cut until the cut began.
For example, on the first attempt, the wood appeared to be perfectly secured on all sides. But, when the cut began, the wood began lifting as the drill bit got closer to its edges.
After two more different configurations of the wood, and four attempts on two wood pieces, we got some advice from Izzie, who saw us struggling, and recommended that we use long screws with the green hold-downs to hold the wood downwards so it wouldn’t shift upwards. This worked fantastically, and we were able to finally CNC our duck into the wood!
Cardboard, hot glue, and duct tape were then used to create walls for the molds. We decided against cutting off the uncut sides of the wood, as it seemed like unnecessary extra work, and were also worried that we would accidentally cut into the pegs we created (as they were very close to the edges).
The silicone parts were then mixed, and poured into the CNC part. We then tapped the mold to release any possible air bubbles.
When releasing the mold, we found that there were areas that were extremely sticky. This was due to the increased amount of hot glue we used for the CNC mold compared to the 3D printed mold. From Dr. Wettergreen’s advice, we decided to place paper towels on the sticky parts. In addition, there was a small air bubble in the mold that created a tiny gap. We decided to continue using our mold for the plastic pouring instead of re-pouring, as the gap created from the air bubble was very small.
Molding and Postprocessing
Then we were ready to mold our ducks! We put the silicone mold halves together and held them in place with rubber bands. Then we poured Part A and Part B of the plastic material (equal parts by volume) and quickly mixed and poured into the mold. It was fun to see how the material got hot due to the exothermic reaction, and how it turned white when it began to cure. We waited for it to finish curing and then demolded our first duck.
The first duck came out pretty well! There was some misalignment of the mold halves that we could see at the parting line on the tail. Dr. Wettergeen advised that we use something rigid on either side of the silicone mold to get more even clamping. With adding wood blocks and also using fewer rubber bands, we were able to get better results in subsequent molds.
After demolding, we had to do some post processing on the ducks. We had to trim the air vent geometry away, and also remove some flash from the parting line. We were able to remove most of the parting line flash with a deburring tool or angle snips, and then we used sandpaper to smooth the last little bit down.
We opted for the classic yellow rubber duck for our colored pieces. Overall, we’re really happy with how our ducks turned out. They are adorable and also very uniform. It’s fun to see the difference between the wooden CNC half and the plastic 3D printed half apparent in the final mast parts.
Cost Type | Cost | Price | Source | Quantity | Total |
Materials and Machine time | FDM 3D print | $8.00/cubic inch | OEDK | ~15 cubic inches
(1 duck, 2x 3D print mold positive) |
$120 |
Silicone | $110.99/gallon | Lets Resin | ~25 cubic inches | $12 | |
Plastic | $145/gallon | Amazon | ~32 cubic inches
(~4 per duck) |
$20 | |
Labor | CNC Operator | $24/hr | Zip Recruiter | 2hr | $48 |
Prototyping Engineer | $38/hr | Ziprecruiter.com | 8hr | $304 | |
Total | $504 |