Tyra: Eva and I immediately knew that we wanted to model a train for this midterm project, specifically one that looked like it was taken right out of a Hallmark Christmas movie. Picking the Mechanical Movement that we would use proved to be more difficult, as we hoped to incorporate the chugging motion into our model. However, the gears that would allow this motion were not able to mesh with other gear well, limiting our options for the handle. Therefore, we decided on movement 55 with a slight adjustment of there being three internal stationary gears as opposed to one.
Mechanical Movement #55
Initial Sketches by Eva
For Gate 1, I generated the gear system we were going to use, while Eva drew our concept sketches. The gear generator was very helpful as it was easy to manipulate the spacing and orientation of the gears so that they worked how we intended. The hardest part was figuring out how to adjust the pitch diameter and diametral pitch to achieve the overall diameter that we wanted.
For Gate 2, we finalized the gear systems and structural pieces for our mechanical model and each laser cut then built a low fidelity version of our model. Eva created and edited the Adobe Illustrator file that we would be using as she had more experience and more accessibility to the application. To fabricate the first version of the low fidelity prototype I started by laser cutting the pieces out of cardboard. Before I committed to the full cut, I cut a small square to ensure I had the setting correct on the laser cutter. Cutting the gears and other pieces went seamlessly; however, the build was a bit more difficult. As with any first build, there were aspects that we had not yet considered such as mounting holes for the axles and the overall size of our model. I pretty much guessed my way through this first prototype, mounting axles where it seemed best, cutting wooden dowels simply to maintain a decent ratio, and putting hot glue everywhere. This prototype did its job by working slightly and proving the overall concept for our model.
The second version of the low-fidelity prototype, made by Eva, improved upon the first by including axle holes and bearing, which allowed the prototype to work much smoother. However, there were still issues that simply came from the cardboard, like the gears catching in-between each other.
Eva: In Illustrator, a challenge I faced was ensuring that every measurement was perfect, down to the thousandth of an inch. If the gears were not spaced correctly or the axle holes weren’t the exact diameter of the dowels, the mechanism would not function. This meant accounting for kerf and imagining the flat images in 3D. To create the buildings, I modified a live joint file and used the pen tool to add detail inspired by the brownstones in New York. I created press-fit joints to secure the buildings to a platform. Printing low-fidelity prototypes allowed us to identify which gears needed dowel holes and which required bearings.
Adobe Illustrator file that was printed on the laser cutter
A low-fidelity print on cardboard
Tyra: For Gate 3, we started with a medium fidelity prototype made from wood. This prototype used the same ai file as the second low fidelity one. While the pieces of the model of the laser cutter, someone booked the machine forcing me to shorten my cut while it was happening. Instead of making four cycles, the machine was only able to make two, causing some pieces to not be fully cut out. The houses and middle-layer external gear snapped while I was trying to force them out of the wooden slab. The external gear breaking basically ruined this prototype as the train was completely unable to move. Even with the broken pieces we were able to get some helpful tips, such as adding more axles and a trainguide that really improved our final.
1st print in wood
2nd print with modifications after 3rd Gate check-in
The final ai file was changed significantly based on the advice we received during Gate 3. We made our model bigger to include a fourth gear on the bottom and to add a guide for the train so that it would change orientation when necessary.
The final cut went fairly seamlessly. I booked the laser cutter in advance to ensure I did not get cut-off, which proved to be even more helpful than I intended, as I had to make two separate cuts to account for the increased size of all the pieces.
Building the final prototype of our model was very time-consuming, but not too bad to do. The worst part was definitely the gluing as each axle needed some, but I could do nothing until the glue dried as these pieces provided the support for everything. I was so annoyed that I got a heat gun in an attempt to make the glue dry faster. Furthermore, all the axles could not be done at the same time as some were on the bottom and others on the top. The other pieces were fine to attach as they were press-fit. Originally, I thought after all that work I was going to have to restart with a new bottom plate as one of the bottom gears kept catching, but luckily Eva realized they just needed to be switched before I began disassembling.
My final step was to make the handle. I CADed the candy cane on OnShape, basing its sizing on what would be most comfortable for Eva and I to turn. Using the dxf file, I cut the piece on the waterjet. The most difficult part of making the handle was using the Cricut. The dxf file from Onshape was unable to be uploaded to the Cricut software; therefore, I had to convert the file to a svg, then import that file in Adobe Illustrator. The ai file then could be uploaded into Cricut. Using the Cricut was fairly simple, although I could not figure out how to ensure the axle hole was not included on the cut. Furthermore, there was no red or white vinyl, which is why the candy cane is gold. Applying the vinyl to the metal was easy as it was basically a sticker, but I did have to use an Exacto knife to remove excess vinyl, including the axle hole.
Eva: The SVG file of the gears incorrectly showed one of the axles as off center, which caused it to get stuck in our prototype. By recutting the gear with a centered axis and moving the position of the dowel, I was able to get our model moving smoothly again. Additionally, I lasercut wooden washers whose inner diameter was 0.31 inches, 0.003 inches smaller than the dowel. Accounting for kerf, this allowed them to fit tightly onto the dowels, acting as stoppers to prevent the upside-down gears from slipping and becoming misaligned.
I was excited to decorate, but it ended up being more tedious and time-consuming than I anticipated. Professor Bisesti provided us with greenery meant for miniatures, but the “tree dust” would not stay no matter what adhesive I tried. Ultimately, I cut wreaths from the green material and added velvet ribbon and pearls to imitate snow. I think the final result is quite warm and festive!
The model functions, but it could be smoother. If I were to create another high-fidelity prototype, I would add another platform to increase the height of the guide rail. Visually, I would decrease the size of the train and I would cut out the windows of the buildings. I would have the external mechanism be a horizontal turn-wheel instead of a vertical pole to make the mechanism smoother and more ergonomic.
Final model (minus the candy cane handle)
Cost Breakdown:
Material:
Four sheets of 3/16″ plywood = $15.88
Seven 20mm metal ball bearings = $11.83
Round 0.3125″ wooden dowel = $4.38
Wood glue = $3.48
3″ x 5″ of aluminum sheet = $29.78
Quarter foot of Cricut vinyl = $0.25
Tools:
Epilogue FusionPro Laser Cutter, Waterjet, Cricut = available free via the OEDK, but thousands of dollars to buy individually (not included in final calculations)
Eva’s Adobe Illustrator monthly subscription = $22.99/month
Labor:
45 combined hours at $15/hour = $675
Total Cost: $763.60
A clean table!
