When starting this assignment, I knew I wanted to print a 3D model with articulated parts, as I knew it would automatically be considered an “impossible object,” given how hard it is to create the internal geometry of those parts without 3D printing. As such, I went on thingiverse to search for an interesting model, landing on an articulated skeleton dragon model made by PenguinParty5. (https://www.thingiverse.com/thing:5346533/files).
Image 1: PenguinParty5’s Model
Unfortunately, when I went to slice it with the Bambu software, it appears the original file was too large to fit on the build plate, let alone within a small gumball capsule; I had to resize it myself. Seeing that the largest dimension of the capsule was about 2 inches, I scaled the longest dimension of the model to be 50mm, which is a bit less than 2 inches. Alongside it are two other iterations, with one of them being 45 mm, and the other being 55 mm. As I went to slice it, however, I noticed all of the supports that would be needed- post-processing this model would be quite annoying.
Image 2: The Generated Supports
As such, I instead decided to print an Articulated Ice Dragon by P1lotz (https://www.thingiverse.com/thing:6972754).
Image 3: An Image of the Printed Model, Provided by P1lotz
Noting that this model appeared to have over 1,000,000 triangles, I accepted the Bambu software’s suggestion to simplify the model before printing three iterations (still 45 mm, 50 mm, and 55mm).
Image 4: The First Three Iterations of My Print
With this first print, it appears as though the 55mm model couldn’t completely fill the capsule; I continued making larger iterations. Given that 55mm was about 30% of the original file, my next three iterations were 32%, 34%, and 36%. Despite the larger sizes, there was still room in the capsule, leading to another iteration at 55%, which ended up filling the capsule quite nicely. Thus, I printed 3 more of the same size using the Bambu FDM machine.
Image 5: The 55% Iteration Snugly Fitting Inside the Capsule
Afterward, I exported the simplified model onto the Prusa Slicer, scaling the model to the same size. Exporting the G-Code onto an SD card, I utilized a Prusa FDM to print my fifth model. Unfortunately, my first model failed due to bed adhesion issues. In attempting to fix it, I tried to use a brim, but it was quite hard to remove when the model finished printing. Therefore, for my third try, I edited the settings of the slicer (increased first layer height, decreased first layer speed, and increased the initial nozzle and bed temperatures of the nozzle and bed) and applied glue to the bed of the Prusa FDM machine. Fortunately, the print succeeded, but the upper layers had undesirable stringing. In attempting to remove said stringing, I broke this iteration, leading to another reprint.
Image 6: The Broken Iteration With Undesired Stringing
This time, in addition to the changes made for bed adhesion, I followed the changes suggested in the relevant Prusa article (https://help.prusa3d.com/article/stringing-and-oozing_1805), leading to a print with a bit less stringing. Learning from my past failure, I opted to use a heat gun to gradually melt the stringing onto the main model.
Image 7: The Iteration, Still On the Prusa FDM
While this method generally worked, it did cause some details to be altered (ex: one of the horns is now bent inwards).
Image 8: The Prusa Iteration After Using the Heat Gun
After ensuring all of my models moved well, I printed and cut out strips of paper with the thingiverse link of the model. Finally, I placed a model and a source strip into five gumball capsules and called it a day.
Image 9: The Cleaned Workspace (Where I used the Heat Gun)
Cost:
Cost of the filament (provided by the slicing softwares): $2.43
1.5 hours of work for $7.25 an hour (Minimum Wage) – $10.88
Electricity cost of using 3D printers is negligible
Total: $13.31
