Hi everyone! In this blog post, I will describe the steps and process I used to 3D-print my impossible objects. After searching on Thingiverse for articulated print designs, I found the perfect print: the cute mini octopus.

Scaling the File

After choosing this print, I was told that this was a popular print in the class and that it usually worked really well. Near the 3D-printer computers, I found one of these that was already printed. After downloading the files and measuring the already completed print with calipers, I found that this print was scaled 100% and printed directly from the file without any scaling. I first measured the gumball capsule size with calipers to determine the scaling factor. I found the dimensions to be roughly 47.5mm x 47.5 mm x 55.5 mm (l x w x h).

Uploading the file into the slicer software, I found that the measurements were based on the length and width of the octopus with its arms fully spread flat, not curled up as it would be in the gumball capsule. Scaling the full file down to the curled form would be difficult to calculate. Luckily, there was a print of the octopus already made at full scale. Upon trying the best fit into the capsule (head down, arms curled underneath the body), I found that the fit was just too big.

Instead of measuring specific dimensions, I estimated that a 95% scaling of the file would be sufficiently small enough to fit in the capsule while also maximizing the space taken up in the capsule. For my initial 3 test prints, I scaled the file to 93%, 95%, and 97% to determine which would be the optimal fit. 100% scaling had dimensions 85.18 x 85.18 x 33.23 mm (x,y,z), and 95% scaling was 80.92 x 80.92 x 31.57 mm (x,y,z). I used the Bambu slicer software and the Bambu FDM printer to make these test prints.

Printing on Bambu

Checking the settings on the slicing software, I found that the settings for quality, strength, and speed were already within the recommended range. The notable settings were 15% for sparse infill density in gyroid pattern and turning supports on. The rest I left the same as they were, and images of the settings are shown below.

Following slicing and printing, I found my 3 test prints ready 6 hours later. At first glance, they all looked great! All that was left was to remove the supports and test each fit to determine the scaling I would use for the remaining prints.

While taking off the supports, I noticed that on the 93% scaled test, one of the tentacles did not print properly at all.

Despite this error, I tested the fits and found the 95% scaled octopus to be the best fit. 97% also fit, but some force was needed to close the lid, so I decided to move forward and print the rest with 95% scaling.

Finishing Remaining Prints

As the assignment required us to use two different machines to make the 5 prints, I decided to print 2 more pieces on the Bambu again and 2 new pieces on the Prusa, leaving a total of 3 made on Bambu and 2 on Prusa. Printing on the Bambu was simple as no one had used the machine or computer since I printed the test pieces, so all the settings were still the same. This time, the print was estimated for 4 hours to print the 2 pieces.

For the Prusa, I input the same settings as I used on the Bambu for the supports, speed, quality, and strength. After slicing on the Prusa software, the supports and slicing looked the same as the Bambu, so I was ready to print. I chose one of the Prusa’s whose PLA looked to be white, but it wasn’t until after the first layer started printing that I noticed the color was actually light pink.

The software estimated the print would take a total of 9 hours and 36 minutes to print the remaining 2 pieces! Fortunately, I started this print at 11PM the night after Beer Bike, and only 2 printers were in use at the time. Despite the long print, it would not cause major disruptions or delays in any other prints since it would be finished by 8:30AM the next morning.

Post-Processing and Finishing the Assignment

The next morning, I came back to successfully completed prints.

All that was left to do was remove the supports. With the Bambu prints, the supports were relatively easy to remove with pliers, and the extra pieces came off easily. However, the Prusa prints felt a lot stickier, and the supports took significantly longer to remove. I used a stitch remover to poke the supports out from in between the small articulated joints.

When cleaning up the pieces, I noticed that one of the prints made on the Bambu had a messed up tentacle again.

Unfortunately, the print did not complete the loop between the segments, so the last 2 parts of the tentacle fell off and broke. Initially, I was going to re-print this piece, however, there were no free printers available for hours. Also, the settings on the computer I used had changed, and when I re-input the initial settings I used, the slicing did not work properly. The top and bottom layers were not going to be printed, and the supports were not being generated properly. There may have been another setting that had changed that I was not aware of, and it was too late to start and finish a print in the time available. Thus, I now have 4 octopi and 1 octopus with a slightly amputated leg.

After removing the supports, I put all the octopi in their gumball capsules with a small printed slip with my name, class, semester, and the Thingiverse user, print, and URL. The final pieces weighed roughly 17g.

Overall, the 3D prints worked well. I did not have much trouble finding the correct settings to print my pieces without error. The main error was a flaw in the printing of tentacles on 2 of my pieces. I would assume that there may be an error in the printer in a specific region of the plate that may have caused problems as this issue occurred twice separately on the same machine, and it did not happen on any other prints.

Cost Analysis and Cleaned Workspace

Total Cost Estimate: ~$10

• Total printing time: 19.5 hours
• PLA material cost (@ $17/kg): ~85g total for 5 pieces = ~$1.45
• 3D printing usage: 10¢/g at Harris County Public Library: ~85g total for 5 pieces = $8.50

The cost analysis is a rough estimate as machine rental costs were not needed since we used the OEDK equipment, and material costs are estimates. 3D printers are free to use at the public library, and the cost is based on the weight of the final product, not on the printing time. Assuming a rate of $15/hr, an additional $292.50 would need to be added to the cost.

Below is an image of the cleaned Prusa printer.