Yo-yo Manufacturing

For Design and Manufacturing II (2.008, year 2012) at MIT undergrad, my team designed and made molds for an injection molded Octopus Yo-Yo. The intent of the project was to teach and expose the team to mass production design and manufacturing, in particular injection molding. I am writing this a few years after graduation, and looking back I realize what the heck they were trying to teach us! When I was a student trying to finish psets and pass final exams, I was not able to connect the dots between 2.008 and the real world. It is nice to have the opportunity and time to reflect on these things after working in the industry for a few years.

Finished Yo-yos

All the fabrication took place at the Laboratory for Manufacturing and Productivity (aka LMP) under the instruction of Professor Sanjay Sarma, and technical instructors Patrick McAtamney and David Dow. Team "Octopus" (Kaitlyn Bailey, Emma Benjaminson, Lucy Du, Tanya Liu, Shaka Thornhill, and myself) split the workload by part, and I was tasked with the body. I will touch on the process of designing an injection molded part (in a classroom setting).

Exploded CAD View

After brainstorming and selecting a concept / design, one of the first tasks was to estimate the shrinkage of the body part. To do this, we found a similar Yo-Yo from a previous year and took several measurements of two dimensions on the part (the thickness and outer diameter). This gave us a bulk average shrinkage factor to uniformly scale our CAD models. There are advanced modeling tools to do this in the industry, and more sophisticated measurement techniques (like volume measurement or 3D scanning), but this approach would satisfy the requirements for a school project and give students an insight into the troubling nature of shrinkage. Something to note is that the measurements show that the part will shrink non-uniformly, which means that scaling the part in CAD uniformly is actually not adequate for high tolerance applications.

Empirical Shrinkage Estimate Table.

This was one of my first CAD projects, and although it was simple, it was a great exercise in DFM. Some of the basics for injection molding are wall thickness consistency, accommodating shrinkage around irregular or non-uniform volumes, features for inserts, and draft angles to allow the part to come out of the mold.

The Yo-Yo was composed of 4 injection molded parts, and 1 thermoformed part. The manufacturing process for each part was unique, requiring the shrinkage estimation, design of parts / molds, programming of CNC machine tools. Here we will look at some of the fabrication processes of the molds.

Machining of the thermoforming mold on a CNC lathe. The program used to cut this part was generated on Mastercam.

Ring mold with lettering

Core Mold, with Ejector Pins

Body Core and Cavity setup on molding machine

The Body mold has a Core and a Cavity side, which press together and then plastic is injected inside to create the part. The holes in the Core are for ejector pins (which push the part out of the mold after cooling). This die set has pre-drilled ejector pin holes available for use by the students. For each part, we had to line up existing locations with appropriate areas on the molded parts. It is expected that the part will stay on the Core due to shrinkage. The Cavity has a magnetic insert with a thread in the middle. This is used to hold a nut insert that will stay inside of the part. The locational accuracy of this nut was critical for the relatively large Body we had designed. It turns out that some Yo-Yos did not spin correctly (wobbly) due to imprecise locating of the nut (my theory).

Setup of the Engel injection molding machine. My fuzzy memory tells me this was a 5-ton machine (small).

After all the molds were made, the process of producing the individual Yo-Yo parts began. Here you will see the Octopus team learning how to setup and run the injection molding machine and thermoforming machine at LMP.

Looking at the first molded legs

Inspection of the first molded legs. After identifying issues with some of the parts. the shot size was adjusted (sometimes plastic did not flow correctly, or not enough plastic was provided, etc).

First cover off the thermoformer

Finished mold for the thermoformed HIPS clear cover. The vacuum holes around the edge are used to suck the heated plastic onto the mold, conforming to its shape (hence "thermo" - "formed").

All the molds with the first run of parts.

From the top in the image above - thermoformed cover; injection molded tentacles; injection molded face; injection molded outer ring; injection molded body. You can see that the first Body shots were short, leaving large voids in the part. The improper shot size was probably due to bad runner and sprue volume estimates.

Inspection and assembly was the final stage. Inspection of critical dimensions of each part was done and checked with tolerances. As you can imagine, the manual inspection of 114 Yo-Yos was tedious and time consuming. In the industry there are automated measuring techniques (such as computer vision and CMMs), and 100% inspection is rare.

Histogram of the OD of the Body part. As you can see this can be approximated by a normal distribution (as expected). This is a good data set to run statistics such as RSS, UCL/LCL, Cpk, etc.

Run chart of the OD of the Body part.

Yo-Yo Rings with color gradient

Finished Product!

Thank you for reading, and thank you to the 2.008 staff for one of the best class projects offered at MIT MechE undergrad!