Our Final Prototype

Our yoyo was loosely inspired by the brass rat which most years features a Boston and Cambridge skyline and is in many ways a trophy of one’s time at MIT.  My team set out to design a yoyo that would look as elegant as possible.  Our yoyos have a laser-etched and then waterjet metal insert piece on either side - one side has a design of the Boston skyline and one side has a design of the Cambridge skyline (Figure 1a and Figure 1b) - which is then surrounded by a wood, laser-cut, ring (Figure 2).  These two components were insert molded to form two insert plates (one of the Boston side and one of the Cambridge side) (Figure 3).  This insert molding process allowed the injected plastic to fix the metal disks to the wooden rings and also allowed the injected black plastic to fill in the outline of the night sky in the metal disk in order to create the dark night’s sky around the metal buildings.  The two insert plates snap fit into the base pieces.  For a full assembly please see Figure 4.  Our base pieces are a dark maroon color and simple in their design (Figure 5).  Their main purpose was to hold the insert plates and our goal was to make these pieces as thin as was possible.  Our yoyo’s stand is a thermoformed structure that was made from clear plastic which we then spray painted blue from the inside of the stand (Figure 6).  The stand served to connect the two sides of the yoyo (Boston and Cambridge) so our stand was inspired by the Charles River and has a subtle wave design on its sides.  Our stand also has “2.008 2017” and a series of our initials on two opposite sides.
Many of the key features of our yoyo are part of the insert plates and the majority of the key features became successes when we were able to manufacture the key features in the way we wanted to.  Most of our challenges arose from picking difficult key features that involved somewhat uncommon materials such as stainless steel and wood.  One key feature is the injection molding plastic filling in the negative space in the metal disks we insert molded and filling in to form a small lip around the outside of the wooden ring.  This was a challenge both in the mold design and the injection molding process.  When designing the molds we needed to make sure the metal disk was held in place securely enough so that when plastic was injected into the mold at great speeds that the metal piece would not waver and allow plastic onto the metal building design.  In order to hold the metal disks in place we constructed our molds to have small magnets set into our mold such that the metal buildings of the metal disk design would line up with the magnets and be held in place during the injection molding process (Figure 7).  We also needed to keep the front face of the metal rings flush with the other side of the mold or else plastic could sneak up onto the decorative front face.  When injection molding we had to carefully optimize parameters to fully fill the mold without forcing plastic up onto the decorative face.
The metal insert disks themselves are a key feature of out yoyo.  A great deal of time went into not only manufacturing these metal pieces but also optimizing the many stages of their production.  From the illustrator files to the laser cutting parameters to the addition of tabs to keep the disks from falling through the water jet’s grate when being cut, there was a great deal of optimization that went into creating a crisply etched metal insert piece.  To see a before and after image of our optimization process please see Figure 8 (the before image is poorly engraved and the outer metal ring was broken while being cut in the waterjet while the after image is well-engraved and has a solid outer ring).
The wooden rings are another key feature of our yoyo that eventually became a success after a large amount of optimization.  We encountered a host of problems from the wood warping during laser cutting to the wood cracking when place in the injection molding machine to the wood splintering when the injected plastic pressed up against the ring to the diameters we needed to support the metal ring not allowing for plastic to travel around the edges to form the lip.  Many of these issues were due to the fact that we were using very thin pieces of wood.  One by one we solved all these issues by cutting the wood in multiple passes, to altering the inner diameter, to coating the wooden rings in oil, and adding notches around the perimeter of the larger rings to allow plastic to travel and form the lip through these hand-made channels (in Figure 9 the image on the left is before optimization and the image on the right is after optimization) (Figure 10 shows flash coming up around the wooden ring when the wooden ring was not pressed tightly enough into the mold.  The image on the right is after full optimization with a ring with a slightly larger outer diameter and the addition of small notches around the outer diameter to allow plastic to flow more easily around the ring).  One of our greatest successes with this yoyo is the fact that most of our wooden rings came out in one piece and with little to no flash around the edges.
There are many opportunities for improvement with our yoyos.  We did not anticipate to encounter many difficulties with insert molding the wooden rings and for each iteration of rings we were not able to make work we had to order new wood.  We were using very thin wood veneer and we quickly found out that for wood marketed as 1/16’’ thickness there was significant variation between orders of the same wood.  The structure and characteristics of the wood were also different with each order and required tweaking of the optimized parameters.  Our yoyos would have been more consistent if we had purchased all the wood we were going to use at once.  We also could potentially improve the yoyo by making it slightly thinner.  A thin, light yoyo was one of our main priorities and while our yoyo is quite thin we could try to make it thinner.
At the beginning of this project we were asked to prototype a piece of our yoyo using a 3D printer.  My group chose to prototype our base piece using a FormLabs printer.  Our piece was created using stereolithography (SLA) (Figure 12).  Due to the manner in which it was made our prototype piece had many bumps from supports that we broke off that were remnants of the stereolithography process.  This first prototype was quite thick and we modified our design after seeing how wide our resulting yoyo would be.  The production time of the SLA piece was a few hours and the production time of the injection molded process for the same piece was less than a minute once we had the molds made and were set-up in the injection molding machine (Figure 5 shows the resulting pieces from the injection molding process).  Our base piece was not complicated and lent itself well to the injection molding process.

Figure 1a: Sketches of the designs to be laser etched and waterjet.

Boston is on the left and Cambridge is on the right.

Figure 1b: The left image shows a grainy laser etching while the

image on the right shows an optimized, clearer etched image.

Figure 2: Shows wooden ring around the metal disks

Figure 3: The resulting insert molded pieces of both the Boston and Cambridge sides

Figure 4: Assembly of both halves of the yoyo that show the wood, metal,

and black insert part snapping into the maroon base piece

Figure 5: Two images of the maroon base piece

Figure 6: The spray-painted blue thermoformed stand

Figure 7: One of the magnets in our mold to hold the metal disks in place  

Figure 8: Before and after images of the laser etched and waterjet metal disks

Figure 9: Wooden ring that splintered and broke when it was insert molded

(the left is the before optimization and the right is the after optimization).  

Figure 10: The image on the left shows flash that leaked up over the wooden ring when the wooden

ring was not pinched tightly enough in the mold.

The image on the right shows the final results after adjusting the outer diameter and adding notches.

Figure 11: The final yoyo

Figure 12: The SLA base piece

Figure 13: The insert plate cavity mold for the Boston side with the
ejector pins and holes for magnets and in the image on the right
the metal disk is inserted

Figure 14: The insert plate core mold that is able to hold the wooden ring in place

Part	Critical Measurement	Design Specifications	Measured Mean	Explanation
Base	Inner Diameter	2.3”	2.281”	The inner diameter is less than designed for because we found that the dimensions were too close for a good press fit, so we brought the inner diameter on the mold in.
Insert Plate (Cambridge)	Outer Diameter	2.31”	2.311”	It is similar, but slightly more because we slightly overestimated the shrinkage.
Insert Plate (Boston)	Outer Diameter	2.31”	2.315”	It is similar, but slightly more because we slightly overestimated the shrinkage.
Stand	Gap Width	1.02”	0.960”	The 3D printed mold for the thermoformed piece was made to be too similar to the design specifications. As the heat was raised, the shrinkage increased and continued dropping until the end of the run.

We only measured these 4 dimensions because these were the critical dimensions for each of our final components. The other components are either based on these components, the machines, or do not need to be changed due to the accuracy of their method of production (ie. laser cutting.) The insert plates came out approximately how we designed them and are kept for the mass production. The press fit for the bases as they were made were somewhat difficult to assemble. However, they were very secure and none of the pieces broke or were completely unable to be assembled. As such, the design specification will be brought down to 2.29”. The gap width of the stand shrank to lower than we initially intended, but we found that being able to press fit the yoyos into their stand felt much better than the initial plan to just have them sit in their stands. As such, the design specifications will be brought down to 0.96” for the mass production. However, it would still be important to better maintain the oven temperature for mass production to avoid the varying shrinkage we experienced. 

The cost of manufacturing the yoyo varies a lot with the processes behind it. As can be seen in table one below. The cost of producing 50 yoyos done in case 1 - production using the processes in 2.008: injection molding, thermoforming, laser cutting and water jetting- is incredibly expensive and more expensive than AM at 50 yoyos. This is mainly due to the cost of the equipment, for case one it is very high and is only spread across 50 yoyos. Case 3, mass production of yoyos, has a unit cost of only $3.79 because the equipment cost is spread over 100,000 yoyos.

Table 1: Cost Comparison of prototyping vs mass production

	Run size [# units]	Material Cost [$/unit]	Tooling Cost [$/unit]	Equipment Cost [$/unit]	Overhead Cost [$/unit]			Total [$/unit]
Case 1: Prototyping	50	8.44	3.60	1117	4.58			$1133.62
	Run Size [# units]	Boston Metal [$/unit]	Cambridge Metal [$/unit]	Wood Ring [$/unit]	Insert Plate [$/unit]	Base [$/unit]	Stand [$/unit]	Total [$/unit]
Case 2: AM Prototyping Different AM Processes Model 1	50	5.418	5.418	48.1432	31.0284	53.703	100.2555	$243.97
	Run Size [# units]	Complete Yoyo [$/unit]	Stand [$/unit]					Total [$/unit]
Case 2 AM Prototyping All Stereolithography Model 2	50	103.8258	100.255					$204.08
	Run Size [# units]	Material Cost [$/unit]	Tooling Cost [$/unit]	Equipment Cost [$/unit]	Overhead Cost [$/unit]			Total [$/unit]
Case 3: Mass Production	100,000	2.97	0.57	1.3590	0.1881			3.79

Graph 1: Production volume vs unit cost of making yoyos in the style of prototyping yoyos in the style of 2.008 - using injection molding, water jetting, laser cutting, and thermoforming with class 105 molds.

Graph 2: Production volume vs Unit Cost for prototyping the yoyo with various manufacturing processes. Making all the parts of the yoyo with FDM is the most expensive while making all the parts with SLS is the least expensive. The second most expensive is making each part with the process that would allow for the correct material to be used. For instance using SLS to get a metal Cambridge plate and FDM with a wood filament to get a wood ring.

Graph 3: Production size vs unit cost for mass production of the yoyo. Notice how the cost of manufacturing a yoyo comes down to less than $10 for production runs roughly greater than 50,000 yoyos.

Graph 4: Production Volume vs Unit Cost for prototyping yoyos in the manner of 2.008 production and using AM Manufacturing. Notice for production runs of roughly less than 400 yoyos it is cheaper to use AM to prototype the yoyos.

There are a couple of manufacturing processes and design elements that we would change for mass production.
Since water-jetting is very expensive, we would use sheet metal forming instead to punch out our stainless steel disks. The etching of the metal disks is currently done by applying two coats of Moly Lube to the sheet followed by etching using a laser cutter. At high volumes, we would use a more powerful laser which does not require to add any spray or paint beforehand. Another alternative to save time during the etching process is using photochemical machining (a process using photoresist similar to photolithography) to etch the details onto our buildings. Photochemical machining is very precise and economical since tooling is inexpensive and quickly produced.
Regarding the thermoforming, it would make sense to have multiple dies to form multiple stands at once. Additionally, we would have a rectangular punch to cut out the stands (for our current stands we used the available circular punch and then cut the rest off after. The spray painting of our stands (from the bottom) can also be automated.
For injection molding at a larger scale we would use multi-cavity molds to injection mold multiple pieces at once. Another constraint of the 2.008 manufacturing equipment is that the gate for injection molding has to be at a specific spot on the side. To improve quality we would put the gate at a spot close to the middle of our parts so the flow pattern is a single source flow and we do not have a weld line (currently slightly visible on our base part).
Our most challenging process was the injection molding of our insert plate which consists of insert-molding both the metal disk and the wooden ring. Insert-molding wood is very challenging since the wood breaks easily (at least at our current thickness of 1/16 “) and varies in thickness which causes flash on top of the wood. Since we need a very reliable process for mass production, we would either find a better, more stable kind of wood with constant (maybe increased) thickness or only insert-mold the steel disk and leaving a groove around the insert plate into which the wood can be fixed in the end using an adhesive.