Summary and Reflection

 The bulk of our material cost is in the ball bearings embedded into the body. The ball bearings cost $2/ea, making up 66% of our yo-yo's $3/ea material cost. Adding in energy and labor costs, and based on our assumptions, our total variable cost per yo-yo is $4.83, which is pretty inexpensive compared to other yo-yos from our class.

Of course the real cost driver in a manufacturing run is amortizing the fixed costs of capital. Injection molding machines, thermoforming machines, and CNC lathes + mills are very expensive pieces of machinery, and we estimate approximately $700,000 could be spend on those machines.

This indicates that were we to enter into production, we'd likely be better off working with a contract manufacturer on their line rather than investing in our own machines, until we get to around 100,000 parts. At that point, we're dropping below $10/yo-yo, and starting to level off, so some consideration could be made about what advantages having our own factory would bring.

A log-log plot of projected cost per yo-yo vs number of yo-yo's manufactured. Fixed costs amortization causes the cost per yo-yo to asymptotically approach the variable cost.

When designing our yo-yo, we did so with the tool capabilities of 2.008 in mind, so we didn't have to compromise much on our designs. Had we been able to accommodate a 3 or more part mold, we could have incorporated some different snap joint designs, but our existing design worked well enough. Embedding more components would improve assembly time, and could be possible with more complex mold tooling. In particular, we could embed a small metal post into the spinner to replace the plastic nub, which would make pressing into the bearing less delicate and smoother. The bearing itself could be insert-molded into the body at the time of injection, which would reduce secondary assembly time. Locating features like pegs and holes could be molded into parts to improve the assembly process. Steel and/or actively cooled molds would improve our cycle time.

2.008 was a tremendously valuable experience for our team diving into the manufacturing mindset, and that mentality will help us be better designers upstream of manufacturing. Simply knowing what different kinds of processes exist and what sorts of parts they're good for is very valuable. Likewise, being able to do a QC analysis on samples to adjust tolerances and/or control limits is critical to having a smooth production run. 

The most value is gained from experience, and if I had to focus on one aspect I think could be improved, it would be the earliest labs introducing how to use MasterCAM. One alternative to consider would be having some tutorial videos as homework before lab, and having lab be the time to generate and verify toolpaths for the paperweights. More experience using the tool to build our own parts could be helpful.

Final Production and Assembly

The Yo! Yo-Yo is Finally a Reality!

The 4 parts of the Yo-Yo manufactured in-house: (Left to Right: Body, Spinner, Window, Ring)

The Body

The body, injection-molded with red polypropylene, is the largest part of the Yo-Yo and has 3 key interfacing features. 

The backside of the body has a small cylindrical recess in it to hold a metal spacer. At the base of the recess, a nut is embedded into the center of the body. This embedded nut allows the two halves of the Yo-Yo to be joined with a set-screw.

On the frontside of the Yo-Yo, a precisely dimensioned outside diameter interfaces with the Ring. An interference of a few thousandths of an inch allows the two parts to mate together and stay together after assembly without any fasteners.

Near the center of the frontside of the body, another small cylindrical recess is precisely dimensioned to fit a 6mm bearing. In final assembly, the small 6mm bearing is press-fit into the body, joining the spinner and the body together.

The body shape worked well and had a relatively quick production time. Even though the injection-molding machine had to be run in semi-automatic mode to insert-mold the nut, the overall cycle time was only 24 seconds. Note the ejector-pin marks on the frontside of the body. These marks are hidden by the spinner.

Though the bearing press-fit easily into the body, the bottom of the bearing recess occasionally interfered with the inner racer of the ball-bearing, increasing friction between the body and the spinner. A secondary recess, or a curved floor of the recess would eliminate this interference. The body also did not shrink as much as we expected it to. This caused difficulties in assembly, because it took a lot of effort to press the ring onto the body.

The Spinner

The spinner has only one key interfacing feature. On the backside of the spinner, a 3mm peg protrudes from the spinner. In final assembly, a ball-bearing is press-fit onto the spinner. When the bearing is press-fit into the body, the spinner and body are constrained together, while still allowing the spinner to rotate independent from the body.

The spinner is originally injection-molded with black plastic. Afterwards, white-out was used to paint the top surface of the spinner and highlight the letters. The backside of the spinner is asymmetrical, causing one side to weigh more than the other. This causes the spinner to naturally orient itself so that the letters are upright.

The 3mm peg on the backside of the spinner did not shrink as much as we expected. During assembly, we had trouble pressing the bearing onto the peg, but with a little help from a press, we were able to do it. Future design iterations should be used to improve this interface. Also, increasing the weight on one half of the backside of the Yo-Yo would improve performance of the spinner so that the Yo! was always upright.

The Window

The Window is thermoformed out of 0.030" High Impact Polystyrene. The window is slightly oversized so that it snaps into the ring and does not move around during assembly. This surface also has a slight draft angle to ease mold removal.

A small lip around the edge of the window is sandwiched between the body and ring after final assembly, keeping everything in place.

The window dimensions were very consistent and worked well in assembly. The window was a little bit cloudy after thermoforming and scratched very easily. It also had a few pock marks on a few of the parts. Though these defects are only cosmetic and do not inhibit the performance of the Yo-Yo, the window could really be cleared up in future iterations.

The Ring

The Ring, injection-molded out of red polypropylene, has 3 important interfacing features.

The window snaps into the inner-most diameter of the ring. The ring has another very important inner diameter, where the ring has an interference fit of a few thousandths of an inch with the body. These two features are important for the assembly of the Yo-Yo. However, the ring has an even more important interfacing feature.

The most important interfacing feature of the Yo-Yo is the outside surface of the ring. This is how the user interfaces with the Yo-Yo and a lot of thought was put into ensuring this surface felt good as the Yo-Yo was used.

The Ring fit on the body and kept the entire Yo-Yo together very well. During assembly, it took more effort to press the ring onto the body than we wanted. Therefore, future design iterations could enlarge that dimension, decreasing assembly time. Though we were able to cut off the runners easily, it would be a great improvement if the runners could be placed on a surface that was not the outer surface where the user touches the Yo-Yo.

	Target	Tolerance	Actual Measurement	Explanation
Complete Assembly
Diameter	2.514"	+/- 0.005"	2.513"	We optimized the shrinkage of the Ring so that this dimension was consistent.
Total Width	1.195"	+/- 0.005"	1.225"	The string gap was about 0.020" larger than the design specifications, which affected the overall width.
String gap	0.075"	+/- 0.025"	0.093"	It appears that shavings of plastic became lodged between the nut and the spacer, probably because the spacer recess was slightly smaller than the spacer OD.
Part 1: Body
Interference Fit Diameter	2.220"	+ 0.000"/-0.005"	2.217"	The interference is slightly larger than designed, because the body did not shrink as much as we expected.
Maximum Thickness	0.350"	+/- .005"	0.351"	This dimension is small compared to the diameter, and wasn't affected by the poor shrinkage.
Part 2: Spinner
Peg Diameter	0.118"	+ 0.000"/-0.005"	0.124"	This dimension did not shrink as much as we expected it to, because it is such a small part.
Peg Depth	0.118"	+/- .005"	0.112"	It appears that air got trapped at the bottom of the peg hole on the mold, because the peg had a concave shape every time.
Outside Diameter	1.860"	+/- .01"	1.835"	This dimension was not extremely critical, and it was ok that is shrank more than expected.
Part 3: Window
Outer diameter	2.030"	+ 0.000"/-0.005"	2.028"	We optimized the thermoforming parameters so that these dimensions were precise.
Inner diameter	1.970"	+ 0.000"/-0.005"	1.968"	We optimized the thermoforming parameters so that these dimensions were precise.
Part 4: Ring
Interference Fit Diameter	2.200"	+ 0.005"/-0.000"	2.218"	When we designed the ring mold, we accounted for shrinkage towards the center upon cooling. However, the ring appears to have shrunk outwards on the ID.
Window Inner Diameter	2.029"	+ 0.005"/-0.000"	2.030"	This dimension has a significant draft angle, so the shrinkage had less of an effect on the inner diameter.

Earlier we discussed the optimization of the Spinner part. This turned out to be the part with the most difficulty in manufacturing, because of the great variation in feature. The most critical dimension was the peg diameter. The average peg diameter was 0.124", which was 0.006" larger than the target dimension. Even though the peg diameter was off by 0.006", there was very low variability in the process. The injection molding process had a C_p value of 6.41 relative to the tolerance range of 0.005".

During our production run, we purposefully introduced a small parameter change. However, it had very little effect on the peg diameter.

For more information on the variation in critical dimensions during our production run of 50 Yo-Yo's, view our statistics report.

Overall, the assembly of the YO! Yo-Yo went very well and the finished product feels nice in the hand, making it a fun toy to play with. We learned a lot about process control and how to make well-fitting interference-fitted parts.

Spinner part process optimization

A short-shot, or incomplete filling.

Of all the parts for our yoyo, the spinner wound up being the most difficult to optimize. The first parts we shot had two defects: short-shots, and "dieseling". Dieseling was a problem we weren't familiar with before, but happens when trapped air in the mold ignites under the modeling temperatures and pressures, leaving charred marks on the part and a very distinct smell we won't soon forget.

An example of "dieseling" on our spinner component.
Our first shots were much, much worse.

In addition to the defects, we also noticed the post we molded to hold our bearing shrunk more than we anticipated.

This gave us three problems to focus on fixing with either our mold or our injection molding parameters:

1. Fix the short-shot
2. Fix the dieseling
3. Fix the shrinkage on bearing post

Fixing the shrinkage problem

We fixed this relatively quickly by re-machining the pocket in our mold to be a bit larger and accommodate a greater shrinkage than we initially expected. No problem.

Fixing the short-shot/dieseling

These two problems actually ended up being related. What was happening in our first batch, is that air couldn't escape the mold, and the trapped air would burn our part. Our first attempt was to vary the pressure and speed profiles to try and get a more complete filling. Unfortunately, higher pressure and speed fixed the short shot, but caused another problem: flash. Flash happens when molten plastic leaks through the parting surface and causes thin little webbing on the final parts. When we backed down on pressure or speed, we got short-shot/dieseling; and when we bumped them up we got flash. So we decided to look at modifying our mold to fix the problem.

A particularly egregious example of flash in our
spinner part

The first attempt to fix the short shot was simply widening the gate. This had a minor effect, but didn't fix the problem.

The next thing we did was use a scribe to cut some small ventilation lines out of our mold for the air to escape through. This helped a little bit, but wasn't enough to fix the problem. Next, we moved on to adding some small diameter, deep pockets for the air to be pushed into. After some trial and error in number of holes, location, and depth, we eventually found a combination that worked. As a result, our parts come out with tiny whisker-like posts as the mold is filled, but these are easier to deal with than either the flash or short-shot/dieseling defects. In theory, we could lower the shot size by an amount to minimize the whisker length, but we haven't tried that and the posts break off very easily so we may focus our efforts elsewhere.

Tuned spinner coming out juuust right. Notice
the lines in the runner from the scribe marks on
the mold used for air ventilation.

Final mold after rework. Enlarged bearing post on the cavity; as well as small ventilation lines
extending radially. Also notice the blind holes drilled as a place for air to escape. These holes
fill with some plastic as well, leaving whiskers on our final part that are easy to remove by hand.

The only dimension that we wound up changing in our final part was the diameter and length of the bearing post to better accommodate the observed shrinkage from our first shots. The other modifications to our mold are to combat defects in the molding process, and don't change the dimensions of our final part. With a tuned mold and parameter sheet, we're ready to enter our production run for this component.

Final optimized parameter sheet for production run.

Spinner Mold Design

            The Core mold contains a pocket for the main disc, as well as pockets for the letters Y-O-! There are also through holes for 6 ejector pins. A ring around the main disc pocket connects to the sprue through a runner. The ring is then connected to the main disc through 6 runners, located over each ejector pin hole.

The Cavity mold consists of one large pocket to add additional weight to one half of the disc. In the middle of the mold, there is also a small circular pocket, with a depressed lip, to create a nub on the backside of the disc.

      We measured multiple parts and found that parts with similar geometry to the spinner disc often had 2% shrinkage. Therefore, all dimensions were scaled by 102% in this mold to allow for shrinkage.

      For the Core, the ejector pin holes were drilled on a mill. Then, the mold was fixture onto a CNC lathe and main disc pocket was cut. Afterwards, the mold was fixtured on a CNC mill and pockets for the letters, outer ring, and runners were milled.

For the Cavity Mold, both pockets were milled using a CNC mill.

            

Process Plan: Spinner Cavity
https://drive.google.com/open?id=0B21e-g0iDwxqaFJ4Mzl4QVFBV1k&authuser=0

Process Plan: Spinner Core
https://drive.google.com/open?id=0B21e-g0iDwxqZHlSUkthWUw0ZFU&authuser=0

Manufacturing Time Estimate

# Yoyos	50
# Prototypes	10
Assumptions
1 hour per injection molded part (optimize)
1 hour per thermoformed part (optimize)
1 hour per ring injection molded part to run 100 parts
2 hours for body injection molded part to run 100 parts
2.5 hours for thermoformed part to run 100 parts

Component	Time/ea (min)	Total Time (min)
002-005 (Thermoform)	13.5	216
# parts	1
Cavity Mold	6	6
Process Optimization	6	60
Production Run	1.5	150
005-002 (Ring)	16.6	130
# parts	2
Core Mold	4	4
Cavity Mold	6	6
Process Optimization	6	60
Production Run	0.6	60
001-002 (Body)	24.2	197
# parts	2
Core Mold	15	15
Cavity Mold	2	2
Process Optimization	6	60
Production Run	1.2	120
003-003 (Spinner)	92.6	260
# parts	2
Core Mold	20	20
Cavity Mold	60	60
Process Optimization	12	120
Production Run	0.6	60
TOTAL		803

Our Initial Yo-Yo Design

                         Figure 1a                                                                         Figure 1b                                                       Figure 1c

Figure 1: Shown above are isometric (a), front (b), and side (c) views of our "Yo!" yo-yo design. The main feature of the yo-yo is the still "Yo!" design which is decoupled from the spinning body via a ball bearing. Upon throwing the yo-yo, the "Yo!" design is expected to stay not rotate with respect to the eye of the viewer.

                                                                             
                                                                             Figure 2

Figure 2: From left to right: an assembled mirror-image half of the yo-yo with the axle still screwed in it (yellow, red, gray), the body of the yo-yo (red), the bearing (green) which is press-fit in the body, the "Yo!" design (teal) which is press-fit into the bearing, the transparent cover (magenta) which will provide a window to the "Yo!" design, and the overlaying ring (yellow) that snaps onto the body and holds the cover in place.

Figure 3 

Figure 3: A transparent view of the body (red part in Figure 2) showing the embedded nut which the axle will screw into.

Our yo-yo will have 4 unique manufactured parts: the body (red), the design (teal), the cover (magenta), and the ring (yellow). The body, design, and ring will all be injection molded and the transparent cover will be thermoformed. As shown in Figure 3, a nut will be placed in the body mold so that it can be embedded in the injection molded part.

The assembly of the yo-yo is described in the following steps:
- Screw the axle into the body
- Press bearing into the body
- Press stem of design into the bearing; embedded nut will prevent deeper pressing of bearing
- Lay cover over the body so that flange covers the rim of the body
- Press ring over cover/body interaction and snap onto body
- Repeat to the other side

The manufacturing process influenced our designs considerably. Thermoforming can't keep tight tolerances like injection molding can, so early in the design process, we realized that we couldn't rely on having the cover snap into the body. Thus we took an action to add an outer ring that would snap over and onto the body and hold the cover from moving through pressure and friction of the small flange on the cover. Additionally, sharp corners have to be removed from the parts to optimize the forming and cooling of thermoformed and injection molded parts. Lastly, thought was put into the removal from molds and thus draft angles were added to parts in appropriate locations.

Project Gantt Chart

We'll use these to plan an manage our timelines and deliverables. As our plans update, so will these links:

* Full project report
* Link to Gantt Chart

Tasks and milestones are grouped into roughly 9 phases:

1) Logistics
2) Part Design, Project Planning
3) Mold Design
4) Mold Rework + Manufacturing
5) Process Optimization
6) Production Run
7) Assembly
8) QA/QC Analysis
9) Post project reflection

Each phase has three large action groups that correspond to the 3 large component groupings of our yoyo. Each component group gets consistent coverage from teammates to minimize need for extensive knowledge transfer :

* Body parts - These injection molded plastic pieces make up the foundation of our Yoyo.
* Spinner parts - These are smaller injection molded pieces that are inserted into the body of the Yoyo.
* Window parts - These are thermoformed plastic pieces that fit together with the body and allow users to see the spinner in action.

Will and Lesley will work on the Body parts, with Will leaning a little more on SolidWorks and Lesley leaning a little more in CAM

Chris and Eric will work on the Spinner parts, with Eric leading more CAM work and Chris taking on more SolidWorks work.

Lucien will lead the thermoformed parts, with Eric and Chris assisting as needed. Since the thermoformed parts are the simplest, it makes sense to have Lucien grab resources as needed and plan on having two people on each of the other parts.

Production and Assembly time shots in the dark at this point, but as we collect data on which tasks took longer than expected, we'll be able to improve visibility and granularity into predictions for how long future tasks will take.

Table of Specifications

		Target		Tolerance		Measurement Device
Complete Assembly
Diameter		2.500"		+/- 0.005"		Digital Caliper
Total Width		1.175"		+/- 0.005"		Digital Caliper
String gap		0.075"		+/- 0.05"		Digital Caliper
Mass		50.00 g		+/- 5.00 g"		Scale
Max rotation speed		2400rpm*		+/- 500 rpm		Laser Tachometer
Moment of Inertia (Izz)		.0536 lb/in^2		-		-
Center of mass (x)		0.063"		+/- 0.005"		-
Center of mass (y)		0.000"		+/- 0.005"		-
Center of mass (z)		0.000"		+/- 0.005"		-
String clearance		0.100"		+/- .0005"		Digital Caliper
Part 1: Body
Inner diameter		2.300"		+ 0.000"/-0.005"		Digital Caliper
Maximum Thickness		0.163"		+/- .0005"		Digital Caliper
Part 2: Window
Outer diameter		2.300"		+ 0.005"/-0.000"		Digital Caliper
Inner diameter		2.000"		+ 0.000"/-0.005"		Digital Caliper
Maximum Thickness		0.050"		+/- .0005"		Digital Caliper
Part 3: Ring
Outer diameter		2.000"		+ 0.005"/-0.000"		Digital Caliper
Lettering Thickness		0.070"		+/- .0005"		Digital Caliper
Maximum Thickness		0.340"		+/- .0005"		Digital Caliper
* Dropped from a height of 39.4 "

With the start of a new semester, another yo-yo team has been born! We are a group of 5 students in MIT's Design and Manufacturing II Class, 2.008, that will design and manufacture yo-yo's over a few months. We are currently deciding between three project ideas: the use of spinners on our yo-yo's, the installation of LED's, or the creation of an asymmetrical, but fully functional, yo-yo. Our weekly meetings are scheduled to be on Mondays and Wednesday, so stay tuned for more updates!

Chris Mills
William Pritchett
Lucien Morales
Lesley Wang
Eric Tu

(Mass Production of Many Wonderful Toys)