The Final Oreo-deal

Drum roll..........The moment we have all been waiting for!

The Oreyoyo is officially ready to meet the world. Check out our promo video!

Team Oreyoyo is very excited to reveal our final prototype. From the onset of the design phase, through the manufacturing and early prototype stages, and now in the final product, it is seen that paying attention to detail was our biggest priority. The attention to detail in our designs’ careful reproduction of an actual Oreo, the added effect of the milk splash base, and the improved performance gained by adding metal washers, are all examples of how our team achieved our goal. 

The heart of our oreo’s design starts with the cookie pieces. The mirrored images of the cookie pattern is distinguished in each oreo by both an ‘Oreo’ and a ‘YoYo’ branded side. Much iteration was required on the Injection molding machine to achieve the required parameters to successfully fill all of the intricate contours and pockets that make up the  distinguishable Oreo design. Careful planning also went into the grooves along the cookies perimeter. As the parting line for the two pieces separated the groove into two parts, any slight offset in the mold design would result in a misaligned pattern. 

Now what is an Oreo cookie without its cream? The silky white, double stuffed goodness in our oreo was, similarly, an injection molded piece. The cream features a critical dimension, with tight tolerances, that allows it to press fit into the cookie pieces.  

Knowing that popular commercials always quote: ‘Oreos are milk’s favorite cookie,’ we decided to incorporate a splash of milk into our design. Our thermoformed piece resembles a splash of milk, while also serving as base that is able to hold the yoyo upright. The piece was constructed out of 0.030’’ thick plastic to allow for a robust base without compromising a small thickness. With the smaller thickness, our team was able to design our base gap to capture the width of the oreo, while also maintaining lower heating and cycle times on the thermoform machine.

Design and attention to detail were closely followed by yoyo performance. After testing our working prototype yoyo, our team decided that the yoyo’s overall performance would be enhanced with added weight. Due to the gap that the cream featured, we could easily incorporate a steel washer within the pieces. Samples of the washers with varying OD and ID by +/- 0.002’’ from the nominal dimensions in CAD were water jetted out and sampled on the prototypes. 

Each final prototype yoyo includes two steel washers, two identical cream parts, an ‘Oreo’ and a ‘Yoyo’ cookie piece, and a thermoformed milk splash. The parts needed of an assembly of one Yoyo are pictured to the right. 

The final prototype features pieces that we believe closely resemble an actual Oreo, while also matching the performance of a mass manufactured Yoyo.

Of course, there is always room for improvement. In our design process, our focus was to achieve as close a resemblance to an Oreo as possible. However, with regards to ergonomics, the shape of the Oreyoyo is too thin for a maximally comfortable grip. That is something we can change by modifying the dimensions of the parts and potentially adding more parts. 

  

To the left is an SLA printed early prototype of the cookie injection molded part. The advantage of SLA is that, unlike FDM 3D printing, there are no visible layers, which allowed us to achieve the intricate surface details on the cookie part. Due to the lack of parting lines and gates on the SLA prototype, it is even smoother than the injection molded piece on the top surface. However, the underside of the SLA prototype is relatively rough from support material. Lastly, while it takes little time to make one SLA piece relative to one injection molded piece (including machine time), injection molding is much faster in mass production.

Comparing Design and Measure Specifications

Here are tables comparing our design specifications to measured specifications:

 Measured

	Oreo cookie press-fit diameter	YoYo cookie press-fit diameter	Cream press-fit diameter	Thermoform gap
Target value	46.80 +/- 0.23 mm	46.80 +/- 0.23 mm	47.05 +/ 0.23 mm	23.00 +/- 0.23 mm
Average measured dimension	47.01 mm	46.99 mm	47.45 mm	23.11 mm
Standard deviation	0.0030in	0.0016in	0.043 mm	0.1514

 Design

	Oreo cookie press-fit diameter	YoYo cookie press-fit diameter	Cream press-fit diameter	Thermoform gap
Nominal dimension	47.0 +/- 0.23 mm	47.0 +/- 0.23 mm	47.5 +/- 0.23 mm	23.1 +/- 0.35 mm

Reason for difference in specifications:
Cookie part 

The average measured values for the cookies are just within the specification.  However, since they are nearly out of specification, most parts that are over that value will be out of specification.  A possible reason for this difference could be that we adjusted for shrinkage too much.  We scaled the molds by 102%, but that appears to have been too much.  Another possibility would be reducing the hold time, so that the parts shrunk a little more.  

Cream part

The cream inner diameter is out of specification tolerances.  Two possible reasons for this are adjusting for shrinkage too much or having too long of a hold time.  

Milk part

The averaged measured value for the thermoformed milk is within specifications.  However the 3 sigma value is outside of the tolerances, which means that many of the parts will be out of specifications

Cost Analysis

 

Now that we have small-scale manufactured 50 Oreyoyos, it's time to think about mass production. The unit costs for producing 50 yoyos we based off of the team’s total budget for the project including materials, lab time, and tooling costs. To extrapolate out to a hypothetical larger prototyping run we assumed that material costs would remain constant, molds would need to be remade every 50,000 cycles, there would be no more design, and the overhead/machine time would not decrease. The graph above show the projected total unit cost, which converges to
$34.36.

We considered the case of using additive manufacturing for mass production, for educational purposes. We assumed that all of our parts were made by Shapeways and the non plastic parts such as the string, nuts, and spacers were sourced similarly to our lab production run. The unit cost did not change for case 2 since no matter what number we put in for quantity their quoted unit price remained at $45.42, which makes sense because there is no distributed cost as production volume increases.

Last but not least, we calculated the cost for mass production of 10,000 yo-yos using a high-volume process using production tooling and non-dedicated equipment. The injection molded parts were approximated together while the thermoformed part was considered separately and then added back in. To make the IM section work we multiplied the mold cost and machine cost by 4 that way we could still keep the cycle time at 30 seconds per set of injection molded parts. Additionally the machine overhead was multiplied by 4 and the machines per worker divided by 4 to account for the fact that the 4 parts were being lumped as 1. The total injection molded unit cost was then added to the thermoform unit cost to give the total yo-yo unit cost. Some things missing from our large scale unit cost would be the cost of assembly and packaging, as well as the cost of the other off the shelf components such as the nuts, axle, spacers, and string. The graph above shows the project total unit cost, which converges to $2.59.

 The table below summarizes our cost analysis:

b.	50 yo-yos made using  2.008 prototyping	50 yo-yos made using AM prototyping	100,000 yo-yos made using high volume production
Unit cost	$34.36	$45.42	$2.59
Material	$1.92	--------------------------	$0.16
Tooling	$7.00	--------------------------	$1.30
Capital	$10.00	--------------------------	$0.01
Overhead	$15.55	--------------------------	$1.11
Findings	The dominant cost components both for our prototyping and for a high volume production run are the tooling costs and the overhead costs.This is because material is a flat rate while tooling and overhead are larger and not as diffuse. Additionally for our prototyping run our material costs are higher because we are not buying components in bulk and our capital costs are high because they are only spread over 50 yo-yos.

Design Changes for Mass Production

1. We modified the surface features of the cookie part to accommodate the smallest diameter mill tool available at the lab. With available resources, we would use the original Oreo surface features.

2. With an injection molding machine that allows us to relocate the gate, we would put it on inside of the cookie where it wouldn't be visible. 

3. We would make cream in different flavors, such as mint, strawberry, etc.

4. (Slightly unrelated) We would make a deal with Oreo to add Oreyoyos as toys that come with Oreo packs :)

Oreos Fresh out of the Oven!

After half a semester of hard work, we finally have an assembly of our yo-yo! In this post, we'll share the process that got us here here.

Injection Molded Cookie Part

We will first talk in-depth about the manufacturing process for the injection molded cookie part, and in the next section we will go into the details of the optimization process for each of our parts.

Mold Design and Machining

The molds used for injection molding the part require a lot of detail-oriented design and machining. We started with a SolidWorks model of our cookie part and used SolidWorks' mold-making function to generate models of our molds.

Cavity Mold

Core Mold

Once we had the SolidWorks models of the molds we loaded the models into MasterCam in order to generate the code we needed to machine the molds on the CNC lathe and mill. Before actually machining, we made sure that we had a solid process plan detailing the machine tools we planned on using. Okay, time to machine the molds. After a quick turn on the CNC lathe and a 2-hour long haul on the CNC mill, our molds were Ore-ready! Checkout how great the surface details turned out.

Injection Molding

Now for the big finale--injection molding. There were a lot of parameters to tweak, such as mold temperature, hold pressure, and cooling time. After hours of optimization, we ended up with a set of parameters that we were happy with. We will explain the optimization process for each part in the next section. With optimization in place, we are finally ready for our production run to make 50 yo-yos!

What We Learned

Injection Molded Cookie Part


Through the optimization process of the injection molded cookie part, we learned that subtle errors in mold alignment can be an eyesore on the plastic part.

As soon as we saw our first injection molded piece, we realized that we had a problem. The two halves of the cookie (core and cavity sides) were misaligned rotationally by 1-2 degrees and vertically as well, as seen in the photo to the left. Being perfectionists, we decided that this imperfection was not to be tolerated. Fortunately we were able to fix the rotational misalignment by re-programming and re-machining the core mold. As a result we had to increase the shim thickness from .003” to .032” and decrease the ejector pin length to 5.570”. We fixed the vertical misalignment by switching to 0.212" and 0.2115" dowel pins.

The second problem we encountered was dishing on the edge of the cookie opposite the injection site. We first tried to fix this by increasing the injection speed parameters, but that had no effect on the dishing. We then increased the feed stroke, which increases the shot size, from 1.20” to 1.35”. This reduced but did not eliminate dishing, and we had capsized on shot size. Our next resort was to increased the holding pressure, but not increasing it so much so as to create flash. We were able to find a sweet spot with holding pressure where there was neither dishing nor flash.

We then evaluated the pressfit between the cookie and cream parts to see if we needed to adjust shrinkage. The inner diameter of the cookie averaged at 1.849", while the outer diameter of the cream averaged at 1.860. This was within comfortable margins for a good pressfit (~0.01"), so there was no need to tweak holding and cooling times to adjust shrinkage.

Most importantly, we had to mix the perfect color for our Oreyo-yo. After experimenting like painters for a bit, we decided to go with a 50:50 mixture of mid-brown and black thermoplastic pellets, as seen in the photo to the right.

With all of our parameters set, we ran the injection molding machine on automatic for 5 pieces to find our average cycle time, which turned out to be a reasonable 30 seconds.

Injection Molded Cream Part

Through optimizing the injection molding process of the cream part, we learned that fixing a problem, such as dishing, can take a lot of tries and tweaking of parameters, but in the end it's pretty satisfying to get it right.

Dishing is the ripple effect in the surface finished caused by shrinkage, as seen in the photo to the left. Our first attempt was to increase the holding pressure, which helps the part cure before ejection. We started from the default of 300 Psi, but even after going as high as 1200 Psi we were only able to reduce but not completely eliminate the dishing. The next thing we did was increase cooling time, which also helps the part cure before ejection. We experimented with cooling times ranging from 25 to 45 seconds but saw no significant change after 30 seconds. Next, we raised the holding time from 2 to 8 seconds, but there was no reduction in dishing, which got us suspicious. We started wondering if the sprue was freezing too early, preventing plastic from flowing in through the gate and causing dishing. Dave, one of our wonderful and knowledgeable lab instructors, showed us a way to find out if our suspicion was right--by weighing the part while increasing the holding time until the weight was no longer changing. It turned out that while the part weighted 8.5g with 2 seconds holding time, it remained at 8.7g after 4 seconds. Aha, this means that somewhere between 2 and 4 seconds holding time the sprue was freezing off.

It was easy once we had diagnosed the problem. We remachined our mold to have a larger runner size, which freezes later As expected, the dishing was gone, and the cream part looked a-cream-azing!

Thermoformed Base

The thermoformed base is the seat for the yo-yo. It's supposed to resemble the milk splash associated with Oreo. Pretty sleek, huh?

Through the optimization process of the thermoformed base, we learned that heating time, which determines how stretchable the plastic is, is very important in thermoforming. Just into our trial runs we found that the platform was too low. This was because our part is taller than typical thermoformed parts. Below is a pictured of a base made with the platform too low. Note how the mold did not have enough room to fully form. 

However, after adjusting the platform height our part still wasn’t fully forming. We decided to optimize by adjust the heating time. Adjusting the heating time allowed us to achieve fully formed parts. However, as a compromise, we had to settle for a longer cycle time. Pictured below is our piece formed with a heating time of 30s, 40s, and 50s.