Tuesday, December 8, 2015

Blog Deliverable 4:

1. Team Video:



2: Description of finished yoyo:


Our team designed a pokéball yo-yo, complete with the classic spherical pokéball shape, striking red, white and black colors, and even a pressable button on one side.  The yo-yo is designed for optimal performance, sloping the surfaces between the two yo-yo halves and adding weight to both sides with shims.


We designed five unique injection molded pieces and one thermoformed piece.  From left to right, on the image below, the pieces are the “ring with button” (injection molded), “button” (thermoformed), “face with button” (injection molded), shim (steel), “base” (injection molded), nut, screw and bushing, nut, base, shim, “face without button” (injection molded), and “ring without button” (injection molded).  


View3.PNG
An additional view shows greater detail on each individual component:
View2.PNG
All injection molded pieces on the yo-yo connect to the base pieces through long pins (a press fit) and in slots around the edges (snap fit).  The use of two different fit types (snap and press) for each injection molded part to the base not only ensures durability over time, but also improves aesthetics, as the snap fit around the edges guarantees that the face pieces will hold tightly to the base piece, without a development of a gap.  The thermoform piece fits into the circular hole in the ring with button piece and is supported at its edges by the two longest pins on the base piece (note that the two base pieces are identical; on the side of the yo-yo without the button, these long pins are not engaged).


Our tooling design for injection molding was unique in that it utilized the ejector pins in the injection molding setup itself to create the long, thin pins extending from the base and face pieces for press-fits.  We also explored unique tooling options with the thermoformed button; we wanted to create a part which was smaller than could be cut with the normal 2.008 die tool, so we added circular ridges around the button shape on the thermoforming die which we could use to position the plastic within a steel punch to quickly and accurately cut out the button shape in the appropriate size and shape.


While injection molding the face, ring, and base pieces, our initial design assumption that we would see approximately uniform shrinkage over all parts (backed by the flexibility in assembly provided by the long press-fit pins) turned out to be correct; after beginning test injection molding runs, the only additional changes to our tooling required for injection molded pieces was a slight widening in the ring pieces, which was purely for aesthetics (though we did some fit-related remachining for our thermoformed piece).  Additionally, we found that the similarities in shape between the button and non-button sides allowed us to use the same injection molding parameters for both.


After successful assembly of our prototype yo-yos and a bit of performance testing (our yo-yo reaches an angular velocity of 6000 rpm!) we are excited to present to you the Yo-yo Bros Pokéball.
Yoyo2.jpg

The parts we made using FDM include:
  • The two external halves that make up the exterior of the yo-yo. The front side (with the button) is different from the backside (no button). Total, this is 4 parts but only 2 unique parts because the top and bottom of the front and back faces are identical.
  • The black curved parts that make up the interior of the yo-yo. Total, this is 2 parts (2 halves) and these are the same.
  • The black horizontal strip on the back. This is 1 part.
  • The black strip on the front (different from the back). This is 1 part.
In total, we used FDM to make 8 parts and 5 unique parts. We decided to use FDM to create these parts to ensure that parts can fit together snugly and to ensure that we have accommodated for tolerances in the fitting together of our parts. By having the physical parts to fit together, we can see the areas that we will need to refine to different dimensions or tolerances.


It is difficult to make a comparison between the FDM parts and the injection molded parts, because the quality is so different. We have much smoother curves and contours with our injection molded parts, but do have a bit of shrinkage still present. For example, there are still small indents on the inner face of the base piece, yet when compared to the FDM part we do not have the poor resolution that the printer has, resulting in a very smooth curve. The pokeball halves also suffer from a slight bit of dimpling on the outer surface which is not present on the FDM parts, but again the injection molded parts are much smoother. A similar comparison can be made for the horizontal parts.

One big different between the parts is that the snap fits on the FDM parts would not work at all directly out of the printer. We had to do some post processing to get them to fit together, something which we didn't have to do with the injection molded parts. This is critical in the production of the yo-yos, and we are satisfied with the way in which our injection molded parts fit compared to the FDM parts.

In terms of opportunities for improvement, one thing that immediately comes to mind is that we could use tighter tolerance ejector pins. We had flash issues where the plastic was coming into contact with the ejector pins, indicating that the ejector pins or reamer used were not as precise as we would like. Along with this, re-positioning of the gates would be another area for improvement, since it did impact aesthetics slightly. Adding ribs to the base part would stop the bowing out of the posts which the pegs are inserted into. Along with this, taking time to remove the dimples on the outer surface would also improve the exterior aesthetics.


3. Table of yoyo specifications vs measured specifications:


Design Specifications


Part Name
Feature
Design (in.)
Avg (in.)
Max (in.
Min (in.)
STD
Variation
Base
Press Fit ID
1.900
1.992
1.996
1.982
.0027
4.842%
Pokeball Half (No Button)
Location Pin Diameter
.1250
.1251
.1260
.1240
.0003
.0800%
Pokeball Half (Button)
Location Pin Diameter
.1250
.1248
.1270
.1235
.0006
.1600%
Divider
(No Button)
Band Width
.2300
.2279
.2295
.2265
.0007
.9130%
Divider (Button)
Button ID
.5575
.5829
.5870
.5710
.0029
2.490%
Button
Outer Diameter
.6000
.5879
.7130
.5600
.0454
2.017%


Base
For this part we measured the inner diameter of the press fit ring. With a variation of almost 5% this dimension changed quite a bit from its intended value. The difference is due to shrinkage of the part while cooling. This could have been remedied by making this (and probably the outer diameter as well) slightly larger features in the mold so that shrinkage brought them closer to our intended value. We did account for about 2% of shrinkage, but were apparently off by another almost 5% for this particular measurement. The part is still entirely effective, however, because the halves and the band that have to fit into this press fit shrunk by similar amounts and the fit is still functional.


Pokeball Half (No Button)
We measured the diameter of the location pins on this part. These pins interface the holes created by the ejection pins in the base. The variation for this part is only .08%, meaning we were able to get very close to our intended value. This is because the shrinkage for this post was easier to account for than for the larger rings in other parts. The pins are so small in diameter that shrinkage doesn’t affect them as much as it does other features on the yo-yo.


Pokeball Half (Button)
The same feature was measured on this half as on the non-button side. However, in this case we found a variation double that of the non-button half. This could be explained by slight differences in the materials or in other subtle conditions since the production runs were done at different times, and since the measurements are still so close to their intended values. This is due once again to the size and predictability of shrinkage of this feature.


Divider (No Button)
We measured the width of the band itself because this is the most critical dimension of this part. If the band is too narrow then gaps will form on the face, and if it is too wide then the parts won’t fit together at all. We found a variation of .9% from the part’s intended width. This is within an acceptable range, and after assembly we have found that our yoyo faces fit together nicely across this band. The slight difference from the intended value here is also due to more shrinkage than expected. The part is thin and wide, which leads to a large amount of shrinkage.


Divider (Button)
For this part we measured the inner diameter of the ring that will hold the button in place on the face of the yoyo. The variation was fairly large at 2.5%. This is because of the shape’s tendency to shrink as well. The ring around the button is fairly thin and large in diameter, so it makes sense that this inner diameter would enlarge as the material is pulled outward towards the opposite edge. The button is still able to be framed nicely within the ring, so the shrinkage is once again not too much of a problem.


Button
The outside diameter is the most important value for this part. This is what butts up against the inner edge of the ring in the divider mentioned above. We found a variation of just over 2% from our intended value due to more shrinkage of the thermoformed plastic than expected. This could have also been accounted for better in the design, but the buttons do still fit within the rings as intended.


Part Name
Feature
Production Spec (in.)
3σ Value (in.)
Variation at 3σ
Base
Press Fit ID
1.900
.0030
.1579%
Pokeball Half (No Button)
Location Pin Diameter
.1250
.0009
.7200%
Pokeball Half (Button)
Location Pin Diameter
.1250
.0009
.7200%
Divider
(No Button)
Band Width
.2300
.0021
.9130%
Divider (Button)
Button ID
.5575
.0030
.5381%
Button
Outer Diameter
.6000
.0300
5.000%


4. Summary of cost analysis:

After performing a cost analysis on the yoyo, it was determined that the cross-over volume between additive manufacturing and the methods we used in 2.008 occurred at about 425 units. Now, this is much more than we made in the class, but in reality we were just getting started with our production run. We had invested all the time in the tooling, but had only used it for a small portion of its life. 

Shown below (apologies for two separate plots) are the unit cost vs volume produced for the 2.008 methods and additive manufacturing, respectively. You can see how high the initial investment is for the method we used in 2.008 compared to having a third party 3D-printing company manufacture the yo-yos. This being said, there is some tweaking to be done with the cost models to get a more accurate representation of where this crossover quantity is. Better quotes should be received from additive manufacturing companies, and more explicit guidelines on what we are actually "paying" for in the 2.008 lab should be set, such as the machines or labor.

This all being said, the graphs still do show the general, very important principle that if you are producing in very small quantities it makes more sense to use additive manufacturing, but it quickly becomes cheaper at higher volumes.

2.008 (injection molding) method: case 1 graph.png

Additive Manufacturing Method:
case 2 g.png

5. Reflection on yoyo design:


              The 008 shop’s tooling only limited our final product in a few ways. First, we were unable to get some cuts deep enough because the available 1/16th endmill was slightly too short. Secondly, we were getting fairly rough final passes on some of our circular molds because of how worn down some of the available tools were. Finally, the shop was not equipped with a die small enough to effectively cut our thermoformed button.

               The lack of a long enough 1/16th endmill mainly affected our two horizontal divider core molds. Since the core mold’s main body extends out from the base of the block of aluminum, cutting deep down into the mold to make the male snap fit parts put us in danger of hitting the base of the endmill against the mold. Dave Dow kindly suggested that he would purchase a longer endmill for us but instead we decided to redesign the depth of the snaps to fit the tools we had available. This simply involved not cutting as deep into the mold, therefore sacrificing a much tighter edge fit for a safer machining process. We were able to compensate for the smaller edge fit by then designing the pin system that we used for the final design.

               A much less acute issue that we faced was the rough finishing passes we were getting in some of our molds from the lathe. This mainly affected our universal cavity mold, which is nearly completely machined on the lathe. The issue probably stemmed from the large number of uses the machine and tools were seeing this semester, which is completely understandable. In order to combat this issue, we simply rubbed a heavy dose of Scotch-Brite along the inside of the mold to smooth the surface.
               Finally, for cutting the thermoformed piece, we did not have a die that was small enough for our button, and we did not feel we had the time to make one. Our solution to this problem was to use the sheet metal stamper in the lab as a thermoformed piece die. This solution allowed us to quickly cut out our pieces and get them into the yoyos without producing a brand new die.


               We have several ideas on how we would adapt our yoyo for mass production that will make assembly easier and make the final product look better. First, we would improve the tolerances on our press fit pins in order to make assembly easier. Secondly, we would put more thought and effort into the production of our button mechanism. Finally, we would implement interior gates, which would vastly improve the quality of the edges of our yoyos.

               To improve our assembly time, we hope to improve the tolerances on our press fit pins. This includes better reaming of the ejector pin holes on the base piece, which should reduce flash that tends to get in the way of the male pins when pressing the outer parts into the base. The resistance we sometimes got when pressing the pieces together could even cause some pins to break.

               We also plan on redeveloping our button mechanism, which tended to have issues with falling into the yoyo’s body. This is because if the button is pressed to hard or at an angle perpendicular to the two posts holding it up, one of the posts would sometimes come out from underneath it. This kind of defect is a difficult one to fix once it has happened, because the entire yoyo needs to be disassembled once it happens. Our planned solutions are to either add two more posts on the two sides of the button that currently don’t have them, as well as to move those posts closer to the interior of the button, with a slightly increased height for a more rigid button press and a tighter grip to the outside wall.

               Finally, we hope to add interior gates to our molds to improve outer edge quality. The gates that we currently have run from the sprue hole to the outer edge of each of our pieces. This means that when we cut the gates off, the piece is usually either left with a small nub of plastic along the outside, or a chunk of plastic cut off along the outside. This gives the yoyo a rough surface finish that could sometimes be quite sharp if not filed down. Our solution is to add internal gates the run to the inner faces of our parts, which means the gates will never be seen, thus we don’t need to be so precise when cutting them and also do not have to file them down afterwards. 

Friday, November 20, 2015

Yo-Yo Bros Blog Deliverable #3

1.      Our most interesting injected molded part is the base piece.
a.      Mold Design
The cavity mold both holds the nut for the pokeball’s shaft and defines the curved back face of the yoyo. This is done by having a curved cavity bored out of the center of the mold, with a center through hole for holding the shaft with the nut.
The core mold is designed to produce the inside face for the yoyo, which includes female snap fits around the outside edge, six press fit pin holes, and two long pins for holding up our thermoformed piece. The snap fits are used to keep our outer ring pieces aligned with the outside edge of the base part and to provide a small amount of support in keeping the final product together. The press fit pin holes are the main source of fastening within the product, with each additional ring part connecting to the base through two of these holes. The holes are made by drilling a through hole in the mold for an ejector pin to rest, and then counterboring a wider hole. This wider hole allows injected plastic to surround the ejector pin tip, creating the pocket that the pins on the ring pieces can press into. In addition, this mold also includes three magnets press fit into holes drilled into the piece. These magnets then hold steel shims around the edge of the base that add some weight to the final product.
b.      Manufacturing process (step by step process plan)
PROCESS PLAN FOR BASE MOLD, Cavity
Step
Operation
Machine
Tool
Justification
1
Face front of mold
Lathe
10
Make front of mold flat and square to fit against core side
2
Cut cavity
Lathe
10
Don’t have any small corners, use large bore tool
3
Cut gate
Mill
9
Cavity side gets gate
PROCESS PLAN FOR BASE MOLD, Core
Step
Operation
Machine
Tool
Justification
1
Face and machine convex curve for profile shape
Lathe
01
Make front of mold flat and square to fit against core side
2
Cut sharp concave corners from step 1
Lathe
03
Finishes off the general profile of the coreb
3
Cut snap fit ring on outside
Lathe
08
Small slot, only needs finishing trepan.
4
Cut bolt insert mold extrusion
Lathe
12
½ inch endmill to cut extrusion for bolt to sit in
5
Drill holes for pegs
Mill
1/16 in drill bit
Cut holes for pegs which support the button from behind
6
Drill hole pattern for press fit pegs
Mill
#8 drill
Drill out lower portion of the hole which is smaller diameter first
7
5 degree taper the hole pattern holes
Mill
15
Gives draft for making ejection easier
c.      Photographs of mold:
    CORE      CAVITY
        
d.      Process Parameters:
Injection Pressure: All 400 psi
Injection pressure was high enough to evenly spread plastic throughout the part; 400 psi was approximately selected initially and worked great through production
Injection Hold Time: 8 seconds
This was an initial parameter set, and worked well throughout testing.
Cooling Time: 30 seconds
This was the biggest issue--during optimization (as seen in the photos), the pin heads wanted to flare out and there was significant dimpling in the part; cooling time was almost doubled throughout the optimization process.
Set Screw Feed Stroke (shot size): 1.6 inches
As seen in pictures, initial shot size was not sufficient to fill the part so during optimization this was raised to 1.6 inchesPhoto 20-11-2015 13 42 00.jpg
Injection Speed Profile: All 3 in/second
Injection speed was initially low but was raised to completely fill the part.
Injection Boost Pressure: 1600 psi
Injection boost pressure was doubled over the course of optimization; the initial value of about 800 psi was not sufficient and caused problems with issues in the shape.
Intrusion Time/Speed: N/A
Screw Feed Delay Time: 17 seconds
This is a high-pressure process for this particular part and to mitigate these effects, screw feed delay time was raised.Photo 20-11-2015 13 40 27.jpg
Ejector Counter: 2
Standard; not modified throughout the course of optimization
⅛” Ejector Pins Length: 5.490 in.
Chosen based on geometry of mold.
Total Shim Thickness: 0
Chosen based on geometry of mold.
¼” Ejector Pin Length: 5.327 in.
Chosen based on geometry of mold.Photo 20-11-2015 13 41 16.jpg
Special Ejector Pin Length/Quantity: N/A
Measurements:
Outer Snap Fit Diameter- This is the most critical dimension on this part once again because this is where it interfaces the other components. This is the base piece that holds the whole yo-yo together, so if this dimension fails in either direction the yo-yo will either be impossible to assemble or will fall apart when used. Shrinkage here ended up being as expected, and was accounted for in this and the other parts so that they still fit together regardless.
Expected Value: 2.14in
Measured Value: 2.138in


2.      Summary of what we’ve learned, and issues we’ve come across:
One significant issue that we faced was that our base piece was experiencing extreme short shot, with as much as 1/4th of the part not being filled with plastic. This lack of plastic was exclusively experienced on the side opposite our one gate in the cavity mold. To fix this issue, we initially wanted to increase the shot size for the part, but noticed that with the current size, there was still a significant amount of plastic left in the screw of the machine. This meant that the plastic in the gate was most likely freezing over before the part was completely filled. To remedy this, we were able to file the gate down to a larger size, which resulted in a nearly perfect filling of the mold. We increased the injection pressure by about 10 Pa to get the last amount of needed plastic to finish the part correctly.  
The base’s second significant issue was that the press fit pin holes on the part were bending outward unexpectedly. This part required the most plastic out of all of our parts, and therefore experienced a significant amount of shrinkage.
The shrinkage from these bases also caused dimples to appear under each of the pin holes.
We were able to minimize the problems from the press fit pin holes and the dimples by increasing the cooling time.

In general, the biggest challenge our group faced during this process was mold machining; we ran into an unexpected ejector pin placement problem 3 days before the deliverable was due and re-machined the majority of our molds; we got very lucky in that parts fit together on the first try (and definitely learned a lot as a team from the experience), and we look forward to production runs and assembly!