Manufacturing Progress
Introduction
The design for the lotus flower yo-yo has been finalized and fabrication has commenced. The first step of fabrication involves machining the molds for all injection molded parts and 3D printing the die for the thermoformed lily pad. After some difficulties and setbacks, described in deeper detail in the subsequent sections, all molds have been machined and most parts have been injection molded. The thermoforming die for the lily pad is now in its second iteration, as the first iteration broke under the stress of the vacuum.
First, we have an analysis of a G-code sample from one of our machining toolpaths. Then the following sections are organized by the part name. Each part section includes the machining parameters for fabricating the mold, the injection molding settings (such as part volume), and optimization thought process.
Snap ring army
Stigma
Outer petals
Outer petals with stigma
Lily Pad Thermoform
G-Code
Stigma
Machining
The machining parameters of the Stigma molds were relatively straightforward, as we were not removing large amounts of material and many of the key features were created simply by deep-drilling operations.
First, we used 3/16” 120° Center Drill to start the five main holes in the mold, which are for both the ejector pins and the shafts of the stigmas. Second, we used a deep drilling operation to drill all the way through the core block. The machining parameters used for these main drilling processes can be found in the table below.
Our remaining passes were the adaptive clearing of the conical top shape of the stigmas and then a finishing scallop pass, followed by the contour pass to create the runners and gates. All of these passes used the maximum spindle speed of the ProtoTRAK of 5000 RPM, and we used the recommended feeds and speeds in the tool library for the ⅛” and 3/32” flat end mills used. We were able to run all passes at 100% feed rates once we visual confirmed that our G code was going to do what we planned, and we got used to brushing away chips at full speed while peck drilling.
Injection Molding
We used 4.25” ejector pins in each of the five stigma holes and achieved a stem length of 0.75”, which is close to the appropriate length to secure the lotus flower into the yo-yo body. To achieve a more accurate stem length we changed the ejector pin lengths to 4.5” and achieved a stem length of 0.5”, which was our desired length.
Injection molding parameters
Shot size: 12 mm
Cool time: 10 seconds
Packing time: 20 seconds
Packing pressure: 1,500 psi
Optimization
In an attempt to hone the injection molding settings for the five stigmas, our team attempted decreasing the part volume from 12 to 8 mm. However, this caused short shot in our parts with only one stigma fully filling. The other four partially filled with plastic, some only existing with the stigma’s dotted top. The part is shown to the left.
As another attempt to hone the injection molding settings for the five stigmas, as well as increase manufacturing efficiency, our team attempted to minimize the packing and cooling time before the part was ejected. We minimized the packing and cooling times both to 10 seconds. However, as demonstrated to the left, the part was often stuck in the mold and unable to be ejected, instead often deforming while held slightly in the mold. After attempting this, we reverted to the previous times of 20 seconds for packing and 10 seconds for cooling.
Our team noticed that the middle of the stigma top sinks in toward the middle instead of producing a flat surface for the dots. The curved surface can be seen to the left. This is occurring because of the part’s non-uniform thickness, where the area under the dotted surface is significantly thicker than the bottom rod of the stigma, and therefore the parts cool at different rates. Although our team enjoys the surprisingly natural appearance of a curved surface, we played with the injection molding parameters in an attempt to reduce the shrinkage.
To minimize the shrinkage in the part, our team adjusted the packing time, packing pressure, and shot size for the stigmas. We tried the extremes of each parameter to investigate if avoiding shrinkage was possible. Increasing the packing time to 30 seconds did not prevent shrinkage, nor did increasing the packing pressure to its maximum value (approximately 2,300 psi). By increasing the shot size from 12mm to 16mm, we produced undesired flash along the parting line.
We will continue to work on tweaking the parameters to reduce the shrinkage. Once we have adequately reduced the shrinkage, our team will evaluate which appearance we prefer and mass manufacture parts accordingly.
Body
Machining
Machining the two molds took approximately 3 hours to complete. Descriptions of the machining difficulties are described in the team’s most recent blog post (on April 7th).
Before the team can injection mold the body mold, a dowel must be press-fit into the center hole in the core mold. This dowel will be used to produce the hole that the stigma will press-fit into. After adding the dowel, we will be able to injection mold our first parts to see how the snap fit feels between the snap ring and the body.
Snap Ring
Machining
Starting with the core, we face the surface of the mold block then use a ⅜” bull nose end mill to run an adaptive roughing and finishing pass to remove a bulk of the inner stock and achieve a flat bottom surface finish. Still using the ⅜” bull nose to prevent losing time changing tools, we ran a circular operation to achieve a nice surface circular surface finish on the outer radial face of the snap ring. Finally, we ran a contour pass with the same tool to remove the inner 0.1" radius fillet and finished off the mold with a quick 2D contour to cut sprue path.
To machine the cavity, we used 3/16” center drill to start the five main ejector pin holes in the mold and then used a deep drilling operation to drill all the way through the core block. After facing the top of the mold, we used a ⅜” end mill, run an adaptive path to remove a bulk of the stock material and an adaptive finishing pass to get to desired mold height. Still using the 3/8" end mill, perform a circular operation on the two circular layers of the snap ring to achieve nice surfaces on the snap-fit interfaces. Using a 1/16" end mill, we removed the remaining corner radius in the snap ring wedge that snaps on top of the thermoformed lily pad.
Although CAM simulations suggested a total machining time of 20 minutes for the cavity and 17 minutes for the core, the actual machining time was approximately 2 hours.
Injection Molding
In our first attempt to injection mold the snap ring, we used 4.875” ejector pins along the circumference of the ring and a single 5.125” ejector pin to hit the inner notch of the snap ring. However, these lengths resulted in legs for the snap ring, as seen on the left, and therefore the ejector pin heights must be adjusted to minimize this effect. The legs were 0.225” long around the circumference of the snap ring, and 0.125” long on the center notch. Our initial shot size was too low (15mm) resulting in a few partially formed parts until we brought the shot size up to 22mm.
Injection molding parameters
Shot size: 22 mm
Cool time: 10 seconds
Packing time: 10 seconds
Packing pressure: 1,000 psi
Optimization
In an attempt to correct the lengths of the ejector pins, we added four white (0.025”) shims between the mold and the mount, then used 5.000” ejector pins along the circumference and a 5.125” ejector pin on the notch. However these were again incorrect heights, as the resulting legs were 0.205” long around the circumference and 0.25” long on the center notch. Our team will be calculating the next attempt more carefully instead of rushing through them, as done here. We will also consider removing the ejector pin from the center notch to assist with ejector pin height matching.
Inner and Outer Petals
Machining
Due to the geometry of these molds, we anticipated that these molds would be extremely sensitive to any z-height differences. If a tool z-offset was even slightly off, it wouldn’t just change the z-height of a feature in the mold, it would also change the radial dimension. For example, if the tool is actually higher than the programmed toolpath, there would be additional stock left in the z-direction, as well as the radial direction. This means that a z-offset would not change just the thickness of the piece, but could prevent the core and cavity from fully closing. To minimize the chances of having small errors in the z-offset we post-processed the operations separately and manually re-confirmed the z-offset by touching off the tool on the part surface after every tool change.
Machining a single side of one mold for each of the petal rings was approximately two and a half hours, majorly due to the careful finishing pass that is shown to the left. When this CAM step was about to start, the feed rate was decreased to 30%. This reduction in feed rate was to ensure that the 1/16” ball end mill was approaching to the appropriate height and successfully broke through its first depth of material left by the prior roughing pass.
During the first attempt at machining the outer petal core, the 1/16” end mill broke when attempting to follow the first petal down its curve, as seen to the right. To fix this, our team added a closer roughing pass that left less material to be removed with the finishing pass. This change allowed the 1/16” end mill to remove the material without putting too much lateral load on the tool and, once a couple of radial passes had been successfully executed, the feed rate was once again increased to 100% and the full speed finishing pass was able to continue.
Injection Molding
When injection molding the outer petals, we used 4.500” ejector pins along the circumference of the runner. This ring is not part of the cosmetic surface, as the runner gets snapped off at the tip of each petal, therefore the exact length of the ejector pin is not critical in this case.
Injection molding parameters
Shot size: 17 mm
Cool time: 10 seconds
Packing time: 20 seconds
Packing pressure: 1,500 psi
Optimization
The top nub of our outer petals core mold is a few thousandths too tall due to the z height issues discussed previously, and resulted in flash along the outer perimeter of our part. In the first couple cycles of injection molding, we incrementally reduced the injection molding volume from 20mm to 16mm to determine how much we could minimize the flash on the parts without causing any shortshots. As the machine heated up after multiple rounds of injection molding, the part began to adhere more to the surface of the cavity mold. Although adding a mold release spray helped the parts to eject more reliably, it also caused an increase of flash to appear around the edges of the petals, despite the lower injection volume. Therefore, we need to shave off a couple thousandths of an inch off the top of our cavity mold before further optimizing injection molding parameters
Lily Pad
Our first die was designed with a shell geometry, that did not leave much material thickness on the top face of the die. When we first started to use the die for thermoforming, we were able to get one good form before the die exploded. This was due to the plastic sticking to the die and then the vacuum pressuring being enough to shatter the die, due to its thin walls. The shattered die and the formed sheet that captured this explosion can be seen below.
After this failure occured, we re-designed our die to be thicker throughout the top face and be printed at a higher resolution. We then conducted multiple trials to determine the best parameters for our thermoforming process. The table below shows the different parameters for each trial and the main observations from each one, while the following figures show the visual results of each trial.
It is difficult to tell whether printing on the white plastic or clear plastic is the better option. We do not believe we will be able to make that decision until after we have all of our pieces ready to assemble together, so that we can see the full perspective and texture of the design. We did notice that the drilled air holes in the die were more visible in the white plastic thermoformed pieces than when using the clear plastic, but the white plastic does provide a much more vibrant color.
Blog post written on April 23rd, 2019