Injection Molding DOE

PLET 323

Team 2

Matt McLouth
Brian Black
Shane Appel
Andrew Cunningham

Section 1

Date Performed: 01/11/10

Date Due: 02/01/10

 Abstract Statement of Theory and Definitions

In this experiment, the primary objective was to Injection molding is a common process for creating
observe how processing parameters impact part warpage. complex parts that require exceptional surface detail and
The parts molded were five sided boxes made of High high tolerances. Injection molding can create a wide range
Density Polyethylene (HDPE). The parts width was being of parts that are only a few micrometers to a couple of
measured to check the amount of warpage. An eight run meters. Injection molding is used to create a large variety
Design of Experiment (DOE) was run in order to evaluate of parts ranging from medical equipment to children toys.
the processing parameters of an injection molding Some advantages of injection molding include high surface
machine. The parameters tested were Barrel Temperature, finish, high tolerances, create complex shapes, quick cycle
Injection Velocity, Pack Pressure, Metering Time, and times, and the ability to quickly make a large volume of
Cooling Time. The processing parameter that demonstrated parts. Part quality depends upon the mold design, part
the largest effect on warpage was Pack Pressure. design, processing conditions, and material.

Introduction
Warpage is a reoccurring defect that continues to
plague the injection molding industry despite numerous
efforts to understand and correct this phenomenon. It is
important to be able to understand and control this type of
defect in an effective and timely manner. If warpage is
occurring, parts could fall out of spec or even fail further Figure 2- Injection Molding Machine
testing. A Design of Experiment (DOE) was conducted in
order to evaluate the effects of various processing
parameters on an injection molding machine and their
relationship to part warpage. The Y direction in Figure 1
was measured in order to quantify the effects of warpage.
The parameters that were adjusted in order to affect
warpage include Barrel Temperature, Injection Velocity,
Pack Pressure, Metering Time, and Cooling Time. The
material selected in this study to evaluate the effect of
warpage was a semi crystalline high density polyethylene.
Figure 3 (Injection Barrel)

The first injection molding machine was introduced in

1872 by the Hyatt brothers. The original injection molding
machines used a plunger to compress the plastic and
heating elements to aid the plastic in becoming molten
more quickly. Since Celluloid was the only plastic material
around for the next fifty years, the injection molding
machine design stayed relatively the same for that time
period. [1]

An injection molding machine was used for this

experiment for several reasons. An injection molding
machine is capable of producing large quantities of parts
with relatively short cycle times. This is important because
it allows the operator to observe changes quickly. Another
benefit of this process is the ease of making process
changes. This aspect is crucial to experiments because the
part may reflect the process change within one shot. This
allows the operator to be efficient and timely. Ability to
Figure 1 – Five Sided Box and Measurement Direction produce high pressures is particularly influential in this
experiment so that the effects of Pack Pressure can be
determined.

1
 Despite numerous advantages of the injection molding next phase is the Compensation or Pack phase. This phase
machine, there are also disadvantages to this process. is where the machine injects plastic at a slower rate to
Drawbacks could include high mold costs; this could be a prevent spiking of the injection pressure and flashing the
problem if a mold has to be created for the experiment. mold. Flashing the mold is when the mold is forced open
Operation can also be increasingly difficult when initially and some material leaks in the opening that is created.
setting up a process and when performing a DOE. Setting Another factor that is accounted for in this phase is the
up and running an injection molding process requires a high shrinkage of plastic materials. Plastic materials shrink
trained operator because there are always unexpected any where from .05mm/mm to1.02mm/mm. This can cause
problems that occur. The operator needs to be as pro-active major problems unless it is accounted for. As the part is
as possible in order to maintain a stable process that cooling it is constantly shrinking. To compensate for this
produces consistent parts. Cooling of the mold could shrinking the packing pressure is kept on after the part is
become a problem and could affect results. If cooling is fully filled. This forces material into the mold as the part is
inconsistent, parts could experience differences in shrinking thus compensating the part. During this phase,
shrinkage. the next shot is being built.

Figure 1 shows the entire injection molding machine Warpage is the result of differences in part shrinkage.
and all of its components. The entire process begins by Non-uniform shrinkage can be caused by many different
adding material into the hopper; this can be done manually factors such as, variations in part cooling, part thickness,
or pneumatically. Material typically comes in pellet form molecular and fiber orientation, and degree of crystallinity.
and in some cases needs to be dried due to their These variations cause residual stresses to be molded into
hygroscopic nature (Nylon, PC, PET, and ABS). The the part. When these stresses are relieved they can lead to
material is then fed from the hopper into the heated barrel warpage.
through the feed throat. Once in the barrel, the material
then begins to move forward by the screw, which acts like There are several process parameters that can affect
an auger to push, compress, and melt the pellets. Frictional part warpage, some of these variables are Barrel
heating and shearing of the material caused by the screw Temperature, Injection Velocity, Pack Pressure, Metering
allows the material to be heated to the processing Time, and Cooling Time. The Barrel Temperature is set in
temperature. Heater bands are used to heat the barrel and order to heat the material to a melt temperature commonly
are located strategically across the barrel. These heater recommended by the supplier. The barrel is heated by
bands are usually placed near the three zones of the screw using heater bands or by electric heated coils. A higher
which include Feed, Compression, and Metering. Barrel Temperature is going to cause the melt to become
less viscous, therefore making it easier to push. As the melt
The Feed Zone initiates the melting process and is heated, it allows for more space between the polymer
advances the material. In this zone, the flight depth (depth chains. The more space there is between the chains, the
of the threads) is the largest in order to soften the material. easier the chains can slide past one another. If the chains
The next zone is referred to as the Compression Zone are sliding easier, the polymer is going to have less
where the material melts. The flight depth of this zone resistance to flow.
transitions into the metering. Finally, the melt reaches the
Metering Zone where the material is mixed to form a Injection Velocity controls the rate at which the screw
homogenous melt. The Metering Zone is where the shot is moves forward and how fast the material is injected into
accumulated, and where the material is pumped into the the mold. Varying Injection Velocity can cause differences
mold. The flight depth is constant for the most part in order in viscosity. A faster Injection Velocity yields a higher
to finish the melting process. Most of the heat is generated degree of molecular orientation thus leading to a less
from the shear and frictional heating between the polymer viscous material entering the mold.
chains; this also causes the molecular chains to become
oriented. The reciprocating screw feeds the material to the Pack Pressure is used in order to lock in the molecular
front of the screw to create the shot. The material in the orientation caused by Injection Velocity. By locking in the
front of the screw is the total weight of the resin that is to oriented state, the material becomes more stressed. After
be injected (often ranging from 20g-20kg).[2] this orientation is established, Pack Pressure begins to
compensate for shrinkage and fills the last 2-5% of the
The screw is then pushed forward at a specified mold. By compensating for shrinkage, the cooling
speed, commonly referred to as Injection Velocity. The becomes more effective because there is more contact
motion of the screw moving forward forces the melted between the mold and the part.
plastic into the mold. Once the material is in the mold, the
material goes through a phenomenon called shear thinning Metering Time or Recovery Time is the time it takes
which could cause imbalances both in the cavity and the the screw to build the next shot. A longer Metering Time
runner system. After the cavities are filled 95-98%, the increases shear heating by manipulating the material
2
longer. An extended Metering Time can ultimately lead to plate. This separation is useful because someone does not
thermal degradation of the polymer. have to degate the part. This prevents more handling of the
parts and helps to remove possible error.
Cooling Time is set in order to allow the part enough
time to solidify before it can safely be ejected. An Another critical piece to the experiment is the
extended Cooling Time allows more heat to be extracted thermolator. A standard thermolator was used to control
from the part. Cooling Time is used to lock in established the mold temperature. This is important because if the
orientation caused by Pack Pressure and Injection Velocity. mold temperature is not controlled then it can rise
uncontrollably and increase cycle times. This rise in
DOE’s are commonly used in order to observe and temperature causes the parts to not cool thoroughly and can
evaluate interaction effects between variables. A major potentially eject parts in a molten state. If a part is ejected
advantage of this analysis is that it provides more to early, the part could experience premature failure.
information quicker than other analysis’s such as One Therefore, a constant mold temperature would help remove
Factor at a Time (OFAT). A DOE intentionally makes one of the possible sources of variation and error. The
large systematic changes to a process by setting high and final critical piece of equipment for this experiment was a
low values for each parameter in order to create the pair of dial calipers.
intended defect. In this case warpage is being measured.
[3] Application of Equipment and Processes
The material used for this procedure was a semi-
crystalline High Density Polyethylene (HDPE). HDPE is a The DOE was setup using an optimized process on the
good material for experimentation because it is a relatively injection molding machine. A high and low was set for
versatile rigid material. HDPE is generally easy to process, each of the factors in order to determine which process
inexpensive, and is highly recyclable. Though HDPE is a settings had the largest impact on part warpage. Table 1
versatile material, HDPE has poor weatherability, low lists all the process parameters used in the DOE, and the
dimensional stability, and low tensile strength. Despite high and low values that were set for each. The second
these disadvantages, they can often be improved with the columns under the high and low settings contain values
addition of additives and fillers. that have parenthesis around them. These values were set
in order to achieve the high and low factors.

Description of Equipment & Process Factor Low High

A Barrel
282 ° 293 °
Temp
The first and primary piece of equipment is the C C
Profile
injection molding machine. The injection machine that was
used in this experiment was the Demag Erogtech (model B Inj.
1.4 (.927cm 12.7 (1.30cm
number: 7144-0017). This machine was chosen because Velocity
cm/s xfer) cm/s xfer)
the machine met the requirements to fill the mold, the mold (Fill Time)
met the tie bar requirements, and the clamp force required Pack 7.6 (20% lo 30.3 (80% hi
C
to hold the mold closed was met and surpassed. While this Pressure MPa inj) MPa inj)
was easily achieved on the machine it is something that Metering
(185 (85
needs to be considered. Time
D 3.1 s RPM 9.3 s RPM
(Screw
act) act)
The next most important piece of equipment is the RPM)
mold. The mold is a simple three plate mold that provides Cooling
automatic degation. The runner system is a standard H Time
E 12 s (~ 35 s) 17 s (~ 25 s)
pattern leading to four cavities. For this experiment only (Cycle
one flow channel was used and the others were shut off by Time)
rotating inserts in order to eliminate cavity to cavity
variations. The plastic drops into the mold through the gate Table 1 (DOE Factors)
which is centered on the bottom of the part. A pinpoint
gate was used to assist automatic degation. This central
gate location allows for the material to evenly flow around
the core and prevent filling problems. When the mold
halves are splitting open to eject the part from the mold the
runner and gate are separated from the molded part. To
achieve this, the mold requires an additional plate. The
other two plates in the mold are the cavity plate and core
3
Table 2 (DOE Matrix)

Table 2 shows a number of variables with a + or –

beside them. The + stands for all high values and the –
stands for all the low values of each factorial.
Figure 5 (Average Chart)
Once the process was optimized the first run was
preformed with ten parts collected. Then each process Each point on Figure 5 shows the average part width
settings were changed to the second run in the DOE for all ten parts for each of the eight runs. The data
matrix. The process was run for five minutes to allow the collected from the experiment showed that all of the
process to stabilize before the next ten parts of the second subgroups were out of the upper and lower control limits
run could be collected. This process was repeated until all showing the data was out of control. This means that
eight runs were completed. 48 hours after the parts were changing the processing parameters has an effect on how
molded the width of the five sided box was measured along much the parts warped and proved that the data collected in
the Y axis (Figure 1). this experiment was valid. [3]

Presentation of Data and Results

Figure 6 (Scree Plot)

The above scree plot shows each of the factors and the
Figure 4 (Range Chart) order of significance. Figure 6 shows that Pack Pressure
(factor C) had the largest effect on part warpage. The other
Figure 4 shows the range chart of the completed DOE. factors Barrel Temperature (factor A) Injection Velocity
The variation within subgroups is on the Y Axis and the (factor B) and Cooling Time (factor E) according to the
subgroup number is on X Axis. All the data points are scree plot had little to no effect, while Metering Time
below the upper control limit therefore the range chart is (factor D) could be considered noise.
said to be in-control. When this graph is in control it means
that there was consistency within the subgroup variation. The two components of a scree plot are the “cliff” and
This showed that changing the process parameters had an the “rubble”. Data that is considered to be in the “cliff” is
effect on the amount of warpage. [3] data that appears higher on the plot. For data to be
considered “rubble” it must not appear to have any height
on the plot. Data that is “rubble” is the result of random
variation or noise. [3]

4
 Another factor that had an affect on part warpage was
Cooling Time. It plays a significant effect on the part
warpage for several reasons. Cooling Time is directly
related to the parts geometry, the more complex the
geometry becomes the more difficult it is to cool the part.
This is especially true when you have a 90° corner. The
inside corner is harder to cool due to heat concentration
caused by the geometry. This heat concentration causes the
inside of the part to be ejected at a hotter state. Therefore,
Cooling Time is crucial to decreasing this warpage. A
longer the Cooling Time permits more heat to be
conducted to the mold. If more of the heat is conducted to
Figure 7 (Estimated Effects) the mold from the inside of the corners then the part is less
likely to warp. If Cooling Time is not sufficient, the part
The estimated effects graph labels each factor, but not may be ejected in a partially molten state. If the part is
in order by their significance. As seen before in the scree ejected under these conditions, the orientation may be lost
plot the estimated effects plot shows that Pack Pressure due to relaxation of the chains. When the chains relax the
(factor C) is the most significant factor. The estimated potential for warpage drastically increases due to
effects graph shows the difference between the average of
differences in shrinkage.
the high and low levels. If the bar has a negative value that
means the lower factor had a greater effect on warpage and Injection Velocity and Metering Time both had
vice versa. The estimated effects graph shows which factor minimal effect on part dimension according to the
has the most significant effect on warpage. [3] estimated effects plot. Injection Velocity had some effect
because a change in velocity alters the orientation of
Interpretation of Data polymer. A higher velocity yielded a larger effect on part
dimension because there is a higher degree of orientation
Warpage is caused by differences in shrinkage within increasing the shear stresses and the potential for warpage.
a part. A couple of examples that cause differences in Metering Time had little to no effect on part warpage
shrinkage include inadequate cooling to some regions of because the Metering Time was not long enough to cause a
the part and differences in wall thickness within the part. temperature effect. The DOE verifies that the Metering
The data shows that Pack Pressure had the greatest effect Time has no effect on warpage, it can be concluded that the
on the parts warpage. To explain why Pack Pressure had results are just noise. [6]
the greatest effect on part warpage there needs to be an
understanding how and what Pack Pressure does during Data collected during the experiment was all inserted
injection. Pack Pressure typically is initiated when the part into the DOE spreadsheet to be analyzed. The range chart
is approximately 95% to 98% filled. Prior to the packing (Figure 4) and the average chart (Figure 5) verify that the
phase the polymer chains are highly oriented due to DOE was successful and the data was relevant.
Injection Velocity. Pack Pressure retains this high level of
orientation by forcing the chains to remain in a highly Possible sources of error may include running the
stressed state. Pack Pressure continues to fill the part out machine in semi-automatic, damaged tooling, and
until the part is completely filled and almost becomes measurement error. Running the machine in semi-
hydrostatic. [4] automatic could have increased the residence time leading
to material degradation. Damaged tooling could have
Although Barrel Temperature was not as significant as caused the part to stick to the mold further deforming the
the Pack Pressure it still caused differences in part part. Dial calipers were impossible to center on the part
dimensions. The parts experienced a classic hourglass warp every time and could have led to poor measurements.
due to insufficient cooling in the corners (see Figure 1). [5]
The inside corners experience more heat concentration Conclusion
than the outside corners due to the limited geometry of the
core half. If the increased heat is not efficiently extracted,
Warpage is a dangerous phenomenon that can cause
the inside corners may shrink more than the outside
serious part deformation and can lead to part failure. To
corners. This difference in shrinkage caused inside corner
determine the processing parameter that had the greatest
to pull the sidewalls inward to form the hourglass shape. A
effect on the part warpage, an eight run DOE was
higher Barrel Temperature compounded this effect by
performed. The results of the DOE showed that Pack
increasing the melt temperature. As the melt temperature
Pressure had the largest effect on the part’s warpage. The
increases the inside corners become more difficult to cool.
5
results determined in this experiment are crucial to References
executing and maintaining a stable process. Without these
results, defects like warpage would be hard to diagnose and [1] "History of Plastic Injection Moulding." History of
correct in a timely manor. Plastic Injection Moulding & Rotational Moulding D&M
Plastics Inc., 16 June 2008. Web. 23 Jan. 2010.
Future work could be done to verify this data and to <[Link]
examine other possible factors. Running a reflected DOE [2] A.B. Strong, Plastics Materials and Processing,
could clear up the data. Another possibility is to run a new Pearson Prentice Hall, Columbus, pp. 435-439 (2006).
DOE and remove Pack Pressure as a factor. Metering Time [3] B.G. Johnson, Plastics Processing and Statistics,
could also be further examined to determine if degradation (2008).
could influence part warpage. Therefore, continued [4] Kramschuster, Adam. Cavitt, Ryan. Ermer, Donald.
research into part warpage is the only way to fully Chen, Zhongbao. Turng, Lih-Sheng. “Quantitative Study
understand this phenomenon and effectively control it. of Shrinkage and Warpage Behavior for Microcellular and
Conventional Injection Molding.” POLYMER
Acknowledgement ENGINEERING AND SCIENCE (2005) : 1408-1418.
[5] J.P. Beaumont, Runner and Gating Design Handbook,
Hanser Gardner Publications, Cincinnnati, pp. 28 -34
The success of this experiment depended largely on (2007).
the encouragement and guidelines provided by the faculty [6] J.M. Fischer, Shrinkage and Warpage, Plastics Design
of Penn State Erie, The Behrend College. Special thanks to Library/ William Andrew Inc., pp. 1-10, 82-91 (2003).
Jon Meckley, John Beaumont, and Bradley Johnson.
Appendix

Raw Injection Molding DOE data sheet (attached

to back).

6
7