Вы находитесь на странице: 1из 3

Product Solutions

Wax Pattern Making: Can Variation in the Wax

Make a Difference?
by Jeffrey Rich, Vice President, MPI, Inc.

t is astounding that investment that are beyond the operator’s control or such variation has on the dimensional
casting has been around for some dificult for the operator to affect. Some stability of a part. Wax manufacturers
5,000 years, yet we still have trouble of these variables are batch-to-batch provide wax with speciic properties,
consistently making high-quality wax variations in the wax, settling of illers, which are documented and supplied
patterns. Of course, for those who nonhomogeneous wax temperature, air with each order of wax. Some of these
understand the nuances of wax and in the wax, and characteristics of the properties include ring and ball softening
wax injection, the problem is clear. die (feed, low path, venting of air, etc.). point, speciic gravity, congealing point,
However, for those who still live in the Environmental conditions may also melting point, ash content, thermal
dark ages of investment casting and affect the pattern-making operation, such expansion, viscosity (usually provided at
think of wax injection as nothing more as wax room temperature and humidity. a given temperature), and recommended
than a wax pump, the answers are a The ICI Process Control course teaches injection temperatures. Each of these
bit more elusive. Wax pattern quality that the best way to achieve a predictable properties has its own acceptable
is dependent on a litany of variables, output from any process is to control the tolerance. The question was, could a
and making the perfect pattern every key process input variables. As you can variation within the wax manufacturer’s
time can seem like pure luck or even see from the above list of process input properties tolerance cause a change in
black magic. Truth be told, it may be variables for wax injection, several are the process output resulting in bad wax
impossible to get a perfect pattern likely important, but they are not under patterns?
every time, but with a process that is in the control of the operator.
control, it is possible to get predictable Recently, MPI’s Pattern Production Experiment Setup
results from each injection. department experienced shifts in • Select wax pattern
The measure of a quality wax dimensional characteristics that appeared • Select wax
pattern falls into two categories: visual to coincide with changing from one batch • Build a measurement jig to measure
and dimensional. Both categories are of wax to another batch. This dimensional sink (cavitation) in wax pattern
impacted by the manipulation of process change in the wax patterns negatively • Conduct measurement
input variables. The variables commonly affected the automated assembly process, system analysis (MSA)
accepted as “key” to the process are in some cases bringing assembly to a stop. • Conduct experiment -
wax temperature, wax pressure, wax MPI conducted an experiment to look Measure multiple patterns
low rate, die temperature, and dwell speciically at the impact of batch-to- from multiple wax batches
time. Additionally, there are variables batch wax variation and the impact that • Analyze results

Figure 1: MSA-graphical results Figure 2: MSA-statistical results

18 ❘ March 2015
Figure 3: Pattern Side A Figure 4: Pattern Side B

For experimental purposes, MPI 24 measurements per pattern, or 120 irst batch. The conclusions were drawn
selected a ive-bar injected runner and measurements per wax batch (Figures 3 from hypothesis analysis using ANOVA
a reclaimed wax. Each batch of wax and 4). of all four batches and multiple runs of
met the same quality standards and was The wax patterns were all injected 2 Sample T-Tests for each batch against
within the manufacturer’s acceptable using the same injection parameters all other batches. The box-plots of the
speciications for each property tested. on the same die on the same MPI wax measurement data also illustrate this
MPI designed and manufactured the jig. injection machine. The recipe used (Figures 5 and 6). Note the distinct
Next, an MSA was performed to prove a wax temperature of 123°F, wax low pattern in the irst batch, compared to
that the measurement system would rate of 4in3/second, wax pressure of 450 the more random measurements in the
provide high-quality, reliable data. MPI pounds per square inch, injection time second batch. Batches 3 and 4 were
did not want the total process variation of 200 seconds, and a die temperature so statistically similar to batch two that
to be signiicantly inluenced by the of 60°F. MPI performed several different they are not shown separately. While
variation in the measurement system. The methods of quantitative analysis on the we intuitively had known for years that
MSA used standard protocol: ive parts, measurement data, including ANOVA, the batch-to-batch variation in the wax
ten measurements, three operators, and 2 Sample T-Tests, descriptive statistics, causes signiicant variation in the wax
three randomized replications. The data and normality testing. Additionally, MPI pattern, up to and including causing
was analyzed using Minitab statistical reviewed several types of qualitative bad parts, it was exciting to have solid
software and the results showed a very tests, including histograms, box-plots, evidence. Now we wanted to know
robust (repeatable and reproducible) and scatter plots. what had changed in the wax enough to
measurement system (Figures 1 and 2). After reviewing the data it was have this much variation show up in the
In the experiment, MPI Pattern determined that, the irst batch of wax wax pattern.
Production measured four different was statistically different from the other It did not take long to determine
batches of wax, taking measurements three batches. The three other batches the culprit. The wax manufacturer had
from 12 points on each side of the were statistically the same. No individual provided all the standard properties
wax pattern (two-sided) for a total of batch was statistically the same as the Continued on pg 20

Figure 5: Batch 1 Measurements Figure 6: Batch 2 Measurements

March 2015 ❘ 19
Product Solutions
and tolerances for the wax. Each
batch had some variation on some
of the properties, but all were within
the speciied tolerances. MPI had
requested full viscosity curves for each
wax to include the paste range (see
Figure 7). The difference in the batches
showed up in the viscosity curves.
Based on the information provided
by the manufacturer, it is impossible
to determine why the shift in the
viscosity curves occurred. It is clear
that the wax from batch 1 resulted in
statistically different patterns than did
the other three batches of wax. The Figure 7: Test Wax Batches Viscosity Curves
viscosity curves for wax batches 2, 3,
and 4 showed only minor variations,
speciically at injection temperature. additional expense. As the tolerances About the Author
Statistically speaking, we are of the delivered parts to the customer Jeffrey Rich is the vice-president and general
dealing with two different waxes. The have become tighter and tighter, casting manager of MPI, Inc. Rich is a Six Sigma
properties measured and advertised the part with little to no machining has Master Black Belt and has used the tools of
by the manufacturer met the tolerance grown even more dificult. The demand variation reduction and lean manufacturing
speciications, yet they were not for the eficient production of inished as an executive at General Electric and
truly the same wax. This is cause for parts has increased. The wax pattern is Textron. He has brought those skills to MPI,
concern if you are striving for a highly instrumental in delivering that casting. where he champions manufacturing and
controlled process. To help ensure that It makes sense that the industry should equipment innovations that have improved
you understand the potential variation focus on continuous improvement in MPI’s ability to manufacture and deliver high-
a given wax batch will have on your the wax room and process controls quality equipment to take advantage of the
process, it is important to require the wax that now allow us to make “net shape” industry’s leading process control standards.
manufacturer to provide the full viscosity wax patterns part of the goal. In order Rich participates in the ICI Process Control
curve. Pay close attention to the wax to continue to be a viable, proitable course as an instructor and has recently
viscosity at the injection temperature. If industry, we must strive to improve accepted a position on the ICI Education
a new batch of wax indicates a viscosity every area of the casting process and Committee.
change at your injection temperature, to educate ourselves on how to control
you should anticipate a change in the our processes and eliminate variations, About MPI
quality of the injected pattern. This even in areas never before considered. MPI is a leader and innovator of wax room
could be manifested as a dimensional Competition continues to drive more equipment, featuring a broad range of wax
and/or visual quality change in the and more automation into the wax room, injection machines that generate higher
injected pattern. and automation demands the reduction casting yields with increased productivity
MPI recommends that foundries of variation beyond the previously and throughput. As the acknowledged leader
require wax manufacturers to provide accepted tolerances in materials, in wax room innovations, MPI has developed
the viscosity curve for each batch of wax. equipment, and production. and introduced more systems and products
By comparing the viscosity curves, you The more your operators and wax than all of its competitors combined.
can determine the wax viscosity at your room engineers understand the process In addition to producing industry-
desired injection temperature. Your wax and the effects of wax variation, the leading wax injection systems, MPI offers
room engineer can use this information more likely you are to produce perfect total wax room automation, from wax
to help optimize your injection parts every time—or at least most of the injection to wax assembly, using integrated
parameters. In our industry, process time! The ICI Process Control course robotics technology. Automation also
optimization through process control has teaches the importance of reducing allows MPI to cost effectively provide
become critical to remain competitive. variation in all aspects of your process. pattern and assembly services, which can
For more than 60 years, the investment I hope this experiment helps illustrate be shipped directly to their customers. A
casting industry has had as its goal the the critical nature of process control and world-class customer service team featuring
production of “near net shape” castings. shows that variation can exist anywhere the Global Support Services program backs
Unfortunately, secondary machining in your process. all MPI equipment. To learn more about MPI,
operations remain a given and necessary visit www.mpi-systems.com.

20 ❘ March 2015