close

Вход

Забыли?

вход по аккаунту

?

Use of SPC methods via data acquisition to control shrinkage and warpage of a microwave oven switch housing

код для вставкиСкачать
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may be
from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted.
Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in reduced
form at the back of the book
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly to
order.
UMI
A Bell & Howell Information Company
300 North Zed) Road, Ann Arbor MI 48106-1346 USA
313/761-4700 800/521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Use of SPC Methods via Data Acquisition to
control Shrinkage and Warpage of a
Microwave Oven Switch Housing
by
Joachim Ruckdeschel
B.S., Technical College o f Rosenheim, Germany (1989)
SUBMITTED IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF MASTERS OF SCIENCE
DEPARTMENT OF PLASTICS ENGINEERING
UNIVERSITY OF MASSACHUSETTS LOWELL
Signature of Author:
!j6w.
Date: H
1
Signature of Thesis
Signatur
Committee Members:^
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
*5?
UMI Number: 1384185
UMI Microform 1384185
Copyright 1997, by UMI Company. Ail rights reserved.
This microform edition is protected against unauthorized
copying under Title 17, United States Code.
UMI
300 North Zeeb Road
Ann Arbor, MI 48103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Use of SPC Methods via Data Acquisition to
control Shrinkage and Warpage of a
Microwave Oven Switch Housing
by
Joachim Ruckdeschel
ABSTRACT OF THESIS SUBMITTED TO THE FACULTY OF THE
DEPARTMENT OF PLASTICS ENGINEERING IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTERS OF SCIENCE
UNIVERSITY OF MASSACHUSETTS LOWELL
1997
Thesis Supervisor: Dr. R. Nick Schott
Professor, Department of Plastics Engineering
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
The objective of this research was to study the effects of different molding
parameters on the product quality. Quality was primarily determined by the part
dimensions of a switch housing for a microwave oven. In a pre-experimental computer
simulation the most significant processing variables were found. Many of the injection
molding processing variables were changed to find the specific effect of each parameter
on the molded part. In this exercise it was seen that the mold temperature had the greatest
influence on the part size. Additional simulations were carried out to find out if it would
be possible to eliminate the annealing process, which was still required after the original
molding process.
Since the housing is a part of an oven, a short time exposure to a temperature of
160°C was required without showing any defects afterwards in order to meet the demands
of the quality assurance. This was the reason to go with PBT (Poly Butylene
Terephthalate), which shows unfortunately a post- mold shrinkage up to 2%.
In conjunction with the research mentioned above, the best operating point was
determined by using online SPC methods via data acquisition. For this investigation three
variables which had the most influence on the dimensions of the housing were varied
(holding pressure, holding time, back pressure).
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGMENTS
I would like to express great thanks to Dr. Schott for all his patience and great help
during the long process of writing this thesis, his guidance and knowledge in conducting
my research at a big electrical company in Germany for the University of Massachusetts,
Lowell. Moreover I also express my thanks to Dipl.-Ing. Axel Wrana, the leader of the
Plastics Department o f the Bosch Siemens company and to all the engineers. They were
all available if I had any questions and gave me many ideas for my work. Most of all, I
cannot thank my parents enough for both their mental and financial support. Without their
strong faith in me, I would not have been able to write this thesis. Furthermore, I would
like to thank Petra Zelger a friend of mine who reminded me all the time to finish this
fantastic challenge. Last but not least I give my special thanks to the people of the Plastics
Engineering Department at U Mass. Lowell who made it possible to finish the work even
after a long period of time.
iii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Contents:
Page:
TABLE OF CONTENTS.............................................................................................iv
LIST OF ILLUSTRATIONS..................................................................................... vii
I.
1.1
INTRODUCTION....................................................................... I
Objectives.................................................................................... 3
0.
QUALITY....................................................................................... 4
m.
FACTORS INFLUENCING QUALITY.......................................4
3.1
3.1.1
3.1.2
3.1.3
M old..............................................................................................
Gating............................................................................................
Cooling..........................................................................................
StifJhess..........................................................................................
3.2
3.2.1
3.2.2
3.3
3.3.1
3.3.2
Maschine...................................................................................... 7
Settings........................................................................................... 7
Power............................................................................................ 9
Operator........................................................................................... 9
Repeatability......................................................................................9
Forgetfiilness.of Operator.................................................................9
3.4
3.4.1
3.4.2
Method...............................................................................................10
Degree of Automation....................................................................... 10
Quality Assurance............................................................................. 10
3.5
3.5.1
Environment...................................................................................... 10
Machine Placement and Temperature............................................... 10
3.6
3.6.1
Material.............................................................................................11
Material Properties.......................................................................... 11
IV.
METHODS OF QUALITY ASSURANCE...................................11
4.1
TQM (Total Quality Managment).................................................... 14
iv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
6
6
7
4.1.1
4.1.1.1
4.1.1.2
4.1.1.3
4.1.1.4
4.1.2
4.1.2.1
4.1.2.1.1
4.1.2.2
Off- Line- Procedure.......................................................................... 14
QFD (Quality Funktion Deployment)............................................... 14
FMEA (Failure Mode and Effects Analysis)..................................... 14
STEM (Statistic Experimental Method)............................................ 16
Capability Determination................................................................... 17
On Line-Procedure.............................................................................20
SPC (Statistical Process Control)......................................................20
QCC (Quality Control Chart) and PCC (Process Control Chart)... 21
CPC (Continuous Process Control).................................................. 26
4.2
Summary Quality-Assurance............................................................ 27
V.
PRIME OBJECTIVES....................................................................28
5.1
5.2
5.2.1
5.2.2
P art................................................................................................
Material.........................................................................................
Production of the Material..............................................................
Material Properties.........................................................................
5.3
5.4
Injection Molding Machine ES 700/200......................................... 33
M old.............................................................................................. 35
VI.
REQUITRMENTS OF THE HOUSING.......................................37
VH.
INVESTIGATIONS AND TEST PROCEDURES.......................38
7.1
7.2
Used Equipment............................................................................. 38
Moisture Content............................................................................... 38
7.3
7.3.1
7.3.2
7.3.3
7.3.4
7.3.4.1
Original Process................................................................................40
Initial Settings................................................................................. 40
Mold Surface Temperature............................................................. 42
Temperature-Distribution during Annealing..................................... 43
Machine Capability........................................................................ 44
Error Analysis................................................................................. 45
7.4
7.4.1
7.4.1.1
7.4.1.2
7.4.1.3
7.4.1.4
7.4.1.5
7.4.1.6
7.4.1.7
Process Analysis................................................................................46
Quality determinig Parameter......................................................... 46
Holding Pressure............................................................................. 50
Holding Time.................................................................................. 53
BackPressure................................................................................. 56
Injection Speed............................................................................... 58
Mold Temperature.............................................................................63
Cooling Time.....................................................................................65
Summary Qualitaty-determining Parameter...................................... 67
V
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
30
30
31
7.4.1.8
Results: Quality-influencing Parameters........................................... 70
7.5
7.5.1
7.5.2
Quaiitaty influencing Parameter during Annealing........................ 71
Temperature and Time.................................................................... 71
Results Annealing........................................................................... 72
7.6
7.6.1
7.6.2
7.6.2.1
1 .6 2 2
7.6.2.3
1 .6 2 A
1.62.5
Oil Heating..................................................................................... 74
Settings........................................................................................... 74
Production with Oil Heating........................................................... 75
Holding Pressure Profile ”6 0 "....................................................... 76
Holding Pressure Profile "70".......................................................... 78
Holding Pressure Profile "75"........................................................... 80
Summary: Oilheating....................................................................... 81
Result Oil Heating............................................................................. 84
7.7
7.7.1
7.7.2
Summary and Conclusions.............................................................85
Material.......................................................................................... 85
Process.............................................................................................. 85
VIE.
RECOMMANDATIONS................................................................87
Literature Cited........................................................................................................ 89
Biographical Sketch of Author................................................................................ 91
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table of Illustrations
No.____________________ Title___________________________________Page
Figure 1:
Micro Wave Ove Switch Housing
............................................... 1
Figure 2:
Using SPC for Optimization.............................................................. 3
Figure 3:
Influence of "6M" on Injection Molding (Ishikawa Diagram)
Figure 4:
Classification o f Process and Product..............................................7
Figure 5:
Source of Errors and their Remdiation.............................................. 12
Figure 6:
Japanese und Western Products from 1950-1982..............................13
Figure 7:
Do it right the first time!.................................................................. 15
Figure 8:
Percentage of Standard Distribution of Standard Distribution
Figure 9:
Quality Control Chart (QCC)........................................................ 21
Figure 10:
Example of QCC with "Trend Development"............................... 23
Figure 11:
Process Control Chart (PCC).............................................................26
Figure 12:
Quality Assurance........................................................................... 27
Figure 13:
Specified Dimensions of the Part.Part...............................................29
Figure 14:
Production of PBT............................................................................. 31
Figure 15:
Waterabsorption of PET as Function of Time.................................. 32
Figure 16:
Injection Molding Machine Engel ES 700/200.................................33
Figure 17:
Process Data Diagram CC 9 0 ......................................................... 33
Figure 18:
Cooling Channel Layout for the Cavity Side.................................... 35
Figure 19:
Cooling Channel Layout Core Side................................................ 36
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
18
Figure
20: Clamping Fixture for Housing........................................................... 40
Figure 21:
Mold Temperature for the Original Process...................................... 42
Figure 22:
Temperature-Profile during Annealing..............................................43
Figure 23:
Position of Distribution and Machine Capability Factor................. 44
Figure 24:
Action Steps to Analyse the Process..................................................48
Figure 25:
Variation of two Parameters X 1and X2 within a coarse and a
precise Experiment...........................................................................48
Figure 26:
Post Mold Shrinkage of PBT B 4520 Original Part....................... 49
Figure 27:
Effect of the Holding Pressure on the Part Weight.......................... 5 1
Figure 28:
Length 1and 1' as Function of Holding Pressure...............................52
Figure 29:
Length I and 1' as Function of Holding Pressure after Annealing.... 53
Figure 30:
Determining Gate Seal off Time........................................................54
Figure
31: Weight as Function of Holding Time................................................ 55
Figure
32: Holding Time and Dimensions..........................................................55
Figure 33:
Dimensions after Annealing as Function o f Holding Time...............56
Figure 34:
Part Weight as Function of Back Pressure........................................ 57
Figure 35:
Dependence of Length on Back Pressure.......................................... 57
Figure 36:
Dimensions after Annealing as Function of Back Pressure.............. 58
Figure
37: Filling Study of the Housing..............................................................58
Figure 38:
General Injection Speed Profile...................................................... 59
Figure 39:
Weight as Function of Injection Speed............................................. 60
Figure 40:
Influence of Injection Speed on the Length...................................... 6 1
Figure 4 1:
Length after Annealing as Function of Injection Speed....................62
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 42:
Effect of Injection Speed on Melt Temperature................................ 63
Figure 43:
Weight vs. Mold Temperature...........................................................64
Figure 44:
Dimensions vs. Mold Temperature....................................................64
Figure 45:
Dimensions Dependence on Cooling Time.......................................65
Figure 46:
Dimensions after Annealing at Different Cooling Times................. 66
Figure 47:
Influence o f Processing Conditions.on Housing Quality.................. 68
Figure 48:
Processing and Shrinkage............................................................... 69
Figure 49
Length vs. Processing Conditions..................................................... 69
Figure 50:
Quantitative Survey about Processing and Product....................... 70
Figure 51:
Length and Width of the Housing as Function o f Annealing............73
Figure 52:
Processing "60" vs. Dimensions A, B and C.....................................77
Figure 53:
Production Run with of Production with.Profile"60".......................77
Figure 54:
Length as Function of Processing with Holding Pressure
Profile“60”
78
Figure 55:
Processing "70" vs. Dimensions A, B und C.................................... 78
Figure 56:
Production Quality with Processing Profile "70"............................ 79
Figure 57:
Length as Function of Processing Profile “70” ................................79
Figure 58:
Processing "75" vs. Dimensions A, B and C.................................... 80
Figure 59:
Production Quality with Processing Profile "75"............................. 80
Figure 60:
Length as Function of Processing Profile “75”.................................81
Figure 61:
3-Dimensional Picture of the Dimensions A,B and C
vs.Processing.................................................................................. 82
Figure 62:
Line Diagram of all Dimension A, B and C......................................82
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 63:
Length of the Housing vs. Processing................................................83
Figure 64:
Part Weight as Function of Molding Conditions............................... 84
Figure 65:
Servo Control....................................................................................88
X
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
¥. INTRODUCTION
This research was done in the Bosch-Siemens Plastics Department, near Munich. In
this department they have about 20 injection molding machines with a clamping force
range between 150kN and 9000kN. Most of these machines were ejected manually due to
the small quantity production runs. However, a certain water heater was injected and the
halves were removed automatically and placed in a chamber where PUR foam was
incorporated and cured. After this process the two halves were welded together and the
final assembly was executed automatically. Production difficulties with a microwave
stove switch housing (Figure 1) were the reasons for this research.
10 cm
Figure 1: Switch housing of a Microwave Oven
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Since this housing was already an existing part of an oven system, the possibility
for major changes did not exist especially not for the geometry. The entire oven concept
was already approved by management and the appliance was built accordingly. This
meant electrical devices had to be placed onto the plastic part and the adjustment of all
system devices required for an easily produceable switch housing were ruled out. The
housing was injection molded in a sprue gated single cavity mold. After a certain
production run the housings were placed manually in an annealing oven and heated. The
annealing process for 20 housings took two hours using a temperature of about 120°C.
After this process the parts were loaded into a clamping frame for approximately 30
minutes to avoid excessive warpage due to the high postmold shrinkage of PBT (up to
2%). For this step one additional operator was required. To find an economical and
technically feasible solution for the overall manufacturing process the current study was
undertaken as a research project which resulted in this thesis.
Quality assurance is becoming ever more important especially, if one considers
product liability and the competition of the market place[l]. In addition to the above
mentioned factors the producer also has the responsibility to provide complete
documentation for the manufactured goods according to good manufacturing practices.
Statistical methods show great promise to improve the manufacturing process. However,
they are only useful when a stable process has been established.
Figure 2 shows the theoretical temporal course of an optimization using distribution
curves which show the normal distribution of a property for a certain product. First, the
curves are broad (still systematic influences), then these influences are eliminated but, we
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
still get a broad distribution. With increased running time the optimization curves
get narrower and narrower and finally meet the target value.
minimized
•-•"cer Control.
c o l sec ;y r.s ie -a :;,
mil m'cs
:cc - 3 -
: I.tr’Linces
Figure 2: Using SPC for optimization [2]
1.1 Objectives
The aim of this research was to use a data acquisition system for the processing
variables during injection molding of a switch housing of a microwave oven. The upper
and lower tolerance limits of the most significant process variables were determined.
Knowing these limits, the optimum processing point can be found to make the injection
process as stable as possible.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Once the optimum conditions have been found it is desirable to operate the molding
process in a closed loop mode to keep process fluctuations to a minimum. The process
measurements are used to carry out SPC (statistical process control) to assure quality and
document the manufacturing conditions.
n . QUALITY
This expression, according to DIN / ISO 9004 is based on whether the properties of
a product meet the requirements of the customer or not [3]. This has to be checked
continuously.
In practice we distinguish two kinds of Quality: "0-hour Quality", the quality which
is noticeable immediately after manufacturing and the "Quality for time", what we call
reliability of the part or service life [4].
In addition measuring product variation, quality assurance also requires that money
be spent to measure and maintain quality. Quality costs are costs associated with
eliminating defects (rejects, refinishing,etc.), finding defects (test.- and evaluation costs)
and costs for avoiding defects (quality management) [5].
Nevertheless it should not be forgotten that quality has to be produced, it is not
possible to "retest” it.
m . FACTORS INFLUENCING QUALITY
Before talking about the quality of a certain part it is necessary to define the
parameters which influence its properties. In this case we are talking about the quality of
a part produced by the injection molding process. This means a lot of parameters are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
influencing the housing. Taking this into account, a lot of different defects can be avoided
by knowing at least the basics of the process.
The injection molding process
is subject to a lot of influences. Since these
influences are difficult to quantify, a
lot of effort has to be spent to
find themost
important parameters, i.e. if injection speed is increased there is not only a reduction in
injection time but also a change in viscosity, melt temperature, and hence, the surface
quality of the part.
To summarize this we can change one parameter after the other to see the exact
influence of this change by having approximately steady state conditions. This can be
achieved by waiting a couple of minutes after the adjustment until the steady state of the
entire system, mold and machine, is reached again.
A good overview of the above
mentioned situation is shown inthe "Ishikawa
Diagram" also known as "6M"-analysis [see Figure3].
r
Gate
Cooling
Stiffness
I Mold
^
^
^
Machine ^
r Milieu
r Method
Process
Automation
Qual.Assur.
v.
Reprocessability
Forgetfulness
Settings
Power
Ejection
>
^
r Material
J
^
Prooerues
Location
-J
Man
Environment
J
Color
^Moisture content ^
Figure: 3: Influence of "6M" on Injection Molding (Ishikawa Diagram) [6]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
3.1. Mold
3.1.1 Gating
Gating of a part determines a lot of its properties. The gate size and the gate
geometry are responsible for internal stresses and the orientation of the molded part. One
also has to make a decision about using a cold or a hot runner system. The corresponding
adjustments to the mold such as dimensioning of the cold runner layout or using a certain
hot runner nozzle will determine if one gets optimum parts.
3.1.2 Cooling
Similar in importance, the cooling system has a big influence on the part quality.
The hotter the mold is, the longer the 2nd stage pressure can work and sink marks can be
eliminated. Most important is the mold temperature for semi-crystalline materials where
the morphology depends very much on the temperature and hence, it influences the
dimensions of a part. When monitoring the mold temperature it is important to see the
real temperature at the mold surface and not only the average temperature of the cooling
liquid. The temperature deviation should be as small as possible to achieve the most
uniform cooling and therefore low stressed parts. To determine the optimum temperature,
thermal equilibrium o f mold, environment and cooling liquid has to be reached. Then it is
possible to see the efficiency of the cooling system which has to be placed between
ejector pins, bolts and the part geometry. In most cases it is necessary to apply different
cooling temperatures to overcome certain effects, such as high molded in stresses at
internal comers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.1.3 Stiffness
The deformation of the mold during the injection of the melt at pressures up to
1500 bar also has an influence the properties of the part. Flashing may occur if the
deflection is too great or if the clamping force is inadequate. Also, the part dimensions
and tolerances increase.
3.2 Machine
3.2.1 Settings
The most important factor concerning the part quality is the machine-settings.
Figure 4 shows the basic influence of the molding parameters on properties of the
injected part.
Holding Pressure
— injection-
p— C o m p re s sio n
: ;i O rie n ta tio n
j
V
W eig h :. D im ensions. S inks
V o ic s, O rientation a: the C e n te r I
Crystallinity
>«
>
r,
.S c a r: o f In jectio n
Figure 4: Classification of processing and the corresponding product quality [7]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
By using a pressure gage inside the mold cavity three different phases can be seen:
1) Filling of the cavity
2) Compression of the melt in the mold
3) Solidification of the molded part
The injection process can be described as follows:
- Cavity pressure
- Melt temperature
- Mold temperature
- Melt velocity
- Thickness of the solidified layer
To see the process under real time conditions, the above mentioned quantities are
different at every point in the mold. However, they are assumed to be constant across the
surface of the part or along the direction of flow.
The cavity pressure can only be determined at a few points in the mold. Melt
temperature can only be measured by injecting the melt into a chamber and using a
thermocouple. Useful results are also given by the Moldflow analysis, which shows the
theoretical filling of a certain part by assuming almost real conditions. The uniformity of
the filling which should be as good as possible can be seen in the first step of a fill
analysis.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
3.2.2 Power
Uniform part quality is not possible if the capacity of a molding machine is reached.
A certain reserve should not be touched. A rule of thumb says that at most only 80% of
the maximum capacity of a machine should be used.
3.3 Operator
3.3.1 Repeatability
Nowadays the operator is not a factor in repeatability. He became a user of very
accurate working machines. Most of the injection molding companies possess machines
which allow one to run automatic cycles. Hence, the human factor in terms of quality can
be eliminated almost completely.
Only in very few exceptions it is necessary to rely on the repeatability of an
operator. Mostly these work steps are very easy and are not worthwhile to be automated.
Basically we can say that the repeatability of an operator is poor because he is
influenced by a lot of random quantities.
3.3.2 Forgetfulness of Operator
In contrast to repeatability, which is mostly based on "physical work", we have to
consider the forgetfulness of the operator. Admittedly there are a lot of technical
auxiliaries to support the operators work, but those are not smarter than the operator who
is using them. To counteract this a lot of strategies have been devised (see also chapter 4).
Two things are endless:
the universe and human
stupidity
but about the universe
la m not so sure.
Albert Einstein
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
3.4 Method
3.4 1 Degree of Automation
This point should be considered in detail before starting work. The possible
solutions of a problem should be compared with the technical equipment available to find
the most economical way to get the work done. Starting with the number of fully
automatic machines which are able to print processing variables and figures showing
ideal and real cavity pressure one can progress to a closed loop controlled machine which
helps to save man hours.
3.4.2 Quality Assurance
The structure of quality assurance is based on the specific demand made upon a
product. Here the most important quality criteria have to be denned exactly. Ideally, those
criteria are the ones which can be measured easily. The dimensions and the corresponding
tolerances have to be determined in such a way that the quality goal is absolutely clear
and attainable.
3.5.Environment
3.5.1 Machine Placement and Temperature
The location of an injection molding machine could influence the product quality
significantly if the environmental conditions are not kept uniform. This means that the
humidity as well as the temperature and thermal equilibrium of the mold and the machine
cannot be guaranteed. In this case one can distinguish between "Summer" and "Winter
parts", especially concerning the cooling unless the system is based on a closed loop.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
3.6 Material
3.6.1 Material Properties
A key requirement for reprocessability is a constant raw material which has to be
ensured by the supplier. These properties include: viscosity, mechanical properties,
constant color, MWD and moisture content. Moisture content is very important for
moisture sensitive materials such as polycondensate polymers. Besides the reduction of
the molecular weight and the corresponding decrease in mechanical properties, a poor
surface quality would also result.
IV. METHODS OF QUALITY ASSURANCE
Figure 5 shows the correlation of costs of producing, discovering and fixing a
failure. It is obvious that 75% of all defects arise in the development state of a new part.
In contrast to this the remediation curve shows that 80% o f all defects get discovered at
the very end of the part history when it is almost ready for market or in the worst case
already in it.
The reason for the scenario shown in the following figure is most often determined
by how quality is controlled in the companies. This is based on a strict quality philosophy
which only looks at the finished part. Only at this moment are you able to approve or
reject it. This means that a large number of bad parts may already have been produced
before the error is discovered. This strategy is old-fashioned and has no use at all in
modem manufacturing requirements.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Repair o f
80% of all
Errors
Cost of Errors
Frequency of errors
Cause of errors
Source of
7 5 % o f all
Errors
c
o
O
0.
Part History
□ C a u se o f errors
■ Frequency of errors
Figure 5: Source of Errors and their Remediation [8]
Modem quality concepts are designed in such a way that failures or possible failure
sources can be eliminated during the development phase before steel is machined.
According to this precept Japanese quality assurance was built in the last four decades
and for the last ten years the European market has been following the same strategy.
The consequences of this history are shown in Figure 6. First the quality standard of
the eastern products was very low, but it improved steadily and very quickly. The
improvement in the European quality standard grew quite slowly. By the end of the
seventies process control also known as STATISTICAL PROCESS CONTROL (SPC)
gained more and more in importance.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Developm ent of Quality Standard
x Europe
o—Japan
1950
1960
1970
Year of Production
1980
1990
Figure 6 : Japanese and Western Products from 1950 - 1982
according to Haward Study (values from 1982 on, estimated)[9]
Since that time the long range planning of new projects, a part of the OFF LINE
QUALITY-ASSURANCE has gained more and more in significance. This planning is a
tool to analyze a new process carefully, to find the key points of a new project, and to pay
attention to them. One is able to evaluate certain points already in the pre-development
state and to stop the development if it is running in a wrong or unstable direction. Thus, it
is obvious to shift the quality assurance steps forward to the development state as far as
possible [10].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1 TQM (Total Quality Management)
The entire effort of effective quality assurance in a company, in all divisions, can be
summarized by the generic term "Total Quality Management" (TQM).
Here we find:
OFF-Line procedures:
QFD; FMEA; STEM
and also:
ON-Line procedures:
SPC and CPC
4.1.1. Off-Line Procedures
4.1.1.1 QFD (Quality Function Deployment),
QFD is responsible for transformation of customer demands into technically
feasible and practical solutions in terms of product and process. This has to be achieved
by having an interactive working team, starting from marketing down to the
specification for an industrial production where the customer demands should have
priority.
4.1.1.2 FMEA (Failure Mode and Effects Analysis),
If a new project has been placed, the FMEA should be executed for all
development steps before mass production is started. The endeavor here is to replace the
detection of failures by avoiding them, as shown in Figure 7.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure.?.: Do it right the first time!
The methods of modem quality assurance must allow us to have a far sighted and
systematic analysis. This requirement is satisfied by the FMEA analysis, because of its
anticipatory nature to find potential defects as well as their causes and their consequences.
Having a rough evaluation, we can compare the RISK NUMBRS (RN) [12], to show the
weak points in a project.
RN=E*A*D
Effect of the failure
E
Probability of Appearance-
A
Probability of Discovery
D
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Only if we carry out this analysis completely and thoroughly by having a critical
look at it will we be able to get reliable results. We distinguish between Design and
Process FMEA.
During the Design FMEA all concepts in the development state of a product are
documented. First the analysis starts with theoretical knowledge and later it will be
continued and completed with results from development.
The Process FMEA is based on the Design FMEA and investigates design features
of a part concerning production, assembling and testing to make sure the final product
still meets the requirements [13].
4.1.1.3 STEM (Statistical Experimental Method)
STEM builds the qualitative and quantitative connection between the product
features and the process parameters. The STEM is a very important element of the
process analysis, since the process can be optimized for a stable processing point and one
can determine which parameters have to be observed to guarantee good results. Before
mass production is started the fluctuations of the part quality features (e.g. dimensions)
have to be minimized for a wide stable processing window.
During mass production important part features can be monitored continuously by
using Statistical Process Control (SPC). This method can be used if a stable and
repeatable process is achieved and the probability is very high to form conclusions from
randomly taken samples that their measured properties are representative of the entire
production. The “high probability" will be explained in more detail in a later section.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.1.4 Capability Determination
Machine.- and Process capability
In spite of the many factors, systematic and random, the process has to fulfill
minimum requirements. This means the fluctuation of the process must be within a
specified range.
The Capability Index gives an idea how good the given process is working in
relation to the given tolerance-range.
We distinguish between the machine and the process itself:
MACHINE:
Single piece of an entire Process or Production
( only few outer influences)
PROCESS:
The entire Process is based on the interaction of
"six sigma"
(see section 3.1 Figure 3)
Machine Capability:
It shows how the quality of a machine or a system is determined by a large spot
check (>50 piece parts.)[14] within a short period of time.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
IS
Minimum Requirement:
The deviation of X +4c is within the tolerance range where X js the
average of all numbers within the spot check. It means that 99.994% of all
produced parts are good ones (see Figure 8 Percentage of the StandardDistribution).
cor
• 9 S .9 S -%
C03% •
•9 9 .;:% . 35.44% -
Turning Point
' - JS
Figure 8: Percentage of Standard Distribution [15]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In addition to this requirement one also calculates the Machine Capability Index, cm.
for a standard distributed feature using the following relationship:
Cm= USL - LSL
6s
The Machine Capability Index cmcompares the Upper .-and the Lower Specification
Limits (USL, LSL) with the 6s distribution o f the machine (s=Standard Deviation).
The position of the deviation to the Specification Limits cannot be seen from the
Machine Capability Index cm. For this reason the Machine Index cmk has to be
determined.
USL -x
Cmk=______________ >1.33
3s
x -LSL
or
Cmk=______________ >1.33
3s
The smaller of the two numbers has to be used for the evaluation.
Process Capability:
If the Machine Capability Investigation leads to a good result, the Process
Capability Investigations will be carried out [16]. The result of this investigation should
clarify if a given process is able to give repeatable results for long periods of time. Thus,
a couple of smaller spot checks have to be taken over a longer period of time. In this
investigation also long-term influences, such as a new lot of material, are taken into
account.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Minimum Requirement: The deviation ^ ±3<T is within the tolerance range,
means 99.73% of all produced parts are good ones (see also Figure 8).
Whereas X represents the grand average o f all spot checks.
The Process Capability Index, cp, considers only the distribution:
cP= USL - LSL
6s
The Process Capability Index, cpk. considers not only the distribution of the process
but also the position of it.
USL -X
cpk=______________> 1.0
X _LSL
or
3s
cpk=______________ > 1.0
3s
The smaller value shows the location of the investigated Process[14].
4.1.2 On Line Procedures
4.1.2.1 SPC (Statistical Process Control)
Via the classic SPC, from a spot check we can make conclusions about the entire
production process. This method has been used successfully during the last years in our
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
production plants. For this relatively "low-scale" test- procedure the standard of the
quality can be guaranteed at very high level by not spending too much effort on it.
To be absolutely sure, a 100% control has to be used, which may be very expensive
and lavish. With this test method one also can see long term trends by comparing the
results of the spot checks. Corrections can then be made very quickly.
4.1.2.1.1 QCC (Quality Control Chart)
Using the SPC method, spot checks will be taken following a predetermined
frequency which has to be entered on the QCC (see Figure 9)
Lapu
Oidltu H o ra n Qsct
partrunr
EK5167
htaclltn radra
USL
Target
LSL
me *
S'
acting Intna
S / 1har
apacnlmLkrt hrilt
USL tE .fl / LSj m
02
Figure 9: Quality Control Chart (QCC) [17]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Where Xi to X5 represent a certain feature of a product. These values are used to
calculate the average x . This number will be indicated on the QCC. If we connect the
points we obtain the "Quality Curve" of the process. The course of this curve shows the
status of the current quality and, moreover, it allows a forecast of the fixture process.
If the fluctuation of the values is very small then the points are very close to the
target, we can say that the investigated process is in control. The predicted quality from
the SPC test method guarantees a good quality standard.
Besides this knowledge o f a controlled and stable process, we can fiirther see the
"Trend" of the process as mentioned above, which might be caused by material changes,
wear of valves etc. [see Figure 10].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
□GQ-QuclitdUragallarta
; d e s rio ilo n
h o u sin g
BHW
j
1
I
— --------------- 1
1 ! ' now
i
1
i
I
“
i
i !
i
r
i
“
O
1
; d ro u in g r u a b e r •
G 41V D 4Q 58
!
i aaplire inUrvdl !
| n - 5 /S ^ n
I
i
i
te s t
tid th
i I
i 'i
1
l
i
i
!
! I
i-1
1
C. J
y
L
r n 'H
fi
i
!
m a c h in e
E S m /2 0 0
re m a rk s
_______ _
1
1
1
J h f
I X
i
l
i
1
1
m a te ria l
FBT
i i
I !
-T
i
1
i
i
!
|
i
X
!
1
1
i
1
|
!
.
!
;
1
'
1
;
i
!
{
i
t
T a rg e t
1
i
t
2 —!—
t
1
!
i
! 1 i k n ii a
1I
u
1
{
!
c
i
i
! ;f
i R
f
M *
1
, /]\
^
;
i
I I
n ■ • i___1
1
i
i Line:
:a i
an
1a^e/aouran
t
Ifora
1
V
:
'
i t
•r . i
K, [ 1 i
* t : i
. '
- 4
|
|
!
’a )
cso
4,
|
|
-
«* Sj
T '
1
i. .
1 1
no
i
!
I
»
;
1
i
!
I
i
*
!
I
I
t
cm
odd
wm
I
i
i
i
i
i
i
1
1
1
1
!
i
;
i
>
1
ii
;'
Figure 10: Example of QCC with "Trend developmental 5]
For this reason Upper and Lower Intervention Limits (UIL, LIL) are determined.
They will show when the process starts to go out of control while taking the reaction time
of the process into account.
For the example of Figure 9, the mentioned values are counted according to the
DGQ Booklet [15].
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
To determine the Upper and Lower Intervention Limits (UIL, LIL) the following
values have to be determined first:
current portion o f bad parts:
Interaction Probability:
Spot Check Range
p [%]
(1'P a) t%]
n
Out of these quantities the Mark-OfF-Factor kA can be defined by using the
"Wilrich-Nomogram".[ 15]
Upper Intervention Limit:
UIL = USL - kA * (s / a j
Lower Intervention Limit:
LIL = LSL - kA * (s / a j
USL:
Upper Spec. Limit
LSL:
Lower Spec. Limit
kA:
Mark-Off Factor n=5, (l-pA)=99.73% and
p=0.27% from the "Wilrich-Nomogram"
s:
Median Number of the Standard Deviation of the
Spot Check Results
an:
Factor for estimating the Standard Deviation
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Target Dimensions for Housing:
Width:
106.10 mm
Tolerance: ± 0.15 mm
USL: 106.25 mm
LSL: 105.95 mm
kA: 4.1
s : 0.016
a„: 0.940
UIL= 106.25 -(4.1*0.016/0.940)
UIL= 106.18 mm
LEL= 105.95 +(4.1*0.016/0.940)
LIL=106.02 mm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.2.2 CPC (Continuous Process Control)
To maintain the quality during production, it ideally has to be ascertained either by
a continuous or a statistical test method (CPC, SPC). SPC is commonly used for the
product quality and CPC is preferably for controlling the process parameters.
This becomes more obvious by knowing the fact that many of the process
parameters can be monitored. Thus, it should not be a big issue to use them in adjustment
calculations or to print them out on a Process Control Chart (PCC), as shown in Figure 11.
In this case we really talk about 100% control of the process parameters.
QtJA'l^T5DATENSTaTISTlK:^ ^ P i i ^ M S S ^ . G ^ A . 0 8 i 2 r ::: j
: rtv c ra u iIK c n u rrts c n a rw iF F i'e s s z a ril:
P o c s ts r ! a n g e :
n : 5 m : IQ A b s ta n c : 100 SZx: 200
SW: . 5 4 .0 O SG : . 70 .0 U SG : .53.0 IVY: 51 1 bar
.305
. .7 2 5
. 9 : 10
1:14
S Z x:
Z eit:
70ZJ;
64 . 0 :
53.0
L
2.00
i-
1.00
M e s s u n g e n : 25
u n te r s p e z : 0
0.0 %
G b ersp ez: 0 0.0 % j
UEGX: 51.0 OEGX: 5 7 .0
UEGS: 0 .0 0 G E G S : 1.50
i_
x : 5 4 .0 0
£ : 1.51
s : 1.40
Cp : 1.23 C pk: 1 .2 2 !
II
j p E c e jC k a ra
S -p e z if.
I ;c *
eh
;
em
} f E j N G a j P F s - 1 |'i c o * -
| Sc s e x z e k
] ( r e c E l..;. 1
flCUPATlONl IKARTER1R?
RAUFTUFMji
;
JC
/ 'c s s - f e x S i C
LSSO «EX
,
• 1C
j I
1!
Figure 11: Process Control Chart (PCC) [18]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.2 Summary: Quality-Assurance
In Figure 12 an overview of the explained scenario is shown. It gives an idea of
how early the quality assurance starts to work. To solve a problem at the very last point
when dimensions are checked is ridiculous.
Industrial Production
J.
J
before
QFD
FMEfl
STEM
FVoc888 Capability
Quality influendng Parcweters
Optima aocbine
settinos
n
r-1
l
dtrino
SPC
.1. '
after
SPC
CPC
Control of pro­
cess paroaeters
weighted response
Control of Pro­
duct quality
(geoeetry, | surface etc.)
J
Figure 12: Quality Assurance [19]
The strategy in injection molding is to replace the quality control on the final
product by using on-line process control or at least to combine both methods to gain the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
best compromise between the quality standards and economics. To get a reliable
relationship between process control and the final product a statistically controlled
process has to be reached first.
V. PRIME OBJECTIVE
The basic idea of this research was to generate data o f an injection-molded part.
Referring to these data, an upper and a lower tolerance limit of the most important
dimensions was to be determined. Based on the outcome of the first investigations an
optimum processing point was to be found which guaranteed a wide and stable processing
window via optimization.
This means that the entire study is an element of quality control, as explained in the
previous chapters. Basically the work alternates between the components of the "Off' and
the "On-Line" procedures. First, work is done in the area o f the STEM (Statistical
Experimental Methods) as representative of the Off-line approach and second, work in
the area of SPC (Statistical Process Control) is representative of the On-line procedure.
Given are:
- Part
- Material
- Injection Molding Machine
-M old
- Method
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
5.1 Part
The part under consideration is a "Switch Housing" for a Microwave Oven, as
mentioned at the beginning (see Figure 1). The part was already improved by changing
the geometry as far as was possible. However, there were still more iterative steps
required to get a good part. In addition the current process for making of the part was to
be investigated for statistical stability. The part is gated slightly off center with a modified
sprue gate. The housing has a length of about 360mm and a width of 138mm. The
dimensions are given in Figure 13 which shows the exact measuring points for this work.
The dimensions are shown in diagrams which follow.
Measuring Points of the Housing:
1
_ .
Bate
10
mm
' -
t
ac
O
CD
18 5
350
mm
mm
r
1/1' = 360 + /- 0.3 mm
fVB/C = 106.1 + /- 0.15 mm
Figure 13: Specified Dimensions of the Part
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
5.2 Material
Based on the high temperature to which the housing is exposed (above 150°C short
time) in combination with the given Injection Molding Machine and the cost of a variety
of polymers, the material PBT (Poly Butylene Terephthalate), Tradename: „Ultradur® B
4520“, was chosen. In addition to this material other ones were also tested, but either they
failed the heat stability test or they could not be processed on the given machine. The
housing was injected in different colors, and the white colored parts showed the smallest
shrinkage. Since the final parts were too large the white color was chosen to investigate
the worst case.
5.2.1 Production of the Material
Ultradur® B 4520 is generated by condensation of Terephtalic acid or Dimethyl
Terephthalate (DMT) with Butandiol-1,4 by using Antimony and Germanium Catalysts
(see Figure 14).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Production of
Poly-Butylene-Terephthalate
(PBT)
HOOC
-C
OOH + H O 4 -CH 2 -C H 2
c —O - f CH 2 -C H 2- h O -
Figure 14: Production of PBT
5.2.2 Material Properties
The amorphous phase softens at about +60°C whereas the semi crystalline phase
starts to melt at a temperature of 210°C. This results in good temperature stability and
toughness within a wide temperature range[24]. The application temperature starts at 40°C and goes up to +100°C for short time exposure it even goes up to 165°C. The
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
reason for this lies the semicrystalline structure of the polymer. A shrinkage of up to 2%
caused by crystallization can occur which might cause problems in terms of warpage.
The appearance of the material without any additives is a white opaque. As with all
polymers PBT also ages (oxidizes). This is accelerated with increasing temperatures.
Polyesters are very sensitive to moisture during processing. A hydrolytic degradation will
take place and the material properties will decrease drastically. Therefore most of the
suppliers deliver this material with a very low moisture content. This allows one to
process material from a sealed material bag without pre drying the resin if the bag arrived
in good shape and was not opened too long. The excellent dielectric and insulating
properties of PBT are almost uninfluenced by water absorption, which is low anyhow.
Equilibrium in moisture content is reached under ambient conditions at about 0.25% to
0.4% and in cold water at about 0.5% to 0.7%. The higher values are valid for PET (Poly
Ethylene Terephthalate) and the smaller ones for PBT (see Figurel5). The maximum
water absorption depends on the level of crystallinity.
JJ
23
13
50
SO
ISO
120
W3
ISO
E x p o su re T im e (days)
Figure 15: Water absorption of PET as a function of time [20]
with permission of the copyright owner. Further reproduction prohibited without permission.
5.3 Injection Molding Machine
The Machine is equipped with a modem CC90 Control unit which makes it
possible to monitor and print out up to 11 Process variables after every cycle. The process
data screen shown in Figure 17, allows to monitor the process for a long period of time for
three variables. Immediately one can see if a variable has a high fluctuation and hence the
process quality will most likely fluctuate too.
i
3265
Figure 16: Injection Molding Machine Engel ES 700/200 (Dimensions in mm)
Figure 17: Process Data Screen CC90 [21]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Technical Data about the Injection Molding Machine Engei ES 700/200
built in 1989
Injection Unit:
Screw Diameter:
Screw Speed:
50 mm
5-240 m hr1
Shot Size:
200 mm
Plasticating Rate
43 g/sec
Injection Rate:
220 g/sec
max. Volume:
392 cm3
max. Shot Size (GPPS):
353 g
Injection Pressure:
1350 bar
Nozzle Length:
300 mm
Nozzle Contact Pressure:
58 kN
Heating Power:
11.3 kW
Number of Heating Zones:
4 + Nozzle
Clamping Unit (Toggle):
Clamping Force:
2000 kN
max. Clamping:
2200 kN
Daylight:
510 mm
Ejection Force:
61 kN
Ejection Stroke:
150 mm
max. Mold Height:
710 mm
min. Mold Height:
400 mm
Tie Rod Diameter:
90 mm
Clamping Platen Size:
710 mm x 710 mm
Clearance between Tie Bars (horiz.)
510 mm
Clearance between Tie Bars (vert.)
510 mm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
5.4 Mold
The mold is a two plate single cavity mold with a hot runner and a conventional
cooling system. Figure 18 and 19 show the cooling channel layout of the cavity, and on
the core side of the mold.
Cooling Channel-Lavout
Cavity Side
E l/A l
E2
E3
A3 E4
Gate
A2
A4
El •
A1 •
E4/A4
E2/A2
E3
A3
Section: Z-Z
Figure 18: Cooling Channel Layout for the Cavity Side
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cooling Channel-Lavout
Core Side
E7
E6
E5 A5
A6 A7
•
i
E8/A8
•
•
E5_____E6
•
•
A6 A5
Section: X-X
Figure 19: Cooling Channel Layout for the Core Side
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VL REQUIREMENTS FOR THE HOUSING
The requirements for the housing are basically very simple because it is not exposed
to high stress and no special considerations for the mechanical properties of the part are
called for. Since this product is used in a kitchen it has to meet some requirements. This
consideration is a part of the "Off line” procedure starting with the material selection and
going to design aspects, both the optical and necessary ones, such as a highly polished
surface which is easier to clean than a textured one. Since the housing is a part of an oven
it must withstand elevated temperatures without showing any changes.
The most important requirements are:
1) Heat stability up to 150°C (if the heated oven is open)
2) Accurate dimensions
3) High gloss surface
4) No visible warpage
For these demands the PBT, Ultradur® B 4520, was the best solution for the time
being. Besides the advantages of this material one also has to accept the bad properties.
Due to the semicrystalline structure, the shrinkage, especially the nonuniform one, will
cause problems. The more complicated a part becomes, the more the above mentioned
tendency becomes true and excessive warpage may cause failures in addition to an
unsightly appearance. Unfortunately, the shrinkage values determined with standard parts
are only partly valid with complex parts and the final dimensions for a difficult part are
difficult to predict.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Moreover, the hydrolytic character of PBT has to be taken into account. It must be
given a close scrutiny. The raw material must be dried or molecular weight reduction
during processing may occur and bad parts could result.
Vn. INVESTIGATIONS AND TEST PROCEDURES
7.1 Equipment
Moisture Content Determination:
Balance:
Sartorius "research"
Dryer:
Hereaus, model: FVT 420
Temperature:
Mold- and Melt temperature: Thermo Element "ebro CTX 1200"
Annealing Temperature:
Thermoelement "Siemens
ThermizetB 4001
Annealing Oven: Memmeret Model: UL 60u
7.2 Moisture Content
Since there is always uncertainty about how much permissible moisture a raw
material contains an investigation at Siemens Laboratories was arranged. The result was
that as received raw material with a humidity content of up to 0.14%!! would show no
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
effect on material properties. This was really amazing because the resin supplier is
recommending a content <0.05%.
To check the current delivery status of the predried material, samples were taken
out of just opened bags and out of bags opened for longer time periods. The weight was
documented and than the samples were dried. The difference in weight represents the
moisture content of the material in the bag.
The granules were dried for three hours at a temperature of 100°C under a vacuum
of 500mbar. The results can be seen in the Table 1, below:
Results: Moisture Content
Table. 1: Water Absorption as a Function of Time
Storage:
Water-content [Weight%J:
new opened bag
0.04
2 hour opened bag production hall
0.06
72 hour opened bag production hall
0.12
These results represent the median number out of three measurements of different
charges. The deviation is about ±0.01 [Weight %]. Based on this investigation the
delivered material can be used without additional drying. It is not necessary to have a
drying oven for this material. For the required high gloss surface the moisture content of
0.12% should be considered carefully, in spite of the Siemens research, where 0.14% was
allowed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
7.3 Original Process
Up to now the housings were injected, stored, manually packed in the annealing
oven for two hours and afterwards clamped with a special fixture to avoid excessive
warpage as illustrated in Figure 20.
Figure 20: Clamping Fixture for the Housing
7.3.1 Initial Process Settings
The settings given below are target values and succumb natural fluctuations.
However, the changes are very small and can be neglected as long as the process remains
stable.
Temperature:
Nozzle
250°C
Heating Zone 1
250°C
Heating Zone 2
250°C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Heating Zone 3
245°C
Heating Zone 4
240°C
Hopper Zone
60°C
Oil
40°C
Hot Runner
250°C
9
40°C
10
60°C
11
40°C
12
45°C
Cooling Device:
20°C
Mold temperature:
Circuit number:
2nd Stage Pressure:
80/70/60/55/45
(Profile)
40/30/40/30/20
bar
Time of 2nd Stage:
5
sec
BackPressure:
10
bar
Injection Speed:
25/18/25/23/30
(Profile)
35/30/30/55/50
mxn/s
Plastificating screw
50
rpm
125
mm
speed:
Shot Size:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
7.3.2 Mold Surface Temperature
The mold surface temperature was checked to see how the mold influences the
shrinkage as shown in Figure 21. It shows that the temperature profile is quite uniform.
Only the comers of the core side should be cooled better to avoid high molded in stresses.
Additionally the long sidewalls of the cavity side should be cooled better to minimize
warpage of the long sidewalls.
Tem perature Distribution in the Mold
during original Processing
Cavity Side:
Core Side:
42
43
43
Sidewall Temperature = 56°C
rnnliint Tem peratures: 40 °OB0°CV400Cy4B0C!/200C
Figure 21: Mold temperature for the Original Process
These temperatures were measured directly after running fully automatic when
equilibrium was reached. Thus, the measured temperatures are very close to the real ones.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
7.3.3 Temperature Distribution for Annealing
The temperature distribution in the annealing oven could promote warpage. As
shown in Figure 22 there is no reason which supports the hypothesis mentioned above.
The temperature profile is uniform.
Temperature Distribution in the Annealing Oven
Temperature setpoint:
Thermocouple:
100°C
see Figure
Oven
_.
104-0
104-0
106*0
m °e
104°C
106-0
105*0
104°C
105*0
104*0
106*0
106*0
107*0
107*0
Housing
_ 107*
Figure 22: Temperature Distribution in Oven during Annealing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
7.3.4 Machine Capability
After reviewing and documenting the original process in detail the attained quality
level of the process was determined. Verifying the attained quality level allows one to
give a statement about the process and whether it is in statistical control and only then an
optimization scheme can be initialized. Therefore, out of the fully automatic production
50 pieces were taken, measured and the results were statistically evaluated. Referring to
this test the calculated Machine Capability Index, cm, should be higher than 1.33 in order
to have a stable process. As shown in a table in the Appendix this value is definitely
higher than 1.33 but the Index, cmk, is negative. The last condition cmk>1.0, illustrated in
Figure 23 which shows that the process is capable but not centered. The reason is, the
injected part does not have the final dimensions at all, it is still too large after the
injection process, but the annealing step is missing (approx. 1% shrinkage). To use this
method new Specification Limits have to be determined, which than lead to cmk>1.0 for
the injection process. Continuing with the original process-settings it is necessary to
predict the quality of the finished parts after annealing.
IL'SL -L S I
Machine Capability Values
cmk =
cmk = -
X - LSL
— 2 1.33
-4 s-3 s-2 s-ls
-Is -2 s-3 s-4 s
cm = 0.6“
cmk= 0.33
-4s
-3s
-2s
-I s
-Is
-2s
-3s
-4s
Figure 23: Position of Distribution and Machine Capability Factor [21]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7.3.4.1 Error Analysis
a) For length 1 and 1'
The length was determined with a caliper having a scale o f 5/100mm. Whereby
5/100mm is the smallest error, which can occur, measuring the length of 360.5mm. The
percentage error related to the total length is negligible, taking also the fact into account
that measuring such a length with a caliper has to be considered carefully.
However, if the difference in length up to (5/10mm) due to changed machine
settings is taken into account, the error cannot be neglected anymore! The difference in
length is related to manually changed machine parameters, e.g. varying the 2nd stage
pressure from 20 to 60 bar. This results in a possible error fp:
fpl= -y- * 100%,
whereAI is the smallest possible error
1is thedifference in length due to settings
fni= —
P
100
*— *
5
100%
= 10% !!
---------
b) Width Measurements A,B,C with Digital Calipers:
AA
fpl=
*100%,
v
A
1
mm
AAis--100
A is Difference due to settings
fp2= — *— *100% = 20%!!
F 100
5
-------Thus, it becomes obvious that an error in measuring the width of the part is larger
than measuring the length of it. This also means that fluctuations within one constant
setting are more related to errors in measuring than to process fluctuations.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
7.4 Process Analysis
The machine capability index shows that a statistically controlled process is
achieved. Due to the annealing, however, the index cmk cannot be used at this stage. As
mentioned before, an adjusted index cmk must be generated to be able to recognize errors
already after the injection molding process and not after annealing where many manhours would have been wasted.
For this reason it should be possible to anticipate the quality before the final
product is produced. A good starting point would be the finished injection molding step.
The next step is based on this consideration. The tolerance range for the injection molded
unannealed parts has to be found which creates good parts (see Figure 24).
Doing so, parts have to be molded, measured, annealed and checked again. The
generated data allow one to predict the quality of the final product if we measure the part
already after injection. In the end this should be possible just by controlling process
parameters which lead to good parts. If deviations are encountered the process can be
adjusted before the final product is ready.
7.4.1 Quality determining Parameters
The parameters are devided into two groups as shown in Table 2. The first group
consists of a maximum five probably most relevant parameters concerning the quality
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
features of the part. They were chosen from experience.
Group two consists of the remaining machine parameters which have to be set according
to the requirements.
Table 2: Classification of Machine Parameters
Group I
Group 2
2nd stage pressure
Screw Recovery Time
Time of 2nd stage pressure
Screw Speed
Back Pressure
Barrel Temperature
Injection speed
Nozzle Temperature
Mold temperature
Cushion
Cooling Time
The action steps are illustrated in Figure 24. According to this an action plan will be
constructed where the second phase in this investigation has been modified accordingly.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
Design oi
P H A S E 2 ExDenmer.ti
PH ASE1
au__s_
[i
Statistical Experiments
• high Experiment
- low Experiment
• mid Point Experiment
a ir
Stipulation of:
- Material
- Mold
- Machine
- Qualicyteatures
m m
wmm
"b
*+ j&i
•=- 3
• -Jv.
Selection of
Parameters
Test for Significance
g?;
Ci
«
I
*
i^A.B.G
I
|
j
iipSScI
rryrf,
Stipulation of optimum j
Machine settings
'M ti
Lonnrmation Trials
Determination
of vananon v
: : :
\7W ,
oerore
arter
si IN - ma_X •
'
Figure 24: Action Steps to analyze the Process [22]
Phase 2 can be executed in different ways. As shown above, two parameters were
narrowed down with a high and a low experiment and then verified with a mid point
experiment as shown in Figure 25.
a*
X
Figure 25: Variation of two parameters XI and X2 with a high and low
Experiment and with a mid point Experiment, a:high Experiment, b:low
Experiment. Center or Mid Point Trial, y: Target value [22]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The method used in this research differs from the described strategy in that no
single parameter was compared but parameter ranges, e.g. the entire profile of the
Injection Speed was considered. However, the principle remains the same. The chosen
parameter ranges have to be determined and the housings produced are measured
accordingly to determine the effect of processing on the product.
Baseline:
Before the process variables were changed the original settings were used to
produce about 20 pieces of almost equal reference parts. After this the parameter change
was carried out and 20 pieces were produced, automatically removed (to achieve best
repeatability of the process) and immediately numbered.
Referring to Figure 26, the pans were measured after two to six hours after ejection,
whereafter no big shrinkage takes place.
P ost Mold Shrinkage of PB T B 4520
107.24
107,22
107,2
E 107.18
E 107,16
107,14
< 107,12
£
■o 107,1
107,08
5 107,06
107,04
107,02
1,03
,1.16
*
!
I
i
1h
as molded
20h
12d
Storage Time
Figure 26: Post Mold Shrinkage of PBT B 4520 Original Part
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Execution
Every trial during the research was discussed with engineers of the department and
prepared according to theory by reading the appropriate literature. After molding with the
original settings a change was executed and parts were produced. A certain time went by
to achieve the new steady state, then 20 samples were collected for measuring. Since the
parts were stacked automatically onto paletts they immediately had to be numbered. If
there was an automatic cycle interruption of the machine the first parts were rejected by
the machine itself to avoid invalid influences, e.g. of material degradation on the results.
Using a fully automatic cycle guarantees the highest possible reproducability and
excludes the operator as a potential error source.
The effect of changing a machine parameter can be roughly estimated in advance.
In this case, however, the exact influence on the housing had to be determined. The
results were all documented and are given in the results section graphically for different
part features, such as weight and dimensions.
7.4.1.1 Holding Pressure
Holding pressure has a big influence on part weight as shown in Figure 27. The
word „original“ means the pressure profile of the original process. This profile had been
used during the first production. In detail: the profile of 80/70/60/50/50/50/45/35/30/25
bar was exerted on the part for five seconds. Since the profile is devided into ten steps,
every step was acting on the housing for 5/10 seconds.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
However, to determine the effect of every parameter the number of required tests
would be too high. For this reason every step of a parameter was set to a constant value
represented by: "H". If the holding pressure has been set to 20 bars throughout one cycle,
this is indicated by 20||.
It was still possible to produce parts with these profiles but up to 50|| bar the
sinkmarks were too severe for real production.
Bolding Pressure
JZ
a>
£
i
origin.
2 0 1|
3 0 1|
4 0 1|
5 0 1|
0 0 1|
7 0 1|
Holding Pressure [bar]
Figure 27: Effect of the Holding Pressure on Part Weight
. However, the dimensions of the part were the important response. Since there was
no difference in trends o f length or width, the length was chosen to show the influence of
processing on shrinkage.
The letters „1“ and „P„ stand for the lengths of the housing. Since one side could be
cooled better, one side had less material than the other and shrank more.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 28 expresses no difference on the effect of different holding pressures.
Holding P ressu re
-
365
“
365
*
364 -----
g 364 x—
0
1 363e
|
------------------
Dimension
-------------
-•_________
* "* |
-----------
_____ -x—----- *--------x
-x— -----* 1
363
origin.
2 0 1|
3 0 1|
4 0 1|
5 0 1|
6 0 1|
7 0 1|
Holding Pressure [bar]
Figure 28: Length I and l'as a function of Holding Pressure
Basically the results of this trial were as expected. However, in this case the exact
influence could be investigated to determine the most critical factors using this machine
and this mold.
In addition to the initial shrinkage after demolding, the shrinkage of the housing
produced with different pressure profiles after the annealing process was checked. Here
the effect on the length is considerably greater than immediately after the injection
process. The trend, however, is basically the same as before (see Figure 29).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
Holding P re ssu re after Annealing
361,60
f 361,40
E 361,20
. 361,00
360.80
§ 360,60
1 360,40
| 360,20
S 360,00
359.80
original
20
3 0 1|
4 0 1|
501
601
701
Holding Pressure [bar]
Figure 29: Length 1and 1’ as Function of Holding Pressure after Annealing
7.4.1.2 Holding Time
Usually this time is held constant during the process and is determined at the very
beginning of the process setup. The gate seal-off time is the determining factor because
only up to this point can pressure be exerted to pack the part. The center of the gate is still
liquid and able to transmit pressure until it freezes. Then the pressure can be removed
because it is useless from this point on.
The principle illustrating how this time can be found is shown in Figure 30 (Weight
of the part on the y-axis and the holding time over the x-axis).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
If the gate freezes off no more material can be pushed into the part, and the weight
remains the same. This is the point how long the pressure should be applied.
_c
3
£
20
Packing Time
(sec)
Figure 30: Determining the Gate Seal Off Time
Using exact the same principle for this part was difficult because of the SemiHotrunner with a hot sprue bushing. In this case it is possible to exert pressure on the part
for a long time and the weight will increase accordingly (see Figure 31). The basics
remain the same but this cannot continued until the weight stays constant. For graphical
reasons this was done but at longer times the parts had flash.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
Weight [g]
{Holding time
270
265
260
255
250
245
5
10
15
20
25
Holding time [sec]
Figure 31: Weight vs. Holding time
The consequences of the holding time can be seen in Figure 32. The steep slope
between 5 and 10 seconds is remarkable. In contrast to the weight, the dimensions remain
almost the same after 10 seconds.
Holding Time
-jr 365,50
JL
365,00
*
S64.50
co
’«
364,00
363.50
eo
363,00
362.50
5
10
15
Holding Time [sec]
20
25
Figure 32: Holding Time and Dimensions
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
The results for part dimensions after annealing are shown in Figure 33 and look
very similar to the results prior to annealing.
Holding Time Effect after Annealing
g
E
1_
2
|
|
|
|
361,80
361,60
361,40
361,20
361,00
360,80
i
360,60
360,40
5
10
15
20
HoldingTime [sec]
Figure33: Dimensions after Annealing as Function of Holding Time
7.4.1.3 BackPressure
The back pressure is responsible for the homogeneity of the melt. The higher the
pressure, the higher is the shear stress, and consequently the higher is the heat dissipation.
Therefore, higher temperatures lower the viscosity of the polymer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
Figure 34 shows the effect of the back pressure on the weight of the housing.
jBack P r e s s u r e
i ,--------------------------- :---------------------------- ;----------------------------:----------------------------4;----------------------------1
5 1|
1 0 ||
1 5 ||
2 0 ||
3 0 ||
4 0 ||
B ack P r e s s u r e [bar]
Figure 34: Part Weight as function of Back Pressure
The length of the part responded slightly to the different Back Pressure profiles.
With increasing Back Pressure the part length steadily increase.
Back P re ss u re
365.00
364.90
364.80
364.70
364.60
364.50
364.40
1 5 ||
2 0 ||
B ack P r e s s u r e [bar]
Figure 35: Dependence o f Length on Back Pressure
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
The same effect could be seen after annealing (see Figure 36).
Back P r e s s u r e - Annealing
^
¥
JL
*
3 6 1 .8 0
3 6 1 .6 0
3 6 1 .4 0
3 6 1 .2 0
3 6 1 .0 0
r
»
3 6 0 .8 0
3 6 0 .6 0
•
3 6 0 .4 0
3 6 0 .2 0
—♦ —Dimension I
—x — Dimension P
1 0 ||
1 5 ||
2 0 ||
3 0 ||
I
4 0 ||
B a c k P r e s s u r e [bar]
Figure 36: Dimensions after Annealing as a Function of Back Pressure
7.4.1.4 Injection Speed
The injection speed profile is varied to achieve an equal melt front velocity as the
part fills. Figure 37 shows a mold filling study which shows the filling pattern of the part.
It can be seen how the melt is filling the part at different time increments.
•
«
«
n
i
Figure 37: Filling Study of the Housing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
It is seen that the velocity profile is especially important A rule of thumb says the
speed should be slow at the beginning to generate a good surface quality without jetting
or leaving a dull injection gate area. Than it should increase in the middle and slow down
at the end to allow a smooth switching from injection to holding pressure (see Figure 38).
Since the part is quite complicated this general rule could not be applied in this
case. Without a specific profile the housing could not be produced without any defects.
:0
is
End of Filling
Figure 38: General Injection Speed Profile
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
The original profile was determined by a preliminary investigation as
25/18/25/23/30/30/35/30/55/50 . This profile was able to fill the part
properly even though it deviates from general rule.
The variation o f the injection speed was restricted mainly between 30|| and 40||
mm/s. Lower speeds lead to short shots and higher speeds distorted the polymer so much
that part cracking during ejection was the result.
Figure 39 shows the effect of injection speed on part weight where the original
profile was executed once with a Back Pressure of 10 bar and once with a Back Pressure
of 40 bar. It can clearly be seen, the part weight depends more on the Back Pressure than
on the injection speed.
Injection S p e e d I
257.50
257.00 _
256.50 .
256.00 .
si 255.50
.
£ 255.00 .
.51
v
> 254.50 _
253.50.
.
_
_
253.00
252.50 __
.. . . . . . . . .
______
30(1 ( 10 bar )
( ) =backpressure
....
__ __________________ _______
.
__
________ ______
_______ ______
40,100A c tio n Speed [rmvs]
Figure 39: Weight as Function of Injection Speed
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
orig^(40)
61
Compared to the weight, the length does not correlate that well to the first curve of
Figure 39. The weight is almost equal at 40|| and the original settings with 10 bar back
pressure. The length is at its minimum at the origin (as shown in Figure 40).
Injection S p e e d
365.00
364.95
E 364.90
E
364.85
c 364.80
0
'*? 364.75
c
E 364.70 .
364.65 :
364.60
.
3011 ( 10 bar)
( ) = back pressure
40II (10 bar)
original (10 bar)
I n je c H o n s p e e d [m m /s]
original (40'
9
Figure 40: Influence of Injection Speed on the Length
For the evaluation after annealing all housings were measured including the cracked
ones which only cracked around the gate area.
In Figure 41 it is clearly indicated that the original setting at 10 bar is between the
two possible injection speeds of 30|| and 40|| mm/s.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
361,80 _
361,60 i—
— 361,40 J—
E
i 361,20 —p.
” 361,00 —
00
c 38080 ! _
3.
\
1 38000^:
E
5 38040 J—
36020 j—
i
38000 |—
30 ||
301|
4 0 1|
501|
Injecdonspeed [nrnfe]
orig'ndQ
arign(4Q
Figure 41 Length after Annealing as Function of Injection Speed
Figure 42 shows the melt temperature as function of injection speed. Up to 30||
mm/s the temperature is rising steadily while from that point on, however, even an
excessive increase in speed does not effect the temperature very much anymore. The
material is destroyed. The temperature was measured after injection through the cavityside of the mold via a thermocouple but not taking gate or additional shear heating in the
cavity into account.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
AMtTenperatra Dependence on Injecfon Speed
2B5 —
•° 275 T>
2 2X3
a 265 ■
200
________________
..
25 .
cjign.
— ...........
10
--
-
15
20
.
-
30
.
40
50
.
60
_____
9*
Injection Speed [rnrfsl
Figure 42: Effect of Injection Speed on Melt Temperature
7.4.1.5 Mold Tem perature
The impact o f the mold temperature on the weight can be seen in Figure 43. The
decreasing weight is related to the increasing specific volume of the polymer at higher
temperatures. Where the values on x-Axis stand for the "Coolant Temperature".
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
255,00
254,00 I
253,00 x
01
I
252,00
251,00
250,00
249,00
40/60/40/45/20
75/92/75/80/40
90 I
Mold T em perature [°C|
Figure 43: Weight vs. Mold Temperature
The effect o f mold temperature on the dimensions is shown in Figure 44. As
anticipated the part dimensions are getting smaller at higher mold temperatures due to a
higher degree o f ciystallinity.
Mold T e m p e r a t u r e b
365,00
364,00 ;
•g 363,00 i
E
j
S
362,00 X
361,00 ...
360,00
359,00
40/60/40/45/20
75/92/75/80/40
Mold T e m p e ra tu re *
Figure 44: Dimensions vs.Mold Temperature [°C]
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
65
7.4.1.6 Cooling Time
The effect of this parameter was investigated by using different coolant
temperatures vs. the original cycle. They were at 75/90/75/80/40°C to get an idea of how
this will influence the part regarding higher cooling temperatures.
Looking at Figure 45, it can be seen that the parts are getting larger with increasing
Cooling time. This is due to better heat conduction within the mold during longer cooling
compared to the insulating properties of the air where the "short time" cooled parts still
can crystallize. On the other hand it must be taken into account that the warpage will also
increase because of the reasons mentioned just now and the fact that there is no fixturing
of the part once it is out of the mold.
jCoding TimBji
362^96
-
362,90
E 362,80
E
362,65-------------------------------------------------------------------------------------------------------------------------
36260 ----------------------------------------------------------------------------------------------------------------36255 j----------------------------:--------------------------46
55
66
1--------------------------- ,
75
Cooling Timofei
Figure 45: Dimensions as a Funktion of Cooling Time
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
66
This effect cannot be noted after annealing because the overall shrinkage is reached
for almost all parts and this value is mostly a material function. Only the parts cooled for
a very long time show a small deviation from this behavior. This might come from the
annealing time which is based on the part produced under normal conditions.
lad in g lirtB Effect after AmedingL
361,00 .------------------------------------------------------------------------------------------------------------
361,20.------------------------------------------------------------------------------------------------------------
E
361,00-------------------------------------------------------------------------------------------------- ;-- -------I 36080 i----------------------------------------------------------------------------------
m
|
;----------------------1-------------------------
— -----
—
36000 i-----------------------------------------------------------------------------------------------------------30040 -____________________________________ ___________________________________________________________________________________________________________
30020 4------------------------------------ ------------------------------------ ------------------------------------------------------------------------- 46
55
66
75
85
Coding Tims [s]
Figure 46: Dimension after Annealing at different Cooling Times
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
7.4.1.7 Summary of Quality-Determining Parameters
An overview of the results is shown in Figure 47. The weight and shrinkage of the
housing are drawn along the Z-Axis, Processing Conditions on the X-Axis and the
Properties of the Housing on the Y-Axis.
Explanations of the Diagram:
1) Dim.Diff.1:
stands for the range in [mm] of a certain
(Range 1)
dimension while varying one parameter
(e.g. holding pressure from 20-70bar)
2) Weight Diff:
stands for the range of the weight [g] while
varying one parameter.
(e.g. holding time from 5-20 sec.)
3) Dim.Diff.2:
stands for the range in [mm] for one specific
(Range 2)
setting after annealing
4) Shrinkage:
stands for the post shrinkage after annealing
except: mold temperature: here only dimen­
sions after injection (common post shrinkage)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
It is easy to see in the diagram that process parameters mainly influence weight
differences which get emphasised even more by two peaks at holding pressure and
holding time.
■ 12-14
□ 10-12
□ 8-10
I ae-8
! ■ 4-6
; □ 2-4
S io
a
a
£
5
total Shrinkage
i Range 2
- Difference in Weight
Range 1
no-2
Figure 47: Influence of Processing Conditions on the Housing
(Dimensions and Weight)
For this reason the weight was eliminated from the next diagram (Figure 48). Clearly it
can be seen that the mold temperature has the highest influence on shrinkage, indicated
by Dim.DifF. 1. The peak is caused by the shrinkage difference between parts out of the
original process and a constant mold temperature of 90°C.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
i
□ 2.5-3 :
□ 2-2,5
□ 1.5-2
i m Q.5-1
totoal Shfikage
Range 2
Injection
Speed
Maid
Temperat
Bade
pressure
Die Head
Press.
HoldngTime
Rangel
coding
Time
Figure 48: Processing and Shrinkage Relationships
The difference in dimensions
after molding (Dim.Diff.l) and annealing (Dim.Diff.2) can
be seen in Figure 49.
2,5-
Inject Speed
2'
£
%
6
o*
m
*S
£C/3
Mold Temp.
1,5-
Press.
10,5-
Processing
0
Range
Cool. Time
Range
Figure 49: Length vs. Processing Conditions
reproduction prohibited without permission.
ith permission o f the copyright owner. Further
Reproduced with
70
Figure 50 shows the Influence of Processing on Product Properties.
|^lchrfFhicaMg|
Ctrins
Tr«
■«-«-»-roans
Tire
BBck
ftesare
^-IjgttiShrleEp
j
tOd
Ftassue
;
I j
i jifta tp l
J
-1- » •
Mid
Tenrp
j
{■
Irjectkn
Speed
0
2
4
6
8
1
0
1
2
1
4
EUcr. [nnj or [g
Figure 50: Quantitative Correlation between Processing and Product Properties
7.4.I.8. Results Concerning Quality Influencing Parameters
The most important parameter in terms of dimensions was the MOLD
TEMPERATURE. Since this effect was so clear it was separated from the other molding
parameters. It was determined that after the investigation an oil heating device should be
ordered to eliminate the annealing process entirely.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
Thus, the most important processing parameters are:
a) Mold Temperature
b) Holding Pressure
c) Holding Time
d) Back Pressure
Since the "Oil Heating" cools the part more slowly the annealing process should be
investigated in more detail.
7.5 Quality Influencing Parameters during Annealing
7.5.1 Temperature and Time
The currently used process was previously established and able to generate good
parts. The settings were given at 100°C for two hours in the oven. The excessive period
of time was the reason for asking how would the result be at higher temperatures and less
time?
For this reason two trials were carried out:
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
1) Temperature at 125°C / 70min
2) Temperature at 150°C / 70min
Based on these data the temperature setting in the oven was switched accordingly.
The actual temperatures achieved were measured via a thermocouple on the surface of the
housings. During unloading the heated oven cools down to a temperature of about 60°C
and one has to wait for a certain amount of time until the desired set temperature is again
reached. However, investigations regarding the structure of the material are beyond the
scope of this work. So, only the effects on the housing were measured. For this five
samples were taken out o f the automated process (65sec Cooling Time and Back Pressure
15bar) and anneald at the above mentioned conditions.
7.5.2 Results of Annealing
The actual temperature achieved at the setting of 125°C after 50min heating was at
about 121°C to 126°C, which could be applied for a further 20 minutes. Doing so yielded
a time saving of about 50 minutes with improved dimensions. At a setting of 150°C the
temperatures after one hour were about 148°C to 152°C again with an overall time of 70
minutes. Now, however, the dimensions were mostly too small!
The result of the annealing trials are given in Figure 51.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
Cooling Time 65sec
Cooling Time 65sec
(5 samples each trial)
(5 samples each trial)
Width [mm]
106,20
106,00
105,90
105,80
361,50
I
361,00
^
c
"
360,50
■C
360,00
§> 359,50
105,70
100
-
E
106,10
120
"
150
I
359,00 -l-
100
Annealing Temperature [°C]
Back Pressure 15 bar
(5 samples each trial)
Width [mm]
150
Annealing Temperature [°C]
Back Pressure 15 bar
(5 samples each trial)
E
E 361.50
106,20
106,10-»
106,00
105,90
105,80
105,70
100
120
361.00
■co 360.50
a 360.00
359.50
359.00
a 358.50
c
120
150
Annealing Temperature [°C]
a
^
|
4-
100
120
150
Annealing Temperature [°C]
Figure 51: Length and Width of the Housing as a function of Annealing Temperature.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
1.6 Oil Heating
Using the fact that higher annealing temperatures could be used, the o-ring seals
were upgraded and the hose system of the mold had to be changed to metal tubes. Due to
a lack of oil heaters the mold could be only heated by two circuits which was acceptable
for determining the principle outcome of this trial at mold temperatures between 104°C
and 112°C.
7.6.1 Settings
Using the results of preliminary trials the most important factors chosen for
variation in this trial were:
1) Holding Pressure
2) Holding Time
3) Back Pressure
The previous investigations showed that the mold temperature had to be set at a
constant value at the above mentioned temperatures. The exact settings of parameters
were discussed with engineers of the department and could be used in a modified way
compared to a production run.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
a)
For the Holding Pressure the following profiles were chosen:
1) "75" =75/60/55/50/45/40/35/30/30/30 bar
2) "70" = 70/00/50/50/40/40/30/30/30/25 bar
3) "60" = 60/55/50/45/40/40/35/30/30/25 bar
b)
Holding Time was set at: 5 and lOsec. At 15 seconds the first
flashing occurred.
c)
Back Pressure was set to 5,10 and 15 bar.
In this row the X-Axis carries the process code, e.g. (60,5,5) which means that this
specific part was produced at a holding pressure according the profile "60", at a time of 5
sec. with a back pressure of 5 bar. The rest of the parameters remain unchanged from the
original ones. The dimensions were determined by measuring 5 parts and determining the
median value.
7.6.2 Production with Oil Heating
By using oil heating the robotic part removal device had to be adjusted. In contrast
to the original cycle were the parts could be stacked on palletts and stored for an uncertain
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
time until they were reheated again in a needless time consuming step, now finished parts
were produced. They immediately had to be clamped on the prepared fixtures to minimize
warpage due to different shrinkage caused by the different wall thicknesses. The fixturing
time should last at least one hour to guarantee a set during cool down.
Since the fixturing units were not available in that number the required time could
not be maintained. The result was warped pieces due to different wall thicknesses and the
uniform mold temperature. The main task of this investigation, however, was just to
prove the possibilty of achieving good parts without annealing. Warpage could be
minimized afterwards, if the oil heated mold worked, by adding more clamping devices
and using a better mold temperature distribution.
7.6.2.1 Holding Pressure Profile "60"
In the first diagram [Figure 52] the correlation with regards to the quality of the
process "60" and the dimensions is shown. Here the influence of the holding time can be
seen as indicated by threee peaks. This effect is amplified by the fact that dimension "B"
is measured close to the gate.
Furthermore, a slight growing of the part can be recognized with increasing back
pressure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
1 0 6 .1 5 g
1 0 6 ,1 0
E
e
■
E
106.05
s
105,95-1
D im . C
10 5 .9 0
to
io
D im . B
D im . A
to
P ro ce ssin g Profile *60
Figure 52: Processing "60" vs. Dimensions A 3 and C
A bar graph of the above mentioned dimensions based on the processing profiles
is given in Figure 53.
106,20
Dimen*, (mm]
106,15
106,10
~
106,05-
D im . A
fi] D im . B
106,00-1
■
D im . C
105,90
Processing Profile *60’
Figure 53: Production Quality with profile "60"
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
Basically the same tendencies are given for the length of the housing [Figure 54].
361,20
361.004
□ Length I ’
360.404
□ Length I
360.20
360,004
359.80-U
60,5.5
60.10,5
60,5,10
60.10.10
Procssting Profile *60'
60.5.15
60.10.15
Figure 54: Length as a Function of Processing with Holding Pressure Profile "60"
1.6.2.2 Holding Pressure Profile "70"
The results o f this molding can be seen in Figures 55 to 57.
106.25
106,20
106,15
1 0 6 ,1 0 4
106,05
4
1 0 6 ,0 0 105,95
105.90
1 05,85
D im .C
D im .B
D im .A
P r o c e s s i n g Prof i l e "70
Figure 55: Processing "70" vs. Dimensions A,B and C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
106,25 106,20-
Dimen*, (mm]
106,15
106,10: ~ ] D im .A
106,05106,00
m
I 0
D im .B
! ■
D im .C
105,95
105,90105.85
Processing Profile ”70’
Figure 56: Production Quality with Processing Profile "70"
361,40-
361,00
□ Length I'
O Length I
in
in
in
in
Processing Profile "70"
Figure 57: Length as a Function of Processing Profile "70"
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
1.6.23 Holding Pressure Profile "75"
The results of this trial are shown in Figures 58 to 60.
106,25
106,20
106,15-.
E
E
m
s
E
5
106,10-4
106,05-|
106,00
105,95
105,90
105,85
D i m .C
D i m .B
u>
P r o c e s s i n g Profile "75*
id
u>
(O
p
*.
im .A
*
—
o
IX
)
Figure 58: Processing "75“ vs. Dimensions A,B and C
106,20
Dimens, [mm]
106,15
106,10
Dim. A
106,05-;
□ Dim. B
106,004
■ Dim. C
106,90
105,85
Processing Profile "75"
Figure 59: Production Quality with Processing Profile "75"
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
361,40361,204
361,004
360804
o> 36040-
□ Lengthl'
E3 Lengthl
Processing Profie "75’
Figure 60: Length as a Function of Processing Profile "75"
7.6.2.4 Summary: Oilheating
Due to a lack of the required number of oil heaters and a too short clamping time
the parts warped. The oil heated mold produced parts which were sometimes on the lower
end of the specified tolerances. A modification to the oil heated mold would eliminate
this potential issue. In particular the mold had to be modified anyway since it produces a
too high deflection of the part during ejection at elevated temperatures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
The results of all oil heated pocess conditions can be seen in Figures 61 and 62.
Since Figure 61 only shows the dimensions B and C clearly, a bar diagram was generated
in Figure 62.
1 0 6 .2 5 10 6 .2 0 > f
I
106,15
:f
■ 106.2 0 -1 0 6 .2 5
■ 106.1 5 -1 0 6 .2 0
o
1 0 6 .1 0 4 -
*
106,05
□
1 0 6 ,10-106,15
106 .0 0 i t
□
106, ,05-106,10
a
105,95 -f
□
106, ,00-106.05
105.90 4- ‘
■
105, 9 5 -1 0 6 ,0 0
105.85
a io5, 9 0 -1 0 5 .9 5
U 105, 8 5 -1 0 5 .9 0
to
10
P r o c a a s l n g P ro flla
Figure 6 1 :3-Dimensionai Picture of the Dimensions A 3 and C vs. Processing
106,25
I
106,20 4
106.15
106,10
Dim. A |
a,' 106,06 * -
Dim. B .
p 106,00
Dim. C ;
<
* ~
106.95
106,90
106,85
IO
to
o
o
IrO
— to
ID O
o
CO
<o O
ID
to
o
to to o to to o to IO
to to IO
to 10 o' o' o o
lO
o
'
to rtc
T o'
o
'
r
d
fs.
v
o
d
r*
<0
CO
CO
o tr“o to
o o' o
IO
d
r* o' rs
o
o
to
r*
to
o*
to
r*
Processing Profile (Holding Pressure Profile, Holding Time, Die Head
Pressure)
Figure 62: Line Diagram of all Dimensions A 3 and C
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
In this diagram it becomes obvious that the dimensional range for the width A,B
and C increases while the holding time gets enhanced from 5 to 10 sec.
It can also be seen that a relativly constant process is reached between the profiles
70/5/10 (Holding pressure, Holding time, Back Pressure) and (75/5/15). Within these
limits a stable process was obtained which would be able to adjust small fluctuations by
itself. Essentially the same can be seen in Figure 63 for the length of the housing.
361 .40
36 1 .20 -U -W
360.80
360.40
□
D im . I
m 1! L illi i H m FT
Figure 63: Length of Housing vs. Processing Profile
A confirmation of the former statement is given by the weight diagram below. It
shows a very uniform weight distribution over a wider process range can be achieved
between the profiles 20/5/10 and 25/5/15 (Figure 64).
».
2
S
2
2
«.
2
£
5
2
) M ( i M
2
2
£
5
i | | l !
S
2
j
P r o c . l n o
Figure 64: Part Weight as Function of Molding Conditions
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
7.6.2.5 Results o f Oil Heating
The results have shown that the annealing process can be avoided by using an oil
heated mold. A time saving of 2 hours per 16 parts will be the outcome. Moreover a
stable processing window has been found which will be able to adjust small process
fluctuations by itself. After finishing this investigation which was executed in a very
similar way as the preinvestigation it was proven that the oil heated mold was capable of
meeting all expectations which were desired in the trials.
In other words the dimensions of the part were most often better than those after
annealing for two hours at 100°C. Unfortunately, the mold temperature could not be
adjusted due to a lack of oil heaters although this was necessary to avoid warpage. Also,
the right time in the clamping unit could not be applied. Therefore, the parts out of these
trials were warped.
Partial solutions have been found by modifying the gate and the clamping unit. The
most important steps could not be executed but were anticipated. These steps could not be
carried out because equipment was not available ( Oil Heating Units and Fixturing Units).
This means after enough fixtures are obtained, the process can run without annealing.
This will save 120 minutes for 16 housings after a certain time for optimization and some
smaller modifications to the mold to allow higher mold temperatures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
After investigating and selecting the most important processing parameters it was
proven that the process was running within a very narrow range even though it was first
out of spec. This range can be pushed into specification limits by oilheating.
A stable process window was found from the profile "70,5,10" (holding pressure,
holding time, back pressure) to the profile (75,5,10). This window is valid for both widths
(A,B,C), lengths 1,1' and weight.
7.7 SUMMARY and CONCLUSIONS
7.7.1 M aterial
Together with processing investigations the conditions for loading new material of
PBT resin have been tested in terms of relative moisture content. New delivered material
can be used without any additional drying. The moisture content guaranteed by the
supplier is low enough for immediate processing.
7.7.2 Process
After documenting the outstanding repeatability o f the injection molding machine
by using SPC Methods, the most important parameters were found by a design of
experiments (DOE). Once the most influential process parameters were found the
optimum operating point was determined (see section 7.4.1).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
The most important parameters were:
1) Mold Temperature
2) Holding Pressure
3) Holding Time
4) Back Pressure
To avoid excessive post mold shrinkage during use in a microwave oven, the part
had to be annealed based on the assumption that the old processing conditions were used.
Therefore, the annealing process was also checked. Here the cycle time of 16 parts could
be reduced by 50 minutes, from 120 minutes at 100°C to 70 minutes at 125°C (see
section 7.4.2).
Annealing as a second production step should be avoided. Hence, parts were
produced by using an oil heated mold to allow higher in-mold shrinkage. The coolant
temperatures of all two circuits were about 120°C. This fact and the insufficient clamping
time in the fixturing lead to warped housings. Nevertheless, the results are valid and give
the real implications of the process running with the right equipment. Here a wider
processing window with an uniform mold surface temperature was determined using
holding pressure, holding time and back pressure (see also section 7.4.5).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
VIII. RECOMMENDATIONS
In addition to the high precision process of injection molding quality control during
the fully automatic production mode will gain more and more importance. However,
pure machine data control will no longer be sufficient. One needs to document good
detailed data relating to every single part and how it was processed. In this case SPC
(Statistical Process Control) and CPC (Continuous Process Control) can be combined to
reach the above mentioned goal. This method as explained earlier in this work stands
basically for Product Quality (SPC) and Process Quality (CPC). However, a real
relationship between Process Parameters and Product Features has to be found first. To
what extent this can be executed for every specific part depends on how accurately the
interaction between process and part can be determined and controlled. After obtaining an
optimized process an ideal curve of a certain process parameter can be saved and a lower
and an upper Boudary Curve" can be determined based on the relationship between part
and process which will produce good quality parts within its bounds (as shown in Figure
65). If the process, and hence, the product quality run outside the specifications the
process will be stopped automatically. The interruption caused by an unexpected
temperature drop or rise of the oil is already a good practice but this is only related to the
process and has nothing to do with the product itself!
This concept is the key to a closed loop controlled machine. As a long range goal
one should be able to record product quality influencing parameters and the machine
should be capable to react to changing product features fully independently.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 65: "Servo Curve" Controlling
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
Literature Cited:
[1]
Masing, W.: Handuch der Qualitatssicherung. 2. Aufl. Hanser,
Munchen 1988
[2]
Kunststoffe 82(1992)7, S.568, Ergebnisse der SPC-Anwendung an
SpritzgieBmaschinen
[3]
DIN ISO 9004, Teil 2: Qualitatsmanagement und Elemente eines
Qualitatssicherungssystems, August 1990
[4]
DGQ- Schrift: Qualitatskosten, Rahmenempfehlungen zu ihrer
Definition, Erfassung, Beurteilung. 4. Auflage DGQ 14-17, BeuthVertrieb GmbH, Berlin
[6]
Thienel, P.: EinfiuB der Fertigung auf die Eigenschafien von
Prazisions-SpritzguBteilen. Seminar VDI- Bildungswerk
"SpritzgieBen thermoplastischer Formteile", Iserlohn, Nov. 1988.
[7]
Gimpel, B.: Qualitatsgerechte Optimierung von Produkten und
Prozessen, Vortrag beim SWO-/GfQS- Seminar " Qualitatssicherung
in der Kunststofftechnik". Aachen 1990
Raulefs, P.: Expertensysteme. Kiinstliche Intelligenz, Friihjahrschule
Teisendorf 1982. Informatik Fachberichte Bd.59, Springer Verlag
[8]
[9]
American Supplier Institute. Inc., Center for Taguchi Methods,
Dearbon/MI, USA
[10]
Kersten, G.: FMEA- Methoden zum systematischen Vermeiden
potentieller Fehler in Konstruktion und ProzeBplanung, S.23, in "Das
optimierte SpritzguBteil", VDI- Verlag GmbH, Diisseldorf 1989
[11]
Ford : Statistische ProzeBregelung, Leitfaden, April 1985, S.51
Forster, K.: Unterlagen Vorlesimg: "Qualitatssicherung", S. 4.5, FH
Rosenheim, 1991
[12]
Bosch: Hinweis und Erleuterung zur Bosch-QualitatssicherungsLeitlinie fiir Lieferanten, Ausgabe 1, Februar 1989, S. 19
[13]
Forster, K.: Unterlagen Vorlesung: "Qualitatssicherung",
S. 4.4-10/22 - 4.4-14/ 22, FH Rosenheim, 1991
[ 14]
Qualitatsrechtlinie Q .101, einschliefilich Leitfaden Statistische
ProzeBregelung, Ford Werke, Koln, 1990
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
[15]
Prof. Forster, K.: Unterlagen Vorlesung: "Qualitatssicherung",
S. 4.4-10/22 - 4.4-14/ 22, FH Rosenheim, 1991
[16]
Lampl, A.rEngel: Fachtagung 1991, Anhang S.7
[ 17]
Printout o f Injection Molding Machine ES 700/200
ENGEL, Austria
[18]
PleBman, K.; MichaeliW.; Koske, J.; Heine, J.; Cremer,M; Gunzel,R.
Klee, D.: Fertigungsparameterbestimmen Formteileigenschaflen mit
SpritzgieBen von Schlagzahmodifiziertem Polystyrol. KunststofFe 81
(1991) 12, S.1141-1144
[19]
Vieweg,R. und Joersten: Kunststoff- Handbuch, Band 8 Polyester,
Munschen: Carl Hanser Verlag
[20]
Screen on the Injection Molding Machine ES 700/200
ENGEL, Austria
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
Biographical Sketch of Author
The author had the opportunity to join UMass Lowell from Fall '91 to Spring of '92,
attending the Masters Program in Plastics Engineering, after he obtained the German
equivalent of the B.S. Degree at the Fachhochschule Rosenheim. He worked three years
at a company in the North East of Munich. There he worked as a project engineer and part
designer for plastic parts. His primary interests lie in Injection Molding and Design of
Plastic Parts. At the moment he is working as a field worker for the packaging industry.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
3 027 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа