E410 PDF

NORTH AMERICAN DIE CASTING ASSOCIATION
DIE CASTING
PROCESS CONTROL
Average
Triangle
Approximating
Histogram
R/3 R/3 R/3

R
E.A. HERMAN
PUBLICATION: E-410
Die Casting
Process Control
E.A. Herman
Although great care has been taken to provide accurate and current information, neither the author(s) nor the publisher, nor
anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged
to be caused by this book. The material contained herein is not intended to provide specific advice or recommendations for any
specific situation. Any opinions expressed by the author(s) are not necessarily those of NADCA. Trademark notice: Product or
corporate names may be trademarks or registered trademarks and are used only for identification and explanation without intent
to infringe nor endorse the product or corporation. © 2012 by North American Die Casting Association, Arlington Heights,
Illinois. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, elec-
tronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system,
without permission in writing form the publisher.
First Printing 1988

Second Printing 1991
Major Revision 2003
Minor Edits Spring 2012
ii
Table of Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Chapter 1: Introduction to Control Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Chapter 2: Measuring Process Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapter 3: Statistical Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Chapter 4: Holding Furnace Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Chapter 5: Plunger Velocity & Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Chapter 6: Clamping Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Chapter 7: Die Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Chapter 8: Release Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Chapter 9: Casting Ejection Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Chapter 10: Cycle Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Chapter 11: Mechanical Die System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Chapter 12: Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Chapter 13: Process Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
iii
Preface
The die casting industry has now progressed to where the degree competitive brands of control equipment. Any specific equip
of process control that can be routinely applied is at least ten ment used to illustrate a point was selected only because of
times (and possibly 100 times) that of just a few years ago. The the suitability and availability of the printable material as
previous edition of this book started here by saying the industry this book was being written. Once a suitable illustration was
was in a “revolution of process control”. That revolution has come found, no effort was made to search out others. This book does
and gone, and practical, in fact necessary, controls must now not describe die casting machine functions, maintenance or
be applied to the process if the die caster is to compete in the fault diagnosis. The assumption is made that the reader knows
market place. Traditionally the control of the process was accom how the machine and die systems work and that everything is
plished by the machine operator. He would inspect the castings working properly and can therefore be adjusted as necessary
as he made them and then make adjustments to the operation to achieve the desired results. The mechanics and engineering
of the process. The success of any particular operation depended design criteria for the various machine systems are presented in
substantially upon the skill of the operator and entirely on the the NADCA course on Machine Systems. This book does not
fidelity of his physical senses. Such manual feed back control explain how to determine what the “set point” operating values
loops do not have sufficient sensitivity, responsiveness or reliabil of the various variables should be. Other NADCA textbooks
ity to maximize the potential of the die casting process. and courses such as Gating, Engineering Die Casting Dies and
Engineering Die Cooling Systems show how to make those
The modern die casting manager knows that to realize the full processing set point calculations. The previous edition went into
potential of the process, he must have both electro-hydraulic- considerable detail on how to establish some, but not all, of the
mechanical feed-back control devices and the process engineer set points and was accordingly titled “Process Engineering and
ing and operating skills to use those devices properly. This book Control.” Some instruments used for controlling the process are
and the associated certificate earning NADCA course have equally useful for diagnosing improper functioning of the die
been created to help develop those skills. Each major process casting machine. So when a particular type of instrumentation
variable is identified, its natural behavior described, and the is discussed in this book, the reader should be aware that there
required control scheme is defined. could be other uses for it.
For definitive cause-and-effect relationships between the The purpose of this book is to show the processing engineer and
processing variables and casting quality, the reader is referred operating technician what type of control method is applicable
to other NADCA courses such as Heat Flow, Die Casting to each of the die casting machine/die/process systems, how
Dies: Designing, Dimensional Repeatability, Gating and to measure the performance of each critical variable, compare
Metallurgy. This book does not describe the set up, calibra actual performance to the desired and to adjust the actual per
tion or maintenance of the process controlling instruments formance to meet the desired. The reader will find this edition
and/or equipment. That information must be obtained from more definitive and focused than the previous editions.
the equipment manufacturers. This book does not evaluate
Die Casting Process, Engineering and Control

2
CH APTER
Introduction to Control Theory
Control Theory lack of a coherent theory of the causes and effects relationships
between specific adjustments, machine/process performance
There are four essential elements in any process control func and casting quality. The machine operator did not have all the
tion. These are: tools (both physical and informational) that he needed. The
1. Predetermined standard (ideal) condition results have been a catch-as-catch-can situation where both
good and bad castings are produced on a somewhat random
2. Measurement of actual condition pattern. Inspection techniques are employed to sort the good
3. Comparison of actual to standard from the bad as illustrated in Figure 1-2.
4. Adjustment to process
Figure 1-1 illustrates these four elements. A typical process may
have many inputs. Some may have a stronger influence on the
output of the process than others. Sometimes one input has such
a strong effect that it is the only one that needs to be controlled.
The accelerator pedal in and automobile is such an example.
There are many inputs that determine the exact speed that the
vehicle is traveling at any instant. Some interpretation by the
driver of the speed limit sign establishes the desired speed. The
speedometer registers the actual speed. The driver then decides
to adjust the speed based on the difference between the actual
Fig. 1-2. Many accepted inspection techniques are invaluable for segregating bad
and desired speeds and adjusts the accelerator pedal accordingly.
parts from good ones. However, such sorting does not constitute process control.
The driver does not concern himself with the exact position
of the pedal. He pushes it down farther to go faster, and that
action overcomes all other inputs. Fortunately, the die caster now has both the coherent processing
theory,1-5 and the equipment6-8 to operate a die casting machine
“in control.” This book explains how.
Generally a process is considered to be out of control when the

parts it is making do not meet the customer’s specifications.
Technically, a process is out of control anytime there is no
positive assurance that a particular process variable is operating
within some predetermined range of conditions that is statis
tically “in control” or “normal” for that process. If operating
within such a predetermined range of conditions produces
product to the customer’s specification, then the process can be
used to make that product. However, if the process is operat
ing within that “normal” span of variation (i.e. the process is
“in control”) and is not (can not be) making the product to the
customer’s specification, then the process is “not capable” and
Fig. 1-1. Processes are controlled by measuring the process variables, comparing
some other process must be used. Such an “other” process can
the actual condition to some desirable standard and making appropriate adjust-
be the original process with upgraded capability through the
ments to the process. Tradit ionally, die casting has been controlled by a person
addition of better process controls. Otherwise it is useless to
performing all three functions.
try to make that product with the process (or machine). The
first step for the processing engineer, is to establish the ideal
Traditionally, the die casting machine operator performed the processing conditions and compare these to the capabilities of
process control function. His effectiveness was limited by the his die casting machines. Then, only if his machines can meet
insensitivity of the human senses, the need to guess the actual the required performance should he attempt to make the par
machine or process performance by observing the casting’s ticular casting. Generally, this book assumes that such process
condition rather than measuring actual performance, and the capability evaluations have been made and that the machine
Die Casting Process, Engineering and Control 5

is capable of the task. However, there is a chapter on the basic Processing variables that have continuous drift in one direc
concepts of measuring process capabilities. tion are ideally suited for statistical process control (SPC)
methods. For example, if wear of a cutting tool (or a surface
in a die casting die) makes a dimension larger, successive parts
Process Control Theory will gradually get bigger. If an appropriate number of parts are
checked for size at appropriate intervals (e.g. five parts each
Process control theory is that if one controls the inputs, the hour) an SPC average and range chart can be plotted to show
output of a process will be predictable and hence the output graphically the process drift. Then corrective action can be
is not routinely measured to “control” the inputs. When all of taken before the process makes parts with dimensions outside
the significant inputs are known and controlled, the output is the pre-established range. Wear of die cavity surfaces and dete
indeed predictable. This concept is diagrammed in Figure 1-3. riorating hydraulic components of the die casting machine can
cause such continuously drifting situations.
More typical of die casting variables is cyclic drift. Die tem

perature is an example of cyclic drift. Over a period of several
hours, the temperature at any one spot in the die will vary
substantially. Sometimes hot, sometimes cold, and ever ywhere
between. Periodic inspection and statistical process control
techniques are not generally effective for controlling cyclic
drift variables. By the time an out of control condition is found
or predicted, the performance cycle may have reversed itself.
Then any adjustment to the process will be wrong. Continuous
control is required for these variables.
A threshold condition is when a cyclic or continuously drift

ing variable causes a catastrophic failure to some critical
Fig. 1-3. Process control theory is that by controlling all the inputs, the output
actual performance value. Often, the variation prior to the
will be predictable. The output, which in the case of die casting would be the cast-
catastrophic event has little or no effect on the product being
ing, is not measured and used in the control function.
made. The temperature differential of the plunger and shot
sleeve is a process variable with a threshold condition. As
these temperatures change (cyclically) the plunger and bore
It often turns out that each input to a manufacturing process change sizes. Within a range of sizes there is no impact on the
has a feed-back control loop like that shown in Figure 1-1. castings being made. But at some point, the plunger will stick
There is some “process” that sets the input to the manufac with dramatic results to the casting.
turing process.
Control of the die casting process is difficult because it has
variables exhibiting all four performance characteristics. It
Behavior of Variables also has mutually compensating variables. For example, the
appearance of the casting might get bad because the plunger
The different types of performance characteristics that can cause a
is slow. The operator might correct the appearance by reduc
process to go out of control are:
ing the flow of cooling water and therefore making the
1. Catastrophic event die hotter. Conceptualization is also difficult, because the
2. Continuous drift process can appear quite different from different view points.
The customer who forces his supplier of die castings to adopt
3. Cyclic drift some specific technique (such as statistical process control)
4. Threshold condition. because of a perceived situation could very well be doing
the wrong thing. The process theory should be understood
before corrective action is taken.
The breaking of a drill is an example of a catastrophic event.
All parts prior to breakage are fine, but the parts after break
age have no hole. The process is no longer operating within
Control Methods
the predetermined range. The process is “out of control.” If
the cost of having parts without the hole is low, periodic There are several methods for controlling the input variables to a
inspection is adequate measurement of the actual processing process. The most common are:
condition. Once the process is discovered to be out of control
• Set and Check
(i.e. no holes), the condition is corrected (i.e. drill replaced),
and all parts made since the previous inspection are repro • Continuous Display – Operator Monitor
cessed. If the absence of the hole is very expensive, then every
part must be checked to insure it has a hole before it leaves • Automatic Control to Set Point
the operation. Small core pins in a die cavity can break caus
ing catastrophic event type conditions. • SERVO Control
6 Die Casting Process, Engineering and Control

• Adaptive Control 3. Injection Velocity
• Statistical Process Control (SPC) 4. Tie Bar Loading
5. Die Temperature
The best example of a set and check control would be the
setting of the hydraulic pressure of the die casting machine’s 6. Release Material
hydraulic system. It is supposed to be XXX psi, and the
control valve is adjusted until that pressure registers on the
7. Casting ejection temperature
pressure gauge. It is now “set.” The gauge is “checked” on
some routine schedule such as the start of each production
shift to be sure that the pressure is what it should be. Oth 8. Cycle timing
erwise there is no attention given to the hydraulic pressure
unless some other problem indicates that there might be a 9. Die wear
pressure fault.
The die is not generally thought of as a process variable, but
Sometimes the tie bar tonnage is displayed continuously where since it can wear (and hence change dimensions on the part)
the operator (or technician) can see and monitor it. The interpre or break, it is included here as a set of process variables. The
tation and reaction to the values displayed are at the discretion alloy content is not addressed in this book. The reader is
of the observer. A thermometer outside the kitchen window referred to the NADCA textbooks, Metallurgy,4 and Molten
showing the outside temperature is a continuous display system. Metal Systems31 for information on alloy control.
However, it is not connected to any “control” function.
The automatic control to a set point is like the temperature con Book Format
trol of a typical home. The thermostat is set to the desired tem
perature. The thermostat continuously monitors the temperature This book is presented in three parts:
of the room. When the room is colder than the set point, the
thermostat automatically starts the furnace. When the tempera I. Statistical Methods
ture of the room reaches the set point, the furnace is turned off.
In the die casting process metal and die temperatures are/can be II. The Die Casting Variables
controlled with automatic control to a set point.
III. Perspective
A SERVO (or proportioning) system is a continuously
variable control system. Unlike the automatic control to a The first part, Statistical Methods, has one chapter describing
set point which simply turns the energy effecter on or off, how to measure and quantify the actual performance of any
the SERVO system measures how far the actual condi single process variable, and one chapter describing the general
tion is from the set point and adjusts the input energy approach to statistical process control. Both chapters are sim
accordingly. The “cruise control” for an automobile is such plifications of the respective subjects, but show the die caster
a system. The accelerator position is adjusted as some the basic concepts as they apply to die casting. The second part,
function of the difference between the actual speed of the
The Die Casting Variables, has one chapter devoted to each
vehicle and the set speed. major variable. Here, the behavior of the variable is described
and the performance characteristic(s) that causes an out of con
Statistical Process Control (SPC) is a system where per trol condition is (are) defined. Then, the type of control system
formance measurements are taken on a random basis. The that best applies to the particular variable is described. Each of
average and range of the measurements are calculated and these chapters is subdivided into the four parts of the control
plotted on a graph. When the plotted values exceed a pre function as listed above. The reader is lead through the step by
determined control value, corrective action is triggered. The step procedures for dealing with each function. The final part,
wear of a cutter in a machining operation can be tracked and Perspective, has a chapter on the mechanics of implementing a
the cutter replaced before out of control parts are made. process control program and a chapter on the theoretical poten
tial of the die casting process.
The Die Casting Variables

There are nine basic variables9 in the die casting process.
These are:
1. Alloy Content
2. Holding Furnace Temperature

CH APTER
Measuring Process Capability
No process variable performs exactly the same all the time. readily useable forms for the die caster and the information
Temperatures drift hotter and cooler. The shot plunger on how the process works and can be controlled (such as this
sometimes moves faster and sometimes slower. This chapter text) are readily available to the die caster. Data on hyd raulic
describes statistical methods of quantifying the performance of pressures, plunger speed, die temperature, and tie bar strain
such process variables. A process variable is literally a process can be readily obtained; and are necessary for meaningful
behavior such as speed, force, temperature, time, or pressure. process capability analysis and for process control. The tech
niques for obtaining those measurements are not discussed
The concept of process capability analysis is to get a math in this chapter, since such discussion is included in the
ematical description of the variable’s behavior. Ideally, chapters of Part II of this book.
the resulting description is simple, easy to understand,
meaningf ul, and readily communicated to others. Statisti
cal analysis provides the necessary techniques to accomplish Collecting Data
these object ives: The performance of a variable is described
by two numbers (i.e. statistics) called the average and the The first step in the study is to collect data. One must have
standard dev iation. The average, denoted by the symbol x, something to analyze, and that something is data. The data
describes the typical or normal performance — literally are numbers and those numbers are measurements of the
the average of all performance measurements. The standard process variable’s performance. The data must be collected
deviation, denoted by the lower case Greek letter sigma (σ), in a way that insures true representation of the variable’s
describes how the variable deviates from the average. behavior. Usually a single measurement does not accurately
represent actual performance, and continuous measurement
It is important to understand that a capability study only records are impractical. A reasonable number of measure
measures actual performance. It does not establish goodness ments is the goal, and that is usually between 50 and 150.
or badness. Nor does the process capability predict potential
for improvement. However, the engineer can use the results Occasionally the behavior of a process variable can be
of the capability study to help make such analyses. The die deduced from characteristics of the product being produced.
caster’s customer may use statistical methods to evaluate the However, the investigator must be able to define explicitly
die caster’s processing capability. His definition of process (mathematically) the relationship between the product’s
capability is significantly different than that used in this characteristic and the processing variable before he attempts
book. He will use measurements of casting quality such as to use such measurements to define the process capability.
the total percent of castings rejected, dimensional varia For example, the size of any particular feature on a die cast
tion, surface finish, or internal soundness whereas the die ing’ is determined by the size of the die, the temperature of
caster would be more likely to measure the process inputs the die when the casting was made, the temperature of the
such as plunger speed, die temperature, etc.. the capabil casting at the instant it was ejected from the die, the alloy
ity computed from the customer’s perspective only provides content and the time-temperature history of the casting
a gross estimate of the die caster’s actual performance. It since it was made. The mathematical relationships between
gives no direction for corrective action. These types of data those process variables and the actual size of the casting’s
can be valuable and mathematic ally legitimate for statisti feature are well documented.1 However, unless all of those
cal quality control (SQC) but NOT for statistical process processi ng conditions are known for a specific casting, it
control (SPC). It is of critical importance that the die caster makes no sense to use the measurement of that casting to
understand the differences between SQC and SPC. There draw conclusions about the fidelity of the die. Fortunately,
is a temptation to believe that since one is using statistical statistical methods can often be used to sort out the essential
methods that the process is in control. Such is not necessar relationships. And, in fact, change of the actual dimen
ily the situation. sions of the die is one type of processing variable that can be
determined from measurements of the casting providing it is
Traditionally, die caster’s have lacked the means to measure done correctly with statistical methods. The engineer should
the critical process variables. Originally, the equipment was thoroughly understand the concepts presented in this book
not available and later the typical die casting manager did before trying to establish any process capabilities for die
not understand the cost effectiveness or the proper use of the casting. The following hypothetical example describes the
measurement equipment. Now, the equipment is available in techn iques for collecting meaningful data.

Reading Value of 30 3rd 10:20 334
Day Time of Day
Number Reading °F 31 3rd 12:40 302
1 1st 9:50 316 32 3rd 1:50 316
2 1st 11:20 352 33 3rd 2:50 334
3 1st 1:40 358 34 3rd 3:40 343
4 1st 3:30 386 35 3rd 4:20 366
5 1st 4:20 361 36 3rd 6:20 374
6 1st 6:10 374 37 3rd 7:20 365
7 1st 7:30 344 38 3rd 9:50 358
8 1st 8:40 337 39 4th 8:10 351
9 1st 10:30 351 40 4th 8:40 354
10 2nd 8:10 348 41 4th 11:10 344
11 2nd 9:30 326 42 4th 12:50 324
12 2nd 11:00 301 43 4th 2:00 318
13 2nd 11:40 329 44 4th 4:10 330
14 2nd 12:10 346 45 4th 5:10 336
15 2nd 1:40 364 46 4th 6:00 366
16 2nd 2:50 399 47 4th 7:20 381
17 2nd 3:20 384 48 4th 7:30 374
18 2nd 4:10 371 49 4th 9:30 387
19 2nd 5:40 367 50 4th 12:00 374
20 2nd 6:00 364 51 5th 10:00 354
21 2nd 6:40 351 52 5th 11:20 356
22 2nd 9:00 376 53 5th 1:20 358
23 2nd 9:30 362 54 5th 3:30 346
24 2nd 9:50 344 55 5th 3:40 336
25 2nd 10:30 345 56 5th 3:50 324
26 2nd 10:50 348 57 5th 6:00 316
27 3rd 8:30 336 58 5th 8:30 326
28 3rd 8:40 316 59 5th 8:50 344
29 3rd 9:20 328 60 5th 11:00 355
Table 2-1. Temperature data from thermocouple in die casting die. (A hypothetical example)

In this example, it has been assumed that a thermocouple 405.5 X
was installed midway between a waterline and the cavity
surface of a die casting die. Temperature readings were 395.5 X
Temperature Measurements, °F
made from the thermocouple and recorded as shown in 385.5 XX
Table 2-1. The readings were made at random intervals 375.5 XXX
during one week of continuous two shift operation. All
365.5 XXXXXXXX
readings were taken during normal operation. The machine
was assumed to be operating on eight hour shifts with no 355.5 XXXXXXXXX
break between shifts. Operator break and lunch periods 345.5 XXXXXXXXXXX
were relieved so there was no break in the operation once 335.5 XXXXXXXXXX
the machine was started in the morning. No measurements
were taken during the first hour each day. Temperatures 325.5 XXXXXXXX
were known to vary widely during that first hour because of 315.5 XXXXX
die “warm up” and few saleable castings were made during 305.5 XX
that time. The time between measurements was selected
295.5 X
randomly in ten minute increments from a minimum of ten
minutes to a maximum of 150 minutes. It is important that Fig. 2-1. A histogram (i.e. frequency distribution) developed by plotting the
measurements for statistical analysis be taken on a random data from Table 2-1. Each temperature value in the table is plotted here as an
basis. Random time intervals eliminate the chance of taking X. These plots show graphically how the variable behaves.
all measurements at the same point of a processing cycle.
And, it is important that they only represent normal operat Next, these extreme values are subtracted to find the total
ing conditions. Die warm up periods and times when some range of all measured values.
thing is not working right (such as when the waterline was
plugged so the die was being run slowly with excess die lube Largest Value 399°F
to “finish off the run”) must not be used to obtain process Smallest Value - 301°F
capability data. (Data from such abnormal conditions can be Range of Values = 98°F
used to establish the degree to which such operating condi
tions hurt the process. But such data is only useful once the
The third step is to establish some reasonable number of groups
capability of normal conditions is quantified. Such studies
(sometimes called cells) into which the data can be organized
deal with the risks associated with uncontrolled processes,
for plotting. Usually ten, fifteen, or twenty cells (groups) are
and that is beyond the scope of this book.)
used. The ideal number of cells depends on the number of mea
surements and on the shape of the frequency distribution. The
general rule is to use:
10 Cells for 75 or fewer measurements

15 Cells for 75 to 150 measurements
Graphical Representations 20 Cells for over 150 measurements
Data as collected (i.e. Table 2-1) is difficult to understand
and to grasp significant meanings from. Even though the The shape of the frequency distribution can also influence the
stat istics of average and standard deviation can be computed choice of the number of cells. If the first plot of the frequency
directly from the raw data, it is usually helpful to first depict distribution looks like a different number of cells would display
the data pictorially. The frequency distribution and time the information better, then it can be re-plotted with a differ
plot are two types of graphs that are particularly helpful for ent grouping. The size of each cell is determined by dividing
understanding the nature of the variable being analyzed. the total range of values by the number of cells desired, and
then adjusting the size larger (if necessary) to achieve a number
that is easy to work with. For the continuing example:
98F°/10 cells = 9.8°F/cell

Frequency Distribution Increase the cell size to 10.0°F for convenience.
The frequency distribution of the data from Table 2-1 is
shown in Fig. 2-1. It shows how the data clusters about the The fourth step is to arrange the cells on a scale that will
average, but that a few readings were significantly higher facilitate plotting the data. The center of that scale should be
or lower than the average. The first step in constructing approximately midway between the extreme data values. For
the frequency distribution is to search out the highest and the ongoing example, that point would be:
lowest values in the data. Readi ng number 16 is the highest
at 399°F and reading number 12 is the lowest at 301°F. 301°F +(98°F/2) = 350°F

The mid-point can be moved as much as half of one cell (i.e. 5 However, most process variables can be treated like normal dis
degrees) when that will place it on a more convenient number. tributions if the frequency distribution is somewhat bell shaped.
Such an adjustment is not required for the ongoing example.
However, the mid-point of the scale should not be a number The frequency distribution will not always be bell shaped.
that could result in a data point falling on a cell boundary. Other shapes that might be found are shown in Fig. 2-2. The
Since all data measurements in Table 2-1 are integer degrees, rectangular distribution (a) is not very common. But, some
the shifting of the scale mid-point from 350°F to 350.5°F will variables might approach it when rigid full time controls are
insure that no such match could ever occur. If there are an used. Two separate distributions (b) might be found on plunger
odd number of cells, the mid-point of the scale will be in the speed. Occasional sticking or valve malfunction could cause
middle of the center cell. The scale is then constructed about the second performance pattern. The double hump bell (c) is
that mid-point as for the frequency distribution in Fig. 2-1. usually a sign that one has actually measured two different pat
terns that are close. For example, the first shift operator might
The final step is to tally the data on the scale. For each mea run the die somewhat hotter than the second shift operator.
surement, an “X” is placed adjacent to the scale between The half bell (d) and the tail only (e) indicate that nothing hap
the numbers that bracket the value of the measurement. For pens until some threshold value is reached. Die blow is likely
example, the first reading (Table 2-1) is 316°F, so for it, and “X” to have such behavior. When these types of distributions are
is placed between 315.5 and 325.5 on the scale. This is repeated observed, the investigator should search out and understand
for each measured value to create the frequency distribution the cause of the threshold. It is likely to be something that is
shown in Fig. 2-1. easily (and even frequently) altered. (In the case of die blow, it
is readily changed with tie bar nut adjustment.) If it is some
The shape of the frequency distribution tells the investigator a lot thing that is likely to be changed, the data do not accurately
about the process. It also helps him determine how to approach represent the process capability. The tail only (e) and skewed
his analysis and any process control action. The outline of the (f) distributions can not be properly analyzed with techniques
frequency distribution in Fig. 2-1 is shaped something like a bell. described in this book, although it is sometimes possible to
Whenever the outline of the distribution has this “bell” shape, approximate the tail only distribution with a triangle which is
it is probably close to what is known as a “normal distribution.” described in a subsequent section of this chapter.
The true normal distribution has a strict mathematical definition.
Fig. 2-2. Some process variables may not

exhibit the bell shaped freq uency distri
bution. Some may have a, (a) rectangular,
(b) two sepa rate, (c) double hump bell,
(d) half bell, (e) tail only, or (f ) skewed
dist ributions.

Fig. 2-3. The data from Table 2-1 are shown here plotted against time. Time graphs such as this are helpful for spotting behavior patterns such as continuous
drift, cyclic drift (illustrated here), and the rates at which the processing condition can change. If the time graph shows the variable to be totally unpredictable
statistical methods may not be applicable for process control.
Time Graph Computing the Statistics

Another useful presentation of the data is a time graph (time plot). The first statistic to be computed is the average which is
The time plot can show patterns and trends that are useful in denoted by “X”. The average is computed by summing all the
establishing the validity of the data and in determining the causes measured values (x) and dividing by the total number (N) of
of variation. The major causes of the variation must be established measurements.
before controls can be specified and designed to reduce the varia
tion. The time graph is very simple. The value of each measure
ment is plotted against time. The data in Table 2-1 is shown on a Total of all values
time graph in Fig. 2-3. It is sometimes helpful to connect the data x = Number of measurements = σx1/N
points with lines as in Fig. 2-3 or to fit a curve through them.
From Fig. 2-3 it can be seen that the process variable depicted The average of temperature measurements listed in Table 2-1
has cyclic drift performance. Peak temperatures occur at is 349°F.
approximately 15 to 20 hour intervals. It also shows that the
temperature can change very rapidly. Such rapid changes are The average, X, shows the expected value. The process
found between 8 and 9 o’clock of the third day, 5 and 6 o’clock variable will usually be close to the average. The average
of the fourth day, and 3 and 4 o’clock of the fifth day. performance can usually be changed by some deliberate
adjustment to the process. The average is also the value that
Because of these rapid changes, the engineer may want to get would be specified by the processing engineer. For example,
four hours of continuous recordings on a chart recorder to better the processing engineer might specify the plunger velocity
understand this variable. Even large temperature excursions for a specific die to be set at 86 in. per sec. The 86 in. per
can happen quickly. The most extreme temperatures recorded sec. is the desired average. The implementation of a process
were within four hours on the second day. There is no evidence, controller allows such parameters to be set (i.e. by the num
however, of any significant difference between the first and bers) on the mac hine as specified by the processing engineer.
second shifts. The time plot also shows that even though the In the ongoing example of die temperature, the average
peak to peak period is long, the temperature drift can be reversed would be increased if the die were to be run faster, or if the
quickly. The ability to have sharp reversals of the performance flow of water through the cooling line was to be shut off.
means that the variable will probably respond well to the cor
rect control technique. But, the random nature of the naturally When the histogram is reasonably bell shaped, the variat ion
occurring reversals indicates that statistical process control (SPC) of performance about the average is described by the second
techniques might have quite limited effectiveness. A smoother statistic, the standard deviation denoted by σ. The first step
profile would improve confidence in SPC. in computing σ is to subtract the average, X, from each mea
surement, X i, as shown in the second column in Table 2-2.
The frequency distribution and time plot help the investigator Then each of these differences is squared, (X i - X). 2 These
understand the behavior of the processing variable. They also squared differences are totaled, which for the example is:
help determine if the data accurately represents the behavior of
the process. And, by studying these graphs, the engineer may
get new ideas on how the process can best be controlled. σ( X i - X)2 = 28,209

That total is then divided by the number of observations to A truly normal distribution will have 99.73% of all data
obtain σ.2 points within (+/-) three standard deviations of the average.
σ2 = {Σ(Xi – X)2}/N And, a process is considered to be “in control” if it is operat
ing within that plus or minus 3σ range. So the normal operat
ing range of the die temperature in the example would be:
Which for the example is:
σ2 = 28209/60 = 470.15 349°F ± 3(21.8F°) or
from 284°F to 414°F
The standard deviation, σ, is therefore:
σ = (470.15)0.5 = 21.8(approx.)
Measurement Difference Difference 334 -15 225

X; X, — X Squared (X; - X)2 302 -47 2209
316 -33 1089 316 -33 1089
352 3 9 334 -15 225
358 9 81 343 -6 36
386 37 1367 366 17 289
361 12 144 374 25 625
374 25 625 365 16 256
344 -5 25 358 9 81
337 -12 144 351 2 4
351 2 4 354 5 25
348 -1 1 344 -5 25
326 -23 529 324 -25 625
301 -48 2304 318 -31 961
329 -20 400 330 -19 361
346 -3 9 336 -13 169
364 15 225 366 17 289
399 50 2500 381 32 1024
384 35 1225 374 25 625
371 22 484 387 38 1444
367 18 324 374 25 625
364 15 225 354 5 25
351” 2 4 356 7 49
376 27 729 358 9 81
362 13 169 346 -3 9
344 -5 25 336 -13 169
345 -4 16 324 -25 625
348 -1 1 316 -33 1089
336 -13 169 326 -23 529
316 -33 1089 344 -5 25
328 -21 441 355 6 36
Table 2-2. The standard deviation, a, required that the difference between each value and the average be found (x, - x) and these differences be squared
(x; -x)2 and summed.

The shape of a truly normal distribution results in the fol
lowing limits.
Percent of total
Limits
measurements within limits
X ± 0.6745 σ 50.00%
X±σ 68.26%
X ± 2σ 95.46%
X ± 3σ 99.73%
The relative effectiveness of various process controlling

techniques is measured by the standard deviation result
ing from the different methods. For example, assume the
above example was from a die that was manually operated
and cont rolled. The processing engineer wanted to evalu
ate a temp erature controller that used the thermocouple
measurements to regulate the cooling water f low through
the waterline. So, he set up the control system, repeated
the process capability study and computed a new standard
deviation, σ, of 2.2. The ratio of the original to the new
Fig. 2-4. It is sometimes possible to approximate a “tail only” shaped histo
standard deviation shows the effectiveness of the control
gram with a triangle.
ling method. For this example the effectiveness of the
thermocouple control would be:
The above suggestions for describing rectangular and tail only
Effectiveness = 21.7/2.2 = 10(approx.) distributions are useful techniques, but not conventional. The
investigator will find them useful for his own use, but may have
So, the new method is ten times as effective at controlling the difficulty communicating his process performance to others
process input of die temperature as the original method. If the when using those statistics. When the methods described in this
standard deviations were nearly equal, there would be no reason chapter are found to be inadequate, the investigator must use more
to incorporate the new system on his die casting machines. advanced statistical methods. There are several good sources for
those more advanced statistical methods as well as methods for
When the frequency distribution is rectangular as in Fig. getting more information from the data.10,11 There are mathemati
2-2a, the variation about the average is best expressed as the cal techniques for testing the validity of data, for determining
half range (R/2). The half range is one half the difference how many measurements are really necessary, and for establishing
between the largest and smallest measurements. The half whether two sets of data are significantly different. The serious
range can be used the same as 3σ in defining the normal investigator is encouraged to become proficient in those advanced
techniques. However, most die casting engineers who are using
operating range for the variable.
statistical methods for the first time will be looking for improve
X ± R/2 = Normal operating range ments so great that advanced techniques will not be required.
If the frequency distribution has the “tail only” shape shown Summary
in Fig. 2-2e, and the investigator is satisfied that the data The ability to control a process depends upon having a math
is valid, it is sometimes possible to approximate the dis ematical description of the behavior of the processing variables.
tribution with a triangle. Such an approximation is shown The behavior descriptions before and after implementation of a
in Fig. 2-4. The centroid of a right triangle is one third of control technique show the effectiveness of the control. Such a
the length of the base measured along the vertical side. The mathematical description usually consists of two numbers (sta
avera ge value will fall on (or near) the centroid. In these tistics) called the average and the standard deviation. To get the
situat ions, the second statistic (describing the variation data necessary to calculate these statistics, the process variable
about the average) can be the range third (The difference must be measured. Usually 50 to 150 repeated measurements
of the largest measurement minus smallest measurement must be made at random intervals during normal operation of
divided by three). Then the normal performance is expressed the process. Then, to best understand the variables behavior and
as the average minus one range third plus two range thirds. to check out the validity of the data a frequency distribution,
(If the frequency distribution shape is symmetrically oppo and possibly a time plot, should be made of the data. Finally,
site the expression would be the average plus one range third the average and standard deviation are computed. The average
minus two range thirds.) defines the normal or expected behavior and the standard devia
tion defines the variation about the average. The expected range
X +2R/3 or X +R/3
of variation is usually plus three standard deviations and minus
-R/3
-2R/3 three standard deviations from the average.

CH APTER
Statistical Process Control
Statistical methods can be used to insure that a process is program must be trained. It is obvious that those taking
performing within its capability, predict that it is on the verge measurements, doing calculations and plotting charts must
of behavior that is outside its normal capability or that it is be properly trained. But, it is equally important that the
actually operating outside its normal capability. Because of the production workers and managers also know what is being
behavior characteristics of most process variables in die cast done, and why. Statistical process control cannot be accom
ing, statistical process control (SPC) is not usually an effective plished if management thinks it is something that “does not
primary control method. Continuous control instrumentation involve” or “must be hidden from” the machine operators.
is usually required. However, SPC is a very good technique To be successful, the operators must take an active part in
for some variables, and is invaluable for verification that the the process. It is the operating personnel who must react
primary control is working. Hard copies of such verification immediately to any out of control situation identified by
can be supplied to the customer as certification. the control process. It is the operating people for whom the
control charts are created. For anyone else, the control chart
The first step in any SPC program is a good “process capabil is only interesting. It is also important that the management
ity” study as described in Chapter 2. It makes no sense to try and marketing personnel understand the program. As stated
to keep something operating within its normal operating range in the previous paragraph, this chapter concentrates on the
if nothing is known about its performance behavior. Tradi technical aspects of design and execution of an SPC pro
tionally, die casters have had only the most general notion of gram, not the management aspects. However, these imple
the actual performance of the variables. The process capabil mentation considerations are critical to the success of the
ity study used as a basis for the SPC program must show the program and are therefore mentioned here.
capability of the process as it will perform for making the
subject casting. The SPC program will not automatically cause The final step in a successful SPC program is continuing com
better performance. It will only help the die caster insure that munication. It is not enough to collect and analyze measure
the process will not perform worse. Any improvement must ments and plot charts. Even when the charted information
be achieved through some physical change to the equipment, actually results in action that causes consistent manufacture of
operating procedure and/or control instrumentation. Such high quality castings, it is not enough. Frequent communica
changes will change the fundamental behavior of the process tions to management, marketing, engineering, the production
and a new process capability study will be required to measure operators and the customers are necessary to maintain the
the new performance. momentum of the program. Little things like the following
memo are very important:
The second step is to design the SPC program for the spe
cific process variable for the subject casting. The design of the “On Tuesday, July 30, 2006, the SPC chart (attached)
program is the primary subject of this chapter. The program for machine number 6 showed that die temperature
must define clearly the measurement method to be used, the could be going out of control. The jobsetter, Joe Smith,
inspection procedure and the inspection schedule. Then the took immediate action and discovered and immediately
charting technique must be designed and the calculations (with corrected a plugged waterline hose. As a result, no sub-
worksheets) specified. Finally, the program must have clearly quality castings were made. Another example of our SPC
defined decision criteria and specific corrective actions to be program at work,”
taken for specific situations.
Such communications should go to everyone.”
The third step is to implement the program. Implementa
tion is a management, not an engineering or quality con
trol function. People must be assigned tasks and be held Average (X) and Range (R) Charts
responsible for performing them. Also, budgets must be The basic tool for SPC is the average and range chart, com
established. SPC programs are not free. And, if one tries monly referred to as the “X and R charts” (pronounced “X
to “bootleg” it through, other priorities will insure failure. and R”). The X and R chart is a time plot of the averages,
(Sometimes test cases to prove the value of the program can X, and ranges, R, of small groups (i.e. subgroups) of mea
be bootlegged, but that is the extent of it.) Finally, there surements taken at intervals (usually random intervals). An
is training. Everyone involved with, or exposed to, the example chart is shown in Fig. 3-1.

Fig. 3-1. This example average and range (X and R) chart is plotted from the data in Table 3-1. This X and R chart is plotted from small data samples (subgroups), of
five measurements per subgroups. When the plotted line approaches a control limit, the process is probably out of control and corrective action must be taken.
Computing X and R surements and insuring “normal” operation of the process

must be followed as described in Chapter 2. The X and R
The X and R chart in Fig. 3-1 was made from the data in Table chart can be developed as data is collected for the process
3-1. Those data are the same as in Table 2-1, but organized into capability study.
sub-groups of five measurements per sub-group.
Each subgroup in Table 3-1 has its average, X, and range
Four or five measurements per sub-group are usually R, calculated. The range, R, is computed by subtracting the
adequate for most industrial processes including most die smallest reading of the group from the largest. These statis
casting applications. Some processes have been accurately tics (i.e. X and R) are then plotted on the graph as shown
monitored with only two measurements per sub-group, in Fig. 3-1. Once a sufficient number of measurements have
and others have required 15 to 20. If the measurement is been taken with the process operating normally, control
obtained from a destructive test, it is sometimes desir limits can be established and added to the chart. It is usu
able to trade off some accuracy to minimize the number ally desirable to have at least 25 subgroups of measurements
of measurements. The same guidelines for taking mea before control limits are calculated.

Day Time Value X R Computing Control Limits
9:50 316
11:20 352 The first step in computing the control limits for the X
1:40 358 354.6 70.0 portion of the chart is to establish the average, X-bar, of
3:30 386 all the X values. That is done by adding the X values for all
1 4:20 361 the subgroups and dividing by the number of subgroups.
6:10 374 For the data in Table 3-1, X-bar equals 348.33. This value is
7:30 344
nearly the same as the average of the frequency distribution
8:40 337 350.8 37.0
10:30 351 calculated in Chapter 2. A line is drawn horizontally on the
8:10 348 X and R chart at the value of X-bar.
9:30 326
11:00 301 The second step is to compute the average, R, of the range
11:40 329 333.2 63.0 values. For the data in Table 3-1, R equals 40.5. A line at
12:10 346 the value of R is also drawn across the chart. Then the upper
1:40 364
control limit for the range (UCLR) and the lower control
2:50 399
3:20 384 limit for the range (LCLR) are computed by using the fac
2 4:10 371 377.0 35.0 tors from Table 3-2 as follows:
5:40 367
6:00 364 UCLR = D4R = 2.11 x 40.5 = 85.46
6:40 351
9:00 376
LCLR = D3R = 0 x 40.5 = 0
9:30 362 355.6 32.0
9:50 344
10:30 345 These control limits are then added to the R portion of the
10:50 348 chart as dotted or dashed lines.
8:30 336
8:40 316 332.4 32.0
9:20 328 Next, the upper and lower control limits (UCLX and LCLX) are
10:20 334 established for the X portion of the chart by using the A: factor
12:40 302 from Table 3-2 as follows:
1:50 316
3 2:50 334 332.2 64.0 UCLX = X + A2R = 348.33 + (.58x40.5) = 371.82
3:40 343
4:20 366
6:20 374 LCLX = X - A2R = 348.33 - (.58 x40.5) = 324.84
7:20 365
9:50 358 360.4 23.0
8:10 351
8:40 354
11:10 344
12:50 324
2:00 318 330.4 26.0 Using the Control Charts
4:10 330
4 5:10 336 Once the control chart is constructed it becomes a work
6:00 366 ing tool. Measurements are made on a continuing schedule,
7:20 381 averages and ranges computed for subgroups of measure
7:30 374 370.4 43.0 ments and the values plotted. Then, whenever either statistic
9:30 387 (X or R) of a subgroup falls outside the control limits, the
12:00 344
process is probably “out of control”. Some corrective action
10:00 354
11:20 356 (such as unplugging a waterline) and/or defensive action
1:20 358 350.0 22.0 (such as 100 percent inspection of all castings made during
3:30 346 the previous period) should be taken.
3:40 336
5 3:50 324 The fourth average value plotted in Fig. 3-1 is above the
6:00 316 UCLX. Since those data were used to create the chart, there
8:30 326 333.0 39.0 was no UCL, to trigger action when the point was plotted.
8:50 344
11:00 355 Under these circumstances, the out of control data should
be deleted from the calculations (not from the chart) for
Table 3-1. The temperature data from Table 2-1 are organized into subgroups
the control limits and the control limits recomputed and re-
and the average and range are shown for each subgroup. The X-bar and R
plotted. Then, the castings made during that out-of-control
chart in Fig. 3-1 is made from this table.
period should be located and inspected.

Number of Factors for R chart
Measurements in Factor for X chart
subgroup Lower Control Limit Upper Control Limit
n A2 D2 D3 D4
2 1.88 1.13 0 3.27
3 1.02 1.69 0 2.57
4 0.073 2.06 0 2.28
5 0.58 2.33 0 2.11
6 0.48 2.53 0 2.00
7 0.42 2.70 0.08 1.92
8 0.37 2.85 0.14 1.86
9 0.34 2.97 0.18 1.82
10 0.31 3.08 0.22 1.78
11 0.29 3.17 0.26 1.74
12 0.27 3.26 0.28 1.72
13 0.25 3.34 0.31 1.69
14 0.24 3.41 0.33 1.67
15 0.22 3.47 0.35 1.65
16 0.21 3.53 0.36 1.64
17 0.20 3.59 0.38 1.62
18 0.19 3.64 0.39 1.61
19 0.19 3.69 0.40 1.60
20 0.18 3.74 0.41 1.59
All factors in table are based on the normal distribution
Table 3-2. Factors for determining from R the 3 sigma control limits for X and R charts. Published from Eugene L. Grant, Statistical Quality Control courtesy
of McGraw-Hill Book Company, Inc.
Periodically, copies of the X and R charts should be forwarded only the most rudimentary, although highly effective, applica
to the customer. If there are no out-of-control points on the tions of statistical methods to die casting. The serious investi
charts, a cover letter could say something to the effect that gator can glean much more from the data, and he can use his
the charts show how well the process is being maintained and initial data to determine how to collect data more meaningful,
controlled. If there are out of control points on the charts, the if he applies more advanced techniques. The NADCA course
cover letter should explain what steps were taken to insure that on SPC provides more explanation of the theory and shows
no defective castings were shipped. how to use the techniques for a larger variety of situations. Also
most colleges and universities offer courses on the subject in
both credit earning and continuing education curricula. There
Additional Information from Charts are also many good books on the subject. The intent in this
The discussions in both Chapters 2 and 3 are only an introduc book is to show the die caster the basics of statistical methods
tion to statistical methods. These chapters give instructions for as they could apply to die casting.

CH APTER
Holding Furnace Temperature
Except for the alloying constituents (which are beyond the furnace temperature specified must be far enough from the
scope of this book), the temperature of the molten metal “alloy limits” that the ± 3σ expected variation in temperature
in the holding furnace is the first process variable that will will not violate one of those limits. For example, if a particu
influence the making of a die casting. And, it was probably lar holding furnace has an alloy that should not get below
the first variable to have had continuous feed back controls 1150°F (620°C) or above 1350°F (730°C) and the 3σ tempera
applied. Such controls must now be considered traditional ture variation of the furnace is 24°F (13.3°C) (σ = 8°F(4.4°C)),
practice in this application. The questions for today’s die then the holding furnace temperature specified for any cast
caster are not if he needs metal temperature controls; but ing should be between 1175°F (635°C) and 1325°F (720°C).
rather, how should his control system be configured to Between these limits, the temperature selected must be that
achieve the desired results. A second question could be: are which is best for making the subject casting. The NADCA
the existing controls actually doing what one thinks they are courses on Metallurgy of Die Casting Alloys and Molten
doing? And finally, at what temperature should the molten Metal Melting and Handling explain the temperature limits.
metal be held for any specific casting. The NADCA course on Gating shows how to determine the
best temperature for making the casting.
Determining the Required Performance For cold chamber machines, the molten metal can lose between
The casting to be made determines the temperature at which 5 and 20°F (3 to 11°C) during ladling and while it is in the shot
the molten metal should be in the holding furnace. There sleeve. The temperature loss to the ladle and shot sleeve can be
can be some trade off between die temperature, cavity filling estimated from the nomographs14 in Fig. 4-1. The die caster
time and the temperature of the molten metal. Also, some should measure the actual temperature loss that happens in his
alloys will experience excessive oxidation (e.g. dross) or other particular situation. If the heat loss is highly variable, the die
alloy degradation if held at too high a temperature, and some caster may need to change the process to stabilize the heat loss
develop sludge if held at too low a temperature. The holding in the ladle and shot sleeve.
Fig. 4-1. The

metal’s temperature
loss to the ladle can
be pre-deter-mined
from the left hand
nomograph, and
the temperature
loss to the shot
sleeve is pre-
determined from
the right hand
nomograph.

The next consideration is the effect of the metal temperature to add circulation to the melt. Electric induction furnaces
variation (± 24°F (13.9°C) in the above example) on the cast create a circulation, but electric resistance and gas fired
ings to be made. Metal temperature can affect the surface furnaces do not. Sometimes a simple paddle submerged into
finish and density of the casting. The process capability must the melt and rotated by an external motor will suffice. Some
be made available to the die and process engineer(s) so that prefer molten metal pumps.9 When the process capability
it can be taken into account as the die gating system and the study is made, it could show that the location of the per
process set points are being calculated. The calculated die manent thermocouple is at a different temperature than the
temperature and the plunger speeds could be effected by the area where the metal is ladled from. Circulation can solve
metal temperature variation. that problem, but it might also be appropriate to relocate the
permanent thermocouple, or to offset the temperature set
Changes in the holding furnace temperature can also cause point by the amount of the difference.
variations in the castings dimensions’. Short term changes in
the molten metal temperature will have a direct effect on the One problem that could arise is that a furnace may never
casting’s temperature at ejection if all other conditions are achieve the specified (i.e. set point) temperature. This phenom
constant. If the holding furnace temperature increases 10°F enon is fairly common where a single furnace is used for both
(5.6°C), the castings will be about 10°F (5.6°C) hotter when melting and holding and the furnace is overloaded. In these
ejected. The final size of the casting is determined by how instances the furnace is simply too small. It must be replaced
much it shrinks after it is released from the die cavity. If the or the throughput reduced. Sometimes, however, the condition
ejection temperature of consecutive aluminum castings was indicates improper burner adjustment or that covers and other
to change 40°F (22. 2°C), a 10 inch (254 mm) feature on insulation should be added. It could also indicate that molten
the castings would change about 0.005 inch (0.127mm). The metal being delivered to the furnace is too cold.
significance of the expected variation in percent solids or
dimensions must be determined by the processing engineer.
If the variation will be intolerable, modification or replace
ment of the holding furnace will be necessary. Solid metal added
for each for each
Alloy 1000 lb -°F 1000 kg -°C
Measuring and Controlling of temperature reduction of temperature reduction
Actual Performance desired, lbs. desired, kg.
Once the processing engineer has established the required Mg 0.678 0.122
holding furnace temperature and the acceptable variation
Al 0.568 0.102
from that temperature, the actual temperatures must be
measured. A ceramic sheathed thermocouple15 is inserted Cu 0.407 0.073
(permanently) into the melt to make the temperature mea Fe 0.278 0.050
surements. The thermocouple is connected to an indicating
Zn
controller that displays the actual temperature and regulates 0.911 0.164
(3, 5, 7)
the heat source. The furnace manufacturers can usually supply
a furnace and control system to meet any situation provided ZA
0.715 0.129
the situation is clearly specified. (12, 27)
Pb 1.191 0.214
Existing furnaces may not have the same performance
characteristics as new furnaces. Some furnaces, may not be Table 4-1. When metal charged into a holding furnace is hotter than the
properly sized for the job. The reader is referred to Chap desired temperature of the furnace, the extra heat can be compensated by
ter 2 for information on how to measure and quantify the adding the correct amount of solid metal (room temperature).
furnace performance capability. When a capability study is
made, the melt temperature should be measured in several
places within the furnace. These places are: where the metal
is ladled from the furnace, near the permanent thermo It is common for the delivery of molten metal to a holding
couple, in each corner (or in three or four places around the furnace to be hotter or colder than the desired temperature
perimeter if the furnace is round), and these measurements of the holding furnace. When such charging of the furnace
should be repeated near the top, near the center, and near drives the temperature unacceptably low it is not a control
the bottom. problem as much as a capacity problem. But, more frequent
charging with smaller amounts will help as will having
the delivered metal hotter. When the charging drives the
temperature too high, the charging can be done more fre
Additional Control Actions quently, or the temperature of the delivered metal lowered.
When temperature measurements from different regions of Another alternative is to add some bars of alloy or trimmed
the furnace are significantly different, it may be necessary runners and biscuits to the holding furnace each time it is

refilled. Table 4-1 shows how many pounds of room tem Each charging is 150 Lbs. (68 kg), so:
perature metal must be added for each 1000 pound-degrees
of excess heat that is added with each refill of the furnace. Temp. diff. x Charge weight = 50°F x 150 Lbs. (28°C)
The pound-degrees are found by multiplying the pounds of (68kg) = 7500 Lb.-°F (1900 kg-°C)
metal added to the furnace by the difference in temperature
From Table 4-1, one finds that 0.568 Lbs. (0.258 kg) of
between the holding furnace and the metal being added. For
example, suppose a holding furnace of aluminum should be
alloy must be added for each 1,000 Lb.-degs. (252 kg-°C),
at 1225°F (663°C), but it is periodically charged with 150 so:
lbs. (68 kg) of molten alloy at 1275°F (691°C). The tempera {7500Lbs.-°F (1900kg.-°C) x 0.568 Lbs. (0.257 kg.)}/
ture difference is: {1000 Lb.-°F (252kg.-°C)}= 4.26 Lbs. (1.93 kg.) of solid
Charge Temperature 1275°F (691°C) metal to be added
Furnace Temperature 1225°F (663°C)
Such additions can be made as a regular procedure and can
Temp. Difference 50°F ( 28°C) constitute a legitimate temperature control practice.

CH APTER
Plunger Velocity & Force
The shot system of the die casting machine is described in plunger must be advanced slowly until the pour hole is closed
detail in the NADCA course and textbook on Machine to avoid splashing of the molten metal out of the shot sleeve. If
Systems. The calculations for establishing the required perfor the system has only two speed controls the closing of the pour
mance of the shot system for any specific casting is described in hole is accomplished at the slow shot speed. If the system has
the NADCA course and textbook Gating of Die Casting Dies. three speed controls, the flow control valve for the first speed
Methods of controlling the machine to maintain the calculated is adjusted until there is no splashing out of the pour hole and
set point conditions are described in this chapter. The primary then left there. The limit switch that signals the initiation of the
functions of the shot system are: slow shot speed (i.e. the second speed) is set so the slow shot
1. Close pour hole speed starts as soon as the plunger has closed the pour hole.
Machines with advanced (e.g. SERVO controlled) shot system
2. Slow shot velocity control might allow the programming of the pour hole close
velocity through the machine controller. The pore hole plunger
3. Fast shot transition point speed and the pour hole close speed to slow shot speed transi
4. Fast shot velocity tion point are not likely to be specified by the part specific engi
neering so they must be determined by the operating people
5. Low impact position through experimentation.
6. Intensification delay
The hydraulic flow control valves are set by manual adjustment
7. Rate of intensification build-up of the actual valve. The instructions for the setting of those
flow control valves should be on a “set-up” chart at the die cast
8. Intensification pressure ing machine. The instructions might be:
Slow Shot Valve: 3 ½ Turns Open
The values for all of these functions are established during
the design of the gating system for the die and must be made Fast Shot Valve: 2 Full Turns Open
available at the die casting machine for set-up and operation
of the machine. The information should be available at the die Some circuits allow the slow shot to be accelerated by gradually
casting machine in the form of a “set-up” chart (i.e. a “visual”) opening the hydraulic flow control valve. An additional flow
that is laminated in plastic and positioned in clear view of the control valve on the pilot circuit of the slow shot flow control
operator/technician. The control of these functions is described valve must be set to achieve such a acceleration.
below first for the cold chamber machines. The control func
tions for hot chamber machines are nearly identical to those The actual plunger speed would be calculated during the design
for the cold chamber machines. The few variations for the hot of the die gating system. For example the slow shot speed may
chamber machines are described later in this chapter. have been calculated to be 13.8 in./sec and the fast shot speed
164 in./sec. However, if there is not instrumentation to actually
measure the speed of the plunger, the set up person can only
Flow Control Functions guess as to the correctness of the speed. In those situations, the
The first five (5) of the above listed shot functions are a set up and/or other operating person(s) must judge the cor
combination of hydraulic fluid flow control and position con rectness of the settings from the quality of the casting. He/she
trol. For these functions the plunger speed is controlled by a will adjust the valve openings to his/her interpretation of the
flow control valve in the hydraulic system. condition. When satisfied, the set up chart instructions will
be changed to what was determined to make the best castings.
These simple systems with no velocity measurement have a
Simple Systems wide variation (i.e. +/- 3σ) and are subject to being set at the
The simpler hydraulic systems have as many parallel circuits wrong condition (i.e. set point) because there are so many other
as there are control functions available to the operator of the process variables affecting the quality of the casting.
machine. The minimum are two as shown in Figure 5-1, one
for the initial slow shot and the other for the fast shot. How If the shot sequence is manually initiated, the plunger movement
ever there can be more. If there are three, the closing of the is started by the operator pressing a palm button. That signals the
pour hole can have a different speed than the slow shot. The machine control circuit to open the slow shot (or pour hole close)

Fig. 5-1. A simple shot
speed control system has
two parallel hydraulic
circuits. One circuit
for the slow shot speed
and one for the fast shot
speed. The directional
control valves are
opened and closed by
the machine sequencing
circuitry and the flow
control valves are manu
ally set through adjust
ments on the valve.
directional valve. The manual initiation is usually used only when The position of the limit switch can be positioned quite
the molten metal is manually ladled into the shot sleeve. When accurately. A permanent scale can be affixed to the machine
the molten metal is mechanically ladled into the shot sleeve, the for such purposes. What is not so accurate is the accommoda
slow shot (or pour hole close) directional valve would be opened tion of the machine’s response to the signal. For example, the
from a signal indicating that ladling is complete. That signal is person setting the switch may not know exactly where in the
most likely to come from a timer that starts when the ladle starts travel of the switch arm the switch actually sends the signal.
to pour the molten metal into the shot sleeve. Then there is the response delay of the control system. If it is a
relay control system, and several relays have to be sequenced,
The transition from the slow shot velocity to the fast shot and it is an A/C control system, there can be electrical delays.
velocity is accomplished by the opening of the directional valve Finally there is the response of the control valve and the inertia
in the fast shot circuit allowing a greater flow of hydraulic fluid. of the hydraulic fluid. All these things and more cause a delay
That transition (i.e. the opening of the fast shot directional in the actual achievement of the fast shot velocity. If the
valve) is triggered by the plunger tripping a limit switch. The calculated position of plunger for the transition to the fast shot
position of the limit switch is also a manual adjustment. The speed is 28.5 inches as for the above example, the limit switch
instructions on the “set-up” chart could be: may have to be set at 27.75 inches to accommodate the delays
Fast Shot Transition Limit Switch Position: 28.5 in. and achieve the required results. The ideal velocities (either as
Fig. 5-2. The calculated

plunger positions,
velocities, and travel
times can be plotted
graphically to aid visu
alization of what the
injection system must
do. The data shown
here are for a hypo
thetical example.

calculated or as determined by experimentation) can be plotted The continuous monitoring systems are only as effective as
graphically against time as shown in Figure 5-2, or (as is more the frequency with which someone looks at them, and even
useful) against plunger position as illustrated in Figure 5-3. then only as effective as what that person does when an “out
The acceleration of the velocity from the slow to fast shot is also of specification” condition is indicated. Manual adjustment is
illustrated in Figure 5-3. still required to correct the performance. However, the person
making the adjustment can see the actual performance and
Both the velocity and the velocity change positions are “set and adjust the valve and limit switch settings to get the desired
check” systems. Once set, the control of the input variable to velocities at the desired plunger positions. Some of these
the casting process is not adjusted until it is manually checked monitoring systems can be set to cause an “alarm” when the
and manually changed. Because there are so many factors influ velocity is out of some acceptable range. The acceptable range
encing the speed of the plunger, actual plunger velocity can can be set into the monitor. Then, the alarm will alert the
vary quite a bit and the changed performance can go unnoticed operator or technician when an adjustment is required.
as long as the castings “look” acceptable.
When the plunger speed is not what it is supposed to be, there is
The calculated velocity and velocity transition positions are a higher than normal chance that an unusable casting was made.
based on a particular plunger diameter. If a different plunger If a continuous monitoring system alarms such a situation,
diameter is used, these values must be recalculated (ref: the alarm signal can cause the robot that extracts the casting
NADCA course and textbook on Gating) and reset on the from the die to set the casting aside. Then, that casting can be
machine. The effect of changing the plunger size from 2.25 inspected and sent on through the processing system if accept
inches to 3.00 inches in diameter for the preceding hypotheti able, and the casting scrapped out if not found to be acceptable.
cal example is illustrated in Figure 5-4. By having a measure of the actual velocities of the plunger, one
will not risk adjusting the plunger velocity when another vari
able is the root cause of some undesirable casting condition.
Monitoring
These simple control systems can be enhanced with periodic
or continuous monitoring. Monitoring systems are com SERVO Systems
mercially available to continuously monitor the shot plunger For the purposes of this book, all advance shot control systems
performance. These monitoring systems display a graph of the that use feed back control to automatically and continuously
performance of the latest shot and can also and simultaneously adjust the shot plunger velocities are included in the category
display a desired reference graph for quick visual comparison. of SERVO systems. (Note: Some manufacturers may not con-
These monitoring systems also display digital read outs of criti sider their particular system to be an actual “SERVO” system even
cal performance values. An example of the screen display from though their type of system is included under the SERVO heading
one such monitoring system is illustrated in Figure 5-5. Target, here.) These systems continuously monitor the shot velocities
or ideal, performance traces superimposed onto the screen vs. the plunger positions. Any difference (i.e. error) between
display are illustrated by the red curves in Figure 5-6. the desired and actual velocities results in some automatic
Fig. 5-3. The veloc

ity curve from Fig.
5-2 is re-plotted
here as function of
plunger position
and with the accel
erations shown.

Fig. 5-4. The injec
tion system variables
are affected by changes
in plunger diameter.
Shown here are the
calculated performance
requirements for 2.25-
in. and 3.00-in. diameter
plungers to achieve iden
tical cavity fill rates.
Fig. 5-5. The screen

display of a commer
cial shot performance
monitoring system is
illustrated here. Illus
tration courtesy of
Visi-Trak Worldwide.

Fig. 5-6. The screen dis
play from a shot moni
toring system shown
here has the target (i.e.
ideal) performances
shown as red lines. The
acceptable deviation
from the “ideal” can
be input and when the
actual performance is
different than the ideal
by more than the accept
able range an indicating
alarm can be triggered.
Illustration courtesy of
Visi-Trak Worldwide.
Fig. 5-7. The PQ2 dia

gram is used to describe
the power curve of the
die casting machine’s shot
system. Point A is the
maximum velocity of the
plunger with no metal in
the shot sleeve. Point B
is the hydraulic system
pressure. The velocity
scale is constructed to be
the square of the velocity
shown. The line between
points A and B is the
power curve.

adjustment to the hydraulic valve. In most such systems the the gating systems of future dies. These data are the capability
adjustment is immediate, so all castings are made at the cor of the machines metal injection system. These values should
rect plunger velocity. The adjustment to the hydraulic flow be re-measured periodically (e.g. every month) to check for
control valve is made electrically rather than manually. The degradation of the machine. If the performance of the machine
control system for SERVO systems should be programmed reduces, it should be repaired.
so that when the control signal to the valve is maxed out,
and the plunger is not achieving the desired speed, there
will be some type of alarm. Pressure Control Functions
The last three (3) of the shot plunger control functions (i.e. func
Impact Control tions 6, 7 and 8) are pressure and time related functions. Once
the cavity is full, the hydraulic pressure will reach the system
When the cavity fills with metal, the plunger must stop
pressure almost instantaneously. As the casting solidifies there
moving. There is no place for the molten metal to go. Imme
is a volumetric shrinkage which results in porosity within the
diately before that instant of the “cavity full” condition, the
casting. High hydraulic pressure on the shot cylinder piston
plunger is traveling at it’s fastest speed. The result of the sudden
applies high force on the plunger, and hence on the biscuit. That
stopping of the plunger is that a hydraulic pressure spike is
high force crushes the solidifying shell of the biscuit reducing
created in the molten metal. That hydraulic pressure spike can
its volume. The crushing of the shell of the biscuit puts pressure
cause the die to spread open and flash. Some machines are
on the molten metal core within the solidifying biscuit, runner
equipped with a control that closes the hydraulic directional
system and casting causing molten (or partially solidified) metal
control valve just before the cavity fills. The result is that the
to move into the shrinkage voids in the casting as they form.
inertial energies of the plunger, hydraulic piston, and hydraulic
There is no evidence to indicate that excessive pressure is del
fluid are absorbed by the trapped hydraulic fluid in the machine eterious to the solidification process. However, excessively high
rather than being transferred to the molten metal. These con pressures will cause dies to spread open a bit and allow flash.
trols must operate very fast and very precisely. These controls Even with high applied pressures, actual internal cavity pressure
are initiated by the position of the plunger. If the hydraulic will drop with time as the casting solidifies. The standard meth
fluid is trapped to soon, the casting will not fill properly. If it is ods of achieving the desired cavity pressures are by means of
trapped to late, the impact pressure and the associated flashing the hydraulic system pressure, ratio of hydraulic piston to metal
of the die will already have happened. Since the control is posi plunger areas, and intensification of the hydraulic pressure.
tion controlled, the amount of molten metal ladled will effect
its usefulness. If too much metal is ladled, the control will
operate too late. If too little is ladled, the control will operate Intensification delay
too soon. So, for such controls to be effective, the ladle quantity The first pressure control function is the delay of the applica
must be controlled as well. tion of the intensified hydraulic pressure. If the pressure on the
molten core of the solidifying casting is increased too soon, it
Fast Shot Process Capability will force the dies open and flash will result. So, it is usually
desirable to allow some solidification to take place before the
Considerable power is required to accelerate the molten metal
high pressure is applied. A delay in the application of the inten
up to the gate velocity. The power required is a function of sified pressure is normally accomplished with a timer. Not all
both the gate velocity and the volume of molten metal that machine controls are equipped with intensification delay capa
must pass through the gate. At high plunger velocities a lot of bility. Also, the process engineering functions do not usually
the available power is used by (lost in) the machine and not calculate what the delay should be. The most common method
available to push on the molten metal. These power losses are of determining the delay time is to experiment at the machine.
pressure drops of the hydraulic fluid passing through valves, Normally the delay is set at zero and then fractions of seconds
around corners in the piping and friction. The machine’s power of delay are added until the die stops flashing. Once the best
performance curve is described by the PQ2 diagram as illus delay time is established, it should be recorded on the “set-up”
trated in Figure 5-7. The straight line connecting points A and chart and the machine control re-set to that delay each time that
B in the figure is the power curve for the machine. Point B is particular die is set in the machine. The example in Figure 5-5
the hydraulic system pressure at the end of the shot plunger above shows no delay. The intensification pressure is shown to
travel. This point B pressure might be less that the pressure start building immediately after the cavity has filled.
setting of the system since some pressure drop is experienced as
the accumulator is discharged to make the shot. Point A is the
maximum plunger speed that can be obtained when all the flow Rate of Intensification Build-Up
control valves are fully open, the system pressure is at its cor Once the cavity has filled and plunger motion stops, the
rect operating pressure, and there is no metal being pushed by intensification pressure will build as rapidly as the dynam
the plunger. Point A is readily measurable if one has a monitor ics and control settings of the mechanism will allow. Some
ing system as described and illustrated above. machine control systems can achieve nearly instant application
of full intensification pressure. Pressure build-up times can be
The operating people must measure the power performance adjusted on some machines. Unfortunately, there is no known
(i.e. the values of points A and B) of their machine(s) and make formula for calculating the optimum delay time, and optimum
sure the information is forwarded to the people engineering settings are usually established by trial and error. The example

Casting
Core Cavity
Ejector
Die Cover
Die Gooseneck
Nozzle
Plunger
Ejector
Pins
Inlet
Start
Inlet Closed
Gooseneck
& Nozzle Full
Sprue and
Molten Runners Full
Metal Die Full
Fig. 5-8. In a hot chamber machine the stroke of the plunger must pass the inlet ports and then displace sufficient metal to fill the gooseneck and nozzle before
metal enters the die.
shown in Figure 5-5 shows the pressure building to the maxi Slow Shot Velocity
mum value in about 500 milliseconds. As with the delay, the The slow shot velocity for the cold chamber machine is estab
rate of pressure build up should usually be as fast as possible lished to control the wave formation in the shot sleeve. There
without flashing the die. Once the best intensification build-up is no such consideration in a hot chamber machine. However,
rate is established, it should be recorded on the “set-up” chart the calculations must account for the plunger displacement
and the machine control re-set to that value each time that required to fill the gooseneck (above the holding furnace metal
particular die is set in the machine. level) and the nozzle. Little or no displacement of molten metal
can occur before the plunger closes the gooseneck’s inlet ports
Intensification Pressure so the first part of the stroke (usually one or two inches) is
non-effective. Venting of air out of the die is probably improved
Some systems achieve intensified hydraulic pressures that
if a slow advance is used to fill the gooseneck and nozzle. In
are four times system pressure. System pressure and/or
other situations, a slow advance through nozzle filling is used
intensified pressure are adjustable on many machines. Inten
in conjunction with an extra hot (900 to 950 deg. F for zinc)
sified hydraulic pressures are usually achieved by one of two
nozzle to superheat the first metal to flow through the die. If
methods. The first method is through the use of differential
one uses the slow shot to preheat the first metal, there must be
piston hydraulic intensifiers. Machines equipped with these
very good temperature control of the nozzle. Some die casters
devices can only achieve maximum intensified pressures that
find a fast shot through the total stroke to work best.
are a fixed multiple (e.g. 2:1, 2.5:1 3:1, etc.) of the normal
system pressure. So, to change the actual intensified pres
sure one must change the system pressure. Changing the
system pressure can have ramifications on the performance
of other shot plunger performances. The second method is
to have a high pressure hydraulic system with its own pump
and accumulator to provide the higher pressure hydrau
lic fluid. These systems can be set to a range of pressures.
However, resetting the intensification pressure for every
die change may not be practical. The reader is referenced to
the NADCA course and text on Machine Systems to get
descriptions of the different intensification systems and how
they work. The most common practice is to set the intensi
fication pressure at one value and then adjust the delay and
the pressure build-up rate to achieve the desired results.
Hot Chamber
The hot chamber injection is shown schematically in Fig. 5-8.
Most of the calculations, measurement of variables and control
decisions are the same as those for the cold chamber process.
However, there are some noteworthy differences. The hot Fig. 5-9. As the holding furnace metal level drops, the empty space in the
chamber machines do not have intensification capability. So, gooseneck passage increases. The increased space requires extra plunger
there are no considerations similar to those for the intensifica stroke. As a result the calculated plunger positions for nozzle full, runners
tion on cold chamber machines. full and die full are changed.

The plunger should be moving at the fast shot speed when the
molten metal reaches the cold sprue. The slow to fast shot tran
sition point is usually set with a limit switch triggered by the
tail rod of the shot hydraulic cylinder. When setting that limit
switch, care must be taken to account for the system reaction
time (as discussed above for the cold chamber machines), varia
tion in holding furnace metal level and variations in gooseneck
and nozzle volumes.
Furnace Level
One feature of the hot chamber process is that the holding
furnace level affects the point on the plunger stroke at which
the cavity becomes full. Referring to Figs. 5-8 and 5-9, the Fig. 5-10. When air is trapped under the gooseneck plunger, the chart record
length of the gooseneck channel above the furnace metal level ing of shot cylinder pressure will show long-term pulsations.
(which is empty of molten metal) increases as the level in the
holding furnace drops. In one example, a change of three (3)
inches in the level of the molten metal in the holding furnace An alternative method is to open the die as the plunger is
changed the plunger position at which the cavity filled by 3/8 returning and before the inlet ports are opened. This is the
of an inch. The position of the plunger at which the molten “open die first” method. With this “open die first” method
metal reaches the sprue is changed by an equal amount. The atmospheric pressure entering the sprue from the die part
slow to fast shot transition limit switch setting must insure ing surface forces the metal down the gooseneck, making
that the fast shot speed is achieved before the molten metal it impossible to dribble. A limit switch actuated by the tail
rod of the shot hydraulic cylinder triggers the opening of
reaches the sprue for all furnace levels.
the die. Opening the die before the gooseneck inlets are
open eliminates sprue dribble but has other disadvantages.
Nozzle One disadvantage is increased filling time for the goose
If a nozzle is replaced, the replacement nozzle could have a neck. There is no vacuum to make the air pressure push the
different length and/or inside diameter than the one being molten metal into the gooseneck. Only the head pressure of
replaced. The change in interior volume of the nozzle will the molten metal in the holding furnace pushes the metal
change the position of the plunger when the metal reaches the into the gooseneck. If the gooseneck is not filled in time,
sprue and when the cavity is filled. The position of the slow to the production rate must be slowed to accommodate the re-
fast shot transition limit switch setting must be checked and filling of the gooseneck.
adjusted if necessary when a nozzle is changed.
A second disadvantage is the tendency for getting air
trapped in the gooseneck under the plunger. Air is drawn
Plunger Return and Die Opening into the gooseneck cylinder when the volume of the cast
One method to sequence the return of the shot plunger and the shot exceeds the total volume of the nozzle hole plus goose
opening of the die is to return the plunger and then open the neck outlet. When air is trapped under the plunger, the
die. This is the “return plunger first” method. As the plunger hydraulic pressure will fluctuate wildly in the shot cylinder
returns a vacuum forms in the nozzle and beneath the plunger as shown in Fig. 5-10. When air bubbles up through the
causing the nozzle to fill with metal as soon as the gooseneck gooseneck it creates dross in both the gooseneck cylinder
inlet ports are opened. Air pressure forces the metal into the and on the holding furnace. The dross in the gooseneck cyl
nozzle to the solidified sprue in the die. The existence of the inder can be driven into the next casting, and the dross on
vacuum causes the full effect of the air pressure to force the the holding furnace must be skimmed off. The extra dross
molten metal into the gooseneck. The result is that the re-filling also represents a cost in greater melt losses. Although open
of the gooseneck is very fast. When the die opens, molten metal ing the inlet ports before the die opens solves the entrapped
(being at the sprue) is likely to dribble down the face of the die. air problem, a better solution is to use a gooseneck with
The metal dribble runs down the parting face of the die. When sufficient volume in the gooseneck outlet passage. It might
the dribble freezes onto the die it must be scraped off, slow be necessary to build a pouch like enlargement in the goose
ing the operation. These frozen metal dribbles are particularly neck outlet chamber below the metal level to achieve the
troublesome for automatic operation. sufficient volume.

CH APTER
Clamping Force
The purpose of the clamping system of a die casting machine is trates all of the control concepts and it provides for the most
to hold the die halves together as the molten metal is injected and possible adjustment. Except for the additional opportunity for
to open and close the die for removal of the casting. This chapter adjustment of the individual tie bars on the tie bar and platen
addresses the holding together of the die halves. type, the two types yield to the same analyses.
There are two types of clamping frame structures used on die The clamp system of a die casting machine is shown in Fig.
casting machines. These are the tie bars with platens type and 6-1 and consists of the platens, tie bars, toggles, and die close
the solid frame type. The tie bar and platen type is used as the cylinder with its attendant hydraulic equipment. The perfor
example in this chapter because it is the most common, illus mance of the clamp system can be analyzed by measuring tie
bar strain, rod end hydraulic pressure in the die close cylinder,
head end hydraulic pressure in the die close cylinder, and the
time intervals between events in the cycle. Casting thickness
variation is the most direct result of improper performance of
the clamp system. Variations in casting thickness obviously
have dimensional effects and they also influence direct material
usage and heat input to the die. The dimensional relationships
are developed thoroughly in the NADCA textbook Dimen-
sional Repeatability of the Die Casting Process. Another, and
rather obvious, result of improper performance is when the
machine will not “lock up”. In that instance the die will close,
but the toggles never straighten so the machine just stalls and
does not continue the process sequence. Measurements of the
critical variables of the clamp system can also contain clues to
problems in the other machine systems.
Computing Required Conditions

The critical control variables of the clamping system are the tie
bar strains (i.e. elongations under load). The required or ideal
strains are established by first determining the loads that must
be carried and then computing the strain that will result. These
loads are calculated in tons of machine clamping force. This
clamping tonnage calculation is one of the first calculations
the die casting engineer makes when considering a new casting
to be made. The hydraulic system of the machine provides the
energy to strain the tie bars. The hydraulic pressure in the die
close cylinder, the diameter of the cylinder and the mechanical
advantage of the toggles create the force to stretch the tie bars.
Computing Tie Bar Load

As the toggles straighten, they first close the die, and then, as
they approach the position of highest mechanical advantage,
stretch the tie bars to develop the holding force on the die. The
close-clamp sequence is illustrated in Fig. 6-1.
Fig. 6-1. The force of the die-close hydraulic cylinder is multiplied by the During the original planning and/or die design stage, the
toggles to stretch the tie bars after the die closes. The amount of tie bar elon expected clamp force should be computed. The required die
gation is directly related to the clamping force on the die. clamp force is:

F = PA/(2000 Lb./ton) (6-1) If the machine has not clamped the die together with at least
the calculated force, the force of the injected metal will force the
Where: die halves apart and the metal will flash between them. When
F= Force required to hold die closed against the injection the die “flashes” it makes a thicker than normal casting. If a
process capability study is made on casting thickness of a die
pressure, tons
that is occasionally flashing it will probably result in a “tail only”
A = Projected area of shot, sq. in. frequency distribution as shown in illustration “e” of Fig. 2-2.
P = Maximum internal pressure of molten metal in die, The above calculations account for the projected cavity area.
p.s.i. Sometimes there are other considerations that dictate addi
tional machine tonnage. One such consideration is the force
The value of P is the most difficult to pre-determine because it necessary to “lock” moving cores in the die. The amount of core
is highly dependent on the machine’s injection system’s dynam locking force required should be calculated as part of the die
ics, injection plunger velocity, intensification ratio, and accu design, 24 and added to the tonnage calculated above.
mulator pressure drop. These variables, and their control, are
discussed in more detail in chapter 5 on the Injection System. Another consideration is the need to flatten the die. Die cast
ing dies warp due to the thermal gradients within them. 24
If the projected area A is not centered between the tie bars, Although the design of the die should accommodate the
the loading required of each tie bar must be computed by the warping, it sometimes does not. In those instances it is some
sum-of-the-moments. The illustration in Fig. 6-2 shows a times possible to clamp the die tightly enough to flatten the
hypothetical off-center situation. An off-center cavity situa die and seal the parting line against flashing. The force needed
tion is simulated in the figure by the three rectangles identified to flatten the die may or may not exceed the force required by
by the circled numbers 1, 2 and 3. The hypothetical cavity is the injection pressure and projected area. The greater of these
positioned in a machine with 6-in. diameter tie bars on 35-in. forces determines the locking force required of the machine. If
centers according to the dimensions shown. The clamping force possible, the locking force of the machine should not be used to
of the tie bar that must exert the greatest force is computed flatten the die. If the die has been so flattened, it springs back
using the sum-of-the-moments. Then the machine is selected, to its unstressed shape as it opens. The solidified casting occu
and subsequently adjusted, so all four tie bars exert that maxi pies completely the space within the cavity, and the changing
mum force. The method is developed in detail in the NADCA shape can pinch the casting making ejection difficult.
course Engineering Die Casting Dies and the course notes for
that course. The point to be made here is that when the cavity Another factor that can make increased machine tonnage
is not centered on the machine platens, a larger machine must required is the fluctuating thickness of the die due to variations
be used than the above equation 6-1 would indicate. in die temperature1. A typical die will vary in temperature +/-
Fig. 6-2. The area rep

resented by rectangles
1, 2, and 3, indicate
an off-center cavity
in respect to tie bars
A, B, C, and D. The
solution of this sim
plified example shows
the off-center condi
tion will increase the
required machine
tonnage by more
than 60 percent to
600 tons. The biscuit,
runners and overflows
have been omitted to
simplify the example.

50°F (27.8°C) during a shift. The die thickness will experience
a corresponding variation according to equation 6-2.
ΔT = (T)(ΔF)(0.0000063in./in.-°F) (6-2)
= (T)(ΔC)(0.0000113 mm/mm -°C)
Where:
ΔT = The change in die thickness, in. (mm)
T = The nominal die thickness, in. (mm)
ΔF = The change in die temperature, °F
ΔC = The change in die temperature, °C
The die thickness, T, in equation 6-2 is the thickness of the

holder blocks. The mounting plates and ejector box rails are not
included. If a die is 10.00 in. (254 mm) thick and experiences a
total temperature fluctuation of 100°F (55.6°C), its size change
is calculated with Eq. (6-2) as follows:
ΔT = (10.00in)(100°F)(0.0000063 in/in-°F)
= (254mm)(55.6°C)(0.0000113mm/mm - °C)
Fig. 6-3. Each die casting machine should have a chart, such as that here
= 0.0063 in. = 0.160mm shown, giving the relationship between clamping force and tie bar stretch unless
the machine is equipped with instrumentation to read clamping force directly.
The die casting machine must be adjusted to clamp with at least
the calculated force when the die is at its coldest (and there size of the die centrally to the platen. The reacting force is carried
fore its smallest) size. Then when its temperature fluctuates to by the tie bars which are out at the corners. The result is that the
the hottest and the die gets thicker, the tie bars must stretch platen must bow. The platens are designed to bow a minimum
an additional 0.0063 in. (0.160 mm). The additional tie bar and tolerable amount with a reasonably sized die. If the die is too
stretching must not exceed the capability of the machine. The small, the platen will bow more than intended and can damage
relationship between strain, stress and tonnage is developed in the machine or die. The minimum die size to be used in a die cast
a following section; but if the machine is the one with the force ing machine should be specified by the machine builder although
strain relationship shown in Fig. 6-3, it will need an additional it is usually two thirds of the area between the tie bar centers, and
100 tons (890 kN) to clamp the thicker die. That 100 tons (890 smaller dies should not be run in the machine. These die size to
kN) must be added to all the other tonnage requirements. machine relationships are developed more fully in the NADCA
course on Engineering Die Casting Dies.
Reducing Tonnage Requirements
The internal separating force in the die casting die is calcu Computing Tie Bar Strain
lated with Eq. (6-1) above. That equation shows the force to Since the “locking” force (pre-load) is created by stretch
be proportional to the internal pressure of the molten metal, P. ing the tie bars. Control of the clamp system depends on
That pressure is inversely proportional to the square of the shot measurement of the tie bar stretching. The most common
plunger diameter. Therefore, increasing the size of the shot sleeve techniques measure the tie bar strain (i.e. stretching of a unit
(or gooseneck bore if hot chamber) can reduce significantly the length) which must be either compared to the required strain
pressure of the molten metal. Increased plunger diameters also or converted to the equivalent stress or load. Therefore, it
reduce the necessary plunger velocity, which in turn reduces becomes necessary to calculate the required unit elongation1
the impact pressure. These relationships are developed more in (i.e. in. per. in.). The unit elongation is a function of total
Chapter 5 and in the NADCA course and textbook on Gating. load, tie bar geometry, and tie bar material, as expressed by:
The tonnage required of the machine is very dependent upon
plunger diameter. If the plunger size is changed for any reason, L = F/AE (6-3)
the machine tonnage calculations should be rechecked. where:
L = Unit elongation, in./in. (mm/mm)
Platen Deflection
The platens of a die casting machine, like the tie bars, are elastic F = Force applied by one tie bar, Lbs (MN)
members, and therefore deflect under load. The moving platen A = Cross section area of tie bar, sq. in. (m2)
may not deflect as much as the stationary platen because the toggle
clamping mechanism is usually placed to apply the clamping force E = Modulus of elasticity of the tie bar material, Lb. /
more directly in line with the die. The stationary platen, however, sq.in. (M Pa) for steel,
is loaded with the clamping force distributed over an area the E = 30,000,000 p.s.i.,(207,000 M Pa)

For the example in Fig. 6-2, the unit elongation for each tie bar and can induce excessive wear in the machine. Some people try
works out to be: to adjust the tie bars unequally to compensate for the condition,
= 0.00034 in./in. = 0.00034 mm/mm but that practice is not recommended.
The total tie bar stretch of a 100-in. (2.54 m) long tie bar would Hydraulic Pressure and Clamping Force
be 0.034 in. (0.86 mm). If a dial indicator is set up to read elon
gation through a 10-in. (254 mm) length it should read 0.0034 Except for the force to accelerate the heavy platen and die,
in. (0.086 mm). the force on the die close cylinder is relatively low until the
die actually closes. Additional movement of the cylinder is
The relationship between tonnage and tie bar strain (elonga then accomplished only by stretching the tie bars. Therefore,
tion) can be plotted graphically for total elongation, unit elon the hydraulic pressure builds rapidly as the tie bar force
gation or actual elongation over the distance of the measuring increases. 23 As the tie bars are stretching the toggles are
device to be used. Fig. 6-3 shows such a chart for total machine increasing their mechanical advantage which, at some point,
tonnages and total tie bar stretch. becomes so great that hydraulic pressure actually drops.
Of course, the pressure immediately builds to the relief
Every die should have a set-up chart on which the die’s tonnage valve setting when the cylinder piston reaches the end of
requirement is given. If the machine shut height is adjusted its stroke. Fig. 6-4 shows the hydraulic pressure and tie bar
until the measured tie bar strain conforms to the value on strain relationship as the die closes and locks.
the calibration chart corresponding to the die tonnage, the
machine is producing only the force required for the condi In Fig. 6-4 it can be seen that the tie bar strain reached
tions. It is not working harder than necessary. only 68 percent of its capacity. That condition would have
been caused by having the shut height adjustment loosened
The example illustrated in Fig. 6-2 considered the situation slightly. If the die requires only 34 tons (302 kN) as indi
where the projected cavity area was not centered within the cated in the figure, it is probably desirable to operate the
machine’s platen area. In such instances the die should be built machine in the “under stressed” condition. The reduced ton
so the die is centered in the machine even though the cavity area nage minimizes wear and tear on both the machine and the
is not. Such die construction makes the die appear larger than die. Such dies are usually small in physical size in compari
necessary. It also makes the die more expensive than if only the son to the machine and high clamping forces can deform the
minimum size is used. If the die is not placed centrally within machine platen around these small dies. However, if the die
the machine, the tie bars will not be strained equally. The condi requires more than the 34 tons (302 kN) per tie bar of clamp
tion twists the machine out of shape and increases substantially force, the force of injection will open the die, and cause
the friction in the tie bar bushings and toggle bearings. The flashing. The opening of the die (i.e. flashing) will reflect in
condition reduces the clamping tonnage that can be generated, the tie bar strain as shown in the illustration.
Fig. 6-4. Hydraulic

pressure builds in the
die close cylinder as
the tie bars start to
stretch. However, as
the toggles approach
their maximum
mechanical advantage,
the hydraulic pressure
will actually drop.
Illustration courtesy of
ILZRO.

Fig. 6-5. When
hydraulic pressure
must build to the relief
valve setting before
the toggles reach their
maximum mechanical
advantage, some fluid
passes through the
relief valve extending
the machine cycle
time. Illustration
courtesy of ILZRO.
As described above, the tonnage reduction is achieved by loosen not be adjusted evenly. For example, a 5-deg. rotation of a nut
ing the shut height. However, once the shut height is loosened on a 0.25-in. (6.35 mm) pitch thread moves the nut 0.0035 in.
and the tie bar strain is reduced, there is a corresponding drop in (0.089 mm) or about 10 percent of the total tie bar elongation.
the instantaneous hydraulic pressure required to accomplish the If one tie bar is overstressed and another is under stressed the
locking of the toggles. Some machines have a metal plate attached twist imposed on the machine can cause binding and friction
which shows the relationship between tonnage and hydraulic pres that will actually reduce the total applied tonnage.
sure. However, such pressure-to-tonnage relationships apply only
to the instantaneous pressure as the toggles are locking. The final Temperature variations in the die and in the machine cause
hydraulic pressure achieved after the cylinder has reached the end changes in the physical size of the die/machine components.
of its stroke has no bearing what-so-ever on clamping tonnage. These size changes will cause corresponding changes in the tie
The die can actually be open and the pressure will be at full system bar strain. The most obvious effect of this type of situation is
pressure when the toggles are fully straightened. when a die warms-up, expands, and hence requires more tie bar
elongation than the clamp cylinder/toggle system can provide.
When a machine is operating at its full clamp capacity, the The clamp system then stalls. All hydraulic fluid dumps over the
pressure-tonnage relationship will be similar to that shown in Fig. relief valve, and the cycle is interrupted. The practical approach
6-5. That illustration shows how the hydraulic pressure builds to then is to back-off the shut height adjustment to compensate for
the relief valve setting before the toggles reach their maximum the larger die. If the die should then cool somewhat, it may flash
mechanical advantage. Some hydraulic fluid bypasses the cylinder because clamping force is insufficient.1
by dumping through the relief valve. The reduction in volume flow
rate of hydraulic fluid flow causes the cylinder to move slower, and Establishing tie bar tension manually through the shut height
therefore the machine takes more time to “lock” as indicated by adjustment is limited to the observation of rather gross devia
dimension “A” in the figure. A partially-loaded machine will actu tions in actual strain (i.e. machine stall or heavy flashing). Also,
ally cycle faster than when it is fully loaded. manual adjustments may be coarse as compared to the fine
adjustment actually required. If tie bar strain is monitored con
stantly, and maintained within very narrow limits, considerable
Measurement and Control economies can be realized in terms of total cycle time, machine
of Clamping Force maintenance and casting thickness. Several techniques are used
or could be used to measure and control tie bar strain.
The most rudimentary method of controlling the machine’s
clamp force is to set the relief valve at a nominal system pres
sure (and leave it at that setting permanently), tighten the shut
height until the machine will not lock, and then loosen the Dial indicators
shut height until the machine will just lock. This technique The simplest form of continuous monitoring of tie bar strain
may insure maximum clamp force for a few shots. is through dial indicators mounted at the rear of each tie bar,
as shown in Fig. 6-6. These indicators must be so installed
Several factors can cause undetectable variation in the tonnage that they measure elongation in a section of the tie bar that is
when the above technique is used. First, the tie bar nuts may actually stressed. In other words, the sensing length must be

Fig. 6-6. Dial indica
tors may be used to
measure tie bar strain.
However, their instal
lation must insure
that the indicator only
registers tie bar stretch
ahead of the threaded
portion of the tie bar.
between the threaded portions of the tie bar such as is illus of major unbalanced conditions and will investigate possible
trated in Fig. 6-6. The installation of the dial indicators into causes such as an out-of-parallel die, an improperly mounted
threadless tie bars is somewhat different in that the indicator die, or incorrect assembly of the die height mechanism.
can be mounted directly onto the end of the tie bar as shown
in Fig. 6-7. The positioning of the dial indicators is such that constant
monitoring is difficult because the machine operator is not
Dial indicators on the tie bars are a form of continuous moni usually at a location from which he can see them. However, the
toring. However, the operator, technician or set up person operator, setter, or other responsible person can spot check the
must read the indicators and convert the readings into ton indicators to catch developing problems.
nage. Then he/she must adjust the tie bar nuts to achieve the
correct tonnage. If the appropriate tonnage vs. strain charts
are available, the setter can adjust the machine to apply the Strain Gauges
correct force. Also, he can balance the force generated so all Tie bar strain can be sensed by special electric coils which
the tie bars carry the same load. If the machine has a motor change resistance when they are compressed or stretched.
ized shut height adjustment it may not be possible to adjust Known as strain gauges, these units can be buried in a hole near
the tie bars individually. However, the setter will be aware one end of the tie bar. Since the gauges and the wires leading to
Fig. 6-7. Dial indicators can be

mounted directly to the end of the
tie bars if the tie bars are the thread
less type. Illustration courtesy of the
Prince Corporation.

Fig. 6-8. Typical tonnage
meters display the desired
tonnage, low limit, high
limit and current actual
tonnages for each tie bar
and for the total machine.
Courtesy of IdraPrince.
them are fragile, it probably is best to insert the units into the can even interrupt the machine cycle. The digital read out can
tie bar, and have the wires protected in flexible steel conduit to be positioned in plain view of the machine operator or techni
a junction box on the platen and then run through pipe from cian. Tie bar mounted strain gauges (load cells) with digital
there to the read out and/or controller instrument. read-out and alarm devices are standard equipment or available
options on most new die casting machines.
Electrical strain gauges have the advantage of perma
nency. Once installed, they will monitor every cycle of the
machine. In fact, if any type of automatic control is to be Full Cycle Displays
used, electrical strain gauges permanently installed in the tie
It is sometimes useful to have a graph showing the tie bar
bars probably are a necessity.
tonnage performance through the entire machine cycle. Such
recordings are generally more useful for diagnosis of problems
Strain gauges simply vary current flow with size changes and
than for routine control. If the machine is equipped with a shot
hence the controlling instrument must be calibrated to compen
system monitor such as that illustrated in figure 5-5, and a digi
sate for variations from one strain gauge to another, as well as in
tal tonnage monitor as illustrated in figure 6-8, the tie bar strain
their installation and then convert the electrical output to tons
gauge outputs can be made to be displayed on the shot system
force. Most strain gauges that are used on die casting machines
monitor. (The reader must be made aware that having such equip-
are packaged into what is called load cells. The load cell is per
ment does not necessarily mean that it has been set up to do what is
manently calibrated to give an electrical output which is propor
being described here. The point is that it could be.) The resulting
tional to force (i.e. tons) and is therefore easier to use with general
display would look like that shown in Figure 6-9.
laboratory equipment such as oscilloscopes and chart recorders, as
well as digital read-out instrumentation and controllers.
Strain gauges mounted in the tie bars can be connected to

Continuous Control
digital read-out instruments which continuously display It may be desirable, or even necessary, to have a continu
actual tonnage to the operator. A commercial readout is ous automatic feedback control system in some instances.
shown in Fig. 6-8. Microprocessor based machine process control computers are
available for providing those control functions. These are an
The digital instruments can be equipped with set point controls extension of the digital read-out systems described above. Such
and upper and lower limits. Then, if tonnage does not reach the systems can provide continuous monitoring and comparison
lower setting, or if it exceeds the upper setting, the instrument with pre-set limits to, alarms, robot controls and central data
would operate an alarm system, such as a light or buzzer, and collection computers. When connected to a robot, such a

Fig. 6-9. Displays of the strain
in all four tie bars will show if
individual tie bars are out of
adjustment. Unequal loading
of the tie bar causes friction
and binding in the mechanical
parts of the machine which
wastes energy Such displays
will also show if the required
pressure in the die closing
cylinder is reaching the system
pressure and slowing the cycle.
system is capable of detecting excessive flash and instructing tie bar is loosened. Once the tight tie bar is loosened, the shut
the robot to automatically reject the casting. The machine can height could very well need tightening. This can all be accom
be automatically stopped if a certain number of consecutive plished automatically.
faults are detected. Three consecutive faults are commonly used
as the criterion. Such systems can even be used to prevent shots
if insufficient tonnage is present and to reverse the machine
Summary
closing mechanism if stress on any bar exceeds a safe value
during lock up. Continuous automatic feed-back control systems are now com
monly used for tie bar load control. Control of the clamp system
Systems are also available that can automatically re-adjust the can help control casting dimensions, reduce the magnitude
tie bars. The instrument stores the out-of-tolerance data until of other variables, reduce machine maintenance, and speed
the machine opens. It then signals the shut-height mecha problem diagnosis. Tie bar load, hydraulic pressure, and cycle
nism to open or close a prescribed amount. Balance of the element times are the critical parameters. The key to control is
load between tie bars can also be achieved. When the load the engineering calculations for required tie bar tonnage and
on a specific tie bar exceeds the set point on its controller and required hydraulic pressure. Then, measurements of actual
if the total machine load was not exceeding its set value, an load are made with electrical strain gauges or dial indicators
out-of-balance condition would exist. The shut height adjust and compared with the calculated requirements to determine
ment should not be loosened, only the single excessively loaded whether or not adjustments are needed.

CH APTER
Die Temperature
The flow of heat from the metal being cast determines the blow (flashing) variations can change the average die tempera
solidification pattern within the casting. If solidification ture or the pattern of the die temperature. It is usually more
progresses properly a usable casting results. However, if the cost effective to control die temperature directly than to control
wrong solidification pattern develops, the resulting casting all, or most, of the other variables that affect it.
must be scrapped. Die casting is therefore a thermal process.
The control of the thermodynamics of the process is all that Traditionally die temperature control was left to the machine
really matters. Much emphasis is placed on gating theory operator. However, the operator was never told what the
and injection control, but even those efforts are actually to temperature should be, now was he given the means to know
establish the best possible initial thermal condition. It can what it was. The operator was expected to look at the cast
be concluded that casting/die temperature control is the ings being made and interpret texture, warping or sticking
most important consideration in process control. A thorough to make temperature related decisions. Depending on how
discussion of the heat flow phenomenon 3 is beyond the scope he interpreted what he saw on the casting, he would speed
of this book. These phenomenon are developed in detail in up or slow down, change the application or release material
the NADCA courses and textbooks on Die Casting Dies: or adjust the cooling water flow. Fortunately there are better
Designing and Engineering Die Cooling Systems. How tools today for the control of die temperature.
ever, the temperature patterns that exist within the casting
and die must be understood and therefore a certain amount
of explanation is unavoidable. Techniques for measuring Determining Requirements
temperature patterns are described in this chapter, as are the
means of controlling those patterns. The die temperature requirements for a specific die are
determined during the die design activity. The preciseness
with which those requirements are established depends on the
The goal is to have control of the transient temperature
technique employed by the die designer. The techniques actu
condition in each casting as it is solidifying and cooling in
ally used are:
the die. Those transient thermal conditions drive the solidi
fication patterns and hence the distribution of porsity in the 1. Casual
casting. The control of heat flow through the die is actually
an indirect method of controlling the transient conditions
2. Informed experience
during casting solidification. 3 3. Basic heat balance
In one way or another all process variables are related to the 4. Finite element analysis (FEA)
transient heat flows during solidification. Cavity fill time and
compaction pressure have been defined previously in terms of Each of these techniques are described in the following para
solidification requirements or behavior. Plunger velocity, tie bar graphs so the process control engineer will understand what he
strain, and hydraulic pressures are actually means to achieve the must look for in each die before planning his control strategy.
needed fill time and cavity pressure which in turn achieves the
necessary thermal pattern or compensates for incorrect thermal
patterns in the solidifying casting. Casual
Die temperature is actually a pattern of temperatures. When The casual approach to the thermal design of the die cast
operating properly, the die will have various temperatures across ing die is to let the die builder drill waterlines wherever he
the parting surface. The sprue or biscuit area should probably wishes. The die builder should not be expected to understand
be somewhat colder than the cavity area. And, the outermost the nature of heat flow3 as that is not his trade. Even when the
extremities of the cavity should be hotter than near the gate. die builder does understand the heat flow considerations, he is
Similarly, the die gets colder at greater distances into the die faced with the conflict of doing what is correct from a heat flow
steel from the cavity surface. The coldest areas are immediately point of view and building the die in the most cost effective
adjacent to the water cooling lines and/or at the outside surfaces. way. Usually cost effectiveness wins. The result is a die with few
water-lines that are too long and go through the wrong places.
The average die temperature is effected by several other vari In many instances a single waterline will have many inches of
ables. Cycle timing, metal temperature, shot speed and die length where no cooling is needed and only a short portion

near a hot spot that really needs cooling. A large percentage method of doing a basic heat balance. As of this writing, relatively
of dies are still built with the casual approach to temperature few dies get a basic heat balance analysis. However, the practice is
control. The control techniques described later in this chapter showing some increase in popularity. Unfortunately, if a basic heat
may or may not be of benefit to these dies. In the worst cases, balanced die is not operated with proper temperature control, the
the only control that can be exercised is through using hot oil die caster may not get the expected performance.
circulating through the cooling lines, the selective application
of the release material (die spray) and/or through cycle timing.
Those variables are discussed in chapters 8 and 10 respectively. Finite Element Analysis (FEA)
Fortunately true finite element analysis techniques are avail
Informed Experience able for analyzing heat transfer in die casting dies and it can be
used to help the designer locate the waterlines very accurately.
Dies designed with informed experience are much more It can also be used to determine accurately the proper set-point
likely to respond well to the control techniques described in temperatures for controlling thermocouples. The reader must
this chapter. In these situations the die designer has a good be aware that just because a die was analyzed with finite ele
understanding of the heat flow phenomenon and places the ment analysis, there is no certainty that the designer used the
waterlines and circuits them so they are generally in the right information to optimize the water cooling lines. The analysis
place and can be controlled individually between regions with only predicts how the design will perform. It does not design
different performance patterns. However, the informed experi the cooling system. So, repeated designs must be created and
ence method has serious limitations and will rarely result in an analyzed to zero in on the best design. The best approach is to
optimized operating condition. use both the heat balance technique along with finite element
analysis to design the water (or oil) cooling lines.
Basic Heat Balance
When the die designer does a Basic Heat Balance he actually
calculates the casting’s heat inputs to every region of the die and Heat Flow Through the Die
places cooling lines in specific calculated locations and of specific Heat is input to the die by the molten metal.3 Cycle time, cast
calculated lengths to remove the actual amounts of heat that will ing thickness, cavity fill time, and metal injection temperature
go into the various regions of the die. The basic heat balance pro influence the rate of heat input. Changes to any of these variables
cedure sounds good, but it is still not a complete solution because will change the rate of heat input. The heat input to the die is
of the complexity of three dimensional heat flow. However, a absorbed into a fairly thin layer of die material (except for very
basic heat balance will provide a die that is highly controllable and heavy wall castings) behind the cavity surface. The heat absorbed
with nearly optimized performance. In many instances, these dies from a single casting is shown diagrammatically in Fig. 7-1.
will perform as well as those with more extensive analysis when
coupled to the control techniques described in this chapter. The The heat absorbed into the surface region of the die cavity
NADCA textbook Die Casting Dies: Designing describes one from one casting must be removed from that region of the
die to bring the temperature of the surface back to the base
value, prior to the next shot. Some of the heat travels back out
through the cavity surface as the parting material carrier (i.e.
water or solvent) evaporates. The remainder flows through the
interior of the die to water lines and to the outside die surfaces
from which it is discharged by radiation or convection.
The rate at which the heat is removed from the surface region
is dependent on the mean cavity surface temperature, type and
quantity of parting material carrier, outside temperature of die,
cooling water flow rate, cooling water temperature, location of
the cooling channel, coating (i.e. calcium deposits) on the cool
ing channel wall, and the ambient conditions. All these factors,
except those related to the parting carrier material, result in
the thermal gradient shown in Fig. 7-1. The gradient actually
determines the heat flow rate. Most of the factors that affect the
gradient are in themselves variables so the heat removal through
the die also is a variable. The role of the parting material carrier
is quite substantial. It is discussed in greater detail in Chapter 8.
When, over time, the heat flowing away from the cavity surface
is not equal to the heat being input from the castings, the
Fig. 7-1. Heat (shaded area) from each casting is absorbed into the die steel
average surface temperature changes. A typical pattern of the
immediately behind the cavity surface. The absorbed heat must be removed
temperature variations over time is shown in Fig. 7-2. Changes
from that region before the next shot is made.
in the average temperature can speed or slow solidification.

Measurement and Control
Control of the heat removal from the cavity surface region
traditionally has been achieved by the operator’s application of
the parting material and his adjustment of water flow through
the cooling channels. However, more consistent die tempera
tures can be achieved with the use of thermocouple control of
the water flow. Commercially available chromel-alumel type K
or iron-constantan type J thermocouples with a stainless steel
sheath are placed between 1/4 and 1/2 in. (6.35-12.7 mm) from
the cavity surface. The thermocouple installation is shown in
Fig. 7-3 and may be 1/8, 3/16 or ¼” in. diameter. Fig. 7-3. Thermocouples are usually spring loaded to hold the sensing junction
(which is welded to the tip of the sheath) tightly against the end of the hole to get
an accurate measurement.
Basic Control
The electrical signal from the thermocouple goes to an Zoned Control
indicating controller which indicates the temperature and In practice, some areas or zones of the die usually have a dif
has an adjustable set-point. When the temperature reaches ferent pattern of temperature variation than others. The typical
the setting, the controller turns on a solenoid valve to allow pattern is a series of generally concentric rings, or zones, around
cooling water to flow through the die. Such a temperature the biscuit or sprue as shown in Fig. 7-6. The area within each
control system is a continuous monitoring system with set zone will have its own unique temperature behavior. At the very
point on-off control. Most systems with automatic control center is the biscuit or sprue. It has the most intense heat flow
have manual bypass systems around the solenoid valve so situation. The die cavity surfaces in this area will be the coldest
that full manual control is possible. Also, the manual bypass before the shot is made; but as the shot is made, will reach the
allows a continuous low flow rate through the die with the highest momentary temperature of any part of the die. The die
high flow rate provided by the solenoid valve and thermo temperature will respond quickly to any change in cycle rate,
couple. When set properly, the low flow rate will remove metal temperature or cooling water flow. Once a stable operat
most of the heat. The high flow rate will be required only ing condition is established and the die is run at a steady rate it
occasionally. The thermal shock to the die is thereby mini is often possible to set the water cooling to be “full-on” when the
mized. The total system is shown schematically in Fig. 7-4. die is running and “off” when it is not running.
By regulating the water flow the thermocouple controlled The next ring, Zone 1 in Fig. 7-6, includes the runners and
cooling system will minimize variations in die temperature gates. This zone is the second most intense and sensitive area of
as shown in Fig. 7-5. Specifically, this type of control sta the die. Also, the die temperature in Zone 1 is likely to affect
bilizes the average cavity surface (or the transient tempera the molten metal flowing through it during die filling. These
ture). Therefore, solidification patterns are stabilized, more effects can show on other parts of the casting since the metal
good castings are made each shift, there are fewer “stickers,” flows on to those other areas. Zone 1 can usually benefit from
and die life is increased. direct temperature control.
Fig. 7-2. When heat input

does not equal heat output,
the base and average cavity
surface temperatures will
increase or decrease accord
ingly. The die temperature
at about 1/4 in. behind the
surface is less than the base
surface temperature, but fol
lows the trend of the surface
temperature. Temperature
trends near a water line lag
those at the surface and are
much smaller in magnitude.

Fig. 7-4. Cooling water flow to the die
is regulated by thermocouple controlled
solenoid valves to maintain the desired
die temperature.
The major portion of the actual cavity features lie in Zone 2. In surfaces that are exposed to the air and machine platens.
Zone 2 the thermal response is sluggish. Die temperature will Except for the extremities of the cavities, the temperature
rise slowly. Several (sometimes even tens of) castings must be of the die in Zone 3 has no direct effect on the casting. The
made to get the die to the proper operating temperature. And, primary function, from the thermal standpoint, of Zone 3
if the waterlines are properly placed, adjustments to the cooling is to provide the proper thermal environment for Zone 2.
water flow will have moderate effects on the temperature drift. The thermal performance of Zone 3 is dominated by the
The casting, however, is most sensitive to the die temperature heat losses from the exposed surfaces of the die, and these
in Zone 2. Therefore, this zone is where it is most important to are constant regardless of the operating status of the die. In
have direct control of the die temperature. extreme instances, its temperature control requirements are
unaffected by whether the die is even running or not. Often
The last region is Zone 3 in Fig. 7-6. The overflows are it is necessary to circulate hot oil through drilled channels
usually in Zone 3. Sometimes the extremities of the cavities or insert electric cartage heaters into this zone to maintain
are in Zone 3 too. Zone 3 contains the bulk of the outer die the desired temperature.
Fig. 7-5. By automatically controlling cooling water flow [bottom of chart] the actual temperature variation of the cavity surface is minimized. The total number
of good castings per shift is maximized by this type of control. Also, die life is increased. The top curves compare the uncontrolled die temperature from Fig.
7-2 to controlled die temperature.

Specific features within a cavity may need their own individual
temperature control. Thick heavy features in local areas of other
wise thin castings will have different temperature requirements
and have different response characteristics than the thinner cast
ing around them. Most dies require from two to four tempera
ture control circuits per die half. Large complex dies may need as
many of 12 circuits. A total system for both heating and cooling
is shown in Fig. 7-7. The names of some companies that supply
temperature control equipment are given in Appendix B.
Zoned temperature control allows control of the solidification

time in various parts of the cavity. By controlling the die tem
peratures, the gradients, and hence the solidification time of
the zones can be regulated in respect to each other to enhance
either the feeding of shrinkage or the dispersion of shrinkage.
Fig. 7-6. Typically die casting dies have concentric thermal zones. Each zone has
a unique temperature requirement and temperature response characteristic. Some
Summary
of the zones may need to be further divided into sectors for control purposes. Heat flow patterns in the die casting die are very complex.
The pattern varies across the cavity surface and through the
die material. Also, the temperature at any one point in the
In addition to the concentric zones described above, the die may die will fluctuate during the casting cycle. All these ther
need to be further divided into sectors such as is also shown in mal patterns change with time. Temperature measurements
Fig. 7-6. It is very common for one cavity to behave differently at selected points behind the cavity surface can be used to
than another. So, the right hand side of the die may need to be control cooling water flow, and thereby control the actual
controlled separately from the left. Or the top of the die may temperature of the die in that area. Control of the heat flow
need to be controlled separately from the bottom. Sectoring is through the casting-die system is necessary to achieve high
most frequently needed in Zones 1 and 2. quality castings on a repeatable basis.
Fig. 7-7. A typical die

temperature regulation
system consisting of: (1)
solenoid valve, (2) needle
valve, (3) water pump,
(4) oil lines, (5) water
lines, (6) temperature
regulator, (7) oil tem
perature regulator, (8)
submerged resistor and
thermostat, (9) sealed
cooling pump, (10) oil
pump, (11) resistor-type
heating chamber, and
(12) air vents.

CH APTER
Release Material
When the die has opened and the casting has been ejected, but perature of the molten metal in combination with the high
before the die closes again, the cavity areas of the die receive pressures cause chemical reactions to take place within the
an application of a release material. The application is usually film of release material during and just after injection. These
made by spraying the material as a fine mist. Sometimes the reactions can be exothermic (i.e. release heat), isothermic (i.e.
application is not made every time the die is open, but only have no heat released nor absorbed) or endothermic (i.e. absorb
after some small number of shots. The purpose of the release heat). If such reactions add or delete heat, they will affect the
material is to keep the casting from sticking to the die. The way the metal flows into and through the cavity.
material is called by a variety of names including “lube,” “die
spray,” “die lube,” “spray,” “parting” and “release.” The chemical reactions described above result in much of the
release material being “burned” away. The burning away is
The- practical die caster knows that the release material can desirable since the material will not then build up onto the die
have a significant influence on the performance of the die cavity (i.e. “varnish” or “scale”) or “stain” the casting. The pig
casting die. Sometimes, by changing the type or brand of ments, if any, will not burn off. Release materials for the high
material or by changing the application pattern or quantity, the temperature alloys such as brass may need to include pigments.
operating performance of the die is improved. The improve The decomposition of the material must happen slowly enough
ments are not limited to the sticking of the castings to the die. that the cavity has time to fill and the casting’s skin form
Often surface finish, productivity, porosity or even warpage is before the release material disappears.
affected. The release material and its application is a significant
processing variable. Some release materials are graded in terms of their “release,”
“anti-solder” and “lubricity” properties. These properties are
probably not available in engineering units, but the rankings
help the process engineer select the best material. The die caster
Determining Required Performance should keep a history of the release materials used, the jobs that
The performance factors25 required from the release material are: each was used on and the performances. Such a history could
be a help in selecting the material for a new job.
Release
Anti-solder The release material is applied to the die by spraying. The
material is first mixed with a solvent (usually water) and
Lubricity (of the moving mechanical parts of the die) then sprayed with compressed air through spray nozzles. The
water boils off the hot die leaving the desired film. The sup
A thin (and hopefully uniform) film of the material is applied pliers of the release materials take great pains to formulate
to the cavity surface and to the sliding parts of the die to the materials to stay in solution with the water (it must not
achieve those performance factors. The film of release material separate in the use tank) and to insure that it will spread-out
is a complex mixture of organic compounds and sometimes (i.e. “wet”) on the cavity surface.
pigments such as graphite and aluminum particles. Except
for lubrication for the moving mechanical parts of the die, the The processing engineer must determine how long the spray
basic function of the film is to keep the molten metal from nozzles must spray to discharge the required amount of water
establishing intimate contact with the die steel. Without con and release before each shot. The total amount of solution to
tact, the casting can not stick to the die. be discharged is the combined total of the water required. For
example, assume that for each casting cycle a particular region
The film also forms an insulating barrier between the molten of a die requires 0.70224 cu. in. (11.51 cm3) of water spayed for
metal and the die. That barrier slows the rate at which heat is cooling purposes and 0.00186 cu. in. (0.03cm3) release material
lost from the metal to the die. 26 The greater the thermal resis for an adequate barrier coating. The total amount of liquid to
tance, the better the casting’s surface finish and the longer the be sprayed each cycle is then the sum of both which is 0.70410
filling time can be. These relationships are developed in more cu. in. (11.54 cm3). If the spray nozzle discharges 0.35205 cu.
detail in the NADCA textbook, Gating. Greater thermal resis in. per second (5.77 cm3/sec), a single nozzle must spray for two
tance, however, may slow solidification somewhat. Increased seconds to discharge the required amount. If the spray nozzles
thermal resistance of the release film will reduce thermal shock are on a programmable reciprocating apparatus that passes
and therefore heat checking of the die steel. 27 The high tem them across the die face as they spray, it must be programmed

Fig. 8-1. The die designer should
supply a water spray diagram like
the one shown here for the process
ing engineer and the operating per
sonnel. The dilution ratio must then
be made to get the correct amounts
for both water and release material
on each area of the die.
to move at the correct speed and to start and stop the spraying be deposited. If the release material is staining the casting or
action of each nozzle at the appropriate time. building up on the cavity surface, less release should be depos
ited. Likewise, if the cavity definition is not being reproduced
The boiling away of the water removes heat from the die surface. faithfully onto the casting or there are cold shot and chill on
When the die designer does a basic heat balance, thermal analy the casting’s surface, there should be more release deposited.
sis or a finite element analysis of the thermodynamics of the die,
he must determine how much heat should/ must be removed by It might be possible to get a fairly good measure of the total
the evaporating water. He will specify that certain areas of the amount of liquid (i.e. water plus release material) sprayed
die surface must have a specific amount of heat removed this way onto the die. One could draw the outlines of the desired
and other areas may have some other amount of heat removed spray pattern (such as is shown in Fig. 8-1) on a large piece
by the water spray.24 The die designer should communicate those of blotter paper. Then weigh and measure the piece of blotter
requirements to the processing engineer and to the operating paper and calculate the weight per square inch (or square cm).
personnel with a diagram as shown in Fig. 8-1. Next, hold the paper in front of the die parting surface and
cycle the spray onto the blotter paper. If there is more spray
than what the blotter paper can soak up, there is probably too
much water (and release) being sprayed onto the die. Then
Measurement and Control cut out each separate region of spray pattern from the blotter
There is no proven “good” way to measure the actual thickness paper and weigh each piece. Measure each piece to determine
of the film of release material. The best measure is the perfor the area, and calculate the new weight per square inch. Each
mance of the operation. If the casting sticks to the die or if fluid ounce of water weighs one weight ounce, and the differ
there is solder build-up, more release material should probably ence between the wet weight and the dry weight will be the

amount of liquid absorbed by the blotter paper. The weight Dilution ratio = Volume of water/Volume of release (8-1)
can be compared to the desired amount of spray and the
nozzles and/or spray time adjusted accordingly. So, if 1 gallon of release is mixed into 25 gallons of water,
the dilution ratio is 25.
As the amount of release is changed to adjust for the casting
conditions, it must be remembered that the cooling effect is Most release materials are already mixed with water when received
also changed because different amounts of water are sprayed from the supplier. The mixture ratio of the material “as received”
onto the die. The interdependence of cooling and release is must then be taken into account before the material is actually
quite unfortunate since both are variables that the process is diluted. The supplier of the release material can best advise as to
highly sensitive to, but quite possibly in different ways. the best dilution ratio for each application of his product.
The dilution ratio of water to actual release material is
established by Eq. 8-1.

CH APTER
Casting Ejection Temperature
A powerful control technique is to open the die and eject the Determining Required Performance
casting when the casting’s actual temperature has declined
to some previously determined ideal value. The technique is As described in Chapter 7 there are many different thermal pat
no different than implanting a temperature probe into the terns in a die casting die. Since there can be only one location
thanksgiving turkey to shut off the oven when the turkey is of the thermocouple that determines when the die opens, its
cooked. Castings, like turkeys, are not all ready to come out location must be selected with care. The first criterion is that it
after the same length of time. The control is of the continu be placed in an area that will solidify later than the other areas.
ous monitoring with control to set point type. Tradition That location insures that the die will not open while part of the
ally, the solidification and cooling time has been controlled casting is still liquid. The biscuit, or sprue, or the runner near
by a timer on the die casting machine. By controlling the the biscuit meet this criterion. However, since different zones
opening of the die casting machine from actual casting of the die have different thermal performance (as described in
temperature, significant improvement in productivity and Chapter 7), it might be better to place the thermocouple at a
dimensional stability is realized. When a timer is used, all thick heavy portion of the casting. It should never be placed in
castings must stay in the die the length of time required for Zone 3 as defined in Fig. 7-6.
the slowest casting. The technique of controlling to the cast
ing temperature stabilizes many of the process variables such Once the location is selected, the thermocouple’s performance
as die temperature. The most dramatic effect is the stabiliza should be predicted and a control strategy established. Both the
tion of casting size. The greatest contribution to variation location of the thermocouple and the control strategy, includ
of casting dimensions’ is ejection temperature variation. So, ing an approximate temperature set point, must be established
when ejection temperature is constant, so is size. during the die and process engineering activities. However, the
Fig. 9-1. The tempera

ture of the metal being
cast drops quickly at
first, as shown in this
plot of calculated tem
peratures for a 0.060-
in. thick magnesium
casting. 3 The rate of
temperature decline
slows as the difference
decreases between the
casting and die surface
temperatures. The curve
flattens (eutectic arrest)
as the solidification
process is completed and
the latent heat of the
eutectic is released.

Fig. 9-2. Heat
absorbed into the sur
face an H-13 die near
the gate during cavity
filling results in these
calculated 3 tempera
ture patterns. The die
surface temperature
increases several
hundred degrees is a
few milliseconds.
Fig. 9-3. Initial cavity

metal temperature, die
surface temperature
build-up, and solidifi
cation rate are all usu
ally quite different near
the overflows than near
the gate.

actual set point of the thermocouple will probably be estab The temperature pattern of the cavity surface follows that of the
lished through trial and error experimentation once the die is casting, although it is always lower. The die surface temperature
actually making castings. also reflects the eutectic arrest. The cavity surface temperature
also drops fairly quickly when the casting is removed, and again
when the parting material is applied. These characteristics are
Temperature Patterns illustrated in Fig. 9-1 and 9-4. The temperature of the die material
at/near the Cavity Surface at a small (e.g. 0.020 in.) distance behind the cavity surface will
be lower than the cavity surface, but will follow the temperature
To properly install and operate the casting ejection temperature trend of the cavity surface. The temperature of the die material at
control, one must understand the temperature behavior of the that small distance from the cavity surface becomes more like that
die cavity surface temperature through the casting cycle. As the at the cavity surface as the ratio of casting material to die material
metal flows into the cavity, the metal temperature in the flow increases. So, a point at the tip of a small core pin in a large casting
stream inside the cavity near the gate will remain fairly con boss will be quite representative of the casting temperature. A
stant until the cavity has filled. Once the cavity fills, flow stops continuous chart recording of cavity surface temperature will show
and the heat losses to the die can no longer be replaced with minimum and maximum temperatures (i.e. thermal shock), the
heat from fresh hot metal. The temperature of the metal in the solidification point, the die open (or ejection) point, the time when
cavity then drops, as shown in Fig. 9-1. As heat builds up in the parting material is applied, and cycle time. The temperatures
the die material below the cavity surface there is a correspond associated with these events are referred to as base, peak, solidifi
ing temperature increase in the die surface. cation, die open, and spray die, respectively as shown in Fig. 9-4.
The die surface temperature increases most quickly and to the If surface temperature recordings are made simultaneously
greatest degree near the gate during cavity filling3 as shown in for two or more critical points in the casting, the differences
Fig. 9-2. The temperature build-up at the die surface near an between the solidification times can be observed as shown in
overflow as shown in Fig. 9-3, is neither as great nor as rapid. Fig. 9-3. If a “shrink-feed” casting is being made, it is neces
Solidification is indicated by a flattening of the temperature sary for solidification to be completed first in those areas of
curve followed by continued temperature drop. The momentary the casting which are farthest from the sprue or shot sleeve.
halt in temperature drop is known as the eutectic arrest and is Solidification should then progress steadily towards the
caused when the solidifying metal releases its latent heat.4,21 sprue or shot sleeve.
Fig. 9-4. Continuous

recording of cavity sur
face temperature shows
the thermal shock,
solidification point, die
open point, application
of parting material,
and cycle time.

surface. Surface thermocouples have been placed successfully
in cores and ejector pins. Commercially-available mineral-
insulated nickel-chromium/nickel-aluminum types with
0.040-in. (1.0 mm) diameter stainless steel, or “Inconel,” outer
sheaths have performed satisfactorily in practice when properly
installed.23 The type that has the thermocouple junction welded
to the outer sheath is preferred.
To obtain the necessary rigid mechanical location and good

thermal contact, the thermocouple must be brazed into position
at the die face. Because it is very difficult to braze the thermo
couple into a large die insert, it is better to locate it in a small
fixed core, if it is in a suitable position, or to make up a special
thermocouple insert. Thermocouple should be brazed so that
the tip is within 0.005 in. (0.13 mm) of the die face. A nickel-
Fig. 9-5. When the thermocouple can be installed into a core pin, the thermo chromium powder brazing alloy has been found suitable for
couple sheath should be no larger than 0.062 in. in diameter and be pushed zinc, aluminum and brass die casting. If necessary, the inserts
tightly against a spherical ended hole in the pin that locates the thermocouple should be hardened and tempered after brazing.
junction no more than 0.020 in. from the end of the core pin.
The choice of position for the control thermocouple will depend
on the requirements of the die. When maximum reproducibil
Measurement and Control ity of dimensions is required, the temperature of the casting at
ejection should be maintained constant and the die may be run
When the controlling thermocouple can be at a thick portion below the maximum rate to avoid casting distortion; a control
of the casting, it can usually be mounted inside a core pin. The point on the casting can be selected for this purpose, and may
core pin should be one that projects well into the thick part of not be the critical position for maximum casting rate. For cast
the casting such as a boss, runner or biscuit. The thermocouple ings which are dimensionally critical, it may be necessary to
can be a type K or type J and be in a stainless steel sheath. The link two thermocouples to one controller and to use the average
thermocouple should be grounded to the sheath for fast response. temperature or whichever temperature reaches the control
The actual thermocouple junction must be pressed securely point last, for the die opening control.
against the end of the hole in the core pin, and the hole in the
core pin should have a spherical radius to match well the end of The actual temperature set-point for the machine opening
the thermocouple. The thermocouple junction should be no more controller must be established by trial and error. It might be
than 0.020 in. (0.5 mm) from the end of the core pin and closer possible to predict accurately the set-point from a rigorous
if possible. The construction is shown in Fig. 9-5 Finite Element Analysis (FEA), but that is rarely done. The
trial and error procedure is not difficult nor is it particularly
Standard H-13 core pins can be used for the thermocouple time consuming.
mounting, but a better response can be obtained with a TZM
core pin. Response time can be improved by use of an infrared The casting rate could be limited by the time required for the
sensor. The thermocouple is replaced by an optical fiber probe biscuit to solidify. The biscuit is most likely to be the limiting
to transmit the infrared radiation from the TZM core to the feature for thin castings produced on a cold chamber machines.
sensor.29 The sensor is mounted in the machine’s control panel. The thermocouple should be located opposite the shot sleeve in
those instances when the biscuit is the last to solidify.
Sometimes the thermocouple or optical fiber probe can not
be installed into a core pin but must be flush with the cavity

CH APTER
Cycle Timing
There is an old adage among die casters: “You have to keep a machine, the accuracy of the set up and whether or not the
constant cycle.” The practical die caster recognizes that the pro process is in control.
cess works best when it is running steadily. Things that make it
slow down (such as operator fatigue) or things that cause a stop
and go operation (such as sticking castings or flash build up) do Determining Desired Timing
much more harm to productivity than just the lost time. The desired timing is dictated by the built in performance char
acteristics of the machine, the type of actuation selected for ejec
Besides the uniformity of the total cycle, the way the cycle tor pins and moving cores in the die, other auxiliary equipment
is divided among the various segments and the uniformity and finally the casting’s processing requirements. The 14 basic
of each segment is also important. The ratio of die open to segments of the die casting machine cycle are charted in Fig.
die closed time is important. Also, the timing of each seg 10-1. That figure shows an operation with moving cores that are
ment of the cycle is a good indicator of the condition of the hydraulically actuated and with hydraulically actuated ejection.
Fig. 10-1. The typical die casting cycle has fourteen segments. Some of these (i.e. machine close, machine lock, decompression, and machine open) as well as the
minimum machine cycle have fixed times as a result of the machine’s design and condition. Others (i.e. ejector retract, cores move in, pour charge, cores move
out, ejection, remove casting and spray release) are constant once the auxiliary equipment is chosen and the die mechanisms designed. Only the solidification
and delayed closing segments are truly process dependent.

Machine Fixed Segments The core and ejector motions can often be operated mechani
cally24 by the opening or closing of the machine. Such mechan
The Machine Close, Machine Lock, Decompression, Machine ical actuation causes those motions to occur simultaneously
Open, and the Minimum Machine Cycle are fixed by the with the opening and closing of the machine as shown in Fig.
design and condition of the machine. The Minimum Machine 10-2. In some instances, considerable productivity improve
Cycle is determined by the time required to recharge the shot ment can be realized mechanical core operation. For the
system accumulator. The minimum machine cycle is inversely example in Fig. 10-1 and 10-2, the increased productivity is 4.5
proportional to the horsepower available to the shot hydraulic sec. per cycle or 16 percent.
system. Generally, these segments will not be the limiting fac
tors to productivity. If the 28 second cycle shown in Fig. 10-1 is required for the die
to achieve the necessary temperature patterns3,24 for the next
shot, there is nothing to be gained by mechanical actuation
Auxiliary Functions of the cores and ejectors. The 4.5 second time saving must be
The cycle segments of Ejector Retract, Cores Move In, Pour added back into other segments of the cycle, possibly segment
Charge, Cores Move Out, Ejection, Remove Casting, and 14. The question of whether such a thermal dynamic limita
Spray Release, are auxiliary to the basic machine functions. The tion exists can only be answered by a thermal analysis24 (and
timing of these auxiliary functions is somewhat effected by the in some instances by the basic heat balance calculations) of the
choices available to the processing and/or die design engineers. die. The thermal analysis would usually be considered part of
The cycle depicted in Fig. 10-1 is one where hydraulic opera the die designing activity.
tion of the moving cores and ejector has been selected. The
ejector must return and close a limit switch before the cores The trade-off between tooling costs and operating costs
move in. When the cores are “in” they close limit switches, is another consideration in the selection of mechanical vs.
and that signals the machine to close. The sequence is reversed hydraulic actuation of the cores and ejector. In general, the
when the machine opens. mechanically actuated mechanisms are more costly to build
Fig. 10-2. The auxiliary functions of ejection and core movements can occur simultaneously with the machine motion if the die is made with mechanical actua
tors for those functions.

into the die. The higher costs are especially the case for large on the basis of time and motion concepts. Anyone that sets up
dies and large cores. Being part of the die, these costs are “up a manufacturing operation with obvious delays is always open
front” money. From a business standpoint, it is usually best to to criticism. So, the time is sometimes added into the cycle by
delay expenditures. So, the extra tooling costs must be adjusted adding time for spraying release. It is sometimes provided, at
for the time value of money (equivalent interest earning power) least in part, by having sequenced rather than simultaneous
and amortized into all the parts to be made during the life of auxiliary functions. The net result is the same. The total cycle
the product. If the resulting increased piece part cost exceeds accommodates the thermodynamics of the die. By putting
the operating costs of the extra 16 percent of cycle time, the extra time into the spraying of the release material or the die
mechanical system should not be used. closed time, additional die cooling is achieved which might
work to shorten the cycle time, but more likely just reduces
The Pour Charge (i.e. ladle), Remove Casting, and Spray die temperature and hence die life.
Release segments are also auxiliary functions, but are some
what different than the die actuations. These relate to auxiliary Sometimes the required cycle time is achieved by adding the
equipment associated with the die casting machine rather than time to the solidification segment instead of having delayed
the die. These time segments can only be changed by replacing closing time. That segment is always considered to be pure
or updating that equipment. One must make sure that if the processing time (and therefore required) and not likely to draw
equipment is programmable that it is not poorly programmed criticism. This ploy may not work if dimensional tolerances are
and is therefore operating more slowly than necessary. critical on the casting. The actual size of the casting is strongly
dependent upon its ejection temperature1 (see Chapter 9), and
the ejection temperature is determined, in part, by the time the
Inject Metal casting is in the die. So, solidification time and total cycle time
must be considered separately to achieve some performance
The cycle segment of injecting the molten metal is process
requirements such as casting size.
dependent, 2,13 but, is such a small part of the total that it is very
seldom a consideration in the sense of the processing variation
considered in this chaper. The timing within the segment is
critical to the process, but those considerations are discussed in Actual Performance Measurement
Chapter 5. Technically, the intensification build up time is part The actual timing of the machine cycle segments can be mea
of this segment although it overlaps solidification. sured directly from the machine’s programmable logic control
ler (PLC). Each segment begins with some type of electrical
control signal. A limit switch must close or open, a timer times
Process Dependent Segments out, or power is directed to a solenoid. If these are not conve
The solidification and Delayed Closing segments are the only nient, the PLC can be programmed to output specific signals
parts of the cycle that are truly process dependent. The solidi for the purpose of measurement.
fication time must be long enough to solidify and adequately
cool all parts of the cast shot. The principle effect of increasing The PLC signals can be processed through a resistor panel to
solidification time is to decrease casting ejection temperature. reduce the voltage as required and then recorded on a chart
The ramifications of casting ejection temperature are discussed recorder or read directly by a micro processor. A micro proces
in Chapter 9. For each size of die casting machine, there is a sor can be programmed to display the segment timing much
minimum solidification time for the biscuit or sprue. In many like the bar charts shown in Figs. 10-1 and 10-2. The micro
instances the biscuit or sprue governs the solidification time. processor could even highlight any segment not performing to
The casting could be removed sooner, if sprue or biscuit were a pre-established time duration.
solidified so the machine could be opened.
A micro processor could monitor the operation continuously
The delayed closing segment is almost always hidden and its and activate an alarm if any segment exceeded its designed
existence denied. The function of the delayed closing is to limits. Then a technician could review the displayed chart of
provide sufficient total cycle time for the cooling water circu all segments as well as other performance measurements to
lating through the die to remove the necessary amount of heat determine corrective action. The micro processor could have
before the next shot. However, if it is put into the cycle as a stored in memory the original performance characteristics
true delay, the machine and operator will appear to be under of the machine. These data could be an important preventive
utilized and managers will try to get the operation speeded up maintenance tool.

CH APTER
Mechanical Die System
Generally the mechanical parts of the die are not considered If the die could be cooled periodically to room temperature and
to be processing variables. Nevertheless some parts of the die measured, changes in dimensions of cavity features could be
are subject to wear and/or breakage and can therefore put the assessed quite straight forwardly. However, it is usually more prac
process “out of control”. When a part breaks, it is a catastrophic tical for many reasons to measure the castings and deduce the con
event. As wear occurs the process exhibits continuous drift. dition of the die from those measurements. Unfortunately, a single
Both of these characteristics were described in the Introduc measurement of a single casting may not accurately reveal the true
tion, Chapter 1, of this book. Continuous drift is the behavior condition of the die. Statistical methods are often required.
that is ideally suited to statistical process control (SPC) meth
ods, and that is the condition discussed in this chapter. When a die casting die is operating under normal conditions and
the cavity is not deteriorating, there is considerable variation of the
casting dimensions. Those short term dimensional variations will
Characteristics of Die Cavity Wear tend to mask the long term deterioration due to wear. The short
term process induced dimensional variations are of four types:
Die cavities wear slowly. Usually it takes weeks or months to
achieve a distinguishable amount of wear. Wear on the cavity 1) Linear
surface is caused by erosion/cavitation, heat checking, and
soldering. Dimensions of features created by two different die 2) Parting Line (Blow)
members (such as dimensions that cross the parting line) can 3) Shift/Match
change because of wear in the mechanical surfaces such as gibs
or leader pins that align the die members. 4) Warpage
Erosion and/or cavitation occurs where gas bubbles entrained The specific type of variation is determined by how the feature
in the molten metal flow collapse with great violence at the of the casting is formed between the members of the die. The
instant of cavity filling. mathematical relationships between the dimensional variations,
die constructions and processing variables is developed fully in the
Heat checking is the most common form of cavity wear. The NADCA Textbook Dimensional Repeatability and the associated
thermal shock of the injected metal deforms the steel at the NADCA course. A statistical process control (SPC) X and R chart
surface of the cavity. The repeated deformations fatigue the as described in Chapter 3 would quantify those short term varia
steel and it develops hair line cracks. At first the resulting tions. Such a control chart would look much like Fig. 3-1 except
cracks form a crazing of the surface. The cracks can either the vertical scales would be dimensions instead of temperatures.
grow into visible crazing, or can let the small pieces of die
steel between them be broken away. If the sprayed on release The short term dimensional variations would have periods
is washed out of the cracks during cavity filling, the molten between peaks of a few hours, and in some instances several
metal can fill them and cause the casting to stick to the die. minutes. Samples should be collected and measured hourly.
When the casting is ejected off such surfaces, the cast metal Within a few days, control limits could be established and an
may shear away from the cavity leaving an obvious trace of effective SPC system put into operation.
cast metal on the cavity surface. This condition is usually
identified as soldering although heat checking is the root
cause. (True soldering is any time the cast metal establishes Establishing Desired Conditions
a metallurgical bond to the cavity surface. The casting is
sheared or torn off the resulting intermetallic layer as it The goal is to avoid having any dimension get outside of the part
is ejected leaving a rough layer of cast metal on the cavity print tolerance due to die deterioration. So, SPC X and R control
surface.) charts should be set up for all critical dimensions plus non-criti
cal dimensions that are expected to be (or subsequently found to
Soldering and heat checking do not necessarily change the be) subject to wear. Initially, such charts are not likely to show
surface dimensionally. Soldering can result in a build up on the the long term drift of wear. Instead, they will describe the effects
cavity surface that can change a casting’s dimension. The normal of the more short term processing variables such as die tempera
cause of change is the polishing that is done to the cavity surface ture. Within a week of steady operation the control limits can be
to remove the heat checked or soldered condition. Sometimes established and the X and R chart will be a working tool. Then is
cavities must be polished daily for such conditions. when the engineer should start watching for wear.

Constructing the Control Chart and LCL) are used on conventional X and R charts to show
when the average of a sample has deviated from some “norm”
The wear of the die cavity will make the dimension get either sufficiently to warrant corrective action. Those limits (i.e. UCL
larger or smaller depending on how the casting features are and LCL) do not imply that parts are out of specification. The
situated in the die24. If the dimension is to become smaller as size limits (USL and LSL) actually suggest that parts will be
material is “worn” off the cavity surface, a steady long term produced outside of specification when the size limit line on
reduction can be expected of the X value for the X control the chart crosses outside the part tolerance.
chart. The concept is illustrated in Fig. 11-1. However, the con
trol limits which actually reflect on other processing variables The upper size limit is:
should not change in their relationship to the average X. So, as USL = X + 3σ = X +3R/d2
shown in Fig. 11-1, both the average, X and the control limits
would slope downward. The control limits would slope upward And the lower size limit is:
if the long term wear caused the dimension to get larger. These LSL = X – 3σ = X -3R/d2
long term wear conditions should not be confused with the
process being statistically “in” or “out” of control. Where d, can be extracted from Table 3-2.
Long Term Trends

Deterioration due to wear will usually progress very slowly,
taking many weeks to become great enough to detect from
the SPC data. To detect such long range trends, the SPC data
must be reduced to fewer data points and to show less short
term variation. One technique is to use only the data from
the subgroups collected during one day (say Monday) of each
week and to average all the X’s for that day. In that way, there
is only one performance statistic for each week. Such statistics
are shown in Table 11-1 as X values and are plotted in Fig.
11-2 for a hypothetical example. The plotted data shows an
obvious trend to smaller dimensions over the 25 weeks.
Fig. 11-1. Cavity wear will cause a long term drift in the average, X, of the X Week N Date X in.
and R SPC control chart. If the rate of drift (slope) can be determined, and 1 I/7/85 10.023
the control limits superimposed with that same drift as shown here, the future 2 1/14/85 10.028
date at which the process will cause out-of-tolerance parts can be predicted. 3 1/21/85 10.021
4 1/28/85* 10.026
The task is to construct a control chart similar to an X control 5 2/4/85 10.025
chart and based on the same data as the X and R chart, but that 6 2/11/85 10.024
actually measures the long term drift from wear. Such a chart 7 2/18/85 10.027
would have the general form of the illustration in Fig. 11 -1. In 8 2/25/85 10.020
Chapter 3, it was shown that the control limits of the X and R 9 3/4/85 10.026
chart were based exclusively on the process performance and had 10 3/11/85 10.024
no relationship to part print tolerances. It was assumed that the 11 3/18/85 10.020
part print tolerances must have a wider range than the X control 12 3/25/85 10.019
limits or the process was not feasible for the part. The control 13 4/1/85 10.025
chart shown in Fig. 11-1 has two additional features. One is
14 4/8/85 10.020
to show the part print tolerance, and the other is to show the
15 4/15/85 10.018
expected size range of the parts. The purpose of the control chart
is to predict when the expected part size, LSL, will intersect the 16 4/22/85 10.023
part print tolerance (Point Z in Fig. 11-1). By predicting point 17 4/29/85 10.021
Z it is possible to schedule the repair of the die. Without such-a 18 5/6/85 10.021
prediction, the first indication of a problem could be the custom 19 5/13/85 10.022
er’s rejection of parts. Then, there may not be time to do a proper 20 5/20/85 10.018
repair and still meet shipping commitments nor will there be a 21 5/27/85 10,017
clear indicator that die wear was the true cause. 22 6/3/85 10.018
23 6/10/85 10.022
24 6/17/85 10.020
Upper and Lower Size Limits 25 6/24/85 10.017
The addition of the Upper Size Limit (USL) and Lower Size Table 11-1. Hypothetical values of X’s for an X and R SPC control chart for a
Limit (LSL) to the control chart is a significant change from 10.00 in. dimension are listed. The subgroup samples are assumed to be for each
conventional charting practice. The X control limits (UCL Monday and were themselves averaged to get the single X value for each day.

When one averages many data points as suggested here, two has a coefficient of thermal expansion of 12.1 x 10 -6/°F (21.8
things happen to change the calculated positions of the UCL x 10 -6°C). So, for the 10 in. (254 mm) dimension, the expan
and LCL. First, there are many more data points, so the range sion factor would be:
of sizes will be considerably greater. But, the confidence in any
calculation made with the greater number of measurements will (10.015 in.)(12.1x10-6/ °F) = 0.00012 in./ °F
be much greater. These factors are compensated for the by selec or
tion of A2 (for this larger sample) from Table 3-2. Typically, the
UCL and LCL will be closer to the X line for the larger sample (254.4mm)(21.8x10-6°C) = 0.00554 mm/ °C
size, and their calculation should be based on the total number
of measurements included in the average, X, calculation. If a particular casting had an ejection temperature of 825°F
(440.6°C) and measured 10.013 in. (254.3 mm), the measured
size must be increased by the amount of additional shrinkage
Compensating Factors caused by the extra 20°F (11.1°C) of unrestrained cooling1.
A graph, such as in Fig. 11-2, might not be showing actual Therefore:
deterioration of the die cavity as it appears. Casting size is Normalized size = 10.013 in.
also sensitive to other processing variables, mainly cast + (20°F)(0.00012in./°F)
ing ejection temperature1. To validate the data, the casting = 10.0154 in.
ejection temperature must also be measured for each casting
that is measured for size. If the ejection temperature can not or
be measured directly (see Chapter 9), it must be estimated
from die temperature, solidification time and metal injec = 254.3 mm
tion temperature. Then each dimension measured must be + (11.1°C)(0.00554 mm/°C)
normalized to account for the different size that its ejection = 254.4mm
temperature would have created.
The normalized size must be used for the data that makes up
For example, if the casting ejection temperature for a 380 the graph in Fig. 11-1.
aluminum casting was found to average 805 deg. F during
the first week of operation, that temperature could be used A long term size drift as shown in Fig. 11-2 caused by
for normalizing the sizes. Also, assume that the cavity increased casting ejection temperature instead of die wear
feature had an average size of 10.015 in. (254.4 mm) on the is a very natural thing. Most efforts (conscious and uncon
room temperature castings during that week. The 380 alloy scious) to improve productivity result in higher average
Fig. 11-2. The averages

of large subgroups
taken infrequently are
plotted from Table
11-1 to show long
term trends of cavity
wear. The statistics
used to show such long
term trends must be
calculated to minimize
the “masking” effects
of short term variations
caused by other pro
cessing variables.

ejection temperatures. So, as such efforts are implemented Then, substituting the value of b into equation 11-1 and solving
over time, average ejection temperature will increase and X N for N = 1 and N = 25 gives:
casting size will decrease.
An alternative to normalizing the data as described above is X1 = X + bN-13

to reset all machine settings to their original conditions and
allow the process to stabilize before castings are selected for X1 = 10.0218 in. + (-0.0002915 in./wks.)(-13 wks.)
die wear analysis. Normalizing the data is the easiest. Either =254.55mm + (-0.0074mm/wks.)(-13 wks.)
way, the investigator must assure himself that the data such = 10.0256 in. = 254.65mm
as is shown in Fig. 11-2 used to construct his control chart are
actually reflecting die wear. X 25 = X + bN-13
X 25 = 10.0218 in. + (-0.0002915 in./wks.)(13 wks.)
Computing the Slope = 10.0180 in.
Although there is an obvious trend to the data plotted in Fig. or

11-2, the trend is not quantified. A common method of quanti = 254.55 mm + (-0.0074 mm/wks.)(13 wks.)
fying such trends is to calculate a least squares best fit straight = 254.46mm
line through the data. The equation of such a line has the form:
So the best fit line will be at 10.0256 in. (254.65 mm) for the
X N = X + bN (11-1) first week (1/7/85) and at 10.0180 in. (254.46 mm) for the 25th
weeks (6/24/85) as shown in Fig. 11-3.
Where:
The next step is to add the upper and lower control limits
X N = The expected value of the subgroup dimensions (i.e. (UCL and LCL) and the upper and lower size limits and the
X) after N weeks. least squares best fit line is obtained, as described previously,
X = The average value of all data points. and as shown in Fig. 11-3.
b = Slope of the best fit line, in./wk.

Control Limit Offset
N = Number of weeks from the N. Plus or minus integers
weeks (N is the week at which X occurs) The equation for the control limit is derived from equation
11-1, and the control limit equations from Chapter 3.
XUCL N = (X + A 2R) + bN (11-3)

The first step is to calculate the average dimensions value, X,
and the average number of weeks, N. For the example in Table X LCL N = (X - A 2R) + bN (11-4)
11-1 and Fig. 11-2:
X = 10.0218 in. XUSL N = (X + 3σ) + bN (11-5)
N = 13 wks. X LSL N = (X - 3σ) + bN (11-6)
The second step is to compute the slope, b, of the best fit line. Where:
N (11-2) XUCL N = The value of the upper control limit at N weeks
Σ (N - N)(X - X)
i i
from N.
i=1
X LCL N = The value of the lower control limit at N weeks
b =
N from N.
Σ (N - N)
i
2
XUSL N = The value of the upper size limit at N weeks
i=1
from N.
And for the example:
X LSL N = The value of the lower size limit at N weeks
b = (-0.3790in.-wk.)/ 1300 wk.2
from N.
= -2.915 x 10-4 in./ wk.
= (-9.63 mm-wk.)/1300 wk.2 A2R = The control limit spacing from X in the X and R
= -74.04 x l0-4 mm/wk. chart as developed in Chapter 3.

Fig. 11 -3. The long
term trend of die
cavity wear is quanti
fied by calculating
and plotting the least
squares best fit line.
The control limits can
be added parallel to
that line and projected
to predict when the
wear will cause out-of-
tolerance parts.
If the value of 3σ is 0.0045 in. for the ongoing example, the Control Decision
equation of the lower control limit (the lower control limit is
The only control decision is to stop production and repair the
the critical concern in the example) is:
die. The question is when to do it. Once the slope of the wear
X LCL N = (10.0218 in. - 0.0045 in.) trend has been established and the control limits determined,
- (0.0002915 in./wks.) (N) the date at which out of tolerance parts will be produced can
= 10.0173 in.-(0.0002915 in./wks.) (N) be predicted. The prediction is actually when the appropri
ate control limit will intersect the part print tolerance limit as
Or illustrated by point Z in Fig. 11-1.
=(254.55 mm - 0.11 mm)
- (0.0074 mm/wk) (N) So, for the example, 59 weeks from April 1,1985, the die is
expected to be worn so badly that it will start making out of
= 254.44 mm-(0.0074mm/wk) (N) tolerance parts. Assume that a new cavity will be required, and
it will take 12 weeks to build and four weeks to install. The
new cavity should be started 43 weeks (59-16) after 4/1/ 85.
Finding the Critical Point Also, a four week supply of parts must be run and inventoried
The critical point Z (Fig. 11-1) is where the lower control limit by 55 weeks (59-4) from 4/1/85.
XLCL N is equal to the part print tolerance. If the part print
tolerance is 10.000 in., equation 11-5 can be written as follows These types of predictions can remove much of the emotional
to find the number of weeks, Nz, between N and Z. trauma from die casting.
10.000 in. = 10.0183 in. – (0.0002915 in./wk.)(NZ)

Solving for NZ gives:
NZ = (10.0173-10.000) in./0.0002915 in./wk.
= 59 weeks

CH APTER
Implementation
Many devices and techniques for controlling the die casting However, once it is decided that implementation of the more
process have been described in this book. Other NADCA publi advanced technology is necessary, the company’s management
cations and courses offer in-depth analyses of the inter-relation must take the second and more positive step. The second step
ships between the process variables controlled by such devices consists of defining the present level of technology existing
and techniques and the influence of those variables on casting in the plant, and defining the goals to be achieved by specific
quality and productivity. The technology detailed in those books dates. Such goals might take the form shown in Table 12-1 of
and programs is useful only to the extent that it is implemented a hypothetical situation.
into production die casting operations.
The third step consists of specifying, ordering, preparing cost
The first step in the implementation procedure is to decide estimates, and doing all the other things associated with the
if implementation of all or part of the control technol first set-up. The goals such as those shown in Table 12-1, can
ogy is necessary or desirable. If the incorporation of more be established at the management level even if an outside
advanced control technology (than that currently used) will consultant’s services are necessary. But, the third step, as well
not improve the productivity of the die casting operation as all the subsequent steps, require trained personnel in the
by an amount greater than the cost of implementation, then plant. The person, or persons, responsible for this third step
the advanced technology should not be introduced. Also, must be bona-fide engineers, and even then must be thor
the short and long range plans of the company may negate oughly trained in the total technology of both the die casting
the need for introduction of a more advanced die casting process and the control devices to be used.
technology. For example, those plans could call for a total
product design change, discontinuation of the product line, The fourth step is the general indoctrination of all respon
or even bankruptcy or sale. In those situations, there is no sible plant personnel. The success of the process control
point in expending a large effort to improve die casting devices depends upon their utilization by the people who
technology in the plant. are engaged in operating the equipment. Everyone on the
Type of Control Finalization Installation Installation

of Equipment on 25% on 50%
Standards and of all Machines of all Machines
exact Form of
Control
Position/Velocity Transducers on Shot Plungers 1/1/00 6/1/00 1/1/01
Multi-Point Plunger Speed and Pressure Readout with
1/1/00 6/1/00 1/1/01
Portable Performance Analyzer
Feedback Control on Shot Cylinder 6/1/01 1/1/02 6/1/02
Tie Bar Stress Monitor 1/1/00 6/1/00 1/1/01
Solenoid Valve Waterline Control 1/1/00 6/1/00 1/1/01
Continuous Monitoring with Central Computer 6/1/03 1/1/04 6/1/04
Process Control From Central Computer 6/1/04 1/1/05 6/1/05
Production Reporting 1/1/04 6/1/04 6/1/04
Table 12-1. A listing of the process control implementation goals with a timetable is an important step. From such a plan costs can be established and progress
can be evaluated.

“team” must be introduced to the equipment, its philosophy, The most difficult problem is the maintenance of a high level
and the details of their specific duties or role. of process control. Little by little, one thing after another will
slip out of control. The effect of each loss may be imperceptible.
The fifth step is to evaluate the results of the first phase of But, collectively, and after a period of time, the gains from the
installation. These results may indicate that changes should process controls will be lost. For this reason, the information
be made to the basic plan, Table 12-1. Also, the experi systems to get details of daily performance (and evaluation),
ence gained should result in formalization of the exact (and back to the line foreman and the machine operators, may be
standardized) type, installation, and procedures that are to more critical than the technology. Standard “labor efficiency”
be used. This step should also include the establishment of techniques are definitely not adequate to the task.
information feed-back control systems to insure the continu
ing function of the devices and learned technology. Process control is an operational philosophy. It includes
operating procedures and the general education of the
The final step is to install the devices and incorporate the oper responsible people. Installation of clever instrumentation
ating procedures on the die casting machines as planned. and “black boxes” is not sufficient.

CH APTER
Process Potential
Many devices and instruments have been described in this ings can be greatly reduced 5 through control of the casting
book along with the operating principles by which these process. And, the control of internal soundness has been the
devices can be used to control the die casting process. The subject of many research projects and papers.
process technology set forth is sufficient for many degrees of
control, which may be classified as: If the critical variables are continuously monitored, potentially
Manual defective castings could theoretically be detected before the die
is open. By implementing the proper measuring devices and
Periodic Checks readout or alarm controllers, the “good” castings could be certi
fied to conform to the specifications. The required measure
Continuous Display and/or alarm ments and control decisions would depend on the particular
Computer control casting and its specifications. The limits of the measured values
could not be pre-calculated in all situations. There usually
would be a try-out period to establish the exact relationship
As any particular installation is advanced from one level of
between casting quality and the process variables.
control to the next, there will be an improvement in total
useable pieces per shift, less scrap, better surface finish,
The magnitude of the process potential for exploiting markets
higher dimensional repeatability, and better internal integ
not generally serviced is great. Now, the means is available
rity of castings. It has been demonstrated 30 that die castings
for extracting the required performance from the process.
can be produced to tolerance ranges of one tenth of those
Together, these two situations present the greatest challenge
commercially produced. Likewise, tests have shown that
ever faced by the die casting industry.
corrosion failure risks of chromium-plated zinc die cast

References
1. E. A. Herman, Dimensional Repeatability of the Die Casting Process, The 17. A. J. Davis, “Effects of the Relationship Between Molten Metal
Society of Die Casting Engineers, Inc., 1982. Flow is Feed Systems and Hydraulic Fluid Flow in Die Casting
Machines,” Transactions, Paper G-T75-124, The Society of Die Cast
2. P. Thukkarum, H. H. Pokorny, Gating Die Casting Die, ing Engineers, Inc., 1975.
The Society of Die Casting Engineers, Inc., 1977.
18. E. A. Herman, “Gating System Design: Now a Simple Science,” Die Cast-
3. E. A. Herman, Heat Flow in the Die Casting Process,
ing Engineer, The Society of Die Casting Engineers, Inc., 1973.
19. P. D. Frost, ILZRO Project ZM-132A Die Casting Process Improvement
4. R. W. Bruner, Metallurgy of Die Casting Alloys,
Progress Report No. 2 (For period January 1 to June 30, 1970), International
Lead Zinc Research Organization, Inc., 1970.
5. E. A. Herman, “The Influence and Control of the Neutral Thermal
20. D. L. LaVelle, “Metallurgy and Production of Aluminum and Zinc Cast
Axis in Die Casting,” Transactions, Paper 54, The Society of Die Cast
ings,” Plating. V. 47, November 2, 1960.
ing Engineers, Inc., 1968.
21. A. L. Bryant, “The General Motors Acurad Die Casting Process,” Trans-
6. Leonard J. Pomplas, “Digital Serve Control of Shot Plunger Velocity
actions, The Society of Die Casting Engineers, Inc., 1966.
and Acceleration,” Transactions, Paper 1572, The Society of Die Casting
Engineers, Inc., 1972. 22. S. D. Sanders, R. C. Bruton, C. W. Rodmann, W. D. Kaiser and P. D.
7. R. J. Larkin, R. K. Beck & K. M. Debe, “Automatic Control of Die Tem Frost, “Die casting Process Instrumentation Study,” Transactions, The
perature in Zinc Die Casting,” Transactions, Paper 62, The Society of Die Society of Die Casting Engineers, Inc., 1970.
Casting Engineers, Inc., 1970.
23. New Technology for Die Casting Zinc,
8. B. K. Dent, R. Fifer, “Product Operation with the ILZRO – Battlle Die BNF Metals Technology Center for ILZRO, 1972
Temperature Controller,” Transactions, Paper 5572, The Society of Die
24. E. A. Herman, Die Casting Dies: Designing,
Casting Engineers, Inc., 1972.
9. E. A. Herman, “The Eight Control Areas of Die Casting,” Die
Casting Engineer, November-December, The Society of Die Casting 25. P. J. Hayes and G. F. Miller, “Evaluation Procedure for Die Lubricants,”
Engineers, Inc., 1972. Transactions, The Society of Die Casting Engineers, Inc., 1981.
10. A. H. Bowker, G. J. Lieberman, Engineering Statistics, 26. C. Kim, G. Ruhlandt, “Computer Modeling of Thermal Stress in Die Cast
Prentice-Hall, Inc., 1960. ing Dies,” Transactions, The Society of Die Casting Engineers, Inc., 1985.
11. E. L. Grant, Statistical Quality Control, 27. C. W. Nelson, “Nature of Heat Transfer at the Die Face,” Transactions, The
McGraw-Hill Book Company, Inc., 1952. Society of Die Casting Engineers, Inc., 1970.
12. E. A. Herman, ed., Die Casting Handbook, 28. S. J. Noesen & H. A. Williams, “The Thermal Fatigue of Die Casting
The Society of Die Casting Engineers, Inc., 1982. Dies,” Transactions, The Society of Die Casting Engineers, Inc., 1975.
13. E. A. Herman, “Instruction Manual,” Computer Gating Program, The 29. A. P. Peterson, “Thermocycling Control of Aluminum Die Casting
Society of Die Casting Engineers, Inc., 1983. Machines,” Transactions, The Society of Die Casting Engineers, Inc.,, 1975.
14. D. Lindsey and J. F. Wallace, “Heat and Fluid Flow in the Die Casting 30. D. G. Gill, W. Babington, “Factors in the Die Casting Process Affecting
Process,” 1968 Transactions, Paper 12. The Society of Die Casting Engi Product Dimensions,” Die Casting Engineer, The Society of Die Casting
neers, Inc., 1968.
Engineers, Inc., November-December, 1967.
15. Temperature Measurement Handbook,
31. D. Apelian, M.M. Makhlouf, High Integrity Aluminum Die Casting,
Omega Engineering, Inc., 1983.
The North American Die Casting Association, 2004.
16. L. W. Garber, “Theoretical Analysis and Experimental Observation of
Air Entrapment During Cold Chamber Filling,” Die Casting Engineer,
May-June, The Society of Die Casting Engineers, Inc., 1982.

NORTH AMERICAN DIE CASTING ASSOCIATION
3250 N. Arlington Heights Rd., Ste. 101

Arlington Heights, IL 60004
P: 847.279.0001 F: 847.279.0002
www.diecasting.org

E410 PDF

Uploaded by

Copyright:

Available Formats

E410 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

E410 PDF

Uploaded by

Copyright:

Available Formats

NORTH AMERICAN DIE CASTING ASSOCIATION

R/3 R/3 R/3

First Printing 1988

Chapter 1: Introduction to Control Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Chapter 2: Measuring Process Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Chapter 3: Statistical Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

Chapter 4: Holding Furnace Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Chapter 5: Plunger Velocity & Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Chapter 6: Clamping Force . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Chapter 7: Die Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

Chapter 8: Release Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

Chapter 9: Casting Ejection Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Chapter 10: Cycle Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

Chapter 11: Mechanical Die System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

Chapter 12: Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Chapter 13: Process Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Die Casting Process, Engineering and Control

Introduction to Control Theory

Generally a process is considered to be out of control when the

Die Casting Process, Engineering and Control 5

More typical of die casting variables is cyclic drift. Die tem­

A threshold condition is when a cyclic or continuously drift­

6 Die Casting Process, Engineering and Control

• Statistical Process Control (SPC) 4. Tie Bar Loading

The Die Casting Variables

2. Holding Furnace Temperature

Die Casting Process, Engineering and Control 7

Measuring Process Capability

Die Casting Process, Engineering and Control 9

10 Die Casting Process, Engineering and Control

10 Cells for 75 or fewer measurements

98F°/10 cells = 9.8°F/cell

Die Casting Process, Engineering and Control 11

Fig. 2-2. Some process variables may not

12 Die Casting Process, Engineering and Control

Time Graph Computing the Statistics

Die Casting Process, Engineering and Control 13

Measurement Difference Difference 334 -15 225

14 Die Casting Process, Engineering and Control

The relative effectiveness of various process controlling

Die Casting Process, Engineering and Control 15

Statistical Process Control

Die Casting Process, Engineering and Control 17

Computing X and R surements and insuring “normal” operation of the process

18 Die Casting Process, Engineering and Control

Die Casting Process, Engineering and Control 19

20 Die Casting Process, Engineering and Control

Holding Furnace Temperature

Fig. 4-1. The

Die Casting Process, Engineering and Control 22

23 Die Casting Process, Engineering and Control

Die Casting Process, Engineering and Control 24

Plunger Velocity & Force

Die Casting Process, Engineering and Control 26

Fig. 5-2. The calculated

27 Die Casting Process, Engineering and Control

Fig. 5-3. The veloc­

Die Casting Process, Engineering and Control 28

Fig. 5-5. The screen

More typical of die casting variables is cyclic drift. Die tem

A threshold condition is when a cyclic or continuously drift

Fig. 5-3. The veloc

Fig. 5-7. The PQ2 dia

Fig. 6-2. The area rep

Fig. 9-1. The tempera

XUCL N = (X + A 2R) + bN (11-3)

N = 13 wks. X LSL N = (X - 3σ) + bN (11-6)