Manufacturing Frontiers

National Science Foundation

Directorate for Engineering

Adnan Akay

Manufacturing Contribution to
US GDP

Source: National Association of Manufacturers,

U.S. Department of Commerce
 Manufacturing Contribution to
US GDP and Employment
Data Source: US Dept of Labor, NAM GDP
calculations using 1982 constant-weighted
price index

Issues & Drivers

!  Competition – emerging economies

!  Enabling technologies
!  Environment, sustainability, resource issues
!  Socio-economics
!  Regulations

EU Manufuture!
 Response

!  New products and service with high added

values
!  New business models
!  Emerging manufacturing sciences and
engineering

EU Manufuture!

Manufacturing Research:
Example Directions

(P.Gouma-SUNY SB)!
!  Improving decision making (tolerancing, fixturing,
tool path optimization)

!  New processes (nanomanufacturing, lithography,

solid freeform fabrication)

!  Metrology and process monitoring (nanometrology)

!  Predictive modeling (incorporation of uncertainty)

(J. Lee-Ohio State U)! (S. Girshik-U. Minnesota)!

Challenge: Manufacturing Across Scales

Manufacturing Miniaturization Trends

Macro!Meso!Micro!!
Nano-Manufacturing!
 ME325/580 Handout: Engineering Materials

Spring 2010 I. Kao

Descriptions and comparisons of the three basic categories of engineering materials (metals,
ceramics, and polymers) and their mechanical, physical, and other properties

Metals Ceramics Polymers

Ferrous and non-ferrous Compound containing Compound formed of
metals metallic (or semi-metallic) repeating structural units
Description and nonmetallic elements called “mers” whose atoms
share electrons to form very
large structure
Structure Crystalline (solid state) Crystalline or non-crystalline Glassy or
(amorphous; e.g., glass, SiO2) Glassy + Crystalline
Mechanical Strong, hard, ductile (esp. High hardness Strength & stiffness vary
Properties FCC) High stiffness
High brittleness
Physical High electrical conductivity Electrical insulation Low density
Properties High thermal conductivity Thermally resistant High electric resistivity
(refractoriness) Low thermal conduction
Other Opaqueness Chemical inertness Carbon +(H2,N2,O2,Cl2)
Properties Reflectivity

In addition to the basic three categories

described above, the composite materials are
typically combinations of two of the three
engineering materials, as illustrated in the
diagram (also called the Venn diagram).
 MEC325/580 Handout: Viscosity for Industrial Slurry

I. Kao Spring 2010

Typical viscosity for slurry used in industrial processes

The slurry used here consists of glycol carrier with silicon carbide grits
at F400 grain size (average grain size is 17 microns). The mixing ratio
is in kilogram of silicon carbide grits vs. liter of carrier fluid. The
viscosity as a function of temperature also resembles the shape of curve
shown here.

20000
ratio=1.25
10000

5000
ratio=0.5
2000
ratio=1.0 ratio=0.75
1000

500

200

100
0 1 2 3 4 5 6 7
Shear rate (1/s)
 MEC 325/580 Handout: Geometric Dimensioning and Tolerancing

Spring 2010 I. Kao

Geometric Dimensioning and Tolerancing (GD&T) is a language for communicating engineering

design specifications. GD&T includes all the symbols, definitions, mathematical formulae, and
application rules necessary to embody a variable engineering language. It conveys both the
nominal dimensions (ideal geometry), and the tolerance for a part. It is now the predominant
language used worldwide as well as the standard language approved by the American Society of
Mechanical Engineers (ASME), the American National Standards Institute (ANSI), and the
United States Department of Defense (DoD). GD&T is the language that designers should use to
translate design requirements into measurable specifications.

The following American National Standards define GD&T’s vocabulary and provide its
grammatical rules.

• ASME Y14.5M-1994, Dimensioning and Tolerancing

• ASME Y14.5.1M-1994, Mathematical Definition of Dimensioning and Tolerancing
Principles
• ASME Y14.41-2003, Digital Product Definition Data Practices

These are often referred to as the “Y14.5” and “the Math Standards,” respectively.

Usage of GD&T and why do we use GD&T

The following drawing is an example for the identification of hole location.

Figure 1: Drawing showing distance to ideal hole location

1
A drawing which does not use GD&T (Figure 2) can be potentially misunderstood and fabricated
incorrectly (see Figure 3 for the illustration).

Figure 2: Drawing that does not use GD&T

Figure 3: Manufactured part that conforms to the drawing without GD&T in Figure 2

2
GD&T provides unique, unambiguous meaning for each control, precluding each person’s
having his own competing interpretation. GD&T is simply a means of controlling surfaces more
precisely and unambiguously. See Figure 4 for an illustration.

Figure 4: Drawing that uses GD&T with unique and unambiguous interpretation

More information and a list of symbols of GD&T can be found in the reference [1].

ASME Y14.41-2003 Standard for CAD

ASME Y14.41-2003 standard is an extension of the Y14.5 standard for 2-dimensional drawings
to 3D computer-aided design (CAD) environments. The standard also provides a guide for CAD
software developers working on improved modeling and annotation practices for the engineering
community. ASME Y14.41 sets forth the requirements for tolerances, dimensional data, and
other annotations, and advances the capabilities of Y14.5. Y14.41 defines the exceptions as well
as additional requirements to existing ASME standards for using product definition data or
drawings in 3-D digital format. [2]

The standard is separated into 3 industrial practices: (i) Models Only: These portions cover the
practices, requirements, and interpretation of the CAD data when there is no engineering
drawing. While ASME Y14.41-2003 is commonly called the “solid model standard,” this is
misleading. The standard was intentionally written for different user levels; (ii) Models and
Drawing: These portions cover what is commonly called “reduced content drawings” or

3
“minimally dimensioned drawings,” where an engineering drawing is available, but does not
contain all the necessary information for producing the part or assembly; (iii) Drawings only:
These portions of the standard allow the historical practices of using engineering drawings to
define a product. However, this standard adds to the practices defined in ASME Y14.5 for
Geometric Dimensioning and Tolerancing with some additional symbols, the use of axinometric
views as dimensionable views, and the concept of supplemental geometry–all of which can help
to clarify the drawing and its interpretation. [3]

Part of the materials in this handout have been taken from the following reference.

Reference:
[1] Walter M. Stites and Paul Drake, Jr., “Dimensioning and Tolerancing Handbook,” Editor
Paul J. Drake, Jr., Ch. 5, McGraw-Hill, 1999
[2] ASME Y14.41-2003 Standard on Digital Product Definition Data Practices, ISBN:
0791828107, 2003
[3] Wikipedia, http://en.wikipedia.org/wiki/ASME_Y14.41-2003

4
 ME325/580 Handout: CNC Machining
Spring 2010 I. Kao

Introduction
Computer numerical control (CNC) is the process of manufacturing machined parts in a
production environment, as controlled and allocated by a computerized controller that used
motors to drive each axis. The CNC technology has been one of manufacturing’s major
development in the 20th century. The controller is designed to control the direction, speed, and
length of time each motor rotates. The operator downloads programmed path to the computer
connected to the machine and then executes the code. The idea of Numerical Control (NC) was
conceived by John Parsons, taken up by USAF, in 1948. This term is used interchangeably with
CNC. The CNC technology not only has facilitated the development of new techniques and
achievement of higher production levels but also has helped to increase product quality.

The CNC technology was developed to:

1. increase production
2. improve the quality and accuracy of manufactured parts
3. stabilize manufacturing costs
4. manufacture complex or otherwise impossible jobs
Numerical control was also designed to help produce parts with the following characteristics:
• similar in terms of raw material
• various sizes and geometry
• small- to medium-sized batches
• a sequence of similar steps was used to complete each workpiece
CNC is now a well-established process, especially with the information and computer
technology, compared to the NC technology first demonstrated in 1952. The CNC technology
has the following advantages over the NC technology:
1. Programs can be entered at the machine and stored into memory.
2. Programs are easier to edit, so part programming process design time is reduced.
3. There is greater flexibility in the complexity of parts that can be produced.
4. Three-dimensional geometric models of parts, stored in the computer, can be used to
generate CNC part programs with tool path almost automatically, thus saving manual
programming labor. This is referred to as CAD/CAM integration.
5. Computers can be connected to other computers worldwide through network, thereby
allowing part programs to be transmitted directly to remote CNC machines.

Training is required for the operator of a CNC machine. The CNC machine also requires
maintenance for smooth operations and extended life.
Two very similar standards are generally followed worldwide: the ISO 6983 and the EIA RS274.
ISO (International Standardization Organization) and EIA (Electronic Industries Association)
developed the main standard for CNC, which used simple programming instructions to enable a
machine tool to carry out a particular operation. The flow charts of CNC processing, with and
without computer aided process, are shown below.

1
 Flow of CNC Processing Flow of Computer-Aided CNC Processing

develop part drawing 3D geometric CAD model

decide machine for the part decide machining ops for the part

choose the required tooling choose the required tooling

design machining sequence use CAM to generate CNC program

calculate the coordinates verify/edit via simulator

calculate spindle speed and feedrate download the part program

write the CNC program verify/edit on actual machine

preapre setup sheets and tool lists Run the program to produce part

verify/edit: simulator/machine tool

verify/edit on actual machine

Run the program to produce part

CNC Programming
A CNC program is a sequential list of machining instructions for the CNC machine to execute.
CNC code consists of blocks (also called lines), each of which contains an individual command
for a movement or specific action. There are two major types of CNC codes, or letter addresses,
in any program. The major CNC codes are G-codes and M-codes.

• G-code are preparatory functions, which involve actual tool moves (for example, control
of the machine). These include rapid moves, feed moves, radial feed moves, dwells, and
roughing and profiling cycles.
• M-codes are miscellaneous functions, which include actions necessary for machining, but
not those that are an actual tool movement (for example, auxiliary functions). These

2
 include spindle on and off, tool changes, coolant on and off, program stops, and other
similar related functions.

Other letter addresses or variable are used in the G- and M-codes to make words. Most G-codes
contain a variable, defined by the programmer, for each specific function. Each designation used
in CNC programming is called a letter address. The letters used for programming are as follows:

N Block number – specifies the start of a block

G Preparatory function, as previously explained
X X-axis coordinate
Y Y-axis coordinate
Z Z-axis coordinate
I X-axis location of arc center
J Y-axis location of arc center
K Z-axis location of arc center
S Set the spindle speed
F Assign a feed rate
T Specify tool to be used
M Miscellaneous function, as previously explained

The Three Major Phases of a CNC Program

The three phases of a CNC program can be illustrated with the following sample code.
Phase CNC program Descriptions
% Program start flag (syntax & format are machine-dependent)
Program

:1001 Four digit program number; up to four digits, 0-9999

setup

N5 G90 G20 Use absolute units, and inch programming

N10 M06 T2 Stop for tool change, use Tool #2
N15 M03 S1200 Turn the spindle on CW to 1200 rpm
N20 G00 X1 Y1 Rapid to (X1, Y1) from the origin
N25 Z0.125 Rapid down to Z0.125
Material
removal

N30 G01 Z-0.125 F5 Feed down to Z-0.125 at 5 in./min

N35 G01 X2 Y2 Feed diagonally to (X2, Y2)
N40 G00 Z1 Rapid up to Z1
N45 X0 Y0 Rapid to (X0, Y0)
N50 M05 Turn the spindle off
shut
Sys

N55 M30 End of program

(1) Program setup: This phase is virtually identical in every program. It always begins with
the program start flag (% sign). Note: The actual setup for each CNC machine may differ.
For example, the CNC in our machine shop uses “0100” to start the CNC code. Line two
always has a program number from 0 to 9999. Line three is the first actually numbered.
G90 tells the controller that all distances (X and Z) are absolute, that is, measured from the
origin. G20 instructs the controller that all coordinates are measured in inch units.
(2) Material removal: This phase deals exclusively with the actual cutting feed moves. It
contains all the commands that designate linear or circular feed moves, rapid moves,

3
 canned cycles such as grooving or profiling, or any other function required for that
particular part.
(3) System shutdown: It contains all those G-codes and M-codes that turn off all options that
were turned on in the setup phase. Functions such as coolant and spindle rotation must be
shut off prior to removal of the part from the machine. The shutdown phase also is
virtually identical in every program.

Using A Programming Sheet

Each row in the following program sheet contains all the data required to write one CNC block.

CNC Programming Part Name: Programmed by:

Sheet Machine: Date: Page:
Setup information:
N G/M X Y Z IJK F R S T Others
Seq Code Pos Pos Pos Loc Feed Rad/ret Speed Tool
5 G90,G20
10 M6 2
15 M3 1200
20 G00 0 0
25 0.1
30 G01 -0.1 2
35 G01 1.5

Block Format
Each block of CNC code needs to be entered correctly. The block comprises of different
components which can produce tool moves on the machine. Here is a sample:

N105 G01 X1.0 Y1.0 Z0.125 F5

N105 Block number Shows the current CNC block number
G01 G-code Tells the machine what to do. In this case, a linear feed
move
X1.0 Y1.0 Z0.125 Coordinate Gives the machine an end point for its move. X
designates an X-axis coordinates, Y/Z for Y/Z-axis.
F5 Special function Contains any special function or related parameter. In
this case, a feed rate of 5 in/min is specified.

There are some simple restrictions on CNC blocks, as follows:

• Each may contain only one tool move.
• Each may contain any number of non-tool move G-codes (e.g., G90 G20 for absolute
system and inch system), provided that they do not conflict between each other (for
example, G42 and G43).
• Each may contain only one feedrate per block.
• Each may contain only one specified tool or spindle speed.
• The block numbers should be sequential.
• Both the program start flag and the program number must be independent of all other
commands.
• The data within a block follow the sequence shown in the above sample block.

4
Preparing to Program
Before you start writing a CNC program, you must first prepare to write it. The steps include
(a) Develop an order of operations
(b) Calculate coordinates and complete a coordinate sheet
(c) Choose tooling with clamping devices, and calculate speeds and feeds

Program Zero
Program zero allows the programmer to specify a position from which all other coordinates will
be referenced. Program zero is also called “part zero” or “machine home.” “Program zero” is
particularly important in absolute programming. In incremental programming (where
coordinates are related incrementally), one has a floating program zero that changes all the time.

Tool Motion
There are three types of tool motions used in a CNC machine. They are:
(1) G00: rapid tool move
(2) G01: straight line feed move
(3) G02/03: circular interpolation or arc feed moves
All cycles such as G71 rough turning are either one of these types or a combination of these
types of motion. These motion command are modal. That is, if you program one of these
commands, you do not need to program the same code again until you want to change the type of
tool motion. The command will be in effect until it is changed or turned off.

Using Canned Cycles

Canned cycles combine many standard programming operations and are designed to shorten the
program length, minimize math calculations, and optimize cutting conditions to improve the
production of the machine. Examples of canned cycles on a mill are drilling, boring, spot facing,
tapping, … etc; on a lathe, threading and pattern repeating cycles. On the lathe, canned cycles
are also referred to as multiple repetitive cycles. Canned cycles also facilitate programming.
You should check out the canned cycles that your CNC control offers. Subroutines are also
available on many CNC controllers. You can utilize these routines to make your own canned
cycles.
Tooling
Separate tools are used for roughing and finishing, and tasks such as drilling, slotting, and thread
cutting require specific tools.
Feedrates, spindle speeds and Cutting Fluids
A good surface finish and economical production rates require proper use of spindle speeds and
feed rates, as well as cutting fluids.

5
CNC Milling

The following tables list G-codes and M-codes.

The G-codes, which include preparatory functions, involve actual tool moves. The following
table lists G-codes in CNC milling.

G-Code name Function Op*

G00 Positioning in rapid Modal
G01 Linear interpolation Modal
G02 Circular interpolation (CW) Modal
G03 Circular interpolation (CCW) Modal
G04 Dwell
G17 XY plane Modal
G18 XZ plane Modal
G19 YZ plane Modal
G20/G70 Inch unit system Modal
G21/G71 Metric unit system Modal
G28 Automatic return to reference point
G29 Automatic return from reference point
G40 Cutter compensation cancel Modal
G41 Cutter compensation left Modal
G42 Cutter compensation right Modal
G43 Tool length compensation (Plus) Modal
G44 Tool length compensation (Minus) Modal
G49 Tool length compensation cancel Modal
G80 Cancel canned cycles Modal
G81 Drilling cycle Modal
G82 Counter boring cycle Modal
G83 Deep hole (peck) drilling cycle Modal
G90 Absolute positioning Modal
G91 Incremental positioning Modal
G92 Reposition origin point
G98 Set initial plane default
G99 Return to retract (rapid) plane
Op: When noted as “modal”, it means that the function remain active until cancelled by another G-code.

* Check your CNC machine manuals for G-codes which are not listed here.

6
The M-codes, miscellaneous functions, include actions necessary for machining, but not those
that are actual tool movements. That is, they are auxiliary functions, such as spindle on and off,
tool changes, coolant on and off, program stops, other similar related functions. The following
table lists M-codes in CNC milling.

M-Code Function
M00 Program stop
M01 Optional program stop
M02 Program end
M03 Spindle on clockwise (CW)
M04 Spindle on counterclockwise (CCW)
M05 Spindle stop
M06 Tool change
M08 Coolant on
M09 Coolant off
M10 Clamps on
M11 Clamps off
M30 Program end, reset to start

7
Letter Address Listing

Letter addresses are variables used in the different G-codes and M-codes. Most G-codes contain
a variable, defined by the programmer, for each specific function. Each letter used in
conjunction with G-codes or M-codes is called words. The letter used for programming are as
follows:

Letter address Function

D Diameter offset register number
H Height offset register number
F Assigns a feed rate
G Preparatory function
I X-axis location of arc center
J Y-axis location of arc center
K Z-axis location of arc center
M Miscellaneous function
N Block number (specifies the start of a block)
P Dwell time
R Retract distance used with G81, 82, 83
Radius when used with G02 or G03
S Sets the spindle speed
T Specifies the tool to be used
X X-axis coordinates
Y Y-axis coordinates
Z Z-axis coordinates

8
Appendix: CNC Programming Sheet

CNC Programming Part Name: Programmed by:

Sheet Machine: Date: Page:
Setup information:
N G/M X Y Z IJK F R S T Others
Seq Code Pos Pos Pos Loc Feed Rad/ret Speed Tool
5
10
15
20
25
30
35

9
 MEC325/580 H ANDOUT: CNC T URNING O PERATION AND L ATHE

MEC325/580, Spring 2010 I. Kao

In this handout, CNC programming for turning operations will be presented. The coordinate references are
discussed with an example of code based on the diameter programming reference.

CNC lathe programming: Before writing CNC code for turning operation, it is important to first identify
and calculate the coordinates of points of transition in turning. The machining may include multiple passes.
We will use a CNC lathe program with the finishing cut of the following lathe part, shown in Figure 1, as an
example to illustrate the concept of CNC programming in turning operations.
+X axis
9 8 4

3 2 3
5 4
2
7 6
1

−Z 1
−9 −8 −7 −6 −5 −4 −3 −2 −1 0 1 +Z axis
Program zero
Figure 1: An example of CNC part for the illustration of CNC turning operation. All units are in inches.

First the coordinate frame with a right-hand coordinate system is shown in Figure 1. The conventional
coordinates of parts in turning operation are in the second quadrant with +X and −Z coordinates. Since a
turning operation is always axisymmetric with respect to the Z axis, the profile shown includes point 1 to
point 9, as shown in Figure 1. Note that the entire profile for programming consideration fits in the second
quadrant. The other half in the third quadrant is a mirror image of the profile shown in Figure 1. All X
values are positive, while all Z values are negative.

However, there are two types of programming references to the XZ dimension. The “diameter program-
ming” relates the X-axis to the diameter of the workpiece. The “radius programming” relates the X-axis to
the radius of the workpiece. Although many controller can work in either mode, diameter programming
is the most common and is the default for most CNC lathe. To change the default, one can enable the radius
programming mode.

Based on Figure 1, the coordinates of the diameter programming for points 1 to 9 are identified and listed
in Table 1, using the diameter programming reference. The workpiece is a cylindrical stock of diameter of
8”; tool #1 is the regular right-hand turning tool, and the tool start position is X4, Z3. The CNC code with
comments is in the following for this turning operation of the finishing cut shown in Figure 1.

N5 ____ Code for program start (machine-dependent)

N10 G90 G20 Absolute, inches system
N15 M06 T1 Tool change to Tool #1
N20 M03 S500 spindle CW with speed 500 RPM
N25 G00 X0 Z0.1 M08 Rapid to X0,Z0.1, coolant on (ready to start)

1
 Table 1: Coordinates of the points identified in Figure 1, under the diameter programming reference.
Point # X Z
1 0 0
2 6 −3
3 6 −4
4 5 −4.5
5 5 −5.5
6 4 −5.5
7 4 −6
8 8 −7
9 8 −9

N30 G01 Z0 F0.02 Feed to point 1 at 0.02 in/rev

N35 G03 X6 Z-3 I0 K-3 CCW circular feed to point 2
N40 G01 Z-4 Linear Feed to point 3
N45 X5 Z-4.5 Feed to point 4
N50 Z-5.5 Feed to point 5
N55 X4 Feed to point 6
N60 Z-6 Feed to point 7
N65 X8 Z-7 Feed to point 8
N70 Z-9 Feed to point 9
N75 G00 Z2 M09 Rapid to Z2, coolant off
N80 M05 Spindle off
N85 M02 End program

Note that block 35 uses G03 to do a circular feed in CCW direction to arrive at the destination point (X6,
Z−3), with center of the 90◦ arc at (I0, K−3), using the notation of I,J,K to express the X, Y, Z coordinates
of the center. In block 40, the G-code G01 is used to do linear interpolation. Since this is a “modal” code,
it stays in effect from block 40 to 70 until it is turned off by G00, another modal code, in block 75. At the
end, the M-code is used to turn off coolant, spindle, and to end the program.

2
CNC Turning Machining Example: A part is to be finished using turning operation, as shown in the
following figure. The information about the workpiece and tool is in the following.
Workpiece size: 4" diameter by 5" length
Tool: Tool #1, right-hand turning tool
Tool start position: X4,Z3
9

8
7 6
0.30 radius 5

4 3
2
1 0.60 1.00 2.00 2.40 4.00

0.20
1.10

1.50

2.30
5.00

1. The coordinates of the finished part, as indicated in the v figure, are calculated using the diameter
programming and listed in the following table.
Point # X Z
1 0 0
2 0.6 0
3 1 −0.2
4 1 −1.1
5 2 −1.1
6 2.4 −1.5
7 2.4 −2
8 3 −2.3
9 4 −2.3

2. The following CNC program is written for the finishing pass of the lathe part.

N05____ Start of the program

N10 G90 G20 Absolute, inches system
N15 M06 T1 Tool change to Tool #1
N20 M03 Spindle on CW
N25 G00 X0 Z0.1 M08 Rapid to X0,Z0.1, coolant on (ready to start)
N30 G01 Z0 F0.012 Feed to point 1 at 0.012 in/rev
N35 X0.6 Feed to point 2
N40 X1 Z-0.2 Feed to point 3
N45 Z-1.1 Feed to point 4
N50 X2 Feed to point 5
N55 X2.4 Z-1.5 Feed to point 6
N60 Z-2 Feed to point 7
N65 G02 X3 Z-2.3 I0.3 K0 Cicular feed to point 8
N70 G01 X4 Feed to point 9
N75 G00 Z3 M09 Rapid to Z3, coolant off
N80 M05 Spindle off
N85 M02 End program

3
 CNC Programming Example (Lecture, Spring 2010)

CNC Programming Sheet

CNC Programming Part Name: Programmed by:

Sheet Machine: Date: Page:
Setup information:
N G/M X Y Z IJK F R S T Others
Seq Code Pos Pos Pos Loc Feed Rad/ret Speed Tool
5 (start)
10 G20
15 G00
20 G17
25 G40
30 G49
35 G80
40 G90
45 M06 T30
50 G00
55 G00 X3.0 Y-0.5 Z0.5
60 M03 S2000
65 M08
70 G01 Z-0.25 F5.0
75 G03 X3.0 Y-3.5 I0 J-1.5 F4.0
80 G01 X7.0 F5.0
85 G00 Z0.2
90 G00 X3.0 Y-0.5
95 G01 Z-0.25 F5.0
100 G01 X7.0
105 G02 X7.0 Y-3.5 I0 J-1.5 F4.5
110 G00 Z0.5
115 X4.0 Y-2.0
120 G81 Z-0.5 F8.0 R0.1
125 X6.0
130 G00 Z0.5

1
 N G/M X Y Z IJK F R S T Others
Seq Code Pos Pos Pos Loc Feed Rad/ret Speed Tool
135 M05
140 G28 Z0
145 M09
150 M30

2
 MEC325/580 Handout: Introduction of Rapid Prototyping
Spring 2010 I. Kao

1. What is Rapid Prototyping?

The idea behind Rapid Prototyping (RP) is to have a machine or machines that can create a
desired solid model directly from a computer-aided design (CAD) file without any human
intervention. A 3-D model is decomposed into multiple 2-dimensional layers with the use of
computer software, and a machine will then manufacture the model one layer at a time. When
the layers are sequentially stacked up and connected, a 3D model will emerge at the end of the
fabrication process.
Since its development in the mid 1980’s numerous Rapid Prototyping1 methods have been
devised. One of the first machine that was devised to fabricate rapid prototypes was the stereo
lithography which used laser to solidify 2D features of a solid model layer by layer in a polymer
liquid bath. Many other methods have since come out. Some of the more common methods
include:
• Stereo Lithography (SLA)
• Selective Laser Sintering (SLS)
• Laminated Object Modeling (LOM)
• Fused Deposition Modeling (FDM)

The RP technique was initially designed to make prototypes quickly for the designers to
evaluate the design and to revise as needed. Owing to the typically high cost of making
prototypes and discarding them after initial evaluation to render the final design, the RP methods
feature the advantages of short time to fabricate and low cost for prototyping. Since then, some
machines, especially those use the laser sintering approach, have been revised and designed to
include metal powder metallurgy with bonder (e.g., resin) and baking process to cure metal
solids with near shape and strength. Nevertheless, this type of efforts still cannot replace the
conventional process of machining. From its inception, the RP technique was not meant to
replace the conventional manufacturing, and it appears that it will not in the foreseeable future.
In many cases where actual parts are required for design consideration, RP is better than the
“virtual” prototyping in which computer is used to view and manipulate solids from different
angles to simulate the design. For example, in the design of ink refilling of an ink-jet printer
cartridge in which the cartridge was decapitated and replaced by a new polymer cap, after
replacing the sponges and inks inside the ink compartments, through ultrasonic welding process.
The physical prototype of the cap design is crucial and needed to actually mesh with the ink
cartridge for the inspection of the statistical range of the parameters of cartridges. A virtual
prototype, no matter how sophisticated it may be, just will not do.
In the following, we briefly explain each of the methods for rapid prototyping.

1
In a broader context, sometimes the “free-form machining” (FFM) is used to refer to these types of
forming processes, distinct from the conventional machining or forming processes.

1
1.1 Stereo Lithography (SLA)
The stereo lithography (SLA or STL) process is based on the principle of curing or hardening a
liquid photo polymer into a specific shape. A laser is used to focus on spot-curing the polymer,
providing the necessary energy to polymerization. Based on the Beer-Lambert law, the exposure
decreases exponentially with depth according to the rule
−z / D p
E(z) = E 0 (1)
where E is the exposure in energy per area, E0 is the exposure at the resin surface (z=0), and Dp
is the “penetration depth” at the laser wavelength and is a property of the resin. At the surface
depth, the polymer is sufficiently
€ exposed for it to gel, or
−C d / D p
Ec = E0 (2)
where Ec is the critical threshold exposure and Cd is the cure depth. Thus, the cure depth is given
by the following equation
€ ⎛ E ⎞
Cd = Dp ⎜ 0 ⎟ (3)
⎝ E c ⎠
The cure depth represents the thickness in which the resin has polymerized into a gel, but it does
not have high strength at this state. Thus, the controlling software will slightly overlaps the
cured volumes, but curing under
€ fluorescent lamps is often necessary as a finishing operation.
Stereo lithography has a vat container with a platform on which the part to be fabricated can
be raised or lowered vertically. This vat is filled with a photo-curable liquid acrylate polymer, a
mixture of acrylic monomers, oligomers (polymer intermediates), and a phtoinitiator. A laser,
generating an ultraviolet beam, is then focused along a selected surface area of photopolymer at
surface to lay out the required feature. As these 2D features are laid, the platform is lowered to
expose a fresh layer of liquid ready for the next layer of 2D features. Successive operations will
render the final 3D solid.
Depending on the capacity of the machine, the cost ranges from $100,000 to $500,000, with
cost of liquid polymer at $300 per gallon. The fume released by the liquid polymer during the
fabrication process needs to be vented out for health consideration. One major area of application
for stereo lithography is in the making of molds and dies for casting and injection molding.
1.2 Selective Laser Sintering (SLS)
Selective laser sintering is a process based on the sintering of polymer and metallic powders
selectively into an individual object. Two cylinders are used in the process chamber: (i) part-
build cylinder which is lowered incrementally to where the sintered part is formed, and (ii)
powder-feed cylinder which is raised incrementally to supply powder to the part-build cylinder
through a roller mechanism. A thin layer is first deposited in the part-build cylinder. A laser
beam, guided by the process-control computer is then focused on that layer, tracing and melting
(or for metal, sintering) a particular cross-section, which then quickly solidifies into a solid mass.
This is repeated for layers after layers of 2D slices of features of the solid. At the end, the loose
particles are shaken off and the part recovered.
Typical cost of a SLS machine is about $500,000.

2
1.3 Laminated Object Modeling (LOM)
Lamination implies a laying down of layers that are adhesively bonded to one another.
Laminated Object Modeling (LOM) uses layers of paper or plastic sheets with glue on one side
to produce parts sheet by sheet. The adhesion process can be done by laser beam through
heating, or simply by gluing. The excess materials are removed manually. LOM usually uses
materials of thickness about 0.5 mm (0.02 in.) although materials as thin as 0.05mm (0.002 in.)
have been used.
One example of such system is the SilverScreen+ JP 5 system which uses back-glue papers
and a cutter on a printer setup to cut (or print onto) the papers into series of 2D features, to be
glued together to make 3D solids. This type of systems can be very inexpensive.
1.4 Fused Deposition Modeling (FDM)
In a Fused Deposition Modeling (FDM) process, a platform is used to move in the vertical (Z)
direction to raise or lower the part to be made, and a nozzle assembly with both model and
support materials controlled by a XY table is used to trace and lay out molten polymer filament
through nozzles to make solid parts layer by layer. The support material is constructed as part of
the slicing algorithm to ensure that overhanging features of the solid part is supported throughout
the fabrication process. More details will follow in the next section.

2. Rapid Prototyping Using the FDM 3000

2.1 What is FDM?
The FDM rapid prototyping machines use the Fused Deposition Modeling (FDM) method to
create prototypes. Fused Deposition Modeling is a process whereby the layers of the model are
created by forcing a special material filament through a heating system, causing the material to
melt into a smooth hot molten paste, which is then forced through a delivery nozzle, and emerges
as a thin ribbon of hot paste. The nozzle is guided along the XY plane, depositing the ribbon at
desired positions to form a layer. After completing one layer, the entire model is lowered. The
machine will then deposit another layer on top of the previous one.
Fused Deposition Modeling uses two different materials to manufacture a model; the first
material is called the model material, and is what the final model will consist of, the second
material is called the support material, this material is laid underneath any overhanging parts of
the model so that the model will not collapse during fabrication. The temperature of the P-400
set includes P-400 ABS model material (Tm=270ºF) and P-400 water-soluble support material
(Tm=235ºF) under an envelop temperature of 70ºF.
Each material is delivered through its own nozzle, which are both mounted on the underside
of the of the FDM head. Apart from two nozzles, the head contains two heating elements, two
thermocouples, two motors and one solenoid actuator. The two motors force the two materials
into the two heating elements, and then through the delivery nozzles. The solenoid actuator
lowers the support nozzle below the level of the model nozzle every time support material is
needed.

3
2.1.1 Motion, the Head and the Z-table
The head can only move in the XY-plain. To accomplish this, the head is mounted on sliders for
motion in the X-direction, while the sliders are mounted on rollers for motion in the Y-direction.
Both X and Y motions are cable-driven.
Vertical or Z-directional motion is accomplished with the use of the z-table. The model is
build directly onto the sponge insert in the z-table, thus when the z-table is lowered so will the
model. The z-table is moved with the use of 4 lead screws.
2.1.2 Materials
A range of plastics and waxes can be used to create models. Different materials display different
physical properties, with each material having its own melting and envelope temperatures. The
envelope temperature is the temperature of the air inside the FDM, and is set to the optimal
solidification temperature of the materials. As a result of the required envelope temperature for a
specific material, all model and support materials come as a couple and should not be mixed.
One particularly interesting model-support combination is that of the P400 ABS model and
P400 water-soluble support. The P400 support material can be removed by means of submerging
the finished part in an ultrasonic bath consisting of a heated-water-chemical solution. The
ultrasound breaks down the support and dissolves it away until only the model remains. This
enable very complicated models to be created, models that could not be created by conventional
machinery.
2.1.3 Nozzles (Tips)
The nozzles come in three different combinations pairs. They are classified according to the
diameter of the outlet opening and are as follows, 0.010", 0.012" and 0.016". All the nozzles are
coded with the use of rings around their lower surface. The codes can be found in the user
manuals.
2.2 Steps Required to Create a Prototype on FDM 3000
The block diagram below shows the basic steps required to create a prototype.

Solid Modeling
(I-DEAS or others)
• Create a Solid Model STL File

InsightV34
• Position & Scale STL
• Slice STL File SSL File
• Add Support and Base
• Edit Curves
• Create Roads
SML File

Model FDM 3000

• Fabricate Model

4
2.2.1 Computer Aided Design and the STL file
First of all, a solid model of the prototype must be designed using any solid modeling package
such as I-DEAS or AutoCAD or ProEngineer. Once this model is complete, it must undergo a
process known as tessellation in order to become an STL file (STereoLithography File).
Tessellation is an approximation of the solid model surface; it is accomplished by breaking the
surface of the solid model into hundreds of small, interconnected triangles, each with a normal
vector pointing outward from the solid. Most solid modeling software will automatically create
an STL file if prompted by the user.
Note: Step by step instructions on how to create an STL file using I-DEAS can be found in
the User Manual.
2.2.2 Procedures to fabricate the rapid prototype with your STL file
Note: You are expected to do this part with the assistance of a staff or TA. You should NOT
attempt the following procedures alone!
(a) Start the FDM3000 machine. Check if the temperatures of the Model and Support are set
correct. Incase they are not, then wait till they reach the required temperature.
(b) In the selection display on the machine, select Model and try to load the model. Confirm
that the Model is flowing smooth. Similarly confirm that the support is also flowing
smooth.
(c) On the PC: Go to ‘Start’, then go to ‘Programs’ and then run the ‘InsightV34’ and select
Insight.
(d) Go to ‘File’ menu option. Select ‘Open’ and select ‘STL’ option and open your STL file.
You would see your model on the screen.
(e) Now go to ‘Orient STL’ option in the software menu. Then rotate the model so that you
obtain the orientation in which you want to make the model. Select the orientation such
that the support needed is minimum.
(f) Click (Slice icon).
(g) Click (Support icon).
(h) Click (Toolpath icon).
(i) Click (Build icon).
(j) Click Build icon one more time.

5
 (k) Next, on the FDM machine press the pause button so that the pause light will start
blinking and the door will open. Now you can set the nozzle position at appropriate place
so that your part is made on desired location on the supporting plate. You can set the X
and Y positions it by pressing the appropriate buttons on the panel of the FDM machine.
You can also set Z by first selected Z axis button. Set Z such that the model nozzle just
touches the surface of the fixture plate.
(l) After setting the nozzle at appropriate position close the door of the machine and press
the pause button. The door will get locked and the machine will start making the model.
(m) Wait near the machine till it finished making the part.
(n) After the solid model is made, the door would get unlocked. Then open the door and
remove the model from the machine very carefully. Using the ultrasonic water bath and
appropriate tools to remove the support material from the prototype. Apply proper post-
fabrication processes, if necessary, such as polishing the model by sand paper or painting.
(o) Don’t forget to turn off the machine after you have finished your work.
(p) Ask the TA to inspect and sign you off.

2.2.3 FDM 3000

The SML file can now be inputted into the FDM 3000. The FDM 3000 will follow the Roads
created during the generation of the SML file.
Although a large part of the fabrication is autonomous, the FDM sill requires an operator,
especially during the first layers of a model. Furthermore, the FDM 3000 requires constant
maintenance and cleaning. Due to the nature of Fused Deposition Modeling, the nozzles, head
and envelope of the FDM require thorough cleaning after every model. Calibration and general
lubrication are also commonly required.

6
 MEC325/580 Orthogonal Cutting Model I. Kao

Vc
Vs Tool
: shear plane angle
chip : rake angle
: clearance angle
V chip or relief angle
t0: depth of cut
tc tc tc: chip thickness
r= t0/tc: chip thickness ratio

t0 t0
workpiece
shear plane
shear deformation cutting edge of tool
to form chips

shear deformation approximated

by a series of parallel plates Tool
sliding against one another thickness of plate
to form the chips shear strain A B
AC AD+DC
(Idealized assumption) BD BD C
chips = parallel shear plates
A B

shear plane
D
C
 MEC325/580 Forces in the Orthogonal Cutting Model I. Kao

R = total force acting on chip

R’= force imposed by the work on the chip
(R = R’)
R”= force measured by the dynamometer
Tool Fs: shear force Tool
Fn: normal force to shear
R F : friction force
chip F N : normal force to friction chip
Fc: cutting force
Ft: thrust force

N
Fs Fc

Fn Ft

R’
R”
 ME325/580 Handout: Copper-Nickel Phase Diagram

Spring 2010 I. Kao

Phase diagram is an equilibrium diagram showing the way an alloy forms. The equilibrium
diagram for the binary copper-nickel alloy is illustrated on the next page. For example, “Monel,”
which resists saltwater corrosion, has 67% nickel and 33% copper shown in the diagram and is
used in packaging beverages and foods. It has a range of working temperatures from –100° to
400°F (–75° to 205°C).
The Copper-Nickel phase diagram shown on the next page is the simplest phase diagram. The
illustration of its usage will be presented in the following with an example. The temperature
under consideration is at T=1260°C. Several states are considered and discussed in the following.
(I) At T=1260°C and under equilibrium condition, there are three possible states, as follows.
(1) When the percentage of Ni is smaller than 36%  Liquid state
(2) When the percentage of Ni is larger than 62%  Solid state
(3) When the percentage of Ni is between 36% and 62%  co-existence of Liquid and
Solid
(II) When liquid and solid co-exist, the following “inverse lever rule” should be applied to
determine the percentage of solid (and percentage of liquid) in the co-existing state.
The Inverse Lever Rule is defined to determine the contents of solid and liquid when they co-
exist within specific temperature range. Here, we use an example of 50% Ni and 50% Cu alloy
at the temperature of 1260°C, as shown in the following diagram, for the illustration of this rule.

LC LC (50 − 36)% 14
The % of solid at the state C is: = = = = 53.8% (1)
LS (LC + CS) (62 − 36)% 26

CS CS (62 − 50)% 12
The % of liquid at the state C is: = = = = 46.2% (2)
LS (LC + CS) (62 − 36)% 26
€
Note the ratios formulated in equation (1) and (2). Based on the results in the equations, the
percentage of solid at 1260°C with 50%Ni-50%Cu alloy is 53.8%, and with 46.2% liquid. For
€ both liquid and solid co-exist from 1210°C to 1316°C, as shown
this alloy with 50%Ni-50%Cu,
in the figure. The percentage of solid and liquid can be determined by equations (1) and (2) at
any given temperature within this range for the 50%Ni-50%Cu alloy. The verification of results
is on the next page.
In the preceding analysis, we consider different compositions of Cu and Ni at a constant given
temperature (1260°C). Similarly, we can also consider a given alloying composition and vary
the temperatures to acquire and analyze different states at different temperatures. For example,
with the alloy composition of 67% nickel and 33% copper, as the temperature cools down from
liquid state, it transitions to co-existence of liquid and solid at 2510°F and becomes complete
solid as the temperature cools further to 2320°F. Thereafter, the alloy becomes entirely solid.

1
Verification of the results:
Because the solid and liquid states at 1260°C have their own respective percentage of Ni and Cu,
we can use the information to verify the answers obtained in equations (1) and (2). It can be
obtained from the figure that at 1260°C

the solid state of this alloy has 62% of Ni and 38% of Cu (point S)
the liquid state of this alloy has 36% of Ni and 64% of Cu (point L)

Thus, the total percentage of the Ni is the sum of the that in the solid and liquid states. That is,

total % of Ni = (53.8%) x 62% + (46.2%) x 36% = 50%

This is as expected since the total weight percentage of the Ni element in the alloy remains 50%
and cannot be changed. Similarly, the total percentage of Cu is

total % of Cu = (53.8%) x 38% + (46.2%) x 64% = 50%

as expected.

2
 ME325/580 Handout: Tin-Lead Phase Diagram

Spring 2010 I. Kao

The equilibrium diagrams for binary alloys can assume different shapes. For example, the
copper-nickel alloy is very simple with only three regions, as discussed earlier. The following is
the phase diagram for tin-lead (Sn-Pb) alloy, which has α and β phases as well as combinations
of them with liquid and solid phases. A few important observations are in order.

(1) The freezing point of pure tin (point C) is 232ºC. As the alloying element (or impurity) of
lead is added, the freezing point is decreased (just like the way that salt lowers the freezing
point of water). This trend of lowering freezing point is shown in CB curve. The same is
true for pure lead (point A; 327ºC) and the curve AB with decreasing freezing point.
The point of intersection between the curves AB and BC (i.e., point B) indicates the eutectic
mixture at 61.9% tin (Sn) and 38.1% lead (Pb). The eutectic composition gives rise to the
lowest possible temperature of solidification for the Sn-Pb binary alloy. The eutectic
temperature of this binary alloy is 183ºC.

1
(2) In the two freezing zones ABD and CBF, the alloy is “pasty.” In other words, they contain
both liquid and solid phases. Taking region ABD as an example, crystals of composition
α will be forming whilst the rest of the alloy is still liquid. These crystals will grow into
dendritic structure (similar to the formation of ice crystals in water with pointy tips) until
such time as the alloy solidifies completely. Then the granular structure will consist of α
dendrite cores in a β crystal mix. Because the β crystals form virtually instantaneously as the
alloy temperature drops below 183ºC, they tend to be small. The resulting alloy is
characterized by a coarse grained structure (large α, small β) and tends to be mechanically
weaker and a poorer conductor. It does have its usage, however. A good example is
plumber’s solder (34%Sn+66%Pb). In this case, electrical conductivity is not an issue, and
the extended pasty stage is advantageous for making “wiped” joints.
(3) The tin-lead binary alloy diagram contains a eutectic point, unlike the Cu-Ni alloy which
does not have one. Electronic solder made of 62%Sn and 38%Pb, or called the eutectic
solder, not only has the lowest melting point and reverts from liquid to solid (and vice versa)
virtually instantaneously (point B on the phase diagram).
In addition, it has lowest melting point at 183ºC, which makes it ideal for electronic grade
solder. The electronic grade solder alloys, or eutectic solder, fall in a narrow band ranging
from 60%Sn+40%Pb to 65%Sn+35%Pb.
(4) The crystal structure of eutectic solder alloys consists of fine equally sized grains of α and β
(because they have limited time to grow) with no evidence of potentially strength-reducing
dendritic core. This fine grain structure also maintains a high degree of electrical
conductivity—a characteristic that lends itself to becoming an ideal electronic solder.
(5) The inverse lever rule discussed previously also can be applied in the regions indicated above
based on the same rules used in the Cu-Ni phase diagram.

2
 ME325/580 H ANDOUT: I RON -C ARBON P HASE D IAGRAM

MEC325/580, Spring 2010 I. Kao

Iron-Carbon or Fe-Fe3 C phase diagram: Iron-Carbon phase diagram (also called the iron and iron carbide
phase diagram), as shown in the following figure, is an equilibrium diagram of iron and carbon that is very
useful in dealing with steel and heat treatment.
1800
steel cast iron

1600
1539oC (2802oF)
δ Liquid
1400
1394oC (2541oF) Liquid+graphite
eutectic
Temperature, oC

γ+liquid
4.30%
1252oC

γ
1200
1154oC (2109oF)
2.11% Fe3C
1000

912oC (1674oF) eutectoid γ+graphite

0.77% (solid)
800
α+γ
α 727oC (1341oF)
0.022%
600
α+graphite
(solid)
400
0 1 2 3 4 5 6 6.67
% Carbon (C)

Figure 1: The equilibrium phase diagram of iron and iron-carbon

As shown in Figure 1, at carbon composition of 2.11% the diagram is partitioned into regions of steel
(%C < 2.11%) and cast iron (%C > 2.11%). Within the region of steel, it can be further broken into two
regions, divided by the eutectoid line with carbon composition of 0.77% (some use 0.80% or 0.83%): (i)
hypo-eutectoid (%C < 0.77%), and (ii) hyper-eutectoid (%C > 0.77%).

The Fe-C phase diagram has one eutectic state at 1154◦ C with 4.30% of carbon (some use 1130◦ C and
4.0%) at which the alloy transforms from liquid to γ-austenite and Fe3 C-graphite. In addition, it also has
one eutectoid state at 727◦ C with 0.77% of carbon (some use 723◦ C and 0.80% or 0.83%) at which the
alloy transitions from one solid (γ-austenite) to two solids (α-ferrite and Fe3 C-graphite).

Remarks:
(i) Phases α and δ are both ferrite (BCC). The α-ferrite (or simply ferrite) is stable at room temperature;
whereas the δ-ferrite is only stable at high temperature. Phase γ is austenite (FCC).

1
 (ii) α → γ: The transition for pure iron from BCC to FCC takes place at 912◦ C.

(iii) γ → δ: The transition for pure iron from FCC to BCC takes place at 1394◦ C.

(iv) Pure iron melts at 1539◦ C.

(v) Fe3 C-graphite is cementite, also called carbide, and is hard and brittle.

Figure 2 illustrates the equilibrium cooling of a hypoeutectoid (which means “less than eutectoid” in
Greek) steel alloy and the changes of phases as it cools to different regions of the phase diagram. The mi-
crostructures of the hypoeutectoid alloy are illustrated at different temperatures, as shown in the figure. The
phase changes from γ-austenite to (α-ferrite+γ-austenite) to (α-ferrite+Fe3 C-graphite) as the temperature is
decreased. While at the state with co-existence of α-ferrite and γ-austenite, the percentage of each can be
calculated using the inverse lever rule. In the following, an example is used for illustration.

1000
γ γ

γ γ γ γ + Fe3C
M γ
900
γ
γ γ α

800 α+γ γ γ
Temperature (oC)

α γ γ
700 O

600 Fe3C
} Pearlite

Proeutectoid α
Eutectoid α

500
α + Fe3C

400
0 1.0 2.0
Composition (wt% C)

Figure 2: Hypoeutectoid alloy in the equilibrium phase diagram and the phase change subject to equilibrium
cooling.

Example: Given the Fe-Fe3 C phase diagram in Figure 1, calculate the phases present for a 1020 steel at the
following temperatures:
(a) T = 1600◦ C

(b) T = 1200◦ C

(c) T = 728◦ C

2
 (d) T = 710◦ C

(e) T = 400◦ C

Solution: The 1020 steel has a carbon content of %C = 0.20% at the hypo-eutectoid region to the left of
the eutectoid state. Figure 1 will be employed to perform the following calculation.

(a) At T = 1600◦ C, it is in liquid state.

(b) At T = 1200◦ C, γ-austenite exists as a single-phase state.

(c) At T = 728◦ C, just above the temperature at the eutectoid point (T = 727◦ C, shown in the figure),
two phases exist: α-ferrite and γ-austenite. The percentage can be determined by the inverse lever
rule.
0.77 − 0.20
α-ferrite = = 76.2% (1)
0.77 − 0.022
0.20 − 0.022
γ-austenite = = 23.8% (2)
0.77 − 0.022
A zoom-in view of the eutectoid region is shown in Figure 2, and can be used for the application of
the inverse lever rule as demonstrated in equations (1) and (2).

(d) At T = 710◦ C, just below the temperature at the eutectoid point, a small amount of Fe3 C-graphite
(cementite) will precipitate following the solubility line from 0.022% carbon at 727◦ C to 0.022%
carbon at room temperature. The percentage of α-ferrite and Fe3 C-graphite are:
6.67 − 0.20
α-ferrite = = 97.3% (3)
6.67 − 0.022
0.20 − 0.022
F e3 C-graphite = = 2.7% (4)
6.67 − 0.022

(e) At T = 400◦ C: Similar to Part (d), the percentages are:

6.67 − 0.20
α-ferrite = = 97.0% (5)
6.67 − 0.0
0.20 − 0.0
F e3 C-graphite = = 3.0% (6)
6.67 − 0.0
The results are not very different from those in Part (d), showing a slight increase in the Fe3 C precip-
itation.

3
 MEC325/580 Handout: Stainless Steel
Spring 2010 I. Kao

Supplementary Lecture about Stainless Steel

The corrosion resistance is imparted by the formation of a strong adherent chromium oxide on
the surface of metal. On the other hand, existence of carbon will form chromium carbide which
takes away ability for chromium to form the shielding chromium oxide. When the amount of
atomic chromium in solution exceeds 12%, improved corrosion resistance and outstanding
appearance are achieved. This category forms what has been commonly called the true stainless
steels.
Several classification schemes have been devised to categorize the stainless steel alloys. The
American Iron and Steel Institute (AISI) groups the metals by chemistry and assigns a three-digit
number that identifies the basic family and the particular alloy within that family. The following
table presents the AISI designation scheme for stainless steels and correlates it with the
microstructural families.
Series Alloys Structure Magnetic?
200 chromium, nickel, manganese, or nitrogen austenitic No
300 chromium, nickel austenitic No
400 chromium only ferritic or martensitic Yes
500 low chromium (<12%) martensitic Yes

The material can also be grouped by microstructural families. In general, there are five main
types as will be described in the following, although new stainless steel alloys have been
developed to meet special needs.
(1) Austenitic (200 and 300 series): These steels are generally composed of chromium, nickel,
and manganese in iron. Nickel is an austenite stabilizer, and with sufficient amounts of both
chromium and nickel, it is possible to produce a stainless steel in which austenite is the stable
structure at room temperature. Known as austenitic stainless steels, these alloys may cost
twice as much as the ferritic variety, with the added expense being attributed to the cost of
the alloying nickel and chromium. Manganese and nitrogen are also austenite stabilizers and
may be substituted for some of the nickel to produce a lower-cost, somewhat lower-quality
austenitic stainless steel. Austenitic stainless steels are nonmagnetic and are highly resistant
to corrosion in almost all media except hydrochloric acid and other halide acids and salts.
However, they are susceptible to stress-corrosion cracking. In addition, they may be polished
to a mirror finish and thus combine attractive appearance and corrosion resistance.
Formability is outstanding (characteristic of the FCC crystal structure), and these steel
strengthen considerably when cold worked. The popular 304 alloy, suited for all types of
dairy equipment, brewing industry, citrus and fruit juice industry, dye tanks, pipelines
buckets, dippers, and food processing industry, is also known as the 18-8 because of the
composition of 18% chromium and 8% nickel (18-8 also refers to 302, 303, 305, and 384).
Austenitic stainless steels are hardened by cold-working. They are most ductile of all
stainless steels, so they can be easily formed, although, with increasing cold work, their

1
 formability is reduced. These steels are in a wide variety of applications, such as
kitchenware, fittings, welded construction, lightweight transportation equipment, furnace and
heat-exchanger parts, and components for severe chemical environment.
(2) Ferrite (400 series): These steels have a high chromium content—up to 27%. Chromium is a
ferrite stabilizer, the addition of chromium tending to increase the temperature range over
which ferrite is the stable structure. If sufficient chromium is added to the iron, and carbon is
kept low, an alloy can be produced that is ferrite at all temperatures below solidification.
These alloys are known as ferritic stainless steels. Such ferrite alloys are also the cheapest
type among stainless steels. They are magnetic and have good corrosion resistance.
Ferrite stainless steels possess rather poor ductility or formability because of the BCC crystal
structure (they have lower ductility than austenitic stainless steels), but they are readily
weldable. They are hardened by cold-working and are not heat-treatable. They are generally
used for nonstructural applications such as kitchen equipment and automotive trim.
(3) Martensitic (400 and 500 series): Most martensitic stainless steels do not contain nickel and
are hardenable by heat treatment. Their chromium content may be as much as 18%. If
increased strength is needed, the martensitic stainless steels should be considered. However,
caution should be taken to assure more than 12% chromium in solution. Slow cools may
allow the carbon and chromium to react and form chromium carbides. When this occurs, the
chromium is not available to react with oxygen and form the protective oxide. Thus the
martensitic stainless steels may only be “stainless” when in the martensitic condition (when
the chromium is trapped in atomic solution), and may be susceptible to red rust when
annealed or normalized for machining or fabrication. The martensitic stainless steel cost
about 1.5 times as much as the ferritic alloys, part of being due to the additional heat
treatment, which generally consists of an austenitization, quench, stress relief, and temper.
These stainless steels are magnetic.
Martensitic stainless steels have high strength, hardness, and fatigue resistance, good
ductility, and moderate corrosion resistance. They are typically used for cutlery, surgical
tools, instruments, valves, and springs.
(4) Precipitation-hardening (PH): A fourth and special class of stainless steels is the
precipitation-hardening variety. These steels contain chromium and nickel, along with
copper, aluminum, titanium, or molybdenum. These alloys are basically martensitic or
austenitic types, modified by the addition of alloying elements such as aluminum that permit
age hardening at relatively low temperatures. By adding age hardening to the existing
strengthening mechanisms, these materials are capable of attaining properties such as a 260-
ksi (1790-MPa) yield strength and 265-ksi (1825-MPa) tensile strength with a 2% elongation.

They have good corrosion resistance and ductility, and they have high strength at elevated
temperatures. Their main application is in aircraft and aerospace structural components.

(5) Duplex structure: Duplex stainless steels contain between 21 to 25% chromium and 5 to 7%
nickel and are water quenched from a hot-working temperature that is between 1830 and
1920°F to produce a microstructure that is approximately half ferrite and half austenite. The
structure offers a higher yield strength and greater resistance to stress corrosion cracking than
either then austenitic or ferritic grades. These steels have a mixture of austenite and ferrite.

2
 They have good strength, and they have higher resistance to both corrosion and stress-
corrosion cracking than do the 300 series of austenite steels. Typical applications are in
water-treatment plants and in heat-exchanger components.

(6) Other stainless steels: Still other stainless alloys have been developed to meet special needs.
Ordinary stainless steels are difficult to machine because of their work-hardening properties
and their tendency to seize during cutting. Special free-machining alloys have been produced
within each family, with addition of sulfur or selenium raising the machinability to
approximately that of a medium-carbon steel.

The following tables shows typical alloy compositions, structure, and usage for the first three
families of stainless steels.

TABLE: Typical Composition (in wt. %) of the ferritic, martensitic, and austenitic Stainless Steels
Element Ferritic Martensitic Austenitic
Carbon 0.08-0.20 0.15-1.2 0.03-0.25
Chromium 11-27 11.5-18 16-26
Manganese 1-1.5 1 2 (5.5-10)
Molybdenum some cases
Nickel 3.5-22
Phosphorus and sulfur Normal (0)
Silicon 1 1 1-2 (0)
Titanium Some cases

TABLE: Popular alloys structures and AISI designation for three primary types of stainless steel
AISI Type
Martensitic 410, 420, 440C
(Hardenable by heat treatment )
Ferritic 405, 430, 446
(More corrosion resistant than martensitic,
but not hardenable by heat treatment)
Austenitic 201, 202, 301, 302,
(best corrosion resistance, but hardenable 302B, 304L, 310,
only by cold working) 316, 321

TABLE: Key purpose and usage for different stainless steel alloys
Purpose and Usage AISI Types
General purpose 410, 430, 202, 302
Automobile parts 301, 409, 430, 434
Hardenable by heat treatment 410, 420, 440C
Hardenable by cold working 201, 301
For elevated-temperature service 446, 302B, 310

3
 Modified for welding 405, 304L, 321
Superior corrosion resistance 316
Catalytic converters 409

Remarks:
(1) Sensitization: Problems with stainless steels are often due to the loss of corrosion resistance
(sensitization) when the amount of chromium in solution drops below 12%. Since chromium
depletion is usually caused by the formation of chromium carbides along grain boundaries,
and these carbides form at elevated temperatures, various means have been developed to
prevent their formation. One approach is to keep the carbon content of stainless steels as low
as possible, usually less than 0.10%. Another method is to tie up the carbon with small
amounts of “stabilizing” elements, such as titanium or niobium, that have a stronger affinity
for carbon than does chromium. Rapidly cooling of these metals through the carbide-
forming range of 900 to 1500°F (480 to 820°C) also works to prevent carbide formation.
(2) Embrittlement: Another problem with high-chromium stainless steels is an embrittlement that
occurs after long times at elevated temperatures. This is attributed to the formation of sigma
phase, a brittle compound that forms at elevated temperatures and coats grain boundaries,
thereby producing a brittle crack path through the metal. Stainless steels used in high-
temperature service should be checked periodically to detect sigma-phase formation.
(3) Passivation & surface treatment: According to ASTM A380, passivation is “the removal of
exogenous iron or iron compounds from the surface of stainless steel by means of a chemical
dissolution, most typically by a treatment with an acid solution that will remove the surface
contamination, it will not significantly affect the stainless itself.” In addition, it also
describes passivation as “the chemical treatment of stainless steel with a mild oxidant, such
as a nitric acid solution, for the purpose of enhancing the spontaneous formation of the
protective passive film.” Passivation is recommended where the surface must be free of iron.
Passivation can also aid in the rapid development of the passive surface layer on the steel, but
usually does not result in a marked change in appearance of the steel surface.
Passivation is performed with acid solutions (or pastes) to remove contaminants and promote
the formation of the passive film on a freshly created surface (for example, via grinding,
machining or mechanical damage). Common passivation treatments include nitric acid
(HNO3) solutions or pastes which will clean the steel surface of free iron contaminants.
Since dangerous acids are involved, only trained personnel can perform such process. In
addition, stainless steel pickling acids are highly corrosive to carbon steel, and should be
thoroughly removed by rinsing the component after completing the process, and/or neutralize
with alkali before the rising. Residual hydrofluoric acid will initiate pitting corrosion.

4
 MEC325/580 Handout: Milling Process and Machines
Spring 2010 I. Kao

Milling—A machining operation in which a workpiece is fed past a rotating cylindrical tool with
multiple cutting edges.

Two (2) basic types of milling operations:

1. peripheral milling: the axis of tool is parallel to the surface being machined; the machining is
performed by cutting edges on the outside peripheral of the cutter.
2. face milling: the axis of cutter is perpendicular to the surface being machined; common type of end
mill found in machine shop.

An illustration of these two basic types of milling operations is shown in the following figure.

Classification of milling machines: Various classifications are used, as follows

1. Horizontal spindle (for peripheral milling) or Vertical spindle (for face milling, end mill, …)
2. Types including (i) knee and column (most common type), (ii) bed type, (iii) planer type, (iv) tracer
mills, and (v) CNC milling machines

Two forms of milling processes: up milling and down milling

Milling is an interrupted
cutting process wherein
entering and leaving the
cut subjects the tool to
impact loading, cyclic
heating, and cycle cutting
forces. Two common
types of milling
configurations are: up
milling (or conventional
milling) and down milling
(or climb milling). The
former has the tool and
workpiece moving in

1
opposite directions; whereas in the latter, the tool moves in the same direction as the work feed.

In up milling, the chip is very thin at the beginning, where the tooth contacts the work, and increases in
thickness, becoming a maximum where the tooth leaves the work.
Advantages include:
(1) The cutter tends to push the work along and lift it upward from the table. This action tends to
eliminate any effect of looseness in the feed screw and nut of the milling machine table and results
in a smooth cut.
(2) Tooth engagement is not a function of workpiece surface characteristics, and contamination or scale
on the surface does not affect tool life.
(3) There is a tendency for the tool to chatter.
Disadvantages include:
(1) The action also tends to loosen the work from the clamping device; therefore, greater clamping
forces must be employed.
(2) The smoothness of the generated surface depends greatly on the sharpness of the cutting edges.
(3) The chips can be carried into the newly machined surface, causing the surface finish to be poorer
(rougher) than in down milling.

In down milling, maximum chip thickness occurs close to the point at which the tooth contacts the work.
Because the relative motion tends to pull the workpiece into the cutter, any possibility of looseness in the
table feed screw must be eliminated if down milling is to be used. It should never be attempted on
machines that are not designed for this type of milling. Virtually all modern milling machines are capable
of doing down milling. Metals that readily work harden should be climb milled.
Advantages include:
(1) Because the material yields in approximately a tangential direction at the end of the tooth
engagement, there is less tendency (than when up milling is used) for the machined surface to show
toothmarks.
(2) The cutting process is smoother with less chatter.
(3) The cutting force tends to hold the work against the machine table, permitting lower clamping
forces, particularly for slender parts.
(4) Recommended for maximum cutter life when CNC machine tools are used.
(5) Suitable for finishing cuts, e.g., on aluminum.
Disadvantages include:
(1) The fact that the cutter teeth strike against the surface of the work at the beginning of each chip can
be a disadvantage if the workpiece has a hard surface, as casting sometimes does.
(2) The teeth may dull more rapidly.
(3) Because of the resulting high impact forces when the teeth engage the workpiece, this operation
must have a rigid setup, and backlash must be eliminated in the table feed mechanism.
(4) It is not suitable for workpiece having surface scale, such as hot-worked metals, forgings, and
castings – because the scale is hard and abrasive and causes excessive wear and damage to cutter
teeth, thus tool life can be short.

2
Probable Cause of Milling Problems and Cures
(cf. DeGarmo, et al. “Materials and Process in Manufacturing”)

Problem Probable Cause Cures

Chatter (vibration) 1. Lack of rigidity in machine, Use large arbors

fixture, arbor, or workpiece
2. Cutting load too great Decrease feed per tooth or number of teeth in contact
3. Dull cutter Sharpen or replace inserts
4. Poor lubrication Flood coolant
5. Straight-tooth cutter Use helical cutter
6. Radial relief too great
7. Rubbing, insufficient clearance Check tool angles

Loss of accuracy 1. High cutting load causing
(cannot hold size) deflection
2. Chip packing, between teeth
3. Chips not cleaned away before
mounting new piece of work
Decrease number of teeth in contact with work or feed
per tooth
Adjust cutting fluid to wash chips out of teeth

Cutter rapidly dulls 1. Cutting load too great Decrease feed per tooth or number of teeth in contact
2. Insufficient coolant Add blending oil or coolant

Poor surface finish 1. Feed too high Check to see if all teeth are set at same height
2. Tool dull
3. Speed too low
4. Not enough cutter teeth

Cutter dig in 1. Radial relief too great Check to see that workpiece is not deflecting and is
(hogs into work) 2. Rake angle too large clamped securely
3. Improper speed

Work burns 1. Cut is too light Enlarge feed per tooth

2. Tool edge worn Sharpen cutter
3. Insufficient radial relief
4. Land too wide

Cutter burns 1. Not enough lubricant Add sulfur-based oil

2. Speed too high Reduce cutting speed
Flood coolant

Teeth breaking 1. Feed too high Decrease feed per tooth

Use cutter with more teeth
Reduce table feed rate

3
 MEC325/580 H ANDOUT: TAYLOR ’ S T OOL W EAR E QUATION
Spring 2010 I. Kao

1 Introduction
The Taylor’s equation for tool wear is expressed in a power-law equation form, as follows

v Tn = C (1)

Equation (1) is a standard nonlinear power equation. In the case of equation presented in equation (1), we
can take logarithmic relationship of the variables and make a linear equation in the log-log coordinates, as
expressed in the following equation
log v + n log T = log C (2)
Equation (2) represents a line in the (log T, log v) space.
In the next sections, an example is given for finding the exponent n, and the coefficient, C, for a tool by
applying the Taylor equation for tool wear.

2 Finding Parameters of the Taylor’s Tool Wear Equation

Equation (2) is a result of taking logarithmic form of equation (1). At least two sets of experimental data of
tool speed and life are required to find the exponent n and the coefficient C. If the two experimental data
sets are given as (T1 , v1 ) and (T2 , v2 ), we can write equation (2) for the two data sets as

log v1 + n log T1 = log C (3)

log v2 + n log T2 = log C (4)

Equation (3) minus (4) renders the following

log v2 − log v1 log(v2 /v1 )

n= = (5)
log T1 − log T2 log(T1 /T2 )
Once the exponent n is obtained from equation (5), the result can be substituted into equation (1) to find the
coefficient C = v1 T1n .

3 Example: Experimental Results

Experiments were conducted to characterize the parameters of tool wear based on the Taylor’s equation.
The experimental results of the relationship between the tool speed and tool life are tabulated in the
following. The points are corresponding to the following figure (the first two circled points).

exp data tool life, T velocity, v

set 1 100 min 400 m/min
set 2 240 min 300 m/min

Applying the data of experiments in the table to equation (5), we obtain

log(300/400) −0.1249
n= = = 0.3286 (6)
log(100/240) −0.3802

1
 Plot of Taylor’s tool life equation
3
10

speed of tool (m/min)

2
10
1 2 3
10 10 10
Tool life (min)
Figure 1: The raw data of experiments are indicated by ’o’. The results of equation of tool wear is plotted
as a line in the logarithmic space of log T versus log v.

Substituting into equation (1), we find C = 1817. Thus, the Taylor’s equation of tool wear is

v (T )0.3296 = 1817 (7)

where the tool speed v has a unit of m/min and the tool life T is in minutes.
The result of the tool life relationship in equation (7) is plotted in Figure 1, in logarithmic scale. The two
circles indicate the two sets of experimental measurements given in the table.

2
 MEC580: H ANDOUT ON THE L EAST-S QUARES B EST F IT A LGORITHM
MEC580 Spring 2010 I. Kao

1 Introduction
In engineering applications, experiments are often conducted with multiple sets of data points for best curve
fitting. If the case when the governing equation between the two parameters is linear, the “linear regression”
algorithm can be applied. However, such linear regression method can not be applied directly if the equation
is in a power equation form, such as that in the Taylor’s equation for tool wear:
v Tn = C (1)
Equation (1) is a standard nonlinear power equation. In the case of equation presented in equation (1), we
can take logarithmic relationship of the variables and make a linear equation in the log-log coordinates, as
expressed in the following equation
log v + n log T = log C (2)
Equation (2) represents a line in the (log T, log v) space.
In the next section, a standard technique for determining the least-squares best fit solution is presented
as a matrix solution.

2 Algorithm of the Weighted Least-Squares Fit for Power Equations

Equation (2) is a result of taking logarithmic form of equation (1). It can be re-arranged in the following
form
log v = log C + ζ log T (3)
where ζ = −n.
With a total of i data sets of (v, T ) from experiments, we can re-arrange equation (3) in the following
matrix form for least-squares fit
y = Ax (4)
where
log v1 1 log T1
   
 log v2   1 log T2  
log C

y =  ..  A =  .. .. x = (5)
   
ζ

 .  . . 
log vi 1 log Ti
The least-squares solution of x in equation (4) can be obtained using the Penrose-Moore generalized inverse
that minimizes the norm of errors in y [3]. That is,
x = A∗ y (6)
where the superscript ‘*’ denotes the generalized inverse. The left inverse is used in equation (6); i.e.,
A∗ = (AT A)−1 AT .
Equation (6) minimizes the norm of the squared errors in y = log v instead of v. In order to compensate
for such discrepancy, we utilize a weighted least-squares fit of the following form
x = (WA)∗ Wy (7)
where the weighting matrix is W = diag[ey1 . . . eyi ] = diag[v1 . . . vi ] that corrects the logarithmic scale
of the norm of squared errors to be minimized.
The equation and derivation presented in this section can also be found in references [1, 2].

1
3 Example: Experimental Results and Curve Fitting
An example is used in this section to illustrate the application of the LS fit equations in (6) and (7). We will
use relationship between the tool speed and tool life, as characterized in the Taylor’s equation in (1). The
experimental results are tabulated in the following:

exp data velocity, v tool life, T

set 1 400 m/min 100 min
set 2 300 m/min 240 min
set 3 200 m/min 820 min

When the three data sets are used to find the LS solution of the exponent, n, and constant, C, we can
obtain the following terms based on the experimental data and equations (6) and (7). We have

1 2.0000 2.6021 400.0 0 0

     

A =  1 2.3802  y =  2.4771  W= 0 300.0 0  (8)

1 2.9138 2.3010 0 0 200.0

The generalized inverse of A and WA are

2.8217 0.6283 8.8962

   
−2.4500 −2.1113 −9.6255
A∗ = (WA)∗ = × 10−3 (9)
−1.0235 −0.1213 1.1448 −3.3497 1.4018 4.5967

Substituting into equations (6) and (7), respectively, we obtain

without weighting: n = 0.3295; C = 1824.3 (10)

with weighting: n = 0.3293; C = 1822.7 (11)

It is noted that equations (10) and (11) differ only slightly. Thus, either solution is acceptable. We will adopt

v (T )0.3293 = 1823 (12)

where v has a unit of m/min and T is in minutes.

If a fourth data set was added via experiments with

exp data velocity, v tool life, T

set 1 400 m/min 100 min
set 2 300 m/min 240 min
set 3 200 m/min 820 min
set 4 50 m/min 54, 000 min

Employing the same procedures above, by including the additional data, we obtain:

without weighting: n = 0.3306; C = 1836.1 (13)

with weighting: n = 0.3298; C = 1827.0 (14)

It is noted that the solution in equation (14) with the weighting matrix renders a result that is more consistent
with those in equations (10) and (11). This is because the addition of the weighting matrix will restore the
scales of v and T in the LS fitting, instead of using the logarithmic scales of log v and log T .
The results of equation (14) are plotted in Figure 1, to compare with the raw data points in a log-log
plot.

2
 4
10

3
10

2
10

1
10
0 1 2 3 4 5
10 10 10 10 10 10
Figure 1: Comparison between the raw data (indicated by ’o’) and the results of the weighted least-squares
best fit using the power equation (1) and LS fit equation (7).

References
[1] N. Xydas and I. Kao, “Modeling of contact mechanics and friction limit surface for soft fingers with
experimental results,” International Journal of Robotic Research, vol. 18, no. 9, pp. 941–950, September
1999.

[2] I. Kao and F. Yang, Stiffness and Contact Mechanics for Soft Fingers in Grasping and Manipulation to
appear on the IEEE Transactions of Robotics and Automation, 2004

[3] G. Strang, Linear Algebra and Its Applications, Academic Press, 2nd edition, 1980.

3
 Taguchi Methods!

Taguchi Methods Lecture (I)

•  “Taguchi Methods” refer to a
collection of
– Non-dynamic methods in product
design
– Dynamic method
– Newest: Mohalanobis-Taguchi
method

Taguchi Methods!

Taguchi Methods Lecture (II)

•  Non-Dynamic Taguchi Method (Handouts on
bboard)!
–  ASME magazine: vol. 14, no. 8, December 1994!
– His visit to ASME after the one-day workshop at Stony
Brook!
–  Introduction & parameter design!
–  Orthogonal arrays!
–  Case study 1: canon example!
–  Case study 2: RL circuit design!
– With Excel spreadsheet for calculation!
– Handout for further details!
 Taguchi Methods
Myth? Or Good Engineering?

Professor Imin Kao!

Department of Mechanical Engineering!
SUNY at Stony Brook!
Phone: 631-632-8308 kao@mal.eng.sunysb.edu!

Taguchi Methods!

Quality Engineering

“Quality Is The Loss A Product Causes To!

Society After Being Shipped,!
Other Than Any Losses Caused By!
Its Intrinsic Functions.”!

!! ! !-- by Dr. Genichi Taguchi!

 Taguchi Methods!

Loss Function
Products That Meet Tolerance Also Inflict A Loss!

where x is the value of the quality characteristic!

! m is the target value!
! k=A/!2 with A= cost to handle a defective; !=tolerance!

Taguchi Methods!

How to Reduce Loss: 2 steps

Starting ! ! Reduce ! !
Place on!
Conditions ! Variation ! Target!
 Taguchi Methods!

R & D Activities

Robust!
Technology!

Taguchi Methods!

Parameter Design

•  Control Factors!
–  Factors you can and want to control during
manufacturing and during use!
–  e.g., gun powder, angle of incline!
•  Noise Factors!
–  Factors you cannot or do not want to
control during manufacturing or during use!
–  e.g. temperature, wind, uncertainties!
 Taguchi Methods!

Sources of Noises
!  Inner Noise!
–  Material or dimensional deterioration!
–  A function of time or usage!

!  Outer Noise!
–  Conditions such as temperature, humility, voltage!
!  Unit-to-Unit Variation!
–  Variation in manufacturing under same conditions!
–  Such as parts in different locations of an oven!

Taguchi Methods!

Comparison Between
Traditional and Taguchi Methods

Traditional Methods:! Taguchi Methods:!

Remove cause of defect! Reduce effects of cause!

Traditional approaches may be

expensive or impossible or may be
more difficult to implement!
 Taguchi Methods!

Case Studies

Parameter Design!
1.  Canon Example!
–  A system with well-known mechanics
model : to design for target distance!
2.  Circuit Design Example!
–  A simple RL circuit: to maintain constant
level of current output!

Taguchi Methods!

Orthogonal Arrays in
Experimental Design
"  L4,
L8, L9, L12, L16, L18, L27, and L36!
"  Recommended: L9 and L18!
–  Number of experiments: 9 and 18!
–  Number of readings: product array!

"  Product Array:!

–  inner array: control factors!
–  outer array: noise factors!
 Taguchi Methods!

Parameter Design:
Definition of S/N Ratios

Professor Imin Kao, Manufacturing Automation Laboratory, SUNY at Stony Brook; kao@mal.eng.sunysb.edu!

Taguchi Methods!

Smaller-the-Better Criterion
Definition of S/N Ratios

Professor Imin Kao, Manufacturing Automation Laboratory, SUNY at Stony Brook; kao@mal.eng.sunysb.edu!
 Taguchi Methods!

Larger-the-Better Criterion
Definition of S/N Ratios

Professor Imin Kao, Manufacturing Automation Laboratory, SUNY at Stony Brook; kao@mal.eng.sunysb.edu!

Taguchi Methods!

Nominal-the-Best Criterion
Definition of S/N Ratios

Professor Imin Kao, Manufacturing Automation Laboratory, SUNY at Stony Brook; kao@mal.eng.sunysb.edu!
 Taguchi Methods!

Case Study of Taguchi Methods:

Canon Example

Professor Imin Kao!

Department of Mechanical Engineering!
SUNY at Stony Brook!
631-632-8308; email: kao@mal.eng.sunysb.edu!

Taguchi Methods!

Cannon Example

Distance ! !y= (F!t/m)2 /g sin2" = k F2 sin2" !

Ball weight & !!t !m = 0.2 kg, !t=0.1sec!
Constant (k) ! !k = 0.02549 m/N2!
Range of " ! !0 < " " 45º!
Range of F ! !0 < F " 170 N!
 Taguchi Methods!

Traditional Approach

Starts with F=130N and apply the !

projectile equation!

y = y(F, ") = k F2 sin2"!

150 = (0.02549)1302 sin(2 ")!

Solve for " = 10.19º!

Taguchi Methods!

Parameter Design
!  Control factors! F1 = 30 N "1 = 5°!
F2 = 90 N "2 = 23°!
F3 = 150 N "3 = 42°!

!  Noise factors! Fi+ = nominal + 10% Fi = nominal!

-
Fi = nominal -10% "i+ = nominal + 3°!
"i = nominal "i- = nominal - 3°!

!  Compound factor! N1 = y(Fi+ , "i+)!

-
N2 = y(Fi , "i- ) for i=1,2,3!
 Calculations of Flying Distance
nominal N1=y(F1’!1’) y N2=y(F3’!3’) y
F1 = 30 N F1 = 33 F1 = 27
!1=5º != 8º 7.65 != 2º 1.30
!2=23º != 26º 21.88 != 20º 11.95
!= 45º 27.76 != 39º 18.18
!3=42º
F2 = 90 N F2 = 99 F2 = 81
!1=5º != 8º 68.87 != 2º 11.67
!2=23º != 26º 196.88 != 20º 107.51
!= 45º 249.85 != 39º 163.60
!3=42º
F2 = 150 N F3 = 165 F3 = 135
!1=5º != 8º 191.30 != 2º 32.41
!2=23º != 26º 546.89 != 20º 298.63
!= 45º 694.02 != 39º 454.44
!3=42º

Taguchi Methods!

Calculation of S/N Ratio and Mean

 Taguchi Methods!

Response Tables

Response table (#)! Response table (mean)!

1st 2nd 3rd
F avg@30N avg@90N avg@150N
6.06 6.06 6.06
! (!=5º) (!=23º) (!=42º)
-0.0386 7.63 10.60
Optimum Condition: " = 42º!
(within chosen range) F = 76.9 N!

Taguchi Methods!

Plot of Response
 Taguchi Methods!

Comparison of Variability
Traditional Solution! Parameter Design!
(F=130 N, "=10.19º) !=45! (F=76.9 N, " = 42º) !=26!

Taguchi Methods!

Conclusions of First Iteration

"  Conclusions From Response Table (#)!
–  The parameter " is more sensitive to parameter
variation so we pick one with largest S/N!
–  F has constant S/N ratio so it is insensitive to
parameter variations!
"  Conclusions From Response Table (mean)!
–  The mean values increase with F!
–  The mean values increase with " for the range of
values in the first iteration!
 Taguchi Methods!

Graphical Interpretations

Increase S/N ratio ! Move mean to target!

Output variations are much! The slope at F=30N!
smaller at "=42o than those! is about 88o so the
at 5o, for input variation! output is nearly linear
of ±3o! when F>30N!

Taguchi Methods!

Summary

•  Quality and Loss Function!

•  Two Steps to Increase Robustness and
to Enhance Quality!
•  Use Signal-to-Noise Ratio (S/N) in
Parameter Design for Robust
Technology!
•  More Robust Results are obtained by
Using Parameter Design!
 Using Taguchi Methods in Circuit Design1
Manufacturing Automation Laboratory (MAL)
Department of Mechanical Engineering
SUNY at Stony Brook

1 Purpose

To design a simple circuit with a resistance, R, and self-inductance, L, so that the output current is at 10
amperes. The loss function, in terms of dollars, is estimated at $200 if the current deviates more than 4
amperes which will cause the circuit to cease functioning.

2 Theoretical Equations to Calculate Current y and Sensitivity

The output current subject to the RL circuit is given by the following equation.

V
y=p 2 (1)
R + (2πf L)2

where V is the input voltage, R is the resistance, f is the frequency, and L is the inductance. The following
terms are also defined:
1
m̂ = (y1 + y2 + · · · + yn )
n
1
Sm = (y1 + y2 + · · · + yn )2
n
1
Ve = (y 2 + y22 + · · · + yn2 − Sm )
n−1 1
The sensitivity is defined in this case (nominal-the-better) to be
1
S= (Sm − Ve ) (2)
n
with the signal-to-noise ratio
1 S m − Ve
η = 10 log (3)
n Ve
The loss function is defined as
A0 2
L= σ (4)
∆20
where A0 is the loss due to malfunction of this circuit, ∆0 is the function limit, and σ 2 is the variance.
1
The example was adapted from the Taguchi Methods – Research and Development.

1
 Control factors Values selected for parameter design
Resistance (R) R1 = 0.5Ω R1 = 5.0Ω R1 = 9.5Ω
Inductance (L) L1 = 0.010H L1 = 0.020H L1 = 0.030H
Noise factors Values estimated for parameter design
Voltage (V ) 90 V 100 V 110 V
Frequency (f ) 50 Hz 55 Hz 60 Hz
R′ −10% 0 10%
L′ −10% 0 10%

Table 1: Values and estimated ranges of control and noise factors

Results S/N ratio Sensitivity

No. R L N1 N2 η S
1 1 1 21.5 38.4 7.6 29.2
2 1 2 10.8 19.4 7.5 23.2
3 1 3 7.2 13.0 7.4 19.7
4 2 1 13.1 20.7 9.7 24.3
5 2 2 9.0 15.2 8.5 21.4
6 2 3 6.6 11.5 8.0 18.8
7 3 1 8.0 12.2 10.4 20.0
8 3 2 6.8 10.7 9.8 18.6
9 3 3 5.5 9.1 8.9 17.0

Table 2: Calculation of S/N ratios and sensitivities

3 Specification of Design Parameters

In the parameter design, we first need to identify the control and noise factors. Control factors are the factors
that we can control or select freely. The noise factors are the ones that we can not or do not want to control,
such as the actual voltage and frequency of input power and the variations in the actual values of resistance
and inductance (assume that the actual values vary within certain ranges).

The control factors are the resistance, R, and inductance, L. Each control parameter is chosen to have
three levels in our analysis. The noise factors include the voltage of power input, V , and frequency, f ,
and the uncertainties of the resistor and inductor components which are assume to vary ±10% from their
nominal values. Table 1 summarizes the factors.

4 Parameter Design

Using the parameter design, we can calculate the signal-to-noise ratios and sensitivities using equations (2)
and (3). The results are tabulated in Table 2.

2
 η S η S
R1 7.5 24.0 L1 9.2 24.5
R2 8.7 21.5 L2 8.6 21.1
R3 9.7 18.5 L3 8.1 18.5

Table 3: Factorial effects

Table 3 of average values are computed in order to compare the control factors of each level using the
S/N ratios, η, and sensitivities, S. From Table 3, we conclude that the optimal design is R3 L1 which has
the highest S/N ratios in R and L, respectively. In order to determine the difference between average output
and the target value, a confirmation experiment is performed. The currents are found to be y1 = 8.0A and
y2 = 12.2A with an average of 10.1. There is nearly no difference between the average and the target value
in this case and thus no further adjustment is needed. If, however, there is a large difference, the output will
need to be adjusted using a factor that has larger sensitivity but affects S/N less. Such a factor is called the
adjustment factor.

5 Loss Calculation

Using the loss function defined in equation (4) as a basis, we employ the following equation to calculate L
with a similar definition in order to obtain figures of loss function for the purpose of comparison. The loss
using parameter design is
A0 1 200 1
L = 2 η/10 = 2 = $1.14
∆0 10 4 101.04
This value is much smaller than $200.

If traditional design is chosen, i.e. R2 L2 , we obtain

200 1
L= = $1.76
42 100.85
The improvement by using the parameter design is $0.62 per each product – about 35% improvement.

6 Conclusion

The parameter design of the Taguchi Methods, when applied to this electronic design problem, yields satis-
factory results. The S/N ratio enhances the robustness of the product and reduces the loss. The quality of
design of this circuit is improved over the traditional solution.

References

[1] G. Taguchi System of Experimental Design, vols. 1 and 2 Quality Resources, Dearborn Michigan, vol.
1 and 2, 1991

3
[2] G. Taguchi and S. Konishi Taguchi Methods – Research and Development ASI press, vol. 1 in Quality
Engineering Series, 1992

4
 Taguchi Methods!

Case Study of Taguchi Methods:

RL circuit example

Professor Imin Kao!

Department of Mechanical Engineering!
SUNY at Stony Brook!
!
631-632-8308; email: kao@mal.eng.sunysb.edu

Taguchi Methods!

RL Circuit Design

!  Design RL circuit such that the current,

y, is at 10 A.!

!  The loss function is estimated at $200 if

current deviates more than 4A!
 Taguchi Methods!

Control and Noise Factors

Taguchi Methods!

Calculating S/N Ratios

•  Apply Nominal-the-Best Criterion!

•  S/N ratio equation for 2 data points!
 Taguchi Methods!

Parameter Design Using

Orthogonal Array

Taguchi Methods!

S/N Ratio: Factorial Effects

!  The
S/N ratios from the orthogonal
array!

!  The optimal design is R3L1 which has

highest S/N ratios in R and L!
!  Confirmation Run: y1=8.0A, y2=12.2A
with average at 10.1A!
 Taguchi Methods

Parameter Design:
Definition of S/N Ratios

Professor Imin Kao, Manufacturing Automation Laboratory, SUNY at Stony Brook; kao@mal.eng.sunysb.edu
 Orthogonal Arrays for
Taguchi Methods
Manufacturing Automation Laboratory
Department of Mechanical Engineering
SUNY at Stony Brook
Stony Brook, NY 11794-2300

Abstract

This article contains the orthogonal arrays that are listed in Appendix 3 of the “Taguchi Methods:
Research and Development”. The book is Volume One of the Quality Engineering Series, published by
the ASI Press.
According to Dr. Taguchi, however, only

 , 
 , and 
are preferred for parameter design.


1 The orthogonal array

The  orthogonal array is for 3 control factors, each with two-level variations. This is denoted as  .

  
No.     
1 1 1 1
2 1 2 2
3 2 1 2
4 2 2 1

Table 1:   orthogonal array

1
 

2 The orthogonal array

The  orthogonal array is for 7 control factors, each with two-level variations. This is denoted as   .

    
No.            
1 1 1 1 1 1 1 1
2 1 1 1 2 2 2 2
3 1 2 2 1 1 2 2
4 1 2 2 2 2 1 1
5 2 1 2 1 2 1 2
6 2 1 2 2 1 2 1
7 2 2 1 1 2 2 1
8 2 2 1 2 1 1 2

Table 2:  orthogonal array



3 The orthogonal array

The 
orthogonal array is for 4 control factors, with 4 three-level variations. This is denoted as  
   .

  
No.        
1 1 1 1 1
2 1 2 2 2
3 1 3 3 3
4 2 1 2 3
5 2 2 3 1
6 2 3 1 2
7 3 1 3 2
8 3 2 1 3
9 3 3 2 1

Table 3: 
orthogonal array

2
 

4 The orthogonal array

The   orthogonal array is for 11 control factors, with 11 two-level variations. This is denoted as       .
    
No.                 
   
1 1 1 1 1 1 1 1 1 1 1 1
2 1 1 1 1 1 2 2 2 2 2 2
3 1 1 2 2 2 1 1 1 2 2 2
4 1 2 1 2 2 1 2 2 1 1 2
5 1 2 2 1 2 2 1 2 1 2 1
6 1 2 2 2 1 2 2 1 2 1 1
7 2 1 2 2 1 1 2 2 1 2 1
8 2 1 2 1 2 2 2 1 1 1 2
9 2 1 1 2 2 2 1 2 2 1 1
10 2 2 2 1 1 1 1 2 2 1 2
11 2 2 1 2 1 2 1 1 1 2 2
12 2 2 1 1 2 1 2 1 2 2 1

Table 4:   orthogonal array

3
 


5 The orthogonal array

The   orthogonal array is for 15 control factors, with 15 two-level variations. This is denoted as        .
     
No.              
           
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2
3 1 1 1 2 2 2 2 1 1 1 1 2 2 2 2
4 1 1 1 2 2 2 2 2 2 2 2 1 1 1 1
5 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2
6 1 2 2 1 1 2 2 2 2 1 1 2 2 1 1
7 1 2 2 2 2 1 1 1 1 2 2 2 2 1 1
8 1 2 2 2 2 1 1 2 2 1 1 1 1 2 2
9 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
10 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
11 2 1 2 2 1 2 1 1 2 1 2 2 1 2 1
12 2 1 2 2 1 2 1 2 1 2 1 1 2 1 2
13 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1
14 2 2 1 1 2 2 1 2 1 1 2 2 1 1 2
15 2 2 1 2 1 1 2 1 2 2 1 2 1 1 2
16 2 2 1 2 1 1 2 2 1 1 2 1 2 2 1

Table 5:   orthogonal array

4
 

6 The orthogonal array

The    orthogonal
 array is for 8 control factors, with 1 two-level and 7 three-level variations. This is denoted
as  

    .

      
No.                 
1 1 1 1 1 1 1 1 1
2 1 1 2 2 2 2 2 2
3 1 1 3 3 3 3 3 3
4 1 2 1 1 2 2 3 3
5 1 2 2 2 3 3 1 1
6 1 2 3 3 1 1 2 2
7 1 3 1 2 1 3 2 3
8 1 3 2 3 2 1 3 1
9 1 3 3 1 3 2 1 2
10 2 1 1 3 3 2 2 1
11 2 1 2 1 1 3 3 2
12 2 1 3 2 2 1 1 3
13 2 2 1 2 3 1 3 2
14 2 2 2 3 1 2 1 3
15 2 2 3 1 2 3 2 1
16 2 3 1 3 2 3 1 2
17 2 3 2 1 3 1 2 3
18 2 3 3 2 1 2 3 1

Table 6:   orthogonal array

5
  

7 The orthogonal array

The   orthogognal array is for 13 contrl factors, with 13 three level variations. This is denoted as        .
     
No.              
       
1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 1 1 1 1 2 2 2 2 2 2 2 2 2
3 1 1 1 1 3 3 3 3 3 3 3 3 3
4 1 2 2 2 1 1 1 2 2 2 3 3 3
5 1 2 2 2 2 2 2 3 3 3 1 1 1
6 1 2 2 2 3 3 3 1 1 1 2 2 2
7 1 3 3 3 1 1 1 3 3 3 2 2 2
8 1 3 3 3 2 2 2 1 1 1 3 3 3
9 1 3 3 3 3 3 3 2 2 2 1 1 1
10 2 1 2 3 1 2 3 1 2 3 1 2 3
11 2 1 2 3 2 3 1 2 3 1 2 3 1
12 2 1 2 3 3 1 2 3 1 2 3 1 2
13 2 2 3 1 1 2 3 2 3 1 3 1 2
14 2 2 3 1 2 3 1 3 1 2 1 2 3
15 2 2 3 1 3 1 2 1 2 3 2 3 1
16 2 3 1 2 1 2 3 3 1 2 2 3 1
17 2 3 1 2 2 3 1 1 2 3 3 1 2
18 2 3 1 2 3 1 2 2 3 1 1 2 3
19 3 1 3 2 1 3 2 1 3 2 1 3 2
20 3 1 3 2 2 1 3 2 1 3 2 1 3
21 3 1 3 2 3 2 1 3 2 1 3 2 1
22 3 2 1 3 1 3 2 2 1 3 3 2 1
23 3 2 1 3 2 1 3 3 2 1 1 3 2
24 3 2 1 3 3 2 1 1 3 2 2 1 3
25 3 3 2 1 1 3 2 3 2 1 2 1 3
26 3 3 2 1 2 1 3 1 3 2 3 2 1
27 3 3 2 1 3 2 1 2 1 3 1 3 2

Table 7:   orthogonal array

6
 

8 The orthogonal array

The  
 orthogognal array   
is for 23 contrl factors, with 11 two level variations and 12 three level variations.
  .
This is denoted as  

         
No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 1’ 2’ 3’ 4’
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
2 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1
3 1 1 1 1 1 1 1 1 1 1 1 3 3 3 3 3 3 3 3 3 3 3 3 1 1 1 1
4 1 1 1 1 1 2 2 2 2 2 2 1 1 1 1 2 2 2 2 3 3 3 3 1 2 2 1
5 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 1 1 1 1 1 2 2 1
6 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 1 1 1 1 2 2 2 2 1 2 2 1
7 1 1 2 2 2 1 1 1 2 2 2 1 1 2 3 1 2 3 3 1 2 2 3 2 1 2 1
8 1 1 2 2 2 1 1 1 2 2 2 2 2 3 1 2 3 1 1 2 3 3 1 2 1 2 1
9 1 1 2 2 2 1 1 1 2 2 2 3 3 1 2 3 1 2 2 3 1 1 2 2 1 2 1
10 1 2 1 2 2 1 2 2 1 1 2 1 1 3 2 1 3 2 3 2 1 3 2 2 2 1 1
11 1 2 1 2 2 1 2 2 1 1 2 2 2 1 3 2 1 3 1 3 2 1 3 2 2 1 1
12 1 2 1 2 2 1 2 2 1 1 2 3 3 2 1 3 2 1 2 1 3 2 1 2 2 1 1
13 1 2 2 1 2 2 1 2 1 2 1 1 2 3 1 3 2 1 3 3 2 1 2 1 1 1 2
14 1 2 2 1 2 2 1 2 1 2 1 2 3 1 2 1 3 2 1 1 3 2 3 1 1 1 2
15 1 2 2 1 2 2 1 2 1 2 1 3 1 2 3 2 1 3 2 2 1 3 1 1 1 1 2
16 1 2 2 2 1 2 2 1 2 1 1 1 2 3 2 1 1 3 2 3 3 2 1 1 2 2 2
17 1 2 2 2 1 2 2 1 2 1 1 2 3 1 3 2 2 1 3 1 1 3 2 1 2 2 2
18 1 2 2 2 1 2 2 1 2 1 1 3 1 2 1 3 3 2 1 2 2 1 3 1 2 2 2
19 2 1 2 2 1 1 2 2 1 2 1 1 2 1 3 3 3 1 2 2 1 2 3 2 1 2 2
20 2 1 2 2 1 1 2 2 1 2 1 2 3 2 1 1 1 2 3 3 2 3 1 2 1 2 2
21 2 1 2 2 1 1 2 2 1 2 1 3 1 3 2 2 2 3 1 1 3 1 2 2 1 2 2
22 2 1 2 1 2 2 2 1 1 1 2 1 2 2 3 3 1 2 1 1 3 3 2 2 2 1 2
23 2 1 2 1 2 2 2 1 1 1 2 2 3 3 1 1 2 3 2 2 1 1 3 2 2 1 2
24 2 1 2 1 2 2 2 1 1 1 2 3 1 1 2 2 3 1 3 3 2 2 1 2 2 1 2
25 2 1 1 2 2 2 1 2 2 1 1 1 3 2 1 2 3 3 1 3 1 2 2 1 1 1 3
26 2 1 1 2 2 2 1 2 2 1 1 2 1 3 2 3 1 1 2 1 2 3 3 1 1 1 3
27 2 1 1 2 2 2 1 2 2 1 1 3 2 1 3 1 2 2 3 2 3 1 1 1 1 1 3
28 2 2 2 1 1 1 1 2 2 1 2 1 3 2 2 2 1 1 3 2 3 1 3 1 2 2 3
29 2 2 2 1 1 1 1 2 2 1 2 2 1 3 3 3 2 2 1 3 1 2 1 1 2 2 3
30 2 2 2 1 1 1 1 2 2 1 2 3 2 1 1 1 3 3 2 1 2 3 2 1 2 2 3
31 2 2 1 2 1 2 1 1 1 2 2 1 3 3 3 2 3 2 2 1 2 1 1 2 1 2 3
32 2 2 1 2 1 2 1 1 1 2 2 2 1 1 1 3 1 3 3 2 3 2 2 2 1 2 3
33 2 2 1 2 1 2 1 1 1 2 2 3 2 2 2 1 2 1 1 3 1 3 3 2 1 2 3
34 2 2 1 1 2 1 2 1 2 2 1 1 3 1 2 3 2 3 1 2 2 3 1 2 2 1 3
35 2 2 1 1 2 1 2 1 2 2 1 2 1 2 3 1 3 1 2 3 3 1 2 2 2 1 3
36 2 2 1 1 2 1 2 1 2 2 1 3 2 3 1 2 1 2 3 1 1 2 3 2 2 1 3

Table 8:   orthogonal array

References

[1] G. Taguchi System of Experimental Design, vols. 1 and 2 Quality Resources, Dearborn Michigan, vol. 1
and 2, 1991

[2] G. Taguchi and S. Konishi Taguchi Methods – Research and Development ASI press, vol. 1 in Quality
Engineering Series, 1992

7
 MEC325/580 H ANDOUT: A N E XAMPLE OF THE TAGUCHI M ETHOD

MEC325/580, Spring 2010 I. Kao

This handout explains the application of the non-dynamic Taguchi method using an example of experimental
design with L4 orthogonal array on a practical manufacturing process with experimental measurements.

Problem Statement: Experiments were conducted for a tile making process using the Taguchi method.
Three control factors are identified as crucial to the strength of the tiles, as described in the following.

A: ingredient #1, B: ingredient #2, and C: temperature of the curing process

with each control factor having two levels. The experiments were conducted with the results tabulated in
Table 1. Note that in Table 1, the orthogonal array employed is the L4 array. (cf. the handout on various
orthogonal arrays) The data measured (or the readings) in Table 1 are the strength of the tiles, in M P a.

Results
No. A B C N1 N2
1 1 1 1 100 250
2 1 2 2 160 185
3 2 1 2 495 295
4 2 2 1 360 313
Table 1: A L4 array and experimental results

1. Since the strength of the tile is considered here as the measure to evaluate the tile making process,
which criterion should you use: larger-the-better, smaller-the-better, or nominal-the-best?

2. Determine the optimal condition of each control factor, based on the parameter design with the S/N
ratios to maximize the strength.

Solution: The experimental design has 4 experiments, each with 2 readings. The terms “experiment” and
“reading” are explained in the Remarks at the end of this handout.

1. Because the strength of the tile is to be maximized, we will use larger-the-better criterion.

2. Once the larger-the-better criterion is chosen, the following equations for calculating the S/N ratios
are employed.
 
2 1 1 1
σ = + (1)
2 y12 y22
η = −10 log σ 2 (2)

where y1 and y2 are the data under the 2 columns N1 and N2 . Substituting the readings in Table 1 into
equations (1) and (2), we obtain the L4 orthogonal array with the calculated data of σ 2 and η listed in
Table 2.
Once the S/N ratios, η, are calculated, as listed in the last column in Table 2, we can calculate the
average S/N ratios associated with each level of the three control factors. For example, the average

1
 Results S/N Ratios
No. A B C N1 N2 σ2 η
1 1 1 1 100 250 0.00005800 42.37
2 1 2 2 160 185 0.00003414 44.67
3 2 1 2 495 295 0.00000779 51.09
4 2 2 1 360 313 0.00000896 50.48
Table 2: Calculating the S/N ratios based on the measurements and readings

A B C
level 1 43.52 46.73 46.42
level 2 50.78 47.57 47.88
Table 3: The response table

S/N ratio for A1 (first level of control parameter A) is the average of experiment No. 1 and 2 (row 1
and 2) by virtue of the designation of levels under the column for the control factor A. Similarly, the
average S/N ratio for B2 is the average of experiment No. 2 and 4. The entries of the response tables
are calculated and listed in Table 3.
The response table can be used to plot the following response chart for the signal-to-noise ratios, as
shown in Figure 1.
52

50

48

46

44

42

40

38
A1 A2 B1 B2 C1 C2
Figure 1: The plot of response chart based on the values obtained in the response table.

Optimal Levels of Control Factors:

From the response chart (or the response table), the optimal design corresponding to the choice of
the combination of one of the two levels of the three control factors is found to be A2 B2 C2 (i.e.,
level 2 of parameter A, level 2 of parameter B, and level 2 of parameter C) which corresponds to the
combination of highest S/N ratios from each parameter.
For example, if the first and second levels of the curing temperatures (C1 and C2 ) are 150◦ F and
200◦ F , respectively, we will choose level 2 with a curing temperature at 200◦ F as our design based
on the results of Taguchi method.

Remarks:

2
 (i) Each row of the orthogonal array is call an “experiment” which represents a set of experimental setup
using the designated levels of the control factors. For example, in the L4 array in Table 1, each of the
4 rows represents one experiment.

(ii) In each experiment, there are “readings”—typically two readings if two compound noise levels are
used, as in the case of the example here with N1 and N2 .

(iii) Note that the largest S/N ratio is always chosen for optimal design levels regardless of the criterion
used. This applies to both positive and negative S/N ratios.

(iv) For other orthogonal arrays, the number of experiments will change and the combination of the levels
of the control factors will also change. For example, a L9 array has 9 experiments and each control
factor has 3 levels. Yet, the methodology of finding the S/N ratios and the response table/chart remains
the same.

3
 F ORGING AND THE L OAD -S TROKE C URVE

MEC325/580, Spring 2010 I. Kao

Forging Problem: In an impression-die forging operation on a moderately complex part with flash made
from a cylindrical workpiece in a cold upset forging, the height changes from h0 = 6 mm to hf = 1.5 mm
(with diameter changing from D0 = 12 mm to Df = 24 mm). Use a value of Kf = 7.0 as the forging
shape factor. The material is low carbon steel, annealed with flow curve parameters: K = 530 M P a and
n = 0.26.

1. Determine the force required to start the plastic deformation; that is, stress beyond the linear elastic
region (assume ǫ = 0.2%).

2. Determine the instantaneous force required for the following heights:

h = 5.75, 5.5, 5.0, 4.5, 4.0, 3.5, 3.0, 2.5, 2.0, 1.5 mm

3. Based on the results in Part (2), plot the load-stroke curve. What is your conclusion about such curve?

Solution : The force required for the forging process is

F = Kf Yf A = Kf (K ǫn )A (1)
 
where ǫ = ln h0
h , Kf is the shape factor (Kf = 7.0, given), A is the area. The workpiece has a volume of

π
V = (12)2 × 6 = 678.58 mm3
4

1. At the starting of yielding (assume ǫ = 0.002 = 0.2%), the flow stress is

Yf = K ǫn = 530(0.002)0.26 = 105.3 M P a (2)

We have from equation (1) the force require for the forging operation
π
F = Kf Yf A = (7.0)(105.3 × 106 ) · (12 × 10−3 )2 = 83, 573 N (3)
4
The force required is 83.57 kN .

2. At h = 5.75 mm the stroke is (h0 − h) = 6 − 5.75 = 0.25 mm. The force is

" 0.26 # 
6 678.58
 
F = (7.0) 530 ln = 192, 686 N (4)
5.75 5.75

Repeat the calculation for h = 5.5, 5.0, 4.5, 4.0, 3.5, 3.0, 2.5, 2.0, 1.5 mm, we can tabulate the results
in the following.

1
 height, h Diameter, D area, A true strain, ǫ Flow stress, Yf Force, F stroke, h0 − h
(mm) (mm) (mm2 ) (M P a) (N ) (mm)
6 12.00 113.10 0 0 0 0
5.988 12.01 113.32 0.00200 105.355 83575 0.012
5.75 12.26 118.01 0.04256 233.247 192686 0.25
5.5 12.53 123.38 0.08701 280.909 242608 0.5
5 13.15 135.72 0.18232 340.482 323464 1
4.5 13.86 150.80 0.28768 383.348 404653 1.5
4 14.70 169.65 0.40547 419.125 497720 2
3.5 15.71 193.88 0.53900 451.324 612522 2.5
3 16.97 226.19 0.69315 481.826 762906 3
2.5 18.59 271.43 0.87547 511.986 972792 3.5
2 20.78 339.29 1.09861 543.120 1289933 4
1.5 24.00 452.39 1.38629 576.977 1827127 4.5

The results of the force-stroke curve are plotted in Figure 1.

2000
1800
1600
1400
load, in kN

1200
1000
800
600
400
200
0
0 1 2 3 4 5
stroke (h0-h), in mm
Figure 1: The load-stroke curve of a impression-die forging process.

2
 E XTRUSION AND A NALYSIS OF P RESSURE D URING P ROCESS

MEC325/580, Spring 2010 I. Kao

Extrusion Problem: In an direct extrusion process, pressure needs to be applied to extrude a billet of length
L0 = 75 mm and diameter D0 = 25 mm with an extrusion ratio of rx = 4.0. The die angle is α = 90◦ .
The billet material has the following parameters for the plastic flow stress equation: K = 415 M P a and
n = 0.18. Use the Johnson’s formula with a = 0.8 and b = 1.5.

1. Determine the ram pressures needed for the extrusion process at the following lengths:
L = 75, 50, 25, 0 mm

2. Based on the results in Part (1), plot the pressure-stroke curve. What is your conclusion about such
curve?

Solution : Use the Johnson’s formula with

p = Ȳf ǫx (1)
where ǫx = a + b ln rx , with a = 0.8 and b = 1.5. Thus,
ǫ = ln rx = ln 4.0 = 1.3893
ǫx = 0.8 + 1.5(ln rx ) = 2.8795
415(1.3863)0.18
Ȳf = = 373 M P a
(1 + 0.18)

1. Use the die angle of α = 90◦ , the billet material is to be forced through the die opening almost
immediately. The ram pressures are calculated in the following at the respective lengths.
At L = 75 mm the ram pressure is
2(75)
 
p = 373 2.8795 + = 3312 M P a (2)
25
where the additional pressure due to friction was added in the term 2(75)/25.
Repeat the calculation for L = 50, 25, 0 mm, we find

2(50)
 
p = 373 2.8795 + = 2566 M P a
25
2(25)
 
p = 373 2.8795 + = 1820 M P a
25
2(0)
 
p = 373 2.8795 + = 1074 M P a
25

2. The ram stroke is (L0 − L). The pressure-stroke curve is plotted in Figure 1.
It is noted that the pressure required for the indirect extrusion is constant, as shown in Figure 1, and it
is equal to the pressure of the direct extrusion at the end when L = 0 mm.

1
 Direct extrusion in solid line; Indirect in dashed line

3000

2500
Ram pressure (MPa)

2000

1500

1000

500

0
0 10 20 30 40 50 60 70
Ram stroke (mm)
Figure 1: The pressure-stroke curves of direct and indirect extrusion processes are in solid and dashed lines,
respectively.

2
 W IRE D RAWING P ROCESS AND A NALYSIS

MEC325/580, Spring 2010 I. Kao

Extrusion Problem: In a wire drawing process to reduce the diameter of a plain carbon steel wire from
D0 = 220 µm to Df = 175 µm in a cold working process, the angle of the die is α = 15◦ and the
coefficient of friction is µ = 0.1. The plastic strength of the material is K = 500 M P a with a strain
hardening exponent of n = 0.25. The tensile strength of the steel wire is Sut = 390 M P a.

1. Determine if the process is feasible?

2. If the drawing process is feasible, what is the force required for the wire drawing process?

3. Determine the safety margin of the drawing force versus the rupture force of the wire. Is this wire
drawing process safe?

Solution : First, we need to determine if the drawing process is feasible, based on the parameters given.

1. The maximum reduction per pass is

A0
= e = 2.71828 =⇒ rmax = 63.2% (1)
Af

A0 −Af D02 −Df2

Here, the reduction is r = A0 = D02
= 36.7%. Thus, the drawing process is feasible.

2. The drawing force is

µ A0
 
F = Af Ȳf 1 + φ ln
tan α Af
π 0.1
2  
= 175 × 10−6 × 3.29 × 108 1 + (0.9719)(0.4577)
4 tan 5◦
= 7.54 N
500×106 (0.4577)0.25 D0 +Df
where ǫ = ln A
A0
f
= 0.4577, Ȳf = (1+0.25) = 329 M P a, D = 2 = 197.5 µm,
(D0 −Df )
Lc = 2 sin α = 258 µm, and φ = 0.88 + 0.12 LDc = 0.9719.

3. The force to rupture, depending on the tensile strength, is

π 2 π
Fr = Af · Sut = Df · Sut = (175 × 10−6 )(390 × 106 ) = 9.38 N (2)
4 4

Thus, the safety margin is

Fr − F 9.38 − 7.54
ns = = = 20% (3)
Fr 9.38
 MEC325/580: Food/Soda Cans
Manufacturing
Facts and Manufacturing Processes

Imin Kao, Professor

Dept. of Mechanical Engineering
College of Engineering and App. Sci.
SUNY at Stony Brook

Prof. Imin Kao

Do You Know?
•  Formaldehyde is added to many food cans
•  The formaldehyde flavor legacy in can-
making
•  You should NEVER cook food with the can
•  A single can-tooling machine spits out 400
cans per minute
•  250 millions cans per day are consumed
(one can per person per day)
 Prof. Imin Kao

Why Using Formaldehyde in Can?

•  To kill bacteria!
–  Steel cans in 1940s use an emulsion (95% water
and 5% oil) for lubrication in mfg process
–  Certain bacteria eats oil in the emulsion, so the
biocide is added
–  Amount is not enough to cause health hazard, but
enough to taste
–  This results in the famous preservative flavor (e.g.,
in Budweiser)

Prof. Imin Kao

Formaldehyde Flavor Legacy

•  Why not use biocides without flavor?
–  Yes, mfg’ers do that in recent years (e.g., Miller
Genuine Draft and other similar brews)
•  Almost every new emulsion formula had to
be made to taste like formaldehyde or else
people aren’t going to accept it.
•  Are there any other things/additives in your
food or beer/soda cans?
 Prof. Imin Kao

Polymer in Your Food Can

•  Polymer (in solvent) is spray-coated inside to
serve two functions:
–  Plasters any microscopic debris (resulting from mfg proc) to
the can wall and away from the food
–  Keeps the food from interacting with can material (e.g.,
tomato acid to corrode the can)
•  Don’t cook the food in the can when you go
camping!
–  Or else you will be eating polymer (since they degrade when
heated)
–  Typical consequence of such a culinary blunder: headaches
and constipation

Prof. Imin Kao

Soda Can Manufacturing

•  Soda can manufacturers are competing with
low-priced plastic and glass bottles
•  A single can-tooling machine can spit 400
cans a minute (i.e. 7 cans per second!!)
•  For one can per a person per day
–  Need 250 million cans per day
•  Employ the “Deep Drawing” process
 Prof. Imin Kao

Manufacture of Aluminum Can

Starting Material: Shipment of
Canstock from a roll can to filler

Cupper Flanger

Bodymaker (inkl. Necker

domer and trimmer)

Washer Printer Printer Inside Inside coating

drier oven coating drier oven

Prof. Imin Kao

Manufacturing Process (I)

Source: J. E. Wang, Texas A&M University

 Prof. Imin Kao

Manufacturing Process (II)

Source: J. E. Wang, Texas A&M University

Prof. Imin Kao

Finished Product of Can

Source: J. E. Wang, Texas A&M University

 Prof. Imin Kao

Can Top: material and design

Source: J. E. Wang, Texas A&M University

Prof. Imin Kao

Video Clip (source: Discovery Channel)

 Prof. Imin Kao

Process & Materials

•  Cost of mfg: $40 per 1,000 cans (4 cent @)
•  Major cost is on the lid of the can
–  Body made of AA3004 aluminum (Al 97.8%; Mg
1.0%) with a yield strength 170 MPa and tensile
strength 215 Mpa.
–  Lid made of a stronger aluminum alloy of AA5182
(Al 95.2%; Mg 4.5%) with a yield strength 395
MPa and tensile strength 420 Mpa.
•  The necking at the lid: to reduce cost of
material

Prof. Imin Kao

What Are You Paying?

•  A typical breakdown of what you are
paying in a can of soda:
–  4 cents for making the can (major cost:
magnesium aluminum alloy with higher ductility)
–  10 cents (or more) goes for advertising
–  Less than 1 cent for the 12 ounces of beverage!
•  That’s why the no-name brand soda sells for
much less in stores!
 Prof. Imin Kao

Summary
•  Formaldehyde smell in food and beer cans
•  Do not cook food with the can when go
camping
•  Addition of polymer inside the can to
protect the can/food, and to enclose mfg
remains
•  How much are you paying in a can of soda
 Chapter 20
Sheet Metalworking
Part V: Metal Forming and Sheet Metal
Working

Groover “Manufacturing Processes”

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

SHEET METALWORKING

!  Three Basic Processes

1.  Cutting Operations
2.  Bending Operations
3.  Drawing
!  Other Sheet Metal Forming Operations
1.  Dies and Presses for Sheet Metal Processes
2.  Sheet Metal Operations Not Performed on Presses
3.  Bending of Tube Stock

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Sheet Metalworking Defined

Cutting and forming operations performed on relatively

thin sheets of metal
!  Thickness of sheet metal = 0.4 mm (1/64 in) to 6 mm
(1/4 in)
!  Thickness of plate stock > 6 mm
!  Operations usually performed as cold working

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Sheet and Plate Metal Products

!  Sheet and plate metal parts for consumer and

industrial products such as
!  Automobiles and trucks
!  Airplanes
!  Railway cars and locomotives
!  Farm and construction equipment
!  Small and large appliances
!  Office furniture
!  Computers and office equipment

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Sheet Metalworking Terminology

!  Punch䇳and䇳die – tooling to perform cutting,

bending, and drawing
!  Stamping press – machine tool that performs
most sheet metal operations
!  Stampings – sheet metal products

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Three Basic Types of Sheet

Metal Processes
1.  Cutting
!  Shearing to separate large sheets
!  Blanking to cut part perimeters out of sheet metal
!  Punching to make holes in sheet metal
2.  Bending
!  Straining sheet around a straight axis
3.  Drawing
!  Forming of sheet into convex or concave shapes

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Sheet Metal Cutting

(1) Just before punch contacts work; (2) punch pushes into
work, causing plastic deformation; (3) punch penetrates into
work causing a smooth cut surface; and (4) fracture is
initiated at opposing cutting edges to separate the sheet

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Characteristics of Sheared Edge

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Blanking and Punching

!  Blanking (a) - sheet metal cutting to separate piece

(called a blank) from surrounding stock
!  Punching (b) - similar to blanking except cut piece is
scrap, called a slug

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Clearance in Sheet Metal Cutting

Distance between punch cutting edge and die cutting

edge
!  Typical values range between 4% and 8% of stock
thickness
!  If too small, fracture lines pass each other,
causing double burnishing and larger force
!  If too large, metal is pinched between cutting
edges and excessive burr results

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Clearance in Sheet Metal Cutting

!  Recommended clearance is calculated by:

c = at
where c = clearance; a = allowance; and t = stock
thickness
!  Allowance a is determined according to type of metal

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Sheet Metal Groups Allowances

Metal group a
1100S and 5052S aluminum alloys, all 0.045
tempers
2024ST and 6061ST aluminum alloys; brass, 0.060
soft cold rolled steel, soft stainless steel

Cold rolled steel, half hard; stainless steel, 0.075

half hard and full hard

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Punch and Die Sizes

!  For a round blank of diameter Db:

!  Blanking punch diameter = Db 䇳 2c
!  Blanking die diameter = Db
where c = clearance
!  For a round hole of diameter Dh:
!  Hole punch diameter = Dh
!  Hole die diameter = Dh + 2c
where c = clearance

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Punch and Die Sizes

!  Die size determines

blank size Db
!  Punch size
determines hole
size Dh
!  c = clearance

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Angular Clearance

!  Purpose: allows slug or blank to drop through die

!  Typical values: 0.25° to 1.5° on each side

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Cutting Forces

!  Important for determining press size (tonnage)

F = S t L = (0.7) Sut t L
where S = shear strength of metal; Sut = ultimate
tensile strength of metal; t = stock thickness, and L =
length of cut edge

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Example of Cutting

Problem: A sheet metal under cutting process has the

following parameters: hole=1”-dia; thickness=!”;
Sut=140,000 psi. Estimate the force required for
cutting.

!  Solution: Fmax= (0.7)(140,000)(!) (" "1) = 38,500 lb

= 19.25 tons = 170,000 Newtons

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Sheet Metal Bending

(a) Straining of sheet metal around a straight axis to take a

permanent bend
(b) Metal on inside of neutral plane is compressed, while
metal on outside of neutral plane is stretched

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Types of Sheet Metal Bending

!  V䇳bending - performed with a V䇳shaped die

!  Edge bending - performed with a wiping die

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 V-Bending

(1) Before bending

(2) After bending
!  Application notes:
!  Low production
!  Performed on a
press brake
!  V-dies are simple
and inexpensive

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Edge Bending

(1) Before bending

(2) After bending
!  Application notes:
!  High production
!  Pressure pad
required
!  Dies are more
complicated and
costly

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Stretching during Bending

!  If bend radius is small relative to stock thickness,

metal tends to stretch during bending
!  Important to estimate amount of stretching, so final
part length = specified dimension
!  Problem: to determine the length of neutral axis of the
part before bending

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Bend Allowance Formula

where Ab = bend allowance; ! = bend angle; R= bend

radius; t = stock thickness; and Kba is factor to
estimate stretching
!  If R < 2t, Kba = 0.33
!  If R ! 2t, Kba = 0.50

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Handout & Example

!  (See Handout)
!  Equations and example to calculate
!  the Bend Allowance
!  the Bending Force
!  the Springback

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Drawing

Sheet metal forming to make cup䇳shaped, box䇳shaped,

or other complex䇳curved, hollow䇳shaped parts
!  Sheet metal blank is positioned over die cavity and
then punch pushes metal into opening
!  Products: beverage cans, ammunition shells,
automobile body panels
!  Also known as deep drawing (to distinguish it from
wire and bar drawing)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Deep Drawing of a Soda Can

!  (See the Soda Can Manufacture PPT)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Shapes other than Cylindrical
Cups

!  Each of the following shapes presents its own unique

technical problems in drawing
!  Square or rectangular boxes (as in sinks)
!  Stepped cups
!  Cones
!  Cups with spherical rather than flat bases
!  Irregular curved forms (as in automobile body
panels)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Other Sheet Metal Forming on

Presses

!  Other sheet metal forming operations performed on

conventional presses can be classified as
!  Operations performed with metal tooling
!  Operations performed with flexible rubber tooling

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Ironing

!  Achieves thinning and elongation of wall in a drawn

cup: (1) start of process; (2) during process

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Embossing

!  Creates indentations in sheet, such as raised (or

indented) lettering or strengthening ribs
!  (a) Punch and die configuration during pressing; (b)
finished part with embossed ribs

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Guerin Process

(1) before and (2) after

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Tools for the Sheet Metalworking

Processes

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Punch and Die Components

!  Components of a punch and die for a blanking operation

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Stamping Press

!  Components of a
typical mechanical
drive stamping
press

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Gap Frame Press

!  Gap frame press for

sheet metalworking
(photo courtesy of E.
W. Bliss Co.)
!  Capacity = 1350 kN
(150 tons)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Press Brake

!  Press brake (photo

courtesy of Niagara
Machine & Tool
Works)
!  Bed width = 9.15 m
(30 ft)
!  Capacity = 11,200 kN
(1250 tons)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 CNC Turret Press

!  Computer
numerical
control turret
press (photo
courtesy of
Strippet, Inc.)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

CNC Turret Press Parts

!  Sheet metal parts

produced on a
turret press,
showing variety of
hole shapes
possible (photo
courtesy of
Strippet Inc.)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Straight-Sided Frame Press

!  Straight䇳sided frame
press (photo courtesy of
Greenerd Press &
Machine Company, Inc.)

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e

Power and Drive Systems

!  Hydraulic presses - use a large piston and cylinder to

drive the ram
!  Longer ram stroke than mechanical types
!  Suited to deep drawing
!  Slower than mechanical drives
!  Mechanical presses – convert rotation of motor to
linear motion of ram
!  High forces at bottom of stroke
!  Suited to blanking and punching

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 Operations Not Performed on
Presses

!  Stretch forming
!  Roll bending and forming
!  Spinning
!  High䇳energy䇳rate forming processes

©2010 John Wiley & Sons, Inc. M P Groover, Fundamentals of Modern Manufacturing 4/e
 ME325/580 H ANDOUT: S HEET-M ETAL B ENDING

MEC325/580, Spring 2010 I. Kao

Sheet-Metal Bending: This handout concerns the sheet metal bending process and analysis. [The material
is from “Fundamentals of Modern Manufacturing: materials, processes, and systems” by M. P. Groover,
Wiley, 2002; and other sources]

Bend Allowance: If the bend radius is small relative to stock thickness, the metal tends to stretch during
bending. It is important to be able to estimate the amount of stretching that occurs, if any, so that the final
part length will match the specified dimension. The problem is to determine the length of the neutral axis
before bending to account for stretching of the final bent section. This length is called the bend allowance,
and it can be estimated as follows:

 
   (1)

where

= bend allowance (in or mm), !  
 #"
$ is the bend angle (in degrees),  is the bend
radius (in or mm),  is the stock thickness (in or mm), as shown in Figure 1(a), and % is a factor to estimate
'&
stretching. Note the term ($) *,+ is the bend angle in radians. The following design values are recommended
for - .
-  00/2/2139 1 if 54768
- if 5:768
(2)

This only applies when the bend radius is small relative to sheet thickness. An illustration is shown in
Figure 1.

F
punch
t
Af
Ai
Rf
Ri

die

 , which is the same

(a) During punch (b) After punch: springback

as that of the V-die, and becomes

;
Figure 1: Illustration of sheet-metal bending V-die. The included angle of the part is
after the springback.

Bending Force: The force required to perform bending depends on the geometry of the punch and die and
the strength, thickness, and width of the sheet metal. The maximum bending force can be estimated by the
following equation, based on bending of a simple beam

<=  ;?>AE @3B DC (3)

< >@ B
where is the bending force (in lb or N), is the tensile strength of the sheet metal (in psi or MPs),

the width of part in the direction of the bend axis (in or mm), is the stock thickness (in or mm), and
E is
is

1
the die opening dimension as defined in Figure 2. The constant ! ;
accounts for differences encountered in

.
an actual bending process. Its value depends on type of bending, as defined in the following

" 

  ;;  0 /22/ 139 1
- for
   

- for
(4)

punch

D
die
D
V-die
wiping die
Figure 2: Illustration of die opening dimensions for V-bending and edge bending.

5 1     ,33

Example: A sheet-metal blank is to be bent as shown in Figure 3. The metal has a modulus of elasticity
, yield strength , and tensile strength .
>@ ! 9,33
1.500 w=1.750
1.0

t=0.125 R=0.187
00

120 o

Figure 3: Example of sheet-metal bending

1. Determine the starting blank size, and

2. Calculate the bending force if a V-die will be used with a die opening dimension
E  / "$# .

Solution: Refer to the dimensions in Figure 3.

B  /&% 9'(# /29   / 3  
. For an

included angle

 6
1. The starting blank has a width of
, as shown, the bend angle is
)  
. Its length is equal to
. The value of - in equation (2) is
0.33 since 54768
. The bend allowance is obtained from equation (1)

 
    0/ 
*%  0/2131+ 0/ 6 9   0 / 6 1,-(#
Thus, the required length of the blank is
.  6 /29 3  0 / 6 1,  6 /&% 1,-(# /
2. The required bending force is obtained from equation (3) using ! ;  /2131 in equation (4) for V-die.

<  /2131  /&% 9   9 /, 33   0/ 6 9  C  6 1 -0/21

Thus,
(5)

2
 Springback: When the bending pressure is removed at the end of the deformation operation, elastic energy
remains in the bent part, causing it to recover partially towards its original shape. This elastic recovery is
called springback, as shown in Figure 1, defined as the increase in the included angle of the bent part relative
to the included angle of the forming tool after the tool is removed. That is,
> = #
;
"

 (6)

where
> is the springback,


is the included angle of the bending die tool, and

;
is the included angle
of the sheet-metal part after it is removed from the bending die, as shown in Figure 1.

Analysis of Spring Back in Bending of Sheet-Metal: The following empirical equation defines the amount
of springback in a bending operation on a sheet metal with the geometry shown in Figure 4.

 ;   
    " 1
   
    
 (7)

where 
and  ;
are the bend radii before and after the spring back,  is the yield strength,
 is the Young’s

modulus, and is the thickness of the sheet metal.

t
A
i
Ri
θi
Rf

Figure 4: Springesback during the bending process of a sheet metal of thickness 

 ;
The springback defined in equation (6) is based on the included angle before and after the springback. It is
useful, however, to use the radii of curvature ( and  
in Figure 4) to represent the amount of springback.
To this end, we assume that the arc length of the curved bend is the same before and after springback, as

     
 ;  ;  8;  8;   ;   
shown in Figure 4; that is,

 (8)

Since

   "  and

#;   "  ; , we can write

#;  "  ;      "  ;    ; 

   (9)

Since  ;  and    4 , the springback is

> 
#;
"
  " ; 

 "
    (10)

3


Equation (10) depends on the ratio of
of the bending die tool .
   , which can be substituted by equation (7), and the included angle
In a design problem with synthesis, the radius of curvature of the bending die angle, , often needs to be  
designed in order to render the final radius at the desired value after springback. In this case, equation (7)
needs to be solved using an equation solver (root finder) or by iteration.
(
 9,33 1    
Example: In a bending operation of a 1010 cold-drawn steel sheet metal of thickness
, with a yield   )
strength of and the Young’s modulus of , the radius of curvature of the bending die
6
tool is .

1. When the bending operation on the die tool is finished, what will be the final radius of the bend based
on equation (7), as the sheet metal is removed from the die tool?

?
springback as a function of the angle .


2. What is the springback as a function of the included angle of the bending die tool Plot the

Solution: Several equations in this section are employed to solve this problem.

1. Applying equation (7), we find

6 ;   9 ,33   9 ,33 
1      
   0 / , 6
0
 
6 
1       
  " 1 6 

Solve for  ;  6 / 939  .
2. Employing equation (10), we have
  " 0 / , 6 0


 "
>
 
If 

 & , the springback is >  0/  % 6  % / 6  . If
  & , the springback is >  0/  6  
6 /   . TheC springback as a function of the bending die angle (included  angle,
 ) is plotted in Figure 5.

4
 0.25

0.2

0.15
spingback

0.1

0.05

0
40 60 80 100 120 140 160 180
include angle in degrees, Ai

Figure 5: Plot of springback as a function of the included angle

5
 MEC325/580 H ANDOUT: VOLUMETRIC C HANGES AND
D EFECTS IN M ETAL C ASTING

Spring 2010 I. Kao

The following lecturing materials are adapted from the textbook [1].

Shrinkage in metal casting: Metals shrink or contract during solidification and cooling processes. Shrink-
age, which causes dimensional changes in casting, is the result of the following three factors:

1. Contraction of the molten metal as it cools prior to solidification;

2. Contraction of the metal during phase change from liquid to solid (latent heat of fusion); and
3. Contraction of the solidified metal (the casting) as its temperature drops to ambient temperature.

The largest amount of shrinkage occurs during the cooling of the casting in factor 2 above. In the
following table, the percentages of contraction for several metals during solidification are listed. Note,
however, some metals expand during cooling, including gray cast iron.

Table 1: Volumetric contraction or expansion percentage for various metals in casting during solidification

metal volumetric contraction

Aluminum 7.1%
Zinc 6.5%
Al, 4.5% Cu 6.3%
Gold 5.5%
White iron 4–5.5%
Copper 4.9%
Brass (70-30%) 4.5%
Magnesium 4.2%
90% Cu, 10% Al 4%
Carbon steels 2.5–4%
Al, 12% Si 3.8%
Lead 3.2%
metal volumetric expansion
Bismuth 3.3%
Silicon 2.9%
Gray cast iron 2.5%

Defects in Casting: Defects are important in casting. Different names have been used to associate the same
or similar defects. As a result, the International Committee of Foundry Technical Associations has developed
standardized nomenclature, consisting of seven basic categories of casting defects, as follows.

1. Metallic projections: This category consists of fins, flash, or massive projections such as swells and
rough surfaces.

1
 2. Cavities: This category consists of rounded or rough internal or exposed cavities, including blow-
holes, pinholes, and shrinkage cavities.

3. Discontinuities: Examples are such as cracks, cold or hot tearing, and cold shuts. If the solidifying
metal is constrained from shrinking freely, cracking and tearing can occur. Although many factors
are involved in tearing, coarse grain size and the presence of low-melting segregates along the grain
boundaries (intergranular) increase the tendency for hot tearing. Incomplete castings result from the
molten metal being at too low a temperature or from the metal being poured too slowly. Cold shut is
an interface in a casting that lacks complete fusion because of the meeting of two streams of partially
solidified metal.

4. Defective surface: This includes defects such as surface folds, laps, scars, adhering sand layers (in
sand casting), and oxide scale.

5. Incomplete casting: This category includes defects such as misruns (due to premature solidification),
insufficient volume of metal poured, and runout (due to loss of metal from the mold after pouring).

6. Incorrect dimension or shape: Such defects are owing to factors such as improper shrinkage al-
lowance, pattern-mounting error, irregular contraction, deformed pattern, or warped casting.

7. Inclusions: Inclusions usually form during melting, solidification, and molding. Generally nonmetal-
lic, they are regarded as harmful because they act like stress raisers and reduce the strength of the
casting. Hard inclusions (spots) also tend to chip or break tools in machining. They can be filtered
out during processing of the molten metal with the environment (usually with oxygen) or the crucible
material. Chemical reactions among components in the molten metal may produce inclusions; slugs
and other foreign materials entrapped in the molten metal also become inclusions. Reactions between
the metal and the mold material may produce inclusions as well. In addition, spalling of the mold
and core surfaces produces inclusions, suggesting the importance of maintaining melt quality and
monitoring the conditions of the molds.

References

[1] S. Kalpakjian and S. R. Schmid Manufacturing Processes for Engineering Materials Prentice Hall,
fourth ed., 2003

2
 MEC325/580 H ANDOUT: M ETAL C ASTING AND R ISER D ESIGN

Spring 2010 I. Kao

An important aspect of design for metal casting using sand mold is the consideration of solidification time,
and the inclusion of riser design to elongate the time to solidification in order to reduce defects or other
failures in the casting process.

The solidification time in metal casting is governed by the following empirical equation called the
Chvorinov’s rule  n
V
TST = Cm (1)
A
where TST is the total solidification time, Cm is the mold constant, VA is the volume-to-surface-area ratio,
and n is the exponent, usually taken as 2. The mold constant, Cm depends on the particular conditions of
the cast operation, including mold material, thermal properties of case metal, and pouring temperature. The
mold constant can be obtained based on experimental data with the same mold, metal, pouring temperature,
... etc, even though the part may be very different. The Chvorinov’s rule suggests that a casting with higher
volume-to-surface-area ratio will cool and solidify more slowly than one with a lower ratio.

The riser design can be used to prevent part of the metal cast from prematurely solidified which causes
casting defects. Risers by its relative position can be side riser or top riser, or by its configuration can be
open riser or blind riser. The following figure shows a closed mold with a complex mold geometry which
requires a riser design. Note the different terminology of the parts of the cast and mold, as illustrated in
Figure 1.

Figure 1: A closed mold in which the mold geometry is more complex and requires a passageway system
leading into the cavity, with a riser design.

Example: Riser design using the Chvorinov’s rule.

A cylindrical riser must be designed for the sand casting mold shown in Figure 1. The passageway leading
to the cavity casting is a steel rectangular plate with dimension of 3′′ × 5′′ × 1′′ . Previous observations have
indicated that the total solidification time (T ST ) for this casting is 1.6 min. The cylindrical riser will have
a diameter-to-height ratio of 1.0. Determine the dimensions of the riser so that the T ST is 2.0 minutes to
allow more time for the flow of metal to cavity for a successful casting.

Solution: First, we need to determine Cm for the casting:

1
 The volume is V = 3 × 5 × 1 = 15 in3 ; the surface area is A = 2(3 × 5 + 3 × 1 + 5 × 1) = 46 in2 .
Given T ST = 1.6 min, we take n = 2 and apply the Chvorinov’s rule,
2
15

1.6 = Cm =⇒ Cm = 15.05 min/in2 (2)
46

Therefore, the mold constant for the riser is also Cm = 15.05 min/in2 .
πD 2 πD 3 D
For the cylindrical
 
riser, the volume is V = 4 h = 4 since h = 1 (given); the surface area is
πD 2
A = πDh + 2 4 = 1.5πD 2 . Thus, the ratio is

V (1) D
= 4 D= (3)
A 1.5 6
Substituting into the Chvorinov’s equation, we have
2
D

2.0 = 15.05 =⇒ D = h = 2.187′′ (4)
6

Therefore, the cylindrical riser with a diamter-to-height ratio of 1.0 should have a diamter of D = 2.187′′ .

Remarks: For the riser and cast, the following comparison can be made.

volume surface area

riser Vr = 8.216 in3 Ar = 22.54 in2
cast Vc = 15 in3 Ac = 46 in2

Based on the table, we have VVrc = 55%. That is, the volume of the cast is increased by 55% due to the
riser if only the rectangular part is concerned. However, the gain in time is 25% that allows the cavity to be
filled more completely.

References

[1] M. P. Groover Fundamentals of Modern Manufacturing: materials, processes, and systems Wiley, third
ed., 2006

2
 Metal Casting
Introduction
Manufacturing Processes –– Podcast Series

Imin Kao, Professor

Dept. of Mechanical Engineering
College of Engineering and App. Sci.
SUNY at Stony Brook

An Introduction of Casting
• Process in which molten metal/materials
flows by gravity or other force into a mold
where it solidifies in the shape of the mold
cavity
• Solidification processes can be classified
according to engineering material processed:
– Metals
– Ceramics, specifically glasses
– Polymers and polymer matrix composites (PMCs)

1
 Metal Casting

Open mold Closed mold

Two Categories of Casting Processes

1. Expendable mold processes – uses an
expendable mold which must be destroyed
to remove casting
– Mold materials: sand, plaster, and similar
materials, plus binders
2. Permanent mold processes – uses a
permanent mold which can be used over
and over to produce many castings
– Made of metal (or, less commonly, a ceramic
refractory material

2
 Solidification Time
• Heat content ∝ volume & heat transfer ∝ surface area
• Solidification time depends on size and shape of
casting by relationship known as Chvorinov's Rule

" V %n
TST = Cm $ '
# A&
where TST = total solidification time; V = volume of
the casting; A = surface area of casting; n = exponent
with typical value = 2; and Cm is mold constant.
!

Mold Constant in Chvorinov's Rule

• Mold constant Cm depends on:
– Mold material
– Thermal properties of casting metal
– Pouring temperature relative to melting point
• Value of Cm for a given casting operation
can be based on experimental data from
previous operations carried out using same
mold material, metal, and pouring
temperature, even though the shape of the
part may be quite different

3
 Chvorinov’s Rule & Cast Design
• A casting with a higher V/A ratio cools and
solidifies more slowly than one with lower ratio
– To feed molten metal to main cavity, TST for riser
must greater than TST for main casting
• Since mold constants of riser and casting will be
equal, design the riser to have a larger V/A ratio
so that the main casting solidifies first
– Riser acts as heat & cast reservoir
– This minimizes the effects of shrinkage

Furnaces for Casting Processes

Furnaces most commonly used in foundries:
• Cupolas
• Direct fuel‑fired furnaces
• Crucible furnaces
• Electric‑arc furnaces
• Induction furnaces

4
 Metals for Casting
• Most commercial castings are made of
alloys rather than pure metals
– Alloys are generally easier to cast, and
properties of product are better
• Casting alloys can be classified as:
– Ferrous: (1) gray cast iron, (2) nodular iron, (3)
white cast iron, (4) malleable iron, and (5) alloy
cast irons (∼ 1400°C or 2500°F) & (6) steel (1650°
C or 3000°F)
– Nonferrous: (1) Aluminum (660°C or 1220°F), (2)
Copper Alloys (1083°C or 1981°F), (3) Zinc Alloys
(419°C or 786°F), (4) others

Post-Solidification Processes
• Trimming
• Removing the core
• Surface cleaning
• Inspection
• Repair, if required
• Heat treatment

5
 Casting Quality
• There are numerous opportunities for things to
go wrong in a casting operation, resulting in
quality defects in the product
• The defects can be classified as follows:
– Common defects to all casting processes: (a)
misrun, (b) cold shut, (c) cold shots, (d) shrinkage
cavity, (e) microporosity, (f) hot tearing
– Defects related to sand casting process: (a) sand
blow, (b) pinholes, (c) sand wash, (d) scabs, (e)
penetration, (f) mold shift, (g) core shift, (h) mold
crack

Common Casting Defects (a) & (b)

Misrun: A casting that has Cold shot: Two portions of
solidified before completely metal flow together but there
filling mold cavity is a lack of fusion due to
premature freezing

6
 Common Casting Defects (c) & (d)
Cold shots: Metal splatters Shrinkage cavity: Depression in surface
during pouring and solid or internal void caused by solidification
globules form and become shrinkage that restricts amount of
entrapped in casting molten metal available in last region to
freeze

Common Casting Defects (e) & (f)

Microporosity: network of Hot tearing: hot cracking
small voids due to local caused by unyielding mold in
shrinkage in the dendritic contraction during cooling,
structure with separation of metal cast

7
 Expendable-Mold
Casting
Manufacturing Processes –– Podcast Series

Imin Kao, Professor

Dept. of Mechanical Engineering
College of Engineering and App. Sci.
SUNY at Stony Brook

Metal Casting

Open mold Closed mold

1
 Riser Design for Casting
• Riser is waste metal that is separated from
the casting and can be re-melted to make
more castings
• To minimize waste in the unit operation, it
is desirable for the volume of metal in the
riser to be a minimum
• Since the geometry of the riser is normally
selected to maximize the V/A ratio (why?),
this allows riser volume to be reduced to the
minimum possible value

Figure 11.1 (textbook) A large sand casting weighing over

680 kg (1500 lb) for an air compressor frame (photo
courtesy of Elkhart Foundry).

2
 Types of Patterns for Sand Casting
Figure 11.3 (textbook) Types of patterns used in
sand casting:
(a) solid pattern (b) split pattern
(c) match‑plate pattern (d) cope and drag pattern

Sand Casting
• Core:
– inserted into the mold cavity prior to pouring
– May require supports to hold it in position in the mold cavity
during pouring, called chaplets
• Desirable Mold Properties:
– Strength; Permeability; Thermal stability; Collapsibility;
Reusability
• Foundry Sands: Silica (SiO2) or silica mixed with
other minerals
– Good refractory properties; Small grain size yields better
surface finish on the cast part; Large grain size is more
permeable; Irregular grain shapes strengthen molds due to
interlocking

3
 Sand Casting Defects (a) & (b)
Sand blow: Balloon‑shaped Pinholes: Formation of many
gas cavity caused by release small gas cavities at or slightly
of mold gases during pouring below surface of casting

Sand Casting Defects (c) & (d)

Sand wash: irregularity in the Scabs: rough area on the surface
surface resulting from erosion due to encrustations of sand and
of sand mold during pouring metal, due to mold surface
flaking off during solidification

4
 Sand Casting Defects (e) & (f)
Penetration: When fluidity of liquid
metal is high, it may penetrate into
sand mold or core, causing casting
surface to consist of a mixture of
sand grains and metal
Sand Casting Defects (g) & (h)
Core shift: the core being
displaced from its intended
position, usually vertical, caused
by buoyancy of the molten metal
Mold shift: A step in cast
product at parting line
caused by sidewise
relative displacement of
cope and drag
Mold crack: a crack develops
due to insufficient mold strength
with molten seeping into the
mold to form a ‘fin’
5
 Other Expendable Mold Processes

• Shell Molding

• Vacuum Molding
• Expanded Polystyrene Process
• Investment Casting
• Plaster Mold and Ceramic Mold Casting

Investment Casting (Lost Wax Process)

A pattern made of wax is coated with a refractory
material to make mold, after which wax is melted
away prior to pouring molten metal
• “Investment” comes from a less familiar
definition of “invest” – “to cover completely,”
which refers to coating of refractory material
around wax pattern
• It is a precision casting process - capable of
producing castings of high accuracy and intricate
detail

6
 Investment Casting

Figure 11.8 (textbook) Steps in investment casting: (1)

wax patterns are produced, (2) several patterns are
attached to a sprue to form a pattern tree

Investment Casting (cont.)

Figure 11.8 (textbook) Steps in investment casting: (3) the pattern

tree is coated with a thin layer of refractory material, (4) the full
mold is formed by covering the coated tree with sufficient
refractory material to make it rigid

7
 Investment Casting (cont.)

Figure 11.8 (textbook) Steps in investment casting: (5) the mold is held
in an inverted position and heated to melt the wax and permit it to drip
out of the cavity, (6) the mold is preheated to a high temperature, the
molten metal is poured, and it solidifies

Investment Casting (cont.)

Figure 11.8 (textbook) Steps in investment casting: (7) the mold is

broken away from the finished casting and the parts are separated from
the sprue

8
 Investment Casting

Figure 11.9 (textbook) A one‑piece compressor stator with

108 separate airfoils made by investment casting (photo
courtesy of Howmet Corp.).

Metals for Casting

• Most commercial castings are made of
alloys rather than pure metals
– Alloys are generally easier to cast, and
properties of product are better
• Casting alloys can be classified as:
– Ferrous: (1) gray cast iron, (2) nodular iron, (3)
white cast iron, (4) malleable iron, and (5) alloy
cast irons (∼ 1400°C or 2500°F) & (6) steel (1650°
C or 3000°F)
– Nonferrous: (1) Aluminum (660°C or 1220°F), (2)
Copper Alloys (1083°C or 1981°F), (3) Zinc Alloys
(419°C or 786°F), (4) others

9
 SME Video Clip

• Next, let’s watch the SME

video clip about Casting
which came with the textbook
for classroom use ONLY…

10
 Permanent-Mold
Casting
Manufacturing Processes –– Podcast Series

Imin Kao, Professor

Dept. of Mechanical Engineering
College of Engineering and App. Sci.
SUNY at Stony Brook

Metal Casting

1
 Permanent Mold Casting Processes
• Economic disadvantage of expendable mold
casting: a new mold is required for every
casting
• In permanent mold casting, the mold is reused
many times
• The processes include:
– Basic permanent mold casting
– Die casting
– Centrifugal casting

Permanent Mold Casting

2
 Permanent Mold Casting

Figure 11.10 (textbook) Steps in permanent mold casting: (2) cores (if
used) are inserted and mold is closed, (3) molten metal is poured into
the mold, where it solidifies.

Advantages and Limitations

• Advantages of permanent mold casting:
– Good dimensional control and surface finish
– More rapid solidification caused by the cold metal
mold results in a finer grain structure, so castings are
stronger
• Limitations:
– Generally limited to metals of lower melting point
– Simpler part geometries compared to sand casting
because of need to open the mold
– High cost of mold

3
 Die Casting
A permanent mold casting process in which
molten metal is injected into mold cavity under
high pressure
• Pressure is maintained during solidification,
then mold is opened and part is removed
• Molds in this casting operation are called dies;
hence the name die casting
• Use of high pressure to force metal into die
cavity is what distinguishes this from other
permanent mold processes

Die Casting Machines

• Designed to hold and accurately close two
mold halves and keep them closed while
liquid metal is forced into cavity
• Two main types:
1. Hot‑chamber machine: Metal is melted in a
container, and a piston injects liquid metal under
high pressure into the die
2. Cold‑chamber machine: Molten metal is poured
into unheated chamber from external melting
container, and a piston injects metal under high
pressure into die cavity

4
Hot-Chamber Die Casting

Cold‑Chamber Die Casting

(2)

5
 Molds for Die Casting
• Usually made of tool steel, mold steel, or
maraging steel
• Tungsten and molybdenum (good refractory
qualities) used to die cast steel and cast iron
• Ejector pins required to remove part from
die when it opens
• Lubricants must be sprayed into cavities to
prevent sticking

Advantages and Limitations

• Advantages of die casting:
– Economical for large production quantities
– Good accuracy and surface finish
– Thin sections are possible
– Rapid cooling provides small grain size and good
strength to casting
• Disadvantages:
– Generally limited to metals with low melting points
– Part geometry must allow removal from die

6
 Metals for Casting
• Most commercial castings are made of
alloys rather than pure metals
– Alloys are generally easier to cast, and
properties of product are better
• Casting alloys can be classified as:
– Ferrous: (1) gray cast iron, (2) nodular iron, (3)
white cast iron, (4) malleable iron, and (5) alloy
cast irons (∼ 1400°C or 2500°F) & (6) steel (1650°
C or 3000°F)
– Nonferrous: (1) Aluminum (660°C or 1220°F), (2)
Copper Alloys (1083°C or 1981°F), (3) Zinc Alloys
(419°C or 786°F), (4) others

SME Video Clip

• Next, let’s watch the SME

video clip about Die Casting
which came with the textbook
for classroom use ONLY…

7
 Manufacturing Automation

Imin Kao
Professor
Department of Mechanical
Engineering
SUNY at Stony Brook

Manufacturing Automation
•  Manufacturing Automation
–  Fixed automation
–  Flexible automation
–  Agile automation
–  …
•  Societal Impacts
–  (name impacts …; See the next page)
•  Positioning System & Accuracy
–  (See Handout)
 Impacts of Automation
•  Increased production rate
•  Reduction of labor (economic impact on
society)
•  Societal impact on labor force
•  Technological innovation
•  Precision & repeatability in production
•  Hostile/hazardous environment
•  24-hour operation
•  …

Concurrent Engineering
•  Concurrent Engineering
–  History & perspectives
–  What is it?
–  Famous case studies
•  Design for X
–  Design for assembly
–  Design for manufacturing/manufacturability
–  Design for XXX
 CE vs. Traditional Prod. Dev.
(a) Traditional
product
development

(b) Product
development
using
Concurrent
Engineering

Programmable Automation Systems

•  Numerical control (NC) and CNC
–  Presented earlier …
•  Industrial robots
–  History
–  Kinematics
–  Robotic programming language (RPL)
–  Workspace or work envelope
•  Programmable login control (PLC)
–  Automation on manufacturing floors
 Introduction to Robotics
•  “Robot” – history and origin
–  1920 by Czech author K. !apek in his play
R.U.R. (Rossum’s Universal robots); from
Czech word “robota” meaning “worker”
–  Webster Dictionary: “An automated apparatus
or device that performs functions ordinarily
ascribed to humans or operates with what
appears to be almost human intelligence”

Robots: Fiction or Reality?

•  Ahead of Its Time?
–  Star Wars: C3-PO like humanoid?
–  The six-million dollar man?
•  Today
–  AIBO pet robot (a dog which can learn)
–  HONDA’s ASIMO humanoid robot
•  Speed: 2 km/h; hand load: 2~5kg/hand; weight:
130~210 kg; height: 160~182 cm
–  Others
 Robotics Research Today
Robotics Research Today:

1.  Where are we now?

2.  Where is the robotics research heading
to?
3.  Different fields of robotics research

Classification of Robots
•  By Coordinate System
1.  Cylindrical coordinate robots
2.  Spherical coordinate robots
3.  Jointed arm (articulated) robots
4.  Cartesian coordinate robots
•  By Mechanism Types
1.  Revolute
2.  Prismatic
Kinematics of a 2-link robot arm
•  A 2-link SCARA (Selectively Compliant
Arm for Robotic Assembly)
–  Kinematics: forward and inverse kinematics
–  Joint space and tool (Cartesian) space
–  Workspace consideration

(see handout: 2-link-robot.pdf )

Workspace of a SCARA Robot

 Workspace Synthesis

Robot Programming Language

•  Robot Programming Language (RPL)
–  High-level macro language to control motions
of a robot
•  Examples:
–  Adept: V++
–  IBM: AML (A manufacturing language)
–  RobotWorld: RAIL
–  …
 RPL Example
PL1:NEW 0; -- PAYLOAD SET TO DEFAULT SLOWSPEED:SUBR; -- Z DOWN & SLOW SPEED
PL2:NEW 15; -- PAYLOAD SET TO 50% ZMOVE(HGT2);
LIN1:NEW 0; -- LINEAR OFF PAYLOAD(PL2);
LIN2:NEW 20; -- LINEAR(20)
ZMOVE(HGT3);
END; -- END SLOWSPEED SUBROUTINE
PT1:NEW PT(-132.95,490.45,0,-27); -- COORDINATE
PT2:NEW PT(-161.55,521.20,0,-27); -- POSITIONS
PT3:NEW PT(-187.05,553.90,0,-27); -- FOR PICKUP:SUBR; -- GRASP, Z UP & LINEAR ON
PT4:NEW PT(-214.95,583.80,0,-27); -- POINTS GRASP;
PT5:NEW PT(-457.50,201.15,0,-27); -- 1 THROUGH 8 DELAY(1.0);
PT6:NEW PT(-483.85,236.10,0,-27); ZMOVE(HGT2);
PT7:NEW PT(-512.70,264.50,0,-27); PAYLOAD(PL1);
PT8:NEW PT(-541.35,297.75,0,-27); ZMOVE(HGT1);
HOME:NEW PT(650,0,0,0); -- HOME POSITION LINEAR(LIN2);
END; -- END PICKUP SUBROUTINE
HGT1:NEW 0;
HGT2:NEW -90;
DROPOFF:SUBR; -- RELEASE & Z UP
HGT3:NEW -184.0;
RELEASE;
DELAY(1.0);
MAIN:SUBR; -- MANUFACTURING SUBROUTINE
ZMOVE(HGT2);
PAYLOAD(PL1);
FASTSPEED:SUBR; -- DEFAULT SPEEDS & LINEAR
OFF ZMOVE(HGT1);
LINEAR(LIN1); END; -- END DROPOFF SUBROUTINE
PAYLOAD(PL1);
END; -- END FASTSPEED SUBROUTINE

RPL Example (cont.)

FASTSPEED; -- MOVES FROM POSITION 1 TO 5
PMOVE(PT1);
SLOWSPEED;
PICKUP;
PMOVE(PT5);
SLOWSPEED;
DROPOFF;

FASTSPEED; -- MOVES FROM POSITION 2 TO 6

PMOVE(PT2);
SLOWSPEED;
PICKUP;
PMOVE(PT6);
SLOWSPEED;
DROPOFF;
…
 K INEMATICS AND W ORKSPACE OF S ERIAL ROBOT A RM

MEC325/580, Spring 2010 I. Kao

This handout explains the kinematics and workspace of a two-link planar serial manipulator, or robot arm.
Analysis of a two-link arm, shown in Figure 1, is the most basic kinematic analysis for serial robots. The
IBM 7545 SCARA robot in the lab has a similar kinematics. Here, we will discuss the forward kinematics
and workspace of such robot arm.

effector point



Kinematics: As shown in Figure 1, the Cartesian coordinates of the end-
can be obtained as follows
Y P(x, y)
L2
    
  !" θ2

   $#&%' ()$#&%
  !" (1)

(2) L1

* is measured from the + axis and  is measured θ1

Note that the angle
,
from the extension line of the base link . With equations (1) and (2),
O
X

we can define the Jacobian matrix which relates the infinitesimal dis-

placement in the Cartesian space ( -. ) to that in the joint space ( - ) as
Figure 1: A 2-link planar manip-
ulator.
follows

/01-.  324 4075 6 4 4095 8 =<?> ''  $#&%'   > 7#&%@

    !!AA > BB$#&%

    !!""  C
- 44 0 : 6 4 4$0 : 8; (3)

where -.
ED -  - GFIH is the vector of the Cartesian coordinates, and - JED 
"K!7FIH is the vector
containing the joint coordinates. The Jacobian matrix relates the infinitesimal displacement - (in radians)

to the resulting infinitesimal displacement -. in the Cartesian space by the following equation

L
 B
/ 0 M
 @
 N < -  P/0Q< - 
-. - - OC - !RC (4)

as that in equation (4) with

/0 - 
An important note is in order here. Note that any time when the angles are involved in direct algebra, such
, you have to use angles in “radians” instead of “degrees.”
S

 $  TU
WVYXZ![\]Z
XZ!$XZ^  _XO`bac
Example of forward kinematics and Jacobian: Two configurations of the 2-link arm are given at
and , with and .
  LXO`bVc
From equations (1) and (2), the coordinates of the end-effector and the Jacobian matrices are:

At

 A$!A TU
WVYX Z [\ Z ed <  C  < XOOX `b`b\]\]fh]f]gYhV C /0 <?> XOXO`b\]`b\]fh]gYf]V h > OX XO`jX`bi]g]h]gYf]a h C



 A
$
 
 A
 T
U
 

X Z $
 X Z e
d <   < XO`bfYX 
/ 
0  < X X
At C X C XO`bfYX XO`bVYX C

that configuration cannot move along

,
Note that the Jacobian matrix for the second configuration is singular. This is because the two-link arm at
direction instantaneously. When this happens, we call the robot
being at a singular configuration. In fact, you should be able to prove that the Jacobian matrix is always

1
with
! _X
singular (i.e., the robot is at a singular configuration) whenever the distal link is aligned with the base link,
.

With infinitesimal angular displacement of -   D XO`jX
XO`jX
FH -  

 (note the angles are in “radi-

ans”), the displacement of the end-effector in the Cartesian coordinates for the two configurations can be
obtained from equation (4), and are
 XO`jX]XhgYf]a  X
-.  < --
 C  9< > OX `jX]Xag[f C c -.  < --
 C  < OX `jX
iYXC c
and (5)



respectively. Again, we find that the two-link manipulator can not move in the -direction instantaneously
at configuration 2 because - is identically zero.

Workspace: Workspace of a robot (or called the work envelope) represents the space within which the robot
can reach without singularity. The boundary of the workspace represents the singular configuration of the
robot. The workspace depends on the angular range of the two angles of the arms, and . The typical
 !
workspace for a two-link arm is illustrated in Figure 2 by the shaded region.
Y

θ2
L2

L1

θ1
X
O
 and  
L L1 2
X  G 

[h]Z and X ! 
hYXZ .
Figure 2: The workspace of a two-link manipulator with link lengths and the range of

2
 Manufacturing Automation:
Programmable Logic Controller
(PLC)

Imin Kao
Professor
Department of Mechanical
Engineering
SUNY at Stony Brook

Programmable Logic Controller

•  History of PLC
–  First introduced in 1970
–  Used for automated factory which provide
system reliability, product quality, information
flow, reduced costs, efficiency, and flexibility
–  Today's PLC are designed using the latest in
microprocessor design and electronic circuitry
which provides reliable operation in industry
applications
 Advantages Over Other Devices
•  PLC offers many advantages over other
control devices such as relay, electrical
timers/counters. Advantages include
–  Improved reliability
–  Smaller space required
–  Easier to maintain
–  Reusable
–  Reprogrammable if requirements change
–  More flexible-performs more functions

Schematic of PLC

I/O CPU &

Memory Programmer

User Supplied
Field Devices Basic PLC Block Diagram
 Nomenclature & Example

(See handout)

Ladder Logic Diagram

•  Example of
programming PLC
using a ladder login
diagram
PLC Logic Table: an example
Logic States
X1 1 1 0 0
X2 1 0 1 0
C1 1 1 0 1
S1 1 1 0 1
T1 1 1 0 1
S2 1 1 0 1
 MEC326/580 H ANDOUT: P OSITIONING S YSTEMS AND ACCURACY

MEC325/580, Spring 2010 I. Kao

This handout discusses the positioning systems and accuracy [1]. Positioning systems, depending on their
control scheme, can be broken into two categories: (i) open-loop positioning systems, and (ii) closed-loop
positioning systems. Furthermore, they can also be either linear positioning or rotational positioning. Most
electromechanical motors are rotary, with some capable of delivering linear motion directly. For closed-loop
positioning system, encoders (both rotary and linear) are typically used to provide the positions for feedback
control. In the following, positioning and accuracy of motorized systems are discussed.

Open-Loop Positioning: Stepper motors are typically used for open-loop linear or rotary systems. For
example, a XY table utilizing two leadscrews in orthogonal directions to index (x, y) position on a plane
can be made an open-loop positioning system. The step angle is determined by the stepper motor—usually
comes in 1.8◦ or 0.9◦ . For example, a 0.9◦ stepper motor has a total of 400 steps per revolution. The
following equation relates the step angle and number of steps of a stepper motor.
360◦
α= (1)
ns
where α is the step angle in degrees and ns is the total number of steps in one full revolution of the motor.
In the previous example, we have ns = 400 and α = 0.9◦ .

Note that a open-loop positioning system counts on the motor to rotate without slip. If the motor slips
and misses counts (which can take place when the load is larger than the rated load and the torque generated
cannot consistently move from one step to another), the positioning error will accumulated.

The resolution (or the smallest linear displacement or control resolution) of the leadscrew driven by a
stepper motor can be determined by the following equation.
 α  lp
r = lp = (2)
360 ns
where l is 1 for single-thread screw, 2 for double-thread screw, · · · etc, and p is the pitch of the leadscrew
measuring the axial distance between two adjacent threads in the unit of inch/rev or mm/rev. Equation
(2) corresponds to one step of the stepper motor, and thus is the resolution of the linear positioning. The
total number of pulses, Np , (one pulse per step) needed for the required linear displacement, x, is
x x ns
Np = = (3)
r lp
The corresponding angle of rotation for the stepper motor is
x
θ = Np α = 360◦ (4)
lp
where the unit of θ is degree. The total number of revolution is
x
Rev = (5)
lp

When continuous motion is required, typically at a constant speed, the following equation relates the
required linear speed with respect to the rotational speed of the stepper motor.
N
v= (l p) (6)
60

1
where v is the linear speed in inch/sec or mm/sec, and N is the rotational speed of the stepper motor in
RPM. Conversely, the constant rotational speed of motor required to keep a constant linear motion is
v
N = 60 (7)
lp

Closed-Loop Positioning: All closed-loop positioning systems require sensory information for feedback
control. Typical sensory information of angular or linear displacement is provided by optical encoders or
potentiometers. DC servo motors are often used with optical encoders for the control of angular or linear
position and speed. Sometimes, tachometers are also used to provide information of angular speed, based
on the fact that back emf is proportional to the speed of rotation. Equations that describe the motion and
analysis are similar to those in equations (1) to (7).

Example: A 1.8◦ -stepper motor is connected to a leadscrew via a coupler connection for motion control of
a platform mounted and carried by the leadscrew. The single-thread leadscrew has a pitch of 5 mm. The
platform is to move a distance of 70 mm at a top speed of 7 mm/sec. Answer the following questions.

1. What is the smallest linear displacement that this motion system can realize?

2. Determine the total angle of revolution of the stepper motor, as well as the number of pulses, required
to move the platform over the specified distance.

3. What is the required angular speed of the stepper motor in order to achieve the top speed of the linear
motion?

Solution: Note the leadscrew is single-thread; thus, l = 1 in equations (2) to (7). Since the step angle is
1.8◦ , equation (1) gives the total number of steps per a full revolution as
360
ns = = 200 (8)
1.8

1. The smallest linear displacement that this motion system can realize is the resolution given by equation
(2). That is,
 
1.8 p 5
r=p = 5/200 = 0.025 mm = 25 µm or r = = = 0.025 mm (9)
360 ns 200

2. To move the entire distance of 70 mm, the total angle of revolution and number of pulses required are
x 70
Np = = = 2, 800
r 0.025
θ = Np α = (2800)(1.8◦ ) = 5040◦ = 14 rev.

respectively.

3. The angular speed of the stepper motors is given by equation (7):

v 7
N = 60 × = 60 × = 84 RP M (10)
p 5

2
Precision in Positioning: Three important measures of precision in positioning are (i) control resolution,
(ii) accuracy, and (iii) repeatability.

Control resolution is defined as the distance separating two adjacent control points in the axis of move-
ment. The control resolution is determined by the pitch of leadscrew, gear ratio, step angles (in the case of
stepper motor), and the angles between slots in an encoder disk. For a stepper motor system without gear
reduction, the control resolution is the same as r given in equation (2). If digital encoders are used in a
control system, the number of bits also affects the resolution–known as the quantization effect. If B is the
number of bits in the storage register (for example, the number of bits in the representation of encoder data,
say a 12-bit encoder), then the number of control points into which the axis range can be divided is 2B .
For example, a 12-bit encoder has 212 = 4096 control points and 4095 equal divisions. Assuming that the
control points are separated equally within the range, we have
L
s= (11)
2B −1
where s is the control resolution of the computer control system in inch or mm, and L is the range of axis
in inch or mm. Similarly, if the range of consideration is angular displacement with a range of angle of Θ
(Θ ≤ 360◦ ), then the angular control resolution is
Θ
sΘ = (12)
2B −1

The resulting control resolution of the positioning system is the maximum of the two values; that is,
CR = M ax{r, s} (13)
where r and s are calculated from equations (2) and (11), respectively. In general, it is desirable that s ≤ r.
In modern sensor and computer technology, this is typically the case.

In an actual environment, many practical factors influence the performance of the control of systems.
These factors include the backlash in the leadscrews or ball screws, backlash in the gearing and transmission,
and deflection and elasticity due to loading. If assuming a normal distribution with mean value being
zero, we can define the random nature of accuracy of positioning systems by the 3-σ principle with ±3σ
encompassing 99.7% of the population.

Accuracy is thus defined in a worst-case scenario in which the desired target point lies exactly between
two adjacent control points. An illustration in Figure 1 helps to visualize the situation. If the target was
closer to one of the control points, then the control system can be moved to the closer control point and the
error would be smaller. The accuracy of any given axis of a positioning system is the maximum possible
error that can occur between the desired target point and the actual position taken by the system, using the
3σ range,
Accuracy = 0.5 CR + 3σ (14)
where CR is the control resolution given by equation (13), and σ is the standard deviation of the distribution
of errors.

Repeatability is defined as the capability of a positioning system to return to a given control point that
has been previously designated. Therefore, the repeatability is given by
Repeatability = ±3σ (15)

3
 distribution of errors
Desired target
point
control control
point point
axis
accuracy=0.5(CR)+3σ repeatability= 3σ
control resolution= CR

Figure 1: Terminology for positioning and accuracy

Example: A closed-loop control system is assumed to have random errors which are normally distributed
(Gaussian) with a standard deviation of σ = 0.004 mm. The range of the workspace is 700 mm with 16-bit
storage register. The single-thread ball screw has a pitch of 5 mm, with a 1.8◦ -stepper motor. Determine (a)
control resolution, (b) accuracy, and (c) repeatability of the positioning system.

Solution: Apply the equations formulated above. For single-thread ball screw, l = 1.

(a) The deterministic resolution defined in equation (2) and quantization in equation (11) are
p 5
r = = = 0.025 mm
ns 200
L 700 700
s = = 16 = = 0.0107 mm
2B − 1 2 −1 65536 − 1
respectively. Since s < r, the control resolution is

CR = M ax{0.025, 0.0107} = 0.025 mm (16)

(b) The accuracy is given by equation (14),

Accuracy = 0.5(0.025) + 3(0.004) = 0.0245 mm (17)

(c) The repeatability is given by equation (15),

Repeatability = ±3(0.004) = ±0.012 mm (18)

References

[1] M. P. Groover Fundamentals of Modern Manufacturing: materials, processes, and systems Wiley,
second ed., 2002

4
 Shop Scheduling
with Many Parts

Imin Kao
Professor
Department of Mechanical
Engineering
SUNY at Stony Brook

Terminology
•  Sequencing:
–  The process of defining the order in which jobs
are to be run on a machine
•  Scheduling:
–  The process of adding start and finish time
information to the job order dictated by the
sequence

Sequence determines the schedule

Assumptions of Shop Scheduling
•  Each job is started on a machine as
soon as the job has finished all
predecessor operations and the
machine has completed all earlier job in
its sequence
•  All jobs are in the shop and ready for
processing at time zero (t =0)
•  Flow time = completion time

Definitions
•  Scheduling process variables:
–  N : the number of jobs to be scheduled
–  M : the number of machines; each job is
assumed to visit each machine once
–  Pij : set up and processing time of job i on
machine j (elements in the time matrix)
 Objectives
①  Minimize average flow time
②  Minimize the time required to complete
all jobs (Cmax = makespan)!
③  Minimize average tardiness
④  Minimize maximum tardiness
⑤  Minimize the number of tardy jobs
Choice of “objective” depends on tasks
and requirements

Permutation Schedule
•  Assumptions:
–  All machines process jobs in the same order
–  Nearly the optimal solution for flow shops
•  Given the sequence, scheduling is:
①  At time 0, the first job is started on machine 1
②  As soon as this operation is completed, the
first job begins on machine 2 & the second
job begins on machine 1
③  Repeat 1 & 2 until the last job finishes on
machine M
 Remarks
•  Permutation Scheduling: Need to
consider (N!) total of job sequences,
where N is the number of jobs
•  As N grows, (N!) grows even more!!
•  The Permutation Scheduling is not
suitable for too many jobs (N < 7~10)

N! = N (N-1)! … !2!1

Example: Shop Scheduling

•  Consider the set of jobs (N=3) and
processing times in the following time
matrix. The unit of time is in minutes.
M Lathe Milling Milling
N (m/c 1) (m/c 2) (m/c 3)
Job 1 2.0 3.5 1.5
Job 2 4.5 3.0 2.5
Job 3 1.5 1.5 5.0
 Example (cont.)
•  Process variables:
–  N = 3 (number of jobs)
–  M = 3 (number of machines: lathe, milling,
milling machines)
•  Gantt Chart is used to illustrate the
permutation scheduling of each job
sequence
•  First example: use the job sequence of
{1, 2, 3}

Example: Job Seq={1,2,3}

M
Lathe (m/c 1) Milling (m/c 2) Milling (m/c 3)
N

Job 1 2.0 3.5 1.5

Job 2 4.5 3.0 2.5
Job 3 1.5 1.5 5.0
Job Sequence = { 1, 2, 3 }
m/c 1 Job 1 Job 2 Job 3

m/c 2 Job 1 Job 2 Job 3

m/c 3 Job 1 Job 2 Job 3

0 2 4 6 8 10 12 14 16 18
time (minutes)
Makespan = 17
 Now, you do it …
M
Lathe (m/c 1) Milling (m/c 2) Milling (m/c 3)
N

Job 1 2.0 3.5 1.5

Job 2 4.5 3.0 2.5
Job 3 1.5 1.5 5.0
Job Sequence = { 3, 2, 1 }
m/c 1 Job 3 Job 2 Job 1
m/c 2 Job 3 Job 2 Job 1
m/c 3 Job 3 Job 2 Job 1

0 2 4 6 8 10 12 14 16 18
time (minutes)
Makespan = 14

About the Makespan

•  The “makespan” must accommodate:
①  The delay before the machine j can begin
processing
②  The total processing time on the machine j
③  The remaining processing time for the last job
after it leaves the machine j
•  Calculate the “theoretical” lower bound
(LB)j based on machine j!
 Theoretical Lower Bound (LB)
•  A lower bound (LBj) based on machine j is
$ j "1 ' N $* M '*
LB j = min%# pir ( + # pij + min% # pir (
i *
i
& r =1 ) i=1 & r = j +1 *)
•  where pij is the (setup+processing) time of job i on
machine j, N is the number of jobs, and M is the
! number of machines. The largest lower bound
amongst all (LBj) is the lower bound for reference.
That is,
LBref = max{LB1, LB2 , ! , LBM }

Calculating Theoretical LB
M
Lathe (m/c 1) Milling (m/c 2) Milling (m/c 3)
N

Job 1 2.0 3.5 1.5

Job 2 4.5 3.0 2.5
Job 3 1.5 1.5 5.0
Machine 1 to 3 (j= 1, 2, 3)
(LB)1= 0+(2+4.5+1.5)+min{(3.5+1.5),(3+2.5),(1.5+5)}= 13
(LB)2= min{2,4.5,1.5}+(3.5+3+1.5)+min{1.5,2.5,5}= 11
(LB)3= min{(2+3.5),(4.5+3),(1.5+1.5)}+(1.5+2.5+5)+0= 12

LB = max{13,11,12} = 13 minutes
 Remarks
•  The theoretical lower bound is 13
minutes. Thus, job sequence {1,2,3},
having makespan of 17 minutes, most
likely is not optimal/minimum
•  There are a total of 6 (3!=3!2!1=6)
permutations of job sequence.
•  The job sequence of {3,1,2} has a
makespan of 13.5 minutes ! minimum
makespan

Procedures of Permutation Scheduling

①  Establish the time matrix based on data/
time of machining, including set-up time
②  Determine (LB)1, (LB)2, … , (LB)M for the
M machines
③  Determine the theoretical lower bound
LBref = max{LB1, LB2 , ! , LBM }
④  Draw the Gantt chart based on permutation
scheduling (a total of N! Gantt charts)
!Note: there may be zero in the time matrix
 ME325/580 Handout: Shop Scheduling with Many Products
Spring 2010 I. Kao

Consider the set of jobs and processing times shown in the following table for three jobs on three
machines. Generate the schedule assuming jobs are processed in the order of {1, 2, 3}. The unit
of the time in the following time matrix table is in minutes.

Milling (machine 1) Lathe (machine 2) Milling (machine 3)

Job 1 2.0 3.5 1.5
Job 2 4.5 3.0 2.5
Job 3 1.5 1.5 5.0

Solution:
The solution of the processing order {1, 2, 3} is summarized in the Gantt chart below. We start
by assigning job 1 to machine 1 at time 0. Since p11 =2.0, the operation lasts until 2.0 minutes.
Since all jobs must go to machine 1 first, the other machines are idle and the other jobs are
queued. At 2.0 minutes, job 1 is loaded onto machine 2 and machine 1 starts on job 2, the
second job in the sequence. Machine 2 finishes job 1, p12 =3.5 minutes later (time is now 5.5
min.). Machine 1 is still busy with job 2; thus, while job 1 is begun on machine 3, machine 2 is
idle, waiting for job 2. The remainder of the schedule is shown in the figure.
Note that the schedule reflects the rule that machine j starts job i when job i is finished on
machine (j−1) and all jobs with earlier locations in the schedule have finished with machine j.
Mach 1 Job 1 Job 2 job 3
Mach 2 Job 1 Job 2 job 3
Mach 3 job 1 Job 2 Job 3
0 2 4 6 8 10 12 14 16

A few observations are in order:

• Machine utilization probably can be made higher with less idle time via change of job
sequence.
• Makespan is 17 minutes for {1, 2, 3} job sequence: machine 1 takes 8 minutes; machine 2
takes 9 minutes; machine 3 takes 11.5 minutes.
• Lower bounds of time span can be established for reference of process scheduling efficiency.

1
Improve the efficiency and reduce makespan

The makespan must accommodate:

(1) the delay before the machine can begin processing
(2) the total processing time on the machine
(3) the remaining processing time for the last job after it leaves the machine

Thus, a lower bound (LBj) based on machine j is

⎧ j−1 ⎫ N ⎧ M ⎫
LB j = min ⎨∑ pir ⎬ + ∑ pij + min ⎨ ∑ pir ⎬
i ⎩r =1 ⎭ i=1 i ⎩r = j+1
⎭

where pij is the (setup+processing) time of job i on machine j, N is the number of jobs, and M is
the number of machines. The largest lower bound amongst all (LBj) is the lower bound for
reference. That is,

LBref = max{LB1, LB2 ,  , LBM }

where the (LBj) terms are obtained from the equation above. This LBref is the lower bound that is
used as a reference in your design.

In this example, the lower bounds can be calculated as follows.

€
Machines 1 to 3 (j=1, 2, 3):
LB1= 0 + (2.0+4.5+1.5) + min{(3.5+1.5), (3.0+2.5), (1.5+5.0)} = 0+8+5 = 13
LB2= min{2.0, 4.5, 1.5} + (3.5+3.0+1.5) + min{1.5, 2.5, 5.0} = 11
LB3= min{(2.0+3.5), (4.5+3.0), (1.5+1.5)} + 9 + 0 = 12

Therefore, the reference lower bound is LB=13 minutes. Since the makespan for the job order
{1, 2, 3} is 17 minutes, we suspect that it can be improved though we may not necessarily be
able to reduce the makespan to 13 minute – the lower bound.
There are 6 (3!= 3x2x1) permutations of the job order. Employ the same method to the other 5
permutations, we find that the most efficient scheduling is {3, 1, 2} with a makespan of 13.5
minutes.

Exercise: Try to follow the above procedure to confirm that the {3, 1, 2} is indeed the most
efficient scheduling with a makespan of 13.5 minutes.

2
 H ANDOUT ON S TATISTICAL P ROCESS C ONTROL (SPC)

MEC325/580, Spring 2010 I. Kao

1 Statistical Process Control and Methodology

The “statistical process control” (SPC) uses various statistical methods to assess and analyze variations in a
process. SPC keeps record of production data, histogram, process capability, and control charts. Two control
charts are most widely used in SPC, which will be discussed in Section 2.

There are two types of variations considered in SPC: (1) random variations and (2) assignable variations.
The former is present if the process is in statistical control; the latter indicates departure from statistical
control. The control charts are used to identify when the process has gone out of statistical control, thus
signaling that some corrective actions should be taken. A process is out of control if there are significant
changes in either process mean or process variability.

2 Control Charts of SPC

The use of control charts is a technique in which statistics computed from measured values of a certain process
characteristics are plotted over time to determine if the process remains in statistical control. The chart consists
of three horizontal lines: a center, a lower control limit (LCL), and a upper control limit (UCL), as shown in
Figure 1. The process is said to be out of statistical control if sample is out of these limits.

Two types of control charts are commonly used in SPC. They are the x̄-chart and the R-chart. The x̄-
chart plots the average measured value of a series of samples, with LCL and UCL bounds corresponding to
3σ standard variation; whereas, the R-chart plots the range of each sample, with its corresponding LCL and
UCL.

In SPC, samples are taken at every designated time period (e.g., every 15 minutes) and certain number of
measurements (or parts) are taken per each sample. The variable m denotes the number of samples, and n
is the number of measurements (d1 , d2 , · · · , dn ) per sample, or the sample size that is designated in Table 1.
Therefore, for each sample, we can compute

i=1 di
Pn
x̄ = (1)
n
R = max{d1 , · · · dn } − min{d1 , · · · dn } (2)

The mean values of x̄ and the range are thus

Pm
j=1 x̄j
¯ =
x̄ (3)
Pmm
j=1 Rj
R̄ = (4)
m

1
Table 1: Constants for the x̄ and R charts. Note that the “Sample size” (n) is the number of measurement per
each sample.

Sample size x̄-chart R-chart

n A2 D3 D4
3 1.023 0 2.574
4 0.729 0 2.282
5 0.577 0 2.114
6 0.483 0 2.004
7 0.419 0.076 1.924
8 0.373 0.136 1.864
9 0.337 0.184 1.816
10 0.308 0.223 1.777
11 0.29 0.26 1.74
12 0.27 0.28 1.72
13 0.25 0.31 1.69
14 0.24 0.33 1.67
15 0.22 0.35 1.65
16 0.21 0.36 1.64
17 0.20 0.38 1.62
18 0.19 0.39 1.61
19 0.19 0.40 1.60
20 0.18 0.41 1.59

The equations for computing the upper and lower bounds are:
(
¯ − A2 R̄
LCL = x̄ − 3σ = x̄
x̄ − chart: ¯ + A2 R̄ (5)
U CL = x̄ + 3σ = x̄
(
LCL = D3 R̄
R − chart: (6)
U CL = D4 R̄

where the constants: A2 , D3 , and D4 are listed in Table 1. Note that Table 1 is listed according to the sample
size n, or preferably called the number of measurement, not the number of samples m.

The procedure for constructing the charts is in the following.

1. Compute the mean (x̄) out of n measurements, and the range (R) for each of the m samples using
equations (1) and (2).

2. Compute the grand mean x̄ ¯ , which is the mean of the x̄ values for the m samples using equation (3).
This will be the center for the x̄-chart.

3. Compute R̄, which is the mean of the R values for the m samples using equation (4). This will be the
center for the R-chart.

4. Determine U CL and LCL, based on equations (5) and (6) and the constants listed in Table 1.

2
 Sample number d1 d2 d3 d4
1 2.46 2.40 2.44 2.46
2 2.45 2.43 2.47 2.39
3 2.38 2.48 2.42 2.42
4 2.42 2.44 2.53 2.49
5 2.42 2.45 2.43 2.44
6 2.44 2.45 2.44 2.39
7 2.39 2.41 2.42 2.46
8 2.45 2.41 2.43 2.41
Table 2: SPC for 8 samples, each with 4 measurements

2.1 LCL and UCL with known mean and standard deviation

For some processes, the mean and standard deviation of the process may be known. Under such circumstances,
the parameters of the x̄-chart can be obtained as follows:

x̄ = µ (7)
3σ
LCL = µ − √ (8)
n
3σ
U CL = µ + √ (9)
n
where µ is the process mean, σ is the standard deviation of the process, n is the number of measurement (or
sample size), and √σn is the standard deviation of the sample mean.

Equations of LCL and U CL this section and the previous section have control limits set at 99.73% of the
samples at 3-sigma range.

3 An Example

Samples are collected from an extrusion process that is in statistical process control, and the diameter of the
extrudate is measured in cm. Eight samples are taken with a time interval of 15 minutes between each sample
for a duration of 2 hours. Four measurements (d1 to d4 ) are performed in each sample. The quantity x̄ is the
average of four measurements in each sample, and R is the range of measurements. The measurements are
tabulated in Table 2. Answer the following questions.

¯ , and the control limits (LCL and U CL) of the x̄-chart.

1. Find the grand average, x̄

2. Calculate the average of R, and the control limits (LCL and U CL) of the R-chart.

3. Construct the x̄-chart and R-chart.

4. Another sample was taken with four measurements as follows: {2.41, 2.50, 2.49, 2.55}. Determine if
the process is out of (statistical) control, based on the previously established control charts.

Solution: We first identify that the number of measurement per each sample (or sample size) is n = 4 with a
total number of 8 sample batches, m = 8. The average x̄ and the range R are calculated and shown in Table 3.

3
 P
x̄ = ( di )/4 R
1 2.440 0.06
2 2.435 0.08
3 2.425 0.10
4 2.470 0.11
5 2.435 0.03
6 2.430 0.06
7 2.420 0.07
8 2.425 0.04
Table 3: The tabulated data for x̄ and R

From Table 3, we find

¯ = 2.435
x̄ R̄ = 0.06875 (10)

1. From the above results, equations (5) and (6), and Table 1 with 4 measurements per sample (or the
sample size), we have n = 4. Thus, we compute:

¯ = 2.435
x̄ (11)
LCL = 2.435 − (0.729) × 0.06875 = 2.3849 (12)
U CL = 2.435 + (0.729) × 0.06875 = 2.4851 (13)

The x̄-chart is plotted in Figure 1.

2. Similarly, we can compute for R-chart:

R̄ = 0.06875 (14)
LCL = D3 R̄ = 0 (15)
U CL = D4 R̄ = (2.282) × 0.06875 = 0.1569 (16)

The R-chart is plotted in Figure 1.

3. See Figure 1.

4. For the next sample with four measurements: 2.41, 2.50, 2.49, 2.55, we find
P4
i=1 di
x̄ = = 2.4875 R = 2.55 − 2.41 = 0.14 (17)
4
Hence, it is out of control (due to the x̄-chart).

4 Control Charts for Attributes

In addition to the x̄-chart and R-chart, two other charts are used for attributes. They are used to monitor the
number of defects present in sample, or the fraction of defect rate, for example, the number of defects per
automobile, existence or absence of flash in a plastic molding. Two types of charts are used: the p-chart and
the c-chart.

4
 x-chart
2.490
2.480 UCL= 2.4851
2.470
2.460
2.450
2.440
2.430
2.435
2.420
2.410
2.400
2.390 LCL= 2.3849
2.380
1 2 3 4 5 6 7 8

R-chart
0.160

0.140
UCL= 0.1569

0.120

0.100

0.080
0.06875
0.060

0.040

0.020

0.000 LCL= 0.0

1 2 3 4 5 6 7 8

Figure 1: The x̄-chart and R-chart

4.1 The p-chart

The p-chart plots the fraction defect rate in successive samples. The “p” stands for proportion which is defined
as
di
pi = (18)
n
for m samples of equal size n, where di is the number of defective items, and n is the number ofqparts in
p̄(1−p̄)
sample. The parameters are calculated based on binomial distribution with standard deviation σ = n .
For m samples of equal size n, the center and control limits are

i=1 pi
Pm
p̄ = (19)
ms
p̄(1 − p̄)
LCL = p̄ − 3 (20)
n
s
p̄(1 − p̄)
U CL = p̄ + 3 (21)
n

If LCL < 0 then use LCL = 0 in equation (20).

5
4.2 The c-chart

The c-chart plots the number of defects or flaws per sample. The c stands for count. The parameters of the
c-chart are based on the Poisson distribution. They are

i=1 ci
Pm
c̄ = (22)
m√
LCL = c̄ − 3 c̄ (23)
√
U CL = c̄ + 3 c̄ (24)

where ci is the number of imperfections or number of events occurring within a defined sample space (e.g.,
defects per car). If LCL < 0 then use LCL = 0 in equation (23).

References

[1] M. P. Groover Fundamentals of Modern Manufacturing: materials, processes, and systems Wiley, second
ed., 2002

[2] S. Kalpakjian and S. R. Schmid Manufacturing Processes for Engineering Materials Prentice Hall, fourth
ed., 2003

6
 Statistical Process Control
(SPC)

Imin Kao
Professor
Department of Mechanical
Engineering
SUNY at Stony Brook

What is SPC?
•  SPC
–  is based on the 3-! principle
–  uses various statistical methods to assess and
analyze variations in a process
–  keeps record of production data, histogram,
process capability, and control charts
•  Two Types of Variations Considered
1.  Random variations
2.  Assignable variations
 SPC and Control Charts
•  Control charts are used to identify when
the process has gone out of statistical
control ! require corrective action
•  A process is “out of control”, if there are
significant changes in
–  Process mean, or
–  Process variability

Control Charts
•  Three horizontal lines
–  Center
–  LCL: Lower control limit
–  UCL: Upper control limit
 Control Chart: Types
•  Two Types:
–  The x " chart
–  The R-chart

Variables for Control Charts

•  SPC takes samples at designated time
interval (e.g., every 15 minutes)
•  Variables:
–  m: the number of samples
–  n: the number of measurement per sample (or
the sample size)
–  Example: takes samples every 15 minutes for 8
hours, each time with 8 parts in the sample !
m = 32; n = 8
 Computing the Parameters
The n measurements are denoted as: d1, d2, … dn

Computing LCL and ULC

•  The LCL and UCL for the two charts
 Constants for the Charts

Note: The tables is

based on the “Sample
size” (n, or the number
of measurement in
each sample), not “m”

Procedures
①  Compute the mean out of n
measurements, and the range R for
each of the m samples
②  Compute the grand mean for the m
samples ! center of the x-chart
③  Compute the mean of range for the m
samples ! center of the R-chart
④  Determine UCL and LCL to complete
the charts
 Example
SPC for 8 samples, each with 4 measurements

Example (cont.)
 Control Charts

UCL & LCL with known Std Dev

 Control Charts for Attributes
•  The p-Chart
The proportion is defined as:

with

Control Charts for Attributes

•  The c-Chart