DESIGN FOR MANUFACTURE (DFM)

AND DESIGN FOR ASSEMBLY (DFA)

Bibliography
Anderson, D. M., Design for Manufacturability Optimizing Cost, Quality, and
Time-to-Market, CIM Press, Lafayett, California, 1990.
Ashby, M. F., Material Selection for Mechanical Design, Pergamon Press, 1992.
Boothroyd, G., P. Dewhurst, W. Knight, Product Design for Manufacturing and
Assembly, Marcel Dekker, 1994.
Corbett, J., M. Dooner, J. Meleka and C. Pym, Design for Manufacture, Addison-
Wesley, 1991.
Eggert, R. J. Engineering Design, Prentice Hall, 2005.
El-Wakil, S. D., Processes and Design for Manufacturing, Thompson Publishing,
1998
Farag, M. H, Materials Selection for Engineering Design, Prentice Hall, 1997.
Poli, C., Design for Manufacturing, a Structured Approach, Butterworth-Heinemann,
2001. (good detailed in-depth explanation of several mfg processes)

OUTLINE
Economic Comments
General Comments Regarding Design
Design for Manufacturing
Design for Assembly
Lean Design and Lean Manufacturing

ECONOMIC COMMENTS
How well parts and assemblies are designed for manufacturing and assembly is literally a
matter of life or death for a company. It is essential for engineers to have a working
knowledge of how to produce a high quality product cheaper, better, and faster.
Customers demand value meaning they need a product that works well for them at a
reasonable price.

The following equation use to be the norm in many industries for determining sales price:

Cost + Profit = Price

In other words, if the cost of producing a product increases, that cost can be passed on to
the consumer (customer). In a highly competitive world, that is not valid. The equation
that drives commerce has become:

Price (fixed) Cost = Profit

In a competitive market, the price is fixed and will likely fall in the future. Therefore, the
only way to increase profit is to decrease cost. This does not mean that the cheapest

DFM/DFA 1
solution is the best solution it means selecting the design that provides the best value.
Value is what customers demand. It is a combination of both cost and quality.

There is tremendous waste in almost every corner of every company. That is the good
news! There are many opportunities to reduce waste. Waste increases costs without
adding value to the product. By reducing waste, costs decrease and profits increase.

The engineers job is to determine the most economical way of producing a sufficiently
high quality product. The good news is that quality and low cost are not mutually
exclusive. Striving for optimal value is the challenge facing engineers.

GENERAL COMMENTS REGARDING DESIGN:

Keep it simple! Simple things are easier to produce and maintain.
Keeping it simple may be difficult (but engineers love a challenge).
Use standardized or interchangeable parts whenever possible
Use off-the-shelf items when ever possible. They are often cheaper and better quality
than you can produce in-house (why?).
Take advantage of vendor expertise. Foundries know the casting business, machine
shops now machining, etc. Team up with them.
Use as few of parts as possible, and where reducing total number of parts may not be
possible, use common parts (identical) where possible.

Common parts and materials saves $$$. Unique parts and materials that are purchased
must be:
- Purchased (a Purchase Order (PO) must be produced and processed)
- Received by someone somehow
- Inspected by someone somehow using something
- Moved to storage by someone somehow using something
- Stored (somewhere floor space costs money, inventory costs money)
- Moved to assembly or fabrication by someone somehow using something
- Handled by the assembler or fabricator
- Installed by the assembler somehow using something
- Installation must be inspected by someone somehow using something
- Inventoried by someone somehow using something (kept track of)
- Paid for by someone somehow (book keeping expense)

Each unique part made in-house must be:

- Designed by someone somehow using something
- The engineering drawings must be maintained by someone somehow using
something
- Engineering drawing must be interpreted by the mechanics and inspectors
- Parts require setup time to be produced
- Must be inspected by someone somehow using something
- Must be transported by someone somehow using something
- Must be tracked through the factory by someone somehow using something

DFM/DFA 2
 - Must be stored by someone somehow using something
- Must be transported to assembly by someone somehow using something
- Must be installed by someone somehow using something
- Installation must be inspected by someone somehow using something

What opportunities exist to make two or more parts into a single part? Good candidates
include (Corbett, et al.)
- no need for relative motion
- no need for subsequent adjustment between parts
- no need to disassemble for service or reparability
- no need for dissimilar materials

However, in the quest for combining parts into a single part, be careful NOT to:
- Create a large part that is difficult to manipulate or move
- Create a complex part that is difficult to fabricate
- Eliminate parts that provide fail-safe redundancies

Example:
While combining parts to make a single new part has advantages, the next best thing is to
use common parts. As an overly simple example, consider the fasteners required to
mount the plate to a support shown below. The service load is known to act only
downward at the end. Bolt A has to carry significantly greater load than Bolt B, and
therefore Bolt B may be smaller. Lets look at the advantages and disadvantages of
selecting the bolts to be the same size or different sizes.

Bolt B
Bolt A
Load
ADVANTAGES
Bolt B same as Bolt A Bolt B smaller than Bolt A
Having a smaller bolt at B may save bolt
costs (small bolts are slightly cheaper,
usually).
Slight weight savings
Plate may be oriented in either direction Would require the plate to be oriented a
(symmetric) specific direction (may be desired, may
not).
There are purchasing advantages to buying large
quantities of one size rather than smaller
quantities of two sizes (bulk discounts).
Only one size of bolt must be tracked in inventory
The mechanic only has to drill one size hole
The mechanic needs only one size of wrench

DFM/DFA 3
The mechanic needs to know only one torque
specification nothing to get confused
The assembly area needs to only store one size of
bolt, nut and washer
Common size would make automation easier
During servicing (maintenance) only one wrench is
required to remove and reinstall the fasteners.
The service department needs to stock only one size
of fastener.
Provides only one fastener rather than two to have
problems with (such as missed delivery date
from vendor, bad batch, etc.).
While the above example might seem obvious and overly simple, it becomes more
complicated when we consider larger systems; for example, an automobile. How many
different sized fasteners should a single car have? How many different sized fasteners
should Ford or General Motors vehicles have? The answer: as few as reasonable.
Obviously, one does not want to use 12mm bolts on a instrument panel when a 2mm
screw will suffice.

How can part count be reduced?

Design parts for multi-function. For example, a tube can be both a structural member as
well as a fluid delivery device.

Design for multi-use. Extending the above example of fasteners and brackets to other
designs within your company can reduce overall part count. For example, can similar
designs use the same fasteners as you have used? What about the bracket? Is this an
item your company uses frequently? If so, can a general purpose bracket be developed
for most of these applications? What about the material used to produce the bracket?
Will one material work for a variety of parts including basic dimensions? If so, you may
need to inventory or only a few materials and stock sizes.

Think about future designs. Are there features you can add now that may extend the
utility of the part to other designs?

Use modular designs. Modular designs consist of subassemblies that can be used in a
variety of applications. For example, printer heads on desktop printers may be made
common amongst many different models. This allows for improvements in the module to
be applied quickly across the product line. It also improves serviceability of the product
(service departments have lower inventory and training requirements.)

DFM/DFA 4
DESIGN FOR ASSEMBLY (DFA)
Why should a design engineer worry about how his or her design will be produced? The
profitability of their employer, and hence their own job, may depend upon proper design
for assembly and manufacturing, thats why. Before beginning detailed part design
(which should involve design for fabrication), the engineer should first consider design
for assembly. This seems backwards how do you design for assembly before
designing the parts? Design needs to be iterative. If the parts are designed first, there is
little that can be done to improve assembly. Concepts should be developed in order to
have a good sense of what the individual parts will be, but before extensive detailed
design is conducted, the concepts of design for assembly should be employed.

As discussed above, one aspect of DFA is to combine multiple parts into a single part
then there is less to assemble. According to Corbett et al., assembly costs can account for
between 40 to 60% of total production costs, therefore, reducing part count is important.
Corbett offers the following suggestions to reduce assembly costs:

Minimize part count (fewer things to assemble)

Use modularization
Orientation should be fool proof and easy
Locating parts should be made easy and fool proof
Sufficient space to allow access with tools and/or hands
Use common parts
Do not have parts that can tangle (star washers, wires, springs, etc.)
Fasteners are a pain in the but they are sometime necessary. They may be
avoided in certain circumstances where snap-together parts are adequate.
Welding can be a viable alternative. Also, there are a variety of fasteners
designed to minimize assembly costs. Make sure one person can install
fasteners fasteners should be able to be installed from one side
Provide lead in chamfers for parts to be inserted
Avoid visual obstructions
Parts should be assembled from one direction.

Anderson offers the following lists:

Errors of commission installing the wrong part, installing the wrong orientation or
position, damaging the part.
Use standard parts
Make different parts obvious
Make sure the wrong part can NOT go into the wrong place
Design the part so it can NOT be oriented improperly
Revisions (changes) to the product (part) are clearly communicated to the
manufacturer and implemented.

Errors of omission leaving out parts or operations

Design so omissions cannot happen use geometries or special features that
prevent subsequent assembly if prior parts are not installed (simple example: a
belt cannot be installed until the pulley has been.)

DFM/DFA 5
 Design so that omissions are easily noticed use of color or geometry to make it
obvious that something is missing. Shadow boards are used to identify when
tools are not replaced, the same can be true of an assembly.
Eliminate process steps that rely on operators memory keep the assembly simple.

Sequence errors errors consisting of sequence of events.

Design the assembly or process so that sequence does not matter
Design so assembly or process steps cannot happen in the wrong order
Design so assembly sequence is obvious
Clearly specify assembly sequence

DESIGN FOR MANUFACTURING (DFM):

The challenge is to design the part to function properly and to be produced economically.

However as always, when defining a problem, lets be careful not to suggest a solution.
So far we have assumed our problem is we need to determine the best design to be
manufactured. A better problem statement might be we need a part that performs
functions xyz, what is the best way to obtain such a part? That begs the question do we
need to manufacture it, or can we purchase it? We should apply the following rule
before worrying about design for manufacturing:

- Do not make what you can purchase off the shelf

o Vendors have expertise you may not have
o Takes advantage of high volume production (vendor sells to others)

- But there will come a time when we do need to worry about design for
manufacturing. Before a part can be designed for manufacturing, the engineer
must first determine what materials are viable choices. Material selection is
critical. It involves performance (loads, environment (corrosion, thermal) etc.) as
well as manufacturing methods. Selection of materials and processes is an
iterative process. Engineers should not work in a vacuum design should be
done as a team. Design teams should include people with experience in
manufacturing methods, materials, purchasing, sales, management, and it often
should include vendors.

Three factors determine the best manufacturing process to be used for a given part: the
material the part is to be made from, the geometric features of the part, and the quantity
of parts to be produced. Brittle materials, for example, cannot be formed by bending or
cold working.

Every part to be produced has a certain amount of information content. Information

content can be quantified as the number of dimensions required to define the geometry.
When designed properly, increased information content does not have a significant effect
on cast parts, but can have a very profound impact on machining costs. However, the

DFM/DFA 6
initial capital expense required to produce foundry dies can be quite high. If only a few
parts are to be produced, die casting does not make sense as the cost of the dies cannot be
justified. A machined part can have relatively high piece cost (cost to produce a single
part) but if only a few parts are to be fabricated, it may be more economical than casting.

The purpose of this section is to provide information regarding what sorts of processes
produce what sorts of parts. The following manufacturing methods will be discussed:
- Polymer Processing
- Metal Casting Processes
- Sheet Metal Processes
- Metal Shaping Processes
- Joining Processes

Polymer Processing
Polymers are generally the least expensive of all engineering materials and provide
highly economical finished products. They are the best choice for many designs.

While polymers may be machined, the vast majority of polymer parts are produced by
molding. Thermoplastics (polymers that melt) are commonly formed with injection
molding, thermoforming, extrusion, or extrusion blow molding. Thermosetting polymers
(polymers that become solid due to cross-linking and do not melt) are formed by
compression molding and transfer molding. In all of these processes, the polymer takes
on the final shape due to direct contact with tooling (die, mold, or mandrel). Heating and
cooling and/or curing (cross-linking) are part of all of these processes, and hence
shrinkage is always an issue.

Design for molding (Eggert, El-Wakil)

Avoid designing parts with thick walls or heavy sections
Design parts without undercuts
Provide generous fillet radii
Ensure holes and similar features do not require complex tooling
Provide appropriate draft
Avoid large changes in thickness (including bosses)
Choose material for minimum tooling, processing, and material costs
Design external threads to lie on parting plane/surface
Add ribs for stiffening

Injection Molding (Poli)

Thermoplastic pellets are heated (melted) and injected into a metal cavity (mold) to
produce the desired shape. A critical feature that dictates processing time is wall
thickness. The part cannot be removed from the injection mold until it has sufficiently
cooled. Thicker walls require longer time to cool and hence, they take longer to produce.
Parts must be removed from the mold, and part features that prevent easy removal
increase the cost of dies and increase process time. An example of such a feature is

DFM/DFA 7
shown below (a cup with a hole in the side). A complex mold is required to produce the
side hole.

{Discuss some dos and do nots}

Thermoforming
While heated, thermoplastic sheets are squeezed between two dies. This process
produces cup-like parts (thin walled, convex/concave parts without undercuts).

Extrusion
Thermoplastic pellets are melted and pressed through an extrusion die to create long
uniform cross-section tubes, rods, sheets, etc. (remember making long star shapes with
play dough?).

Extrusion blow-molding

DFM/DFA 8
Extrusion blow-molding starts with a hollow thin-walled thermoplastic part (typically
produced by extrusion), entrapping it between two halves of a larger mold, and
expanding it while hot into the final shape. This is used for making plastic drinking
bottles and similar parts. This is analogous to blowing up a balloon inside a container
while under pressure, the balloon conforms to the shape of the container. If the balloon
could then be frozen in shape it would be an example of extrusion blow-molding.

Compression Molding
Thermosetting polymers are often actually partially thermoplastic (copolymers). They
may be solid (but soft) at room temperature and become softer upon initial heating. After
extended heating, the polymer cures (cross-linking) and becomes strong. Compression
molding is a forming process used with such materials.

A piece of thermosetting polymer (called a charge) is placed in a cavity and heated. The
mold is closed, squeezing the charge so that it flows and fills the cavity. Heat is
maintained to cure the polymer. The process takes between 20 seconds for small thin
parts and up to 24 hours for very large thick parts.

Transfer Molding
Transfer molding is very similar to compression molding. The primary difference is that
the charge is placed in an external cavity, and once it is soft it is forced into the mold
through a sprue (similar a metal casting).

DFM/DFA 9
DFM/DFA 10
Metal Casting
Casting is well suited for parts with complex geometry,
parts with internal features that are difficult to machine,
and for moderate to high volume production. Casting is
effective for materials that are expensive and difficult to
machine as little raw material is wasted. It generally is
not competitive for parts that can be made from
extrusions or sheet metal, nor for very high melting point
metals like tungsten.

There is much similarity between metal casting processes and polymer processes such as
injection molding; therefore, the same general comments are applicable:

Design for molding (Eggert, El-Wakil)

Avoid designing parts with thick walls or heavy sections
Design parts without undercuts
Provide generous fillet radii
Ensure holes and similar features do not require complex tooling
Provide appropriate draft
Avoid large changes in thickness (including bosses)
Choose material for minimum tooling, processing, and material costs
Design external threads to lie on parting plane/surface
Add ribs for stiffening

Design for Casting; Poli and el-Wakil add the following comments:
Part geometry should allow for smooth flow to fill cavity evenly.
Cooling (solidification) should be quick to reduce cycle time and uniform to reduce
warpage
The above two comments suggest that avoiding abrupt changes in geometry. They
also suggest a balance on thickness too thick and the part shrinks substantially
and cools slowly, too thin and material may not flow sufficiently (recommended
minimum is 0.25 inch walls but 0.06 inch for investment casting).
Use reinforcing ribs to provide strength and stiffness where needed in webs (or
similar) rather than resorting to increasing overall part thickness.
Hot tears are caused by tensile stresses forming during cooling. These can be large if
the part is self-constrained or constrained by the mold. Within the geometry of the
part avoid over constraining.
If casting low ductility metals (eg. some cast irons) avoid projections that could be
easily broken.
Minimize features that need to be subsequently machined
Consider cast-weld construction (welding together cast parts) to avoid complex
expensive coring.

Common defects associated with castings include:

Inclusions (sand, slag, other foreign contamination)
Voids, porosity (caused by shrinkage)

DFM/DFA 11
 Lack of fill (molten metal not filling the cavity)
Poor microstructure (due to cooling rate, and mold-quenching)
Residual stress (due to differential cooling rates, abrupt geometry changes,
microstructural changes)
Hot tears (irregular crack with heavily oxidized surface created during cooling)
Cold shut (internal or surface crack likely caused where the flow of two molten
streams meeting while relatively cold hence not flowing into each other).
Shrinkage (surface sinks)
Quench cracks (created by subsequent heat treating of steel castings)
Defects may be found with X-ray, CT-scans and visual (for cracks)

Metals and alloys

Most metal alloys can be cast. But due to grain structure effects, shrink rates, high
viscosity (low fluidity), etc. many are not. Specific alloys have been engineered for
the sake of casting. Alloy elements to improve grain structure (grain refiners allow
for more smaller equiaxed grains) and fluidity are often employed with casting alloys.
In order to achieve desired properties, subsequent heat treating may be required.

There are three main casting processes:

Sand casting (sand is formed around a pattern, the pattern is removed and molten
metal is poured in its place).
Investment casting (a plaster (or similar) mold is formed around a wax pattern, was
is melted and molten metal is poured in its place)
Die casting (very similar to injection molding: the molten metal is injected under
pressure into a metal mold)

Sand Casting
Sand castings are typically used for large parts. Due
to the time required to pack sand, and to cut sprues,
runners and risers, sand castings are typically used for
low volume production.

Sand (and binder) is packed around a pattern. The

pattern is typically made from wood or metal and has
the same geometry as the completed part is to have.
To compensate for shrinkage, the pattern is made
slightly larger than the desired final part. The patterns can be a single piece (if one side
of the part is flat) or two pieces (cope top half, drag bottom half). For hollow parts, a
separate sand core is positioned inside the mold cavity.

Once the sand is packed in place, the pattern is removed. Passage ways must be cut in to
the sand to allow molten metal to flow into the mold cavity. A sprue is a cylindrical hole
cut through the cope into which the molten metal is poured. Runners allow metal to flow
from the sprue into the mold cavity. The entrance from the runner into the cavity is
referred to as the gate. Risers allow metal to flow out of the cavity and to some extent act
as a reservoir providing extra material to help compensate for shrinkage.

DFM/DFA 12
Due to the low conductivity of the sand, cooling rates are relatively low. This results in
large dendrite crystal formation that will affect material properties. Also, porosity (small
voids) is common in sand castings. The surface finish of sand cast parts is relatively
rough, and if a smooth surface is required castings must be subsequently machined.

Investment Casting
Used in relatively low volume production (less than 10,000
pieces). The process involves pouring wax into a mold,
cooled, and removed. The wax is then covered in a plaster
slurry (or similar). The plastic hardens and the wax is
melted and removed leaving a cavity in the plaster. Molten
metal is then poured into this cavity. Once it has hardened,
the plaster is broken away. Very good surface finish is
achievable and most parts do not need subsequent surface
machining. The image at the right is an investment cast
turbine blade.

Die Casting
This process is very similar to injection molding. Molten
metal is injected into a metal cavity (mold). It is an
economical process for high production volumes. Part
shown at the right are produced from die casting.

Forging
Forging is a bulk deformation process performed on
metals at elevated temperatures. Under large compressive
force, the metal is forced to fill a cavity.
It is a viable alternative for many
castings. However, if an internal cavity
is required in the part, then forging is
not likely to be an option. Forging
generally produces parts with higher
strength and ductility and less defects
than castings making the parts more
robust against impact and fatigue.
Castings generally are more isotropic
where as forgings have direction
properties due to elongated grain flow.

Design for Forging:

Due to high cost and limited life of tooling, forging is generally more expensive than
alternatives.
Avoid large section thickness changes (as with castings)
Incorporate large fillet radii
A 5 to 10 degree draft is required (similar to castings)

DFM/DFA 13
 Use easily formed materials such as aluminum alloys and copper alloys
Steel parts may be forged, but not as easy as aluminum. Very soft steel
(spheroidized) is more easily forged, but the subsequent heat treating will remove
any cold working effects.
Avoid external and internal undercuts. These features may not be possible to produce
with forging.

Sheet Metal
Sheet metal forming consists of shearing, bending, and/or deep drawing. Shearing is
performed to shape the outer geometry as well as to cut holes or other features within the
sheet. Multiple cuts can be made simultaneously. Bending is performed to shape the part
analogous to folding paper. Deep drawing pulls or stretches the sheet to form the
part in ways the bending cannot. With drawing, flat sheets can be formed into
hemispherical or other such geometry.

Sheet metal forming is limited to highly ductile materials in the form of thin sheets
(typically less than 0.25 inches, although thicker sheets may be fabricated). Typically,
rolled sheets of metal are used hence, even before being formed, the material will be
anisotropic. Since this is a plastic deformation process, once the part is bent or drawn,
elastic unloading will occur. Tight dimensional tolerances, therefore, may not be
achieved.

Design for sheet metalworking

Minimize manufactured scrap
Avoid designing parts with narrow cutouts or projections
Keep side-action features to a minimum or avoid complexity
Reduce number of bend planes

Defects in Deep Drawn Parts:

Tearing
Wrinkling
Orange-peel or Luders lines (surface texturing)
Punch and die marks on surface

Metal Shaping
Other metal shaping processes include rolling, drawing, and extrusion. Rolling is done to
reduce thickness substantially. Drawing is the process of reducing the diameter of a
wire, bar or tube by pulling it through a die of similar cross-section (Poli). Extrusions
form long parts with uniform cross section.

From Poli:
Rolling is usually the first process used after casting an ingot. Cast ingots are
rolled to form slabs (thick plates, say 40mm thick), billets (long thick rods with
square, rectangular, or circular cross-sections), and blooms. Slabs are then rolled

DFM/DFA 14
 into sheets, plates, and welded pipes, and billets are rolled and drawn into bars,
rods, pipes, and wires. Blooms are roll formed into structural shapes such as I-
beams and rails.

Machining
Machining processes remove material (by cutting) from a work-piece in order to produce
a desired form. There are many different machining processes, but the following are
basic:
Turning/Lathes diametral features are cut into the part with a semi-stationary cutting
tool while the work-piece is rotated.
Milling Machines Slots, pockets, recesses, holes and other features are cut into the
work-piece with a rotating cutting tool.
Boring and Drilling Drilling is used to cut holes, boring can enlarge existing holes and
do basic milling operations such as cutting grooves.

Design for machining, general comments

Use standard parts as much as possible.
Tight tolerances and smooth surface finish increase costs
Design the part to minimize quantity of material removal. Chips cost money to make,
and the material in the chips that you have purchased is wasted.
Workpiece must have a holding feature for turning operations, this would be a
uniform diameter, for milling operations a flat base.
Radii in finished parts should equal cutting radii of the tool.
Thin parts can cause problems with machining. For thinner parts, beware of
deflections that can be caused by cutting forces.
Use raw material available in standard forms (bars, sheets, rolls, etc.)
Employ standard features (holes, slots, chamfers, fillets, etc.)
Avoid sharp internal corners on turned parts
Allow for run out
For drilling, the surface should be perpendicular to the hole to be drilled.
For tapped holes, it is not possible to tap the entire length of a blind hole.
Internal features are generally more complicated to machine than external features.
The golden rule for designed parts to be machined is never deviate from the
primary axis (Corbett, et al.). This means machined features should be added
using one axis only (for milled parts, features should be added from one side only,
for turned parts, features should be diameter only)

Joining Process
Joining is done for one of three reasons: to combine parts that were not fabricated as a
single piece (due to complexity, weight, differing materials, fail safe design, etc.), to
allow for adjustments and/or relative motion, and to join parts that are designed to be
disassembled (for service, etc.). Some joining processes are permanent (welding,
brazing, adhesives, rivets, etc.) and others are removable (threaded fasteners). For load-
bearing structures, most joining processes involve a change in the nominal geometry and
therefore introduce stress concentrations.

DFM/DFA 15
Fastening is a very important part of almost every design and the design of joints is
critical. Fasteners typically transfer loads from one part into another, they must hold the
two (or more) parts in fixed position, they must withstand temperature changes and
vibrations which work to loosed threaded connections, and they may incorporate
differing materials (which can greatly enhance corrosion problems). Although fasteners
themselves are generally inexpensive (not always, some fasteners can cost thousands of
dollars) fastened joints add significant expense to a part. Holes must be drilled, they may
need to be de-burred or countersunk, fasteners must be inserted, wrenches applied to both
the nut and bolt, and finally tightened to the proper torque. Not only do these steps add
cost to the part, they introduce the possibility of defects at each step of the process.

Rivets
Riveting involves inserting a slug (cylinder) of ductile metal through holes
cut/drilled/punched into the parts to be joined. The slug is plastically deformed forming a
head and tail on the rivet that captivates the joint. Holes inherently are stress
concentrations, however, if the rivet is over expanded to not only fill the hole but to
expand the hole, compressive residual stresses can be produced in the hole to
significantly increase fatigue life.

Threaded Fasteners
Bolts (threaded fasteners tightened with a nut) and screws (threaded fasteners inserted
into threaded parts) are used for both for joints intended to be permanent as well as
removable. A wide variety of designs exist for threaded fasteners both for bolts and
nuts. Design problems arising from threaded fasteners include stress concentration at
the hole and loosening of the joint. Methods used to prevent loosening include thread
lock (adhesive) and mechanical locking (deformed threads, safety wire, cotter pins, etc.).

As with rivets, compressive residual stresses can be introduced into the hole. This can be
achieved by cold-expansion of the hole prior to fastener insertion or by press-fit fasteners
(specialty tight-tolerance shanks that are larger than the hole diameter) that cold-expand
the hole during insertion.

Other Fasteners
There are hundreds if not thousands of different types of fasteners; each intended to help
improve the performance of the joint or to reduce assembly costs.

Welding and Brazing

Welding is the process of melting the base metal and filler material (if used) to create a
solid joint. Brazing is the process of melting the filler material (braze) and creating a
metallic bond between the braze and base metal (base metal is not melted). Both
processes require substantial heating of the base material that can alter its microstructure.

Common Welding Defects

Due to high localized heating and subsequent microstructure changes, high residual
stresses, distortion and cracking are potential defects in welds. Distortion and cracking
are detectable with non-destructive techniques (x-ray, ultrasound) but residual stresses

DFM/DFA 16
remain hidden. The only way to know the effect of residual stresses is through
destructive testing. Additionally, inclusions (slag) porosity, undercuts, lack of fusion,
and lack of penetration can all degrade the structural quality of the weld.

Depending upon the alloy being welded, microstructure changes may include elimination
of precipitation hardening, annealing (eliminating cold-worked properties), and creation
of martensite, creation of dendrite crystals. To mitigate the effects of these typically
undesirable consequences subsequent heat treating may need to be performed. Heat
treating may be done locally around the weld, or the entire part may have to be re-heat
treated. Even then, it may not be possible for the weld to have the same properties as the
base metal (for example, if the base metal was coldworked).

Adhesive Joints
Adhesive bonding can be a strong inexpensive joining process. The base metal must be
cleaned and in instances roughened. Depending upon the adhesive material, the joint is
formed either by microscopic interlocking mechanisms (the adhesive flows in rough
areas and when solidified become mechanically locked) and/or through intermolecular
bonding.

Strength of an adhesive bond depends upon area of contact and the loading direction.
Adhesive joints are strongest in shear, but have low tensile and pealing strength.
Therefore, lap joints are generally used for these joints.

Shear Load (Good) Tensile Load

(not good can peel)

Tensile Load (not good)

Adhesive bonding has advantage over other joining techniques including room
temperature (no thermal effects such microstructure changes and residual stress), can
bond differing materials such as ceramics, and can bond dissimilar materials such as
polymers to metals, metals to ceramics, et cetera. Disadvantages include difficult to
inspect bond quality, lower strength than other methods (although it can be comparable
strength), degradation over time, and lower service temperature limits. Additionally,
cracks can form in the adhesive and propagate rapidly without detection.

DFM/DFA 17