Design Principles

In this section, basic design principles and general Figure 8.1 Specific strength
recommendations are presented to assist the design
engineer in designing components using KetaSpire ® KT-820 CF30
polyetheretherketone (PEEK) that meet the cost/
KT-880 GF30
performance requirements of their applications.
Guidelines are given on the effects of stresses caused KT-880 NT
by assembly, temperature changes, environmental
KT-820 NT
factors, and time as it relates to creep.
Zinc ZA8
Ultra-performance polymers, such as KetaSpire® PEEK,
are ideal candidates for metal-to-plastic conversion due Steel stainless 301
to their superior mechanical and thermal properties and Steel carbon 1018
higher specific strength. For many applications, weight
Magnesium AE42-F
reduction has many benefits, such as reduced friction,
lower energy usage, lower dynamic stress on Aluminum 380
the assembly, and lower system cost. Aluminum 2011-T6
The concept of material efficiency can be illustrated 0 20 40 60 80 100 120 140 160
by calculating specific strength and modulus. Specific Specific strength [ kNm/kg ]
strength is the material’s strength divided by its density
and can be related to the weight of material required to At times, designing with plastics may appear more
support a load. Table 8.1 lists the specific strength and complicated than with metals. But the diversity of
modulus of several common metals and selected grades products, conversion processes, and secondary
of KetaSpire® PEEK; the specific strength data is graphed operations (welding, inserts, printing, painting, metalizing)
in Figure 8.1. Specific strength can be used to compare the that are available with plastics give the designer
mass of material required to support a load. For example, unprecedented freedom.
KetaSpire® KT-880 GF30 has a specific strength of 106
A designer may be tempted to make a PEEK part that
kNm/kg (425 klbf-inch/lb) and carbon steel’s is 35 kNm/
merely duplicates the dimensions of a metal part without
kg (140 klbf-inch/lb), suggesting that for equivalent strength
taking advantage of the versatility of the material or the
the PEEK part would weigh about 1 ∕3 the weight of the
design freedom offered. This approach can lead to
steel part.
inefficient designs, parts that are difficult to produce,
For a successful conversion it is highly desirable to or parts with less than optimal performance.
understand the differences between metals and plastics
and the design flexibility of injection molding.

Table 8.1 Specific strength and modulus of common metals and KetaSpire® PEEK

Material Density Tensile Strength Specific Strength Tensile Modulus Specific Modulus

g/cm3 MPa kpsi kNm/kg klbf-in./lb GPa kpsi MNm/kg Mlbf-in./lb

Aluminum 2011-T6 2.83 270 39.2 95 383 70.0 10,200 24.7 99.3
Aluminum 380 2.76 159 23.1 58 231 71.0 10,300 25.7 103.3
Magnesium AE42-F 1.80 145 21.0 81 323 45.0 6,500 25.0 100.4
Steel carbon 1018 7.87 275 39.9 35 140 205.0 29,700 26.0 104.6
Steel stainless 301 8.03 205 29.7 26 102 212.0 30,700 26.4 106.0
Zinc ZA8 6.30 290 42.1 46 185 86.0 12,500 13.7 54.8
®
KetaSpire PEEK
KT-820 NT 1.30 95 13.8 73 293 3.5 510 2.7 10.9
KT-880 NT 1.30 100 14.5 77 309 3.7 540 2.8 11.4
KT-880 GF30 1.53 162 23.5 106 425 10.8 1,570 7.1 28.3
KT-820 CF30 1.41 201 29.2 143 572 19.7 2,860 14.0 56.1

64 \ KetaSpire® PEEK Design and Processing Guide

The following sections discuss those areas of mechanical The maximum stress (σ) occurs at the surface of the
design and stress analysis that relate to designing with beam furthest from the neutral surface and is given by:
plastics, comparing metal to plastics and discussing
factors that are specific to plastics alone. σ = Mc = M
I Z

Mechanical Design Where:

M = Bending moment, Nm ( lbf – in )
The use of classical stress and deflection equations may
c = Distance from neutral axis, m ( in )
provide starting points for part design. Mechanical design I = Moment of inertia, m4 ( in4 )
calculations for KetaSpire® resins will be similar to those Z = I = Section modulus, m3 ( in3 )
used with any engineering material. As with all plastics, c
however, the analysis used must reflect the viscoelastic Figure 8.3 gives the cross-sectional area (A), the moment
nature of the material. In addition, the material properties of inertia (I), the distance from the neutral axis (c), and the
can vary with strain rate, temperature, and chemical section modulus (Z) for some common cross sections.
environment or with fiber orientation for fiber-reinforced For other cross sections and/or geometries, the design
plastics. Therefore, the analysis must be appropriate for all engineer can consult stress analysis handbooks or
anticipated service conditions. For example, if the service employ finite element analysis.
condition involves enduring load for a long period of time,
then the apparent or creep modulus should be used Limitations of Design Calculations
instead of the short-term elastic modulus. Or if the loading For stress calculations, the application of the classical
is cyclical and long term, the fatigue strength at the design mechanical design equations are useful as starting
life will be the limiting factor. points, but some critical factors are simply not adequately
considered by these analyses. The viscoelastic behavior
The initial step in any part design analysis is the
of polymeric materials limits application of some of the
determination of the loads the part will be subjected to,
design equations to, for example, low deflection cases.
and calculation of the resultant stress and deformation or
Often the calculation of maximum stress contains a
strain. The loads may be externally applied or result from
number of simplifying assumptions which can diminish
stresses due to temperature changes or assembly.
the credibility of the results.
An example of an externally applied load is the weight
Also, the impact resistance of a design is directly
of medical instruments on a sterilizer tray. Examples of
related to its ability to absorb impact energy without
assembly loads are the loads on a housing flange when
fracture. It is difficult to predict the ability of a design to
it is bolted to an engine or the load on the hub of a pulley
absorb energy. In addition, even armed with the energy
when a bearing is pressed into it. Thermally induced
absorption requirements, practical toughness constants
stresses can arise when the temperature of the assembly
for engineering resins don’t exist. The results of laboratory
increases and the dimensions of the plastic part change
testing vary with the type and speed of the impact test,
more or less than the metal part to which it is attached.
even for fixed geometries. Therefore, the ability of the
Using Classical Stress-Strain Equations design to withstand impact must be checked by testing
the impact of prototype parts.
To use the classical equations, the following simplifying
assumptions are necessary: Similarly, fatigue test results will vary depending on the
cyclic rate chosen for the test, the dynamics of the test,
• The part can be analyzed as one or more
and the test specimen used. Therefore, they should only
simple structures
be used as a rough indication of a material’s ability to
• The material can be considered linearly elastic
perform in a fatigue application.
and isotropic
• The load is a single concentrated or distributed Deflection Calculations
static load gradually applied for a short time To determine the deflection of a proposed part design
• The part has low residual or molded-in stresses using classical equations, a modulus of elasticity value
is required. It is important that the appropriate value
While all of these assumptions may not be strictly valid for
be used. The value must represent the modulus of
a particular situation, the classical equations can provide
the material at or near the temperature and humidity
a starting point for analysis. The design engineer can then
expected in the application. Please contact your Solvay
modify the analysis to take into consideration the effects
representative for details.
of the simplifying assumptions.

A variety of parts can be analyzed using a beam bending

model. Figure 8.2 lists the equations for maximum stress
and deflection for some selected beams.

KetaSpire® PEEK Design and Processing Guide / 65

Figure 8.2 Maximum stress and deflection equations
Simply supported beam Cantilevered beam ( one end fixed )
Concentrated load at center Concentrated load at free end

F
FL F FL
σ= σ=
4Z Z
( at load ) ( at support )

FL3 FL3
Y= L Y=
48EI Y 8EI
L Y
( at load ) ( at load )

Simply supported beam Cantilevered beam ( one end fixed )

Uniformly distributed load Uniformly distributed load

F ( total load )
F ( total load )
FL FL
σ= σ=
8Z 2Z
( at center ) ( at support )

5FL3 FL3
Y= Y=
384EI 8EI
L L
( at center ) Y ( at support )
Y

Both ends fixed Both ends fixed

Concentrated load at center Uniformly distributed load

F ( total load )

F FL Y FL
½L σ= σ=
8Z 1Z
( at supports ) ( at supports )

5FL3 5FL3
Y= Y=
192EI 384EI
L ( at center )
Y ( at load )
L

66 \ KetaSpire® PEEK Design and Processing Guide

Stress Calculations To calculate the deflection, a value for E, the modulus, is
After the designer has calculated the maximum stress, required. In this case, the beam will bend under load and
the value is then compared to the appropriate material the most appropriate modulus is the flexural modulus.
property, i.e., tensile, compressive, or shear strength. The Values for the flexural modulus of KetaSpire® PEEK grades
comparison should be appropriate in terms of temperature can be found in the Typical Properties Tables or the
and humidity to the requirements of the application. product data sheets on the Solvay Specialty Polymers
website. There are two values of flexural modulus shown
To illustrate the use of the equations in Figures 8.2 and for each grade, one obtained by testing according to
8.3, the following examples are presented. ASTM D790 and one obtained by testing according to
ISO 178. Although these test methods measure the same
The case chosen for the examples is a cantilever beam
property, they vary in some of the details, and the values
with one end fixed and a concentrated load at the free
obtained are not identical. For this exercise, the value
end as shown in Figure 8.4.
obtained by ISO 178 was arbitrarily chosen.
Figure 8.4 Beam used for examples For this example, the deflection at room temperature will
F = 1 Kg ( 2.2 lbs ) Cross section be calculated for KT-880 NT, the high-flow, unreinforced
6 mm grade.
( 0.24 in )
The value for flexural modulus by ISO 178 is 3.90 GPa.
Assuming a load of 1 kg on the free end, the force
y becomes 9.8 N. Substituting these values into the
deflection equation:

L = 100 mm ( 3.94 in ) 25 mm
( 0.98 in. ) ( kg × 9.8 )( 0.1 m )3
Y= = 1.86 mm ( 0.073 in )
3( 3.90 GPa )( 450 mm4 )
From Figure 8.2, for a cantilevered beam with rectangular Repeating the calculation using the moduli of the glass-
cross section, the deflection is given by: fiber reinforced and carbon-fiber reinforced grades, the
deflections are shown in Table 8.2.
FL3
Y=
3EI Table 8.2 Deflections for a 1 kg load on a cantilever
beam
Where:

Y = Deflection KetaSpire® Flexural Modulus by Deflection,

Grade ISO 178, GPa (kpsi) mm (inch)
F = Force on the free end of the beam KT-880 NT 3.90 (566) 1.86 (0.073)
KT-880 GF30 10.6 (1,540) 0.68 (0.027)
L = Length of the cantilever
BG20
E = Modulus of the beam material KT-880 CF30 21.5 (3,120) 0.32 (0.013)

I = Moment of inertia of the beam

Table 8.2 shows that the calculated values for deflection
are inversely proportional to the flexural modulus; the
From Figure 8.3, the moment of inertia of a rectangular
highest modulus CF30 grade has the lowest deflection
cross section is given by:
under this load or in other words has the greatest
stiffness.
bd3
I=
12 As stated previously, the properties of plastic materials
are a function of the temperature at which they are
Where: measured. To provide guidance regarding the behavior
I = The moment of inertia at elevated temperatures, the modulus of KetaSpire®
PEEK has been measured at several temperatures. The
b = Width of the beam deflections at elevated temperature can be calculated
d = Thickness of the beam by using these moduli.

For this example, KetaSpire® grades KT-820 NT and

For the beam shown in Figure 8.4: KT-820 GF30 BG20 were chosen, and the values of
flexural modulus obtained using ASTM D790 were used.
25 × 63
The moment of inertia = = 450 mm4
12

68 \ KetaSpire® PEEK Design and Processing Guide

Table 8.3 Deflection as a function of material Designing for Equivalent Part Stiffness
and temperature Sometimes, a design engineer wants to replace a metal
part with one made of plastic, but still wants to retain the
Temperature Flexural Modulus Deflection
°C (°F) GPa (kpsi) mm (inch) rigidity of the metal part. From Table 8.1, it is apparent
that the modulus of the listed metals is much higher
KT-820 NT
than the modulus of the polymers. Depending upon the
23 (73) 3.77 (547) 1.93 (0.076)
polymer grade and the metal, the metal modulus is from
100 (212) 3.44 (499) 2.11 (0.083 2 to 20 times higher than the polymer modulus at room
150 (302) 2.35 (341) 3.09 (0.122) temperature. When the temperature is increased, the
200 (392) 0.32 (46) 22.69 (0.893) modulus of the polymer will decrease, while the metal’s
KT-820 GF30 BG20 modulus is essentially constant causing the ratio to
23 (73) 10.0 (1,450) 0.73 (0.029) become even higher.
100 (212) 9.34 (1,350) 0.78 (0.031) According to structural mechanics, the deflection of the
150 (302) 6.60 (957) 1.10 (0.043) part is inversely proportional to modulus. That means
200 (392) 2.26 (328) 3.21 (0.126) the deflection of a plastic part would be greater than the
deflection of a metal part in the same design. Naturally,
The results of these calculations are shown in Table 8.3. the difference in modulus has to be managed to achieve
Unfilled KetaSpire® PEEK resists serious deflection until the same deflection. It is easy to compensate for the
the temperature exceeds 150 °C (302 °F). The calculated modulus difference by changing the geometry of the
deflection at 200 °C (392 °F) indicates a serious loss of part. Ribs offer a very economical way to stiffen the part.
stiffness suggesting that the unfilled grade may not be In short, it is essential to redesign a plastic part for a
suitable for highly stressed applications at 200 °C (392 °F). structural application.
However, the glass-fiber reinforced grade continues
There are two fairly simple ways to maintain the stiffness
to resist serious deflection even at 200 °C (392 °F).
of a part when substituting one material with another
Reinforcing Fiber Orientation Considerations material, even though the materials have different moduli
of elasticity.
When designing with plastics, especially filled plastics,
the designer must be cognizant of the effects of the fillers In the first method, the cross-sectional thickness is
and reinforcing fibers on the mechanical properties of the increased to provide stiffness. In the second, ribs are
plastic. The processing of filled plastics tends to cause added to achieve greater stiffness. An example of each
orientation of fibers or high-aspect-ratio fillers parallel to approach follows.
the direction of flow.
Changing section thickness
Since the design of the part and the processing are In reviewing the deflection equations in Figure 8.2, the
interrelated, the designer should consider what portions deflection is always proportional to the load and length
of the part are likely to be oriented and how the properties and inversely proportional to the modulus of elasticity
will be affected. Shrinkage, strength, stiffness, and and moment of inertia.
coefficient of thermal expansion will differ depending
on the aspect ratio of the fiber (the ratio of its length For example, when both ends are fixed with a uniformly
to its diameter) and the degree of fiber orientation. distributed load, the deflection is determined by:
Perpendicular to the fiber orientation, the fibers act
more like fillers than reinforcing agents. FL3
Y=
384EI
When molding polymers, there are instances where melt
fronts meet (commonly known as weld lines) such as Therefore, to equate the stiffness using two different
when the plastic melt flows around a core pin. However, materials, the deflections are equated as follows:
the reinforcement in the plastic, if present, often does
not cross the weld line. Thus, the weld line does not have FL3 FL3
=Y=
384EI metal 384EI
the strength of the reinforced polymer and at times can plastic

even be less than the matrix polymer itself. These factors

must be taken into account when designing parts with
reinforced plastics.

KetaSpire® PEEK Design and Processing Guide / 69

Since the load and length are to remain the same, the FL3 if the moment of inertia of the ribbed design is equal to the
becomes a constant. Removing the constants from both plate design, the parts will have equivalent deflection and/
sides of the equation yields: or stiffness.

Equation 1 [ EI ] metal = [ EI ] plastic From Figure 8.3, the I for a ribbed section is selected. By
assuming that the section width “b” is the same for both,
If the metal part were magnesium with a modulus of
the Irib has to be equal or greater than the Iplate.
elasticity E of 45 GPa (6.5 Mpsi) and the thermoplastic
chosen to replace it was KetaSpire® KT-820 CF30 with Assigning a section width “b” = 25.4 mm (1.0 inch), the
a modulus of elasticity of 19.7 GPa (2.85 Mpsi), then moment of inertia of a ribbed construction that will satisfy
the moment of inertia must be increased to maintain that condition can be calculated.
equivalent deflection under load. The moment of inertia
The moment of inertia for the plate design is:
can be increased by increasing the thickness or adding
ribs. Substituting the E values in equation 1: bh3 25.4 × ( 34 )3
Iplate = = = 86.9 mm4
12 12
( 45 × 109 ) I metal = ( 19.7 × 109 ) IKetaSpire® PEEK By choosing the arbitrary rib design shown in Figure 8.5,
and working through the calculations, the moment of
For equivalent deflection, I must be increased by a factor inertia is found to be:
of 2.28.
Irib = 1379 mm-4
The moment of inertia for a rectangular section is:

bd3 Figure 8.5 Rib design example

I=
12
1.9 mm ( 0.075 in )
where b is the width and d is the thickness of the section,
substituting into our equation to determine the required
thickness yields:

2.28 d3 metal = d3 KetaSpire® PEEK 12.7 mm ( 0.50 in )

If d for the original metal part is 2.54 mm (0.10 inch) then

the required thickness in KetaSpire® PEEK is given by:

D= 3 2.28( 2.54 )3
25.4 mm ( 1.00 in )
d = 3.34 mm ( 0.131 inch )

To achieve equivalent deflection, the thickness has to be 2.54 mm ( 0.100 in )

increased by approximately 32 %. However, ribs can be
Placing one rib every 25.4 mm (1.0 inch) will increase the
used to effectively increase the moment of inertia more
moment of inertia approximately 17 times. The designer
efficiently as discussed in the next section.
can alter the rib height and rib width as well as the number
Adding Ribs to Maintain Stiffness and spacing of the ribs.

In the last section, it was determined that if a metal Designing for Sustained Load
part were to be replaced with a part molded of KetaSpire®
Up to this point, the stress-strain calculations and
KT-820 CF30, a part thickness of 3.34 mm (0.131 inch)
examples have dealt with immediate stress-strain
would be required to equal the stiffness of a 2.54 mm
response and therefore short-term properties. If the part in
(0.10 inch) thick magnesium part.
question must sustain loads for long periods of time or at
By incorporating ribs into the KetaSpire® PEEK design, elevated temperatures, apparent (creep) modulus values
the wall thickness and weight can be reduced very must be used to account for the additional strain and part
efficiently and yet be as stiff as the magnesium part. deflection that may occur.

To demonstrate this, the moment of inertia (I) of the new

rib design can be equated with that of the 3.34 mm (0.131
inch) thick plate design. Selecting the same material,
KetaSpire® KT-820 CF30, the modulus of elasticity
remains 19.7 GPa (2.85 Mpsi) in both cases; therefore,

70 \ KetaSpire® PEEK Design and Processing Guide

 Considering Stress Concentrations Using CAE to design a plastic part can provide more
Classical mechanical design may result in a component efficient use of the polymer, better moldability, and better
design which fails prematurely or at a much lower stress performance of the final product. Cost savings can be
than predicted due to stress concentration. Stress achieved by reduced development time. More efficient
concentrations may occur at sharp corners, around holes, design results in a lower cost part that costs less to
or other part features. Impact and fatigue situations are produce.
especially sensitive to stress concentrations. The basic CAE package for the polymer industry includes
Minimizing sharp corners reduces stress concentrations injection molding and mechanical simulation by finite
and results in parts with greater structural strength. element analysis (FEA). For more complex cases, software
exists that is capable of bridging the gap between the
To avoid stress concentration problems, inside corner injection molding and the mechanical simulation for
radii should be equal to at least half of the nominal wall filled polymers by using predicted fiber orientations in
thickness. A fillet radius of 0.4 mm (0.015 inch) should the FEA. This allows simulation of the effect of different
be considered minimum. molding options (gate position, injection time, etc.) on
the mechanical behavior of the final part, hence further
Figure 8.6 shows the effect of inside corner radius on the
improving the accuracy of the numerical models.
stress concentration factor. For example, if the nominal
wall thickness is 2 mm (0.080 inch) and an inside corner
Injection Molding Simulation
radius is 0.5 mm (0.020 inch), then the radius to thickness
The application of injection molding software for a plastic
ratio is 0.25 and the stress concentration factor will be
part has become a standard practice at a number of
over two. A stress of x will have the effect of over 2x on
manufacturers and suppliers. Such software provides an
the part.
understanding of mold filling, packing, and cooling before
Figure 8.6 Stress concentration at inside corners the tool is designed and built. Using the software, one can
evaluate gate positions, weld line locations, air traps, and
2.6 fiber orientation. It can also predict the injection pressure
required to fill and pack the part along with many other
Stress concentration factor

2.4
critical issues that impact part moldability and part quality.
2.2
2.0 Some of the most common results of a flow analysis are
the flow of the plastic vs. time (fill pattern), the pressure
1.8
required to fill the part (pressure profile), weld line
1.6 locations, and fiber orientation. Often, a warpage analysis
1.4 that incorporates the effects of crystallinity, cooling, and
1.2 fiber orientation is also performed. This type of analysis
takes into account many important aspects of the molding
1.0
process, such as temperatures, packing efficiency,
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Radius / thickness ratio shrinkage data, and fiber orientations, and ultimately
allows an estimation of the final dimensions and shape
Outside corners should have a radius equal to the sum of a molded part.
of the radius of the inside corner and the wall thickness
It is critical to recognize that the simulation of the injection
to maintain a uniform wall thickness.
molding process includes complex equations involving
thermodynamics, fluid dynamics, and heat transfer.
Computer Aided Engineering (CAE) The analysis is based on a number of assumptions
Designers can take advantage of the flexibility offered by and simplifications. Correlating simulation results with
injection molding to develop plastic parts with complex experimental evidence is recommended to gain a better
shapes and functions. Modern CAD tools allow designers insight of the accuracy that can be expected from such
to put flexibility into practice by building virtual models of simulation and achieve higher confidence.
the parts instead of building real prototypes. CAE tools
add the capability to perform various types of numerical
simulations on these virtual prototypes, such as injection
molding simulations or mechanical stress analysis.

KetaSpire® PEEK Design and Processing Guide / 71

Finite Element Analysis (FEA) elastic FEA is sufficient and allows quick evaluations
Mechanical simulations using FEA estimate the stresses of different design options. However, for unfilled grades
and strains of a part under a given set of service conditions. or materials in high temperature or moisture conditions,
There are many limitations regarding the use of textbook one can easily observe that the stress-strain curves
formulae to calculate the mechanical response of a become more and more nonlinear. This usually means
complex part. that the linear elastic assumption is still valid but only
within a reduced range of low strains and low stresses.
The FEA involves building a numerical model of a In that case, it is advisable to use nonlinear FEA packages
particular mechanical testing situation. This process that permit more accurate modeling of the material,
requires the knowledge and availability of different data: capturing the nonlinearity (for instance elasto-plasticity).
This covers the commonly used short-term, quasi-static
• CAD model of the studied component(s)
simulations.
• Material properties: The mechanical behavior model
of the chosen polymer material, taking into account More complex simulations can also involve creep (long-
the temperature, moisture content and time-scale of term deformation) or strain-rate dependent response to
loading (quasi-static, long term or dynamic); thermal rapid loading such as impact or drop testing. Nonlinear
expansion information is also often needed if elevated simulation packages generally offer material models that
service temperatures are to be considered. allow more accurate representations of polymer behavior
• Boundary conditions of the test case to be simulated, in a molded part.
which covers fixation points, contact area, etc.
Coupling the Injection Molding and Mechanical
• Loads, such as forces, moments or temperature fields Simulations
The material properties of the plastic part are extremely New sophisticated micromechanical simulation software
dependent on the molding process. For filled polymers, based on homogenization rules offers an interesting
fiber orientation during the flow leads to anisotropic alternative to the isotropic mechanical assumption. These
mechanical properties in the molded component. tools allow the fiber orientation data to be imported to the
However, the classical approach for the mechanical FEA package from the injection molding simulation. This
analysis of such parts is based on the assumption that software allows the definition of an anisotropic nonlinear
the material properties are isotropic in the part (identical behavior model using this fiber orientation data. To access
everywhere and in all directions). The tensile data this technology, one needs to perform appropriate and
typically available for this simulation is generated using specific material characterization based on real testing
standardized molded tensile bars. Indeed, these molded for different sets of orientation and reverse engineering
tensile bars exhibit very good fiber orientation, hence techniques. Please contact your Solvay representative
leading to favorable mechanical properties along the to access data and simulation assistance.
test direction.

The use of this tensile data as such to define isotropic

Designing for Assembly
behavior of a real molded component typically leads to Interference or Press Fits
an over-estimation of the mechanical response; therefore, One of the most economical methods that can be used to
it is common to weight the published tensile stress-strain assemble two parts is a press fit. The joint is achieved by
curves obtained from testing by a factor close to 0.7 for pressing or forcing the shaft into a hole whose diameter
isotropic numerical simulations. For filled polymers, the is smaller than the diameter of the shaft. The difference in
isotropic assumption and the choice of the weighting diameter between the hole and shaft is referred to as the
factor is arbitrary and very much dependent on the type diametrical interference. The force maintaining the joint is
of part, the injection process parameters and material primarily a compressive stress on the shaft resulting from
formulations. the hoop stress in the hub created by the insertion of the
Another important aspect linked to the material behavior shaft. Depending upon the relative moduli of the shaft and
is the choice of the material model: Is a linear elastic hub materials, the compressive stress in the shaft can also
assumption valid or is a nonlinear material model needed? contribute to maintaining the joint. The stress holding an
This is very much dependent on the polymer and the test interference fit will exhibit relaxation over time in a manner
conditions. For instance, a highly filled polymer used at that is analogous to creep because the apparent modulus
low temperatures and low strains usually exhibits a linear of the polymeric material decreases over time.
stress-strain response. In that case, performing linear

72 \ KetaSpire® PEEK Design and Processing Guide

Calculating the Allowable Interference Figure 8.7 Designing for mechanical fasteners
The allowable interference between a shaft and a hub Poor Better
can be determined by using the general equation:

SdDs F + υh 1 – υs
I= +
F Eh Es

And the geometry factor, F, is given by:

High bending stress Added bosses with small
2 as bolt is tightened gap, when bosses touch,
Ds
1+ stress becomes compressive
Dh
F=
2
Ds Flathead screw Truss or round head screw
1–
Dh

Where: High stress from wedging Recessed design avoids

I = Diametrical interference action of screw head wedging stresses
Sd = Working stress
Dh = Outside diameter of the hub
Ds = Diameter of the shaft
Eh = Modulus of the hub material
Es = Modulus of the shaft material
υh = Poisson’s ratio of the hub material Standard screw can allow Shoulder screw limits
υs = Poisson’s ratio of the shaft material high stress on tightening stress when tightened
F = Geometry factor

Mechanical Fasteners
If the shaft and hub are made from the same resin Mechanical fasteners provide an economical method of
grade, then: joining dissimilar materials together. Fasteners frequently
used with injection molded plastic parts include screws,
Eh = ES = E
bolts, nuts, lock washers and lock nuts. When using metal
and the interference is: mechanical fasteners, good design practice should be
used to prevent the plastic parts being assembled from
Sd F+1 becoming overstressed.
I= D
E s F
The most obvious procedure for preventing a highly
If the hub is made of KetaSpire PEEK and the shaft
® stressed assembly is to control the tightening of the
is made of metal, then the interference is: mechanical fasteners with torque limiting drivers. When
torque cannot be controlled, as might be the case with
Sd Ds F + υh field assembly, shoulder screws will limit compression
I=
F Eh on the plastic part. Other alternatives may be to use
flange-head screws, large washers or shoulder washers.
When a press fit is used with dissimilar materials, the Figure 8.7 presents some preferred designs when using
differences in thermal expansion can increase or decrease mechanical fasteners.
the interference between two mating parts. This could
increase or reduce the stress affecting joint strength.

A press fit can creep or stress relax over time. This could
cause a decrease in the retention force of the assembly.
Therefore, testing the assembly under its expected
operating conditions is highly recommended.

KetaSpire® PEEK Design and Processing Guide / 73

Figure 8.8 Boss design for self-tapping screws Figure 8.9 Torque developed during screw
installation
Boss o.d. < = 2 × Pitch diameter
Stripping torque
Pitch diameter

Radius 1 mm ( 0.039 in ) Clamp

Torque
load
Head makes contact
Thread with captured material
forming
torque
Driving torque

Penetration depth

Self-tapping screws Tightening torque

A common type of mechanical fastener used with plastics Figure 8.9 shows how torque changes as a function of
is a self-tapping screw. A self-tapping screw cuts or forms screw penetration. Tightening torque is the recommended
threads as it is inserted into the plastic and eliminates installation torque for a given application. It must be high
the need for molding internal threads or the secondary enough to fully engage the screw threads and develop
operation of tapping the thread form by machining. clamp load but lower than the torque that would cause
The major types are thread-forming and thread-cutting. failure of the threads, known as the stripping torque.

The modulus of elasticity of the plastic material plays an The optimum tightening torque value can be calculated
important role in deciding what type of self-tapping screw from the average driving torque and the average stripping
is most suitable for the application. For plastic materials torque using the following equation.
with a modulus less than 3.0 GPa (440 kpsi), such as
most unreinforced resins, thread-forming screws are 1 3 1
best since the plastic is ductile enough to be deformed TT = T + T
2 2 D 2 s
without cracking or shearing. For glass and mineral filled
grades, thread-cutting screws are preferred.
where
For optimum strip-out torque, the hole diameter of the TT = Tightening torque
boss should be equal to the pitch diameter of the screw. TD = Average driving torque
The outer diameter of the boss should be equal to two or TS = Average stripping torque
three times the hole diameter and the boss height should
be more than twice the thickness of the boss. Some self-tapping screws have been designed specifically
Figure 8.8 illustrates the basic boss design for use with for use with plastics and these have the advantage of
self-tapping screws. having a greater difference between driving and stripping
torque than the typical screws designed for metal. These
To avoid stripping or high stress assemblies, torque special fasteners can provide an additional safety factor
controlled drivers should be used on assembly lines. for automated assembly.
Improving torque retention
To minimize the loss of torque due to creep, reduce
the compressive stress under the screw head by:

• Increasing the screw head diameter

• Using a large-diameter flat washer
• Reducing the clamping torque
• Using a spring or spiral washer
• Using shoulder bolt to reduce stresses
on plastic part
• Using a metal bushing

74 \ KetaSpire® PEEK Design and Processing Guide

Figure 8.10 Boss design for ultrasonic inserts Figure 8.11 Cantilever type snap fit
Insert diameter

Boss diameter =
2 × insert diameter

b
d

t L
0.7 t Y

Pull out force calculation

Molded-in threads
The strength of a joint can be characterized
One of the benefits of using plastic materials instead of
by the amount of force required to pull-out a screw.
metals is the ability to mold thread forms directly into the
The pull out force can be estimated by using the
part. This eliminates the secondary machining operations
following equation:
needed with metal parts to form the threads. molded-in
threads can be either external or internal. In the case of
F = π SDL internal threads, some type of unscrewing or collapsible
where core is required. External threads can be formed more
F = Pull out force easily if the parting line of the mold is perpendicular to
S = Shear strength the thread.
D = Pitch diameter
L = Thread engagement length Designing with Snap Fits
The use of snap fits in plastics is very prevalent. All snap
When repeated assembly and disassembly are required fit designs require the plastic to flex like a cantilever
or expected, threaded metal inserts should be used spring as it moves past an interference that is designed
instead of self-tapping screws. on the mating part. Once the flexible arm moves past the
Threaded inserts interference, it returns to its normal unflexed, unstressed
position. Usually a step or protrusion has been designed
Threaded metal inserts can be used to provide permanent
on the cantilever that engages and locks into the mating
metal threads in a plastic part; a wide variety of sizes
part, creating a simple assembly method without
and types are available. Inserts are usually installed in
additional parts. This is shown in Figure 8.11.
molded bosses whose internal diameter is designed for
the insert. The most commonly used metal inserts are Each cantilever arm must deflect a distance “Y” in order
either molded-in or ultrasonically placed in the part as a to be inserted. The key to proper snap design is to not
secondary operation. In the case of the molded-in insert, exceed the strain/stress limits of the material being used.
the insert is placed in the mold and the plastic is injected A snap fit design that has been used for a ductile, low
around it. Stress will develop when the plastic cools modulus plastic will probably not be suitable for a highly
around the insert. To reduce this stress, heat the inserts reinforced, very rigid plastic.
to the temperature of the mold.
For rigid materials, the length of the cantilever may be
The ultrasonic insert is pressed into the plastic by melting increased or the interference deflection “Y” reduced.
the plastic with high frequency vibrations, generated by Adding a “stop” can prevent over deflection of the
an ultrasonic welding machine. The ultrasonic welding cantilever during assembly.
melts material around the metal insert as it is being
installed, forming a bond between the insert and the
plastic that is usually strong and relatively free of stress.

Figure 8.10 depicts the recommended insert and boss

designs for use with KetaSpire® PEEK resin.

KetaSpire® PEEK Design and Processing Guide / 75

Figure 8.12 Cantilever beam - concentrated load Solving the deflection equation for F, the force needed
at free end to deflect the beam can be calculated as follows:

F Equation 1

3YEI
F=
L3

L Y The modulus of elasticity, E, is defined as:

σ σ
E= and therefore ε =
Straight cantilever beam equation ε E

The relationship between maximum deflection and strain substituting that in the cantilever stress equation:
for a straight cantilevered beam was calculated as follows:
From Figure 8.2, the cantilever beam was chosen and the Equation 2
drawing is repeated as Figure 8.12. The maximum stress,
is given by the following: FLd
ε=
2EI
FL
σ= Using the relationship of equation 1 and substituting
Z
for F in equation 2 the relationship between strain and
Since this beam has a rectangular cross section, deflection is derived:

Equation 3
bd2
Z= and
6
3Yd
ε=
2L2
bd3
I=
12
This equation allows the designer to calculate the
Therefore: strain required from the maximum deflection of a design.
Table 8.4 summarizes maximum strain recommendations
FLd for several grades of KetaSpire® PEEK resins.
σ=
2I Once a suitable grade of KetaSpire® resin has been
selected, the basic equations can be used to find the load,
The deflection of the beam, Y, is given by:
F, needed to deflect the cantilever the required amount.
FL3 In a cantilever snap fit, the stress and strain are maximum
Y=
3EI at the base of the cantilever and become proportionately
lower toward the tip where the load is applied. In fact
Table 8.4 Strain recommendations for cantilever the stress and strain at any point can be calculated by
snap-fits substituting a different L, the distance from the load
toward the fixed end. Therefore, if the cantilever thickness
One-Time Multiple
Assembly Assembly was gradually reduced from fixed end to tip, the beam
KetaSpire® Max Design Strain Max Design Strain will be able to deflect more than in the fixed thickness
Grade (%) (%) cantilever without incurring higher maximum stresses. In
KT-820 NT 5.0 3.0 this way, the capability of the material can be maximized.
KT-880 NT 5.0 3.0
KT-820 GF30 2.0 1.5
KT-880 GF30 2.0 1.5
KT-820 CF30 1.5 1.0
KT-880 CF30 1.2 1.0

76 \ KetaSpire® PEEK Design and Processing Guide

 Figure 8.13 Snap fit design using tapered beam Designing for Injection Molding
Many of the applications for KetaSpire® PEEK resins will
L
be manufactured using the injection molding process.
An engineer who has designed a part to meet the
performance requirements of the application must also
take into account the fact that there are elements in the
part design that can influence moldability. These factors
include wall thickness and wall thickness transitions, draft,
ribs, bosses, and coring. The effect of these factors on
Y maximum hL moldability should be considered by the design engineer
h0 before a mold is built to make the part.

Wall Thickness
Figure 8.14 Proportionality constant (K) for In general, parts should be designed with the thinnest
tapered beam wall that will have sufficient structural strength to support
2.4 the expected loads, keep deflection within design criteria
limits, have adequate flow, and meet flammability and
Proportionality constant [ K ]

2.2 impact requirements. Parts designed in this manner will

2.0 have the lowest possible weight, and therefore the lowest
material cost, and the shortest molding cycle.
1.8

1.6 Wall Thickness Variation

1.4 Part designs that contain uniform wall thicknesses are
ideally suited for the injection molding process. They
1.2
minimize molded-in stress, reduce the potential for
1.0 sink marks on the surface of the part, and eliminate the
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 potential for voids in a molded part. However, structural,
Ratio of hL to hO appearance, and draft considerations may require varying
wall thicknesses. When changes in wall section thickness
Tapered Cantilever Beam Equation are necessary, the designer should consider a gradual
For the tapered design shown in Figure 8.13, hL is the transition, such as the tapered or gradual designs shown
thickness at the free end. The value of the proportionality in Figure 8.15.
constant, K, for a tapered beam design can be found in
Sharp transitions may create problems in appearance
Figure 8.14. The maximum strain can be calculated from:
and dimensional stability, because they may result in
Equation 4
differential cooling and turbulent flow. A sharp transition
may also result in a stress concentration, which may
adversely affect part performance under loading or
3Yh0
ε= impact.
2L2K

For example, if the beam thickness has been gradually Draft Angle
reduced to half its fixed end thickness; the ratio of hL To aid in the release of the part from the mold, parts
to h0 would be 0.5 and K (from Figure 8.14) would be 1.6. are usually designed with a taper. The taper creates a
Therefore, the maximum strain and the corresponding clearance as soon as the mold begins to move, allowing
stress would be multiplied the reciprocal of K, 0.625. the part to break free from its mold cavity. The taper is
The strain will be reduced by about 40 % of the strain commonly referred to as draft, and the amount of taper
of a constant thickness cantilever beam design with as draft angle, as shown in Figure 8.16.
equal deflection.

KetaSpire® PEEK Design and Processing Guide / 77

Figure 8.15 Wall thickness transition Ribs
The stiffness of a part design can be increased with
Poor Sharp properly designed and located ribs, without creating
thick walls. Proper rib design will allow for decreased wall
thickness. This will save material and weight, and shorten
molding cycles. It will also eliminate thick walls, which can
Tapered cause molding problems like sink marks on the surface
Good
of parts or voids on the inside of parts. Ribs that are
correctly positioned may also function as internal runners,
assisting plastic melt flow during molding.

Best Gradual In general, the following guidelines should be used when

designing with ribs. The thickness at the rib base should
be no greater than 60 % of the adjacent wall thickness.
Figure 8.16 Draft - designing for mold release When ribs are opposite appearance areas, the width
should be kept as thin as possible. If there are areas in
the molded part where structure is more important than
appearance, then ribs are often 75 %, or even 100 %, of
the outside wall thickness. Whenever possible, ribs should
be smoothly connected to other structural features such
as side walls, bosses, and mounting pads. If there are
several ribs in a part, they need not be constant in height
or width, and are often matched to the stress distribution
in the part. All ribs should have a minimum of 1/2 ° of draft
per side and should have a minimum radius of 0.8 mm
(0.03 in.) at the base to reduce stress concentrations
and sink marks.

Figure 8.17 shows the recommended rib size relationships.

Draft angle
Coring
Figure 8.17 Draft - recommended rib design Proper design practice should include uniform wall
thickness throughout a part. Heavy sections in a part
½° to 1½° draft can extend cycle time, cause sink marks on the part
surface, cause voids within the part and increase
molded-in stresses.
t
Heavy sections should be cored to provide uniform wall
thickness. For simplicity and economy in injection molds,
cores should be parallel to the line of draw of the mold.
t = 0.6T Cores placed in any other direction usually create the
need for some type of side action or manually loaded
R = > 0.8 mm ( 0.03" ) and removed loose cores.

Cores which extend into the cavity will be subject to high

T pressure. For blind cores (cores that are unsupported) that
have diameters greater than 1.5 mm (0.060 in.) the core
lengths should not exceed three times the diameter, while
Adequate draft angle should be provided to allow easy blind cores with diameters less than 1.5 mm (0.060 in.)
part removal from the mold. Generally, the designer should not exceed twice their diameter in length. These
should allow a draft angle of 0.5 ° to 1 ° per side for both recommendations may be doubled for through cores
inside and outside walls for KetaSpire® resins. However, (cores that telescope into or shut off with the opposite
in some special cases, smaller draft angles, as low as side of the mold). Draft should be added to all cores,
1 ∕8 ° to 1 ∕4 °, have been used with draw polish on the mold and all tooling draw polished for best ejection.
surface.

More draft should be used for deep draws or when cores

are used. Textured finishes increase draft requirements by
a minimum of 1 ° per side for each 0.025 mm (0.001 in.)
of texture depth.

78 \ KetaSpire® PEEK Design and Processing Guide

Bosses properly designed rib. Bosses should be designed in the
Bosses are projections from the nominal wall of a part same manner as ribs. Heavy sections should be avoided
that will eventually be used as mounting or fastening to prevent the occurrence of sink marks on the surface
points. The design of bosses is largely dependent upon and voids in the interior of the part.
their role in a given part. Cored bosses can be used for
press fits, self-tapping screws, and ultrasonic inserts. Figure 8.18 Boss design - general guidelines
Each of these will exert stress on the wall of the boss. O.D. = 2 × I.D.

As a general guideline, the outside diameter of a boss O.D.

should be twice the inside diameter of the hole, and the
wall thickness at the base of the boss should not exceed
60% of the part wall thickness unless structural concerns I.D.
override appearance requirements. Figure 8.18 illustrates
these guidelines.

Additional forces imposed on bosses may tend to be 0.3 T T

transmitted down the boss and into the nominal wall.
For this reason, a minimum radius of 25 % of the wall
thickness is required at the base of the boss to provide
strength and reduce stress concentration. A boss can
be further strengthened by using gusset-plate supports
around the boss, or attaching it to a nearby wall with a

KetaSpire® PEEK Design and Processing Guide / 79