08 Package Engineering Design Testing PDF
08 Package Engineering Design Testing PDF
PREPARED BY
PRESENTED BY
WESTPAK, INC.
83 Great Oaks Blvd., San Jose, CA 95119
(408) 224-1300 FAX (408) 224-5113
www.westpak.com
________________________________________________
I. INTRODUCTION
The purpose of this paper is to assist engineers and designers in the analysis of
requirements for protective package systems. The basics of product fragility
testing, package design, and package testing from a dynamics standpoint will be
covered.
There are numerous sources that one can turn to for precise technical data on
the behavior of products and materials in a dynamic environment. But how does
one use this information to help design a better package? How does one
determine what is important and what is not? How does one go about designing
an optimum package system or even recognizing when that system has been
designed? Finally, how does one know if the final numbers are believable or if
significant questions exist which would benefit from further analysis? These and
other areas will be investigated.
1
II. DEFINITION OF TERMS
Packaging Dynamics refers to active forces; that is, it implies motion rather than
static forces. Types of dynamic input include SHOCK or IMPACT which is
defined as a sudden severe non-periodic excitation of an object, and
VIBRATION which is defined as oscillation of an element or system about some
fixed reference point. While other inputs including temperature, humidity,
compression and static electric discharge may be important for a particular
design, they are not dealt with here. Refer to Appendix I for a complete definition
of terms list.
Note that metric units are cited as primary values with English units included in
brackets whenever feasible.
SHOCK
2
acceleration is also the peak or the high point of the acceleration vs. time pulse.
Note that DECELERATION is negative acceleration. The two terms are often
used interchangeably although acceleration properly refers to an increasing rate
of velocity change whereas deceleration describes a decreasing rate of velocity
change. (Acceleration is the differential of velocity with respect to time.)
VELOCITY CHANGE (∆V) is another unit often used in dynamic packaging work.
It refers to the difference between initial and final velocity and can be thought of
as a measure of energy dissipated at impact. It is equal to the area under the
acceleration vs. time pulse (the integral of the pulse).
∆V = Vi - (-Vr ) = Vi + Vr = (1+e) 2 gh
Velocity change is a crucial concept which can help determine the accuracy of
test results and help predict the required shock response characteristics for a
given package system.
VIBRATION
3
FREQUENCY is a measure of the number of cycles per time period, typically
cycles per second or hertz (Hz).
PERIOD refers to the time necessary to complete one cycle. This is the inverse
of frequency.
FIGURE 1
SINUSOIDAL AND RANDOM VIBRATION
Time Time
Acceleration
4
III. THE CONCEPT OF A PROTECTIVE PACKAGE
5
Figure 2
The relationship between these areas has been expressed by the equation
PDE = PR + P (Physical Distribution Environment = Product Ruggedness +
Package). Thus the job of package design for fragile products amounts to
defining and quantifying the variables in this simple equation.
Since the product and package must work together as a system to resist the
forces of the distribution environment, it is obvious that a tradeoff can be made
between the amount of ruggedness built into the product and the amount of
protection designed into the package. The exact tradeoffs between product
ruggedness and package protection should be a matter of economic analysis
between the product designers and the package engineer. In an ideal world, this
tradeoff would be based on economic considerations where the total delivered
cost of the product is a minimum. This concept is graphically demonstrated in
Figure 3.
6
Figure 3
O
TOTAL SYSTEM COST P OLD OPTIMUM NEW OPTIMUM
T
I
M
D U D
O M O
L L
L L
A A
R R
S S
E INCREMENTAL E INCREMENTAL
INCREMENTAL INCREMENTAL
X DAMAGE X DAMAGE
PACKAGING PACKAGING
P SAVING/LOSS P SAVING/LOSS
COST COST
E E
N N
D D
E E
D COST OF
D
IMPROVED
PRODUCT DESIGN
One of the most severe physical input that a protective package must mitigate is
the shock input associated with drops or other mishandling of a packaged
product. In this case, the job of the package system is to transform the relatively
high peak G short duration input typical of dropping a package onto a rigid
surface into a long duration low G shock pulse which is below the fragility level of
the product. See Figure 4.
7
Figure 4
SHOCK INPUT
G's
TIME
The package does this generally by means of a cushion system which deflects in
response to the mass of the product and the deceleration produced by the
impact. The cushion can deflect in compression, in shear, in torsion or any other
spring mode, although generally the compressive mode is used in packaging
design work. All cushion systems work in this way; namely, they deflect and in
doing so, trade peak acceleration for duration.
8
IV. A BRIEF HISTORY OF PROTECTIVE PACKAGE DEVELOPMENT
During the 1950's, considerable attention was focused on the general area of
shock response testing as well as the equipment and techniques useful to
describe the phenomenon of shock and shock response. The Firestone
Aerospace Division was active in designing and testing cushion systems
(primarily rubber airbags) for military applications. One of the big drawbacks was
the lack of reliable fragility information on various military hardware. Another was
the inadequate sophistication of equipment used to determine shock fragility.
To determine its feasibility and to simplify the procedure, Dr. Robert Newton at
the Naval Postgraduate School in Monterey, California, was asked to formulate a
test procedure which would utilize shock response spectrum analysis for
commercial products with an eye towards improving the packaging procedure for
these products. The result of his effort is the now famous Damage Boundary
Theory for product fragility testing.
Michigan State University then ran a lengthy series of tests on a wide variety of
consumer products during the late 1960's. Equipment to run this testing was
9
leased from Monterey Research Laboratories and the results were published in
Technical Report Number 17 from the multi sponsored research group at
Michigan State University. The results showed that the theory was indeed
workable and did provide an accurate means of assessing product fragility.
The Damage Boundary Theory was simplified and put in an easy to follow five
step procedure by the MTS Corporation, which had acquired the Monterey
Research Laboratory facilities in the late 60's. The "five step" also incorporated
dynamic cushion testing, which had been developed through the efforts of the
ASTM D-10 Committee on Packaging.
It was therefore significant that in the early 1970's, for the first time, package
development could seriously be considered an engineering discipline. The tools
and procedures were now in place to effectively and efficiently design protective
packages. Refinements have occurred since that time but nothing rivals the
significance of the Damage Boundary Theory. Vibration testing for both products
and cushion systems has also been added to the package design and test
procedures.
10
V. DEFINE THE ENVIRONMENT
Note that the environment must be quantified in terms of all potentially harmful
inputs. These inputs may include temperature and humidity extremes,
atmospheric pressure changes, compression, shock, vibration, and electrostatic
discharge among others. Only the effects of shock and vibration are covered
here. However, the designer must be aware of all likely hazards in the
environment and quantify them in terms of their ability to cause damage to the
product.
SHOCK ENVIRONMENT
Most shock inputs occur during physical handling, especially the loading and
unloading of transport vehicles. Defining this environment amounts to
quantifying the drop height experienced by packages. Studies have attempted to
define drop height as a function of the package size and weight. (Figure 5)
Much literature is available with similar information.
Other methods of obtaining this data are direct observation and measurement of
the environment with appropriate devices. Whether taken from published
literature or direct observation and measurement, this data has a certain
probability associated with it and must be properly interpreted. Arbitrarily
increasing the severity of the "assumed environment" to achieve a hypothetical
increase in the confidence level leads to costly overpackaging.
11
FIGURE 5
TYPICAL DESIGN DROP HEIGHT CHART
(CAUTION: This chart shows trends only, not actual values)
1. Severe drops are rare. Most packages are subjected to many low
drops while relatively few packages receive more than one drop
from greater heights (See Figure 6).
2. Unitized loads are subjected to fewer and lower drop heights than
individual packages.
3. Most packages are dropped on their base; over 50% of the total
number recorded.
4. The heavier the package, the lower the drop height.
5. The larger the package, the lower the drop height.
6. Warning labels such as "Fragile" and "Handle with Care" have little
effect on package handling.
7. Hand holds on the sides of packages appear to reduce the
incidence of higher drop heights.
12
FIGURE 6
DROP HEIGHT vs. PROBABILITY
(Source: FPL22)
VIBRATION ENVIRONMENT
The vibration environment is complex and random. The primary source of this
vibration input is the various transport vehicles in which products travel from the
time they are produced until they reach the final consumer. To quantify it, the
acceleration vs. frequency profiles (spectra) of transport vehicles must be
determined.
Once the likely transportation mode has been defined, envelope or composite
spectra such as that shown in Figures 7 and 8 can be assigned to the
environment. This information can be used to program a random vibration
spectrum for package testing. It is also useful to note at which frequencies the
highest dynamic inputs occur. A well designed package system will attenuate
vibration input at these frequencies.
13
In summary, defining the distribution environment for shock and vibration
amounts to defining the design drop height and the vibration profile likely to be
encountered. Information presented here is not intended to define the total
transportation shock and vibration environment. Rather it is meant only to give a
brief overview of the formats used to present such data.
FIGURE 7 FIGURE 8
TRUCK SPECTRA ENVELOPE SPECTRUM
A A
C Peak C
C 1.0 C 1.0
E E
L L
E E
R R
A 0.5 A 0.5
99% T
T
I I
O 0.4 O 0.4
N N
0.3 0.3
G’s G’s
0 0
to to
P 0.2 P 0.2
E E
A A
K K
90%
0.1 0.1
(Source: FPL-22)
14
VI. DETERMINE PRODUCT FRAGILITY
SHOCK FRAGILITY
The Damage Boundary is the principal tool used for product shock fragility
assessment. The Damage Boundary Plot takes the general shape of that shown
in Figure 9. It defines an area on a graph bounded by Peak Acceleration on the
vertical axis and Velocity Change on the horizontal axis. Any shock pulse
experienced by the product which can be plotted inside this boundary will cause
damage regardless of whether the product is packaged or not. (Remember this
is a product test.)
15
FIGURE 9
DAMAGE BOUNDARY
To run a Damage Boundary test, mount the product on the table of a shock test
machine (Figure 10). Support the product by a fixture similar in configuration to a
package. The fixture should be as rigid as possible so that it does not distort the
shock pulse transmitted to the product.
16
FIGURE 10
SHOCK TEST MACHINE
Set the shock machine to produce a low velocity change pulse with a duration of
approximately 2 msec (a half sine waveform is generally used for this test). After
the test, examine the product to determine if it is damaged. If not, set the shock
machine to produce a slightly higher velocity change and repeat the test.
Continue this process with small increases in velocity change until damage
occurs. The last non-failure shock input defines the critical velocity change
(∆Vc ) for the product in that orientation. (Refer to Figure 9)
Fixture a new test specimen to the shock machine and set the machine to
produce a trapezoidal pulse with low acceleration and a velocity change of
(2)∆Vc. After the shock pulse, examine the product to determine if damage has
occurred. If not, set the shock machine to produce a higher acceleration level at
constant velocity change. Repeat this process with small increments in
acceleration until the failure level is reached. The last non-failure shock input
defines the critical acceleration (Ac) for the product in that orientation. (Refer to
Figure 9.)
17
The Damage Boundary may now be plotted by drawing a vertical line through the
critical velocity change point and a horizontal line at the critical acceleration point.
The intersection of these two lines (the knee) is a smooth curve as Figure 9
shows. A rectangular corner approximates the damage region.
The use of the trapezoidal wave results in a (near) linear abscissa on the
Damage Boundary. This means that it is necessary to determine only one point
to define the critical acceleration for the product in that orientation. Other
waveforms result in critical accelerations which are a complex function of the
natural frequency of components within the product. Figure 11 shows Damage
Boundaries for various waveforms.
FIGURE 11
DAMAGE BOUNDARIES FOR VARIOUS WAVE SHAPES
TRAPEZOIDAL PULSE
RECTANGULAR PULSE
VELOCITY CHANGE
18
The Damage Boundary is a valuable and powerful tool. Critical velocity change
is related to freefall drop height from the formula:
∆V = (1 + e) 2 gh
Critical velocity change tells the designer how high the unpackaged product can
fall onto a surface before damage occurs. If this drop height is likely to be
exceeded in the distribution environment, then the product must be cushioned.
The performance requirements of the cushion are that no more than the critical
acceleration be transmitted to the product.
FIGURE 12
DROP HEIGHT vs. PRODUCT VELOCITY CHANGE
D e = 0 ( No rebound) e = 0.5 (50% rebound)
E in cm
S 100 254
I
e = 1.0
G
(100% rebound)
N 80 203
D
R 60 152
O
P
40 102
H
E 20 51
I
G
H 0 0
T
0 100 200 300 400 500 600 in/sec
0 254 508 762 1016 1270 1524 cm/sec
PRODUCT VELOCITY CHANGE
(Source: ASTM D3332)
19
The Damage Boundary also tells the engineer that at low velocity changes,
infinite accelerations are possible without damage. Conversely, at low
acceleration levels infinite velocity change is allowable without product damage.
This means BOTH critical acceleration and critical velocity change are necessary
to properly characterize product fragility.
The shock pulse used to determine critical velocity change (∆Vc) may look like
that in Figure 13A. While it is often called a half sine pulse, its shape is more
characteristic of a versed sine rather than a true half sine. Figure 13B shows the
trapezoidal pulse used for the Critical Acceleration test. Note that this pulse is
often called a square wave or rectangular wave. In reality, it is a trapezoid
because the rise and fall times are not infinitely short. Shown together on the
same scale these two pulses would look like those shown in Figure 13C, with the
velocity change pulse on top and the acceleration pulse below.
FIGURE 13
DAMAGE BOUNDARY WAVEFORMS
A B
20
Before running the Damage Boundary test, the engineer must define what
constitutes damage to the product. Damage may be catastrophic failure or less
severe damage modes which make a product unacceptable to the customer. In
some cases damage can be determined by looking at the product or it may
involve running sophisticated functional checks. Once the determination of
damage is made, the definition must remain constant throughout the test and
must be consistent with what is unacceptable to the customer.
In general, Damage Boundary tests must be run for each axis in each orientation
of the product. In the case of a rectangular product such as a television set, this
means a total of 12 specimens for a rigorous test (6 for critical velocity change
and 6 for critical acceleration). However, since the testing is normally done in the
prototype stage, rarely is this number of product available for a potentially
destructive test. As a practical matter, much information can be gained from a
limited number of units.
It is imperative that the fragility level of the product be clearly established prior to
designing a package system. In general, the amount (thickness) of cushion
increases exponentially as the fragility of the product decreases linearly. The
result may be a tremendous waste of material if an engineer decides to use
21
20 G's as the "assumed fragility" of the product just to be conservative when the
actual fragility of the product is 30 G's or more.
VIBRATION
At (and near) the product resonant frequency, the response acceleration can be
very much greater than the input, causing the product fatigue and failure in a
short time. Amplification occurs in this frequency band. The purpose of
vibration sensitivity assessment is to identify those critical frequencies likely to
cause damage to the product.
22
FIGURE 14
PRODUCT RESONANT FREQUENCY PLOT
The Resonant Frequency Search Test is run by fixturing a product to the table of
a vibration test machine and subjecting it to a sinusoidal low level constant
acceleration input (typically .25 to .5 G's) over the frequency range of the
distribution environment, typically 3 to 500 Hz (cycles per second). Random
vibration can also be used for this purpose. The response/input ratio
(transmissibility) is plotted as a function of frequency. This ratio reaches a
maximum at the component resonant frequency. The test usually involves
monitoring many components in each axis of the product in order to characterize
its overall vibration sensitivities.
23
Not only is vibration input a certainty, but its damaging effects can be severe.
This is particularly true if a package cushion amplifies vibration input at the
product natural frequency. This can result in a rapid build up of acceleration
leading to component failure in a very short period of time. Thus, it is possible for
an improperly designed package to actually destroy the product it is
intended to protect. Without adequate vibration data on the product and the
package, it is impossible to know if this situation exists prior to shipment.
Real products behave more like complex spring/mass systems rather than the
single-degree-of-freedom (See Appendix I) model implied by the simplex
transmissibility plot of Figure 14. Actual data shown in Figure 15 reveals the
more normal interactive nature and the constructive and desctuctive
interferences that are typical for electronic products. Sub-harmonic peaks are
normally benign and should not be mistaken as a true product resonance.
Likewise, harmonics will often occur at integer multiples of the primary resonant
frequency.
Recent studies have found that random vibration excitation of products results in
better and more predictive data from a resonance search test as compared to
sinusoidal vibration. The reasons appear to be associated with the fact that
random inputs excite all resonances simultaneously at approximately the same
levels as will occur on transit vehicles. Normally, this random vibration
resonance search results in lower amplification levels and slightly lower resonant
frequencies as compared to sine vibration. This is probably due to normal
destructive interferences between spring/mass systems within the product.
24
Care must be exercised when attaching transducers (accelerometers) to a
product under test to avoid loading the monitored components and thereby
altering the true resonant responses. Use the lightest accelerometers
appropriate for the measurement to reduce the effect of the instrumentation on
the test results.
25
VII. CUSHION MATERIAL PERFORMANCE
SHOCK PERFORMANCE
The resulting cushion curve shows peak acceleration on the vertical axis and
static stress on the horizontal axis (static stress = weight/bearing area). Each
curve is drawn from a minimum of 5 test points (static stress levels) and each
test point is the average of the last 4 of 5 acceleration readings (impacts) of the
cushion material.
26
FIGURE 15
80
2 cm thick
60
40 3 cm thick
20
0 2
0 25 50 75 100 125 150 g/cm
.35 .71 1.1 1.4 1.8 2.1 lb/in 2
Most cushion curves have the general shape of those in Figure 15. The left-hand
portion shows a relatively high deceleration transmitted through the cushion. In
this area the static stress is low because of the light weight on the cushion; the
object (platen) does not have sufficient force to deflect the cushion and therefore
the effect resembles dropping a product onto a rigid surface.
In the center portion of the curve (where the cushion is being used effectively),
the object has sufficient force to deflect the cushion and cause the deceleration
to be spread over a longer period of time. The result is a lower deceleration
level.
On the right-hand portion of the curve, the cushion material is overloaded and the
object continues right through the cushion (it bottoms out) and impacts with the
surface on which the cushion is resting. Thus, it approaches using no cushion at
all resulting in, once again, high deceleration levels.
27
It is desirable to use cushions in the lower portion ("belly") of the curve where
performance is optimum. When the product critical acceleration, weight and
design drop height are known, the usable static stress range of cushion area can
be determined for a given material and thickness.
VIBRATION PERFORMANCE
Figure 16 shows the possible test setups and Figure 17 the transmissibility plot
for a typical cushion. The mass of the test block is changed in order to vary the
loading on the cushion material and the test is repeated. Different plots are
obtained in this fashion (see Figure 18). A series of 5 vibrational sweeps at
different loadings are recommended to construct the Amplification/Attenuation
plot shown in Figure 19.
FIGURE 16
TYPICAL CUSHION VIBRATION TEST SETUPS
METHOD B
METHOD A
CUSHION
RESPONSE
ACCELEROMETER MASS
MASS INPUT
CUSHION ACCELEROMETER CUSHION
VIBRATION TABLE VIBRATION TABLE
28
FIGURE 17
CUSHION RESONANT FREQUENCY PLOT
(TRANSMISSIBILITY)
FIGURE 18
MULTIPLE RESONANT FREQUENCY PLOTS
(different cushion loadings)
0.1
B A C
0.01
1 10 100 1000
FREQUENCY - Hz
(Source: Westpak, Inc.)
29
FIGURE 18
AMPLIFICATION/ATTENUATION PLOT
FREQUENCY - Hz
100
90
C ATTENUATION ZONE
80
C
70
A C
60
C
50 AMPLIFICATION
A ZONE
40 B
A C
B
30 A
B A
X
20 B
UNITY ZONE
B
10
0
0 10 20 30 40 50 60 70 80 90 100 110 120
The plot in Figure 19 may be interpreted as follows: for a given frequency, low
static stress levels result in the same acceleration transmitted to the product as
the input. In other words, the response/input ratio is approximately 1. As the
loading increases, there is a range over which the cushion material amplifies the
vibrational input. In this region the response/input ratio is greater than 1. At
higher static stress levels, the cushion material attenuates (reduces) the
vibrational input and the response/input ratio is less than 1.
30
The producers and users of cushion materials should be familiar with cushion
curves (impact) and Amplification/Attenuation plots (vibration) both in terms of
the data and how it is obtained. This is important. For example, information
obtained from procedures like ASTM D1596 will likely be different from the
Enclosed Test Block method described in ASTM D4168 for testing of foam-in-
place materials.
Users of cushion materials should insist on shock and vibration data when
designing with a given material. Without adequate performance data, the
package designer cannot optimize performance. It is possible to design a
package system that will destroy a product rather than protect it,
particularly if the cushion amplifies vibrational input at product critical
frequencies.
31
VIII. PACKAGE TEST PROCEDURES
The reason for establishing the test procedure before designing the package
comes from the characteristic of many cushion materials where they transmit
higher levels of deceleration with increasing drops. For example, a close look at
Figure 20 shows that the first drop will often result in lower transmitted
deceleration than succeeding drops. This is especially true for "semi-resilient"
cushions such as expanded polystyrene (EPS). Heavier loadings necessary to
achieve lower deceleration levels will also have a negative effect on the cushion's
ability to withstand repeated impacts. On the other hand, more resilient materials
such as polyethylene foam generally show very little change with repeated
impacts.
The result is that the design process is dependent on the material used and the
test procedure. An EPS pack for a single impact verification test procedure
would be different than if the test procedure required multiple impacts on the
same face. However, if one were using polyethylene foam cushions, it would
probably make little difference if the test procedure called for single or multiple
impacts on each face. The end result of this step should be a clearly established
test procedure.
32
FIGURE 20
CUSHION CURVE SHOWING THE EFFECT OF MULTIPLE IMPACTS
DECELERATION G
100
5 IMPACT NUMBER
90 4
3
80 2
70
60
1
50
40
30
20
10
0
0 25 50 75 100 125 150
There is subtle, though distinct differences between the performance and the
engineering design characteristics of a package system, both in terms of design
criteria and the testing to verify compliance. Package engineering design refers
to the ability to mitigate shock and vibration to levels below product fragility.
Package performance refers to the ability of the package system itself to
withstand the normal forces involved in the distribution process. It is very
possible to design a package which has the proper engineering design
characteristics but will not withstand the forces typical of the shipping
environment.
33
Vibration design verification refers to the ability of a package system to attenuate
vibration input at and near product natural frequencies. It is tested by subjecting
an instrumented package to vibration input in a sinusoidal sweep test (ASTM
D999) or similar procedure. The normal vibration performance test involves one
or more resonance dwells at package resonant frequencies. Random vibration
testing may be used in place of or in addition to sinusoidal resonance search and
dwell tests.
34
IX. PACKAGE SYSTEM DESIGN
∆x = 2h / (A - 2)
where:
∆x = cushion deflection in cm (or inches)
h = drop height in cm (or inches)
A = the required deceleration level (G's)
This gives the theoretical deflection necessary, not the overall cushion
thickness. In general, materials such as expanded polyethylene foam will
compress approximately 40 to 60% of total thickness before "bottoming out"
starts to occur. More flexible materials such as polyurethane foam will compress
up to 80% before it bottoms out.
DESIGN EXAMPLE:
35
MATERIAL OPTIMUM TOTAL THICKNESS
STRAIN %
cm inches
The optimum static stress loading (weight/bearing area) for a given material,
thickness and drop height combination is determined from a cushion curve.
(Figure 21) Theoretically, any portion of the cushion curve that lies below the
product fragility level will define a static stress loading capable of transmitting
less than the critical acceleration to the product. For optimum material usage, it
is normally desirable to load the cushion to the highest static stress allowed by
the curves. However, many designers find it desirable to load the material at the
low point (belly) of the cushion curve where transmitted deceleration is a
minimum.
Note that the procedure used for running cushion curves may have a significant
effect on the usefulness of the information.
The end result of this step should be the optimum static stress loading for the
material thickness determined earlier.
36
If vibration data will be used in the design (and it certainly should), the next step
is to draw a horizontal line across the Amplification/Attenuation plot for this
cushion material, tangent with the product natural frequencies. For most
designs, the lowest product resonance in each axis is the most important. This
plot must describe the same material and thickness as that described in the
(shock) cushion curve.
FIGURE 21
DESIGN FOR SHOCK PROTECTION
75 cm (30 in) DROP HEIGHT ("DESIGN" DROP HEIGHT)
DECELERATION G
100
90
2 cm THICK
80
70
60
50 3 cm THICK
40
30 PRODUCT FRAGILITY = 28 G
20
FIGURE 22
CUSHION VIBRATION DESIGN
AMPLIFICATION/ATTENUATION PLOT
FREQUENCY - Hz
100
3 cm THICK CUSHION
90
ATTENUATION ZONE
80
70
60
AMPLIFICATIOPN ZONE
50
40 LOWEST PRODUCT
NATURAL FREQUENCY = 37 Hz
30
20 UNITY ZONE
10 MINIMUM LOADING FOR
VIBRATION PROTECTION = 68 g/cm^2
0
0 10 20 30 40 50 60 70 80 90 100 110 120
37
The minimum static stress loading is determined from the intersection of the
attenuation boundary and the lowest product critical frequency (See Figure 22).
Higher static stress loading will result in greater attenuation (which is desirable)
while lower static stress loading may amplify vibrational input.
The end result should be a static stress loading which will give good results for
both shock and vibration requirements. Cushion thickness may have to be
adjusted in order to achieve this goal.
38
C. CONSIDER THE USE OF RIBS
It is instructive to investigate why ribs have been used in package cushions for
many years. In general, the use of ribs will result in less material in the design
and therefore, higher loading on the material which remains. Ribs can also result
in greater deflection from a given cushion thickness.
39
FIGURE 23
VARIOUS RIB CONFIGURATIONS
RIB COMPRESSION
0%
25%
T .6T
50%
It is interesting to note that most rib designs are trapezoidal in cross section and
most literature treats this as a "standard" shape for ribs. From a theoretical
standpoint the best rib design is a pyramidal cross section. A rib with a
hemispherical cross section also is a good theoretical design. The reason is that
at zero deflection, the static stress loading is (theoretically) infinite and therefore
deflection occurs very rapidly at the onset of a dynamic input. As deflection of
the cushion material continues in response to the input, the static stress
decreases as the area of the cushion increases. Ideally this deflection and
change in loading will occur at a rate which is optimum for the shock
performance of the cushion.
40
D. PACKAGE DESIGN SUMMARY
Once the total thickness, static loading and rib configurations are determined, the
package must be designed using these numbers. This is the point where both
the performance and the integrity requirements of the package system must be
addressed. Certainly numerous other factors enter into the process of
determining the best package design. These include fabrication requirements,
end user constraints, ecological considerations, flammability and a host of others.
The important requirement for dynamics is a static loading in each product axis
which satisfies the product shock sensitivities and does not result in vibration
amplification at product critical (resonant) frequencies.
41
X. PACKAGE PROTOTYPE TESTING
For design verification testing of package systems, flat impacts (as opposed to
corner or edge impacts) are generally used with the deceleration transmitted
through the cushion measured by accelerometers mounted on the product. The
test procedure should be that previously agreed to, but in most cases will follow
ASTM D5276 or similar procedures. Take care to ensure flat impacts. This is
important! The difference between a flat drop and an "almost flat" drop can be
very drastic in terms of response deceleration.
It is also important that the monitored location (where the accelerometers are
mounted) be as rigid as possible and ideally as close to the product/cushion
interface as possible. The reason is to determine the package input, not the
product response characteristics. In may cases these are difficult to separate.
If the product were a solid uniform mass, it probably wouldn't make any
difference where the accelerometers were located; the input from the cushion
would be identical to the response of the mass. However, most products have
suspended masses and other flexible components which will be excited (put into
motion) by a shock input. The response of these various suspended
components can cause such things as "chattering" or high frequency noise on
the response waveform (see Figure 24).
42
Often the response peak deceleration is well above the input of the cushion. For
example, a primary cushion response waveform may have a peak of 40 G's with
superimposed high frequency on top of it which may double that number. It is
sometimes important to identify the difference between package input and
product response.
FIGURE 24
UNFILTERED AND FILTERED RESPONSE WAVEFORMS
UNFILTERED (LEFT), FILTERED (RIGHT)
This is one of the most common problems in package response testing. Several
methods of reducing this problem include:
2. Understand the use of electronic filters and how they can reduce
the apparent affect of high frequency ringing superimposed on
the primary response waveform. Exercise care to avoid
overfiltering and distorting the response data. (See Figure 24)
Remember that the best filter is no filter at all.
43
3. If possible, restrict flexible elements within the product in order
to make it as homogeneous and rigid as possible. (It is
sometimes instructive to perform two drop tests; one with the
flexible elements unrestrained showing the high frequency
response and the second with flexible elements restrained
showing the difference this has on the product response
characteristics.)
Package shock performance tests typically involve a series of corner and edge
impacts such as those called out in ASTM D4169. This procedure is perhaps the
most up-to-date method incorporating much of the environmental input studies to
date. This standard is highly recommended for package integrity testing.
44
mechanical shaker) amounts to a series of repeated impacts with very short
intervals between events. It may be referred to as a repeated impact test, a
bounce test, a fatigue test or something else.....but it should not be mistaken for
a vibration test.
If the package system meets all its requirements, then the job is finished. If not,
further package system refinements are necessary. The following sections
should help with those refinements.
45
XI. INTERPRETING PACKAGE RESPONSE DATA
A. WAVEFORM ANALYSIS
The response waveform generated during a package drop test contains a wealth
of information useful to the package designer. A good designer should definitely
learn to interpret this information.
The first piece of information taken from a response waveform is the total velocity
change (or what can be thought of as the energy released during the impact).
This can be determined by integrating the waveform. The integral, as described
earlier, is the area under the deceleration vs. time pulse. This area can be
estimated by multiplying the peak by the duration. Use the following formula:
∆V = Ap x g x D x .6
The ".6" is a factor to account for the shape of the waveform which is generally
someplace between a halfsine and a haversine. The resulting estimate of
velocity change should fall somewhere between the minimum and maximum
lines on the drop height vs. velocity change chart shown in Figure 12. If it
doesn't, there is something wrong with the test and it should be investigated.
In general, the rise time of a shock response pulse (the time from onset of the
pulse until peak acceleration) should be 1/3 to 1/2 total pulse duration. If the rise
time is shorter than this, it generally indicates that the cushion is too stiff or the
loading too light. If the rise time is greater than 1/2 pulse duration, this generally
indicates that the material is too flexible or is overloaded. In a similar
46
way, if a sharp spike is seen at the very beginning of the waveform, it generally
indicates that the cushion material is too stiff or too lightly loaded. Conversely, if
a sharp spike is seen near the end of the waveform this indicates that the
material is too flexible or too heavily loaded. Refer to Figure 25.
FIGURE 25
B. INCONSISTENT DATA
In some cases, repeated drops will produce different results with the same drop
height and accelerometer location. This generally indicates that either the drop is
not flat or that the product is rotating within the cushion on impact. One way to
resolve this is to use two accelerometers at different locations within the product.
Product rotation upon impact is normally an indication that the cushion material is
not properly distributed in relation to the weight of the product. Another method
of determining this is to use a triaxial accelerometer (3 accelerometers in one) to
measure the cross axis deceleration during a package drop test.
Instrumentation problems can also cause inconsistent data during a drop test.
Be sure that the acceleration monitoring system is connected properly and that
there are no shorts or intermittence in the accelerometer cable(s). Of course,
only recently calibrated instruments should be used for any serious testing.
47
XII. DESIGNING WITH NEW OR COMBINATION MATERIALS
As most people are aware, moldable polyethylene and polypropylene have been
introduced worldwide. These two materials promise to revolutionize at least part
of the foam cushion industry in that they are capable of being molded in
configurations similar to expanded polystyrene and are easily recyclable.
Although they have excellent shock and vibration characteristics, their high costs
necessitate optimum material usage in order to be economically feasible. This
often requires the use of flexure in a design situation rather than compression as
in traditional cushion designs.
This means that standard cushion curves may no longer serve as the only
guideline for determining proper static loading. Rather, the designer will first use
a formula to determine the required deflection (and therefore overall cushion
thickness) and following this, will use waveform analysis of a drop test in order to
further optimize the cushion system. This will likely result in a substantial amount
of trial and error until newer tools like Finite Element Analysis (FEA) and
Statistical Energy Analysis (SEA) become available and widely used.
Another interesting feature of cushion systems of the future will likely be the
increased use of combination materials; for example, polyethylene and
polyurethane foam used together, or polystyrene and polyethylene foam used in
the same package system. Currently it is rare to find a design which uses
combination materials, either in series or in parallel. On the few occasions that
one finds it, the results are normally not encouraging.
The reason is that there are few guidelines to help the designer produce an
optimum package system. Rather, it is generally a series of trial and error efforts
that may or may not produce fruitful results. However, all this is changing.
48
Through the use of deflection equations, waveform analysis and microcomputers,
the designer can more quickly determine if combination materials make sense in
a given situation over conventional techniques.
Of the designs done to date, it appears that combination materials used in series
make more sense than those used in parallel. It has also been noted that the
more successful designs use materials that have approximately the same spring
rates. On the other hand, those that use materials with vastly different spring
rates such as polyurethane and polystyrene generally produce unsatisfactory
results.
49
XIII. CONCLUSION
All cushion systems work in the same way, namely they trade peak deceleration
for duration; that is, they trade a high peak short duration shock pulse for a
longer duration lower peak shock pulse (See Figure 26). The longer duration is
in response to the deflection of the cushion. This deflection can be the result of
compression, shear, flexure, or other motion of the material. In any case, the
results are the same, namely, the material must "give" in order to change the
shape of the deceleration vs. time pulse delivered to the product. The nature of
this deflection is controlled by a series of simple physical formulas.
FIGURE 26
INPUT AND CUSHIONED RESPONSE SHOCK PULSES
SHOCK INPUT
G's
TIME
It is likely that package design and testing will become more technical in the
future. However, the increased sophistication will simply involve adaptations of a
few basic techniques explored herein. The designer is encouraged to learn why
and how cushion material do their job and to use this information to design better
package cushion systems.
50
APPENDIX I
DEFINITIONS OF TERMS
ACCELERATION A vector quantity describing the time rate of positive
change of velocity of a body in relation to a fixed reference
point. It is usually expressed in G's which are multiples of
the gravitational constant. Deceleration is the time rate of
negative change of velocity.
EQUIVALENT DROP HEIGHT The height of a free fall required by a body in a vacuum to
attain a particular instantaneous velocity at impact.
i
FREQUENCY, RESONANT The frequency at which a spring-mass system displays its
maximum response.
POWER SPECTRAL DENSITY A term used to describe the intensity of random vibration
in terms of mean squared acceleration per unit frequency.
The units are G2/Hz.
ii
SHOCK MACHINE A device for subjecting a system to a controlled and
reproducible mechanical shock.
iii
VIBRATION, RANDOM An oscillation having an instantaneous frequency and
amplitude that can be specified only on a probability basis.
iv