|
STEP 2 - PRODUCT FRAGILITY ANALYSIS
Just as the weight of the product can be measured using a scale, the product
ruggedness can be measured with dynamic inputs. A shock machine is used to
generate a damage boundary curve, and a vibration system is used to map out the
natural frequencies of a product. SHOCK: DAMAGE
BOUNDARY
The damage boundary theory is a testing protocol which determines, in an
engineering sense, which shock inputs will cause damage to a product and which
will not. There are two parts of a shock which can cause damage, the
acceleration level and the velocity change. The velocity change, or the area
under the acceleration time history of the shock, can be thought of as the
energy contained in a shock. The higher the velocity change the higher the
energy content. There is a minimum velocity change which must be achieved before
damage to the product can occur. This level is called the critical velocity
change. Below the critical velocity change, no damage occurs regardless of the
input acceleration level. In essence, there is not enough energy in this region
of the damage boundary to cause harm to the product. Exceeding the critical
velocity change, however, does not necessarily imply that damage results. If the
change in velocity occurs in a manner which administers acceptable doses of
acceleration to the product, the velocity change can be very large with out
causing damage. However, if the critical velocity and the critical acceleration
are both exceeded, damage occurs.
 A typical
damage boundary curve can be found in Figure 6. It is a plot of the damage
causing parameters of a shock pulse and defines the region where certain
combinations of acceleration and velocity change will cause damage. If the
combination of acceleration and velocity change fall in the clear band outside
the damage region, no damage occurs. For example, if the velocity change of the
input is below that of the product's critical velocity change, then the
acceleration level of the input can be in the 100s G's, 1000s G's, 10,000s G's,
or even infinite without causing damage. In practical terms, to achieve these
very high accelerations means that the duration of the shock must be very short.
If we think of it in a graphical sense, we have a certain amount of area or
velocity change which we can re-arrange to produce a shock pulse. If we decide
to make the pulse very tall, i.e. very high in acceleration, then because of the
limited amount of area that the pulse can contain it must also be very short in
duration. In fact, the duration is so short that the product cannot respond the
acceleration level of the event, only the energy input. Because the input
velocity change did not exceeded the critical velocity change of the product, no
damage has occurred.
When the velocity change of the input exceeds the critical
velocity change of the product, however, the only way to avoid damage is to
limit the input acceleration to a level below that of the product's critical
acceleration. This is usually one of the functions that a cushioned package
performs. It translates the high acceleration events experienced on the outside
of the container to lower acceleration events experienced inside at the unit.
To determine a damage boundary requires running two sets
of tests. A step velocity test is used to determine the product's critical
velocity change and a step acceleration test is used to determine the critical
acceleration.
1) STEP VELOCITY TEST
 To run the step velocity test, the unit is fixtured to the
table of the shock machine and subjected to a short duration pulse with a
relatively low velocity change content. It is important that the duration of the
event be so short that the unit cannot respond to the acceleration level of the
shock, only the velocity change. Most commercial machines intended for use in
this type of testing produce a half sine pulse 2-3 msec in duration. Following
the input, the test unit is examined for functional, physical and aesthetic
damage. If none has occurred, it is given an input with a slightly larger
velocity change component, but with roughly the same duration. This process is
repeated until damage to the unit has occurred. The last non-failure input
defines the critical velocity change for the unit in the orientation in which it
was tested.
The critical velocity change can be equated to an
equivalent free fall drop range (EFFDR). These numbers describe how far the
unpackaged unit can fall onto a rigid surface before damage will occur. The
process for calculating the EFFDR is demonstrated in figure 7. In essence, it
describes the ranges of heights that the bare product may fall before damage
will occur, based upon the type of surface it impacts. Undoubtedly, if you drop
the unit onto something soft, you can drop it farther without damage than you
could if you dropped onto something hard. This is why a range of heights is
defined, rather than one specific height above which damage will occur and below
which damage will not occur.
1) STEP ACCELERATION TEST
For the step acceleration test, a new unit is fixtured to the table of the shock
machine and given a low acceleration square wave pulse with a relatively large
velocity change content. The velocity change of the input must be at least 1.57
times the critical velocity change defined in the step velocity test. This will
ensure that the test is conducted past the knee of the damage region. This knee
occurs where the product shifts from being velocity sensitive to being
acceleration sensitive. Following the input, the test unit is examined for
damage and if none has occurred the unit is subjected to a slightly higher
acceleration level with roughly the same velocity change. This process is
continued until damage to the unit has occurred. The last non-failure input
defines the critical acceleration for the unit in the orientation in which it
was tested.
The designer now has all the information necessary to set
the shock protection requirements of his product. The critical velocity change
shows the designer the maximum drop height the bare product can be subjected to
before damage will occur. If that drop height is less than the design drop
height specified in STEP 1, a package or interface material is necessary. If a
package system is necessary, it must transmit less than the critical
acceleration to the unit, when dropped from the design drop height.
In a rigorous testing program, damage boundary curves are
generated for each orientation of the unit. To do this requires damaging 2 units
per orientation, one for the step velocity test and one for the step
acceleration test. Rarely are this many units available for destructive testing
in the prototype stage of a product's life, the most beneficial time to do this
type of work. Compromises are often made to limit the number of units which must
be damaged. In certain situations it is possible to perform the testing only
along the three orthogonal axes. In addition, the unit can often be repaired
between tests, so that one unit and a few spare parts may be used to perform all
of the testing.
 It should be noted that the square wave used to determine
the critical acceleration provides conservative results. In general, a square
wave of a given acceleration and duration is the most severe waveform possible.
It contains not only the fundamental frequency associated with the pulse
duration, but also all the higher harmonics associated with the quick rise and
decay of the waveform. What this means is that square wave can cause damage, at
a given acceleration level and velocity change, while other waveforms do not.
Figure 8 displays examples of the damage regions which various waveforms may
produce. As can be observed form this plot, the square wave encompasses all
damage regions produced by other waveforms. This is useful to the package
designer, because in the early stages of development it is not known what shape
waveform the package will transmit to the product. By using a square wave for
this test, we gain confidence in our final package design. We know that if the
unit can pass a square wave of a given velocity change and acceleration level,
we can pass any waveform that our package may produce of the same velocity
change and acceleration level.
In addition to the engineering reasons, there are also the
economic and practical reasons for using the square wave to determine the
critical acceleration. As can be noted from figure 8, the square wave produces a
flat horizontal line to bound its damage region while other waveforms tend to
make scalloped upward sloped shapes. Because of the square wave's flat line, we
can define the critical acceleration with just one set of tests. This is
something that can be done in an afternoon. To define the actual shape of the
damage region for other waveforms, requires performing tests along the entire
length of the velocity change axis. This approach can damage hundreds of units
and require weeks or even months to complete. Although it may provide more
precise results for a given waveform, it says nothing about how other waveforms
may cause damage. Usually the expense and effort to define the critical
acceleration in this manner is not warranted because no practical benefits are
gained.
VIBRATION: RESONANCE SEARCH & DWELL
It is generally agreed that damage due to vibration is unlikely except at those
frequencies where the product is most sensitive. The identification of those
frequencies, therefore, becomes critical in designing a package system. The
purpose of the bare product vibration testing is to identify the natural or
resonant frequencies of the critical components within the product.
 To run the vibration test, the unit
is secured to the table of a vibration test machine and subjected to a low level
sinusoidal input over a broad frequency range. The product can be observed for
resonances either visually, audibly, or fitted with response accelerometers
attached to its critical components. If the unit has been instrumented, the
table input and component responses are monitored through out the test.
Typically, the ratio of the component response to the table input acceleration
is plotted as a function of frequency and is called a transmissibility plot (see
Figure 9). The transmissibility ratio (response divided by input) reaches a peak
at the natural or resonant frequency of the component. The plotting of this
ratio comprises the resonance search portion of the testing.
Once the product resonant frequencies have been
determined, the vibration system is tuned to those frequencies and the product
is forced to dwell there for a predetermined length. This will identify those
frequencies which are prone to induce damage or fatigue.
If the unit is likely to be shipped in more than one axis,
the vibration sensitivities of the product in those axes should be investigated
as well.
|
