At-a-glance chart of the J1879 Robustness Validation Standard

Download Report

Transcript At-a-glance chart of the J1879 Robustness Validation Standard

J1879 Robustness Validation Hand Book
A Joint SAE, ZVEI, JSAE, AEC Automotive Electronics Robustness Validation Plan
Robustness Diagram
Trends and Challenges
Failure
Mode A
VERY ROBUST
ROBUST
Robustness Curve - Sufficient Robustness
If the Area Of The Robustness Curve (blue ring in figure 1) lies
outside the specification limits, the robustness of the component
is sufficient. In general, the target is met if the robustness curve
lies outside the specification limits.
LEAST ROBUST
Robustness Margin
Application II
Parameter B
The current qualification and verification methods do not provide
statistical evidence that a device under test will meet customer
demand of parts per million failure rates. The correlation between
applied stress condition and lifetime under use condition is not
established. The result of a qualification is qualitative, under the
same requirements, it is possible to select the worst choice.
Analyzed returns from the field demonstrate that some tests are
not detecting deficiencies during the qualification process. In
consequence, a more pragmatic validation approach must be
introduced. Robustness validation seeks to define the guard band
between the outer limits of the specification and the component's
actual performance. Robustness is the degree to which a
component or system is impervious or resistant to factors
which can effect its function, performance or other identified
attribute or quality characteristic.
Robustness Decision
Application II
Robustness Curve – Insufficient Robustness
Strategies for improvement of insufficient robustness.
In case of insufficient robustness, improvement measures have
to be defined and implemented during the development phase.
Corrections for Unexpected Failures that are discovered after the
development phase should be corrected and added to the
knowledge base.
Specification
Solutions for Improvement
Robustness Validation Components
Robustness Validation has four key components
• Knowledge of use/application conditions - Mission Profile
• Knowledge of failure mechanisms, failure modes and their
interactions – Captured in the Knowledge Matrix database
• Acceleration models for the failure mechanisms needed for
defining and assessing accelerated tests
• Testing of the part until failure or determination of End of Life
Robustness Validation results in a product being qualified as “fit for
application”, not “fit for standard”. This approach requires more up
front communication and explanation between customer and
supplier than a stress-based qualification.
Parameter A
Figure 1.
The robustness diagram is a way to demonstrate robustness validation
results graphically. The figure above gives an example for two
parameters, i.e., Temperature and Voltage. The guard band, or safety
margin, between the limits of the specification and the component's
actual performance determines the component’s robustness as
indicated by the three robustness curves.
Qualification Process Flow Using
Robustness Validation
Input
Knowledge Matrix - A knowledge matrix is needed to capture
the basic mechanisms behind each potential failure
mechanism, the root cause(s) of each failure mechanism and
the effects of failure to the electronic component, product
performance and application. The knowledge matrix is the
database for generating a qualification plan based on the
application profile and for generating the reliability performance
numbers under use conditions.
Mission Profile - The mission profile determines, within a
specific group of applications, the range of environmental, life
time and manufacturing conditions to which the device is
exposed during its life. This life time includes the whole supply
chain (storage, shipping, processing, operating and nonoperating). Based on the mission profile the potential risks to
fail in the application together with the potential failure
mechanisms can be defined.
Robustness Assessment - A robustness assessment has to
be done separately for each failure mechanism. The overall
component robustness can be estimated by the failure mode’s
statistical data.
Quality Engineering
requirements
Environment defined by mission profile
Performance in the application spec
Potential risk and failure
mechanism
Failure mechanism
Known potential failure mechanisms
Unknown potential failure mechanisms
Qualification setup
Perform stress tests
Review stress set up
Design for Reliability
Screening strategy
Technology solution
Application review
Mission profile review
Improvement
Reliability Characterization
Robustness Assessment
NO
Qualification plan
Reliability tests, test conditions, duration &
acceptance, test vehicle, Sample size /
number of lots
Stress tests according to quality plan
Parameter change over time of stress
Fail distribution
Model used for extrapolation
Compare data to requirements
Compare delta to robustness target
Robustness
Sufficient?
YES
Monitor plan set up
Production
Figure 2.
Intrinsic reliability monitoring
Extrinsic reliability monitoring
Frequency of monitoring/sample size
Actions on deviations in monitoring results
Before implementing the solution for improving robustness, the
solution must be reviewed with respect to several other aspect,
i.e.:
• Does the expected improvement meet the robustness
target?
• Does the improvement solution influence the robustness
of other failure mechanisms?
• What is the implementation risk (probability that the
device fails in it’s implementation)?
Robustness ~ (1-Implementation Risk) It should be taken into
account that, in general, from a statistical point of view, nearly
everything that reduces the risk increases the robustness.
Expected Outcomes
• Mindset Change – Rather than relying on the current reliability
system of probability prediction, i.e., 95%Reliability /
90%Confidence (50,000 ppm), Robustness Validation provides
more and better data to prove whether or not a component of
system will achieve very high reliability in its intended
application.
• “Smarter” testing which is quicker, better and less expensive
than traditional past and present methods eliminates the wastes
associated with low-value testing/analysis and the risks
associated with insufficient testing/analysis.
• Better more efficient, effective, reliable methods for
validating changes to the design, process and application.
• Address the “real” problems, including NTF (no trouble
found), HW/SW interfaces, system issues and “soft” failures
such as out-of-spec performance.
• Provide methods for preventing and capturing problems
caused by unintended and undisclosed design/process
changes.