TestProNews Q2 - 2005
Change Test Data into Information
Learn about the most common reports used in design and manufacturing to improve
designs and processes
One of engineering and manufacturing managers' greatest challenges is making sense
of the large amounts of test data generated by automated test. This morass of test
data, often distributed throughout a company, has value only after it is turned
into information that engineers can use to validate and optimize their designs,
and managers can use to improve their company's design and manufacturing processes.
Enterprise test data management (TDM) systems generate
test data analysis reports and test performance analysis reports that change test
data into this valuable information (Figure 1). Test data analysis reports display
trends in measurement of parametric values, perform computational analysis on test
results, and show the most common reasons for failures. Test performance analysis
reports evaluate data from the automated test systems themselves, including yield,
cycle time, and station utilization.
|
Figure 1: Test data analysis reports and test performance reports change test data into valuable information that engineers and managers can use to improve their designs and manufacturing processes.
Enlarge Image
|
Test Data Analysis Reports
The simplest form of test data analysis is to manually retrieve and review raw test
data and reports that were previously generated by automated test systems.
An enterprise TDM system removes the complications of finding test data and gives
nearly unlimited flexibility in retrieving test data and reports based on:
|
- descriptive information, such as data from a specific product serial number
- property values, such as test date, location, station and operator
- failure data, such as data from all products that failed a particular test
- parametric data, such as all data for tests run at a temperature greater than a
specified temperature
- any combination of the above criteria
|
Once engineers retrieve the data that meets their criteria, they import the data
to interactive analysis software such as Microsoft Excel or National Instruments
DIAdem. The design group of a company that makes mixed signal ICs, for example,
uses
Arendar
Enterprise TDM software to capture, store, organize, and access their test data,
and then import data that meet specified criteria into Excel for interactive analysis
using pivot tables (Figure 2).
|
Figure 2: A mixed signal design group of a company that makes mixed signal ICs uses Arendar to capture, store, organize, and access their test data, and then sends the data that meet specified criteria into Excel for interactive analysis using pivot tables.
Enlarge Image
|
Another type of test data analysis report is the parametric data report,
which is typically used to analyze measurement values to determine products' specifications,
characterizations, reasons for failure and limits for maximum yield.
One type of parametric data report available in enterprise TDM systems that characterizes
the measurements from multiple products or tests is a histogram report. The histogram
aggregates the number of occurrences of a parameter to identify the most frequently
occurring values of the parameter. For example, a large defense contractor's design
group uses Arendar's histograms to evaluate the frequency of occurrence of the power-on
voltage on military equipment (Figure 3).
|
Figure 3: A large defense contractor's design group uses Arendar's histograms to show the frequency of occurrence of the power-on voltage on military equipment.
Enlarge Image
|
The histogram has two commonly used derivatives, the Pareto chart and the capability
histogram. A Pareto chart orders the number of occurrences from most frequently
occurring to least frequently occurring and adds a line chart to show the cumulative
percentage of occurrences. Engineers and managers use a Pareto chart to identify
what measurements have the most impact on their designs and processes. For example,
Figure 4 shows how a power supply manufacturer uses a Pareto chart to identify the
most common reason for failure.
|
Figure 4: A power supply manufacturer uses a Pareto chart to identify the most common reason for failure.
Enlarge Image
|
A capability histogram adds calculated distribution and control limits to
a histogram. Engineers and managers use the capability histogram to identify the
distribution of occurrences of a parametric value and can determine if a process
is in control. By using Arendar and a capability histogram, a materials development
company is able to determine the frequency distribution of materials used in RF
antennae (Figure 5).
|
Figure 5: A wireless broadband company uses Arendar and a capability histogram to determine the frequency distribution of antennae.
Enlarge Image
|
Other types of parametric data reports available in an enterprise TDM system identify
trends in the measurement data. With a parametric x-y plot, users visualize trends
in measurement values such as noise vs. temperature, input voltage vs. output voltage,
and frequency vs. power supply voltage. Figure 6 shows how a company that makes
analog ICs for communication devices plots jitter versus average temperature to
characterize their prototype ICs.
|
Figure 6: A company that designs analog ICs for communications uses a parametric x-y plot to analyze jitter versus average temperature to characterize their prototype ICs.
Enlarge Image
|
(Continued on next page)
|
|
|
|
Read about the latest information for integrating automated test systems throughout the enterprise. |
|
|
|
 |
Download PDF |
|
|
