Microwave Journal
www.microwavejournal.com/articles/1267-automated-test-and-data-management

Automated Test and Data Management

May 10, 2006

Spartan Automated Test and Data Management is a value-added software application for automating the measurement and evaluation of electrical performance parameters of RF and microwave components, subsystems and systems. Spartan adds value to the organization through efficient test automation and a built-in data management system.


Although the initial test implementation for Spartan has been S-parameter measurements, its architecture has been designed to serve as a test bed framework that can be expanded to accommodate a wide variety of tests that might be performed on RF and microwave devices. For example, in addition to S-parameter tests, Spartan has implemented a DC attenuation test capability, and passive intermodulation will be added within the next few months. The significance of this capability is that test results for items that are evaluated for a variety of performance parameters can be correlated. Furthermore, if tests are performed at various stages of product development or manufacturing, the results can be compared and variations evaluated. An example of this would be the testing of high reliability devices that are measured before, during and after high temperature burn-in.

Value Added in Test Automation

Companies large and small are looking to gain competitive advantage through cost reduction, faster delivery, specification improvement, product quality and customer support. To achieve these goals, companies employ programs such as value stream mapping, lean manufacturing and Six-Sigma Quality. Spartan can be an important contributor toward these goals.

In lean manufacturing and value stream mapping, the objective is to reduce waste and streamline the manufacturing process. A guiding principle in the design and development of Spartan is that it must add efficiency to the test process and users must be able to put it into use with a minimal impact on their current process.

Six-Sigma programs aim at identifying the causes of deviations and eliminating them with the ultimate benefit a decrease in manufacturing defects and improved reliability. Spartan supports Six-Sigma goals through the use of consistent test procedures and performance criteria. Additionally, the data mining features facilitate design and manufacturing enhancements and process improvement.

A Production Example

Spartan is being used by Polyphaser to perform production testing of their lightning protectors. Because Spartan is implemented as a Client-Server application, where various databases reside on the server and are accessible over a secure network, the test procedures are created by Polyphaser engineering in Nevada, and available for use at their production test stations in both Nevada and China.

In their production area, Spartan has been implemented with a bar code reader for data input, which minimizes the chances for error, and a touch screen monitor to keep the test area clear of computer accessories (keyboard and mouse). Testing proceeds as follows:

  • The person performing the test scans the part number bar code on the work order; this recalls the test procedure and acceptance specifications from the test sequences database

  • Spartan automatically sets up the vector network analyzer (VNA) and verifies calibration status

  • Prior to testing the first item, a second bar code is scanned to enter the Work Order Number. Because Polyphaser uses the Auto-Increment Serial Number feature, it is not necessary to enter the serial number. Spartan checks the database of prior tests and enters the next available serial number.

  • The first item to be tested is connected to the VNA and the operator presses Run on the touch screen

  • The measurement is performed and the operator is given a pass/fail status

  • The test operator disconnects that part, attaches the next part, and presses Run

It is useful to note that no knowledge of the test is required by the operator, so labor of all skill levels can be used. Also, the test process is reduced to the barest minimum to cut time and cost. The operator never touches the test equipment, so the test is run exactly as defined in the test procedure (see Figure 1) and the acceptance criteria are consistently applied. This improves data accuracy and reduces errors. In addition, all data is stored for future processing. If the network analyzer goes out of service for any reason, a different network analyzer can be substituted, and even if it is not the same model or from the same manufacturer, the tests can proceed exactly as before. The operator interface and test process are completely independent of the instrument.

An Engineering Example

The previously described process works great for production test, but would leave product development engineers unsatisfied. For product development or troubleshooting, an interactive version of the automated test exists. This allows the engineers to track the performance data on various iterations of the test and annotate the results as desired. In this mode, the analyzer can be running in a continuous sweep mode to monitor performance and then captured on command. Note that this same mode is useful in production testing where parts may be tuned to optimize performance prior to final test.

Test Data Management

Modern automated test systems generate a significant amount of data and within this data resides valuable information beyond a simple pass or fail decision, although that is certainly a first tier need. This information can be used to identify manufacturing trends, track first pass yields, match current production items to prior ship sets, improve product design, evaluate system performance models and more. The ability to readily access that data for analysis is a key feature of Spartan (see Figure 2).

Each time that a test is performed, Spartan saves the raw test data along with numerous variables. Some of these variables are system controlled such as time, date, model number and serial number, while others are defined by the organization using Spartan. For example, maybe it is important to be able to keep track of connector types on the product; this can then be added as a database variable.

Because Spartan includes a built-in data management system, authorized users can mine the data and filter it against any of the variables in the database using the Spartan query tool. If new test variables are added at some point in the future, the existing data is not impacted and the sort routines will work with both the old and the new data. This is an important advantage over systems without a data management system.

Data Mining Examples

Sorting, categorizing and analyzing data is extremely important. Some examples of the data management functions facilitated by Spartan include:

  • Trend Analysis – Compare performance between production lots to identify changes — good or bad.

  • Sort Parts by Performance – Categorize parts based on performance results and save the best parts for the most demanding requirements.

  • Compare Design Iterations – What are the variables affecting performance? Can cost be reduced through design changes? Material changes? Process changes?

  • Determine First Pass Yields – How often do parts need to be reworked? Are some manufacturing lines more consistent or producing better parts?

  • Phase and Amplitude Match to Prior Shipments – Compare and match electrical performance to ship sets previously supplied.

The Make vs. Buy Decision

Modern test instruments and software development tools make it easier than ever to realize an automated measurement capability. So why consider an investment in Spartan?

  • Usability – Summitek’s products have long been praised for the quality of the operator interface and ease of use.

  • Test Efficiency – The Spartan design is focused on streamlining the testing process.

  • Flexibility – Spartan provides flexibility in the types of tests that can be performed and the way in which to perform them.

  • QA Process Control – The Spartan Test Sequence Editor is used to define how measurements are to be performed and the acceptance criteria to assure everybody follows the same process each and every time.

  • Test Data Management – The ability to readily access test results for additional analysis and statistical processing is a key product differentiator.

  • Professional Software Development, Code Control and Long-term Support – Extensive testing before product release means the user is operational with minimal impact on his or her business. In addition, by participating in the Software Maintenance Program the user can stay current with Spartan test capabilities and features.

  • Cost – The investment in Spartan is readily justified relative to any alternative, including continuing to operate in a non-automated mode.

Conclusion

Spartan Automated Test and Data Management raises the bar in the test and evaluation of RF and microwave components and systems. Implementing Spartan in production facilitates lean manufacturing goals and will provide access to the data that can provide a competitive edge. The flexibility and power of Spartan allows the user to confidently outsource his or her test automation, so that the focus is on deriving benefit from the information contained within the data rather than the data collection.

Persons interested in Spartan can evaluate the product prior to purchase by contacting Summitek or visiting www.SpartanTest.com.

Summitek Instruments Inc.,
Englewood, CO (303) 768-8080,
www.summitekinstruments.com