Proprietary Test Software Enhances Repeatability and Trends Analysis

OLYMPUS DIGITAL CAMERAProducts with complex test requirements and legacy products often create challenges in outsourced manufacturing. Are the latest bench test procedures in place? Are legacy components creating unacceptable fallout levels? Are multiple tests being performed in the correct sequence? Is enforced routing in place to ensure repaired assemblies complete all retest steps? The team at Burton Industries wanted to automate their processes to the point where the answers to these questions was always “yes.” In early 2019, Burton Industries’ Test Engineering and IT departments teamed up to develop test software that enhances the test team’s ability to support customers with complex test requirements and/or legacy products.

“Some of our customers require an in-circuit test (ICT), followed by a functional test and then finally a read of four-to-five voltages with a digital multimeter to make sure it hits certain values. Initially we wanted to automate the bench test process with software programming that would ensure that every test was performed, exactly to the defined customer requirements,” said Adam Waite, Test Engineering Manager.

This program has reduced the potential for human error on two levels.

Utilizing the system, the technician types in the number registering on the tester or multimeter and the program checks it against the pre-defined range. It then automatically records that in the defect yield tracking system and sends this to the continuous improvement database so that trends and needs for improvement can be analyzed.

As the test operator passes units, the quality assurance inspector scans all bar codes to convert work-in-process (WIP) to finished goods for shipment. The program automatically verifies that all routed test steps have been completed and that any defects logged against the board have been cleared.

The enforced routing that this system drives ensures that no steps get missed and that defective printed circuit boards assemblies (PCBAs) are properly handled. However, it doesn’t stop there. For example, if a customer narrows a test range, the automated system also helps enforce change management because it is verifying that ranges entered match the latest test specification revision. That is very important with new products, because there can be rapid changes as a new product evolves. It isn’t unusual to see three versions of test from prototyping to preproduction to production.

The program supports any aspect of ICT, bench test or functional test. An unanticipated benefit of the system is its contribution to root cause identification and better continuous improvement trends analysis.

“The program allows us to quickly gain access to first pass yield at the sequence level and see patterns in the level at which that failure occurs. That gives us a better idea of why a technician may be testing a circuit several times. This is very important in building legacy product where use of older last buy parts may result in a higher component failures. We have no choice but to use those parts. Understanding trends can help us structure a more efficient test process,” added Adam.

For example, in one legacy project an end-of-life ASIC IC had frequent failures. The customer had instructed the team to test it up to seven times which could lengthen test times to 20 minutes. The test database was able to generate a trends report that helped convince the customer that pretesting these parts was more efficient. The end result was shorter, more predictable final tests. This improved on-time delivery performance. Additionally, since bad components were screened out before assembly, fewer PCBA failures translated to less rework and the associated thermal stress that occurs with multiple reflow processes.

“We originally wanted to give our team a soft poka-yoke tool to minimize human error in test. However, now we are seeing the power of data. We are looking at the top three assemblies that are causing excess troubleshooting time in test to determine root cause. We’ve found that 80 percent of the time when on-time delivery metrics are off, the cause is extra troubleshooting in test. Consequently, we see this as a powerful tool for delivering better customer service,” said Adam.

 

Q3 2019 Newsletter

Burton Q3 2019 Newsletter

Download Burton Industries’ Q3 2019 newsletter here.