Test is critical to the board manufacturing process. Effective test ensures quality and customer satisfaction both for the OEM (original equipment manufacturer) and the CEM (contract electronics manufacturer). By isolating defects before product shipment, test minimizes returns and related costs. But test takes time, and the cost can be prohibitive.
In today's outsourced global manufacturing environment, collaboration among board design, layout, and test group —who must work together both to specify test coverage metrics and to define and check any design modifications that are required to achieve testability — is challenging. Testing today's dense boards requires the CEM or test house to develop complex complementary test strategies using a variety of machines. After a test strategy has been developed, test programs have to be generated and debugged for different equipment.
A number of tools exist to streamline the execution of these tasks. But in many cases, tools like point systems aren't helpful because complementary test strategies typically use testers from more than one vendor, and many of the upstream process steps for test (e.g., for recovery of CAD design and bill of materials data) also are required for assembly. In-house systems that are intended to eliminate duplication are extremely costly to develop and expensive to maintain, even for a large CEM.
In order to collaborate effectively, cross-functional design and test groups jointly responsible for test strategy development need to be able to view design data and added manufacturing data anywhere. One must be able to automatically recover all the board design data (including PCB layout, bill of materials, and circuit schematics) needed to drive manufacturing and test processes all the way from assembly-line balancing and test strategy definition, through generation of machine programs and factory floor documentation.
The test solution should be able to display the PCB layout graphically, just as the layout designer sees it on a PCB CAD system, together with multi-page circuit schematics. It should also display test data such as probe locations, assignments of components to different test systems, input lists, and documentation. Moreover, if a board design is revised to address a testability issue, the differences between the current and previous revisions should be easily evident, making it easier to verify that old problems have been solved and that new ones have not been introduced. Given visible design and manufacturing data, a test strategy can be built based on coverage, speed, and cost requirements.
Developing a test strategy for testing the components on a board today is a much more demanding task than it was in the days when close to 100 percent test coverage could be achieved with an in-circuit bed-of-nails tester. Increasingly complex and dense boards often have nets that are physically inaccessible for probing. A complementary test strategy needs to balance tests between different test systems that detect overlapping but nonetheless different failure ranges, minimizing duplicate tests for the same potential defect, while at the same time meeting test coverage and quality goals.
A reasonable first step in defining a complementary test strategy is to specify the preferred test system for each part number on a board based on test coverage and speed requirements, confidence in different test methods for different part numbers, and budgetary constraints.
Once you have completed your tests, PCB test software helps to eliminate the test bottleneck, facilitating the development of less costly complementary test strategies, and helps to check the viability of these strategies and identify design for testability issues more quickly. Once the test strategy has been finalized, the software speeds up the generation of shop-floor documentation, input lists, and target test systems, together with data required for fixture manufacture.
Quality engineers, plant supervisors, and managers need to see faults reported by test in real time, and must receive alerts if the faults exceed specified limits. Quality engineers need to determine whether the faults reported by test are caused by defects, and they need to be able to drill down to the root causes of defects so that corrective actions can be taken. If a defect has been introduced during the manufacturing process (e.g., a solder open), the process step responsible for introducing it needs to be identified and corrected to prevent recurrence.
It should be easy to determine whether a component defect is general in nature, if it only affects a particular batch of components, or is an isolated case. To enable the repair loop, the repair operator needs to be presented with board graphics that pinpoint defect locations on dense boards, together with historical data documenting successful past repair actions. Post-repair actions have been taken into consideration as well. All of these downstream processes can be managed by the new product information (NPI) and execution systems components of manufacturing process management (MPM) software. MPM's focus on process flow makes it particularly appropriate for test.
While there are several approaches to test process design and implementation, the wider range of applications, greater integration, and level of maintenance provided by independent NPI and executions systems software merit consideration in comparison with point solutions from equipment vendors and in-house solutions. Greater MPM value usually can be achieved by a more comprehensive independent approach that eliminates duplication, and is developed and supported in response to the needs of the global manufacturing industry.
This article was written by Antony Caulcutt, director of product marketing (NPI), and Santiago Cabezas, senior marketing manager, Tecnomatix Unicam, a subsidiary of Tecnomatix Technologies, Ltd. For more information, visit www.tecnomatix.com .