-
- News
- Books
Featured Books
- smt007 Magazine
Latest Issues
Current IssueBox Build
One trend is to add box build and final assembly to your product offering. In this issue, we explore the opportunities and risks of adding system assembly to your service portfolio.
IPC APEX EXPO 2024 Pre-show
This month’s issue devotes its pages to a comprehensive preview of the IPC APEX EXPO 2024 event. Whether your role is technical or business, if you're new-to-the-industry or seasoned veteran, you'll find value throughout this program.
Boost Your Sales
Every part of your business can be evaluated as a process, including your sales funnel. Optimizing your selling process requires a coordinated effort between marketing and sales. In this issue, industry experts in marketing and sales offer their best advice on how to boost your sales efforts.
- Articles
- Columns
Search Console
- Links
- Events
||| MENU - smt007 Magazine
Modern PCB Test Strategies
December 31, 1969 |Estimated reading time: 14 minutes
As automated test equipment is integrated into the assembly process, DFT must not only encompass traditional hardware access issues, but also include knowledge of test equipment diagnostic capability.
By John W. Ledden
Design for test (DFT) is not the job of an individual, but a group with representatives from design engineering, test engineering, manufacturing and purchasing. Design engineering must specify the functional product and tolerance requirement. Test engineering must produce a strategy that gives the highest possible first-pass yield (FPY) at the lowest cost with the least amount of rework. Manufacturing and quality must provide input on production cost, information on what has and has not worked in the past for similar products, as well as help with design for volume (DFV) ramp-up issues. Purchasing must provide information on available devices, especially reliability. Test engineering and purchasing should work together when buying devices with on-board test hardware to ensure these devices are available and easy to implement. Process improvement by using the test systems as sensors to collect relevant historical data should be a quality group goal. All these functions should be completed before placing/removing any nodal access.
Parameters
Preparation and understanding are key before setting policy in a test environment. Parameters that affect test strategy include:
Accessibility. Complete access and large test pads are always the goal when designing a board for manufacture. The four reasons normally given when full access is not provided are:
- Board size. Designs are smaller; the issue is the amount of "extra" board room for test pads. Unfortunately, most design engineers assume test pad access is the least important thing on the printed circuit board (PCB). This is not the case when the product must be debugged by design engineers because the simpler diagnostics of an in-circuit tester (ICT) cannot be used. Test options are limited if full access is not provided.
- Functionality. Lost performance in high-speed designs affects board sections but can be worked around to minimize its effect on product testability.
- Board size/node count. This is when the physical board size cannot be tested on any available equipment. Fortunately, this problem can be solved with a budget increase either for new test equipment or using an outside test facility. When the node count is greater than available ICT, the issue is harder to resolve. The DFT group must have knowledge of test methods that will allow manufacturing to ship a good product using the least amount of time and money. Built-in self-test, boundary scan (BS) and functional block testing can do this. The diagnostics must support the unit under test (UUT); this can be done only with an in-depth knowledge of the test methods used, available test equipment and capability, and the manufacturing environment's fault spectrum.
- DFT rules not used, followed or understood. Historically, DFT rules were enforced by one engineer or a small group of engineers that understood the manufacturing environment, process and functional test requirements, and device technology. In a manual environment, the process is long and requires communication between design, computer-aided design (CAD) and test. This tedious and repetitive work is prone to human error and is often rushed because of time-to-market pressure. There has been a move in the industry to use automatic "producibility analyzers" to evaluate CAD files using DFT rules. When contract manufacturers (CM) are used, multiple rule sets can be catalogued. Rule consistencies and error-free product evaluation are advantages of this approach.
Test Equipment Availability
The DFT group should be aware of existing test strategy. With OEMs moving toward greater CM dependence, the equipment used varies from site to site. Without clearly understanding the manufacturers' process, too much or too little test could be used. Existing test methods include:
1. Manual and automated vision testing verifies component placement on a PCB using visualization and comparison. This technique has several implementations:
Manual vision is the most widely used in-line test, but it is becoming less viable as manufacturing volume increases and board designs and components shrink.
Its main advantages are low up-front cost and no test fixture, while its main disadvantages are high long-term cost, inconsistent fault detection, difficult data collection, nonelectrical test and limited line of sight.
Automated optical inspection (AOI) typically is used before or after reflow and is a relatively new approach to identifying manufacturing defects. It is a nonelectrical, fixtureless, in-line technique that uses "learn and compare" programming to minimize ramp-up time. Automated vision does a good job on polarity, presence and absence of components as long as the second-source components are similar to the original learned ones.
Its main advantages include easy to follow diagnostic, quick and easy program development, and no test fixture. Its main disadvantages are poor short detection, high false-failure rate and it is not an electrical test.
Automated X-ray inspection (AXI) currently is the only method that tests ball grid arrays (BGA) for solder quality and hidden solder balls. It is a nonelectrical, non-contact technique that finds process faults early, reducing work-in-process (WIP). Advancements in this field include pass/fail data and component-level diagnostics. There are two major AXI methods currently available: 2-D, which looks through the complete board, and 3-D, which takes multiple pictures at different angles.
Its main advantages are that it is the only tool available for BGA solder quality and embedded components, and the lack of fixture costs. Its main disadvantages are low speed, high false-failure rate, difficulty of testing reworked joints, high per-board cost and a long program development time.
2. Manufacturing defect analyzers (MDA) are a good tool for a high-volume/low-mix environment in which test is used to diagnose manufacturing defects only. Repeatability between testers is an issue when a residual subtraction technique is not used. Also, MDAs do not have digital drivers and therefore cannot test parts functionally or program on-board firmware. Test time is less than with vision so MDAs can keep up with the line beat rate. This method employs a bed of nails so the diagnostic output can be followed.
Its main advantages are lower up-front cost, lower WIP, low programming and program maintenance cost, high throughput, easy-to-follow diagnostics, and fast and thorough shorts and opens testing. Its main disadvantages are that it cannot verify that the bill of material (BOM) matches the UUT, no digital verification, no functional capability, no ability to load firmware, usually no test coverage indication, repeatability between board runs and lines, fixturing costs, and access issues.
3. ICT will find manufacturing defects as well as test analog, digital and mixed-signal devices to ensure they meet specifications. Many have the capability of programming on-board memory, including serial number, pass/fail and genealogy data. Some make program generation easier by imbedding tools in easy-to-use graphical user interfaces (GUI) and storing the code in a separate file to enable multiple version testing and easy firmware changes. Some testers have sophisticated instrumentation that will verify the functional aspects of the UUT, as well as interface with commercially available instruments. There are testers that have an embedded CAD interface and a non-multiplexed environment to shorten development time. Finally, some testers provide in-depth UUT coverage analysis that details the parts that are and are not being tested.
ICT's main advantages are low test cost per board, digital and functional capability, high throughput, excellent diagnostics, fast and thorough shorts and opens testing, program firmware, fault coverage report, and programming ease. Its main disadvantages are fixture, program and debug time, fixture cost, up-front expense, and access issues.
4. Flying-probe testers have gained popularity in the last few years because of advancements in mechanical accuracy, speed and reliability. Additionally, the current market requirement for a fast-turn, fixtureless test system for prototypes, low-volume manufacturing operations and verification has made flying-probe a desirable test option. The best probe solutions offer learn capability as well as a BOM test with guards added automatically during the test generation process. The probe's software should offer an easy way to port CAD data because X-Y and BOM data must be used in program generation. Because nodal access may not be complete on one side of the board, test generation software automatically should generate a split program with no redundancy.
Probes test for the connection of digital, analog and mixed-signal components using vectorless techniques; this should be done with capacitive plates available to the user on both sides of the UUT.
The main advantages of flying-probe testers are that it is the fastest time-to-market tool, automatic test generation, no fixturing costs, good diagnostics and programming ease. The main disadvantages are low throughput, limited digital coverage, capital expense and access issues.
5. Functional test, which could be considered the first automated test philosophy, has seen a resurgence in importance. It is board- or unit-specific and can be performed by various equipment. Some examples are:
Final product test is the most common functional test method. Testing the final unit after assembly is inexpensive and cuts down on operations with few faults. However, diagnostics are nonexistent or difficult, which adds cost. With test only in the final product, there is a chance that it could be destroyed without the software or hardware protection that automatic test provides. Final product testing also is slow and typically uses a large amount of floor space. It typically is not used when a standard must be met because parametric measurements usually are not supported.
Final product test's main advantages are lowest initial cost, one-time assembly, and product and quality assurance. Its main disadvantages include low diagnostic resolution, lack of speed, high long-term cost, FPY, board or machine damage because of undetected shorts, expensive rework, and no parametric testing capability.
Hot mock-ups generally are positioned at various assembly stages, rather than just at final test. Diagnostically, this is better than final product test, but it is more expensive because special test units must be built. Hot mock-ups can be faster than final product test if the program is fine-tuned to test one specific board only. Unfortunately, because of a lack of protection, it is possible to destroy the test bed if shorts are not diagnosed earlier in the process.
The main advantage of hot mock-ups is the low initial cost. The main disadvantages are inefficient floor space use, cost to maintain test equipment, damage to UUT because of shorts and no parametric testing capability.
Software-controlled, commercially available instruments are commonly called "rack and stack" test because the instruments are purchased separately and then connected. The software that synchronizes the equipment usually is completely customized. Commercially available instruments are inexpensive compared to integrated solutions and, if done correctly, allow an independent UUT validation. But these "homemade" systems typically are slower, and engineering changes and production floor support are difficult because these applications are under-documented.
Their main advantages are protection against UUT damage, faster throughput, less floor space required, and independent/industry-accepted verification. The main disadvantages are that they are time-consuming, difficult to support, update and use at remote facilities.
Commercial, custom integrated systems couple software and hardware on one test platform, e.g., IEEE, VXI, Compact PCI or PXI. Documentation, software support and standard manufacturing concepts make these systems easier to use and support. The up-front cost is higher than an internally built solution, but this cost is justifiable because of higher performance, throughput and repeatability. It also is easier to support on the production floor and during new product development.The main advantages are fast throughput, low floor-space requirement, easiest to support and reconfigure, best repeatability, and provides independent industry-accepted verification. The main disadvantage is the high initial cost.
6. Non-contact test methods such as laser systems are the latest development in PCB test technology. This technique already is proven in the bare-board arena and is being considered for populated board testing. The technology finds faults with only line-of-sight, non-masked access. At less that 10 milliseconds per test, it is fast enough to use in a volume production line.
Fast throughput, no required fixturing, and line-of-sight/non-masked access are its main advantages; not production tested, highest initial cost, high maintenance and access issues are its main disadvantages.
Table 1 summarizes the described test methods.
Test Methods and Fault Coverage
It is important to understand available test methods and fault coverage before developing a test strategy. There are many electrical test methods with various amounts of fault coverage.
Shorts and opens. MDAs and ICT are good at detecting shorts they have bed-of-nail access to each electrical node and can measure resistance between nets for shorts. Bare-board testers use a capacitance-to-ground technique that, if restricted to bare boards, is effective and fast. Flying probes use both the capacitive technique and a proximity shorts technique; the former is not repeatable enough for most manufacturing facilities and lacks good diagnostics. The best proximity tests use raw CAD data to identify the trace location and allow the programmer to select the maximum distance between test nodes. This gives some control over test speed; however, it is recommended that functional equipment have current-clamping or fold-back supplies to prevent board or tester damage, as low-impedance shorts through devices cannot be detected during a shorts test alone.
Passive analog ensures acceptable process quality by verifying that the UUT is soldered to the board and the correct value component was installed. This testing often is done when there is a minimum amount of WIP so the problem can be corrected before a great deal of bad product is built. Power is not applied to the board and selective passive or active guards are used to zero-out parallel current paths. Bed-of-nail access is needed for the UUT and surrounding guard locations.
Vision systems provide device-level diagnostics. They use a golden board and compare that to the UUT without electrical tests. MDAs provide an electrical test and component-level diagnostics, again compared to a known good board. ICT tests electrically and provides device-level diagnostics comparing value and tolerance to the BOM. Functional testers test to a designer's specification (often a known good board). If functional tests are thorough, it ensures the product can be shipped. However, if FPY is not extremely high, the manufacturer runs the risk of bad product buildup, waste and expensive manual diagnostics/rework.
Active analog. ICT, functional testers and flying probes that apply power to the board do a good job of detecting bad active analog components. ICT and flying probes, while providing pin-level diagnostics, do not measure some critical manufacturer specifications (e.g., bandwidth, input offset current, etc.). Functional testers measure output characteristics without providing pin-level diagnostics. MDAs help with vectorless techniques and vision systems prove presence only. X-ray provides solder quality diagnostics.
Digital and mixed-signal component testing. Vision, X-ray and MDAs only diagnose opens and shorts. ICT uses various methods depending on the device, circuit and accessibility. It can use a vectorless technique for continuity only, BS for continuity and device identification when full access is provided. Creating a model for a particular component by manual vector generation can be time-consuming and may not cover enough faults to justify the effort. A combination strategy of a vectorless technique for connectivity and a limited vector test to ensure device operation could be used to maximize coverage while limiting development time.
Functional systems test the circuit/module to a design specification but lack the pin-level/component-level diagnostics needed for inexpensive repair. In most cases, functional test does not provide the in-depth data necessary for process improvement. Both functional and ICT program on-board flash, in-system programmable and serial on-board memory devices (Table 2).
Manufacturing Test Strategy
No one strategy will or should suit all manufacturers. Numerous variables must be considered when developing a test and process improvement strategy.
Manufacturing fault spectrum identification should be factory- and product-specific. These data, if relevant and dependable, lower personnel and scrap cost and increase customer confidence. Fault data should be collected, compiled and discussed at regular meetings driven by the quality group. The data also should be used to develop a test strategy that detects common and preventable faults. These data should include both factory and field failures with date identification. New products should be monitored for faults, while mature products should be monitored to improve FPY and supplier quality. Fault data should be compared long-term and short-term internally and in conjunction with other sites to improve total quality. Weather conditions, staff, supplier and line-change data should be tracked, as these often are hidden quality factors.
Two important quality factors are relevant data collection and distributive test. A sensor's ability to collect data that will serve to improve quality, and the data manager's ability to deliver that data to the correct group, impacts current and future products. The definition of correct data depends on the facility and product. Testers act as sensors and monitor the process. An effective distribution test strategy identifies process problems as close to their source as possible, reducing the quantity of bad products produced.
This article is adapted from a presentation originally given at NEPCON West 2000.
ACKNOWLEDGEMENTS
The author expresses his appreciation to Alan Albee for his contributions to this paper, as well as colleagues at GenRad for their assistance in the review process.
JOHN W. LEDDEN may be contacted at GenRad Inc., 7 Technology Park Dr., Westford, MA 01886-0033; (978) 589-7000; E-mail:leddenj@genrad.com.