Issue link: https://iconnect007.uberflip.com/i/1000349
44 SMT007 MAGAZINE I JULY 2018 ity problem that wasn't detected during initial testing but became widespread in the service environment, demonstrating how expensive it can be if key reliability tests are not performed when appropriate. Testing for mechanical performance is primarily driven by the alloy, so any manufac- turer that desires to replace a SAC305 paste with another SAC305 paste can rely on past testing in most cases with certain market segments as possible exceptions (e.g., medical, aerospace, automotive). Examples of mechanical reliabil- ity tests include vibration, drop/shock, ther- mal cycling, and high-temperature endurance. Vibration and drop/shock test mechanical stress robustness, thermal cycling tests fatigue life, and high temperature endurance tests resistance to creep. The actual parameters for these tests, such as temperature, test duration, vibration amplitude, and drop height heavily depend on the product and its intended service environment. Testing for electrochemical reliability is of key importance for no-clean processes. Elec- trochemical reliability testing requires the use of a temperature/humidity/bias (THB) test condition. At the material level, suppliers use various surface insulation resistance (SIR) tests to demonstrate that their material meets the minimum level of performance required by the specification. However, these tests all use one specific bias, design, and environmental condition set so they cannot replicate all possi- ble design combinations. Therefore, testing on assemblies while powered and exposed to appropriate environmental conditions is criti- cal to validating the suitability of any no-clean flux for use on an assembly. T esting for potential electrochemical reli- ability defects is also important for processes that employ cleaning after soldering, but those processes require a focus on ensuring clean- liness after cleaning. Processes that employ cleaning can typically rely on extraction test- ing (e.g., resistance of solvent extract, iono- graph, ion chromatography) on assemblies to demonstr ate the effectiveness of the clean- ing process. The THB testing described for no-clean assemblies can also be performed on cleaned assemblies as a means of assur- ing that the cleaning process results in reliable assemblies at the test limits used for extraction testing. The reader is cautioned that extraction testing on uncleaned no-clean materials, used alone and without correlation to parallel THB testing on assemblies, is not suitable as a reli - ability evaluation test. Finally, any good test plan has a focus on quantitative measures developed under controlled and consistent conditions. If anom- alies are observed in the data while it is being analyzed, the data should be considered suspect and an effort to determine the cause of the anomaly and eliminate it should be performed. This may also require performing re-testing in some cases as a good evaluation is one where special cause factors are not unfairly present on a subset of the pastes under test. Do not be afraid to reject bad data, even if that is an unpopular choice! You can follow good data where it leads you and expect success- ful use of the chosen solder paste for years to come. The focus should be on performing the testing well and collecting good data on the last attempt, rather than requiring that the test is done perfectly on the first and only attempt planned. SMT007 Jason Fullerton is a customer technical support engineer at Alpha Assembly Solutions.