Issue link: https://iconnect007.uberflip.com/i/710139
16 SMT Magazine • August 2016 to quickly retrieve data and eventually deduce a statistical record of defects and their possible causes. Samples may be subjected to metallog- raphy and other destructive tests, which will be often the most important when considering the disposition of hardware. These dispositions will often fall under the categories: • Suitable for use • Suitable for use only after article has been repaired or refurbished • Not acceptable, return to supplier • Not acceptable, scrap Again, keep the records in a machine-read- able and searchable format. Hardware that has been deemed to fail in ser- vice may also be forwarded to the laboratory in order to try to establish the mode of failure; typi- cally, these will fall under the broad headings: • Material defect (poor choice of material, wrong composition, brittle inclusions in metallic alloys) • Processing defects (soldering, brazing and welding defects, incorrect curing of organic materials) • Human error (machining defects, poor workmanship, wrong units of measurement) • Service defects (resulting from over testing, environmental mechanical loading (shock, vibration, thermal cycling fatigue, corrosion and eventually, wear-out at end of expected life) The most important documents generated by the laboratory are the laboratory or technical reports. Again, keep the records in a machine- readable and searchable format. Too often during material or failure review meetings between manufacturers, suppliers and customers, the sales or technical support staff are unable to provide data concerning accep- tance or reject rates. When failures do occur in service there is often insufficient readily avail- able data available for participants. Customers need to know the frequency of such failures, the probable failure modes and the environ- mental conditions that may have contributed; sheet for any unusual or pre-release parts used. There are so many new parts introduced in this business, and sometimes a customer specifies one not yet in general circulation. Finding out process information on some of these exotic sensors is next to impossible without the right contacts, so please be transparent about the in- formation available regarding new parts. Dave Cusumano VP Engineering SALINE LECTRONICS In a perfect situation, we would prefer that our customers supply ODB++ files with their Gerber files and a BoM in MS Excel format. This will allow us to be more efficient when de- veloping the manufacturing process (with less chance for error) for each customer's assembly. Barrie Dunn, PhD Professor, School of Engi- neering UNIVERSITY OF PORTSMOUTH Companies engaged in the production of high-reliability electronics will have invested in capital equipment required for performing fail- ure investigations and documenting findings as laboratory or technical reports. Laboratories and personnel tasked to carry out these activities also exist within the user industries. Suspect or 'non-conforming to specification' items received at in-coming inspection might be electrically tested then examined by visual, scanning elec - tron microscopy, acoustic imaging or by a host of other non-destructive methods. It is essential to keep the findings as company reports. Such documents need to be stored in machine-read- able and searchable formats. This is essential in order for lessons to be learnt and staff to be able VOICES OF THE INDUSTRY X X