Issue link: https://iconnect007.uberflip.com/i/1345210
MARCH 2021 I SMT007 MAGAZINE 29 reliability, soldermask registration, and imped- ance control. Results are typically available in an online report or downloadable for pro- cessing. We compare the raw test data against an internal grading criterion to establish pass/ marginal/fail limits based on company-specific needs. As the quote model concept evolved, the original low/medium/high stratifications were supplemented with more than a dozen param- eters related to specific board features. e smallest feature geometry on which a sup- plier can consistently meet the internal criteria drives that supplier's rating for that feature. By making PCQR 2 submission an annual require- ment, vendors regularly demonstrate their ability to maintain their processes and the tan- gible results of their capital expenditures and process improvements. Using PCQR 2 Results ese results can be quickly and objectively applied to our quote model. Figure 4 shows an example of quote model database logic. Bool- ean criteria are based on specific features the suppliers did or did not implement on their chosen submission, as well as other parame- ters such as ITAR capability and those estab- lished by audit. Minimum/maximum values are objectively established by PCQR 2 perfor- mance for each feature. For example, consider the internal conduc- tors of the 24VH-E panel. at test vehicle has internal conductors of widths 0.05, 0.075, 0.1, and 0.125 mm (2, 3, 4, and 5 mils). Most ven- dors are unsuccessful with 2-mil traces when compared against our grading criteria, while some succeed with 3-mil traces, more with 4-mil traces, and all with 5 mils. A shop that does not meet criteria at 4 mils but passes at 5 mils would be rated for a minimum of 5, while one that fails at 3 mils but is marginal at 4 and passes at 5, might be rated better (such as 4.5 mils) depending on the marginal performance. A flexible, customizable quote model can sup- port other exceptions, such as that shown for Supplier 5 in Figure 4 where sequential lamina- tion capability requires consultation with the PCB commodity team. Collating the results of multiple submissions enables trend analysis. Data from four different PCQR 2 submissions are shown in Figure 5. e data in this example could be used in a trend analysis over time at Supplier #1 Site A, or between Supplier #1 Sites A and B, or between Suppliers #1 and #2. e results of such a trend Figure 4: Example of a quote model database.