Issue link: https://iconnect007.uberflip.com/i/1345210
26 SMT007 MAGAZINE I MARCH 2021 Introduction Our company provides products and sys- tems for a wide variety of applications. is requires a supply chain that can support and thrive in a high-mix, low-volume environ- ment. With over 1,000 unique PCB designs, one of our challenges is sourcing boards to the "right" shop with the best mix of technology, capacity, and cost. Compounding this chal- lenge, PCB sourcing is decentralized, so each R&D team has the ultimate decision on which shop is awarded the business. A data-based approach to supplier qualification using the IPC-9151D Process Capability, Quality, and Relative Reliability (PCQR 2 ) Benchmark Test Standard, built into a company quote model, guides each request for quotes to the correct subset of suppliers based on the board's tech- nical requirements and allows R&D teams to focus on price and delivery when making the final vendor selection. Background Prior to PCQR 2 , the PCB quoting environ- ment was haphazard. Many sourcing decisions were of a subjective nature, based on individ- ual R&D teams' past experience with a given vendor regardless of PCB technology require- ments. Quote requests were sent in various formats (spreadsheets, documents, handwrit- ten, email, etc.), sometimes lacking data perti- nent to the PCB cost. Every supplier received every quote request, so some shops were being asked to quote, and in some cases, build boards for which they were not qualified. When this resulted in failures during prototype valida- tion, R&D was set back and the product release delayed. In other cases, the experience of employees at the suppliers, working above and beyond the capabilities of their processes and equipment, could deliver boards of sufficient quality for validation and production. Such a production methodology—where key workers produce "art" instead of following a process to yield a product—represented a risk to supply continuity that presented itself as a challenge to our PCB commodity team. Quote Model To solve this challenge, the team needed a dynamic, automated solution, one that would scale with and adapt to changes in the sup- ply base, technology, and market. e solu- tion chosen was to develop a company quote model. A standard quote form, filled out by the PCB designer with the board specifica- tions, expected volume, contacts, etc., cap- tured data in a uniform format and ensured completeness. e PCB specifications could then be fed into the quote model database to determine qualified suppliers. Initially, suppli- ers were stratified into low, medium, and high technical capability, and only those suppliers deemed qualified to build a board, given its technology, would receive a request for quote. e R&D team could be assured that each and every received quote was valid, and they could evaluate based on a simpler subset of criteria: prototype or production costs, and lead times. How, then, should we determine the profi- ciency of each supplier at each facility? Most vendors state their process capabilities on their websites. ese are useful to determine rough competence, but they are also never 100% accurate or reliable. Onsite audits were and continue to be helpful, but these are still some- what subjective, and using them exclusively to qualify suppliers could require us to retain or hire content experts above and beyond those required for any other commodity. A custom qualification board would allow shops to dem- onstrate their capabilities on the specific tech- nologies required for the next few generations of company products, but this would require the commodity team to develop test vehicles and manage the evaluation of each sample, and it would be limited to only those suppliers who had received the test vehicle and built it. A bet- ter option was to use a third-party test vehicle, and the PCB commodity team chose the IPC- 9151D test standard.