Issue link: https://iconnect007.uberflip.com/i/1000349
42 SMT007 MAGAZINE I JULY 2018 depth and breadth of any test plan will depend on the available resources being engaged effi- ciently but not overtaxed, such that a result takes an inordinate time to determine. Now that there is an understanding about the general scope of the test plan, the next step is to determine the specific focus of the test- ing that will be performed. There are two key areas that any evaluation should investigate: quality and reliability. Quality is defined by the American Society for Quality (ASQ) as "the characteristics of a product… that bear on its ability to satisfy stated or implied needs." In this case, the focus is on the key process output variables (KPOVs) from the solder print and reflow processes that are used to ensure the quality of the resulting printed circuit board assembly (PCBA). This varies by product function and design, so these key factors vary by assembler. The best place to find the KPOVs for any process is the process control plan, if one exists. These factors are the factors that are controlled for or inspected in the current process and represent the effort to make the product to specification. The most obvious of the KPOVs come from solder paste inspection (SPI) systems: print volume, which is typically normalized as transfer efficiency (actual print volume divided by aperture theoretical volume, expressed as a percentage). Other SPI-based outputs include area coverage and height, which are best used as a supplement to volumetric measurements. It is important to analyze SPI data as a func- tion of stencil area ratio for each aperture, as the distribution of transfer efficiencies will be a function of area ratio (A/R). Combin- ing all the data together will result in an over- all data distribution that is a combination of many different sub-distributions. For example, if testing a paste with a Type 4 powder size distribution (per IPC/ANSI-J-STD-005), the transfer efficiency for apertures with an A/R above 0.8 should be very close to 100% and have a distribution with low variation. The same paste, when tested on an aperture with an A/R of 0.50 will have a very different distri- bution with an expected lower transfer effi- ciency and higher variation. Combining the data from multiple A/R apertures can mask the