Issue link: https://iconnect007.uberflip.com/i/871322
September 2017 • The PCB Magazine 23 PCB PROCESS ENGINEERING: DETAILS FROM ONE OF THE ORIGINALS Figure 6: Statistical process control and Cp can lead to enhanced design rules and capability. means. But everybody that buys their tools is going to be part of this supply chain where we each have to supply to the next stage in the Val- ue Chain until it gets in the customer's hands. What designers really need is feedback from manufacturers, fabricators that can recognize the design-related interactions with many fac- toring processes, so that they can feed back to the designers. "When you come to this branch in your design, go this way and not this way be- cause this way is going cost you less and you're going to deliver the better product, cheaper, faster. But these other two branches, although they look attractive, are going to lead you down a path that's going to make this thing very dif- ficult to manufacture." And it's typically not as simple as "we can't make it." It's that we're go- ing to have a 3% yield loss or a 7% yield loss, not a 95% yield loss. It's all these subtleties that process engineers spend their days working on. Because this is such a complicated manufacturing process that has so many variables in it, but the biggest vari- able, which we can't quantify, is the variable of the design itself. That's why I created a new parameter called, "The Complexity Index," so I could rank designs by how complicated they were. That's given me an independent variable to look at important factors in manufacturing. Otherwise, you're looking at just the number of holes, hole diameter, number of layers, type of material, size of the panel, size of the board, etc. You end up working with one independent variable to look at all the significant results, in- stead of ten independent variables; it's pretty difficult to relate how that many affect manu- facturing. Figure 7 shows how the Complexi- ty Index is used to model first-pass yield in the manufacturing process. One of the "25 Essential Skills" is a thing called dimensional analysis [4] , which refers to the dimensionless parameters used in engineer- ing as pseudo-independent variables, like the Reynolds number or the Mach number. I figure that we have as much right to invent these as anybody else.