Issue link: https://iconnect007.uberflip.com/i/899995
November 2017 • The PCB Design Magazine 19 I think the companies that embrace team- work in all facets are going to win. When you start talking about the designers, do they know that the prototype is going to go into big pro- duction? In some cases, they do, but if they had a system to always call the fab house to have that relationship, why not call the assembly house and get all the feedback before the design is completely made? Then you have a winner in all facets and you're not chasing your tail after the fact. You're saving time and money upfront. The goal is to work together as a team and share our expertise with one another. Matties: That's a great vision. One of the things that we started talking about some years ago was DFP, design for profitability. If it's a smart design, everybody makes money. What do we do to solve that? Bird: I've been in this industry since '79 when Happy and I worked together at HP, and the things that we care about now are so different from the things we cared about then. For in- stance, who cared about the height of a trace or a trapezoidal shape back then? Now we're try- ing to squeeze gaps down, and all a sudden, we get this metallic bathtub where we are hit- ting the spacing limits because we can't etch out the gaps cleanly. We use EN- EPIG a lot. But now we're getting to a point where nickel, the thickest metal in the stack, is too lossy. It just goes against our density and performance goals. Therefore, we are going to be evaluating EPIG, actually more accurately EPAG (electroless palladium/au- tocatalytic gold). We're looking at metal thicknesses and trace fea- tures that I never thought we would see during my career. Also, as PCB technologists, we have to be willing to go out and scour the globe, not just within our approved vendor database, for what's up to date and then next year do it again and then do it again, be- cause things are changing so quickly. The best thing we can deliver to the engineering com- munity? Put together a package that is properly perfect circuit, but it has to be manufacturable, and that's where our input comes in. Matties: One of the things that Happy men- tioned in a more recent conversation is design for automation. Happy, why don't you talk a little bit about that. Holden: Well, with design for automation, when you start investing in very expensive au- tomated equipment, you can't afford to replace that equipment over time because designers keep trying to innovate and doing anything they want. Instead, if you've got to get payback and have a cost-effective product, this is the capability of the automated equipment. You somehow have to design within the constraint of that; otherwise you can't run it down this really great automated line. We're back to hand- crafting it which is not going to be inexpensive. Automation is great, but it's not like humans. You can't reprogram it with a training course or a set of instructions. You're now constrained by the boundary condition of that automated equipment. Matties: Do you all overhear designers talking about design for automation at all? Torres: If I could put in my two cents from a fab standpoint, I would have the same comments as everyone else: Suppliers very rarely get in - volved early. If I can be so bold, I see it as a comment on busi- ness in general that everybody is just concerned with their own responsibility, their own exper- tise. You've got to start acting like partners, not only with suppliers, but with your internal customers as well. If that designer understood the cost problems of the purchasing de- partment, one phone call to the fab manufacturer before the design is completed could save a lot of time and money. If you de- sign something for manufacturability, it's going to be more cost-effective for their company, and obviously if you go one step further and design for automation, that's even better. THE HDI ROUNDTABLE EXPERTS DISCUSSION Paul Petty