Issue link: https://iconnect007.uberflip.com/i/1213413
86 PCB007 MAGAZINE I FEBRUARY 2020 our challenge is to supply laser drilling sys- tems that support high-yield, cost-effective volume manufacturing. One of the hurdles is that companies often have their own IP around the technology, so it's hard to generalize any solution. Add to that the plethora of compo- nents and component requirements for ther- mal and structural integrity, and things can get tricky quickly. Having a very flexible tool for processing such a potentially large range becomes crucial. ESI's technology delivers the necessary power control, accuracy, and beam steering needed to enable a reliable intercon- nect and ensure the customer's product func- tionality. Johnson: What if an embedded component doesn't work? PCB fabricators are not accus- tomed to being a test house. Ryder: I hesitate to believe that many PCB or substrate suppliers would make themselves the component tester. Embedded resistors would be conceivably easy to test, even within a test circuit. However, if you had 150 embedded ca- pacitors, how would you test for capacitance within the total embedded structure? And if it's an RF sender, a transmitter, Bluetooth, or some sort of processing component, you're not testing a simple digital response; you're test- ing a full functionality. This would be too risky being outside of core competency. In addition, active component testing is often subject to a chip manufacturer's own IP and test methods. Johnson: Maybe we exposed an assumption I didn't know I had. For the manufacturers in- volved in active component embedding, how far does a typical PCB fabricator go? Ryder: Essentially, the end-product should be a final substrate or a PCB, ready for surface mount. Johnson: Who does the fabricated board test? Ryder: When it comes down to a full final func- tional test, this would have to take place post- SMT. Johnson: They're taking a flying leap on that. Ryder: That is the risk, Johnson: Which takes us all the way back to the beginning of this conversation. Ryder: That is part of the decision-making pro- cess on whether to embed an active compo- nent or not. This demonstrates one aspect of what limits the technology from becoming more extensive than it is right now. Johnson: That's why it's not a widely accepted approach that shows up in your typical proto- type process. Ryder: A single function module would be a good way to start. Some of these modules can be included in a SiP or even on the motherboard. To me, that seems more where the technology is right now. For large, high I/O embedded actives for complex processing, we're a little ways out before widespread implementation. Johnson: What do you think needs to happen to create that hockey stick in the forecast that keeps moving out? You've already alluded to infrastructure. Ryder: First and foremost, there needs to be a kill- er application that justifies the investment from an end-product perspective. If it means X perfor- mance improvement or Y dimensional improve- ment that differentiates your product and drives margin, the stakeholders will make it happen. There are players in the market that know how to execute, and they will drive demand down the supply chain and make it possible. It's getting there. I see a lot of movement and development in the market, but most of the movement is sort of early technological adoption in late R&D proj- ect phases. But that's not to say that there is no volume manufacturing. But the scope of applica- tions is still somewhat limited. Johnson: What you're saying seems to match up with what we've been talking about the past couple weeks. Happy Holden said that