Issue link: https://iconnect007.uberflip.com/i/1330321
JANUARY 2021 I PCB007 MAGAZINE 35 I was very pleased to see our overall perfor- mance. Our team provided our customer base with excellent technical support and ultimate- ly shipped the same number of machines this year as last despite periods of full quar- antine. During the lull of the quarantine, the team was provided with the opportunity to work more creatively in ways that otherwise might not have occurred previously. You do not get opportunities like that as a compa- ny to get everybody on board and have all arrows pointed at a project. We have rolled out our new technology at a few key custom- ers deploying this new technology and we are preparing for a big unveil sometime in March or April. Johnson: So how does this new offering direct- ly deal with our "X=X c – 1" model? Hogan: Total process control means that, for the imaging steps at least, the bounds of cir- cuit production will be better controlled, un- derstood, and inspected to an increased cer- tainty. It targets transitioning "science project" panels to production panels through improved data collection and integration. It also captures and recognizes process missteps earlier in the process such that material scrap rates should be reduced. Johnson: This is more important when you start dealing with the weird, non-linear mate- rials that don't scale well. Hogan: Yes. The idea is that we are at a point in digital imaging where we have the ability to alter a process on the fly. You cannot do things like that in develop or in etch because the pro- cesses are not responsive. You can do it in drill, in imaging and, theoretically, you can do it in CAM. So, the idea of data collection to create a process control that is meaningful is, I think, the key to reducing the failure modes for first pass panels. Johnson: And these concepts came out of the Heterogeneous Integration Roadmap at the UCLA CHiPS program? Hogan: Yes, the fundamental issue being solved with the UCLA CHiPS organization as well as a similar program at the University of Stuttgart are the problems the PCB fabrication market will experience in the next decade. At CHiPS we produce 2-micron features to directly re- ceive the pads of the die itself—without pack- aging. The registrations issues are daunting. The method used to solve that unique problem set us on a line of discussion that will now be applied to the PCB fabrication segment. Johnson: With the technology and the ap- proach you are describing, doesn't the inter- poser become the new breadboard? Hogan: That's right. And downstream, the in- terposer just disappears because the interpos- er becomes a silicon-based substrate like the PCB but at much higher resolutions: X=X c – 1. If you think about a classic motherboard on a computer, you get your memory devices, your copper line and then you've got a controller. Well, there's lag that happens. The controller sends—it's microseconds—but it sends a sig- nal across copper where the transition of ma- terials causes a latency so that now the travel distance might only be an inch, but the travel distance and materials induce latency. Now, if you've got it all on silicon, it's all similar ma- terials and there's no latency. The memory be- comes faster, you can use cheaper memory for certain applications and all of these other flex- ible tools come into play. Higher reliability, in- creased performance—it is the direction we all must go. Johnson: When you started your career did you ever think we would be at this point in manufacturing? Higher reliability, increased performance—it is the direction we all must go.