I N T E R V I E W : Joel Scutchfield, Koh Young

Koh Young has been at the forefront of SPI and inspec-tion technologies for over two decades. They invest heavily in research and development and are busy creating new solutions using AI to handle the increasing automation, miniaturization and sensitive components and substrates. Trevor Galbraith interviewed Joel Scutchfield at SMTAi 2023. So my guest here today on Tech Talk is Joel Scutchfield. Great to see you, Joel. Great to be back. I want to use this Tech Talk to discuss the technology and the general direction of inspection and the Koh Young Roadmap. Inspection is essentially the eyes and ears on the factory floor. How do you see the technology changing and how do you see the roadmap and developing for Koh Young to adapt to that? It’s a great question. I like to refer toourselves as smart factory enablers. That’s really where we’re going. We’re providing solutions that enable the realization of smart factory, which is what we’re all trying to achieve. We’re using our data set, which we know is a little bit different than others, it’s very measurement-based and it uses the copious amounts data that we extract to basically fire our proprietary AI engines and make our systems much less dependent on human intervention. And that dependency is not just a case of reducing the number of human interactions,
it’s also a case of improving the quality? Absolutely. There’s so many tentacles to this thing, Trevor. We’re looking at this from a lot of different angles. The enablement of smart factory is one piece, but as we all know, the forecast for the labor market globally, is playing a part in this as well. So we need a higher level of automation across the board for that smart factory realization. To do that, again, we have to utilize all of the tools that are available to us to make these systems able to judge, adjust and accurately provide that pathway to reduce the dependency on people. These reshoring activities, semiconductor activities, EV battery activities are certainly going to take a lot of people, or a lot of automation. One of the things that we’ve been working on for the last year or so is a complete overhaul of our software structure. And
why are we doing that? Well, again, we have to be flexible, we have to be fast, we have to be able to handle a lot of data. And ideally from a human intervention perspective, boy wouldn’t it be nice if when you walk up to an SPI machine, you walk up to an AOI machine, you walk up to the DPI, dispense process inspection system and the software interface is the same on all machines. They have to be the same. They have to be intuitive and use a common interface. Exactly. And it all ties back to, how do we reduce the need for human intervention. So the AI piece is certainly a big part of that also. And as we’ve talked
many times, we’ve made a very significant investment in that arena with our team in San Diego that’s focused strictly on that, creating engines that use our dataset specifically to reduce human intervention at the programming level for AOI. And certainly in the case of print process control, use the KPO tools to look for trends, to look for anomalies and be able to automatically adjust the squeegee speed, pressure, stencil snap off, et cetera. And we’re doing that. And that obviously, as you know, has taken some time. It’s alot of data collection, it’s a lot of analysis, it’s a lot of algorithm creation and adjustment, retesting, regathering of data, re-analyzing, re-spinning of algorithms, and this is how it works. This is the generative path forward when trying to apply AI in this space. You’ve got the research facilities you mentioned down in San Diego, so are they working on AI paths to try and make all of this work faster, more seamlessly and with less human interface really? That’s exactly right. It comes back to these tentacles that I was referring to earlier. Everything is somewhat intertwined to a degree, but at the end of the day, it’s all being done with the vision of knowing that to be successful manufacturing in the Americas, and certainly in Europe and other high labor cost areas, we have to have this automation. The automation is different than automation we’ve utilized in the past. Going forward, it’s really AI driven automation now, and being in a position to take advantage of that we feel it gives us a significant differentiating advantage. So I mean, you’ve got lots of automation solutions at the moment, certainly on the line, but I think one of the most labor-intensive areas within the factory is the box build. So what are you planning to do about that? Is that on the roadmap? Yeah, it absolutely is. And again, that’s back to the people conundrum. The frontend of the SMT lines today is very much not labor-intensive. We’ve gotten things to the point where a single operator can oversee multiple lines from the screen printer up through the reflow oven, but on the backend, that’s not the case. That’s where we still have a lot of people. You’ll be hearing more soon about a tool that basically takes our probe construction and puts that in a format that’s adaptable to alternate access points, let’s say on the backend, whether that be mounting it over a conveyor, potentially on a robot arm. Various applications that go on at the backend that need that flexibility and adaptability. So yeah, that’s very much a big part of it. Another trend that’s happening in the factories, Joel, is increasingly we’re seeing contract manufacturers putting in clean rooms. There is a shortage of OSATs for final assembly here. And so some contract manufacturers are starting to bring some of that backend packaging activity into the EMS factory. How are you adapting to these requirements? So we see that as a huge initiative as well. A little bit of a different conversation as it applies to the reshoring of semicon-ductor manufacturing in general, which as you alluded to, brings a completely different set of challenges. People, infrastructure. We just don’t have it here. We really don’t have it in Europe right now. And so our direction here is to again, adapt our technology for the needs of that space as it grows. And it’s going to take a little bit more time than people originally anticipated. Do you have the technical capability to be able to inspect all these very small devices, shiny devices in such a small sizes? On some of these devices, you are looking at a million connections. You’re absolutely right. And so again, back to the respin of the software to be able to handle that increased volume of let’s say a million plus pads on a surface. But then again, you get into that clean room, what are we doing? We’re positioned now, Trevor, to basically handle everything post fab. So I have a completed wafer, thography’s finished and that OSAT’s now going to take that and run with it. They’re going to dice it and add their elements. They’re all going down the heterogeneous path and will move from 2D to two and a half to 3D and Stack and do all these things with interposers and silicon glass. With the Meister Series, we feel we have an advantage because we’ve been doing that and doing that very well for a long period of time on the SMT side. So when we get into those heterogeneous, hybrid type applications, now we have with the Meister D+, the capability of handling the shiny reflective devices for micro cracks and FODD and other anomalies in addition to all the other inspections needs. You mentioned that dye are starting to go in the Z direction with the stacked dyes which of course are very difficult for you to do with optical inspection. Have you thought about moving into other areas of inspection or test to give an end-to-end solution? Yes. So again, something you’ll be hearing more about soon is adapting that Meister system to handle all the various forms of packaging that exist. So once that dye comes out of the wafer and now it’s placed in a tray and other things are happening, the needs change for inspection. The handling needs to change. Substrates are getting more fragile? For sure. Thinner. I was at a conference a few weeks ago, a lot of discussion, silicon or glass? Which way are we going to go? And there’s advantages to both, but again, handling is different, potentially, for both. So we have to be ready for all of that. Well, you’ve certainly got a whole lot of challenges ahead of you Joel, but it appears that Koh Young is definitely in the right place at the right time? We like to think so, but we always have our ear to the ground as to what’s coming next. That’s really what it’s all about. And fortunately, we have a very broad spectrum of ustomers that are very involved in a lot of these initiatives. Well, thank you for bringing us up to date with the roadmap and we look forward to seeing you execute it over the next few years. That’s the challenge. I look forward to continuing to provide updates as these steps are taken. That’s the exciting part of what we do. Thank you for the opportunity to talk.
–TREVOR GALBRAITH