Judy Warner (00:01) Hi Steve, thanks so much for joining us today. It's been a pleasure to get to know you and I'm excited to share you with our audience and today we're going to talk about Glasscore technology. So before we get started, why don't you introduce yourself to our audience, tell us a little bit about your technology, background, and your role as Samtech. Steve Hillerich (00:22) Yeah, so I'm Steve Hillerick. I've been at Samtech 26 years. I'm an engineer, ⁓ although I don't do much direct engineering anymore. ⁓ Product manager for Samtech's Glasscore technology, which for Samtech is the tip of the spear technology-wise. And so my role is to help sell and promote, market, ⁓ talk to customers and sort of act as an application engineer as well. ⁓ And, you know, to some degree, help drive, you know, maybe what we should develop next and why and what's interesting and you know what the market tells us. Yeah, you know, no, no, I was just didn't know if you wanted to get into why glass I think was a follow up question. Judy Warner (00:59) Right, exactly. Well that's something. Go ahead, see. Yeah, it is. But before we do that, like, why don't you give us some context of, ⁓ I'm going to redo that. Okay. Hold on. One thing I really appreciate about Samtech is that, that you are actually shaping product based on the feedback you're getting from the real world. So let's go ahead and jump into it. So let's talk about what Glasscore, what you mean by Glasscore. And I know when I first met you, was scratching my head a little bit because I always think of Samtech as, you know, super high speed, high performance cabling and interconnects. And I never really thought about glass cores and interconnects. So why don't you unpack that for us? Steve Hillerich (01:57) Yeah, there's a lot there. You know, the first point I think is that ⁓ the customers are absolutely what drives pretty much everything we do. know, first and foremost, we're a service company. ⁓ We've tried and I believe we have for the most part become a technology leader as well in the last 10, 20 years. And to your point, you know, that's led to the ⁓ generation of a lot more high frequency types of cables and connectors. And ⁓ so, you know, how we get to glass, ⁓ step back about 15 years or so ago and Samtech makes an acquisition of a optical engine company. And at the simplest form, if you know nothing else about how, about how an optical engine works, it's a high speed long haul cable, you know, so it fits very nicely into the catalog. ⁓ but it is an active device. And so that was our first kind of foray into yeah, microelectronics and active devices. And we start, we want to control the assembly. So we start packaging. Judy Warner (02:49) Okay. Steve Hillerich (02:56) these devices. And ⁓ we essentially started working with external customers as well as packaging our own products and really weren't trying to take over the packaging world per se, but listen to our customers understand where the bottlenecks are for say next generation packages and such. And what kept coming up over and over again was that, you know, if Moore's law is seemingly slowing a bit ⁓ that, you know, as you move from say a single sensor on a substrate, which together on another board makes an entire system. And you move to then multi-chip module type systems that as we look forward, ⁓ and we do have silicon-based substrates, but we're hearing that glass might be the next sort of generation of substrate that enables devices to continue down the path of miniaturization and higher frequency, which is ⁓ the constant goal almost all the time of electronic devices. ⁓ So we get into Glass with ⁓ acquisition of a company that has a recipe for a TGV through GlassVIA. And at the time, I think we didn't know what we didn't know. And ⁓ let's try and listen to customers, but take a bit of a gamble. And in the end, it ends with us sort of controlling the entire ⁓ system, the entire creation of a substrate, how you drill a hole matters, how you fill the TGV matters. Judy Warner (04:05) Okay. Steve Hillerich (04:25) how you integrate to the RDLs, the redistribution layers matters. And so to end up with that complete metalized glass core substrate, it took us years to integrate and develop processes, ⁓ but we've basically pulled it off. And in the end, I think we would argue that it is an interconnect. know, ⁓ like, Symptek likes to think at the highest level of the entire signal chain. from the way a signal moves out of ⁓ a silicon die, through a substrate, through a board, through into a connector and into the other, all the way back through to the other silicon device. And so, as much as a connector or a cable is part of that interconnect chain, we think that the glass substrate is as well. It's just as close to the chip as you can get. Judy Warner (05:06) Mm-hmm. Mm-hmm. Yeah, so do you think of it as substrate or do you think of it as an interposer? how do you and how do your customers sort of think about that particular interconnect? Steve Hillerich (05:35) That's a good question. And ⁓ the truth is the difference between an interposer and a substrate is in the details. ⁓ some of them call it one thing, some of them call it the other. But there are differences. We find that most customers are using it more as a substrate, mounting their dye directly on it in a 2.5 or 3D stocking type of application. ⁓ Interposer, it's used somewhat as that. Judy Warner (05:44) Ha ha ha. Steve Hillerich (06:05) Um, really the terminology, the lingo is a function of, you know, kind of who we're talking to. kind of call it whatever you want to us. It is an interconnect, you know? Uh, but the, way that the, the, um, folks that are say designing connectors and cables and PCBs, they use one set of lingo and, uh, the folks closer to the silicon kind of use another set of lingo. so, uh, depending on who I'm talking to, it's, and it's usually someone closer to the silicon. It's a substrate. Judy Warner (06:11) I see. Right. true. Right. Steve Hillerich (06:34) generally, you know. Judy Warner (06:35) Right. Well, it's an interconnect. Maybe that's the way that both of those types of folks understand it's really an interkec. Okay. I, at the risk of going down a really geeky rabbit hole here, what, how do you attach, how do you metalize glass, Steve, especially if you're drilling vias into glass? Like I come from PCB world. So, ⁓ Steve Hillerich (06:39) It is. Judy Warner (07:04) I'm hardwired with thinking about plating those types of substrates that have things that will bite and ways to metalize that. How do you metalize glass? Steve Hillerich (07:19) Yeah, that's the tricky part. honestly, it's not even that the trickiest part is metalizing the glass. It's doing it in a way that's reliable, you know, under thermal cycling, mechanical shock. But in the longer answer is that this part of the answer to this question, a big part of it is in this secret sauce, right? That we won't totally, but I can tell you, you're right. It's not a PCB. We're not using a mechanical drill, you know, and Judy Warner (07:29) Mm. Right. Okay. Steve Hillerich (07:48) It's very much closer to a semiconductor fab tooling set, know, the types of tool lithographies, um, you know, sputter tools, they're, they're, the drill tools are lasers and chemical etchings to make holes and other tools and still plating tools. Um, and then lithography, um, but in a, in a clean room, you know, you wouldn't, uh, not necessarily have put a PCB in a cleaner room, but, uh, what we found is that we can essentially find. Judy Warner (07:53) Okay. I see. Mmm. Right. Steve Hillerich (08:18) a lot of existing semiconductor tools, but the glass being for variety of reasons, one of which is that it's clear, ⁓ things like alignment become a little bit trickier, right? Maybe aligning a mask or something. So we've sort of had to repurpose some machines in some ways to some existing semiconductor equipment ⁓ in some ways and change recipes to drill holes and fill TGVs and do it reliably, you know. Judy Warner (08:21) Okay. Steve Hillerich (08:48) But that is the real trick, you know, and I guess I would also add beyond that, the other thing we've learned is that I think glass is a very fragmented supply chain right now today. I think the cat's kind of out of the bag on glass and it's, you know, how it's going to be used. And I think we're at the beginning of something really big, not just Samtech, but the market. And so, you know, we've learned that all of those elements to create a fully glass core circuit board need to be controlled. Judy Warner (08:59) Mm. Right. Mmm. Steve Hillerich (09:17) upstream and downstream. know, today, if you don't work with us, let's say, and you wanted to end up with a similar product, you can go to the glass suppliers and they will they want to sell wafers or, you know, panels. And so they'll even give them to you with holes in them. And then you can go to somewhere else and they'll fill it with a TGV. And then you go to somewhere else and you add the RDL and you can probably get something that works even though it's expensive and slow. But then when you start to thermally cycle it or mechanically shock it, That's when you really learn the constraints. so for SandTech, it was important over the last eight or nine years to ⁓ focus on integrating that all under one roof, know, control everything from start to finish, you know, that reliability is the trickiest. Judy Warner (09:56) Right. Right. Yeah. And I imagine that was a steep learning curve and obviously it took you a long time. So let's talk about things that our audience will care about. So that's sort of the foundation of here's what it is. Here's what can accomplish. So tell us why Glass and what are the performance or the speed or I know that you work a lot in the high bandwidth area with these, with this Glasscore technology. So What does it do for engineers, you know, as you're out there dealing with your own customers all the time? Steve Hillerich (10:38) Yeah, OK, so the first thing I would say off the bat is that we are still learning like almost daily, it seems like, of new reasons why glass matters, you know, ⁓ which is exciting and challenging at the same time, you know. But the first thing in general is that it's a substrate interposer substrate. It's generally a substrate. so, you know, and back to the nature of miniaturization and higher frequencies, you know, it's ⁓ Judy Warner (10:47) Okay. Steve Hillerich (11:07) very fine features, very small features, the via diameters and trace and space of the RDL. They're much closer to what you could get out of a silicon die than they are PCB. So you just start with size, making smaller devices, shorter signal paths, you know, that are less lossy kind of thing. And then you, then you move on to things like ⁓ that. It's very uniformly thick across the wafer and it's very smooth and it's very flat. And so those are the kind of properties that lend itself to Judy Warner (11:24) Mm-hmm. Steve Hillerich (11:37) reliably 2.5 or 3D stocking, know, copper pillar attached or normal compression or even some kind of anodic bonding or something. So everybody cares about those two things at first. Then of course, there's always a CTE match to the device, know, over, know, coefficient of thermal expansion of the device, you know, of the glass of your choice, because glass is a family of products. It's not just one material type, you know, how does it match your system and under what conditions. Judy Warner (11:41) Mm-hmm. Steve Hillerich (12:05) And then from there, what we've seen so far is that there are usually one or two other elements that really differentiate ⁓ their widget. And it really comes back to, at this point, if you don't need glass, you shouldn't use it. And that sounds like a dumb thing for a glass, gotta say, but the PCB in the ceramic world, it's pretty mature, pretty commercialized. So if you can get a PCB to work, you should use it. It's just cheaper. But for our... Judy Warner (12:21) Right. Right. Steve Hillerich (12:33) our customers that we find wins with, ⁓ it is part of the thing that differentiates their widget. Whether they have a new sensor or a MEMS device or you name it, ⁓ it really won't work like it should without some function of the glass. so beyond those first couple of variables or feature size and such, usually it's things like, ⁓ could be the hermeticity of the vias. We do a helium leak check on our Narvias and they are hermetic. so depending on your device and where it's going, whether it's maybe in the body or in space or whatever else environment, maybe that matters to somebody else. Maybe it doesn't. The big one we see lately ⁓ is the electrical properties. And specifically when you consider a material set like fused silica or synthetic quartz ⁓ has an excellent dielectric constant, loss tangent. And we've got customers that they're pushing well over 300 gigahertz through the substrate. So yeah, and so now you're getting back to getting shorter signal paths that are less lossy and then generally require less power because of all of that together. And so that's a big deal ⁓ lately. For others though, it's also sometimes just as simple as the clarity. Judy Warner (13:37) Wow. Steve Hillerich (13:57) the fact that it is glass after all. And that may offer sort of unique assembly methods ⁓ that you couldn't consider with silicon in a substrate or interposer in a way. But ⁓ the high frequency portion of it is definitely the fastest growing ⁓ section we find. And so just to take it a step further, what we've seen is that customers have learned how to characterize ⁓ our processes and these material sets very well electrically. And the modeling matches the reality just a spot on almost unbelievable. So. Judy Warner (14:37) which is phenomenal, right? Like when do get that? Steve Hillerich (14:42) Yeah, that's right. And so, I mean, we almost thought there were some mistakes in the simulations, you know, because it just matches. so, you know, then it leads into things like reduced cycle iterations. You know, you don't have to get there is the first thing, but separately and as importantly, or maybe more, customers have also learned how to embed passive devices into the glass substrate. You know, so just using the patterning of the TGVs and the RDLs, you know, they're creating whether it's a filter. Judy Warner (14:55) Right. Steve Hillerich (15:12) or a wave guide, you know, and you could consider a crossover or a Wilkinson divider or, you know, a capacitor even. And what they're finding is not only do have a substrate that is small, that is low loss, you know, that enables all these wonderful miniaturization features, but then they can also pattern embedded RF components into it rather than having them sort of discreetly attached to the top of the bottom. And then you just continue even further. Now that your device... has ⁓ less component count, less assembly steps, lower loss because there's less of an interconnection through a solder joint or something else. And that specific area of the high-frequency realm is honestly really where we see the most growth today. Judy Warner (15:58) So tell us a little bit again for our audience to give them sort of context of where this might be a win, right? Or something they need to look at. ⁓ What are some of the applications as you deal with customers out there? Because obviously this is like leading bleeding edge stuff you're talking about. So what kind of applications ⁓ are you seeing across your sort of your customer profiles? Steve Hillerich (16:27) Yeah, well, you know, it's a little bit of everything. You know, sometimes the applications can be a little sensitive, of course, but ⁓ it's ⁓ like I often high frequency, ⁓ anything from sort of telecom to data com to backhaul types of systems, you know, that are pushing not just 6G, but thinking about 7G infrastructures ⁓ through biomedical devices and implantable devices, you know, and where miniaturization is key and Judy Warner (16:34) Right. Mm. Steve Hillerich (16:57) biocompatibility is key and while implantable, least invasive as possible is ideal. ⁓ All the way through handheld devices. You can imagine as things get smaller and smaller, DNA sequencing or some device like that, which maybe you have to have a sample that you take back to a lab today, but tomorrow you could envision having this handheld device in the field. Judy Warner (17:26) Hmm. Steve Hillerich (17:26) where you put the sample on the device and sort of measure the DNA from that all the way through layers of MEMS devices, where hermeticity of the VIAs matter for some kind of mechanical device. And it's military devices, which we definitely can't talk about, but just a variety of next generation. Judy Warner (17:51) right. Steve Hillerich (17:55) sort of devices all over the place. And that's kind of really what we find is, again, back to if you can use a PCB, you should, you know, but we, yeah, it's the, somebody who doesn't want just a little bit faster widget, they want, a whole generation, you know. Judy Warner (18:03) Yeah, absolutely. It's, yeah. Right. Well, or they can't get there without it, right? So obviously you're adding complexity to your design and all of that. yeah. Well, you know, to play the devil's advocate a little bit, you know, as I think about our listeners is who else is using this kind of technology and how do you differentiate say what Samtech's offering was sort of the competitive landscape. Steve Hillerich (18:45) Yeah, that's a good question. It's fair. ⁓ and, and the trick is that back too early, whether we were lucky or smart, you know, we got into the space eight or nine years ago. And so we have a headstart and you know, we're still learning of course, you know, but the competition's coming, you know, so I think it is back to a few key things in that there's really only a few companies in the world that we know of really that have everything under one roof, you know, kind of all of the processes from Judy Warner (18:54) you Steve Hillerich (19:15) start to finish. And again, that's a very big deal. But we also know that there's a lot of money and energy being poured into changing that the competition is coming, you know, and so there's a sweet spot of volume for that matter. You know, you hear about some companies that are pouring hundreds of million dollars into a large panel, glass panel manufacturing, chasing very specific markets, working on wafers, as we work on six inch and eight inch wafers. Judy Warner (19:39) Mm-hmm. Steve Hillerich (19:45) And, you know, so the types of devices we target maybe are a little different than somebody who is targeting a more consumer grade end product, you know, that would require a glass panel, you know. ⁓ I do think it's only a matter of time though before we have more competition that has it all under one roof. And for that matter, I really think the market is still figuring out what it wants. You know, if you check the news ⁓ and just kind of poke around on this stuff, you can get a lot of mixed opinions, you know, about. Judy Warner (19:55) Got it. Right. Steve Hillerich (20:15) So certain companies are really excited about it and then they're not excited about it and different formats of panels and wafers. I guess I would add one thing that's very different and matters for our technology is that the via shape is a straight walled, fully filled and planar via. know, one of the most common shapes you'll see is an hourglass shaped via. And even when attempting to fully Judy Warner (20:38) Mm-hmm. Steve Hillerich (20:43) fill with a plating method, you you can get a pocket in there or you can get a sort of groove in the top and the bottom. And essentially when you start to do post-processing, you might have outgassing or liquids trapped in there. So our via recipe and technology is different. I'm not sure we'd know of anybody else doing it that way, but. Judy Warner (21:08) You're getting dead in, you know, perform simulation versus performance. So to me, that says a lot. Like everybody would love to have that. So if you're there, I would say you're in a really good position. ⁓ Steve, I know that we're just a couple weeks away from IMS, which is going to take International Microwave Symposium, which takes place in San Francisco. In regards to this technology and maybe whatever else you want to talk about, what's going to be your focus and what are going to be the fun stuff that you're showing off as we go into that show? Steve Hillerich (21:53) Yeah, so ⁓ the fun point is that for the first time in a few years, we're kind of back to proactively selling this product. If you looked around our website, it's kind of hard to find. And the truth of the matter is we told our sales force to quit selling it a couple of years ago because we have more work than we can handle. ⁓ And so we've been running 24 seven shifts for years now, but we started an expansion to our clean room ⁓ several months ago that we expect to be ready sort of towards the beginning of the next year. And these projects typically have a little bit longer gestation cycle than a set connector or a cable would. So now is a good time to sort of get back into the market, start to tell this story, proactively sell Glass again. So yes, we'll be at IMS. And so we have a couple of demos, one of which tells a really cool story about the entire, like I described earlier, the signal chain from Judy Warner (22:40) Mm. Steve Hillerich (22:50) ⁓ silicon to silicon. And so there's a demo we have that, you he's going to use a glass based substrate to generate a ⁓ radio frequency. And it essentially is going to travel through a variety of all of our connectors and cables and interconnects. And we're going to measure the signal at a variety of points and demonstrate, you know, how our products, you know, work as well as they do and offer such clean signals through the entire system. You know, ⁓ the second demo that we're really excited to show at IMS is a phased array antenna module in the millimeter wave space. And those are kind of two buzzwords in general, millimeter waves, know, frequency range and phased array, but it's going to demonstrate a 3D stacked assembly with 64 die on it and on one piece of glass and all of the RDL and how it generates the, you know, Judy Warner (23:31) Right. Steve Hillerich (23:49) the frequency with the antenna and the embedded waveguides. And we have some simulations to show and then the reality of the device right there that's being measured real time and all of the eye patterns and all that fun stuff. So that's going to be really cool. And I think that demo by itself almost tells the complete story. Like I said, there's much you can do with Glass, but that tells the whole story. Judy Warner (24:15) I, you know, I'm a geek. So I really love your guys' demos because they're so like, you can sort of see that end to end story and you can see the measurement going on and it's, good eye candy. You guys do a good job with your demos. So I know. Yeah. Steve Hillerich (24:35) We have a great marketing team, you know, it made me think of another really important point to the end to end, you know, as these systems are getting smaller and maybe even specifically faster, it's not enough anymore to just think about, you know, the Silicon, then someone else thinks about the package and then someone else thinks about the system. You kind of have to consider the system holistically and that's not easy to do. Judy Warner (24:54) Mm-mm. Steve Hillerich (25:00) you've really got to consider the whole signal chain through the system and I think that's maybe the biggest value Samtech adds on its entire portfolio. Judy Warner (25:01) Mm-mm. That's how I came up with the, that it was that idea that I was observing that caused me to come up with the idea that ecosystem is hopefully in this podcast, I'm showing people their upstream and downstream stakeholders so they get that perspective. And this is certainly, you know, a more novel, interesting side of that. So I appreciate you coming on and talking about us. think last year, I don't know, I've lost track of time. Steve Hillerich (25:18) Yeah. Judy Warner (25:39) But I know that one of the things that other members of your team were talking about the orange nitrowave cable. ⁓ Is that something that you're continuing to do? do you have any input on sort of where that is for our RF engineering audience? Steve Hillerich (25:57) It is. Sadly, it's not exactly my expertise. I know it is the next generation sort of waveguide cable ⁓ and ⁓ tons of demand for it. ⁓ I'd be lying if I told you I had the deepest knowledge of its sort of performance and offering, but it will be also demonstrated at IMS. we will have ⁓ several engineers there who can tell you more about it. Judy Warner (26:08) Okay. Okay, I'll put the link in the because I knew that was a big part of other members of your RF team. I know they were really excited about that. So I'll just link it for our audience in the show notes. So Steve, where do you go from this from Glasscore Technology? ⁓ Gosh, it just kind of blows my mind how fast we're going and how quickly technology is moving. So Steve Hillerich (26:41) It's... Judy Warner (26:48) Do you have an idea of where you're going? And then why don't you tell folks where they might be able to find you either this glass core technology, either online or at iMess. Steve Hillerich (26:59) Yeah. So ⁓ you're right. It's amazing how we got here. And when I started at SEMTECH 26 years ago, I would have never guessed that we'd even be having this conversation. But ⁓ really what blows me away as much as anything is it's not even just the glass, but it's what it enables. It's these end applications that are, mean, Judy Warner (27:10) Right? Right. Steve Hillerich (27:22) It's a, know, we may not have flying cars like they promised us in the 1980s, but some of the things we have, I would have never dreamed of. And every day I almost talked to somebody who's got an idea or they have a widget in process and you just, sometimes scary, you know, we jokingly, we jokingly say, baby, we're enabling the end of the world. I don't know. know, it's not that or something. Judy Warner (27:44) Yeah, I know. say that too. well, we're good or for bad, we're running headlong into this complete unknown. And I know, you know, I've been around 30 years plus and it feels like we're living in a science fiction movie. Like I'm with you. If you stick around long enough, it feels like I would have never predicted this 10, 20, 30 years ago. It's fun. Steve Hillerich (28:02) Yeah. Yep. It's true. Judy Warner (28:15) But like you said, hopefully we're not building AI to kill us or whatever it is we're doing, but it is fun and it is fun to see what engineers can do with these new tools. So, well, thanks for the conversation. ⁓ Is there anything else you want to talk about before I let you go? There's a lot here, but I want to make sure we send our listeners to go dig that out. Steve Hillerich (28:44) Yeah, so we'll be at IMS booth 1453 this year out in San Francisco, like you mentioned, you know, the webpage to find more about these products is samtech.com slash GCT, Glasscore Technology. Or just contact me and we can, we try to be easy to work with and straightforward, you know, so if you reach out, I'll answer the call. Judy Warner (28:49) Okay. Okay. Okay. sounds good. One last thing that I had on my notes that I forgot to ask you about is the process design kit. You want to talk about that for a second before I let you go? Steve Hillerich (29:20) ⁓ Yeah, good call. So this is something we're still early on. We're very excited about. And it's likely a year-ish before it's really ready for live action. But the point is, all of these Glasscore circuit boards are some kind of custom ⁓ routing, right? ⁓ Design. And so that's tricky to put in a catalog. Or for that matter, we're dealing with different material sets than customers are used to. We're also looking. Judy Warner (29:27) Okay. Steve Hillerich (29:51) Fusilica, synthetic board, Sapphire, right? And so ⁓ while I can give them data sheets on the material properties and maybe some S-parameter or electrical data from things that we've generated ourselves, how do you continue to push the envelope? And for that matter, how do you continue to ⁓ accelerate the design cycle? And so we're working on a PDK, Process Design Kit, and know. Judy Warner (29:51) Right. Steve Hillerich (30:16) where it lands fully, it's hard to say, but the goal is that if you're a designer and you want to utilize glass, that it's novel material with these novel methods, and rather than starting from scratch, you very easily will be able to pull in elements, filter, wave guide, whatever, into your design. And then sure, maybe you got to work on some routing to some level, but we hope to even have little blocks of sort of virtual blocks of, ⁓ routing chunks that you could sort of design a system, a substrate for your system, much easier than you could starting from scratch with some basic design rules. Yeah, and then know it's going to work because the simulations do match reality so well. Judy Warner (30:50) I see. That's cool. Very cool. Well, thanks for geeking out with me today. It's been really fun to learn from you and learn about this technology. ⁓ Again, I will send our audience to the website that you told us about and then IMS booth 1453. Steve, great to talk with you. For our audience, thanks for joining us today. I trust you enjoyed this conversation with myself and Steve Hilarik. Make sure you go check out the show notes. We'll see you next week. Until then, remember to always stay connected to the ecosystem.