#399 – An Interview with Steve Kreuzer

Chris Gammell: This is the Amp Hour Podcast. Released July 15th, 2018. Episode 399. An interview with Steve Kruiser. Welcome to the Amp Hour. I'm Chris Gammell of Contextual Electronics. And I'm Steve Kruiser of Exponent Consulting and Engineering. Welcome, Steve Kruiser, former roommate and high school buddy and college buddy and person who entered the same industry I did. How you doing, Chris? You're good, man. It's been too long. It's been too long.

Dave Jones: It has. It has. No matter how long it's been, it's been too long.

Chris Gammell: Exactly. Exactly. So, I've known Steve since the first day he showed up in third grade at Smallwood Elementary School. We hung out and he lived up the street from me. And ever since then, we've been like in parallel paths towards the same industry, at least. Yeah. That's been pretty crazy.

Dave Jones: It has. I don't think there's anybody else who has lived in the same quantity of places that I've lived in at the same time.

Chris Gammell: Right. Exactly. Right. But you ended up on the other side of the coin a little bit. So, you are a PhD in mechanical engineering, definitely further than I decided to go and also mechanical. So, kudos on both those things.

Dave Jones: Oh, thank you.

Chris Gammell: Yeah. I mean, how did we get here? What is your background story? Since I'm only just telling you the part I was part of.

Dave Jones: Absolutely. Yeah. So, after leaving Buffalo, you know, we both went to Case Western and there got into the mechanical engineering line of study. A lot of that came from just my background at home, working a lot of sort of construction and remodeling stuff with my dad and family members and just kind of doing odd sort of mechanical type stuff around the house and elsewhere. But after studying a case in mechanical, got involved, actually was looking for research opportunities just to try out what grad school would be like, got hooked up with a lab down in North Carolina at Duke doing biomechanics type stuff. And that really, you know, boggled my mind, the idea that biology could be controlled by mechanics, which was, you know, something that was abstract in my classes at the time, but really took on a really concrete meaning in an interesting way when I started looking at cartilage and some muscular skeletal mechanics type stuff. So, that got me interested in the biomechanics realm. And then I decided I wanted to go to grad school based on those experiences. Also did a co-op working on dishwashers, which in contrast to cartilage and whatnot was. Yeah, yeah. So, it was down in Kentucky, which was a lot of fun. Went to the Derby, all that sort of good stuff, but didn't end up really stoking my interest all that much. That's surprising. Who doesn't like a good dishwasher?

Chris Gammell: Right. And actually, that's a good thing, though. GE has since sold that division, right? Indeed. They're now a Chinese industry?

Dave Jones: They are. And that was after they had a huge fire that burned down one of six warehouses, I think, in that location. So, yeah, it was kind of interesting. You know, it was a nice introduction, I guess, to sort of corporate America. But what it taught me was, even though I was in their sort of research and development wing, where they were doing sort of, this is an embarrassment to the term to call it this, but their sort of skunkworks type approach to dishwashers and refrigerators and stuff. I like that. I like that.

Chris Gammell: I like the idea of, like, being in the skunkworks dishwasher division when things go wrong, you know? Yeah, exactly. Imagine, like, that scene of, you know, Willy Wonka, where, like, the devils are just shooting out everywhere. That's pretty much what it was. Yep. Okay, great.

Dave Jones: That satisfies me. Minus the candy, I guess. But so, yeah, so that was kind of my introduction to corporate America. And I decided I wanted to get away from that. Weird. Yeah, right? But ended up going down to, well, looking into grad schools, wanted to go to a bigger school. So, as you know, obviously, and some listeners may know, Case Western, while a good school certainly is in Cleveland and is in sort of the Northeast, which is where the Great Lakes is probably a better way to put it, which is obviously where we're from. Yep. So, I wanted to get away from that, at least just for life experience purposes. So, explored the Southeast and just sort of looking at colleges around there and got hooked up down in Texas at UT Austin, which, you know. Hook them. Yeah, hook them indeed. And you can't really call that the Southeast. Anybody who knows anything about Texas would barbecue you pretty quickly if you tried to do that. Right, right. So, got down to Texas, you know, spent a little bit longer than I wanted to down there, but had a good time. Austin, as you know, is a great place to spend a little more time than you were anticipating. Right, right. So, there are worse places to be. There's things to keep you occupied outside of school, right?

Chris Gammell: Indeed, indeed. And you were still doing bio stuff. I remember something about like the force of acceleration on cells or something like that.

Dave Jones: Yeah, yeah. So, that was one aspect of it. And actually, yeah, to tie it back to the experiences I had in undergrad and mentioning the sort of cartilage mechanics. So, that's kind of at the cellular level where you're looking at the impact or the role of mechanical forces in driving cartilage maturation. So, that's really, you know, that's an interesting level on its own. But what we ended up doing or what I did for my dissertation work was related to molecular dynamics where we were performing, you know, supercomputing simulations on the atomic level for individual proteins. And looking really at the role of mechanical load in unfolding those proteins, which is kind of an inverse problem to what a lot of people talk about with respect to protein structures and things where really, you know, the holy grail is to be able to take a DNA sequence and turn it into a three-dimensional protein structure. And while that's getting closer, we were going the other direction and we were really looking at for disease states, how you might be able to or how the body and how biology might have designed certain disease states to kick off based on load application where you might get a partial unfolding of a protein that exposes a binding site for some other molecule that then leads to a cascade of chemical effects that then, you know, causes disease or something like that. So, it was all very interesting. And, you know, obviously, we were looking at some really sophisticated modeling approaches and things. But as that description probably invokes, it was a little bit pie in the sky and a little bit far away from actual applications.

Chris Gammell: Right. I would think that, like, by the time that a protein starts unfolding, it's in pretty bad shape just otherwise.

Dave Jones: Yep. Yep. There's definitely something going wrong. You know, in some cases, your proteins are going to unfold during normal use. But really, you know, if you think about what application that might have, you know, that could potentially have drug delivery or drug design applications down the road. But it was so far down the road that, you know, it was tough for me to really see where that would actually be used and kind of got, you know, I found myself farther and farther away from, you know, actual hands-on application type engineering. Right. Right. Certainly. Right. So.

Chris Gammell: Which, and I'm curious about that, too, because, you know, I was actually thinking through, like, we haven't had too many people get to the PhD level on this show. You know, we've had Dr. Howard Johnson on the show. We've had Dr. Greg Sharvata on the show. Dmitry, Dr. Dmitry Netaspasov on the show. So, we've had some, but, like, oh, actually, I guess, Bunny Huang as well. Dr. Bunny Huang. But, like, most of the time, we are very focused here on, you know, the more hands-on, the more applications side of things. So, I am interested in that. Before I do that, I want to make sure that we still have people held here because we're getting there. And the crazy thing about all this stuff and the thing that I was actually surprised about when me and Cruz, I call Steve Cruz sometimes, when we reconnected was that Steve is actually in the same industry. He's in electronics. So, we are getting there, but let's keep going down this path. Yeah, exactly. Just keep people hooked in. Hopefully, they're still there.

Dave Jones: Hopefully, you are still there because I do think it gets pretty interesting. And really, you know, the off-ramp from grad school into industry came from, you know, when you're in a PhD program, you're often exposed to funding sources and the difficulty getting funding. Yeah. And that's really where things went off the rail for me because, you know, I didn't want to be in a situation where, you know, obviously, any occupation you get into is, you know, subject to capitalism and so forth. But what I didn't like about grad school type stuff was, you know, now you're subject to funding levels dictated by Congress and that can change, you know, by term and all that sort of stuff. But, you know, certainly not to get into politics, but, you know, I didn't want to be, you know, sort of beholden to those sorts of funding mechanisms. And really, I wanted to focus more and try to get back to more application level work. And so, yeah, and it's interesting.

Chris Gammell: What is the split you'd say? I mean, like, so, I mean, you've probably, you know, you had cohorts when you were in a PhD program. Did you see people also do the same thing of that off-ramp that you mentioned? I mean, moving into industry or moving into getting paid, hopefully? You know, that's what I always wonder too. It's like, yeah, you're in school a long time. How about money? You guys like money?

Dave Jones: Well, right. And that's another, you know, obviously really important aspect is, you know, when you start looking at some of the salaries. And obviously, if you're in a really good position at a really well-known school and you're bringing in a lot of research money and so forth, you're probably doing pretty well. But, you know, as you see in news stories every once in a while and probably more frequently than we'd like to see, it's very difficult to get tenured positions. It's getting more difficult. So, you know, of the cohort of folks that I was in school with at UT in mechanical engineering, you know, I'd say a good number of them, and I'm not sure on percentages, but it wouldn't shock me if it was 50% or so, you know, went into something that was more application-driven and more industry-driven. Where, you know, they probably very much like me probably got in with sort of high ideals in terms of, you know, I want to be a professor. I want to live that life that, you know, maybe you see on TV of just sort of… I want the elbow patches. Exactly. The elbow patches. Totally. You know, and I, to this day, I don't have elbow patches, but… Well, you know. Yeah, there's always time, I guess, but maybe an adjunct position or something. Right, right, right. But…

Chris Gammell: You get extra patchy for adjunct because they pay even less for that. That's right. Your patches are tiny. That's… No, they're everywhere. Oh, they're all. Okay. You need more patches because your clothing is falling apart. You got to make up.

Dave Jones: Right, right. Yeah. Yeah. So, I would say a significant number ended up going into industry when they really started to look into the economics of what it looks like to slog through a postdoc, maybe multiple postdocs. You know, get an associate position or an adjunct position or something and really have to live your life off of that. So, you know, in terms of my motivations and how I got into electronics and really kind of working in that field, you know, a lot of it was enjoying the research that I did but recognizing the lack of applicability in the short or medium term of that research to anything that, you know, would matter for things that I see on a regular basis. And then also, you know, just looking at the pure, the idea of, you know, what would be comfortable, what would be something that, you know, I wouldn't have to be scraping by, you know, in that role. And that's not to take anything away from the people who choose to do that. Obviously, that's a great role for a lot of people, but it just wasn't something that I was looking to do. So, when I was looking to…

Chris Gammell: So, you were just kind of like looking more towards industry, you're saying in general. Yeah. So, you were just… You were like, okay, well, where are the jobs? Go towards that and see what kind of pops out.

Dave Jones: Yeah. And, you know, because I did a lot of simulation work, there is a good bit of… And really, I'd say that the primary industry where molecular dynamics and sort of atomic simulations show up, at least that I saw at the time, was in drug discovery and sort of the pharma approach and realm. And so, I initially contacted a bunch of companies out in, you know, this Boston area as well as other places where you have hotbeds of biotech and pharma and so forth. And so, you know, my research was even, you know, my research was even, you know, a little bit too far afield for those folks. But interestingly, the company that I worked for and ended up getting a job for, Exponent, you know, that we are the type of place that sort of collects people with unique backgrounds and people who work in interdisciplinary research fields. And so, you know, the stuff I was doing, the protein unfolding, the protein simulations, that's not anything that we're working on. And, you know, obviously, as being on this program, we're talking about electronics. But, you know, it's the sort of place where if you have good fundamentals and you understand your field and you understand and have the ability to sort of contextualize fundamental mechanics in different areas, you know, you can make a good progress and start your career in that area. So, I ended up getting hooked up with some folks at Exponent, talked to a bunch of different people, first in, you know, sort of a polymers realm. And so, proteins are basically just polymers, but ultimately got directed into a mechanical engineering position out in our Menlo Park office, which is our headquarters, where I ended up doing or starting in on doing a lot of computational work, where, you know, on that scale, instead of looking at atomic motions, you know, now you're starting to look at continuum mechanics, where you're into finite element analyses, things like computational fluid dynamics and things of that nature on a much larger scale. And this is what led me to realize that I had made a move that I had made a move that I was happy with and satisfied my goals in terms of leaving academia was, you know, one of the very first projects I worked on was something related to a consumer electronics device that I'm not really at liberty to discuss exactly what it was, but it was something that, you know, I saw on a regular basis. And so, we were working on this project for this company. And so, like, the very first day I walk in, I'm like, bam, there's something that I'm working on. It's a problem for people. It's a problem for this company. They're trying to solve this problem. And I'm helping in a not insignificant way to solve that problem for them to help people, help consumers, help the company, you know, all these sorts of, you know, really hands-on, really applicable outcomes of my work.

Chris Gammell: That's great. Yeah, that sounds like, it sounds like where, whereas most, I mean, PhD research to me, it seems like the way I've heard it described to me is you keep finding, you know, the shallower or the, sorry, the narrower and narrower field. Then you become an expert in that. So, you're just diving, diving, diving down into this, this hole, not a hole, but like a, like a tiny drill into the, you know, into the crust of the research earth, right? That is what you're trying to do. You're trying to get as deep as possible. Yeah. And it sounds like the thing that you've been able to do in this, this kind of like this consulting level in the PhD with, with the PhD, but kind of at that higher level is now like kind of backing out and being like, oh, well, how does all this stuff tie together and kind of zooming out from the cloud level? Exactly. And then, but then also being able to, to, to drill down in these small areas and figure out where you need to drill almost.

Dave Jones: Yeah. Yeah. And so, you know, one of the models for thinking about what a PhD really is, is, is that it's, it's a demonstration that you're able to, you know, suffer a little bit certainly, but also you're able to, to sort of see a project through. You're able to, you know, ideally you do advance the field, you know, the, in theory, you don't get the degree until you've done something that's novel. But, you know, really the, the big focus is on understanding the fundamentals of how your research relates back to some of the, you know, more well-known aspects of whatever your field is. And where we really try to capitalize on the experience of our consultancies in, you know, encouraging them to take their fundamental knowledge, the sort of stuff that you would get in your graduate level classes, even your undergraduate level classes in your, your boards or your sort of oral exams that you take prior to qualification for the, you know, getting your, your degree. And, and, and really use that level knowledge in support of solving problems for people and companies and things.

Chris Gammell: That's good. That's good. So let's get a, let's get into the exponent level then. So, so you'd mentioned consumer electronics, obviously that's the, the side of things we're going to kind of play in. We're going to obviously talk about electronics, how this stuff relates to electronics and the mechanical side of things, but what, what, who would actually end up calling exponent, right? So like, what is the size, what is the size of company, the type of company, the, the type of product that would end up calling? We, we, you know, like you already said, you know, you're not gonna be able to talk about details, but we'll talk about, you know, generic company types and company product types and, and situations.

Dave Jones: Totally. And, and, and really the answer to this question, I think is one of the things that, that most satisfies me about the position because there isn't really an answer. It's, it's kind of all companies, which is really exciting. Yeah. Cause we, you know, as, as I mentioned, we were working that first project that I alluded to was for a company that makes a laptop. So, you know, if they're making a laptop or a, you know, that scale of a device, they're probably going to be a pretty large company. Right.

Chris Gammell: Well, yeah, that's kind of what I was getting at too, is like, what is the, the scale of company almost like, so, so we see, is, are there small businesses too, though? Like, I mean, like, so, and, and mostly wondering, because if people are listening, you know, are they going to be able to like, so say they run into a problem. Are they going to be like, they call you guys up and you're like, well, you know, it's a quarter million dollar retainer. Yeah. I don't, I don't know anything about the money, but you know, like just accessibility too.

Dave Jones: Totally. Totally. Yeah. So we, we do run the gamut. I mean, that's, those companies are the larger end of things, obviously, but we work with a lot of startups where, you know, and, and I think actually our model can plug in really well with startups in a lot of ways. Because we have, you know, in some sense, we have a surge capacity to overcome problems. So you can imagine a startup, you've got the founders, you've got the sort of initial employees, you've got sort of the, the, the core of your team has, you know, certainly, or presumably really strong capability in whatever their area is. But they often, you know, need to branch out beyond that. There's aspects of a product design that they might not be super familiar with, but they need help with it in order to really accomplish what their primary objectives are. So we can sort of plug in, in those circumstances where, you know, maybe you're a bunch of electronics folks, but you need a mechanical engineer to help you with your design aspects. Or maybe you need somebody who has materials expertise to understand something about whether you should make your enclosure for your device out of V0 plastic or some other type of plastic or, you know, all those sorts of things. So we do run the gamut of these really large companies where we're plugging in a very specific levels on a project or a program level within those companies, all the way down to startups where, you know, they've, they've got a problem that they just don't have the internal capabilities or expertise to handle.

Chris Gammell: Okay. So like, what is the, I guess, how well, how well defined of a problem though, too, because, because I, because you say that, you know, we plug in a lot of these places and I do believe that. But, but, but I also wonder about like, is it like, oh, I need a board made, you know, can I call exponent or is it more like I need, I I'm already making, I'm making 5,000 boards a year, you know, and I have a specific heat problem in this one spot and I can't figure it out.

Dave Jones: And I need to deal with that. It's, it's definitely the latter. So one of the things that we, you know, make sure to, to be clear about with companies is that, you know, there's a, there's a positive aspect to this, but we don't do any design work. So we don't stamp any drawings or anything like that. But, you know, what that ultimately does is it, it gives us a plugin with startups and companies that are concerned about their intellectual property because we don't retain any of that. It's, you know, it's all yours. We work out half of you, but, but really it's more on the level of sort of the classic place where we interface is that some company has an issue. They've got a problem. They don't have the internal capabilities or the time, which is a capability in some sense. They don't have the capabilities to deal with that problem. And so, you know, maybe it's a thermal issue with respect to their board. What are the consequences of that? You know, we've seen a lot of different types of projects so we can come at it from, you know, ideally having that level of expertise to answer the question and handle the question, but also having that experience level of working in this field, seeing a lot of different problems and ideally being able to diagnose the problem hopefully quickly to get you your answer. Yeah.

Chris Gammell: Yeah. Yeah. That makes sense. Yeah. Cause I would imagine like if I make a board and so let's just go through a quick design exercise, I suppose. So I make a board, it's got a microcontroller on it. It's got a battery, whatever. And the product starts on fire at some point. Right. And it happens after, you know, it's being shipped to the customer and the customer opens it. And the first time they turn it on, it turns, it starts on fire. Yeah. Like that's something that is a very specific problem that I might be like, I don't know, I could probably go through and try and replicate things. But at a certain point I might be like, well, I'm guessing there's some kind of failure here. Then that's kind of the call situation or what? Yeah.

Dave Jones: Yeah. It's, you know, we get called when there are problems, when things are on fire, kind of the, the oh snap type moments where a company doesn't necessarily know what's happening. They need the help and they're under some sort of time pressure. So, you know, another sort of classic example would be if a company is facing a recall scenario where, you know, they've developed something, they've got an issue, they've got a product out there and they need to decide whether the problem is widespread enough to justify a recall. The CPSC, so the government's kind of breathing down their neck maybe and trying to figure out or demanding answers as to why this is happening.

Chris Gammell: You know, and maybe they're telling people not to take stuff on planes. They're like saying you're going to get, yeah. Yeah.

Dave Jones: And that's actually an example of a project that we can talk about because we did work. Oh no, really? Yeah. Well, so we were one of the, we do quite a bit of battery work, particularly in this office here. And one of the things that we've done that we are able to have publicly disclosed is, is our work on behalf of Samsung in dealing with their battery issues there. Oh my. Yeah. And so that was a scenario where, you know, they, they were looking to make sure that their consumers or their customers were safe basically. Right. You know, they had these issues and, you know, they came to us because we have, you know, in that field and that sort of battery realm, we've got a lot of experts who have really well-known expertise, I guess. And so we got involved in that. It was a short, a quick, pardon the phrase, but a quick burn type scenario where we needed to get answers on behalf of, of Samsung in order to guide them and provide sort of advice in terms of how to approach that. And really, you know, where can we, can we identify what the problem is? And that was the initial goal of the project was just, what's the problem? And then once we figure out what the problem is, then, you know, in concert with all the stakeholders, decisions can be made about how to mitigate that problem. Sure.

Chris Gammell: Right. And, and I'm guessing the consulting fees compared to that was probably to the actual recall was a drop in the bucket. Oh yeah. But I mean, like, I mean, given what they did, I, I mean, like, you know, obviously they caught a lot of flack for it, but like I, they, they did probably the best they could from what the, you know, what ultimately was a design flaw. Right. Like, but they, nobody had eyed. So that's good.

Dave Jones: You know what I mean? Yeah. That is definitely an important metric. But, you know, they, they did see quite a bit of, of company value erased as a result of that, you know, at least initially. And, and, you know, nobody likes to see that happen. And, but they, they survived it, obviously, you know, they've got a lot of good products and, you know, they're. And a lot of money.

Chris Gammell: That helps too. That does help too. Yeah. Yeah. For a circumstance like that. What was the problem? I, I'm trying to remember, was it like a, it was a mechanical spacing issue or something with like the closure?

Dave Jones: So, yeah. So, um, there actually ended up being a pair of, um, you know, yeah. And actually, I guess I, I should, I'll have to, uh, defer on that question because I'm not sure while we did work on it. I'm not sure. I'd have to review a little bit. You're saying you didn't work on it, but that's fine. I did actually work. I personally worked on, which was kind of, which was a cool experience, but, um, what exactly we publicly disclosed in terms of the causes, um, I should probably.

Chris Gammell: So here, I can say, stay what I remember. I remember there would being like a gap between, uh, but like there was a machined out gap or something like that. And I thought it was like when there was certain amounts of pressure based on the, the, the battery pack, it just, there wasn't just quite enough room. That's what I remember. Yeah.

Dave Jones: It was, it was a spacing issue and it related, uh, to the packaging, uh, of the batteries, um, and how they were constructed. Yeah.

Chris Gammell: Okay. Well, so then let's, let's talk a little bit about like stepping back a step from that. Obviously that was an unfortunate situation that did, did not work out well from the, you know, the bank account of Samsung. But is that something that should have been checked before? Like, I don't even know how you would have caught that. You know what I mean? Like other than like, just, just on a scale side of things, like would a, would a consumer company like that, or just a generally a company like that, are there certain audits that they should be doing or simulations like you were talking about that they should be doing that could have caught that?

Dave Jones: Well, sure. Yeah, definitely. And, you know, what we see a lot is, and what we, we are able to help our clients with often as well as just sort of reliability, um, audits and things of that nature where, you know, you're, you're going in, um, you know, in some cases for us, it manifests as, uh, factory visits and factory audits where, um, you know, we're actually walking around the factories and checking, uh, for, you know, anything out of the ordinary, that sort of thing. Um, so there's that level to it. Uh, there's certainly a ton of simulation. One of the things I didn't really appreciate before I got into, uh, the field and really, uh, before I got a few years deep into it is just the amount of simulation that's done on, uh, anything, anything and everything. I mean, everything from, and kind of getting outside of electronics, everything from, um, potato chips to, uh, you know, potato chip bags, you know, everything. Yeah.

Chris Gammell: You know, cause they're looking at Steve, are you hungry right now? I am a little bit hungry. You said everything. And then you said potato chips twice in a row.

Dave Jones: So I do, I do like a good potato chip, but, um, there's, there's a lot of stuff that, that gets simulated. And so, you know, going back to electronics, you know, all of these companies, um, anybody who's large enough to be putting, uh, millions of products out is going to do a ton of analysis. But, you know, really, uh, you're looking at extremely low failure rates. And so catching that is, it almost, you know, it becomes more of a statistical exercise than anything else. So, uh, we certainly, and, and as a standard in the field, there's accelerated aging tests that you can do where, um, and not necessarily even just aging, but it just accelerated, uh, testing where you might be trying to replicate what a device is going to experience. You're looking at some, uh, like subcritical metrics. So something that's not actually going to cause, uh, for instance, a battery fire, but something that's indicative of a potential fire down the road or a thermal event down the road. So you accelerate that testing. Um, but yeah, it's, it's extremely difficult and these companies do a really, really good job of it. Um, generally speaking, I mean, there are, we do hear about the extreme events obviously, but, uh, there's a lot of work that goes into making sure that you don't hear, you know, that there aren't more events to hear about. I should say. Sure. Yeah, exactly.

Chris Gammell: That's, that is the idea is that like, if, if that testing wasn't being done, we'd hear about it every other week.

Dave Jones: All the time. Yeah. All the time. Cause I mean, you think about just a standard battery, the amount of energy that's in there. Um, but the fact that, uh, things are engineered well enough to, uh, withstand some of the damage that we expose them to both, you know, in the field, uh, as consumers, but also in our testing as a mechanical engineer, I do terrible things to batteries and, and, and have difficult time getting them to, to do anything fun. Magical firebags right there. Yeah. Yeah, exactly.

Chris Gammell: Yeah. That's, uh, that's, that's tough. So like, so what should be, so, okay. So now back to, back to the listening audience, what should they be considering when they are putting a battery into a product then? I guess, I guess that's kind of part of, part of the, the consideration here is yes. A lot of products have batteries, lithium ions cheaper than ever. There are obviously some risks in there, but what should they be doing as general guidelines? And then when they, should they be doing it or what is this simulation that happens at the end?

Dave Jones: Yeah. Yeah. Great question. So there are a lot of decisions related to, and, and some of your listeners, uh, you know, certainly know this better than I do in terms of the specifics, but battery management units, um, you know, the, the electronics that you're actually connecting to the battery is, you know, are really, really critical with respect to, um, shutoffs. Um, shutoffs and, and various, uh, conditions on overcharging and, and trying to avoid those circumstances. Um, so certainly the, the software aspect of that is, is a critical component. Um, from my perspective, when I start thinking about the mechanics of it, um, there's, there's really two aspects of it. Um, and it, it relates to a thermal event of the battery that you really want to, um, as you consider the, the role, uh, the interplay between the cell, uh, or battery. And the environment, what you really want to do is you want to protect the cell from the environment. Um, so by that, we mean, you know, you want to make sure that whatever enclosure it's in is for whatever the application is, uh, is stiff enough or, um, protects the battery in enough ways, uh, to, to really prevent the sort of mechanical damage that could lead to a thermal event.

Chris Gammell: When you say cell, sorry, do you, do you mean the entire, the overall, like, uh, an actual, you know, 2000 milliamp hour battery pack? Or do you mean a, a subsection of that? A subsection.

Dave Jones: So, um, I would, I would refer to a battery as the assembled components, uh, which may include multiple cells and, and would include also a management unit of some kind and the connections between those individual cells. So some of the cooler things, um, along these lines is like, if you go and you open up, um, you know, some of the articles on Tesla, um, battery packs, they have, uh, just sort of the public stuff. Uh, they have a number of teardown pictures where you're looking at these little 18, six fifties or close to 18, six 50 cells, which is just describing a form factor of a lithium ion cell. Um, and they're all aligned in rows and, you know, connected in various ways. And so each of those little units is going to be a cell and then connected together is going to form the battery pack as a whole.

Chris Gammell: So, okay. I was, I was thinking about more like the pouch, like the square pouch style. Yeah. If it was like internal to that, if that was a single cell or if you met at the 18, six 50 level, I, I guess I understand that one.

Dave Jones: Yeah. And either way, really, I mean, you could have a pouch cell, um, where you have a soft polymer, um, pouch around the electrodes. Um, you could have a prismatic cell where you have a harder, uh, enclosure for the cell, uh, for the electrodes, or you could have a 18, six 50, uh, where it's, or similar, where it's a cylindrical design, um, or button cells, you know, a lot of different form factors.

Chris Gammell: So really just anode, anode cathode electrolyte is a cell, right?

Dave Jones: Yeah.

Chris Gammell: That's kind of the definition here.

Dave Jones: That's yeah. Yeah. That's one way to think about it. And you really want to understand, um, from a design perspective, um, you can imagine if you're somebody designing, um, a cell phone, for instance, your requirements or your design goals for protecting that cell are going to be very different from somebody who's designing a drone where, uh, you know, a cell phone might be dropped, but it's probably in a stiff enclosure. Uh, whereas a drone, uh, you know, if it's flying over a residential unit, taking pictures of fireworks or something, uh, and suddenly it loses control and falls down, you know, you may have a pouch in that drone, uh, and, and it's subject to crashing from 40 feet or something. So, uh, you want to know what sort of, um, potential mechanicals from my perspective, at least the potential mechanical conditions that your cell is going to be exposed to not only during intended use, but more importantly, during unintended use. Um, you know, if somebody takes a nail gun and shoots it at your drone, um, you know, what's going to happen.

Chris Gammell: Well, I laugh, but this is like part of, part of your work is like, you have to be like, well, what about nail guns?

Dave Jones: What about hawks? Maybe not hawks, but we did have a project where I was, uh, fortunate enough to rig up a pneumatic nail gun to shoot into a battery and see what happens.

Chris Gammell: That's so awesome. It was pretty cool. How are you not doing this on YouTube as like a, as a promotional thing for the company?

Dave Jones: We, we have talked about doing some of those things. Certainly. Right. It's finding the time to do it, I guess, but. Yeah, I guess so.

Chris Gammell: Okay. So you said, uh, the first thing you have to consider is protecting itself in the environment.

Dave Jones: Was it, what was the second thing you had said? The second thing, and this is, this is something that's definitely overlooked is protecting the environment from the cell. Um, so, uh, you know, there's, there's the concept of, you don't want the cell to go off, but, uh, as with, and this touches back on something we, we talked about a few minutes ago. Um, you know, these things are, are statistically driven. And they're very low probability events. Um, but no matter how good your process is, you can never engineer out all of the potential failure. So you have to assume that something's going to go wrong. And when that something goes wrong, you want to make sure that your device and your enclosure is sufficient to contain that event. Um, such that somebody is not going to be shot with molten aluminum or something.

Chris Gammell: Right. Right. Right. So that's the idea of like, and I guess that's kind of also UL testing as well. It's like, it's when failure happens, does it fail properly? Right.

Dave Jones: Exactly. Exactly. And you really want to, again, and this gets back to there's, there's sort of UL style tests where you have some sort of standardized way to test these things. And then there's considering the application and what is that cell and that device really going to experience when some teenage kid gets ahold of it and, you know, wants to drive his truck over it or something, you know, what is, what's really going to happen there? Okay.

Chris Gammell: Well, so let's, that's, that's as good an example as any, right? So, so that, so let's say that the truck driving over device is part of the, the expected failure mechanism. Is that something that you would then go and actually simulate or like how, how do you actually, aside from actually just physically testing it? Yeah. Do you simulate that piece or what happens then?

Dave Jones: We do to a certain extent. So, you know, there are various approaches for simulating the, the thermal runaway events of batteries that gets into a lot of different kinetics where you, you have a lot of chemical reactions occurring and, and, you know, some very detailed analyses that can be done. Um, but I think the, where you would get the most benefit is just, uh, from, again, looking at, from my perspective is, is considering the mechanics of it and really looking at taking your device as a whole, um, where you're simulating, um, the stresses and strains and deformations and so forth of your entire device under this truck loading. Um, and really looking to see, does the battery experience a certain type of stress, uh, in a certain location? Right. So one of the areas that might be sensitive for a battery is where the windings are exposed at the top, right? So you want to evaluate, okay, if I drive a, uh, F three 50 over this thing, is the top going to get smushed or is the enclosure strong enough to protect, uh, and minimize any deformation there. And when you do that sort of thing, those are much more established, um, approaches than, you know, getting into the kinetics of, of, of the actual, um, chemical reaction occurring inside.

Chris Gammell: Right. Cause you already know some, I mean, I'm guessing there's probably a bunch of literature around, around that, like the chemical stuff. It's like, okay, when it hits this certain point, bad things are going to happen. Just assume it's going to get worse and worse. Right. Right. But it's about how do we prevent that getting to that bad point in the first place.

Dave Jones: Exactly. Exactly. And you can get a lot of information just from doing, um, modeling where, you know, really even as simple as considering, uh, a battery is just a brick, like with some homogenous properties and just put it in there. But you're, you're ultimately want to evaluate the impinging forces and stresses on, uh, the cell itself.

Chris Gammell: So that was, so the output of that would be like, you would end up seeing like a vector basically that's kind of going into that brick. Then you'd be like, well, if the truck drove over in this way, you'd see forces here and here and here.

Dave Jones: Yeah. Yeah. And you see hotspots of stresses. And so one of the things we love to make are these color maps of stresses and they make beautiful pictures. And sometimes they tell you good information, uh, with respect to, you know, what type of stress and locations of peak stresses and things of that nature.

Chris Gammell: So what is, what are the programs that you're, what are the types of programs that you use to do that?

Dave Jones: Yeah. Uh, the program of choice that we use at least is this, uh, is Abacus. It's one of the, um, premier, uh, finite element software packages, just commercial packages. And so, um, we rely on that because it's already, you know, from a software level, at least it's been validated and so forth. So, um, you know, we don't have to worry about software level validation there. Um, but there are a number.

Chris Gammell: Oh, it's made by DeSau too. So is that part of solid ties right into solid works?

Dave Jones: It does tie into solid works. Um, absolutely. Yeah. So there's a lot of nice applications that way. Um, there's a, you know, certainly Ansys is another option. Comsole is, is one that could potentially, um, play in, although its strength is really in some other areas. But, um, you know, finite element, the nice thing about, uh, finite element analysis relative to, you know, going back to what I was talking about with my graduate work was, you know, finite element analysis has been around for 30, 40 years. It's a, it's a really well-known approach. Um, so it's trusted, uh, it's used, uh, certainly in, in any sort of major engineering project is going to involve, uh, some level of finite element analysis. So we can leverage the trust that the engineering community from a solid mechanics perspective has in those, um, those software tools, uh, to, to really get some good results.

Chris Gammell: So what is finite element analysis for us? Yeah. For us, electro weenies out here. Um, what is, I know this is a big thing in mechanical world, but like, yeah, I'm, I'm looking at the Wikipedia page, like, Oh, your Stokes differential equations. I remember those.

Dave Jones: Yeah. There's some math involved, but, um, probably the best way to think about it is, um, and, and most likely people have seen things like this, where you have a volume, uh, let's just even think about a brick. Like if you think about the analogy we were talking about, where you take a battery and just assume it's a homogenous brick, you divide that up into a number of different volumes. So, um, and we call those elements where, uh, you essentially solve those differential equations on an element by element basis. Um, and you're tracking the, the deformations, um, of the, the nodal coordinates that form the boundaries of that element. So you're really just looking for the movement of the elements where that movement is governed by, uh, material properties like stiffnesses, um, various things like that, that you plug in. Um, and it essentially just is a, a matrix solution type approach, uh, for getting you those deformations. So it's like spice for spice for stuff.

Chris Gammell: Yeah. I mean, I, I, please don't write in, you know, I know that it's not spice for stuff. I'm just saying it's, it's math being applied in a generic way to, yeah. Okay. Yeah. Uh, it is a reverse solver for that kind of thing. Okay. So how, so is it like, uh, so if, so you're assuming a battery is a brick, is it like you have a book of numbers that is like, well, batteries are four, you know, seven. Yeah. Right.

Dave Jones: Yeah. So, so that, that gets into a lot of the, the, um, the testing aspects of things as well. So, uh, you have, um, so for some generic things like aluminum, for instance, if you're thinking about an aluminum device, uh, most of those material properties are very well known, um, at least for, uh, your basic like stiffness and, um, you know, other sort of generic material property strength, et cetera. Um, for something like a, a battery, um, that's where, you know, we often get involved, um, where we want to measure the stiffness of the battery so that we can then use that information in simulation. So, uh, it's sort of a two pronged thing where we'll go and do some mechanical tests on the batteries, not to kill them, but just to understand how stiff they are and then take that information.

Chris Gammell: So like, does it, does it flex, you know, a certain amount of force and you're measuring basically how much it flexes at a certain point. And then you back calculate that into your model. Totally. Yep. Yep. Exactly. And does that always go okay? No, I guess they're, are they, are they stable enough that it doesn't matter that much or what does that go?

Dave Jones: You know, actually, I think you'd probably be surprised at how stable they are. Um, they're very, at least for, for most of the manufacturers that we work with, they're, they're very well engineered products. Um, in the sense that you can, you can do some significant deformation to a battery and, and get away with it. Um, and now it really does depend on exactly how you're doing that. Um, so for instance, there's the concept of just bending it. Um, but you know, the, the radius of curvature of that bend is, is potentially going to drive how explosive or not that event is. Um, so, you know, it depends on how you do it, but, um, it, it can get really interesting. And so a lot of those tests, um, we'll do, and you know, one of our facilities, um, we have, uh, it's actually a restaurant hood, but it's this big air handling unit. Uh, and we do all the tests under that hood, uh, within cinder block enclosures. So if anything exciting happens, yeah, it, you know, it's contained, it's safe, all that sort of thing.

Chris Gammell: So, um, yeah, that's good lab procedure stuff, right? Yeah. Don't want to kill myself. Yeah. Right, right, right. And, and is, so like you, you as the engineer on a project, are you actually doing that? Is it like each test is different enough that you need to basically go and like run experiments on this stuff? Yeah.

Dave Jones: And, and, you know, Exponent's model as a whole, uh, really is that we shouldn't be doing anything that's standardized. Um, so, you know, we try to advise our clients, anything that's standardized, you know, go and do it at Intertech or some of the other labs that are, um, you know, set up to do high throughput, um, tests in that, that manner. But for things that we get involved in, typically there's something unique about it, or many, maybe multiple things that are unique that justify, uh, you know, our role in, in setting up that test. And, and so, you know, going back to, um, this notion of, um, this notion of sort of collecting people who have done interdisciplinary things and have gone through a PhD program where they've done research, you know, everybody who we bring in ideally has, um, you know, they, they experimental chops to reliably run a test.

Chris Gammell: Um, and you know where the, the method design of experiments, that kind of thing.

Dave Jones: Yeah, exactly. Right. Exactly. So.

Chris Gammell: That's important. Yeah. So how do you then go, how do you go back, calculate that stuff to, okay. So again, we're talking about a, a, a battery as a brick, right. And you know, now, okay, there you go through your finite element analysis, you know, that there's going to be a hundred newtons of force at point X, Y, Z, whatever. Or, but then you, you, you, you, you, Steve, the, the, you know, the PhD smarty pants, uh, sorry, that sounded way negative, more negative than it should have been. Uh, you, the consulting engineer, uh, you, but you, you, at some point someone has to say, oh, this is bad. Right. So how do you know? Like, so I, I hear a hundred newtons of force. I don't know that I made that up, but like, at some point you have to say that is going to be problematic. So how do you calculate when that's going to be a bad event?

Dave Jones: Yeah. So that's, that's sort of the third level of analysis where you've done your first sort of material property testing, where you just want to understand what are the, what's the stiffness of the brick? Uh, then you go and you do your truck simulation, you drive a truck over it and you say, okay, you're getting that a hundred newtons. Then you say, okay, well, if I'm getting a hundred newtons in that location, is that something I need to worry about? Right. So you set up another test where you apply that load in that location, um, under the conditions where you're trying to match what you are seeing in the simulation in terms of the worst case scenario. And you're trying to recreate that worst case scenario in a lab setting to understand, uh, whether or not that's going to cause something that you should be concerned with.

Chris Gammell: Oh, okay. So I was actually thinking it was going to be like another, uh, like you'll have to excuse me here. I thought there's going to be like a book that's like, well, if there's a hundred newtons of force, batteries explode. But obviously I know that doesn't actually exist, but I just mean that at some point I thought there would be like a calculation that's like, don't, don't have too much. Yeah. Like, like if the liquid expands too much at this point, it's going to cause a heat.

Dave Jones: Yep. Yep. Yep. Yeah. So there, there are certain, you know, sort of rules of thumb that you'd want to avoid. Um, but that's much more well established for something like a generic metal. Um, so you can think of, um, you know, there's this, this website called mat web. That's just got these lists of all these different materials and it, it has their, um, you know, mechanical properties is one of the things that's included there. And one of those mechanical properties are, or maybe a few of them are the strength of the material. So for your generic aluminum, you know, at what stress is it going to break? Um, so that's very well known for something that's homogenous and established like an aluminum. But when you get into something like a battery, for instance, or solder or anything, um, that's maybe got, uh, uh, less well established material properties, um, particularly with a battery where you have chemical kinetics and you have other things going on. It's very difficult to a priori say a hundred newtons is going to cause this stress and this stress is, is the failure point. You really should be doing some sort of testing because your chemistries are going to be different, um, potentially for different manufacturers, even with the same specifications. Your chemistries could be different. Your kinetics could be different. A lot of different things could happen that you need to test out.

Chris Gammell: Got it. Okay. And so, so that's why at that point you're saying that you would, so you're saying here's a worst case scenario because that's going to be the worst thing. We're going to try and replicate this a hundred newtons of force. And again, we're making up all these numbers here. Oh, a hundred newtons. Uh, but if you're going to be able to then apply it repeatedly in some controlled environment, and then you're going to have a statistical model, like you might age that battery again to see if there's a, a negative impact of that kind of thing and throw whatever you can at it at that point. Is that kind of the idea?

Dave Jones: Basically. Yep. Yep. And, and you know, those, those sorts of things, because of, you know, how involved it can sound, uh, you really only want to do on, on some of your more significant events like a truck. Um, but yeah. Right. Ultimately it's, it's something where as a designer, you need to make an evaluation of what is my device going to experience. And, you know, maybe some of that you, you get from product knowledge where it's been out in the field and you've gotten some returns and those returns have tire tracks on them. You know, something, or it's got a nail through it. Yeah. And you know, we laugh about some of these things, but you know, we also do a lot of sort of, uh, failure analysis on things that have gone wrong and we see some crazy stuff. So, well, I mean, that's the thing,

Chris Gammell: like, especially with consumer stuff, like I've never worked in the consumer space, but I can only imagine, like just the, the sheer number of like events that they have, let alone the really crazy ones, you know, like totally, totally. And then some of the people that try and return them for refunds too, I'm sure is like, Oh, Oh, what happened? We don't cover that one. Yeah, exactly.

Dave Jones: Exactly. What do you mean? There's, there's a 16 gauge nail sticking through this thing. I know we're

Chris Gammell: not covering this. Yeah. Right. Right. Yeah. That, that one didn't, didn't make the cut. Sorry. Okay. So what is, what pops out all this then? So is it like, uh, again, I know we were making like a crazy scenario here, but is it like, uh, you have a report that's like, well in the truck, the truck driving report or what? Yeah, basically. So, I mean, how does it get, how does it get fed back into the next design cycle for the companies that you might be working for?

Dave Jones: Sure. Yeah. So, um, in, in that scenario, you know, we would, we would let them know what the results of the tests are, um, you know, and, and typically where that plugs in for them is that they might say, okay, well, under this worst case test, you know, we see that we need to enhance the stiffness of the enclosure, uh, in this location. And so they'll have their, um, you know, CAD guys go in and modify, um, what the enclosure looks like. Maybe they'll add a new component to provide some additional stiffness or, you know, whatever the case may be. Ideally, they're going to take that information and they're going to understand their vulnerabilities and then they're going to modify their product to overcome those vulnerabilities. Well, I'm guessing that the,

Chris Gammell: you know, in the case that the exponent is being hired as well to, to do something, it's probably in risk, you know, it sounds like a lot of this stuff is reactive versus proactive because in a proactive scenario, it's like, there's thousands of millions, unlimited number of things that could happen to a product. How would you choose it? You know, you'd have to have as many, you know, it'd be like monkeys and typewriters type thing, right? Except. Yeah. The monkeys are not typing, they're shooting nails into things.

Dave Jones: Yeah. Those are crazy monkeys. You see any of those, you run away, but.

Chris Gammell: Right. Exactly. That's right. Yeah. Never trust a monkey with a nail gun.

Dave Jones: Nope. Nope. Yeah. Uh, you know, and you bring up an interesting distinction there between reactive and proactive. So, uh, you know, we classically work on, uh, you know, the company was founded as sort of a reactive firm, um, that, you know, back in the day was doing things like airplane crashes and bridge collapses and stuff like that.

Chris Gammell: Yeah. Um, calculating what happened in the first place, right? Like what actually, how, what went wrong? How do we not do this ever again?

Dave Jones: Exactly. And, and, you know, we talked about the Samsung example. That's a, that's a scenario where that was, you know, what we would call reactive because something happened and, you know, we want to understand what happened there, but we do see a lot in, in many, you know, in fact, most of the things we can't talk about are things that are proactive where, um, you know, we're involved in the design cycle and while we don't design anything, you know, we're advising on design choices. So, um, should you include that structural rib in that location or should you move it, you know, things like that. And really the motivation for that, uh, often what I see at least is that it comes out of companies, internal reliability testing where, you know, again, uh, the amount of, of testing that these folks do is really impressive and gives you, gives me at least a lot of confidence in products, but, um, they will identify some issue where it's, you know, it's maybe it's, it's some stapler and they're dropping it in, you know, 10 different orientations and it always breaks or ejects a staple when it's dropped in the seventh orientation. And so, um, you know, there it's a question of, okay, well, why is the seventh one worse than the sixth or the eighth one?

Chris Gammell: That makes sense. Okay. And that, that is, that is good too, because it sounds, it sounds like that's the time to call, you know, uh, you know, uh, not, not that they wouldn't call you sooner, but I'm just saying that at that point, if you, if you are doing these, you know, ad hoc, uh, real life situations, uh, at a certain point, you, you, you could dive into yourself. You might have someone internally at a company, but that's also something where you might call a consultancy, like the exponent to do that

Dave Jones: kind of thing. Right. Right. And, and typically in those scenarios, um, you know, there, there's something by the time we get the call, they've tried to fix it and they are running into some sort of roadblock. So, you know, maybe it's something that requires a different discipline, uh, you know, knowledge of, of glass mechanics or whatever. Um, so they may call us, or it may just be that they don't have the bandwidth. So, you know, everybody's tied up on something.

Chris Gammell: Yeah. Yep. That makes sense. Cool. Well, uh, before, you know, we're starting to run out of time. I amazingly, we've talked about enough nail guns, I suppose, but, um, but I did want to talk about the other piece you had mentioned. So you had mentioned CFD, which is computational fluid dynamics. If I remember my intro classes, uh, uh, what, what is that stuff? Cause you had mentioned BGAs as well, you know, again, specifically to the electronics industry. Yeah. And we got off into the,

Dave Jones: into the battery tangents, which is, you know, something that's really interesting, obviously, but, um, the BGA side of things and, and looking at, uh, solder performance and things is a strong component as well of, of a lot of the electronics work that we do at least where, um, you know, now you're looking at on a board level, um, you're trying to understand failures of products. And in these circumstances, more often than not, uh, what we've seen is that a company has put these into, maybe it's a thermal cycle, um, where they're just, you know, um, baking these boards, um, as just part of the normal reliability testing. Um, and they start to see some sort of failure. Well, like the

Chris Gammell: same thing of like Arrhenius equation, trying to do like, like trying to do accelerated life testing,

Dave Jones: that kind of idea. Yeah, exactly. And so, um, you know, some of it may be motivated by returns, but you know, more often than not, it's, they catch it before that it gets to that point. Uh, and they've seen some sort of reliability issue and then they want to understand, okay, well, why is, why is my solder ball in this one position, um, failing? Uh, you know, what is, what's really driving that? And so that's where we get into these ideas of heat transfer, um, where CFD is, you know, going to be a prime, um, tool to evaluate that as well as finite element analysis for sort of other applications. But, um, you know, the, the concept of your heat transfer paths and, um, you know, really what happens when you have a certain board location heat up, how does that change the stresses in the solder on an adjacent chip, for instance? Um, those sorts of questions, yeah, can really come into play. And I was just listening to, I think it was maybe six episodes ago, uh, was your conversation with, um, with Dave about the, uh, the BGAs and you guys were talking about solder masks and those sorts of things. Um, you know, you're really getting into, uh, those questions of layout, uh, and where I come in and think about that is, you know, you, you have this, uh, potentially a massive chip, uh, if you're looking at like a motherboard for a laptop, right. Really can, can dominate things and everybody, you know, you have an FR4, which is going to be, uh, you know, it's a composite with some copper layers and things. So it's going to have a stiffness, um, but you're going to have silicon chips in there. You're going to have, um, over molds. You're going to have all sorts of stuff going on. Uh, and where, you know, from a mechanical engineering perspective, where that gets really interesting is as things start to heat up, you have different, uh, expansion coefficients of those materials. And so, you know, one material wants to expand a lot. Another one doesn't want to expand, you know, and the poor solder is in the middle and it's got to deal with the fact that the board is expanding maybe more than the chip is. And, you know, it's cupping or something or saddling. And what is that really going to do to the stresses in the solder balls? Um, and are you, are you saying this about the actual assembly

Chris Gammell: process or you're saying after the board has been assembled and it's just, you know,

Dave Jones: trying to keep the chip connected? Yeah. So we've done a number of projects where we've looked at their whole reflow process of, uh, putting chips on and, and what really, uh, those are, what those balls are experiencing. Um, and that gets into, so there's, there's this concept of, um, fiscal elasticity where you have some sort of time dependence of your material properties and you have the ability of stresses to relax out. Um, so you want to, in some cases you need to incorporate the fact that you've got a chip that you've applied to the board, uh, and then you're putting it through another reflow process to put another chip on, um, you know, and you're deforming the board and now you've got a board that's, um, because of the expansion process, because of the cooling process, as you're taking it down from, um, out of reflow, you know, now you're, you're really, um, you've got a deformed board, you've got solder ball stresses. Um, what does that do over time? Do those relax out or do those results in cracks through the solder that then can lead to, you know,

Chris Gammell: uh, drop signals of some kind. Right. You know, actually, uh, I have dealt with maybe not specifically this connect, this situation, but I have talked to people. I work with a lot of like modules that are actually like chip on, uh, chip on modules, like they're, they're pre-certified RF modules and they talk about like reflowing those and like the need to bake them out again. If you, so like, say you want to remove it off the board and solder it back down, you're like, don't do that. You know, that's a bad idea. And I think it's because, and if you look at it, it's like sandwich on sandwich on sandwich, right? These days, you know, they, you put down a, an RF module that's effectively a tiny, not probably not FR four, but some, you know, fiberglass control impedance board with an RF chip on it. It's been tested. It's got a metal can around it. And then I'm going to solder that to like an ESP 32, right? That's, I guess maybe that's not a good example, but like a, you know, a little Bluetooth module or one of the ones with a metal can around it, you go and solder that to another board in a totally different way. And, uh, and now you've got stresses that are just stacked up and up and up and up. So exactly, exactly. And then once you

Dave Jones: start looking at, okay, well, uh, you've got those stresses baked in now, you know, potentially you're doing this reliability testing where you're, you're putting it in an oven and you're running it through some thermal cycles, or maybe you're taking that board and then you're exposing it to a drop scenario where you take a device, you drop it on the ground. Uh, what kind of, uh, stresses is the, are the balls going to experience? And as a result of that, um, you know, it's possible that you end up with, uh, due to the manufacturing process, you end up with something that's close enough to the failure point that all it needs is just a little more stimulus, which it may see

Chris Gammell: during the drop process. Well, now that we've scared the absolute pants off everyone listening, uh, who's made boards, let's talk about some practical, like what people can do for best practices or maybe when also when they should call someone like you, right? Cause I, I mean, like for me, like I've done this before and it, you know, most of the time it works, right? I'm obviously not making millions of anything, so I don't need to worry about it that much. If it doesn't work, I grab the next one. Not, not saying that's my design method, just that that's my practical way to get through certain prototyping issues. Yeah. And somewhere between prototyping and making a million, when do people call you and why should people call you? Yeah, that's a great question.

Dave Jones: So, you know, I think there, there are a number of, of good practices. Uh, I, you know, from the perspective of board layout, that's really a question that, um, is not something that I'm in a good position to answer, but, uh, certainly with respect to supports and, and considerations for how you mount your boards, um, you know, that's something that you really want to think about, um, what sort of deflections you're expecting your board to experience where, you know, you know, again, really we're looking at, or I'm considering scenarios where you've, uh, exposed your device to some sort of mechanical perturbation, like a drop or an impact or something like that. Um, and so that's like a

Chris Gammell: board mounted inside an enclosure that that's going to be hit the ground at some weird angle.

Dave Jones: Yeah, exactly. Yeah. And, and we do enough of these simulations and, and one of the benefits of them is you can see what happens on sort of a microsecond, um, timescale where you're watching the board vibrate. Um, and so, you know, those it's, it's really interesting. Um, but you know, that speaks to the importance of supports. Um, you know, it speaks to the importance of if you have a large chip that may dominate the board, um, you know, trying to make sure that you've got, um, your sort of subordinate chips, um, spaced far enough away. If, if you have the luxury of space, um, spaced far enough away from that larger chip, because you would expect that that larger chip is going to keep things in its vicinity pretty flat. And then you're going to get large deformations coming off of that. If the board as a whole wants to move. Uh, so it's this notion of just sort of chip spacing. So, but really, uh, you know, when you should be looking for, uh, assistance is, is a scenario where, um, you've designed something that is, is working fine in isolation. Um, you know, if it's not under some sort of, uh, stimulus environment, but you really are expecting it to experience some kind of stimulus or, um, you know, you're, you're noticing that during its normal use, um, maybe it's a thermal cycle event. That's, you know, you weren't expecting it to be all that intense, but, uh, you're leaving it in the car or maybe it's a, some sort of connected device or, you know, maybe it's a LIDAR for instance, where, um, you are, you know, microns could be really important with respect to resolution. So you need to make sure that you understand what sort of thermal, um, deformations you expect to see in your device, uh, in order to keep the sensors true to whatever

Chris Gammell: the datum is that they're supposed to be at. Got it. So like, so like, uh, so if people are doing like tolerance stackups and they're doing, uh, high reliability situations, if they're doing high stress situations, those kinds of things are where the, this really, really comes into play. It seems like exactly. Exactly. Also, it sounds like large, larger chips, chipsets, right? So like, just because it, the thing that's planar and needs to be flat expects to be flat. And you're talking about like leaving chips away from that, that big mama chip in the middle. Right. Right. How does this play back into the, the computational fluid dynamics side of things?

Dave Jones: I mean, like, right. Yeah, yeah, yeah. That's a good point. So we talked about this idea of reflow, um, and, you know, in, in those circumstances, typically it's homogenous temperature. Um, but the interesting aspects of the CFD for instance, is when you start, um, running chips pretty hot. So, um, or exposing them to environments where, uh, heat transfer is going to be really important. So, uh, if you have, um, you know, uh, let's say, uh, an outdoor wireless router of some kind,

Chris Gammell: you know, maybe that's going to be. I always think of automotive too. Totally. Like automotive, like they're running hot, they're, they've got crazy processing and then they're like, oh yeah. And this can work up to like, you know, 85 C no problem. So that means you have temp rise on top of 85 C it's like, Holy crap. Yes. And you know, that's insane to me. Yeah. Yeah. And so the,

Dave Jones: the CFD aspect of it and where the simulation side, um, uh, plugs in is you want to understand how easily the device or the board or whatever is shedding that heat that's being generated. Um, so you, you need to look at, uh, you know, maybe the dye is heating up quite a bit. Um, but it's got heat transfer pathways, you know, down into the board, uh, and through the traces in the board, as well as it's convecting up, um, in a way based on whatever the airflow patterns might be. So, um, you know, there, there are different scenarios and different cases where you might look at a forced convection where you've got wind flow over or airflow over your, your board, but more likely you're just looking at sort of ambient, um, temperature distributions and, and sort of, um, non forced convection, just sort of natural convection away from the board. Um, and there you're calculating, um, you're ultimately just trying to determine how much heat am I getting away from this chip? And as a result of, of not being able to get all of it away, what's the resulting

Chris Gammell: temperature of the components? Right. Well, I can imagine too. Like, so I've had, uh, I remember I used to have a cell phone that was like, anytime you took a video, it got just like roasting hot and you know, like, cause image sensors just heat up a lot and they take a lot of power and blah, blah, blah, blah, blah. But then, you know, then tying that back to earlier where now you have a chip that's generating a bunch of heat and then it's on the backside of a battery. It's like, Oh, okay. Well now you have to then go and double check to make sure you're not heating up a battery in a bad way or doing that kind of thing too. Yeah, you do. And all those things tie

Dave Jones: together. And, you know, you also have the sort of the human factors aspect of it. So you mentioned this idea of it getting hot. Um, but potentially if it's a cell phone, for instance, you might be holding it against your head and you know, everybody's experienced the scenarios where the phones really heat up. Um, you know, it's, it's possible that something like that is going to, you know, it's, they're engineered to the point where they're not going to result in any sort of uncomfortable, um, experience, but, uh, you do want to consider that.

Chris Gammell: Unless we forget the, uh, the Sony laptops that were known for, uh, for heating up a little bit too much in the, uh, the old, the lap region, the lap region, indeed. There were some bad, bad times had by all. And that's right. So they, that in that scenario though, like they, they didn't do that calculation or they didn't know what was going on. And they, I think they had a recall on that too.

Dave Jones: Right. Right. So the, you know, it's possible that, um, so obviously your devices are going to have some fan of some kind and, you know, maybe it was just underpowered or it wasn't working properly. I don't, I don't particularly know in that circumstance, but you know, ultimately you're not shedding the heat. Um, and it may be too, that they were expecting it to be elevated off of a surface. Um, so maybe there's right, right, right, right. But if it's sitting on your lap, your body's not very good at conducting heat away. So it's just going to absorb it and it's not going to transmit

Chris Gammell: out. And so that's part of the human factors, like you're talking about of like actually seeing it used in situ understanding what's happening. And it's like, I guess that's like a product testing

Dave Jones: as much as anything else. Right. Yeah. And you know, there's user studies and there's things that, you know, some of the larger companies do to just understand how is this going to be used? And in those circumstances, you know, what is that doing to all the potentially vulnerable components?

Chris Gammell: I mean, how much of the time are people coming to you because they're like, well, look, we have to, we have to operate on the edge of this, you know, like there, there are specs and it, at least for me, like I'm, I'm very rarely pushing specs because why would I be, I'm not designing anything crazy anymore, you know, but like sometimes like a car company might be pushing a specification because they have to be right. They're operating in Dubai and automobile, whatever. Um, how often would you say customers are coming to you and being like, look, we just have this scenario and it, it mandates that we're on the edge of the specs. And then, and then we're going to calculate if we can get away with it. Is that not get away with it, but you know what I mean? Like if it'll be sufficient margin to operate. Right. Yeah. I, it's not as often,

Dave Jones: thankfully it's not as often, uh, as you might think. Um, so there's usually pretty good safety factors built into a lot of these things. Um, you know, and I, I think where they're operating, um, to the edge, uh, there's usually going to be a, um, yeah, there's going to be some sort of sensor issue where it's, you know, it's really the functionality, not necessarily the, um, anything dangerous, um, that they're, they're operating under, uh, thankfully, but, you know, for instance, um, part of the march of, of lithium ion batteries in terms of getting, um, higher energy densities and so forth is, is, you know, shrinking some of the layers, um, that are involved in that process where, you know, you're, you're making things thinner, uh, in order to pack more into the same volume. Um, and, and that's a circumstance where, you know, really ideally you're doing a lot of testing to verify that any changes you're making in order to pump up your capacity are not limiting other aspects of functionality. And that might end up, and hopefully that's also being done at the

Chris Gammell: manufacturer level too, so that when they specify it, they're like, well, now we've changed our boundary box because we are going thinner. We're going to give you a thinner battery. That's what you asked for, but don't do a, B and C now because it's the, you know, the, the margin has shrunk for our side, right? Yeah. Yeah. So yeah, that's lots of, there are, there are a lot of things,

Dave Jones: uh, to think about certainly keeps us busy, but, um, yeah, but that's good. That's good. Right. But it, you know, it's also, um, really, you know, when you start to see the interiors of some of these larger companies, it gets, I do have a more of a sense of ease, I guess, because, uh, there's a lot of really smart people working on, uh, you know, really important problems in terms of trying to get some of the more, what may seem functionally to be a minor gain, um, in performance. You know, it's all, it can be a result of, of, as you mentioned, pushing things a little bit, uh, in terms of thinning things out or whatever the case may be, but tests are being done and people, you know, really are paying a lot of attention to reliability because it doesn't take much, uh, to show up on the front page of CNN, uh, for, you know, something that, you know, ultimately may not be your fault, but the, the social media costs of

Chris Gammell: something like that, uh, you know, maybe large, uh, nightmares. Yeah, exactly. Exactly. Okay. So what, uh, when, when you do have an issue that's like, so again, to, to, to kind of talk about what is, what is the stuff that is output of this analysis, right? So the CFD analysis, what is the output? But then is it like, Oh, move this via or like, what, what is it that you would then tell them

Dave Jones: to do? Right. Um, so some of it, uh, might be, yeah, as specific as, as move a via or, you know, move their chip a little bit in this direction or that direction, or give yourself a little bit more spacing or something along those lines. Um, other aspects might be, uh, something that we see often is, uh, stiffeners. So sort of board stiffeners that you might put in to protect components or to keep things planar. Um, some of it might be, uh, some of these underflow or under fills that you might put under a large chip, um, to, to provide additional support between the board and the chip, um, so that not all of the mechanical compliance has to come from solder. Um, so it might be a matter of, of changing up the formulation for that, um, and using a different underfill for instance.

Chris Gammell: So yeah, that's, that's good to know. I mean, cause I guess, I guess ultimately if like the, if you look at what the, what the normal outcome of like, uh, of recommendations is, I know that there won't be standard, but if you look at the outcome, then that can also gives an idea of like what some of the issues were in the first place. So like, like you were talking about flexure there and you're also talking about like heat transfer and, uh, you know, all those other things. So I could imagine there's unlimited things you could suggest, but, um, it's good to know kind

Dave Jones: of what has been suggested in the past. Yeah. And typically, you know, the results of one of these analyses is, is most often a specific mechanism for why something did what you didn't want it to do. And so those specific mechanisms are typically treatable in some way.

Chris Gammell: Right. It's a very bounded problem that's difficult to solve, but, but obviously you are solving them and it's like, okay, so there's going to be a solution for a bounded problem, which is good, right? I mean, that's science. Yeah, exactly. Yeah. Science. Uh, speaking of science, uh, to take it back to the beginning. So like, is this something that, that exponent and you and, and, you know, your colleagues, is it something that you publish about or is it because it's being paid for by the companies that it all stays internal?

Dave Jones: More of the latter. Um, one of the things that's nice about the company is we are encouraged to, uh, still stay in the scientific and engineering community with respect to conferences and publishing, um, things. So where we might publish, um, and, and typically do publish is, um, something that we want to potentially educate the, um, you know, the, the discipline about. So it might be something that we're seeing in specific projects for a client where it's, you know, a particular component is causing issues across the field. So, you know, there's a lot of different companies facing these issues, or maybe it's, um, something that we see as a potential issue down the road. We may want to alert our clients with respect to, you know, the, and, and all of them potentially that, Hey, you

Chris Gammell: know, we've seen this issue. Um, let's stay away. And even like alerting your coworkers in a public, where I guess you have to be kind of quasi public, you know, internally, at least to be like, if you see a, you may also see B and watch out. Yeah. So there's some institutional knowledge.

Dave Jones: Yeah, exactly. Exactly. Yeah. But yeah, we do, we do try to publish and a lot of that relates to just staying current in the field. Also, you know, um, that's, that's obviously a very important component of what we do is making sure we know what the latest technologies are and so forth. So some of that relates back to, um, you know, publishing and staying in it. But a lot of the stuff, as you pointed to, you know, is, is funded by clients and, you know, we, we take that confidentiality, uh, very, very seriously. So we don't often publish things that are

Chris Gammell: specific to clients. Yeah. Right. It'd be like if it was public enough and if they were willing to publish it as well, that kind of thing. Yeah. What are some of the conferences that you would

Dave Jones: go to for this, this kind of stuff? Yeah. So, um, I, I get involved in a lot of different areas, um, with respect to the, the battery work, there's a, a few, um, larger battery conferences. Um, I forget some of the names right now, but I often actually do a lot of, I still work in the medical device field as well. Um, and I go to a number of conferences that are specific for like cardiovascular modeling and things of that nature that are a little outside of what we're talking about here, but, um, also, um, work in some of the additive manufacturing areas where we're going to like polymers conferences and talking about properties related to that. Yeah.

Chris Gammell: That's cool. Yeah. Okay. All right. Well, if you think of any, we'll add them to the show but, uh, yeah, that's yeah. Um, so y'all hiring? Yeah. Yeah. We're all, we always are. Um, so,

Dave Jones: you know, one of the, we do have a pretty, I'm not sure I'd say strict, but we do, uh, try to, um, look for folks who have their PhD in, in some field that's appropriate for engineering. So, um, we, we try to set the bar at that. Um, yeah. So what, like you talked about too,

Chris Gammell: it sounds like a pedigree type thing from, uh, you know, you're, you're approaching clients as like, Hey, we've got all, you know, the stable of, of PhDs and this type of method. So that, that makes sense from that perspective. Yeah. But I'm not too put off by that.

Dave Jones: Yeah. It's, it can, depending on folks, they, it can be a little bit, um, difficult for some folks. We are always looking for anybody who's got a battery experience or who's got, um, electronics experience, obviously in mechanical engineering, we, uh, approach and delve into a lot of different fields. So, um, people who have mechanical PhDs, um, this, especially if they're listening to your program, they're probably pretty interdisciplinary. So, um, you know, we'd like to hear from them.

Chris Gammell: Awesome. How do people get ahold of you? And like, if they've got a problem or if they want to talk about, you know, the product or generally how do they get ahold of you?

Dave Jones: Yeah. Uh, probably the easiest way is, uh, just shoot me an email. Um, so, uh, I'm sure you'll include in the show notes, but it's just ask S cruiser at exponent.com. Um, and you want to check the spelling on that. Cause the last name's a little bit non-intuitive, but, um, that's probably the best way. Uh, other methods include LinkedIn. I try to be pretty active on there. Um, you know, it's a, it's a good opportunity to share ideas with the professional field. Um, but, uh,

Chris Gammell: I was going to say, if you follow Steve on Twitter, you're probably gonna get a lot more Sabres than you are, uh, Buffalo Sabres. We're both from Buffalo. He's, he's much more of a sports fan than I am, but you'll get a little bit more, uh, go bills and go Sabres than you

Dave Jones: will. Yeah. Yeah. For no good reason really. But, uh, well, you know, I, I, I was very luckily

Chris Gammell: cured of my, uh, sports, uh, at a young age. I don't know what happened to you, man. Yeah. I don't know either. And in a lot of ways, I don't know what, well, Steve, I'm, I am continually, uh, amazed that you and I keep crossing paths and, you know, like obviously went to the same high school, the same college. We ended up in the same house, obviously as roommates. Uh, but then now that we're in the same frigging field too. It's like, it's a, it's fate, buddy. It is indeed. And it, it sounds like you are doing some cool stuff. So I'm really glad you came on the show. Yeah. Thanks for having me. I appreciate it. Uh, I had invited you, hopefully you will come to town for IMTS. You said you might or might not do that, but if you do, we'll be, we'll hang out. I think I might. It looks pretty awesome. That's coming up in September. That's machining show here in Chicago. If anyone else is going to that, let me know. And past guests, uh, John Saunders will be there as well. So he's doing a, uh, uh, event there. Okay, cool. Well, thanks, Steve. We'll talk to you soon. Sounds good. Take care.