Impedance Control Demo w/ NCAB Group Field Application Engineer Ryan Miller

In this episode we welcome back NCAB Group Field Application Engineer Ryan Miller. The first time we had him on the show he gave us a lesson on what impedance control is, in what situations is it necessary, and more.

This time we recorded a video demo going through how he would create a stackup the requires controlled impedance.

In order for this episode to make sense watch the video version:
https://www.youtube.com/watch?v=XwsyTUTN1z0

pickplacepodcast.com

Transcript

0:05

Welcome to the PickPlace Podcast, the show where we talk about electronics manufacturing and everything related to getting a circuit board into the world. This is Chris Denny with Worthington. And this is Melissa Hugg with CircuitHub. Wait, I forgot we were like recording a video. Can we start over? No, this is the best part. This is, this is the show. This is the show, Melissa, where we don't, you know, we do very limited editing. says the guy who doesn't do any editing. Alright, well, um, if you're watching this, then, uh, Melissa has successfully edited our very first video podcast. And if you're not watching this, how did you get this feed? Uh, so great. Yeah. So we, we've had a hard enough time and we've been busy enough where it's been difficult for us just to get some audio recordings out. So we thought, Hey, what better way to get the podcast going again to try than to try our very first video podcast, right? Exactly. Exactly. Yeah. And if you're wondering why we haven't been making, um, podcast episodes, uh, listen to the previous episode. Yes. Big announcements in the previous episode. So, um, if you are listening to this, uh, I assume we're going to put out, uh, an audio feed of this episode as well. But if you're listening to this audio only, it's going to get really confusing because, uh, today's episode is going to be a video episode for a very specific reason, which we'll get to in a moment. So I highly recommend that you either pause this, um, check your show notes for a link to the... Uh, to the video so you can watch it on a computer because this is going to be a different format for us and it will probably be incredibly boring if you try to just listen to it. Unless you are one of those listeners who use the podcast to fall asleep, then this one will be perfect because it will make absolutely no sense, even less sense to you than any of the other episodes. Well said, Melissa. Well said. Uh, so yeah, so today's episode we had, um, uh, recorded episode 56 was an episode where we brought on, uh, Ryan Miller from NCAB to discuss impedance control, super popular topic. We have people reach out to us all the time, asking for more information on impedance control. And, uh, that was a really well received episode. Got a lot of good feedback on that one. Um, but if you'll recall, if you did listen to that episode, we had so much more to talk about, but it was almost a little too difficult to talk about, uh, an audio only format. So we thought, uh, you know, at the time we said, Hey, let's, let's do a video episode where we can, uh, get a screen recording and, and, uh, learn even more. So that's what we're going to try to do today. Um, there's a lot of great information out there about impedance control. Uh, you can just do some Googling, you can find some tremendous stuff, but hopefully this is just another, uh, valuable resource for you. And, uh, if you happen to come across this video, uh, on YouTube and you're wondering why we're blustering through this, please, uh, listen to our podcast. Cause that's what we normally do. We don't normally do, we don't normally do these video episodes. So I highly recommend Google for a pick place podcast and subscribe and, and follow and all that good stuff in your favorite, uh, podcasting app. So, but, um, to get things started, let's, let's invite Ryan back. Uh, he's very patiently waiting in the green room for us. Um, And, uh, yeah, welcome back to the show, Ryan. I think you're going to have to unmute yourself because I no longer seem to have the ability to do it for you. That's okay. Uh, thanks for having me back. It's always, uh, a good thing to get invited back. Right. That says a lot, right? Yeah. So all of our, all of our guests who we've only had on one time, that's because we hate you and we never want to have you back. Is that true, Melissa? I think that's what Melissa said. She told me that the other day. Oh, yeah. Yeah. No, no, absolutely not. Absolutely not. We've loved all of our guests. We just had so much more to talk about with Ryan and we wanted to get him back on the show. So, um, yeah, great. So, uh, if you are listening to this or watching this and you didn't listen to episode 56, Uh, it's kind of a prerequisite. I think you're going to be a little bit lost as we go over these things here, but to just bring you up to speed a little bit, that episode was all kind of like a very early introduction to what impedance control is, um, why engineers need it, uh, a little bit about why engineers need it, but more importantly, how a fab actually does their work to make sure your impedance control, uh, matches, uh, what you need it to be. And that's where, uh, Ryan comes in and we'll work with customers to help them, uh, understand how to do that. So episode 56, highly recommend go back and listen to that before getting started with this episode. Um, and Ryan, so we, last time we talked, you had, you had thrown out, uh, a few different types of impedance can, uh, control and or controlled impedances. I don't know. I don't know these terminologies. I'm not an engineer. Uh, I just play one on TV. And I didn't even stay at a Holiday Inn Express last night. So at this point, I'm not sure what I'm doing. In that case, how about I go back and explain, uh, what we had talked about a little bit last time. Yes, please. That would be, that would be much appreciated for sure. Yeah. So the last time I talked about why we need impedance control, you know, some of the different. Connections like, uh, for antennas, for, uh, semiconductor devices, um, the data sheet will tell you if the, uh, impedance transmission lines are, uh, required for the device. Um, some of the structures that are common, uh, that I talked about was the coded microstrip, the embedded offset strip lines, uh, microstrips. Um, You know, those are all the technical terms of the, the traces themselves, but really when it comes down to it on a day to day basis, um, what I get is requests, something like this that we're seeing here on the screen where it says six layer rigid, O62 thick, FR4, you know, six layers, 50 ohm single ended, And, uh, 100 ohm differential pairs, they're on layers one and six, it's very basic. So I got questions already. Okay. Um, so I understand six layer rigid, this is a rigid PCB, it's not a flex PCB, got it, that means there's six layers of copper that have to be considered, uh, 0. 62 inches thick, understood, that's a standard thickness of a PCB, FR4, that's the, that's the material the PCB's made of, the woven fiberglass and the rating of it. Yeah. Now, when it says layers 1 and 6, it's pointing out that there's a 50 ohm single ended trace that you need and a 100 ohm differential pair that you need. Now, when it says layers 1 and 6, does that mean there's one on each layer, or does it mean that there's, that the 50 ohm is on layer 1 and the 100 ohm is on layer 2? Well, that means that, uh, on layer 1 and layer 6, we're gonna have a 50 ohm single ended, which would be, uh, what? I called a coded microstrip in the last podcast and a 100 ohm differential pair, uh, which is what I called, um, where's my notes? Last time I called it an edge coupled coded microstrip. Okay. So we're going to have one of each trace on, uh, layers one and six. Okay. Okay. Got it. Understood. So this is where the video podcast comes into play. Because it's so hard to talk about that without video. Yes, as we learned in our previous recording. Yeah. What I have here in front of me, um, is a basic six layer stack up, uh, 063 thick over solder mask right here. Oh. Um, and what we're looking at. So the solder mask adds a little bit of thickness, just like one and a half mil, maybe. Yeah, here I have it set to a half a mil per side. Um, it, it varies, you know, I've seen when I worked in the factory, we used a half a mil. Uh, some other factories I've seen are seven tenths of a mil, but, um, it's not going to affect the thickness, you know, going from five tenths to seven tenths is not going to really have a large effect on those impedance traces that are on the outer layers here. So what we're looking at here is a stack up that I put together. This is a cross section of it. is a basic six layer stack up, um, using FR4 materials and I got it to be 0. 62 thick over copper. Okay. As we mentioned, 63 over solder mask. Um, and you know, on a daily basis, that's a good place to start. Uh, you know, when I put together a stack up, I, I have a starting point in my brain. Um, you know, I'm going to start with this. Basic setup, and I'm just going to add to it. So, when I get to this point. This part you just clicked here, this is the, this would be considered like the top layer. Where you clicked one foil, copper foil, that's the top layer of copper. Yeah. Okay. Yeah, so, let's walk through that. Um, we have, uh, the top layer of copper here. Um, layer two is here. Okay. With layer three. Yep. Now, in between layers one and two. We have prepreg. Right. This, uh, this comes to the manufacturer that's going to build the PCB in an uncured form, and it's just glass and epoxy fabric. Yep. It's the same exact glass and epoxy fabric that this core is made from. The only difference being, uh, this core is a lot thicker. In this case, I use, uh, an 18 mil core, which is a common off the shelf thickness. Um, but this is already cured. This is one solid piece of material. When we put the board in the press, it's gonna heat up to, uh, temperature and it's gonna ramp up to gel. It's going to come back down and while all that's happening under the tremendous pressure that it's under, this prepreg is going to cure out and it's going to be solid. Just like the core when it's all done. So basically you're almost building cores in between, much thinner cores, but a little bit with that prepreg almost. Pretty much, yes. But, you know, there have been some instances where we had to build custom cores in the factory, but those cores were, uh, not UL certified, so... There you go. You know, we, we can do a custom core like that, but they're not UL certified. In this case, it's, it's not really, it doesn't, um, it doesn't, uh, count. Sure. I should say, for lack of a better term, it doesn't qualify as a core. Um, it's just prepreg. Yep. Understood. It's just basically bonding. It's creating an insulation between layers one and two and bonding them together. That's exactly what it is. Yep. Bonding and insulating. Gotcha. Okay, cool. And then, and then rinse and repeat as we go down from three to four and five to six. Yeah, so when we think of the stack up, we, we think of the overall stack up. You know, we got layer six, five, four, Three, two and one. Mm-hmm.. But when I think in terms of controlled impedances, it's like, uh, you know, how do you eat an elephant? One bite at a time. That's right. Kind of thing.. Yes. Yes. We're gonna talk about just part of the stack up. Okay. Uh, and we call it an impedance structure. Okay. So if we go back to our requirement, you know, we wanna add these impedance traces mm-hmm. and. One thing, the first thing I want to demonstrate is, you know, just adding a 50 ohm trace, but I also want to demonstrate how to adjust it. First we're just going to start with the dielectrics we have here, and I'm going to add A single ended 50 ohm trace and a 100 ohm differential pair. You have no idea how excited this is for me to see. Oh, really? Because I'm experiencing the end result of this all the time. And I love to see how things are done and how things are made. It's such a black box for me. It's such a mystery for me. So when you just click that like new button and you're selecting things, it's like, Oh, here it comes. I'm finally going to understand what, what in the world is happening here. So very excited. I'm going to try to explain it. Um, it really, if I can do it, anybody can do it. It just takes time to study. Yeah, sure. Yes. It's not difficult. Uh, it's just, it's very basic math and, uh, you know, For someone who's interested in this, I would encourage, uh, more study into stackups and materials and what's going on during the press cycle. Sure. Yeah, visit a PCBFab. If, if you're really, um, curious about these things, find out if there's a local PCBFab and ask them if you can tour because that was one way I learned quite a bit. I will tell, I will warn you. You, you may leave there more confused than you entered. However, I highly recommend it because you will learn something. I remember going into the press room and being like, Oh, now I get it. Now I understand registration, why things don't line up, you know, absolutely perfectly because you're trying to line them up in this registration process before you press them. I get it now. Yeah. So, and, and you'll learn a lot about that kind of stuff. If you can visit a local fab, you'd be surprised if you live in the United States somewhere. I guarantee there's a, there's a fab within a couple hours drive from your house. There's quite a few of them out there. And, and many of them, uh, are all about giving tours. Yeah. Especially to their customers. Yep. But just make sure you don't wear your good shoes that day. Wear simple shoes. There's a lot of chemicals on the ground, yeah. Yeah. Yeah. But, uh, getting back to how I'm setting up the impedance traces for this board. Uh, it just by default, this software puts in a five mil trace width, the lower trace width, but if it didn't, excuse me, that is where I would start to begin with, uh, whether I'm doing, you know, the single ended traces, differential pairs, um. Coplanar waveguide, which we're going to get into a little later. I'm always going to start with a five mil trace with one exception. If the customer tells me, yeah, this is, this is, uh, an application where I just need a 17 mil trace, and it really has happened, uh, more with, uh, the coplanar traces and it does the single ended and differential pairs. Okay. If a customer tells me that they're targeting a trace that wide, then, uh, we'll go ahead and shoot for that. And so, and real quick, that, that sort of trapezoidal shape of, of the combination of the lower trace width and the upper trace width, that's just, that's a normal. shape of, of a trace. You do have a little bit of a, uh, of a trapezoidal thing going on when, when you fabricate a PCB and the traces. Yeah. Yeah. I'm, I'm glad you brought that up. So when we etch the traces, um, they're going to go through the etching solution and. In a nutshell, more etching solution is going to hit the top side of the trace here than it does the bottom side of the trace. Yeah, sure. When we're laying out impedance traces in our layout software, we want to always, always, always go by the lower trace. Okay. The upper trace width is something that we're really only concerned with in certain RF applications. Okay. I've seen But, for the purposes of controlled impedance, uh, calculations, it's important to know. Um, since this is an external layer, it's going to start with half ounce copper and plate up, uh, after final plating. Uh, we're figuring about a mil in the hole. We're actually going to... at about 1. 9 mils thick. And that's because when you're, when you're plating the holes of the PCB, you're, you're, you're also adding copper to the traces. It's just a side effect of that process. True. Yeah. So when we do final plating, what we want to do is we want to add the final amount of copper in the holes. To bring it up to one mil thick in the barrel. And we also want to enhance the pattern on the external layers. So instead of just being the base copper, it's going to be finalized copper. I see. Oh, okay. I see. Very good. But anytime we etch, we're always going to have, um, that trapezoidal effect. So in controlled impedance traces. What I'm going to say here is kind of contradictory from, from the bottom to the top, we're in this case, I'm figuring, uh, about a mil of loss. Uh, if it were strictly a print and etch and plating wasn't involved, I'm going to say you're going to experience about one mil of loss per ounce. So, um, if it's, uh, an internal layer, which are print and etch only, uh, you know, we. Say you have a 5 mil impedance trace, you're going to have a 4 mil loss, or a 4 mil topside width when it's done. Uh, if it's 1 2 oz copper, then I would say it's going to be 4 1 2 mils. Now here's where the contradictory part comes in. The outer layers are plated, and I'm still going to use the 1 mil per ounce. So really what I'm saying here is, you know, this is going to... You know, we say it's gonna finish at one ounce copper, but it's really it's a little bit over. Okay So what is one ounce copper thicknesses? Would you say that's 1. 9 mil? That's one ounce copper thickness? One ounce copper off the shelf is 1. 4 mils thick. Okay, so, you know, we're only 5 tenths of a mil away. Yeah, which if you think about it is almost Half ounce copper because half ounce copper is 7 tenths of a mil thick. So we do the impedance traces. So I'm gonna stay with the the one mil loss and we calculate see where we're at and Wow, we're really high. We're at 73 ohms. Sure. Okay, so there's two different ways I can I can change this. I can change the trace width maybe say Go way up and just it's kind of arbitrary, you know Changing up to 20 mils. Sure. We can go You know this time we hit low and this is a good opportunity for me to bring up something that's going on in my brain Let's go back to well, we'll talk about it with the 20 mils. You know, we're low this time one of the rules I always remember is Uh, more impedance, less copper. So, let me explain. So here, we, we need the impedance to increase. So, we need to decrease the trace width. Sure, yep. And that's what we're doing here. We're just playing with the trace width to get it there. Um, we can go to 15, try that. So the, I see you're not entering any length in this calculator. It's just width that you're playing with so far. That's true. And that's a very good observation. Um, I wouldn't have thought of that one on my own. So I don't lay out very many boards. Yeah. The length is not a variable in this scenario. Okay. Um, we're only concerned with, uh, you know, trace width, the copper thickness. thicknesses. Or when I set up the stackup, it takes all of these variables from the stackup, but yeah, you're right. Um, TraceLength is not there. Okay. That's something I had no idea. I always assumed that was part of the calculation. Good to know. Yeah. So, what I'm gonna do is I'm gonna cheat a little and, uh, do what they the software's gonna actually, based on these dielectrics, tell me what would get me at 50 ohms. That's crazy. So you guys... You guys will target tenths of a mil like that. Like you might try a 12. 125 mil. Like you're, it seems almost unachievable to nail that, that trace width. I would imagine you have a margin of plus or minus something there. Yeah, yeah, on the, you know, in a manufacturing scenario, the IPC gives us a plus or minus 20 on the trace width just for manufacturing purposes. Plus or minus 0. 2 when you say 20. Plus or minus 20 percent, I'm sorry. Okay, 20 percent, gotcha, gotcha, gotcha. But in this case, we're not really concerned with, um, We are concerned with the trace widths and, and their tolerances, but we're not thinking about it in those terms just yet. Um, but you know, a 12. 125 mil trace is a, an extremely wide trace for a single lone, a single ended trace on any layer. Okay. Um, but the 0. 125... Yeah, 12 mil is pretty wide, yeah. Yeah, the 0. 125, um, you know... What I'd like to see is this, you know something more a little bit more simplified Yeah, and you know, we're still gonna get there. Oh, yeah But what I encourage you to do what I encourage your viewers to do is something like this Okay We go you can see we're still we're still getting there. We have maybe a 1. 2 percent error, which is a little bit more than I like. Okay, so Let's do this. Okay. That should get us, get the error down. Okay. The error is down below, uh, one. So 1%. So I'll take that. The reason being is because, you know, when we lay out circuit boards, we have, uh, traces all over the board. We have five mil traces. We have eight mil traces. We have 10, 12 mil traces. Uh, sometimes. It gets convoluted and it's hard to find those 12 mil impedance traces among all the other traces. Sure, of course, yeah. So, this differentiates it. That tenth of a mil makes it stand out? Exactly. Yeah, it makes it very easy for the factory to find. And, um, you know, I was just, since I've, I've been doing this for the past 11 years, um, You know, in the factory scenarios, I saw quite a bit where it's really, really hard to find those traces. So wait a second, you don't get a screenshot from the engineer saying these are the traces right here. They just tell you, they just tell you look on layer one and look on layer six for my impedance controlled traces. Sometimes we do very rarely, but most of the time, uh, this is what we get. Does it help to have a screenshot showing where the. It does. Screenshots always help. Oh yeah. Um, screenshots help anything in PCB. Yes. There's two things that are always definite when we're talking about PCBs. The first is screenshots explain so much and in some scenarios they're almost mandatory. And the second one is we're talking about PCBs so the answer really depends. Yeah. That, that's always the case. That's always the case. Uh, can I do this? Well, it depends. But here's what I want to do. I want to, I'm sorry, go ahead. I was just going to say the same thing on the assembly side, man. Like the more screenshots and markups and pictures of your, of your product, um, 3d renderings, anything you can give us, like we'll take it. You know what I mean? Like, we'll take anything you want to give us, uh, that we can use as a reference and, you know, we don't, we don't always need it, but if you give it to us, it's going to make, it's going to make us happy for sure. Yeah. I would say, you know, more often than not, we'll get a stack up drawing on the fabrication drawing. And the fabrication notes in conjunction with that stack up drawing will, uh, spell out all the impedance requirements. And that's really enough. Okay. But, uh, yeah, we, I did this just so viewers will know that we don't have to have the fancy drawings because not every board comes with one. Yep. Okay. Very good. But what I'd like to do next is get these 50 Ohm traces down to something a little bit more reasonable. Okay. So I'm going to go back to. Uh, 5 mils, but in this case, we're just going to go straight to it, 5. 1. Um, we're still a little, we're still really high when we go back there. Yes. So what I want to do is dive into changing dielectrics instead of, when I set up impedance traces, and I know I'm going to have Uh, single ended traces on the same impedance structure, which is, uh, you know, um, this right here. I was talking in terms of breaking the elephant down. Yep. We're just thinking in terms of these dielectrics right here. Sure, sure, sure. Okay. So I want to make them thinner in order to, uh, bring this impedance down. Thicker dielectrics require, uh, more, uh, require wider impedance traces sometimes. And they require, uh, they are going to produce, um, higher impedances. I never would have, never ever in a million years would have thought that that would have anything to do with it. That's fascinating. Yeah, but we have to, we have to be careful here in this, uh, The scenario for an extreme environment circuit board, you know, something that's military aerospace is going to see a lot of shock and vibration. Uh, we don't want to, uh, do what we're about to do here. And that is. Uh, we're ultimately going to end up with a single sheet of prepreg here. Okay. And those extreme environment boards that experience that shock and vibration, we want, always, always want double ply between, uh, every structure. Okay. And that's just so it has, uh, plenty of bonding strength to withstand those shock and vibration events. Okay. Now, on a typical circuit board... I would go with, uh, you know, something like this, where we have just, uh, one sheet of prepreg between the outer layers and the next innermost layer. That's very acceptable because really all we're doing is bonding this one layer to, uh, this one layer here, right? Now, to make the point, if you look at the stack up here, we're bonding these two layers together. Mm hmm. This layer here is toothy and has traces and features on it. Um, and it's going to require a little bit more prepreg resin to, to bond these together. Whereas this layer here, this side of the foil is just all smooth. So it's not going to experience as much Uh, resin loss during the press cycle. Of course, I never thought of that. So then, help me to understand then, for layer two, uh, you don't have the, you don't have the toothy stuff, is that because layer two is more of a... More of a copper, uh, you know, it's all filled with copper. It's not a bunch of traces. Yeah. Layer two, in this case, I've set up to be a plane. Yeah. Um, but let's, let's make it more realistic. Um, it could be a mixed plane signal. In that case, we're going to have a little bit different loss. Uh, what I'm doing is, um, um, I'm applying this finishing to the stack up. So it, uh, calculates. The resin loss. You saw the prepregs change here. The core stayed the same. So what that does is simulate the press cycle at the factory where the resin loss occurs. Interesting. Okay. But, you know, we can go with this either way. Um, you know, if we were back to just planes, um, That's actually what I see most commonly is just a straight plane. Yeah, let's go with bread and butter then. Let's not, uh... Yeah, let's start off small. So, Okay, what I want to do is I want to decrease this dielectric here. The first thing I'm going to do is I'm going to start switching out the types. If you, if you look at this 2113, it's a little bit thicker than the 1080 here. The difference in those two prepregs is, you know, the the 1080 has a different ratio of glass bundle to epoxy than the 2113 is. Or has. Um, the 20, the 1080 having actually a little bit more resonant in class. Okay. As compared to the 2113. So what I want to do is, uh, I want to change that 2113 to a 1080. Um, uh, and I'm really just copying and pasting it here and I'm deleting that. Okay. Yep, I see. Yep. And, you know, without doing anything else, we're going to just, uh, calculate and see. Uh, where we're at, and you know, we've dropped a little bit, um, you know, the next thing I could do is, uh, start changing out types again. I can swap this out for, uh, a 106, and that's going to make this total dielectric from here to here, uh, just a little bit thinner. No, wrong button. Sorry. So you can see we did drop a little bit, but nothing really substantial. So at this point, I'm thinking, okay, I want to go a single ply prepreg on this structure. So what I'm going to do is swap this out. And when I do go single ply, let me. Let me say this, you know, you can see that this single sheet is 2. 4 mils thick, this one is 2. 7 mils thick. Now IPC 6012 says that if it's not specified, if dielectric thickness, minimum dielectric thickness on the fab drawing is not specified, then it's going to go to three and a half mils. Okay. So I, I try to stay above that because I'm setting up. Uh, something for a customer and I don't know, always know all the details about their product. So, I want to respect that. Um, so I'm not going to use either of these two sheets as a single plot. Yeah, because they're not thick enough. Yeah, I want to swap that out for something that's thicker. And I'm going to go back to the 2113 that we had in there before. And I'm going to just simply delete the 106. And we can see now that we're, we're getting into range here. Um, we really want to get that impedance down a little bit more. So what I like to do is for now, keep this impedance structure the same and just change the trace a little bit. So I'll go up in a one mil increment. Okay, now we're a little closer. So maybe I can go for, uh, six and a quarter. That's perfect. I can probably even, uh, push it if I want to. Sure. There are some high levels of OCD going on right now. Yeah. I'm, I apologize. Oh, you're two hundredths off. No, I'm just kidding. Perfect. So when I get that, that trace close, I do go double check in this case, we got it down to 0 percent error. I usually shoot for anything less than 1%. Yeah. Uh, as acceptable. So, okay, now we got the, the single ended trace set up. Yeah. And I, next we gotta go set up the differential pair. That was my next question. But I wanna Now you, you've been influencing the calculation of the differential pair this whole time by everything you've been doing with the prepreg and, yeah. Exactly, yeah. So it's a lot easier to, if we had set up for the differential pair. Then tried to put in the single-ended trace after that, it's, um, gonna be a little bit more difficult. Okay. Because, uh, it takes, uh, a little bit more adjustment of the trace width and dielectrics to dial in a single-ended trace than it does a differential pair. Mm-hmm., because then the differential pair, you have the trace separation as a variable. It helps out tremendously. So, um, set up your single ended traces first, and it'll be easier. And I'm sorry, did you have anything else? Well, so what I was going to say is just the fact that, like, you have an additional variable now. You know, because of that trace separation, you have an extra lever to pull on this that you didn't have on the single ended strip. Yeah, exactly. Yeah. So, yeah, I, when I first started out in my career doing this, you know, I was just as lost as anyone else and I started setting up differential pairs first because I thought they looked so much cooler than the single ended traces, right? It's right. It's funny how engineering comes down to human factors so many times. It does. It just looks cooler. Yeah, it looks cooler. It operates faster, generates more heat, uses more power, but we need it. This is the same reason that all of our customers love to use black PCBs even though we want them to use green. Yeah. The black... When I started out in this industry, I started out in solder mask. I was actually a solder mask technician for the first year. So you can call yourself an ink stained wretch. Oh, you should have saw my lab coat. It wasn't as bad as some of the guys, but yeah, it had solder mask of different colors all over it. I believe I wasn't ink stained. But I digress. I haven't, Oh, okay. Here's where I was going with this. Yeah. Yeah. The green solder mask is, is pretty transparent when you compare it to the black solder mask and the black solder mask can cover up. And the Blacksauta mask does not play well. with, um, light sensors in manufacturing operations. So we use light sensors on all of our, uh, automated systems to detect when there's a PCB present. And the black oftentimes does not reflect enough of the light back into the sensor that it's projecting. So there's on the sensors, they're projecting a small amount of light and they're waiting for that light to reflect back into itself to know whether or not a PCB is there. And green and blue, these things reflect really, really well, but black. It's black. It's like inherent in the color that it doesn't want to reflect light. So it doesn't reflect the light back into the sensor. And inevitably the board just, they zoom through like, what, wait a second. Where's it go? Yeah. Yeah. Something you would have no, most customers have no clue. That's something we have to deal with. So we had to like swap out a bunch of our sensors and stuff to make sure that they were compatible with the black solder mask because people, that's what they want. How do you get through that day without pulling out your hair? Uh, it's a miracle I still have it. I'll be honest with you. Really it is. What I did there was, uh, I just made a quick change on the video. I noticed that I forgot to apply the finishing when I made this, uh, modification down here. Okay. So that's recalculated real quick. Um, we're still at less than 1 percent error. We're still good. So. Yeah, now we got to dial in. This demonstration, we're in good shape, I think. Yeah, for the, yeah. Well, if this were a customer stackup, I would say, okay, it's, it's good. I might, you know, my OCD might try to get it a little closer. Sure. But, yeah, as you were saying before, we're almost there. So, you know, in this case, we're, I start out with a 10 mil pitch. I have a five mil trace width and a five mil separation between traces. Um, and that's just arbitrarily, you know, if you want to start out with a wider pitch or, uh, a finer pitch than that, you may do that, but this is kind of like a good B starting point for me. Um, for some reason, five and five, when you're talking about PCBs, it's always a good place to start, except for some of those really small semiconductors. 5 and 5 is a dream. There's a lot of those now. Yeah. It seems like every day I'm, I'm dealing with at least one or two PCBs that do that. Yep. There, I'll tell you, it's funny. There's a big difference between a half millimeter pitch device and a 0. 4 or even a 0. 3 millimeter pitch device. Big difference. Uh, difference in the cost to, to get the stencil made, uh, the, the, the, The trace widths and the technology needed to get the traces right. The, um, the accuracy of the equipment, it, it's, it might seem like a 10th of a millimeter, but it's, uh, it's a lot more than that. Yeah. And you know, it on, on the PCB manufacturing side of the world too, it does the same thing, you know, the, the smaller traces and spaces are going to. Uh, require a more advanced factory, which, you know, are more expensive to get boards from. Yep. Yep. And I, as a matter of fact, I was working on redesigning a product for a customer and they're like, Hey, let's go USB C because that's what everything's moving to. And I'm like, okay, we can, we can do that. That's fine. Yeah. But the, the semiconductor that controls USB C is. Four times more expensive. The, the jack itself is like six times more expensive. The P C B FAB is gonna be X more expensive, it's gonna take an extra 20 hours of engineering. It's like all you're doing is just a very like 9,600 BO communication. Like you don't need . Yeah. U S B for the U S B C for this. Trust me. So we went with that old school U S B uh, connector. Yeah. As a daily consumer. The old school USB connector was just fine for me. Oh, now there, there's where you and I were going to have some fighting words. Cause I can't tell you how many times I've tried to plug in the old USB A connector and it's been upside down. And as soon as they switched to USB C where it could go either direction, I was so thrilled with it. So from a connector, from a physical connector standpoint, it's fantastic. From an implementation on the software side, it is a mess. That's a popular topic on the PickPlace podcast. Yeah, but I always just, I, yeah, I do too. I, that is one thing that pains me about that. That connector, but I don't want to go too far off topic. Let's go back into the traces. Okay. So as you mentioned before, we, uh, we're almost there in this case, we need a little bit more impedance. They're kind of low, so I can't change the dielectrics because as you mentioned, uh, these dielectrics are already dialed in for, uh, the 50 Ohm traces. So in this case, our only choice is to. Um, change the trace width. So, uh, I'm going to go at, because there are differential pairs, you know, I went at one mil increments for the single ended, but because these are differential pairs, I'm going to go at half, at half mil increments. So four and a half, and we're going to stay with the one mil, three and a half. And then we're going to keep the same pitch. So I want to adjust that as well. And, Yeah, I would say, you know, we're right there at the 1 percent mark. It's really interesting. I would say, you know, if this were my stackup that I want to send to the factory, I would leave it. Okay, perfect. So we have those, those traces set up. Now here's something that I see in the field a lot. We'll have an existing design like this And we have to rev spin and add a new device that's going to add new impedance traces. Um, so we can put in some 85 Ohm differential traces. How about that? The catch is they have to be at, uh, A certain width. So let's arbitrarily say that because it's already an existing design, we only have so much room. So we have to have them at, let's say, three and three. Okay. And I chose three and three because I just know it's not going to work, but I see this in the field all the time. Uh, so we're just going to add one more differential pair and we'll go with. So now I see on the top of your software, you're, you're on the, you're on the impedance control three of three. So now we're editing the third. Yep. Yeah. So we want to have a three mil trace width with a three mil separation. That gives us a six mil pitch, um, which is, you know, the typical of what we have to work with. Uh, matter of fact, I'm, I'm going, uh, back and forth with, uh, one of my customers now who's in this exact same scenario. Um, and their board has been at a stopped point for more than a week now, just because of this scenario. So we have some impedance traces that, uh, we can't fit in here. And not only can we not fit them in here, but the impedances are really high. Um, in this case, you're going to have to go with what the calculator shows you. So I'm not even going to try to go through. And arbitrarily go up and down through the numbers. Depending on your software, you may have to. A lot of software is different. I'm just going to goal seek here. And find out that, you know, with that pitch, it can even achieve a proper trace width to get the target impedance that we're looking for. Right. Not achievable. I see that. Yeah. And, you know, so really the first thing I'm going to do is I'm going to start changing the pitch. And I'm going to do one mil increments, and I went too far. Too much. Yeah, so, um, we're going to go with a pitch that's a little bit more reasonable. So back to my five and five. But I thought you were saying in this scenario, sometimes the customer is telling you they need it to be three mil. Yeah. Yeah. That's the point. You know, this board that I'm working on is jammed up because of that. Um, What it's really going to come down to is if you're going to use this stack up structure and this combination with the existing, with the previous design these are the traces that we're going to have to use if you want to achieve 85 ohm differential impedance. So you would just go back to the customer and you would say, hey, I know. You told us you need a 3 mil trace here, but this is going to be your differential impedance if, if you, if you force us to do that. However, if we can change it to this, then we could get, you know, you give them the option and all right, you decide what you want us to do. Well, I'm going to go back to the customer and I'm going to, going to explain, Hey, you know, here's the calculation with the 3 mil trace width and space that you wanted. It's way off. In order to get 85 ohms, this is the calculation that we have to go with. And I know that you can't fit that into your design, um, but you're probably going to have to go back to, uh, play with redesigning a little bit to fit that in there. Yep. Yep. Otherwise, what we're going to have to do is rework the whole entire stack up, uh, redesign everything, um, maybe from, maybe from the ground up. And, uh, it's quite possible that some of the other impedance traces that we already had dialed in from the previous build are going to change. That's right. And everybody that, uh, that does PCBs and layout work knows that that's not a good thing. Right. So, it's really a hard situation to be in. You know, it's, we can't change the, the physics involved in manufacturing, um, that's just something we can't do. Mm hmm. I mean, you'd be super wealthy if you could. We could. If I had that magic wand, um, I would set up a booth somewhere in a very public place with a big sign that says, Bring me all of your PCB problems. That would make a lot of money. Yep. Changing physics, one PCB at a time. So, okay, let's, let's go into, um, another scenario, uh, that is really quite common. And, you know, after we've done what we've already done, it won't be very hard to understand. We're going to add some impedance traces here to, uh, this, this inner layer here. Before we do that, we want to go back. These dialectics here are most likely going to change. We want to go back and check our board thickness. Quick question. Right here. If you, if you change the prepreg that's between layers 3 and 4, will that have an impact on the impedance of, of what you've already calculated for layers 1 and 6? Good question. No, it won't. Okay. Um, when I was talking in terms of thinking of this structure here, Uh, the only dielectrics that infect, that affect the impedance traces on this top layer here is everything between layers one and two. Okay. As soon as this, as soon as the power switch is turned on and this board lights up, the impedance trace, well, every trace on the board is going to have some kind of transmission. Sure. But the impedance traces, uh, they're going to have a transmission where they're looking for their reference plane, which I forgot to mention. I apologize. This plane right here. Apology accepted. Yeah, so let's go back for a second. You can see when we're calculating those, the impedances are referencing a reference plane here. Yes. So really what's happening is the electromagnetic energy is flowing through the trace and it's looking for this reference plane. I see. And when it hits that reference plane, um, It pretty much flows back up into the loop. It doesn't go, uh, it, it. Some of the electromagnetic energy is going to flow into this adjacent dielectric, but it's going to, uh, stop at a certain point. It's mostly going back into, uh, the loop. Uh, and that will, you know, the little bit of, uh, electromagnetic energy... Uh, that does flow, is absorbed by this adjacent core, does not affect these traces here that we're about to set up. Gotcha. So, while we're talking about that, in this case, uh, when we're putting impedance traces here, we're gonna have, uh, a reference plane here. Now, when an impedance trace is looking for its reference plane, it's looking for the reference plane that's closest to it. Well, alright, so then if you have a core that's 17 mil thick, that's a lot further away than your prepreg, which right now is only 6. 6 mil thick. Yes. So it's looking to reference this right here. Now, if we had it set up so... Uh, the, and, and it does happen very rarely. That's another thing that, uh, that scenario of adding the, the 85 ohm trace can, can throw it off if we have it set up. So the adjacent, uh, core or the adjacent dielectric is the closest impedance or the closest plane, but it shouldn't be the reference plane. Um, it's going to throw off the calculations and, um. You know, the board most likely, it may or may not work. I don't know. I see. So, so when you are using the term reference plane, you're, you're referring to the fact that layer two is just a big copper pour as opposed to a layer with a bunch of copper traces. And so it's going to seek that out as opposed to layer four, which is a bunch of copper traces. Yes. I see. Okay. All right. I understand now. Gotcha. And, you know, when we're setting up the layout, when we're laying out the board and software, you know, that's where we're going to set up. It's automatically going to go through our VIA connections that are made when we're doing that. So it's not something that's special that you have to set up. Uh, it's, it's just, uh, you know, I'm talking a little bit about the physics of what's going on inside the board when we turn on the light. Um, Another thing to note here is we do have those, those two, uh, layers with traces on it. So the dielectrics, um, here are going to bond together and to each layer. Um, now since this core here, you know, this is an 18 mil core. It's really thick in this case. It's not going to, um, reference any place, but this, because from here down, it's a lot thicker, right? But just by my experience, I can, I can look at this and know that if we're going to put some phrases here, uh, the dielectrics are going to have to change. I'm a little confused by what you just said because when I look at the core, I see that core as being 17 mil thick, well almost 18 mil thick, is that correct? Yeah, we'll call it an 18 mil thick board, or core. And then you said that the prepreg, you're, you know, that's only what 6. 6 mil thick at the, at the current, the current stack up. So technically it's, it's much closer to layer three is much closer to layer four than it is to layer two. When you look at it from that perspective, yes. Okay. But what, what I'm talking about is. This prepreg here, um, in conjunction with this core. Oh, because, because you're talking about the plane. I get it. And so it's much further from layer five. I get it now. Okay. Yeah. So I haven't got into it yet, but this plane over here is not a mirror plane or is not a reference plane. It's a mirror plane. So we're going to reference this plane. And this plane is going to come into play a little bit too. As you can see, let's, uh, well, let's just do it. Let's jump right into it. We'll add, um, we'll do a 50 Ohm trace. And you can see here, uh, in this software, just that it automatically selects the structure that we need. In, in other softwares, you're going to have to. Find out what structure matches the impedance structure that you're working with. Okay. And yeah, let's, let's throw a 90 ohm differential pair in this time. So now we're just focusing on terms of the structure. Again, the structure is comprised of the reference plane for this layer here, but You know, let's flip it on the other side. If we're going to put traces, impedance traces here, then this becomes the reference plane. Yes. But for these traces only. Yes. And then this becomes the mirror plane. Right. All right. Yep. Um, so yeah, I'm just going to start at 5 again. Hit the wrong button. So I'm at 85 ohms. That would be great if we're doing the 90 ohm differential pairs. But, uh, let's do this. Let's go ahead and swap out. This core for something thinner, and I'm going to go pretty drastic here. Um, I'm going to go to a five mil core with a half ounce copper on each side. I don't like the way they have these buttons set up here. That dangerous button there, I wish it were all the way over here. Yep, I know. I've used a million pieces of software. I've had the same sentiment. You put okay and cancel right next to each other. Really? Really? Yeah. That's pet peeve of the week right there. Oh yeah. Well, I'll tell you about it later. Never mind. I don't want to digress again. I digress all the time. That's okay. So that's, that's the name of the show is digressions with Chris Denny and Melissa. Poor Melissa. Well, anyways, we, we got a lot closer going from going to a five mil core. And I chose a five mil core because, um, you know, through my experience, I've learned that, uh, five, four and five mil cores are great. When you're working on in penis traces on internal layers, um, it doesn't mean you have to use a 5 mil core. You can use a 12 mil core. I, you know, going back to the 5 mil rule, I, I start out at 5 mils. But in this scenario, in this case, you know, I think that, uh, a 6. 45 mil, uh, trace width, um, well, nope, we're on the wrong structure. In this case that, uh, we still did get down there, you know, the, the five mil penis trace is great. We're going to stay there. Um, no, we're going to dial it in a little bit more. Okay. Cause if, if we go, let's see, which one is it? Which one? Yeah. It has a little bit more error than I would like it to have. So it's a single ended trace. I'm going to. Go in terms of, uh, one mil increments, but because we're on half ounce copper and it's an internal layer, I'm only going to do a half mil loss. Um, going back to the one mil per ounce on an edge layer. Yep. Makes sense. So we're getting there. Um, you know, I'd call it good. We can dial it in a little bit more, but I'm gonna leave these here for the time being. Okay. I just want to take note of the impedance for the viewers. We're pretty high. We are gonna dial those in, but as you mentioned about the stack up thickness, and I, as I mentioned before, you know, before we even started these, I wanted to go back and check the thickness. So I know that we're gonna need to increase the thickness. I want to go to the thicker prepreg The thickest prepreg as I can fill into this, uh, area here starting out. Um, changing this dielectric is not gonna ha is gonna have very, very little effect on the impedance traces on this structure here. Now, when you change the dielectric between the impedance, uh, layer and its reference plane, it will change a lot. I see. That makes sense. Okay. As it did before, because you had 18, you went to 5, and that had a huge impact on, on the calculation. Yeah. Yeah, it did. And so now I'm just gonna swap these out. Um, as I mentioned, go with the thickest prepreg you can. Um, that is... Um, available. That's how we want to say it. And in this case, well, I already had 2113s in there. I want to go to 2116s, which are, uh, you know, the next step up in thickness. And when we get there, we find out that, that, you know, adding that little thickness, um, did have a little effect, but not as much. Um. So my goal right now is to forget about the impedance traces just for a moment and get the board thickness back up. Um, by my, you know, we don't want to add any more than, let's get out of this. I don't want to add any more than three sheets of prepreg here on this particular structure. If we do that, Um, most likely what's going to happen in the press at the factory is the registration is going to be, uh, thrown off a little bit. Of course. Yeah, that makes sense. Yep. Because there's too much prepreg gel in between those layers and they, you know, when the press, when the, when the boards go into the press anyways, just because of the prepreg that we do have here, when it gets up to temperature and gels up, there is going to be a little bit of movement. The more gel we have in there, the more movement we're going to have. Sure. And the registration is going to be gone. So I want to stick with no more than three sheets. And as we do that, we can see that, um, you know, it's just not there. So there's two things we can do. One thing is we can increase these dialectics and accept wider traces, right? And in most scenarios that that would work. Um, but if you have to have thinner traces, what we would need to do is kind of, uh, put a blank core in here. And that's going to make for a more expensive, uh, circuit board because it's more material we're using in the stackup. Of course. Yep. So let's try it. First, let's try to see, you know, what we can do with the cores. When we had them up to 18 mils before, you know, the, the traces were kind of too wide. So we're going to leave these three sheets of thick prepreg in here and I'm going to start swapping out. Um, probably I'm going to move to, uh, an eight or a 10 mil core. And I'm going to see where that gets us. If you haven't noticed, um, much of this is, uh, trial and error going back and forth and back and forth. And, but I suspect like after doing it for 10 years, you kind of know what, what you're going to do to reach that target, like you're at least. Not totally guessing. You, you, you kind of have a, a hunch a lot of times. Yeah, you know, uh, it is a lot easier than it used to be when I first started out. Um, you know, I, a few things I mentioned, you know, I started 5 mils here and there just because it's a good starting place for me. Um, as you get more experience based on, you know, your applications. You'll know what numbers you need to start out at. If you're doing applications that require, you know, some heavier copper, sometimes you might need a little thicker traces, a little wider traces. Or if you're doing low power and high speed, you might need to start out at three mils. Really, just Comes down to gaining that experience and finding out what works for you. In this case, I'm starting now. I'm, I've been giving stack ups to the PCB community for a while. So 5 mils seems to be a good place for me to start. And sorry, if I'm rambling on about that a little bit too much. So where were we? Okay. We were setting up. The dielectrics here. So we've gone to, uh, an 8 mil core and we're getting there, uh, but we're not quite there. Yep. So I turned off the mirror. So now I got to swap out two at a time because I can't turn the mirror back on. Okay. Um, I'm going to go with, I'm thinking about 10, but I think I'm just going to shoot for 12 and see what we get with 12. Okay. So yeah, the impedance traces are crazy high and we're still not there on the thickness. Well, you also haven't changed, uh, the core between four and five. Yeah. Well, right now, um, I just meant from, from a thickness calculation. I want to, yeah, we haven't, let's do that. I gotta find out, I gotta get this thing. Uh, back in shape so I can turn the mirror back on. Okay. There we go. So now let's get up. I like their icon for the mirror, the little DaVinci, uh, Yeah, that's good. They, they have some, some funny little icons here. Uh, you know, what we're doing right now is we're setting it up, uh, the traces for layers one and three, but when we're done, we're going to hit this, uh, mirror build icon. And it's going to actually just copy everything we've done over to the other traces. Gotcha. Uh huh. I didn't want to mention that because it's... Specific to Polar software. It's a really nice feature, but not everybody has it. Sure. Um, so we've gone back. Uh, So we're at 74 ohms. Uh, our thickness still isn't quite there. So I'm just going to keep going up. And it looks like we're getting back to the, the 17 mil or the 18 mil, um, cores that we were having before, that we had before in this case. I think if we try a 14 mil core, it would just about get us there. Not quite. So we'll keep swapping. Um, looks like we might be back up to the 18 mil. We'll try 16 mil core. Off a glitz. Getting closer and closer. The only reason I'm not choosing thicker prepregs than a 2116, we do have that option, but. Uh, the 106, 2116 are the most common, uh, off the shelf options that factories have. Yeah, so you want to stick to your bread and butter. Yeah, for sure. Um, we're just going to go back to the 18 mil cores. This little red light up here, uh, is so distracting. It's so annoying. Okay, so we're, we're pretty much there. We're 63 mils over copper, 64 mils over solder mask. Um, so we're going to leave it there and try this out. We're 82 ohms. We need 50. Um, so we're gonna have to, uh, to get there by increasing. Um, and since we have to go so far, I'm gonna... Jump ahead. Uh, okay. Okay. So we still, yeah, we're getting some fat traces there. But sometimes that's what you have to do. And so I think I, you know, the lesson I take from this is that when you're designing your PCB. You want to leave some room around your impedance control traces because they, they may have to get porky. Yeah, you do want to leave, uh, a little room, uh, not only for that, but, you know, the factory, when they get your stack up, um, And your impedance requirements, they're going to tool up your board for their machines. And so they're going to actually have to adjust them a little bit too. It's, it's nothing out of the ordinary. And, uh, is it, is it normal for like, after they make all their adjustments for you to go back to them and say, Hey, look, in order to meet these requirements, we had to change your stack up, you know, to this and get approval before you go to, to, to fabricate. Sometimes that does happen. Um, I wouldn't exactly call it normal, but it does happen quite a bit. Um, so when the factory has the stack up in it, you know, they can change the dielectrics by a plus or minus 20%. And I, you know, I wouldn't say they can't change it. Uh, let's, let's rephrase it to what it really is. It's a manufacturing tolerance. Yeah. Well, what I mean is not, not so much at the factory. What I mean is like when you're done with your work here and you said, Hey, look, I had to change your stack up to all this before I send this to fab. Are you okay with these changes I've made? Most certainly. Yeah. Um, if I have to make a drastic change, then I'm going to always contact the customer. And most of the time they're coming to me because they can't do it and they're okay with it. Yep. Anyways, but yeah, you know, there are some times when the customer says, I can't do that because I don't have the space for that. Yep. And sometimes we can't do it because. You can't, uh, you know, the trace width needs to be three mils, but the finished copper is two ounces and we can't, uh, do a three mil trace on two ounce copper. Gotcha. Um, so there's so many, like so many intricacies and so much minutiae that can go on here. It's crazy. The more I learned about PCBs, the more I realized I really don't know much about PCBs. Me too. Um, yeah, for sure. And now I guess, um, Like, so your, your process then is to just kind of take that text file or fab, fab note and, uh, keep playing with these numbers, keep playing with the stack up until you find the right cocktail that gets the job done for the customer and you meet all the target impedances, you meet the PCB thickness, you meet the manufacturability requirements because That's a factor too, as you mentioned earlier. You can't go with four layers of prepreg between these two cores, otherwise you're going to have registration issues. So, you've got to take all of these various things into account in order to get this, um, you know, to meet this customer's requirements. What this is honestly telling me, you know, in a big picture standpoint, Ladies and gentlemen, do not try this at home. Like, this is what I'm getting out of this. You know, like this is something where, you know, you, you, you know, the requirements for your design, the EE has, has laid out the requirements for the design. They know the chips they're using. They know the connectors they're using. They know the RF, you know, um, blah, blah, blah, blah, blah, blah, blah. The, when it comes time to the, you know, the actual manufacturability of it, like how exactly we're going to make this product for you, it's, it's kind of so much better to leave it up to an expert who lives this each and every day, because there's so many factors that go into it. That the, the EE who messes with this once every six months is just not, he's, he's not going to be able to remember all this the way that, that you are. Plus you're using, you know, this software that's probably thousands of dollars and he's going to use, you know. Probably not that, you know? Yeah. You know, uh, you know, this software is expensive and, and it has a lot of bells and whistles. Um, when I started out in the factory, I started with a software that was not anywhere near as good as this. Um, and actually, you know, if this is something that interests you, uh, and you want to learn more about it, I recommend get some free software, um, just to play around with. Grab a datasheet for, uh, you know, a common material and just start playing around with it. Yeah. But yeah, if you have a relationship with your factory, um, you should be able to contact the engineers at the factory and say, Hey, I need to stack up with impedance traces and here's my requirements. Um, if you're a paying customer, I can't imagine them saying no. Yeah, exactly. Yeah, even if you, you know, it's just for the next board that, you know, we don't have a PO for yet, but it's coming down the line. Uh, when I worked at the factory, I would put those customers, um, priority, you know, like any factory does. Um, if you're not a paying customer, uh, You still may get some luck there, but I wouldn't expect him to get back with you in the next day or so. Yeah, yeah, yeah, yeah, yeah, but still, like, I, I think, I, I guess my, my point is though, um, you know, it's helpful to, I think, see, the, the, the whole point of this show, and, and I, I'm going to sound like a broken record here for, for repeat listeners, um, we're trying to peel back the curtain in the manufacturing process. And this is one of those examples where, uh, it's, it's, it's pretty obvious to me looking at what Ryan's demonstrated here that, you know, this is really like with a, like, okay, a good analogy for longtime listeners. This is like stencil design. Like, do not try to design your own stencils. Like, your assembly house is so much better at this, and they just need you to do the generic stencil file, and we'll take it from there. Because it's so complicated to nail the stencil exactly right. Um, and I feel like it's similar with, uh, impedance design. It's so complicated to get it exactly right. And, and Ryan knows his factories too, right? He knows, Oh, you know, we plan on sending this to our shop in Florida, whatever, you know, and in Florida, I know that they stock these materials there. And I know that these are the right. This is going to be the stack up that's going to work well for them. And things are going to go smooth as opposed to, uh, I'm going to work with my fab in France. And, you know, I know over in Europe, they don't use as much isola. I don't know. Maybe they do. I'm just throwing that out. No, actually you're, uh, in some foreign countries, they don't use as much isola. You're right. Yeah. So, um, I just see this as an example of where, you know, just having this knowledge and just understanding. You know, what is taking place can help you to become a better engineer because you're just going to make better decisions about, you know, how to go about getting your product made, you know, ladies and gentlemen, do not try this at home. Yeah. And here's the best thing about having that, you know, develop that relationship with the factory engineers and get that stack up from them is when you give them your data with the PO and they tool it up. Uh, they already know that that stackup and those traces are gonna work. They've already dialed it in. They don't have to dial in anything else. As opposed to, you know, if we come to, uh, stackup software and produce it ourself, they're gonna go through it and check it and then they're gonna dial it in. No, we already got that done. Gotcha. Yeah, so smart. Brilliant. Yeah. Um, you know, uh, uh, this, this is, this has been a long episode so far. We, we're not afraid of long episodes, don't get me wrong. We love to go deep and, and I think we have here so far. Um, but, uh, before we... Before we put, uh, poor Melissa under so much effort to do all this editing. Yeah. You know, um, is there, is there any other further like, uh, specific things? Like, I think we've seen a good demonstration of what it takes to play with your stack up, how to, how to, you know, the impact that the plane has on, on the, uh, impedance control traces. The distance between the two, the materials. And, and, and I was surprised to learn how the, the pre preg material and the distances at the core and everything plays such a huge factor in how all this is done. Um, uh, that was all just mind blowing to me, really. I have no clue about any of this stuff. Um, but before we get into, which is, you know, my. The whole reason I do this podcast, the pet peeve of the week. Is there, is there any further like, uh, like, um, you know, details you wanted to go into or, or would you just say, you know, Reach out to Ryan for further, uh, further questions. Yeah. You know, people can always reach out to me, um, for further questions. Uh, Ryan Miller, ryan. miller at, uh, ncapgroup. com. But, you know, if I had one piece of advice to give, I would say, this is a good start. We can get. Much more complex. You know, we didn't even get into the co to the, uh, coplanar waveguide traces. Um, we didn't get into H D I structures and oh gosh, you know, you know, you got another variable. Um, man, we can, we can dive so much deeper, probably enough for another video podcast later. Sure, yeah, sure. I'd imagine. Yeah, if anybody has any questions, um, please feel free to contact me or, you know, you can always contact Chris. He can get you in touch with me. Yep, absolutely. Yeah, yeah. Yeah, that's that. I would call it here. I'd say that's a really good start for basic stack up design. And, and for me, that is As a person who loves to know details, I, I, I have scratched an itch that I didn't realize I had to understand, to understand this impedance control. It's, it's fascinating. I know I'm never going to need to use this, right? I just, in my role and what I do and, and, you know, operations of a factory, I know this is not something that I'm going to. Need to get involved with, but I know my customers do, and I'm happy to be able to understand a little bit more about it. And so when they ask me, Hey, Chris, what's the standard stack up of your fab? Now I know why they're asking, because they're trying to match their, they're trying to match their design to that standard stack up to make sure that they have enough room around their traces. So that when the fab, uh, calculates the impedance. They have the room they need to get that, to get that done. And they know, you know, everything's going to work together. That's really cool. It really answers a lot of questions for me. Yeah, I, I'm happy that you guys had me back. Um, like I said, it's always wonderful to get invited back. Yes. Well, I'll tell you what, Ryan, uh, if, if you want to make sure you get invited back a third time, you better hit us with the best pet peeve of the week. You could possibly think of no, just kidding. No pressure. You just, you just lay it on us though. It's I'm, I'm ready for it. I am here for it. I'm primed for it. Pet peeve of the week. You know, I was in traffic the other day. I live in the Atlanta area. And if you know, the Atlanta area, you know, the traffic is terrible. I've heard all the time, but, uh, Rush hour times. It's even more worse. Yeah. So, if you're at a red light and you're turning right, and you get a chance to go, you better go. You better. I agree. So, my pet peeve is that person at the red light that I'm behind that won't turn right. Oh, that is my mother. That is everybody who's ever been in front of me, ever. I feel like... It's like, you, there is no, now, now I will say in all fairness, look for the sign that says no right turn on red. And if you see one. We do have a lot of those in Atlanta. Yeah. Cause that means you've got a blind corner and you're putting somebody in danger if you're to turn right on red. Uh, but if that doesn't exist, you better get moving. There ain't no traffic coming. Let's get out of the way, buddy. Yeah. Or people get so angry sometimes. I was going to say, or, uh, if it's like a. You have to yield to turn and you don't pull into the intersection. Yeah. Oh, that's the worst. And they stay behind the line. Yeah. Cause it turns yellow and you're like, come on, go. I used to live in LA a long time ago. And yeah, you, you would not even get to turn ever if you did that. That's right. I will say, um, I'm a, I'm a big, big fan of rotaries. I, they, there is a, there is this infamous, um, intersection, uh, here in western Massachusetts that leads to, uh, uh, three different towns off of a major highway, Highway 91. And this intersection, if you get off Highway 91, it leads either to Northampton, or Hadley, or Amherst, and um, there's a lot of businesses, there's a lot of universities, there's a, there's just so much going on in these towns. And everybody's taking 91 to get there, because there's no other great road to get there. Now, there used to be a stoplight there, and like you're saying, Ryan, So many people would not turn right on red, and it got so bad, traffic would back up into the highway, a 65 mile an hour highway. bad. Remember? Yeah, I remember. And it would back up like a mile, I'm telling you, it would, it was crazy. They made one change to this. The only change they made is they turned it into a rotary. Now I said it's one change, it's a big change, I get it. But they changed it instead of a stoplight into a rotary, never any backup on the highway anymore. Ever. There's never a backup. Nice. I know. And... So I'm a big fan of the rotary. It just keeps everything moving. Everything just keeps flowing. And I feel like they're much safer because you have to slow down and you have to go through the intersection slowly. So if you do have an accident, you're only having an accident at 15, 20 miles an hour, as opposed to you're coming off a highway at 65 miles an hour. And you're blasting through that intersection at 45, 50 miles an hour still. And somebody else isn't hitting the brakes, you know, and you're, you know, you go through the red light or something, boom, you're, you're hitting somebody at really high speed. So I think the rotary is safer too. That's, we could talk infrastructure all day this day, this, you know, this weekend infrastructure will be our new podcast, Melissa, what do you think? I got scared by those, um. I took a trip to Ireland for a week once and well you're on the wrong side of the road first of all yeah first yeah that was my first goal is to learn how to do that but coming out of the rental car place immediately there was one there and then down the road two more and They have them a lot in that country. Yes. They're very popular in Europe. So of in this circle and I'm scratching my head going, man, how do I get in and get out? But there were no accidents, you just use your turn signal, people let you in, you get to where you're supposed to go and people let you out and it was straining. Yep. I highly recommend, uh, putting one of those, um, Plastic sticks that goes between your window and the frame of the frame of the vehicle, uh, that has the American flag on the top of it. Just so everybody's aware, here in Out of Towner, steer clear. Yeah, and In Ireland, yeah, you'd probably, people would be like, okay, yeah, just stupid Americans. Get out of this guy's way. It's funny though, we've had some folks from, uh, we have folks that live in the UK that come to visit. And, uh, sometimes they'll drive instead of us. And, uh, you see them as they're approaching an intersection. Their brain is working overtime trying to figure out are they going to make the turn into this intersection on the right side of the road. Well, yeah, yeah, because I used to spend a lot of time in South Africa and I didn't even drive there. But yeah, coming back, um, whenever I would encounter a rotary and have to like remind myself to go the correct way around it because my brain just was like, oh, yes, go left. Like, no, don't go left. Yes. Yes. That would be very bad. That's a good one, Ryan. Man, man, that's a good one. Now, I will say I grew up in New Jersey where that's where I started to learn how to drive and you want to talk about traffic. It's, it's up there with Atlanta and, um, oh, the Jersey Turnpike, the Garden State Parkway. These highways are infamous. Like everybody knows them for, for being traffic y. It's like, It's like, uh, is it I 5 in L. A.? There's some famous highway in L. A. I think it's I 5 that everybody talks about as being super traffic y. And that's, that's the Jersey Turnpike and the Garden State Parkway for us. And, um... And, and so traffic, like, it's funny when I hear about people that I live with out here in Western Massachusetts now that talk about traffic, I'm like, you, you have no idea. You have no idea. Yeah. Like a three minute delay in your, in your day for traffic is nothing. We're talking like an hour most of the time. Three minute delay can cost you an hour. I used to leave at work, uh, when I, when I did work. Uh, at the factory here. I would leave five minutes after five because if I left right at five, I would be in traffic for an extra like 45 minutes. It's crazy how it works. I believe it. That's a good one, Ryan. Um, well let's, let's pause it there. Um, I'm certain there's, there's more, uh, more we could talk about. We could probably go on and on and on, but I think, I think we're at a good stopping point now. Um, as always, we do like to remind our listeners that, uh, uh, you can get in touch with us, uh, probably the easiest way is just contact at pickplacepodcast. com. Many of you do use that channel. Great. Continue to do that. Um, I say tweeted us, maybe you're supposed to X us now, I'm not sure how, what the verbiage is at this point, post at us, something, send us a message, uh, at CircuitHub or at WAssembly, um, you know, so long as it's named X, uh, you know, until it's named Y next week, I'm sure it'll just... I'm getting into my grumpy old, grumpy old man phase of my life at this point. So, but, uh, if you've enjoyed this discussion and, and you thought this was helpful, please share it with somebody, uh, Oh, Hey, it's YouTube. So I can say smash that like button. Oh my God. Smash it. Don't subscribe because I don't, I don't know how often we're going to do these sorts of things. We do have plans for, for doing more videos. We're going to. Big plans ahead, depending on how well this goes, we're going to do some more stuff with walking around Worthington Assembly and showing how different processes are done. Um, so maybe do subscribe, who knows? Uh, but either way, listen to, uh, listen to the podcast. That's, that's going to continue to be the primary means that we distribute information and, uh, keep sending us, uh, some show suggestions. We have a few queued up as we keep mentioning and we keep not getting to, but I promise we will, we will get to them. I promise. Yeah. And if you have any, um. Demos like this that you would be interested in having us do? Yeah, definitely let us know about that as well. Absolutely. Yeah, if you wanna see more detail on, oh gosh, I, I get so many, I get my, I, I get nervous just, just thinking about the demos we could do and, oh, there's so much. Yeah. Yeah. So much. Alright, thanks for listening to The Big Place podcast if you like what you've heard or podcast. Thanks everybody. Thanks Melissa. Thanks, Ryan. Oh, and thanks little kitty. Oh, can you see her? Yeah. That's Charlie again. Hi Charlie. Very good. Thanks everybody.