WEBVTT 00:00:00.000 --> 00:00:03.060 Okay, so let's just unpack this. Look around 00:00:03.060 --> 00:00:05.459 you right now. Literally anywhere. Yeah, right. 00:00:05.540 --> 00:00:07.320 Whether you're listening to this on a phone or 00:00:07.320 --> 00:00:10.939 maybe a laptop or you're in your car, inside 00:00:10.939 --> 00:00:15.199 every single one of those devices is a hidden 00:00:15.199 --> 00:00:17.480 city. The literal city. You've got your highways, 00:00:17.660 --> 00:00:19.660 your power plants, your residential districts. 00:00:19.859 --> 00:00:22.059 Exactly. And we usually just ignore it. I mean, 00:00:22.079 --> 00:00:24.379 we treat our devices like these magic black boxes. 00:00:24.420 --> 00:00:26.359 You press a button and a thing happens. Right. 00:00:26.640 --> 00:00:29.420 But if you were to crack that box open, which... 00:00:30.079 --> 00:00:32.340 Honestly, most of us are way too scared to do 00:00:32.340 --> 00:00:35.359 because of the warranty. And the glue. So much 00:00:35.359 --> 00:00:38.100 glue these days. Right. So much glue. But if 00:00:38.100 --> 00:00:39.920 you did, you'd find the fundamental infrastructure 00:00:39.920 --> 00:00:44.840 of the modern world. Today for our deep dive, 00:00:44.920 --> 00:00:46.640 we were talking about the printed circuit board 00:00:46.640 --> 00:00:50.759 or the PCB. It really is the nervous system of 00:00:50.759 --> 00:00:52.719 everything we own. But you're right. I mean, 00:00:52.740 --> 00:00:54.420 most people just see a green square with some 00:00:54.420 --> 00:00:55.960 black chips on it and they don't think twice. 00:00:56.259 --> 00:00:59.109 And that's our mission today. We are pulling 00:00:59.109 --> 00:01:01.530 from a massive stack of source material covering 00:01:01.530 --> 00:01:04.230 the history, the composition manufacturing, and 00:01:04.230 --> 00:01:08.129 even future tech of PCBs. We want to move beyond 00:01:08.129 --> 00:01:11.030 that whole green square with chips idea. Definitely. 00:01:11.250 --> 00:01:13.989 I want to understand how we went from literally 00:01:13.989 --> 00:01:17.629 hand wiring radios to these insane microscopic 00:01:17.629 --> 00:01:20.829 metropolises. Because when you look at the research, 00:01:21.090 --> 00:01:24.489 the PCB isn't just a holder for parts. It's a 00:01:24.489 --> 00:01:27.719 map of engineering history. It is. And what's 00:01:27.719 --> 00:01:29.680 fascinating here is that the evolution of the 00:01:29.680 --> 00:01:32.000 PCB dictates the evolution of the device itself. 00:01:32.200 --> 00:01:34.500 You literally cannot have a smartphone without 00:01:34.500 --> 00:01:36.739 the specific manufacturing breakthroughs we're 00:01:36.739 --> 00:01:38.280 going to talk about today. It's really a story 00:01:38.280 --> 00:01:39.959 about density. So let's start with the before 00:01:39.959 --> 00:01:42.109 picture. Yeah. Because I think... To appreciate 00:01:42.109 --> 00:01:44.569 the sleek green board, you kind of have to understand 00:01:44.569 --> 00:01:47.189 the absolute nightmare that came before it. Nightmare 00:01:47.189 --> 00:01:49.069 is definitely the correct term. If you go back 00:01:49.069 --> 00:01:51.409 to the early 20th century, so say the 1920s or 00:01:51.409 --> 00:01:53.930 30s, electronics were built using what's called 00:01:53.930 --> 00:01:56.150 point -to -point wiring. The rat's nest. Precisely. 00:01:56.150 --> 00:01:59.829 The rat's nest. Imagine a heavy metal chassis, 00:02:00.049 --> 00:02:02.590 so basically just a steel frame. You'd screw 00:02:02.590 --> 00:02:05.939 these big... bulky vacuum tubes and transformers 00:02:05.939 --> 00:02:08.000 onto it. And then to connect them, you'd have 00:02:08.000 --> 00:02:10.580 a human being sitting there with a soldering 00:02:10.580 --> 00:02:13.960 iron and a spool of copper wire, physically soldering 00:02:13.960 --> 00:02:16.900 a connection from point A to point B. I've actually 00:02:16.900 --> 00:02:20.120 seen photos of the insides of old guitar amplifiers 00:02:20.120 --> 00:02:22.840 from the 50s that still use this. It looks like 00:02:22.840 --> 00:02:25.500 a copper spaghetti explosion. It does. It's beautiful 00:02:25.500 --> 00:02:28.219 in a chaotic way, but from an engineering standpoint, 00:02:28.500 --> 00:02:30.789 it was terrible. First off, it was heavy, wood 00:02:30.789 --> 00:02:33.310 bottoms, metal frames. Second, it was incredibly 00:02:33.310 --> 00:02:35.770 labor intensive. You needed highly skilled craftsmen 00:02:35.770 --> 00:02:38.349 to build a single radio. But the absolute biggest 00:02:38.349 --> 00:02:40.490 problem was reliability. Well, yeah, because 00:02:40.490 --> 00:02:42.770 if you shake a box full of loose wires, something's 00:02:42.770 --> 00:02:44.270 going to touch something it definitely shouldn't 00:02:44.270 --> 00:02:46.870 touch. Exactly. It was fragile. A short circuit 00:02:46.870 --> 00:02:49.289 wasn't just a metaphor back then. It was a constant 00:02:49.289 --> 00:02:52.430 daily risk. If two wires vibrated against each 00:02:52.430 --> 00:02:54.270 other and the insulation wore off, the device 00:02:54.270 --> 00:02:57.680 died. Or, you know, caught fire. So was the move 00:02:57.680 --> 00:02:59.939 to the circuit board just about making things 00:02:59.939 --> 00:03:02.639 tidier? Or was there a specific catalyst that 00:03:02.639 --> 00:03:05.180 pushed it? Well, there were some early pioneers 00:03:05.180 --> 00:03:08.360 playing with the idea. A German inventor named 00:03:08.360 --> 00:03:10.979 Albert Hansen actually had a patent way back 00:03:10.979 --> 00:03:13.740 in 1903 for flat foil conductors on a board. 00:03:14.099 --> 00:03:18.060 1903? Yeah, very early. And then around 1936, 00:03:18.280 --> 00:03:21.240 there was Paul Eisler, an Austrian engineer working 00:03:21.240 --> 00:03:23.939 in the UK, who actually put a printed circuit 00:03:23.939 --> 00:03:27.259 into a radio set. But the manufacturing tech 00:03:27.259 --> 00:03:29.360 just wasn't there to make it cheap at scale. 00:03:29.719 --> 00:03:32.419 The real force multiplier, the thing that forced 00:03:32.419 --> 00:03:34.909 the... entire industry to change was World War 00:03:34.909 --> 00:03:37.370 II. It always seems to come back to the war effort 00:03:37.370 --> 00:03:39.590 in these tech deep dives. It usually does, yeah. 00:03:40.090 --> 00:03:42.469 Specifically, the U .S. Army had a massive problem. 00:03:42.569 --> 00:03:44.349 They needed something called the proximity fuse. 00:03:44.669 --> 00:03:46.310 Okay. I was reading about this in the source 00:03:46.310 --> 00:03:47.909 material. This is for anti -aircraft shells, 00:03:48.090 --> 00:03:50.669 right? Yes. The goal was to shoot down planes. 00:03:51.479 --> 00:03:53.659 But hitting a moving plane with a dumb bullet 00:03:53.659 --> 00:03:56.560 in the sky is really, really hard. So they needed 00:03:56.560 --> 00:03:59.199 a shell that had a tiny radar set inside it so 00:03:59.199 --> 00:04:01.219 it would detonate when it just got near the plane. 00:04:01.479 --> 00:04:04.360 But think about the physics of that. You are 00:04:04.360 --> 00:04:07.280 taking a delicate radio, putting it inside a 00:04:07.280 --> 00:04:09.719 bullet, and then literally firing it out of a 00:04:09.719 --> 00:04:12.219 cannon. The g -forces are immense. We're talking 00:04:12.219 --> 00:04:15.439 thousands of g's of acceleration in a fraction 00:04:15.439 --> 00:04:17.579 of a second. Right. So if you use the old rat's 00:04:17.579 --> 00:04:20.569 nest wiring inside a cannon shell. the shock 00:04:20.569 --> 00:04:22.910 of the launch would just snap every single connection 00:04:22.910 --> 00:04:25.410 instantly. It would be debris before it even 00:04:25.410 --> 00:04:28.050 left the barrel. Exactly. So the Army worked 00:04:28.050 --> 00:04:29.910 with the Central Lab Division of Globe Union 00:04:29.910 --> 00:04:32.389 to find a way to make electronics that were completely 00:04:32.389 --> 00:04:35.329 solid. And they came up with a ceramic plate. 00:04:35.810 --> 00:04:37.949 Instead of copper wires floating in the air, 00:04:38.089 --> 00:04:41.129 they screen printed metallic paint directly onto 00:04:41.129 --> 00:04:42.829 the ceramic. So they were literally printing 00:04:42.829 --> 00:04:45.649 the wires. Yes, printing them. They used a silver 00:04:45.649 --> 00:04:48.069 paste for the conductors and a carbon ink for 00:04:48.069 --> 00:04:51.089 the resistors. was effectively the first mass 00:04:51.089 --> 00:04:53.569 -produced printed circuit. It was rugged. It 00:04:53.569 --> 00:04:55.930 was solid state. And it could actually survive 00:04:55.930 --> 00:04:58.149 being fired from a gun. And this wasn't just 00:04:58.149 --> 00:05:00.610 some cool new gadget. It was classified top secret. 00:05:00.810 --> 00:05:03.769 Highly classified military survival tech. The 00:05:03.769 --> 00:05:06.029 U .S. didn't release this technology for commercial 00:05:06.029 --> 00:05:09.990 use until 1948. But once they did, it triggered 00:05:09.990 --> 00:05:13.350 an absolute revolution. By the mid -50s, companies 00:05:13.350 --> 00:05:16.449 like Motorola were dumping millions into these 00:05:16.449 --> 00:05:19.449 plated circuits for consumer radios. They realized 00:05:19.449 --> 00:05:21.629 they could make them lighter, cheaper, and way 00:05:21.629 --> 00:05:23.990 more reliable. Okay, so that gets us the concept. 00:05:24.430 --> 00:05:26.769 But let's talk about the reality of the board 00:05:26.769 --> 00:05:29.029 itself today. Because when I look at a motherboard 00:05:29.029 --> 00:05:32.230 now, or even just the board in my toaster, I'm 00:05:32.230 --> 00:05:34.529 not seeing ceramic and silver paint. I'm seeing 00:05:34.529 --> 00:05:37.209 that iconic green stuff. Right, the modern PCB. 00:05:37.290 --> 00:05:39.610 It's quite a bit different from the ceramic fuse. 00:05:41.550 --> 00:05:44.430 Not as a single object, but as a laminate. It's 00:05:44.430 --> 00:05:46.529 a sandwich. I always appreciate a good food analogy. 00:05:46.850 --> 00:05:48.589 Walk us through the layers of this sandwich. 00:05:48.810 --> 00:05:51.769 Okay, so the bread of the sandwich is the substrate. 00:05:52.069 --> 00:05:54.870 The industry standard is a material called FR4. 00:05:55.110 --> 00:05:57.949 Which the notes say is woven glass epoxy. Correct. 00:05:58.009 --> 00:06:01.129 The FR actually stands for flame retardant. It's 00:06:01.129 --> 00:06:03.410 basically fiberglass mixed with a really tough 00:06:03.410 --> 00:06:05.810 resin. It provides a mechanical spine of the 00:06:05.810 --> 00:06:08.509 board. It's stiff. It doesn't conduct electricity. 00:06:08.910 --> 00:06:11.439 And crucially, it doesn't... melt when you get 00:06:11.439 --> 00:06:14.000 it extremely hot okay so that's the rigid structure 00:06:14.000 --> 00:06:16.819 then you have the copper the meat of the sandwich 00:06:16.819 --> 00:06:20.279 you laminate a thin sheet of copper foil onto 00:06:20.279 --> 00:06:23.000 that fiberglass but here is the part that people 00:06:23.000 --> 00:06:24.920 often misunderstand and it's really the core 00:06:24.920 --> 00:06:27.879 of the whole manufacturing magic we don't add 00:06:27.879 --> 00:06:31.459 the wires to the board We use subtractive manufacturing. 00:06:31.800 --> 00:06:33.480 Right. I really wanted to clarify this point. 00:06:33.600 --> 00:06:35.720 All right. Because the sources emphasize this. 00:06:35.779 --> 00:06:39.000 It's like a sculptor chipping away stone to reveal 00:06:39.000 --> 00:06:42.300 a statue rather than a painter adding paint to 00:06:42.300 --> 00:06:44.480 a blank canvas. That is the perfect analogy. 00:06:44.660 --> 00:06:47.060 You start with a board that is completely covered 00:06:47.060 --> 00:06:50.000 in a solid sheet of copper. Then you use a process 00:06:50.000 --> 00:06:52.480 called photolithography. Which sounds a lot like 00:06:52.480 --> 00:06:54.949 photography. It's chemically very, very similar. 00:06:55.009 --> 00:06:56.689 You cover the copper with a light -sensitive 00:06:56.689 --> 00:06:59.149 chemical called photoresist. Then you take a 00:06:59.149 --> 00:07:01.329 mask, which is basically a negative image of 00:07:01.329 --> 00:07:03.610 your circuit traces, and you blast the whole 00:07:03.610 --> 00:07:06.470 thing with UV light. The light hardens the chemical 00:07:06.470 --> 00:07:08.189 over the parts of the copper you want to keep. 00:07:08.370 --> 00:07:10.649 So the artwork or the blueprint protects the 00:07:10.649 --> 00:07:12.949 copper lines you need, and the rest of it remains 00:07:12.949 --> 00:07:16.089 soft. Exactly. Then you dunk the whole board 00:07:16.089 --> 00:07:19.269 into an acid bath. The chemical etching dissolves 00:07:19.269 --> 00:07:21.209 all the exposed copper, the stuff you didn't 00:07:21.209 --> 00:07:24.350 protect. When you wash it all off, the only thing 00:07:24.350 --> 00:07:27.029 left is that intricate web of copper highways. 00:07:27.410 --> 00:07:29.850 That perfectly explains why the lines on a circuit 00:07:29.850 --> 00:07:33.329 board are so incredibly crisp. It's chemical 00:07:33.329 --> 00:07:35.230 esching. It's not someone trying to draw with 00:07:35.230 --> 00:07:38.009 a tiny pen. Right. It allows for precision down 00:07:38.009 --> 00:07:40.290 to the micrometer level. But then you have a 00:07:40.290 --> 00:07:43.540 new problem. Copper oxidizes. If you leave it 00:07:43.540 --> 00:07:46.459 exposed to the air, it turns green and corrodes. 00:07:46.720 --> 00:07:49.180 Plus, if you try to solder components to it later, 00:07:49.360 --> 00:07:51.899 the hot solder might just flow everywhere across 00:07:51.899 --> 00:07:54.139 the copper and connect things that definitely 00:07:54.139 --> 00:07:56.959 shouldn't connect. Solder shorts. Exactly. So 00:07:56.959 --> 00:07:59.389 you need a protective condiment. on your sandwich 00:07:59.389 --> 00:08:01.389 that's the solder mask and that is the green 00:08:01.389 --> 00:08:04.910 color yes it's a layer of polymer or epoxy that 00:08:04.910 --> 00:08:07.089 covers the entire board insulating all those 00:08:07.089 --> 00:08:10.129 tiny copper traces and leaving holes only where 00:08:10.129 --> 00:08:13.189 you explicitly need to solder a component is 00:08:13.189 --> 00:08:15.490 there a scientific reason it's green or is that 00:08:15.490 --> 00:08:17.949 just a tradition thing mostly historical actually 00:08:17.949 --> 00:08:20.810 the original two -part epoxy resins they developed 00:08:20.810 --> 00:08:24.240 were just naturally a brownish green color But 00:08:24.240 --> 00:08:26.720 it's stuck around because green happens to be 00:08:26.720 --> 00:08:29.660 very easy for human eyes to inspect. Our eyes 00:08:29.660 --> 00:08:32.059 are most sensitive to green light, so quality 00:08:32.059 --> 00:08:35.179 control workers could spot tiny defects way easier 00:08:35.179 --> 00:08:37.200 on a green board than, say, a red or blue one. 00:08:37.500 --> 00:08:39.519 Though I have seen other colors lately. Yeah. 00:08:39.559 --> 00:08:41.919 Like you see black boards in gaming PCs or purple 00:08:41.919 --> 00:08:44.100 boards from those custom prototype manufacturers. 00:08:44.100 --> 00:08:46.159 Oh, yeah. You can add any pigment you want now. 00:08:46.279 --> 00:08:48.519 But green is still by far the cheapest and the 00:08:48.519 --> 00:08:51.360 global standard. Got it. Then the final layer 00:08:51.360 --> 00:08:54.500 on the very top is the writing. The R1, the C3, 00:08:54.679 --> 00:08:57.360 the little brand logo. The silkscreen, it's literally 00:08:57.360 --> 00:09:00.100 just ink printed on top. It's the legend for 00:09:00.100 --> 00:09:02.139 the map telling the humans and the machines where 00:09:02.139 --> 00:09:04.519 each specific part is supposed to go. So we've 00:09:04.519 --> 00:09:07.639 got the complete sandwich. FR4, copper, solder, 00:09:07.820 --> 00:09:11.159 mask, silkscreen. But a bare board is just a 00:09:11.159 --> 00:09:12.620 board. It doesn't actually do anything yet. You 00:09:12.620 --> 00:09:14.620 have to populate it. And looking at the history 00:09:14.620 --> 00:09:16.740 we pulled, this seems to be where the second 00:09:16.740 --> 00:09:19.080 massive revolution happened. The shift from through 00:09:19.080 --> 00:09:21.529 hole to surface mount. Oh, this is the shift 00:09:21.529 --> 00:09:24.750 that truly gave us the modern digital age. For 00:09:24.750 --> 00:09:27.629 decades, through -hole was king. And it's exactly 00:09:27.629 --> 00:09:29.809 what it sounds like. You drill a physical hole 00:09:29.809 --> 00:09:32.690 all the way through the FR -4. The components 00:09:32.690 --> 00:09:35.289 say a standard resistor looks like a little pill 00:09:35.289 --> 00:09:38.649 with two long wire legs. I remember these from 00:09:38.649 --> 00:09:41.169 shop class back in the day. You stick the wire 00:09:41.169 --> 00:09:43.649 legs through the holes, you flip the board over 00:09:43.649 --> 00:09:45.570 and you solder them to the metal pad on the back. 00:09:45.809 --> 00:09:48.570 Then you take clippers and snip off the excess 00:09:48.570 --> 00:09:51.139 wire. Right. And it's mechanically very strong. 00:09:51.139 --> 00:09:53.139 If you're building a heavy guitar amp or a power 00:09:53.139 --> 00:09:54.799 supply that might get kicked around a stage, 00:09:55.019 --> 00:09:57.860 the Ruhold is fantastic. But think about the 00:09:57.860 --> 00:10:00.179 manufacturing constraints of that. You have to 00:10:00.179 --> 00:10:03.059 drill thousands of tiny holes. That eats up expensive 00:10:03.059 --> 00:10:05.460 drill bits. It creates a ton of fiberglass dust. 00:10:05.759 --> 00:10:08.440 It's just slow. And you're wasting so much space. 00:10:08.559 --> 00:10:10.259 If I have to poke a leg completely through the 00:10:10.259 --> 00:10:13.240 board, I can't put another chip on the backside 00:10:13.240 --> 00:10:15.620 of that exact same spot. I've used up that real 00:10:15.620 --> 00:10:17.659 estate on both sides of the board for one component. 00:10:17.799 --> 00:10:19.759 Exactly. You're completely blocked. Enter the 00:10:19.759 --> 00:10:23.120 1980s and surface mount technology or SMT. The 00:10:23.120 --> 00:10:25.649 legs are gone. Completely gone. Instead of long 00:10:25.649 --> 00:10:28.330 wires, SMT components just have tiny little metal 00:10:28.330 --> 00:10:30.909 tabs or caps on the ends. You don't drill any 00:10:30.909 --> 00:10:33.549 holes for them. You apply a precise paste of 00:10:33.549 --> 00:10:36.149 solder to the pads on the surface. A machine 00:10:36.149 --> 00:10:38.269 places the component right on top, kind of like 00:10:38.269 --> 00:10:40.470 laying down a tile on a floor. And then you run 00:10:40.470 --> 00:10:42.690 the whole board through a giant baking oven to 00:10:42.690 --> 00:10:45.029 melt the solder. This sounds like it completely 00:10:45.029 --> 00:10:47.649 changes the physical scale of everything. Drastically. 00:10:48.149 --> 00:10:51.230 SMT components are a fraction of the size, sometimes 00:10:51.230 --> 00:10:53.929 one -tenth the weight of their through -hole 00:10:53.929 --> 00:10:56.009 counterparts. And because you don't need to drill 00:10:56.009 --> 00:10:58.690 holes, you can suddenly populate both sides of 00:10:58.690 --> 00:11:00.330 the board independently. And this is the great 00:11:00.330 --> 00:11:02.289 enabler for mass automation, isn't it? Yeah. 00:11:02.309 --> 00:11:04.070 I've watched videos of those pick -and -place 00:11:04.070 --> 00:11:06.570 machines. They move so incredibly fast, they're 00:11:06.570 --> 00:11:10.570 just a blur. That's the real key to SMT. A human 00:11:10.570 --> 00:11:13.799 with a soldering iron. can maybe do a few hundred 00:11:13.799 --> 00:11:16.100 joints an hour if they're highly caffeinated 00:11:16.100 --> 00:11:18.899 and super fast. A modern pick -and -place line 00:11:18.899 --> 00:11:22.100 can handle tens of thousands of components per 00:11:22.100 --> 00:11:24.840 hour. The machine grabs a resistor, which is 00:11:24.840 --> 00:11:27.480 literally the size of a grain of sand, now checks 00:11:27.480 --> 00:11:29.500 its orientation with a high -speed camera and 00:11:29.500 --> 00:11:31.740 slaps it onto the board with perfect precision. 00:11:32.080 --> 00:11:34.120 Which is exactly why you can buy a smartphone 00:11:34.120 --> 00:11:37.279 for a few hundred dollars today. If we were still 00:11:37.279 --> 00:11:39.559 using through -hole technology, an iPhone would 00:11:39.559 --> 00:11:42.120 probably be the size of a briefcase. Oh, at least. 00:11:42.220 --> 00:11:44.980 And it would cost $10 ,000 to assemble by hand. 00:11:45.599 --> 00:11:47.659 SMT is what allowed us to make complex electronics 00:11:47.659 --> 00:11:50.240 cheap enough to be almost disposable. Speaking 00:11:50.240 --> 00:11:52.220 of smartphones, though, the boards inside them 00:11:52.220 --> 00:11:54.279 don't really look like the standard green rigid 00:11:54.279 --> 00:11:56.919 squares anymore. I opened up an old broken phone 00:11:56.919 --> 00:11:59.000 recently, and the board was, first of all, it 00:11:59.000 --> 00:12:01.399 was incredibly dense. But there were also these 00:12:01.399 --> 00:12:03.840 thin orange ribbon cables everywhere connecting 00:12:03.840 --> 00:12:06.240 things. It's the next evolution. We are moving 00:12:06.240 --> 00:12:08.559 beyond the rigid board. Those orange ribbons 00:12:08.559 --> 00:12:12.419 you saw are flexible PCBs or FPCBs. I've heard 00:12:12.419 --> 00:12:14.320 of Kapton tape before. Is that related to this? 00:12:14.360 --> 00:12:17.330 Yes. Captain is a brand name for polymide. That's 00:12:17.330 --> 00:12:20.509 the substrate they use. Instead of rigid FR4 00:12:20.509 --> 00:12:22.889 fiberglass, they use this high -performance plastic 00:12:22.889 --> 00:12:25.529 film. It's highly heat -resistant, so it can 00:12:25.529 --> 00:12:27.389 handle the soldering ovens, but it's completely 00:12:27.389 --> 00:12:29.850 flexible. Which allows you to essentially do 00:12:29.850 --> 00:12:32.490 origami with the electronics. Exactly. If you 00:12:32.490 --> 00:12:34.730 look at a modern folding phone, or even just 00:12:34.730 --> 00:12:37.210 how the tiny camera module connects to the main 00:12:37.210 --> 00:12:39.809 motherboard in a regular phone, you need that 00:12:39.809 --> 00:12:43.470 circuit to bend and fold. FPCBs allow the electronics 00:12:43.470 --> 00:12:46.299 to fold. flow into whatever complex 3D shape 00:12:46.299 --> 00:12:48.620 the enclosure demands. That makes sense. And 00:12:48.620 --> 00:12:50.480 what about the density aspect? Because you mentioned 00:12:50.480 --> 00:12:53.259 microovias in the research notes. This is HDI 00:12:53.259 --> 00:12:56.120 or high density interconnect. Remember how we 00:12:56.120 --> 00:12:58.200 said drilling holes mechanically was slow and 00:12:58.200 --> 00:13:00.340 took up too much space? Well, mechanical drill 00:13:00.340 --> 00:13:02.360 bits can only get so small before they just snap 00:13:02.360 --> 00:13:04.940 under the pressure. For modern processes like 00:13:04.940 --> 00:13:07.000 the core brain of your phone, we need connections 00:13:07.000 --> 00:13:09.120 between layers that are microscopic to handle 00:13:09.120 --> 00:13:11.500 the sheer number of signals and the raw speed. 00:13:11.899 --> 00:13:13.399 So how do you drill a hole that... that small 00:13:13.399 --> 00:13:16.519 if the bits keep breaking. Lasers. Lasers, of 00:13:16.519 --> 00:13:19.700 course. We use lasers to blast microscopic holes 00:13:19.700 --> 00:13:21.820 called microvias to connect the different layers 00:13:21.820 --> 00:13:24.200 of copper. And that's another thing. Your phone's 00:13:24.200 --> 00:13:26.200 motherboard isn't just one layer of copper on 00:13:26.200 --> 00:13:28.440 top and one on the bottom. It might be 10 or 00:13:28.440 --> 00:13:30.919 12 independent layers of copper stacked up inside 00:13:30.919 --> 00:13:33.519 that thin board like a lasagna, all connected 00:13:33.519 --> 00:13:36.120 by these microscopic laser tunnels. It's just 00:13:36.120 --> 00:13:39.340 incredible engineering. The level of precision 00:13:39.340 --> 00:13:42.299 required to align 12 layers of copper at a micro... 00:13:42.320 --> 00:13:45.059 Microscopic scale is mind -boggling. And the 00:13:45.059 --> 00:13:47.240 sources also briefly mentioned metal core boards, 00:13:47.360 --> 00:13:50.360 too. Yes, specialty boards. For things like those 00:13:50.360 --> 00:13:53.200 super bright LED lights, we use metal core boards 00:13:53.200 --> 00:13:55.279 with an aluminum or copper backing instead of 00:13:55.279 --> 00:13:58.679 fiberglass. FR4 is a terrible conductor of heat, 00:13:58.720 --> 00:14:00.740 so if you put a high -power LED on it, it would 00:14:00.740 --> 00:14:03.259 burn itself out. The metal core acts as a giant 00:14:03.259 --> 00:14:06.059 heat sink to pull the heat away. Ah, okay. So 00:14:06.059 --> 00:14:08.840 we've got all these amazing advancements, HDI 00:14:08.840 --> 00:14:12.980 flex boards, metal cores. But and here is where 00:14:12.980 --> 00:14:14.879 the conversation kind of takes a turn and the 00:14:14.879 --> 00:14:16.820 source is this incredible complexity comes with 00:14:16.820 --> 00:14:18.899 the cost. And I don't just mean the literal price 00:14:18.899 --> 00:14:20.799 tag. You're talking about the environmental and 00:14:20.799 --> 00:14:23.519 repairability issues. Right. When I had that 00:14:23.519 --> 00:14:26.279 old rat's nest guitar amp, if a capacitor blew, 00:14:26.539 --> 00:14:29.279 I could actually see it. It would be physically 00:14:29.279 --> 00:14:32.419 leaking or charred black. I could grab a soldering 00:14:32.419 --> 00:14:34.799 iron, snip it out and put a brand new one in 00:14:34.799 --> 00:14:37.740 for 50 cents. Correct. It was totally serviceable 00:14:37.740 --> 00:14:40.889 by an end user. But with an HDI multilayered 00:14:40.889 --> 00:14:43.850 surface mount board, if one single microscopic 00:14:43.850 --> 00:14:46.909 component fails out of thousands, what happens? 00:14:47.110 --> 00:14:49.850 The map is simply too complex for a human to 00:14:49.850 --> 00:14:52.730 read. Troubleshooting which of the 5000 micro 00:14:52.730 --> 00:14:55.429 components failed is incredibly time consuming. 00:14:55.730 --> 00:14:58.850 And physically trying to replace a chip the size 00:14:58.850 --> 00:15:01.789 of a dust moat without accidentally melting the 00:15:01.789 --> 00:15:03.980 plastic connector right next to it. That requires 00:15:03.980 --> 00:15:06.500 high -level skill microscopes and very expensive 00:15:06.500 --> 00:15:08.700 rework stations. So the manufacturer just says, 00:15:08.700 --> 00:15:10.960 we can't efficiently fix this. Just swap out 00:15:10.960 --> 00:15:12.840 the entire board. Or even worse, they just say, 00:15:12.879 --> 00:15:15.019 buy a new phone. The entire industry shifted 00:15:15.019 --> 00:15:17.279 from component -level repair to board -level 00:15:17.279 --> 00:15:19.720 replacement. And it is highly economically efficient 00:15:19.720 --> 00:15:22.080 for the factory. They don't want to pay a skilled 00:15:22.080 --> 00:15:24.600 technician $50 an hour to hunt down a $0 .10 00:15:24.600 --> 00:15:27.200 broken resistor. But the downside is it creates 00:15:27.200 --> 00:15:29.330 a massive mountain of e -waste. Which brings 00:15:29.330 --> 00:15:31.330 up an interesting point from the sources. They 00:15:31.330 --> 00:15:34.730 extensively cover this acronym ROHS. I actually 00:15:34.730 --> 00:15:36.909 see this printed on electronics boxes sometimes. 00:15:37.750 --> 00:15:40.230 ROH. That's related to the materials in the e 00:15:40.230 --> 00:15:42.669 -waste, right? It is. It stands for the Restriction 00:15:42.669 --> 00:15:44.990 of Hazardous Substances. It's a major piece of 00:15:44.990 --> 00:15:47.110 legislation that started in the European Union 00:15:47.110 --> 00:15:49.590 back in the early 2000s. And what exactly were 00:15:49.590 --> 00:15:52.460 they trying to ban from the boards? The big target 00:15:52.460 --> 00:15:55.720 for PCBs was lead. For decades, the standard 00:15:55.720 --> 00:15:58.059 solder used to attach components was a specific 00:15:58.059 --> 00:16:01.659 alloy of tin and lead. And lead is actually amazing 00:16:01.659 --> 00:16:04.360 for manufacturing. It melts at a nice low temperature, 00:16:04.539 --> 00:16:06.679 it flows beautifully across the pads, and the 00:16:06.679 --> 00:16:09.019 resulting joints aren't brittle. It's a joy to 00:16:09.019 --> 00:16:11.220 work with on an assembly line. But it's also 00:16:11.220 --> 00:16:14.039 a known neurotoxin. Right. And when you start 00:16:14.039 --> 00:16:16.159 throwing billions of discarded circuit boards 00:16:16.159 --> 00:16:19.299 into global landfills that eventually leaches 00:16:19.299 --> 00:16:22.299 out into the groundwater. The EU stepped in and 00:16:22.299 --> 00:16:25.399 banned lead in consumer electronics. And we're 00:16:25.399 --> 00:16:27.000 just impartially reporting on the legislation 00:16:27.000 --> 00:16:28.899 from the sources here. But it's a fascinating 00:16:28.899 --> 00:16:31.980 case study in how a regulation in one specific 00:16:31.980 --> 00:16:35.500 region can inadvertently set a global standard 00:16:35.500 --> 00:16:38.929 because. a massive company like Apple or Samsung 00:16:38.929 --> 00:16:41.350 isn't going to set up two different factories 00:16:41.350 --> 00:16:44.350 to make a clean phone for Europe and a leaded 00:16:44.350 --> 00:16:46.850 phone for the U .S. Exactly. It's way too expensive 00:16:46.850 --> 00:16:49.830 to run two entirely separate global supply chains. 00:16:50.250 --> 00:16:52.750 So even though the U .S. didn't have a sweeping 00:16:52.750 --> 00:16:55.710 federal ban at the time, manufacturers just unilaterally 00:16:55.710 --> 00:16:58.409 switched to lead free solder globally to comply 00:16:58.409 --> 00:17:01.389 with the EU. It definitely makes manufacturing 00:17:01.389 --> 00:17:03.389 slightly harder because lead free solder needs 00:17:03.389 --> 00:17:05.809 higher oven temperatures and the joints can be 00:17:05.809 --> 00:17:08.119 a bit more brittle over time, but it removed 00:17:08.119 --> 00:17:10.599 a massive amount of toxic heavy metal from the 00:17:10.599 --> 00:17:13.119 global waste stream. Wow. So pulling this all 00:17:13.119 --> 00:17:14.819 together for you listening, we've gone from the 00:17:14.819 --> 00:17:17.559 rat's nest of bulky, fragile, hand -soldered 00:17:17.559 --> 00:17:21.660 wires to classified military ceramic fuses to 00:17:21.660 --> 00:17:24.140 the mass -produced green FR -4 sandwiches we 00:17:24.140 --> 00:17:26.559 all recognize, and now to these laser -drilled, 00:17:26.599 --> 00:17:28.680 flexible, high -density marvels in our pockets. 00:17:29.160 --> 00:17:30.880 It's a relentless trajectory of integration. 00:17:31.420 --> 00:17:34.039 Every single step has been about packing more 00:17:34.039 --> 00:17:37.579 functionality into less space, removing the slow 00:17:37.579 --> 00:17:40.599 human hands from the assembly process, and drastically 00:17:40.599 --> 00:17:42.859 increasing the reliability. But looking at the 00:17:42.859 --> 00:17:44.660 cutting -edge research you brought into the sources, 00:17:44.900 --> 00:17:47.500 it seems like we aren't even close to done shrinking 00:17:47.500 --> 00:17:49.599 things. We are now seeing things called embedded 00:17:49.599 --> 00:17:52.480 components. Yes, this is the absolute bleeding 00:17:52.480 --> 00:17:56.019 -edge frontier of PCB design right now. So usually 00:17:56.019 --> 00:17:58.589 components sit on the surface of the board. But 00:17:58.589 --> 00:18:00.390 you're saying we are starting to put them inside 00:18:00.390 --> 00:18:03.390 the board itself. Correct. To save even more 00:18:03.390 --> 00:18:05.869 precious surface area and to actually improve 00:18:05.869 --> 00:18:08.230 electrical performance by shortening the signal 00:18:08.230 --> 00:18:10.869 paths, manufacturers are figuring out how to 00:18:10.869 --> 00:18:14.150 bury tiny capacitors and resistors directly inside 00:18:14.150 --> 00:18:16.369 the inner layers of the fiberglass substrate 00:18:16.369 --> 00:18:19.069 itself. We are quite literally merging the component 00:18:19.069 --> 00:18:21.250 with the infrastructure. We are. The physical 00:18:21.250 --> 00:18:24.109 distinction between what is the board and what 00:18:24.109 --> 00:18:26.630 is the chip is starting to completely blur. So 00:18:26.630 --> 00:18:28.259 here is the provocative. get a final thought 00:18:28.259 --> 00:18:31.359 I want to leave you with today. As we move rapidly 00:18:31.359 --> 00:18:34.460 toward this future, where the electronics aren't 00:18:34.460 --> 00:18:37.599 just attached to a board, but functionally, are 00:18:37.599 --> 00:18:40.279 the board what happens to the whole concept of 00:18:40.279 --> 00:18:43.460 ownership and repair? It's a deeply valid philosophical 00:18:43.460 --> 00:18:46.740 and economic question. When the battery, the 00:18:46.740 --> 00:18:49.079 processor, and the memory are all fused into 00:18:49.079 --> 00:18:52.200 a single indivisible block of material repairability, 00:18:52.480 --> 00:18:55.660 it doesn't just get difficult. It becomes basically 00:18:55.660 --> 00:18:57.859 physically impossible. Are we just moving to 00:18:57.859 --> 00:19:00.279 a world where we don't really own our physical 00:19:00.279 --> 00:19:03.180 devices anymore? We're just essentially leasing 00:19:03.180 --> 00:19:05.539 the functionality until the internal glue finally 00:19:05.539 --> 00:19:07.900 fails? It certainly seems to be trending that 00:19:07.900 --> 00:19:10.480 way. When you can't access or fix the map, you 00:19:10.480 --> 00:19:12.740 just have to replace the whole world. The PCB 00:19:12.740 --> 00:19:15.680 is transforming from an open city into a sealed 00:19:15.680 --> 00:19:18.000 vault. That is definitely something to chew on 00:19:18.000 --> 00:19:19.619 the next time you're forced to upgrade your device. 00:19:19.779 --> 00:19:23.220 Indeed it is. So do us a favor. The next time 00:19:23.220 --> 00:19:25.819 you happen to see a bare circuit board, maybe 00:19:25.819 --> 00:19:28.539 inside a broken TV remote, or if you walk past 00:19:28.539 --> 00:19:31.319 an e -waste bin, don't just see a boring piece 00:19:31.319 --> 00:19:33.920 of green plastic. Look closely at the traces. 00:19:34.000 --> 00:19:36.140 Look at the complex layers and the tiny vias. 00:19:36.180 --> 00:19:38.500 It's literally a history book written in microscopic 00:19:38.500 --> 00:19:40.700 copper. Thank you so much for joining us on this 00:19:40.700 --> 00:19:41.900 deep dive. See you next time.