MEP EP #169: Is “I’m an Engineer” a valid defense?

Podcast Transcript

Host 3     00:11

Hello, and welcome to the macro fab engineering podcast. We are your host, Stephen Craig and Parker

Host 1     00:17

Dolman. This is episode 169. So Parker, you got some you got some stuff to tell us, right?

Host 3     00:24

Yeah. It's been a couple episodes since we've done a me a new podcast. Yeah. So we got some fun stuff. Yeah. So I've been working more on the pennant tour. So I we brought that up when Ben was on the podcast, he's helping him he do the software side and kind of the peripheral design on the penetrator. So we decided to kind of go with the at SAM D 21. G 18. A microcontroller is the first time I've ever done a SAM D microcontroller. So it's kind of weird, because I'm used to like mSv, 430s. And in what's the other one, the Silicon Labs, Silicon Labs, the the sleepy bees, the lows and like the parallax propeller. So I've never really dealt too much with Sam D's before. I've done a lot of pic 32 stuff. What's interesting is, you couldn't use a 32.768 kilohertz clock like you would use for like a real time clock that said system clock. And then it has like a PLL that boosts it, or I guess that boost is the wrong term accelerates the clock to 80 megahertz phase locks it a gazillion times up to 80. It's kind of crazy, because like, I mean, there's plenty of of processors that run on that, or they have options for an RTC like that. But I haven't seen it before where they use that as the main clock. Yeah, the sleepy bees from Silicon Labs can, but they only go to like 20 megahertz on that. I mean, this thing's going to 80 megahertz on that clock, which I guess it's it's only four times more, right? Yeah. Every time that this little crystal clicks, a whole bunch of stuff has to happen. Right? Yes. Many more times faster than it clicks. Yeah, that's the whole point. Have you ever looked at phase lock loop stuff? And I have not. That's a Black Magic Art for me. Yeah, it's crazy. It's It's really weird. I've only looked at it enough to, like, answer the question. Like, what is a phase lock loop? I don't even know that. Oh, we should have someone on the podcast. come on and talk about phase lock loops. Yeah, that would be invigorating. Yeah. It starts with Hello, this is you. So what's the phase lock loop? Just like, break right as the first question. Does that just That's it and then like, answer that. And that's it. So whenever, how about we should actually come up with a list of questions that are like that, to ask electrical engineers when they come on. Oh, yeah. But they have to be like super hard questions. Well, yeah. Like, what is the phase lock loop? Like, or implement phase lock loop in FPGA? Like it like, bullshit, like interview questions that you would get asked? Yeah, yeah. Huh. That's always the first question. And they immediately leave the Hangout session. Yeah. Just like that's it, you're on your own for this. Yeah, we have a product at work that uses multiple phase lock loops. So I've looked into it. And it's, it's an oscillator that that is different from many other oscillators as in like, one oscillator controls another via phase lock looping kind of thing. So but in terms of, so like, on low frequency stuff, they're a little bit, I guess, easier to understand. But on this, like, how do you get 32 kilohertz to ring up to 80 kilohertz, you know, effectively, that's 80 megahertz. Oh, did I say kilohertz? Yeah, megahertz. Yeah. It's, you have to go fast. Yeah, gotta go fast. I go fast. So yeah, and one of the big things with this pennant tar is trying to reduce the cost of the pin hex system. Yeah. Because we did a pretty good job at lowering the cost the pin neck system, but we want to take it a step further and reducing the cost. So it's like, you know, going with a cheaper microcontroller, and so instead of like the $10, pic, 32 or going down to you know, a $3 samdi chip. And how we're the difference between the pic Three, two, and the Sandy is they're about the same speed, but the same DS a lot smaller die, and it's got less IO. And on the pic 32 We use like all the IO so it has like 100 GP IO. So to get around that, we're using some, like IO expanders that are spy addressable. And that way we can expand how much IO we can have. So we can interface with like all the solenoids, the solenoid MOSFETs, and all the switches and that kind of stuff.

Host 3     05:21

So, with the with it being all on spy? Do you have any kind of have you had to put any thought into how fast you address things? Or are you putting like faster things on GPIO and slower things on the spy? expander? Not, we don't really have to worry about that too much. Because like pinball, our old pinball kernel on the pin heck ran at like 1000 hertz, which is plenty fast for you know, human interaction. Yeah, but you know, the picture two's running at 80 megahertz. So 1000 hertz is like nothing. So it's doing all the background stuff, managing all the DMA, and all the all the game management stuff. And then it just handles the IO stuff at, you know, 100 megahertz at 100,000 Hertz. Yeah. 100 megahertz would be insane.

Host 2     06:12

And so, basically, we're going to run the spy at kind of like the fast as it can go and not really worry about it. It's basically like, if we'll get to it, basically, because we're running the colonel at 100 Hertz. Man, I'm 1000 hertz. So we don't really have to worry about that too much.

Host 3     06:29

Sure. Yeah. Do you know what your spy right speed is? Not right now. Okay. It's up there. Right. Yeah. Well, we haven't gotten that far yet. But it's going to be fast enough. We did some calculations mainly on like the lights. Yeah. Because we are. We're doing sir lights. And so we calculate like how fast we can drive those that serial light string, right? Like, okay, we have to drive it for this amount of time. And you have to make sure you keep all your clocks synced and lots of stuff. And so you're filling up that buffer and stuff like that. And we basically made sure running zero lights wouldn't interfere with like normal, you know, game management, and it won't got it. Yeah. Yep. And the big thing with the pin atari from the pin hack is reducing like voltage rails, because the pin hack had 12 volts, 50 volts, five volts, 3.3 volts, and then ground. Right. So it was a six layer board. Yeah, that's that's a lot of that's a lot of expense in just regulation. Right. Yeah. And so the big thing with the Well, I wonder with Pretoria's reduce how many voltage rails we're using, like on the board to just 3.3 volts. So sure, we'll have to have 50 volts for like the solenoids, and we'll have to have higher voltage, like five volts for all the external like lights. And, and we're doing serial switches. So we have to send five volts down for those two, but those are just on the IO, the peripheral, right? So you just run a big fat trace to it on one layer. So you don't have to have its own internal plane layer taking up, you know, your your stack up. And so we're hoping to get down to basically a four layer board. So we'll have signal ground 3.3 volts signal again. Yeah, well, we can be jerks and like, do 3.3 volts on top and then signals inside and then ground. No one can see the traces or you can never repair it. Yeah, but you can say it's there for shielding, right? Yes. For shielding. Yeah. Like, like a pinball controller needs that. You know? Exactly. It's on a big steel. backplane. So yeah, it's got all the shielding needs. Right, right. So on the pin hex system, you had mounting holes and you use pen studs, to kind of mount it to like a little bracket. Are you going to do something similar to that on this one? Yeah, we're probably going to get off a pen studs because the machine shop that we used for that didn't like pen studs, because he had to pay for the pen stuff tooling and they didn't like that. Yeah, there's like an apple and stuff you have to buy for it. Yeah, yeah. And personally, I was like we build enough of these it's the cost of it wasn't an issue but it shouldn't you shouldn't be but it was we'll probably figure something else out. But I like to have a I'm basically looking so like when we sell the penetrator to people like to because this is not the sell will probably sell to individual like hobbyists and stuff but it's it's designed to be sold to like spooky pinball like small pinball manufacturers that are going to be buying like 1000 of these a year. Right? And so we'll probably sell it with the backplane this time. And so instead of just like specifications of what the backplane is, you buy it and it's already attached to the backplane so you just put it in the back box just to reduce like board flex and shipment and bunch of other issues that pop up when you when you try to ship a ginormous board. It's got through hole components that poke out. And so you start stacking them into a box and they start flexing and yeah, it gets nasty. Sure. So so the backplane does that just screw into the wall of the? Yeah, they just take like self tapping screws and just rip it right into the plywood

Host 1     10:23

Or pinball enclosures made out of plywood or I would have thought they would have been made like MDO or something like that.

Host 3     10:30

It's it's MDO Oh, but just plywood. It's just glue and sawdust. Oh, that's MDF? Isn't MDO the MDO is what they make signs out of which is plywood with with its waterproof plywood with paper on it. Oh, that's yeah, that's the one that has veneers on on either side. Right? So it's real smooth. It can be veneer, but most of its paper, actually, it's a sheet of paper on it. Okay, so it's very smooth. It's waterproof. Got it actually built the back. Like I have like a drawer system in the back of the Jeep. I made a MDO because it's waterproof. And you can glue like carpet to it really well. Okay, cool. So yep. And then for the voltage because I want to like simplify like what you what voltage you apply to the board. Because like the old system, the pin hack, you had to give it 12 volts and five volts. And so you kind of had to have like a multi rail power supply to handle it. And those are kind of expensive. And so I'm hoping to get to just supplying you just give it 12 volts and 50 volts, because you have to give it we like I like to separate out the 50 volt supply from the 12 volt supply. It kind of makes it a lot easier to handle like crosstalk between the solenoids and the low logic system. Right? Because the only really the only way that the 50 volt interacts with any of the other things is through switches right through MOSFETs. Yeah, through MOSFETs. And so the 50 how they're connected on the pin heck is we had a small ferrite bead that connected the grounds together. So so your, your potential was the same on the grounds, but any kind of noise was suppressed through the ground plane. Oh, really? Did you find you had to do you needed that? No. Well, we never experimented with just like putting a trace there or anything but it worked. Oh, they worked really well. Yeah, okay. Yeah, we never had issues with it. So we just said, Okay, we'll just run that. Yeah, leave it got it? Because that's technically the correct way to do it in quotes. Is there? Yeah, I got it. It's like training if like, if you had a really noisy like digital plane, and then a quiet analog section. Yeah, you don't really kind of want connecting together, but you need them to be referenced the same. So putting something that will kind of suppress spikes and stuff is kind of what you need. Yeah, a little ferrite bead kind of works. I guess you could technically put a trace there or like two pads and and figure out what frequencies you need to suppress. And then put the correct ferrite bead there. But I was like, Yeah, throw 100 Like, oh, is it up? It's one of the part number I use for USB shielding to ground? Well, it's my designs. It's like 100 megahertz. Fair beat I think or something like that.

Host 1     13:25

Got it. So yeah, I mean, the so effectively your star grounding with a ferrite bead in the

Host 3     13:31

Path in between, Yep. Cool. It seems to work pretty well. And so for the pinna tour system, we kind of want to move to like serial switches and lights to kind of simplify wiring. That's the one of the other areas that you can like save costs when building pinball machines. The most of the money and labor is put into like attaching the wiring harness to like all the solenoids, and all the switches and stuff. And like because you have to hand solder all that. There's not a machine that can do that. Yeah, they have eggs for like flipping over, like the playfield. And like people spend, I watch one documentary where people spend like multiple days on a single play field, like one person is dedicated to a play field. It takes stern pinball, which is the largest pinball company in existence right now. It takes longer to build and stuff a play field that does Ford to build a car from start to finish. Yeah, that's how much it's I think it's more of a testament of like how much labor savings like giant robots have? Well, but begs the question, is stern really, really, really inefficient? Or is Ford really really, really efficient? Like I think it's, I think pinball machines are just really inefficient to build. Yeah, currently. Yeah, sure. And so what we want to do is make it faster so we want to build it Serial switches. So you just have to plug cables into each one. Yeah. All right. And so I started looking at like serial, LEDs, the smart LEDs like the Ws, ws 2812 B's. The problem with those smart LEDs is they can't reflow at all at lead free temps right there Max temperatures like 235 C, you reflow them, and then they basically their lenses crack, and lenses crack, and then the gold bonds pull off their die. Yeah, and it's so they fail. Yeah, and we've talked about these all the time on the podcast, and they have like a 30 to 40% failure rate and lead free reflow which kind of sucks. Because I remember, there was a really good picture on one of the previous podcasts of me wearing like a welding helmet, replacing these LEDs and testing them. So I was wearing the welding helmet to protect me eyes from how bright they were. Yeah, it was just like looking at them when when you turn them on full bright white. They're just unbelievable. Like, there was like, it's like, staring at an eclipse. Yeah, they're ridiculous. Right? I remember I remember used to sit at the bench for hours upon hours replacing those. I don't remember. It's like the product 1000 or something like it wasn't it wasn't like an array of these lights and like, yeah, inside the array, like two or three of them were out. And the thing is, since their cereal, you had to replace the first one and the chain that was bad. So you fix that one, clean the board, wash it, dry it tested again. And then there'd be another one in the chain dead. Do the whole process over again. So you do like them in batches. So you'd place one, do a whole batch wash them all. But it was not fun. That was an entire week in my life was replacing 3000 LEDs. Yeah, that's. And then we basically after that we put a blanket statement of like, we're not doing these LEDs ever again in production for people. Right, but but so I started doing more research and these LED in the smart LEDs actually go into the original, like manufacturers like APA electronics. So looking through the catalog, I'm like, Okay, it's been like five years since that incident. They've had have like, found their way and started making like a lead free version that can like reflow at like 250 or higher Celsius. Right. And they have according to the datasheet at least Yeah. And so the APA 102 C 5052 56 Can reflow it 256 degrees Celsius, go figure, go figure. And then they also have the APA 104 260 5050, which reflows at 260 degrees Celsius. And so one of those uses like the single wire, ws 2812 method of communication. And the other one uses data clock. I can't remember which ones which. So it makes it it makes you wonder nowadays, like why would any manufacturer make a a surface mount product that can't handle the industry standard? Like, I know, it's really curious. And it's not like it's not an isolated event? Like look all over the internet for ws 2812. And all you get is these are really cool, but you know, you can't use

Host 3     18:31

Them. Yeah, if you Google like ws 2012, lead free reflow. Everyone just like no, yeah, just don't even try. Everyone's had the same results. Yeah, it's pretty bad. So I contacted the their sales rep here in America about them. And I'm like, Hey, so what's the deal with these parts? That she says with reflow? And they're like, yeah, these are actually kind of new. And so what they happen was a, I guess the guy told me a big enough customer came in and ordered enough of them and told them we have to reflow these lead free reflow specs. And so they changed their entire product line to be lead free. Reflow compatible. That one company Okay, so first of all, that one company is the biggest order they've ever had. So they make a bunch of money. And now they're going to have a lead free process. So that mean that one company is their Savior in a way it kind of I guess at least for their the savior of other people using these LEDs. Well, okay, so we also talked about it in the caution a long time ago. But we say in that using a vapor phase technology, these could have been soldered properly with lead free, correct Yeah. So I mean I Okay, so there is a industry standard way that you could use them. You just have to find a cm that has a vapor phase oven. Yeah, right. Exactly. So, but you're not talking mass volume there. So, no, because this is a vapor phase kind of slow. It's batch. Yeah. Yeah. So, all right. It's not like it's not a conveyor style reflow it is, you put them in there for, you know, 15 minutes or however long you're reflow profile is and now 15 minutes is way long, so six minutes. But um, yeah, it's not a continual stream. So you got to put them in, cool down, pull them out, etc. You guys have a hilar 17. Something? Yeah, something like that. 17. It's one of the big guys. I have reprofiled Rs, just not that long ago. In fact, we did some reprofiling. Today, actually, just two, we had some boards that have some pretty tall electrolytics on them. And the electric it's, it's interesting, because the electrolytics, depending on the direction that the boards went in the oven, they either soldered well or had some issues. And it's it was due to the fact that you have hot air blowing on it. And those electrolytic were actually shielding their pads if they were going in one direction. So in other words, kind of hot tip from the Mac read engineering bug. If you have a say a line of electrolytics, you want to try to penalize the boards such that the electrolytics are in a line that is parallel with the conveyor belt, such that the hot air can flow around the electrostatics as opposed to them being like a shield in a way. Yeah, so like one side gets refloat and the other in the in the trailing side does not yet you have like a heat shadow in a way. And the thing is you can actually get like any passives that are in that heat shadow. Gotta watch out for that. Yep. Interesting. Yeah, actually never really thought about that. But I've never designed anything with really tall electrolytics yet. Well, and if it's one electrolytic it's one thing. But if you have like four or five in a straight line, and they're all you know, what if they're all like five times taller than a, oh, 603 there's a good chance that first of all, the electrolytic pad won't solder, but maybe also anything close by. So keep an eye out for that. But you can just crank the oven. Yeah, that to you weren't an issue sometimes. Right? So I ordered two reels, one of each of those. And when they arrive, I'll do some testing. I basically I'll put them in my inventory at macro fab and then have the production team just make some test boards and then test them and see how well will work. If I get 100% yield, then, you know, I'm thinking about making these house parts for macro fab, which would be pretty nice for because I think one of them is a drop in replacement for ws 20 twelves. Nice, which would be nice. Yeah. I actually had a customer just asked me the other day about this part number. And the thing is, I didn't recognize the part number. All I read, all I did was I pulled up the datasheet. And I saw that led and I was like no, no, no, no, no, you're like, Oh, hell no.

Host 1     23:11

And he was like, what it says lead free temperature. He's like, look at the temperature. I'm like, I don't know. We'll have to find out.

Host 3     23:21

Was it like SK something something? I don't remember off the top. There's some of them that say like they do 250. But they can't they just lie on the datasheet. Yeah, but I made sure I talked a lot with the sales rep. And he's like, these can do lead free 260. So we shall see. Nice. So APA electronic, you're going to be on the hook for this one. All our listeners will know. And then actually, at the FAB, we got some new equipment. We got a my chronic my 300 and a micronic. My 700. So the 300 is our new pick and place. And they're working on getting that I think it's actually placing parts already this week, or last week, actually. And so the big thing over that our old one was a might 200 It's basically twice as wide. And it has a larger multi pick head, which is cool. So it goes a little bit faster. And it's quieter. So they got some new maglev technology in it or something, huh? Okay. Yeah. And I think the coolest feature is it's got a it's got electrical verification now. And so what that enables you to do is when you build the tooling for the machine that the program, you say like hey, the part that goes on on our one is a 10k resistor with this tolerance, because I already have that information. And so, yeah, that's all fine and so people will load the real into the clip Pool, which is what they call the feeder on the micronic. And then you put it in. And the clip says, Hey, I'm a 10k. But the clip doesn't actually know it's a 10k. It only knows it's a 10k, because a human told it was right. And so you refer, you're basically banking on the fact that a human, put a 10k resistor reel into the clip, right. And so what this does is gives another, like, we have some checks and verifications for that, but you know, stuff still happens. And so this gives us another level of just in case, you know, just in case someone actually loaded 100k in there, or a 1k, or whatever, into that machine. Or you could actually find, hey, this is supposed to be a 1% part, but this is actually a 5% part. Oh, it'll check tolerance, it will check tolerances to that's really cool. Will it check capacitance also? I think it does as well, I think it does capacitance and diet. Oh, so it'll tell you if you're placing it the wrong way. Yeah. Which which is nice. Which it's rare, but it does happen occasionally, that a polarized component can be flipped inside of the tape. I've actually, we've received parts and tubes from like DigiKey and Mauser parts backwards. Yeah. Yeah. Well, a lot of times they're they're tubed and reeled by Mauser or whatnot. And so that's another level of a mistake could be made. Yeah, exactly. So and I've seen actually parts, jump in the tape. And so like, wait, because you don't you have a couple of components exposed in the clip. And so when it moves forward, a part can jump and then kind of just like, rotate around. So I've seen actually, I've seen diodes flip backwards. Oh, yeah. Because of that, for sure. And it's insane to think about that. That can happen. But yeah, it does happen. Oh, it's just why. It's why manufacturing is never 100% yield, no matter what, and stuff like that. Yeah. If you're making something you shouldn't expect 100% yield. I mean, if you're asking for one than maybe, I would say if you're asking for one you should be expecting 100% yield. Yeah, for sure. Low volume. Yeah, you should be expecting 100% yield, but for high volume stuff, stuff like that. Just like there's like a percent chance something will happen. And so it you know, you will get one that that happens on? Yeah. I've actually seen an soI see eight flip upside down in tape. Just from the tape being advanced. A slot. Okay, that's not once, but it has happened. Yeah, yeah. Kinda crazy. And then we got a, my 700 which they were working on the the micronic. Guys were there this week to set that machine up, which is our new upgraded pace jet machine. So it will be able to do smaller dot sizes for parts and it's dual head so it's faster. No, no, that's gonna be cool. So the my 200 and the my five 300 It was a 500 year old paste yet, right? Yeah. So we still have those machines in the store running. What are those going to be dedicated for? There will be dedicated to our 10 Day prototypes class.

Host 3     28:19

Okay, got it. Got it. Yeah. So the your master stuff goes on one line. And production ish. Stuff goes on the other line? Yeah, low volume production is stuff will go there. Cool. Yeah. And the crazy thing is when you look at the my 500 through the 700 it lost weight. Yeah, cuz if you remember the 500 It was the 500 is such an insane machine when they like when you look at it. It's the chassis is made out of granite. Yeah, because that thing holds s. So these will be really stable, really heavy and stable. But the 700 looks like a normal like stencil machine. Hmm, really? So like, yeah, so the 500 looks like nothing else in like SMT. Machinery. But at 700. It looks just like a stencil printer would look like. Yeah. And it's not made out of granite. It's really it is really heavy. But I haven't dug into it yet. So I'm like, how do they make it that heavy? I'm gonna guess like it has led bricks underneath it. Or it's made that conch as concrete blocks underneath, because for some reason, because it's just like a sheet metal outside like a normal machine would be. Oh, yeah. The My 500 had like a, like a really thick exterior. Like it was Yeah. Yeah. Yeah. I can't remember what the name heavy heavy robotics. We had them on a year and a half ago, two years ago. And we were talking about their robotic arm offerings. And I remember something that they mentioned about like CNC and motion. And you could pick accuracy and stiffness and rigidity as like those are There's like, knobs, three knobs you can play with. Yeah, yeah, the knobs you can tweak. It always felt like the My 500 just tweaked all three knobs up. You're like, yes, it's rigid as hell. It's unbelievably accurate. And it's really fast. But I guess I guess the one knob that didn't didn't tweak is how much it cost? How much it cost to manufacture, right? Yeah. So the 500 was the first generation of that kind of pace jet machine. So the 700 is actually the third generation. So we skipped the 600. Got it. And so I think, basically, I think, but the 500 they didn't really know how to design it. And so they kind of just they did what they you said, they cranked all the knobs and made sure it worked. And so they could actually sell the technology. And now they're backing it off. And then now they're tweaking it and say, Okay, now we need to make money. So it's, it's optimize all that stuff. Maybe he doesn't need to have, you know, a granite exterior. Yeah, yeah, for sure. Yeah, it's heavy as hell. And when your pace jeder cost more than your pick and place. or I shouldn't say patient when your paste application device. Let's put it this way that the one of the paste machines we use at work, I think, was 20k. Whereas a paste jetter is 10x that, you know, so like, yeah, it gets up there. Yeah, get the job done for a lot less money. Yep. Yeah, the problem is stencils are eaten up your your art, your and our ease. Oh, of course. Yeah. Yeah. And with the way that you guys manufacture stuff, stencils would be awful. You you'd be you'd be having five or six stencils arrive every day. And they're only used for that day. It'd be like, four times that number. But yeah, yeah, probably. Yeah. We have like, I think at the moment, we have, like 30 or 40 different panels show up every day right now. That's crazy. That's different panels. Right? Yeah. Yeah. That's insane. How many different kinds of jobs we run on those machines? So I'm so glad that we have new machines now. So it's based I'm looking forward to seeing the the lead times and stuff start to you know, drop down. Yeah, that's the goal, right to get 10. Day less than 10 day. Yes. You heard it here first. That's not that's not really a new secret that's been macro Feds thing for like, ever. Ever. Yep. Faster, faster, faster. Right. And be affordable. Ish, though. Pretty good for US manufacturing. I would say, Oh, you're you're in the middle. Yeah. Like, you're not trying to be the cheapest guy. And you're not like just obscene ly expensive. You're just Yep. So that's it. Yeah. Cool. So. So I've been working a bit on the macro amp, which we talked about the last time we had a you and I podcast. And that's like the second oldest macro, grab podcast projects, or like at least top five. It's yeah,

Host 3     33:17

It's way up there, actually. Okay, so it's interesting, because Parker and I were talking about this just the other day, we dog on the SSPs all the time, the super simple, or super stupid power supply. We dug on that as like this, like huge, unfinished project in our first one in the first episode, stuff like that. I actually, so I started digging into it the other day, because I was actually thinking about redesigning some chunks of it. And I maybe I didn't even remember this, but like, the last place we left, it wasn't working. Like we left it functioning, like it was totally functioning. The core of it was the Energon que, right and, and that function, like we have video of it working, I have an entire blog post I wrote with the end of the blog post was like, here's the thing functioning, all we need to do now is kind of like continue to validate it and then put it all together. And that's where we left it like my mind. Like I've tricked myself with how many times we've talked about this thing where I'm like, hi, that things like not even functioning No, like it's straight up works. works and works fine. Yeah. Although Parker and I were talking about potentially looking at it in a different manner and considering some different topologies because even though it worked like there's there's probably some better ways of doing it. It is not the most efficient machine. It might actually be a runner for one of the most inefficient Yeah, it's I always think we should change the name of the super simple entropy machine.

Host 1     35:00

Hmm, I see. I see. Yeah. Yeah. And it just gets worse as it gets older. Right. Yeah.

Host 3     35:05

That's a good one. So yeah, I don't like I'm no, that was a tangent one. But it was it was worth bringing up that like, No, we were we actually almost finished that project. It's it's way further along than I thought it was. So, but the Mac ramp. Yeah. So the Mac ramp was a new tube amplifier that I've been designing for a long time. And I got the boards made, but I had never actually just tested them, like literally, they were built, just never applied power. And just a while ago, I installed some new tubes in it and put in the J FETs. That that I just had in a bag like those were the last things to solder in. So I started those in and fired it up. And lo and behold, everything functioned. Everything asterisk, what because one of the J FETs was actually bad. It's just DOA, which kind of suck. But in fact, I thought it was one of the new tube that was bad, like from the get go. So I clipped out a new tube and replaced it with another one. And it still was bad. And I was like, Oh, well, I guess I just wasted a new two. But, ya know, I found out that a J fit was bad, which you know that those three like those to 92 style transistors like, never had a bad one before. A brand new bad one, I should say. But I guess there's a first for everything. But yes, so I got the I fired it up, everything seems to function well. I've got way, way more gain on tap than I actually need. And in the preamp itself, I don't have any attenuation control, like I like a volume control the will within the structure of the preamp, the really the idea is to have the attenuation control at the front end of the preamp. So the preamp is always basically running full bore and just you choke off what's gets sent to it. So right now, I followed the general structure that was set forth in the datasheet for the new tubes. And just the way that it turned out, I got 42 times gain in this preamp and 42 is a lot if you think about the fact that I've already got it set up for line level input. So 42 is way more than enough to drive the power amp to like saturation and beyond, you know, way, way, way beyond. So what's the normal gain for line level to like a power amp? What's the normal gain you'd be looking at. So that depends on a ton of factors. But in general, I would say somewhere in the 10 to 12x range, because for this style of amplifier, they the power amp will distort somewhere in the 10 to 12 volt input range.

Host 1     37:57

Maybe a little bit more, but I would I would probably leave it about there. And that's, you know, with a 10 volt input to the power amp, dimpled RMS, I'm gonna get the full, you know, 12 to 20 Watt output. So with 42 times gain, it's just going to be like an awful mess. So the good thing is gain is easy to get rid of, it's a lot harder to increase. So I just need to chop off some gains somewhere. In other words, probably just increase some resistor values here there. So that all worked out? Well. I did have some other component values that I realized that I had just some bad values on there's a power supply resistor that I chose to be 2k. And that that would be a totally fine resistor values a 2k three watt. And but I realized that if he guy, yeah, cuz I'm dropping some juice through it, but I but I realized that I pretty heavily miscalculated how much current is flowing through that resistor. So that resistor supposed to drop somewhere in the vicinity of a few volts, but it's dropping like 20 to 30 right now.

Host 3     39:13

So I basically what happened is all of those J FETs. I was talking about they're a J FET. J FET follows the same convention as a MOSFET. Right? You have source drain and gate right? I would say yes, I yeah, I don't mess with so the, these are en channels. So that means that the source, so I haven't set up as a buffer. So the source is spitting out to a resistor that goes to ground and I set that resistor to ground far lower than needed. I think I haven't a 10k or something like that, but you have six of them. And you have 30 some odd volts on that 10k resistor so you know, follow whatever current that is and then mult But by six and then multiply that by the 2k power resistor, and you start dropping a ton of juice. So in other words, all of those 10 K's in my buffer circuits, I just, I can arbitrarily boost them to 100k or 220, or something high, it doesn't, frankly, it doesn't really even matter that much, I just need some juice flowing through the J FET. I don't want it to just be like micro amps, but you know, it doesn't need to be five milliamps per per buffer, like, that's just way too much juice. So basically, I need to update that. And then, and then I should have everything working pretty. So you're gonna change the power supply resistor, or we're going to change the J FET. Like loaders, both. I'll start with by changing the loaders, and then see where I sit. And then if I and then I'll probably have to bump that 2k resistor down a little bit to just get a usable headroom. And that, so that's all good. And then I had originally designed in an ri, a filter, the RIA filter, which is basically the it's like the inverse of what they cut a record at. That's right. Yeah. So I haven't had a chance yet to actually test if that filter is doing what it needs to however, I have put a signal into it. And I'm totally getting signal out of it. So so another In other words, like it's functioning, I just don't know if the polls are where they need to be. The way I'm going to test it is I'm going to put a record through it and see if it sounds okay. I'd love to say is I actually I have a a test record that has certain frequencies on it. No lie. Yeah, I do. So we I can bring that up to Colorado, and we can pop it on your record and then make sure we get like a kilohertz out. Oh, nice. Yeah, yeah, I liked it. That's, that's great. You know, I looked it up once reference microphones on Amazon. They're not that expensive, actually. So they're not intended to sound good. They're intended to be a reference, they're intended to be like, You know what, it detects this SPL at this frequency. It's that SPL at that frequency. So it'd be really fun to measure one meter from a speaker because that's usually the standard, and then see if we're getting that. So we could just talk course, of course, we'll have to have a couple of beers before we do that. But so so the data will be you know, accurate, only as good as how many beers

Host 1     42:30

Within three beers of accuracy. Yeah, there we go. Yeah, like that. In fact, so have you have you heard of the Gosh, what are they called? escaping me now? They're called Audio exciters. They're called Audio exciters. They're seven

Host 3     42:45

Is that where you call the front row at a concert? Like oh, that's groupies. Yeah, yeah. So these audio exciters think of a speaker without the cone. So it's just a, it's the coil that moves a magnet. It's a coil that moves a magnet. Yeah, exactly. So these things are available on believe the website is parts Express, but you can probably get them on Amazon too. They're like 17 bucks. And basically what it is, is it's just a vibration motor effectively that that is tuned or it has an impedance that works with the amplifier output. So like eight ohms.

Host 1     43:23

There's some there's some information available on YouTube and a handful of other sites where you can basically take these audio exciters and stick them to things and just turn anything into a speaker. I've seen this Yeah, really, really freaking cool. Especially because like you can try different materials. And you can try different locations on the material and different XY sizes and thicknesses and you can tune your speakers by like cutting them effectively. So my my idea right now is after I build and finish the macro amp, I kind of want to make some foam speakers with these audio exciters and make a real like Goofy wanqi Looking record, like amp space in my basement where it's like, this is my audio area where I have like these really goofy hifi speakers that are made of pink foam

Host 3     44:18

Cheapest material at Home Depot so maybe that's something we'll have to play with when you when you get up here. You know, I actually think would be really cool if you made a false wall. Yeah, that was the whole wall was the speaker. Yeah, so that would be pretty sweet. I've seen some people do that. And it's it's pretty cool. The thing that I want to do that I think is super wacky, it's like way beyond and I love this.

Host 1     44:47

You use magnet wire to connect to the audio transducer, the audio exciter so magnet wire is what carries your signal, but you also hang the speaker from this Dealing by the actual signal wires. So it's suspended in space. And it's basically held by nothing other than the signal wires itself. And how wacky is that? You know, just like zero gravity speakers hanging in space your gravity speaker, there's

Host 3     45:16

For the most perfect. Come into a Kickstarter near you. It's just a spool of wire a big foam. See real cheesemakers. Hey, I bet you ever used white like gloss white foam? It would sell billions. Oh, all the apple people would buy it. Yeah, they'd be called the the was it the? What was the name of the floating speaker? The Zero G speaker. Yeah, I zeros. Oh, yeah, lowercase A uppercase I would just be that would be dumb. No. No one would buy it. Yeah, I can just spray paint them white. Does that work? Sure. I get well you can't spray paint. Oh, good, you know. So yeah, so yeah, the the the macro amp, it's functioning, I've got the enclosure all drilled out. I'm just gonna kind of slap everything in there and start wiring it up. So one of the other things I've kind of been hammering on is an amp that I've, frankly, this is the third time I've built this amp. And I don't mean like three different amps. This is the third time I've built this specific one. Like, chop it apart, rebuild it, like the Johnny Cash song. Which will not a Johnny Cadillac buy parts or something like that. I you know, I heard that I heard he had 1500 individual like unique songs. Like that's insane. Absolute, like hits? No, like it was called one piece at a time. It's a he works at like a Cadillac like, plant and he takes a piece at a time to build his own Cadillac. He steals a Cadillac one piece at a time, basically. Yeah. But yeah, I was thinking that amp is like my jeep. Oh, yeah. Cuz you've been up and down building, like, rebuild it every three years. Yeah, well. And so the thing is, like, I had no intent of actually rebuilding this app. And in fact, before I rebuilt this entirely, I actually service to the amp and got it up to like, full functioning capability. And then I was like, I literally just got out clippers one day, and I was like, just Nope, we're doing we're doing this and I just chopped it all apart. But what's still original? The Transformers? Like the I was gonna say the knobs? No, no, the chassis is entirely different. The Wire, nothing the use of transformers. And the reason why is because they spent $500 on the Transformers 12 years ago. But unlike what although I did take all the Transformers apart and I repainted them. I got I'd like took all the rust off of them. And they they look gorgeous now. But actually, oh say that that amplifier is less the original employer than my jeep actually is though. Oh, yeah.

Host 1     48:05

Yeah. Well, and it's funny because like it is now actually closer to what it was originally based off of, which was an amp that came out in 87. Kind of, sort of, sort of tightly. And yet also loosely based it off of that amp tightly as in like, I follow the general locations of where parts go, but loosely as demography. Yeah, right. But loosely as in like, I designed all the boards for it. And I used an entirely different wiring scheme. And I added a whole bunch of upgrades and a whole bunch of other stuff too. But back in episode 155, we talked about a new footprint for a coax connection, a coax PCB connection that I've been wanting to try for a

Host 3     48:54

Lot and you say 155, but I remember you like the second or third week at macro fab. You had this idea? Yeah. Yeah. And it took you 155 episodes to actually try it. Well, I'll get no it took me 155 episodes to have an excuse to try it. And the excuse was, I looked at this amp and I was like, I could use that footprint to do something new I yet no. Looking back now, I designed this or rebuilt an entire amp around a footprint in a PCB broker.

Host 1     49:31

And so the magic of around that is, I've always wanted to try building an amp where either 100% or nearly 100% of every wire was coax shielded, because I wanted to see like, okay, all the extra effort of every wire now having a shield that you have to ground or do something with, does that make a difference? And the answer is, yeah, 100% It absolutely does make a difference. So I actually got the datasheet for this coax. that I used throughout the amp and found that it has about 18 pico farad of capacitance per foot. And that's most of the runs of all the signals on here are three inches or less. So we're probably looking at, you know, an average of an extra four or five pico ferrets for any given run, that's enough that it's not going to have much, or if any, have an impact on the actual sound of the amplifier, but isn't going to have an impact on crosstalk and lower noise and things because on these kind of amps, like you have a huge chassis and you're passing giant gains around like the cumulative gain of this amp is 72 to 73 DB, which is over 4000 times gain, you know, so it takes 340 micro volts worth of input to get a distorted output on the out outside, like, you breathe on it and it's distorted, you know, like it has so much gain. And these things are like, notorious for noise and oscillation and feedback and squeal and all that jazz. So I wanted to see like, okay, if I'm, if I design my, my boards, do it my way with my footprint, can we get difference in the answers? Yeah, 100%. Like I said, this is the third time I've built it. And it's never sounded like this. So it's really interesting. And this is a lot quieter. There's very little hum and buzz. I mean, it still exists. Because it's pretty much

Host 3     51:30

Got it. Oh, you can sit it's got 4000 times gain. Right, right. Yeah, you have to. Right, it's got to pick up vibrations on the input. Just from like from the air. Yeah, exactly. So you know, a you live with it, but at the volumes that it has noticeable gains. Those are also the volumes that make you go deaf. So it's one of those things where like, okay, yeah, sure, it has noticeable gain, but the signal to noise ratio is through the freaking roof. So, yeah, the the footprint works. I'm super happy with it. And I'd like to potentially use it in the future. It's not something I'd ever use in like a professional setting. Because it's just, I mean, it's really like hand crafted kind of thing. Like you have to do it in awareness. And it doesn't make sense to do it in a production sense. But on a single one off app like this. Oh yeah, it's awesome. And so super happy with that that came out well, artisan crafted amplifiers by Steven Craig. Yeah. Wangka fires. Yeah, I came up with a term today it was Maga which is make amps. No make analog great amps. So that's that. I need to start wearing a red hat that says Maga right. Yeah, be careful with that one now. Yeah, gotta be real careful with. All right, on to the RFO RFO. So I gotta get a cool topic today that I saw it pop up on Hackaday. It's called hack my health UL certifications and turning the lights on with an ESP 8266. So there's some there's some interesting stuff behind this article. And I, I was really interested in reading through it, because I've had this same question before, where it's like, Okay, so we've got all this really cool IoT stuff, where like, I can unlock my door from my work, if I wanted to, or I could, I don't know, I could turn on my pool cleaning robot if I had a pool from work, like, that kind of stuff. But but at the same time, so I, you came up with two examples that are like, completely worthless. I do have an IoT socket. Yeah. With my garage, air conditioner plugged into it. And so, so if I'm at work, and I'm like, Hey, I'm like, hey, I want to go home and work on my wagon. But it's 100 Fucking degrees outside. I'm like, Okay, I'm gonna spend so I know I'm going to do that. So I don't have anything planned to whatever, I go on my phone and click Turn on AC. Yeah, and it will turn on my wall unit in the garage. And by the time I get home, it's like 74 degrees in my garage. Nice. So there's one application that because I don't want it on timer, because not every day I'm going to come home and work it there. Right out there. Right so I want to make a decision about two hours before I get home. Yeah, because we all know it's going to take two hours to drop in Houston from Yeah, it's well it's like a 30 degree difference. Yeah, for sure. Well, okay, so that brings up an interesting question your your outlet or your socket is that UL certified? I don't know keep talking I'm actually gonna grab Okay, he's gonna go I'll yeah, he's he's actually just in the room right next to his garage. So.

Host 1     54:51

So the UL certification when it comes to IoT kind of stuff. What's what's interesting about it is like if you know we have have all these IoT devices that do all this magical stuff. But as soon as we start connecting them to the mains grids, Will, you're no longer playing with just your IoT little widget, you're starting to play with some other stuff that usually requires some kind of certification. And it's the the issue comes up with what if you install something and burn down your house or your garage? And it did not have the UL certification? What happens? What you know, are you are you Sol for it? Are you Sol? What is your insurance company going to say with that? Is your insurance company even gonna care? And so this article kind of talks about that a little bit and brings up some some good points. But as I was reading it, I remembered my buddy Roz, who was our guest on The Last Star Wars episode that we did, he actually works for an insurance company. So I hit him up and asked him questions on this with the the UL certification, and what that would mean for a homeowner's policy. And he actually pulled up basically the people

Host 3     56:03

Who would pay you if your house burned down. Right, right. So he researched it, and he pulled up their standard homeowner's policy and dug through it and what was really interesting, you know, I'm not gonna give away the the name of the insurance company, but I assume in general, they're all fairly similar. What's interesting is, is in the his situation, there was no clause that said that they could deny coverage for appliances that did not have a UL certification. In fact, there wasn't any clauses about devices plugged into mains requiring the market to be any kind of certification. Right, right. UL or ETL. Or any so this this socket thing? Yeah, it does have a it's the other scumbag company. ETL ETL. Yeah, yeah. Yeah. Avi Eagles at the mark of the beast. That's the mark of the beast because he thinks it's inferior, but it's the same thing. It's the same thing. They take your money and they put a sticker on your on your product. Like that's yeah, I see as well. But the thing is, it doesn't have the F because it's Wi Fi. It doesn't have the FCC ID on the outside like it should.

Host 1     57:16

Is it required in I guess it actually I don't know. Because actually, my phone doesn't have on the outside either.

Host 3     57:23

But But if Okay, so does your phone have a removable back? Nope. It does not. Okay, this is a Google Pixel two and doesn't have the FCC. I remember one of my phones had that on the back that yeah, the ID. But maybe that's not actually required. Maybe it's not I'm pulling the case off my phone right now. I have a Galaxy S nine. And on the back. I see the don't throw it in the trash can symbol and the CEE but no FCC. Actually, what's funny is mindsets. I have this I have the G logo for Google. Yeah. And it has just a C. The GTA, I guess let's grant that thought out of Google. That's great. So yeah, okay. So back to the insurance stuff. So my buddy looked through the homeowner policy and basically said, yeah, like, if you plug in an appliance, and your house burns down because of that appliance, it's not guaranteed to be your fault. Or it's not like inherently your fault. In other words, like, if you go to Amazon and you buy some garbage, whatever it is, the consumer is not required to inspect validate the product, exactly. The consumer is not not even supposed to necessarily have knowledge of that. That's more Well, I'm not gonna go down that route. But But regardless, so so that would not be on them. Now, the question is, okay, great. So if I buy it from someone else, and it doesn't have the mark on it, that's one thing. But what if I build a little circuit, and I shove it in a J box and put it in the wall, and I burned down my house? Now we're starting to get into some different territory there. And what it comes down to is, maybe is what he came down with. So what's interesting is it's not like a guaranteed inherent Yes, like, if if the insurance company found out that I made a circuit that burns out like this amp right here, if my amp or you're just like cut for your 5000 or five kilowatt, you know, Bruce's Hey, hey, 10 kilowatt.

Host 1     59:31

Thank you. Yeah, if that burned down my house, that doesn't inherently mean that I would be denied coverage. What it comes down to is the legal term negligence, which I'm not even going to pretend to know how to interpret that. However, what it really comes down to is the insurance company proving negligence, like was there negligence on your behalf? Or did you just pull up I Instructables thing and follow it to the tee. And and say, Well, okay, I follow what was given to me online. So there's a bunch of questions that you would very likely be asked like, what do you do? Are you aware of electrical code? Are you a electrical engineer like these? Like, yeah,

Host 3     1:00:16

Do you use? I'm an engineer as a valid defense. On your device burning your house down? Yeah, right. Maybe not actually. Because that means you're a bad engineer.

Host 1     1:00:27

I don't know. Or an inquisitive mind one or the other. Right? Yeah. Will this burn down my house? Let's find out.

Host 3     1:00:36

I was just testing the scientific theory.

Host 1     1:00:39

There we go. Yeah. Test. Test complete. Yes. That's complete. Hypothesis. false false. Yeah. So So I guess the takeaway with all of this,

Host 3     1:00:53

Then I was kind of insurance companies want to make sure you didn't do it on purpose. Yeah. Well, they don't they don't want you committing arson. Right. Yeah. On your own property. Right. I'm pretty sure. Arson to your own. I don't know, legally, I'm not going to pretend to know them. I mean, if you burned down? Yeah, I mean, that's just insurance fraud. At that point, I think you burning down your house is called you burning down your house? Like that's just sort of how it goes? You're right, maybe arson means that you're doing it maliciously, to not your own property. Yeah, I think that's true. I think you would just be insurance. Making sure you're not committing insurance fraud is what they're worried about. Yeah. Well, okay, so

Host 1     1:01:35

All said and done. Like, I didn't know that. And I was a little bit surprised that, you know, if you I always thought that if you plugged in something into the wall that doesn't have the UL cert, then you are taking a risk. And you sort of are but you're not taking as much of a risk as I thought you were. Yeah. So

Host 3     1:01:52

All those all those Hackaday comments are wrong? Well, I haven't looked at them. Oh, they're like, Oh, if you plug in a device that's not UL certified, you're going to, you know, void your insurance policy, you know, okay, so actually coming up here soon, I have scheduled because we bought some new equipment at work. And in order to get the city of Denver to sign off on our electrical permit, we have to get Intertek to come out and do an inspection of the installation. Especially because this is European equipment. So we have a 240 to 400 volt, three phase transformer. So we drive on the left side of the road.

Host 1     1:02:31

Yeah. So I have the inspector come in, in a week or two, I'm going to ask him about that. Also, like, if it doesn't have, you know, the mark of the beast, and we plug it into the wall, like what kind of risk are we taking? I mean, obviously, that I'm willing to take a lot of risk with the stuff that I have here. But but just like a general consumer.

Host 3     1:02:59

I don't know. Be interesting. Yeah. Cool. So I found another topic on Hackaday that I thought was a really cool thing to talk about. And it's called component shelf life, how to use all that old junk, which is basically you have lots of old junk. Yeah, well, and that's, it kind of hits home a little bit. But it's really a question of how old is too old for a component? And how do you know that? And how do you deal appropriately with it? So macro fab goes through components? Pretty damn fast. So like, you don't have lots of old components? Right? Yeah, we have, like, no old components. Yeah. Like your, your component life is measured in days. Yeah, week at most. Right. Right. So yeah, it's, I mean, maybe your house parts, you have stock of house parts, but those are all things that are known to be accepted. Good for a long time. Yeah. Right. Like resistors. And stuff like Yeah, so we always check the data sheets for that. And like, the main thing with component lifes, is when you check data sheets, is they mainly care about how you're storing them. are you storing them in a cool, dry place? Are they in vacuum bags? Are they ESD? Protected? All that good stuff? Right? Yeah, yeah, for sure. Well, so we made moisture. So we make sure we store parts correctly. The problem you run into is like surplus stores. That's like when you go to a surplus store and I've been to a surplus store here in Houston and pick the like, electrolytic capacitor out and read the date code, and it was 30 years old. It's like what's the older you are? part that's almost as old as I am. And it's for sale. It's not any cheaper than buying one on Mouser. Yeah, yeah. Well, and that's that's an interesting you know, trap because like, if you go to a surplus store to buy Something that you want to be reliable. You have no guarantees. No, you know, it's like plugging a non UL Mark thing into the wall right? into your PCB. Right, right. Right. So, okay, so what are components to avoid if you're if you're looking for long life like in terms of like buying at a surplus store to avoid electrolytic capacitors 100%. Yeah. And I'd say look at like the leads of parts and see how badly oxidized they are. The good thing with like, through hole parts is there's usually enough like meat there you can like, you know, file off any x oxidization off like leads and stuff, but surface mount parts. The good thing with surface mount parts is usually the reel, they come on at least find the reel. So like if you're doing tape, they will have a date code. And I don't know, I mean, each manufacturer is different on what they recommend. So you'd have to look up like how long this x part is, okay, being exposed to air. If your hand soldering, like s&t parts, it's probably fine to just like, pull them off the tape and, you know, sobre Mon, but if you're going to refloat them, especially because a lot of times where I go to surplus stores, they just have the reels just like in a cabinet. And so it's like, okay, you can't reflow those because they're probably explode like popcorn. All the moisture, especially here in Houston. Oh, yeah, they're just sponges. Yes. So yeah, yeah, I you know, the old resistors I guess they're not even really that old. But like metal film resistors that have that like, ceramic outside on them? Those things last forever. I mean, those things are really light and the concrete result, I call them concrete. The, you know, the, the white stone looking? Yeah, they got ceramic goop on the outside, though wire around forever, old, old inductors like core wound inductors. And things like those will last a long, long time. Some film caps do have some life to them, because they are moisture sensitive. I guess it's worth like, look at how the component is constructed and packaged. Like if it if it's a dip kind of thing, then maybe, maybe not. But if it's got like a ceramic outside, then you're probably fine. You know? Unless you reflow it. They could be bad. Because ceramics are really good at absorbing moisture.

Host 3     1:07:32

Sure, yeah. Then you can get cracking and things like that. Yep. Especially when reflow but otherwise, they're fine. But this reminds me of buying parts off third party. third party sources, like for the open env cameras from way back in the day. Yeah, that's like four years ago, right? Something like that. And so the customer, they they ordered their cameras, sensors, third party, because we couldn't find them at all, from a authorized dealer at all. Right? And actually, the, the, the OEM manufacturer for the Park said that these were end of life, like three or four years before we were doing this project for them. Oh, yeah. When we were asking for them, like they were like, there wasn't even a trace of them in their factory. Yeah. And so our customer found these parts, and we're like, Okay, well use them because you're supplying them, but you know, no warranty. And they just would not reflow and they will just pop off the board. And these are like image sensors that are so the BGA is basically Yeah, and what we found out was the balls on all of them had oxidized. So they were the sensors were either they could have been either old stocks or like old sock and removed off of old boards. But since they were really consistent, we think they were just like poorly stored for five years, six years or however long they've been, you know, out of stock for. And so all the the balls were just all oxidized. Right? And I actually Hackaday did an article about this story, way back then. And the comments were amazing, like how to get rid of the oxidization. And the one that sticks out was someone told us to like sand the balls with sandpaper. I'm just like, Yeah, we got to take like a micron layer off of these balls. And we're gonna do with sandpaper for like the entire production run. Right, right. Yeah. And basically what we ended up doing was we sent we it's actually Steven Steven re bald, like 20 inch shop and they all worked and then we shipped them off and got them professionally rebuilt and then they all worked. Yeah, but man that was trial and error. Just like going down, like the eighth level of hell or something like that, for most factoring, you know, if you don't have the right equipment to reball, BGA, even even the BGA is they have a smaller number of pins. It's actually it's sort of easy, but not at all. Like the the actual actions you take with your hands. The physical motions is is easy, but getting everything to just be perfect is kind of hard. It's basically easy, and you just have to get lucky. Yeah, sort of. Yeah. Because because I did it. I did. It was like a little tray. And like no machinery or anything like that. It was like, like solder wick, and a little laser cut piece of something. I don't even remember what might have been stained, or something like that. Was it stainless? There was one was stainless and one was mylar? Yeah. Okay. Yeah. Cuz you had you. We built a little jig to hold the part. Right. I think we 3d printed that. And then you work the balls off. Yeah. And then you put the stencil down. And I think it was a stainless stencil and you put new paste down. Yeah. And then you had a bigger stencil that was mylar that you would basically shaped balls over and they would fall into the holes. And then you would shape the rest of the balls off into the tray, and then lift the mylar perfectly. None of the balls can move. Right. And then you had recalled your BGA. Right. Right. And then you had to carefully put that back on a new board and reflow that board and pray to God that it worked. And it work. Yeah. On your run. You had 100% yield. I was impressed. Yeah, yeah. It was a pain in the ass. That was an absolute. And then we're like, Oh, see, we just did 20 He can do the next 1000 of them. Right? There was 1500 that was what it was 1500 and we had to keep telling the customer it's like, yeah, our yield is like 1% You know, like, this is not gonna work. Yeah. Right, we got done. Yeah. Yeah, you know, it all worked out in the end. But ya know, that's another story from the depths of the the gray market like you the thing is, you don't know. Okay, so if the if the manufacturer of the parts themselves are like, I have no clue wherever these have been, they could have sat outside in some hot warehouse, you know, open warehouse in some country somewhere, you know? Yeah. Who knows? In the rain? Yeah. You have no idea walking up to snow both ways to school.

Host 1     1:12:37

Yeah, exactly. So, don't don't buy stuff from the gray market. What's the point? There's no point. Don't design

Host 3     1:12:43

A product that needs something from the gray market. That was the problem was is they were using an outdated sensor? Well, okay. Then for their next iteration, they upgraded to a new sensor. And guess what we had, like almost 100% yield. Okay,

Host 1     1:12:57

So here was the argument, and I can see it, but it I think the lesson was learned here, that particular sensor that we had to reball had JPEG compression built into the sensor cran. They, they didn't want to move to another one because it did not. They discontinued the JPEG compression. That's why they were so adamant about that. And they wanted that in this particular revision.

Host 3     1:13:22

So yes, that was the reason why, yep. But their next product, they use a modern sensor that the company made and they had the compression. So Right. Yeah, they just implemented it elsewhere. So okay, whatever. Cool. So our next topic is EU bands lead solder for real this time. And so back in, like, oh, five or Oh, six, the EU came up with Rojas much to the demise of electronic products everywhere. So I'm gonna go a little bit of a rant on Rojas. Do it now. Do it, okay. So Rojas it means good things get led out of the environment. LEDs, bad. LEDs bad get the little LEDs bad for the environment. The problem with Rojas was how much how lead gets into the environment is the majority of it gets in through car batteries being tossed into landfills or on wherever. car batteries go instead of being recycled, right? Rojas exempts car batteries. So you got like 80% of where your lead comes from into the environment. Yeah, it's not even been touched.

Host 1     1:14:42

Yeah, yeah. And so, you know, Rojas is more than just lead. I think there's what 13

Host 3     1:14:50

That? Yeah, it's got some heavy metals but the big thing was lead. And that's my biggest thing is like okay, you could have pushed at this was actually going to be good for the environment, you would have pushed car manufacturers in that industry to change off of lead acid batteries. The change is something new, or something different. But they didn't. So it's I hate getting political, but it was really a feel good political wake for Okay, so I'm actually, so I'm on Rojas guide.com right now and on the FAQ page and there is a question, does Rojas applied to batteries? And like immediately under he goes, No. And although although it gives a give some, some qualifications here, it says all batteries, regardless of type or application are covered under the EU battery directive, which there was one that came out in 2006. And another one that came out in 2013. That basically restricts certain heavy metals lead mercury and cadmium. Yeah, but but still, yeah, no, you're right, like Rojas? Like, yes. almost guaranteed. Rojas was driven by lead. And, and it's not attacking one of the main sources. Yeah, it wasn't. It wasn't addressing the big elephant in the room. It was attacking the mouse's, right? Right. So regardless, the thing is, hobbyists and commercial users could still buy led for soldering. after that. It just couldn't be in consumer products. Fair enough. But as of March 2018, the EU reclassified lead as being toxic to reproduction. And because of that designation, it can never be sold to the general public. And so you can only be a commercial user. And that went into effect February 28 of this year. And so there's a Reddit thread on the electronics subreddit of a guy went to his local electronics store slash surplus store, and they were completely out of leaded solder. Hmm. So like, I can still go to Home Depot or to EPO, which is a local surplus store here in Houston and I can still buy leaded solder. I can still buy online. Oh, yeah, yeah, I you can buy it on Mazur. Yeah, so EU they first banned memes and now the banning or leaded solder. I feel that there should be a meme about banning leaded solder when we can make one but they just can't view it. Yeah, they can't. They're not. When does that go into effect? That's articles. What it was article 13. And then they omit they changed it now. It's article 17. Something like that? I don't know. It's actually about memes. It's about copyright. But the best meme is it's about memes. Yeah. Which is. So again, my 20 pounds of vintage caster 4060 or 6040, leaded solder just keeps going up in price. Yeah, maybe I can sell it in the black market and EU sell it in like six. And we already said, Oh, I mean, black mark, it's gotta be metric, what six inches in metric.

Host 1     1:18:23

Like 12 or 13. I only use freedom units. I live in America. So it's like 14 centimeters, something like that. Maybe 15? Actually, I'm doing I'm doing 100% of everything in in metric. Now. we base it

Host 3     1:18:36

On your board designs? So I do. I do my. So on board designs. I do my because most parts are in metric now. Like the ICS. Yep. So I do. I do my placements, like where I put my parts, I do that in metric. And then I tried to make it like, I put my parts in placements of like, what the pitch widths are. So it's like point one millimeter a placement so that way, when I route that level works, then I do my fan out and metric. And then I switched to Imperial for everything else. Especially if it's connectors. I do Imperial for connectors, like their locations. Oh, okay. Because I've yet to do fixed Drean or anything like that metric. Imperial is gonna be a lot better for like people building stuff in America to handle

Host 1     1:19:27

Oh, for sure. So, okay, here's the thing. I realized a little while ago that

Host 3     1:19:35

I the only thing I was using Imperial for was my board outline, like my actual board size. Because I had a good feel for that. Everything else it did metric because every data sheet was in metric everything like and I was like This is so dumb. Why am I doing all these conversions when I could just do one conversion or just kind of learn gut feel for metric. So I do metric everything. However, the only thing I don't do that The trick is tolerancing now because I don't have a good feel for like I can I have an understanding of what three thousandths of an inches or 2000s but like two thousandths of an inch in metric is like I have no clue like like it's not that I don't have a clue it's just like I don't have like a an understanding or gut feel for that you Parker's I can tell he's googling it right? point oh five millimeters. Oh thank you. Thank you. Yeah, so like but if someone were to come over to me it'd be like is point oh five millimeters good enough to over or undersized this hole? I'd be like, I have no freaking clue and they actually were three cuz it's the metric system you shouldn't be using decimals

Host 1     1:20:42

Should be Oh, two mil should be 50,800 nanometers. Oh god. Yeah, they're gonna get that rant. Yeah, I'm not touching that. I'm not touch how many European listeners are we going to lose? On this episode? Hey, they should be gaining because we're we're we're admitting that the metric system is superior in almost every way other than tolerancing

Host 3     1:21:09

Board outlines no just whatever you used to know I mean that's why I do I do the I do the fact that the reason why I place electrical components on metric is so that fan outs work well. Yeah. And metric and then components and stuff that interface with the outside world. Most my boards are in our in, you know, America, they don't go buy pinball machines don't go international. And so it's like, okay, you want connectors that are you know, metrically or Imperioli space, so you can have fixtures that work with it and imperially space that just sounds like it's got some gravitas to it, you know? That's like that. Yeah. metrically spaced doesn't matter just doesn't have as much ring. And we can even go deeper. It's in purely space like in the old EU. It's ours. Oh, yeah, there we go. But no lead right. Now and the older you can have led Oh, okay. Well, but in America, we can have led spacing right now. We could still have led spacing. You know, actually the crazy thing is a way tangent before we finish this podcast. Yeah. Is it's interesting to look at leaded gasoline and violent crimes. You've looked at this before. That sounds like something that you'd see on Freakonomics. Actually, I think it might have been a Freakonomics where so lead in small doses causes people to be more agitated, and aggressive. Okay. And so when we basically the EPA, I think it was the EPA ban leaded gasoline to go unleaded. There is a sharp drop. And like the year that follows in aggression, like in violent crime, like noticeable drop, it's a huge nervous, like, it's actually like one of the greatest, like, changes in violent crime as from changing off lead to gasoline. That sounds like a correlation without causation. Kind of thing. But but it's such a steep change, and we know lead causes aggression. Sure. So yeah, that's an eyebrow razor. For sure. And then the other big one that's coming up is changing. Like street lamps in like public space lighting to blue lighting from yellow sodium lights. Okay. Yeah, because they use sodium lights because they were really efficient at making a lot of light, right, with one bulb. But the problem is, they're usually you know, that yellow tinge, right, slightly yellow, reddish tinge. Whereas when that actually causes more aggression, humans as well, whereas blue lighting does not it causes a more passive effect. And so when you ever you go to like, if you look at like, night limited bus stations and stuff, yeah, modern ones are blue, like led cool, lit. Cool. This is a cruel in bus station. So calms down your aggression, huh? Yeah. Okay, get this. nother tangent. This I was I was dating a girl back in college, who lived actually in Baltimore. And she lived in a not so great neighborhood in Baltimore. And I went and hung out with her was that the entirety, entire place of Baltimore, I, hey, I'm not going to offend the entire city of Baltimore here. Their city is great. None, no, get this. At that time, this was 2008 or nine. Baltimore had some interesting, I don't know, ideas on on handling crime. And one of the things they did was they set up on the street lamps, these blue blinking lights at intersections that had high crime. And so like if there was a murder, or if there's something that happened there, which that was a fairly common thing. They put a blue light there in an effort to like, kind of like, ward it off and say, hey, there was crime at this location. What it ended up just becoming were crime beacons where people would be like, yeah, just meet under the blue light. Like, it ended up attracting tons of crime. It's like, well, that's where the Fight Club goes. Yeah, it was a horribly failed broad, like, you know, initiative that they tried it was it was actually kind of funny. That's interesting. Yeah. So you go blue lights. unfortunate, but that is a very interesting project.

Host 1     1:25:32

Yeah, maybe they have still do leaded gas there who knows? Yeah. Or they just eat leaded circuit boards maybe we just need to go the way the EU and just get rid of it. All right.

Host 3     1:25:48

I mean, we build I don't know of anyone that builds products using leaded solder unless you absolutely need it like NASA or some aerospace stuff. You know, the last place that I even heard that still, you to use LED solder told me last year that they're switching over. And they did safety products and things like that, but they were like, yeah, we're just we're just, it's easier to get Rojas stuff now. So like it's actually harder to find a cm that does lead. We actually in our entire partner network. We have one one cm one partner that will do let it Yeah, they basically I have an entire facility that all it does is lead right and it's because the moment you put it in the same building, you're gonna have cross contamination. Yeah, it's game over. It's like, here's a tangent for you. It's like sour beers. If you if you if you ferment sour beers in the same building as you do like standard beards, your standard meals will become sour just because that lactobacillus yeast mean it's gonna get better, right? Well, yeah, I

Host 1     1:26:55

Mean, I like sours but yeah, so you can't you know, cross contamination. Yep. So before we're getting onto another tangent, let's end this podcast. Yeah, like Yeah. All right. So that was the macro fab engineering podcast. We were your host, Stephen Craig and Parker Dolman. Take it easy. And don't eat lead kids don't eat the lead chips

Host 3     1:27:30

Thank you. Yes, you our listener for downloading our show. If you have a cool idea, project or topic or lead let Stephen and I know Tweet us at Mac fab at Longhorn engineer or at analog E and G or email us at podcasts at Mac femp.com Also check out our Slack channel. And if you from the EU, we'll be posting some really cool pictures of leaded solder reels. If you're not subscribed to the podcast yet, click that subscribe button that way you get the latest episode right when it releases and please review us wherever you listen, as it helps the show stay visible and helps new listeners find us

Transcribed by https://otter.ai