- Hoppes #9
- for the printer seems to work well
- Will it hold up long term?
- Still waiting on fans for the printer and other various parts
- Jeep Fan Evolution
- Connector Woes
- CNC radiator
- 5 gallons of DI water with automotive antifreeze
- Warmed up after 3 hours of cutting
Rapid Fire Opinion
- The Myth of the three capacitor values
- Could this be the next project for the MEP?
Special thanks to whixr over at Tymkrs for the intro and outro!
About The Hosts
Parker Dillmann is MacroFab's Co-Founder, and Lead ECE with backgrounds in Embedded System Design, and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. He also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components. Parker graduated with a BS in Electrical and Computer Engineering from the University of Texas.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Host 1 00:00
Hello and welcome to the macro fed engineering podcast. We are your host, Stephen Craig and Parker Dolman. This is episode 220. Oh, yeah. 220 episodes. So guess guess what I've been doing this week CNC work. Well, a little bit of that, but also searching for connectors, which oh no done on I'm like the worst thing ever in fact, so filled G on our on our Slack channel even suggested the other day like making a shirt. Gosh would I now know is escaping me what is it? Oh connectors destroy bonds or connectors break bonds or something like that, which I feel like that's freaking awesome. I love the fact that that because it's so obscure, so ridiculously obscure that only an engineer would know it. And that's sort of like a thing that we've talked about multiple times on the podcast. So it's like, pandering to like 500 people maybe? No, I think most electrical and mechanical engineers, they would get it they will get that because mechanical guys deal with connectors all the time, too. Because you had, you know, mechanical guys make the connector. Yeah, they made the connectors against Yeah. Yeah, I think they would both get it. I also like your suggestion was, what was it? Can you scroll up? Oh, it was a shirt that said fuck connectors. You know, I think that might have a little bit wider spread, like, appeal. But I don't know. I the thing is like, Phil, if you made that shirt, I would I would wear what I would wear. I would absolutely wear that just because it would be a weird conversation starter. And then it would also be a conversation Ender, like guaranteed like someone's like, why are you wearing that? I'm an engineer. Okay, well, they're walking away.
Host 3 02:09
All right. So why why are you looking for connectors this weekend? Okay, so this weekend, Ben? Okay, so we've been working on some, some designs with a buddy of mine. And one of the things we were doing for a while is, is some PCB mounted potentiometers. And so like there's, there's a bunch of great offerings for PCB mounted potentiometers that are 90 degree, so you stick them on the end of the board. And then so your main board fits on one face of the chassis and then 90 degrees, so that your, your potentiometers poke through some side of the chassis. And there's nothing wrong with that. I've done plenty of designs with it. But we saw some images of some guys who were doing panel mount potentiometers, that have just three wires soldered to the lugs, and then they go off to a small connector that pops to the board. And I really liked it from a manufacturing standpoint, because it makes things a whole hell of a lot easier, you don't have a 90 degree angle to worry about things lining up with holes, it also doesn't force you to have to design around specific standoffs for your mainboard. Because if you think about the stackup of everything
Host 1 03:21
And manufacturing process, because if you have those 90 degree potentiometers, or any kind 90 u connector, you got to put your board in at basically a 45 degree angle into your chassis so that the connectors hit the front plate and then tilt down and slide in.
Host 3 03:36
Right. Right, it doesn't really work that well, especially if you have items on the main board that are having to go through the top of the chassis. So like it just, it's that's not a particularly good way of doing it. It totally works. But it also thinking, you know, a lot of people, in my experience, don't really put a ton of thought into this. But thinking down the line, if anyone is ever going to have to pull that board out for any reason, like if you need to service it, I've found so many designs where they're really well thought out except for servicing. Like they are just not serviceable. In fact, we were just talking about that frequency kind of the other day like it's not serviceable at all you have to basically break it to get it out and then repair it to get it back.
Host 1 04:21
We've totally taken things apart where like in your in your example potentiometers were a 90 degree potentiometer were soldered in after the board was put into it. So you can't actually get it out without having to basically de solder this component. Yep, yep. Yep. So so all said and done. I liked the idea, even though it's technically more work. I like this idea of having a potentiometer that has three legs, or three wires coming off to a pigtail that just connects into the board. You know, another another kind of manufacturing I guess. I don't know, like hack, I guess you could say, but something that's just been going through my mind is if you have different values Use of potentiometer you can have different colors of wire. So if if it's three green wires coming off of this pot, that means 100k. If it's three black wires, it's one meg like you just don't ever mess them up. And if you have them all in bins next, you know, for whoever's doing the assembly, they just grab the black wires for this one. And I love that too. Because when it comes down to like, creating instructions, or even drawings, for assemblers, you just make your drawing, put an arrow to it and say, green wires. You don't even have to tell them put 100k pot here, you just say put the green wire one here. Like I always like the idea of like, it's not it's not an idea. It was not the concept of like, dumbing things down because people can't understand them. It's dumbing things down so that no mistakes are made. And there's just there's not really a need to for an operator to know what 100 Gay pot means. But they can certainly know what a green wire pot is, you know, or green wire assembly. So, so what this all boils down to is, I love this idea. Now I gotta go find some connectors. And I'm seriously like, oh, shit, like, I hate I hate looking for connectors. So the worst. But okay, so one of the things, well, two things that I knew I wanted with this, well, first of all, they're potentiometer. So there's, there's only three legs on this. So I need a three position guy, I want it to be foolproof. So I want it to have a indicating tab on the connector, because you can't put it in backwards. And then And then other than that, it just needs to generally be cheap. You know, the, it would be kind of nice if the connector was black, because we do have like a color scheme going on with our assemblies. And White is not part of that. So it'd be really cool if if it was black, but I don't care. If you order enough of them. You can get them in any color. Probably. Yeah, we're not at that kind of quantity right now. So So regardless, one of the first places I actually started because I knew what the connector looked like, but I didn't know the name is Amazon. And the reason why is because like, Okay, so you've seen a bazillion of these. You've seen a bazillion of these, but I'm gonna I'm gonna ask you like, Do you know the name of it? Like, okay, so think of an RC, anything helicopter, car, plane, whatever. Those battery packs that you get with them usually have two, two wires coming off of them. And they terminate and a little white connector. What is that called? So? I don't know the answer, but I know what not the answer. Okay. Yeah. All the other connectors. Yeah, no, because people call those JST connectors. Yep. But that's the manufacturer or the original manufacturer of them. Right. Right. Because I bet you billions of companies over in China probably make them now. Yeah, I mean, it's header pins and and a plastic shroud. Right. Yeah, doubt there's a, ya know, they were making tons of them. And that's exactly it. Okay, so everyone calls them JST connectors, like that's a type of connector. It's not a type of connector. It's, it is an original manufacturer. But like the problem is like, if you go searching for J S T on Amazon, you'll find a ton of JST connectors. But then if you go over to Mouser, DigiKey, and you search for j is T connector, that doesn't exist, a manufacturer called J S T exists, right? But then it's really kind of difficult to find, to find specifically the connector you want because J S T makes a ton of connectors, and hence like, well, this is the normal path of searching for connectors, right like you're trying to search for for shit and then it gets
Host 3 08:47
Up just a giant pain in the ass. And I think I think you've actually talked to them before. At my first job in Houston. I did talk with Sam tech a lot because there was a SAM tech rep there. And I remember Sam tech came to Mac Feb a few times. They have like gorgeous catalogs. I love their catalog. You could just leave through it and there's like full page color with like 3d renderings of all their connectors and stuff.
Host 2 09:15
Well the best thing is their website is because it has a you basically pick your family and then you can just be like, I want these settings and they're like, here's your partner that you can order. And you can order the weirdest stuff and I usually have a couple in stock that you can get samples of which is like bizarre.
Host 1 09:32
Oh dude, we got samples at my first job of connectors where they didn't even exist. They hadn't they had them in the catalog, but nobody had ever ordered them. So they literally just sent them from their engineering lab like they were like they're not the same color is that okay? We're like it's fine whatever. Yeah,
Host 2 09:52
There are. There are only downside is they are a little more pricey than other companies but all their connectors are all everything I've Use from them is awesome. Yeah, there's just got to be like, Well, you said they expect, you know, it is going to cost a little bit more, but for good connectors, that's just
Host 3 10:09
That's, that's how it's always going to be. In fact, you know, like, when we were talking about connectors bus bombs that shirt like, that's kind of why like you kind of want to spend money on connectors. And I don't know, as as with as with so many things.
Host 1 10:27
In fact where I think we're going to talk about this later. Yeah, it all depends on your design. And it all depends on how deep you want to test everything, how much effort you want to put into it, because like, if you're fine with just testing the hell out of a bunch of really cheapo connectors, and you're fine with whatever failure rate those connectors will come up with. And, you know, that's up to you. Okay, so regardless, like I'm going to samtec looking because I know I want a jst style, but you can't just type in jst at some place. So I go to Sam tech, and I looked through all of their fancy pictures, and they, of course, they don't have this kind of thing. They have like tons of like high speed like 5000 pin connectors and stuff like know what I'm talking about analog potentiometers stuff, like how the sandtex naming schemes are like Tiger connector and so like he was like, What's that even mean? The Joe exotic connector? Yeah, the joke's Tiger King connector. Yeah, cuz like, they're they're kind of art scheme is
Host 1 11:31
What? tiger stripes all over the place. Yeah. Maybe it's a different company, because I'm actually on sandtex. website. I
Host 1 11:39
Don't see Tiger connection. Maybe I saw that somewhere else. No, no, I've totally seen that too. Yeah. Okay. Like Tiger bite speed. Oh, I think it is Tiger bite. Yeah, it's all kinds of stuff like that. Yeah. In fact, one of the first projects I worked on we samtec made custom connectors for us, for our ribbon cable that connects to a touchscreen. And men. To be honest, I'm not just trying to, you know, praise Sam tech all the time here. But like, they were super easy to work with. And like, I thought getting custom connectors would be like, Oh, my God, this is going to be the end of the world difficult. And it was like, no, we want a custom connector. And they're like, Okay, here you go. Cool. So, regardless, so I did a bit more searching. And it's funny because I ended up finding the sort of general part number I wanted, from an image from a kit off of Amazon of a crimp kit that had JST connectors in it. And they have like, they had some lettering on it that I was like, Wait, that's gotta be like a family. And then I went over to jst and found out that the x H Series is kind of that like, that battery? That one that everyone knows, but doesn't know what it is? Yeah, when you take apart some cheap electronics, the battery has that connector on it. Yeah, that. Yeah. And that's just the thing. Like, I know, my specs for this thing, are really minimal. I don't need a very special connector, I need it to be cheap. I need it to hold. And I need it to be able to potentially take up to like 200 volts. And these connectors do exactly that. So it probably won't it most of its life will probably sitting more around like 10 volts, but it is possible that they could go up to 200. And luckily, this x H Series is a 2.5 millimeter pitch. It comes in multiple pin arrangements. And it does exactly that. So
Host 3 13:43
How much isolation you need for 200 volts? Oh, these these things have like 1000 volt isolation. Okay, so yeah, that's no problem. And, and usually I use like either 22 or 24 gauge wire in these kinds of assemblies, which these are meant for that also. So basically, I'm looking at battery lead connections.
Host 1 14:05
I just thought it was funny that like,
Host 3 14:08
What 99% of the time when I search for connectors, I know the connector that I want in my head, I just have no clue what it's called. And it's so hard to find. And and on top of that I'm sort of doing a little bit of like, I'm establishing some groundwork for future designs also. So I don't want to just pick some rando connector that could go out of stock or when I kind of want to pick a family that I'm happy with that
Host 1 14:34
I could build some footprints around and then in the future, if I ever want to pull them out, I know that they will always be there. Also, something interesting to note, if you go to JSPs website, it looks like the website came out or like I don't know 1996 or something. It's not that old because I have another connector company called for you remember for UConn. Oh yeah. Their website looks like comes out of the mid 90s There was this like or Early 2000 JSPs fine, maybe I'm not giving them enough credit. The only problem is if you go to J's T's website and you're like, show me your products, they're like, Okay, here you go. And there's a list of them. And it's like 47 pages of just Yes. But like, like individual lines have connectors like, wow, like, how am I supposed to search your connectors? Am I supposed to just go through 47 pages and be like, That one looks good. That one looks good. What what? I wanted to have a catalog. It's like a master thick catalog of just connectors, you know, with the way that their website looks, I would guess that they probably actually have a really badass catalog. And they still rely on that to get their their sales because the website, maybe I didn't find where I needed to go on the website. I didn't spend a ton of time because it was just like, Oh, good. God, I don't have enough time to search pages and pages of, you know, just random little connectors. And and once again, I complained about this a few weeks ago, I don't know your parts came. So like, if you just have like, one two BH three dash 470 h x? I don't know what that they have 10,000 of those in stock for three cents apiece. Right. Okay, cool. So now I have to look at data sheets and be like, is this gonna work? I'm actually on their website. Right now I'm looking for a 1996 or 2000. Interesting, I don't they don't have like a because usually, a lot of times companies like this is their like, here's our catalog that you can get. They don't have anything like that. Symantec has it really close, in my opinion to be in, right? Because if you go to their website, they're like, find a connector. And then they show pictures of things and just like does your connector does the connector you want look like this? And then you hover over and click and then it goes into that family? And it's like, do you want it to look like this, this or this? Like you can like visually search and use that as ideas or you already know what type of connector you want, you know, so it's like, it feels like JST connectors are like, it's like the Xerox or Kleenex of the character world. I think that's why people just use that name. Like, the what
Host 1 17:11
They really mean when they say when someone says JSC, they really mean JSC x H.
Host 1 17:17
A, basically. Yeah, and that's the so here's the thing, JST connectors are big enough to have their own Wikipedia page. Which Okay, so when you say jst style connectors, you think
Host 3 17:33
Like, you think that jst means something where no, it doesn't say, in fact, let me see here, I just looked it up right before the Yeah, if you go to Google and just type in JST connector, there's a Wikipedia page for it. In fact, I'm gonna put the link in our show notes right now.
Host 1 17:52
Straight up and and the Wikipedia page is infinitely more helpful than so the Wikipedia page has like charts, and it has data sheets for all of these different models, like, go to the to the Wikipedia page first, and you'll be able to get to what you want. It's like, go to the Wikipedia page, measure the 10. Pitch what you want, and then you can find what you want the Yeah, yeah, it's so it's actually really helpful. It's super helpful. Actually, so it's funny. I kind of wish that JSPs page just linked to the Wikipedia page. Yeah. Ah, are you looking for a JST? Connector? Check here first. I like how it says, For the x h use in mini RC batteries. Yep, yep, exactly. Yeah. And they also say not 0.1 inch pitch. Oh, yeah. Because it's 2.5. It's two mils. Too close. Have you ever run into that before? Oh, yeah. I'm sure I'm sure at macro fed that happens quite a bit actually where people accidentally designed for the wrong pitch. Not as often as you would think. You would think that'd be a bigger issue but I think most engineers look out for that. Everyone that more and in hacking stuff together? Oh, like you have 1.5 versus 2.5 for it. Yeah, basically. Because in pinball we actually use 396 A lot which is 156 mills. So at VH series, what was called the motor, the molex KK type connectors. Okay, you know what? Okay, so that's another one real quick. I'm gonna dog on that for just a quick second. Because there's this nebulous term JST connector and in your head, you know what that means. But it doesn't actually mean that when people say MOLEX connector, that that's the exact same thing, right? Oh, yeah. He's the builders. They just think it's the four pin. accessory. What 12 volt ground ground five volt connector is what that's what they call a Molex connector. Right. And in fact, you know, MOLEX connector probably has its own Wikipedia page too. And I'm sorry, it's not an X just did the same thing. It's actually the molex kk 396 is what we use a lot in pinball. Yep, yep,
Host 3 20:22
Straight up. MOLEX connector has its own webpage. And it says MOLEX connector is the vernacular term for a two piece pin and socket interconnection, which, you know, I'm not not to just get all like hyper anal about being correct, but it's, we should be calling them pin socket connectors. And in fact, okay, so that's just a thing. If you go to Mauser, in fact, I'm going to do it right now. If I'm going to Mauser and I go to connectors, and let's see here. Okay, so I just clicked on connectors, and I'm looking at all of the options that they have on here. They have a Molex, no, Molex is not there and jst isn't there. So
Host 1 21:06
You can search for both of those brands in their filter categories. But if you're like, oh, I want a Molex connector. There's not a category for that. So it's confusing, but they do have a category for pin and socket connectors. Which are they're not the Xerox or Kleenex or connectors yet. Not yet. They don't have a category or, or something named after Molex is a is a producer of molex connectors. But a lot of other producers are also producers of molex connectors. When you get someone on the podcast that is a connector from a connector company a connector head. Yeah, yeah. Be good because we had, you know, capacitor. You don't want to we haven't had resistors yet either.
Host 2 21:53
Because we have semiconductors, a lot of semiconductor designers.
Host 1 21:59
And capacitors, but nothing, nothing else yet. You know, I one, one thing I would love to discuss is the noise floor in various types of resistors. And like, when when you should like, what's it? What's a good rule of thumb for thick film versus thin film? So that's actually what I was about to ask is like, when doesn't matter? Because in low volume, they're the same price. And high volume. There's a slight difference in price. It's like, okay, I, in my mind, it's only like, if you're building a billion of these things, then it matters. Unless there's a performance reason. There are performance reasons, but I'm not going to pretend to know all the details behind that. Yeah. Yeah, that would be fun. Yep, we're gonna have to maybe Panasonic, that'd be cool. The erh or the E, what is the
Host 1 22:54
BR J? Er J? That's it. Yeah. Or is that the no er j is is resistors. Yeah.
Host 1 23:04
So many times. Um, so a couple of weeks ago, I was working on that. The Zebra printer stuff. Yeah, I finally deployed that live to some of our operations guys, and certainly out really good.
Host 2 23:20
Host 1 24:45
It's basically the gloss right over over the structure. The shellac, yeah, the shellac over the structure, so it makes everything look fancy. It actually does make it look like a like, ICC took like 10 minutes to make it look like that. But compared to a the, like, just straight HTML, it compared to JSPs webpage. Yeah, it actually looks,
Host 2 25:11
There's a huge difference. Like, it's mainly about, like,
Host 1 25:14
Host 1 27:08
Usually, it's like a circuit board. And you're like, oh, what does it do? It does blah, blah, blah. And like, okay, cool. Like, yeah, like, cool. Yeah. But like, this is like, it was like, wow, that looks like it looks like a modern UI. And I'm like, Yeah, I just slept 10 minutes to paint on it. Nice.
Host 2 27:25
But it was like, how do you get semantic UI to work with a local application?
Host 1 27:32
Like this, too. So super, super practical knowledge? Yeah, very practical stuff. Like if you do it this way, and why it's done this way, as opposed to just like, actually, that's the thing. There's not any tutorials like that out here. You know, I'm surprised because like, you would think that that would be something that at least somebody's throwing together. But I mean, you're the guy who's doing it. Yeah, it's like, I need to I, this is what I need to do. I think a lot of times, it's most of the I think we've talked about this before, but it feels like software developers assume too much about the people who are using their stuff. Yeah.
Host 2 28:15
Whereas hardware developers, like, especially if you look at data sheets, like the we talked about data sheet is like, you if you read this, you understand exactly how this thing works. That doesn't really exist in the software world.
Host 1 28:31
It goes back to what I was saying earlier about, like, you know, an assembler might know what 100k pot is. But if you're writing documentation for them, don't say get the 100k potentiometer. Say get the one with the green wires. Yes. Or the one in the blue bin. Right. Right, right. Yeah. Now, there's some stuff that's different. Like there's some really good documentation for depending on the API's and stuff. But sometimes, some API's like, that I've been working with, like, I just have to hit it, hit the API endpoint, and get it and be like, and then change my parameters a bit and get something else different and be like, Okay, I changed something. And it resulted in this different outcome. Why? Because that's not documented. And a lot of times, that's not documented at all. Yeah. You know, okay, so real, real quick tangent, but still on point. I was I was talking to Roz the other day, who's been on this podcast before he Josh Rogier. Josh Rosen. Yeah, he's, he's been getting into a lot of designing circuitry on his own now. And he's doing a fantastic job for someone who, who has just basically self taught, and he's making a bunch of his own boards. And he he's really got a good grasp on circuit theory. But one of the things that he's kind of anal about in a good way is like, Oh, I'm focusing on like, where do I ground? What circuits? How do I, he's he's starting to crack that shell on like, Oh, I get what my circuit is doing. But there's this magical realm of zero that I don't understand. Right? like this huge like, playing in world of like underworld of zero. And so he and I have been working on schematics together and he's like, Hey, so I've designed this circuit. And I know because I'm the one who designed the circuit, I know where these grounds go. And I know how I want them to go on the layout. But if I'm the one designing the schematic, and I'm handing it to you to lay out, how do you know where I want my grounds to go? Because I designed this, this schematic and was like, Hey, that's a really great question. The answer is, I don't. And the answer is also like I sort of do at the same time, because like, it's kind of up to me. Like, if I'm the layout engineer, it's up to me to interpret the schematic. And in a way that makes the most sense, electrically, but also, if you're the schematic engineer, you need to portray that to me properly. And we started getting into these conversations about how most EDA tools are excellent at one or the other, you know, they're excellent at like, Oh, this is my schematic, and it looks really, really beautiful. Or here's my layout, and it looks really, really beautiful. But a lot of times like do you know of an EDA tool that would instruct a Layout Designer on like proper grounding schemes? If you were given a schematic? No, but how the schematic how I would treat that is on the schematic end is what I do for myself, because a lot of times I'll do schematic like, a week or two weeks before everyone will touch the layout, because I'll go through a couple of different iterations on the schematics. Yeah, that makes is, is you call them different nets. So there are different nets? And then you so let's say we were doing a star ground? Yeah. So you would have all your different modules would be a different ground, or different sections, and then you would have a node that everything those nets go to. So you could and then you could actually write notes like this is a star ground, and should be pretty obvious. But if you're doing planning plans just need more ground, because then that's what they are, they're just going straight to a plane. Sure. So I would say call them in different nets, then maybe having some notes on like, what kind of grounding scheme you expect. That should be sufficient, least from the projects that we do. And I think that's a totally reasonable way of handling it, especially if you work on a close. If you work closely with your engineers, then like you guys can all establish, hey, if we see XYZ thing and a schematic that means this. But it's like the difference between an and like, some stuff I do you have an analog and digital ground. And you call them two separate things. So when you go to layout, you know, oh, analog grounds, usually are grouped together into digital grounds are grouped together. So you start separating out the parts and that kind of stuff, right? But here's the thing, a lot of EDA tools don't distinguish between those. It just says, oh, that's all net zero, which is great. No, you would you would call the nets, analog ground and digital ground in this case, and then you would have a a bridge between them that would connect them. Right? Correct. But but a lot of EDA tools don't force that bridge.
Host 1 33:09
Oh, no. Well, that you let you call any net and net, right. So but that's just the thing, if you if you were to individually net, each chunk of ground, let's say you had four or five different grounds you wanted to have, you know, chunked together, you could do that. But then your EDA tool isn't going to say like, Hey, at the end of the day, these all need to be connected, it's happy, because it's like, oh, all of these are individually separate, it still would still be up to you to connect them. Yeah, in the schematic, you'd have to connect them. Right. But as soon as you connect them, then they're not individual anymore. You can connect them through at least an eagle, you can connect them through a node that doesn't treat them separately. It's treat them separately. And then they are connected through a like a pad basically. So would that be like a full on? Like you connect them through a component? Kind of? That's okay. A node is what I would call it or like you can connect it to a via they're connected to one via Oh, so they still have to be connected to an element of some sort? Yeah, yeah. Okay. So in my, in my ideal situation, you could have like, say you have your analog and you're talking about like you want to be able to have virtual nets. i Yes. I want to have two nets that come together to make one net, but it still treats them separate. separately. Yeah, that would be super ideal. And I've never run into an EDA tool that doesn't Yeah, you're right. I haven't run it. So in that case, what you would need is you'd have to have I don't know how the trace calls him but rat nests. Like local retinas. Well, you would have a different retinas you'd have at hierarchical retinas then yeah, yeah, you'd have redness, which was all the nets and then you could have a top level Net? Excuse me? That's right. And that top level net would want the bottom level nets to connect one in with one net with one connection basic. Well, it could be depending on your granting scheme, because this is exactly so. So all of this, we're arguing about something that we even know. And this is exactly what Roz and I have been talking about, yeah, you know, 1000 miles apart, like, he's like, How do I tell you these things? And the whole point of me bringing up this whole tangent was that like, well, I can just look at your schematic and I know what you're getting at. But the only reason I know that is because I've done it 100 times, you could, you could, on a schematic draw spots, or boxes, because I've done this before to like, like, these, I'll draw like, let's say had an analog front end and a whole bunch of ADCs. And like, okay, these grounds, I'll draw a box around them. And like, I'll draw the schematic, you're talking about the same thing, though. It's like, I can look at your schematic and see that right, right. But I'm like, okay, these grounds need to be their own isolated analog ground from this analog ground, because this is like the front end. Yeah, that's like exposed to the world. And it shouldn't be directly connected to the back end of the ADCs. Right, you don't want weird, we had ground flown all over the place. Yeah. And you know, what I've done in the past. And what we actually ended up agreeing on is, we did dip trace where, you know, groundless ground, we have one ground net, but on the schematic, what we did was we just put numbers next to the ground symbols. So it's like these foreground symbols all have the number one next to them. So they all get connected together, these next five have the number two next to them, and they get connected together. And then the group one and the group to get connected via one trace. And we daisy chain that way. And so you know, that kind of goes back to what I was saying is like, in your group of engineers, you can come up with whatever scheme you want. But it's also like, it's not a scheme that everyone else is using, there's not really a good scheme for these kinds of things. Yeah.
Host 1 37:01
I bet you're like, Cadence it does it for for a $3,000 add on package. or multi node ground package. So I've been doing a little bit more on the CNC also. Yes. And I actually, this weekend, I cut a three hour cut on the machine, I did a whole bunch of different various pieces. And I ran into something that I wasn't necessarily expecting to run into. It's not necessarily an issue, but at the end of the whole cut, I was like, cool, this is great. And I put my hand on the spindle and the spindle was relatively warm. I wouldn't even call it hot. It was just warm. It was above room temperature. Yeah, yeah, it was above probably blood temperature. You know, it was probably like it was probably like 110 or something like okay, Fahrenheit. And, and it was one of those things where it's like, normally my spindle is like icy cold, because I'm pumping water through it. So I went I went underneath my machine and popped open my water reservoir, it stuck my hand in there. I was like no, this is the exact same temperature. It's like oh, okay, so great. Basically, what I found is after about three hours of cutting my five gallons of water and antifreeze is not really doing a fantastic job of cooling anymore. You're you're trying to dissipate through a plastic bucket with no surface area well but I'm also relying on the fact that water you know, water is pretty good at carrying away heat and and so like all the all it boils down to is that I need a radiator if I'm going to do anything over a two hour cut, basically so that's sort of my next little adventure is to get a radiator and stick some fans on it and just so I don't do that I don't know how expensive because back in the day when I did a lot of watercooling for PCs Yeah, this is before like there were there was this huge aftermarket aftermarket support for like watercooling PCs where now you can just like go on Amazon you can buy a radiator that you can mount 120 millimeter fan one
Host 2 39:14
Heater cores for cars is what we used it's because you can buy a heater core for like 10 bucks yeah
Host 1 39:23
If you want me to I can go check up in my my part collection up in the attic and see if I have a heater core that I can just ship you oh yeah sure whatever I mean, I was gonna do exactly what you just said like they have tons of radiators on Amazon that you just slap 120 millimeter fan on it and I would not be surprised if a you know a single 120 millimeter fan and a radiator was enough to keep this cool. I would agree. So I mean like I don't even I'm not even like my flow rate is not even really that high. So I was just gonna slap that on there and just have that go to town so you can buy one for $15 radiator you can buy, you can buy one I'm not shipping you. Yeah, well, and the funny thing is I have I have a fan that runs on on 220 just lying around, so I'm just gonna slap that on it and be done. You know, okay, also, I'm gonna I'm gonna pull something out of the grave here. If I really wanted to up in my garage, I do have the SSPs sitting up there. And there's two giant radiators in there. That's where your graveyard set? Yeah, yeah, I could I could cannibalize one. For the sake of science, you know? So we'll just do that. That way. You don't have to just don't have to order anything. Yeah, yeah. So I'll probably I'll probably just end up doing that. But I was, I mean, there was no like downside, they CNC, you still ran fine. I wouldn't go much more than that. Because it probably would start ramping up in temperature pretty quick. But, you know, one of the things I was I was thinking about, I was like, I ran three hours, I can guesstimate what my temperature was before. And I can guesstimate what my temperature is after hell, I could even measure it if I want to do is like, maybe I should do the calculation to find out like, what the what the amount of heat energy was being dumped from the radio. And I was like, no, no, I'm not going to do. Like, I would totally what is what is my instantaneous heat dissipation due to the water flowing through my spindle and see how efficient it is? Because there's a lot there's a ton of thermal mass, like I do, like a lot of thermal mass in five gallons of DI water, you know, to heat up that much. I'm sure that okay, so I was running my my spindle at 10k rpm. I don't know what that what that what the like the duty cycle is for that actually. Wait 10k RPM would be a 50% duty cycle, or a little less than 50% duty cycle. So is it the max speed? 20k? The max speed is 24k. Okay, well, I mean, 24k is probably not 100% duty cycle. So I don't I don't know exactly what so yeah, it's probably a little less than 50% duty cycle on three phase. So I could guesstimate how much juice was being dumped into the machine. And then guesstimate how much I'm taking out of the machine? It probably depends on Load To. It does because it's a constant torque. spindle. So I don't know. I was doing one inch a second at point one inch depth of cut into MDF. That's really not that heavy of a load. Don't know that's that's, I feel like it my math hat on. I'll go on to that. So I did a lot more work on the Jeep fan evolution to a Body Control Module. Well, I say work on I bought things. Hey, that counts. That counts, right? Yeah. So Fabio and Roger on the macro app, Slack channel suggested the NRF series of microcontrollers for this. Specifically, I bought the NRF 52 Dash DK Arduino compatible development board. It's like 39 bucks. So I bought two of them. And the main thing though, is I went and found there's a configuration for Arduino IDE for it. So it should be all groovy when they finally show up. Basically, I want to do a proof hopefully by next podcast, I have a proof of concept basically have like, a unit in low power mode. And you press a button and a light lights up on the other one. If that works, and I'm like, good, I can start doing a layout. Go straight to Layout mode schematic and then lay out yeah, basically I'm like, okay, the software is good enough to make that work. Then I can do everything else. Yeah, you can be pretty confident. Yeah. So that's, that's, that's the funny thing is that's my level of prototyping at this point. Like, with like jumper wires and stuff. Like, it's really like, okay, the software proof of concept works. Let's just go get boards built. That's
Host 1 44:21
Just the evolution of Blinky. Yeah, right? No, okay. How many times have we talked about that, like, the very first thing you do is just make sure that your your processor can come up and blink an LED and you're like, Okay, I'm 90% done with this project. I think that's a better title than why I came up with this podcast is JSG the Xerox of connectors, or is Blinky 90% of your project? Yeah, that's a that's a that one's good. Cool. What about what about your 3d printer? Um, I think I did. I talked about this last week. I don't remember. Now, so we could go man. I don't remember at all 220 episodes they all run together. Yeah, they did. So I was talking about taking it apart and fixing it still waiting on the parts. The main one is the fans and so we know basically I want to get all the parts in before I take it down and do a rebuild. But I did experiment with some hops number nine which is a a gun. A gun oil slash cleaner. It's a it's a built has a cleaner. It's like two in one shampoo. Where it cleans and conditions. Time it just says it doesn't either. Really well. You shower with detergent, don't you? For tight on your head. Just don't eat the Tide Pods kids. Yeah. So hops. Number nine is like a beginner gun oil. Right? Because it comes in all the cleaner kits and stuff. But it's really it has a solvent in it. So it's really good at cleaning stuff. And it has a a protective oil leaves behind. It is really it is the tune. It's a conditioner, shampoo and conditioner. Yeah, head and drums. And so I actually tried to and my printer is running right now. It is quieter than it's ever been sees the work. Would I recommend it? No, maybe you're not sure. My sample size is one printer. Right? Because so I think we I think we talked about using hops? I don't know, two weeks ago, something like that. And so like it's still running. So let's just call this a long term test. Yeah, long term test, basically, maybe that'd be like, once a month, we'll be like, is it start running? Yes, yes. Okay. And then after, I don't know, six months, it'll be like Hops is good. The main, the way to really do that would be to use it. And then actually check the were under like a microscope. But that and to like, filter out the oil and see if there's any like, particulate in the bearings. Yeah. But both of those things were not capable of doing. No, they're not. No, actually the funny thing is I do send my engine oil to our lab. My cars do you really on on one of them? I do. My, my Jeep Wagoneer or not wagon, Euro Wrangler, because it has a new engine. And so I this is like the first time I've ever owned a vehicle that like I know the entire lifespan of something. Because I buy all my vehicles use so I'm like this engine is brand new. So I have like charts of like all it's weird numbers. A nice yeah, yeah, having fun with it. I like it because like you can see like, like, when I when I redid the oil pan on it, you can actually tell by the numbers of the oil change afterwards, because it had a slightly higher slug and calm content from the gasket. bleeding out some silicone. Yeah, into the oil. Damn. Crazy what you can find out.
Host 1 48:05
Just going a little too deep into it, man. Hey, you're having fun with it? Yeah, it's a hobby. So given that you know how much money I put into that jeep? I'm like, Okay, I
Host 1 48:16
Want to know, what I'm doing. Maintenance wise, is keeping that engine alive as long as I can. Sure. So you can tell that by the word numbers. So cool. RFO time? Yeah, let's hit the RFO. We got a really cool article this week. One that that? Parker actually shared with me earlier in the week. And I read through it today. And I was like, oh, we need to talk about this. Just it's it's kind of in 20 Enough, I only read like half of it and you've read the whole thing. And I was the one who shared it. Well, last week, you were like, Hey, we should talk about this. I was like, I know, I need to read it. You know, you sent me like a summary and everything, like a couple hours ago. And I'm like, Well, I guess I need to actually read this. So this article comes our way from signal integrity journal.com. And it's titled The myth of the three capacitor values. Which one? Okay, so what's interesting about this is, I didn't even know that that myth existed previously, maybe there maybe I'm too young of an engineer. But this is basically this article is talking about using multiple different bypass capacitors on your power pins of your components. So you know, I so let's just put it this way. I didn't know about the myth of the three capacitor values. But I knew I guess what this article is talking about the myth of using multiple capacitors on a bypass pin. I like the title because usually when it's like the myth of blah, blah, blah. It's usually like the scary little children. Yes, maybe this is supposed to scare like engineers. You know, it's actually in my opinion is sort of, you know, the capacitor boogeyman is gonna get you all three of them. So, so what What a lot of this is talking about is using. We've all probably seen this before. But you know the impedance chart of a capacitor, where at some point a capacitor stopped acting as a capacitor and starts acting as an inductor. And you can kind of basically plot that based off of its value, but also some other characteristics like its component size and things like that. So if you want to have the bathtub curve, bathtub impedance curve of low impedance from low frequency way out to, you know, the bazillion hertz, then you use multiple capacitors in parallel of lower and lower values. So like, you know, one micro farad, point one micro farad and point o one micro farad, and you kind of smear the entire impedance chart, or, you know, you get low impedance out to infinity. And what this what this article is talking about is, that's not necessarily true. And that idea of using three capacitor values is has been a rule of thumb, and I think they were mentioning things like that's been a rule of thumb for like 50 years or something like that. And engineers just do that. In fact, there's, there's some examples of modern data sheets, where they're showing like power pins on modern processors where there is suggesting us multiple different values on there. And what this article kind of breaks down or starts to talk about is mlcc capacitors nowadays. So in your standard, oh, 805 or 6030402 packages, kind of they're not shouldn't say kind of their package, the argument is that their package definition or their dimensions have much more of an impact on the impedance curve than the value itself. Therefore, these mlcc components actually do a better job per capacitance value of having low impedance out to high frequencies, then doing multiple capacitors in parallel. So multiple capacitors in parallel have some significant drawbacks, because you can actually get peaking based off of resonances in between the three capacitors. So the author of this article, is that short little trace between them? Yeah, yeah. Well, and and, you know, the whatever ESL and ESR and, and decided multiple caps. Okay, right. Right, right. Well, and the thing is the, that rule of thumb with using three capacitors, stems from old school through hole designs, where you use a big a physically large capacitor for the high value, and then the next one down is lower in value, but also lower in size. And then the one down from that is lower in value and lower in size. And so not only are you changing value, you're changing the physical dimensions of the capacitor. But that doesn't necessarily apply if you're using 30603. capacitors. Hmm, gotcha. Gotcha, gotcha. I see what you're saying. So it's interesting, the argument at the end of the article is not to use three different capacitors, but to use three capacitors of the same value.
Host 1 53:10
Interesting. I wonder how because how I design a lot of my audios, and this stems from the older idea, and this is from knowledge that's been passed on me by other designers. I've never actually tested this and it's actually a really good idea is I actually we should we should old couple weeks ago, we had a podcast an RFO about testing power supplies. Yeah. So we should probably hit up. We should probably hit up a a, one of these, like, testing manufacturers to make a little scopes and stuff so we can get like really good lab equipment, like Keatley? Yeah, like easily. And then because what I do is a lot of my elbow This is mainly for LBOs. Low Voltage drop out regulators is sort of its lean linear regulation is not switchers switchers are pretty much like I'm like, doing the textbook what the datasheet says, oh, yeah, and most of the time they tell you like just do this. Yeah, but for LBOs is a little bit different. So what I typically do is I will use a this is for like normal like five volts or 3.3 volt regulation under an amp is 1206 10 microfarad cap, ceramic cap surface mount. Oh 8051 microfarad oh six oh 3.1 microfarad. If I had the space that's my stack up so I'm changing the size and the values going down because of you. The main thing that you want to do is to make sure your crossover on basically your frequency because you know how there's a the frequency charts on their on their capacity levels were in this See becomes an L, you don't want them to be the same, right? But Trump's to be the same. So changing the values and the size, pretty much always guarantees that they're going to be not crossing. So you're going to have capacity. It's so so the thing is, and I'm not saying that this is true, just because this is the first time I've seen it, and I'm still raising an eyebrow to it. But if you read this article, this article is basically saying that's garbage. Yeah, I'd love to test it. Yeah, it would be really cool. And and, you know, the funny thing is, we were talking about this earlier in the podcast, and I said, I was gonna bring it up later, what the kind of the whole thing of this article is like, at the end of the day, you just have to test your your circuit, like, you know, like, if you do it, if you do it one way or the other, it's sort of doesn't matter. Just if your circuit works. It works, right? Yes, yes. Now, I had to go look at the article. And so it would be basic. So if I had 10 1.1. So I basically have an effective 11.1 microfarad capacitor there. And so the article is saying I should pick 11.1, micro farad, capacitor, three of the same size. Actually, here's what the article says. And this, this is something I totally don't agree with. For a lot of good reasons. But it's it was basically saying, like, if you if your component that you have available is, oh 402, pick the largest value that you can find in oh four, two, which is like 10, micro farad. And it goes put three of those in there. It basically says are expensive, it will and so like whatever the biggest you can afford, I guess, whatever. But like that's the argument that this article was saying is like, you'll get better response from picking the largest the largest value within the particular size you're going with and just using three of those than you would with doing the whole 123 separate values. Now, I was actually talking to my boss about this today, in our designs, I was like, you know, that's that's kind of interesting and kind of curious, but like, we pepper our boards with point one micro farad caps for bypassing, like if we bumped all of those two to 10 micro farad. And did three of those on every power pin.
Host 1 57:20
Do you know what the short circuit turn on current would be during rush? With just paper?
Host 1 57:26
I mean, like it wouldn't be so ridiculous. So I don't know like the once again, it boils down to like, you have to be the arbiter of your design. You can't just like make choices, like ridiculous choices with all these things. I think we haven't had a, a, a project on the podcast for a long time, mainly, since we're in two different states. Now. We are like we were social distancing before it was cool. I think this should be our first our first, like, project like team project for the podcast. We come up with a project every podcast. I know. We don't do them. No, we don't. And we usually come up with projects that are like my project or your project. Yeah, no, this is this is a good joint. This is a good one. And because this is the thing, you're really good at the simulation stuff. And I can get them built like right away. Now do the layout. Well, okay, so if we were to do this, we would have to come up with a a way that makes the most sense to actually test if it's worth if it makes a difference. So I would say we have to come up with a test what we want to test how we test that? Yeah. And then the different inputs, like basically all the different layouts we want to do, like I've gotten my opinion, we just stick with LBOs. So we pick like one lto. And then or maybe we're gonna do the same footprint, but different LBOs we can pop in there. And then different capacitor configurations. And then yeah,
Host 1 59:10
I mean, what happened, I love the idea, it would be really cool to in simulation,
Host 1 59:17
Populate non ideal components. So every capacitor has an R and L Annecy. And then, you know, actually simulate that this sounds like a really great situation where we can kind of leaned back on our Slack channel, because there's a lot of really smart people in there that probably would be able to come up with a great test plan on how we can actually say, conclusively, this is worthwhile, or this is not, you know, one one way or another. So and I think another way to go about it too, is also one of the things I want to test is price. Ooh, that's good. Because that's that's the thing about this article is it's like Yeah, put the biggest value in an oath or OTO package, because that's like, you know, the smallest package. So you have lower ESL ESR. But 10, micro farad, ofce, for twos are expensive, you know, and at the same time, what I'm what's kind of going through my mind right now is like, okay, so take whatever LDL you have, and let's say you have a pretty heavy load, let's say you're pulling 500 milliamps off of an LTO at five volts. That's pretty hefty, right? So that's one situation where you have a really high current application, but then choose a whole different application where that LTO is powering a processor running at 300 megahertz, completely different type of load, really high frequency pulses. And so like, I'm thinking, it needs to be tried in multiple applications like low voltage, low current, you know, high frequency voltage and current high current, like, you know, different different applications. And we would need, we will need a program will load that can do that, then, yeah, or just plug it into a processor, plug it into a giant resistor. So now like, I like having a because I assume, if we can get a testing, manufacturer on board, we can get programmable loaded, we can say, pull 300 milliamps at this frequency. You know, yeah, an active cycle between like, you know, 203 100 Yeah, yeah, it's kind of simulate a processor. That's just like flipping a bunch of LEDs. Actually, I bet you you could probably just do like a square wave load basically. Like you were saying, do that do the load but then see how the impedance like flattens out the curves on the on the on the current pulses. We got the lean a lot on our Slack channel, that slack channel comes up with a test that you want to test. Yeah, we'll make it. We'll make it we'll make it work. Yeah, that's cool. Let's let's all argue about on our Slack channel. Yep. Yeah, this is cool. I kind of have Sorry, I'm gonna call another person out here. I kind of have a feeling that Tom Manish is going to have some thoughts on this one. He seems like the kind of guy that will know like, this is what we're gonna test it. Yep. Yeah. Man, I was trying to end the podcast but Well, that was the macro engineering podcast. We were your host Stephen Craig and Parker Dolman. Take it easy. Later, everyone.
Transcribed by https://otter.ai