- The ole ordered a 0603 and it was metric instead resulting in an 0201 imperial chip component.
- 0.6mm x 0.3mm = 0.024” x 0.012”
- 0201 imperial chip size only size that does not match the actual size
- Imperial 0201 is 24mil x 12mil… not 20mil x 10mil like the name implies
- Imperial 0603 is 60mil x 30mil
- 0603 metric is the only standard metric size that has the same size code as an imperial one!
- Does anyone actually use the metric sizing for chip components?
- Hardware v software – Features and their impact
- needs a lot of motivation to make a feature. It is slow and deliberate. Removing features is a big deal
- features by default cannot break the device unless extreme cases are encountered
- Software is “easier” to flip the switch on features.
- Have to consider all the ways this will break things.
- Kicad – Panelization
- Moving between layers has hotkeys – yay
- Middle mouse now by default pans and wheel is zoom
- Snap to points is huge
- Creating arrays actually creates arrays
- There are pluses and minuses to this.
- Ref Des is maintained!
- Changes to one board are not copied!
- Holes are not a function…..They are a footprint and therefore need to be created.
- Most tutorials suggest that Mouse bites be made a part….
- Buying new machines
- Sometimes being an engineer is annoying but you still have to do it
- Learn from my experience
- Air Hose
- Power Cable
- If you are a machine provider also consider these things
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 2 00:11
Welcome to the Mac fab engineering podcast. We're your hosts, Parker, Dolman. And Steven Craig. This is episode 257. So I see a really juicy topic there, Steven. So I
Host 1 00:25
Spent the entire day today playing with KY CAD. Which the last time that I downloaded and spent some time with KY CAD, or is it chi CAD, I know somebody's gonna be yelling at me right now. But whatever, I'm going to call it KY CAD from here on out. If it's chi CAD to you just replace that in your mind. So, so I spent the entire day playing with ki cat. And the last time I really put some effort into KiCad was many, many years ago when I was at the fab. And I think both Parker and I downloaded it at one time to just kind of like play around with it and figure it out. Because it was like, oh, there's this new free EDA tool. Maybe it wasn't new at that time. But it was new to me at least.
Host 2 01:06
Yeah, that was around version four of KitKat. I think,
Host 1 01:10
Yeah. What's it on? Now?
Host 1 01:12
I don't revive something. Yeah, it's
Host 1 01:14
Something like that. So I actually downloaded it this morning. And the reason why is because we got a customer who sent us some boards that were KY CAD, and I wanted to prep them for manufacturing. So a lot of with some of our customers, we, if we've been working with them for a while, we will just work out like manufacturing bugs with them. But if they're a new customer, we tend to just say, Hey, if you can send us your files, then we'll do things like adjusting the board outline for panelization, or adding fiducials if necessary, or just a handful of other things that are like housekeeping jobs that need to happen, and 99% of the time don't affect the end result, but they just make it so it's possible to manufacture. Because a lot of times like a designer isn't thinking about the machines that this thing has to go through. They just think about the rectangle that their world exists within
Host 2 02:13
A my rectangles have corners
Host 1 02:15
Might have chamfers. So yeah, a lot of times, I have to massage people's boards to get them ready. And that's what I was doing. Today we had a customer, we're like, okay, cool, I've got a handful of boards. And these boards are very tall and very thin. So they don't really work very well to have tabs in between the boards in an array that work really well for having rails on the top and bottom, which is great, because that's how they go through our, our assembly. And since they're so thin, when we penalize them, we need to penalize them in like, probably a minimum of five, I would say five to 10, something like that. With these, I think I'm gonna go with five, it allows us to not put too much stock in one array. So okay, so I kind of want to do a little bit of a side note, because as I was playing around with this, I've, I've thought, Okay, I'm literally going from scratch here, I'm downloading an EDA tool. And as opposed to a lot of other situations. In this situation, I know exactly what I want on the output. So I had to make KiCad do exactly what I wanted. I knew my end goal. And I knew my starting goal. I had to figure everything out in KiCad. So I as I was playing with KiCad, I wrote down notes on things that were like, Oh, this is cool. And oh, like, Oh, why is it this way. And I thought it'd be fun to kind of talk about that today. Actually, and before I start, I wanted to give a quick side note about something that I feel is unique to the the idea between hardware versus software. And I and I apologize to anyone, I'm not trying to boil down either one of those into too simple of a concept here. But to make a change, or to add a feature to hardware takes a lot, in my opinion. What I mean by that is, if you want something in hardware, you have to be pretty damn sure that it's worthwhile. And worthwhile can mean a lot of things like worthwhile isn't like the end user, like finds value in it. But it could also mean like, Oh, I just think this is cool. And I like it, therefore it's worthwhile. But making a decision, even to something as simple as a button in hardware requires a lot. You have to spec a button, you have to find a way to get it into your processor, you have to physically add it to something that in my opinion is a little bit different than software. And I'm not trying to say that software is easier by any means. In fact, I would make the argument that quite a bit, not easier. But let's say you already have a program, and you already have like a title of buttons up at the top, then my argument is adding another button is adding another button. There's a lot of code that goes behind that. But it's not as much of like a huge issue to say like, in terms of like hardware, where it's like, if we add another button, that means more physical pins and more connections, and all these other things that have to happen behind the scenes. So I kind of feel like feature creeping in software is almost more dangerous than it is in hardware. Because it's a little bit easier to just say, like, oh, I can just do that, you know,
Host 2 05:42
I mean, look at the software beersmith. That is definitely what you're talking about is like, Oh, we're gonna add this new feature with this button. And now it has like eight zillion buttons on the homescreen? Well, because the framework for the button is there. So they're just like, oh, all it is, is just a snippet of code, right?
Host 1 06:03
So why am I even talking about all of this? Well, it kind of, it's a little bit in reference to KiCad here. So my goal today was to jump into key CAD, open up my customers files and massage them into a panelized array. And I noticed some pretty interesting things. So first of all, I don't remember this or recall this from previously. And this is this is a feature that I'm really, really sold on with dip trace is moving between layers really, really rapidly. In dip trace. It's funny, because I kind of relate this to first person shooter video games, like Parker's kind of smiling here, in first person shooter video games, how do you access your different weapons? How do you do that? For me, it
Host 2 06:56
Depends on the game. It depends on the game. But you know, some I do quick keys like 1234. And above was,
Host 1 07:03
Right. Or you could use the mouse wheel right? But But in general, the default is the the above was do this, you got your number keys. That's how dip trace works. Layer One is your top layer. Layer two is your bottom layer. If you add internal layers, they just go in sequence. So if you need to, like scan, scan between layers to see like, hey, how do I sneak a trace around in dip trace is just your hammer in three layers. And I love that because it allows me to see my board very quickly. And in a lot of other EDA tools, I've found that switching between layers is way more onerous, you have to go to a menu and you have to say I want to show all the top layers. And then I want to show all the bottom layers or whatever. Or you have you build layer sets and things like that. It's not as fast as going between layers. So it's sometimes it can be a little bit more difficult when you're trying to see like, oh, I need to take this layer this trace to this layer. And then that layer or this one is better than that. I love dip trace for that. And key cat has that which I don't remember it having that before. It's a little bit more awkward, because it uses like page up and page down which no problem there. But it's not as easy as 1234 For me, but and then it uses some of the F keys for internal layer. So you kind of have to walk around, it's still generally easy. And I really appreciate that super cool. By default now, I think we talked about this a while ago, or I Parker mentioned it a while ago, the mouse wheel is like and using air quotes here proper zoom, where like, it follows the same rules as a lot of other programs. So you don't have to rewire your mind in order to use it.
Host 2 08:51
KiCad has this is a big problem with the default way this happens on KiCad.
Host 1 08:57
Well, it is no longer the default or the awkward way is no longer default. Yeah, I
Host 2 09:02
Well, I knew they added the option where you can like check a box in the setting that makes it like every other program uses it must have realized everyone checks that box now.
Host 1 09:13
Yeah, possibly. I don't. Yeah, maybe it reported back to the mothership. And they're like, oh, yeah, 99% of people kicked that box. Yeah. So it's that way now. And middle mouse click is scroll now, which is awesome. That's so good.
Host 2 09:27
Good. That's how it should be.
Host 1 09:30
Yes, yes. That is how it should be. So one of the biggest takeaways from downloading KY CAD now is in a few hours. Like I was doing everything I needed to do. It was it's super user intuitive. Now, it feels really nice. Like everything does what you kind of think it would do. And if you're like, Hmm, I need to find this feature. It's they're typically worded fairly well. Oh, and it's just easy to navigate. Previously, I remember thinking like, Okay, this feels like a drafting program, like the way old KiCad was, it felt like, okay, there's those oddities, that once you figure them out, you're really, really fast at it. But now they kind of got rid of the oddities, and just made it feel nice, the user experiences is better. I also think that Keith had spent a lot of time with just like, making their buttons look kind of cute, you know, like, everything has like, kind of nice graphics behind it. And there's, you know, I don't care about that, necessarily, but it's, it's also like, Oh, that's cool, you know. So something that dip trace just added recently, that I think is like, I don't know why it doesn't exist by default at the beginning, but snapping the points, like KY CAD has snapping to the endpoint. So if you draw a line, and you need to draw a line from that endpoint to somewhere else, you can snap to the edge of it and just go for it. And that is so critical. I can't tell you how many times I've run into programs that don't have that and you think you've closed a rectangle or something like that. And then it's not close. In fact, Google SketchUp had a lot of problems with that back in the day, I'm sure that it's much better now. But like closing a polygon, without Snap Points is, like, maddening. So I think that's super awesome. Now, here is something that is in reference to what I was talking about earlier with hardware versus software. That, I think is kind of interesting. So in terms of like navigating around the software, key cat is excellent, in my opinion. In fact, if someone was like, hey, what program should I get? I'm going to start making PCBs, I'd be like, just go get KiCad and learn it. Because as soon as you learn that, like any other EDA tool should feel very similar. I feel like it's a great starting point. Oh, and not that like starting point as in like, you would need to go to something else later, you could start and end in KiCad. I feel there is some interesting quirks, though, because it doesn't feel full featured when it comes to manufacturing. Like you can still you can get Gerber's out, you can you can do all your plots, and things like that. But when it comes to creating arrays, there's some unique quirks where I'm like, huh, and I was doing a handful of research on Google earlier. And it seems like a lot of other people tend to agree with this, where it's like, yeah, this is maybe a future update kind of thing would be good. So in order to create an array in KY CAD, you basically select on a panel array, a panel, right, yeah. So to take your PCB, and make however many number you want in X and however many you want, you want and why and then tell it the spacing, and it just spits it all out. Right? That's, that's great. There's some unique quirks to that, that
Host 1 13:03
Are, can be both positive and negative. And, and I do not envy software, guys. Because software guys have so much power under the hood, that they have the control where it feels like you could get lost just trying to fix everyone's edge cases. I feel like in hardware, there's a lot less edge case hunting, because in many ways, it's limited. And a lot of people that are already using hardware, just know that it's limited, so they're not going to approach those edge cases. Whereas with software, it's a lot easier to ride the the fine line between like this is an edge case, who knows if they're gonna do it? And if they do it, is it gonna break everything? I'm not sure. So. So back to penalisation. Take this as an example. With KY CAD. When you go to panelized something, you select all the items that you want panelized you right click, and then you just say array, and then it gives you some prompts about how many do you want? How do you want them space, blah, blah, blah, and you press go and tada? Like they're arrayed, and that's super awesome. You still have to massage some things because it's when it's a writing them. It's literally just copying and paste, copy pasting. Yeah, and you still have to create like the PCB rails on the side and things like that. One of the cool things though, is that it when it copies and paste it copies the ref does. So if you have C four and you do it five times, you'll have five instances of capacitor C four. And that's really great. And that's required for manufacturing, you have to have that. Dip trace, on the other hand doesn't handle it that way. Dip trace fights the user in every possible way to never create duplicates. And there's really there's a lot of positives and negatives behind that. Because if you create a duplicate of something else It can get confused really quickly. And when you say which C four are you talking about? Or do you start doing things like C four underscore one to mean the first array or blah, blah, blah. The way dip trace handles penalisation is it doesn't make copies of things, you tell it how many things you want to copy, and you tell it which direction you want it to copy and the spacing and blah, blah, blah, and then it creates these white boxes that represent the copies, but doesn't actually show you the copies. Now, KiCad actually physically creates a full on copy of the other thing, which I think is is great. There is a downfall to that, however, so the rift has been maintained is great. But the changes you make to the original board don't get translated to any other board. So you have to make sure the very, very last thing you do is array, because let's say you you build a board and you array it and you spend two hours to massage all the board out ones, and then you find a problem. You have to start over like you forget to put a fiducial on Yeah, yeah, yeah. And let's say you, you decided to do a big array like a 25 by 25 or so you're not going to go in and put fiducials on each and every one. And it's like, ah, like, you're creating an array, but they're not linked in any way. And that's that's one of the benefits of doing like a linked connection like a dip trace, you can actually you can the very first thing when you make a PCB and dip traces, you can drop down your board outline and then set your, your panel, you could do that the very first thing I'm sorry, set your array. I don't I don't remember how Eagle does it, how does Eagle do penalisation
Host 2 16:47
Very similar to KiCad. Where so it will make it makes copies. But what it does is instead of making two copies of C four, it will make a new designator for it. But then it does a copies the designators to a new layer. I think it's like layer like 200 and some odd, whatever, right? It's just some random layer. And it makes that C four. So when you go to export it, it changes the silkscreen to that copied. That's it's actually a really, that way you have you have independent parts still. But when you make you silkscreen, it's correct. Is
Host 1 17:30
That is that the squash command or something like that?
Host 1 17:34
No, but the squash command? I don't think the squash command doesn't exist anymore actually
Host 1 17:38
An eagle? Oh, it doesn't. Okay, no,
Host 2 17:40
Everything's squashed. By default. There's a new command called, well, those command that was always there, but because used to be you, you could so back in Eagle, eights or prior or whatever, anyways, before I noticed that that feature went away. When you put down a part and Eagle, your designator, excuse me, excuse me, all the texts that describe the component, like a designator was connected relative to the components. So if you move the component around that doesn't fit with it, and you couldn't do anything about it. Now, you could squash the components which unlinked that and so you can move the designator around in relationship to the part so like, you could disassociate it and who's, which is really, that's like, the last thing you do right is clean up your substring. But they got rid of it and just made everything like well, the designator is relative to the component, but you can move the designator around wherever you want.
Host 1 18:46
So if you place, say, a resistor, and then you move the the designator, and then you move the resistor does it move
Host 2 18:56
It's keep that relation moves and moves doesn't enter with it. Okay, in that same relation? Yeah, same relationship. Okay. So it's basically everything is just squashed by default. And then if you mess something up, there's a function called like, restore position, and it like restores the silkscreen.
Host 1 19:13
Oh, it doesn't suck it back to the position that's in the footprint. Yeah. Oh, that's cool. I like that.
Host 2 19:19
Yeah. But yeah, it's interesting to squash basically, because like, everyone ends up squashing every single component because you have to tweak all the designator smokescreens eventually, right? And, yeah, so it was one of those. Oh, everyone does is 100% of the time, so just make them
Host 1 19:36
Make it by default squashed.
Host 1 19:37
Yeah. Yes. So no, it's not squash. It was Smash. Oh, that's
Host 1 19:43
What it is. Yeah. I mean, same, same thing. Like what does that even mean?
Host 2 19:47
Smash squash. So Eagle. Eagle is really weird and how it names things because it was, it's written by or at least originally was written by Germans, right? That's why it's based is cream? Yes. Yeah. So paste translated into German. And then translated back is cream. This is solder paste.
Host 1 20:11
Those are some things that I'm curious about because like, Why leave it that way?
Host 2 20:17
I think he's just convention. It's still called cream. Yeah, it's still called, even after Autodesk bought it.
Host 1 20:24
No, I can remember. I was looking at something else. Yeah. Oh, you know, I was looking at a stencil printer stuff. And it's Korean. And they call it a cream too. Which, okay, cool. If I translate that, that's fine. It's the first time I saw that it was like, Wait, well, I agree. So so back to keycap. So I got everything. arrayed up. I wanted to add mouse bites next, because in general, we use mouse bites, because they're easier to deal with. So we have some default methods of doing mouse bytes. So I just wanted to copy those over. Like I said, earlier, I knew exactly what I wanted out of this array, I just need to dump it in. Here's where one thing just baffles me. And I'm curious about the mindset behind this. And I'm not saying it's bad, it's just like, that's a really interesting choice to me, there is you can't just place a hole in KiCad, like, there's no function to just place a hole. And I actually spent a good chunk of time like searching around, I was like, Hey, where's my button to place a mechanical hole, and it just doesn't exist? That's because KiCad wants mechanical holes to be considered a footprint. So in order to place a hole, you have to create a part. And then you have to bring that in. And this This goes back to those those educators, things were like, okay, the software guys who were developing KiCad sat down and had to make that an active decision to not include a hole. And I'm curious why they why they did that. I mean, I guess there's like, the idea of the purity of the link between schematic and PCB, like if something goes on the PCB than it exists on the schematic doesn't matter what it is, you know, I in fact, I wouldn't even be surprised if like, you say you wanted to put a piece of artwork on the PCBs in the silkscreen, then it needs to be a component. And that, yeah, I watched a YouTube video earlier, where a guy was suggesting that perhaps you might create the schematic, but you might not be doing the layout. And you need to tell your layout guy that, hey, this board has five mounting holes. So you put five mounting holes on your schematic such that when he creates the board, they're just there, right? And that makes sense. I could I could get behind that it's just coming from a land where like, you want to hold click the whole button and put down like, it just I was like, Wait, now I have because I haven't used KiCad, I have to go figure out how to make a part. And then suck it in, well make a library, make a part, suck it in and just replace a hole. And all I wanted was a hole that's a mouse bite, like just just a little drill hit, you know, it's like so inconsequential. But I but I get it. So there's that like, one, every EDA tool has a workflow and a workflow where like, there's ways to break it. And then there's some workflows where like, I don't care at all, you have to do it my way. And it seems like they made it like that. And it kind of going back to like my, my button thing, like a whole could have been a button on on the tab. But like someone made an active decision to just be like, this is not possible, you cannot do this. And I kind of feel like maybe that's better. Maybe there's something more positive about that. Because then you have to be deliberate about every little thing you do. And there are some default libraries out there that are called like mounting hole. And what's nice about them is you don't choose based off of like, this is a point 157 inch hole. It like the way they're labeled in there's like M 2x, blah, blah, blah hole. And and so there's a little bit more trustworthy. trustworthiness behind those. It's like, Oh, I know I have this fastener. I can go look at this hole. And it's set for a clearance fit of blah, blah, blah there for like, this is a good hole for that. And I suppose it kind of prevents you from making some mistakes there. And I see some positive stuff behind that. But it's also like, at the same time in my situation, I just wanted a half.
Host 2 24:49
You could be like eagle and do it both ways. You can have your whole button and do holes in a library.
Host 1 24:56
Well yeah, and dip trace does the same thing to like if I I have some holes on my PCBs that I want to physically connect a chassis. So I make them plated through holes, but I put them on my schematic and I connect them to my chassis ground. Therefore, I know when I'm laying them out, that's my chassis ground connection. And but but I had the option, all the other holes, I can just make mounting hole and put them wherever I want. It's like in
Host 2 25:21
Autodesk fusion, where you can use the whole tool. Or you could draw circle, draw a circle in your schematic and extrude it or an extrude it, sink it down, right cut cuts. That's right. That's that. Oh, that would be the technical term. Okay, which
Host 1 25:39
One do you use more often?
Host 2 25:41
I try to use the whole command more often.
Host 1 25:44
I think. I think I've used the whole command one time,
Host 2 25:48
I tried to do the whole command more, because it's easier to go back in history and change stuff later. It feels like Yeah. And also gives you more options. Like, okay, I need to, like you can actually adjust, like clearances and stuff with the whole tool, whereas doing it with the extruded tool, you kind of have to like guess
Host 1 26:09
The whole tool in fusion is it feels more proper, it feels more like the official way to do Yeah. Well, it also
Host 2 26:17
Has like tapers, and you can do a bunch of stuff like that built into that tool.
Host 1 26:22
Yeah. Like, if you're just drawing a circle and extruding it. That's not following a real operation. That's like, but if you create a hole, it's like you're telling fusion, take it over to a drill and drill it. And I get the I get the value behind that. It's like thinking in terms of operations of how things get made. So the so one of the interesting things about key CAD, when it comes to all this penalisation stuff is I started on Google with typing in key CAD panelized boards, because I thought there was just going to be like, Oh, press these buttons and do these things. And no, I got like 18 Tutorials, where every single one was different. And like, not even like, like slightly different. Do it this way, do it that way different as in like, Oh, I've written a script that does it or this other one is like export your board and bring it into this Gerber panelizer. And then another one was like, Oh, you need to create a footprint that has all the mouse bites and all your outlines and stuff. And like everyone was like, viciously different, like a huge amount of difference. And funny enough, I ended up doing it differently than all of them. I mean, I did it manually, because I wanted to kind of learn it that way. And and that is the part of KY CAD that still feels hobbyist, where like some guys like, Oh, I've written this magic script that does all the stuff like that feels very hobbyist, as opposed to like, here's, here's the way that we've all agreed to do it, because it accomplishes the same goal.
Host 2 27:59
So instead of having like, go into key cad.org or.com, or whatever it is, and they have an official way of doing penalisation, where you can look up like, this is how the software developers intended this to work.
Host 1 28:14
There's almost like, there's not a way that they intended it to work. They just gave like, it's a free canvas and you choose how to do it. So I don't know, it's cool. The nice thing is, like I said I, it was easy to pick up and very user friendly and intuitive. Because I would have been like massively frustrated if I was having to fight the user interface, just to find out that it was difficult to do what I wanted to do. But it's actually really easy to do that. So it's, it's not, it's not so onerous that you're just like, ah, all I want to do is draw a line. Except for the whole Wii U series, I have to create a part that's a whole but whatever. I mean, that's that's not a big deal. And actually a testament to keep that as soon as I learned that I had to create a part that was a whole I went and did it and I didn't have to look up anything. I was I figured out within minutes like here's my footprint library, here's how to create a new footprint here's how to create a hole and here's how to suck it in and I didn't have to like look anything up. And I was there in minutes. I think that's a good testament to like, Okay, your user interface is intuitive. That's huge. So, key kids cool. I don't know if you haven't used it, give it a shot. I would say you know, if you're getting if you're new to the game, like pick up key get for sure.
Host 2 29:41
If you're also new to the game, listen to all 256 episodes of the backlog of the microwave engineering podcast.
Host 1 29:47
Hey, by the time this is out, there's 257
Host 1 29:50
That's true, but not with already listened to 57.
Host 1 29:54
Well listen to the rest of the myths and then go listen to all the other ones.
Host 2 29:57
Correct. Okay. So I was looking through the electronics subreddit and saw that oh man, someone had committed the mistake that all electrical engineers who do any kind of purchasing the components has done. And that is the Oh ordered and Oh 603. And it was metric instead of an resulting in a O to A one Imperial chip components. Pitfall, I think every single electrical engineer has fallen into this trap. So I actually went into this and was like, Okay, why does this happen so much? And why is it only? Oh, 603? And, oh, 201 This happens with? So basically, what I found out is the oh, 603 metric component, which is point six millimeter by point three millimeters, is the only standard metric size that is, has that shares the same code size as the Imperial one.
Host 1 31:02
The curves cross at that point? Yeah,
Host 2 31:05
They do at that one spot. So yeah, if you look at the whole chart, and oh, 60603 is the only metric one. So like, there's no like 12 Oh, 605 metric size? I mean, there probably is, but it's not a standard. So this is my question. Does anyone actually use the metric sizing for chip components out there? Slack Twitter, let me know. Because, yeah, I don't know if anyone actually does. Are those
Host 1 31:40
Are those crickets and hearing in? Yeah, I've never I've never heard anyone using that. Yeah. Now, not even data sheets that are always metric. Or they virtually always be it depends on the manufacturer. I mean, for the for like footprint diagrams and drawing.
Host 2 32:04
Okay. Yeah. But I'm also I was looking at the charts and the actual, like sizes of what they are the only Imperial chip size that doesn't match the actual size is? Oh, 201. Because the, like 0201. What that should mean is 20 mils by 10. mils. It's actually 24 mils by 12 mils. Hmm. So it's not identical, like 0603 is 60 mils by 30 mils. An Imperial chip size.
Host 1 32:41
Why Why would they do that, though?
Host 1 32:44
I don't I don't know.
Host 1 32:45
Well, I guess whatever. Probably the manufacturing of the part dictates the actual size. Yeah, power does to actually go to 20,000 by 10,000 is probably significantly more difficult. Probably. So I gotta I got an interesting thing. Quick side tangent, I was actually considering doing this over the break. This is ridiculous. But I want to ask, I want to ask you, and, and all the rest of the people on slack if your situation is the same, and I was thinking about doing statistics on this, okay. I don't do a ton of hand soldering of components. But it seems like every time without fail that I do hand soldering of SMD components, and I dump my resistors out into a little pile, like 98% of them flip upside down. Like they never, ever flip up. Right?
Host 1 33:42
The butter side of toast
Host 1 33:43
Test. Yeah, I really feel like this is a this is an amp hour, Dave Jones kind of thing. Like it feels like he would do this. And I mean that in a really positive way. Like, I remember one time he measured the resistance of like 1000 resistors. And then did like a Gaussian showed a Gaussian curve over there or whatnot. But I was thinking about getting, like 5006 or three components and like spilling them out on a table and counting how many who are upside down and see if if they kind of tend towards flipping upside down because I swear to God, every time I have to solder them, I'm flipping over most of them. And it's annoying as hell and it just, I don't know if that's your experience or anyone else's experience. But um, but in the Slack channel,
Host 2 34:26
I would say it definitely happens more often with LEDs. SMT LEDs always look upside down.
Host 1 34:34
Even even like SOT six packages, I was playing with a with a six leg package, and all of them were upside down. I had to solder like 20 of them and I swear,
Host 2 34:44
It's it's partially, they don't. They're not like dice or they're not equally, like they're not equally sized. Right. And so they don't want to roll when they hit the ground. But you're flipping them from right side up and tape out upside down. Like you're tossing them.
Host 1 35:06
Yeah, but I swear every time I get tweezers and I go and I grab them, and I flip them upside down, and then I opened my tweezers, they they just flip back to being upside down. I swear to God, like, maybe it's just me. I don't know. Maybe I'm a weirdo.
Host 2 35:21
I did find more my favorite things I ever did was I built that vacuum pin.
Host 1 35:27
Oh, with the little chip on it.
Host 1 35:29
Remember that little vacuum pin? I
Host 1 35:30
Host 2 35:32
Yeah, that was the with a foot actuated switch. Right. Best thing ever pick up
Host 1 35:38
Qf peas really easily.
Host 2 35:40
Just everything just like that was actually more of like hand building with paste, not like actual hands soldering cream? Cream.
Host 1 35:52
Yeah, I don't know if there's a if there's enough people who are like, Hey, that's my experience. Also, maybe I will actually roll the dice on on like a handful of Oh 603 components and see if I can prove that they tend to flip upside down. It would be
Host 2 36:08
Really fun to build like a machine that will try to test this.
Host 1 36:13
Doesn't micronic have a machine that like you can just steal SMD components onto it and vibrates them towards the head and it'll pick them up?
Host 2 36:20
I don't think it's my chronic. But yes, there is a company that builds a feeder for a pick and place like that. Yeah. And it's in your components don't come and tape. They come in like ketchup packages. Like dipping, like, not like the kind that's like a foil bag. But like the, like the dipping style ketchup packages where you peel the top off. And it's like a blue plastic tray. That but full of like 10k Owino. Five, like resistors actually there was I think they only come in like, oh, four, two, but yeah, and you like load those in to like the cartridge. It's kind of funky.
Host 1 36:57
I could see that. Now Mauser has a drop down where it's like real cut tape ketchup package.
Host 2 37:06
But yeah, it has a cool ultrasonic table that kind of like bounces, they just
Host 1 37:10
They just kind of like march towards the machine. And then it has a vision system where the head will like, know where the part is, and then just go and snag it.
Host 2 37:19
Yeah, and only picks up ones that are right set up. You know, okay,
Host 1 37:23
So I specifically looked this up IPC 610. A, well, or whatever rev is the most recent, they, they do say that it is acceptable to install a resistor upside down. Like that is not a requirement for class one, two or three,
Host 2 37:42
Like your electrons don't spin the other way.
Host 1 37:45
Well, first, I thought like, that's kind of interesting, because the resistive element on an SMD component is on one side of that ceramic substrate. So if you have it upside down, does its power handling change? And I would think that IPC would take that into account but they're like, a bazillion times smarter than me if they say it's okay. Is that what classes that under? Like? Like I said, it was said it's unacceptable for one, two and three. Interesting. Yeah, I could I could be 100% wrong on that. But
Host 2 38:17
I would be bored if my resistor was put on upside down, though.
Host 1 38:21
Yeah, yeah. Maybe, you know, I maybe I spoke incorrectly. Maybe that was class one and two. But I swear it was all three classes. Yeah. Yeah, it's a cosmetic thing we did we do training every once in a while at work. And we consider our stuff to be we call it class two plus. Because our stuff doesn't need to go into space. So it doesn't need to be class three. But it's, we have a higher level of texture that enters the space. Actually, whenever whenever we do QC on a board, and it comes out like flawless. We always just say send it to space. Because it because it means class three, but our class two plus that we came up with is all the class two electrical rules, but the minimum cosmetic rules of class three. So it has to look fantastic because our people actually look at our PCBs. So they have to look fantastic, but just function in the weight class to requires. That was a tangent.
Host 2 39:24
Yep. So does anyone actually use metric sizing for chip components? Let us know.
Host 1 39:30
And do your chips all flip upside down like mine do or did they just hate me?
Host 2 39:34
So just bought them next time?
Host 1 39:38
Just just say, say a prayer before and ask them very nicely not to and that's never worked before. Okay, so I've got something that is an experience of mine that I want to I want to just kind of pass on to people. And perhaps this is also been your experience, but just learn from mine. This is not necessarily a mistake, but it's just something to consider when when buying a new machine for your shop, which, you know, that's just making an assumption that you do do that. But if you ever have to do that, just be make sure that beforehand, when you're buying it, you're super explicit about everything. And that whoever you're buying it from is super explicit with you about everything. And when I say everything, I mean everything. And the I haven't experienced because we just bought a new stencil printer. And it arrived. Will it actually arrived like two weeks ago, but we did the installation yesterday. And we're doing training today. And this Gosh, what is it it's a ESC U S dash 2,000x a screen printer. So good old fashioned squeegee style by the stencil, throw it in there and it it wipes cream onto your boards. And so the thing about it is this this machine, it was manufactured in Korea, and it has some unique quirks to it. First of all, it runs on 220 volt, single phase not a problem. I've got that available. But they don't send it with a plug that is acceptable for us plugs.
Host 2 41:23
Is it just a bare end or a different plug?
Host 1 41:26
Now it's the European two prong with the ground tab on top. That's what they sent us. So I wasn't well all of us on our on our purchasing chain. We're not explicit with them about that we were explicit about like, this is the power that our factory has available. Will this work? And they're like, Yeah, sure, that'll work. Yeah, everything's great. Cool. We made the assumption that it would come free with it with a cable for that, because we were told that hey, well, it comes with a worthless power cable. Cool, great. Just go to Home Depot and pick up the right plug, though. Oh, of course. Yeah. No, that's it's not necessarily a problem. Because I was able, I actually just pulled the cable off our old Stacey Sensenbrenner and soldered it into their weird power connector that goes on the side. Not a problem. But that's something that like we had, we didn't fully know. And we needed it right away for them to be able to do the installation. So I have to go and splice cables and do all this stuff. I feel like that's something that we could have known. And we weren't fully explicit about that. And they weren't either. Another thing that showed up is this, this machine requires compressed air. It has an air hose that's metric, and it has a quick disconnect on it. But it is a Japanese cue style. Quick quick disconnect. Yeah, Parker Parker's faces like what the hell ever heard of that? Yeah, no, go look it up. So the regular quick disconnect that's you can buy at Home Depot is called an M style quick disconnect. Whereas there's this this Japanese style that's called Q style. And it's used in there's a handful of applications for it, but it's just not stature. And here's the thing. It also doesn't have standard threads on the back. So it's it's a finer thread than our NCPT. Right, so you can't just hit a tapered thread? I think so. But it's a really it's like twice as fine as
Host 2 43:27
Just keep throwing on more. More dope or more. Teflon tape man. Yeah, I mean, like, right that thing on.
Host 1 43:34
It's a simple fix, though, because like, I mean, I just had to go out and buy a BB and, and some hose clamps and stick it on, because we're not talking about super high pressure here. So not a big ordeal. That's another thing that I wish I would have known though before. First of all, it's something I wish I would have known before the machine showed up. But even more than that, I wish I would have known before the technician showed up to start the install and then start training because his time is limited. So it's like, Ah, you got to be kidding me. Okay, cool. You know, and another thing, I'll raise my hand saying, you know, I probably didn't do all my due diligence on this one. Our old stencil printer had an automatic cleaning system where it basically has, for lack of better words, it has a toilet paper roller in there, and it sprays up a solvent on your stencil, and then it wipes the stencil. Nothing particularly special there. And we knew this machine has the has a very similar system. So we've got we've got generic solvent there and the guy shows up and he's like, Oh, that stuff. We know that stuff eats the lines, and it destroys the system. He's like, Are you kidding me? So like, it needs IPA, which isn't a problem because we have we're an electronics manufacturer, we have IPA, but in order to fill the whole tank, I had to use all of our IPA in the shop. So I mean, once again, not necessarily problem because I just got some IPA on order and a guy down the street delivers it, what it's like, these are three things that I feel we could have had buttoned up beforehand, that they're all simple things. And so learning from my example here is just being an engineer. A lot of times a goal with something is to have everything buttoned down is to have everything in control. And that's fantastic. And all, it's also somewhat exhausting. Because you have to think about everything, you have to put everything with a situation like this where I just want to buy a thing. It's brand new, like it's supposed to work, we asked all the right questions, you know, we have compressed air we have you know, it has a cleaning system, it we have 240 volt, one phase, like, we asked all the right questions, but like, we didn't go deep enough into all of those questions. And I think at the same time, there's also the idea of, if you're selling a machine, it's worth also asking those, like, if you know, your machine comes with a European plug, and it's going to America, and you're sending a guy out, it might be worth an email just been like, hey, we know this machine comes with a book like this, or at least just have it in your spec sheet. Right. Right. Which Which, by the way, the solvent being IPA for the cleaning system, not in the manual anywhere? Not at all, I actually scoured the manual, it's not there. So it's like, okay, we wouldn't even known that if we knew if we hadn't known to ask, which we didn't know to ask. And, and it's funny, too, because like, the technician shows up, and he holds up the power cable and is like, oh, yeah, most people just snip the end off. And I was like, Cool. Like, I can do that. But like, I gotta go find the cable now.
Host 2 46:48
Well, the thing is also, it's like that plug just got shipped halfway around the world as well. Yeah. Why didn't you just leave it off in Korea?
Host 1 46:57
It was a molded plug. I so
Host 2 47:00
Like it had to be cut. Still, they could have cut there and shave saved $1. And in carbon credits shipping across the planet.
Host 1 47:09
And it was funny too, because after he showed us the power cable, and he picks up the air cable, and he literally goes, Oh, yeah, no one ever uses this connector. And I was like, Why do you supply it with that then? Like you literally hold it up and be like, well, this is your problem now.
Host 2 47:25
Yeah, exactly. Why didn't the installer have the he knows all these things he should have. He should have came with the right stuff.
Host 1 47:32
And we've been talking to these guys for like a month. So it's not like this is like a brand new thing that we just did yesterday, or anything. So it's like, and that's
Host 2 47:42
You know, you don't want it is this little snide, but what is he doesn't tell them on purpose. So he gets a waste a couple hours a day? Oh,
Host 1 47:53
No, that's not it. I don't think there's any conspiracy by this or anything like that. But maybe the A Yeah. So just like, if you know, something, don't assume that other people also know that. Like, if you know that these cables, you're going to have to replace it. Reach out and be like, Hey, guys. Yeah, we're going to have to replace the end of this connector. Maybe it'd be a good idea for you to have that when you show up. You're not because I don't have my truck doesn't have it? Well, I mean, this guy flew in from California, he's not flying with it. And Okay, another thing. My, we bought a CNC two years ago, something like that a little bit under two years ago. We, we went through the entire checklist on like a video call of like, here's all the things we're going to need when the technician shows up. And it was funny, because like, we had everything buttoned down, I had the whole area cleaned. I had tables out for everything. And the guy shows up. And he's like, oh, yeah, you're supposed to have a block of aluminum that's this by this by this. And I was like, Was somebody going to tell us this? They just assumed that we had a random block of aluminum lying around that just like oh, it just would manifest itself because we bought a CNC therefore we just have random aluminum. We just have stock of aluminum. Yeah. And in order to do the training, we had to have this block of aluminum. So I had to overnight a block of aluminum to our place. And it's not a huge deal because it wasn't like massively expensive or difficult or anything like that. But it sounds like you knew that we needed that and you didn't tell us that we needed that. You just assumed that we had it like really? I don't know. Like I probably make those mistakes with customers all the time where it's like, come on, you should have known this. Whatever we all have. Thanks for letting me vent Parker.
Host 2 49:47
I think you need to thank all our listeners for letting you allow them or them allowing you to vent to them. Yeah. We'll see how many episodes how many people don't download next week's episode.
Host 1 50:00
This is not normal. Or his
Host 2 50:04
Thinking is pretty normal. So that was the Mac five engineering podcast. We're your host Mark Dolman and
Host 1 50:12
Steven Gregg. Later everyone take it easy
Host 2 50:23
Thank you. Yes you our listener for downloading our podcasts and listening to Steven rant. If you have a cool idea, project or topic, let Stephen and I know Tweet us at Mac fab at Longhorn engineer or at analog EMG or emails at email@example.com. Also check out our Slack channel. You can find it at macro fab.com/slack
Transcribed by https://otter.ai