Stephen is on the hunt for the next step in his electrical engineering career and shares the shifts in the industry and what employers are looking for.
Why is there such a disconnect between component datasheet drawings and EDA footprint layout tools? Stephen and Parker dive into this on this podcast.
What are the pros and cons of becoming an electrical engineer? With great power comes great responsibility or just the lack of sleep!
Figure 1: PCB Surface Finishes. Left to Right: HASL, OSP, ENIG.
Figure 2: Bench Design Stephen is working on.
Parker is an Electrical Engineer with backgrounds in Embedded System Design and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. The following year he designed and produced an Atari 2600 video mod to allow the Atari to display a crisp, RF fuzz free picture on newer TVs. Over a thousand Atari video mods where produced by Parker from 2006 to 2011 and the mod is still made by other enthusiasts in the Atari community.
In 2006, Parker enrolled at The University of Texas at Austin as a Petroleum Engineer. After realizing electronics was his passion he switched majors in 2007 to Electrical and Computer Engineering. Following his previous background in making the Atari 2600 video mod, Parker decided to take more board layout classes and circuit design classes. Other areas of study include robotics, microcontroller theory and design, FPGA development with VHDL and Verilog, and image and signal processing with DSPs. In 2010, Parker won a Ti sponsored Launchpad programming and design contest that was held by the IEEE CS chapter at the University. Parker graduated with a BS in Electrical and Computer Engineering in the Spring of 2012.
In the Summer of 2012, Parker was hired on as an Electrical Engineer at Dynamic Perception to design and prototype new electronic products. Here, Parker learned about full product development cycles and honed his board layout skills. Seeing the difficulties in managing operations and FCC/CE compliance testing, Parker thought there had to be a better way for small electronic companies to get their product out in customer's hands.
Parker also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components.
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.
Special thanks to whixr over at Tymkrs for the intro and outro!
Hello, and welcome to the macro fab engineering podcast. We're your hosts, Parker, Dawn and Steven
Greg. And this is episode number 5959.
And we have an announcement, right? Yeah.
So macro fab and Mouser electronics have teamed up to create a free monthly meetup in Houston for anyone involved with hardware and electronics engineering or manufacturing. Yeah, we're
going to be doing this the last Wednesday of every month. Yep. So starting in April, right.
Yeah, yeah. Yeah, that's that's the start. Starting month.
Yeah. So the link to it will be in the show notes. We'll probably tweet about it. I think it's on an event bright. Yeah, yeah. So you're over there and just probably search either Mouser or macro fab and bam, pop up.
Right. And it's free to attend. Yep,
we'll have food, drinks. So be good time. Other things this is like the list of things that our marketing people want us to say. What to Expect networking, fireside chats and q&a. Individual project sharing and discussion. door prizes. Hmm. Unfortunately, I don't think we can win one of these door prizes. But I wish we could. That'd be awesome refreshments and free parking, which I looked at. And it was like, really nice. Like, it's
bad. I like I'm free parking is like a perk. Yeah. So So yeah, no, it's just gonna be a meetup where we can kind of hang out and chat and talk shop and, and have fun at the fab.
Yeah. And the like, I guess the people that were hoping that show up or basically small to medium sized businesses, like the engineers that work there, and was it electrical engineers and that kind of stuff? Yeah, product designers.
Yeah. If electronics is your thing, if you're an engineer, please swing on by and hang out. Parker and I will be there. We can have some food and some drink. Yep. So it'll be it'll be a good time. And I'm sure we'll continue talking about this. In future podcasts, especially as we get closer to when the event
happened. So what is the last Wednesday of April?
Give us your computer phone part. I got my computer phone out already. Oh, this is a this is a an 18 plus age event. So Correct.
That will be the 26th of April.
Great. Well, we hope to see. We hope to see every listener out there. Yeah.
I think would fill up the building.
Yeah, it would probably be passed by her code.
Yeah. So with that out of the way, I guess, the required the required part of this podcast, we're required to say, right, right.
Marketing walks in engineering, just before the podcast. Here's a piece of paper, say this stuff on? No, no, we're actually really excited about this. It's gonna be a ton of fun.
We've been working on it for like the last two months, almost. Yeah, yeah, setting this stuff up. And it's gonna be really cool. Especially since you know, Mausers kind of sponsoring it. Yeah. I'm hoping it's gonna we just get a lot of people that just show up and just want to talk about what they do. And
well, and we've generated a ton of really cool ideas for the future. So this is going to evolve as we go along. Yeah. And hopefully, it'll have a bunch of good information for people. It'd be a good way to meet, you know, other local engineers. And just a good way to hang out with the with guys in the area.
Yeah. Then. So the past two episodes, we've had guests, so we haven't really been able to address listener questions. Yeah. So Christian writes, could you talk about differences in surface finishes or on PCBs? I'm especially wondering what difference between Enoch and other surfaces
right other than High Castle or
right bare copper etc, etc.
Right so um, yes, we can talk about that. I guess we will Yeah, we do have some information on that Christian so actually, so I have I have a little cheat sheet here for Parker and I that has four different finishes that are common finishes right right yet for for common finishes that pretty much every board manufacturer is going to
offer. I don't have hassle on here, though. Will
because in his question, he said everything else then. Okay, passable. And, I mean, let's talk about hassle real quick. It's Like the cheapest Yep. And basically the second cheapest Well, I guess the cheapest is just bare copper but bare copper
but everything corroded the by the time you get it,
yeah, I mean unless they like cleaned it and vacuum sealed it right away. Now yeah no, nobody really does bare copper. So hassle is the cheap one. Basically they just melt lead solder onto the board and then they take what's called an air knife and they run that across the board and blow off as much excess solder as possible. And it leaves it leaves lumpy solder all over your pads. So the finish is really crappy it's not very level and it's leaded solder. So I mean there's a ton of disadvantage they have lead free passes as well. Well okay you can perform the same thing but lead free starts to get into the same cost as these other flat surface one so it just it sort of doesn't make sense correct? So hassle has its place if you're doing ultimate cheap stuff yeah through hole stuff that that it doesn't matter if it's leaded great you can you can get away with or hassle
you're basically the only thing you're looking at is if you if it doesn't matter that your board is flat and we'll go into that
yeah
what why you want your board flat. Then hassles fine lead free hassles fine. Basically, through whole stuff, I would say hassle is completely acceptable.
It is kind of nice to solder.
Well, it's because it's already got solder Earth, the solder already on the board pretty much
well, and a lot of and a lot of ways. Guys who get other finishes will pretend pads, you're effectively doing hassle in that in that way. So okay, so let's take a look at some of the other ones. In fact, I have I have four down here. And I'll just name off the four and then we'll kind of go through what what we got on here. So there's immersion tin, there's one called OSP, which is organic solderability preservative.
That sounds like it's organic. Can you imagine walking to the grocery store? It's like oh yes, we have organic solderability preservatives in our jam
over at whole PCB. Yeah, whole PCB. So then we also have E NAIG, which is electroless, nickel, immersion gold. And then another version of gold card called
hard goal. And there is also another version of immersion, immersion silver, right, which is similar to Enoch and an immersion 10.
So really, actually, what you got with that is you have the material, you have tin, silver or gold. And then the type of finish or how it's applied is immersion, or plating or things like that. So so let's let's, actually, before we get into kind of like pros and cons of these, which ones have you used in the past?
I've used immersion, silver and gold, hassle, passive lead free. And apparently I have used OSP as well. Yeah, didn't didn't realize I didn't realize it. Yeah, at the time, but I have used it.
So I yeah, I've used a good bit of hassle. I've used OSP. And I've used EEG and probably at some point in time, worked with hard gold. So there's there's reasons for each one, that but I guess let's let's go into that. So immersion 10. What about immersion? 10 is special?
Do you know asking me I'd never used it before.
So Okay. On our cheat sheet, we kind of have some advantages and disadvantage of these. So the advantages of immersion tin, the fact that it comes out flat
it, I think you're pretty much any immersion technique is going to do that. Sure. Sure.
So So actually, all of these techniques that we just explained, come out, generally flat, yes. Well, they come out flat, it depending on the process, you can get flat and controlled thickness. So some of them actually don't aren't super great at controlling the thickness. Gotcha. In most cases. It doesn't matter though. So yeah, so in fact, if there's if we have advantaged flat surface on here for just assume that's on all of them. So immersion tin has no lead. So it's just a tin on top of copper. It's easily workable. And it works really well for pin insertion for press fit connectors. So that actually just it's a little bit more durable. So if you've ever seen there's there's some high pin count, press fit connectors that you actually press in with an arbor press. Immersion tin actually worked well for that.
Probably also that it's 10 is like soft enough or something
where the pins can dig into it again. Yeah. Also, it also, it helps if you're doing something like that to have combat compatible metals. Yep. So if you're pushing something in, you want to sort of be the same as well that in
you don't have dissimilar metals creating a charge, like on a, I guess that'd be good for if you're doing like thermocouples. Mm hmm. So that way, you keep similar metals in the chain out to your thermocouple. Because thermocouples work by two dissimilar metals soldered together. And that creates a a, a voltage differential that you measure, right. And so if you add more dissimilar metals in the chain, well, you just creating more thermal junctions coming back to your amplifier,
right. So you either have to cancel those out, either through signal conditioning on analog side, or on the digital side, you have to physically remove offsets. And that just gets dirty.
Because like, if you actually get the connectors for thermocouples, that use the same material that's in the wire, right for the contacts and then that you know, connector goes onto your board, and now you're going from whatever metal that is to, you know, sack three or five or or sn 99, you're 10 100% or whatever, okay, to whatever finish you using, then to copper, then always, you know, copper trace a little cross, then back through your finish into the paste or solder right then into the leg of the chip. And then on to the gold wire lead
on all of these have
offset its materials all the way down. Yeah, right, right.
So, you know, with with tin, there's a couple things, I won't mention all of the disadvantages, because there's a handful, but if you can damage it fairly easily. It's susceptible to tin whiskers, which that sucks in terms of shorting. On find pin pitch. So immersion tin, to be honest, I haven't seen it very much other than specific applications like connectors, and things like that. So honestly, the big one out of this is the neg Yeah, that's that's what you're gonna see the most. So the next one in the list is the organic solderability. preservative. Yep. Now this one, this one's interesting, because both Parker and I, we had, we've dealt with this in the past, we've seen this, but when we were doing research for the podcast to put together this list, I you know, I saw this as an option. I'm like, What the hell is that? And then looking at pictures, it's like, oh, my gosh, I've seen that like 100 times. Yeah, you've seen before. Yeah. And work with it. It's basically
the stuff they put on. It's like this very specific. The really cheap perfboard. So you got Radio Shack?
Yes. That's what this goes on. Yes. It's kind of got like a, like
a flat copper finish.
It's a flat copper, but it's almost like a pink fleshy color.
Yeah. It's a good way to put it, but it's really flat. Yes,
it's yeah, it's really, really flat. Once again, it's an advantages. There's no lead in it. It's, it's generally cheap. I don't I don't know if the organic aspect makes any difference to anyone but free range?
Sure. I mean, it is organic in terms of the compound has carbon on it? Sure. Yeah, of course. Not. It was, you know, artists and created in a field in the middle of Wyoming,
grass, grass fed PCB, grass fed PCBs. That might be the title of there it is done. Episode 59 grass fed PCB or
organic grass fed PCBs.
Yeah. So I mean, to be honest, this one broken down all the way to its its bottom line. Here's this cheap, it's cheap. If Radio Shack has it on their perf boards, it is the cheapest option. Yeah, where was that? Whenever they manufactured however many billions they made. And how's
it apply? Cuz it just spray it on?
It's organic. Man. It's grown. This is cool. I actually, I don't know I the thing is, I've never actually made a board with this. So I don't know the process behind it. But you know, we should actually post a picture of it. Because I guarantee you a lot of our listeners will look at a picture and just like oh, yeah, I've seen that a bazillion times. No, I've actually I read somewhere that it's, it's also easy to rework, and it's good for the rule.
You know this that's funny because you said the same thing about me. version 10 It says remarkable, but then this advantage is you wrote not good for multiple reflow assembly processes. So you can't would I guess for immersion 10 would be okay for like reworking through hole that's Yeah, but then working SMT That's right. Gotcha. That's
right. Cuz I think immersion 10 When it comes to its adhesion, down to the board, on like, surface mount pads is not fantastic. Gotcha. But in terms of through hole crap, it's awesome.
Then OSP is not very good for Thirroul. But you see it on perfboard.
Well, that comes down to GB, GB GB. Yeah, yeah. So, so that's a OSP. Honestly, that's more of an exotic one. You really won't run into that much. You just get that in a spray can. Oh, SP can always pee on aerosol, not aerosol.
Aerosol, American, propane, propane and propane. Gotcha. Yeah. Okay. I see where you're going with that one. Yeah.
Okay. So here's the big one. EEG. EEG is actually the most common finish that is out there. Now.
It's what we have we only do in egg boards, right?
I you know, yeah, no, we do 100%, you know, on everything. And, and actually, it used to be one of the most expensive, but But now, just because it's so prolific and out there everywhere. It's actually incredibly cheap.
I think it's just the basically, the fact that SMT is the way of life now. Yeah. You have to go with a flat surface, a flat finish. Yes. And one of the best ones is a NIC. And so just the nature of it being one of the best ones. It just drives the cost down?
Well, and okay, so gold, you have fantastic corrosion resistance. You have, it's very easy to apply using a What does it electroplating? Well,
they don't electroplate it? Well,
I'm sorry. You're right, you're right. Not in this version.
I mean, he means electrically, you're right.
My bed, it produces a very flat surface. It's easy to work on. And it's highly conductive. Yep. So it kind of just meets all the requirements there. And in general, it's fairly good at controlling how thick you you want to apply it. It's not fantastic at that. I was actually reading some articles on that. The control process has a lot to do with temperature and time and, and and how the border actually dipped into the the solutions. Regardless, with with enough processes, and it's been around for long enough, you can control the thickness fairly well. Yeah, most people don't care though. Now. So
in a similar vein, there's electroless immersion. Silver, right, which is the only difference is that you silver instead of gold, so it's slightly cheaper. You see that on some kind of like, you see that on, like, fast turnaround boards? Yeah, they'll use silver immersion. I don't know what the actual differences a silver is not as good. Rejecting corrosion as gold. But I don't know if there's any other difference between besides that.
Yeah. Well, and what's like, is
gold better for keeping that level? For silver? I don't know.
Oh, I don't know. I mean, technically, silver is a slightly better conductor than gold. But they're, but when if you look at like the, the list of good conductors, it's like silver, gold, copper, and there's a couple other up in there. And there's also close to each other and how good they are that copper does almost as good as silver and gold. So you're really not getting a lot by putting the other guys I honestly, I'm not entirely sure what the massive benefits of silver are. Yeah, it might be other than cost.
Yeah. It might be just cost and then it probably goes down to corrosion where gold has a way longer shelf life than than the sulfur does. Right.
And EEG is getting so cheap nowadays. Yep, that just works out really well. And really the the amount of gold that's on the board. I mean, it's it's in the couple microns thick. A couple items that Yeah, it's really, it's not that thick. You're not really getting a lot of gold on there. But it just creates a protective layer across the top that's incredibly flat. So there you go. And then the last one we have on this list is hard gold, which is effectively You take enough EEG and make it thick. Yeah. And you have a hard goal. And the reason why that's even an option is because he NYG even though it's fantastic, it still can succumb to physical dangers. Yeah, physical damage. So hard gold is basically a thick version of EEG. And it works really well for card edge connection, like PCI connections in your computer. That would be hard gold. Yep. So if you have a part of the PCB that's going to receive abuse on a regular basis, you would go with hard
gold. Yeah, basically edge connectors for cartridges. So like NAS cartridges, that kind of stuff. They use hard. They use hard gold on that edge. And usually it's applied in a certain area. It's not over your entire board. That's right. So usually you have inagh, almost, you're bored, and you say, this area needs to be hard gold, right?
So they mask everything off and do a slightly different process in that area.
Or just leave it in the tank longer, right. But why would you want a flat surface? Why isn't hassle? The best thing in the world?
Ooh, okay. Yeah. Now you're getting into some fun stuff. Yep. So why not park it's
all about clip being co planar. That's right. for surface mount parts. It's even more so for like BJs and stuff. But even for like SOSC packages and stuff, when you put paste down and put a part down. A lot of times for hassled you'll have a couple pads that are really thin or light on on tanning. And if it reflows, there might not be contact with you or your your, your toes of your parts might be floating above the pad. That's right. And you won't get really good reflow it's you also get a lot of tombstones of chip packages, like Oh, four twos, oh, six threes, etc, etc. They will actually tombstone a lot faster. Yeah, or not faster, but you'll have more of them?
Well, and it kind of depends on who does your hassle. I've seen some hassle where it comes out fairly flat. And I've seen some where you run your hand across it. And it's like Braille. Yeah, it's wavy. And so surface tension, when it comes to solder is absolutely our friend in the contract manufacturing, yes. Because when when apart, goes through a the reflow oven, it's usually pulled into the most optimal position. Yep, for all the pads. In fact, when you know, in some cases, we hand placed components on our PCBs, and a human hand is never going to be as accurate as our pick and place machine. And so you know, they might be slightly cockeyed or a little bit angled off, but surface tension of the solder is going to pull it in with with hassle. Maybe not, yeah,
the neck goes down too, because we can control how much paste goes down. That's right with our my 500 machine or with a stencil, you know, pretty accurately how much paste is going down on those pads. If you have like ENIAC finished, which is flat, you know that that's the only material on the board, that's going to create surface tension. And so both sides of the part will have equal surface tension, and they pull the part straight. That's right, stoning, and no tombstone. And if you have a hassle finish, you will have a Moses time you will have an unequal amount of tin on each pad, you add your equal amounts of paste, we still have an unequal amount of basically wedding ability, right of that of that area. And so it will pull one way harder than the other way.
So you know, if you're, if you're building boards by hand, or you're doing all through whole stuff, hassle might work well for you. Yep. But one one situation where it's absolute garbage is if you have like a Power Pad or a thermal pad that's underneath the chip. And if you can't control the thickness of that pad. Oh, yeah, you're almost guaranteed to have floating pins. Yeah,
that pad is though that partial is going to be floating on that. That's underpad right.
So so once it reaches 200 And what 17 Whatever, wherever
solder melts, depends on the
depends on the edge, somewhere above 200. See, in that range. Oh, even even if you know the solder actually fully melts, the parts still might rock and roll on whatever. So it's just there's just no guarantee with hassle.
And I went on way longer for just one question. It's a good question, though. And if anyone else has any more questions, send them in podcast at macro voices.com Or at macro fab on Twitter. That's right. Alright, so projects that we've been working on. Yep. I've been working on the side project with spooky, Pinball, trying to get our new system working. And since we're kind of going towards, like HD video for the the dot matrix display, so getting rid of the dot matrix, we put in an LCD with the previous version. And we're driving that with a FPGA to simulate a dot matrix. Okay, basically, because it's cheaper than just drawing a dot Yeah, was drawn and dot matrix on LCD screen share cuts, it's less expensive than a actual dot matrix display.
That's actually really funny.
Yeah. It's actually the main reason that it's still dot matrix on that is there's a memory. The parallax pillar, which is what we're using for the audio video, part of the pinball machines, doesn't have enough memory to do higher resolution graphics. And so we're basically it was like, a crutch to hobble along on. So we're fixing that by basically accessing the parallax propeller and going with a Raspberry Pi three to do the audio video. And that's great. And outputs, HDMI, all that good stuff. Problem is the LCD screens we use, we had to get a converter board to go from HDMI to the LVDS signal the panel once Yeah. And those are like about 25 bucks. So not too expensive, but still 25 bucks, right? So we're trying to cost that out. Okay. And you can make your own know, the Raspberry Pi three through its GPIO supports a parallel display interface. Oh, really. And so it's basically a ginormous wide parallel, parallel bus that's got all your red, green and blue signals. Yeah. In parallel. So we're the panel we use the LCD panel is actually an RGB 666, which is six bits of that, excuse me, six bits of red, six bits of green, six bits of blue, and then it spits out and then also needs like, horizontal seeing vertical sync and a pixel clock. Yeah. Well, the Raspberry Pi happens to supports RGB 666 natively. So all we had to do is go from that parallel display interface to an LVDS. Chip. And bam, we're done. Right? Okay, supposedly,
that seems really easy. Suppose deceptively easy to easily.
So I made some boards, to basically it's like it plugs. What are they called? Raspberry Pi add ons,
hats, hats.
It's lame. Why can they come plates?
Why are they called shields on an Arduino?
Well, because of the pie plates.
No, continue on. Continue. That was That was terrible Parker.
So the chip I'm using is the Ti d s 90 c 365. A, which is a basically an RGB 666. Parallel display interface to LVD. Yes, converter chip. Oh,
so you just get in power and just dump all the signals into it into it. And Bob's your uncle in like, Flynn. Oh, electronics are just becoming Lego blocks, man. Just find whichever color and size you need and drop it in,
drop it in. So we'll see if this works. I'm hoping it does. I got the board's I'm still waiting on the parts should be good.
How much is the how much is that? Chip?
$1.15 quantity? That's not bad. Verse $25. So when are you making
a little plate? for it? For the for the pie? Yes. Okay. Make it a plate for the PI. Great, that's cool. Yeah, cuz
all you got to do is like, tweak the config. txt file that's in the Raspberry Pi and reboot it and boom, it'll start doing it. So it
all the GPIO is already defined for it.
Yeah, it's actually these are alternate functions for the GPIO ports.
Oh, so all you have to do is route them to the chip?
No, you basically just tell the Linux kernel on boot right to basically use those assets and display interface
for you just you just grab whatever GPIO
know the ones that that are have the right. You have to route them the right way. Yeah, yeah,
of course. But I mean, that's, that's stupid. So
basically, it goes straight from the Raspberry Pi enter into this chip, and then out of this chip into our LVTs connector. That's it.
Oh, wow. And what's the what's the power supply range on that? Oh, watch one on the
the piazza chip 3.3 volt. Nice.
Yeah. Can you power it right from the
No, because the LCD screen needs 12 volts, right? So 12 volts come to the board, and it's got two regulators. One is five volts that spits it out to the PI. And then it has its own on board 3.3 volts to power that ship and the screen logic. Okay, so I kind of separated a bit to hopefully kind of keep current paths. Nice and clean.
Yeah, I would think that the 12 volt you'd want separate from the PI. Yep. As much as you can. Yeah,
I did pretty good on isolation. It's a four layer board by put all the signals on one layer. Yeah. So it's got signals on top. Actually, probably should have sandwich them. Anyways, I put signals on top, then it's ground. 3.3 volt ground. Okay, so I should have put it inside. Yeah. Posted on ground signal ground. 3.3 volts. Yeah, then, I mean, you need to be pleat shielding. Yeah. Because and the thing about it is we're only running like a 16 megahertz pixel clock. So it's not crazy fast that we need, like, all this quietness of the board. But I'm like, it's only like an extra 20 $30 to get rush for layer verse rush to layer boards done.
Right. Yeah. Well, cool. Yeah. So this week? Well, actually for for past couple weeks. I've been working on I think we've talked about it before, I've been working on kind of updating our manufacturing line. And one of one of the aspects of that is getting new benches out on the floor for us benches. No, yeah, no more use benches. We what's what's kind of nice is we've we've got a little bit of a new budget for our operations team. And part of that budget is going into updating our benches and getting automated and having some conveyors for things
and more conveyors than we currently have. Right. Like there be converse, basically for the benches.
Well, conveyors for the entire PCB, a line from start to finish, we've always had conveyors up to the reflow oven. So the pace yet the printer, I shouldn't say always. But for a while we've had, we've had conversation. Now we're having an app a full, proper conveyorized line. So but along with that comes benches, and one of one of the new kind of directives is to make things look nice, effectively. So aesthetics come into play with this. So I designed some some benches in Google SketchUp. Because Google SketchUp is awesome. And the the material of choice here is 8020. So 8020 is not particularly cheap, but it is absolutely awesome. To work with.
Yeah, I was about to say it's like what was the difference between buying an off shelf nice bench versus designing your own? Why? Why did you choose that path? Okay, so it can't just because it looks nice? No, no, actually,
it it wasn't. And you know, it's funny, because when if you look at a bench, it's like, come on, it's a bench, you know, four legs and a top right? What do you need? So I actually, I actually went out to our manufacturing floor, and I pulled all of our operators together, because, well, I had I had a chat with with everyone, basically, to figure out what they need to get their job done every day. What are just basic functions that makes an operator, regardless of your task, what are your basic functions to make your life easier, and most of our operators, and in fact, all of them probably have worked at other contract manufacturers. Correct. And so they've had experience with things like various types of benches, and I got a cross section from everyone of, hey, this feature this little thing, this, you know, this or that makes the difference. And so basically, I grabbed everyone's ideas, kind of compiled them all together, and found that the best way to meet all of their requirements, plus make it look nice, was to build them from scratch. So also, Chris, our CEO, kind of really likes 8020 Yeah, he does. And so I and I do too, I think I think it looks really cool. And
me a welder and a square box to me and I'll make you the world. Yeah,
right. But 8020 will look nice and clean and shots true.
And my square boxing will look nice if you paint it.
Well, yeah, yeah. I mean, you got to put a bunch of but put a bunch of labor time into that grinder. Right. grinder and paint makes me the welder right. Yeah, that's right. So so yeah, so I've been building building benches, which is a bunch of fun. So So I've have some stats. Yeah, you're
taking over our entire engineering area,
and that will change probably tomorrow. So all 8020 Connect construction Every bench has a one and five eighths inch plastic laminate, tabletop, okay, which is fairly beefy. And can, the goal is for these benches to last a long time, especially because they're not particularly cheap. So they're 60 inches wide by 30 inches deep. And they have an adjustable height from 30, and five eighths inches to 32 and five eighths inches. So your two inch height adjust actually makes a big difference. So you know, a difference in how your chair can work with it. They also all have a bar across the middle for foot rest, which is, you know, it seems dead simple, but it's actually a big thing for people out on the floor, you're working eight hours a day on a little rolly chair, having one bar across to put your feet on makes a big difference. So I ordered all of the 8020 and got it in and we got a ton of boxes of 8020.
Yeah, so the thing is, you filled up one room with crates and pallets of stuff, and then filled up with engineering with basically ginormous structures are at 20.
That's right. So the total length of 8020 that goes into one of the benches. Now this is just the bench part. This is not what sits on top of the bench, there's more 8020 that goes on top that basically forms a like a shelf for for lights and fixturing and fume extraction and things like that. But just the the bench part is 40.33 feet of aluminum extrusion per every bench. So these things are pending, of course, right? Of course. Leroy Jenkins approved benches.
Well, it's hoping that we they would our listeners would maybe know that reference, but
I'm pretty sure yeah, you just I'm pretty sure a large amount of our listeners know that reference. So, so each each bench has 42 angle brackets, and 84 nuts and bolts. So I, I did the tracking number on every package that came in and tallied up all the weight of all of the boxes that came in based off of what UBS gave us. And then I divided that by the number of benches. So each bench without the top just purely in aluminum extrusion nuts, bolts, and angle brackets is 68 and a half pounds.
Yeah. And they feel around that way when you pick them up. Yeah. So it's a pretty heavy. So yeah, we'll post some pictures. Of just I guess we have like you built like a model bench.
I yeah, I built I built one bench to completion and I put the ESD mat and everything. And I set that out in the main area so people can see kind of what's coming down the pipe. Right.
So this is what Steven has been working on for the past month.
That and a lot of other things but as a joke, yeah.
Okay, cool. So the RFO Yeah, we have one or for this week. This one's I thought was really interesting. I saw the article and I'm like, This is gonna be either really click Beatty or not. Okay. Particles from outer space are wrecking havoc on low grade personal electronics.
Oh, geez. i This sounds clickbait. Yeah. So,
but there's actual science. While not really science, I couldn't find the paper that they were talking about in this because there's no, there's no, I couldn't find a source. It was on phys.org. So at least has some maybe some credibility. I don't know. I don't technically anyone can register a.org domain though.
So no credibility. Yeah. So so what what is the article said?
Basically, it went into the fact that cosmic radiation and particles from out of space, can with enough energy, hit your device that's not shielded? Flip a bit and cause your OSD crash or your device to crash.
Well, okay. So you may, but But absolutely, that could have Yeah,
and actually that some of these signals they what they talked about, originally outside of our solar system to Oh, absolutely. Yeah. And they call a single event upset. Really? Yeah. And
you Oh, my gosh, they have an acronym for se you. Oh, but they called
it an upset.
But how can you detect that? So
apparently they did. There's some testing. They did. The I cannot pronounce About basically Cypress Semiconductor. I can't pronounce the people who did it. But basically, they did some testing to see how prevalent this would be. And the results were published in the 2004. Electronic Design News. I haven't found that article yet either. But that's like pre YouTube. So it's really hard to find stuff on the internet. That's pre YouTube. Okay. A simple cell phone. This is a simple cell phone 2004. So it's before the iPhone. Oh, yeah. It was the Nokia brick. Yeah. Nokia bricks. That was the name of the game. Back then. With this is funny. 500 kilobytes of memory, which is like nothing we have now in our pockets, right. Would have one potential area error every 28 years.
Toy under? What is that? Meaning that there's enough cosmic radiation such that you would hit one bit, or one would have some
errors? Error? Regarding the A S EU? Yeah. A single event upset caused by outside influences? Uh, huh. Once every 20 years. Okay, okay. So and that was the first one. A router farm. This is actually really funny router farm because it all the way around. Hang
on. I gotta say something real quick. i i Last month, I turned 30. I have not had a piece of electronics for 28 years of any sort. Exactly.
Well, that's just saying that the given the odds, it would take 28 years to get a events on average. Well, yeah. So you could happen the first time you turn on your phone?
Well, within 28 years, the clock is ticking effectively.
No, no, it could happen. Right? When you turn your phone, right. Just the odds are it will happen within 20 years.
Right. Right. Yeah, I see what you're saying. But But that sort of feels like those, those specs you see, when you look at like EEPROMs when they're like, you can only write to this 100,000 times or you know, that
actually is true. I'm actually worn out he prompts before. However, we're okay. Okay, we're trying to Yes, wrote wrote normally took actually like, three days or so to burn out 10,000 rewrites
wrote, yes, but okay, live. Okay. How about a better one? How about an SD card for a guy on using it on the camera? You know, like that? Actually, that would take away?
No, actually, actually, a lot of professional photographers take especially sports, go through SD cards, because they get really close to the limit. And they just tossed
them. Well, okay. Yes. But what I'm saying is it's difficult to do. It's not like every day,
and actually like solid state drives that you put in your computer. People are already hitting the right. Limits
on those. Well, that's actually kind of easy. If he asked me,
just suppose you have a swap file or page file on your computer? Yeah. Well, okay, back to this thing. Yeah, right. Router farm, which they don't use anymore. That's like DSL and dial up. Yeah. A router forum used by that service provider. With only this only 25 gigabytes of memory. Only 2004. Yeah, we'll actually get an error in the memory once every 17 hours. Oh, that's actually up there. Yeah, that's a potential networking error is what they're talking about. So they would get a blip, probably somewhere in a wire that causes a zero to go to one or one to go to zero in the communication. But the thing is, TCP has is pretty robust. So it's like, you lose a packet. Who cares? Right? That's kind of the whole you'll get the next one. Yeah. Well, it's kind of the point of the of the Internet Protocol. TCP is that it doesn't have to get everything.
Right. Right.
And then a person flying an airplane at 35,000 feet, where radiation levels are considerably higher than at sea level. Yeah. Who is working on a laptop with 500 kilobytes of memory? may experience one potential error every five hours. Okay, that's, that's significant if you're a Trans Atlantic flight or transcontinental flight. Yeah, you probably experienced an error due to cosmic radiation. Well, you know, so yeah, what I won't make your computer crash depends on what but it flips.
And every time I fly internationally, I bring my 500 kilobyte laptop, and I make sure that it reads the entire memory all at once and then reads it.
Well actually, it would be just to check but even worse now because your laptops got like 16 gigs in it. Windows alone uses like seven. So just that alone, your computer is probably more susceptible.
You turn it on and Windows is like thanks. Yeah,
I guess it depends on what bit it flips. Yeah. Because it could flip a part of memory that's not being used. So it doesn't matter or can be flip something that's, you know, has an air check. Yeah, so that's fine.
It's flipping the ACH bit. Oh, yeah. Yeah. And then and then everything gets.
Yeah. Basically, when you finish reading the article, the engineers bottom line. So what is your bottom line?
Well, could it be possible? Wait, get this for for a cosmic radiation Ray, to flip a protection bit in an MCU? And then you can actually read back? Technically, technically, that's possible. Yeah. So
but yeah, the party a lot higher energy event, then flipping something that's supposed to be temporary.
Well, and and the rarity is you gotta think about probability of that happening, like
DDR RAM is basically a charge, there's a tiny little capacitor that they either charge up or don't charge, and that makes it a one or zero, effectively,
the gate of FET of a CMOS FET is the capacitor that's just basically stored up and has nowhere to go. Yeah.
And so that's why like DDR RAM, you have to continually refresh it. Yeah, you have to read and write, or you lose all your ones. So that's why it's more susceptible. That's why they keep talking about memory, not hard drive, because hard drive, you have to magnetically flip that bits. Well back then. And so it's very hard to get that much energy to do that. By cosmic radiation. Probably the same thing inside your CPU. You have to drive a gate with some energy to make it open or closed. Yeah. Same thing. Oh, but
a solar flare event, or if a pulsar or something were to kind of sweep by the Earth. That would just cause everyone to be screwed up. Oh, yeah. Just go. Yeah, everything
they talk about on like those shows of like, the ways the
world war? Yeah. The history channels like Apocalypse Apocalypse Now. Yeah, right.
They talked about same thing where like a quasar or something, but high, high burst gamma radiation rents. And they been able to measure these out there. Mm hmm. But if one struck earth were boned. Well,
it's gonna, ya know, if one of them were gonna strike Earth, it's not that it's not the electrical grid would go down. It's just the entire atmosphere will be stripped out. Yeah.
All those zones gone. Like okay, we were fine here in the bombshell. Right? We wouldn't even know to run outside immediately got a sunburn.
And everything, just nuclear holocaust.
Oh, let's go back downstairs and drink some
more beer. Cheers.
Okay, um, so they say the engineers bottom line. This is a major problem for industry and engineers. But the general public shouldn't worry about it. I guess they haven't worried about sort of, yeah, yeah, sure. But it's one of those things. Do you worry about this when you design your your synthesizer? No,
no, no. And actually, I took a class. Gosh, I don't remember exactly which one it was. But but our professor, we looked at how many transistors were in a processor. And we looked at an error rate of transistors working or not just generally a failure rate. And it's pretty much guaranteed that there's transistors that are just not working on any processor at any one point in time. So you have way more of an error rate due to just silicon process you have in cosmic rays. Exactly. To be honest, I don't care. It's not gonna affect me.
You know, I actually just came to the crazy idea was that, okay? So you know how they have. This is kinda like an atomic clock where they have like, you know, they are counting the pulses of electrons on what cesium atom, that atomic clock works rubidium or CD, cesium. Yeah, something like that. Well, what if and then going off is cosmic radiation because the first thing on my mind like what have you put a cosmic radiation generator in your synthesizer? And I'm like, Parker, that's absolutely stupid. Yeah. cosmic radiation generators, terrestrial generator. So I'm like, Okay, what if instead of that, my thinking my brain, what if it's usually Reno detector? So neutrinos that just bombard Earth all the time? Yeah. Yeah. So what if you just put a neutrino detector and that created your clock cycle for your synthesizer?
Well, okay. Yeah.
So how rad would that be?
That would be that would be red. If you had a neutrino detected that was small enough to do that. I wouldn't be the richest person on earth. I don't know how big one is. There's, there's one, I believe it's in Japan. And it's like, this monstrous chasm under the earth that's like, like 10 miles under the earth is huge.
You need to submit, like a 20 page paper of like, I need a clock signal from this for my synthesizer. And that's like, 20 pages of like, a doctorates that you get. Right? And I bet you if you did that they would approve it.
Yeah. And they'd say, like, please go find some neutrinos for me.
Or something like that. That's like, you're actually counting cosmic radiation. And that becomes your pulse feed into your I mean, just you could do it with a sensitive Geiger counter. Yeah,
this cosmic radiation isn't predictable enough. That's the point. To just have it unpredictable. Yeah. I guess for just random noise.
Yeah. So one of the commenters on this article, yeah. Has a posted a link from Intel until I mean, look real quick. Yep. Until so back in the day, back when it was like semiconductors was a new thing. And they were trying to Intel was trying to figure out why they were getting random errors. And their their CPUs, and this goes back to your, you know, sometimes transistors just don't work, right things. And so they created the world's largest led safe. And their idea was, if you make two board, two computers, put one in the lead safe one outside, the one in the lead safe, it's protected by cosmic radiation effectively, and will have less errors. And they found out both have the same amount of errors. And there's actually alpha particles being generated by the encapsulation of I think, was one of their transistors. Let's see. Was it was the transistor breaking down? No, the transistor. The encapsulation was made out of a radioactive material. Oh, so it's bombarding alpha particles. Yeah, it's a way we but here's the thing. Damn, yeah. Yeah, it was alpha particle emissions from thorium, which was in the was inside the encapsulation material. So like, what they put on the package, they had some three minutes. And that was causing alpha particle bombardment of their CPU, die.
We okay, this is what's really getting to me right now about this. Like, I would love to check out the validity of this comment. I mean, I sure it could be true, totally true. But the thing about it is like the management at IBM, he said, oh, sorry, Intel, Intel must have trusted their engineers so much, because there had been some engineer who was like, Yeah, we need to build this really big lead box. Put stuff in it. Because, well, the magic space rays are coming in messing up our computers, and we need to test that.
I mean, that'd be cool to do. Totally cool. But
like management must have been like, Yeah, sure. That sounds great. Let's do it. No, well, it's crazy.
But this way is they built that lead box, they can just tell their customers, hey, we can't do anything about it because it's constant radiation. But they actually found out was thorium is in their encapsulation material and they can't get rid of it. Because thorium is kind of an I guess with enough money, you can purify the encapsulation material enough. But basically, they just decided to design memory that can withstand the alpha particle bombardment of the capsulation material. Hmm.
So that's crazy. Yeah. Engineering. Yeah,
that's it that's, I put away the early they tronics the 60s guys pretty crazy time.
It was the wild west
of ideas of how to test
especially like you build this entire room of lead and then the results are negative,
negative. Yeah, or Yeah, so guess how much this safe weighed? I just read it off the page.
Come on. I was gonna look really smart. You can mess it up. Something like what I don't know. 25 tonnes?
Yeah. I want to how much lead that actually is minus 25 tonnes Tony. Ah, I guess with that, that was the engineering Mac fab engineering podcast. Yeah, Episode 59. Yep. Oh man, almost number 60. We are your hosts Parker DOMA and Steven Craig later everyone take it easy. Sorry for the long episode was good fun though.
Why is there such a disconnect between component datasheet drawings and EDA footprint layout tools? Stephen and Parker dive into this on this podcast.
Stephen is on the hunt for the next step in his electrical engineering career and shares the shifts in the industry and what employers are looking for.
What are the pros and cons of becoming an electrical engineer? With great power comes great responsibility or just the lack of sleep!