#559 – Occam's Engineering Razor

Chris Gammell: This is The Amp Hour Podcast. Release September 26th, 2021. Episode 559. Occam's Engineering Razor.

Dave Jones: Welcome to the Amp Hour. I'm Dave Jones from the AEV blog.

Chris Gammell: And I'm Chris Gammell of Contextual Electronics. Two weeks in a row. Yeah. Yeah. You know, we got things to say, possibly. Do we? All right.

Dave Jones: Well, you know, I can.

Chris Gammell: Electronics and things.

Dave Jones: Well, I can start off. I can start off. As usual, you have no idea. I'm going to say this. It just popped in my head. Have you ever had, I'm sure you have. It's kind of a rhetorical question. It happens to all of us. Have you ever had one of those moments where, well, it's not a moment. It's a whole saga of where your engineering brain kicked in. You've got some problem, right? Something's gone wrong. And your engineering brain kicks in and just thinks of all of the possible things that could possibly be. And you go and investigate all of those. And it turns out to be like, no, the simplest, easiest solution. Because you break. Occam's razor. Yeah. Anti-Occam's razor. Yes. The engineering Occam's razor. Yes. The engineering razor. Have you, you know, obviously you've had that, right? I mean, that goes without saying. Oh, 100%. Yeah.

Chris Gammell: It's like the curse of knowledge type of thing. Yes. Yes. Curse of knowledge. Yes. Yes.

Dave Jones: You know too much. It's.

Chris Gammell: So what happened? Obviously you have a story here.

Dave Jones: Yes. Yes. I have a story with that. That's probably going to, I'm going to go edit a video after this showing it. But so maybe it'll be LinkedIn by the time we do this. But yeah, it's my home solar system.

Chris Gammell: I'm looking at the tweets about it. Yep.

Dave Jones: Yep. I recently. I recently. Yeah. So you probably know what the answer is. I recently installed a new five, five kilowatt system. Right. And, but that supplemented my existing three kilowatt system. So what I did is I moved my existing three kilowatt system from the more ideal roof where we wanted to put the bigger one. And we put it on the other side of the roof, which was less ideal. In fact, it's fairly poor roof angle to put it on, but it's better than nothing. Right. What am I going to do? Sell it? Scrap it? No, I might as well just move it. Right.

Chris Gammell: Peen it up in the backyard against the fence or something. Yeah, no,

Dave Jones: exactly. Right. So yeah, I thought, yeah, I'll still get, yeah, and I might get half the, half the energy out of it that I used to because it's on the other side of the roof, but who cares? Right. It's already paid for itself. It's free. Might as well move it. Right. So I moved it and then everything looked to be going fine, but then just like a week ago or something, I looked at some perfect days, which we've been having recently. Right. And if you've ever seen like the solar curve on a perfect cloudless day, it's just this smooth.

Chris Gammell: Yeah. You have a video of it that we'll share on your Twitter. Yeah. Yeah.

Dave Jones: Yeah. Right. Yeah. It'll have a beautiful curve. And then I noticed that, no, it had all these little, it had these dips in it and stuff. And I thought, oh, that was just like a cloudy day. The cloud came over. But then I was like comparing data with my old new system. And I noticed that my new system showed a perfect curve for that same day. And my old system, which I've moved, didn't. It showed this big, you know, wiggles in them, as I call them, right? These big dips and everything. And so I started looking into the data and sure enough, every perfect day, which showed up on my new system. So there were, there were no clouds at all in the sky. I should have got a perfect curve on my other one. And I did it. Right. So there's something weird going on with my three kilowatt system that I moved. And then.

Chris Gammell: And you start diving. So to take it back to the thing, you, yeah, you start, well, you start thinking like, well, it could be how the internals of the inverter or whatever. Right. All these other things.

Dave Jones: It could be the inverter, but the inverter wasn't. Okay. You know, so we can go through all the different scenarios, right? We can go through. Yeah. The troubleshooting. Right. Obviously. And, okay, let's go for the most obscure complex one. Let's go. Oh, they, you know, these guys on, you know, these contractors on the roof were physically thrown around these panels, moving them and stuff like that. And, you know,

Chris Gammell: I should measure the, the, like that, that thing that Martin did where he showed like the cracks in solar panels. I need to go and measure every solar panel. Yeah.

Dave Jones: Has these micro cracks and things. So I thought, you know, look, there could be something wrong. You know, one of the panels could be damaged. And because they're all in series, right? It's a big series string of 12 panels. If one of them goes bad, it, it will get bypassed and take down the whole, you know, it can drop your whole system. Right.

Chris Gammell: Yeah. Right.

Dave Jones: And so, yeah, so it could be that, or it could be, you know, a dodgy connection or something that's, you know, heating up some sort of, or it could be, you know, thermal issues, everyone, you know, so I was putting all these theories on Twitter and everything and everyone's coming up with, Oh, it's, it's gotta be your inverter or it's, you know, so it's, it's, it's thermal. It's, you know, this,

Chris Gammell: it's more useful than having people suggest things. Oh, on Twitter. Yeah. That's why I put it on there. Yeah. Yeah.

Dave Jones: Because it's just, I just love to give people bait to then just, you know, kind of just throw wild theories out there. Everyone said it's the inverter getting hot. Post your temperature logs of your inverter. It's a, well, the inverter hasn't moved. It's in the same shaded spot. It's been for the last seven, eight years. So, you know, it's not that. And mains voltage, eight out of 10 engineers said it was mains voltage.

Speaker ?: Right.

Dave Jones: It was your mains voltage is getting too high because you've installed a new system on there in parallel. It's causing the main voltage to boost up and it's causing shutdown of the inverter, like temporary momentary shutdown of the inverter and stuff, you know? So that was the leading theory. So, you know, I got my data logger out. I haven't actually looked at the data yet from the data logger, but it obviously wasn't that. Right. And then it turned out as,

Chris Gammell: no, before you get to the conclusion here, cause I think it's important. Yeah. I actually had a conversation with my wife about this recently. So how rigorous were you in your measurements? Right. So I saw you took a video and you had that.

Dave Jones: I took time-lapse video. Yes. I took time-lapse video. I went and installed my GoPro up there to see that there wasn't any shading or anything like that. Yep.

Chris Gammell: We said, you also had data logging and you.

Dave Jones: Doing mains, mains data logging as well.

Chris Gammell: Yep. Did you keep like a journal of like, as you were trying these things or was it kind of more ad hoc?

Dave Jones: No, this just happened over a day or two. You know, this only happened over a couple of days. This wasn't months or anything. No.

Chris Gammell: Got it. Okay. Okay. Proceed, please. I like to come back to that, but the rigorousness, but yeah. Yes.

Dave Jones: No, it's, it's, and you know, it's something. So anyway, I set up a time-lapse camera on top. Cause I thought the first thing I thought was that,

Dave Jones: it's shading because you know, it's, it's on a poor angle and there's tree, you know, there's lots of big trees around and stuff like that.

Chris Gammell: So maybe it's just coming into the, it's coming into springtime there. Right. So it's like, yeah, the sun is getting higher in the sky and you'll have a better

Dave Jones: higher in the sky and stuff like that. And yeah. But the other thing I noticed is that by analyzing the data is that it seemed to happen about the same time every day, about 9 45 AM. It started to happen. That's why I thought, Oh, okay. Shading. The sun's coming behind a tree that's in the neighbor's yard or something. And, and I don't know unless I'm up there. Right. At that exact time. So I set up a time-lapse camera and didn't really show anything. Right. It was a perfect day. And a time-lapse camera didn't really show any problems. So I thought, Oh, it's not the shading. Yeah. Wah, wah, wah, wah. It's the shading folks. Yeah. It turns out that I've got a TV antenna up there. Right. Which once again, the, the shadow of the mast on this big TV antenna, it's right next to the, and yes, it is between. Tell me,

Chris Gammell: tell me, Grampy Dave, what's a TV antenna?

Dave Jones: A TV antenna. It picks up this newfangled digital signal that,

Chris Gammell: when I was a boy. You see kids, when you don't want, when you want to watch Netflix, but it's always on. Well, the internet. And then in the middle, there's these things between the, between the, the, in the middle of the program, these things called commercials come on.

Dave Jones: And TV's filled with those anyway. Yeah. Yeah. Right. But, but I, I knew that the shadow of the mast didn't come over until like, like, you know, 12 o'clock or something like that. So I thought it couldn't be that. And I didn't see it on the time-lapse. Right. Although I, I'm going to have to view it in high def. It might be with hindsight. Now I might be able to see it, but anyway, it didn't correlate with the time. Right. It was like nine 45 every day. Boom. It starts cutting out. Right. Which, which made sense for the voltage being an issue, right. As the sun's coming up, like everyone on my street has solar right now. Like I'm like, I don't think there's a single house on my street that does not have solar.

Chris Gammell: Right. So you're saying the, the grid grid starts to get packed.

Dave Jones: Yeah. The voltage starts to get boosted on the grid. And I was already high. It was already borderline. Right. I was already, I knew I was at already like 248 volts. Right. And the maximum grid voltage is 253. Right.

Chris Gammell: Oh, interesting.

Dave Jones: And I knew it did actually get to like the, the, the inverter has a voltage display on it. And it says like 251 at times, you know? So I thought, yeah, yeah, you know, it's pretty darn close, you know? So the leading theory on the interwebs was that, yeah, it was, it was the voltage getting too high and the inverter's been tripped out and they don't give you that information in the inverted data sheet or whatever. So anyway, yeah, no, it's the shading, but how it worked is of course, the antenna mast has to have guy wires, right? It's got to have these very thin steel cable guy wires. And there's a coax of course, running up one of them as well. Right. There's a coax cable, but I thought that's, that's, that's too far away. Right. This giant sun up there, right. You've got this giant light source up there and you've got this little piece of coax, like five meters away from it. Surely that's not going to cast a sharp enough shadow.

Chris Gammell: Sun's more like a point source though, you know, like that's. Yeah,

Dave Jones: it is. But compared with the, like the distance away of the coax, like, and, and, and sure enough, the, the shadow of the coax is this like just this fuzzy little thing you can barely see, but that was enough to drop the output by like 20%. Wow. Of my entire system. Just this tiny little slither of a fuzzy coax cable coming over the panels, shaded thing. It's like, it, it wasn't a sharp shadow. Like it was just this fuzzy little thing. And yeah, dropped at 20%. And that was it. That was it. So after all that, it was, yeah, it was the simplest solution. It was simply that by that, that was the first thing I thought of. Oh, my antenna shaded it. And then I thought, nah, it can't be. It's too far away. It's, you know, the mask doesn't come over. I know, I know the shadow of the mask doesn't hit it until 12 or something like that. And surely that's not going to drop it by 20%. Right. Right. Right. You know, it's like, yeah, but there, there you go. So a tiny little imperfection in the insulation coming on just one panel, just one panel was enough to take out my entire string by about 15, 20%. Remarkable.

Chris Gammell: I had a, I had a scenario the other day where I was, I forget who it was, but I was like guiding someone along and I was like, Oh, you know, you really want to like, uh, it was like they were setting up like some new system and they, they, they had a sensor in there. I was like, Oh yeah, you really want to like validate the sensor and like, make sure you really understand what it's doing and the output, the inputs of that sensor. And just make sure you understand it as an isolated thing. I was probably like echoing when Philip Johnston was on the show. Cause he was talking about it. And then I was thinking like, I don't, I don't do that.

Dave Jones: Right.

Chris Gammell: That is some BS advice I just gave, you know, like do as I say,

Dave Jones: not as I do. Yeah. That's right. Yeah, exactly.

Chris Gammell: Yeah.

Dave Jones: Yeah. Yeah.

Chris Gammell: And so it must've been someone younger. Cause that's like the kind of thing. Right. Yeah. Yeah. Yeah.

Dave Jones: You're going to the grandpa mode. Yeah. Now that you're a daddy, you know, you go. Yeah, exactly. Well,

Chris Gammell: and so that, that's what I was going to come back to as well. So like the, so this guy, this is a conversation I was having with my wife the other day about, you know, teaching our daughter and just like things that I hope that she'll learn. And, and one of the things was like, you know, creativity and resilience was like one of these things that we were kind of talking through. Yep. And just like, I forget, I forget the, the path of the conversation, but I know at some point I brought up Ben Krasnow and the video where he did the, the laser etching of traces on a plastic. Do you remember that video?

Dave Jones: I think so. Yeah.

Chris Gammell: Yeah. So he was basically trying to replicate the demo, which actually was the, the company that Andrea Spies helps found the, the Swiss accent guy from YouTube has also been on the show. So we'll link all of them then. But basically it's like they, you go to like trade shows and they put, they put traces directly into plastic and Ben did this experiment. And so I've definitely mentioned on the show here before, I think when Ben was came back on, but like when he says, he's like, well, on, trial 66 and experiment 66, I started to, I started to finally get some, some, some traction or something. Yeah. And then it was like, but, but it was actually experiment 83 where I finally got it to work perfectly. And it's like, that is what I would like to impart to my daughter. I think I was sort of talking about was like wanting to have some kind of like human biological way to like, like to be rewarded for not knowing. I think this also came up because there was, there was a post where someone was talking about like, they, they had a friend, they were in grad school. The friend dropped out, went to another career because they said they didn't like not knowing. And then they talked about like how that's all of grad school and engineering, science, really just anything where you're trying to make new discoveries. And it is, it sucks. That is a terrible, terrible feeling. But the people that are really good at it, Ben being one of them, other, you know, obviously other researchers, people that are just very resilient in that way. Like you have to be able to sit in that discomfort and like, keep trying new things. Like all of the troubleshooting stories that we tell on here, obviously people have heard all of our troubleshooting stories, save the one you just told, I think at this point, but like that troubleshooting stories are stories of resilience. And I think that, you know, there there's, there's a balance between like, you know, ramming your head into a wall until you break through. There's, there is that kind of resilience. And sometimes that works. Sometimes you're just going to crack your head, but the, the rigor and like, that's why I asked about rigor as well. Yeah. Right. One thing that I've in my younger self would just keep trying new things. And that's kind of the banging your head against the wall without actually doing anything. I think once you start to take notes and you start to be more rigorous about it, like Ben in that example or other, you know, like just scientific method in general, like that. It's like, once you get into that mode, I just feel like you, you just kind of level up as a, as a scientist, as an engineer, like just as a troubleshooter. And like, that's something I still struggle with. You know, that's something I still want to get better at, but people are really good at it. It is so impressive.

Dave Jones: Motivation plays a role as well. Like you talked about researchers and stuff like that. Like if you're doing your PhD, right, you've got to persist because you don't want to piss away your three years. You've been working on this PhD, right? You don't, you don't just want to quit. It is the sunk cost. But if you're a Ben Krasnow and you're just doing this as a hobby thing on the side, that, that takes a different kind of person just to, to do that as a, like as a hobby, for example, there's, there's a big difference between a, you're being paid to do this or B, you've got some sort of, you know, reason because you're, you know, you have to finish your university degree or your PhD or whatnot. Right. And just know you, you have no reason to do this. Right. But you're going to do it anyway, because you just love the challenge. I, that, that you can't really, I, I don't think, I think you'd have a hard time trying to teach somebody that.

Chris Gammell: Well, and that's what we, that's what we started talking about too. Right. So my wife and I were like, well, how do you actually teach? You know, like, it's not really something, yeah, you can't teach it, but you can encourage it. Right.

Dave Jones: You can encourage it, but whether or not they take it up, that's probably dependent upon their own personality and whatnot. And there's tons of other factors. There's a million other factors. Sure, sure, sure.

Chris Gammell: Nature versus nurture type of stuff. But like one thing that my wife talked about was like her own, her upbringing and her mother, like inspired. Confidence in them. And just, you know, really supported that. You know, I think feeling supported, feeling like you have the ability to fail. Super important. Right. I think having low stakes, but also, you know, like personally when I'm in like super high stakes, if I'm like, I started to get into thrash mode, that's what I've talked about on the show before. I, I start thrashing. I'm just like, Oh, I got to try this. I got to try this. I got to try this.

Chris Gammell: Yeah. Where almost always the right way to do it is to like slow down. Yeah. And like analyze things.

Dave Jones: Analyze the data and say, you know,

Chris Gammell: go for the most likely scenarios. What really happened here? Right. What, what is the most likely thing? Like start to make a list and order things out. And it's just like, right. And that is so hard when you're, when you're just stressed. And like, No,

Dave Jones: because engineers are usually under time pressure and we, that's, that's what we do. Thrash mode is our bread and butter.

Chris Gammell: Right. Totally. Well, and I think, and most problems, right. So like most problems are not, you know, there, there might be something you're missing that, you know, like something's example you gave, right. It's like a, a small thing, but like something that's, you know, it was up on your roofs. Not like it's super easy to get to. Right. But like, it was a non-obvious thing. Right. Yeah. Yeah. And if it was something more complex though, right. Then you might keep trying simple things, but you'd have to then increase in complexity. And it's just like, again, it's just like that rigor. That's really hard to impart. I would love to hear from people if they, they know ways to like, either they've made themselves better or they helped teach others. Like that is just, that is something that we could all benefit from.

Dave Jones: And also, cause I was doing this in real time on Twitter essentially. And I couldn't devote all my time to it. Cause either I, I was here at the lab or be, even when I was home, the kids were doing the home schooling thing and they, they had other stuff. I had to monitor them. Right. So I, I literally couldn't go up on the roof. Cause I was, I was inside. You're doing your homework on the roof now. Right. Yeah. Well, actually that did happen. I actually went up there on the roof. I said, I'm going up on the roof for five minutes. Call me if you need anything. And it turns out that of course he needed me and he shouted out to me and I didn't hear him. And, Oh God, you know, it was like, you know, and I felt really bad. Cause I spent five minutes on the roof. You know,

Chris Gammell: that's what he's going to be taught sitting on the therapist couch. And then dad was on the roof. So,

Dave Jones: and that would be the opportunity you have to investigate that path. I mean, some things are just easier to investigate, even though you might not think, you know, Oh, they're not, that's doesn't stand much of a chance, but it's easy to do. So I'll do it. Right. Right. And then you'll do a bunch of those when you should have actually, you know, looked at, okay, which is the most likely and thoroughly and went to the effort to investigate that one. Yes. Right.

Chris Gammell: And it's like this upfront investment that you do. So like rigor is an upfront investment or like writing down your hypothesis and like, okay, I'm going to try this thing now. And then how did it work out? Yes or no. And like, that is an investment that like, it does not, you know, if it is always these simple things, it's not going to pay off most of the time. You know, you're going to spend more time than you would if you just kind of like guess that it got it pretty, right? Whatever. Yep. Same thing with like setting up experiments, setting up. Yeah. Yeah.

Dave Jones: Well, the solar thing was a classic, right? It was easier for me to set up a timer. Like I thought, Oh yeah. Shading. I'll definitely check that. Right. That is the most likely thing. Right. So I set up a time-lapse camera and I put it up there. Right. And because that was easy. Cause then I could set up in the morning, go to work, come back and it's, you know, right. And it's done. Whereas, and, and it didn't show anything. So I wasn't, you know, I didn't rigorously go, well, what if there's a little bit of shadow that I'm not seeing? I, and I needed to be up there right at the, right at 9 45 AM when it started happening. Right. And then only when I had time to actually do that, like, like a day, a day or two later that I thought, right, I'm going to be up there at 9 45 AM and I'm going to watch it. Sure enough, the coax cable shadow just barely comes, you know, you can barely see it. Bingo. That's got to be it. I wasn't thorough enough.

Chris Gammell: I think another thing that's in there is like consistency of setup. Right. So that's like another thing that really is helpful. Right. So in your case, it was, that was a really good move. I think to like put a camera up there. That's like a consistent setup. You get to monitor over time,

Dave Jones: but that only got gross shadows. It, it only, it didn't have the resolution required to see the detail that actually mattered. Right. But I didn't know that at the time. I thought, you know.

Chris Gammell: Right. See, but you could call that automation. You could call that like, you know, you could call that monitoring systems, that sort of thing. Like if it's, it's a less manual process, I think that just kind of, again, that was another investment where you took the time to set it up instead of like, yeah. Yeah.

Dave Jones: You know, it's one of those things. I was, I was bound to come a gutter like, you know, and it was, yeah.

Chris Gammell: Yeah. I think, I think there's the consistency of setup thing though, too, is like another one for me that like I struggled with in the past because I would, you know, so when I was working at a, the coworking space in Chicago, I had the setup, I tried this thing, I get to the end of the day. I'm like, tear it all down. And then now I'm starting all afresh and there might be new errors that are in there. So like, this is where I feel very lucky, very privileged to have like a workshop now, but like that can really, that can change a lot of things too. Oh yeah. Well, we've talked about that many times where a dedicated, you know,

Dave Jones: parts on your bench. I talked about this in the mailbag video release yesterday. I was going through ex devs who's on the EV blog forum. I went through his bench, right?

Chris Gammell: Well, that's the last, last week we talked about the, yes, yes, yes. Yeah. Same guy. Yeah.

Dave Jones: Anyway, I'm going to release it as a second channel video as well. Just a tour of his lab and you know, everything's compartmentalized, right? He's got these different stations for different things. He's got ones just for burning, you know, long-term temperature testing. It's set up. There's his thermal chamber. There's all this long-term stuff where he logs everything and does everything. Then he's got his repair station. Then he's got his, you know, other hacking station. Then he's got his test stand, a test standards station. And it all just, it's all nicely set out. So he's got that dedicated space where, yeah, he doesn't have to cobble something together for long-term testing. It's all there. And it's permanently set up. That's right. Right. So he can just put it in, go leave it for three months and he's got his results. Right. So, you know, yeah. Whereas me, if I have to do this shit, I've got to cobble something together. It's like, you know, yeah. All right.

Chris Gammell: Yep. Well, that brings me to the, to another link on our, our subreddit. Jay Carlson did a tour of his lab past guest, Jay Carlson. And Jay is someone that I was talking to when I was asking about Twitter, about, about workshops and stuff. Jay uses the trade. Right. And this is a couple other people have done this. I think Lewis was another person on Twitter.

Dave Jones: Is this the project tray method you're talking about?

Chris Gammell: That's right. Yeah. So instead of a project box, it's a tray. And basically you just pull it out, work on top of it and you're done. And that's awesome. Yeah. I mean, I remember you've done, I've done,

Dave Jones: I have physical boxes. Yes. Cause trays are, you know, you've got to have quite, you know, cause my tray in quote marks or my box, that includes all the parts as well. So it's, you know, a tray is not the best kind of, you know, if it's just a board and a couple of other bits and bobs, then yeah, you can put it in the tray, but you know.

Chris Gammell: Yeah. I think it's, it's different if it's like an assembly versus a trouble. I think tray for troubleshooting a box for assembly. And it's got,

Dave Jones: and by talking about that, like it would have, like, you might have like a, a little debug program or hooked up to it or something. You leave it hooked up that sits in the tray and it goes away with the tray. Like, and, and then all the connections are all hooked up, right? You don't have to dick around. We're trying to find the connections and get them wrong. Right.

Chris Gammell: Somewhere where I've had a lack of consistency in the past. Why is this not working? Oh,

Dave Jones: it's because I've put the header cables on the wrong way. You know, I swapped a pin and you want, and you spend half a day bloody troubleshooting that thinking it's your project. That's fine. When it's actually no dumb ass. That's right. Dave has just swapped the, you know, transmit and receive on the bloody debug header or something, you know?

Chris Gammell: Right.

Dave Jones: It's like, Oh God. Yep. Yep. Yep.

Chris Gammell: I know that one. Yeah. And I think that the problem too, is like when that is code involved, I've been trying to get better at this too, but when there's code involved, I almost always default to the being my fault. I did something stupid. Whereas the reality is it's like, it's some externality. That's like, you know, like why would the code won't just change? I mean, bit rot's a, you know, something we laugh about, but like, it's not something that actually happens, you know, like bits don't just flip. Rare occasions. Cosmic rate. It is incredibly,

Dave Jones: it's so rare. You could rule it out, you know? Like, yep.

Chris Gammell: Yeah, exactly. Yeah. Yeah. Yeah. You get there, you've won the lottery basically. Yeah. Yeah. Yeah. And so, and, and I, so then I start, like you said, I just start to tear things down and then it's hard to get back to where I was unless I've been very diligent about revision control and everything like that. So yeah, it's a, it's, it's good to take a deep breath when that, you know, like, I think one thing is like recognizing the feeling, like, I feel like that manic feeling of like, Oh, I gotta, I gotta try this and try this and try this and just be like, yep. Yep. And you waste so much time and you get in such a state,

Dave Jones: you work yourself up into a, you know? Yep.

Chris Gammell: Yeah. Yeah, exactly. Yeah. And then, you know, take a walk. That's a great one. You know, just step away, come back, whatever.

Dave Jones: Oh boy.

Chris Gammell: Yeah. Drink waters, you know, all, all useful things. A water glass is a great troubleshooting. Rubber ducky is a great water shooting tool or troubleshooting tool.

Dave Jones: Speaking of shop tours.

Chris Gammell: Yeah.

Dave Jones: Adam Savage on tested, just posted a tour of Grant Imahara's shop. Rip Grant Imahara.

Chris Gammell: Yep.

Dave Jones: Yes. One of us.

Chris Gammell: Yeah.

Dave Jones: Oh, it was a year. It was a year actually. I just saw,

Chris Gammell: I just saw a thing about that. Oh, see,

Dave Jones: I thought it probably would have been longer than that, but then again, everything seems long during COVID,

Chris Gammell: you know? No, maybe I saw a, actually, maybe I saw, I saw a thing that was like a remembrance video. Oh, it was 11 months ago. It's possible. Maybe it was. Yeah. But, Oh, it's very,

Dave Jones: he meticulously designed it from the ground up. It was, yeah, it's, it's very compartmentalized. It's very, it's, and it's set up. So it's optimized for efficiency and stuff. It really is quite nice. So, yep. It's, you know, he's got like big CNC machines and, you know, stuff like that. It's very cool. So, and, and it's still exactly the same a year later, they haven't actually touched it. So,

Chris Gammell: yeah. Yeah. You kind of wonder about, I mean, like when all that stuff kind of, stuff kind of goes away, you know, like, you know, like your family. Oh yeah. No,

Dave Jones: all my stuff is just like, yeah, it'll, it'll, but yeah, they'll sell it for a song on eBay or something. Yeah. I'll give it away. I don't know. Yep. I don't care. I'm not really fussed over what happens when I'm dead to my stuff when I'm dead, really. You know? Yep.

Chris Gammell: Right.

Dave Jones: I do not. Yeah. My kids will get,

Chris Gammell: there was something else about Jay's, Jay's, uh, workshop tour that I wanted to, I forget what it was though. Like he does a really nice work, work down of the whole thing. Obviously Jay's in Nebraska, lots of room there. but I think, yeah, I mean, it's just, you know, Jay always writes great articles about embedded and everything else anyway. So it's, it's the same kind of thing, but I cannot remember what it was at this point.

Dave Jones: Anyway, it's easy to look organized when you have space.

Chris Gammell: Yeah. I agree with that. Yeah.

Dave Jones: Yeah. It's, it's, yeah, it's so easy. Like I would love to have like, instead of a tray per project, I would love to have a bench per project.

Chris Gammell: Yeah.

Dave Jones: Like seriously, like if I was super rich and I had like, you know, 500 square meter space, like I had an entire top floor of a commercial office building just to myself, you know, I would literally have a separate bench for each project. And there the project would sit on the bench with, with the required test equipment. Cause I've got so many bloody scopes and meters and stuff. It wouldn't matter. Yeah. And everything would just sit on that bench and that would be the project bench. So I could just walk up to the project bench and spend an hour on it. You know, that'd be, yep. Which is a very inefficient way to do engineering, by the way, is to just spend an hour here, hour there. It doesn't work because it takes you an hour to get up to speed, you know,

Chris Gammell: or a startup cost.

Dave Jones: But anyway, that'd be cool. Right. I'd love just to be able to walk up and tinker on, you know, a thing, but now I've got to know I've got to put it away. And then once you put it away, it's gone. Right. It's like, yeah.

Chris Gammell: Yeah. I've, I've been actually feeling that on the, the 3d printing space. I've been printing out some cases. I've been posting about some of my progress. So just like, so I've been doing these off the shelf modules. I just got like some cheapo, like external MOSFET modules and then like breakout boards and all these other things. And I'm basically kind of like, almost kind of like how a PLC wiring cabinet is. I'm kind of like replicating that. Right. It's just like wires. So just like point to point wiring for modules and stuff like that, instead of doing a board. But I've been trying embedded. I actually, I succeeded in embedding M two and a half and M three. Yeah. Right. Into the print. And so that I had captive machine screw threads. And that was an, an exciting experience, but not ultimately something I think I'm going to do that often.

Dave Jones: And yeah, but don't you just make it the thing to a certain size and then just shove them in. Um, are these the press fit ones?

Chris Gammell: These are not the press fit ones. So I, I, I had some press fit ones, but the wrong sizes. And so in the meantime, I was like, Oh, I can.

Dave Jones: Right. Okay. Cause they are the best solution.

Chris Gammell: Well, I think the other thing too, is that I was, I was remembering putting together an old 3d printer and like, you know, like there are certain formats where you can like slide in like a nut and that works great. You know, basically the slots already there. You slide, you, you know, you basically drop it in, you drop in a nut and then, uh, and then you're able to screw in like kind of like horizontally to it. But the thing that my brain didn't realize is that that's because of how it's printed, right. It's printed with like an opening at the top. And then it basically becomes like a Plinko thing. You just drop the nut into. In this case, I'm actually trying to do it horizontally. So I actually like built up these supports, pause the print, inserted a nut, and then tried printing over top of it. And that is not the way to do it.

Dave Jones: No, you don't want to be doing that. No, no. The, uh, fluted press fit. Yeah, exactly. Yeah.

Chris Gammell: And that's what I, that's what I landed on as well. Like, yeah, like the heat set M3 press fit things like very easy. You can get, you can even get like a, a weather tip. That's basically just like a brass. It's just like a brass piece of brass. Basically it's been machined. And you just like use that and heat it up and goes right in and no big deal. But, uh, you know, just, uh, hold downs and cases and all that other crap that comes along with assembling larger things.

Dave Jones: I, I kind of miss the days when everything was in 19 inch racks. Yeah. Yeah.

Chris Gammell: I mean, that's, yeah, I think starting from a large space is definitely a nice, a nice start. As you get more and more complex, it's definitely, I think it's good to start from the big, the big side of things too, and then kind of shrink down later. That's another lesson I, I keep forgetting.

Dave Jones: I'm like, Oh, probably like two thirds of the, of the production test equipment that I've designed over the decades. It has been rack mount, you know, it's been nine inch rack mount stuff.

Chris Gammell: Yeah. That's nice. Yeah. And then it's, yeah,

Dave Jones: because it's nice to have the luxury of, you know, just like not having to think about the cases. Like, of course it goes into a nine inch rack. Cause we have all the test instruments in there. We have the nine inch rack mount PC to control it. Where else would it go?

Chris Gammell: There's no space on the bench, obviously.

Dave Jones: And all the wiring's at the back. Like there's, there's no connectors on the front unless you're absolutely need it for some, you know, use a plugin test reason or something, but generally, yep. It's all of your decisions made for you. Nine inch rack, all the connectors on the back, you know, it's like, okay. You know, it's really simple. Yeah. Yep.

Chris Gammell: That's the way to do it. It's good. I think just having like the, kind of like the mental model around like a mechanical assembly is, is good to have these days, you know, as, as things keep getting smaller and like, I don't know. I still, I don't print anything. That's pretty. That's for sure. It's all still like black square rectangle. You know,

Dave Jones: I've done videos on this and I'm sure we've mentioned on the Ampair over the years is that often a good way to design a project is to start with the case. Yes, we did.

Chris Gammell: Because you made fun of me on one of your videos without mentioning me.

Dave Jones: Did I? Right. Did I? Yeah.

Chris Gammell: Cause I was asking about, I was asking on here. I asked you on, I took a chance, asked you on air for some help sourcing a case. And then you, you know, you look at the camera like, Oh, someone couldn't figure out there was a battery holders or something like that. And one of your videos. Yes.

Dave Jones: You can buy cases with battery. Yes. I know that. And then design and then design your project around the case.

Chris Gammell: I think that's right.

Dave Jones: That's, you know, that's a real, you know, it's obviously not for everything. I think in, yeah. But if you don't know, like if you're like, if you've got the idea for a project, like, okay, it does this thing and it's got to have some sort of usury form factor kind of stuff, but, but you're not, Oh, I don't know. What does a case look like? Then just browse the catalogs until you find someone that looks good. You know, if you need batteries, go search cases that have a, you know, if you need a nine volt battery, go and search for ones that have a nine volt battery snap holder built into them. Yeah. You know, like, and then work from there. And it's, it's great. Cause once you have the physical thing in your hand, you go, Oh, you know, all the dial, all the dials and switches in your brain light up and you go, Oh yeah. Okay. Yeah. You know? And yeah, it's great. Works, works wonders.

Chris Gammell: I have a, I actually, to take it back to the, I just remembered from one of the projects that I have, that's in a nice case. I had, I had a troubleshooting story this week and it ended up Occam's razor, same kind of thing where like, I had like multiple, uh, selectable like feedback resistors is like a boost converter. And it was like five, 12, 24 volt output. And so like, basically the dividers had things coming back. And I'm like, man, this thing is just not out. It's just like outputting, you know, it was like three and a half, three, three and a half volts input. And then like three and a half, three point three volts output. And I was like, Oh, am I, Oh, maybe I'm configuring this wrong and you know, whatever, whatever, whatever, whatever. Yep. Nope. Hadn't, hadn't turned on the enable pin, you know, just one of them. Yeah. Classic.

Dave Jones: Yep. Yep. Or you leave it floating. Oh, that's a classic. Yes. Because you forgot to solder that you, you, you miss that solder joint and it's just making intimate and contact and floating. Oh goodness. You know? Yep. It's just, yep. That'll, that'll,

Chris Gammell: high quality troubleshooting stories right there. Oh yeah. Oh boy. Yeah. A lot of fun. A lot of fun.

Dave Jones: Love it. Yeah. All right. What else we got on the list that we didn't cover last week? Cause we didn't know we'd do it this week. That's right. You should.

Chris Gammell: What else is on the list? I had sorted by top. Oh, you know, just Intel lying. Lying? Well, okay. There's been all this to do about like, Oh, you know, the U S is pulling chip design back from overseas and blah, blah, blah. Who's going to do it. Oh, Intel is going to do it. And then there's this video, there's this article on, on fortune.com, which, you know, starting from a bad place, but it's basically like, uh, what the link is car makers need to get out of the semiconductor stone age. And they're like, basically we don't make anything bigger than 16 nanometers. And it's like, of course you don't like your Intel. Like, but why would you, but you know, nothing is, you know, there's actual like analog reasons to not have, you know, yes, you can operate at these lower. So if people want to go and listen, when, uh, uh, Adrian Tang was on the show, he talked about like, there is possible. So Adrian was the, uh, the JPL chip designer. He's like, yes, it is possible to do this sort of stuff. Most of the reason it's not done is because the package is not there. Like the it's, it's not available. You know, like you can't get access to the analog design package for the lower stuff, but also there's like tons and tons of leakage at 16 nanometers that you have to do all this crazy shit there. And like, if you're, you're trying to make like a trench fat, you're not going to use 16 nanometers to do it. You know, like you want, you want millimeters effectively. Not really, but like millimeters. Damn it. Yeah. I mean, like, it's just like, just the, it just says a really, I don't know, tone deaf kind of thing. Like the chief executive of Intel says I'll make as many Intel 16 nanometer chips as they want. It's like, okay, cool. But that's not what the car makers want.

Dave Jones: Cause you know, the, the, the average car, does the average car have anything? Yeah. I think vision, like vision type chips these days, you know, entertain, oh, vision chips. Probably infotainment, our favorite,

Chris Gammell: one of our favorite words here on the amp hour. Yeah.

Dave Jones: Infotainment. But, but for most of the other stuff, that's just running and controlling the car and sensories and stuff like that.

Chris Gammell: They need to do load dump testing and they need to do like, you know, they need to buck things down. You know, they need to, you need to operate off, like you're operating off a 12 volt rail and it needs to be able to survive a load dump. It's like, okay, cool. Like, can you do that Intel?

Dave Jones: Doesn't the average car have like over a hundred micro control? There are a lot. And I'm not talking micro processors. I'm talking little micro controllers that don't need to be 16 nanometers. Yes, there are.

Chris Gammell: There are a lot. I, apparently that number is shrinking. They're starting to consolidate. Some cars are starting to like the Tesla's of the world are starting to consolidate more, pull stuff more back into the centralized, you know,

Dave Jones: one, one central processing.

Chris Gammell: Yeah. I'd say like wireless, what? No, no, no. Like one or two, like basically having stuff kind of more centralized. Right. So yeah, well, I'm sure people will get mad.

Dave Jones: So just sensors transmitting data directly back rather than that sort of thing.

Chris Gammell: So I,

Dave Jones: we don't know anything about cars here. Local process. Other than, of course,

Chris Gammell: there's a lot of other stuff in a car than the vision type chips that Intel wants to make. So it's like, I don't know. This is where I personally, I just always kind of like bristle at the idea of like, Oh, Intel is going to do it. It's like, man, Intel couldn't even make like a, you know, little shitty makery chip for more than six months before someone killed it. It's like, they are not, they, this is not the business they want. Are you still like high margin? Like, okay, maybe they're, you know, they're moving into the, the foundry business. Maybe they'll figure it out. But like, if you're, if you've moved up through the ranks and you've, you know, built your career based on like these, these business, business ideas that have built Intel to what it is, and it's done very well. Then like when someone comes and says, Oh, we want to pay 10 cents for a, you know, a regulator module. Yeah. No, you know, like, no, we don't want to do that.

Dave Jones: But I, I don't know why they can't be in every market, right? Why they can't cater, you know, like even, even spin off, you know, part or, I think it probably comes down to like,

Chris Gammell: again, this is conjecture. I don't, I don't know the economics of fabs. Or anything like that. But like, I would imagine that like the, the, there's probably some baseline of like how, how many wafers and like, basically like the cost per square foot in a, in an Intel fab, because they're just, they are leading edge. Right. And, and I think, I think actually, you know, to their credit in this article, I think they talk about, they're not going to move old equipment back in. Like that makes sense.

Dave Jones: Right.

Chris Gammell: Okay. Something's got to be done there. So it's like, do you go and invest? This is kind of, I think we talked about this a couple of shows ago.

Dave Jones: Why did Intel have to do it?

Chris Gammell: Because it's expensive, you know, and like, and like they're a big name and like, there's, there is expert. There's a lot of expertise there. Like, so like, yeah, there are other people that could do it, but it's like, are they going to be able to spin up fast enough? Are they going to be able to like lobby? Right. Right. Governmental type, you know, this is a lot of government stuff too. And it's just like, Oh yeah,

Dave Jones: of course.

Chris Gammell: So it's, it's a, it is an interesting time. Yeah. It's, I don't know.

Dave Jones: Well, how about the chip fab at the national security agency? Why don't they, they start churning out voltage rings. Yeah, maybe. Yeah. I don't know. I think for those who don't know, they have their own in-house chip fab because they don't want the IP to go out anywhere else. They don't want it to leave the wall. So, yep. I, I don't know the details of it though. I don't know. I'm sure it's out there. Maybe. Dave,

Chris Gammell: Dave has been removed from this podcast. Uh, yeah. Oh boy. Yeah. Uh, you know, it's like, it's just, it's big stuff, right? This is all big money, big business, like government involvement. Like they're like, and even without government involvement, just like automaker stuff, like automakers are just monstrous. Like you just, you know, like at the scale they're talking about, it's really tough. And so I think, you know, one of my predictions is, okay, Intel may be getting into the foundry business, but I think there is going to be more, more market for a global foundries, or even like the skywaters of the world to expand more. I think they will see more business, but actually not from traditional chip companies because they're all buying each other. Anyways. I think, I think car companies are going to be like, well, I guess we got to make our own chips. We got to like, you know, I think there, you know, there is that know-how out there. Apple's going vertical, you know, like all, all these other companies are going, not all these other companies, Tesla's vertical, right? They have their crazy ass like M one or.

Dave Jones: Yeah. But Apple don't have their own fabs, do they? No,

Chris Gammell: they don't. That's, I don't think that's going to change, but I, I think, you know, Intel is moving in the foundry business. I think other foundries that are out there, I think they will continue to, you know, do well, grow, grow at a higher rate than they might have otherwise. But I think it'll be because some of the chip knowledge will move in house. What I'd love to see is if there was that, you know, if it drove more adoption of open tools, I don't actually think that's going to be as much of a large scale thing, but that would be super cool if it was, if like, you know, if a Ford is like, well, we got to design a chip module and do we want to pay mentor graphics for like a full license or whatever? Yeah. If, you know, they probably will, right. That's, you know, it's a drop in the bucket compared to otherwise.

Dave Jones: You're getting down into the details. But like,

Chris Gammell: but what I'm saying is I would love if like a Ford was like, well, if we have to build a new chip company or a new chip division, why don't we do it with these open tools that are out there? I don't think that's going to happen because the people they hire are probably also used to using the mentor graphics of the world. Correct. Yeah.

Dave Jones: Yeah.

Chris Gammell: But, you know, a guy can dream, Dave.

Dave Jones: That's right. I think it's just, this is an open, you know,

Chris Gammell: this is just big, big money, big tech stuff that's way above our pay grade.

Dave Jones: Indeed. How many, like how many companies in the U S we'll take the U S as an example, can actually produce their own, you know, the bread and butter chips,

Chris Gammell: like in their own fabs.

Dave Jones: Yes. Fabs in the U S is what I'm talking about.

Chris Gammell: I think microchip can still do it. I think TI still has some limited, but I think they're, you know, they, they're, they got burned because they moved a lot of stuff out to TSMC as well. Who else is out there? I mean, AMD is at global foundry. So like, but they're not like, they're like, they're all over. I mean, like all these companies are like multinational anyways. So no,

Dave Jones: yeah, exactly. It makes it hard.

Chris Gammell: The chip, the places that I know that chips are made in the U S is Intel's up in Portland. Microchip's down in Arizona. Intel's also down in Arizona. TI, I think still runs a fab in Richardson, Texas. There's a bunch of stuff in. Here we go. There is a wiki, of course,

Dave Jones: list of semiconductor fabrication plants, and you can search. Yeah. You can sort by country. Oh, cool. This is good.

Chris Gammell: That's great. Can you make that to me, please?

Dave Jones: There's two in Australia.

Chris Gammell: Yes, I will. Interesting. Are they, I bet they're at university. That's the thing too. Like it's also about scale, right? So like having, having the ability to make one wafer is a lot different than being able to like, Oh yeah. No, crank wafers out. Right. I mean, it's just, it is.

Dave Jones: But at this point, the automotive manufacturers, for example, they would probably pay anything. If they could just get the chips, they would happily, you know, pay 10 times the price.

Chris Gammell: Oh, maybe. Yeah. I don't know. Well,

Dave Jones: if it may, if it means the difference between shipping a car and not shipping a car and, and actually making a sale, I think they would. I don't know.

Chris Gammell: Yeah. Germany's got a lot. Germany's got a few. Oh, Japan's got a lot. Those recalls are pretty expensive there, Dave, as you may recall.

Speaker ?: Yeah.

Dave Jones: Interestingly, Japan's probably got as many as China or more. Actually, Japan probably have the most fabs.

Chris Gammell: Oh, interesting.

Dave Jones: Of any country.

Chris Gammell: Yeah. I mean, they still have a lot of like, they were still very big into Rome. That's interesting. I was, I was thinking, I was thinking like memory and drivers and stuff like that. So yeah. Sharp, Fujitsu. Yeah.

Dave Jones: UK has got quite a few. US actually have a lot. Wow. Actually, I could as, almost as many as China.

Chris Gammell: Oh, interesting.

Dave Jones: Wow. Infineon have two, four, six fabs in the USA. I assume that these are all still valid.

Chris Gammell: Yeah.

Dave Jones: Yeah. I'm sorting by country. Interesting. Looking at the US. USA. Intel has five fabs in the US.

Chris Gammell: Yep.

Dave Jones: All in Chandler, Arizona.

Chris Gammell: That's not true. They have some up in Hillsborough as well. Okay.

Dave Jones: Are you saying Wikipedia is not?

Chris Gammell: No, no, it's further down. It says right there. Oh, is it? D1C, D1B, D1D.

Dave Jones: Oh, sorry. Yeah, yeah, yeah. No. Okay. So it didn't sort by. Right.

Chris Gammell: Yeah. So like Malta, New York, Ithaca, New York, Albany, New York. That's all. I kind of just put that all in like the East fish kill. Like that's like global foundries. IBM was there, you know, like global foundries spun off from AB, AMD, I believe. Uh, Cree, Cree is in Durham. Uh, so that's, that's right near me. Yeah. That's, I mean, yeah. Skywater is the one we've talked about before that are up in Minnesota. That's a North by DigiKey. Yeah. And so, I mean, like, I mean, I think it's really, do they have capacity on here too or no? No, they have processing nodes. So on the right side, on the far right, it's saying that Apple,

Dave Jones: Apple have a, actually do have a fab, but I don't know what its capabilities are.

Chris Gammell: Do they say it's in Texas?

Dave Jones: Yeah. No, San Jose. Oh, interesting. Apple formerly Maxim. Ah, okay. And then formerly Samsung. So it's an old. Got it. Yeah. They did a lot of. Yeah. It only goes down to 90 nanometers.

Chris Gammell: So after, after I was at Samsung, which was done in Austin, they had moved up into the LSI business. So like large system integration. I don't know what LSI stands for, but like they did a bunch of Apple stuff there basically. Right. But Apple, yeah. I mean, like Apple is a, I think, I think a lot of people are going to emulate Apple, especially given like, okay. So like you're looking at the marketplace right now. You're like, well, we can't get any chips for five years. What are we going to do? Again? And I hope it turns into.

Dave Jones: Yeah. But Apple, Apple aren't making them though. That's the thing.

Chris Gammell: I think you're right about that. Yeah. But I think.

Dave Jones: Yeah. Yeah. Apple are, yeah. They're designing their own, but they're not making them. So it's like, you know, it's not necessarily a thing I'd follow.

Chris Gammell: Yeah. But if you're like sitting around, if you're waiting around versus like being able to do something and like, you know, source it from a foundry, you know, if you can, if you can get time at a foundry, I guess, I don't know. I think this is a good example of like maintaining infrastructure. I, you know, I was not a fan as people remember when, when analog devices talked about outsourcing a lot of their stuff to TSMC. They even said the process technology, we're just going to do what they do. And it's like, that's just, I, you know, it's a very, I think it's a very corporate. I'm not going to say short sighted kind of thing. You know, it's just like, yeah, it looks better operationally. Yeah. Your operational budgets are lower. It's like, yeah. Expensive as hell. To do that. But that, that is, that is your moat. That is like how you are maintaining your dominance. If you have it. And I don't know.

Dave Jones: Yep. It's yep. No, that was, yeah, it's dumb.

Chris Gammell: I'm not going to be a semiconductor executive anytime soon. So that part's not a big deal.

Dave Jones: Exactly. Maybe we should get one on the show.

Chris Gammell: Sure.

Dave Jones: Maybe we should attempt to get a, a fab CEO. A fab CEO.

Chris Gammell: Yeah. I don't, I think that would be a boring.

Dave Jones: Yeah. Well, no, if they're an engineer, we talk about it.

Speaker ?: Got it.

Dave Jones: Got it. And we can do a, you know, so we mainly talk about the engineering and then we have a side discussion on the politics of it all. Oh, sure.

Chris Gammell: I'm sure they'd be, yeah. CEOs are known for being, showing their hand.

Dave Jones: Oh yeah. Right. Exactly. Oh boy. Fascinating. Anyway. Yeah. We'll link that.

Chris Gammell: We'll definitely link this list. And this is a good find. So we'll link this in. It's great.

Dave Jones: Right. So, so the easy solution is to just simply make your own chips. Yeah. On your desk. That's just with your chip. Yeah. Great killer. That's coming this year because Chris predicted it like 10, 11 years ago or whatever.

Chris Gammell: That's great.

Dave Jones: Yeah.

Chris Gammell: I'm just saying it would be really useful by now. I'm just saying, Dave. Yeah. Yeah. It would, wouldn't it? Yeah. Yeah.

Dave Jones: So with those Mars bases and you know, exactly. That's what I'm saying. Right. Oh boy.

Chris Gammell: This is another link I put on here. Demos over deadlines. This is something my friend's company does. And I really liked this idea. And so basically it, you know, like this whole article is talking about like, Oh, software is so hard to, to predict in terms of timelines and stuff like that. And there's obviously some things you can't get out of in terms of, uh, if you've promised it by a certain date, but I really liked the idea of the demo as like a measurable piece. You know, this is tied a lot into like, what's it called? Uh, lean, not lean. What am I saying here? A sprint. It's a sprint is part of not extreme programming. Anyways, my brain's not working.

Dave Jones: What are we talking about? The demo for, we're talking about demo for investors. Are we talking about a demo for customers? No,

Chris Gammell: for one another. So like, instead of like, instead of like your boss being like, you must have this on my desk three weeks from the day.

Dave Jones: Yes.

Chris Gammell: They say, show me a demo as soon as you can of every piece that you think is relevant. And so, you know, it might be in software. It's a little easier.

Dave Jones: Well, that's called a Gantt chart. It's called it, you know, you've got, that's called a milestone on a Gantt chart. Show us this thing where you're, okay. It's, it's week 20. You're supposed to be at this point. Show us this thing working. Sure.

Chris Gammell: Could be. Yeah.

Dave Jones: You know, it's like, that's,

Chris Gammell: I feel like there's always the focus on the engineering. Yeah. I mean, but I feel like the demo is like a very, it depends how big it is. Right. So like for hardware, if it's like the board is finished and working, that's a lot different than like, Oh, let me show you the data from ADC and how linear it is, which is like the thing I did this week. And that's kind of like the difference is like, you kind of break it down into smaller parts. And I, I've been at companies that have done that sort of thing, like the, on the hardware side, but I think it's really hard in hardware to do that sort of thing. Because if it's just one little piece, like if it's just the LED lighting up, like that doesn't really feel like a lot of accomplishment. And yet if you make that culturally part of a thing, you're like, well, actually, it's the tools work. Yeah, exactly. The tool chain.

Dave Jones: Yeah. The tool chat on the back for everyone. Yeah. Yeah. Yeah. Tool chain works. Cake in the lounge. Cake in the lounge. Yeah.

Chris Gammell: I just really like this as a, it's like a kind of framework for, for building out teams. And like, I don't know, I'm sure you've, you've done this sort of thing in your past work as well though. Right. Of course. Yeah.

Dave Jones: Oh, I think I, I think I want to know is, are Gantt charts still a thing?

Chris Gammell: I think so. Yeah. I mean,

Dave Jones: leave it in the comments down below. I, you know, cause they were all the rage in the, you know, nineties.

Chris Gammell: Yeah. Are they still a thing? Agile. That is what I was trying to think of. I think agile is the, you know, that's like the new thing, but I think that there's,

Dave Jones: hang on. I'm going to, I'm going to ask this on Twitter right now. I'm going to set up a poll. Oh yeah.

Chris Gammell: This is fun. This is good. Good radio, Dave.

Dave Jones: And now I've got to figure out how to spell Gantt.

Chris Gammell: Oh yeah. Is it two N's or two T's? That gets me every time. Yeah. I know there's two or something.

Dave Jones: Yeah. Yeah. I know. Yep. Same here. G A N double T. Yes. I thought it was double T. Okay. Yeah.

Chris Gammell: I think it is. I think there are some things that are still pretty linear. Right. I mean, I think that's, that's what it really comes down to, but when you start to like try and compress stuff, you need to show something. Right. So like, again, if you're building out like a, a full, my, my past experience was like, there was a six month development cycle. You're to get to Reve. It was six months on one project that I was on. And it's like, that's really tough because if you get to the six month mark and you're like, Oh shit, this doesn't work. You know, that's like, it's like teammates might not know what you're doing. You know, like, yeah, you're going to have like status updates, but like showing it is really important. I think it's getting easier to do that. But like, again, from a cultural perspective, it's tough because I don't know, like everybody working remote too. Like, how do you actually show, you know, it's tough to show like hardware in situ as well. Like, Oh, the led lit up. And well, I could fake that. I specifically know how to fake that.

Dave Jones: It's the old smoke and mirrors demo. It's like, never give. And well, okay. There's a positive and a negative to giving an engineer a demo deadline. Right. One is that they're going to fake it with a smoke and mirrors. You could. Right. Just to shut you up because they know, don't, you know, because the engineer knows where they're at, but they've got this pain in the ass deadline for this stupid demo. And that's not how they want to do things in that order. I realized,

Chris Gammell: I realized you didn't read the article and that's okay. But the idea is that there isn't a deadline, that there isn't a deadline. You just show it when you have a thing.

Dave Jones: Yeah. But I'm, yeah, but I'm talking about, you do get, I've had these deadlines, like, like show us a demo. We want to demo next Friday.

Chris Gammell: Yeah. That's, that's rough. That's basically like a mini delivery, right? That's, that's, yeah.

Dave Jones: It's, it's a mini delivery. That will lead to smoke and mirrors. That is the wank word, bingo buzzword, deliverable. Yes. Right.

Chris Gammell: Deliverable.

Dave Jones: Not to a customer. It's to me, the manager, deliverable, please. I snap my fingers, servant.

Chris Gammell: And,

Dave Jones: and I will not be above saying like, yeah, I've done smoke and mirrors demos just to shut them up. Yeah. Because I know where I am in the project and you're, you guys are just wasting my time by making me meet this demo deadline. So you're going to get a fricking smoke and mirrors. Sure. Sure. Right. And yeah, no, screw that. Would you, whereas other times it's good.

Chris Gammell: In that same like a scenario though, like the same, maybe even the same project, would you offer information otherwise? Or was it just like culturally you were like, Oh, well I'm not going to show anything. Cause it'll just mean that more is asked of me.

Dave Jones: If they come around and they want to see what I'm doing, that's fine. I'll, I'll, you know, I'll, I'll tell them where I'm up to and stuff like that. But you know, don't judge me. It's like, no, go away. I'm, I'm the one who's organizing how to do this entire project. It's just me. Right. And you just, you know, come in and want some, you know, you're trying to push some schedule that, you know, is just, I, I know that's the wrong way. Cause you're not the one doing it. I'm the one doing the project. So that's why I always did. If possible, I always did my own Gantt charts. Right. I would not let, I would not let them push the Gantt charts on me. Interesting.

Chris Gammell: Yeah. Turn it around.

Dave Jones: Because I'm the one who has to do the work. Right. I'm, I'm the one with the experience and the one who has to do the job.

Chris Gammell: I am the one who,

Dave Jones: I'm not saying that they're not experienced managers. Right. They might be, but they don't know the nitty gritty detail. Whereas almost off the bat, like, yeah, if you're an ex a really experienced engineer, right, right off the bat, you're going to know, okay, this part's going to take me X amount of time. This part will take me X amount, you know, like, Oh, this part's a real hurdle. You know, I've got to allow, you know, a big amount of time for that and, you know, stuff like that. So, whereas the manager doesn't necessarily know that the engine, poor engine, schmuck and engineer tasked with the task with the thing, task with the task. Yeah. Yeah. Getting the job done is the one who should be set in the Gantt chart, in my opinion.

Chris Gammell: Yeah. I think that's right. I mean, I think, I think you basically just restated like some of the agile stuff. Like, you know, it's like, Oh, okay. That's like a big thing.

Dave Jones: And you should have the skills. Yes. Young engineers out there. Like seriously, if you did not get a, a class, you know, if you did not get, you know, a class on, you know, project management, not from, you know, like, no, it's, I think one of the things is when you need to know the skills of creating Gantt charts as well. But you also need the wisdom. You need the knowledge behind it to be able to put reasonable estimates on there.

Chris Gammell: Right. Right. I think so. Dave, if you, if you update your vocabulary slightly, you could move into a whole new paradigm of agile. It's not a Gantt chart. It's a burndown chart.

Dave Jones: Oh, a burndown. I've never heard of a burndown chart. If you look it up on many agile things. A burndown.

Chris Gammell: But actually, that is one of the things that I like, the, that, that I do like. You know, so I read about agile and I've been part of teams that have tried it.

Dave Jones: Oh, so you want to get down to zero because you've got your, you've got your deliverable on that day. And oh, so you're either, let me guess you're either above or below the trend line.

Chris Gammell: Uh, I think there's really no judgment on the trend line. It's more of just like, so like, you want to bet. Yeah.

Dave Jones: If you can be judged, you will be judged. Yeah. That's true.

Chris Gammell: I think, you know, so like in a best case scenario, like when I've read about this stuff, when I've heard friends talking about it, you know, like in a best case scenario, people are giving like real estimates on like how long it takes to do things. And like, there's a whole bunch of BS out there. There's scrum masters and all the other things that are out there. And this is, this is many, we are many, many, many years behind. You know, this is 20 plus years old.

Dave Jones: I know. Yeah. We're, yeah, no, we're totally, like both of us have been out of large corporate business for a long time. Right. But yeah,

Chris Gammell: but startups love agile as well. So like, and like I said, there are some things that I really like about it. Like the things that I think one of the most powerful things of it is like, if you can't like, if you can't estimate something in a good way, if you say it's going to take two weeks and it's like the actual work's going to take two weeks, it's like, well, no, you can't, you're not allowed to do that. You got to break it down more. Like that is as a general rule. If you were just using that as a guiding principle of like, well, you got to break it down more. It, you know, this, you know, you need to have it in like half day chunks at a max. It's like that, that alone is a guiding decision, especially in hardware. Assuming you're not like putting, you know, putting in the timing of like waiting for a board to come back or waiting for a board to be assembled or whatever. Breaking it down more actually makes a lot of sense. I think, you know, like, oh, I got to check it. Linearity of this ADC. I got to, you know, implement a new four to 20 controller. I got to do all this other crap. And it's like, yeah, okay. So then you start to like add it all up and you're like, and that's effectively what a Gantt chart is too. Right. I mean, like you're just breaking it down more and more and more. If you're just saying.

Dave Jones: Yeah. But again, again, charts also show dependencies as well. Like you, you cannot proceed beyond this point unless you've got X, X, X line. Yeah. Yeah. Meet up. Right. You know? So yeah. Somebody, we've, we've got feedback from, from Twitter. Already. Oh, live. That I put up. Shura Lubresky.

Chris Gammell: Boy, Dave does great with those Eastern European names.

Dave Jones: Oh, I agree. Yeah. Yeah. Again, Gantt is bullshit. Use pert charts. Oh, we've got a pert fan boys. Jonathan man says, sadly. Yes. I thought full on Gantt charts full of every subtask are pretty unwieldy. He much prefers the roadmap really is a high, which is a high level. We're into methodologies and frameworks now. And sprints. Oh, sprints. There's another wank word. Bingo. That's the agile thing. Planning day to day task. Okay. Yep. Yep. So anyway, 66% of people have said, yes, Gantt charts are still a thing.

Chris Gammell: Yeah. I think, I think hardware is, I think it's hard to remap a linear thing like hardware where it's very defined. It's like, you, you can try and work ahead and whatever, but like until your boards and your desk, you know, there is no effect. Like, you know, there's no, there's no way to do a three mil space trace, like high quality circuit board in your lab. I'm like, I know they keep trying it. And it's like, it's not, it's just like, send that out. And it's going to be great.

Dave Jones: I've done six, six. Yes. Right. Yes. And like, you know, there's a bit, how you do it?

Chris Gammell: And like the dragonfly MX or whatever that thing, not MX, but whatever the dragonfly, like people keep talking about that one. It's like, yes, there's some very high end tools that will have some result that might approximate it, but probably not 12 layers or whatever else you need, you know? Like, so it's just like, okay, so there are some like true just bits versus Adams kind of dependencies. And you can't get around those. And that's when I think that's when a waterfall chart does start to make sense. I all know.

Dave Jones: I think it's just, yeah, it's,

Chris Gammell: it's tough to, you know, make this stuff work, right? I mean, there's whole, Jack cancel made a career on like managing software projects at this or firmware projects. And like, you know, like the same thing for hardware projects and software. It's just like, there's so much stuff out there.

Dave Jones: I think just all you've got to do, give me a date, leave me the fuck alone. What? If I'm allowed to drop the F bomb on the iPad. Not usually. Not usually, but we're going to make the exception. Yeah. That's probably the first time you've heard me drop that, you know,

Chris Gammell: no, not the first time I've heard you. That's for sure.

Dave Jones: Anyway. Yeah. Yep.

Chris Gammell: Yeah.

Dave Jones: Give me a deadline. Leave me alone.

Chris Gammell: I think it's tough when it's like bigger teams, right? Any, anything gets hard as you have. No, no, no,

Dave Jones: of course, you know, it did. Of course, it totally depends on the project, but you know, I'm, I'm, I'm so used to being the only person on the team, you know, it's like, yeah, it's, yeah, I've worked in larger groups, but it's, I've done a ton of stuff. It's just been me. Does not play well with others. Does not play well with others. You know, no, no, that's, that's why they put me down in the dungeon. That's right.

Chris Gammell: Yeah. You move my desk again.

Dave Jones: My stapler. My stapler. It was my red stapler. I rewatched it recently.

Chris Gammell: It holds up.

Dave Jones: Oh yeah. I haven't. I'm going to have to do it. Yeah. Yeah. Of course it does. Of course it does. It's classic. Yes. The, yes. Those who don't know, tell them.

Chris Gammell: Office space. Yeah. Office space. Mike judge. What a genius. What a genius. Yep.

Dave Jones: Fantastic. See the one who also did. Oh God. The name's escaping me. Where he goes into the future and everyone's dumb. And he's the smartest guy in the world.

Chris Gammell: Yep.

Dave Jones: Idiocracy. Idiocracy. Is that it? Yes. Yeah. He did Silicon Valley.

Chris Gammell: I mean like, yeah, man, he's, he's the prognosticator of prognosticators. Right. Yeah.

Dave Jones: So he's the Scott Adams of movies, is he? He's like the Dillbert, the Scott Adams of. Yeah.

Chris Gammell: Better, better than that. But yeah. Yeah. Right. Yeah. Man. Yeah.

Dave Jones: Well, it didn't, didn't Elon Musk famously said, if it's in a Dillbert cartoon, it's banned here at SpaceX. All right. If it, yes. Yeah. Yeah. That's what he said. Yes. If, if that thing, if, if, if that is a thing in a Dillbert cartoon, we are not doing that here.

Chris Gammell: Must be tough. Not being able to drink coffee and at SpaceX or Tesla. Can you imagine that? If you just like, if you like band, so like you have a thousand engineers on staff and then on just any random Tuesday, coffee machine breaks, coffee is verboten. You are not allowed to drink it ever again on campus. How many people quit that day? Give me a percentage.

Dave Jones: Well, I can only speak for me. You'd be fine. Yeah. Yeah. I get it. Yeah.

Chris Gammell: I, I, I'd, I'd, I'd quit that day. Right. 85, 90%, maybe. Yeah.

Dave Jones: Yeah. Yeah. Right. Oh boy. I don't know. No, they'd just be, they'd just be down on them outside. Like a smokers. Yeah. Right. Right. Right.

Chris Gammell: They'd be like,

Dave Jones: yeah. Huddle them together in the alleyway outside the building.

Chris Gammell: Right. Everybody's going off campus just to, just to get their fix.

Dave Jones: Down on those durries. Yeah.

Chris Gammell: Yeah.

Dave Jones: Oh boy. Yeah.

Chris Gammell: Well, if any of you listening have a boss that tells you you can't drink coffee, let me tell you right now, they're not allowed to do that. You're going to be fine.

Dave Jones: And if they ask for a demo, just tell them to bugger off.

Chris Gammell: Yep. Yep.

Dave Jones: Yep. Just waste my time. Go away. Yep. Speaking of which, we should go away because our amp power is up.

Chris Gammell: Our amp power is up. All right. Well,

Dave Jones: that's it.

Chris Gammell: Hope you don't have to do any demos, even though you do them voluntarily every, every week on your, YouTube channel. Yeah. Demo machine.

Dave Jones: Catch you next time.

Chris Gammell: See ya. Bye.