#88 – Yonderly Yodeling Yobbos

Chris Gammell: This is the Amp Hour Podcast, recorded March 25th, 2012. Episode 88, Yonderly Yodeling Yabos.

Dave Jones: Welcome to the Amp Hour. I'm Dave Jones from the EEV blog. And I'm Chris Gammell of Chris Gammell's Analog Life. What's happening in Cleveland, Chris?

Chris Gammell: I am a cave troll once again, back in the basement. Nice. Yeah, I have descended. Why? Do tell. Oh, it's warming up, you know, might as well ease the burden on my wife and, you know, I'll emerge in six months. Sorry. She's happy. She sounds happy about it.

Dave Jones: Yeah, exactly. Boot you out, get rid of you.

Chris Gammell: Yeah, and it's good, too. You know, I'm back at the bench and it's easy to lose focus when you're away geographically, right? Yeah. I mean, I'm sure even when you're at home and you probably walk out to your garage, you're like, oh, I used to have a lab out here.

Dave Jones: Yeah, I still do that. I go, yeah, I wander out there and go, oh, shit, no, my lab's actually 15 minutes away. Yeah.

Chris Gammell: I know.

Dave Jones: Yeah, so that's why you've got to have some stuff on your work, you know, on your desk. You know, you've got to have a multimeter. You've got to have a, you know, a ruler for measuring stuff. You've got to have, you know, things for just doing basic stuff and storming ideas and stuff like that. So you can't actually work when you're displaced from your lab.

Chris Gammell: The worst thing is to be like, oh, I can't do this today, right? Yeah. I can't think of, I can't, you know, chunk on a problem, right? Yeah. That's pretty, yeah, that sucks.

Dave Jones: The problem I've got is when I'm working on a project, right, because I'll work at both the lab and at home here in my study, and I've got to bring my prototype back and forth. I've got to constantly bring it back and forth. And, you know, and sometimes I'll leave it in a place, you know, I'll go to the lab and then I'll go, oh, no. Oh, I left the damn thing at the, oh. Yeah. I left the damn thing at home. It's really annoying.

Chris Gammell: Yeah, that sucks. That happens with my laptop, actually. Oh, right. That is something I've, I've just started leaving my laptop at work because, you know, I've got a longer drive than I used to now. Yeah. Since I changed jobs. And, man.

Dave Jones: How far, by the way?

Chris Gammell: It's about half hour. It's a good.

Dave Jones: Oh, right. Yeah, that's. It's a good half a podcast. That's probably the limit. But, you know, half hour was my irritation limit to actually driving to work, you know. So.

Chris Gammell: Yeah.

Dave Jones: Yeah. So, when I was searching for a job, it would always be, you know, oh, is it more than half an hour each way, you know. Right. And if it is, well, you know, I don't care how good the job is, piss off, you know.

Chris Gammell: Yeah. Exactly. No, exactly. I turned down a job exactly for that reason. Although. Because I knew the traffic pattern, too. Yeah. I knew that not only is it a half hour normally, on like two out of five days, it's, you always hear on the radio just, oh.

Dave Jones: Oh, that's horrible. Yeah.

Chris Gammell: You know, horrible traffic on this road, yada, yada, yada. So.

Dave Jones: Although I did not take my own advice there. When I took the job at Altium, I.

Chris Gammell: I thought you said you had like an hour drive, right?

Dave Jones: That was. Something like that. Yeah. But the good thing is you could make your own hours. So, you go like, you know, two hours after peak hour or something. And it was fairly easy. So, you could save. Save significant time. But, yeah, I just wanted to get out of my previous job. So, I just took the Altium one.

Chris Gammell: Yep. Yep. That'll do it.

Dave Jones: Yep.

Chris Gammell: Yeah, but if you forget your computer, I mean, like, I don't know. Yeah. I don't know. It probably wasn't always like this, you know. But even just reading data sheets now, you know, I can't just like be like, oh, can I borrow your computer, neighbor? You know, I can't walk a cube over, you know. So, if I forget my laptop, that's an hour round trip mistake right there. Yep. And, yeah, that's the thing. So, I've just started keeping that thing at work. And I got enough computers at home. But I'm sure the same thing goes with prototypes, right? Yep. You forget a prototype. It's a 15-minute walk both ways, right?

Dave Jones: Yeah, exactly. Exactly. So, yeah.

Chris Gammell: And that's enough to kill motivation for a day, too. It's like, ah, I'll just... Oh, I know. Yeah, it's annoying. I'll do it tomorrow.

Dave Jones: Well, I just, I forgot yesterday. I went to the lab and I forgot my card, my swipe card. So, I couldn't get in. So, here I am at the gate and I'm patting myself going, where's my wallet? Where's my wallet? Where's my swipe card? Oh, man. You know, I had to drive back home and get the bloody... Oh, it's just...

Chris Gammell: Yeah, it kills the motivation. It does. Yeah, it definitely does. So, it's nice having... You know, being back down at the bench is good, you know?

Dave Jones: How do other people handle that? How do our listeners handle that? Do you just go, right, well, I don't do anything at home. I do everything at the lab or whatever. Or I do everything at home and I, you know, I'm even my labs in my, you know, study at my home. So, everything's there and, you know, just the act of, you know, working... Because if I'm working on projects, we've talked about this before, you know, do you use the GitHub approach to store all your project files on? How do you do that? You know, I bring them back on memory stick, you know, sneak in it, as they call it. And, you know, and I go, oh, have I overwrite? Did I, you know, was my last updated version on this machine or the one at the lab? And, oh, you know.

Chris Gammell: Yeah.

Dave Jones: That's tough.

Chris Gammell: That is tough. But, you know, you can't put hardware up on a network either. No, exactly. Until they create that materializer and dematerializer, whatever you want to call it. Right. Yeah. Until the transporter's around, really. Beam me up. You're right. Right. Oh, boy.

Dave Jones: It's too tough. What have we got on today's show? Lots of electronic stuff, I hope.

Chris Gammell: I think so. We'll get some shout-outs. Yeah, they do. They keep asking about that. What's up with that?

Dave Jones: Well, we had a shout-out. We wanted to thank one of our sponsors, one of our donors. Yes. Maurice. Thank you very much, Maurice. He's one of our more generous benefactors. And we really appreciate it.

Chris Gammell: We do. A platinum sponsor. A platinum sponsor. And what does he win?

Dave Jones: A free shout-out on the Amp. Yeah. Yay. Yeah.

Chris Gammell: So, if other people are interested in donating, we do have a page on our website. But if not, we appreciate listeners all the same. So, thanks again, Maurice. Thank you very much. We really do appreciate that.

Dave Jones: We don't use the money. We only use the money for the server and stuff like that. So, if it goes over, it just sits there and accumulates. And we'll use it for... So, we don't siphon it off and go and spend it on beer or something like that.

Chris Gammell: Right. Especially because you don't drink beer. So, that helps. Exactly.

Dave Jones: That helps a lot. Yeah. And you brew your own.

Chris Gammell: I don't. No. I'm not one of them.

Dave Jones: I thought you were a home brewer.

Chris Gammell: Nope. Nope. All I home brew is electronics, Dave.

Dave Jones: Right. Excellent. Good answer. Yes.

Chris Gammell: Yes.

Dave Jones: All right.

Chris Gammell: Another fun thing on this shout-out list we got is... This is a bit of fun. This is a meeting ticker. I don't remember where I saw this. But if you got like Dilbert-like meetings, you can go over to this and you can put in... It's just a real simple script on a website. You put in the number of attendees. You put in how much you think everybody makes on average. So, if it's like 100 bucks an hour. I know that a lot of engineering firms bill anywhere from 100 to 150 bucks an hour in terms of like when they're keeping track of accounting. Oh, minimum. Yep. Yeah. Right. And then you say when the meeting starts. Yep. And you hit go. So, you got 150 an hour. You got two people in this meeting with me and Dave. And then... Where did it go? We hit go. And it says... Stop the bleeding. The meeting has cost. And it just goes up like a gas station, basically. And there's a big button. Stop the bleeding. Yep. That's great. So, if you've got a meeting that usually runs long, this could be a really good useful tool for you. Put it up on.

Dave Jones: We're just hemorrhaging cash. That's my favorite term at a company. We're hemorrhaging cash. Yep. Just spending it like a drunken sailor. Oh, boy. It's good stuff.

Speaker ?: Yeah.

Dave Jones: That's a lot of fun. Yep. I know. Somebody actually did that as a physical project. Like, they had like... Wasn't it available for sale somewhere? Perhaps on one of those gadget websites you could buy like a meeting.

Chris Gammell: Like ThinkGeek or something like that?

Dave Jones: Yeah. I don't know. Something. But yeah, you could buy it. You know.

Chris Gammell: I'll have to look for that.

Dave Jones: Yeah. So, it had to... Yeah. The meeting time. The meeting cost clock or something. Yeah. Like that. How much is costing you? And yeah. And unfortunately, managers don't get the joke. You know. They just... Yeah. Because many times it has been brought up in meetings I've been in. It's like, you know how much this shit's costing us?

Chris Gammell: Right. Especially when you have like 10 or 20 people sitting in a meeting. It's just... Yeah. It's... You know. And that's an interesting point just in terms of engineering is that the biggest cost of a project is almost always people. You know. Oh, yeah. Yeah. Yeah. And manufacturing and everything else. Yep. It's not that... It's not that large of a cost really. It's almost... Well, materials are usually a big cost, right? But the... It depends on what you're working on. Yeah. Right, right. But just you can't ignore the cost of the people. And if you look at a budget, that often is the biggest thing. More so, especially if you compare it to like, you know, the cost of buying scopes or, you know, outfitting a lab and stuff like that. The people. It keeps coming up. So, yeah.

Dave Jones: And that's what happens. Like, you know, sometimes you... If you've ever worked at that company where you have to walk around and get five signatures to buy another $100 multimeter just to... Yeah. You can spend all day and then they um and ah and, you know, rub their chin. You know, I don't know. You know, I don't know. I forgot the budget for it. It's like, oh, dude, I've already wasted five hours.

Chris Gammell: Yep.

Dave Jones: You know, oh. Not to mention accounting's time because everything has to be in triplicate. And then they've got to raise a purchase order. And it was like, one company I've worked at, it was like, they actually, they worked out the cost of it. And it was actually a minimum of $200 or something in company cost to actually raise a single purchase order. Just the, you know, all the paperwork and everything else. Just even thinking about it. Yeah. Just the paper, no, this is just the paperwork involved in the triplicate and doing everything else and somebody's time to, just the accountant's time to handle all that was $200. So buying a $1 transistor on a, you know, it would, would cost you $200 because it was, you know, so who cares if Farnell charged $50 for a bloody resistor or something like that? It doesn't matter because of the fact that you had to raise a purchase order and do it as a, cause a lot of people may not realize that that's how it used to work with, you know, big companies like Farnell's DigiKey and stuff like that. A company would actually raise a purchase order in triplicate with carbon paper. Then they'd actually fax it. Then they'd have to do a fax and then they'd fax the purchase order to the company. Even though you could buy with, buy with credit card online at that stage, most companies didn't get into that until late, you know, especially at larger companies, they, um, they, they still had to operate the old fashioned way and everything had to be done with paper. There's none of this credit card rubbish doing it instantly. Yeah.

Chris Gammell: And that's 30 and all the other payment times and all that other junk, right?

Dave Jones: And they fax away the purchase order and then, you know, you'd eventually get your parts. Whereas it was much quicker if you just whipped out your credit card and ordered online, you'd have it next day.

Chris Gammell: But, you know, it's... Yeah. You've, you've talked about that in the past, why, where you've worked places that, that have that minimum. And I think that's a, a great policy for any managers, you know, like if, if you should just have some small limit, you know, like a hundred bucks, it should be no brainer. Just do it. Right. Because even, even going over to your boss and saying, Hey, can I buy this $50 thing? It's like, why are you bugging me right now? Yeah, exactly.

Dave Jones: You just wasted 50 bucks of my time, you know. Right. Exactly. It's not worth it. And that's why Altium gave everyone a credit card or they gave the hardware engineers a credit card because, you know, they, because we needed to do stuff like that. And it meant we could, you know, buy anything we wanted at any time for anything, which just saved the company a whole lot of money. It sounds like it's, you know, it's a, you know, it's a free ticket to splurge on everything. But, you know, most engineers are sensible and they buy, just buy the stuff they need and they don't have to hassle anyone.

Chris Gammell: Right. And there's always, you know, whenever you have to fill out expense reports or anything else like that. I mean, you could, I mean, yeah, that's the bad part, right? But really, you can only get so far, you can only buy so much stuff. And even if you spent $50,000 over time, you know, if you spent $50,000 and they said, what are you buying here? You're fired, right? Yeah. Over time, you're likely saving $50,000 by not going and asking someone and raising POs each time. Yeah. So, yeah, it's self-correcting. Yes. It's a control system.

Dave Jones: Even if you are buying those farting novelty toys and those remote control tanks for the office. You know? Right.

Chris Gammell: Exactly. Exactly.

Dave Jones: Oh, boy.

Chris Gammell: So, one last thing involving money here. Jack Gansel recently had a salary survey come out and we've had Jack on the show before and that was a lot of fun. But he had a salary survey and it's always interesting reading his stuff because a lot of his...

Dave Jones: Does he do the survey himself? Yeah.

Chris Gammell: He does it with just like an online survey tool. Oh, right. Okay. It's lower numbers, but it's interesting.

Dave Jones: So, it's based on his audience, is it?

Chris Gammell: Yeah. And that's why it's interesting because a lot of it's consultants and you see, you kind of start to see the upper ranges of what people make and it's sort of like, whoa, I could be making that someday. You know, it's embedded as well, right? Yeah. Yeah. You know, some people are charging 300 bucks an hour for their time and that's... Yeah. Ooh, damn. Yeah. That's serious. That's a lot of dough.

Dave Jones: And if you can get away with it, well, you know.

Chris Gammell: Right. Yeah. Why not? It's specialized knowledge, you know? Like if you're...

Dave Jones: And a company who doesn't care. They're used to paying that sort of stuff, you know? Yeah. And that's what happens. If you work at some big company and you know what they're willing to pay, then you quit and you go contract and then you charge back at these absorbent, you know, absorbent rates of $300 an hour or something. Right. And you know the company's just going to pay it because it's all, you know, it's just a drop in the ocean.

Chris Gammell: Yeah. So... And the interesting thing about that too is that a lot of times, you know, even if you're getting paid that much, they'll just lap it up because you're off the books basically then. Like there's always, there's this weird, you know, I always talk about the NBA playbook and it's, I think it's part of like lean everything, you know, everything's lean these days. Yeah. And I think that's part of it where you just offload all the risk and all the inventory in terms of people's time here, obviously. But for some reason, because, you know, this consultant isn't part of the organization, there's some reason that's okay versus hiring someone on. And because you can easily, you know, it's more liquid in terms of people's hours, you can easily let a consultant go, even though you probably won't, you know, like for some reason, that's agreeable to upper management. So yeah, if you got it, work it, man.

Dave Jones: I find it interesting. He's got a graph of presumably full-time salary. I'm not going to read all the text here, but versus industry experience. How many years have you been in the industry? And it starts out at just over 60,000 for zero to four years. And then it peaks at about 20 years. And after that, it actually goes down a bit. You know, so yeah, really, I've kind of found that, you know, once you hit that, probably the 10-year mark, actually, it's after that, it doesn't matter whether you've got 10 years experience or 40 years experience, you're going to earn the same, you know, so.

Chris Gammell: Roughly. I think part of that, though, is you see a lot of exodus because of people going to management. So what happens is the supply shrinks. You'd think that that would make salaries go up because there'd be fewer people. But in reality, they just start hiring younger people then. So that's what makes it flatten out eventually, I think. So, you know, I've written about that kind of stuff before, you know, like expectations of salaries and stuff like that. And I can link that article in if people are interested in it. But yeah, it's...

Dave Jones: Oh, he's got an Australian New Zealand one. Yeah, that's not far off the mark. I'd say it's probably on the low side, actually. They're like a zero to four years is under 60,000. In fact, in 2012, it's dropped to like 36,000, 37,000 or something. What? Wow. Well, this is a small sample size, too. It was up near 60 in 2006, 2009. Yeah, it's, you know, how many... I think that's what really plays into this. Yeah, it's Australia.

Chris Gammell: Right. Yeah, I think that really plays into it. But yeah, it's still an interesting thing to look at. And as we said, when Jack was on the show, Jack has a great newsletter for any, you know, especially for young players, you know, like this is a great... He has a great newsletter to read. I was suggested it by my mentor and, you know, a lot of great insider knowledge with it. So I highly suggest people sign up for it.

Dave Jones: Right.

Chris Gammell: Yep. I agree. All right. What else we got other than the shout outs?

Dave Jones: We've got some news. Where is this Sony power thing?

Chris Gammell: Oh, the thing that Marjorie was talking about? Yep. Yeah. So Marjorie Connor on EDN. This is kind of a remote rant. This is a...

Dave Jones: It's a news slash rant. Sony have come out with this. Well, you tell it.

Chris Gammell: So basically, Sony has this new idea. Oh my gosh. We're going to solve the energy crisis. We're going to solve distributing power to devices in the public and such. And basically what they're saying is...

Dave Jones: And make money at the same time, folks.

Chris Gammell: Yeah. Well, there it is. Hey, what do you know? Sony's going to sell you not just the outlet. They're going to sell you the authenticated plug that you have to plug into that outlet. And Marjorie rightly so. I mean, she sounds pissed the way she writes about it. But it's because it's ridiculous. I mean, like, who's going to pay for this thing? Basically, you'd have a power plug that's no longer a standard plug. It'd have to have this chip inside of the... The receptacle would be a wall receptacle like normal. But then it would have an authentication with whatever you're plugging into it. So if you didn't have one of those, you don't get any power. It's like...

Dave Jones: Who's coming up with this stuff? Well, it's biggest Brian Dead idea, really.

Chris Gammell: I know. And I think the idea is, you know, you'd sell these outlets to, you know, Starbucks and places where people go and plug in. And really, it is a problem that needs to be solved where you, you know, power isn't really free, right? I mean, if you go to an airport...

Dave Jones: How much power are you going to use at, say, a Starbucks to power your notebook is trivial?

Chris Gammell: Well, yes. I think eventually this will work its way into, like, electric vehicles and stuff like that. And I was listening to a Science Friday podcast about distribution electricity for electric vehicles today. And, you know, they didn't really talk about the... You know, they talk about charging stations and everything like that. But they didn't really talk about, you know, if you're going to have to pay for it when you go to work or if you're just charging on the street. But that is going to be a big part of it, I think, eventually. And so this might actually have some kind of validity when you're charging amps. You know, lots and lots of amps.

Dave Jones: Oh, a huge amount. Yeah, exactly. But apart from that, for a regular PowerPoint, it's just ridiculous.

Chris Gammell: Right. Because they said, like, the... Yeah, the capacity of a car is like 31 kilowatt hours, they said. I think for one of the cars, maybe the Leaf or the Volt. But a battery on a laptop is what? Like amp hours? Yeah. You know, like maybe 10 amp hours, something like that?

Dave Jones: Yeah, 10 watt hours or something. Yeah, it's not much.

Chris Gammell: Oh, yeah, watt hours, right. Watt hours. It's got to have the capacity of the battery. Right, yeah. Yeah, so...

Dave Jones: And what about the reliability shit like this? You know, you're there. What happens in 5, 10 years down the track? Like, how reliable does this electronics has to be? You just want power, right? Right. You know, it's just ridiculous. Is this thing still going to work in 20 years' time? Like a regular PowerPoint? I doubt it.

Chris Gammell: Well, if you buy a big supply of these plugs that talk back and forth, right? I mean, it's like... That's the other thing. I think it's just like, oh, look what we can do with RFID. You know, like, ooh, it's so cool.

Dave Jones: Yeah, it's cool. But it's bloody... The wastage is massive.

Chris Gammell: Yeah.

Dave Jones: The amount of electronics you have to produce just to handle the distribution of power is ridiculous. Back in my day, we did that with two wires. Oh, yeah, exactly. A switch. Yeah, we're saving the planet by introducing all this advanced technology that needs billion dollar fab plants.

Chris Gammell: No, no, we're saving Sony from their rapidly declining margins, really. Right.

Dave Jones: Oh, boy. Yeah. Oh, it's just got brain dead written all over it. Fail.

Chris Gammell: Yep. Boy. Oh, boy.

Dave Jones: But it's great. You know, I can just picture how brilliant this thing is at the Sony product design meeting. You know, they'll just go in apeshit over this idea. Right.

Chris Gammell: With 20 people sitting in the meeting, 18 of whom got their little counters running, seeing how much money is being wasted. Yeah.

Dave Jones: Oh, boy. Yeah. Oh, that's so bad. Anyway. Fail. Move on.

Chris Gammell: All right. How about a little bit more fun here? We had a suggestion a while back. This is an old piece of equipment, or an old piece of video, but I thought it was fun, and I know you would think it's fun. Did you see that someone turned a quadcopter into a Back to the Future DeLorean?

Dave Jones: I want one.

Chris Gammell: That's so great. End of story.

Dave Jones: That's brilliant.

Chris Gammell: Basically, if people haven't seen it, they use the quad rotors, you know, they dressed up the middle of it so it looked like a DeLorean, but then the quad, the rotor parts, they actually put LEDs in there, and then it looked like the DeLorean.

Dave Jones: It actually looked like the flying DeLorean, yeah. Yeah.

Chris Gammell: That's great.

Dave Jones: And it actually sort of hovered just like it did in the movie. You know, when it's flying up there and it hovers around in that storm, you know, and it sort of had that wobbly, same kind of wobbly effect. I love it.

Chris Gammell: Yeah. It was awesome. Great stuff.

Dave Jones: Brilliant. Ah, boy. What else have we got? We've got, where was it? Where was it? Speaking of videos, there's a video. I love that you put this on here. The video of making the read switches. Oh, yeah, yeah, yeah.

Chris Gammell: That was great. That was from Reddit, I believe. Right.

Dave Jones: And it's from Mida, who make read switches. That's pretty much all they do, actually. They're one of the big manufacturers, worldwide manufacturers of high-quality read switches. You've got to watch this video. It's a very corporate kind of, you know, video. But it shows you the steps. I lost count of the number of high-tech steps which went into making a bloody read switch.

Chris Gammell: Right. And we should, for people out there who don't know what a read switch is, basically. So it's a relay, basically. You have a coil of wire.

Dave Jones: Well, you don't even have that. It's just a glass tube. If you're just talking about the read switch itself, there is no coil. It's just the glass tube with the two contacts.

Chris Gammell: Right. And this is before you put it into a relay.

Dave Jones: It's just the two contacts inside a glass tube. And when you put some sort of magnet near it or a coil, then it pulls the two contacts, just like a regular relay. And, yeah, I couldn't believe the amount of steps which went into manufacturing this thing. They use a class 1000 clean room. And all these different layers are put in 10 microns of sputted, you know, tungsten or something over the contacts. And then, oh, man, the amount of steps. Dozens of them. And they've got these automated machines all inside the clean room. You know, I thought, well, what the hell do you need a clean room to make a read switch for? But, you know, apparently it's important. If you want to make a real high-quality read switch or a read relay, then, you know, now I can understand why they cost, you know, 10, 20 bucks each for a good read relay. That's what I've paid in volume. You know, I've had to do a real high-quality, you know, switching matrix. And some of the relays are 10, 20 bucks a pop.

Chris Gammell: Yep. And you really only want to use those when you're, you know, you need really high isolation. And that's basically the price is itself right out there with, you know, a $10 part.

Dave Jones: Yeah, you want very low leakage ones and very high reliability and all that sort of stuff. And, yeah. I've got to ask, how would this factory differ from the Wanhang Lo factory in China spitting out read relays? You know?

Chris Gammell: You know, that probably shows up in your, you know, if you do a test on the actual isolation of it, so it'd probably be a difference in impedance, right? You can see terra ohms of impedance on a good read relay, right? But if you have some kind of leakage inside that tube because of dirt or whatever's in there. Contamination in there. Yeah, you're just going to see a really, you're going to see a lot of leakage through the tube as opposed to, you know, on the board or anything. You know, those kind of resistances, you start to really see, you see problems everywhere, right?

Dave Jones: You see leakage across your PCB material, across your board. So you've got to actually cut slots into your PCB. I've had to cut slots between all my relays to actually get the leakage down because then it's got to go right around the slot. You know, I can't just travel over the surface if you get dirt or moisture on the surface of your board or something like that. Right. And moisture, incidentally, can get trapped under the relay. So if you're using your board, you're designing a switching, you know, some board which uses, you know, very high leakage, very low leakage read relays, you know, and the humidity changes you're using in some environment, you know, in some tropical region. And it's changing all the time. Then you can get moisture trapped under your relay and that can ruin your entire product, you know? Yeah. So you've got to take that sort of stuff into account when you're laying out your board. You've got to think about that sort of stuff. Uh-huh. Can moisture get trapped? Right. I'm going to put a big slot. Whoosh. You know, try it now, you bastard.

Chris Gammell: Yeah. And sometimes even that doesn't matter because you can see, you'll see during processing, you know, even if you do have a slot and it's designed all right, you know, sometimes the processing itself will deposit that kind of dirt or moisture underneath the relay. So you have to watch out in the actual manufacturing process.

Dave Jones: And after X amount of actuations, the material can start rubbing off and floating around inside and, you know, things like that. So it's very important to get a high quality read relay if you really care about the reliability of something.

Chris Gammell: Yep.

Dave Jones: So, yeah. Anyway, watch the video. You'll be absolutely amazed at the amount of technology, manufacturing technology goes into a humble glass tube with two contacts in it. And it's filled with nitrogen and it's all laser cut. They've got all these custom laser machines which cut and seal everything. And, ah, it's just incredible.

Chris Gammell: Yep. Another basic component, transformers. There was actually a great, great article on EDN this week about using a power transformer. I'm not sure who the – it wasn't like a normal writer. I think it was like a submitted article. Ghost writer? Maybe. Maybe ghost writer. Probably just an engineer who wanted to talk about transformers. Right. But a really interesting article because he talks about actually using a 60 hertz transformer or a 50 hertz transformer at different frequencies and then the D rating you can do in order to use it at different frequencies and voltages and stuff like that. And basically, though, it becomes a – this article was just a really great basics of transformer theory and that alone is what did it for me. I mean this is a good refresher course on how transformers work and how you can design them and a lot of the equations behind it. So if – I don't know about you but my transformer knowledge came much later in my career.

Dave Jones: Yes. I've done it as required pretty much. That's right. I have to – oh, I've got to do a custom transformer for a switch mode power supply. Oh, shit. I better go. I better go learn how it all works again. Yeah. Yeah. Yeah. And that's – Yep. Yep. And if you're not doing it often – Because that's the shit you don't remember. Right. Well, some people do. If you remember all your transformer theory and all the – you're quite – man, you're just weird.

Chris Gammell: Well, you're probably doing a lot, really. Oh, you're probably doing a lot. That's probably the real thing. Yeah, exactly. Right.

Dave Jones: You typically only tend to remember the stuff that you do reasonably often or you have an interest in.

Chris Gammell: Yeah. Well, I find with like transformers, I remember like the turns ratio stuff, right? Like just the basics so I can do quick – Oh, the very simple stuff, yeah. Yeah. Quick calculations. That'll come back.

Dave Jones: That'll pop back in your memory after 20 years maybe but –

Chris Gammell: Right. Yeah. Right. But yeah, it's – you know, I don't think a lot of people are doing custom transformers these days because there's so many off the shelf available but – I don't know.

Dave Jones: In DC to DC design, it's still very common to sort of do your own custom flyback transformer or something like that if you're doing a flyback supply.

Chris Gammell: There's so many off the shelf – I mean I'm actually working on something like that now.

Dave Jones: Oh, there's not a – well, yeah, there is but they're quite expensive, you know, if you get them from the regular manufacturers and usually if you're doing a DC to DC converter, you want to keep the cost down probably and – I don't know, man. It depends on what you're shooting for.

Chris Gammell: I find some great off the shelf stuff. Yeah. And the thing is like it's so – I'm looking at such low profile ones that it's – you know, you kind of have to go off the shelf because getting something that small and custom is pricey. So I don't know why – I think –

Dave Jones: It depends how many – how much time you've got and how many spins you want to do too.

Chris Gammell: Well, yeah, that's the main thing, right? But I think the – I think off the shelf is the way to go always. I mean 80% of the time you should be trying to go off the shelf.

Dave Jones: Oh, yes, you could.

Chris Gammell: But you should be – Still in understanding it and still in trying to actually do the design and, you know, work through the equations of a flyback design, yeah. You still need to know a little bit about transformers.

Dave Jones: You do need to know your stuff so you can pick the right transformer. Yeah. That has the right properties, you know. You can't just go, oh, that's a flyback transformer. It fits my size. I'll choose that, you know.

Chris Gammell: Yeah.

Dave Jones: It doesn't quite work that way. There was actually an output. It'll generally work, but your efficiency – you'll be sitting there scratching your head wondering why your efficiency is 60% at 1 amp instead of the 90% you wanted, you know. I mean, because you've got to target a lot of this design. If you're designing a flyback transformer, you've got to target a specific output voltage and specific output current, you know, or an output power. So, you know, it's optimized around that particular output power. And if you don't do that, if you don't do the calculations and you pick some generic thing, then your efficiency is going to be shit house. So you've got to be careful. It's really hard, though, when you're trying to design something wideband. Like if you're designing a switch mode that needs to power anything from, you know, one hilly watt through to 10 watts, you know, it's really hard to get good efficiency over the whole range.

Chris Gammell: Yeah. And that's a big trap, too, because the top line, you know, marketing spec is always, oh, we're 95% efficient. Yeah. It's like – but then, you know, they should always have to put in parentheses underneath if it's perfect.

Dave Jones: Or at one output current. And they typically show you the actual graph. They'll show you the efficiency graph versus output current, you know, and you know it's going to peak at one current. And that's – Yeah.

Chris Gammell: Yeah. But you've still got to – you know, it's – I don't think – they don't always put the graph at the top. Sometimes you've got to dig for it.

Dave Jones: No, it's buried away. Yeah. So, yeah. Exactly.

Chris Gammell: Oh, marketing. Yep. Isn't that fun? Well, speaking of power supplies, you've got some power supply stuff ongoing, right? I mean, you're deep in your own design cycle right now.

Dave Jones: I have. I've got two designs that I'm working on. One is the legacy design. Well, one is the battery-powered power supply, which you've seen. Another one is – well, another one, which I haven't talked about yet and which I might get into eventually. But I'm working on two. The new one I'm doing in the new CAD tool, I've chosen DipTrace. So, I'm trying to use DipTrace for the first time, and it's working pretty well for me. So, I thought I'd try an alternative commercial package. And, yeah.

Chris Gammell: I think you should have full disclosure here, too. Why are you really using it, Dave? Because it's a commercial package.

Dave Jones: But, no, I am in their reseller program. If you buy it, you can do it through my website, and I get a commission.

Chris Gammell: Yeah. Okay. There we go. Full disclosure.

Speaker ?: Yep.

Dave Jones: Full disclosure. There you go. But I'm not – you know.

Chris Gammell: I just meant because you got a license from it. So, that's what I thought it was.

Dave Jones: Well, I got a free license. Hey, it's the only other commercial package that I've got a license to. So, it makes sense for me to use it. If someone else wants to give me a license, I'll happily try their package out, too. You know?

Chris Gammell: The open source community would love to give you a free open source license for a package.

Dave Jones: One free open source license, please, to KICAD. No, I fully promote KICAD because it's open source. I just want to – in fact, I'm probably going to end up using two packages. I want – I'm going to support KICAD because it's open source, and I'm going to use and support an alternative commercial package to Altium as well. So, you know, there's room for both. Because everyone always asks me, you know, which PCB package do I recommend? And, you know, I don't just want to recommend KICAD to everyone. That's just – you know, it's just not realistic. So, people ask me, well, what commercial package do you recommend? I don't want to use an open source one for whatever reason. Fine. So, you know, I want to be able to recommend something other than Altium.

Chris Gammell: Yeah. You said that there's no libraries, though, in this new one?

Dave Jones: No, this new one has a shitload of libraries, but I can't get any of the parts. That's right. I put this on last week on the list. Murphy's Law, every time, it is the age-old problem with CAD. No matter how big your component library is, I think DipTrace say they've got like 100,000 parts or something. I don't know. It's a lot. It's very extensive. You look through and, oh, they've got hundreds of TI parts or something. Do they have the one you actually want? No. Oh, no, no. Definitely not.

Chris Gammell: Why would they? Of course not.

Dave Jones: You know. Murphy gets – and I did that for like five different manufacturers. I got, right, I've got a TI DC to DC converter. You look through the list and you start to salivate because you see all these TI DC to DC converters. And you get through and, oh, they don't have the one you want.

Chris Gammell: It's heartbreaking. Yeah, that's the worst is you scroll down. I've done that before where you're like scrolling down. You're like, all right, 456, 457, 450. Oh.

Dave Jones: I know. What? What? I know.

Chris Gammell: Yeah, sometimes you've got to watch out for that though because sometimes you'll have cross – you'll have compatible packages. So 456 might be the same pinout as 458 even though it's a different whatever. So, yeah, that is –

Dave Jones: And I tried to – once again, I thought I was being very smart. I even went – I knew that there was a national instruments pin compatible and schematic compatible one to the part I wanted. So I went, oh, I'll go check out if they've got the national one because then I can just place that in there and I get my footprint. And I get my schematic and all I've got to do is just rename it. That's it. Bang. But do they have the national one? No. They have everything but – Right. And it doesn't – this isn't peculiar to DipTrace or any – it's the same for every CAD package I've used. It's just –

Chris Gammell: Yeah, yeah.

Dave Jones: Yeah. Can't win. It's not even possible. So anyway. So, yeah, I've started to use it. It seems really good and easy to use, but I haven't – gotten to the point where I'll actually recommend it fully yet because it might suck when I get through the whole process. So I don't know. I'm always reserving my judgment.

Chris Gammell: Yep.

Dave Jones: Yep.

Chris Gammell: So how is that affecting your board? I mean, you're just doing the layout right now for the boards and such?

Dave Jones: I'm doing – I'm still at the schematic stage. So I haven't tried using the board layout side of it yet.

Dave Jones: Okay. Yep. Getting those parts on the schematic. So, yeah, I could find that the PCB layout side of things could suck completely. So, yep.

Chris Gammell: Yeah. Well, speaking of CAD and packaging and stuff – or sorry, part layouts and stuff, I've been talking to a couple guys, one being Stuart Allen and another being CJ Gervaisi. And I think we mentioned those guys on here before, but we've been talking about like what kind of – because we've been talking about KiCat on the show before and there's really not a lot of – even like the KiCatlibraries.org, it seems like it's kind of minimal. We were actually talking about maybe doing a GitHub way of like putting a part library on there or even individual parts and then having people be able to put something up there and then you can actually do modifications and forks and stuff like that because that seems like a pretty good use of that kind of software. CJ actually started a new site. I forget where it is on here. But he started a new site called PCB Exchange and that's something different where it's actually just uploaded directly to the site. But – Right. You know, we're trying different things. So, if people are interested in participating or if they know of other stuff that's out there, I'd be interested from the KiCat side of things because I've already run into that and, you know, first thing you do when you're – you know, first thing you do when you hit that roadblock and you're like, oh, it's not already in the native libraries. You're like, yeah, type it into Google, you know, like tap, tap, tap, tap, tap. Has somebody already done it, you know? Right. And I haven't found it yet but that doesn't mean my Google foo is strong. Yeah, right. Yeah. So, we might be doing a KiCat GitHub account.

Dave Jones: There's lots of logistics involved in how that actually works, you know, how it will actually work in practice with people integrating their parts. Say you've got a library of, you know, Atmel parts. How do you put that one in the existing Atmel library and stuff like that? Yeah. Or should you have your own personal library? You know, here's Dave's EEV blog library with a bunch of random parts in it, you know?

Chris Gammell: Right. And that's usually what happens on people's hard drives, right? I mean, you end up just having the parts that you've used in the past there. Yeah, exactly. As you need them, you pull them in.

Dave Jones: You know, I'm not going to have my own personal libraries and then split them into every manufacturer. You know, I've just got the EEV blog library and that's, you know, that's it. That's the parts I've used.

Chris Gammell: Right. And CJ's idea of having like a centrally vetted body of people that are actually looking over the part footprints is important too, right? Yeah. But then someone's going to do that. So, yeah, in the end, I think, I mean, I think if the world was perfect and I had lots of cash and as a perfect service existed, I would say, you know, it would be best to have a company that's in charge of keeping track of that. And really, there are some companies that do that out there.

Dave Jones: There's an online company. You can pay the, you can go, well, I need this part and you pay them 10 bucks or something and they will do that part for you.

Chris Gammell: Right.

Dave Jones: In whatever package you want. You know, they will give you the schematic and the PCB footprint for it. Right.

Chris Gammell: Yeah. And if you're in a company, right, and you have that clock ticking on your screen in your meeting, right, that makes sense cost-wise sometimes, you know, like because it can be quite cumbersome to do, you know, footprints and stuff.

Dave Jones: Yeah, but a lot of time you're forced to because you've got your own company standards, you know, that you want to use a specific footprint because that's what your specific manufacturer can handle. Or you want your schematic to be split up into multiple blocks instead of one big thing with the pins in sequential number order or something like that. You know, there's many different requirements.

Chris Gammell: So probably we'll have lots of things to complain about ad nauseum, right, on here.

Dave Jones: Oh, there's just, I don't know. Is there, there's really no, there can never be a universal answer to this.

Chris Gammell: No, probably not.

Dave Jones: PCB question. The PCB library question. There just can't be. And anyone who thinks they can solve it is probably living in la-la land. Yeah. Anyway. Oh, well. It's always good to have. One last thing about. Speaking of the. Oh, go ahead. No? Okay. Speaking of the power supply thing, some people have suggested that I make my, the power supply design I'm working on, make the front panel swappable for right and left-handed users. You know, actually put the board behind there with multiple footprints like upside down so that if they want, they can solder the components in one way or the other. And have the binding posts on once on the left-hand side instead of the right-hand side and the knobs on the left instead of the right. And it's like.

Chris Gammell: Really, I feel like that's someone pulling an early April Fool's prank on you and being like.

Dave Jones: Right.

Chris Gammell: Hey, watch me make Dave take his project deadline out two months.

Dave Jones: I can see the reasoning behind it. I can see it too. Like, I'm a lefty. It's a storm in a teacup. Like, I don't think there's actually a problem, an issue to be solved there.

Chris Gammell: No, I don't think there is either.

Dave Jones: When I'm designing my scientific calculator watch, for example, it's going to matter which side I put the buttons on because, you know, you've got to use it right or left-handed. You know, it's one of those things. And if your fingers are over the screen when you're typing on the keypad, that thing matters. You know, that sort of stuff matters. But for a general purpose instrument like this, you know, does it? I don't think it matters.

Chris Gammell: I could say I'm left-handed. I don't know if you knew that, Dave. Oh, you are. Okay. Yeah, I am. Weird. You just get used to it. Yeah, weird. Yeah. But you just get used to it. I mean, and I personally think that's, I mean, aside from the statistics of it, too, I mean, like, what, 10% of people are left-handed? So you'd be designing for 10% of your customers. That right there is a bad idea. No matter what you're doing, unless they're paying a lot more money, no offense to all the other southpaws out there, unless they're paying more money, it's not worth it. Because you've got to design for the big group out there. Because otherwise, if you spend all your time and all your money specializing for that little group, you know, like, you'll just never get it done.

Dave Jones: It depends on what the product is, though. I don't think so, man. Sometimes there can be an advantage there. I can see there. Explain, please. Maybe be an advantage there. What's the advantage? You know? Well, for example, my scientific calculator watch, right? That is, you have to wear the watch on one hand or the other. And, you know, I would design it for me because I wear my watch on my left-hand side, right? So I'm going to put the keys on the right-hand side and the screen on the left so when I bring my right hand over. You know, but that will almost make the product unusable for anyone who likes to wear a watch on their right arm. You know? Right.

Chris Gammell: But here's my counterpoint to that is that you are right-handed, as is, you know, 90% of the population, right? So the first thing you should do is get that product out there. That's the idea. You get it out there. Oh, yeah, for sure. When all the lefties are like, oh, Dave, I can't use your watch, then you do it because then the first product is already out there. And then you can roll in all the fancy add-ons that you're going to want to do anyways. I do agree with that. Yeah. So you always got to get that first product out there. You know, like get it out there because waiting around for that 10%, no offense, lefties. I mean, I'm not buying Dave's watch either.

Dave Jones: And you might want to put some thought into it. How can you better improve the design to make it left, right-hand neutral perhaps? That's another aspect. Yeah, exactly. Why not put the screen above? That might be worth putting a couple of days of design effort in, you know? Right.

Chris Gammell: And that's more philosophical though, right? I mean, you have – everybody has limited information when they start out with the design. Yeah. But, you know, you do the best you can. But eventually you got to push something out there. And when – I think it's always better to push fast and fail. Unless it's some like big – you know, there's examples where you don't want to do that, right? I mean like if you're doing a huge consumer product where you get one introduction, well, you have to be fast and awesome, right? Yeah. That's right. Yeah. But most of the time you're not going to be able to. So I think, you know, as long as it doesn't hinder you later, then go ahead for it, you know? Like, I don't know.

Dave Jones: Yeah, I totally agree. So the answer is, sorry, no, neat idea. But no, I don't think I'm going to go to the extra effort to make it all swappable. You know, and Mike from Mike's Electric Stuff had the idea of that, of course, you would actually silkscreen both sides of the board so that when you flip it left, right, you flip the panel over and you'd have the silkscreen. And it's novel, right? It's a neat, elegant engineering solution.

Chris Gammell: Is it warranted?

Dave Jones: I don't think so.

Chris Gammell: Well, I've done a design with a mechanical design where I tried to make everything symmetrical. And my engineering, like, the engineering center of my brain was just like, yes, this is awesome. You are doing so great.

Dave Jones: Look at the symmetry. Oh.

Chris Gammell: And the thing is, though, eventually, you know, it started out real basic and it started out symmetrical. But eventually you drill down and then you do something. And then it's like, that is the point where it all falls apart. And you could spend a lot of time, you know, then making that stuff symmetrical. But there's, like, I think it was like, you know, like an LED had to then also be on both sides of the board or something like that. And at a certain point, it's just like, well, you know, unless you're really dedicated to the symmetry, it's tough. And with circuit boards, you'd almost have to route signals symmetrically and just, oh, just, yikes. Why not? It's not optimal. It's, you know, like the engineering part of my brain's like, Chris, what are you saying?

Dave Jones: But engineering is the art of compromise often. That's very correct.

Chris Gammell: Yeah.

Dave Jones: You know?

Chris Gammell: Yeah. That's a fun thought project.

Dave Jones: It is a fun thought process to think about. I enjoy thinking about the design aspects of doing something like that. But whether or not I'd implement it is a different thing entirely.

Speaker ?: So.

Dave Jones: Yep. Yep. Yep. Sorry. And there's one more thing on the power supply, if I may. The AM Spire Richard from the forum. That's his, AM Spire's his handle. He's actually taken my PSU design and he's trying to do it, redo it with all discrete components. You'll love this, Chris. Very, lots of analog shit involved. And yeah, he's going through all the simulations, all the designs, and he's doing it for BJTs and MOSFETs. And he's trying to get it stable across all the loads. And it's really interesting to follow. Yeah. So, and it'll show you actually how hard it is to make something simple like a power supply. You think, okay, I'm just going to do a single transistor power supply. Pretty simple. You know? A follower. But no. To make it stable across all loads, there's a lot of art in that. So, please follow along with Richard's design quest there. It's very interesting.

Chris Gammell: Yeah, that is interesting. That's a lot of the stuff I used to do, a lot of that analog control loop stuff.

Dave Jones: Control loop stability and things like that. Yeah. Yeah, there's some...

Chris Gammell: And it's not easy. No, it's not. And there's some good books out there for that, though. That one Sergio Franco book that I really like is a good book for that. It's art and... Oh, crap. Designed with operation amplifiers and analog circuits or something like that. I'll try and link it in there. But that's a good book for that kind of thing. So, if people are interested in analog control loops, it... Oof. It is. That's a hell of a topic. Yeah.

Dave Jones: Oh, boy. So, that's enough on the power supplies.

Chris Gammell: Yeah. I think. Yeah. Well, no.

Dave Jones: Actually, we do... We have a chip of the week later. Oh, yeah. Should we do that last?

Chris Gammell: Well, before we do that, let's talk a little bit about parts because we've been following along a little bit as Ian from Dangerous Prototypes. He actually... We saw his Akihabara video. He actually went to the... This is a couple of weeks ago now. But he went to the market in Seoul. And this is another great video just about these huge electronics markets. So, I highly recommend people look at this video as well because this one's actually even bigger than Akihabara. And this is in South Korea.

Dave Jones: I haven't looked at it yet. I'm going to have to watch it.

Chris Gammell: Okay. Yeah. It's pretty great. And, you know, like they... You know, it's a lot of small vendors and small stalls and stuff like that. But you don't... I never realized. They're just everywhere, man. I mean, like Asia. Asia's where it's at. Yeah.

Dave Jones: Well, there's lots of electronics in South Korea. It's a very big industry there.

Chris Gammell: Yeah. Yeah. Well, I've been there. But I didn't get to go to these markets. So, that's why I'm kind of jealous, actually. So... But, yeah. Great video. So, I highly recommend people check it out. Should we talk about parts, though? And talk about your chip chip? Chip of the week? Chip chip chiparoo? Chip of the week. Yeah.

Dave Jones: Da-da-da-da! Is the... Where is it? At the bottom. Here it is. It is the Texas Instruments slash... Well, Burr Brown. Come on. Let's still call it Burr Brown.

Chris Gammell: Slash national slash...

Dave Jones: No, wrong. Oh, God. Yes. Right.

Chris Gammell: The megalith of TI.

Dave Jones: Yeah. Yeah. Anyway, it is branded a Texas Instruments part, but it still has the old Burr Brown part numbers. I still like. Oh, yeah. So, it's the INA. You know, when you see that INA, you go, oh, it's a Burr Brown.

Chris Gammell: Oh, yeah. That's got to be good. Yeah, no. It's got to be good, right? Yeah.

Dave Jones: And it's the INA-219. And I'm using it in my new power supply. So, you've probably already seen it. So, it's a bit of a spoiler. But it's really neat. It is a zero-drift, bi-directional, current power monitor with I2C interface. That's the full title of it. And it's basically... That sounds nice. Exactly. Yeah. It's a little SOT8, S08 package device. You can even get it in a SOT23-8. Is that going too far? SOT8 pins on a SOT23, six is enough. Eight on a SOT23 is a tiny little blob of nothing, you know?

Chris Gammell: It can get smaller. SC70 is smaller. Yeah, I know. SC70-8 would be like...

Dave Jones: Pain in the ass. Yeah. Anyway, use the S08 package. It's nicer. And it basically... What it is, is it's got a built-in programmable gain amplifier. Or it's actually attenuator. I might get into that. It's got a built-in 12-bit analog to digital converter. And it can monitor... It's basically designed for use with an external shunt, current shunt, and a high-side current shunt. And it can measure the power that your product is drawing. So, it can measure the voltage and the current, because it's got an input switch. So, it can measure the input voltage and via the current sensor resistor, the input current. And then it's got a power register built in. It calculates the power based on a factor you put in. You can program in the shunt register values. And it will actually output via the I2C interface. It'll output the power your product's using or the voltage of the current. And you can measure it. And it's cheap. It's like a buck or something like that. And, you know, considering it's got a 12-bit ADC in it, and the programmable gain amp for a buck is, you know, it's pretty good. And it's got a low offset voltage. That is pretty good, yeah. And, you know, it's got like 10 microvolts offset voltage, typical. Yeah. Max of 40. And it's really neat.

Chris Gammell: You know, I tweeted about this the other day. But I have, like, designer guilt when I use these kind of parts. I don't know about you. But I'm like, ah, you know, I should be doing this in transistors. And I should be designing. Yeah, you think you're selling out to the man. I should. I do. I mean, I feel like you could throw down, like, three of these and monitor, you know, you can monitor the power of your house. You know, like, and talk back to a simple micro. Right? And I'm thinking, oh, I should do this in analog. But, man, sometimes it's just so easy and neat. And it's, you know, it's all on TI. If they mess it up, you get to yell at them. And it's like, ah. Yeah. I don't. Yeah. But I do it anyways. I mean, I'm using parts like this.

Dave Jones: Well, that's engineering. I mean, you use what's available. There's a reason there's 200,000 parts on DigiKey is because, you know, that's how you do advanced. No one would be able to do advanced modern systems and designs without using these devices. Right. And we should bring that up, too. In the meantime, it'd be five years.

Chris Gammell: Exactly. We should bring that up. Because we actually, there was feedback on that survey we did and said, oh, well, you guys are always talking about system engineering and you should talk about chip engineering. It's like, well, okay. But this is, I mean, first off, the system engineering is the stuff Dave and I know. But, you know, it's like, what are you going to do? You know? It's tough.

Dave Jones: You've only got six months to design the product or something or, you know, I've got a week to do this module. You know? Shit. Well, you're not going to, you know, roll your own, are you?

Chris Gammell: Right. Of course not. You've got to look for the chip that does the job. Even cost. You know, if you buy 10 transistors, you might get up to that same dollar price, right? Or, you know, an I2C interface.

Dave Jones: And you need a voltage reference for your ADC because, you know, it's like, yeah. You could easily spend more than a dollar rolling your own solution here with a 12-bit ADC and a programmable gain amp and registers and I2C interface.

Chris Gammell: Or as you said with that SOT23, the size is another big thing too, right? The size, yeah. You really just can't even do that these days because, I mean, you can squeeze a lot more onto silicon than you can on a board with a three-leaded transistor, right?

Dave Jones: Sometimes you won't even have room for the bypass cap for the bloody single, for this little SOT23 bastard, let alone another whole bunch of discrete components. It's not going to happen.

Chris Gammell: Yeah. So. I don't know. I just, I still have, I'm going to still have designer guilt. I know I will. Yeah, I know. I don't know if that's a real thing, but I feel it. I feel like I should be doing more, you know, but thank you. Thank you, chip designers. You help me do what I do. So, I appreciate you guys.

Dave Jones: None of our modern designs would be possible without using the system on chips we have these days. It just wouldn't. Oh, yeah. And the microcontrollers and everything else. It just. Right.

Chris Gammell: And you look at, like, we were talking about the cost of engineering before too, right? I mean, for the same kind of thing, I think about, like, DC to DC supplies. You used to have, you know, like, racks of DC to DC converters doing that kind of stuff, whereas now you do it with, like, a, you know, a tiny little inductor and a six megahertz switcher chip, right? It's like, you don't get to do that back then. You know, you were, you lost efficiency. You lost time designing this thing, and it cost, you know, 20 times as much because you had all these different components on there. So, yeah, I get it, but I don't know. It's just a part of engineering, right?

Dave Jones: Yeah. Some people commented on my power supply design why I didn't just integrate the switching controller into my microcontroller. I've already got a microcontroller. Why don't I just use that to, you know, drive some external MOSFETs and switch this, you know, do my own switcher. You need to do a PWM or something.

Chris Gammell: Yeah. Sorry? Like, do a PWM or something like that.

Dave Jones: Yeah, do my own PWM and use it. You know, it's got built-in ADCs, and it's got, you know, I can use it to bloody do this thing. Well, God, why would you bother? You know, all that effort in the code and everything else to try and get your own switch mode controller working in them, you know, it's just ridiculous. Right. Just use an external controller chip you know that's going to work and be done with it.

Chris Gammell: Right. We're describing basically the competitive advantage of every chip manufacturer out there. Right. I mean, like, they do stuff so that we don't have to. Yeah. That's right. Yeah. But at the same time, we should mention there is risk, right? I mean, first off, as everyone probably has felt before, when there's an obsolescence, you are sometimes really up a creek without a paddle, right? That's a big deal. Or you just can't get it.

Dave Jones: Right. Yeah, exactly. It's not obsolete anymore, but some bastard's gone and sucked up all the stock, you know?

Chris Gammell: Yeah, exactly. Exactly. And you're screwed. Yeah, that's definitely a risk there. The other risk is, you know, you just lose, you don't have that knowledge for a time when you do need to customize something, right? That's, that's like a really soft one that people don't think about. You know, like, like transformer design we were talking about before. If you have a company that's using all transformers, like just doing transformers offline and then going to a linear supply or something like that. And then you, you get rid of all those people because you're like, oh, DC to DCs are the future. Well, when you go back and you have to design a transformer, guess what? You're paying that consultant $300 an hour, just like we were talking about, you know? Just because the knowledge isn't there anymore. And yeah, you just, that's a risk. I mean, that's a risk with you using integrated components like that. So, I don't know.

Dave Jones: But that's why we get all our modern technology, I'm afraid. Yeah. So there's no way to avoid it. One thing I like about this, these Texas Instruments slash Burr Brown data sheets are that they have on the front page, mind you, a little box, which has the related products. And they're very similar to the functionality of this device. But, you know, so if you pull up this data sheet and you go, oh, this is almost perfect, but wouldn't it be good if I had another, a different chip, which did this? Well, it's there in the related products. There's the INA209 if you want one that has peak hold and a built-in comparator. It almost does my entire power supply functionality in one chip. There it is, the INA209. You jump to that one and it's got some extra stuff built in. Works exactly the same as this one, but, you know, it's got a couple of extra features. And then there's a whole bunch of other, they name a bunch of other ones that if you want it in a smaller package, you know?

Chris Gammell: Yeah. And as I said before the show, this is kind of like the Wikipedia way of browsing parts. Which is really nice, right? I mean, you're not going to make any huge jumps, but you're going to make these incremental jumps among similar topics, really. So I really like it, too. And the fact that it's actually hyperlinked in the PDF.

Dave Jones: It's hyperlinked in the PDF, yeah.

Chris Gammell: I know that's not much to expect from a chip company, but I am thoroughly impressed because I've seen much worse. Yeah, exactly. Exactly.

Dave Jones: And it's a good way to design because often you might miss chips doing the parametric search thing, either on DigiKey or on the manufacturer's website. You can easily miss parts doing the parametric searches.

Chris Gammell: I was complaining about parametric searches earlier this week because it doesn't seem optimal. I mean, I know optimally you talk to someone who just can snap to and say, oh, no, you need this part, right? Yeah, because they've used it before, yeah. Right, exactly. But it's just not – I hate parametric searching. I really do. I love it sometimes.

Dave Jones: Sometimes it's incredibly powerful because you can find, oh, there's 30 chips which do what I want. Freaking brilliant, you know? Yeah, I guess so. And then you have choice. Choice is good sometimes.

Chris Gammell: Choice is good sometimes. But, you know, the problem is your situation is when you know exactly what you want, right? My situation was I wasn't really sure what I needed and I need, like, a really general basis, right? And that's when it is good to talk to, like, apps engineers and stuff like that. But, you know, like, I was like, oh, I need to do DC to DC switching and I think I need to do it isolated and I think – you know, there's all these soft requirements. And I'm like, yeah, where do I really go for that? And because if you just drill down into a parametric search, you know, it's like, oh, well, you're looking for this part, right? And it's like, well, I don't know if I'm looking for that part. So it kind of stinks. Some people suggested the Adafruit because Adafruit keeps a Wikipedia type or Wikipedia page of parts that they like. And that is a good kind of resource. I like that as a – because then it's like, oh, power. Okay, so then you have – you know, you kind of drill down but it's still very generic and then you have a list of parts.

Dave Jones: It's like a potluck thing going to somebody's – you know. It's like you're using somebody's component, CAD component library, you know. The odds of the device you want being there are pretty darn small.

Chris Gammell: That's true. But it's something, right? And the people that were suggesting it to me were well-intentioned, probably assuming that I was doing it for hobbyist stuff. But I was really talking about for stuff at work, right? So I was very specialized. So in the end, my solution was go talk to a vendor, you know, and see what they've got. And so that made the vendors feel good because they feel like they're still needed these days. But we'll see. It's just that that's the best I could come up with so far. So as long as I'm buying a lot of parts, they'll keep talking to me. How am I supposed to stop?

Dave Jones: Do we have any dregs?

Chris Gammell: The dregs of the amp hour, eh?

Dave Jones: Dregs, the end trials.

Chris Gammell: Well, since we talked about transformers earlier, there was a really cool video on Reddit, actually, about transformers. But not just your run-of-the-mill, you know, 120 down to 5 kind of transformer turns or 120 down to 12 or something like that. This was the other direction. This is going 120 up to 10,000 and actually disassembling microwave.

Dave Jones: Well, that's going to wire it up backwards, isn't it?

Chris Gammell: Well, yeah, it is wiring it up backwards. But this is actually harvesting microwave transformers in order to get that high-voltage transformer. And so this is a really cool how-to video on how to, like, you know, heat it up and pull it apart. And because there's these really standard...

Dave Jones: How to kill yourself, yeah.

Chris Gammell: Well, they say not to do that. So I think that's... Right, what? They say not to do what they say or not to kill yourself? They say don't kill yourself.

Dave Jones: Right, okay.

Chris Gammell: Yeah, and we should state that here, too. If you're playing with transformers on the back of a microwave, you can kill yourself. So really, transformers in general, you know, they're not current-limited. Even if you have a fuse in there, enough current can slip through to really mess you up. So be careful. So... But... Really cool way of... You know, if you're doing, like, a high-voltage project, you know, like a Tesla coil. Or a... We saw that one video where the guy was doing a high-voltage wire in front of a blackboard to hold a piece of paper by electrostatic force. That was a fun one, too. So if you need high-voltage, if you know what you're doing, this is a great video. So I like that one.

Dave Jones: Cool. I'm not really into that sort of stuff. I like looking at it, but I'm not... You know, I like looking at Tesla videos and, you know, what people have produced and stuff like that. But, eh, I'm not really into that sort of thing.

Chris Gammell: Don't know about you. No, I like it, too, from an outsider's perspective. Yeah. Yeah, I mean, personally, normal voltage scares me. I don't need high-voltage scaring me. Yeah, exactly. Anything over 12-volt DC is HV, as far as I'm concerned. Yeah. I mean, yeah, it's... I just really... I think it's a comfort thing, really. You know, like, people kind of work up to that. You start at 100 volts, and, you know, you get popped a couple... I haven't popped before by 1,000 volts, and it wasn't fun, but I'd work on it again. You learn, you know, safety techniques and ways to work better. But, yeah, you work up to that.

Dave Jones: Or you just avoid it completely, and you enjoy a career, you know? Yeah.

Chris Gammell: That reminds me of my... One of my favorite... My favorite Jack Handy quotes is, I bet the beginners in a yodeling class just go and try and yodel right off. No, you see, we work up to that. Right. Oh, so good. All right, well, people can find us on Twitter, on G+, on Facebook, on email.

Dave Jones: Is anyone using G+,?

Chris Gammell: I don't think so.

Dave Jones: Did it fizzle out?

Chris Gammell: It's... They're trying to work on it, but... They hired Kevin Rose, the guy that started Dig at Google, to work on it. So we'll see if that does anything. But you can find us all those places. You can find us on email at theampire at gmail.com. And, yeah, I guess I'll talk to you next week, huh, Dave? Absolutely. All right, see you then.

Dave Jones: Bye. Bye.