#713 – Rubber Duck Incarnate

Chris Gammell: This is The Amp Hour Podcast. Released January 25th, 2026. Episode 713. Rubber Duck Incarnate.

Dave Jones: Welcome to the Amp Hour. I'm Dave Jones from the EEV Blog.

Chris Gammell: And I'm Chris Gammell of Contextual Electronics.

Dave Jones: Did you notice my one octave lower voice today?

Chris Gammell: I feel like I should drop my voice a little bit. And I'm Chris Gammell of Contextual. Everyone should be like, what speed am I listening at? Do I have it on the wrong RPMs? Yes, this is CNN.

Dave Jones: Yeah, yeah, there's something wrong for my voice at the moment, which has lowered.

Chris Gammell: I should have just finally made good on that promise and, you know, put you into a vocal replicator. A replicator, major voice AI. Right, yeah, right. Can you say anything I wanted to say?

Dave Jones: Am I real or am I AI? How have you been? It's been a while.

Chris Gammell: It's been a while, yeah. Yeah, I think. I mean, we took it easy. I mean, I took it easy on the first half of the holiday. Then you were traveling with, you know, no internet.

Dave Jones: So it's been, I'd say, two months. Well, that was the thing. I was hoping that we'd do something while I was away. I went on a three-week jaunt road trip out in the sticks, the back of Burke, you know, whoop whoop, as we call it here. I was assured by she who planned everything. By the tour director. By the tour director. Yes. That, oh, we're staying at a really nice place and it's going to have rock solid internet. No worries whatsoever. You can do all your business. You can do, you know, upload all your videos. You can do everything else. You can still stay, no, no, fail, fail. And we didn't have mobile reception as a backup because of places we went. We had practically no mobile reception. And so we were reliant upon the, you know, the house that will stay in internet connection. You know, we get like an Airbnb house because there's more than just our family stay in there. We go in a big group and, you know, we hire a big house and everything. And it's, oh, no, it's a big, fancy, expensive, real house. And they've got real internet. And nope. Nope. We even went to one place. We were talking about this before the show that had Starlink. And I thought, oh, yeah, no worries, right? We'll have, I'll get hundreds of megbits or even 50 megbits or something, you know. Everyone raves about how great Starlink is.

Chris Gammell: I mean, I had it when I was on a boat. And I was in the middle of the ocean, you know. It was like wild.

Dave Jones: It's basically limited by the number of nearby users using it, really. So I thought, yeah, we're in the middle of the sticks, you know. We'll have, you know, fantastic. Have it all to myself. Have a whole satellite to myself, you know. And, well, as they rush in and out every five minutes. Oh, remind me to talk about that after this. So I go, oh, great. You know, I do the speed test or whatever. And I, like, the page wouldn't even load. And I'm going, what, you know. And then I finally get the thing, you know, I finally use other tools to try and get the speed. And it was, like, 400k bits per second. And then it was maxing out at 800k bits per second. That's K. That's not meg. That's, like, you know, it's, like, with modern websites, it's borderline unusable. So we couldn't even stream, you know. We'd take our Amazon Fire Stick with us, you know, so that we can stream all the 10 million different bloody plans we signed up for.

Chris Gammell: Yeah, of course.

Dave Jones: You know, to keep the kids happy and everyone entertained.

Chris Gammell: That was good data energy right there, you know. Yeah. What am I paying for?

Dave Jones: Yeah, exactly.

Chris Gammell: Bloody kids.

Dave Jones: Yep. So, yeah. Couldn't, you know, you can't even download, like, a 240p potato vision version. Can't even stream that on 800k bits.

Chris Gammell: You know, even more so than the, you know, what you could or couldn't do, it's always the expectation. You know, it's, like, I was told. It's, like, you had prepared. And, you know, you had, like, brought all your media on the disk. Yep. And you're, like, you know, and you're just, like, all right, I'm going to the, you know, the back of beyond. Who knows where I will be? There's no chance I'll have a connection. But if there's even the hint of a connection, I'm, like, oh, no, I'm cool. I got, I can reach the world. And then when it doesn't work, then I'm pissed.

Dave Jones: Right. I'm not sure. Yeah, you don't follow me on X. I said, well, you don't use X anymore, but I posted a Twitter on X of me working. I took the entire morning off to go catch up with work because, you know, I had, like, you know, I have to release videos and have to do stuff and catch up on 10 zillion emails and do whatnot. So I had to ride to the nearby beach that actually had mobile reception where I could actually get 50 megbits per second via my mobile. So I, you know, I was sitting there off my laptop next to the beach doing, you know, it was tough, you know. It was, like, rough. I mean, worst place is the beach, right, of course, yeah. I got bloody burnt. The Australian sun's unforgiving. I also posted on X that I got burnt. I was sitting, like, I found, like, you know, one of those huge, like, shelter hut things with, like, a table and everything and chair, you know. Oh, no, was it a pergola? Yeah.

Chris Gammell: Pergolas will get you.

Dave Jones: Yeah. Pergola?

Chris Gammell: No, like, the ones that, like, look like you have a roof, but they're- Pergola. Not pergola. You think Italians say pergola? Pergola. Pergola. Pergola, pergola, pergola. No? Is it me wrong? I don't know. Pergola. Okay. You know, this is why you're southern, you're below the equator, sir. Stay down there with your pergola.

Dave Jones: Right. So I found this big thing. So I was sitting under there for four hours, right, in complete and utter shade, and I got burnt because of the reflected sun or whatever. I don't know. Yeah. Oops.

Chris Gammell: Yeah. Yeah. It'll get you.

Dave Jones: Yeah.

Chris Gammell: It'll get you.

Dave Jones: Don't do that. Yeah. Australian sun. Very unforgiving. Especially in summer. Yeah. Anyway. So, yeah, that was fun.

Chris Gammell: I don't know. I think you might be- Pergola. I've been saying it wrong, apparently.

Dave Jones: Oh, you have. Is that the official, where did you get it from? Just random Google?

Chris Gammell: There's some Google thing. Yeah. Oh, I say it in the British pronunciation. Oh, really? Oh, yes. I'm very high class, Dave. I have been watching Downton Abbey again.

Speaker ?: Right.

Chris Gammell: You know, I've been watching a lot of Bluey, too. You'd think I'd pick up on that, you know? You're right.

Dave Jones: Oh, boy. Yeah. Anyway. So, yeah, that was fun. Yeah. So, we're talking about, I should have bought a Starlink Mini before I went. You know, one of those little backpacky jobs. And can't you get, like, a $5 plan or something that, like, keeps you alive, but, you know, and then you pay as you use or something? I'm not sure how it works, but-

Chris Gammell: They will get to your wallet, I'm sure. I'm sure there's, you know, it's just, like, the drip feed of, yeah.

Dave Jones: Oh, well, I'm sure they will. But, yeah, I think it's a $5 plan to, like- Yeah, keep it active. It's designed for this exact scenario. You buy it, and you keep the plan alive, but you don't use it until you need to. And then you pay $5 a month or something to keep your plan going. But then when you use it, though, I assume you ramp back up to, you know, the normal rates, which is fine, which is exactly what I want it for. So, yeah, I'm thinking about getting one now, but it's bloody too late.

Chris Gammell: We're going on a big yearly trip. Sounds like you should have thought about it a month ago, yeah.

Dave Jones: Yeah, exactly. Well, I did, but I procrastinated, and she said, oh, no, wait, trust me, we're going to have rock-solid connection.

Chris Gammell: Can't blame the misses, man. Sounds like you let go of the wheel, you know?

Dave Jones: Yep, yep, I did. I should have just ponied up the money and bought the Starlink dish. I wanted to complain. I wanted to complain about Starlink. On the amp hour?

Chris Gammell: How dare you, sir?

Dave Jones: I know. It's unprecedented.

Chris Gammell: Okay. Complain. Go.

Dave Jones: I've been doing a lot of night gazing lately, you know, just stand out there in the dark and just watch the sky every single time. Satellites just whizzing over as bright as the freaking moon.

Chris Gammell: You know those are Starlink, right? I know. I have an app I was showing my kids. It's called Stellarium, I think. Stellarium. Oh, right. It's like a follow the thing, and you can actually, it's great because you get, you know, satellites come across. It's got all that mapping data. And yeah, let me tell you that the blanket of Starlink is, it's not small.

Dave Jones: Yeah. And if you see one, just wait 30 seconds and another one will come right behind it in the same line. So I'm not sure if they were being like new ones being sort of deployed and still sort of getting into their orbit or whether or not they do actually orbit in a line or in various lines overhead. I'm not sure, but.

Chris Gammell: Well, it really does look like a blanket. It looks like a, like a diamond shaped blanket. If you, if you look at the maps, I mean, they're like crisscrossing each other. That's just to get coverage. But the line I think about, like when they are deploying. Yeah. You, it looks like a train, like pulling out of the station. Yeah. Yeah.

Dave Jones: Yeah. It's a train. Like I was like, I was counting. I had my stopwatch out like, yeah, like every 47 seconds or something. And a new one would come to that exact spot. I was picking the same spot in the sky and a new one. No, no, no.

Chris Gammell: I think that's, that's just normal operation. I think I'm talking about like, if you saw like, if you saw like 15 in a line, that's like a deployment. Oh, okay. No, right.

Dave Jones: No, I didn't know. It was just a new one, but it was the exact spot. Like I actually got a tree, you know, there's no wind. So I got a tree. I was using that as a reference point. So I put that exactly over one and then I time it and another one will fly in exactly the same tip of that tree at, you know, 47 seconds later or something.

Chris Gammell: I don't know if you heard this, but people get upset when their internet accesses.

Dave Jones: Yeah, right.

Chris Gammell: So you need a lot of satellites, you know? It's, yeah.

Dave Jones: And they're going to lower them even further. They've just announced that they're going to lower them another 50 kilometers or something from like, I don't know, 350 to 300 or something.

Chris Gammell: Here's what I want to know. You know, they're not the, they're not the only game in town in theory. They are really the only game in town right now, but like longer term, you know, Bezos is trying to put up a cancellation. There's other ones. Holy crap. Like how much stuff can we put up there? Like there are practical limits at some point.

Dave Jones: Yeah. It's just nuts. It's crazy. I personally know people who want to put up more, but they won't be in low earth orbit. They want to put up thousands of them in higher geosync or whatever, sun synchronous or whatever. But yeah. So do you want to know why they're lowering these satellites?

Chris Gammell: I don't have any good guesses. Sorry. Yeah, go ahead.

Dave Jones: Okay. Well, it's rather interesting. It's because the sun has greater output at the moment and that does something to the earth's atmosphere that compresses the earth's atmosphere or expands it or something, whichever way. And apparently the Starlinks have to be, they can't be too high, not only for the latency thing, but also because if they do fail, they want them to burn up fast, right? You know, they want them to, you know, drift out of orbit and burn up quick. And apparently because the sun is changing its solar output, that changes the atmospheric pressure. It's probably not the right word or something like that. So that, yeah, they have to lower the satellites in order to keep the same ratio of the ability to deorbit in a certain amount of time. There might even be a requirement for that as part of their license or something. That might even be a requirement. Oh, yeah, they have to burn up in X amount of time. I don't know if that's a federal requirement or whether or not that's, you know, just their own personal requirement. But, yeah, they are actively changing the orbit of 10,000 Starlink satellites to lower it.

Chris Gammell: So your latency is going to improve. The lifespan of those things will go down too because then you have, what, more atmospheric drag?

Dave Jones: I don't know. It's going to be the same. The atmospheric drag has gone down because the sun is increasing its pressure. So they want to go back to where it originally was. So they're not increasing the drag. They're just going back to where it was originally designed to be, if you know what I mean.

Chris Gammell: I don't, but I don't really get much about space, to be honest.

Dave Jones: What, should I follow all the cool space YouTubers and stuff?

Chris Gammell: I don't, no. It's great. Yeah. Yeah, I mean, it is. It's a wild time. No doubt about that. I just, you know, how many things can you follow?

Dave Jones: Please tell me that you know that humans are going back to the moon in a couple of weeks. I did know about that, but part of that is because you've been talking about it. But I thought you weren't following me. The only place I post is X.

Chris Gammell: You did a live stream the other day, as you may recall.

Dave Jones: Oh, right.

Chris Gammell: And you said, I will be up at whatever time in the morning it takes. Yes, whatever time in the morning. Yeah. Yeah.

Dave Jones: I mean, yeah, rightfully so.

Chris Gammell: Yeah. Artemis is going, right?

Dave Jones: Yeah. Yep. Artemis 2. Yep. Yep. Sending four humans back to the moon. And they're not going to land, of course, because they haven't even started building anything to bloody land yet. Yeah. Apart from Starship, which is probably not going to land. So, you know. You still haven't read Delta V, have you?

Chris Gammell: The Daniel Suarez book?

Dave Jones: No. No, I haven't.

Chris Gammell: You got to read that. Before they go, you got to read that, man. Yep. Yeah. All right. I remember reading about how there was this big dip. I think it's 70s into the 80s. There was like a dip in like interest in space just because it was like so, you know, there were a lot of missions, a lot of space shuttle missions, stuff like it. It's not like there was no interest. Right. But it also wasn't like Apollo interest. Yeah, exactly. It was like it became business as opposed to like groundbreaking science and, you know, and like world news. And that's good, actually. And I feel like we're not back to that, but we're there's a lot of good things happening and a lot of consistency. And that's actually all good things. So not that has any that I have any excuse in this space. I just don't pay attention. Right. Okay.

Dave Jones: Well, no, I love following all the space news on X. It's one of the big, you know, I'm always real. I think today I've retweeted three different space things or something like that. So, you know, yeah, yeah, yeah, space, space, space, space, space, space, very excited. So, yep, yep. I'll be watching that whenever it happens. So, yeah. Unfortunately, it's costing you as the US taxpayer an awful lot of money, but, you know, heck, humans are going back to the moon. So kind of cool, I guess, even though, yeah, we don't have any real concrete plans to follow it up. But, well, they do. There's the whole Artemis plan, which Artemis 3, 4, 5, 6, 7. But, you know, if you've been following it, you know that it's kind of a bit of a mess.

Chris Gammell: So, yeah.

Dave Jones: Yep.

Chris Gammell: Yeah. I feel like all things are kind of up in the air when it comes to space funding right now. Yep. Governmental space funding.

Dave Jones: Yeah. Because they abandoned all the cool practical solutions to land on the moon and they went with Starship. And, of course, Starship just keeps blowing up and, you know, it's like trying to land the Starship on the moon. Have you seen what Starship looks like? You know?

Chris Gammell: Yeah.

Dave Jones: It's bloody huge. And they're just trying to land that thing when everything else on the moon they're trying to land is just falling over. It's like, yeah. Anyway, so they, I'm not sure if you've heard recently, but just in the last three months or something, they've announced that their SpaceX are not the only plan now for the human landing system, the HLS it's called. So they're allowing Blue Origin and others to compete now. It's like, yeah, they should have done that from day one. Well, they should have just chose the most practical solution. Yeah. Anyway, no, no. It's the US. It involves politics and involves state politics and involves funding. You know, jobs in states and things like that. So, you know. But, hey, the Artemis rocket's cool. It's like, it's awesome. It's just a one-off thing and it's super duper bloody expensive. It's a bloody cool rocket, though.

Chris Gammell: Yeah, of course. Of course. You know, big exploding things going up and not blowing up. Yes.

Dave Jones: Fantastic. It's great.

Chris Gammell: It's great.

Dave Jones: Yeah. Science! And they could have put humans on the first one. They could have put humans on Artemis 1 and sent them around. That would have been fine.

Chris Gammell: Well, that's the science part, Dave. But no, wusses. Yeah, I was going to say.

Dave Jones: Bloody OH&S people.

Chris Gammell: Maybe the PR people, too, you know.

Dave Jones: Yeah, maybe.

Speaker ?: Yeah.

Dave Jones: But no, they would have survived and it would have been great. So, you plan on watching it or?

Chris Gammell: I don't, you know, I don't know. I guess so.

Dave Jones: Dude, first time in like over 50 years, humans going back to the moon and you're just.

Chris Gammell: I mean, it takes a while right from the launch to the actual landing. So, like, I'll watch one. Oh, yeah. No, no.

Dave Jones: You can't watch. Well, there is no landing. There is no landing on the moon. The orbiting, rather. The orbiting.

Chris Gammell: Yeah.

Dave Jones: So. But no, you have to at least watch the launch, dude.

Chris Gammell: No, I don't really. No? No.

Dave Jones: So, I can't get you excited? There's nothing I can do to get you excited about it?

Chris Gammell: I, could you get some of the spaceship designers on the Amp Hour? That would be exciting. I'll do my best. Okay, good.

Dave Jones: I can get Billy.

Chris Gammell: Who's Billy?

Dave Jones: You haven't seen my video about Billy. I'll have to link in my video. He's what makes Artemis go up. Billy. Hang on. EV blog. It's on my second channel. EV blog to Billy.

Chris Gammell: Is this going to be a joke? If this is a joke, I'm not sure.

Dave Jones: No, no. It's not a joke. Here we go. You can watch it later. Don't watch it now. But Billy. Billy. Billy is the unsung hero of the first Artemis mission. And basically, he's been working there. He's the height. Did you even watch the first Artemis launch? No. Probably not, right?

Chris Gammell: No. I don't think so, Dave.

Dave Jones: No? Oh. Okay. Anyway, the first Artemis launch, they're about to launch it. And then like at 40 seconds or a couple of minutes before launch, they found, oh, no, there's something wrong with a hydrogen valve or something, right? So they're about to launch the thing. The rocket's fully fueled. And so these three hydrogen techs called the Red Team or the Red Squad or something, they had to go out under the fully fueled rocket with their giant spanners and fix the hydrogen leak or whatever it was. Oh, no. And this team was led by Billy, who's been the head hydrogen tech at NASA for whoever he, I'm not sure. Does he work for, yeah, he probably works for NASA, yeah. And he's been the head hydrogen tech for like 40 years or 30 years or something like that. And he's made every space shuttle go up. You know, he's like, anyway, so they had this conference. So they're all heroes. They got them and it launched and they actually went out there, fixed the problem and it launched, right? So these guys were heroes that they went out there and fixed the Artemis 1 launch. And so they got them on the podcast. Like, you know, they're doing the live feed podcast show of the launch, right? This is the official NASA feed. So they got the three of them on there to, you know, talk and talk about how they fix it and all that sort of stuff. And there's Billy just sitting in the background. They didn't give him the microphone. And it was a tragedy. He was like-

Chris Gammell: So he's a folk hero. That's what you're saying. He's a folk hero.

Dave Jones: He's a full hero. Yeah, Billy. And he sent me a, him and the red team sent me a signed autograph photo of him with Bill Nelson, the NASA administrator, the former NASA administrator, Bill Nelson. So, and ironically, Billy worked on Bill Nelson's space shuttle launch in the 1980s because Bill Nelson, who was the head of NASA and he was a former astronaut shuttle astronaut. So, yeah. So anyway, Billy, Billy makes Artemis go up. Got it. Yeah. Anyway, it's my meme video. Oh, five and a half thousand views. Oh, has it been three years? Has it been three years? Oh my God. Since Artemis one to Artemis two. Oh, sorry. Yeah. I'm just, yep.

Chris Gammell: Just wistfully wondering how, why governmental launches take so long. Yeah.

Dave Jones: That's, yeah. Oh boy. Anyway, SpaceX has launched, you know, 500 rockets in that time. Exactly. You know. Yeah. Yeah. It's crazy. Like one, two per day or something. Anyway.

Chris Gammell: Cool. What's new in the electronics side of the world then? How about that?

Dave Jones: Well, electronics. Well, news. We should, you know, considering that this is like a, you know, newsy type podcast. News-ish. News-ish. You know, da-da-da-da-da-da-da. I don't think. Culling straight off the telecine.

Chris Gammell: If this is the thing you'd sent me, I don't think it's really news, but okay.

Dave Jones: Well, it's news in the community. It's that Art of Fruit and SparkFun have kind of like had a falling out. You know, the two big open source, you know, maker icons in our industry. They've had a very public falling out. It's kind of everywhere on X and the EV blog forum. And it stems from a, because Art of Fruit were selling the teensy, which is, which SparkFun make on behalf of the guy who designed it. Paul somebody.

Chris Gammell: Paul Stoffergen.

Dave Jones: Yes. Yeah. Anyway, he used to make it himself, but then he decided, no, I don't want to make it myself anymore. So he got SparkFun to make it. And then Art of Fruit simply resell the teensy. And SparkFun have just announced, yeah, that they're not letting Art of Fruit resell it anymore because of codes of conduct violation. Anyway, they've had like a private, sort of a semi-public private spat kind of thing. And as a result, Art of Fruit are no longer allowed to sell SparkFun stuff. So anyway, it's, yeah, it's all kind of, it has nothing to do with the teensy, by the way. And Paul has nothing to do with that. But it raises an interesting question. I didn't know the teensy was not open source. All of the libraries and stuff are, but it uses a proprietary NDA bootloader, which Paul wanted to make it open source, but he couldn't. That's right. Because he signed an NDA with NXP, I think it is, who make the chip that's on the teensy.

Chris Gammell: It is NXP, yep.

Dave Jones: Yeah. So it's like, it's a closed, technically a closed sourced product.

Chris Gammell: Yeah. It's interesting because you see, so like I've seen people who've designed it in and they're like, okay, well, I'm buying this teensy. I want to like put it on my board. And then you can't do that because of, you need the bootloader.

Dave Jones: Well, you still can. Yeah. Sorry. But you need to buy the bootloader. You need to buy the bootloader. Yeah. Yeah. You've got to physically buy the board. Yeah.

Chris Gammell: Yeah. Oops. You could buy the chip as well. You could buy the bootloader chip. So there's a separate bootloader chip that you can buy from PJRC.

Dave Jones: Oh, right. Paul's own.

Chris Gammell: So if you think about that as like an IP block going onto a secondary microcontroller. And that's basically the bootstrap to get the IMX working.

Dave Jones: So if you design your teensy into your product and then you want to refine your product and reduce the bomb cost, you don't want the whole board soldered in. You don't want the headers and all the rest of it. Right. If you just want to put the chip on your board and still run your teensy software or whatever, you know, for want of a better term. Yeah. You can buy the chip. You have to buy the chip from Paul. Yeah.

Chris Gammell: Exactly. Yep.

Dave Jones: Right. Okay. What? So some people just weren't aware of that. They just built it in, designed it into their product without knowing that. Is that what you're saying?

Chris Gammell: I think people figured it out. I'm not sure.

Dave Jones: Well, I'm sure they would when they went, oh, I can't just download the. Yeah. Right.

Chris Gammell: You can't just buy an IMX RT 1062. Yeah. Yeah.

Dave Jones: And then you go, oh, I can download the bootload. Oh, no, I can't. Yeah.

Chris Gammell: Yeah. I don't know what it is. Maybe the register maps or something like that were under NDA or, you know, who knows, but it doesn't really matter.

Dave Jones: Oh, it's part of the security or something. I think Paul did a great explanation on the forum. I will link it in. Paul did come on the forum and do a very extensive post. I'll try and find it and I'll try and link it in. But yeah.

Chris Gammell: Paul has been listed under my, you know, I have a list of guests for the Amp Hour. Paul has been under the heading White Whale. What? For basically all of the Amp. He's my Moby Dick trying to get him under the Amp Hour. He is the most, he's very good online, but boy, he is, you know, he does not like the spotlight in that way. Like he's, you know, he's the nerd's nerd. I love talking to Paul. Yeah. Like I get to, when I get to see him at like Hackaday events and stuff, man, I love hanging out with that guy. Got it. Paul, if you're listening, you are welcome at the Amp Hour.

Dave Jones: There you go. I've sent you the direct link to Paul's post on the forum and it explains the whole thing and what goes into it. Yeah.

Chris Gammell: Yeah.

Dave Jones: I'll actually read the appropriate paragraph here. Teensy does have a closed source bootloader. The reason is simple business reality of operating a tiny company serving a small niche with a larger, larger, anyway, market with several large, comparative large companies. They mentioned Art of Fruit, SparkFun and all the rest. We, hang on. No, where does he mention NDA? He does mention NDA somewhere. Sorry. No, that's not the paragraph I wanted to do.

Chris Gammell: I mean, if you think about it, like it is a way to encapsulate, it's basically a way to encapsulate IP as well, right? So if you think about all the work that Paul and his team have put in, you know, on making the Teensy what it is, like it's basically the Teensy magic black box, right? Like literally magic black box.

Dave Jones: No, it's more power to him. It's like, yeah, here he says, comments have been made about the NDAs. We did indeed sign an NDA with NXP. The chip security features are the main thing covered by the NDA. Teensy comes in a standard version which doesn't support the security stuff and a lockable version which offers encrypted firmware and secure firmware update process. The lockable version would not have been possible without the security info NXP provided under the NDA. So there you go. But also the reason is partially that, well, he put a lot of work into this and he obviously wants to make a return on it. And you can't do that if you just give away the bootloader because who's going to buy it from you?

Chris Gammell: Well, it's interesting in this case as well. Like, you know, the Teensy. Have you used the Teensy before? No. No, I haven't. Basically, it's a compatibility layer for like, say you were using an Arduino Uno, right? Like an old school Antmega 328p, whatever, right? You know, basically, it's like your board is taking cocaine. You know, it's like, you know, basically an IMX RT part on there is, you know, it's something like 4600x faster. So a lot of people, basically, they maybe are running a super loop and they want it to, you know, they just want to juice that thing up. Man, the Teensy is an easy and accessible way to do so. And that is before you get into all the other things that Paul has kind of loaded in there as well. Like you think about the audio stuff that's in there as well. There's like a whole audio toolbox and things like it. So it is really, it's an accessibility layer on top of, so it's like you have Arduino IDE and you're starting from that. And then basically, you know, all this stuff can compile down to a Teensy board. But then there's also all this other toolkit on top of it to, you know, so like Tim, I think Tim's been on the show before. Let me double check. Trash 80. Yeah. Trash 80's definitely been on the show before. That is a great episode where Tim was talking about how he uses the Teensy to make his episode 292. So my brain still works post-holidays. And Tim still makes some of his boards based on top of the Teensy. And so I know he knows Paul as well.

Dave Jones: So got it. Yeah. So I don't blame Paul whatsoever. And this is just like a private spat really between.

Chris Gammell: Yeah, I think those are two different paths really. Yeah, yeah, yeah.

Dave Jones: Yeah, yeah. They're two entirely different things. But anyway, I thought it was interesting about the whole Teensy thing.

Chris Gammell: SparkFun is saying that they are not transacting with Adafruit at all anymore. So I think it's like, you know, they did cross each other's SKUs and whatever.

Dave Jones: Yeah, now they're not going to do that for personal reasons. Yeah. The Teensy just happened to be the collateral damage really. So, yeah. Well, it's not really. I mean, it doesn't really stop anyone from getting Teensy's. You just have to buy it from SparkFun instead of SparkFun or Adafruit. So it's like, you know, meh, meh.

Chris Gammell: And that's when the great distribution wars of 2026 began.

Dave Jones: Wars of, yeah. And, yeah, speaking of which, you know, because we'll discuss in this on the forum, this thing about, oh, keeping, you know, like making the chip secure or something like that so that you're the source of it. This is not new. This has been going on since before open source, since before the internet, back in the magazine project days. Right? It was common back then that you would not publish or allow to download from a BBS back then or whatever. And, you know, dial up BBS at three other four per second, you know.

Chris Gammell: Right. You tune in the AM radio to catch or capture on your cassette tape.

Dave Jones: Because the magazines back then, they ran their own BBS and you could dial in and you could potentially down. You know, if they used one of these newfangled microcontrollers in a project, right, you could dial in and you could potentially download the, you know, the binary for it so you can program your own chip. But a lot of people. You wouldn't download a car, Dave.

Chris Gammell: Come on.

Dave Jones: You wouldn't download a car.

Chris Gammell: Still one of my favorite.

Dave Jones: Oh, God. I was watching an old DVD just a couple of months ago and that stupid, you know, FBI warning came up.

Chris Gammell: Oh, my God. Yeah, motion picture of the institution of America. Yeah.

Dave Jones: This is the FBI. You wouldn't download it. Is it? It was an FBI ad or something. Oh, my God.

Chris Gammell: I don't remember who the ad was from. I just remember the ad.

Dave Jones: Yeah, it's got, God, I physically bought the DVD. I have it. Can you not put it on the, you know, like. And they force you to sit through it. You can't skip it. That's right, Dave. It's like, oh, my God. Unbelievable. Anyway, what was I saying? Yeah. Yeah. So, it was common back in the day to either do it with, say, a microcontroller binary or even a PCB file. Because back in the day, you would publish the overlay for the printed circuit. That was published in the magazine for the PCB so that you could etch your own, right? So, they printed it as a one-to-one thing or sometimes a two-to-one you had to scale it or something. And so that you could reproduce and make your own PCB. But some people didn't well know. I put a lot of work into this. I want to make money from it. I want to be the sole source for the microcontroller or the board or whatever. And there was nothing wrong with that, really. And, you know, it's just a choice that an individual creator made, whether or not, you know, they want to turn it into a business or not. And, yeah, so that's been happening since, you know, the early 80s, really.

Chris Gammell: Some of the stuff these days is just the availability to make stuff is way easier.

Dave Jones: And, geez, back then, I mean, you know, it was $1,000 or something just to get a developer programmer for the freaking, you know, whatever microcontroller you wanted to use. That was a big thing about, like, the basic stamps. That's when you – do you ever remember the basic stamp?

Chris Gammell: I was a one-time user, so I never really got a one.

Dave Jones: Right. Okay. Yeah, it was – it basically had a, you know, it was actually programmed in basic, but it had this bootloader that you could download via the serial, like a serial port, a serial connection. So it was like that was game-changing. It's like, what? You don't have to buy the $1,000 manufacturer programmer for this microcontroller? You can just do it via a serial port? You know, that was like – yeah, huge.

Chris Gammell: Yeah. Man, I am abusing the crap out of a Tag Connect right now. Do you ever use a Tag Connect 10-pin?

Dave Jones: Oh, right.

Chris Gammell: Yeah. Yeah. I got some extra pins. I'm like, yeah, you know what? We're going to use this one to shove power in. We're going to use this one to measure power like usual. We're going to add some serial ports. Yeah, no problem. I'll make a little breakout board.

Dave Jones: Oh, you're not physically abusing your – I mean, I'm sure I'll do that at some point as well.

Chris Gammell: Yeah, feature abusing.

Dave Jones: Oh, right. Feature abuse. Okay. I thought, yeah, well, like don't you normally like hook it up and leave it hooked up? But you're – no.

Chris Gammell: In this case, I am – instead of – you know, a lot of people are like, well, why aren't you using – it's like a tiny board. And they're like, why aren't you using a 6-pin? Yeah. It's like because the 6-pin cable I don't like. Or there's also now more recently there's like an edge, kind of like a nibbler kind of that connects to the edge. Kind of cool. But I've got a circular board, so that's not going to work. I'm just taking the extra pins. You know, the 10-pin is really meant for like a J-tag, like a full J-tag. And I'm just using an SWD because I only really use arm parts. And so I got a couple extra pins and I'm using it to shove power in and get serial out. And I've seen it done one other time before. Izurio does the same thing. So I'm just kind of following their lead, that kind of idea.

Dave Jones: And somebody down the track is going to actually complain, why the hell did you do this? It's non-standard.

Chris Gammell: Yeah, exactly. Yeah. I think the key thing is you make it so that it's only forward compatible, right? So it's like not breaking. If you don't have it, it's not breaking anything. But if you do have it, then you get some extra features. Right. Yeah, I think that's the key.

Dave Jones: Then you get some extra features. Yeah, right. All right.

Chris Gammell: Yeah. And now I have to add this secondary board to like break out all these pins onto my order. But I'm like pretty close to ordering a new board. Nice little six-layer jobby. I think I've shown you the old version.

Dave Jones: Yeah, yeah, yeah. You have shown me the old version.

Chris Gammell: Yeah. I've shown you the kind of the ring version. Yes. That's the old version. And then this one is basically the size of a CR2032 version. So yeah. You're right. Got it. A couple more features.

Dave Jones: And you had to go to six-layer because of the sheer density of the stuff.

Chris Gammell: Yeah, it's so cheap too, honestly. It's nuts. It's nuts how cheap it is. Yeah. Yeah.

Dave Jones: It's not blind via so.

Chris Gammell: It's not blind, no, but it is via N-Pad now. So like the JLC process is default epoxy filled. So you can just... It's bad. I'm getting spoiled by it. You know, it's...

Dave Jones: Yeah, I know. Yeah. Everything's... Yeah.

Chris Gammell: It's bad.

Dave Jones: Everything's changed, dude. Yeah, I know. Yeah. I know.

Chris Gammell: But yeah, I make small stuff, so that's nice. Yeah. I'm going to run headlong into like the RF stuff because it's just not that big of a board and I'm not that great at RF. So like at some point I'm going to be like, oh yeah, yeah. It looks like you need like more than the size of a CR2032 ground plane or like just better RF capabilities than I will have. So it probably won't be like, you know, it's not going to be winning any awards in the RF domain, but that's okay. Yeah.

Dave Jones: Right. Have you tried to buy RAM recently? Oh, God, no.

Chris Gammell: Oh my gosh. What a... I'm like in the market. I'm like lately in the market for a new laptop and it's like... Yep. Screw it. Nope. Because I'll try another two years, you know.

Dave Jones: Yeah. But can't you just take the RAM out of your old obsolete laptop and whack in the new one?

Chris Gammell: A lot of the laptops I'm buying these days, it's soldered down RAM, Dave.

Dave Jones: Oh, okay. Oh, right. Okay. So you can't just buy it RAM-less. No.

Chris Gammell: Yeah.

Dave Jones: Yeah. You can... Like, is that something you can still do with like the likes of Dell and stuff? You used to be able to like buy it with no memory or something? Yeah.

Chris Gammell: There's expansion ports and stuff like it, but I mean, you still have to, you know, you probably have, you know, if you were doing like a two or three year like cycle hop, you know, you probably have three years ago RAM versus the latest and greatest, you know. Yeah. And density. Yeah.

Dave Jones: We just tried to do that at home. We had... I thought I had DDR... Like, we opened the machine and tried to put the RAM in. Yeah. It's like, oh, this is DDR4. This machine needs the 5. Needs the DDR5, you know. And they're physically not compatible. It's not like you can actually plug the thing in and it works slower. No, it's just, you know.

Chris Gammell: Right. It's just connectors are different. Yep. Yeah.

Dave Jones: Yep. The little tab on it. Classic. You know. Yeah. Yeah. Bastards. Yeah. So, I'm wondering if my old 128 gig of RAM that I used to have in my Dual Xeon server is now worth something. But it's probably so old it's not, right? Probably not.

Chris Gammell: Honestly, I mean like... DDR3 or something, probably. You can buy old server stuff too. Like, you can buy it a lot of places, but it's... I don't... You know, people are looking to do the latest and greatest machines, right? Right. So, yeah. Or, you know, and they don't want to backboard old stuff too, so...

Dave Jones: I'm sure you can get some sort of adaptory board or something if you're really desperate and you have to use DDR3 memory in a DDR5 machine or something. Maybe there's some way to do it. I don't know. I don't keep up with this crap, you know.

Chris Gammell: I guess you could have a LLM help you to, you know, crosswire it.

Dave Jones: Ironically. Which is the cause for the RAM.

Chris Gammell: Right, right, right. Right, it is the cause. Yeah, the cause and the solution to it.

Dave Jones: Because, yeah, the big news...

Chris Gammell: I have been successfully vibe coding a little bit on a embedded device. And for the first time, that is the first time that's worked for me. And I have like a new guideline. So, this is like... I already had the hardware working. All of the hardware stuff is... This is in Zephyr. So, everything like all the device trees all kind of sussed out, whatever. I'm able to turn everything on and off. But I think the real differentiation in this use case, not saying it's broadly applicable. In the past, I've been like, oh, help me like to get this sensor going, whatever. It's crap. It's still crap as far as I could tell. But once you have it working, now it's like, okay, well, help me build a menu system with this, you know, LED display that I have. And, you know, help me utilize pieces of the real-time operating system to like improve timing and, you know, kind of user interaction and stuff like that. And like that has actually been quite solid. You know, I switched over using Cloud Code for that. And it's been good. Not like perfect, but good. Probably as good as a junior firmware engineer and probably better than, you know, I'm not that good. So, anything's better than me. Yeah.

Dave Jones: Yeah. That'd be the same for me these days. Like, anything is better than my meager programming skills, you know.

Chris Gammell: Yeah. And I think, honestly, in my mind, a lot of this is like just getting you over the hump, right? Most of the time when I find myself like in the thick of it.

Dave Jones: That's all I want these days. That's all I can take it from there. Just get me over that bloody hump. You know, for those who don't hump know that hump is like you're sitting there at a blank editing screen. And it's like, oh, I've got to start typing code. It's like, no, just give me something that at least I can like semi compile. And then, you know, I can go in and tweak to make it work.

Chris Gammell: Yeah. And there's like the bootstrapping nature of it all. Just like, yeah, starting from like someone else's example, you know, even external examples. That's how we always start is usually someone else's example. But I also think just the psychological aspect of it. Just like, yeah, I'll probably be able to figure it out. Or I'll have a friendly, you know, available friend to ask dumb questions to. You're like, that part's not bad. I, you know, if there's nothing else that comes out of this AI boom, if there's that, you know, once in a while asking a dumb question really can help. You know, it's rubber ducking, you know, incarnate. Right. That's nice. Yeah. I don't have huge expectations on the rubber duck waking up and telling me all the answers. Oh, rubber duck incarnate. Yeah. Yeah. I like it.

Dave Jones: Yep. Rubber ducking. Yeah. It's pretty good. Yeah. Yeah. Yeah. Yeah. Yeah. It's right.

Chris Gammell: But it would be like, you know, I think a lot of people expect the duck to wake up and write all the code for you. I think you should still temper your expectations about that. But, you know, it might have a couple of, might get in a good quack or two.

Dave Jones: Sure. We can get AI to make an image of a rubber duck too. I'm sure that's not a problem. Oh. Yeah. That's just trivial. Yeah. No problem with that. I rubber duck wearing a bikini.

Chris Gammell: I might skip on that one.

Dave Jones: No, you might skip. Haven't you, have you heard about all that ruckus?

Chris Gammell: I have. I'd like to skip.

Dave Jones: Yeah. Yeah. Yeah. Well, that made Australian politics. I believe it. Yeah.

Chris Gammell: Yeah. Yeah. I had an interesting conversation the other day about CAD software and someone was saying like, this is of course in my wheelhouse of like, well, what if because key CAD is, you know, flat and open and whatever, that because a lot of people are turning the LLMs to like. What do you mean by flat? Like it's not vertically integrated like a lot of, a lot of CAD programs are right. Because they're built by big companies for big companies. And so they're vertically integrated. Like they want you to buy the whole solution. Whereas key CAD. Yeah.

Dave Jones: They want you to buy the signal integrity. They want you to buy. Yeah. Exactly.

Chris Gammell: And like, and in certain cases you should, right? I mean, like there's very good reasons to in certain ways, but it makes it less. And I should say this, this is mostly my friend's idea. So, you know, he's listening, just, you know, it's his idea. But like the idea that it might actually not turn the tides, but just I think there'll be more interesting, more public things that'll happen as a result, much like other open source software kind of enables, you know, MCPs are out there, that sort of thing for key CAD already. But people building tooling on top of it because it's just a little bit easier to, and maybe a little faster as a result. Cool. Yeah. The reason I'm bringing it up as well is because key CAD just, you know, did their end of year drive email, but they also.

Dave Jones: I heard something about that.

Chris Gammell: Yeah. They also lined up all the, all the things that they're working on right now in 10 that's supposed to come out in early 2026. So.

Dave Jones: Oh, I thought it already just come out or something. No.

Chris Gammell: I should probably know that. I don't believe so. Right. You know, it's kind of the features march on and they probably look a little bit more like those vertically integrated ones. So it's, you know, it's nothing, you know, it's not like, you know, snap your fingers and everything's done, but it continues to get better.

Dave Jones: Yep.

Chris Gammell: Multi-threaded PCB net inspector. I can make a lot of mistakes. And if you could check my mistakes multiple times, you know, synchronously, you could tell me all my mistakes faster. Yep.

Dave Jones: The actual progress of KeyCat over what the 15 years we've been doing this show has just been quite remarkable. But it has taken, you know, that 15 years kind of thing, you know, and it's still going.

Chris Gammell: Don't have to tell me twice.

Dave Jones: Right. Yep. Yes. Back when I famously did my review video of KeyCat in quote marks and they were, a lot of people were angry about it because I was, you know.

Chris Gammell: Was it me?

Dave Jones: Was I mad? No, no, no. It was, you know, it was the, well, actually it was more the community was angry at it. The actual KeyCat people themselves said, thank you. This video is very, very valuable for us. But the community, eh. They kind of went, oh, Dave, you didn't even, what the hell is this? You didn't even read any documentation before you started. It's like, yeah, that's the point. It's like, yeah, they didn't quite get it. Anyway.

Chris Gammell: Yeah. Yeah. Yeah. Some of that stuff has actually gotten a lot better in the, you know, the ease of use and things like it. So. Yeah. It's nice. It's nice. It is interesting seeing like all the, bringing back to the LLM thing and the AI tools, like the amount of like tooling that I see, not just for KeyCat, but really just in the electronic space. I don't know if you've like catch it. I mostly catch it. And I'm like, huh. Okay. You know, another one. And I'm just like, who's writing these tools? Like some of them are getting funded. Yeah.

Dave Jones: There's just infinite number of people writing infinite number of tools. Yeah.

Chris Gammell: It really is. And it's like, you know, on the one hand, great. You know, like there are a lot of problems to solve. I'm not against that. But the other one is like, I feel like sometimes they're then getting these like funded ideas. And then, and then I'm just wondering like, who, who do they think is going to use this? You know, like, not. And not that there won't be users. There will be users. But there won't be enough users that the VCs will be happy. And that's the part that I, I've never understood. I, the best I have going for it is like all the VCs have gotten together and they're all like, well, eventually like, you know, one of the big CAD vendors will buy them. Like that's the best, the best I've gotten. So. Yeah. I have no idea if that's true.

Dave Jones: Of course. That's obviously all there after. Yeah. You know. Right.

Chris Gammell: Guaranteed returns. Right. Yep. Yeah.

Dave Jones: They want their 10X return because it was bought out by a bigger fish, you know. Yep. And they're not holding the bag anymore. That's right. Yes.

Chris Gammell: Yeah. They got to make their numbers, you know. Got to make them numbers. Yeah.

Dave Jones: When that old music stops, you don't want to be the last one without the chair. Right. It's like. Yeah.

Chris Gammell: Yep. Yeah. Okay. Last thing about LLM tooling.

Dave Jones: Oh God. More. I know. I know. I know. I know.

Chris Gammell: I know.

Dave Jones: Oh, come on.

Chris Gammell: There was a Hacker News article a while back called Doom Coding. So it's like when you're like on the train and you're like on the bus, you know, out and about and you're either Doom Scrolling or Doom Coding, they called it. And I really liked it just because it was like, it was kind of like an easy setup option. This is actually how I found that I finally tried Cloud Code in any legit way. It was basically someone was like, well, you can like remote into your home machine and like chuck ideas at the LLM while you're moving about. And it'll actually like be in your programming environment and that sort of thing. Ultimately, like very, what's it called? Rise and grindy. You know that term? No, but. Tech bros who are like rise and grind, bro. Oh, right. Yeah, I can. That's kind of what I imagine. It's like that like always working mindset sort of thing. It's like, yep, yep, yep. But I do. I mean, like I learned some new tooling from it and, you know, I've got Tailscale going now. You know, nothing like novel in any other person's thing. But like now I can like, you know, remote into my machine from anywhere, command the computer to do things.

Dave Jones: But you told me before the show that it completely screwed up your system.

Chris Gammell: Tailscale? Oh, yeah. It totally. I don't. I'm not sure what I'm doing. Yeah. So I was like, I had like an internet. I had like a Wi-Fi blip and it's like, I don't know what any website is anymore. It's like, oh, OK. Nothing of restart won't fix, Dave. You know. Right. I can do that remotely now. Excellent. Actually. Oh, this is actually interesting, though. So that's what I was like thinking about. I was trying to think of what actually I figured out from this. One was Cloud Code, fine, whatever. But the one that I actually really, really liked, I had been like looking around for terminal programs for this sort of thing on a phone. And there's a ton. There are just a ton of like SSH things. But there was one that like came highly recommended called Terminus. In Terminus, it had an option in there. So it's like SSH to this, SSH to that, you know, SF, F2P, whatever to that. And then it had a serial option. I'm like, no frigging way. And Dave, I plugged in a Zephyr board that had a UART on it. And I plugged it into my phone, USB-C to USB-C. And I just started seeing log outputs. And I don't know if I've been just in the dark about this forever. But like being able to have basically like a portable like debug machine, you know, I could actually like interact with, you know, Zephyr shell. I can type stuff in. It'll actually do commands, whatever. But now from my phone, I can like get outputs over the shell. It's like spitting out sensor readings and status and whatever. And then I can send actual commands to it as well. So I can like just like plug in my hardware, USB-C to USB-C. And I can control it from without like pulling out a laptop or anything like that. Have you ever seen that sort of thing before?

Dave Jones: No, but you have to have, wouldn't you, to do this, you would have to have support into the product that you're debugging to be able to send that serial stuff over the USB.

Chris Gammell: Right? Yeah, correct. Yes. Yeah.

Dave Jones: Yeah. So it's not like you can just use it to debugger, you know?

Chris Gammell: No, no, no. This isn't debug anything. This is debug my hardware. This is like, this is like Chris is usually awkwardly standing at a meetup with a laptop balanced on one knee and like a board balanced on the other. Is this because you're using Zephyr? This is because I'm using a UART. But yeah, I also then, I use the shell.

Dave Jones: Yeah, but where is the UART physically in your product? How is it getting to the USB? Like what, what wrapper layer is it? You know, what software is doing this?

Chris Gammell: I'll tell you about the exact board it is. It's an NRI-52-840. There is a USB CDC device in there. And that is basically a virtualized UART. Sorry, that's the virtualized UART that goes straight out to the D plus D minus USB 2.0 pins that are on this part, right? So that's all the Zephyr stuff there. But there's, I also tested it and I plugged it into a CP2102 USB to UART driver, right? Which is basically just doing the same thing in an external chip. Same, same. I have a NRF9160, which is a UART only output talking through this, through this chip to USB, right? So it's USB to UART, UART to the microcontroller. Same thing. Same thing, you know?

Speaker ?: Right.

Chris Gammell: All right. So like now I can like debug and see all the, I've always wondered because it's like, how do I show someone if I'm like, hey, this thing's, it's doing things. And they're like, yeah, okay. I see the LED blinking. That's not interesting. But I can be like log messages, readings, interact with it, whatever. Like, you know, I can do that from the cloud, but now I can do that from my phone. I mean, I've always been able to do it from my laptop as well. Yeah.

Dave Jones: But is the product you're debugging and showing off, does that, is that your code that's output into the serial console or whatever? Or is that part of the Zephyr system? Yeah.

Chris Gammell: So Zephyr has a built-in shell, right? So the shell is like a, you know, it's a menu system effectively over UART. So let's just say it's just going to be USB instead of saying UART. So it's, I've got the USB plugged in for my phone or even for my laptop, right? I'm plugged into this device. I can connect to the USB device. It will then, you know, if I hit enter, it'll give me a prompt. If I type in help, it'll show all the options. If I turn on certain menus, so I can have like a, there's an ADC shell. And so you can just like arbitrarily take ADC readings. Okay, right. There's a sensor shell. You can just arbitrarily take sensor readings. Yeah.

Dave Jones: Cool.

Chris Gammell: Yeah. You can make a whole product and it's like other people build shells into their programs as well, but usually the bespoke, this is just like a standardized subsystem.

Dave Jones: Right. So if you're running Zephyr on X microcontroller, it's got to be tailored for that microcontroller or does Zephyr only run on the one type of micro?

Chris Gammell: No, it runs on a bunch. It runs on a whole bunch. So like on my desk right now, I've got a NXP part. I've got a ton of Nordic parts. I've got Espressif parts. I thought so.

Dave Jones: So each build of that is tailored for that micro. So it knows how to read all the UART stuff. It knows how to read the analog pins. It knows how to write. Okay. So it's got all that base functionality.

Chris Gammell: Each vendor basically kind of makes their own compatibility layer for their hardware.

Dave Jones: It makes their own version of build of Zephyr for their parts.

Chris Gammell: No, no, not their own build. No, it's just the, so like the build is always the same. The kernel effectively is all the same.

Dave Jones: Oh, okay.

Chris Gammell: And each vendor basically contributes their code that says, hey, so we just made ABC123 this new part. We're going to enable the Bluetooth subsystem and the ADC subsystem.

Dave Jones: It's a config file, basically. It's a config file. A bunch of them.

Chris Gammell: Yeah. And C code as well. Yeah, right. So it's gluing together APIs, basically. Right. Gotcha. But because the sensor, sorry, because the shell is a standardized feature in Zephyr, basically you have a UART on your device. You can hook it up. And then if you have a, so say you have an ADC on your device and it's running the ADC subsystem and you turn on the ADC shell as well, those things work together. And then they kind of just links it all together and pushes it out. Got it. Yeah. Okay. Yeah. Cool bananas. Yeah, no, it's great. It's great. It's a huge, not huge, it is a large learning curve, but if you get up it, like, man, it's great for that kind of stuff. Right. Like, especially troubleshooting. Like, you've done troubleshooting like that too, right? It's like, you're used to, like, setting a break point on the register for an ADC. It's like, I don't want to freaking do that. You know, like.

Dave Jones: I know. I know. It's real. That's real old school.

Chris Gammell: Yeah. Right. And it's like, you know, most of us end up printf debugging at some point.

Dave Jones: Yeah, you would dump things to a serial port so that you can, you know, just, yeah, you know, read English things. Oh, yes, it's now, I'm now in this part of the code because I got, you know, it spat out this.

Chris Gammell: Yeah, and it's like super easy to make your own, like, shell commands as well. So, you could, like, say, like, you had a specialized calibration command for, like, an ADC. You could basically write your own shell command that's just, like, calibrate, you know, and then it goes and runs just a function. Right. But it's just super easy to then call it from the command line. Zephyr. Dave, is this the year of Zephyr for Dave Jones? Oh, my God. Maybe. It's momentous. Maybe I'll play with it. Yeah, you should. I mean, like, honestly, the only way to do it is to find a board that works really well already. Yes, yeah, right. Just follow the examples. Don't do anything else.

Dave Jones: I'm not going to do it from scratch. No. No. Yeah.

Chris Gammell: Oh, my God. That's the problem. It's like most people who are, like, bottom-up embedded, they're like, yeah, I'm, like, I'm going to build it. You know, I'm going to build free RTOS all the way, you know, from the bottom-up. It's like, dude, you're going to, it's a bad time. Start from, like, a known example.

Dave Jones: When I was a boy, you did pay a penalty for just, you know, randomly outputting your UART, you know, something to the console, to the actual serial port. Because, you know, I learned very early on that, yeah, okay, you include a printf in your C code so that you can output something to the serial port, right? And, oh, the printf code takes, like, 1K of your memory. You know, it actually compiles in that routine. And it's like, oh, my chip only has 2K or it only has 4K. And I'm using half or a quarter of that just so I can output some nice little strings to the serial port. It's like, oh, my God, I can't afford space, you know? Yeah.

Chris Gammell: Yeah. Yeah. Exactly. Exactly. Yeah, I mean, and that is one significant change for, I mean, not always, right? I mean, there's still a lot of parts out there that are insanely small, insanely cheap, whatever.

Dave Jones: Yeah, K2K, something like that.

Chris Gammell: Yeah.

Dave Jones: Yep. How much room does the Zephyr core take on a micro?

Chris Gammell: How much, like? Well, it's all configurable, right? It's like you just leave stuff out until it fits sort of thing. You know, like much like when they say, like, oh, we put Linux onto, like, a USB drive sort of thing. It's like, you know, like a little tiny device.

Dave Jones: Yeah, they've stripped out the GUI. They've stripped out the, you know.

Chris Gammell: Yeah, exactly. Exactly. Yeah. Nothing really. Yeah. But, I mean, I guess it's flexible in that way. That's not how I'm using it. You know, I'm always using it on large parts with large memories. And luckily getting larger, actually. I'm very excited about new parts continue to have more flash, more RAM. So.

Dave Jones: No, don't mention RAM. But it's all built into the silicon. So, it's fine. They're making it themselves. It's not. Yep.

Chris Gammell: Yeah.

Dave Jones: Oh, boy. What else we got to finish off the show? What else happened while I was away? I totally missed this. I probably would have done a video on it. And did you hear about the Donut Lab solid state battery? No. It was at CES. It was the Consumer Electronics Show. Oh, this Donut Lab, this company claimed, we've developed this solid state battery. And it can work from, you know, zero degrees to 120 degrees. And it charges in five minutes. And it has no, it won't catch on fire because it has no ability to, you know, it's all solid state woo-woo. Yeah. Right. Yeah. It's like, and it basically, they went through, they did a big promo video for it that went viral. And they basically, we've solved every aspect of what's wrong with batteries in EVs today or whatever. And they're tied in with this EV bike company or something that is going to be used in this EV bike or whatever. And it's like, and everyone's going, and even the large battery companies are going, no, this is bullshit. This is absolutely bullshit.

Chris Gammell: I mean, like literally, it's just like definition of too good to be true, right? Yeah.

Dave Jones: It's definition of too good to be true. It's like, they went, we've had a hundred PhDs in battery chemistry working on solid state batteries for the last 10 years. And no. And you guys with like, apparently like four employees or something. And just one person working on the chemist has suddenly developed this thing.

Chris Gammell: This is interesting to think about. Like if you had a true groundbreaking discovery, like the best way to do it was to be to actually just rip it out and just be like small incremental changes. Because like, because no one, if it was truly a truly breakthrough, right? Like, yeah. Like just leaps and bounds ahead. People like us would just be like, yeah, right. You know, like, yeah. And you'd say, oh, well, we'll prove it. Right. But you're not going to get enough groundswell. You'd have to like find some other method. You know, humans are just skeptical in some ways and completely lacklif skepticism in other ways. It's really quite a duality.

Dave Jones: Yep. And as part of their video, I just posted this on X yesterday. They showed this as part of the promo video. They showed a single pouch cell, you know, it's sort of like a small pouch shell. And they were charging it at 270 amps. And they were showing how it only rose like 10 degrees Celsius or something like that. Charging the thing at 270 amps. And everyone went, how is this possible? And I just mentioned, well, they didn't say that the whole 270 amps is going into that one pouch cell. They just let you assume it.

Chris Gammell: Yeah, yeah, yeah.

Dave Jones: So it's like, no, they've got all the others in parallel off camera.

Chris Gammell: That's right. That's right. Hey, no attention to the cells behind the curtain.

Dave Jones: Yeah, the cell behind the curtain.

Chris Gammell: Or the dummy load behind the curtain. Yeah, exactly. I think that's right. What is that smoking back there? Oh, that's our space heater. We're very cold in this Las Vegas convention hall. Air conditioning is off the charts. Did you see that video of the guys that were making, trying to make a cannonball run with Rivian? No. Do you know what the cannonball run is?

Dave Jones: I know the term. I don't know the exact. So it's New York to Los Angeles.

Chris Gammell: So all the way across.

Dave Jones: I knew it was between two destinations.

Chris Gammell: As fast as possible, basically. And it's like the. Right. Okay. I think the current record is like 23, 25 hours. Like it's insanely fast for how far. And this is all on public roads, right? Yeah, exactly. Like if you do the math.

Dave Jones: But aren't there speed limits on public roads?

Chris Gammell: Oh, they sure are. And if you do the math, the average speed of the current record holder is like I think 114 miles per hour average. Right. So it's physically illegal.

Dave Jones: Oh, yeah.

Chris Gammell: Yeah, exactly. They should be arrested on site.

Dave Jones: Right.

Chris Gammell: But, you know, fine. You're going across Nebraska at 200 miles an hour. Who cares? Right. But like, you know, if you're doing that in L.A., like it's not great. Anyways, these guys are trying to do it with a Rivian. But the, you know, the actual thing doesn't matter. The digging into the packs and seeing them. These are actually Rivian engineers like doing it on their weekends as well. Right. And like digging in is so cool. So I'll link that into as well. Okay, cool. It's a two-parter. They had to basically build. They like scrapped a second Rivian and put the battery pack in the trunk or in the bed of the truck. Right. Right. And it's like, oh, my God.

Dave Jones: Oh, so they did it on one charge.

Chris Gammell: Oh, God, no. No, no, no. I think they got like. No, no. I know I'm speaking in freedom units here. But it was like in a theoretical like 80 or 900 mile range. Whereas the Rivian is usually like less than. It was like, I think they did a double pack. So it's like, I think it's like 350 mile range normally. So like significant jump. But then they had to build on charging infrastructure as well. Anyways. Oh, God. Fascinating stuff. Yeah.

Dave Jones: It's great.

Chris Gammell: Yeah. And then, you know, they're still like, well, we're going to drive the speed limit, you know. Yeah. Right. But the hard thing is like refueling. Yes, officer.

Dave Jones: We're definitely driving the speed limit.

Chris Gammell: Yeah. Well, I think the real thing, you know, like just charge infrastructure is tough. And like you're going through the Midwest and, you know. Right. It's not terrible, actually, surprisingly. But it's like, it's tough to charge batteries, you know. And like, so a lot of interesting challenges, a lot of like very, you could tell these are very good engineers and they're just doing it in their side time too. So it's like. Yeah. Right. Yeah. Cool. Yeah.

Dave Jones: Yeah. I think I'd enjoy that.

Chris Gammell: Yeah. Oh, you'll love it. You will love it. Yeah.

Dave Jones: All right. I'll watch that during lunch today. In the land of the free, do you guys have average speed cameras? Like on long highways and freeways and stuff, like there'll be a camera at the start and there'll be a camera at the end and they'll take a photo at the start and the end and they'll Oh, interesting. You know, fine you if you do more than the average. Yeah, right.

Chris Gammell: Because it's like the math theorem, right?

Dave Jones: Is that a thing over there?

Chris Gammell: Yeah. I remember my math teacher in high school told me about that and he was like, you know, it was like all about proving the theorem and just like, yeah, scaring the crap out of us. I don't know any of them personally.

Dave Jones: Okay. Right. Interesting. Oh, they're here everywhere. If you drive like, you know. Interesting. State to state. They're like, you know, I passed probably on my road trip that I just did. I probably went through, I don't know, five, six, seven, eight average speed cameras or something like that just on the long highways and stuff.

Chris Gammell: Well, did you get any tickets?

Dave Jones: Well, I don't know yet. I've only just gotten back. Maybe they're still in the mail. I don't know. Yeah, right, right. Yeah. Yeah. And do you guys have any unlimited speed roads in the US? Because we do here in the outback, I think in Northern Territory, we've got unlimited speed.

Chris Gammell: Like the Autobahn. Yeah. I think there is effectively that in my, I think in Montana and stuff like that, like basically in like the mountainous.

Dave Jones: Okay. Right. It's a state thing, right?

Chris Gammell: Yeah. So it's not going to be any like population areas, but I think there are, I think Montana's popping out in my, in my mind, but I'm not sure if that's actually right. Right. Okay. And also they got moose, man. And if you hit a moose, you're dead.

Dave Jones: Oh yeah. Well, same thing for kangaroo. You know, if you hit a kangaroo at 200 Ks an hour, you know, you're going to know it.

Chris Gammell: Basically you're just toast. Yep. You know, take that, take that into consideration for your, your next cannonball run. Right.

Dave Jones: Right. Okay. Yes. Yeah. Yeah. I was invited to do the solar trip, you know, the cross Australia solar race thing. I've been invited on that a couple of times, but it's a big effort to sort of, you know, go along and, you know, like hop in the chase car and, you know, like camp at night and, you know.

Chris Gammell: Yeah. It'd be fun. It's like a lifestyle thing, right? You know?

Dave Jones: Yeah. Yeah. Yeah. It's a total lifestyle.

Chris Gammell: A lot of, a lot of slow times and fast times and slow times. Yeah. Yeah. Yeah. Yeah. You know, you know, present company excluded, of course. I'm not a huge fan of road trips and that sounds like a big one. Yeah.

Dave Jones: You weren't. No, I love road trips. I love it.

Chris Gammell: Yeah.

Dave Jones: We did go on a road trip together. We even shot, did we shoot an episode while we were on the road that. We were record, we didn't shoot it, but we, we recorded it.

Chris Gammell: Yes.

Dave Jones: We actually recorded it on the microphone. Yeah. I think we turned on our little recorder and.

Chris Gammell: I will link that in as well. Let's see. Probably Canberra in the title. I don't know.

Dave Jones: Yeah. Maybe. Or Canberra deep space, you know, tracking. I don't know.

Chris Gammell: Something like that. Yeah. Right. Right.

Speaker ?: Right.

Chris Gammell: Yeah. There it is. Yep. 343 road trip to the deep space network.

Dave Jones: There you go. Yep.

Chris Gammell: Yeah. This came up actually. I was wondering in, it was definitely a different site, but that new show Pluribus. You watch that one yet? No. What's that? Oh man. It's so good. It's Vince Gilligan. The guy, Vince Gilligan. No, Vince, Vince something. The guy who did Breaking Bad. He has a new like sci-fi show.

Dave Jones: I know you're going to hate me, but I haven't watched Breaking Bad.

Chris Gammell: That's fine. He's great at making, he's great at making compelling television, but he's doing sci-fi. He's doing like quasi sci-fi now. And it's just like, it's the human element, of course, right? Of it all. Right. And it's so good. It is just so good. And there's like a scene with, at the beginning with a bunch of space telescopes and. Oh, okay. Cool. It's so good. Oh man. Excellent. So good. Yeah. Speaking of one last thing that is so good. This was a link from a friend, but it was about how the RP 2350 PIOs work. One, one ROM it's called. Was it a mystery? Not a mystery. It's just, it was a mystery to me and it's like a walkthrough. And so this is Piers Rocks.

Dave Jones: Why is it a mystery? It's just, that's like basic fundamental stuff, isn't it?

Chris Gammell: What is a PIO, Dave?

Dave Jones: Seriously? Like, what am I like? What am I? Like, it's just a parallel IO.

Chris Gammell: No, that's the thing. Not as programmable IO on the Raspberry Pi stuff. So it's got a state machine in there with 32 bytes of machine code you can put in. And like how those all work.

Dave Jones: Oh, I didn't know that. Oh, okay. Right. So that's what's interesting about it. It's got its own little state machine processor. Exactly.

Chris Gammell: So you could basically write like a little listener for serial stuff, whatever. So it's all how that works and how the ROM stuff. That makes sense. Yeah, exactly. Yeah. Parallel IO. Yes. Right. Of course. You know that one.

Dave Jones: Well, that's a, yeah. Because I'm old school. That's what PIO means. It means parallel IO.

Chris Gammell: And this is why we clarified, right? Yeah.

Dave Jones: Right.

Chris Gammell: This is a great deep dive, how it works, how you could do like DMA with it, stuff like that. So cool. Great explanation. Check it out.

Dave Jones: Okay. Oh, so the state machine can pull straight from the memory into the-

Chris Gammell: Yeah. And there's multiples of them on board. So like, this is why you see like Raspberry Pi, RP 2350, RP 2040s. They're all doing really cool things. It's getting pretty popular, isn't it? Oh, yeah. Yeah.

Dave Jones: Yeah. Yeah.

Chris Gammell: Yeah. It was interesting. So Adafruit, you know, in their spat, they said, you know, we're not doing the Teensy anymore. Fine, whatever. We're putting the Teensy form factor for the 3.1 with an RP 2350.

Dave Jones: Yes, they are. They're making their own variant, which people are arguing that it's not really going to be a Teensy, but it's-

Chris Gammell: Misses the point a little bit, but yeah, fine, whatever. I think it is such a versatile part that like, if there was, you know, like in terms of like dropping in for a lot of things, like the 2350 is a great part, no doubt. Yep. Downsides are just that it's, you know, kind of bring your own everything, you know, like that's the downside, but it's also the upside. So- What do you mean, bring your own everything? Well, there's no Flash internally. There's no, you know, it's just like you have to-

Dave Jones: Oh, really?

Chris Gammell: Yeah, yeah, yeah. Perhaps you should listen to the Raspberry Pi team being here three different times.

Dave Jones: Maybe I should. Okay, right. I just assumed it had- What's it? Yeah, right. So it only has RAM internally. That's right. So it's dumb when you power it on.

Chris Gammell: That's right.

Speaker ?: Yep.

Dave Jones: Oh, I didn't know that. I never looked into the details. I just assumed.

Chris Gammell: Everything with Raspberry Pi, and I think, you know, I think the guys, if they're listening, I think they would agree with this. Because everything should be seen with Raspberry Pi, the silicon team at least, is through the lens of cost optimization. And like-

Dave Jones: Well, it does make the silicon cheaper if you don't have to build the Flash process into your silicon. Yes. Because that's a different process, you know.

Chris Gammell: Yeah, exactly. And just like other people are going to- You know, microcontrollers are great because they have Flash on board. But like, let me tell you, the Flash makers are really good at making Flash.

Dave Jones: Then, technically, it's not a microcontroller. That's right. Because the definition of a microcontroller is that it has on board RAM and ROM.

Chris Gammell: Yeah, I think so. Yeah. I think it maybe has some kind of like cash ROM thingy on there. Don't quote me on that one, but-

Dave Jones: Right. No, well, I believe that's a microprocessor then.

Chris Gammell: Yeah. Not a microcontroller. Yeah, I think that little S- If you look at the little blob next to the big blob on the RP, on the Pico, that is Spy Flash. Ah, right.

Dave Jones: There you go.

Chris Gammell: Yeah.

Dave Jones: Yep. But is it seamless? Like, it's like, it appears as though it's internal and you just hook the chip up? Is that how it kind of works?

Chris Gammell: I have not done a design with it, so I don't know. But I- Yeah, I'm sure it's configuration.

Dave Jones: Well, that would be my first big question would be, oh shit, do I have to include bootloader code in my program to actually load from the external bloody Flash memory before it actually does anything?

Chris Gammell: I think if you're pointing at 0x0000, you're going to be fine because it's going to hook it up. Like, where else is it going to go?

Dave Jones: Right, yes. I mean, if I was designing that chip, I would have made sure that it included a state machine in there that simply read from a spy bus external ROM with it, like, completely seamlessly. Otherwise, that would kind of suck.

Chris Gammell: They were here last February, so they probably got another year or two before they're back, but they better come back.

Speaker ?: Yeah, right.

Dave Jones: Okay.

Chris Gammell: I'll link that one in as well. That's 687 RP 2350. Got it. When they came back.

Dave Jones: There you go.

Chris Gammell: Yeah. Good group of folks. And yeah, I think it's really interesting, like, every decision viewed through that, I think they would agree with that too.

Dave Jones: Oh, yeah. Totally. As they should, right?

Chris Gammell: I mean, that's why they're winning a lot of these things.

Dave Jones: Oh, yeah. I would be surprised if they didn't do that. I'd be going, well, what the hell are you doing then?

Chris Gammell: Right.

Dave Jones: Right. Cheaper.

Chris Gammell: We want cheaper hardware.

Dave Jones: Yeah, exactly. Come on, guys. But ironically, everyone now complains that the Raspberry Pi is so expensive that, like, little... You know who's not complaining? People have done comparisons.

Chris Gammell: Jeff is not complaining. Jeff Geerling loves making videos.

Dave Jones: Yes. Jeff Geerling loves that. Yeah, yeah.

Chris Gammell: And we love him for it.

Dave Jones: He loves that. He makes infinite videos on, yeah, on this thing. Yeah. And how people are saying, well, it's cheaper to buy one of these mini form factor PCs than it is to buy a Raspberry Pi these days.

Chris Gammell: I go a little upmarket for the Mac Mini. I'm all about that Mac Mini these days. Their software's still hot garbage sometimes.

Speaker ?: Right.

Chris Gammell: But man, that hardware's good. Yeah. Yeah. Apple Silicon, guys, if you're listening, you know, you're never allowed to... You know, I think you'll actually get, like, shot. Like, Tim Cook will actually put a bullet through your head if you come in the Empire and talk about it. Right. But if you want to... Oh, yes. If you want to tempt that guy...

Dave Jones: If you want to suicide that way, yeah. Yeah, right. By all means. We're all for it. Come on down. All right.

Chris Gammell: Well, we're well over time. All right. On to 2026.

Dave Jones: Oh, shit, yeah.

Chris Gammell: Yeah. I know. I know, man. Time marches on.

Dave Jones: Catch you next time.

Chris Gammell: See ya. Bye.

Speaker ?: Bye. ! Thank you.