#94 – Gnomic Gazumping Gobemouche

Chris Gammell: This is the F-Hour Podcast, recorded May 6th, 2012. Episode 94, Gnomic Gazumping Gobamush.

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

Chris Gammell: And I'm Chris Gammell of Chris Gammell's Analog Life.

Dave Jones: What's up, Chris?

Chris Gammell: What's up, Dave?

Dave Jones: That was a very American, what's up?

Chris Gammell: Oh, yeah. G'day, mate.

Dave Jones: Not really. G'day. Should we start by talking about who we've got on next week's show?

Chris Gammell: Next week, yes. We could start with that.

Dave Jones: We have got Oivind.

Chris Gammell: Oivind. Oivind. We hope we're saying that right.

Dave Jones: We do hope we're saying it because it's got funny characters in it. Yes. You know those...

Chris Gammell: I don't even know what that is.

Dave Jones: Non-ASCII. I don't even know what the character's called. But apparently, it's a very common first name in Norway.

Chris Gammell: Right.

Dave Jones: And we have Oivind. He's the co-founder and CTO of Energy Micro on the show next week. So please get your questions in.

Chris Gammell: Yeah. We'll have a separate post on the website. So if you want to pop over there and drop a question, we will try our best to get that in front of Oivind. Yes. Should be interesting. They're a fabulous semiconductor company. They are. If you haven't heard of them. Dave has done reviews of the Gecko board. Is that right, Dave?

Dave Jones: The Gecko. Yes. The Energy Micro Gecko. Micro controller.

Chris Gammell: So we're going to talk to them just about, or talk to Oivind about, you know, what's it like to be a fabulous semiconductor company? And a startup one at that. Yeah, exactly. Yeah.

Dave Jones: How much cash does it require, I guess? Right. Yeah. What effort's required? Yeah.

Chris Gammell: How do you go to TSMC and just say, hmm, can I make a chip now? Or whoever they're using. I don't know. Yeah. I don't know.

Dave Jones: But it'd be interesting. And he's been in several startups. So he should know about all that sort of stuff. Yeah. That'd be great. So please get your questions in. We love having CTOs on. Yep.

Chris Gammell: Speaking of the future, I will be talking at Maker Faire. I just found out this week. And so if anyone's going to be out there, I'll be giving a talk. It probably won't be that good, but I'll be out there anyways. A talk on what? It's stuff. It's in there. You've got it all clued up, right? It's submitted, but it's not written yet.

Dave Jones: So were you invited to give a talk or did you?

Chris Gammell: No, I applied. Right.

Dave Jones: You applied. Okay.

Chris Gammell: Cool. Yeah. The overarching topic is going to be really boring. Documentation. I'm going to challenge myself to see if I can make that somehow.

Dave Jones: Interesting. Well, that's a challenge.

Chris Gammell: Yeah. It will be.

Dave Jones: It's an interesting topic though. Documentation is important.

Chris Gammell: Well, they must have thought so. So we'll see if I can do any good with that, but that'll be in two weeks.

Dave Jones: It can make or break a company. Like a company like Artifruit, for example, have made their rep on their personal service and the quality of their project documentation and stuff like that. That's a great point. So, you know, there are-

Chris Gammell: We could just spend the rest of the show just to help me write my speech if you want to. That's it. No one has to go. It makes it a lot easier. I can just sit on stage drinking beer. Right. You guys heard this already. You can just press play and, you know, it's all done. Yeah, exactly. Yeah. Yeah. And here's me and Dave presenting on documentation. Hope you enjoy.

Dave Jones: So how does that work? Is there like an auditorium there where like they give speech, like where there's presentations all day? Like I know they have the fair with the stands and everything. Yeah. But how do these- There's like stages. A stage. Oh, right. Yeah.

Chris Gammell: So there's like a whole bunch of those and I'm not sure where I'm going to be. Okay. I'm very late on details right now, but- Right. Apparently I will be there and I know it's going out there, but I'm getting excited about it, so.

Dave Jones: Excellent.

Chris Gammell: I'll be hanging out with my DM. You know, me and Jeff will be hanging out, so. Yep. It's going to be great. I'd love to do some jealous.

Dave Jones: I'd love to do something like that.

Chris Gammell: You could have flown over, man. You know you could have.

Dave Jones: Oh, yeah. Right.

Chris Gammell: Maybe next year. Yeah, maybe. I think that might be the final time when me and Dave finally shake hands and start arguing with each other.

Dave Jones: And the world implodes, yeah. Yeah. Oh, boy. Well, that sounds like fun.

Chris Gammell: Yeah. Enjoy it. It should be. Yeah. It should be.

Dave Jones: Are you going to record it? Are you going to have a video set up?

Chris Gammell: Oh, probably. I guess so.

Speaker ?: Yeah.

Dave Jones: Please do. I think they might do stuff. Upload it to your YouTube channel.

Chris Gammell: Yeah. Yeah, I'll try to. I'll try to. But I got the usual. I'm actually sticking around. I'm going to be there early. I'll go to the Computer History Museum finally. Ah, awesome. Surplus shopping, you know. Put it around the area.

Dave Jones: Isn't Jim Williams' thing closed by then?

Chris Gammell: Yeah, that's going to be good. I think that's going back to LT. Bummer. There's a lot of other cool stuff there. Right. Cool. Very much looking forward to it. What about last week? We mentioned the incubator last week. I didn't really hear from anyone, but we should mention that again real quick. Yeah. The Lemnos Labs. So, if people didn't hear last week, there is, I think there's one week left on the submission. Lemnos Labs is doing a hardware accelerator program. So, if you've got an idea for a company, you want to go get some funding, get some support, be kind of grouped in with other people, that's, you know, the kind of whole accelerator idea. Or the incubator, rather.

Dave Jones: Yeah. It's all the rage, isn't it? Hardware is. Doing your own startup. Not only hardware, but software. Doing your own startup and these incubator programs. And these, it's like every man and his dog. I personally know, like, three or four people doing. Yeah. That have, you know, joined a startup, quit their job, whatever, you know. And they're having a go at it. It's just, and that's just here in Australia. So.

Chris Gammell: That's a heck of an undertaking. So, I'm always impressed with people who can do that.

Dave Jones: Yeah, same here. Because it is a lot of risk. Because most of them fail. You know, if you just look at the odds, you wouldn't do it. You know, because if you just look at sheer statistics.

Chris Gammell: Yeah, I think it's a calculated risk in a lot of people. Because I think a lot of people look at it and say, well, it's six months or however long. You know, it's a good idea to set that and say, I've got six months. And the nice thing about engineering, at least. I mean, if you were like a, you know, if you were writing plays or whatever and you wanted to go say, all right, I'm going to quit my copy editing job and go try and write a play for six months.

Dave Jones: Right.

Chris Gammell: That's a tough market to get back into, I'm sure. With engineering, I mean, if you're in a decent job market, and I think right now the job market is pretty decent. Maybe I'm wrong. I mean, at least in my area.

Dave Jones: It's not bad.

Chris Gammell: It's certainly been a lot worse. That's for sure. Right. Especially the past couple of years. And, yeah. So, there's a lot less risk in that regard. You know, six months is still a big chunk of income to take out of your pocket. But if you've saved up and you're willing to take that risk, the upside could be really awesome. It could be. And you could meet a lot of cool people. Yeah. I mean.

Dave Jones: And if you're going to waste someone else's money, all the better, right? Well, yeah. Instead of someone else's coin.

Chris Gammell: Yeah. I mean, well, you never seem like you're a big fan of taking money, though, venture funding.

Dave Jones: No, I am not a big fan of taking money. I'm a big fan of doing it on your own.

Chris Gammell: Right.

Dave Jones: But then again, I'm a big fan of bootstrapping. I think doing it on your own and putting all your money on the line, you know, is just dumb. I think that's just very foolish. Very foolish indeed. Yeah. Yeah. You know, especially in hardware and software, like software, all you need is time, right? Time and talent. The two T's. You don't need money, really. Well, you have to eat, right? But that's about it. And hardware, you know, unless you're doing something massively expensive, you know, you're only talking thousands. You aren't talking, you know, tens of thousands or hundreds of thousands of dollars. Right. So, you know, it's not a huge risk. And it's certainly possible to save up the money to do something like that. It's possible to work for, you know, three or four years, you know, if you're out of college or something, work for three or four years and then save up the money. You can easily save up, I don't know, you know, 50 grand or something. That'll keep you going for a year or whatever, you know.

Chris Gammell: Dave's on the ramen diet, yeah.

Speaker ?: Right.

Chris Gammell: Yeah, no, that's an important time, though. And I think a lot of people use that time because, you know, it's a low cost of living. You're used to that. If you're in college, you're used to that, you know, low cost of living. And you can, you know, you go from making nothing to making 50 grand or whatever people are making these days. Yep. If you're usually spending 20 grand a year because you're a college student, yeah, why not? You know, save that other 30 grand and use it for something cool or even, you know, if you're into buying houses or whatever else you want to do with it.

Dave Jones: Yep. Save up that screw you money, you know.

Chris Gammell: The screw you money. There you go. Yeah. Right.

Dave Jones: Screw you. I'm going to take six months off and work on my startup, you know.

Chris Gammell: Right.

Dave Jones: But as I said, you know, the best, as we've talked about before, the best ones or the best and the safest ones, you start up in your garage and you work on it part time. And then once you know it's going somewhere, then you quit your job and you, you know, it's a bit foolish to sort of have an idea and then just quit your job and go, right, I'm going to work on my idea. Yeah.

Chris Gammell: Yeah. Yeah.

Dave Jones: And you haven't actually done any background stuff to see if it's A, viable and, you know, and B, you actually like doing it and it's going in the right direction. Yeah.

Chris Gammell: You know, a good example of that, I saw that the evil mad science labs folks, Lenore and Wendell, they just made the plunge themselves. So, I don't know if they listened, but congrats to them. It's, you know, it's like they're doing kit business stuff.

Dave Jones: Yeah. I thought they were already full time at that.

Chris Gammell: I had thought so too, but.

Dave Jones: Okay. No, they were doing it part time from the garage. Yep.

Chris Gammell: Oh, they have a shop, but yeah.

Dave Jones: Oh, right. Okay. Yeah.

Chris Gammell: Yeah. It's cool. That's, that's got to be exciting, you know. So, I hope to do that myself someday.

Dave Jones: Yep.

Chris Gammell: We'll see.

Dave Jones: I can highly recommend it. And please support people who do stuff like this because that could be you one day.

Chris Gammell: Exactly.

Dave Jones: So, if you support them, then they'll support you when you go into your venture and, you know, it's one big happy community.

Chris Gammell: Look at that. Dave. Dave, the old softy. I like it. Getting soft in his old age. I am indeed.

Dave Jones: Oh, boy.

Chris Gammell: Well, speaking of spreading the love, we have a ad hoc workbench of the week. We haven't done that in God knows how long. True. But our buddy, Alan, he did a lab tour in a very different kind of manner.

Dave Jones: As in Alan from Tektronix, not Alan Yates.

Dave Jones: Yes. Yeah, sorry. Who we've had on the show.

Chris Gammell: Alan Wolk. Yeah. From Tektronix. Or Alan at Tech on Twitter. But he did a cool video. And he's got a great YouTube channel as well. He does a lot of scope stuff and a lot of fun stuff like that. And he did his tour of his shop through the screen of his old tech scope.

Dave Jones: That's all I know. It's great.

Chris Gammell: Yeah. So he set up a circuit to do horizontal and vertical sync. And then basically he put the audio or the video out from his camera, you know, just a regular video out. And then he synced it up with this circuit. And then he actually was able to display the video. Yeah.

Dave Jones: Using the Z-axis input for the modulation, for the intensity modulation, which not all scopes have. Not all scopes have a Z-axis input.

Chris Gammell: Right.

Dave Jones: So there you go.

Chris Gammell: Yeah, it was really cool. I mean, and then he kind of walked around the shop and you get to see all his gear. And obviously he's a collector of old tech stuff, which is fun.

Dave Jones: And he did constant current sources using transistors. And, you know, it was really quite an elegant, simple circuit.

Chris Gammell: Yeah.

Dave Jones: But he did use an off-the-shelf, like, you know, a national semiconductor sync circuit, which are getting... I think they're getting harder to get these days. Might these things be dying out? Because nothing's, you know, composite video anymore. Right.

Chris Gammell: They're all moving to HDMI or even just DVI stuff, right? So it's tougher to find. I like it.

Dave Jones: But yeah, that was good. You know, it's not a new concept. People have been displaying video on their scopes for donkey's years. But it's always fun to see. I love to watch.

Chris Gammell: And showing a lab tour. And showing a lab tour. Yeah. Yeah, that's great.

Dave Jones: Well done, Alan.

Chris Gammell: All right. Wish I'd thought of it. Yeah. Well, next time, buddy. Yeah.

Dave Jones: I'm not even sure if my Tektronix 2225 has a Z-axis input. I can't remember.

Chris Gammell: Yeah. I don't remember what model he was using.

Dave Jones: It was a Tekt 485. It was an old 485, I believe. Okay. The old workhorse.

Chris Gammell: Yeah. I've got a 425 that's kind of sitting on my bench. It's not doing much, but... Is that a Valve one? Is that a what?

Dave Jones: A Valve one. Oh, no. No? It's Solid State. No, it's a Solid State. Okay. Yeah.

Chris Gammell: It's the military model, though. So I was gifted it, and I haven't really played with it much, but... Cool. Yep. It's fun. I like it. Speaking of playing with stuff. Yep. Yeah. Yeah. What's up? Speaking of playing with stuff.

Dave Jones: I finally got my freaking MakerBot working.

Chris Gammell: You did? Oh, my. How many months has that been?

Dave Jones: I know. A long time. And yes, people will say, I told you so, I told you so. And I never got around to checking it, because you had to take the damn thing apart. But it was the problem. The main problem was the motor current. I didn't set it up correctly. I didn't follow the instructions. Turns out I had a different brand of motor, which requires three times the current of the MakerBot brand motor. And... Did you buy it separately, or did it just get kitted that way? Actually, no. That's how the kit came. But apparently, there's different types of kits or something. I don't know. So there's two instructions.

Chris Gammell: The importance of documentation right there, right? Yes, exactly.

Dave Jones: And it was there. Granted, I should have checked. I should have read it. Yeah, it was there. When you go back, it was obvious. Yeah. But the thing is, it worked, right? It almost worked. Yeah, it started to work. At one third of the current. Like, give me a break, right? If it didn't work at all...

Chris Gammell: That's the worst for troubleshooting, right? When you have, like, an indicator. You know, you expect, like, a very discreet on or off kind of behavior.

Dave Jones: On or off, fail, or work or fail, right? And that's what I expected.

Speaker ?: Right, right?

Chris Gammell: Those marginal cases is terrible. It's terrible for finding stuff like that.

Dave Jones: Yep. I know. So Murphy's Law, every time. Getcha. Oh, yeah.

Chris Gammell: Oh, yeah.

Dave Jones: So, you know, I thought I had a working maker bot because the motor moved around and everything, you know, it made the right noise and it squirted out the stuff. And, you know, we know. And everyone was throwing all sorts of things at me. Oh, you've got to print directly from the SD card because it can't stream from the USB. It loses the buffer sync and, you know, buffer underruns and screw things up. And, yeah, and other people said, oh, you know, your temperatures are all out and all this sort of shit. Oh, your belts are all slipping and, you know, different alignments. And, oh, it was... So everyone's throwing all this info at me, you know.

Chris Gammell: Yeah, I was going to say, did you go through your usual troubleshooting, thou shalt check the power supplies and all that other stuff too?

Dave Jones: Thou shalt check the power supplies? Well, I checked the voltages for the motor drive currents, you know, and they were, as per the instructions, I just happened to be using the wrong instructions.

Chris Gammell: Aha. Okay.

Dave Jones: So I did thou shalt check voltages. Yes, I did.

Chris Gammell: Yeah.

Dave Jones: And it all looked good. And it all almost worked, you know. So I thought, okay, so it made sense that, oh, okay, there could be like a little slippage on the belt or something like that. So that made perfect sense. That's the problem with troubleshooting.

Chris Gammell: You get in these modes and you kind of, your brain constructs, oh, it could be this. You know, you hear something, you get the seed of an idea and you're like, oh, what if it's this, this, this, and this? And you go down this whole path and then finally get to the end, you're like, oh no, that's not it at all.

Dave Jones: No, it was the bleeding obvious. I should have gone back and, you know.

Chris Gammell: Oh, that's rough, man.

Dave Jones: I know. It sucks. Anyway.

Chris Gammell: Well, at least it works now. That's good.

Dave Jones: It works. And it prints. And the first thing I tried to print was a little camera microscope, camera adapter for my stereo microscope, right? So I can put a video camera over the thing. So I measured it all up with my calipers. I put it into the 3D CAD software and I generated the model and, you know, it was very simple. It was just like a two-stepped cylinder, you know, with two different internal outer diameters. There were no overhangs or anything, you know, real dead easy to print. And I finally got the thing to print and it printed perfectly, except it won't print without the raft on the bottom, which it lays down, you know, it lays down a strip of plastic first and then builds on top of that. And you can't separate the bloody thing. It's impossible. And everyone says, oh, don't use a raft. And well, I don't use a raft and it doesn't work. And now everyone's saying, oh, no, that's because you haven't cleaned your capped on tape before you started. Well, okay. Learn something new again, right? There's all these little tricks to getting these 3D bloody printers working, right?

Chris Gammell: So if you were flying in the dark where you didn't have like a forum or Twitter or anything like that, I mean, what would happen?

Dave Jones: Oh, you'd be trial and error, you know, forever. Yeah.

Chris Gammell: It would just sit there basically?

Dave Jones: Yeah. I'd sit there and scratch your head and go up 10 different blind alleys before you, you know, you'd eventually crap it. Crap it? Crack it. You'd eventually crack it, I guess, but.

Chris Gammell: Because it reminds me like how you said you grew up doing electronics like that where you didn't have, I mean, I guess you had like manuals and stuff like that, but it wasn't the same.

Dave Jones: You only had the books in your library. That was it. You didn't have anyone to talk to. I didn't have any friends who did electronics. I didn't know anyone who did electronics. And before I started work, I didn't, you know, have anyone. Yeah. I was completely on my own. I had my books and my parts and my multimeter and that was it.

Chris Gammell: So does that mean you like were ripping stuff up more often? I mean, like, you know, because like usually when I'm doing a circuit, like let's say I'm just like breadboarding something. I'll put down parts or whatever and I'll try this and that. Is it, was it more of a, ah, screw this and you rip everything out and you start over? Was that more of the, the, the MO of working like that where you didn't have any, you know, troubleshooting?

Dave Jones: Sometimes you did that, but often you'd just try and brute force it and you'd, you know, you'd keep tweaking things. Switch out this part. Yeah. You switch out this part until you, until you figure out what the hell's happening and you build up a broad picture of what's, what's wrong or whatever and why your circuit didn't work.

Chris Gammell: And yep. Yeah. Yeah. And you know, and there's value in that too, because that's, I mean, really that's, I think about like labs that I've been in and just even doing stuff at home and that's when you really learn it, right? I mean, that's. You really learn. You really, you have to, because you can't just be like, oh, I'll just tweak this value. It's like, no, I need to know what this value does first.

Dave Jones: And there were no simulators back then, right? Or there were, but geez, you know. Yeah. You didn't have access to it. I didn't even have a computer until, you know, late in the game, let alone a simulator. Yeah. So yeah, you, you build up and trial and error and lots of smoke and fire and, and lots of circuits which just do nothing and you scratch your head and, you know, and, and you do learn a lot when you start substituting components and troubleshooting. And as I said, I, I hope your next project doesn't work and you have to troubleshoot the

Speaker ?: thing.

Dave Jones: Yeah. Cause if it works, like if that maker bot for worked first go, I wouldn't have learned much. Right.

Chris Gammell: Right.

Dave Jones: But now I've learned about motor drive currents for them, built slippage, backlash. I've learned about, you know, printing from SD cards is better. Cleaning, you capped on tape and doing this and setting this and, you know, unbelievable.

Chris Gammell: So if you, if you just ordered an off the shelf working printer, like the other ones that are out there and you think, what happens then?

Dave Jones: I wouldn't have appreciated it as much. I guess I wouldn't have appreciated the, you know, the workings of it. Um, but I guess I, maybe I wouldn't have cared if these products just worked, if these 3d print, if you could buy a thousand dollars 3d printer that just worked, you know, then

Chris Gammell: I wonder about that, you know, like, because there is value in it, but at the same time, you know, you hate, you hate the, that frustrating feeling that you almost have to be primed for it. You have to be like, all right, I know this is going to be frustrating and I know it's not going to work. Right. Exactly. And I'm going to, I'm using this to learn and there's a lot of value in that, but sometimes you also just want that other one right next to it and be like, no, I'm just going to do this, right?

Dave Jones: Yeah. I do this because I want my case or my front panel or my widget that I'm printing. I just want it now and I want it perfect. And anyway, my microscope, I printed it, came out very nice except the raft stuck to the bottom, but it was nice and round. And, you know, I was quite, I didn't think it could, you know, do a better result than that. I was quite impressed. And I go to put it on my microscope and my camera and wah, doesn't fit.

Chris Gammell: Yeah.

Dave Jones: You know, it's-

Chris Gammell: Bad measuring, measure twice.

Dave Jones: No, I measured perfectly. And I even allowed half of, I thought, well, one millimetres might be a bit much. I allowed half a millimetre bigger. So if I wanted a 42 millimetre diameter in a hole, I made it 42.5 in my software, right?

Chris Gammell: Yeah.

Dave Jones: So I catered for that, knowing that, you know, it might turn out a bit smaller, but it turns out like two millimetres smaller than what- So if I program in 42 millimetres into my CAD software, I print the thing and measure it and it's 40 millimetres inside diameter. Oh, geez. And it's like, that's significant. That's very significant.

Chris Gammell: That's a pretty big calibration constant in there. Yeah.

Dave Jones: And like, why can't that be in the CAD software? Like, how much expansion do you want? Or it should automatically know and add that fudge factor in. Like, I shouldn't be tweaking. I shouldn't have to know that and then tweak my CAD drawing to do that. It should be the other way around. I should do the CAD drawing exactly the dimensions I want and then be able to print. And the printer should know that, okay, it's going to ooze out. You know, the plastic's going to ooze out a bit and, you know, circles are going to be a smaller diameter than what you've programmed in. So maybe it is in there and I'm missing it. I'm missing the option, but I can't see it. So I don't know. And is that unique to the MakerBot or does that happen on other low-cost 3D printers? I'm sure the really high-end, you know, your $10,000, $20,000 ones don't do that. I'm sure they're very precise. I use them and no, they don't. Yep.

Chris Gammell: No.

Dave Jones: And then you get into things like overhangs and stuff like that. You know, you are limited to what you can print with these things as opposed to the big commercial ones. They're much better because they'll print with different materials and then you drop them in water and all the stuff that holds it up will dissolve away and, you know. And yeah, very nice.

Chris Gammell: Yeah, it's awesome. I love that.

Dave Jones: But these low-end ones, you know, you're quite limited.

Chris Gammell: Hey, man, you're using a CPLD where you need like a Vertex 6 FPGA, you know.

Dave Jones: Right.

Chris Gammell: And that's what you're asking for, you know. And there is. There's the same thing happening in electronics too, you know. Like sometimes you're just going to, if you're using a 12-bit ADD, right, when you need a 20-bit one, it's like, well, what do you expect, right?

Dave Jones: Yeah, well, exactly.

Chris Gammell: But sometimes you don't know you need that, right? You don't know what you need until you actually start using it, right?

Dave Jones: Well, that's the thing. I thought this thing would be more useful than what it currently is, you know. And you start using them and you figure out the limitations and, well, you know, these low-end ones, I don't think they're all they're cracked up to be. They're useful, you know. If you just want to do a flat front panel or something like that, nice, you know. There's no reason why it can't do that. Just watch out your hole diameters, huh? You know, you try and shove your pots in your front panel and you find that ain't bloody fit.

Chris Gammell: Yeah.

Dave Jones: You know, and jeez, you may as well got out a plate and drilled it yourself, you know. Right, right. I don't know. Anyway, yeah, 3D printers, they need to get a lot better at this lower price point. And I'm sure they will. But what's involved in actually getting, you know, good rock-solid performance into a $1,000 printer? Is it even possible to get that? Or are there physical limits to, you know, how cheap you can make all of the precision gears and things like that?

Chris Gammell: Right.

Dave Jones: Hmm. Well, I don't know.

Chris Gammell: I don't have one.

Dave Jones: I hope not. I really hope. Please, if you've got like a, you know, a cheap low-end, a different brand, a 3D printer that actually works. Yeah. Then please let us know.

Chris Gammell: Speaking of low-end printers, there was a thing on Hackaday a while back and it was on home, it was used an old Epson printer, apparently, because they use, I think we've maybe talked about that on here before or something else, but this was like really slick because he actually used like the sheet feeder as like feeding this entire sheet of copper, you know, copper-clad PCB into this thing. He prints the resist on there. And I was thinking between that and like, you know, like maybe a, you know, a timer or an agitator, right, where you're using like a chemical bath, you could do a pretty decent home job, but I'm not sure, you know, like you can make a decent board at home and, but I'm not sure, it might be the same thing where you have, you might just run into it and say, oh, you get all set up, you get all the way through it. You're like, oh, this is, this is just okay.

Dave Jones: I've been there. I've done that. I've tried to make, you know, I used to make all my boards at home or most of them. And it's, it's once again, there's lots of limitations. Often I will have to change my design knowing that I'm going to home make this thing because I can't do the plated through vias, right?

Chris Gammell: Yeah.

Dave Jones: And if those plated through vias are, you know, under SMD chips or something like that, right, that's a pain in the ass.

Chris Gammell: If you're doing SMD chips, you're probably not doing it at home, right?

Dave Jones: Oh no, I've, I've, I've done SMD ones at home. No problems. I can get down to 7th hour, 7th hour at home.

Chris Gammell: Really? Yeah.

Dave Jones: That's not bad. I know. It's certainly possible. But, you know, and then you'll get, you know, over etched or, you know, overexposed or something like that. It is tricky. Yeah.

Chris Gammell: And this isn't with exposure, but yeah. So you always did the exposure method where you're actually. Yes.

Dave Jones: Yeah. Yeah. The photo exposure. Yep.

Chris Gammell: Right. Oh, that's interesting too, because they, you know, that's actually what they have to do for, for chips for when they're doing photolithography. They have to make those same kind of corrections. Mostly because they've been stuck at the 193 nanometer level. So if people don't know when you're, when you're doing photolithography, you know, there, there's a certain wavelength of light. And if you don't go any small on the wavelength of light, you really start to limit yourself on resolution. Well, they were saying that for years and years and years. And right now they're what, I think Intel just pushed a 20 nanometer node. Yeah. I'll talk about that in a second, but you know, like they're, they're on 20 nanometers and I think they finally had to switch over to EV or 13 nanometer light.

Dave Jones: I was going to say, what, what, what light are they using to do that?

Chris Gammell: I think with 20, I think they finally had to do it, but it's, it's, they've been stalling it for years and years and years. So there's all these tricks they do. If you actually look at a mask set for, for a chip, you know, they'll actually, they'll, they'll have these really high paid consultants that'll, that'll actually, and a lot of software too that'll.

Dave Jones: I was going to say the software would automatically, they would tweak their software to automatically compensate for that, for their process. Right.

Chris Gammell: Right. Yeah. And, and, and what they do is what ends up happening is, you know, you'll, you, when you're like turning a corner, say you're doing like a trace and you're turning a corner, they'll actually set up these interference patterns so that the interference patterns will actually, when you're finally doing that exposure, it'll actually create the feature you want, which is, it's really crazy. Nice. You know, it's like, it's like slotted lines and stuff like that. And, and actually one of, one of the writers over at engineer blogs wrote about it at one point. I think that's how I learned some of it. So I'll try and, I'll try and link that in, but yeah, it's, it's just, you know, crazy stuff you'd never think of, but it's all because it's the same thing. It's that same limitations of, in their case, it's the, you know, the, the, the light and, and for you it's, you know, if you wanted to do overhangs or 3d printing, or you wanted to try and get some better resolution, you'd have to do fanciness. I'm sure in the, in the CAD model as well, or, or same thing with the, the home PCB etching. You have to do these, these little tweaks. You have to compensate. Yeah. And, and really that's, I mean, that's a lot of what manufacturing engineering or process engineering can be sometimes is knowing those tricks and applying them and, and really taking advantage of these quirks and, and just trying to eke out that extra 10% or 5% or whatever you need in order to, to get, to get better, you know.

Dave Jones: But in an ideal world, you shouldn't have to, you know, you shouldn't have to be and have to tweak your source stuff to match your process. You know, your, your, your source stuff should be exactly what you want. And then your process should be able to do it.

Chris Gammell: You know? Yeah. Yeah. That, that is an ideal world. You're, you are correct, Dave.

Dave Jones: And, and that is what the, say the consumer world expect. So if you want these 3d printers to take off in the consumer world, right. You know, everyone has, you know, every mom and dad has one to print, you know, little, little Johnny's toys, right. Then, you know, it needs to be that dumb ass.

Chris Gammell: Yeah. Yeah. Maybe. And, and, but maybe that would be a firmware update on, on one of the, like in that case, it might be a firmware update. So that's, it's blind to the user. Yeah. They need to get a bit smarter. Right. Yeah. So, so yeah, that Intel thing I was talking about though, I think we've, we've mentioned on here before Intel is on 20 nanometers right now, which is an entire process node. So their chips are entire process node above anyone else. Yep. That's unbelievable. I mean, like the amount of money it costs to do that. So, I mean, like you were saying, oh, well it should be in an ideal world. It should be, the process should match reality. Right. Yeah. And the reason they don't is because of money. That's exactly. Well, yeah. Because money.

Dave Jones: And they're on the bleeding freaking edge, you know.

Chris Gammell: Well, yeah. And the money allows them to be in the, the bleeding freaking edge. And, and yeah, it's, it's, it's just unbelievable because now they're, they're already into FinFets and all the other crazy 3D, 3D based transistors and everything. And, and that's how they're even holding off ARM. You know, you wonder, you see ARM everywhere for all these microprocessors. You're like, well, how is Intel still around? Because the desktop business isn't doing that great. But I think eventually, and I think I've seen this a couple other places too, Intel's eventually just going to be like, well, our technology is great. Our processor is doing okay. We'll keep making those. But, you know, they've already started talking about getting into the foundry business where other people can come in and start printing stuff there.

Dave Jones: And they're ideally placed to do that as well. Yeah. Because they, they have the world's best fabs, right? Bar none. Right.

Chris Gammell: Oh yeah. Yeah. Yeah. Hearing, hearing, when I, when I used to be in the semiconductor industry, hearing, hearing, hearing the vendors talk about Intel was, there was usually fear in their voice. Yeah. Right. Just because of the demands they made. And, you know, any big company makes those kind of big demands and they had to capitulate. But yeah, apparently Intel was one of the, was one of the top. Nice. So, yeah.

Dave Jones: Speaking of printers. Yes. I, this is rather interesting. There's a company in Australia I may have mentioned before, Silverbrook.

Chris Gammell: Silverbrook.

Dave Jones: Research. Silverbrook. They're a, a run by a guy called Kia Silverbrook, who founded it, you know, 15 years ago or something. And he is the world's largest patent holder of useless patents. Yes. He is the world's largest by double. He is like twice, he's got like four and a half thousand patents. Right. And that's his goal in life is to have the most patents. And, you know, he, and he does by, by two times. That Japanese guy, I forget his name. But anyway, we'll link in the, a, a link to Silverbrook. And what happened, right? Mysterious people, mysterious companies have been funding hundreds of millions of dollars into Silverbrook's company, Silverbrook Research over the last decade or more. And they haven't produced anything to date. Really? They're into, you know, they're into MEMS printer, MEMS, you know, a new revolutionary printing technology that prints out, you know, a page, you know, 10 pages per second or something, you know, at high volume, low cost, you know. And it is marvelous technology.

Chris Gammell: Wait, like you're talking about like actual paper printing?

Dave Jones: Actually paper printing. Yeah.

Chris Gammell: People are investing millions of dollars in paper printing. Oh, yes. It's still a big market. Perhaps they should go talk to Kodak. I mean, it's, Xerox. They're not, they're not doing so hard right now.

Dave Jones: Anyway, they, they do lots of stuff and lots of other stuff as well. And, um, the latest news report, I'll have to link it in. Um, and what's actually about two or three weeks old. I just found it the other day and tweeted it. They have like 300 employees in Australia. So they claim in 200 of them are PhDs, you know, over 200. Like, no, sorry. Over 70% of their employees are PhDs, right? So they've sucked up every PhD in the country, literally, right? And they're into, you know, into, uh, polymers, uh, polymer research and micro machine research and, you know, all sorts of avenues of technology. So they've sucked them all up. But anyway, the interesting story is, is that there's this guy in the US and I forget his name cause I don't have it in front of me. Um, but he's been put in, he's put in $300 million of his own personal money over the last 10 years into Silverbrook research. And, uh, all this info has just come out. It's just been made public. And he has finally had, you know, he has had jack of it cause they haven't produced anything. You see no return on his money. And he started to sue them, um, and, and actually withheld it and withdrew his money and support. And what happened is, um, Silverbrook had to, uh, put all of their employees on a permanent leave or, or actually sack them all. Depends which report you read. And they're going to form another company to avoid the lawsuit and then rehire everyone.

Chris Gammell: Oh, geez.

Dave Jones: I know, really messy business, right? Really.

Chris Gammell: That's what, that's what you get when you put money in Australia. That's all I'm saying.

Speaker ?: Right.

Dave Jones: Anyway, um, yeah, I don't blame him. You know, he's seen no return on his investment for 10 plus years and Silverbrook are famous for not having actually produced anything after all this time. You know, great place to work. They pay huge money to all these PhDs who wouldn't be able to get any jobs anywhere else. Right. Where else can you earn hundreds of thousands of dollars doing pure research? You know, you can't.

Chris Gammell: So, um, yeah.

Dave Jones: And so apparently it's a dream job for them. Um, but yeah. Right.

Chris Gammell: But you don't have any output.

Dave Jones: They have no output. Um, they have finally, uh, demoed something at the CES show or something like that, but still, you know. Right. And, uh, anyway, I just thought it was an interesting story about how, you know, they haven't produced anything. Their backers pulled out and they had to sack everyone and, you know, and, oh, nasty. But.

Chris Gammell: See, now that's crazy because they, they've got millions of dollars, right? They're working on this, this MEMS printing thing. And then I see this thing where this guy takes a laser scribed DVD drive, right? And he starts printing. He has, he has graphene discs and he just prints super caps onto them with this commercial printer, right? Nice. Yeah. Yeah. It's just, so, so if people haven't seen this, it's, it's basically, you know, laser scribe is the thing where you can etch, you know, like if you burn a CD, so you have a CD, a writable CD, and then you want to like write the name of it on top. And for some reason, you don't, you have a Sharpie around or whatever you want to use this fancy, this fancy printer. Basically it's a, it's like a laser cutter that basically scribes into the top of the DVD so you can make this nice looking cover. Well, this guy, he doesn't have millions of dollars. He has this laser scribe thing. He puts a carbon disc on top of a normal CD and then he actually uses this laser scribe and that, he, he prints a super cap using that, right? Awesome. Yeah. It's really cool. Well, I'll, I'll link the video in, but it's, it's, it's insane. You can get these really precise patterns. And I think that it's to the point where even like, you know, listeners could do this kind of thing at home. Yeah.

Dave Jones: So because they really are advanced. The optics and the drive systems inside CD writers are incredibly advanced.

Chris Gammell: Right. And so the, I guess the larger question is, I mean, do you need the hundreds, hundreds of millions? I mean, I guess certain things you do need the hundreds of millions for, right? But I guess when they're misplaced like that, good Lord. I mean, that's a crazy story that they're blowing money like that.

Dave Jones: I mean, yeah. Yeah. The, I guess the only downside to using that is they're based on a spindle system. You know, it has to be something that's been, you know, if it was XY based, oh, imagine the possibilities.

Chris Gammell: Yeah. You know? That's true. Wow.

Dave Jones: If you have this laser head on an X, on a real precise XY bed.

Chris Gammell: Yeah.

Dave Jones: Woohoo.

Chris Gammell: Right.

Dave Jones: You know? So that's neat.

Chris Gammell: Yeah. Yeah. That's really cool.

Dave Jones: Anyway, the followup to the story, which just, the Silverbrook story, which just came out yesterday was that apparently they might have resolved all the issues. They might've appeased in some way, secret deals went down and I don't know. So the employees might be back, but yeah. I don't know. Does anyone else know of a similar story where, you know?

Chris Gammell: Yeah, you do? Yeah. EE store. Do you remember those guys?

Dave Jones: Oh, right. Yeah. You've talked about them before.

Chris Gammell: Those guys are full of crap. I mean, it's the same. It's actually a super cap thing. It kind of ties these two together. It's a, it's a company that was in Austin. I was really interested in them when I was living in Austin. I wanted to go work for them because I knew they were hiring process engineers when I was doing that kind of thing. Yep. And super secretive. No one knew what was going on. No one, it was a private company. Yep. And they said for years, oh, well, we have this amazing product. It'll be great for electric cars. And they just kept taking people's money basically. You know, it's the same kind of thing.

Dave Jones: That's exactly the same thing. Yep.

Chris Gammell: Yeah. I mean, it's the oldest. I mean, and that's, that's the problem when you take, you know, that's the flip side. You know, like sometimes when you're, when you're starting companies, like we talked about at the beginning of the show, sometimes you're going to, you're going to be in a tough situation because, you know, like investors are going to want to dupe you. Right. Yeah. But I think a lot of other times, you know, companies are out there trying to dupe investors, right?

Dave Jones: Dupe investors. Yeah, exactly.

Chris Gammell: That was a lot of concern with that, that jobs act that was in the U S where the, you know, the, you know, now, now that jobs act passed in the U S we talked about on here. Yep. You can, you can invest up to like 10,000 bucks into small companies up to a thousand people can do it. So that's a million dollars into any company. Right. Well, the thing they worried about was people, you know, like taking advantage of snake oil salesman. Yeah, exactly. It's basically, it's, it's basically selling nothingness and, and, and taking people in

Dave Jones: and, and, and that, and the problem is that could be the average Joe who doesn't know about technology. Right. Right. They could get, you know, sold on, you know, some over unity bloody energy. Oh yeah. Right. Easily.

Chris Gammell: Right. I mean, take that. I mean, it could just be a glitz and glam presentation, you know, like really, really slick marketing folk that, you know, like just sell the hell out of something and, you know, sometimes people just forget the whole too good to be true thing. I've, I've, I've been taken in before. That's stupid, stupid guy flying video. Right. I got taken, I got taken so hard.

Dave Jones: Ah, you're gullible. Young and gullible. Yep.

Chris Gammell: Yep. Yep. Oh boy.

Dave Jones: So what does it, uh, uh, what was I going to say? I've got a mental block.

Chris Gammell: All right.

Dave Jones: I'll just go bang my head against the desk here and, uh, think about what I was actually going to say. Cause my mind wanders. It happens.

Chris Gammell: All right. Well, let's turn, let's turn the page completely and talk about something totally different. Uh, I wanted to call out a DigiKey. Have you seen, you know, we've talked about it on here before about how I wanted to have like a thing that was like a visual catalog and, and, you know, being able to look at connectors and flip through a catalog and being able to see stuff. DigiKey did that. Are they listening or is that like their new thing? I mean, they, I don't think so. Uh, but yeah, DigiKey has a new catalog. Like, so, I mean, a lot of people on here are fans of DigiKey, but they have like a graphical based catalog.

Dave Jones: Yep. That's awesome. So it's like flicking. Cause that was one of the advantages of a paper catalog, right? You could flick through. And especially if you've got connectors or something like that, you're looking for a particular type.

Chris Gammell: Right.

Dave Jones: And you can't really do that with a name based search cause you don't really, you know, there's different names and it's hard, but flicking through a visual representation. That's, that's what you want.

Chris Gammell: Right. Yeah. And so it's, I mean, it's awesome.

Dave Jones: It's, uh, it's handy. Very.

Chris Gammell: It's, it's not like super advanced yet cause it doesn't get all the way down to it. Right. I mean, like it doesn't get all the way down to the, and, and, and it only has certain things that it's useful for, right? If you're, if you're looking for a cap, right, you might know you're, you're looking for like a, you know, an 80, 805 cap. And then at a certain point you need to actually say, all right, well, I need like a 10 microfarad cap or whatever, whatever you need, you know, like, um, so at some point it does stop and say, okay, well, let's get to the parameters now. But I don't, I don't think for the stuff where you would need to go visual the whole way, like with a connector, with a cable or something like that. I don't think it goes all the way to the bottom, but, um, you know, but it's still very useful. Yeah. It's a, it's a great start. And I hope everybody else copies it because that would be awesome. That would, you know, like I still get, I still get a lot of paper catalogs. I don't know if you do, but, um, well, yeah, I still do.

Dave Jones: I still have them. Well, uh, not at DigiKey, they've actually stopped printing after my video. They contacted and said, no, we're not going to print the paper catalog anymore. It's gone.

Chris Gammell: Oh yeah. So.

Dave Jones: Yep.

Chris Gammell: Well, I meant like, uh, these are actually specific vendor catalogs. Like I've got a bunch of connectors, vendors that send me catalogs and stuff. And, and like for that, I'm like, well, yeah, I mean, I still want it. So, but maybe this is the end of that too. You know, like it's, I don't know.

Dave Jones: Well, that's, that's what I wanted to talk about with the printing thing. Back to it. Have they chosen? Well, no, cause it's a similar thing, right? You're like papers go in the way of the dodo. Right.

Chris Gammell: Well, we keep saying that. I mean, it's not, it's not there yet. But yeah.

Dave Jones: You know, in general. Right. And maybe it's, you know, and there's no. Right. And there's no money in these things anymore. Like how, like there's no money in making printers. Right. It's all in the consumables. Right. You can buy a printer for 50. You can buy a laser printer for 50 freaking dollars in Australia.

Chris Gammell: Yeah. Right.

Dave Jones: It's crazy. It's insane. And you know, they, that almost has to be cost price. Right. So there's no money in that. So with Silverbrook making these, you know, this new printing technology, well, has it passed, is it already passed its own prior? Right. Before it even comes out.

Chris Gammell: Who would invest in that? Maybe, maybe because they started 10 years ago, it sounded good back then.

Dave Jones: But yeah, it sounded good back then. But now it's like, well, no, sorry. All the models changed. The market's changed. That's the same thing with that storage cap, right? Right. The storage cap has that passed. It's, you know, has the technology moved on? No. The other different types of memory, everyone was working on FRAM and stuff like that, right? Right. Yeah, they still are MRAM, FRAM. And then Flash came along and well, their market suddenly just vanished overnight. No, no, no, no.

Chris Gammell: I think, no, I don't think with Flash, that was a different thing. I mean, Flash, I guess Flash kept growing, if that's what you meant. Like, they kept getting, like, higher and higher densities, but...

Dave Jones: Well, no, back when they originally started, like the FRAM thing, I don't even think Flash was not, even not around or very immature. And then all of a sudden, bang, they figured out how to mass produce them. Oh, I didn't know FRAM has been around that long.

Chris Gammell: Okay.

Dave Jones: No, yeah, because FRAM started in Australia. It was, it went way back, way back to the 90s, I believe.

Chris Gammell: Oh, interesting.

Dave Jones: Oh, yeah, yeah, it's been a long time. They were working on that for a decade.

Chris Gammell: Yeah, they're still working on other types of memories, though, too. That's what I thought you meant. Yeah. Because, I mean, they have FRAM now, which is... Yeah, they do have it. There's a niche for it, you know. There's a niche application for it.

Dave Jones: But it's not the big, ultimate thing that would be in every consumer device. That ended up being Flash and not their technology, not their FRAM technology. But that's what they sold the investors, right? Ah, I see. Is that it was going to be in every mobile phone, every computer, every, you know. And it never happened.

Chris Gammell: Yeah, and there's probably a lot of people that were really wrong about, you know, like, there's a lot of probably, I doubt they were PowerPoint presentations. They're probably transparencies, you know, in the trash bins of the world. Yeah, trans, yeah. Overhead projectors, you know. Right, exactly. And those probably had about the great benefits of FRAM. And, yeah, they were wrong. I mean, they're just, yeah.

Dave Jones: People are wrong every day. Well, not necessarily wrong, but they were passed by. With hindsight, they were right at the time, you know. But they just, you know, it was either harder than they thought or somebody else came along and actually gazumped them or another technology came along and gazumped them. Gazump. Or the market changed like the... Gazump? Gazump? You don't know what gazump is?

Chris Gammell: No.

Dave Jones: It means to get one up on somebody, like to actually beat somebody. You actually come from nowhere and beat them. You outbid them at the last second or, you know, something like that.

Chris Gammell: I imagine like a gazump button on eBay, you know, coming at the last minute. Gazump.

Dave Jones: Yeah, it's like a snipe. That's kind of like a gazumping kind of. It comes from the housing market. Technically, it's something different. And it means that you've placed a deposit on a property and they're not supposed to take a deposit from anyone else, but they do. And they gazump you, you know.

Chris Gammell: Gazump.

Dave Jones: Yeah. And there's actually laws against that in certain respects. Yeah. Yeah. There's actually anti-gazumping.

Chris Gammell: Yeah. There's a bunch of lawmakers. Gazumping has gone on long enough. No more gazumping. I'm stumping on gazumping.

Dave Jones: I bet that happens. Like, who would have foreseen that printers would become, practically give them away, right? Everyone, I'm sure all the printer makers back in the day thought that they'd be selling $1,000 printers, right? And that's what people would pay.

Chris Gammell: I mean, yeah, HP was based. I mean, HP continued their growth path on that, right?

Dave Jones: Yeah. And now it's only a niche market where, you know, a huge, big volume, you know, office printers will cost $10,000, you know?

Chris Gammell: Yeah.

Dave Jones: But, yeah. That's crazy. It just changed. So, you know, you've got to be careful of that if you're doing your startup or whatever. You can get gazumped. You can get to pass by. Anyway, that's why if you spend five years working on something, you know, the longer you spend on it, the more chance that somebody's going to come along and beat you or your market vanishes.

Chris Gammell: Yeah. Yeah. Here's something interesting. Speaking of almost not working, I saw an article on, what the heck is this site? CompoundSemiconductor.net. Basically, it's talking about, like, the doping in an LED. So, we've had John Edmund on this show before. Yep, from Cree. Cree. He was awesome. He was the co-founder of Cree. He taught us all about, like, you know, working with silicon carbide and then the gallium nitride layer on top of that. And basically, gallium nitride and magnesium are the reason that you can have white LEDs these days, right? I mean, there's this whole huge pending revolution, I guess. I mean, I don't know if it's even pending anymore. It's probably...

Dave Jones: Well, it's another one. Here's another one where it could get gazumped, right? Some new technology someone's working on might, hey, be ten times better for one-tenth the price.

Chris Gammell: Right. Who knows? Anyway. People out there get gazumping. I think we found another ridiculous name to go on our title.

Dave Jones: Oh, yes. It might be G. Yep. Oh, yeah.

Chris Gammell: Gazump. But basically, it was... This article talks about the blue emissions. I mean, if people don't remember from the show with John Edmund, LEDs are actually... The white LEDs we all see, they're all blue LEDs with, like, filters over the top.

Dave Jones: Phosphacoded, yeah.

Chris Gammell: Right. And basically, this article is talking about how it almost doesn't happen. Like, the doping within an LED. And granted, like, that's way outside my deal, right? Right. But basically, the doping necessary in an LED is almost not possible. That's what this paper is talking about. Right. And they're just finding this out now. It says it's... I had the quote in here. Magnesium and gallium nitride only accidentally behaves as a shallow acceptor. It's truly lucky that the electrical level lies close enough to the valence band to generate enough holes for p-type conduction. Otherwise, blue LEDs and solid-state lightning would not be possible. That's, like, one of those things where it's just, like... What's wrong with that? There's nothing wrong with that. It's just crazy that, like, you know, like, one tweak in, you know, in universal... Not even universal constants, right? But just, like, it's so on the edge, you know? And now there's an entire industry being based upon it, you know?

Dave Jones: Based around this on-the-edge physical, you know, physics process.

Chris Gammell: Right. And it's not the only thing like this, right? I mean, like, there's other natural phenomenon that's like this. But it's just, when you hear about it, you're like, oh, man. It's like when they're studying about, like, you know, life starting on the planet Earth, right? And, you know, like, all the different things that are necessary. Yeah, sure, it happened like that. But it probably shouldn't have, right? You know, you and I should not be here talking about electronics, right? That's right. You know, it's like all this stuff. And, yeah.

Dave Jones: Some freaky shit happened. Right. Yeah, freaky shit.

Chris Gammell: Right. And, yeah, so that's what I was... So, yeah. LEDs are possible because of some weirdness. So, cool. I love it. Yeah. Cool article. Definitely people should check it out and learn more about LEDs because I need to myself.

Dave Jones: Oh, that's brilliant. Yeah. See, I don't know. I just, like, I find it fascinating, but I just want to buy them. And so I want them to work, you know?

Chris Gammell: Yeah. Yeah. Yeah, we've talked about that before.

Dave Jones: They're just a component. Yeah. Like, you want to buy relays and you want them to work.

Chris Gammell: I do. I do. I've been chasing relays. Anyways, so I've got a project I'm working on where I need, like, a low impedance path. And I need it not to vary with time or temperature, right? And so I think the first thing people would say, well, you could throw a MOSFET in there, right? You could drive this MOSFET. And I could, but it's, you know, like, there's no real guarantees. You know, either I'd have to, like, heat it and then there's self-heating stuff. And basically, the best...

Dave Jones: Is it that critical that you need the on resistance to be consistent or something?

Chris Gammell: Yeah. Yeah, I really do. It's... Right. Okay. Right. Yep. And so it's for a measurement, basically. And that's why. It's in a measurement path. So right there.

Dave Jones: It matters. Bingo.

Chris Gammell: Right. Yeah. It would go into my error budget, right? And I've mentioned error budgets before. Yeah.

Dave Jones: That's a big deal.

Chris Gammell: Right. Yeah. So, and if it's a big source of uncertainty, more than anything, you know, error is one thing. You can always calibrate out error. But uncertainty, then you just start... You just explode in terms of, you know, being able to be confident. And drift as well. Right. Exactly. Exactly. So if you don't know...

Dave Jones: Uncertainty and drift are... Yeah. Yuck.

Chris Gammell: So those two things. And noise, too. I mean, noise is another thing, right? But that all kind of adds in there. And it does.

Dave Jones: Yeah.

Chris Gammell: Right. So I have this... I need like a really low impedance path.

Dave Jones: It's called a piece of wire.

Chris Gammell: Right. Yeah. And that's one thing.

Dave Jones: Which is like a jumper link, right? On off.

Chris Gammell: And that's... I've said before, a jumper is just a really, really slow relay, right? Exactly. With a meat controller. A meat controller. Yeah. Someone said that about... I was talking about control systems and using these... Well, no.

Dave Jones: If you wanted to get fancy, you could program in a little robotic arm into your thing, which lifted up the jumper and disconnected. But that's called a relay, right?

Chris Gammell: Right. Yeah. The little robotic arm. That's like... Right. That's getting into... One of those machines called where they're like super complicated, like a mousetrap machine.

Dave Jones: Oh, a Rube Goldberg machine. That's it. Yeah.

Chris Gammell: Rube Goldberg. Yeah. So, yeah. Right. A relay seems to be the best thing. And so, I started doing some research just on relays and, you know, what did I need to do? And my thing is I needed to be small. I needed to be cheap, which is... Well, that's...

Dave Jones: Is it a one-off thing or is it a matrix?

Chris Gammell: It's a matrix, really. Right. Yeah. It's many relays. So, that's why I needed to be small.

Dave Jones: One thing you've got to... For those youngsters out there playing along at home, one trap with relays is that if you want to do matrix, you know, matrices of relays and you want to pack them close together, then they can actually interfere with each other. The magnetic fields can interfere. So, you've got to get a shielded... You've got to get a magnetically shielded relay if you want to pack them tightly together. You know, if you've got a big card and you're, you know, switching 48 inputs to 16 outputs or something like that in any combination.

Chris Gammell: So, you've had experience with that, right? I have indeed. Yeah.

Dave Jones: It's a big deal. And you can buy special magnetically shielded ultra miniature relays just for these, you know, they used to go in the phone switching exchanges and stuff like that. You know, you had to switch 10,000 lines to... Oh, yeah. Yeah. And these relays cost like 10 bucks a pop in volume, some of them. You know, they're really, you know, they're really high spec relays and they've got massively low leakage because when you put a thousand of them in parallel, right, they each might have two gig ohms, like, you know, off impedance, right? Right. You know, leakage. But you put a thousand two gig ohm resistors in parallel effectively and, well, you know, that's a lot of leakage.

Chris Gammell: Right. Yeah. It starts to really affect your stuff.

Dave Jones: Oh, yeah. And that varies enormously with temperature and time. Let me tell you. I've had huge issues with matrices which vary with time. And I had to work on projects where I was actually measuring hundreds of mega ohms to a gig ohm, right? So, through these relays. So, wow. It was... Yeah, man.

Chris Gammell: It was complex shit. Yeah. Getting into high impedance stuff. I mean, that's what I used to do in my old job. That's nasty. And that's... Yeah. Yeah. You start counting electrons.

Dave Jones: And it's bad enough with one or two relays, yeah. And you put a dozen, a hundred together and you're in deep shit.

Chris Gammell: So, we should step back. Anyway. In case there's people out there... You know, we forget this sometimes. We should step back and talk about relays just at their base as well. A relay is basically you have a strip of metal and you use a magnetic coil and you move... You physically move that coil or that piece of metal to contact another piece of metal and that creates a connection. So, it is a mechanical process, right? And like Dave was saying, you sometimes need to shield that because stray magnetic fields can then actually make those pieces of metal come together. But, yeah. It sounds really simple. And yet, it really can be crazy. And we've talked about the video on here before where they're making reeds for reed relays. Dave, you love that video, right? Oh, yeah.

Dave Jones: See, reed relays, if you don't know, are the duck's guts for low signal level stuff. Right. So, if you're doing low signal level, reed relay is what you want.

Chris Gammell: Right. And that's another good distinction too. So, there's signal level relays. There's power relays. Power, yep. Power relays can be huge. And they're necessary too because when you're switching power, you... Yep. You can't necessarily put... Unless you have a really fancy expensive MOSFET or you're willing to burn some power in your device, you need a low impedance path. And often, that's shoving two pieces of wire and metal together in this case.

Dave Jones: But they have to be a... Sorry?

Chris Gammell: No, go ahead.

Dave Jones: They have to be a special type because usually they actually have to break huge amounts of current, right? So, the arcing is phenomenal. So, they have to be special types of metal. They have to be multiple contacts. They've got to separate a certain... You know, a much larger distance than a signal reed relay and things like that. And reed relays have to be, once again, like a gold contact one or something like that. So, you get low voltages generated from different types of metal actually contacting. Right. Because these things can end up like a point contact diode if they were bad enough. Right.

Chris Gammell: If you get arcing or anything like that, you start to pit there and you change the whole material, basically.

Dave Jones: Your relay can turn into a freaking diode. Right. It's crazy, right? It can have non-linear effects. It's crazy.

Chris Gammell: Yeah.

Dave Jones: Right?

Chris Gammell: Oh, man. So, yeah. It seems like this really boring topic. I know. Come on. It kind of is. I mean, it is a boring topic. It kind of is, yeah. But it's so important because these things are everywhere, right? If you're doing a- I know. You're trying to hold off a high impedance path or you're trying to create a low impedance path for power. Like, these things really, really can matter. Yep. So, yeah.

Dave Jones: And because they're a physical thing that changes, you know, or the rated life of them is a big deal. How do they change over time? Right. You know? Right. Wow. It isn't like a MOSFET. A solid state semiconductor can, you know, be used an infinite amount of times in theory, right? In theory. Because there's no wear and tear.

Chris Gammell: Right. If you don't, like, exceed heat ratings and everything, you should be able to switch it. They should be infinite effectively.

Dave Jones: Yeah.

Chris Gammell: Right. But if you smack two pieces of metal together enough times, you're going to start wearing on it. Exactly. If there's any air in there, you can create oxidation and yada, yada, yada, yada, yada.

Dave Jones: I'm looking at my alarm sensor up in the corner, right? I've got a house, you know, one of those infrared, passive infrared sensors up in the corner. And they've actually got a relay in them. And they've got a very high reliability relay rated for 10 million changes, right? Because every time I wave my hand here, here we go, I'm waving my hand. You can't see it. It's great on radiators. And the lead turns on in there, right? And the relay switches. So every time you're walking around your house, this relay is going off, on, off, on, off, on, right? Right. And it's doing that 24-7.

Chris Gammell: Right. And if you're running around your house like a maniac, or like Dave's child will be doing in a few months, I'm sure.

Dave Jones: Well, he's waving, little Sagan has now, he loves the red LED comes on. And now he knows that if he waves his hand, you know, it comes on. So now he thinks he can control every light everywhere by waving his hand.

Chris Gammell: Why is that kid waving at the lamp? That's weird. Well, just sit him down, have him listen to this show. We'll tell him about relays.

Dave Jones: I need to give him the talk. Oh, yeah.

Chris Gammell: The relay talk. One of many. Sagan, today we're talking about MOSFETs. Yeah. We need to talk. Son, step into my office. Come on. Yeah. Oh, my goodness. Oh, boy.

Dave Jones: Yeah. I'm like, yeah, but this seemingly innocent topic is, you know. It's a rabbit hole. You could write a PhD paper on relays, right? Right. Same thing with anything.

Chris Gammell: And so, you know, I'm looking for this situation where I need really small, I need it to be low cost, which is, you know, I've already given up hope on that. But I start kind of digging into it. I asked on Twitter. I get some nice responses. So thank you to everyone who helped me out on Twitter. And started poking around. I found one on Mauser, which is this tiny Omron.

Dave Jones: Mauser.

Chris Gammell: How do you say it?

Dave Jones: Mauser.

Chris Gammell: Mauser.

Dave Jones: Yeah. Mauser. That kind of sounds Bronx kind of accent, doesn't it? Like, you know.

Chris Gammell: I say it like Schnauser. Mauser. Schnauser.

Dave Jones: Mauser.

Chris Gammell: Yeah.

Dave Jones: Yeah.

Chris Gammell: And so, basically, we were poking around on Mauser. And we found one. And so, this is actually, it was one of those cases where, you know, you're on a site and you're looking for a relay. You know that, right? But you can't necessarily type in low cost and small and get all the things you need. Sometimes there's that one search term. And in this case, it was a very wank term. It was MEMS. It was a MEMS relay. Ah, yes. A microelectrical mechanical system, right? That's what MEMS stands for. And so, we found one. And it is. It is very tiny. It's an LGA package, right? It's a single. I know, right? It's a single pull, double throw. And it's crazy. And it's exactly what I was thinking about. You know, it's basically, I was thinking about it in terms of how a lot of these vendors now are putting coils into their packaging, right? And so, to step back from relays, another way you can isolate is with these isolator chips. I know that a bunch of people are doing them now. Yep. But they actually, they have two chips in there. And then, in between, they put two coils. And basically, it's a transformer.

Dave Jones: It's a transformer. And so, they do them for like an isolating USB and stuff like that. Right, right.

Chris Gammell: And they're actually doing even low power switchers now. Like, they're DC to DC power switchers, right? And basically, it's just two chips in there with these coils. And then, it's a bunch of polymide, like the actual black stuff the chips are made out of. Yep. And so, I was figuring the same thing. I was like, well, you could put these little coils in there because basically, that's what a relay is. You could make very tiny coils. And you could energize those coils. And then, you could, you know, pull two capacitive plates together. And, you know, try and actually make contact like that. You know, like basically, just a tiny relay. It's the same idea, tiny relay though. And, I'm pretty sure that's what this is. And, it's, you know, low power. So, you don't...

Dave Jones: How low power? How much drive current does it need? Because a good low power read relay will take like five milliamps or something like that, you know?

Chris Gammell: Right. Switching current is... Oh, wait. That's switching current.

Dave Jones: A live data sheet readings here first.

Chris Gammell: I know, right? Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.

Speaker ?: Yeah. Yeah. Yeah. Yeah.

Chris Gammell: I think it's like a five volt switch though. So, it's, you know, it's low voltage. But, the thing is, it doesn't even matter. Because, this is all an RF relay. And, so, of course, it costs like, you know, it's a 10 gigahertz relay. That's what it turns out to be, right? Right. And, so, it's like, well, all right. I'll see you later. I mean, I'm just... I'm never using these. What, it's a hundred bucks, right? Yeah. It's like a hundred bucks in one quantity. So, it's the same thing, you know? And, so, eventually, I came to the conclusion that the whole balance point of low power and low size and low cost and it just, you know, I'm asking for too much. And, sometimes, you hit that point where you're just like, all right, you're asking for too much. Yep. But, then a couple days later, someone, I think it was someone, I think it was Paul on Twitter, he pointed me to a Spectrum article that just came out, IEEE Spectrum. And, basically, they're now actually doing this at the chip level. And, so, what they're doing is they create these really... There's this great graphic for it. Basically, though, they create like this... They basically create a relay on a chip where they create a metalization layer over top of, you know, like two contacts, like two metalized type contacts. And, then they etch away everything in between, so then it's an air gap. But, it's, you know, it's suspended by this, quote unquote, squishy area. And, then, basically, though, there's now a pressure put down upon this thing. It's actually making a contact. So, you'll have to look at the picture. I'm sorry, this is...

Dave Jones: I'm looking at the electron microscope photo of this. And, it's cool. We'll have to post it in the show notes, I think. Because, it's just a cool photo.

Chris Gammell: Right. Yeah. It's definitely a MEMS device. Like, this is an ultimate MEMS device. I'm not sure if this is the one I was looking at on the Mouser. But, yeah. But, the thing is, certain applications, you just can't get away from needing a relay. And, so, this is kind of like taking an old technology and pushing it into, you know, the 21st century, right?

Dave Jones: Into the Solastide era.

Chris Gammell: Yeah. Right. Exactly. And, now, so, who knows? Maybe this is... I mean, they're still calling it a relay. It's kind of like, well, maybe you guys should call it something else. But, you know, just some of the parameters you need to get. You know, like, how do you do it other than, you know, actually pushing two pieces of metal together? I don't know.

Dave Jones: Well, and the problem here is, right, it is 7.5 square microns, right? Right. That's how big the relay... You're limited by the freaking package, right?

Chris Gammell: Right, yeah.

Dave Jones: To actually mount this thing. The relay... But, hey, it could be used... Like, so, if you're talking a one-off relay here, it's almost pointless, right? If you're talking, okay, you have one of these MEMS devices in a package. Well, the package is going to be a million times bigger than the actual, you know, relay element itself. Right. But you could create massive matrices with these things. Right. And actually have them on a device.

Chris Gammell: And that's actually the drive beyond this whole thing in the first place. If you actually read further on in the article, that's actually why they're doing this. It's actually not for... It's not for me. And it's not for you, Dave. Right. And it's not for anyone needing a single relay because this thing can't handle any kind of significant power. However, what it can do is what a lot of devices can't do. I mean, we had a... You know, there's a new TI part out there that's like the successor to the MSP430. I'm not sure if it's still on our list or not, but it's called the Wolverine, I think. Right. And they're talking about like, you know, 600 microamps of standby current or sorry, nanoamps or something like that. And, you know, there's all these sleep modes and everything else like that. But the thing is, you know, as you get smaller on your transistors, you know, talking about fabs again, right? As your geometries get smaller and smaller, physics starts to run... You start to run smack into physics, right? Oh, yeah. Runs right. Right. And so basically, as you get smaller and smaller transistors, they need less power to switch on and off. However, the leakage in them, like you were talking about, Dave, it goes way up.

Dave Jones: I was going to say, what is the leakage in one of these suckers?

Chris Gammell: Well, no, that's just in a transistor, right? So CMOS transistor, that's just the basic tradeoff. You get smaller, you get less power, but your leakage is way worse. So, you know, you have to turn on sleep modes and you actually shut off whole parts of the circuit with bigger CMOS stuff. You know, you should shut down the power rails, basically, to different parts on the chip. But the idea behind this nano relay thingy is you make everything out of these. It switches way slower, right? So maybe you can only do 100 megahertz, but...

Dave Jones: Only 100 megahertz.

Chris Gammell: Right, which isn't bad, right? I mean, there's FPGAs running that pretty regularly. But the idea is when it's off, it is just, it's totally off, right? It's not like, you know, you could just turn off a switch and now...

Dave Jones: Well, no, there's always leakage and I want to know what it is, damn it.

Chris Gammell: Oh, well, okay. It's a lot better than a tiny CMOS circuit, right? And that's kind of why they're actually doing this in the first place. And so my journey for finding a relay actually found this, ended up at this... Nice. ...really interesting kind of take. It's like this remix of old technology trying to push it into this new era. And it might end up creating, you know, you think about sensors and a lot of the sleep modes and everything they talk about. And even using like energy harvesting and how you can power stuff off, you know, vibrational energy. Yeah, vibration, yeah. Or thermal energy, something like that. And if you have a circuit that has effectively zero standby current, you know, that could really change some stuff. And it could change how people make chips and everything else. So I thought it was very exciting. I think it's pretty cool.

Dave Jones: It is very exciting. I totally agree.

Chris Gammell: I want to know if these little chips, like if you have a whole array of these things, I mean, you can hear like a, you know, like an old Pentium processor. You can hear those things clicking away. But if you're actually like, if you're actually making context now, is that thing going to sound like, you know, tap dancing?

Dave Jones: Is it going to be a high-pitched squeal as you...

Chris Gammell: Right, yeah. So it'll be an, it's an interesting technology to keep an eye on, I think. So.

Dave Jones: Totally agree. We've gone way over, dude. We don't even have a special guest. Oh, really? But we do next week.

Chris Gammell: We do. Oh, man. We're done?

Dave Jones: We're done. You were right. An hour and eight minutes.

Chris Gammell: Yeah. Oh, man. Okay. Well, I got really excited about relays there for a bit. As you do. I do. Yeah. All right. So usually we tell people to subscribe to Twitter. This week we were going to do something a little bit different, though. So the most important thing is obviously listening to the show, we think, and hopefully you think so as well. And we realized, though, that sometimes people don't know how to subscribe. Right. Yeah. So subscribing is very important. There's links on our website. You can go to iTunes if you're into that sort of thing. But the thing we wanted to bring up, so subscribing is important. Dave, what's your favorite podcatching program?

Dave Jones: I am using Miro. M-I-R-O.

Chris Gammell: Okay. So basically that, and I use BeyondPod. I use it on my phone because then I can plug it into my car and I listen on the way to work. These are like really easy ways to listen to podcasts. Podcasts. If you're listening on the site right now, if you just click and play, that's cool because you're probably the ones commenting, which we love. But if it's easier to subscribe, we want to make sure people knew about these other things. So Miro, BeyondPod, Google Listen, there's a whole bunch of them out there. iTunes, if you're into that whole.

Dave Jones: And all you've got to do is type in our feed address, which is The Amp Hour.com slash feed, is it?

Chris Gammell: Well, it's easier to go to our site, but you could use that if you wanted to.

Dave Jones: But anyway, you can copy and paste that into your program and it automatically catches all of our programs and new ones will pop up as well. And you can choose to play them or download them.

Chris Gammell: Yeah. And you can set up auto-downloading sometimes as well.

Speaker ?: So.

Dave Jones: Yep.

Chris Gammell: Highly recommended.

Dave Jones: It can auto-download to your iPod, to your connected iPod, or it can auto-download to your MP3 player or a USB stick, which then you can play in your car and stuff like that.

Chris Gammell: Yeah, that's why I like it on my phone because it just does every night at 2 a.m. I'm getting more podcasts. And so. Right. It's awesome. So subscribe. We'd love it if you did. If not, we'll take you listening on the site too because you're probably commenting. Awesome. Looking forward to next week. We are. Yes. See you then. See ya. We'll see you next time.