#290 – An Interview with Mark Morin of Nufern

Mark Morin: This is The Amp Hour Podcast. Recorded March 9th, 2016. Episode 290. An interview with Mark Morin of Newfern.

Dave Jones: Welcome to The Amp Hour. I'm Dave Jones from the AEV blog. And I'm Chris Gammell of Contextual Electronics.

Mark Morin: And I'm Mark Morin, Principal Engineer at Newfern.

Chris Gammell: Welcome, Mark.

Dave Jones: You were going to call yourself Jedi, Mark.

Mark Morin: Yeah, I thought he was going to be. I gave you guys the slip on that one, I guess.

Chris Gammell: I guess we're not going to edit in the lightsaber noise now. Sorry, man. That's okay.

Mark Morin: Hey, Mark, thanks for joining us. Oh, no problem. Thank you, guys. So, what is Newfern? We are a fiber optic company in Connecticut, you know, in the United States. Uh-huh. I feel like you guys have an international op. Ah, right, yeah. That's good. So, we make actual fiber. We have an on-site draw. Nice. Make the best specialty fibers in the world. If you go to any lab anywhere, there will be Newfern spools in it. And then we also have a smaller group that makes lasers. And we have manufacturing lasers for, like, material processing. And then there's higher power lasers that are kind of an ITAR thing.

Dave Jones: What's an ITAR thing for those who don't know?

Mark Morin: International trafficking on arms regulation. That means that we're not going to talk about them.

Chris Gammell: He can't tell you, Dave.

Mark Morin: He can't tell you.

Chris Gammell: Because you're not American.

Mark Morin: Yeah, past that. We're not talking about it.

Chris Gammell: That's what that means. That's okay. I'm sure that, you know, lasers on their own or lasers of all types are interesting. So, we don't need to dive into the ITAR stuff.

Dave Jones: Sure. And everyone wants to hear about lasers, of course. But, you know, poor old fiber optic doesn't get much love, does it? Right. And it probably should. There's a lot of tech which goes into fiber, I'm led to believe. Yes.

Mark Morin: Tons, I hear. Tell us. Tell us. I'm certainly not involved in that on a day-to-day basis at all. So, you know, they have, I guess they start with a preform. And I'm going to, if anybody here listens to this, I'm going to get butchered for my simplistic explanation. But they start with a glass preform.

Chris Gammell: Well, there's clear cables. That's about it. Basically. Yeah.

Mark Morin: And then they put it in this flaming tower and it just kind of drips down and turns into fiber, you know. Oh, really? That's cool. Yeah. I mean, it's super simplistic explanation, I suppose. Yeah.

Dave Jones: But what makes your stuff better than the rest? You said, like, the best optical labs in the world will have your fiber. What makes it better? Is it lower, I don't know? Is it lower jitter? Is it lower blah, blah, insert? Super pure. I don't know.

Mark Morin: I think our polarization maintaining fiber is one of the fibers everybody stands on.

Dave Jones: So that's a key parameter, is it? Polarization of the light as it travels.

Mark Morin: If you're trying to maintain it, yeah, it's a huge thing. So there's kind of like if you imagine a flag on top of every photon and the flags. I like that. Right? That's novel. Yeah. As the flags, as the photons are going down the pipe that is the fiber optic, the flags are all randomly oriented. So if you take this polarizing fiber and you coil it up correctly, you can get all the flags to kind of orient themselves over into the same direction over a length of fiber, right?

Dave Jones: That's what like a polarization fiber does. Please don't tell me that makes your audio system sound better. No.

Mark Morin: No. Excellent. That certainly does not. Honestly, I probably, I've built the lasers. I could measure the polarization. I couldn't explain to you what they actually do with a polarized light when it comes out of the laser. It's the systems I've worked on. That's what the scientists are for, right? Sure.

Dave Jones: Yeah, that's a big deal to the scientists who are doing fiber optic research and stuff

Mark Morin: like that, right? Some communications applications I've heard and then I kind of move on to the next project before any of the implementation is actually done.

Chris Gammell: I'm really interested in, I mean, so like it feels, you know, I already know it's not true before it comes out of my mouth, but it feels like fiber is kind of solved. Like, you know what I mean? Because obviously we're talking to Dave right now and even we're talking to each other, you know, from across the state.

Dave Jones: I have a fiber, yeah.

Chris Gammell: So like some of that stuff feels solved, but obviously there's going to be increased in technology. So like even on both the laser side and the fiber side, what is this current state of technology even? I honestly have no idea.

Mark Morin: Oh God, neither do I. The fiber we're talking over that's the internet is like SMF28. It's like a standard telecom long distance communication fiber. And that's like the polarization maintaining fiber I was talking about as a specialty thing

Dave Jones: for, you know, labs or-

Dave Jones: For science experiments and stuff like that.

Mark Morin: Some kind of laser that would be a point to point thing or-

Dave Jones: Well, that's the thing I was going to ask. Like, you know, people talk about, oh, just lay fiber and that's it. It's done. And then we'll be able to use the bandwidth of it for the next hundred years, you know? Kind of. Yeah. Like, is it like, yeah. But, you know, you're saying it's kind of like a standard kind of, you know, fiber that they lay down. Is that good enough for the traffic? Or will they have to rip up all these international fiber links in 10 years because, oh, well, we've got all this electronics and optics. Yeah. Like, you know, like we can't- What are that for? You know, we need it. It's- Will they find it's not good enough maybe in 10, 20 years?

Mark Morin: I think for telecom communications, they try and exploit bandwidth that they're not currently using. So if you have a- You know, like the standard EDFA for telecom is a 1550 central wavelength. So you would try and make some kind of out-of-band laser, you know, whether it's ytterbium or thulium. And then you would use those bands to try-

Dave Jones: You can't just throw those words in there.

Mark Morin: Also, can you spell them?

Chris Gammell: Because I'm taking notes here and I have no idea what you just said. Yeah.

Mark Morin: If you go way down to the bottom of the table of elements, there's a couple of, like, BMs. They end in BM. And they're all the rare earth elements that they kind of dope the center of the fiber with, right? So when you pump the fiber with a wavelength in the 960 range, so you shoot that into the cladding of the fiber, right? That's like the outer layer.

Dave Jones: Oh, it's in- Oh, okay. So it's in the cladding. It's not in the glass itself.

Mark Morin: Well, the elements in the glass in the- Well, it's all glass, but- Right. The element is in the center of the glass. And then you kind of shoot a wavelength like 960, the pump light, right? You're trying to pump up the inversion level of the rare earth element atom. So you shoot in, like, 960, for example, and then the index refraction of each layer of the fiber kind of keeps the reflected waves, right? Because the wave is never really going to enter perfectly, like, maridonial to the fiber. So-

Dave Jones: I have experience in that game.

Chris Gammell: I'm going to take a quick break here, guys, and just say, I have no idea what you're talking

Speaker ?: about.

Chris Gammell: I don't expect to, and I just wanted to show some solidarity with some of our listeners right now who are just staring at their podcast players saying, what the hell is this?

Mark Morin: This is great. So a maridonial ray is a ray of, like, a photon that goes directly down the middle of the fiber. It's like a theoretical perfection that never really gets achieved, right? So a skew ray is more at the angle tangent to the side of the fiber. So you launch down some non-zero angle into the core of the fiber, and then it would bounce around in there, right? Between the walls of the fiber.

Chris Gammell: And then there's, like, losses and stuff like that because of the bouncing or what?

Mark Morin: Because of the bouncing. And if you turn the fiber, your loss is increased because the angle of incidence is higher. Yes. But the fiber is constructed so that the outer layers are more reflective concentrically as you get to the middle. And then, so if you launch light in, it kind of finds its way to the core, right? Right. So with, like, a yterbium or erbium fiber, so the core is doped with this rare earth element, and you shoot photons into the cladding, and then that finds its way to the core. And then when those photons, like, hit the electron that's actually, like, hanging in the valence layer, the outer valence layer of the rare earth element, it'll knock it up one valence layer, right? Mm-hmm. Ah. And then you just kind of—

Chris Gammell: Because those things have a lot of layers, too. Like, the stuff at the bottom of the periodic table has tons and tons and tons and tons of layers. Right.

Mark Morin: There's hundreds of electrons in these things, right? Right. Right. So you bump a couple of electrons up, and when they're at a higher energy state, they either can fall on their own, which is called ASC, like, automatic stimulated emission, or you can seed them with a signal. So if you shoot a second set of photons that are the exact same wavelength of the rare earth elements, like, natural emission down the fiber, you'll just happen to knock down these electrons and produce, like, clones of the photon you sent in at the same wavelength, right?

Chris Gammell: So— And this all has to happen immediately? Oh, it's all— Like, they're chasing each other, almost? I mean, like, at the speed of light kind of thing?

Mark Morin: Yes, literally. Yeah. So, like, yterbium is—1064 would be your central wavelength target for a yterbium laser. So you shoot a 1064 photon down the center of a yterbium-doped fiber that's already pumped up, and then every photon that you hit with a photon that you seeded will knock down—or every electron you hit, I should say. When that knocks down a valence layer, it releases, you know, 5 or 10 photons. So the amplification happens there, like, on a particle-by-particle basis. Wow. So just over the law of large numbers, because you're just pumping a ton of photons into this fiber, you will hit a bunch of other electrons and emit a ton of photons. So that's kind of how the amplification works. So your laser, like a fiber laser, is just a couple of coils of these things arranged in series, kind of like a circuit. And each bank would have its own pump light that would pump up the rare earth element to a higher, you know, the electrons to a higher valence level. And then because your previous state already has the amplified light coming through, you can pump more electrons up in the sequential states, right? So each stage can have more gain than the last because you're seeding with more photons. So that's basically how the fiber laser works. And then at the end, you put it into some useful, either focus it on a material or focus it into a fiber to send a signal. You know, your seed photons could be a digital pulse, right?

Chris Gammell: Yeah.

Mark Morin: Like some carrier wave. And then you just let your signal ride on that.

Chris Gammell: So you're talking about basically doing this. This is in a gain stage. It's all outside of like a transmission, actual transmission piece of fiber, right? So just to simplify that. So if I'm sending a signal from, I'm just sending a single pulse from me to Dave over the transatlantic cable. And I know there's a lot of stuff in between there. But if I'm sending a single pulse, I would use this kind of coil of this Atterbian stuff at the 1064. And then that would basically act as an amplifier. Is that right?

Mark Morin: Yes. Well, yeah. So like you're in a simpler version of this is your hometown's fiber network. On some telephone poles, there's going to be a box with a little amplification stage in it. That's just going to crank up the signal enough for the hundred homes that are around that box. And then the line, you know, you go down to the next neighborhood, there's another box that just cranks the signal up again. All right. So they kind of daisy chain them around. The transatlantic, I think there's repeaters in the middle. Yeah, I think there is too. Oh, yeah.

Dave Jones: Every hundred, 50Ks, 100Ks or something, isn't it?

Mark Morin: Yeah. And the long haul fiber, though, is like passive, I'm pretty sure. Right. Okay.

Dave Jones: Now, you mentioned it literally happens at the speed of light. But isn't there a fiber optic latency?

Mark Morin: There is some.

Dave Jones: Right. Some, how much?

Chris Gammell: It's not a air kind of thing or what? Order of magnitude.

Dave Jones: Dude. Yeah. Is it like, yeah. Can you give us a rule of thumb figure?

Mark Morin: That's, I'm going to guess that it's like tenths or single percentages delay.

Dave Jones: Right.

Mark Morin: Of the speed of light. It's not half the speed of light or anything like that.

Mark Morin: Got it. And I'm pretty sure that depends on the material in the core. Absolutely. Yes. So you'd have to pick which fiber you're talking about and then you could figure that out. We should have a follow-up episode and actually have one of the optical scientists come on here.

Chris Gammell: So, but basically when you're going from an electrical to an optical like that, you're saying that there's going to be a laser and then there's going to be an amplification stage possibly. And then it's just photons all the way down until you get to a receiver kind of thing.

Mark Morin: Well, you could have a laser on its own. That's not really high speed like an internet signal would be different, but just a laser. You could just pump the medium and get the ASE signal and then amplify that or use some gratings to set up the, well, there's like gratings are basically mirrors that are tuned to a wavelength that are inserted in the fiber. So you could use gratings to set up the wavelength you want in a small cavity, you know, a small coil, and then you could send that down, send that signal down the fiber to a higher amplification stage.

Dave Jones: Isn't the, oh, sorry, go on.

Mark Morin: That's a, like a, you'd call it a CW laser, right? Continuous wave.

Dave Jones: Right.

Mark Morin: If you wanted to send packets of like internet, you would have some kind of diode that you would modulate super fast. Yeah. And usually they just dither that diode's intensity by a couple hundred milliamps, I want to say.

Chris Gammell: Oh, really? They don't take it all the way?

Mark Morin: So, no, it doesn't, it never goes to zero.

Chris Gammell: Yeah. If you, it's a lot. I guess you want to keep it biased, right? So that would make sense.

Mark Morin: Right. And the diode, like LEDs and laser diodes have a threshold where they start emitting. So there's, you know, you'll get to 600 milliamps and there'll be no light. And then at 700, you have a bunch of light. So you could really even just dither around that lazing threshold. But there's spiking involved with that. And I think real companies would just go well into the lazing and then kind of move around and make a one this many dB and a zero is below this dB, right? Right.

Chris Gammell: Yeah. Because then you could test if it's just broken, broken instead of off for a split second kind of thing, right? Right. Exactly.

Dave Jones: That's interesting because I figure that a lot of people out there would think that lasers just, you know, on, off, on, off, you know, it'd be ones and zeros down here, but it's, it's more analog-y than that.

Mark Morin: Yeah. Like everything's analog. Yay. Damn it. You know, it's, it's, it's all digital on paper and then analog and you try and do it in real life.

Dave Jones: But now isn't, isn't the holy grail of fiber optics, um, to have completely, um, optical amplifiers like optical charge pumps or whatever the term is so that you don't have to in, in these repeaters you've been talking about, you don't have to convert it to convert optical to, um, a signal and then convert that back to optical again. Right.

Mark Morin: I, I, I would think that they do that in these repeaters.

Dave Jones: Right.

Mark Morin: I would think that they just take, they take the fiber in, have an amplification stage that just runs and then they take the fiber out.

Dave Jones: That's, that's a purely optical amplifier. Because you don't want to have to convert, um, from optical if you don't have to.

Mark Morin: You don't want to convert the signal. No. From optical to electrical. You'll have to make the pump light with, with electrical, I would guess, unless there's some really wacky thing they can do. But yeah, that I wouldn't think that in just like a long haul amplifier that you would convert from signal to electrical and then back to signal.

Dave Jones: Now, do you have any experience in this? Because, um, I used to work at a company where the holy grail of, I used to work in the seismic industry, which is underwater sonar, you know, searching for oil and submarines and other things. The holy grail of that and everyone, it was always, oh, it's a decade away. It's a decade away. What I thought was a new grail of that is using fiber optics as vibration sensing elements. Have you heard about that or been involved in that?

Mark Morin: Now, if you go on our applications page.

Dave Jones: Ah, there you go.

Mark Morin: click on geophysical. Wait, you said you're an engineer, not a salesperson. Come on. Yeah, well, hey, I'll take it where I can get it. Yeah, that's right. But you go on geophysical, and then there's downhole instrumentation, and there's oil and gas production monitoring.

Dave Jones: Ah, there it is, geophysical, yep, yep.

Mark Morin: So we're 10 years later, then, it sounds like.

Dave Jones: Oil and gas.

Mark Morin: It's probably been 10 years since you were saying that. So how does that work, then? I'm pretty sure they can sense the temperature and vibration down the fiber.

Dave Jones: Oh, I didn't know about the temperature. That's interesting. I know the temperature.

Mark Morin: That's on, like, one of our little sales brochures in the lobby. That's for sure. Let's see if I can.

Chris Gammell: What's, how does it sense that? Just, like, how the medium changes or something like that?

Mark Morin: I think they put a, another rare earth element or doping in the center of the fiber, and that has a, it'd be like having a resistor that you built with a purposely high, you know, a thermistor, right? Yeah, right, right. A high temperature coefficient. Yeah. So you'd have kind of the same, it's not resistance, but it's a, you know, you'd have some. Yeah, yeah, index or refraction or coefficient that would, and then I'd assume you'd shoot pulses down, reflect them, and see what you got back for intensity and delay to kind of figure this stuff out. Yeah.

Dave Jones: Now, I'm looking at your list here of your, like, geophysical products, fiber-type products, and there's a category list, and the category list has, like, one of them has the element YB, which I won't try and pronounce. Uturbium. Uturbium. Yeah, that's the uturbium. Yeah. Uturbium. Right. Right? Yes. But then it's got other categories, like 1.0 micron, or it's got select cutoff. Like, what are these categories about?

Mark Morin: Well, like, 1 micron is the wavelength for a CO2 laser, so if you had a gas laser, and then you wanted to launch that into a fiber, you would buy, like, a 1 micron fiber, I think.

Dave Jones: Got it.

Chris Gammell: So I'm going to take a little step back here. Yeah, sure. How the hell did you get into this?

Mark Morin: Well, I was in the high school library one day, and they said laser electro-optic technicians were going to be the highest, like, the highest demand field in the next 10 years. And I said, well, I'll just go to school for that then.

Chris Gammell: Wait, this was mid to late 90s, is that right?

Mark Morin: 1996, I think. Exactly.

Chris Gammell: Right when all the fiber stuff was getting rolled out. Sure. Yeah, yeah.

Mark Morin: And they happened to have a major at our local community college that was laser electro-optics, so I just decided I'd enroll in that, and that was the beginning. But I was in a trade school for electronics before that, so, you know, my high school was a trade school. Oh, gosh. So I already kind of had the electronics bit handled for the most part, so it was a breeze to go through. The electronics half of this major was not that much of a challenge, you know? Actually, I think I was able to absorb some of the higher concept classes, like, you know, you take a linear circuits class, and there's kids that are still trying to figure out whether, you know, what series and parallel means. Oh, yeah, yeah, yeah. They're shoving op amps down your throat a month later, and, you know, having the bassist from a trade school was actually a big help in kind of figuring that stuff out. Oh, yes, huge.

Chris Gammell: That's great. So the thing that really strikes me about this already, I mean, obviously, this is a very practical application of, you know, chemistry. Obviously, the periodic table, stuff like that. And that is great. You know, I mean, like, obviously, in the electronics field, we don't hear about that as much. But, you know, like, valence electrons matter, right? I mean, it happens in a lot of stuff, but it's...

Dave Jones: Well, it happens at the semiconductor physics level. Right, exactly. In our industry. But apart from that, no, we don't really, you know, most application engineers, you know, don't... Which is the level we're at.

Mark Morin: Would you ever buy, like, a germanium diode to get the right forward voltage, you know, like...

Dave Jones: Oh, yeah, yeah, you would.

Mark Morin: Like, germanium or silicon is maybe the only time you think about it. Yeah, exactly.

Dave Jones: That's about it, yeah.

Chris Gammell: Yeah, and I mean, that's just kind of... That's almost like a branding thing as much as anything else. It's like, oh, I need this type instead of I need it because the valence, you know, you have to drop the different energy levels and stuff like that. Sure. No, it's really cool, though. And, like, that... I've talked to someone... Actually, one of my old bosses, he used to work on really high-speed oscillators. And he used to sort of... Kind of like the atterbium-type stuff. Or maybe it was atterbium. I don't remember at all. But, you know, like, just those really interesting, exotic-type materials that you need to use. So that... But that also drives, of course, the price up and very specialized equipment and everything. So...

Mark Morin: Yeah, I can imagine.

Dave Jones: So what are you doing at the... So what are you actually doing there?

Mark Morin: Oh, I design the circuit boards that control all this stuff. Basically.

Dave Jones: So he is one of us, folks. He is one of us. Yeah.

Mark Morin: Domain knowledge around the other stuff. Just because I've been immersed in it for 15 years, I've absorbed all this nonsense. Right. But day-to-day, I'm driving a seat of Altium. Right. Nice. You know, or I'm down in the lab actually cobbling together some prototype board to try and prove out whatever we're trying to do that week, you know? Yeah. For the most part.

Chris Gammell: So you said there's two different pieces to your business here, right? So there is the actual making of the fiber. But then you said there's, like, the amplifier side and the stuff that you're building. So what are some of, like, the device types that you're building at a high level?

Mark Morin: So we have a NuQ laser, which is a pulsed laser. So if you've bought an 0603 chip resistor, this is the kind of laser that etches the die and then actually, you know, scrubs off the carbon resistive element. Oh. I've been to the – Ah, right.

Chris Gammell: So the laser trimming you're saying. That's what – yeah, like the trimming.

Mark Morin: Oh, yeah. Yeah. Yeah. I've been to the factory where they do that. It's really neat to watch for about three minutes and then it's incredibly boring for the rest of the time. Yeah, right. But it's a little ceramic wafer kind of gets vacuumed in onto a tray and then the laser goes blam and then everything's kind of etched into a grid, right? And then they deposit carbon on it and then the thing sits there and, like, two flying lead probes come down and actually measure the resistance while it etches away carbon. It's really neat.

Chris Gammell: Oh. So it's, like, that dynamic where it actually is –

Mark Morin: Oh, it's unreal. Yeah. How, like, live it is.

Chris Gammell: I always assumed it would be, like, with resistors and stuff like that. It was just an amount of material type thing or is this for, like, just the super precise ones?

Mark Morin: To be fair, I think those were, like, 0.01% resistors.

Chris Gammell: The stuff that we're paying for out the wazoo anyway. Right.

Mark Morin: So the – I think the – even the 1%ers are just – it etches whatever it etches for 100K and then it just breaks the wafer up and they put leads on it, right?

Chris Gammell: Yeah. That's pretty cool.

Mark Morin: Yeah.

Dave Jones: And then they sort them afterwards because it's just – you know, it's just not worth trimming, you know? Sure.

Mark Morin: And then there's, you know, applications like metal engraving and etching. You know, if you have a serial number in the side of something, they don't do that with an engraving tool anymore. They just do that with a laser.

Speaker ?: Huh.

Mark Morin: Take zero time.

Dave Jones: I've actually seen that. Like, if you look at those micro photographs of dyes, right, you can actually see that the – that, like, you know, there might be some logo or something. It's, like, obviously been, like, laser etched into the dye rather than chemically etched. Sure. Because it just looks different, you know? It doesn't look chemically processed. It looks like it's been burned in, you know, with the laser.

Mark Morin: And sometimes you can see the dots. It will almost look like a dot matrix printer. Mm-hmm. Right. Put it down, you know, if it's moving really fast where the laser pulses are coming out at a rate that's comparable to the speed of the actual, you know, transom or whatever they're using. Oh, it's a – well, a system that uses an X-Y table, right, where you have a fixed optical piece and then a table that moves around under it, that kind of stuff is not fast enough for that. But there's a – Of course. A device called a Galvo that is basically three mirrors that move really fast, right? So your scan speed can be millimeters or centimeters per second. That kind of stuff, you can get dot resolution out of these lasers.

Dave Jones: Wow.

Mark Morin: And it actually looks pretty neat. So, like, we have a test pattern that's a bunch of dots and they're all arranged in a perfect grid so you can tell if there's, you know, dots off from where they're supposed to be. You can tell that your electronics design is messed up somewhere.

Dave Jones: Yeah, exactly. And a dummy question, but is that – I've got one of those too. Shoot. Is that every laser or is that a specific pulsed laser?

Mark Morin: That's the specific NuQ pulsed laser. Right. Does that. Okay. So, you know, your operation range –

Dave Jones: Because most people think about, like, laser is just on. That's it. It's not pulse. They think it's just on.

Chris Gammell: Yeah, lasers are meant to kill James Bond, right? That's what they're supposed to use.

Mark Morin: Exactly, right. You run it across a table at a spy's crotch. Slowly, slowly, though. And walk out of the room after you've told them your entire plan, right? That's right.

Chris Gammell: That's how you haze your interns. Is that right? Yes, exactly.

Dave Jones: And follow-up question. Why do you pulse the laser?

Mark Morin: Oh, the peak power. So if you're – like the NuQ, for example.

Dave Jones: Ah, chip out, right, yeah.

Mark Morin: Right. Your average power is 20, 30, or 50 watts depending on the model. Got it. But if you're running at 50 kilohertz, your peak power is kilowatts, you know? Right. So the damage threshold of the surface of the material you're attacking is some number, right? And you want to melt it. And if you can just shoot a pulse in a short enough time, it'll just melt the surface or evaporate it. It just goes away.

Dave Jones: Oh, I know one thing we can get into that a lot of people might –

Chris Gammell: Can I ask my dumb question first, though? You can ask your dumb question, then I'll get into it. So you mentioned the galvos and the mirrors. I've always wondered this. How does a laser bounce off a mirror? I know that's a really basic question, but like why doesn't it etch the mirror? Is it about the focal length or something?

Mark Morin: Well, because the mirror is made of a material that's reflective at the wavelength you're using, hopefully, or it'll wear out pretty fast, right? Yeah, you're right, yeah. The material you're shooting at, you know, like copper at 1064 is very reflective. So if you're trying to etch copper, you have to get a lot more energy. But like aluminum is pretty susceptible, right? So you can etch aluminum very easily.

Chris Gammell: And what determines the material's susceptibility to different wavelengths then? Is that just basically – I'm glad. Yeah.

Mark Morin: That's a whole other science on its own. That's another podcast as well. Yeah. It's all empirical testing to me, but like I'm sure there's people that can figure that out.

Chris Gammell: Yeah, that one blew up.

Mark Morin: Exactly. But like, hey, this isn't doing anything. I wonder what's going on here, right? Huh. Okay.

Chris Gammell: No, that's really interesting stuff. I just have no idea where that comes from. So you could have something that just isn't susceptible at a certain wavelength. Like if – so if like a plastic wasn't susceptible at a certain wavelength, I realize that's probably not the case. But if it wasn't – No.

Dave Jones: A silver mirror, for example, like a regular mirror might be silver-backed and silver might not be, you know, absorb-ive. Yes, exactly. Absorb-ive, I guess you'd say. If that's a word, at that wavelength. Right. So it might just reflect it.

Mark Morin: Yep. So like plywood is a good example. Or CO2. Right. But like your plywood sheet, that 8x4 sheet you pick up at Home Depot is cut with a CO2 laser. Yep. In almost every factory. There's no saw cutting this, right? And the CO2 can burn plywood. If you put it under that Nu-Q laser at 1064, it's not going to do anything. You know, you could run it for an hour and maybe start seeing a spot. But there's not a high damage threshold there.

Dave Jones: Got it. Even though that's a big-ass 50-watt kilowatt peak laser, it still doesn't do anything.

Chris Gammell: Right. Yep. So it matters power plus wavelength. Wavelength. Right.

Mark Morin: Yes. Interesting. So you're – yeah. It's like you have some thermal coefficient of absorption, right? And you get enough heat in the surface and it starts going away. And if you're not absorbing it at a high enough rate, you're just not doing any damage.

Dave Jones: Ah, physics for the win. Sure. Yep. All right. Laser drivers. Yeah. Because there's a lot of – yeah. Because there's a lot of – it's almost like a black art. Oh, laser drivers. They've got a – you know, there's special requirements for laser driver circuitry. And I'm sure you could – I'm guessing you can tell us all about that. Sure. Well, something. Yes. So, yeah. What is the problem, first of all? Like, everyone – like, I see this on the forum all the time. I'm trying to design this laser driver, but you've got to have super high peak currents, blah, blah, blah. Tell us what the issues are.

Mark Morin: Well, it depends what you're trying to do. If you wanted to make a – you know, you go online and you find all these, like, eval boards for laser drivers and –

Chris Gammell: Mm-hmm. Sorry. What is a laser driver? I guess I'm sorry. I don't –

Mark Morin: Well, you just – it's – well, there's laser diodes, right? And they – you just pump current through it and you get photons out is the most basic explanation. You know, that's like a transducer, I suppose you'd say. Sure.

Dave Jones: So they're current-driven devices, just like – Yes.

Chris Gammell: So electrical current turns into photons through this diode. Right. Yeah.

Mark Morin: And there's a bunch of challenges with switching on and off laser diodes. Like, when you cross the lasing thresholds, you get this relaxation oscillation. That creep makes a little spike, right? So if you just want to shoot a laser diode into a wall, it's really not that complicated. But if you want to start amplifying it in a fiber system, you can't have all these spikes coming out of the laser every time you turn it on and off, right? So you have to have a pretty well under control current source without overshoot or – Right.

Dave Jones: Because it's basically a current source you're switching off and on. And when you switch a current source off and on, you can have issues, especially if it's a large current source. Oh, yeah.

Mark Morin: And you're talking on the low end, you know, 10 amps or so. And on the high end, a lot higher.

Dave Jones: Yeah. And you've got to switch them fast too, you know.

Mark Morin: Well, this is where you get all the problems, right?

Dave Jones: Everybody can turn one on and –

Mark Morin: And one switch –

Dave Jones: Leaving it on. One hertz, right? You switch it with a switch and on.

Mark Morin: We have a 500 millisecond rise time and it's critically damped and everything's beautiful. But if you want to, you know, switch it kilohertz or faster, then you have to do a lot of things. And I'm sure you guys understand we can't drill too deep into this. Oh, yeah. That's okay.

Dave Jones: No, no. We just want generics. Like, you know, do you have to use – you know, is it hard to get special purpose – do you have to use special purpose transistors to end up doing it? Because, you know –

Mark Morin: Not really. It's a – No? You can use almost a standard DC to DC converter – Right. Architecture to make your power supply and, you know, a linear – Almost an audio type transistor would give you a – Okay. You want to have a nice analog region you can control in your – Mm-hmm. In your transistor, right? You don't want to be just smashing into saturation. Right. Yeah. Depending on your circuit that you have set up to do this, but –

Chris Gammell: Does that mean – well, that would require like – so if you had – I'm just like imagining this in my head. So you have a FET that's switching into a diode like that and then you're going to want some kind of feedback there. Does that mean you have feedback that's – you're using nonlinear feedback then because you're basically monitoring the diode voltage then? Is that the idea?

Dave Jones: Well, you've got to generate a constant current. You've got to remember it's not a –

Speaker ?: Right.

Dave Jones: It's a constant current generator.

Mark Morin: You would use some kind of current feedback loop into your –

Chris Gammell: Okay.

Mark Morin: Control.

Chris Gammell: Gotcha. Gotcha.

Mark Morin: Element, right? Whether you're using an op amp or something else. And you would feed the current back and then use your feedback to kind of dampen your signal and hopefully not dampen it so much that your rise time is useless. Right.

Dave Jones: Yeah.

Chris Gammell: But at the same time, you have to – We've replicated a 56K connection on this fibrillistic system.

Dave Jones: And look, this loop is really stable but – oh, geez, the rise time is half a second. Yeah. But it's really stable. Super stable. You know, it's not going to oscillate.

Mark Morin: Yeah. Yeah, exactly. And that's fine for some stuff. You know, some laser sources you just want to turn on and leave on all the time and that's fine. And then there's some things that you want to go fast and you have to have really good control over everything. So –

Chris Gammell: That is great. Yeah, that's interesting stuff.

Dave Jones: Now, you've been involved in lots of – you're doing production and qualification testing and burning and stuff like that as well.

Mark Morin: So we had this – well, I don't know if you want to lead into this RF design story here. Yeah. Sure. Why not? We have a –

Dave Jones: This is an electronic show.

Mark Morin: Yes, which is – I kind of wrote this out because I was thinking that this would be electronics focused as we could get here. We had this – well, in the new Q laser, the way you switch it, right, to get a pulse is there's an element called an AOM, which is an acousto-optic modulator. So it's a crystal. You put in 150 – Sounds like something that Ghostbusters would use. That's right. Something like that. It's on the proton pack somewhere. That's right. You know, you put 150 megahertz in it at a couple of watts and you get a lower optical insertion loss, right? So it's like you – it's kind of like a FET in electronics.

Chris Gammell: Could you – so – Wait. Optical insertion loss? Could you define that too, please? Right.

Mark Morin: So you're kind of like a CB radio, like the insertion loss of your cable to your antenna, right?

Chris Gammell: Gotcha. It's like the matching of two things, that kind of thing.

Mark Morin: Right. It's like 60 dB when you have no RF going into it and then the loss goes down to 1 dB when you put a bunch of RF into it. Do you get what I'm saying here? I think so. I probably not. If you're putting 10 dB in or 10 dBm in, right, and you have the AOM switched off, you're going to get minus 50 dBm out. And then when you switch the AOM on, you're going to get 9 dBm out.

Chris Gammell: So it's like a – it's a matching system. That's the idea. So it –

Mark Morin: It's more like a gate, to be honest with you. Oh, interesting. It's just – it's stopping photons from flowing through when you don't give it an RF signal. Oh, interesting. So that crystal basically controls the entire laser. If that thing shuts off for too long and all your pump diodes are running, all of those fibers will just burn themselves up because they'll have too much energy. Wow.

Chris Gammell: Do you ever feel like a wizard where you're like, oh, yes, today I was controlling the crystal and then I –

Mark Morin: If you start explaining it like this, yeah, it sounds like that. I put on my robe and my wizard hat. Day to day, it's a lot simpler than it sounds.

Dave Jones: And I walk around with a cane with a big crystal on the end of it. Sure.

Chris Gammell: Sorry. We're just super immature and have no idea what you're talking about. So this is what we do when we're confused and scared.

Mark Morin: So basically you drive this thing with a CB radio circuit, right?

Chris Gammell: Interesting. OK.

Mark Morin: So it's an old Motorola transistor, an MRF 553. Went obsolete in 2012, right? Uh-huh. And we started to try and – Everyone panicked. Oh, yeah. I'm sure they did. And everybody was at the same time trying to get their last time buys in. Yeah. Right. And we got some and we immediately started to redesign the board to find a replacement for this thing because it's kind of – I don't know if you remember these things, it looks like a cross where there's like four leads coming out.

Dave Jones: Oh, yes. Yes, I know. And there's a little circular package in the middle. Yep. Right?

Mark Morin: Yep. Yep. There's a hole drilled in the board that the thing kind of sits in the pocket.

Dave Jones: Yep. Uh-huh.

Mark Morin: So –

Dave Jones: Which is a very common package for RF amplifier type chips. You know, if you open up anything, you know, old school RF, yeah, you'll find these packages.

Mark Morin: Well, it seems like it used to be anyway.

Dave Jones: Well, yeah.

Mark Morin: You know, there's not a heck of a lot of variety on DigiKey nowadays to find, you know, some replacements that drop in. So we were scouring the internet for a replacement for this thing. We couldn't find anything. We ended up finding one by a, you know, a fly-by-night obsolete transistor manufacturer. Time to call the broker. Sure. We bought some from one of the big three distributors, you know, and tried it out.

Chris Gammell: Did you buy it in a darkened alley or anything like that?

Mark Morin: No, no. You'd be surprised how large of a company was carrying these things. Okay. And that's kind of what led Credence to the company that was making them, right? So we got them.

Chris Gammell: Oh, so this is like a secondhand, like somebody in their garage using an old chip manufacturing equipment.

Mark Morin: I think somebody bought the remaining guys from somebody that made these before, right? Right. And they just decided we'll crank out as many as we can before the whole thing falls apart.

Chris Gammell: I've always said that my retirement plan is to become a chip broker and just find a part like this and then just stockpile them and just charge someone like you through the roof kind of thing.

Mark Morin: Buy them at $1.20 and sell them for $10.

Chris Gammell: Oh, more than that. Come on, man. $10? Yeah. Let's stop there.

Mark Morin: Well, yeah, the gravy train never ends. That's right. Until it hurts. So this thing basically passed all of our qualification tests. We did dummy load testing and AOM insertion loss testing, built lasers, ran them for thousands of hours and burn in. And then we order a reel from this company, like Factory Direct.

Chris Gammell: Oh, no.

Mark Morin: I know what comes next. So we send those off to our CM, right? And they load them up on a bunch of boards and we get the boards in and absolutely nothing works. Nothing works at all. Yep. And you're like, what is going on here? We just tested these. No, no.

Chris Gammell: You tested the good ones.

Mark Morin: Right. Exactly. They apparently made like 25 good ones before they made this reel. And so we took one out of the old package we had originally ordered, soldered onto a board, snaps right on, you know, turns on. So we're all like, this is – if there's any kids out there listening, thinking.

Chris Gammell: Gather around, children.

Mark Morin: Electrical engineering is just the gravy train and it's really a collection of these stomach-sinking moments and then a few days of digging yourself out, you know? Uh-huh. Yes. And so we start going back and forth with them, like what happened. This is obviously different than the other ones we got. And it took them a while to admit it. But you could see that the package physically changed. The old one had a big fat lead on one side to kind of mark pin one. And the new one had like a skinny short lead to mark pin one. Oh. So they were like, oh, well, what happened is we bought the remaining dyes and packaging and then like there's no more of those anywhere and the tooling doesn't exist anymore. So we couldn't make them anymore. So we just started making it with, you know, a newer package. Whenever we had LUT, right. Right. But it changes the, you know, input impedance of the transistor. So the whole circuit is you'd have to adjust. So he's like, oh, yeah, just change some caps and inductors. I'm sure it'll work. You know, blah, blah, blah, RF stuff. Yeah. This is their sales guy on the phone, you know? Right.

Chris Gammell: Well, that's a little better than if the engineer was saying that. I mean, that would be something else.

Speaker ?: But.

Mark Morin: Yeah. I don't know if they had engineers to be honest with you. Oh, jeez. But it was a. Yeah. So we started over from scratch after a year of intermittent work on this. So who ended up bearing the cost for that? Oh, I'm pretty sure we ate it.

Dave Jones: Right.

Chris Gammell: You just ate it. What are you going to do, right? I mean, you're dealing with people that, you know. Don't care. Yeah, exactly.

Mark Morin: They're like, yeah, we sold you the real. See ya.

Chris Gammell: You know. That money order doesn't come back, you know. Right. Exactly.

Dave Jones: Because they weren't guaranteeing specs based on a data sheet, were they? They were just.

Mark Morin: Well, the kicker is they said, oh, if you look at the bottom of our data sheet, there's an asterisk that says all specs are subject to change.

Dave Jones: Change, yep, yep.

Mark Morin: And, you know, it's a lesson learned. You can't trust anything that's even written down on paper. The other thing is their data sheet didn't have the S parameters in it ever, right? So they're like, oh, yeah, it's an unspecked feature. That's the most important part of an RF transistor.

Dave Jones: That's funny. Can you tell us about the S parameters for those out there who aren't conversing with such things? Right. In general?

Mark Morin: So there's an input.

Dave Jones: Generic, because you can do an entire course on S parameters.

Mark Morin: There's basically an input impedance and then a reflection coefficient and then an output impedance and then like a reflected power coefficient. Right. So there's four S parameters, S1, 1, S1, 2, S2, 1, and S2, 2.

Dave Jones: For any particular device or server. Right.

Mark Morin: So you'll see this funny looking navigational map in the back of an RF transistor data sheet with a squiggly line drawn on it by some guy. And then he scanned it, right? And that's your... That's your... That's data right there. Yes, that's data. This is a data sheet, obviously. So there's the data. Exactly. Exactly. So they're kind of telling you on that little... That math how to match the transistor to the input and output, sir.

Dave Jones: Got it.

Mark Morin: This is a short version, right? Yep. Wow. That's all we need.

Chris Gammell: So the... So what happened? So then... Okay, so you're year in, and I think that if... I'm guessing you guys did great on the procuring enough of the existing ones. That's what happens with last time buys.

Mark Morin: So you were probably okay for a while. At this point, we called and begged for a second last time buy.

Dave Jones: That never works.

Mark Morin: No, it totally worked, Chris. It did? You'd be surprised. We got some. Wow. Not a full reel. We got a partial. Right, okay. Yeah. And then we went to a hole-in-the-wall Southeast Asian distributor that claimed they had some, and we bought a big reel off of them.

Chris Gammell: Yep.

Mark Morin: And then we were testing every piece. Well, we threw some on boards to see if they worked, and it was about a one in ten shot.

Dave Jones: And we started... Oh, no.

Mark Morin: Wow.

Dave Jones: We started testing... You didn't selectively test them, did you? You were that desperate.

Mark Morin: We started testing. We tested the first hundred off the reel in the lab just to see what the results would be, and it was like one out of ten. And we decided to just throw the whole thing in the trash. Wow. And then there was a huge stock of non-roast compliant parts out in the world, right? Mm-hmm.

Dave Jones: Oh, yeah. Yeah, it's tempting just to go, look, just nobody mention. Just use them. Just use them.

Mark Morin: You know, we didn't even talk about that. There's de-letting services, you know? And we were thinking about... Oh, okay. No. Buying leaded parts and sending them off to a de-letting service, but it just turned into like, how do you guarantee... You know, we're putting roast compliant on our data sheets, so we can't... Right. ...can't even mess with it.

Dave Jones: Yeah.

Mark Morin: So we ended up just probably doing what we should have done in the first place, going to a brand new transistor... Yeah. ...working out the whole thing again, and then at the end, the new transistor is a higher bandwidth than the old one, so it amplifies all this junk on our clock.

Dave Jones: Which can be a problem. Oh, yeah. Oh, yeah.

Mark Morin: So you end up getting spikes...

Dave Jones: All of that crap on your power rail that your old transistor was too slow to, you know, to amplify... Oh, it was beautiful. Natural filtering. Yeah.

Mark Morin: It's all coming through, right? The old one had a roll-off frequency of 175 megahertz, and we're using 150, so it works out perfect, and the new one was, you know, 0 to 500 megahertz, so it amplified all the sideband garbage. Oh, man. Yay! That makes the resultant laser pulse coming out of the AOM have little spikes in it, so...

Chris Gammell: So what, you had to put extra filtering in then just to make sure you cleaned everything up and stuff?

Mark Morin: We ended up changing the input clock to a pure sine wave crystal oscillator. Right. And that cleaned it up. Right.

Chris Gammell: Man, that's crazy, though. Like, that is... I did very similar work in past jobs as well, and sometimes you're just like, yeah, why didn't we just start here, you know? Yeah. Yeah.

Mark Morin: Well, because you didn't want to do any of the work in the beginning, right? No, exactly.

Dave Jones: You wanted to see if you could get away with it. Right.

Mark Morin: We have real products we want to develop and not stuff that's already basically done if we could just find more transistors. So, you know, there's no functional reason to get rid of that thing. It was fine. It always worked. So, you know, it was just we couldn't get them anymore was the only reason we wanted to change it at all.

Chris Gammell: Also kind of calls out just like how... I mean, this is not... I don't think it's unique to you guys either. It's just like how dependent we are on... This isn't even single source. I mean, it is single source, but, you know, just dependent on the specs of certain types of parts. You know, like it's just... They're like these linchpin specs effectively that, you know, we just... Our products are based on them and it's scary. It's almost like you have to do like a risk audit around each product and be like, that's the part right there. That's the one that will be screwed from.

Mark Morin: You know, yeah, you could. You know, you get... You'll never sleep at night, but yeah. Right. Well, you need a huge staff just to do that. That's true.

Chris Gammell: Yeah.

Mark Morin: The other... You know, if you have 1,400 parts on a board, which one are you looking at? That they could just change the spec on you and decide that you're done making lasers for a while, you know? Yeah.

Dave Jones: And we've joked about this on previous episodes, Chris, but the, you know, sometimes you'll get... In a company, you'll get promoted to component obsolescence engineer. You know? But it's... At some big companies, that's a real job. Yeah. You know? Because as you said, like there's hundreds and hundreds of parts and, you know, and with critical specs or whatever. And, you know, one of them goes belly up, then...

Chris Gammell: At the end of the day, a manufacturing company makes their money from making stuff. And if you can't make stuff, you're screwed. So, yeah.

Mark Morin: That's the bottom line. Yeah. But it's just... It's like that project is funny. How many times you go to the... You're at the top of the mountain and you crash back down and then you climb it up again. Oh, yeah. And then you crash and it's just like two years of basically, you know, intermittent work because you send a board out to spin and it takes a while to come back and all that. But, you know, when you get back on it, you're like, oh, yeah, we're totally, you know, totally smoked again. Just, you know, it was deja vu all over again like four times. It was pretty funny.

Chris Gammell: Yeah. So, you mentioned 1,400 components on a board. Like, what are some other common... So, if someone like Dave cracks open a device that has a laser in it and stuff like that, what are some common circuits that people would see in there? So, you were talking here about the diode and the drivers and stuff like that. What else is in there?

Mark Morin: Oh, you know what? There's kind of a... There's a Bob Pease book that I think most of the laser industry used to design most of the lasers. I could link it on your subreddit. Sure. Yes, please. Yeah. It's basically... There's a, you know, a photodiode transimpedance amplifier.

Chris Gammell: Okay.

Mark Morin: Yeah, yeah. And then there's a current source, right, to drive the laser diode, which would have an op amp as a control system. Maybe not in our particular systems, but I think most lasers you'd open up that you bought off of eBay would have that. Mm-hmm. There'd certainly be some kind of micro or FPGA in there kind of handling everything.

Chris Gammell: Yeah. And then, like, is it like an ADC as well? I mean, just a...

Mark Morin: Yeah, well, depending if you got a... If you have a built-in ADC and DAC in the micro, I guess you could get away without it.

Chris Gammell: Oh, I see. I see. Okay. Yeah.

Mark Morin: And then, you know, you'd probably have some TTL connector on the front that would control everything.

Chris Gammell: So this is on the receiver side, though? That's what we're talking about here? Is, like, transimpedance amp to actually detect and...

Mark Morin: No, no, no. That's in your laser, right? You have gain stages, and if one shuts off, you know, your electronics short out and the... Well, even, like, the customer signal goes away, right? To amplify, you have to shut all the laser diodes off before the laser kills itself. So there'd be transimpedance amplifiers monitoring photodiodes at every gain stage to kind of make sure that there's a seed signal coming in before you pump watts and watts in, you know, over... You know, this stuff all happens in hundreds of microseconds timescale. So if you're pumping a couple watts for a couple seconds, it's a game over if there's no input signal. Yeah.

Chris Gammell: Because it'll just rail, and then that'll just pump all that junk into the lines and stuff like that. Yeah.

Mark Morin: And that energy has to go some way. Right. That ASE term where it, like, automatic stimulated emission, it just kind of happens. And then once it happens... Overstimulation. You start bumping those electrons off with random wavelengths in random directions, and then boom, it's all gone. It's over.

Chris Gammell: Especially, yeah. And what happens if we... So if we did that, what would it look like? I mean, does it actually go up in smoke or what?

Mark Morin: I bet if you go on YouTube and... I will be doing this. And we look up self-laced fiber or, like, plasma ball, you'll see a bunch of videos where there's, like, a blue plasma ball traveling down the center of a fiber in, like, a coil, right? So it just, like, kind of looks like it's going around a racetrack like a sparkler. And it just... It's like... Cool. It sounds like a snake hissing, and you have the evolutionary reaction of that and, like, jump out of your shoes, man. It's scary when it happens. Right. But luckily, we don't want to get that often. That's a rarity.

Chris Gammell: So do you get spec then on, like, how fast you have to shut stuff down, or is it just kind of, like, just as fast as possible? I mean, is it... Oh, yeah.

Mark Morin: That's a thing.

Chris Gammell: That's, like, something you would sell to your customers. You'd say, we shut down and blah, blah, blah, blah, blah.

Mark Morin: Well, that's more like a safety feature inside the laser that would manifest itself as, like, a downtime for the customer, right? So you'd just be saying, like, hey, if something happens, the laser saves itself, and then, you know, you cycle power, maybe it recovers on its own, and then the world is right again. You know, nothing... You really want the customer to never notice any of this. It all just happens. Yeah, right. You know, and...

Chris Gammell: What you could do is just instead, you could have... You could go cheap on the components and then put, like, a printer on the back, and every time it blows itself up, it prints a coupon for half off the next one, you know? Sure. Yeah. That's something. Yeah. Yeah, that's marketing genius right there. Yeah, that's good. So you mentioned multiple diodes as well. So are those photo diodes you're saying, or are there actual multiple laser diodes?

Mark Morin: There's both. For every gain stage in a fiber amplifier, you'd have pump diodes to...

Chris Gammell: And what are those?

Mark Morin: It's a laser diode, and it's usually at a wavelength that is... That the rare earth element in the fiber will absorb efficiently, and not necessarily be the wavelength that would transmit down the fiber as data or material processing light, right? So this is off-band, I guess you'd call it.

Chris Gammell: And that's just to juice up the energy in the line? Is that kind of the idea, or what?

Mark Morin: Yeah, well, it's providing all of your amplification energy for the laser.

Dave Jones: That is so crazy.

Mark Morin: It's pretty crazy.

Dave Jones: Yeah.

Chris Gammell: And so how do you actually sum all of these together then?

Mark Morin: That's the very highly profitable, patentable, optical component that everybody wants to be able to build, right? This, whatever you call it, the combiner, I guess you'd say.

Chris Gammell: Are there any... I mean, patents are public. Are there any public patents we could go look at to figure it out, or no?

Mark Morin: Probably. You're not going to tell them to us. I've either worked for the company that did it, or we have an NRE with a company that's doing it now, too. So I don't want to...

Chris Gammell: Okay, no, that's no problem.

Mark Morin: Spill too many beans, but...

Chris Gammell: Well, just to give us an idea, though, this is like a separate element outside of the actual circuit board. Is that the idea, or...?

Mark Morin: Oh, yeah. So the laser diodes kind of are soldered to the circuit board or wired to the circuit board, depending on the system. You pump the current through those. There's a fiber coming out of the laser diode, right? There's an optical component inside the laser diode that launches the light out of the die into the fiber, right?

Chris Gammell: Sure.

Mark Morin: And then that fiber splices onto this combiner component. And that... There's, you know, a bunch of fibers hanging out of one side and one fiber coming out of the other side, right? Oh. So you have the one single-mode fiber on either side for your signal, right? That's the central fiber you're going down. And then you have all these multi-mode fibers, which is like a larger core, so it's easier to launch the light into. And you would splice your laser diodes onto the multi-mode fibers, and then somewhere in the combiner, all the magic happens, where it takes the core light of the multi-mode fibers and puts it into the cladding of the single-mode fiber so that it can then reflect down to the core where the rare earth element is to get... Wow. ...to turn into amplifying energy, right?

Chris Gammell: Yeah. That is so cool, though. Like...

Mark Morin: It's pretty cool. Yeah. Man, that's...

Dave Jones: Can you explain the difference between multi-mode and single-mode for those playing along at home?

Mark Morin: Well, it depends on the wavelength. There's a number of modes in a fiber. Like, the wavelength is actually the, like, spatial... The core size of the fiber can only hold a certain amount of a wave, say. And if your wavelength is a lot shorter than the core of the fiber, then you can have more than one wave in there, right? So if you're multi-moded, it means that if you took this fiber, like, with a red, like, pen laser going down it, right? Mm-hmm. If you had a multi-mode fiber and you shot a red pen laser down it and then you look at it, the output of the fiber on a piece of paper, right? So you can imagine there's, like, a divergent cone of light coming out of the core of the fiber. You would see, like, five spots on the paper because the fiber core is too big for that wavelength. So that's what they call multi-mode fiber, right? So I guess the larger core diameter fibers are just called multi-mode because they're so big they will have multi-modes in most wavelengths. And then the smaller ones, you're technically single-moded or maybe just not single-moded depending on the actual wavelength of the laser you're launching into it.

Dave Jones: And from memory, because this was in a galaxy far, far away from me, single-mode fiber had lower loss versus distance. Yes. Generally. Generally.

Mark Morin: Because you're not interfering with your own...

Dave Jones: Yeah.

Mark Morin: I don't know if it actually interferes or if there's just enough room in there for it to refract. Because of, like, angles and stuff like that? Yeah.

Chris Gammell: The idea, like, how it bounces to the...

Mark Morin: You know, we're getting into the territory of fiber where I'm more in generalities than knowing the actual specifics of how you're... I mean, we're masters of hand-waving here, so...

Dave Jones: I have some experience. Yeah, but multi-mode... So everyone might be asking, well, why don't you just use single-mode all the time? It's because I think the interface is more critical. Right. If you're trying to, you know, like, you use that fiber on the back of your PC to connect your audio. Right. You know, whatever the TOS link, whatever it's called. Whatever it is. Yeah, yeah, yeah, the SP-DIF thing, right? That's going to be multi-mode, I believe. Yes. Because it's just much easier to align. It's much less critical. Right.

Mark Morin: And you're just trying to launch, A, like it's audio, so the bandwidth is nothing.

Dave Jones: And the distance is meters, tens of meters. It's not, you know, 100 kilometers.

Mark Morin: Right, so you're optimizing for the launch angle of a crappy red LED, right? Right. And the core diameter is a lot more critical in, like, an optical cable than anything else to get as much light as you can.

Dave Jones: That's right. Yeah, that's why we use multi-mode fiber in our seismic cables for data transmission. You know, they'd be 150 meters long. But you've got to join these things, these big cables together in the middle, on the back deck of a boat, tossing around in the middle of the North Sea with all sorts of gunk and crap and everything in there. It's like, eh, you can't exactly get the ideal mating that you want every time, you know? Yeah. Yeah, we need, like, the most, the least critical optical connection we could get. So, multi-mode it was. Right. And we didn't care about loss down, like, a single 150 meter length where it would be repeated, so.

Mark Morin: Exactly. Exactly.

Chris Gammell: So, what about, what about, I mean, so, like, kind of what Dave's talking about here, what about connectors? I mean, so, when, if we, again, if we're looking at one of the boards that you work on, are there specialized connectors or is it literally, like, shoving a piece of fiber into a plastic housing and that does it?

Mark Morin: There's, for the most part, no optical connectors on the board. We'd use electrical connectors to go to whatever transducer component. And, like, the front panel of our laser, you know, is a DB25 and a DB9. So, it's not, we don't really have, I guess, in, like, telecom, if you were launching the signal down the fiber, you'd probably have either a connector or you'd solder the seed diode right onto the board.

Chris Gammell: Gotcha. So, really, you're making boards that then are driving transducers. Is that right?

Mark Morin: Yeah, or we're reading back the photodiode light from a, you know, we're reading back some signal tap from a photodiode.

Chris Gammell: Gotcha. And that also has its own interface and stuff like that. Right. Yeah.

Mark Morin: And then from that, we make all these decisions on what the laser needs to do at a given moment. Hmm.

Chris Gammell: Are there interface problems there at all or no? Or is it pretty straightforward at that point? Because I imagine these are high-frequency signals as well, right?

Mark Morin: Somewhat. It's surprisingly low. You know, everybody looks in an oscilloscope and says, oh, well, when we look at the pulse here, we want to do this. And you're like, well, that's nice if you have an oscilloscope front end on the board to read it with. But that would, you know, be a $100 chip at least, right? So, you know, we are manufacturing things and the cost is part of this. So, you can't go – your bandwidth of your input circuit is kind of limited by dollars anyway. So, you don't really have to do all this ultra-high bandwidth. Yeah. You know, cabling or anything. But if we had to do something like that, I'd probably just solder right onto the board with the optical component, I'd think.

Chris Gammell: But, no, I mean, that's nice that you don't have to even think about that. So, I mean – or not think about that. But it's less of a manufacturing problem for the board then and more of a system-level integration type thing.

Mark Morin: Oh, yeah. There are fiber connectors. I mean, if you've ever seen these telecom amplifiers, they're covered in them. It's crazy. Yeah, I haven't. Oh, okay. Well, you know.

Chris Gammell: I'll try and add a link to something. I don't know. Maybe you could send me a link to a generic one after we – Oh, sure. Yeah, fair.

Dave Jones: And, like, ours, for example, on our seismic cables, we effectively manufactured our own custom optical connector. You know, we would – yeah, we would have, like, a brass ferrule that the two fiber cores, which we would, you know, terminate, we'd use a termination machine, and then you'd polish them down. You know, so you'd – so the fiber itself, the actual glass fiber would go through, like, a ceramic, you know, ferrule, and then you'd actually glue that in there, and then you'd polish. So the fiber's, like, sticking out of this. Maybe I can get some photos, but the fiber is sticking out of this. Then you'd actually polish it down. So you'd put it in a little jig, and you'd actually polish it on this really fine, you know, sandpaper polishing stuff. Sure. And then you'd feed that ceramic thing into a brass ferrule, which would be inside our custom connector, which would have not just the optic, it would have the, you know, all the power cables and all the data cables and everything else in there as well. So it was combined electrical and optical connector, and you'd join those together, and bam, it just mated the two fibers together. And there'd be a little spring mechanism, too, that would push the two ferrules together. Yeah. Yeah. That's cool. Because, you know, because we had to have the spring thing, because otherwise your tolerances would be too tight. You know, you wouldn't be able to match it up every time. And, you know, when this thing's waving around in the ocean, you know, lots of stress and strain on it, you want the fiber to take that out and the two fibers to sort of gently push together all the time. So, yeah. So a lot of those things are totally custom like that.

Mark Morin: There's definitely, that sounds like what they call like an FC connector, if it was just fiber.

Dave Jones: It was an FC, yes. Yeah. Okay. Yeah. There was FC. There's FC and SC. Yes.

Mark Morin: That's a square one. Yep. Yep. And then my favorite's the S2000, which has got like a little doghouse that automatically opens when you connect them.

Dave Jones: Oh, yes. Yeah. It's fantastic. Yeah. It pops up. So it keeps dust and crap and crud out.

Mark Morin: Beautiful, but they're like $400 a piece and nobody ever wants to use it. Right. Yeah. Yeah. Yeah. Yeah. So you look at it and you're like, now somebody put some thought into this thing, you know?

Chris Gammell: So you have a theory that you – Oh, yeah. One last thing that we want to talk about this and then one other thing, but what is the theory that you have? Right. You've sent us this before.

Mark Morin: We've been talking about this Q1 theory where the most commonly failed component on everybody's board has to be Q1.

Chris Gammell: So that's how you actually will just go straight to that or what?

Mark Morin: Yes. And then I'll justify it. Right. So the – you would say like why isn't it R1 or actually why isn't it Q9? Right. Because every board has – the annotation starts at 1. Right. So Q1 is on every board that has a Q. Right.

Chris Gammell: And it's going to be the transistor that breaks. Is that kind of the thought here?

Mark Morin: That's where I'm going. Yeah. Like the – if you're using a discrete transistor, you're usually handling a little more power. You know, if you didn't have that much power, you'd have a – you know, an opto output or something like that.

Dave Jones: Because it would be the most critical part too. This comes from the design side of things. When you're designing something, you do that critical part first. So it's probably going to be given the designated Q1.

Mark Morin: And this is the left to right up to down annotation does that to you, right? You draw your most critical circuit first and then you annotate it. And that page has all the low annotations on it. So, yes, this is my theory. And then, you know, it just – it keeps popping up because we're like, well, we're going to start annotation at like 11 or something. So we just don't have Q1s anymore or just keep a tally of our annotation and just, you know, okay, we used up to 83 on the last board, started 84. Just so we never – Right.

Chris Gammell: A running tally. Yeah.

Mark Morin: Right. You'll have somebody be like, oh, Q1 failed. And you're like, which – like again? And then it's a totally different system, you know? Yeah. Yeah. All right. So you get this – you get the – like the hairs stand up on the back of your neck every time you hear Q1 after a while.

Chris Gammell: Yeah, especially if you have to go and replace it with a now obsolete component. You're burning another one, right? Right. Yeah, that's true.

Mark Morin: And it's never good. But I always thought that was – it's kind of like the – they say most accidents are within two miles of your home. It's because to go anywhere, you have to drive the first two miles. So every time you get in your car, you're driving those two miles. It always seemed like a dumb stat to me. And it's the same – like same idea with the Q1 failure. It's like it's always going to be Q1. It's statistics, yeah. Yeah. You'd figure. Yeah. Like a resistor. There's so many chip resistors on a board that they're automatically going to diffuse themselves. But yeah, this is – that's my running theory about the Q1 failure.

Chris Gammell: I like it. So there's one other thing. Actually, the reason that you and I got in contact with each other – Oh, yeah. You were mentioning to me about some of the jobs that you guys are hiring for. Sure. And I was like, well, you've got to tell us about lasers if we're going to talk about the jobs too. Oh, yeah. But you guys are in Connecticut. We are in Connecticut. I say it as a Clevander.

Mark Morin: So for the – yeah, if you want to flee the cleave and come on out here, then we'd be more than happy to interview you. But I don't know what your statistics for Connecticut or Western New England listeners are. But I figured if there's a few, then it's worth throwing these links out there. We have an opening for an electrical hardware engineer. So you'd be using like Altium to design circuit boards for lasers. If this last conversation has been thrilling to you, by all means, give me a call.

Chris Gammell: Remember how they haze their new people. Don't forget that. Right.

Mark Morin: You will be strapped to a table with a laser point. Laser. Yeah. Yeah. Get ready. Don't worry. You won't absorb it, whatever the frequency is. Yeah. Well, that's the joke, right? Yeah. Right. You find that out.

Chris Gammell: Where are you guys in Connecticut? I mean like where –

Mark Morin: Oh, so we are two and a half hours north of New York City, I guess. Mm-hmm. Okay. It's the closest big landmark. We're north of Hartford. If you've ever been to Bradley International Airport, we are – Nope. We are within – you could throw a snowball and hit Bradley Airport from here. So we're on Route 91, which is like our major highway down here. And anybody from the area will know exactly where I'm talking about. But we're in the town of East Granby. If you just want to go on Google Maps and see how far your commute would be if you applied for this job.

Chris Gammell: Or how far it would be from New York and Boston. Right.

Mark Morin: Yes. That too. There's clusters of engineering out in Boston, obviously.

Chris Gammell: Yeah.

Mark Morin: And out here it's a little more sparse. But there's a lot of big defense contractors up on 91 here. Oh, interesting. There's a Northrop Grumman and what – it used to be Hamilton Sunstrand. I think it's Hamilton something else now. And I think there's an electric boat down near New Haven.

Chris Gammell: Yeah. I know that former multiple guests of the show, Greg Charvat, lives down south there now. I'm not sure where. I know he lives on the water somewhere now. Sure. But, yeah. Yeah. It's interesting.

Speaker ?: It's funny.

Chris Gammell: It must have been some kind of tech center at some point, right? I mean, because of, like you said, all the defense contracting stuff.

Mark Morin: There's a bunch of it. I don't know if it just pops up on every major highway or what. Because it's just – it's ever-present, it seems like.

Chris Gammell: Yeah. Well.

Mark Morin: But, yeah. And then we have another opening for an embedded software engineer, you know. Wrong show. Wrong show. Yeah. Sorry, guys. Sorry. If anybody wants to make a career change and stop committing all your sins to copper and then just be able to change your switch statement to fix a problem, then, you know, you could hop on over to embedded software.

Chris Gammell: Ain't no obsolete components in embedded software. No. No.

Mark Morin: Yeah.

Chris Gammell: Well, that sounds cool, man. That's a lot of interesting stuff. I mean, like I said, I have very little knowledge of the whole optical side of things, but it's big. I mean, like there is a lot of need for this stuff. Oh, there is.

Dave Jones: Yep. And there's a huge amount of science and physics behind it that, you know. Yeah.

Mark Morin: It's certainly interesting stuff to work on, that's for sure. Yeah. Well, thank you very much, Mark, for joining us. Thank you guys for having me. This was great.

Chris Gammell: We encourage people to check out the lasers and check out the company. That sounds cool. Thanks, Mark. Thank you. Cool. Thanks, mate. Have a good one. Catch you next time.