MEP EP#17: System-in-Package (SiP) Platforms with Greg Sheridan

Podcast Transcript

Host 1     00:09

So welcome to the macro fab engineering podcast. I'm your guest, Greg Sheridan.

Host 3     00:15

And I'm Steven Craig and I'm

Host 1     00:17

You forgot to do where your hosts Oh, you're right. You mess up the intro was me. So anyways,

Host 3     00:22

We're your host, Steven Craig and Parker Dolman. We're going with that. Yeah. That's That's it. It's done. It's fixed.

Host 1     00:29

Good. Yeah. Usually the guest Actually, no, the previous guests have always been perfect. So you did pretty good, Greg. All right. Yeah. Good. Steven messed up.

Host 3     00:38

Yeah. Well, I guess I'm a guest today too. Okay,

Host 1     00:46

So Greg is with

Host 1     00:49

Octo Octavio CEVO. System cable systems. Yeah.

Host 1     00:53

Which makes the BeagleBone on chip.

Host 1     00:57

That is that is one way of describing it. That is how we like to market it. It's technically the OSD 3358 Dash 512 M dash bas.

Host 1     01:08

That's actually impressive, because that's not written on the sheet either.

Host 3     01:10

Well, yeah, that was purely for memory.

Host 1     01:13

I that was that was my part number. And there was a lot of discussion internally, I got a lot of pushback from making it that long, but what you created that part number. Yeah, that was that was my stream. Those actually mean something.

Host 1     01:27

Nothing is that part number? It's like a. It's like a microchip part number four pick. Yeah, they drew attention just keeps going. Yeah.

Host 1     01:37

Well, again, there's hopefully this is the first in what will be very many parts. So we need all that. Yeah. But yes, I mean, so that basically what it is, yeah, it's we've taken the core of the BeagleBone, black, single board computer and put it onto a single chip.

Host 1     01:55

Cool. And the so How do y'all do that.

Host 1     01:59

So what we've done or what our technology is, and is really, it's called our we're calling it System in Package, also known as multi chip modules. But basically, what that is, is we take the die from the different components that are on the BeagleBone, in this case, so the am 3358 processor from Texas Instruments, as well as the P MC, and the LTO and the DDR memory. And we take those di, or guests just for the three ti components and the DDR is package. And we put those onto a substrate. So we've designed a substrate that we put those dye on. And then we actually bond use bond wire. So I see type manufacturing techniques to bond those two, Bond those chips out. And then we add in all the passives and all that other kind of stuff as well. So it's really versus like an SOC, which is a single, die, single die thing that has all of the components in it. What we've done is we've actually taken individual die, that are more specialized, and packaged those all into a single single device that is hopefully very easy to use.

Host 1     03:19

So do you get these dies then for like the actual process, you get them from TI.

Host 1     03:24

So yeah, we we come from TI we have one of our founders was a Principal Fellow at TI he was actually one of the engineers that did the Speak and Spell back in the back in the day. So he's, you know, decades with ti so we have a lot of relationships. But yeah, we were able we getting wafer wafer wafer from TI and then saw them up and package them all

Host 3     03:52

Up. Okay, cool.

Host 1     03:54

I just imagine like a hacksaw and wafers from

Host 1     03:58

Apparently they're diamond drills or diamond saws diamond tip saw? Yeah.

Host 1     04:03

I figured it'd be a lot more involved than just a hacksaw.

Host 1     04:06

Yeah. Same concept, just a little bit more precision and more

Host 1     04:11

Precision and clean. Right? Yeah.

Host 3     04:13

The the substrate. What does that entail? Like what is it? Is it a PCB? Kind of? I mean, yeah, if

Host 1     04:20

You look at it, it's, it's really it's fr for board. But what it is, is we've taken all the complexities that would go into, you know, laying out these components separately. Sure. On your own board, and we've now done it in here. So this is a six layer, basically PC board, yeah, that were given you in a single BGA. So now when you go ahead and build your own board, it you can get down and we've got reference designs, or there's reference designs out there using this chip that are only two layers.

Host 3     04:51

So you take a lot of the difficult work out of the equation,

Host 1     04:56

That that's the goal. I mean, that's that's kind of what Teva was founded on and that's definitely why we did this first ship was to, yeah, help bring pull out some of the complexities and some of the hardships that go into electronics make it easier for people to get their hands on and use these kinds of powerful processors,

Host 3     05:15

Well, it's not trivial to if anyone has laid out RAM before and write at really high speeds, it's not as simple as just connecting them. And the placement of every component has an impact on the operation. So right, you guys take it all, and put it on this one board? Right? That effectively ends up being a considered a chip in its own right, right.

Host 1     05:41

I mean, yeah, to the, to the outside world, right? It's supposed to look, I mean, our goal is to make this thing look kind of like a microcontroller. So bring, make it as easy, you know, people with, like, Arduinos and stuff like that, right? They've taken Arduinos and built 1000s of different projects, yeah, around them, because microcontrollers are easy to use, and easy to design,

Host 3     06:02

And are doing it makes it even easier. Right, exactly.

Host 1     06:05

But you know, you look at the different single board computers that are out there. People use them, but very few people actually go and create their own projects around that. Sure.

Host 1     06:14

Yeah. No one, no one makes a embedded Raspberry Pi device. Exactly. They just glue the Raspberry Pi on to the project, right.

Host 1     06:22

So and one of the reasons we believe that is is because of the complexities, like you just mentioned, of connecting the DDR right to the processor dealing with power sequencing, and then all the different passive components that go along with that.

Host 3     06:38

Well, all the all the prototype that you would have to do prototyping, and all the manufacturing runs you would have to do to figure out these issues. You guys have done that. Exactly. And so it's a drop in port for Yeah, exactly. So

Host 1     06:51

Our goal and our hope is right, that people will take the BeagleBone platform, you know, continue to develop on that, use that as their prototypes, then when they're going to production, they can just drop in our device, and focus on really what makes their product, you know, special. So you don't have to worry about the tedious DDR interface of the power supplies or placing 140 Different decoupling capacitors and stuff like that, right? Yeah,

Host 1     07:19

The thing I liked about motes actually got I think I got the question answer when I asked on Twitter, was the fact that it runs the same kernel and build of Linux that the BeagleBone Black does. So you don't have to compile anything special, right? You can actually actually use what you designed on the go and black and over to your design.

Host 3     07:40

People can port their designs that they already have over to this

Host 1     07:44

Exactly. And we've been working with BeagleBone since the beginning. I mean, they're very close partner of ours. And, and yeah, so that this board, this device was really designed with that in mind is that you start with a BeagleBone. And that's your development platform. And then when you go into production, you know, you can just drop this thing down, use the same code and just go.

Host 1     08:05

That's awesome. That's really cool. I got a question, since it's pretty much a encapsulated PCB board. Correct. That's, uh, you're only down a little bit. How so? How well does it handle since mostly a maker are geared towards makers and small designers? That kind of stuff? How old does it handle multiple reflows so I mean, it's a BGA so pretty much only got a couple shots at it, right when you rework one of these things, right.

Host 1     08:34

But now the nice thing about this device is that we've intentionally designed it to have very wide pitch, a very wide pitch BGA so it's a 1.27 millimeter

Host 1     08:46

It's actually one of the largest pitch BGA I've seen. Yeah, it's

Host 1     08:49

Big and that's we get that a lot. It's like wow, I can't believe it so big but we did that on purpose specifically to to help out the maker type market and the people that might not have the most sophisticated manufacturing abilities right so

Host 3     09:05

Toaster oven soldering.

Host 1     09:08

That's one of the things I'm I want to get a video together on is to actually show that yes, you could reflow this in a toaster oven. Yeah, no, I haven't died yet. I'm pretty sure it can be done. I haven't done it yet.

Host 1     09:18

I can do it on a skillets scale. I'd be impressed.

Host 1     09:21

I'm going to start with a toaster oven but no I mean this it's a again it's all pretty much standard. I see packaging technology again another one of our founders is a did 2025 plus years at TI principal engineer in the packaging for TI so I mean he's

Host 3     09:41

So he knows yeah

Host 1     09:44

Yeah, he's he's assured me that this thing is just as good as just about any other IC you'll you're going to put on your wall that's cool.

Host 1     09:50

Yeah, I wonder how long it's gonna take till we see someone who has it super glued upside down on some perfboard and with Oh, he did bugs in someone dead bugs it that would be pretty cool. Yeah, I can't wait till that picture pops up on Hackaday Yeah, so.

Host 3     10:03

Okay, so we talked about thermal reflow. But what about the thermal performance? Because you got everything now in a really close proximity, you got this thing screaming that. I don't want

Host 1     10:14

It. So it runs, it runs at a gigahertz. Okay. Yeah, um, but I mean, really, if you look at it, none of these devices are particularly high power devices, right? And the am 3358 is a fairly low power device for the type of performance you're getting. The P MC in here is a couple of AMP. But you know, really, it again, you guys, I have an example here a clearer picture that you guys can see. But I mean, you see, it's fairly well spaced apart. So again, given that we chosen to make the balls as big as they are, yeah, as wide as they are, we actually have a lot of space. So we actually have a lot of area to dissipate and dissipate the heat. But again, you look at all these individual components, none of them really dissipate that much heat by themselves. And, you know, we've run plenty of thermal models and stuff like that. And, yeah, we're not concerned. Now, as we get in, hopefully, you know, as we get into higher performing processors, like if you look at BeagleBoard, they've recently announced BeagleBoard x 15. And the processor itself on there has a heatsink. Okay. So, you know, if we're starting to put stuff like that in here, then yeah, then thermals become a lot more

Host 1     11:24

Could put a big thermal pad in the middle of it.

Host 1     11:26

Yeah, yeah. I mean, they're the, this is very, there's a lot of really cool, exciting stuff we can do with this technology. Again, this is kind of the first one to get out there, you know, we really want to enable the people that are playing with BeagleBone, black to really get out there and be able to make their own product. But yeah, we're very excited about where we can go with this stuff.

Host 3     11:48

I just find that super cool that now you have that much power in your hand with something that you can just, you know, plop down into your into your design. And just, it's not that you don't have to worry about the thermal issues that go along with it. But I mean, it's one of those things where you're not you're not the designer themselves, being a maker, whatever can just plop it down on the board and not really have to consider that because it's already taken care of. That's just I don't know, I find that super cool.

Host 1     12:15

Yeah, yeah. I mean, again, the goal is to make it as easy for people to design with as possible take as many of the different variables out. I mean, obviously, this is still a complex device. I mean, just an eight processor just still has a lot of complexities that come with it. But hopefully, we've made it easy enough where again, somebody can doesn't feel intimidated with it and can take it and really go build their own.

Host 3     12:36

Well, let's, uh, let's recap that. What's the part number on it again? Is there an annotated partnumber?

Host 1     12:44

So right now we only have one device. So if you go search for Octavio systems, or the OSD 3358, okay, you know, you'll you'll find this device. This current version has, again, the gigahertz am 3358 processor has 512 megabytes of DDR three memory, the power supplies and all that kind of stuff, we will be releasing a version with a gig later this year. And then, you know, we'll also be going and looking to fill out the processor portfolio as well with the other versions of the TI processors in the same family. Cool.

Host 1     13:17

So the gig would have a 1024. And the package, I think I'm actually

Host 1     13:21

Just going with one G

Host 3     13:22

One G one, G, O G. G, yeah. What's the size of the package?

Host 1     13:29

So it is a it's a 20 by 20 BGA. So 400 balls, okay, which is fairly, can be fairly daunting. But again, we've hopefully made it big enough that it's easy to use, but it's a 27 millimeter by 27. millimeter.

Host 3     13:43

Okay. Just over an inch button. Yeah. Yeah.

Host 1     13:46

I mean, if you stick a quarter on top of it, it's really it's about the size of a quarter. So yeah, it's really not that big. Yeah. I mean, again, if you compare it to, like you said, a BGA. You're used to seeing I mean, this is a a monster.

Host 3     13:58

Yeah. We placed a 512 pin. That was on a point five pitch the other day and right, and it was a beast. Yeah. Right. So yeah, no, this looks massive in comparison. Right. Alright. Next

Host 1     14:11

Question is. So it's going to be, you know, just plopping down, not have to worry about it, right, like thermals, and that kind of stuff. But what about like, the guys that are actually starting to do small volume stuff and actually have to start worrying about like FCC and CE certifications? Is there. Y'all have like tech docs for that kind of stuff or recommendations for like making sure this device isn't going to radiate and right.

Host 1     14:37

Alright, so we haven't done any kind of FCC testing on this device yet. Again, you know, this is this is not going to be a radiator. Yeah. The traces that come out, are tiny. Yeah, I mean, the traces on the board are so much tinier than what you're actually going to have on a normal PCB if you're doing something like this, right. So I mean, and really what's going to affect it is the traces that are coming on board in your, in the designs that people are doing. It's on our checklist of things to go through and just to kind of prove to everybody that it's okay. Yeah. But we don't expect any any kind of problems again, mainly because the what would be antennas on here are so much smaller than what would be on a PCB with this same design, actually

Host 1     15:25

Speaking that what's on that six layer PCB? What is the trace width? Average trace width?

Host 1     15:30

Whoo. All right. Now you're, you're getting over my head here. I can

Host 3     15:34

Tell you that really, really small, relatively small.

Host 1     15:37

Really tiny, really tiny, but yeah, I bet

Host 3     15:42

You it's six mil or less?

Host 1     15:45

Oh, yeah, it has. I know, there's definitely some in there in the Quran range.

Host 1     15:50

Yeah, definitely got bond. It's got bond wires on it. Right. So we'll actually we'll have a picture of the epoxy sample that Greg here has.

Host 3     15:58

Oh, you can see it's got a window to see everything inside.

Host 1     16:01

Yeah, you can actually see the bond wires, which is really cool.

Host 1     16:04

Yeah. That's cool. You got a picture that can actually see the bond wires? Yeah. Nice. It's really cool. Looking. Nice.

Host 1     16:11

And I guess, so What are y'all looking forward to go to the future, then? You can't just keep like, you know, this make it faster or more RAM. Right? Yeah,

Host 1     16:23

I mean, we can, right? I mean, the, you know, again, we are we look at ourselves as mainly system integrators, okay, so we're not going out and designing silicone or anything. So we're not going to make a faster processor, or something like that. What we are looking to do with this particular device is look to see what else we can integrate into it. So what else can we do that's going to make a designer's life easier. So there's a couple of things that if you look around, you know, that are going on most designs that people really want in systems, you know, we're looking at integrating those types of things. Beyond that, you know, we're obviously looking at different processors, different families, different subsystems that we can build. And again, what we're really looking to do is to make the entire system development process as easy as possible. So, you know, we're not just focusing on processors, memory and stuff like that. We're looking to other subsystems as well, you know, power subsystem, you know, different analog front ends and things like that. So what are these different blocks that, you know, couldn't be kind of a pain in the butt for designers to deal with that we could simplify and make it easier to make kind of standard building block components for

Host 3     17:40

Them? Well, and that's cool, because you kind of see that in the industry nowadays, where it's less of a game of who can make the fastest thing anymore. It's more about what features what peripherals, what can I provide as a semiconductor manufacturer that makes your life easier?

Host 1     17:58

Yeah. And then what's cool about what we do versus being a semi cut semiconductor manufacturers, we're actually able to pull from like the best process technology from the different areas, right. So as a semiconductor manufacturers typically looking like, Alright, how good I've got a microcontroller. I want to build Yeah, so it's an SOC. So you've got the one di so what's the best ADC I can put on here? What's the best power supply? What's the best Core i can put all on here? But I've got the constraints of this silicon process. Yeah, this one silicon process, what we're able to do is we're able to go pick the best ADC from the best ADC silicon process, the best processor from the best processor, so it can process the best power supplies, so on so forth, because they're always going to share the same process, right? I mean, if you look at where semiconductor technology has gone, has gone, you know, you had Moore's law really driving it. And that's great. But while Moore's Law has driven it, they've it's also driven. Further specialization in the different process technologies. So your best ADC process no longer is compatible with the best process processor? process, right? So really, if you want the best performance out of your processor, and your ADC, the only way to do that is with two discrete chips. Oh, that's cool. So what we're able to do is we're able to give you the same performance of two discrete chips in a single package. Wow.

Host 1     19:28

That's, that's pretty cool. Yeah, that's all fine. We

Host 1     19:31

Think so we're pretty excited about it. And yeah. Neat. Just speechless here.

Host 1     19:41

I kind of really want to build something with

Host 3     19:43

Word parking and come up with neat. Well, actually, you know, I realized, Greg, we didn't we jump right into right into the chip, I guess. Can you tell us just a little bit about your background, and what you do? Yeah,

Host 1     19:59

Yeah. So, me personally, so I am the sales and marketing manager for Octavio systems. We are a small startup based out of Austin. But with offices around around Texas, we've got a large group here in Houston. And we were founded by a couple of X ti guys, like I mentioned, with a lot of experience. And really, once they kind of retired, they saw this need and basically needed a complement to the Moore's law, like I said, Right? Where you can give the integration, the best performance with the best integration, and how do you do that it System in Package. But this technology has been around for decades, but it's only been in the hands of, you know, people like Apple or people like TI and done for very specific purposes. Sure. So the whole concept of octava is how do we take this really cool technology that's very powerful? And how do we put it into the hands of extremely innovative people and just give it to them and see what they can go and create. So that's, that's what we try to do, you know, is try to make electronics easier for people to use and be creative around. And again, we leverage this system and package technology to do that. Neat. Ah,

Host 1     21:27

So Greg, so what do you do at

Host 1     21:31

So I'm sales and marketing. So anything on the website, anything that goes external to, to the company is me. So if you see spelling errors on the website, or any of the agonal I do the social media, I do the flyers I do the podcasts I do I do anything that anybody would actually see.

Host 1     21:50

So actually, he was this morning while I was actually typing up everything for for this podcast. And I'm like, You know what, I never actually looked to see what Greg did. I looked up, I'm like, Oh, no sales. Then the next line is like, you know, BS in computer engineering. So I'm like, okay, good. He's gonna fit in just right.

Host 1     22:11

I'm a I'm a computer engineer by schooling computer nerd at heart. Went to went to Georgia Tech, spent a lot of time and money there got a BS and an MS in engineering. And then and then I decided to go to the dark side, and start doing more of the business stuff. So I went and got an MBA and then did sales with Texas Instruments for a couple years and then, you know, found, found the startup life, which is a lot more fun, a lot more exciting. And they let me play with the electronics again, they let me get back into the weeds of stuff.

Host 1     22:44

The great thing about start startup kind of stuff is you just get to do lots of things. Yeah.

Host 1     22:48

Yeah. So yes, you know, what is it? I do? My technically what I do is, is that but it's whatever, whatever's necessary to be done, right? Yeah, exactly. Yeah. Yep. Very cool. Neat. Today, today's word kid is

Host 3     23:03

Brought to you by

Host 1     23:06

Uh, so, uh, Steven, do you have any more questions for Greg? I think I'm done.

Host 3     23:11

I I think we covered a lot of territory there. Yeah. All

Host 1     23:14

Right. Good. Hopefully I didn't, didn't bore your audience here. No,

Host 1     23:17

It was. That was that was that was awesome. Cool. So yeah, the RFO section. So there's this really interesting project that's on on Crowd Supply right now. It's called the ello computer. And it's basically a low power computer that runs basic, that is built only out of thin PCBs. The entire thing, the entire thing is made a PCB so they don't have any kind of molds or anything for this thing. So even the keys are PCBs, that flex and contact another PCB to make the keys work.

Host 1     23:54

Okay. So I thought what's actually running? What's what's behind it? What's the process? It's

Host 1     23:58

A it's actually a pic. 32 MX? Okay. I think it's the MX Yeah, check. Yeah. And it runs.

Host 1     24:03

And it just runs basically old. Runs old

Host 3     24:06

School basic. You can program in basic, but you have access to the pins.

Host 1     24:10

Yeah. So it actually has a prototyping area where you can plug stuff your hardware in. Okay, that's pretty cool. And actually, it was I thought was interesting. It's like, what if you took your chip and combine it with this idea?

Host 3     24:22

Oh, that would be that would be really cool. Yeah. Then you have a

Host 1     24:25

Gigahertz Linux computer. Yep. In this. It looks about 10 inches by six inches.

Host 1     24:33

Yeah. Was that a size that you need for a keyboard? Yeah.

Host 3     24:36

Yeah, it's almost like a tablet, but it's entirely made of Fr four.

Host 1     24:40

Yeah, that's crazy.

Host 1     24:42

That's really cool project. Um, I've actually I backed the just getting the PCBs for

Host 3     24:49

That. That's cool. I can't wait. I want to see that one. Yes. Cool.

Host 1     24:53

So the speaking of PCBs there was out at Maker Faire last weekend. There's this company pocket CNC. Are you guys familiar with them at all? I heard the name. Yeah. So they've got this like awesome five axis CNC machine very precision milled. It's like a couple 1000 bucks, but it runs a BeagleBone. Black as its processor, but they've also I don't know if they announced it before them but they built a like a $400.03 axis CNC machine out of PCB material. Oh,

Host 3     25:24

Yeah, I saw that. Yeah, the frame the table, every

Host 1     25:27

Everything is all PCB. Obviously, you've got the servo motors. Yeah, actually,

Host 1     25:31

It comes from you order. I think it comes from OSHPark. The boards do, I think, I saw that I saw I was I didn't get to go to Maker Faire this year. So I was reading Oh, like the new stuff. Yeah, I saw it and it comes in, you get it in giant sheets, it comes in a big panel, right, break it out. together.

Host 3     25:50

So it's like, it's like the old die cast, not die cast, but the old plastic model like airplane models. Yeah.

Host 1     25:59

It's foreigner bucks. And it's a full three axis CNC machine, it's, it's pretty neat. Again, just kind of the cool stuff that people can do with just fr for for creative stuff. Well,

Host 3     26:10

It's, it's, it's interesting like it when you when you go to get something milled at a machine shop, you have to be really specific budget tolerance, blah, blah, blah, what you don't really see very often but but it's it's true of PCBs is that they're pretty high tolerance, just stock, you know, what you get from your manufacturer is going to be really close. And so using it as an actual using fr for from a PCB manufacturer, as as a mechanical building aspect. You know, it can actually work pretty well.

Host 1     26:44

Does something cool. I had never seen that done before. So that was that was pretty neat for me. But again, it seems like the same line of what what you were talking?

Host 1     26:52

Yeah, um, apparently this keyboard or this computer's open source. Okay, I did take a quick look at the GitHub account for it. And looks like most of the stuff is up there. Yeah, up there. I basically looked at the assembly docs and stuff and see how do they make the keys work. And I think it's just around out with just a little tab and the tab flexes.

Host 3     27:12

Oh, diving board. So yeah, diving board style.

Host 1     27:16

That's still pretty cool. So really cool.

Host 3     27:18

Um, maybe we should tag team and get one of these chips in there.

Host 1     27:21

Yeah, yeah, I think I'd like I'd like to see that. Yeah. So

Host 1     27:25

That's really cool project. Um, I think stuff like that, like using fr four is gonna become even more and more commonplace. You you see it a lot in in, in building like, fixtures to program stuff, where people will build the FR for like, a thing that holds their boards with with f4. But I think we'll see more projects like this going forward, mainly because fr for that whole PCB assembly or PCB process has gotten so inexpensive, right, in the last, you know, 10 years, even more so in last five years.

Host 3     28:04

Well, and you can use it as the front panel of whatever your device is, because you're already getting pre silk screened. And

Host 1     28:09

Right now, yeah, it looks nice. Yeah, it looks good.

Host 1     28:13

It would be interesting to see is a how well since this thing doesn't actually have a real enclosure is how well does it handle ESD and that kind of stuff?

Host 1     28:21

Yeah, but this is one of those projects that you're just doing just because it's cool. Now. Be a rugged design, you're not going to go stick it in any kind of industrial environments or anything like that.

Host 1     28:32

You'd be surprised what you find, like if you go on eBay, and search for Raspberry Pi, industrial enclosures. I'm like, oh my god, people using these things.

Host 3     28:42

I'm pretty sure I'm not gonna take a computer that just allows me to program basic out to the oil rig out in the middle of the Gulf. You know, like, I'm pretty sure that's not gonna happen.

Host 1     28:53

I don't know DIN rail Raspberry Pi's kind of scare me. Do they actually have those? Yeah, really? Yeah, you can buy an enclosure for a Raspberry Pi, or to put it on your DIN rail, industrial stuff, but the fact of the USB will well this is true with the old older Raspberry Pi's were the USB we're just randomly reset the device sometimes that you pull too much power

Host 3     29:18

Or you saw on the Raspberry Pi to be or whatever it was or the plus or I can't remember what it was. If you if you took a picture of it with a camera that had a specific bulb, the the frequency would actually cause it to reset.

Host 1     29:34

Yeah, really? Yeah. It was on the it was the LTO I think yeah the LTO

Host 3     29:39

Was an IT WAS A what was a die bonded to the PCB and and something about the the frequency of light that hit it would cause the LDL to reset

Host 1     29:50

So the the silicone the dye was actually exposed Yeah, so no no no

Host 3     29:54

The dye was flipped upside down okay and bonded directly to the PCB. What but the light

Host 1     30:00

Yeah, I think it was just through the package. No, I'm

Host 3     30:02

Pretty sure. I think Dave Jones did a whole video about on Eevee blog. Okay, I looked that up. Yeah. Cuz he actually shows he takes a picture of an bam,

Host 1     30:11

Yeah, this silicone is and that's it, you know, this little clear sample that I have here would be cool. It technically could work. But if it gets any kind of UV light on it the the silicone process actually goes goes crazy in it. So yeah, if that's Yeah, yeah, sort of deal there.

Host 3     30:28

Well, and it's interesting that CMOS technology, I mean, it's it's sensitive to certain frequencies.

Host 1     30:33

Yeah. Well, we only have, you know, 35 nanometers separating your gates. Yeah, you can find much for electron can get

Host 3     30:41

Those gates excited with a little bit of a little bit of energy. Yep.

Host 1     30:46

Cool. And it's not neat.

Host 3     30:49

No, it's just cool.

Host 1     30:52

Oh, well, I'll say the next thing is awesome. Oh, but so we like talking about IoT crap. Hear me making fun of it.

Host 1     31:02

You talking about IoT is the future IoT is everywhere. Yeah, I

Host 1     31:05

Do there. What is it?

Host 3     31:06

Nobody knows. Cloud. It's the cloud is a

Host 1     31:11

Marketing term Makarov uses cloud manufacturing brought to you by macro fab. And that is the only advertising whatever do here. Anyways, a nest who's owned by Amazon, right. So Google, Google. That's right. Google know, alphabets alphabet has, right? Yeah, that's what they're called. Now. That's right. That's right alphabet. They are releasing a version actually think they already did a version of their protocol called Open thread. So they're trying to, I think they're trying to do like a land grab kind of thing. We're trying to get everyone in their ecosystem kind of thing. For IoT devices, basically use this protocol. So all devices could talk to each other.

Host 1     31:59

So what is the protocol give you what is the protocol do?

Host 1     32:02

Um, you know, I don't know. I think it's just not it's like, it probably

Host 1     32:06

Define just how much you talk to the nest, or I think

Host 1     32:09

It allows, you know, I don't know, maybe the servers. Now, no,

Host 3     32:15

The biggest question I have is, and I don't want to sound like a jerk here. But who cares? You do? Well, yeah. I mean, like, I think, what do they what is their endgame with it? I

Host 1     32:26

Think it's, it's so like, all your IoT devices, say, your nest or your frigerator or your coffee pot can all talk to the same server.

Host 1     32:36

Same way, my nest phone app now can also work with my fridge and my toaster and coffee, you

Host 3     32:44

Can use your nest as the funnel for everything. Yeah, that okay, okay, I get it.

Host 1     32:48

So okay, just because the nest is that is a learning thermostat, but knows all this personal data about what you do? It could be Hey, around this time, you know, Parker gets up and wants coffee, right? So it would automatically figure that out and start brewing coffee at 9:30am, which is when I get up

Host 1     33:13

Early. This is more more for Google to get more and more personal information, right? Yeah, probably

Host 3     33:19

The data mining, data mining. So it sounds like they're trying to make the beginnings of Hal 9000.

Host 1     33:27

I mean, really what every IoT discussion always turns into. Or sky, or song, right?

Host 3     33:37

Well, someone actually asked me the other day, about, about macro fab and the fact that we have online user interface where you can purchase or upload your designs and they said, What happens when computers start uploading designs and making their own boards are due Are we beginning Skynet?

Host 1     33:54

Computers making computers? There you go.

Host 1     33:57

Freaky. Computers, telling computers to tell humans how to build boards. Be like it'd be like The Matrix. Oh, yeah. Oh, we're like, Wait, like, rip cable.

Host 3     34:12

Sorry, everyone. We're looking at the end.

Host 1     34:16

That's cool. You guys are on the cutting edge.

Host 1     34:20

We pretend to be

Host 1     34:25

Yeah, did that stuff. I mean, it's, it's cool. You know, the more open that people make it, the the more standardization that can happen around that is going to help make it you know, again, just more accessible. And I mean, I can't wait to get in my house and have everything talk to my iPhone and do everything that I want all from my phone. So if somebody is actually going to make it easier to do that, then that sounds cool.

Host 3     34:54

That brings up a huge problem though. And I'm not trying to make things drag out or anything like that. But the word was like As soon as we can standardize it, every company out there says I want my standard be the standard

Host 1     35:06

Was 15 standards, we have to make a standard to standardize.

Host 1     35:10

Eventually it's going to play out right. I mean, you had blu ray versus HD DVD. DVD. Yeah. I mean, yeah. Betamax. I mean, it will there will be a standard eventually somebody will win somebody will get tired of putting money into it. And then

Host 3     35:22

I Tripoli will come around and write it into their books right

Host 1     35:25

Now they'll just IEEE will make their own standard, which is completely different from everyone else's. Right, right. Are we IEEE six, one IoT or something like that? So does that anyone else have anything else that for this podcast? I think it's good for our info section was a little short, but you know, slow news week where you're going to

Host 1     35:51

Know I appreciate you guys having me on and this was a lot of fun. Yeah. Thanks for coming on.

Host 1     35:56

You want to sign us out?

Host 1     35:58

Oh, sure. I'll do my best. So thank you for listening to the macro fab engineering podcast. I was your guest, Greg Sheridan.

Host 1     36:07

And we were your hosts, Parker Dolman. And Steven Craig. There we go. We got that time. All right. Gotcha. Later, guys. Take it easy. Josh took a really good nap.

Transcribed by https://otter.ai