Why Mesh Networking Matters – Brandon Satrom

Welcome to the macro fab engineering podcast. I'm your guest, Brandon satrom.

And we are your hosts Parker Dolan

and Steven Gregg.

This is episode 151.

Brandon satrom Is the developer advocate for particle an IoT platform company. Brandon is also the founder of carrot pants press, a company dedicated to educating makers. on episode number 122. Brandon discuss the particle photon and built an IoT breathalyzer with Parker live on the podcast. Brandon loves to talk about code, open source, microcontrollers, robots, and whatever new shiny tool or technology that is currently distracting him.

So Brandon, what is currently distracting you now?

Mesh Network NeoPixels. I have spent way way too much time messing with. We've all done stuff with ws 2012 over over the years whatnot, and they're fun devices. But as I have I've done more lately with particles, new third generation devices and mesh network. I've spent a ton of time messing with NeoPixel illustrations and animations as a way to, to experiment with network latency and things like that, which has been a lot of fun. So it's the day job, but that's also been the thing that's been super distracting for me right now. So it's been it's been fun.

Well, blinking LEDs are always distracting. Absolutely.

I could just stare at them. You have the best kind of distraction, because the distraction is what you get paid for.

Well, that's very true. Yeah, you're right. That's right. You can't beat it. Can you like update your like company profile that said, like, blinking engineer or something like that? I should do that. You should? Absolutely

should do that. Yeah,

that sounds like yeah, that sounds LinkedIn ish.

And it's, it's, it's very domain specific, too. So if anybody else is like, Hey, can you work on this thing? And be like, I'm sorry, if it doesn't involve blinky lights? It's not on my job description. So I can't do it. Sorry, I have to pass.

Yeah, I'm, like, blink blink, president of bling.

President blank. Yeah, that definitely does sound like a like the job title.

So, Brandon, why mesh networking? Why does why does this matter? Yeah.

So you know, I know, Parker, when we when we talked before. Earlier in the podcast, we had a bit of a conversation about the IoT and sort of the bad reputation that it has. And it's worth reiterating, again, that I think a lot of the reputation that the Internet of Things has in much of the industry is it's for a couple of different reasons. But one of them is the fact that people tend to just put Wi Fi radios or cellular radios on every single device and call it good, right? Smart Home vendors, industrial organizations, I mean, it tends to be a very common thing. It's the cheapest, it's the easiest way to create a IoT device by putting a Wi Fi radio on it and calling it

good. Yeah, cuz everyone's got a Wi Fi modem at their house now.

Right, right. Absolutely. The problem with this is that Wi Fi networks are inherently brittle, right, when a Wi Fi network goes down, every device on the network loses its ability loses its IP address, and it loses its ability to communicate with other devices on the network. And so if you're doing a serious installation, or you need reliability, a lot of companies will tend to turn to something more like cellular but the problem with cellular is that radios tend to be quite a bit more expensive, almost twice as much, in many cases. And so can be harder to do to create a network installation at scale. And you're not moving away from that problem. Still having every single device connected to the internet. And every device with a public IP is an attack vector. It's an opportunity for another soldier in the next button at army. And so, you know, mesh networking is an interesting thing to me. And it's an interesting thing to particle, because I think it signifies where the space where the connected hardware space is maturing. And rather than putting a Wi Fi or cellular radio on every single device, we're trying to be a little bit more nuanced and create networks of devices where a select few devices have that connection to the open internet, or maybe none do at all, but that there's an ability for a local network of devices to coexist and can come in who can communicate with one another without having to expose every single node or every single endpoint to the open Internet. And I think that that's fundamentally more secure because obviously not every device in that case is uniquely addressable. But it also ends up being cheaper in many cases because the devices don't have to have that Wi Fi cellular radio on or to be a part of the network.

It's not just dollar wise it's probably also power cheaper as well. Is you have you have less power. Need for high powered radios and stuff to Go talk to the base station.

Yeah, big time, I mean when you start talking about depending on the on the mesh radio that you're using, but when those sorts of cases, you're talking about devices that can now be powered off of coin cell batteries, or can sleep for longer periods of time, and you can get, you can get a lot more out of them than just running, running off of mains power.

So particle wasn't the first to do mesh networking. So why, why is particle moving into this space now.

So I think a big thing, you know, particle and you know, full disclosure, I was a particle customer, three ish years or so before I joined the company earlier this year. And I think I came to particle the way that a lot of a lot of developers, and IoT builders tend to, and it's because it was, you know, inexpensive, accessible hardware. I mean, when I first started working with devices like this, your your options for getting a maker device online, was buying a $35 Uno and then a $65 Wi Fi shield that would bolt on top of it. And so, you know, in the early days, moving to particle was a $30, investment, seven $90 investment. And that was sort of the cheaper option. And over the years, this particle is added cellular, and now mesh, it's really been in response to where customers where customers are going and what they're asking for in many cases, but it's also in recognizing that we want to make each of those connectivity options as as easy as possible. And so, you know, particle prides itself on having made Wi Fi really easy and accessible for makers and for prototype errs. And it's the same thing with cellular as well. And then it's the same thing with mesh, right? If you look at a lot of the mesh networking standards out there, you're sort of left to your own devices, when it comes to setting up gateways and defining roles in the network and determining what your communication mechanism between nodes is going to be you're on your own and figuring out how to backhaul data to the cloud, if that's something that you're looking to do, or to handle secure security between devices and the process for provisioning new devices on the network and things like that, there's a laundry list of things that you have to do to roll your own mesh network using some of these other standards. And so what we wanted to do was actually create something that was developer friendly, just as easy as Wi Fi or cellular, you can get one or two devices out of the box. And you can immediately have a mesh network that already has networking provision at the device wellness level, the ability to do all of those things that I just said are actually just come sort of for free, as a part of using that hardware in and of itself.

Yes, and what kind of technology is the mesh network for particle using them.

So the big thing, there's a few pieces, the the actual, the hardware that we're using is a Nordic NRF 52 841 of the threads certified radios and there's a couple out there D Link has one NXP has one, Silicon Labs has their their mighty Gecko, which is another interesting, interesting one as well. But you know, we chose to use the Nordic because they you know, NORC has actually done some pretty amazing stuff in the context of both thread, but also Wi Fi and cellular with radios that can be sort of multipurpose, you have a thread radio that can also sort of coexist with Wi Fi or coexist with with cellular as well. And so the chip that's underlying all of our all of our mesh based or third generation devices is a NRF 52 840. And those devices give you the ability to implement sort of the open the open thread stack. And that's what you mean by a thread radio is Yeah, so I jumped ahead of myself a little bit there. Yeah,

yeah, I had,

I was trying to recall, recall what that what that meant. Yeah,

yeah. So I mean, in the context of mesh networking, sort of the the older, or the existing standards, or your your Z wave and ZigBee ease. And then Laura can be sort of lumped into that group as well, because it does provide some interesting connectivity options at larger distances. But ZigBee and Z wave have definitely been around for a lot longer thread is an interesting case because it was actually created by by the nest team at Google. And the whole reason that they went down this path was actually is an interesting use case. And part of why particle decided to choose an open thread base for it and why I love a lot of other vendors have as well because when nest moved from thermostats to smoke detectors, they were moving from higher powered like mains power devices into battery powered devices and smoke detectors and they didn't want to use Wi Fi because Wi Fi was too power hungry. And also Wi Fi is brittle, right if the network goes down and your smoke detectors can alert for fire, that's a big problem. So they couldn't use Wi Fi they looked at Bluetooth but Bluetooth is point to point and they needed to be able to multicast messages across the network so that multiple smoke detectors could receive could receive a message so so that led me nested down this down this path of creating this eight to 15, four bay standard called threads that they then released as an open source library called Open thread that created a working group around this particles part of it, NXP Scilab, TK, Nordic D link, there's a ton of companies that are part of this working group as well, that's really designed to evolve the standards where not only can it work well with devices from a single vendor, but work between vendors and even start to work with other other standards. So there's actually some work happening in that working group now to allow ZigBee because it's also an open standard tool as ZigBee and thread devices to talk to one another on the same network as well, which is pretty interesting. And this is fascinating to me, because from a from a smart home context, one of the big challenges I've always had with Wi Fi, is it's given a lot of vendors that's excuse to try to lock you into their ecosystem, whether they lock you into a protocol to lock it, they lock you into a protocol or to have to use their app or to you know, the the echo sort of broke this in a couple of ways, because they created this convergence of devices that wanted to be featured on their platform, but you still have this radio level difference to where devices are communicating across the open Internet just to do something two feet away in the same room. And so thread and ZigBee. And these other standards make it really interesting where you can actually start to have this interoperability across vendors and across standards, which is pretty cool.

Yeah, that's been my biggest, I guess hinderance in adopting like any kind of home automation stuff is the whole, I don't want to be locked in, you know, forever on one thing. Yeah.

Yeah, same, same here. And I even like a couple of years ago, I kind of tried to roll, roll my own sort of thing where I was running, I was running a Raspberry Pi and a in a Samsung Arctic 10 as sort of my own my own sort of home automation hub. And I was doing some stuff with with Node red actually running, running on that hub to, you know, facilitate communication between devices. But what I ran into was those cases where vendors only allowed me to communicate via their API. And then that was just a cloud based. It's just a cloud based communication. So I'm back back to where I started. There's no, I don't really have a local network. I'm just proxying things through the internet, which is annoying. So yeah, same thing.

You want to be able to open the door when you get home. And your internet's out when the power is out. Right?

Exactly. Yeah. Yeah. And I think this is why a lot of consumers in the smart home space have been really reticent as well, like we'll never have outside of outside of hobbyists and people that are just really hardcore about this stuff. We'll never have general acceptance of these kinds of technologies as long as those constraints exist.

Now, I think it will basically for home automation to really take off it's got to be just like a, how a normal light switch or socket works now is you can go to Home Depot or Lowe's and buy something off the shelf and you wire it in. And it just it just works with your what existing network you have. Yeah, yeah, absolutely.

Well, also, whatever happens in home automation needs to be protected from the servers going down or the company going out of business or the API just not working for whatever reason.

So there's a there's a little side tangent on that one. Steven, I don't know if you saw the article where this one person bought for his this is a really sad story, I guess. Bought a one of those like Life Alert things? Oh, yeah. His, his grandmother. Yeah. And apparently that company went out of business. And the device couldn't report anymore back up to you know, the mothership. And they found her, you know, I guess, pass in her apartment or wherever. And it was within her reach or something like that. So they don't know if was able to use the device or something like that.

Oh, that's awful. And she got no zero notification whatsoever that it was no longer operable.

Yeah, like the company went out of business, and no one was notified of this.

So this is again, like, because Wi Fi tends to be the cheapest, most expedient option. I think a lot of vendors do this, and then they don't think about the long term. Or they'll do sort of seditious things like use DRM, on the firmware in their devices, and DMCA takedowns. To prevent security researchers and even individual hackers from being able to monitor devices, there was the Google had acquired one of these smart hubs that went, went offline. And then as soon as they took they turn all the servers off, the device becomes completely non functional. And that's a consumer case. It's not life and death. But that's, that's a great example. I mean, there's lots of these kinds of things where, by relying on Wi Fi connections to the open internet, we're asking for these sorts of cases where companies go out of business. And we as consumers can't take control of how the devices function even on our own, even on our own home networks.

So let's talk about something a little bit more happy. So you've been working on a mesh networking project,

right? Yeah, yeah. I have. So the project in this is sort of foreshadowed this a little bit earlier. But you know, in the context of mesh networking, one of the things that's fascinating to me is not only do you care about sort of the range of devices and your ability to communicate between devices on the local network, but also the speed at which you can communicate, I've been thinking a lot about latency. And you know, when I talked earlier about particle wanting to make it really easy for developers to implement mesh networks, a big part of it is actually hardware, things that just work. But another piece of it is API's that make it really easy for devices to communicate. So one of the API's that that particle is provided with the new hardware is mesh is Pub Sub and a mesh network. So you can do publish and subscribe of messages. And under the covers, it's just UDP UDP multicast messaging. So super fast, very low latency. And I wanted to test for myself exactly how low latency could it be. And so the idea that I had to do that was to create a mesh network of NeoPixel strips that are wrapped around a bunch of tiny little 12 inch Christmas trees that I got from Amazon. So I got three of these trees, I got three NeoPixel strips. And I used the I use the the party badge that I created for, for spectra for our conference in October, which is, you know, I spent just like many of us is I spent a good bit of my year and trenched in the badge life world. So we've created a badge that has sort of a built in Piezo Buzzer and since that was already there, I just loaded it up, there's a UI for selecting songs on the on the device, I just loaded up some Christmas songs. And then anytime a note was playing a tone, or no tone call, was playing against the people I was firing off, I'd fire off a mesh dot publish to the devices on the network. And then in that device was just a mesh only device. So I guess we haven't covered this yet. But in the context of particle devices, there's there's three types of devices. There's the argon, which is a Wi Fi device, the boron which is cellular, LTE, Khadem, one device, and then the xenon and the Xenon is are the mesh only devices. They're the the workhorse devices. They're the cheaper devices that can run as a notes off a coin cell batteries and whatnot. And so I have an argon that sort of is the gateway but the badge and then the actual Christmas trees with the NeoPixel strips are all powered by Xenon ions. And the three Xenon are just listening for they're just they just have a mesh subscribe call. That's that's passing off a handler that will then light up the NeoPixel strip. And what I found, you know, when I went into this, not knowing how it was going to look, I was a little worried that it was going to seem laggy that it would, it would just sort of touch off, but it actually has worked out really great. It's been almost instantaneous to see, to actually see the the pixels on the strips actually all laid up almost instantly as I'm hitting the button. Just pretty fascinating. That's cool.

out of out of curiosity, have you gone down to I don't know what the scope or something and actually measured the time?

I have not. That is a that is a break, that is a project for my Christmas break to actually see. Yeah, to actually do those measurements to see because I mean, it's a, I'm quite confident that it is a naked eye effect, like it looks to me to be synchronized, when in fact, there probably is a very, of course, there's a very miniscule delay, but I'd be really interested to know exactly how far it is. And I also haven't really done anything yet where I just have the trees kind of closely grouped together. Because I wrote a blog post about this last week. And so it's really more about significant work. I would like to actually test the edges of it now to sort of move the trees apart, sort of see what happens.

Yeah, be really cool for you. Yeah, put it in like a long hallway and see if the message is like, I mean, it's, it's they want it, the electromagnetic waves go at close to the speed of light. So you're gonna be able to measure something like that. I don't know when you're in your home lab. Yeah, probably

not. It's not that sophisticated. But I mean that that idea of the hallway, like there's some interesting things you could do with that same demo to mess to mess with sort of the unique features of a mesh network. And one of these is that when, when a device sort of sits in a certain plays in the network between two other devices that also can function as a repeater. So it has the ability to take a message that comes to its node, respond to it, but then also pass it along. If it sees that there's another node that may be out of range of the main device. And so whereas right now I have the bad sort of multicast is sort of broadcasting to all three, if I were to spread them out outside of the range where they were part of that, that where they had that connection to the main device, it would actually bounce the messages one to the other. So if I did that correctly, you'd get sort of a, you know, bounce, you get the delayed effect as well, which would be, it's gonna be kind of interesting.

Yeah, that would be the more interesting thing, I think. Measure. Yeah.

Because then you're also you're also doing a bit of a range test, which we've had, it's interesting, we've actually had some folks in our community that have done some pretty, some pretty interesting things with range testing, you can run these devices with an antenna or without. But with the antennas, if you have both sort of your gateway, and your end device with an antenna. in a business setting, there's some people that are seeing 50 meters, between devices. This is in places that have concrete floors, you know, metal framing, things like that busy Wi Fi. In a home setting, they're seeing something more like 100 meters. And and one guy actually the other day did a line of sight test where he actually was able to get 500 meters between devices. And this is out in the open with no other ambient RF around but 500 500 meters on lines. This line of sight is pretty interesting.

And that's battery powered right now it's

battery powered. Yeah, yeah. Very cool. Yeah. Yeah. That's all that is all due credit to the to the Nordic, the Nordic semiconductor crew. And, and just opened turn itself. That's pretty cool. Pretty cool to see that.

You're the I wonder if I don't know if you've ever looked into the called is called cantinas? Where they take a like a Pringles can or Folgers Coffee can and they have you put your Wi Fi antenna in it. And it kind of like makes a directional

like a, like a transmission beam.

Yeah, what? Oh, yes. called.

I don't remember what they're called. Have I seen that before at that? Yeah. Because like tubes are. They have like a metal coating or something like that in?

Yeah, I can't remember what they're called.

That's pretty cool, though. that would that would certainly it's can something that would make your your mesh network look pretty pro at that point. Yeah.

Yeah, the first link I found is how to build a tin can waveguide Wi Fi antenna, and then it has a, a tan of beef stew that he eats to turn it into this thing. I think I have to make this. Yeah, I think you need to, like see how far you can

push it. I saw a guy build a he, he built a device that that was basically one of those cans, that was a was an antenna. And it was on a rotary table with like an XY gantry. And he would turn and scan the horizon with this. And basically make a heat map. So you could see Wi Fi signals with this with this device. And it were it would take hours to scan the environment. But then he could reproduce a map of it. I couldn't have a Wi Fi. It was really cool. Oh, wow. It's even a simpler name. It's can and antenna combined. So it's can tena cantenna Go vague. That's cool. Go figure.

Along those lines, I have actually been really fascinated with you know, latency is sort of the first, the first thing I've been messing with sort of testing API's that we have, but signal strength and being able to almost triangulate device location without having, you know, GPS coordinates or anything like that was able to create sort of a topology map of a network based on signal strength or based on where messages are coming from. And things like that is something that I'd like to spend more time playing with. We don't have a lot of stuff documented around that right now. Because I think a lot of it is just, you know, customers, the developers playing with it and seeing the things that they can then come up with, the baby would be really neat to create sort of this virtual map of devices based on what you're seeing and behavior and activity.

That's actually one of the questions someone asks in our Slack channel. Oh, cool. So if were you done with your project? Or is there something else like the future of your project?

Well, the so the future of it, I have to share because I went on, there's a few things that happened as a result, all of this messing with, you know, messing with lighting up the trees based on button presses, or based on music that was coming from, from the devices I went down this path of, and this may be an area where Steven can educate me a little bit more, but I went down this path of of could you represent, can I create sort of a color of sound algorithm where a note that comes across via the device gets mapped to a gets mapped to a color as well and so you so you get sort of a uniform based on the notes that are coming across with a badge you get more of a uniform color as opposed to random which is what I was doing it first of all, And I was prototyping and, you know, recognizing obviously, that, that, you know, sound is based on air molecules and compression waves. And light is based on electromagnetic waves that there's, there's definitely a difference there. And even though, but they do still both share this concept of frequency. And so I started doing a bunch of research on what this looked like, again, all just I went down this whole rabbit hole of like, oh, this is interesting, I can light up things. But can I actually make more of a pattern to it that could be, quote, unquote, pleasing to the eye. And I went down this, this path of realizing there's a lot of folks that have actually done some research on sound and color. People have done there was actually a scientist that created a mapping between discrete pitches and colors. Because he had a condition called synesthesia, which is a condition when you experience one sound, or one senses a different sense, and he perceived sound as color. And so he basically, he basically created this thing called the, this Grob scribing, correspondence, where or sort of the keyboard with lights where he basically took the sounds that he was hearing, and then map them to a color on the spectrum, which is, which is pretty crazy stuff. And so I started researching this, and I found some people that actually created some JavaScript and Python code to do a mapping of wavelengths to RGB values. And I just created a C based version of it on my own as well. And the long and the short of it is that if you assume that the note A occurs at about 440 hertz, and you double it about 40 times, you end up getting something into the range of light that we can see. So you get sort of a red, orange, yellow, green, blue colors at that point. So like, right, like an F sharp, four is dark, red, a four is orange, a C sharp, five is lime, etc. It's all it's arbitrary to a certain extent, but it's really interesting music creates a sort of consistent mapping, as well. But it also has this other side effect of now what I was doing on the actual devices, through the processor was I was adding computation, it wasn't just receiving a message and then lighting up the NeoPixels. But now I had a computation step where I was taking this frequency, and I was basically doubling it 40 times and then mapping it to a wavelength and then interpreting the RGB values. And it still ended up being very like, right on. I mean, I was getting, you know, getting the animations to pop up, you know, as quickly as I could push the button from what I can see.

So you were you were doing the computation on the mesh side. And so you were sending a frequency to them. Now instead, just on off,

yeah. So that was Yeah, yeah. So I was basically redoing the calculation across all three end nodes. And that was more of a just to see what happened. Yeah, the more the the saner thing to do would have been to do the calculation actually, on the sender side, and then send it out that way. But that was really Yeah, it was a bit of a test to sort of see what would happen on the end devices, when I added that extra computation between receiving the publish receiving the message, that UDP message and then actually lighting up the pixels. Just kind of cool. So and again, I don't know if I will ever be done with this project. Because every time I finish something, I come up with another idea, but Adafruit actually just sent out their latest ADA boxes this month, and that actually included the cool, cool new device that they've created called the Adafruit Neo trellis, which is a basically it's a sandy 51 base board that has a bunch of NeoPixels built in. And the kit comes with some, some nice, I don't what do you call those little moldy, little soft buttons that you can press. And so what I what I was, the other thing I was messing with here is because that's a different microprocessor. Using serial communication with a mesh base device to then join it to the mesh network. So sort of mesh unifying a non mesh device by just doing serial, and then sending serial commands of the buttons that are pressed, and then lighting up the trees that way. So just different different areas of sort of poking at the edges of this entire project and seeing what happens with it, I probably need to do more to write up, you know, write up the entire thing because it's, it's been a lot of fun to work with. So that's sort of the next piece and I have that I actually did get that demo piece with a trellis working earlier today. So where there's a correspondence when you when you put it in sort of basic MIDI mode or key press mode that you will press one of the buttons and it actually will use a sort of an Adafruit wheel animation and I'm sending this I'm sending that same wheel position via serial over to the xenon and the Xenon is doing a mesh publish to to the other to the other devices in there seeing that wheel position and then lighting up the entire strip with that color.

Cool. Yeah

Do you have to show it offense to the podcast?

Oh, yeah, I will definitely it's all working. And there's no way there's absolutely no way that it won't work. Of course, after the podcast.

I would totally say we would do it now. But like, this is an audio only podcast. Yeah, I don't think I don't think shifting light frequency down to audio range, like the backwards way is going to

be pretty fast. I will, I will post a video of it tomorrow morning. And if you want you can put it in the show notes or something. Yeah.

You know, maybe, maybe I'm missing something here, I get some a little bit confused. But the whole shift of color down to audio or the other way around? That just seems like it's a lookup table effectively, it seems like there's not much of a calculation there. Like, it's sort of like everything's already done for you. Right?

It is. And I actually started, that's a great question. I actually started with that place of, I'll just do the lookup table across the notes that I know that I'm sending, because because of course the notes are actually all mapped in arrays on the bad side. And so you know, those frequencies are coming from the same coming from a place where I could very easily just add a second array or additional values that map to the actual RGB colors. And so that's again, that's probably the saner way to do it. This was another exercise in over engineering, which I guess I get to do, because no one because none of this stuff goes to production. So I get to I get to over engineer it all at once. But yeah, that's definitely a case where that'd be the saner alternative.

But I love your solution. That's fantastic.

I know this is one of those cases where I couldn't stop myself like I'm like, Yeah, I'll just go and do the lookup table, that's fine. And then when I went down this research, and I found a site where a guy had actually created a calculator where you could enter a fundamental note and a pitch standard and a bias and, and the environmental temperature and humidity and he would use all of that to give you a color of light, resonant with that pitch. And I'm like, oh, okay, I have to replicate this. That was that was ultimately what did it.

You know, one of the things is, I see I see these devices being for some reason, this just keeps coming to mind. But entertainment purposes on a on a stage seems like these would be absolutely phenomenal for like controlling, or for something that needs to be assembled or disassembled on a regular basis. And then you know, reconfigured on the fly can think it seems to me like if you had some kind of fancy stage lighting setup, that you could talk about replacing DMX. Well, actually, no, maybe not even replacing it, maybe maybe you could have a communicate over DMX. But But I'm talking about like the actual physical construction of it, you can have this whole setup with I'm using quotes here. No wires, you know?

Oh, yeah. Well, I think there's a lot of cases, when you start sort of spreading your devices out. There's a lot of cases where solutions can have have those sorts of qualities, whether it's an entertainment case, or whether it's a sort of a large environment, like a sporting event or something like that, or even, we have, you know, a customer based out east or potential customer that's prototyping right now. And they're looking at using their devices in sporting events and trying to figure out that problem of how you beat the fact that most stadiums or any place with a ton of people has terrible Wi Fi terrible cellular, it's impossible to communicate reliably, where you're actually trying to create something that's a little bit more resilient. And so it'll be interesting to see if they're able to pull something like that off, because it could be a cool, a cool alternative. But yeah, any environment where you've got a lot of RF got huge interference, we've got a ton of people, or even where you're actually dealing with quick break down, spin up sort of cases, could definitely be a cold case for that.

So that's a good segue into the first question. And I think most of these are from Jared from the slack. Cool. And so the first question is, could you have a sports game of a sword and hand out each like I got sent a device to a spectator, where has a little LCD or LED with a message chip? And so you have like, he says, 10,000 I mean, the stadium that I go to has, like 100,000 people that go to it. So I'm gonna say 100,000 systems attempting to figure out where they're at in the stadium, connecting to base stations, and then he's easily you're not making a ginormous like, display Allah. Like the I'm thinking like, the People's Republic of North Korea, people like North People's Republic in quotes. When they have like the giant like, people, screens where they flip the cards around and they like dance with the cards. I've seen those videos. They're actually pretty impressive, like the coordination between like 1000 people

You know, what's actually coming to mind is a really, really lazy wave where you don't have to stand up, it just goes around in a circle, Hey,

yo, you just hold up all of your little device and the wave goes, just goes. Yeah, yeah. Yeah, I think those are that's absolutely I mean, those are interesting sorts of use cases. And I love first of all, I love the whole receiving color information and turning the audience that used to play because that's what I've been playing with right now with this idea of sort of showing the power of communication via color via display. View absolutely cut. I mean, this is the thing that that people are, are prototyping with right now, the challenge that you have to figure out is, network design is everything in this case. And I think this is definitely an area where as more and more people start to create real solutions with mesh networking. I think the state of the art around network design is going to evolve in terms of what sort of recommendations people make based on how far depart to place nodes or how to deal with repeaters and gateways and things like that. It's the kind of thing that that I would say, yeah, absolutely. You could do something like that. The devil is in the details of how you actually configure that network and how you set everything up. And where you place gateways and things like that. But

you just have all 100,000 of them be repeaters. Right.

Right. Right. Exact? Yeah, absolutely. Yeah. So

yeah, if you have 10,000 of these systems, how do you like power them.

So that's one of the areas where mesh actually shines in the case of a lot more, a lot more devices at scale, because you can actually start to talk about powering for longer periods of time via conventional means via batteries or but also moving into smaller power form factors like a coin cell battery. You know, sleeping devices, of these kinds tend to run more at the micro amp level when they're not really doing that much. And so you can actually shut these things down, respond at intervals, you can run over over longer periods of time, without having to worry about burning through battery life too quickly.

Because Jarrett's got the question where he's got a indoor garden that he's doing logging metrics, Jarrett, we're not going to ask what you're growing in your indoor garden. But he's got a bunch of devices that he's logging, like the temperature and I guess like light levels, and maybe even soil, humidity, soil moisture, or something like that. But so right now, he's got a whole bunch of like, he says, a million and quotes, phone chargers plugging everything together. So I imagined like this ginormous like, spaghetti mess of like, like plant and vines, but like cell phone lines like intertwine, right, it will look like that show. What like, what, like humans are dead. And so it's like, nature's, like overgrowing cities and stuff. That's all I'm imagining here is like,

imagine, like been overtaken by USB micro cables.

You mean? You mean it looks like developing anything on a Raspberry Pi? Yes, basically.

Just a massive cables.

Yeah, I mean, those are the kinds of cases where I'm guessing those metrics are happening at a pretty, you know, pretty infrequent intervals, it's probably not something you need to do every 10 seconds or 30 seconds, every several minutes, maybe. But you could definitely, in that case, power something off of a single double A or A couple double A's or even a coin cell and run a project over the course of several days or weeks without having to change out your power. Now, you do still have to change that power. So he says indoor gardens, I don't know exactly what this would look like. But those low power draws make it a lot easier to turn to something like solar and have to work without having to worry too much about whether you're getting enough current draw for the panels, as well

say Yeah, it's like a solar kind of thing. Yeah,

yeah, that's that's what I would do. I mean, I've messed with solar you know, I've done a bit with solar before and I'm no pro here but it's always been a challenge when I'm running when I would run previously off of like, you know, five volt microcontrollers or something like that, but having something that's actually got a much lower power draw would be a whole lot easier to mess with.

Also see if the timings not super critical, I could see a circuit that just basically monitors how much charge has built up from the solar power and when it has enough it then fires up sends out a message and then re re does the cycle.

Oh yeah, I can imagine that being like like you had to have like some margin of error in that so like it fires it off but doesn't have enough power to go back into like it's sleep monitoring mode so that it just you know, it's suicide it gave out its last information ever knew about the last message it just sends out goodbye world. That's always been my problem with doing stuff like like wireless. So used to run some like wireless like security cameras. And it's like, well, you still got to run power to him. So they're not truly wireless. Yeah. But the thing with mesh networking, I think is it's the low power having batteries, which is where it makes it shine. I guess exploring more into solar would probably be the best bet for something that's like you don't have to worry about Yeah.

A lot of the I mean, our particles, third gen devices all have all have lipo ports built on. So you can also go through that case of sort of using a small lipo as your backup or sort of have have, have solar charged the lipo and use the lipo and then sort of monitor how things are out, you know, the draw throughout throughout the day, things like that.

Yeah. Or Jared, you could develop a bio cell that grows and somehow turns the soil energy directly. And then adding a solar cell is probably way easier, though. And

then Sally, I did a Dave Jones, because he loves that kind of stuff. Right?

Yeah, he loves. Actually, I think it was Dave Jones That was talking about a, it was a cell phone charge system, like a wireless cell phone charger that use lasers for infrared lasers. And so it would be so you'd go to like your coffee shop, right? And they would have a hub that would be in the center of the hall room. And it would, I guess it would figure out where your cell phone would be at and it would beam energy into a port on it. Like it would be infrared energy and charge your phone up. Because you would have like a I guess you'd have like a case on it that that had a sensor that could see that. I thought it was a pretty cool idea.

But it's cool. It sounds cool, but also sounds dangerous. Yeah,

he proved that like at levels that it wouldn't be dangerous. You could, you wouldn't be able to charge your phone. But that's kind of cool idea. Like it says like, if you had this like big indoor area. And you could just track where stuff sat and up up tie with like computer vision, and then just beam energy like microwaves down at like the robot.

That's yeah, that's really cool. But I could see a situation where you're like sitting at a table reading a book. And then for some reason, you just start feeling really hot. And it just starts getting hotter.

And you look over at your coffee and your coffee starts boiling. No, right. I'm thinking like more of like a warehouse like an industrial setting where you have a bunch of robots. And you know, when a robot needs to charge, it's on downtime. So instead of that you just be microwave energy into it from, you know, the ceiling.

That's cool. Out of a Pringles can right.

Oh, that's great. And so one last question. Go. Brandon is an update on care. Pants press. So when you were on Episode 122, you had a new Kickstarter coming out, and a new book. So yeah, let's have an update. Yeah, so it's good. Yeah, it is. Yes, it is good.

You wouldn't have asked if it wasn't right, I would hope? No, yes, the Kickstarter went well. We did launch a Kickstarter for the second book. And I'll get a little bit background. I mean, what what care pants press primarily does is our goal is to expose kids to the world of electronics through story. And so we create illustrated children's books that are designed to sort of teach basic electronics concepts. The first book was called Edie gets his power back was about a green LED named Edie figured out how to light up his led in the town of red Bordeaux. And the second book, The second book is called Le saves. Today it was about Elliot blue LED. The circuit here got a little more complicated, but this was actually she was with her friend Ed working at a at a railroad crossing and needed to figure out how to blink her and Ed's LEDs back and forth with the help of a five by five timer chip, before a train came barreling down the track. And so for that second book, what we hit what we did in the first book as well. And what a lot of a lot of our backers really enjoyed was having sort of the book with a breadboard and components to actually build a circuit from the book, we created sort of this 50 page PDF guide book that you could walk along, you know, walk walk through kids between the ages of, you know, for upwards of 10 or 11 actually walk through the components and explain what they do and sort of build it step by step. We did for the second book actually introduce sort of a custom PCB so started to introduce the world of printed circuit boards and having components that are and what actually did in this case for this for this individual kit for backers is we had the PCB made, but instead of actually soldering components into the PCB, I soldered in sockets and headers for every single component. And so what kids will do when they go through this second book is actually insert the components. And then they can swap out different resistors in different capacitors to see how that changes are the blink rate, and the time that an individual light goes on or off. And that's been an interesting thing. So we actually just the Kickstarter was successful. We promised that we would get the books it to everybody by December by mid December, it is mid December, and we have shipped everything. Which is Yeah, which is pretty fantastic. So I can I can Kickstarter to keep some exactly I can claim to be a two time Kickstarter that had to on time on time projects delivered, which is, which is pretty fun. So we got the books out, and people are just starting to get their hands on them. But you know, one of the things that I really love about this is that kids seem to really love the idea of of getting a sort of an engaging tale. I mean, it's a story first, but they're almost learning learning things by accident you could put it put it in one way but they're really sort of learning what these components are and what they do and I've seen this in my kids and and others as well as they can instantly recognize what they what these components are and how they work and it really gets them excited about actually being able to create things with electronics and explore further.

Oh, cool. So Brandon, where can people find out more about carrot pants press particle, your mesh, Christmas trees, all that kind of stuff.

So So for for particle you can find out you if you go to blog dot particle that IO so it actually has the mesh Christmas Tree Project out there. I will on my Twitter account at Brandon satrom post a video of the Neo trellis demo and some of those other things, other things tomorrow and that carrot pants were actually on Twitter as at carrot pants pub, or care pain. studios.com is our website. We're actually in the process. We're gonna put the new books up for sale on our other sites tomorrow. After I got all the Kickstarter books out I had that three or four days where I didn't want to ever see another shipping label or box ever again in my entire life. And I've gotten past that and now I'm poor putting the books up online so that other folks can get them and

so that actually be the day this podcast comes out. Yeah, I

was about to say tomorrow would be when you're okay. All right. Yeah, so actually go see it. Go there right now if

you're listening to this. Got it.

Bye bye bye.

helped me get close to break even on this Kickstarter.

So with that, Brandon, you wanted to sign us out? Yeah.

That was the macro fab engineering podcast. I was your guest Brandon satrom.

And we are your hosts Parker

Dolman and Steven Greg.

See you later guys. Take it easy Thank you. Yes, you our listener for downloading our show. If you have a cool idea project topic or mesh networking idea. Let Stephen and I and Brandon know Tweet us at Mac fab at Longhorn engineer with no O's or at analog e Ng, or at brandom sat room or email us at podcasts at Mack fab.com. Also check out our Slack channel. Brandon is a member of our Slack channel and hangs out there as long with Steven tonight. And if you're not subscribed to that podcast yet, click that subscribe button. We also have a new podcast email, like newsletter, so go check that out. You can subscribe to that on Mac fab.com/blog/podcast And please review us wherever you listen. It helps this show stay visible and helps new listeners find us