Embedded - 152: Dodecahedrocopter.com

Episode Date: May 18, 2016

Chris and Elecia chat about hobbies and respond to listener feedback and questions. Chris was on an episode of Let's Drone Out, you can listen to it here or search in your favorite podcast platform. It is recorded and broadcast live every Thursday at 8 P.M. (UTC+1) onPowering On. Chris' new quadcopter is a Vortex 285. It runs Clean Flight, an open source flight controller software package. While we had various opinions about RTOSs, we were both interested in the one Alvaro suggested to us: Zephyr Project. As for other embedded podcasts, of course you know about The Amp Hour. And we had Saron of CodeNewbie podcast on, that show is mostly software and people. How aboutMacrofab Engineering? Or O'Reilly's HW podcast?

Welcome to Embedded, a show about the many aspects of engineering. I'm Elysia White. I'm Christopher White. And we don't have a guest this week. It's just going to be us talking to each other. And the cat refused to join us. I know. She's very, very difficult.

And we are mostly going to be addressing listener questions, emails, whatnot. And then talking about Chris's hobbies. What? That's the plan. All right. Didn't you get the memo? Maybe whatnot will run long. Maybe not.

In fact, that makes me tempted to start with, weren't you on another podcast recently? Weren't you cheating on us i was on a podcast it wasn't recently it was about a month ago but it just recently dropped on itunes well are we gonna just jump in like that just jump right in just is this a show there's no small talk uh what you want me to lighten around you sure pretty funny when married people do it title yes i was on a show called let's drone out i was invited by some nice folks in the uk uh and we talked about uh rc aircraft and drones for about an hour it was great fun there were more of them than i expected so it was different than a one-on-one interview it was more of a panel yeah yeah um and uh yeah it was different than a one-on-one interview. It was more of a panel. Yeah.

And yeah, it was a fun show. So you were invited because you started playing with your airplanes again. Yes, playing with my airplanes again. I'm sorry, that really didn't come out right. You had added a GPS module. You had added a cameras and all kinds of stuff to you know a large-ish airplane. But RC controlled

electric airplane. And then after that you immediately needed a quadcopter. Well, what? No. What? That's not what happened. Are you sure? I was bullied into it by my brother. Because he has one. I see. Yes. That is true. But I mean, part of the

reason the quadcopters are interesting is because they're a little easier to deal with in terms of taking off and landing once you get proficient enough to do so without crashing into yourself. Ah, nice caveat. But planes, taking off isn't such a big deal because you can hand toss

them, but landing, you do have to have some approach distance and some form of field and no trees and stuff. So I got a little, I don't know, a little nervous about flying in various places that I thought I would fly, but probably won't end up doing because there's too many people and not enough room. So quads are interesting in that you can kind of fly them anywhere. And they're small. I mean, a plane is a big plane to carry all those electronics. It has maybe a one and a half or a three quarters meter to a meter wingspan and lots of big stuff to haul around.

But a decent quad is a foot long. And hopefully it won't get stuck in a tree. Yeah. Has less cross-section for getting branches stuck into it, yeah. So this quad has like seven ARM processors? Yeah, so what I ended up getting is this is a Vortex racing quad and they're uh the code name is is a two it used to be nothing but then they brought out a 250 and that's just the i think it's the millimeters from one end of the diagonal to the other and this one's a 285 so vortex 285 um and i i kind of wanted to build one of my own

but then i kind of wanted to get used to flying them first. And so I kind of compromised and bought one that's basically pre-built and ready to fly, but it's pretty hackable. The flight controller computer runs open source software called CleanFlight, which is pretty cool and does all sorts of stuff that I haven't even scratched the surface of

and this one has a bunch of extraneous other fun stuff like it has an led panel on the back which is looks a lot like uh what do you call it neopixels yeah neopixels which has its own microcontroller driving it so you can play games with that it has an on-screen display which goes on to the uh the video input so as you're flying these things you actually have either goggles or a screen on you're not watching them because it turns out they're so fast and so small that in about 10 seconds they're outside of your field of view unless you're an eagle so you actually fly them first person there's a camera up front. It transmits down on a 5.8 gigahertz band to either goggles you're wearing, which display it in front of your eyes, or an LCD screen,

and you fly like you're flying in an airplane. And so it has another microcontroller, which muxes in a video overlay, which gives you all sorts of information like artificial horizon, your altitude and speed, and whether your battery's about to die and good stuff like that. So it's pretty fun to fly. Once you get used to it, it's kind of like having your own little jet airplane. Because you, well, I think you did better once you started sitting down. Yeah, there was some orientation, disorientation issues initially. Because this is really like your hands are moving the rc controller which is pretty

complex and programmable right and then this is your whole field of view there is no i'm looking at the quad quadcopter and now i'm looking at the field no it's just all quadcopter if somebody wants to clock me over the head and take my wallet i won't be able to see them coming it was easy to sneak up on you yeah so um so initially i just you know i'm having fun flying it but then you know i've got a lot of ideas for playing around with the firmware um putting a raspberry pi on it and maybe doing some uh machine vision experiments uh so you know it it's and it's amazing it's another one of these demonstrations of what's happened with the low cost of microcontrollers and the arm revolution, because this thing has seven arm cortex processors on it.

And, and we know that because it was part of their advertising, which just, I'm like, who would make that part of your advertising? Right. And it's not like it's really expensive. I mean,

it's a few hundred dollars, but it's not like a thousand dollars where you'd have, you know, Oh, we have so many microprocessors doing stuff. It's, you know, it's not like it's really expensive. I mean, it's a few hundred dollars, but it's not like a thousand dollars where you'd have, you know, oh, we have so many microprocessors doing stuff. It's, you know, it's just, it's so cheap to just throw one in places that you might not have thought to throw one before where you'd have an analog board in the past. Like the speed controllers for the motors.

There's four arms and each in each arm is probably a Cortex M0. I know it's a Cortex something, but I'm not sure which part. And their job is just to do the current control for the motors commanded by the central flight controller. So instead of having the flight controller have to talk to an analog board and continuously output a PWM or something for each arm. Given the motors around here, if it's doing it all analog and moving off over the space, it's going to be noisy. So this lets it have good signals.

It just issues a command and says, well, okay, motor four, you're at 75%, and leave it there until I tell you to do something else. And then the little Cortex-M0 in the arm, okay, I'm just driving this motor. So from that standpoint, it's a cool kind of decentralized architecture to have all these little microcontrollers.

So you can have one that's master control, but you can have little ones at the leafs, at the edges, that are just taking commands and doing the analog stuff right at the edge. So, I don't know, it's kind of amazing. Well, in looking at this LCD controller, which is one of the ones I can see,

it's got a USB jack. That's the main USB that goes to the flight controller. So I don't know how that's routed exactly. I think it probably goes to the bigger Cortex-M3 that's the flight controller, and then it can deliver firmware updates to anything else. And then you've got this camera on here. As you said, you actually, it's sort of a VR system, except what's the opposite of VR?

It's like reality, but not your reality. Distance reality. Distance reality. Projected reality. Astral projection. Yeah, this is very much astral projection into a quadcopter. Yeah. Your soul now is here.

Yeah. Does that bother you? A little bit, especially when I bounce it off the ground. Yeah. This actually has two cameras. It's got the camera down there, which is just a regular analog camera that goes to the radio transmitter. And then that orange thing is a video camera, so I can record flights at higher quality.

So you have two cameras on here, as well as four flight controllers, or four motor controllers, and a flight controller, and an LCD controller. And you must have a couple more. I mean, I suspect there's an inertial system in here that feeds into the flight controller.

I think it's got a Bosch IMU. Okay. And there is an optional port for GPS, which I ordered from eBay, but disappeared. Ah, the postal system. Games I can play with GPS just by plugging it in. So yeah, it's a pre-built system, but it's hackable enough that, you know,

if I want to play around with the firmware, I can do that without building something from spare parts. So the flight controller is open source? Yeah. And that's just kind of crazy to me because there's just, you can do a lot with a flight controller. There's a lot of different open source flight controllers.

This is just one of them. And what is it called? CleanFlight. And have you looked at the code at all? I've looked at it just a little bit. I've poked around the GitHub, but there's a lot there. It has a lot of features, some of which this doesn't handle, and some that aren't quite finished yet. So a lot of the GPS kind of functionality, like return to home, which is kind of a nice feature if you get lost or if the radio cuts out. I'm not sure that's quite done yet. So there's areas to contribute there.

It's pretty stable. I mean, I still have this bias that open source software is not stable. It's running most of the world. So if that's not true, we're in deep trouble. I know, but it's still... Well, things like this that are clearly hobby-centric, I think, are usually less stable. But this seems pretty stable.

I don't think there's any issues with it. I haven't... You know, I don't see a lot of people complaining, oh, things fall out of the sky. And you're not going to blame that for when you did fall out of the sky no that was because i was uh showing off reaching beyond my capabilities yeah i thought well i can i can do a roll why don't i try a loop and then it ended aiming straight up and i didn't know where it was so i just turned the engine off and let it fall

to its death which wasn't a big death uh uh, I mean, when you have a lot of enthusiasts contributing stuff, I mean, people into this do racing and things. I saw the videos for the racing. Oh, my goodness. They're cruising through forests at 70, 80 miles an hour on these courses, avoiding obstacles, you know, and they're like three feet off the ground. They're not flying around like, you know, you take your DJ out, phantom up 100 feet

and take pictures of your house or something. These guys are cruising around at 80 miles an hour and this will do that. I'm nowhere near good enough to do anything with it. Yeah, I mean, you watch these videos on YouTube. It's like watching the forest, you know, the- You can be an Ewok. The speeder.

Yeah, well, Ewoks never played. But you can be the people on the speeders. It's pretty much doing the speeder thing in Return of the Jedi. So it's pretty amazing. So yeah, I think if people are passionate about something, they want it to work well. And I think they're willing to contribute and make sure it works well. I think the projects that have the most trouble

are the ones that don't have a lot of contributors. Well, it's hard. I mean, it's hard to get contributors because there are other people giving up their time for free. And then if you have a good idea, there's always a possibility that somebody else has it, and then you're splitting your contributor base. Right. I mean, I suspect that CleanFlight probably has some sponsorship by the quadcopter vendors. I have no idea.

I don't really know how that works. I mean, they don't... It's never really free. Nothing's ever really free. Yeah. And these guys, the one who makes my quad, I think they have to, you know, qualify CleanFlight before they put a version, a new version on their site, they make sure that it works

or they make small tweaks for their products. So there's some of that going on too where it's like, okay, here's the base CleanFlight and you can run it if you want, but if you want specific updates that are tuned to this quad, then wait for us to add in the specific stuff. I haven't really dug into the software enough to

really say. But you got started with Raspberry Pi. You took my Raspberry Pi out of the box and then you complained a lot about how I'd set it up. What? No. No, it just was in bad shape. What's the difference?

It wouldn't boot because it had bad blocks on the SD card so I'm not sure how this works oh we're going to blame the SD card it's not your fault at all it just was broken yeah go on believe in that anyway and then you hooked up the camera and did Python and OpenCD

yeah so I had this dumb idea to we're really just going to talk about my dumb ideas oh yeah that was that was that was really my plan i didn't i didn't write a lot of show notes this time i wanted to learn a little bit about machine vision stuff so i had this dumb idea way back before i got the quad that i would stick a raspberry pi on the quad and have it fly around and then you know aim itself at some fiducial marker like a red ball or maybe some sort of QR code

or something it could see and figure out easily. I think it would be cool if you could have a red ball in your hand and then dance with the quadcopter and then a blue ball in the other hand, a different quadcopter, and you can control them by making them mimic you. It would be very cool. That would be very cool. Many things would be very cool. Okay, so are you done yet? Anyway, the original idea was to have it pick up a ball That would be... Mimic you. It would be very cool. That would be very cool. Many things would be very cool. Okay.

So are you done yet? Anyway, the original idea was to have it pick up a ball and then drop it into a basket or something. Ah, robot soccer. And then I realized the limits of my abilities with all of this stuff. And I've downsized to a robot instead of a flying device to start with. So now the idea was to stick it on one of these insect-like hexapod robots that I have and just, you know, have it follow something. You know, you're holding a red ball, have it turn to face you or whatever. So yeah, I loaded up the Raspberry

Pi. OpenCV and Python are fairly friendly to get up and running on there. I played around a little bit, but it's painfully slow on an original Pi. I mean, I'm maybe doing almost no processing is a couple frames per second. So it's not... You were looking for circles. Yeah, trying to auto-detect circles. And depending on the parameters, it was extremely slow. Well, Pi loading the full-size images was slow for me.

It wasn't full-size. I downsized it. It's just too computationally intensive. Or Python loading images. Well, I don't know if it's Python's fault or not, because I expect it's loading, it's running the OpenCV normal library from bindings, so I don't think it's running in Python.

So if that's that slow, it may just be that slow. But then a couple of days days ago I started saying, well, let's port this to C++, which isn't working that well. What's gone wrong? So I didn't want to have the whole development environment on the Pi because it's painfully slow. And that's actually where I ended up last time.

I think we had the same problem. So I tried to set up a cross-compile environment. I spooled up a little Ubuntu VM, which is fine. Followed most of the instructions for bringing a cross compiling environment up on the Pi, which was fine. Ran Hello World. That all worked. It's getting the libraries to work. So there's a library

to run the camera called RaspiCam or RaspiCam or however you want to say it. And it installs itself into one place. And the make file system that they want to use is CMake, which I'm familiar with but not excited about. And things get installed in places with absolute pass. And long story short, I can't make it work. Hours of searching.

Yeah, I just can't make it work. And I'd have to go in and hack something in the make files that i don't really want to touch and so this is another one of those things where and i think i tweeted a snarky comment about this but you get an idea in your head to do something and then you start examining the reality of it well i, I think this is worse. You get an idea, you realize it's possible and that you actually have what you need to do it. And then you get shot in the foot by CMake, which is not what you want to spend your afternoon doing.

You're going to have to learn a build system. But this is true of everything, right? There's always some piece of the system that you don't know you have to learn when you embark on a project that's mundane and stupid. Well, not necessarily stupid, but mundane and bookkeeping and tedious and you know i remember this from from grad school too there was always oh i'm gonna i'm gonna do some fluid dynamics simulation and then you spend four days trying to solve an integral or or you know simplify some big math

expression which is all just bookkeeping and it's not what you want to be doing it's getting over that not what you want to be doing part that's not such a big deal when it's your job because somebody's paying you i guess and you have to do it you kind of have to do it and it's it's part of you know step by step we're doing a personal project i don't know know, for me, I feel like there's this, the motivational factor is much more, oh, I want to see this working. And the longer and further out that gets, the harder it is to have, you know, intermittent milestones where, oh, this is working a little bit. Oh, this is working a lot, a lot more. Oh, this is, this is finally working the way I i want if that first step from zero to that

first milestone is like two weeks of figuring out something that i really am not that interested in like a build system i'm just not going to continue so i understand in some ways where people who stick with arduino yeah and python and things where you can, you know, not have to worry about some of those issues are coming from. So anyway, frustration. But you have a long list of potential projects. Yes. And one of them I sort of mentioned at some point long ago uh that involved

ocean things will you describe that idea i know you're not working on it but i think i did on the podcast at some point but the basic idea was i watching some folks on twitter um stewart mcandrew apologize if I've got the name wrong. He's been working on CubeSats for a long time. And I thought, well, that should be applicable to terrestrial stuff. So what if you made a CubeSat that just

floated in water and had a GPS and iridium radio and you toss it out of a boat or something and have it float and just follow where it goes and maybe take some uh take some environmental data temperature and that kind of stuff but that was a basic idea and then you know if if somebody found it if it washed up it would say take a picture email tweet whatever and then throw me back right um it would be neat to have a little

wanderer yeah so that was the that was the idea um i have lots of dumb ideas i just don't do any of them you have more ideas i mean i tend to not make a list it's it's funny i think i always have a to-do list that's usually very long and then I just look at it and choose a couple of tasks, and that's all. I don't think about it much. I just use that to make sure I don't forget about stuff or to realize that it's been on my to-do list for a week, so it must not be useful, and I should just cross it off. And you made this list of all the projects you want to work on yeah and because i'm paralyzed by indecision okay let's go through this list didn't help well this list includes things that are not technical things these are other hobbies that i have that i've let go or things like that as well yeah we really do need more surfing lessons, don't we? That's not on here. Why not? I forgot.

You should add it. All right. Go surfing. So I'm just supposed to read this list? No, I figured you'd read one and then I would... How about the listeners tell me what to do and I'll just do the one they tell me. Okay, go ahead.

Big list of things to do slash projects. That's the title. Okay. Open CV experiment. O- slash projects. That's the title. Okay. OpenCV experiment. O-P-E-R-B-L-O-R-P. OpenCV experiments with Hexapod and Raspberry Pi. Okay, we talked about that. GPS experiments with the drone.

And clean flight. You just need the GPS to come in and then to play with that a little bit with what they have and then to start writing other software. Yeah. What do you want it to do? Do you want to have like geocaching but with quadcopters? I have no idea what that would mean. I don't either. Just have

one hovering until somebody finds it? Sure. Well, since the battery only lasts four minutes, I don't think that's going to work. Okay, okay. This is is gonna take a long time um learn to fly drone hey you actually did pretty good so i can almost check that off because you i mean earlier you you did there was a little bit of crashing but the last time we went yeah there was no crashing uh remote 3g based uh weather monitor with the electron uh single board thingy that's a cell motor the cell mode yes the from particle and we're thinking about putting that on our local beach and having it tell us when the when everything is just right for us to head down there instead of so that's actually probably a weekend project uh practice

with bare metal bring ups on dev kits because I haven't done this very much and I feel like it's hard and I feel like I should get some practice on it for just continuing development kind of education. And we have all these dev kits around, so it seemed like I could bring them all up in IR. But then you told me that that was easy, so I got less interested.

Well, I mean, it's easy because you do it a lot. And because most of the vendors now have example kits. I mean, you use the systems. You're not unfamiliar with IAR or the compilers. All too familiar. Yeah. But I actually had to bring up a Kinetis K60, which is no longer Freescale. Now it's NXP.

I was a little confused, but then I remembered. Anyway, I had to bring that up today. And I looked at the board, and I know it came with demo code, but I didn't really care what the demo code was going to do. So I just went online and found their big box of bare metal examples and pulled it down to my local drive found the example that was closest to what i wanted to do compiled and ran it okay and now i have to strip out all the cruft and then make it so it doesn't compile with warnings what the hell people why do you

examples compile with warnings see i wasn't thinking about just grabbing some examples and throwing it on there and i was thinking of making it work from scratch. Well, that's the thing. I don't, I have made things work from scratch, but I don't hardly ever anymore. There's always some example. I wasn't saying it was something I wanted to do as a, you know, oh, now I'm going to do this every time I have a new project. It was a learning experience. It was a learning experience.

It was a refresher course. I mean, I guess I had a client who had to port some startup and hard fault code from Kyle to GCC, and we were kind of baffled with the idea that there are multiple assembly dialects. Right. And yeah, okay. I guess I knew that because of startup.us

being an important thing. But... I'll just take that one off. That one doesn't sound like fun. None of them are fun. Is anybody paying you for these? No. Next one, write new theme song.

Oh, yeah. And outro. Oh, okay, fine. And outro. Yeah. these no next one write new theme song oh yeah and an outro oh okay fine and outro yeah great no i was just going to use the old theme song of the outro next one re-record a song that i've been working on for years so that my brother can play guitar over it okay next one relearn piano well enough to play beethoven sonata that I was working on a year ago and then stopped again. Continue in the video game course, which we have not been continuing on because somebody keeps saying, I don't want. No, yesterday I was willing to and you weren't.

That will be taken out of context. Next. Again? But you did do some Unity stuff. Sort of. And that's the next one. Unity physics experiments. Make a drone simulator, question mark.

Little 2D physics things, question mark. You did a little tiny drone simulator. It was pretty cute. It was a drone simulator in the same way that... I'm drawing a blank on a metaphor here. In the same way that that was a metaphor? The wind-up bath toys are an aircraft carrier. From scratch flight controller.

And that includes learning the physics of drone flight control. Don't set the bar very high there. Well, that's not going to be good. But, you know, I found some papers on the physics, and my physics is getting rusty, and it might be fun to go through that and fly badly, or at least have it hover. Hovering's hard. Hovering's hard? I would think so.

The stability there, it just... What's the whole point of a flight controller is to keep it stable. Oh, all right. Well, then I guess so. Ocean uh ocean probe project which we talked about audio processing experiments with dsps and music kind of stuff learn guitar take online course on something you can see i'm sort of kind of flailing uh go back and do some more astrophotography get my equipment together for that which i haven't touched in years no and i foolishly when we moved put the telescope near our bedroom so that you could take it out and then you told me you can't take it out on the deck because it's too wobbly

yep and the last two i just came up with uh it's not that heavy uh last two i came up with uh this week spherical drone yeah nobody tell me if these have already been done because i'm really excited about these spherical drone uh which involves not a quadcopter but probably a dodecahedrocopter good dodecahedrocopter i just came up with that and i'm keeping it you should go get dodecahedrocopter.com right this second dodeca dodecahedrocopter.com okay uh and then it's it's probably its predecessor which is spherical servo driven rolly ball but this isn't like bb8's spherical no no i wanted to have tell me about this it has little feet it's a sphere of like uh it's hard to describe. You know, some of those lamps, they have like radial,

it's like from a central point and there's radial, let's say sticks coming out of equal length. So if a starfish is 2D, this would be 3D? Sure, go with that. A 3D starfish. And then on the end of each rod is a foot, maybe like a Nerf dart kind of thing. With the sucker at the end? Sure.

Okay. And then those are connected to servos, and they can push or pull. And then it has an Arduino or some microcontroller in it that makes it roll by pushing the right ones in the right sequence. So this would be like a rolling stiff kush ball. Yeah. All right. Do-deca-hydra... Ball.

Ball. No, that doesn't work. Do-deca-hydra walking robot. Yeah. So these are the things, and I'm leaving several thousand things off here, but these are all in my head all the time. So when I go to start one,

I spend four minutes on it and then think, well, I should really be working on the other thing. And then I think I should be working on the other thing. And then I go and surf Twitter. And so you made a list so that you could prioritize and stay focused? Yeah. Yeah. Yeah. It's not working at all. Well, I mean, you made progress on the Pi. I made enough progress on the Pi to know I hate everything and don't want to use computers anymore. Well, yeah. I have been so in that state for weeks.

What I really should do, and probably the answer is, since those roadblocks bother me, is I should eliminate them. And if I don't want to learn enough to force the cross-compilation environment to work by hacking cmake or whatever then i should just punt and not do that and get one of the quad core raspberry pies and just build on directly on it i think i think i'm starting to come to the

conclusion that i don't need to be a first principles person on things. This is what happened to me in school too, is I often would get bogged down in first principles instead of just taking what they were selling, you know. Okay, here's the equation, do this. Yeah, but where'd that come from? And I'd spend three days in the library, you know, trying to figure out the fundamental physics behind something that I didn't have the machinery to do yet, which is why they were just giving me the formula. But that happens to me with this stuff too. And I was like, well, you know, I feel really guilty.

I should learn CMake well enough to be an expert in it. But why? I just want to make this thing go. Well, I mean, this is why a lot of things are sort of circles i mean you get started with arduino because you can make a lot of progress and then if you like it you can go to something where you get more flexibility and you learn more but the constant measurable progress is really important yeah but i'm supposed to be an expert I know and I feel like with our hobbies sometimes we try to be experts

well I mean not with my latest project we'll talk about that in a second yeah and that gets me bogged down in why don't I understand this well if I'm an expert I should be able to learn this in an hour. And then I go read, you know, docs for CMake.

I'm like, what is this thing? I get it, but come on. Yeah, I mean, I get CMake on the surface, but in this particular instance where somebody nine libraries down hard-coded a path, I don't know how to fix that because I don't even know how to figure out. Right.

So anyway, probably bashing CMake more than it deserves. But yeah, it's hard to get started on stuff. And self-motivation is tricky, especially when you do this for a living too. It's hard to come off a few hours of fighting with something or hacking with something and then go do something else frustrating or that becomes frustrating that's supposed to come off a few hours of fighting with something or hacking with something and then go do something else frustrating or that becomes frustrating. That's supposed to be your fun project.

Or like I coded for a lot of Sunday and then I got up Monday morning and I was like, I don't want to do paying work. No, I don't want to do this. Yeah. So if anybody, you know,

wants to tell me which of those many things I should do, if anybody you know wants to tell me which of those many things I should do that'd be great if anybody would like my suggestions what are your suggestions

I don't know it's just a matter of you know repaint areas and rehang pictures and do gardening crap oh

your list is practical my list for you is practical my list for myself is not always. I see. Anyway, well, that's that. Okay.

But that was started with the drone show thing, which we should mention again. Right. I didn't mention all of them. What's it called? Let's Drone Out was the show, and they're live every Thursday at 8 p.m. GMT plus one on a website called, I believe it's called poweringon.co.uk.

Okay. Those will be in the show notes along with a link to Christopher's particular episode. Yes. Okay. So my project, which was the first thing that sounded interesting to me after weeks of, I don't want to do anything, I'm just going to do what I have to do to get by, is for Bring a Hack. Which, I mean, it's Bring a Hack. You're supposed to bring a hack. This is one of the things that happens around Maker Faire.

Post Maker Faire. Osh Park sponsors it, and Lane is fantastic for doing it and it's Sunday night after Maker Faire closes it's still at BJ's Pizza I think BJ's Pizza in San Mateo

and I think it is from 8 to midnight 6 to midnight and it's free I mean Osh Park really sponsors it and lots of people come and I think it's a great way to meet other engineers, because we don't tend to get... And hobbyists. Oh, yeah, well, yes. But it's, I don't know, when you get a lot of engineers in a room, sometimes it's either clicky, or it's just, everybody gets awkward. But because you have your own, since you are supposed to bring a hack. You have something to talk about.

You have something to talk about. And I just think it's a genius idea that you have something to talk about. Yeah. So I wasn't really going to bring anything to bring a hack. Well, I mean, I might have. We're getting a new logo soon. I might have managed to bring t-shirts, but we'll see if the new logo is done by then. So, I mean, my plan for bring a hack was you know embedded FM stickers I wasn't I don't know I might pull out a ring or something

I don't know but then as I was feeling kind of I hate all technology we were watching cartoons and it was the evening after I had broken down a bunch of boxes, which is something I find very soothing. Like, Amazon sent us boxes, and then I use a box cutter and I destroy them. And there's just this constructive destruction that I find quite satisfying when I'm in a bad mood, which I had been for a while.

And so we're watching cartoons and John Johns, the Martian Manhunter guy goes up to this robot and he like puts his hand in and he pulls out his fist full of wires and the robot like acts differently for a little, little while and then it dies. And then later in that same episode, Superman turns to Johns and says, are you having a good time? And even though this cartoon guy is usually very, very serious,

he doesn't have much of a sense of humor, he said that he was in fact having a good time causing all this mayhem. I think he just said yes. He just said yes, but all the rest was very implied. So, I was thinking about these ideas of constructive destruction and fistful of wires and being amused by that. So, I decided that my bring a hack is called fistful of wires and it's going to be a bag

and you can reach in and pull out some wires and it will change the behavior of the audio motor and light subsystem inside the back and i i'm just really amused by this it's's terrible. I mean, it's got everything I hate. It's got audio. It's got so much hot glue. It's got flashing lights. And yet I am so amused. Well, see, now this is the kind of thing I think I should be doing. This is more silly, fun stuff that doesn't really have a point.

Because maybe we do things that have a point too much. And it's a good side way. I mean, you've learned to do some audio stuff. It hasn't been high levels of learning. I mean, because I'd already played with the NeoPixels and putting it on an embed-enabled platform was no big deal. I did finally get to load up my Adafruit sound effects board, which has been fun.

But, you know, that took all of an hour. And then the six hours choosing audio effects. Right. But I like it as a craft project more than anything else. Because right now, it's still single wires you pull out. But I have this plan to make it so that even though you're only pulling out one single wire a year, you still get sort of a handful. Sure, sure.

So yeah, that's what I've been working on. I think that's definitely in the spirit of bring a hack. Yes. And although I think some people are going to be like, I don't understand. Tough. Why is this cool? And I'm like, this is like performance art, man. Those are the people you have reach into the other bag that's got the charged capacitor.

I know. So, yeah, I don't, I mean, I guess I have some projects like you do, some projects that I want to do that I just have, I don't have time for, but I don't have a big list. I think the other thing that's been on my list was a natural language processor, Twitter bot, that I could just replace myself on Twitter with a Twitter bot. All right. It would read my Slack channel, and anytime something looked amusing, it would post it to Twitter. Yeah, nothing could go wrong with that. No. All right. Is it time for follow-ups?

I guess so. Let's talk about listener stuff. First of all, thank you to Trigva, who wrote us about the Amp Hour show where he was hoping to hear about a chip designer. No. No, no, I'm sorry. He was hoping to hear about a ship designer, but it turned out they were talking to a chip designer, and he was disappointed. And I did not know why you were writing to our show about that, except for the possibility that if we had said ship designer, we would

have had a ship designer, which then made me go look for ship designers. So next week, folks. But apparently the Amp Hour in their post statement at one point said to send us comments. You are welcome to send us comments about us or the Amp

Hour. Or anything else. Whatever. Although if you're going to send us comments about us or the amp hour or anything else whatever although if you're going to send us feedback of the i hate that sort um it goes a lot further if you manage to structure it in the form of i didn't like that so much i think you needed more information here and here and and I'm an expert, and I would be happy to help you either on or offline. Those sorts of emails I really like. Your speaker spoke too fast, or your guest had nothing to do with Embedded despite the fact that she hosts a show called Embedded. Yeah, yeah. Do you guys understand? I got to interview an NPR host. Who do you think we do this show for? I mean, like, really, who do you think we do this show for? Checking our subscriber count and it is declining. A hundred people a minute now.

Soon our only listener will be... My phone. Exactly. Yeah, no, I did enjoy doing Kelly's show, and I thought it was amusing that I asked her technical questions, and she replied with political answers, and it just cracks me up every time I think about it. As for the embedded show, I've been listening to it.

It is more political than, I don't know if political. It's more sad. I'd rather hear about technology. Which is another question we got. Can you recommend other podcasts podcasts look at embedded engineering besides the amp hour spark gap which i think has gone hiatus yeah embedded engineering um andre had one that he listens to oh no that was uh chris chris gammill said one just started up uh well and there there's engineering commons for general engineering stuff yes um they're not specific embedded but they're they're nice people um saran about software i was pretty pleased with her podcast it's it's

very easy to listen to it's not super technical technical. And yet, it does creep in, often at odd moments. Yeah. You listen to other software ones. Not really. Nothing you'd recommend? Not for somebody looking for embedded stuff. So yeah, I don't know a lot. If you other listeners do know some,

I would be happy to talk about some of the others. Because it is nice to have podcasts. Although if you are looking for one, or looking for some, some of the ham radio ones get into embedded systems. Again, not directly, but it's still a lot of the same overlapping information. So look for the ham radio, as well as, of course, our sworn enemies, the Amp Hour. Let's see, feedback on Tori's show. Yes, so we got some feedback

that some of the things about liability were perhaps a little overstated. So when we talked to our bookkeeper and tax preparer, Tori Chavez, we did mention that if you have a corporation, it's sort of an umbrella to hide under if somebody sues you. Of course, we mostly deal with it in the form of service agreements. If we take a contract that works out very badly for some reason, and they sue us because they don't believe we've, we've never done this, but they don't believe we've fulfilled our end of the contract, then we only are liable for what's in our business.

They can't go after our personal finances. And yet, if I went out, even if I was driving around going to a radio shack or an existing actual store. Doing something for business. Doing something for business and hit somebody with my car. No, that person can probably sue me and my business. So there are different levels of liability. And we are not lawyers. Yeah.

Insurance is a good thing. And lawyers. And contracts are a good thing.. Insurance is a good thing. And lawyers... And contracts are a good thing. Contracts are a good thing. So, I mean, the idea is, yes, corporation, being a corporation can protect you. It cannot protect you from everything. Certainly, being a corporation is better than not being a corporation. If you're dealing with clients on a regular basis or contract law.

Yeah. I think leaving it there is probably the best thing. And if you have questions, consult a lawyer. Yes. Okay. Do you have any favorite RTOSs? Or are we doing a lot of bare metal work these days? Almost everything I do as an RTOS.

ThreadX is the one that is kind of the IBM of the RTOS world. I don't think anybody ever got fired for using ThreadX. They're pretty pricey. Yes, they're pretty pricey. So what happens a lot of times is, you know, you'll do a bake-off between a bunch of products, and there might be one that's free-ish, and somebody will say, well, what about support? And so the company is more likely to spend a lot of money on an RTOS that might not be any

different or better than one that's free, just because they've got somebody to yell at when something hypothetically goes wrong. Now, the number of times it's actually been useful turns out to be really small. Well, and a lot of times those bake-offs look for one particular weird, tweaky feature involving stacked semaphores on top of flags or whatever. Or some impossible-to-reproduce-or-prove latency

in a very synthetic situation. And so you go with the expensive one because it ticks a box that nobody really cares about. But I have to say the ones I've used and seen, and you may differ, I don't see that there's lots of huge, gigantic, gaping differences. Well, I have recommended PumpkinSalvo because it is very similar to FreeRTOS, and yet for like 800 bucks, you get somebody to point at. Yeah.

And so it's a cheap way to get a little bit of support. But they're all providing kind of the same APIs. They all have a messaging system. They all have a scheduler. They all have some concept of threads and thread synchronization. And maybe you get more with Ethernet or USB stacks. Sure, sure. But those are usually bolt-on things.

They're not necessarily part of the core product. But it's nice if you can buy them all from the same person, if they're all quality. I've used some heavier weight stuff. I've used Green Hills on bigger systems. That was quite nice, but very sophisticated. I mean, it did things, you know, they had process protection, memory protection and things

while still being an RTOS. They had concept of virtualization even five, 10 years ago where you could take non-mission-critical systems, parts of your system, and wall them off. So you could have an infotainment part of your software that was running Linux, but then have an RTOS, and it'll all be on the same system. And the RTOS would keep Linux in a little box.

So there's varying levels of sophistication, but those tend to be extremely expensive. Whereas know thread x license might be a few thousand dollars you know greenhills license might be tens of thousands um so it depends on what you need and i i don't have any that i'm just jumping up and down to say yes always use this i would certainly look at artems i would certainly look at RTEMS. I would certainly look at FreeRTOS and go from there. And if somebody's saying I need to have a support license, well, that opens up a different discussion sometimes. It's not necessarily about the quality of the system. It's about business relationships.

But there aren't any that I'm looking at for fun. And I do a lot of development bare metal still. It started to go that everybody wanted an RTOS, and now I feel like I'm sort of penduluming the other way that nobody wants an RTOS. It doesn't make sense. With all the Internet of Things things and Bluetooth and all that junk? It's sort of like your quadcopter has a bunch of processors.

It seems like people are buying a processor to do this. You load it with their code and you never change it. No, I have no idea if that has an RTOS on it. I should look. Be amusing. Alvaro did write in to suggest... Zephyr Project. He was looking forward to playing with.

Yeah, Zephyr Project is a new-ish RTOS. Actually, I don't know if I'm lying about its age. It's new to us. New to us. It's from the Linux Foundation, is that right? Yeah. It's got the backing of the Linux Foundation, and it's a microkernel RTOS with the usual features.

I don't know if there's anything specifically unique about it, but you can check it out. It's effortproject.org. They have a nicely organized website. It sounds like they are trying to keep things relatively straightforward. They have a driver model that makes sense.

I've read through that. It may not be what you're used to if they're trying to do some new modern things, but that might be cool. Yeah, that would be on my list of things to talk about or to look at more. Let's see, on my list of things that I don't know much about, which includes Zephyr,

Platform.io. A couple people have emailed and asked about that. Yeah. I keep looking at it and thinking I've seen this before, and it's because people keep suggesting it and I keep looking at it. It's not platform.io. That's something totally different. Yes, it is platform.io.org.

And it appears to be an IDE for the Internet of Things, if you must. It reminded me of the embed compiler, or there was another one. I want to say code builder. I'll have to look it up. But there are a couple of online compilers. And so it looks like that, but I haven't tried it. Yeah, you know, it's like a lot of these IDEs are trying to solve this problem of how do you take the vast number of frameworks and libraries that are now available

and make it so they're integrated in your development environment so you can pick and choose and drop them in quickly. So that's one of its features. Codebender. Codebender, that was it. Codebender. Codebender. Yeah, I'm glad somebody's out there thinking about these problems.

I'm not sure this is... I don't know. If you work at platformio.org or you know somebody who does, connect us. It might be interesting to talk to them to find out what they're doing and why and all that. So, yeah, if you know somebody there. Also, if you know somebody at SpaceX. No, no, no begging for guests.

You can beg for guests. If it works, who cares? Speaking of Green Hills, that was like 12 topics back. Dennis Jackson emailed us and asked... What does that have to do with Green Hills? I'm getting there. He asked, what do most people use for their development operating system? Not their RTOS, but what do you program on?

Oh, okay. And I said, well, it's 50-50OS, but what do you program on? Oh, okay. And I said, well, it's 50-50 here, but we don't know. I think a lot of the compilers run more under Windows. Or Linux. IAR and whatnot. But GCC is building a lot of momentum and that's really cool.

It is building momentum. GCC ain't the new kid on the block i know but it is for embedded i guess i mean if you want a a business ready compiler not like arduino okay um what other compilers would you say? compilers? for embedded cross compilers Kyle, IAR, GCC so two of those are Windows only I think Clang is still that would be Mac but it's probably not

it's not just Mac it's just a compiler runs on everything yeah anyway It runs on everything. Yeah. Anyway, having not gotten a good answer from us, Dennis emailed Jack Ansell. I see.

Who did not know the answer either, so that actually made me feel a lot better, but said that he asked at Greenhills, and they say most of the developers use Windows. At Greenhills? That does not surprise me. No, no, no. For most of the people who use their product. Does their product run on Mac OS?

That is sort of a chicken egg confirmation bias sort of problem. Do most people watching movies in Windows Media Player use Windows? Probably. Dennis was specifically asking about FDA, FAA security.

I know it's what you're used to because at most of the places that I can remember over the last three or four jobs, it's been mostly Mac. And either running Linux virtual machines to do things that couldn't be done on the Mac directly or Windows virtual machines. My current client is mostly Mac. A few people run Windows, but most people are running VMs.

Yeah, the medical place I was at a while back, they were mostly pcs dells and such but most of my clients seem to use pcs sometimes vms um i had one client who did a lot of linux development and that was cool but then they still had to have windows in order to do lower level stuff. So, yeah, I don't know if there's an answer. I don't know if there's a study. The next person I would ask would be Michael Barr because he seems to take a lot of surveys.

Take a lot of surveys or give a lot of surveys? Take the results of a lot of surveys. Give a lot of surveys. All right. English is a terrible language. Especially when you use it incorrectly. Okay, so emails. Someone was a little concerned about jobs in emails.

If you email at embedded.fm for a contest or whatever, the only people who see it are me and Chris, and occasionally Andre or Chris Zwevek. Andre Csicak. Sorry, I guess he gets his whole name. And so, if we are doing a contest, we are not getting your email for any reason. You do not go automatically on our newsletter. You do not get sent over to whoever did the contest for spamming later. Yeah, we don't share emails with anybody.

It says that on the website. Well, yes, but if people actually email show at embedded.fm, they may not see the part of the website that says we don't share. We should hire a robot that takes those mails and have a third-party company send the mails back saying we don't share it with anybody. Yeah, so when we run contests, we're not farming for emails. And some of you may have noticed that if you win a contest, I email you and say, can I share it with the providing company?

Because we really don't share emails. Yeah, we're not doing this. We don't get paid by the contest sponsors. We're just running the contests for fun, usually. Yeah. It's not a backdoor sponsorship system. We're not getting paid by the email. Yeah.

We're not getting paid. So, yes. And then I wanted to say about the newsletter the newsletter also if you sign up for that your email you get once a week you basically get the rss feed from the blog and from the podcast the show notes and um i only did a newsletter because some people really like to have it in their email and if you do sign up for the newsletter, if the art comes in on the right day, I will probably put it on the newsletter there. So there might be a few benefits, but nothing that isn't, you know, you get it in the next 24 hours. Let's see what else.

The blog is still going. We usually manage three posts a week. Andre is still writing about Embedded Wednesdays and getting together what C is and how to use it. He hasn't started on a platform yet, but I think he's still headed

towards the STM32 platforms. Chris Veck is writing about microprocessors and has started his platform which is the MSP430 that's right no there should be more letters after that and I don't remember which it is

but he's blinking lights and talking about GPIO peripherals and it's pretty cool because it's a very inside the chip view and Andres is more of an outside-the-chip view. Yeah, they go well. And then I'm writing about toys or whatever I decide to rant about that week.

And every week I edit a podcast. And you often edit one or more of these. By the way, three a week, and we are all cross-editing and writing and it's a lot of work actually. Okay, so my next thing on my list is VR glasses. VR glasses? Yeah. Why?

I don't know. We can just take that off. Okay. Nothing to say there? Well, I guess one of the other things to do with the blog is the Hackaday Open Judging, which is to say, at one point on this show, I said, if you have a Hackaday project and you want to know what a judge thinks of it, since I'm not judging this year, but I'm supposed to be doing this advisor thing that nobody defined for me,

I figured I'd advise you about your project. So if you want me to for me, I figured I'd advise you about your project. So if you want me to judge it, I will. Pending, not losing, not everybody signing up. And then I'll write about it. And there's one of those up there on our blog if you want to see. I am a mean, mean mean cruel person but you know if you want to win if you want to know what what's missing in your project i'll help you because it would be kind of cool if our listeners won uh surprisingly i'd like me to insult you and your project that it was not no i i didn't

uh okay but it is true that that i haven't gotten a whole lot of people saying me me next me next i can't imagine why that would be um i did get one person who's who's seriously considering it and yet hasn't agreed yet um well that kind of covers my list. Do you have anything else? I had a couple of things. There was, oh, on the subject of podcasts, I found the thing that Chris Gamble mentioned, which was microfab.com has a podcast. Oh, Raleigh Hardware has a podcast, now that you say that. Yeah, and Microfab's podcast is the Microfab Engineering Podcast.

It's microfab.com. And they seem to talk about little electronics things. So it might be exactly what they're looking for. The other question that just came in on Twitter was, is there any podcast about electronics in Spanish similar to the Amp Hour or Abeta.fm? And I do not know the answer to that. So if anybody does... We do have some listeners who I know are based in South America,

so if you have any suggestions for that person, or if you don't know... You can be from anywhere and answer this question. That's right. You know, sometimes typing is easier. The other thing I had was the Salier job ad, which I think I'm going to just not say anything about. I don't think we can leave it there.

So let me summarize. Salier has a job ad for a fairly junior position. Support. has a job ad for a fairly junior position support and it wants thousands of hours of all levels of development and they are either going to find someone who is the shiniest unicorn in the universe or they're going to hire a liar and it is such an example of how not to write a job ad. So without going into extreme details, I guess we're talking about this, but it had every kind of development.

So first of all, it's a support, it's a phone support job. Let's get that out of the way first. It's a phone support job. You're going to be answering calls from customers and helping them with their problems. And they want thousands of hours of Java development. But they want, it wasn't Java development, it was every kind of development.

Yeah. They want you to be a software developer with experience. They want you to be an analog electronics engineer with experience. They want you to be a front-end developer, which means website stuff, with experience. They want you to be a UI and UX designer with experience. Did I miss anything?

I'm sure you did. If there was a discipline, you had to be at least a journeyman in it. And on top of that, they had a survey which asked questions about, you know, how many thousands of hours you've used JavaScript, you've written in JavaScript in the last two years, how many hundreds of hours you've used their device. And I guess the reason that bugs me is they're free to look for whoever they want of course yes i think they're not going to find this person because they finding you they don't exist but if they did why would they want your freaking support job right you're not going to find people looking for a phone support job who have these qualifications so i'm very confused as to what they're thinking. And first off, this comes from a place of,

I want Saylea to do well. I like their product. I use it. And we have used it. We've had them on the show. I would like them to succeed. And it looks like they're trying to navigate some direction changes.

So I understand that they're trying some new things. When I look at that job posting, as a senior engineer with 20 mumble years of experience, I get insecure. I feel stupid. I feel dumb. And I think... Is this what support people now can do?

If I think it's going to... isn't just a problem with with this particular job ad i see this a lot with job ads and so my point is not a specific one to this job ad it is a industry-wide problem where there's kind of an inflation of the kinds of qualifications that people want uh for various jobs to the point where if you don't have multiple degrees, it's not possible for you to fulfill them. And so people are either lying about what they're asking for and just hoping that, you know, by casting a wide net, they're going to get all kinds of people or they're, you know, delusional. And when I look at a job like that and I get insecure, I can only imagine what people coming out of college or with two to five years experience who are changing jobs feel like when they see jobs like this, they're not going to

apply. And so you're losing potential talent who might do really well at the job you're hiring for, who aren't even going to come in to talk to you. Because they're going to look at that and say, well, that's definitely not me. I may be an expert C and C++ embedded developer, but I don't have any analog electronics experience. And so that bothers me. It bothers me because it discourages people and it perpetuates this whole rock star, ninja, superstar mentality that everyone has to be just a unicorn in order to succeed in this industry. And you know what? Most people are just good people and good engineers. And some people are unicorns for a year, and then they're a good engineer for the rest of their career. And some people are just good engineers for five years, and suddenly they get onto some project that they're really hot about, and they're a unicorn.

And it's not all that we're all geniuses all the time. I wish that some of these job ads would be written for, if you don't know these 10 things, you are willing to learn them. Well, that's what we were talking about with Paul. Yeah, and if you, so here are these 10 things we want, how many of them are you good at? And if it's more than three, apply for the job. Yeah. Because somebody's going to apply for the job that has two, and is just going to use bluster and confidence to get through where you probably have what they need. You just, you can't tell from that impenetrable wall of we want everything.

Yeah. And I, you know, I tried not to be guilty of this in the past. I fought back to previous job ads I had written and I always tried to say, well, what do I not need to put on here? What's on here that isn't really necessary? Oh, I don't really need anybody with x86 assembly experience. That's not really going to come up. That's more of a gotcha. And this thing that's in a requirement, that's actually a nice to have.have. I always tried to load my nice-to-haves with a lot more skills than my requirements. My requirements were always pretty big in general. My nice-to-haves were

lots. And sometimes the nice-to-haves were, have you heard of this thing? Do you know what this acronym stands for? You don't necessarily have to have used it, but you have to know where to go to look it up and become familiar with it further in practical use so anyway salia write a better job ad yeah we said so that's all i got okay well um yeah well uh

yeah i should go uh figure out how to pull out more wires and make them do more sounds and blinks and i haven't i mean I've got the motor running, but you know, it just sort of zaps you when you nevermind, um, projects later. Uh, okay.

So thank you to Christopher for producing and co-hosting. And of course, thank you for listening. Please also check out our blog and newsletter. Are you a robot? Yes. You can find it all on Embedded.fm on the website, along with a contact link if you'd like to say hello,

or you can send us email show at Embedded.fm. We will be here next week, and here's a final thought to tide you over. Since it is a Christopher and Elysia babble with each other, that does mean that we should be reading Pooh. This is Winnie the Pooh, and from the last point I read, once upon a time, a very long time ago, about last Friday, Winnie the Pooh lived in a forest all by himself under the name of Sanders.

What does Sanders mean? asked Christopher Robin. It means he had the name over the door in gold letters and lived under it. Winnie the Pooh wasn't sure, Christopher Robin said. Now I am, said a growly voice. Then I will go on, said I. Embedded FM is an independently produced radio show that focuses on the many aspects of engineering. on, said I.