UM LIVESTREAM: 2026.18 - ESP32-S-WHAT?

[00:00:32]Heat.

[00:00:36]Heat.

[00:01:15]Oh my. Always with the dramatic music.

[00:01:18]Sean, what are you doing? Cheers everybody.

[00:01:24]I'm not ready to start the stream. I needed like five more minutes.

[00:01:29]Also wake up with cracked lips and they're really sore. Anyway, hi. Ah. Um, hello everybody. How are we all? I really wish I could put my camera. So, I look at the screen, right?

[00:01:47]But my camera's here. I wish I could put the camera here so I could look at the camera, but then it would be in front of my screen and I won't be able to see the screen. Awkward. Like, someone should uh find out how to put a camera inside a monitor. That would that would be awesome.

[00:02:02]How is everyone?

[00:02:04]Uh, wacky wacky today. Warning you all. I'm a bit uh all over the place. Um, so, uh, Austin was saying that, uh, time goes past so slowly and I'm thinking it goes past really fast cuz I blinked and it was two weeks ago that I streamed and um, I've done nothing since. Well, that's not true. I've done so much stuff since. HMS says, "If only there was a tripod, you could 3D print for it." E, >> but it would still be I still need in front of my screen. Like literally in front of my screen where I look at the pixels. Anyway, um I've done stacks for the last two weeks, but none of it my own stuff, which is just really, you'll see. I'll show you some um I'll show you some things that I've been seeing here for two weeks, and you'll understand.

[00:02:59]And it's a bit frustrating, but um get a teleprompter. I don't need a a tele Hello AJ. A teleprompter.

[00:03:07]So So I You mean so I can look at the screen at a weird angle? No. Look, it's fine. It's okay. It's just awkward. Do you all think it's not awkward that I'm not looking at you? Like I'm not looking at any of you right now. Now I am. Now I feel like that's a bit bizarre. Anyway, um I could probably hang the camera from Anyway, we're not even going to go here now. Let's Let's stop talking about the camera. How's everyone's weeks been?

[00:03:33]Hang on. I need to check something. Uh yeah. No, I don't think so. Um sorry, I'm just uh having a bit of an argument with Claude. It's uh it's pretty funny.

[00:03:48]Um and when I say it's pretty funny, it's tragic. Claude is tragic. Um Claude has regressed so badly, I got to tell you. Anyway, um not we're not going to have a claude bashing day.

[00:04:00]I need to just check that something works so I can show you and okay. So cool.

[00:04:12]Oh my god. Sorry. I told you I wasn't ready. I needed like 5 minutes. Um but it's too late. Too bad. So um yeah.

[00:04:26]I'm so sorry. I'm so sorry.

[00:04:29]Breathe. Start again. Should we start again? I'll I'll start the intro music, right? And then we can uh pretend that that whole portion never happened and we can just start from scratch. No. So, okay. So, two weeks ago or just after my stream, I think it was the Thursday, the uh my new hub boards uh arrived. USB hubboards arrive and I've not even opened them. It's been two weeks.

[00:04:59]They're still in the packet because I've had zero time to myself, zero time for my own things. Um other than, you know, either 5:00 in the morning at home hacking away at some code for the display or evening when I get home, but like no time here to do stuff. I planned to. I tried. I just could not get get this going. So, I've not even opened it. I'm gonna open it now to show you all what they look like. If I can find where I put my knifey knife. There it is. Oh, that is dirty.

[00:05:38]Here's a new blade. Probably. Um, let's not cut my fingers.

[00:05:44]So, I decided I didn't want to hand prototype another board there. There's a lot of parts, a lot of repetitiveness on it. I' I've built three of them now, and I didn't want to do anymore by hand. So, I thought, stuff it. I'm just going to panelize it. I'm pretty confident with this design. And if it's wrong, well, I've just loan some money. But if I'm right, then I have production boards.

[00:06:10]Get out.

[00:06:13]So, I've not even looked at them.

[00:06:16]Wow.

[00:06:18]Okay. This just be a moment.

[00:06:21]The uh the Bible gets stuck between the the the things. It's stuck together and so Here we go.

[00:06:31]No, stuck on that one. Destroying it.

[00:06:37]Wower. So, these are the USB boards. How good do they look? And I went black, of course, cuz whatever.

[00:06:47]Smart hub ESP32S3 6 port USB hub.

[00:06:53]So this had all the design changes I spoke about the other week.

[00:06:59]It um yeah so I I need to but these need to be I'm going to pick and place these. I'm going to at least pick and place all the passives all of the all of the parts that I've got on feeders already or parts already used which is a fair chunk. And then the rest I'll do by hand, but that'll take care of like 2/3 of the parts on the boards. So, um, I only got five panels because that's the, you know, minimum they do. That's 10 of them. So, if they work, I'll have 10 of these that I will put into stock once I work out how much they're going to be.

[00:07:35]And then my and if it works, I'm I then have to go through the tidy up all of the schematic and the PCB stuff and then publish that. So, as I said to you, this will be I'm going to I am going to release the files for these. So, you'll be able to make your own or you'll be able to change it, do whatever else you want with it. Move the ports around for those that want all the ports on one side and all those things, right? So, um, yeah, but I need to validate it first.

[00:08:05]So, there's that. I also got my first production boards for uh for my so it does have a name. Oh, you are going to kind of see it anyway because it's going to be on the web thing. So, they're called pixel racks. It's a pixel rack.

[00:08:27]And so these are my controller boards for the pixel rack. And um this was a another risky panel to make because I I had a validated board.

[00:08:44]So this was my final. Is that my R seven?

[00:08:51]I think it is. No, this one. Right. This was my final design. my R seven board.

[00:09:00]Oh, look at that.

[00:09:03]Set fire to my soul.

[00:09:07]Anyway, um and working great. And so I was This is the one that I showed. This was running the thing I showed the other week. And I know we were talking about all things we could, you know, people with suggestions and stuff. So after the stream, I had a chat with Quindor. Hi, Quindor. Quindor's here. Um, they had a name for a while, but I just wasn't saying anything. Um, and Quindor convinced me to add another port, one more port. It was a challenge. I had to really push some stuff around, but I I added a second port.

[00:09:47]Um, so this is the this is the new controller board. They look the same almost.

[00:09:53]almost but there's now two STEMIQT ports and the second one is not uh so the first one's obviously I squ C on the I squed C bus it's okay it was a good idea but the second one is breaks out two other IO and I guess technically it can be any type of expansion I don't know what I'm going to do with them yet but um well beyond the ability to connect.

[00:10:24]Um, what were those sensors? DS. I don't know what the model number is. Those are one wire sensors. Apparently, people use those use those DS. Was that one something something mental blank um in the service rack to in their server racks so they can measure temperature different places and stuff. So, the point is to add external sensors for those people that want to use one of these actually in a rack embedded properly properly in a rack or whatever.

[00:10:55]Just it's just more expansion. But the other thing was, and this is um Austin's fault. I don't know if Austin's fallen asleep if he's watching. Um but he pointed out to me that I actually had the um the CH390, so the Ethernet chip on here was configured in the wrong mode.

[00:11:15]It didn't matter. It's only got one mode and I configured it not for that mode, but it falls back into that mode anyway, which is fine, which is why my the boards were working before. Um, so you can tell that you can normally tell a an Ethernet chip that the the power is it's going to be in either in current mode or um voltage mode. The CH390 only supports voltage mode, but I had it configured for current mode, which is fine. It's just the way you tap into the uh the center pin on the RJ45.

[00:11:46]Um so I changed that as well, which is fine, but there two changes I made to a board. And the the extra port actually was a lot of shuffling of parts around on the board. And so it was pretty risky to make a a production board off that without prototyping. But again, I didn't want to prototype by hand on these ones anymore. So I ran production boards and they're great. I've got now one in a rack and I've got three other ones that I haven't soldered the through hole on yet and more panels. So that's that's locked off. That's pretty cool. Um that is but the controller board is all I managed to do for myself this whole uh two weeks apart from obviously manufacturing stuff like had some lots of reflow master pros cuz I've been out of stock or I've been low on stock and stuff but it's all been client work. So, that's been a little frustrating, but I do have some stuff to show you. I do. Uh, yes. What? What AJ's saying?

[00:12:46]What's the best he thinks? Nice effect.

[00:12:49]Oh, which effect? Which one? The edge.

[00:12:51]Yes. I wonder who came up with those that edge stuff. AJ, sorry, my coffee's getting cold. I didn't say hi to everyone in chat. Look how messed up I am. Um, uh, Austin, if you're still around, hello. Uh, Graeme, good morning. Mr. England, hello. Windor, yes. Spooks wood. Infamy 4K. Uh, David Essa, hello. Uh, Desipu, I think Apple did that, but they made a big black hole in the screen for that.

[00:13:25]Yeah. Now, I was I was I've got a there was a joke about the camera in the monitor.

[00:13:32]Not comparable. The cameras though, unfortunately. Um, Johnny, hey, Rat Madness. Hello, Rat Madness. Unexpected heavy industry. Not so heavy.

[00:13:42]Heavy as my daughter used to say when she was little. Um, Mr. One Spring.

[00:13:48]Hello.

[00:13:50]Super Bruce is here. I've really lost track. I can't uh I know. I'm sure I've missed people.

[00:13:58]I'm doing a really bad job today.

[00:14:01]Um, yeah. Okay. Well, there you go. So, I have some S31s on the way. Well, I've got some S31 modules samples on the way.

[00:14:15]Um, I'm still waiting for actual silicon. Apparently, that's being I'm in the batch. I'm in the queue to get from the batch that's happening in the next 3 or 4 days, apparently. and then they'll ship them to me. Um, I don't know how many I'm getting. Probably only five chips. I don't know. I don't even know which ones, the 16 or the 32 uh megabyte um SPI RAM ones. Uh the sorry, the PS RAM ones. I I don't know what I'm getting. Um and I've had to fight I've really had to fight to get in at this point in the queue. Um, it's been a little frustrating. Anyway, but I I do have two modules on the way.

[00:15:02]Um, obviously I don't really use modules much in my projects. Very hard to make a small board with a module, but the S31 is big bigger than or it's not as big as the P4. Um, but it's it's bigger than the S3.

[00:15:15]So, it will not fit on pretty much all of my designs already. So, if I'm going to use the S31 on a like a microcontroller, a dev board or something, which I I want to look at doing, um I might have to make a new board for it.

[00:15:32]Um but it yeah, it's an insane chip to ship. It really is an insane ship and I'm looking forward to using it. Um but there's nothing out for it yet. I mean, the data shed is still at.5 um pre-release and so not so tiny. S31, no. Um, I don't think I can get it on an edge board either. Um, I might be able to squeeze it onto a Feather, one of my feathers, but not definitely not on the Pro and definitely not on the Tiny or the OMG or the Edge. It's just a big chip. I mean they they could have gone the same LGA package as the S3 and there are pins on the S3 that they could have exposed to to have more IO but that would have been like a half step but they couldn't fit the other stuff that wanted inside architecture uh is P4 alike but you will need the newest IDF. Yeah. Yeah. Sure. Sure. Um but the advantage of the S31 is it doesn't need all the other power stuff that the P4 needs. like it doesn't have MIP CSI and all those things. So, it doesn't need the multiple voltage rails and whatever.

[00:16:39]It's literally it should be as easy to implement an S as an S3. Um, yeah, the P4 is only just released.

[00:16:50]Well, it's actually discontinued and then re-released. You all know that, right? There's a new version now, the X version, and they they've dropped the first release they did because there was problems with it. So the the P4 went through many revisions um and they actually had a couple of candidates for mass production that they pulled. So the P4 took a long time to come out because it yeah I mean you think about it every time they go to the fab to make a um to make samples to see if this is the one you know that takes time. They got to wait for their slot and cost a lot of money and they actually went through a lot of issues. Surprisingly, the S31's only gone through one round. Like that was it. Their chip that they they're launching is the chip that they went to samples and, you know, engineering samples on. So, um, according to Google, SP32 is hundreds of times faster and vastly more powerful than Intel 808. Yeah, I'm not surprised.

[00:17:51]And that was the original ESP32. So, for me to look at what I can even fit on a board, I need to actually make a footprint for it because there isn't one and a schematic symbol because there isn't one. So, that's what I was going to do this morning on stream. But before I do that, I want to show you some stuff that I have been working on, which is cool cuz I I I listen to almost all of your No, I listened to all of your feedback two weeks ago when I showed my u pixel rack.

[00:18:20]This is the So, by the way, this is the pixel rack currently, not in its frame.

[00:18:25]And yes, it's it's elastic bands because the the frame is what holds it together.

[00:18:31]Oh no, bad timing. Bad boy.

[00:18:37]Sorry.

[00:18:39]Unexpected spoiler. Um, unexpected spoiler. Bugger bugger.

[00:18:47]Actually, if it would have said anything, you wouldn't have even known.

[00:18:49]Anyway, so yeah, this is it not in its rack face. But this is a case from JLC.

[00:18:56]This is in nylon. This is the case I'm going to be using or I'm going to get JLC to make or one piece cases in uh nylon. And um this is actually one I wasn't I wasn't wasn't this I was going to show. And where did I put my And these are my custom rack plates from JLC as well in steel and they are great.

[00:19:30]So I've solved all of that. I'm just hacking away at the code, but I listen to everyone. Uh how is the process cost quality? So the cases are not cheap uh because of their size and these are not 3D printed. They they they use MJS or whatever it is um to do them and they're an awkward size and they're they're not cheap. They are um unfortunately but the quality of them is superb and um yeah I did get a 3D printed version from them as well just to compare in ABS and it just looks it's a nice quality 3D print don't get me wrong but it looks pants compared to the this nylon case it looks like a um production case right and the the finish on it matches the powder a coating on the front. So, um yeah. Uh so, they came out really really well. The I was really impressed with the rack, the the the steel face plate.

[00:20:34]Perfect. Um so, that's exciting. So, I'm going to I want to show you some things I did in the software. Um I'm not going to show it to you on the actual display. I got no way I'm mounting right now, but but you'll be able to see it in the actual Let me switch. Let me switch and you'll Here we go. This the interface mostly the same as what you saw last time kind of. There's some new things I've done I wanted to show you because it's pretty cool. So, and some of them were suggestions from people. So, I'm trying to think where to start. Um, I'll start with the something that is is far from complete. In fact, the the the redo I'm doing on it right now I completely broke. So, it's not on this device here. But the sprite editor now has So, this isn't quite the view everyone's going to see. Um I obviously have uh high permission levels. you'll understand in a moment, but the um there's a an online gallery of um I'll be able to my my plan is to be able to push new content to a an like an endpoint online that these will ingest if you want to to get new things like sprites and stuff and icons and all sorts of things. Um, so you know, you've got your sprite editor like before where I can go in his pinky. I I can edit sprites, save my own sprites and stuff, but I've got the ability to also publish them to the cloud, a uh uh non authenticating whatever cloud, right? Um, and so you'll be able to browse a gallery of stuff and you'll be able to download sprites and stuff if you want.

[00:22:32]But the the the version of this this it's kind of in a broken state at the moment. Um so but my plan is to be able to do that for like sprites, but sprites now are also um I mean I I can do a sprite that's the whole size of the screen, right? So, you know, there could be custom backgrounds, custom, you know, animated backgrounds and all sorts of stuff, but that's pretty cool. Um uh Okay. And then um people want to upload. No, no, no.

[00:23:06]People will No, people can't publish.

[00:23:09]No, no. You'll be able to download content from it. You won't be able to push content. I can push content. You can't. Uh we are we are not going to be doing um dword uh sprites.

[00:23:25]So, no. If you want to make your own for your own device, absolutely do what you want to do. But you won't you won't be able to publish, right? It's just a a place where I can publish new content so you guys can get extra visual content without having to, you know, do a firmware update or, you know, to to to download new firmware every time I want to put more visuals in there. Does that make sense? Yeah. Um, so says it'll still buy one.

[00:23:58]>> Thanks.

[00:23:59]Um, so, uh, in settings, check this one out.

[00:24:06]Discoverable.

[00:24:09]So, I've added a way that you can um on your local network, you can find other pixel racks. So, if you have a if you got two of them, they can find each other. So, quite a few people were talking about um if you got multiple of them, you can sync them together or you know share content as as in make them all do the same things or whatever else. Well, this is the the foundation for that and it'll all be behind a secret. You you pick your own secret.

[00:24:44]They'll have to obviously match secrets.

[00:24:46]So, um would also allow you to put them into subgroups if you wanted to like a work groupoup name, right? Um and but yeah, you'll be able it's all discoverable and and right now it's using um um MDSN multiccast.

[00:25:03]I might change it to UDP. I don't know.

[00:25:06]Um uh I'm a bit worried about the multiccast thing. Anyway, it's um but I've started the basis for that. So you will be able to connect devices and technically like I'll have to obviously implement some type of interface be able to maybe manage all your devices from any device right um pull content in and out and stuff like that and speaking of pulling content in and out there's a file manager which so this is looking on the little FS flash um and you can obviously look at stuff so like the the scheme schema. Here's the schema of everything where everything runs from. Um like here's the JSON that holds all the the modules. So you can view files obviously you can go into folders and you can download files and um you'll be able to also upload. So you'll be able to back up and back up all your content off a rack, whatever else. Even hand hack the the uh the JSON if you want to. It's all just JSON, right? Um, you know, like here's a uh this is a sprite, right? I mean, there's obviously all the the data encoded to make the pixel data be as small as possible, but it's just JSON.

[00:26:31]Um, so yeah. Um, we haven't even got to the cool stuff yet. Uh, nothing. There is some things there but not published yet. Okay, so back to display. So here's the cool stuff. Um, let's start with Okay, you saw I showed you the other week that I had um might just move this up.

[00:27:06]Yeah, that I had this canvas editor that I was working on. Right.

[00:27:11]um and that had a timeline mode, right? Well, this is all working now. So, you've actually got full timeline control if you're animating stuff. Um you see that, right?

[00:27:26]And then you you've actually got um a root canvas. You can actually animate the canvas. So, the canvas itself is everything.

[00:27:34]And so, you can animate the canvas moving and opacity and stuff.

[00:27:39]Everything's working. adding adding um and deleting key frames and everything else is great. Um and obviously in the editor you can you know click on things. Oh, so I have a new pallet system. So this is the color palette. It's not a operating a system palette picker anymore. Um it's a fixed palette. Um the reason I did a fixed palette is firstly all of these colors have been really vetted to work on the display. There are some colors that just don't work great on an RGB LED, like a like a pure blue. Um, doesn't look great. It It's actually really hard to read. Um, and anyway, so these colors are colors I've spent quite a bit of time tuning. I'm still tweaking them a bit. It's the full color palette. Um, no brown. Well, obviously, have you tried doing brown on an RGB LED? Doesn't quite work. Um, you got close. I mean, you know, this is a pretty close to a brown when you're looking at it on device.

[00:28:43]So, you got your main colors here, and then you've got a, you know, anyway, the point is it's a fixed palette now. And anything that uses color palettes uses this. Um, the advantage of this is it's for the sprites, it's index, and shipu. Um, for the sprites it's indexed, right? So I don't have to store the color. I just store the index into the pallet. Um, anyway, so that's pretty cool. Um, but but but check this out.

[00:29:23]Zoom.

[00:29:29]What?

[00:29:31]What? Sorry. So originally this pallet, you can also use your scroll wheel.

[00:29:39]Originally this was designed to do a single frame page of content. So the you've got the other systems you can do scrolling text and stuff I showed you, but this was designed to compose a single frame, but it'd be nice if you can animate things coming into that frame. And to be able to animate things coming into the frame, you need to be able to position them off the frame.

[00:30:00]So, how do you position them off the frame to animate them in? Because this is fully animatable, remember? So, the you can now zoom in and out. So, you can actually build content that'll sit off, right? So, there. I don't know. I'll put that over there. And then in the timeline, I go over here. Oh, that's already animating in. Which one's that? my multi row.

[00:30:30]Okay, move that over there. Um, I'll put here. Obviously, it doesn't fit. I'll put it whatever. So, you can now animate stuff coming in and out of the frame.

[00:30:46]Or you could on the left and right is the full frame. So you you could actually build a whole frame with the content on the left and the middle and the right and just animate between them all by actually animating the um the root.

[00:31:11]All right. The point is you can zoom in and out now of the content.

[00:31:19]Why is that? Oh, interesting. I don't have my scroll wheel working in timeline mode. Okay, I have to fix that. It's only working in the editor. Uh, whatever happened to I really want to go high resolution screen. Whatever happened to that? H. So, I'll show you what happened to that in a minute.

[00:31:46]I'll show you what happened to that in a minute.

[00:31:49]So anyway, you saw most of this anyway.

[00:31:51]So this is all working now. Um, but something else I added. So it occurred to me that remember I showed you, you know, if you want to put some data on the screen, you can, you know, pick from a data source.

[00:32:08]So this is the open weather data source.

[00:32:10]I can pick the current temperature. I can put some text in the front, text in the back, an animation, how it's going to look visually. I added some new new visuals. So, you can now do like rainbow characters like this. Um, that's where it's it's animating per character. You can do rainbow scrolling. Uh, hang on.

[00:32:29]Let me just static center, right? You can scroll horizontally, left or right. God, that's really fast. Um, I need to change some of the defaults.

[00:32:42]There we go. You can so or vertically.

[00:32:46]Um, and then there's also rainbow text which just does the whole text.

[00:32:51]I've added some new effects. Does it dounner?

[00:32:56]Maybe. But but but the problem with this was it's really great. You can do a static. You can we can fade it in and out. We can scroll it left to right. We can scroll it and hold it. So it'll scroll in. Hold for 1.5 seconds and scroll out. The problem is you can only show one bit of information.

[00:33:16]And so if you wanted to do for open weather, open weather actually gives you a description, a weather description.

[00:33:23]Um, uh, well, there's a feels. No, that's not feels like. What is it?

[00:33:30]Uh, I added a whole bunch more things as well, but um, currently overcast, right? So, you can't do that and the temperature at the same time.

[00:33:45]So, how am I going to solve that?

[00:33:48]Well, I everything you saw there is going to go.

[00:33:54]I'm not going to ship it.

[00:33:57]What I've done instead is I've So, that was a sequence. I've now got a V2 sequence just for now. I've got both in here. And this is what a V2 sequence looks like. You ready?

[00:34:09]Check that out, baby. Look at that content. That's all one sequence.

[00:34:15]Everything you see on there got icons and sprites, got different colors, different text effects, different data, all together. And it's all drag and drop.

[00:34:35]So this line here is the construction of everything. And so you can see here I've got an icon. Oh, I've got a new icon picker. Going to be shipping a whole bunch of icons, right? They're all two-tone icons. So this is just my my starting ones. Um, but you can see that one was the skull and that one there's that star thing. And you can just pick a color and it'll do the two tones, whatever. So if I want to change that one to, I don't know, a locked box, you'll see it when it goes past again.

[00:35:10]um sprite picker, color picker, and then effects. We'll get to those.

[00:35:19]So, basically, if I go, let me just close that. I'm going to go a new Let's do one up here. Let's go into info. I'm going to go add a sequence.

[00:35:29]So, this is just going to be uh net work info, right? And I literally just drag these. It shows me all my tags for all the different things that I can add instead of that drop down. So, I can say um uh I'm going to put a color. So, you see I'm just dragging in, right? I'm going to put a color.

[00:35:50]Uh it's going to be I know red. Why is that there?

[00:35:58]Okay, that's a bug. That shouldn't be there. Um so, okay, red. And I'm going to drag in the IP address. Yep, there you go. And I'm going to, this is currently set to static center, which is fine. And I'm going to now um stick an effect in here. And I'm going to make this effect a no vertical gradient.

[00:36:19]Uh yeah, red and blue. Why not?

[00:36:22]And then I'm going to Oops. Um subnet.

[00:36:28]Why is sub not filled in? Okay. No, we'll pick something else. M address.

[00:36:33]No. Why not populated?

[00:36:37]Interesting. Okay, I'll have to look at that. So, and you can insert text. I'm space here. I'm going to go um up time and then I'm going to stick an icon.

[00:36:50]It's literally drag and drop. And then you can go actually no, I want the icon to be over here.

[00:36:55]You can construct the whole thing.

[00:36:59]And let's make it uh scroll right to left.

[00:37:04]See it real time live. And um yeah, just build up the frame with whatever content you want. I'll have to look at why these aren't being populated.

[00:37:15]Is last boot last boot time. Is that being populated?

[00:37:21]Unknown boot reason. Okay, so I picked a really bad data source.

[00:37:27]What can I tell you? Um the point is you can now construct and join any data together. So you've got you'll have two ways of doing it but you can do combination of color and whatever you want like this or you can use the canvas editor to do really complex things. So this is all about a single line that you can scroll and you can construct and you can honestly make it as long as you want, right? You can put as much info as you want and it'll just keep scrolling.

[00:37:56]Hey Alan, it'll just keep scrolling if you're in scroll mode. Um, obviously if you do a static center and it doesn't fit, it doesn't fit.

[00:38:06]That's on you. But, um, yeah, you can add as many icons and sprites, whatever you want to do on it, or you go the canvas editor, which you can do really complex constructions.

[00:38:18]So, this is the new sequence editor, and the old sequence editor is going to go.

[00:38:21]all this stuff like this is going to disappear because it's just it's it's kind of cool to do something very quickly simple but you you're really limited in what you can do um and everything that you can do I have to hand code and to add stuff I need to ship firmware but with the V2 editor you can construct how you want it right you can make any text any color any effect um yeah So now the uh what am I going to do about the high-res display? Well, look at this high-res panels. This is just for me testing right now. If I click high-res panels and I click save is going to reboot.

[00:39:13]This is how I'm switching just for the moment.

[00:39:16]And look at this. We got high red panels.

[00:39:23]That's a joke cuz these are not high-res panels.

[00:39:27]But this is now high-res panels.

[00:39:34]So now everything in here is working on a double or it should be. Oh, hang on. I didn't refresh it right now. You see here everything is double pixels which is why I uh had to do this stuff as well to be able to lay stuff out properly.

[00:40:01]Anyway, we can you can now do two lines of text every so everything is already set up that you can do the high-res. For those that don't remember or or uh don't know what we're talking about, I've currently got on the way to me new LED panels that fit inside here.

[00:40:25]They fit in in here, but they will make the pixel density twice as many pixels.

[00:40:33]Not twice X, twice Y, just twice as many pixels. So instead of 11 pixels high, it's going to be 15 high and 175 across a lot of pixels. 25,560 LEDs in a 19in rack.

[00:40:51]So, it's trying to show the high-res stuff with the low res panel, which won't work obviously. And I don't know if I'm going to be able to ship it because I don't know if I'm going to be able to cool it. I have enough power. I don't know if I have enough cooling to keep it cool, but to know I need to build them. So, the panels are on the way and I need to be able to design on them.

[00:41:17]Yes. Your side has 320 x 240. Correct.

[00:41:20]Doesn't work so great in a uh rack though, does it Alan?

[00:41:26]So yes, the editor now fully supports either not real time because it's a complete buffer shift and everything else. Now to get this working though was not easy. I had to port the whole entire project into a hybrid Arduino IDF. I almost went full IDF and just dropped the Arduino stuff completely almost. But I had to go hybrid um because there was not enough SRAMM to do this with a straight Arduino project.

[00:42:06]Arduino meaning platform IO, right? PIO Arduino. Um because I had to cuz so all the screens are driven from DMA through the parallel interface and so to do that with the extra pixels, I needed to double the the memory used by the DMA buffer which put me down to uh not enough contiguous memory to do any TLS at all. Zero. And the only way I could get around that was to get access to the IDF level uh SDK configs to tweak everything everything to crawl back enough um SRAMM to ensure that my max block available stayed above 32. You need 32 to 35k to do a TLS call. And if you don't have in a contiguous block that much, you'll every single um endpoint call will fail. Um so once you go full, never go back. So no, look, I've got I'm managing projects, client projects in full IDF. I I used to be intimidated by IDF, I guess, years ago, whatever. Um I'm I'm not now. I'm quite comfortable in IDF land but it would have been a significant rewrite like significant rewrite because you know not not like everything about it would have been new new written code because everything's different like none of the API so none you know none of the function names are the same none of the API calls the same all all the construction of stuff and everything else so I went the hybrid first the hybrid works and I just have enough like you see here sitting at might be able to see sitting on 36. Um if I look at the log um I've got running um free heap stuff everywhere right to be sure but I'm not getting any fail. So if you look at all the things like so um so what do we got? Uh endpoint hits so one so four endpoint hits here. Uh, no.

[00:44:26]Five, six end points here. 7 8 9 10 11 12 13 14 15 16 I mean, look, look at them all in a row. Bang. Hit hit.

[00:44:36]Yeah. And um my memory usage ended at 37K for max block available.

[00:44:44]started I mean obviously it okay I started at 4098 went down to 3787 and here all the way through never drops below this was a challenge this took me about three nights to get working and the problem is when it's not working it crashes hard and takes down the whole thing whole project so so gradually port stuff from IDF to if needed you mean Arduino to IDF is that what you meant no this is still Arduino code, but it's using the um platform IO hybrid where you do it both. It's not using Arduino as a component in IDF.

[00:45:24]It's doing that weird thing where um it lets you recompile IDF in every time you make a config J.

[00:45:33]So, it compiles both.

[00:45:35]The downside is my compile times have gone up massively if I need to tweak anything on the IDF side because it literally has to recompile both. But um the advantage is I have now better control over the the memory usage. I can turn things off like I can turn so in Arduino you can turn Bluetooth. You can disable Bluetooth in that it doesn't turn on but you can't compile Bluetooth out. I'm not using Bluetooth for anything. So Bluetooth's literally compiled out now. There's no Bluetooth at all capability in this code. It's it's not compiled in, right? Stuff like that. I I can um change the uh the RX and the TX buffer size for my TLS stuff.

[00:46:20]Like Arduino uses a big buffer, enough buffer for the receiving so you can get really big payloads, but a matching sending. So, so I think they've get like 16k for in and 16k for out. Well, my messages going out is this big, right? With maybe some header stuff. My messages going out needs 4K, right? I don't need 16k. So, or I think I've set it to 8K now, right? So, 8K out, 16K in, that gives me 8K back instantly, right?

[00:46:54]You can't do that with Arduino. You got to do it with the other way around. So, it was it was a massive job. Um, it's stable now, but it wasn't stable for quite a few days. Um, so yeah, it's pretty it's coming along.

[00:47:10]It's exciting, but I'm just not getting getting to spend enough time on it, which is a bit frustrating. Um, but I'll get there. So, I'll have the those screen those new panels will be here this week, but I'm on site end of the week, at least Thursday, at least tomorrow, maybe Friday. I might not get a chance to build them till the weekend.

[00:47:30]So, I don't even know when I'm going to get a chance to put one together and um see if it melts or not. But there you go. Um oh, one last thing I want to show you except I need to put it back to uh what am I doing? I need to put it back to normal res. So, that and save.

[00:47:54]Reboot. I mean, this is temporary, right? This won't be a thing that you'll be able to do. It'll boot in the correct mode. Um, chat disconnects and reconnects. Any ideas? No, it would be it'd be YouTube on your not you, but YouTube on your side somewhere that Google having a problem with the server or something, I would say. Um, sorry.

[00:48:20]No idea. Um, so one thing I want to one last thing I want to show you. Um, there was the live. So, this is kind This is live off the device, right? Um, I showed you a couple weeks ago the like MQTT stuff, which I'm I still haven't progressed that yet. I will, but something else I added. Um, so Oh, John's John Clayton's not here, is he? No, that's okay. So, John Clayton was talking to me about, you know, pushing stuff onto a screen for a boat for his project, and we spoke offline, like on Discord afterwards in length. Um, kind of working out how he would connect his stuff together. And it occurred to me that um I could I could offer an easier way for that to happen that doesn't require MQTT sitting in the middle. And so, what I've started, now this is just a start, right? But what I started working on, so right now what you're seeing is that's the display, right? It's it's literally whatever's on the screen here is on this on there, right? So I've made a thing where I can do this watching two horses walk into a bar.

[00:49:42]Long face. Anyway, I didn't want to put the whole um it's interesting. I'm live. Unexpected Maker is live. That shouldn't look like that. Anyway, so what I've done is I've created an endpoint that you can hit that you can send it a JSON payload and it will put a message on the device. And and right now it's just text with color. text with color text with color. But I um because that's all I've added. But the plan is to add the ability that you'll be able to um set how it's going to scroll in and out.

[00:50:23]Like you'll be able to do maybe 50% of what the V2 um what this V2 sequence thing will be able to do, right? Maybe 50 or 60%. I don't know. It it depends on how I'm going to do parameters. But you'll be able to construct a payload that has formatting and stuff in it of data of any data of anything you want and you'll be able to push it to your rack and it'll display on the screen. So you you don't need MQTT there. You don't need whatever else. in any app or website or anything that can access, you know, that can push something to an IP address on your local network, you'll be able to push this to. I'm just using curl at the moment in my command line just to to push to it. And um I thought that would be a really cool solution for people that want to get data onto their thing, not through this interface, right? there. I The whole point was to allow people to to connect their own things to this as well to show data without having to, you know, add it as a user module or whatever. Like you can use my system to build content for the rack, but you can also push stuff to it externally through an admin portal or through a uh you could associate color with temperature.

[00:51:52]Well, you could do or you can do that yourself, Dave. Like you you can decide how you want to um show and display the data like what the what the content's going to be and what colors that content's going to be and whatever and you'll be able to just push that to the rack uh to the display, the pixel rack. Um so that's still work in progress. That was just the first pass I started working on. Uh it's just text and color right now.

[00:52:22]Um, they're supposed to be showing one at a time. That's a bit of a mistake.

[00:52:27]Via what protocol device is listening on a port. No, it's a it's an endpoint.

[00:52:35]So, it's it's like a REST API. It's just an endpoint.

[00:52:39]You you hit an IP address and with a URL and um you pass it the um yes, HTTP endpoint. you pass it a JSON blob and it uses that JSON blob to build what it is it's going to show and then it shows it.

[00:52:59]Um, and so I will add with the option of of you know having a passphrase or something that you need to provide with it. Um, yeah. Well, it's just it's rest. So yeah, it's a usual. Well, no, it's not post. It's it's just a get it's it's REST API.

[00:53:26]Um I mean you could do it from a a string if you want to.

[00:53:33]You know what I mean? You can do it from like a you know a query string, but um that's kind of pointless because that requires a like a browser maybe. I don't know. Anyway, um right now it just Yeah, it's still being worked on. I use an ad block original light in my browser.

[00:53:58]I have no idea, Mr. One spring. I don't use any ad blockers, but it could be.

[00:54:02]Okay, trying to use it.

[00:54:06]We did our own security. Well, right now, I mean, there is no security. It's I'm trying to make this fully open in that you can do whatever you want to do with your rack in terms of pushing content to it. Um, so you know, and I'm trying to give everyone as much flexibility as possible to do that. So I mean, if you think about it, the with with the discoverability stuff, which is still something I'm working on, right?

[00:54:33]Like you can't actually I don't have another one running right now. If I did, I could actually click discoverable, right? Well, I can do it now. Um, so this is now broadcasting over the uh uh casting over MD MD MDNS. Oh my god, tongue twister. So, if I had like another rack sitting here doing the same thing, they'd both see each other in this list. There'd be a list there, right? And you can connect them. they both have as long as they have the same secret. So if you had that, you could be pushing content from this rack to that rack using that same endpoint system if you wanted to. But you could also do it from Python. You could do it from whatever you want. You could long as your app can send a uh you know a REST API call to an endpoint, you can push content to it.

[00:55:28]Wow, that was a demo that went way longer than I had planned.

[00:55:33]H couldn't that they go what saying what did I what did I say what what did I do Johnny probably okay if I I I don't remember what I said that was funny but I apologize or What?

[00:56:03]There was a whole discussion last time about syncing and whatever between two pixels. There was that's why this is so I I had a good think about it afterwards. Um I'm still not sure about the syncing how that would work because syncing only works if you've got content on both. Um no, you said no. What I said was I saw your post. You wrote that you use it origin and you don't have a problem with it.

[00:56:34]So that's why I said okay Johnny uses it.

[00:56:39]I I wasn't saying you probably use it.

[00:56:43]I was saying you use it and not have no issues. So you would probably know is what I meant.

[00:56:49]Okay. Anyway, I wasn't I'm not clairvoyant, mate. I don't know what software you machine. Well, I I did hack you last night and check out all the stuff that was on there. Sure. But, you know, uh I was looking for your prawn.

[00:57:02]Anyway, um sync. Yeah. S the syncing thing. I I don't understand how that specifically would work from a content on there, but syncing from a I want to send this message to two of my racks at the same time. That endpoint system does that, right? You can you can that can go out to multiple, right? You just need the IP addresses of them. Anyway, uh do you have a widget that looks like GitHub commit history?

[00:57:35]You know, the green squares widget.

[00:57:42]Oh, you could make uh I don't um I mean you could but how to do what though?

[00:57:53]you wanted to actually show you your git commit history like that. Can you actually get that? Can the data that generates that in git, can you actually pull that data off that they use to generate that? Because if you can, then um sure I could make a a module that does that I guess. Um add a time stamp field to frame data.

[00:58:22]Then a high-res network time can do a high-res network time can do display.

[00:58:32]What can do? But what's going to do the syncing though?

[00:58:38]But no, I'm not adding timestamps to frame data or there is no frame data.

[00:58:42]No, no, it's not built that way.

[00:58:46]Um, so yeah. No. Anyway, this this I'm not saying how do I get two displays to show the same thing. The question isn't that.

[00:58:56]The question is what is it you're syncing?

[00:59:00]Like where's the content coming from that you're syncing?

[00:59:04]You know, do you have the same scene list on both devices?

[00:59:10]Exactly the same scene list showing exactly the same things. If that's the case, sure. I can keep them in sync so they both show every every scene starting at the same time all the way through. They're in I can get them in lock step, right? But only if the content's the same. But if the content's not, then what am I syncing? Right? It's not about how do I sync two devices or three devices. It's what am I syncing and and and how does that work in terms of making sense? Anyway, Kad anyone that that that demo went much longer than I thought it would. Um, yeah. Okay. K. Should we do should we build a just random idea uh display say network traffic in that form?

[01:00:02]Ah.

[01:00:07]Hey, Tech Nless.

[01:00:12]You just need a master slave.

[01:00:15]Right. You're talking about so whatever's on this device here when you're displaying it, push that live to this device. I can do that. Sure. Sure.

[01:00:26]But I can't do all the things like you know what I'm You know, I can sync it this way. I can sync it this way. I can sync it this way. I can go this way. I'm not going to be adding all the different ways that it can be synced to cater for every single person that might want to sync it a particular way because that's just at the end of the day if I give you the ability to push content to any number of your racks, you sync it, you do it. Like it's not something that I need to support in the actual rack interface, I don't think. Um because I'll never be able to cover every single scenario that every single person wants.

[01:00:59]Um, so you know, could I put a master slave setup thing going? Sure. But, you know, if it ends up that I do all this work to do that and 90 99% of the people who buy one only ever buy one, then I've kind of, you know, put a lot of effort into something that no one's using. So, um, catch you later, Dave. Thanks for joining. I hope you're well. Have a good night.

[01:01:27]Okay. Are we am I going to do a footprint for the S31?

[01:01:32]It's I've not even looked at the current data sheet. The first the last data sheet I looked at was a 0.1 data sheet.

[01:01:39]I think I downloaded a 0.2 or I was sent a 0.2 but never really went through it and ended up to 0.5.

[01:01:49]Um, so pre-release 0.5. This is their data sheet.

[01:01:56]So it's 8 by 8 mm. The current chip is 5 by five. And you think, okay, there's only, you know, an extra 3 mm, but there's an extra 3 mm and all of the uh passes that have to go around it that you need to get access to. Um, so yeah, we're not going to spend the day looking through the data sheet. What I want to do is look at Oh jeez.

[01:02:28]Um, why did that not zoom?

[01:02:32]Okay, so you can all read it as well.

[01:02:36]Um, >> see chat has disconnected.

[01:02:42]>> Okay. And back. Cool.

[01:02:46]Look at geese. At geese.

[01:02:50]So, I don't like uh I get why they do it. I really hate how this modern chip design stuff sprinkles power rails all the way throughout.

[01:03:06]What's that one? Here.

[01:03:09]Here.

[01:03:14]That one. That one.

[01:03:16]That one. The problem is it makes it really hard to get access to I/IO, especially if you're using chunks of IO to get it out. It's not about the layers. What is what? Sorry, but why is everyone's solution to any problem? I say layers. This is not a layers problem. It's a getting to the pins from the side problem when you need to put caps and inductors and whatever else on power rails as close to the chip as possible which blocks access to the pins for your traces or you know stuff like that, right?

[01:03:59]Um anyway, it was just a a comment about this this new thing of getting them sprinkled everywhere. It makes it really difficult. Um, you know, why can't they put these two power rails at the end? Like, why can't they have one here and one there? Okay, then I can put the power content in the corner here and the corner here and not block access to any of these IO.

[01:04:27]That's what I'm asking. Like, this might be more convenient for them or whatever else, but there's a Yeah. Anyway, um well, FBJs are quite different because you don't access them from the side. You access them from underneath, right? I mean, sure, some people do that, you know, tricky using balls that aren't being used to link things up and whatever, but it's a different thing.

[01:04:51]It's a different a different issue for of accessibility on an FPGA. Um will be it could be power stability. It could be power domains. Maybe some of these are on a different domain or it could just be the most convenient spot to tap into the power inside the chip, right? I mean, the all these different um power pins have got a different purpose.

[01:05:14]Um some are connected to digital, some are connected to the analog side, some are connected to peripherals and some are connected to just main, you know what I mean? So, I I get it, but it's still it's frustrating.

[01:05:29]Um, well, I don't know if this is going to be frustrating yet cuz I haven't tried laying it out, but I know that on the um the C61 there was a it was a little frustrating.

[01:05:41]Um, so a and the C5, which I have abandoned doing anything with because it just C61 C5 don't have enough IO. The So I can't I'm not going to be making Oh, I should actually talk about this very quickly.

[01:06:00]I'm not going to be making any more C C6 tiny C rubbish guy. Everyone hear that? Dang.

[01:06:09]I'm not going to be making any tiny C6s anymore. I'm going to be moving them over to a tiny C61.

[01:06:16]Um, so they'll be exactly the same boards, fully drop in compatible in every way, but they'll have the C61 on them instead. So they'll have uh 2 meg of PS RAM and the newer chip. Um, yeah. So that's the thing. I don't want to go 0201 just to be able to get things close enough to those pins.

[01:06:39]where I need to to like it's it's yeah it's fine for you grind because you're not making your boards but you know design for manufacturing is important for me because you know I not only do I need my design to be good and do everything I want it to do but it's got to be as painless as possible for me to manufacture with um you know with the highest you know pass rate as possible.

[01:07:05]Next generation will have wireless pins.

[01:07:08]I'll buy one of those to shipoo for sure. So anyway, I'm going to Yeah, but I'm not doing not doing any C5 stuff. I I Yeah, but S31 for sure I want to do.

[01:07:26]Um, so I guess I need to make a footprint, don't I? Um, we need to make a schematic symbol first.

[01:07:38]Got a coil on the package so you can power the chip by induction. Sure. And that could also oscillate through all of the power rails and RF and uh you know could signal aliens in Mars. I don't want to do a symbol first. Do I want to do I don't want to do really you know what I want to see if I can do I haven't tried any of this yet.

[01:08:00]Um, I want to see if I Okay, so Oh my god, it's been so long since I've even used the symbol editor.

[01:08:09]I'm going to go a new symbol. I need to pick a library. Do I add a new? No.

[01:08:18]Where my libraries?

[01:08:20]I I want to pick my current.

[01:08:27]Okay.

[01:08:29]What? Anyone know how to use this?

[01:08:33]Uh, panels library tree. That's what I want. Thank you. Um, where am I going to add it to?

[01:08:49]You know what? Stuff it. New library.

[01:08:54]and clad libraries.

[01:09:01]I'm going to call this I don't know um speed 32 S31.

[01:09:13]Why not?

[01:09:17]Okay. Then then we can go new symbol symbol name. Why is it underscore one?

[01:09:32]Um or you just create symbol with I've not made a symbol from scratch in such a long time. Why don't you know what?

[01:09:40]That's just really stupid of me. Why don't I go and grab a symbol for an ESP32 that's already done to start with?

[01:09:51]My brain is so the S3 is probably the closest we should go from.

[01:10:00]I really don't like that.

[01:10:04]Um, where are my ones?

[01:10:18]Um, I'm so used to not working this way.

[01:10:30]Um, okay. I don't even know where to start from.

[01:10:55]I can't I don't know where mine are.

[01:11:00]I'm just going to start with this.

[01:11:02]Wow. Um, they could include Oh, right.

[01:11:07]Sorry, that was a that was a new message. Okay. So, I'm going to going to copy this, I think.

[01:11:16]And go new symbol.

[01:11:21]Been up all night and press US.

[01:11:24]Good night, AJ. Have a good sleep.

[01:11:27]32 S31.

[01:11:30]Okay. And move that cuz you move the U.

[01:11:41]This might just be a bad idea doing it like this.

[01:11:45]I don't know. Um.

[01:11:51]Oh my god.

[01:11:56]How many pins on each side is it?

[01:12:00]Excuse me. 20 pins on each side.

[01:12:05]I tell you what I'm going to do. I'm not going to do my the symbol first. This is a pain in the backside. I'm going to, you know, footprint editor.

[01:12:21]Okay. New library.

[01:12:27]No, I don't want to do that yet.

[01:12:30]Um, available footprint generators.

[01:12:32]What?

[01:12:41]Let's do a QFN wizard. [ __ ] Okay. Where's the info though?

[01:12:51]What pitch is it?

[01:12:53]Probably 4 mm.

[01:12:56]I mean the package size at the end won't it hopefully okay it is lead pitch 35 [ __ ] eat.

[01:13:26]Really?

[01:13:34]Why? Why do they do that?

[01:13:41]Okay. Um, body size. So, it's 8 by8 and the EP size is 6.5x 6.5. So body is what?

[01:13:59]Okay. Package width. So it's 8 by 8.

[01:14:09]That was 6.55.

[01:14:14]6.5. I've never used this at at Wizard before. I I don't even know what it's what we're doing with it yet, but um pads are okay.

[01:14:42]So, E. Yeah. Five and Where's it showing the pad width and height?

[01:14:54]72 XL. What does that mean?

[01:15:05]Tip to expose.

[01:15:09]Does anyone know how to read this?

[01:15:16]Or am I just lead width or lead length and width L1?

[01:15:23]Oh god.

[01:15:25]4 by.17.17.

[01:15:36]4 Where's it showing? How many? Where's the uh Oh, here. 20. That's what NX is, is it?

[01:15:51]2020. Right. That's starting to look a little bit more like the right thing.

[01:15:59]Um, I really don't like the curved on those.

[01:16:04]Where's the curve? Oval. Not oval.

[01:16:10]Well, I want oval, but not I don't want square, but I not that. Why does it keep resizing it?

[01:16:18]Bloody hell.

[01:16:20]Is that the text?

[01:16:25]Um, offset length for pitch 35. Okay. E- pad. I'll have to manually fix that. And um, yeah.

[01:16:36]Package What does everyone think?

[01:16:43]And then how do I save that file?

[01:16:59]How do I get out of the wizard? Export footprint to editor.

[01:17:05]Okay.

[01:17:07]Okay. Wow. Has anyone done that before?

[01:17:11]I've not done that before in Kyad.

[01:17:15]I generally just build them from scratch. Um, interesting. Okay. So, they've got their EP has got that the corner here, pin one. Okay.

[01:17:30]Um, we can do that.

[01:17:37]Maybe um Can I create change the shape? No. Can I shape modification move corner to Can'ter the corner?

[01:17:55]Really?

[01:17:58]Okay.

[01:18:03]I'll fix that corner later then.

[01:18:11]I guess this is it then.

[01:18:19]Okay.

[01:18:21]So file save as unexpected maker ESP32 S31.

[01:18:38]Okay. Few things I don't like. Um okay let's go fab.

[01:18:47]Okay.

[01:18:54]I don't like how thick they are.

[01:19:01]And I don't like these.

[01:19:06]You can go.

[01:19:09]You only care about the corner point. Catch you later, Shipu.

[01:19:27]Okay.

[01:19:31]So, from a size point of view, um, should we see how bad it looks on a, uh, tiny tiny board? Let's have a look on a tiny board and see exactly how screwed we would be.

[01:19:53]So if I added the SP32 S31 H. You can see the difference in chip size maybe.

[01:20:10]Wait, what just happened?

[01:20:21]Um, yeah, you can see the difference in chip size is pretty massive.

[01:20:30]And I still have to, you know, I'm like, I don't know. Maybe I I could squeeze one on.

[01:20:40]Could we turn freaking locking off?

[01:20:43]Oops.

[01:20:48]Um, I mean, I could maybe squeeze them between pins.

[01:20:57]I don't know. Can you imagine if I did?

[01:21:00]If I could squeeze one on, that would be epic. But I don't know.

[01:21:08]So, I need So, it's got built-in PS RAM, but I need external flash. I it's almost pointless sticking it on a tiny board cuz you don't get all of the IO that you know maybe on a pro.

[01:21:26]Okay, let's let's let's let's uh open project S3 make more sense on a pro board because of all the IO discard.

[01:21:56]Let's try that again. Add I mean that's a this this board is insane in terms of routing already. Um and with an even finer pitch I mean look at the pitch difference.

[01:22:32]Okay. So the question is what does it need?

[01:22:36]What's the minimum 16 layer?

[01:22:42]Why stop at 16?

[01:22:45]Okay. What does it need? Let's um I need to see a bare minimum of layout if they've got one, which they don't.

[01:23:00]No way.

[01:23:08]Do they not have a basic schematic and they got lots of tables.

[01:23:23]Oh, son of a [ __ ] Um, I mean, they must have something online.

[01:23:34]Oh, we should look at this as well.

[01:23:39]Open in Safari.

[01:23:44]All right. Is there anything yet?

[01:23:49]Bugs and affected chip revisions.

[01:23:55]Fector revisions version zero.

[01:23:58]Flash boot fails at 1.8 volts. Don't care. Signature with invalid R and incorrectly accepted. Okay. Near the Ver worry for me. Okay, that's good. Um, but yeah, I don't want this, you know. Okay.

[01:24:21]Didn't really want the module schematic.

[01:24:22]I just wanted a uh you know how they just do it like a basic uh like um Do I have one open?

[01:24:38]uh window right so like on the S3 data sheet this is the just the S3 they include you know I'll get there eventually data sheets become so long it's not funny they used to include just a layout they not do that anymore Am I imagining things?

[01:25:13]I must be imagining things. They don't do it anymore.

[01:25:19]This is show just a chip with, you know, a couple of caps and crystal and inductor.

[01:25:24]>> Says no, but sure they used to also.

[01:25:27]>> Wow. Okay. They're just they're moving everything online, are they?

[01:25:32]Um BCB layout design actually. Yeah. So this is what you're talking about with the um the module layout, right? Like this. They used to show a picture like this always in every data sheet.

[01:25:53]That's really frustrating. They don't right. So this is their that's what they say is their bare minimum. Right.

[01:25:59]There's no way I'm fitting all of this uh in the uh Well, maybe. Let's Let's see what Okay, let's break it down. Right.

[01:26:13]Don't need that. Don't need that.

[01:26:16]External crystal. Fine.

[01:26:20]One cap on power pin there. One on here.

[01:26:24]So, two decoupling caps there. Um, why are they pulling those up?

[01:26:33]Not biggest. Don't know. So, cap there, cap there, cap there. Okay. So, so far it's pretty much just the same. Um, so what are these used for?

[01:26:54]just need caps for those, right? So, that's just for the PS RAM and stuff.

[01:27:00]Just external capacitors. We're not actually providing power to those. That happens internally, I'm pretty sure.

[01:27:04]Cap. Cap.

[01:27:07]Yeah, this is all pretty much the same as the S3.

[01:27:11]Um, where's the reset line?

[01:27:15]What do they have on that? If I can find it. VF chip PU. Chip PU has got where's chip chipu uh activate chip lo pin not be floating great a um crystal inductor an inductor is mandatory at exter line is mandatory Sorry.

[01:27:51]Holy moly.

[01:28:01]And and okay, so you know they've got oscillation problems on the chip when they force you to stick an inductor on your crystal line.

[01:28:13]Wonderful.

[01:28:15]Yeah, they'll have oscillation. So, they've obviously got something from the the chip creeping through the crystal.

[01:28:23]Um, I wonder if we're going to have mega harmonics issues like we do on the S3.

[01:28:30]The S3's got such bad RF harmonics. So bad. Um, I can't see anything connected to chip PU like their standard reset >> says. Oh, speaking of remember those boards that were failing went for third run today passed fully. Very happy.

[01:28:45]>> Oh, well done. Well done.

[01:28:49]Passed magically or they passed because you you you made the change to part to make them pass. It sounds like they just happened to pass the third time all by themselves. How funny. Oh, hey Matt.

[01:29:03]Morning.

[01:29:09]Yeah. Okay. Can anyone see anything that's connected to the chip PU? You know that you pull down whatever blah blah blah. Uh oh. So, what's what's Bruce pointing me to?

[01:29:24]Documentation content. Isn't that what I'm currently looking at?

[01:29:31]Oh, that's right. That's just their wiki version of it all. Okay.

[01:29:38]Yeah. I I mean, I need to read the data sheet obviously. Like I I need to go through here. Road ship enable need some time stabilizing before.

[01:29:51]Okay, so just standard RC that we use normally flash and RAM. Yep.

[01:30:02]External crystal clock source only supports 40 MHz crystal.

[01:30:09]This crystal shown here. Note the accuracy of selected crystal should be within it's - 10 ppm. Yep, no problem.

[01:30:16]Series inductor must be added. Exteral clock trace initially suggested to use inductor.

[01:30:23]Those inductors required for normal system startup even if RF function is not used. Yeah, you see that they have got issolation problems in their chip.

[01:30:31]Inductor is still required.

[01:30:45]Wow, look at that. So, select text mode using that tool. Observe the 2.4 GHz signal radio communication analyzer.

[01:30:53]Attain the actual frequency offset and adjust the frequency offset within 10 ppm. So you you need to adjust the crystal to uh to not sync and oscillate with the RF. Wow.

[01:31:08]They're just packing too much in their chips these days. Like I I just feel like there's a um there's a reference schematic. Okay. Thanks, Bruce. I Anyway Anyway, going to shut my mouth.

[01:31:25]Um RF circuit mainly composed of three parts. RF trace on the PC board. Yeah.

[01:31:31]Yeah. Yeah. Yeah. 50 ohm impedance chip matching circuit must be placed. Yep.

[01:31:36]Yep.

[01:31:38]Wow. CLCC really shown antenna and antennas.

[01:32:13]Wow. Okay. Um, what is this? Let's see what this is.

[01:32:22]Oh, that was just the same page. But how do you tune it? Yeah. Okay, I know how to do that. Thanks. Um tension random place a pull up system pin do not add high value capacitors or the chip may enter download mode pins two channels Ethernet MAC City camera controller. Whoa.

[01:33:05]Wow. Chunky.

[01:33:11]Got 888 interface.

[01:33:16]Okay.

[01:33:17]Camera.

[01:33:19]Yep. Next.

[01:33:24]That's the PCB layout.

[01:33:29]And I love it. I love it.

[01:33:33]Little bit of meandering happening here.

[01:33:45]That is hilarious.

[01:33:49]Sorry.

[01:33:53]I mean, look at that. Clearly, no one cares much about like clearly the manufacturing process of PCBs is so good these days that that there you see what I'm talking about. This there's a hole there. Hang on. Um, open image in a new tab.

[01:34:16]Look at that. The this I mean that would normally be I know they're the same, but that would normally be such a red flag.

[01:34:26]like etching deposition stuff like you that would just be a manufacturing fault always but there's nothing it's not touching anything it's the same trace but like I don't know I look at that and go see this is the problem you see all these packages is that even on the screen.

[01:34:55]No, I'm sorry. Hang on.

[01:35:00]So, this is why I'm going to have a problem putting on one of my tiny pro boards is all of these.

[01:35:08]It's the extra space required to get all of these in. And I'm not going to go 2011 everywhere to do all that.

[01:35:16]Um but anyway and the RF that's just a zoom in. I want to see the Oh, excuse me. So this is the RF area. Open in new window. Not new window. Open in new tab.

[01:36:02]So they're not using any stubs. Oh, they are. They are. Why?

[01:36:15]So, you'd figure if they had this this many matching components to a PCB antenna, you'd think they'd be able to match it well enough that they wouldn't need to stub any of the components.

[01:36:31]You all know what's what this stubbing is. There's I'm happy to explain if you want me to explain.

[01:36:42]Um here.

[01:36:49]Okay, that's just buy.

[01:36:53]So, is it a stub or have they or is that just cuz they want to get to the ground?

[01:37:02]Cuz No, that's got to be That's a stub.

[01:37:07]So, you might like I'm sure many people have heard or or whatever like um when when you're tuning an antenna, I'm not talking about RF to an MCU, but when you're tuning the wavelength of an antenna, you can um so like so this antenna here, right? So you don't have to redesign this antenna to match 200 uh 2.4 4 GHz you the this is designed for a particular wavelength right a tenth of the wavelength which you're trying to match and the way you can tune it forget about I'm not talking about tuning it to the chip but just the antenna itself is this leg here right you can make that longer or shorter and as you make it longer or shorter you're kind of you're tuning it like this to try to fit the wavelength stuff Have they mentioned the stub in there?

[01:38:06]Okay, I'll have to go back and have a look. Right. So, that's how you like, you know, you tune like a longer the wire, the different the frequency, right? So, you you're tuning the length that way. So, what these stubs do is the stub is is is going from a component in in parallel in this case to ground. and the longer the path to ground. It's a similar thing where you kind of you're tuning the component value with the stub.

[01:38:35]Um so sometimes you just can't get a a a cap or an inductor at the exact value you want or you know from a tolerance point of view or something. Um, and so you can stub the ground with a a thin a little thin trace. And the longer you you pull it, the you can tune it and it it kind of helps you not quite suppress it. It it lets you reduce harmonics and and do all sorts of things, right? It's um I'm I'm surprised like on the S3 they they did it that they've got a harmonic problem. uh add a stub on the ground pad of the ground capacitor near the chip side in the matching circuits press the second harmonic right it is approved feral to keep the stub length 15 mil right so it's literally to yeah it's um so they got harmonic problems still when I say problems I mean all all chips have harmonics right second third fourth harmonics of whatever the frequency But they they normally if you're at 2.4 GHz and your next harmonic half or full harmonic, it's supposed to oscillate down, but the S3 oscillates up and gets bigger, which is a fail, right? Because, you know, not only does your frequency have to kind of match and not be over a certain threshold, but your harmonics have to be not be over a certain threshold. And a straightup S3 chip fails like because of it. Um, well, it's still trial and many error.

[01:40:18]It it's it's a band. It's a I don't know if it's really it's not a band-aid. Look, it's it's a way to tune, right? But it's not something that you can tune live like it's a PCB design change. Like swapping a component, you can tune live, right?

[01:40:36]Stick it on your your um your VNA spectrum analyzer. Look at the frequencies. You can you can watch like I can on my spectrum analyzer I can look at the frequencies right I can look at the harmonics and so changing a component so my boards have all got um five components for matching right they do re for the antenna matching yeah but you can't do well stubbing on 0201 is a problem and and o201 for antenna matching no the S3 recommends not O201, although they do O21 on their modules for space. Um anyway, uh please components and place them close to the pin as zigzag. Two capacitors should not be entered in the same direction to minimize interference.

[01:41:25]Sure. So my boards, let me just go.

[01:41:33]Okay, these are a bit different now because I've got the RF switch in the middle, right? switches between the two outputs. But if you think about it, I've got here I've got a matching network here and that's to tune the antenna to 50 ohms and then I've got matching here which takes the 50 ohms and brings it back down to 30 J5 or whatever it is right at the pin. Yeah, I know you are Bruce. Yeah, I get it. So these do two things, right? So we've got an inductor inline inductor and a um so series inductor and a shunt capacitor. They do two things. They shift the fre the the uh impedance from 50 to as close to 30 J5 or whatever it is as possible. Can't remember what it is. Um but they also it also works as a filter to suppress the harmonics. Now, sure, technically I could have done a um you know, I could have not had that pad do this, right?

[01:42:49]I I could have called this um something else. And then I I could have put a little trace on here back to ground pin and I could have tuned it that way to suppress harmonics. But that meant every time I wanted to change it, I have to get another PCB made, right?

[01:43:08]Um, on the S3, they gave you no tuning details from it. But it's interesting that the stub, they've said 15 mil.

[01:43:18]Um, and they asked to get the stub, right? Different piece of material, right?

[01:43:25]So, I I I just never bothered. And I did my filtering in here which and I managed to suppress the the second and harmonics onwards by like 45%.

[01:43:36]But the cost of doing that is I also lost a little bit of my gain at the 2.4 GHz point, right? Because you you know you can trimming it to only suppress at these frequencies but it's impossible to not affect everything, right? like that comes down a little bit for this to come down a lot.

[01:44:00]Um, but it turns out it I didn't drop it enough.

[01:44:06]So, I dropped it enough before I went for the series D, but at full power transmission, like 20 dB on the chip, which is way too strong, um, there's enough oscillation still coming through harmonics wise, that sometimes the RF chip will just say, "No, I'm not passing this through," and it won't pass the signal through. Um, obviously that that got discovered after I shipped a lot of uh series D boards. Um, and the solution is don't run it at 20 like drop it down to 18 or or 15 dB dBm dBm, sorry, not db dbm and um go straight through the chip here and you see almost you see negligible performance differences. So, I was hoping that the S31 I was hoping they would have learned from the S3 and the S31 they would have uh tuned their LNA a bit better, but it looks like they haven't.

[01:45:11]Um, you have to use their exact stack up and layout. The values will work, sure, but then you're you're just making their module and you have to use their PCB antenna.

[01:45:23]You know what I mean? Like it's it's all tuned together. they're not independent.

[01:45:27]Um, so yeah, I I I always laugh when they they publish their component values and people stick them on their own board and their own design and they don't work and they they don't understand why. So that's a bit disappointing. Well, I don't know. We'll see. We'll see. I'll have to measure one and find out. Um, so bad time to be designing a board for flash that needs flash cuz you know, not great in the availability realm and price, but it looks like, look, I might be able to squeeze what I need to do. I need to just get a basic test board up, but I don't even have silicon yet. Um, yeah, but I might be able to squeeze one on.

[01:46:35]Where's the pin? Where's um antenna is actually pin one.

[01:46:45]Interesting. Okay.

[01:46:59]So, I could probably I'll try I'll try to get it on a pro like a Progress 31 if I can.

[01:47:21]Um, and where's the issue? The is going to be where's the antenna and where's the crystal? So, ground power.

[01:47:38]So, it's this other corner.

[01:47:43]So, it's kind of the same. Okay. which is not good.

[01:47:49]Maybe it means I won't be able to get a drop in replacement like you know this battery stuff can't be here because the crystal and all its friends has to be over here.

[01:48:13]It really depends. If I do a if I don't do a series D board, if I just put like a a straight up onboard antenna on it, then I could probably keep maybe the footprint the same.

[01:48:32]All the chips going to have to get moved right over. And I'll have room to move it because I don't have flash and PS RAM. So that's flash and that's PS RAM. We don't need the PS RAM. So the flash can be moved over buttons have to be moved. I could probably do it. I think that'll be the first attempt would be to um get it on here, but I don't think I'd get it. I doubt I'd be able to get it on the tiny board.

[01:49:03]Not enough horizontal space. I'd have to make it longer, which will drop compatibility.

[01:49:09]Um, okay. So, what I what I need to do is do the freaking schematic.

[01:49:17]Uh, yeah. But you know what?

[01:49:22]I don't think I got the head space to it this morning to be honest with you. Um, I want that layout schematic. I hate how they like they'll do a picture like this in their design guide stuff and then the schematic they give you looks of the chip is like a totally different layout.

[01:49:42]I hate that. Can this can I um see I want I want this in a PDF so I can grab the if I do it I want to be able to grab the text.

[01:49:58]If I grab this text, I I think I can fan out a um new symbol editor. Let's go back to Did I save one? I did. All right. So, I don't want that. I want um a square one that looks like this.

[01:50:30]And I've got that.

[01:50:33]Maybe I just rip off my S30. It's this one.

[01:50:43]Edit.

[01:50:44]Edit symbol.

[01:50:46]This one. This is my S3 one.

[01:50:52]Okay.

[01:50:54]and close your symbol editor.

[01:51:00]Try this again.

[01:51:03]Open.

[01:51:07]Wait.

[01:51:13]View library 31.

[01:51:25]Start with my one. Much better.

[01:51:30]Mac has OCR in the image. Yeah, that's not image preview. That was that's that's their website.

[01:51:37]So, yeah, I'd have to open it up into image preview and I can I should be able to grab it from that. Um, but I need to make sure I've got 20 pins first.

[01:51:47]Um, well, actually, maybe I don't. So, hang on. Save that. So if I open this uh if I put in downloads and then I open it in preview.

[01:52:05]Can I grab I can. Oh well, let me grab them all though. Yeah, cuz they're a funky shape. See that?

[01:52:16]It's I don't know if it's missing them or not.

[01:52:24]I don't think it's grabbing them all.

[01:52:27]Copy.

[01:52:30]Um, and Kadh, I should be able to go. So, what is it?

[01:52:39]Isn't there a fan?

[01:52:42]Is it um where's that mode?

[01:52:56]There's a where you can just add a an array of let's post.

[01:53:08]Yeah, like that. Um, does anyone know what I'm talking about? There's a what tool is it where you can actually tell it like it'll build it from an array. Like you copy and paste a bunch of names and it'll been build pins for you. But I don't know.

[01:53:42]What tool it is?

[01:53:45]Preferences. Inspect. View. Place. Place place. Draw pins.

[01:53:51]No.

[01:53:52]Uh, edit panels.

[01:54:06]Does anyone know?

[01:54:09]I'm sure there was a a way that you could say.

[01:54:26]Obviously not.

[01:54:29]Okay. Um, I mean I can type them all in, but no one no one know what I'm talking about.

[01:54:45]Um, okay. No, let's uh kad schematic. Not schematic. Uh symbol editor was it called fan out pins or something creating fan out to create sequential pins simulator use the insert key shortcut to auto uh or use that there was a there's a way to fan out pins Um at uh paste uh uh pins, right? Use the pin table, the insert key or spreadsheet copy paste.

[01:55:45]All right.

[01:55:48]Bulk placing symbol pins.

[01:55:51]Use the in I keep look at the people insert keys. Which one's the insert key in your keyboard?

[01:56:09]You are you for real?

[01:56:13]That is crazy.

[01:56:14]Um, okay. Okay. So, looks like I have to just use the insert key.

[01:56:25]It's going to be easy for me just to use the pin table, I think.

[01:56:29]Um, okay. So, this is going to be fun. I don't know how I'm going to show this at the same time for you, but um hopefully I can.

[01:56:51]Um, really?

[01:57:00]Come on.

[01:57:02]Okay, maybe you can see what I'm doing here. I don't know. So, it's going to be V D A3 D4 GPU.

[01:57:32]Is it those? They're also what? Uh zero and one. Oh, interesting. Okay, so they are zero and one. They put them on here.

[01:57:40]I go zero one. Okay. 2 3 4 five.

[01:57:56]I'm on the right chip. Yeah. Okay. Good.

[01:57:59]Five. Uh so pin 11 is B DD PST talking about having PST PSD.

[01:58:12]Um is it coffee time yet? I six up to 11. Really? Six.

[01:58:23]Why can't I hit enter on it to edit it?

[01:58:30]8 9 10 11. So pin 18 is Vref touch >> super.

[01:58:50]>> Thanks Matt.

[01:58:54]>> Thank you Matt. Can I now pay you to actually go and get it for me?

[01:58:59]Is that Is that Is that too much to ask?

[01:59:04]Um, I need to So, I I'm going to copy that text.

[01:59:12]Copy. Oops. So, 14, 15, 16 up up to 19.

[01:59:27]18 19 >> M says of course if the weather holds I should be able to get you something resembling a coffee by about 2 p.m.

[01:59:38]Rolling on the floor laughing.

[01:59:41]>> Thanks. You mean 2 p.m. tomorrow, right?

[01:59:44]But thank you.

[01:59:47]That's wild. Awesome. Okay. 27 is SDIO data zero.

[02:00:00]SDIO data one SDIO data 2.

[02:00:06]Whoops.

[02:00:09]This is so much fun.

[02:00:14]VDD PS RAM 1 P 8 1.

[02:00:25]It's a fun name. SD IO data 3 SDIO block.

[02:00:41]S IO command.

[02:00:46]Oh, that one. But number two, I know I need to fix these electrical type things. I will do that after once I get past this pain.

[02:01:05]BDD.

[02:01:10]Yeah, I wish someone else was doing this for me or I had already done it.

[02:01:16]I used to do this all the time with the early espresso chips when there was nothing around and they weren't providing stuff. Um 1 P.

[02:01:31]Um SP I CS uh SP IQ.

[02:01:41]No S P IQ.

[02:01:46]SPI WD SPI SPI HD SPI clock SPI data B C A B D PSD 2.

[02:02:19]Uh, okay.

[02:02:22]USB positive. USB negative.

[02:02:34]IO 33.

[02:02:40]IO 34.

[02:02:44]85. I'm going to run out of pins soon.

[02:02:49]I 36. I go 37.

[02:02:52]All the way to 40. I go 38.

[02:02:57]39. I 40. Is that all? 34 5 6 7 8 9 40.

[02:03:02]Okay. Pin 54 is DD PST3.

[02:03:12]I 042 I43 56 57 58 59 60 61 62 63 64 65 66 14 more. 1 2 3 4 5 6 7 8 9 10 13 14 6 9 70 72 73 74 75 76 77 78 79 80 Wow. Okay.

[02:04:03]IO43, IO44, IIO45, IO46, IO478, IIO 49. Now VREF ADC VRF DC VD PST 4 IO50 so much IO one 2 53 50 No.

[02:04:50]Mt cdi MTMS.

[02:05:04]Okay. U0 DXD U0 R XD GPI 60 to I don't 61 77 is X and external E B D A2 and ground.

[02:05:44]Now the problem's going to be what happens to the actual Yeah, that looks right. That's legit. Ship it.

[02:05:55]I'd love to see people do that. Work that out.

[02:05:59]Um, why aren't they showing pin numbers?

[02:06:14]H.

[02:06:16]Um, okay.

[02:06:24]It goes up to 13.

[02:06:30]No, it's not a power input. I know.

[02:06:38]We're going to make this bigger.

[02:06:41]That is hilarious. This here is gold.

[02:06:49]We should try to route that for sure.

[02:06:55]this these ones for sure. Okay. Um, how big are we making this?

[02:07:14]I don't know how big are we going to need it is if I put space between them all.

[02:07:31]I will center it maybe.

[02:07:42]Um, I it's going to be big enough although it's always going to I always don't live leave leave enough room and things crash into each other. Um, okay. Let's just space these a bit so I've got more room.

[02:08:18]They're going to be his IO.

[02:08:21]Okay.

[02:08:25]So, 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20.

[02:08:46]Whoops.

[02:08:53]Yes, I know. Terrible, terrible, terrible.

[02:09:00]Um, how am I going to separate these?

[02:09:04]This is going to be funny.

[02:09:08]Okay. 80 79 78 76 54 3 2 No, didn't grab it. Two. One.

[02:09:25]No. Which ones are those?

[02:09:30]That's not right.

[02:09:33]Okay, that one. 71 69 68 7 6 5 4 3 2 1.

[02:09:52]Is that 20?

[02:09:56]I think that's 20.

[02:10:00]No. Where's the 80? Oh [ __ ] I missed it. Oh, I only need to change their direction. Okay. 80s this way.

[02:10:10]Okay. And then how do I get these in order? I can't see that. Why? These are showing pin numbers and these aren't.

[02:10:16]Where's the show pin?

[02:10:28]Why is it not properties visible?

[02:10:36]Where's the show pin number?

[02:10:39]Ah, right. Number text size. That would be why 50. Ah, okay. So, these have all got no size on their pins for the Okay.

[02:11:05]Number text size 50.

[02:11:10]There we go.

[02:11:13]Now we can see 41 42 46 What do you do? Uh, I usually do the pin layout first, then just build the box around it. Sure.

[02:11:40]Sure. There's lots of things I could do better.

[02:11:43]Debree. Okay. They're all on top of each other.

[02:11:51]Oh, interesting. Why aren't these loading up?

[02:11:56]What they do now?

[02:12:13]This is just the the fun side of doing uh PCB stuff that they just never tell you about. It's just awesome.

[02:12:43]Wow. Okay, that's a big thing.

[02:12:47]Needs to weaken. Oh, these aren't even the right spot. What have I done?

[02:12:55]Okay, that's 80 pins.

[02:13:09]Right.

[02:13:16]Oh, very cold. Very not nice. That's not nice. Cold ice coffee. Um, yeah. I was always a glutton for punishment. Went straight into this every time.

[02:13:31]Catch later, Mr. Onespring. Have a good sleep. And to those I lost during all of that cuz it was so riveting. I apologize.

[02:13:39]Yeah, few people dropped off. I don't blame them. Don't blame you. Um, okay.

[02:13:47]So, I kind of want to group groupy things together. Uh, if I can.

[02:14:01]I always want to keep my power separate if I can.

[02:14:10]Um, doesn't really matter. Just a way to distinguish things easily.

[02:14:30]Um And I'm going to be out of space on this side. It looks like if I'm going to space them out.

[02:15:11]No. Maybe. No. Yep. I need one more space.

[02:15:24]Okay, they're spaced out. Oh, these are not almost.

[02:15:53]So yeah, this has more I got more room.

[02:16:02]That's like keeping stuff in groups.

[02:16:04]Although that power that power belongs to that anyway. So I'll leave that.

[02:16:09]It's just so it's easy to find things amongst the mess of stuff.

[02:16:16]Okay.

[02:16:24]It's not great.

[02:16:33]I It's usable, but it's not great.

[02:16:45]But this is going to be I'm going to start getting a little bit pedantic now, I reckon.

[02:16:54]But you know, you only live once. Might as well waste it doing mundane things like this.

[02:16:59]I did say I wasn't going to do this now, didn't I? I I said I was didn't think I did the it. But looks like I am doing it now. Um it looks like I did it now.

[02:17:10]Crazy is that it didn't take that long.

[02:17:16]Not that long. Um V32S31 inte says poof suddenly an hour is gone.

[02:17:30]>> Yeah.

[02:17:32]Yeah. Always goes when you have so much fun. Okay. So, what I need to do is take my ProS3, duplicate it, and rip the sheen the sheeels. Rip the sheeels out of it, and start trying to lay this in. Oh my god. See, I I need to work out what I'm going to use for what. And I feel like So, here's the thing.

[02:18:04]I love my boards. I love I love my Ranger boards. I do, but they're an absolute pain in the backside. So, because I'm I'm now constrained and stuck. So, like on the Pro S3, I can only use that many IO. I can only break out that many IO and all the power and grounds and LDO2 and whatever else all has to be in the same spot to make it compatible, right? The same with my tiny boards. Like, they got to stay compatible if I'm making a tiny board.

[02:18:33]So, I feel like the S3 is an opportunity to make a bigger board with more IO and a new design, right? Something new that I could then, I guess, effectively put a P4 on. I I never finished a P4 board because I don't have any boards where I can fit a P4 on it. It would have had to have been a new board.

[02:18:56]And because there's no radio on the P4, I know you can do a co-processor, but that's not the same. Not. It's not the same because I don't want to add a cop.

[02:19:04]Like if I'm going to stick a squeeze a P4 onto a board and not make a new design, then it's got to it won't have a co-processor. It won't have radio because there's no room to squeeze a P4 and that as well.

[02:19:18]But I didn't want to make a new board cuz everyone's made P4 boards with a co-processor. Big deal, right? Um, this is different because this has got the radio and everything already on it.

[02:19:32]But if I designed a new board for this, I could also then make it a board that can fit fit a P4. So yeah, I I don't know what I just don't think the world needs another microcontroller board to be honest. Um, I mean, another tiny board or another pro whatever board maybe cuz it's not like a new ecosystem. It's not like a new whole whatever. Like it's just another in the range which I you know I'd consider like I'm I'm pretty sure I can get on a feather like but here's the problem. Put on a feather and what does it give you? The feather's got way less IO than the Pro S3.

[02:20:23]So, you're not going to get any benefit of extra IO. You're just going to get the S3 S31 on a featherboard. Um, and I still have to make the featherboard compatible, but it's a wider board, so the more chances of getting onto a feather.

[02:20:40]Uh, you know, but either way, I want to get this onto a board that I can have a a prototype PCB ready so when my silicon comes, I can stick one on and make sure it works. Um, but what to do with it is the question.

[02:20:57]I um this is screaming Squixle because you know I started looking at the P4 for Squixle but this gives me all the benefits of the P4. Second, it's got a separate fast PS RAM bus that isn't joined with flash.

[02:21:17]So, no flash contention and is faster and has way more PS RAM buffers and stuff. Like this would allow me to do a uh a much more featurerich squigle, but it wouldn't allow it to be backward compatible.

[02:21:37]Um, this could possibly drive the 720 x 720 screens, but I don't know. I I didn't sell enough squiggles to really justify any of that to be honest with you. uh for the amount of work development work and stuff involved. So, and this would be even less way less. So, yeah, I don't know. It's a gone of the days where, you know, 90% of my sales were microcontroll 90% of my sales are still microcontroller boards. My sales are only 10% of what they all used to be.

[02:22:14]But um I think people are less interested in microcontroller boards now than before because the market's just flooded. Like it's just so many different boards out there. Um way too much choice, very little distinguishing them all.

[02:22:33]Um you know, there are boards on the market now that have got RF switches on them.

[02:22:39]you can, you know, it's like I I'm running out of things that I can put on a board to try to be novel or first and um yeah, it's an interesting dilemma.

[02:22:58]I mean, people that are already buying a tiny board, for instance, or a the ProS3, like the the people that buy Pro S3 and put them into integrate them into the things that they're making. um you know to do a pro S31 where they could literally just drop in a new thing into their existing design would be of way more benefit to them and me in terms of barrier to entry into market but would take no advantage of the S31 at all except for a slightly faster two faster cores that's all it would give you is two faster cores and faster PS RAM which won't necessarily be a thing I can't break out any more IO because the it's fixed to the footprint So yeah, although how many how many?

[02:23:54]Ah, that's a stupid idea. That's a very How many um wait uh how many peripherals?

[02:24:06]How many how many there um LCD LCD peripheral?

[02:24:24]Yeah.

[02:24:28]control module.

[02:24:33]It connects to external C drive. You can use any GPI on the matrix.

[02:24:41]Um, it's a lot of IO.

[02:24:48]So, wow.

[02:24:54]I wonder how many concurrent rack matrices I could rack panels I could run at once with the LCD peripheral.

[02:25:13]Intermitt tech says poof and the IHO is gone. He >> Yeah, sure. Doing what you do. I'm not talking about that. I'm talking about.

[02:25:21]So, the magic of my rack, one of the magics of my rack, my my pixel rack is that I am driving driving it through five IO. These connectors at the top, each one of those is dedicated power, ground, and IO for five different panels inside, which is why I can get the frame rates I'm doing and get the number of pixels I'm doing. And each one is independently level shifted and has independent power. Not power control as in onoff, but power control as in their own power feed that with their own, you know, bulk, capacitance, whatever else to ensure that each panel gets all their power. And on the S3, I'm limited to six IO.

[02:26:26]And and I just did I did five because it the maths worked out. Five for a 19 in and three for a um a 10 in, right? Using the same panels, but I could have the num the size of the panels and double the IO.

[02:26:46]So I could drive each one each one would be less pixels I'm driving and I could so I could drive them faster.

[02:26:57]doesn't really benefit. It would if I went for a 2 U rack.

[02:27:03]So, maybe the next step after we get the one UU racks out is we do a 2 U rack display that's got 5,000.

[02:27:12]How many would it be? Hang on.

[02:27:16]This is stupid. 15. So, it be uh Oh, interesting. Be 31. 31 * 20. Uh, so 35* 5. Yeah. 5,425 LEDs in a 2U rack.

[02:27:46]Wow.

[02:27:48]That's going to suck a bit of juice. You might need two uh PoE ports. might you might need two Ethernet connectors uh for two lots of power.

[02:27:59]Let's not go there. Let let's let's let's not do that.

[02:28:04]No, I'm just being stupid now. Like I'm trying to hoodwink the S31 into something which is kind of stupid.

[02:28:11]>> Intermittent tech says 82.3 BT90W baby.

[02:28:17]>> Yeah, but who's got hardware that'll deliver that? like none of my you know I've got I think I've got two ports that can do no four ports I got four ports that can do PoE plus and four ports that can do just PoE so 30 W and 7 watts whatever it is um obviously you can buy other little smaller boards I've got one up in my power above my production line so I've just got one Ethernet powering that which then powers whatever But um yeah, like you who's got and I know you can get injectors.

[02:28:56]I don't even know if you can get injectors that work like that. Um can you get injectors that can do 90 watts?

[02:29:04]God, think of the heat. My god. And then I need I need the PoE modules like that can do that.

[02:29:12]I know lots of switches are PoE++++.

[02:29:15]I'm talking about what most people have.

[02:29:17]like I don't have one, right? Um I'd have to get a dedicated switch to be able to deliver that power. And um I I know high-end data centers or or not even high-end data, I get it. That stuff's been available for a long time.

[02:29:32]But, you know, for any of you that are are looking at possibly getting one of these, like a 10 in or a 19in rack of these, like a pixel rack at home, like you need PoE plus, like you have to have a POE plus port or an injector. So, I need to know if it isn't the same as the current ones. Can't remember when I got this, but this is a a POE plus UniFi injector, which is not for this. It's for my front security camera that is still not being powered. Like, I've still got separate power and separate Ethernet. I got this to do that, but I've never got around to uh um redoing it. So, yeah. I mean, you could do it like this, sure. And um but do UniFi have a the big one, like the 90 watt one, do they do a POE++ in an injector?

[02:30:38]It doesn't matter. I'm not I'm not making one right now. That's okay. But yeah, letting you all know anyone that wants to be using a rack that's going to be at home that wants to be able to power it, right?

[02:30:50]It it it's got it. It's Oh, you can use PoE, but you just you're not going to get it's too much power. Like if I go um the extra pixel the high-res display 100% POE plus on both 10 in and on um the 19in rack and I'll probably disable like if it detects less than PE plus I won't let it run or I'll like really limit your brightness and how many pixels can be on the screen at once.

[02:31:23]So just a warning. Yeah.

[02:31:26]Um, and your Switch is still limited. Dellex and watch total.

[02:31:33]Oh, right. Okay. Yeah.

[02:31:37]Uh, I think it's hard to develop down to a cost cuz you're competing against the Chinese. Yeah. Is there any way you can say no?

[02:31:46]Well, I already have that on my boards.

[02:31:48]That's been my I guess quality and reliability of my boards has been key from the start. I mean, there are people who still rocking and contact me who are still rocking and using their tiny picker they got from Crowd Supply.

[02:32:03]Um, I mean, I've still got one one or two lying around here that I use occasionally that got, you know, the old antennas on them, my original ones. One of them still got the blue clip, the plastic 3D print clip I used to put on that was just there for transport.

[02:32:14]You're supposed to take it off when you use it, but everyone kept them on because they look cool. And I kept mine on as well. Um, so yeah, I mean, you know, I've always tried to build reliable and quality, but you know, anyone can do that now as well. So POS+ is $55.

[02:32:34]A POE++ injector is $55 Canadian.

[02:32:40]I'm sorry. Uh, Amazon.com.au.

[02:32:46]Um, I will switch it over in a second. A uni P O E++ injector code. Okay. Unifi. No, that's not unifi.

[02:33:03]Ubiquity POE++. Okay.

[02:33:07]10. Yeah. 90 bucks. Okay. I thought they were more. I'm pretty sure I paid a lot more for my current one. I got it, as I said, a while ago, but um interesting.

[02:33:30]And you know, you can get obviously much cheaper ones that look ugly. Look at that lo vision.

[02:33:36]Sorry, that's just Sorry.

[02:33:45]That's high from the store is 55. Yeah, I'm on Amazon in Australia and you're comparing Canadian dollars.

[02:33:56]Um, yeah. I mean, you can get cheap PoE injectors, but I don't know. There's something about pushing that power through that I don't want to trust anything other than a uh a UniFi Gigabit PoE plus injector 30 W 48 volts $57.

[02:34:18]Don't know what brand it is.

[02:34:24]Anyway, so I got myself one of the not that switch, whatever the I can't remember what the switch is. The 60 watt, no, the 120 watt little um 120 watt. That one the I got one of these at home now.

[02:34:46]Yeah, the USW flex mini.

[02:34:50]Yep. See? Purchased.

[02:34:53]Um, no, that's the one I've got up top.

[02:34:55]No, no, no, no, no. That's the one that's on my rack in here. The other one. The So, it's just a flex. The new Whoops.

[02:35:08]Flex. Not flex. Flex.

[02:35:18]Maybe it's not the flex.

[02:35:21]They come in a 60 watt and a 120 watt and um Oh, yeah. This one. Oh, the Ultra. Okay.

[02:35:36]Right. So, I got the 60 watt one of this because um I upgraded one of my uh APS in my in my house to a a 7, but the power that was originally going to it was um just PoE and it needs needed POE plus. And this was actually the cheapest way for me to do it to extend it to get I didn't have any POE at home at all.

[02:36:05]They were all using injectors and they were all, as I said, POE injectors. So, um, I I literally converted six of my got rid of six of my injectors and used one of these and it was great.

[02:36:19]Um, infy 10 GPE++ to 90 W. You do here in the UK, but it's in the UK.

[02:36:33]Doesn't help me if it's in the UK. No, I get it. Yeah, I know. Lots of options.

[02:36:37]Lots of options.

[02:36:40]Oh, I've got a 210 watt.

[02:36:43]When did that happen?

[02:36:47]Anyway, okay. So, I I'm going to try to get my hubboard set up on the pick and place today.

[02:36:59]I'm going to do my best to have a working hub or two for next stream with a new case designed already and um yeah with this validated.

[02:37:16]So yeah, hopefully next week you'll be looking at that one. These ones working and we can discuss it. Um and then I have to get one in in a yellow PCB for John Oxer.

[02:37:35]No, it's not the Anyway, confusing.

[02:37:39]Okay, folks.

[02:37:43]You're right. UK is England, but you wrote UK.

[02:37:47]Here in UK, I know you're not in the UK, but you wrote UK.

[02:37:54]We can complain about USB hubs versus PoE++. Yes, we can.

[02:38:00]No, it didn't make any sense. Folks, thanks for watching. Um, nothing another nothing stream except I I do have a symbol and a footprint. I can I can now that tedious work is always what prevents me from jumping in on a new chip. But it's done now. So I can now put together a I can start massaging it into a design.

[02:38:23]Your screen says DK. My screen says UK.

[02:38:28]Are you Oh, interesting. If I hover on it, it says DK. And if I take the thing away, it says UK.

[02:38:38]What?

[02:38:40]Let me show you this.

[02:38:43]I'm not lying.

[02:38:45]Oops.

[02:38:47]Look, this is the line. It says here, look above this line. Sayary whatever 90 watt AUD 108 here in UK. It says it says UK, but if I hover over it, it changes to DK. Did you edit it or something?

[02:39:06]I'll get that bigger.

[02:39:08]There you go. Now you can really see it.

[02:39:10]Look at that. It says UK, UK, DK, UK, DK.

[02:39:18]That's why I commented about the UK. You see that?

[02:39:27]That's definitely says UK.

[02:39:32]Auto translate from English to English.

[02:39:38]But see, now you're US and messing with us or messing with US. You're in Canada.

[02:39:42]You're not in the U. Okay, I'm just kidding now. Okay, folks. I'm going.

[02:39:48]I'm going.

[02:39:50]Bye, everyone. Have a fantastic week. I will catch you all uh next time.

[02:39:57]Bye. Coffee time. Coffee time. Thanks, Matt, for the coffee. Money. Bye.