 and welcome to Disco Lady Aida. Hey everybody and welcome to a Sunday night Lady Aida o'clock as it were with me Lady Aida at my desk with me Mr. Lady Aida on camera control. I've got some fun Disco Lady Aida news, some hardware updates to the whole bunch of PR testing and a hardware building this weekend. So it's all good and I will also have a great search at the end so jump in with your questions, comments, well wishes, free donuts and more. Mr. Lady Aida do you have any updates, news you want to tell people about? Last week we moved the show to Thursday, watched our Thursday show, that was the Ask an Engineer show, watched it, great show, went over why and a bunch of other things and also if Lady Aida can go to the computer real fast. So we put out a call for help. We're talking to our community that's on the ground in India and we're gonna find out if we can send PPE. We already got a hold of a few folks. Turns out they don't really need PPE, they need oxygen and there's a couple projects that are going on. So take a look at the links that I got back from our friends that we normally just do like projects with in India and more. But there's a couple like DIY and maker-made oxygen, I guess they're called oxygen concentrators. Yeah, they work by using zeolite to absorb nitrogen out of the air and when you absorb all nitrogen, what you have left is the oxygen. So it's like, it's oxygen concentrating that by removing the nitrogen part, they require a compressor and cylinders. They're very mechanical contraptions that they can be built. So we'll see what happens. I mean, we got a chance to help folks in the US when ventilators were needed. And then ultimately there was this other things like face shields and we did those and then electronics for, you name it, we were building stuff. But we'll see what we can do with rallying the community. The thing that we can do immediately is send stuff over to India. If the hospitals need it and we have someone looking into it now, we'll ask our carriers, our shipping partners, DHL and UPS, can we get these through imports in India for customs because anything as PPE gets stopped. So whatever it is, we'll keep our community updated, but no, we're on it. We always, this is Ada for it, this is what we do. And our thoughts are with everyone over there. This is, it's every day, it's been one of the highest case counts in the world and oxygen, it might also be distribution, not necessarily the creation of oxygen. It might just be the distribution of what's needed. So we're gonna find out more, but we're watching this. We'll let the community know if there's things that we think that everyone can do or even not do. You can just keep an eye on the blog post and we'll continue to update it. Okay, with that being said, Lady Ada, what is on your desk this week? So this week we did a lot of work to add parallel capture support to Circuit Python. It's something that we actually added to Arduino. We had Arduino support for, for the same D51, but the RP2040 also has the ability to do parallel capture and upcoming IMX M7 ships also have a parallel capture capability and they do like fun IO. So basically they have a way of getting a large amounts of data through an 8-bit parallel port into the device. And these are used for camera modules. When we talked about those camera modules, last week we showed off the header and I showed off my design for a camera module breakout. But actually getting the support into Circuit Python, actually the Seles presence also, by the way, they did have camera support that they added for taking snapshots. And a lot of the work with camera is actually the configuration of the camera. Like there's so many like things that you could turn on and off and move and configure and adapt. So a lot of that was not just setting up the peripheral to capture data, but setting up the camera to send the data over the peripheral. So the same D51 has a camera capture peripheral. It has the ability to like DMA fast data into a RAM buffer. The Raspberry Pi RP2040 does it through PIOs and DMA and timers, but it also do it. So let's take a look on the overhead at some hardware built examples that I wanted to test out these PRs. So this here is a Grand Central with one of our TFT shields on it. And so the camera, you know, we tried to make it as much as possible, like we tried to line up the pins as much as possible so you could plug in the camera, but you do have to wire to the red power and black ground wires do have to be separately wired because the mega, which we based the Circuit that the Grand Central off of has five volt power and this is three volt. So the upshot is, you know, and there's also SDA and SCL, whatever the upshot is that we, you do have to bring the wires over. But other than that, it's very nice to plug it in. So you just cut the two pins off and then plug it into the end. And then this TFT is just an SPI display. So when this boots up, you'll see it's, hold on. Let's see if you can see real fast. It is, no, it goes straight. It is running Circuit Python, although you can't see it, but this is, you know. Is it because Circuit Python is too fast? It's a little too fast in this case. Oh, oh. It booted up and it's. Oh, it's too fast now. It's too fast. Oh, it's too fast now. I thought because it was a scripting language on a microcontroller, you may as well just like, like why aren't you cutting your teeth the right way with a, you know, assembler. So I can get it a little bit dimmer. Okay, so now you can see me. Hi, I'm in here. That looks pretty fast too. Yeah, so this is like eight frames per second or so, which isn't too bad. And, you know, getting the data, it's actually, I think a combo of getting the data and then you have to convert it to five, six, five format and then piping it out to the display. But it's just, you know, it's a good way to demo. It's interesting. It has a little bit of a flicker effect. I think the camera, camera, TFT camera. These are three cameras. It's making it a little flickerier. Like you see a weird, it's a flicker effect. Yeah, everybody knows that. And you don't see that. Everybody knows when you look at a camera on a camera. Yeah. But when you tilt it down, you know, kind of, you can kind of tell something. Yeah, so. So here's, let me just translate this for y'all keeping score at home. Yeah. All right, so this is what this means. Yes. See how I can see me watching me, yeah. Yeah, no flicker. So this is what this means. Yeah. So if you're doing stuff in Python and you want to do microcontrollers and you know Circuit Python or you don't even know Circuit Python but you want to do machine learning or if you want to make your own digital camera or if you want the ease of use of something like Circuit Python for microcontrollers and machine learning, this is the first step. We're there. So we'll be able to do things like an animated GIF camera. You can make yourself. You could do things like recognize a bird. You could do things like edge computing. Is that a face or not? Yeah. Unconnected to the internet. Not sending your stuff to the all over. On microcontroller. Not cloud. On Circuit Python. On Circuit Python. So yeah, we've done some machine learning type things on microcontrollers but we never actually got to the camera stuff. And I was like, well, if we're going to do it I want to really get all the camera stuff working really well. So this is the first step because I think with Circuit Python Circuit Python would be a really good match for tiny ML and TensorFlow Lite because all the stuff that takes a lot of computation happens and see in the background. The setup step, which is the most complicated, like loading the model into RAM is very annoying to do and see because you have to include it as a header or like it's on file system. It should show up as a drive on your computer and you should just drag and drop the models in. Like we do with our machine learning kits. Yeah. Yeah. And then the computation still happens in the background, right? And something, this is the same as few ones. So this is not even overclocked. This is just like the standard 120. You mean it could be faster? Okay, it could be probably like... Well, you know what? Like let's just not do that. Stop, so fast. Okay, so then we also, like I said, we also are working on the RP2040. The RP2040 is still, it's not like buggy but there's still a little bit more we have to do. So this is the same camera. Okay, I saw a little Blinka Circuit Python. So I know for sure this is Circuit Python. It's also, it also is very fast. Okay, so this one is, I will say it's not as fast because we're still working on it. Let's say, you know, it's a couple of frames per second and it has a little bit of a posterization effect which you may not be able to tell but there's something with the bit shifter. So this is a live camera view of a live camera going into a Raspberry Pi RP2040 Pico running Circuit Python. Right. So this is not a SIMD, it's not nearly as fast as SIMD51 and it doesn't have floating point support and it doesn't have SIMD support. This is definitely not gonna be nearly as fast as the SIMD51 on the other hand. Which is already too fast. It has two cores. So like, it won't be fast, but maybe it'll be fast enough. So I think we could still probably do some basic vision recognition projects. You know, maybe like tracking a red ball or something. So as to talk to the Circuit Python team tomorrow about what we wanna do with this and how we wanna integrate, you know, machine learning stuff because OpenMV did a really good job with their MicroPython TensorFlow Lite implementation. It's not CircuitPython-y, it's MicroPython-y, which like only like people who've really messed with core and library and they're like, oh, they ate the way the API structured. I'm like, okay, this is very MicroPython-y. It's very good. It works very, very well. It's also very specifically for the STM32. So... Yeah, I have a repo called the Many Forks of MicroPython. And this is both the strength... It's a great fork. This is both the strength and one of the challenges. Everyone takes Circuit Python, they fork it, they make it their own and then it never gets back merged into upstream. Yeah, I think it's a disparate fork. It's an excellent fork, right? OpenMV fork is awesome and I really love the IDE. So I wanna see like, is there a way that when we can include it as a submodule, sort of like we did with MicroLab? So I'll take a look at that. Doesn't it sound like really cool stuff? Okay, I looked at their code. I looked at the code base they have. So all the camera, they have tons of camera drivers. All the camera drivers are in C. So they did something interesting rather than like what we did with our code, you can even see it in our GitHub is we made the camera control libraries in Pure Python and the image capture in C, obviously, because it has to be, you have to do it instantly and like DMA this whole buffer of 256K. But that said, by the way, in case you're like, hey, how come, these cameras are like 640 by 480, but a lot of the visual recognition stuff, by the way, doesn't need that high resolution. It runs at like 160 by 120 or even lower, I think some of the TensorFlow Lite for microcontroller projects are like 160 by 120 or 120 by 80, they're very, very small. So even this is lower resolution than what the camera can do. Like this image is not nearly as big as, some of these cameras can do like 300 by 200, so 3000 pixels by 2000 pixels there, but still higher resolution. I think there's something here that we can do, we can use this. And of course, you can configure the camera to send you thumbnail images. So that's this vision stuff. So this is actually pulled into core. So if you want to try it out, the 751 is in the latest build, which is now 7.00. We've also been doing micro Python catch up PRs, Scott's on a roll. I think he did 1.10 and he's doing 1.11 micro Python. So we're going to slowly catch up. So if you've been missing the Walrus operator, I think it will be merged soon. Okay, so it's camera stuff that's happening, and of course we'll make our own breakout and maybe we'll make a feather wing with camera stuff. We'll do something cool. Okay, so that's the camera stuff. All right, well thanks. So the next thing, also circuit Python-y, is I saw this in the newsletter and I thought it was cool. So let's go to the computer so I can show this off. So this is U2IF, which is an idea that I've kind of had and I just never got to it. And I love it when people do things that I meant to get to. I really hate people who are like, I had that idea for us. Like, I had that idea for us maybe, but I don't care. Cause like you did it and I don't want to do it. And like you did it for me. And this is a great open source software. Love it. Like, the less good I write, the happier I am. So this is a firmware for the RP2040, specifically for the Pico, where you see here the computer, this is a computer. Computer is a Python program. It has MicroPython modules, which we can ignore that. It's basically just the interfaces is MicroPython compatible through the U2IF interface through USB to a Pico, which then has GPIO, SPI, I squared C, NeoPixel, et cetera. What does it sound like? It sounds like Blinka. Like our Blinka support for the FT232H and MCP221. So this is basically a firmware that turns a Pico, $4 mark controller into a USB to hardware interface, where then you write code on your computer with CPython. So Carter did a great job the last few days and added support for U2IF to CircuitPython library support via Blinka. So the same code that runs on a Raspberry Pi, that runs on the FT232H, that runs on the MCP221, that runs on the Pico when it's running CircuitPython, the same U51, same U21, all these chips, same library, same code, will now run on CPython. You just load the UF2 file, right, drag and drop the UF2 file for the firmware for U2IF. And now your Pico is like a little, like a hardware dongle Swiss Army knife, which is sweet because it's so inexpensive. And it's not Formata. And anything that's not Formata is awesome. I've just been burned by Formata so much. So this actually just got merged and you can see the demo image that Carter made, which is like a breadboard with like so much stuff. So first up, he's got a button. So digital input works with pull-ups. This is a little potentiometer, can barely see it. So he's testing analog inputs. He's got a BME680, so it's an I-squared-C sensor, LIS331, another I-squared-C sensor. This is a Neopixel ring over here. A BMP388, so you got like a lot of sensors, I think a servo going on over here. So it's PWM output. So basically he tested everything and it all works. He did a really good job. The only thing that it does that we didn't implement is I2S because we're like not 100% sure how we want to expose it because it's like, it's not normally exposed. We're not sure exactly how we want to implement that and we don't usually provide a blink of layer for I2S. But I-squared-C, hardware SPIs, this TFT is a hardware SPI display, GPIO input output, PWM and analog input. So that's cool. So added support for it and then we went back and forth and we did some fixes and blah, blah, blah. And basically we got it merged in like two hours ago. So let's go to the overhead and I can show my little- We're making things too easy for people to do whatever they want with their computers and their electronics. I know, that's what we do. So here I've got, so this is a penchiometer. I mean like, by the way, for the longest time, all folks wanted for like art installations or sensors or whatever is like, look, I just want to run Python on my computer. Can I just plug electronics into it and so they can communicate back and forth? I just want to get- That's all I want. They're like, I just want to get the sensor data into my computer. I have a Jupyter notebook thing. I have a spreadsheet thing. But they're processing. I have all sorts of things that are already on my giant computer that could do anything. I have Python on it. I have my 15 different version of Python. You know, just in case I have to like do different things, save them in a little pocket universe. I actually remember like it would be a final project at MIT, like a MIT student would have a final project of like, I want some step sensors that cause a video to play on a computer. Like that was really, really hard. Like incredibly hard. Like- Here you go kid, PhD for you. Yeah. And so- Yeah, PhD. But it would be like a final project in an art class. Like, okay, how to get, make a step sensor, right? It's like two pads that you press on and it makes a contact. And then get that into a computer and have it play a video. And this was like art installation. So this is- Hard. This is super easy. This is now super-duper easy. I think things with Jupyter notebooks, people are gonna do a lot of needs up. Yeah. So this is an LED, this is the PWM output and then this is an analog input. So even though it seems like, oh, this is just the resistor for this LED. It's not. I'm actually reading the potentiometer value. Wait, hold on, I'm gonna write this. All right, botics keep guys says, every time I want to interface a microcontroller to Raspberry Pi, a regular computer, I think we're from on and deciding not to go there. Exactly. And QIF will change everything. Yep. Yeah, nobody, nobody's a big fan of- Does CPython run on Raspbian? Yes. Whenever you run Python at a command line, like Linux or Mac or Windows, that's called CPython. Yeah. Because it's actually written in C. Yeah. Normally we use to call it Python, but like Python is a language. CPython is the thing that you, when you type Python, that's the program that's running. That makes sense. And then would you be able to like get a STEM involved in this? Yeah, we had the STEM of 2040 that we showed off, I think last week or the week before. Yeah. So one of the things that I saw about this was like, ooh, I'll totally have a port for that. And that way you can just plug in your STEM aborts. And that'll be a way to add any STEM. Anyway, so, yes. So this is an analog input. And the analog input is converted to a PWM output. So this is actually testing two things here. And then over here I've got, let me turn this down, so right. This is a distance sensor. And then as I move my hand closer and farther away from the distance sensor, you can see I'm making the neopixel lights. This is my neopixel test. So, and very fast. And then you can see the onboard LED, that's data. So that's every time the message is passed back and forth. So this is very stable. I mean, this has actually been sitting on my desk for a few hours. And I kind of like every few minutes I sort of like noodle with it. And it's very stable. So the person who did the U2IF code did a great job. So check it out. All right. And then blink and we haven't done a release yet. So you'll have to wait till tomorrow when like folks wake up and we'll do a release of Blinka. And then you'll be able to use it and we'll also eventually have a guide. Do check out the pin out for the, you can't just use any pin for anything. The ITC pins are kind of like fixed. There's only a couple, you know what I mean? Like they don't, it's not, it's not as flexible as circuit Python where it's like, use any pin for like anything you want. You have to, there's some pins for some things. It also does a UART mode also. We have to add, I don't know what's more. Anything else or you want to do some searching? Okay. So that's, yeah, that's my desk. And so let's, let's head into the great search because I just shut off all my hardware. Where are you? Great search is brought to you by Digikey. Thank you Digikey for helping make the great search happen every single week. Lady Edie uses her powers of engineering and more to navigate the Digikey site, to search for all the parts and components and we take requests and speak enough. Speaking of, we have two requests. We have two requests. So we're going to get to do one this week and one next week. So somebody posted on the forums and they asked, okay, they want a latching power relay like the Omron G6C. So basically they said, look, they want a power latching relay feather wing. This is what we've got now. This is the power relay feather wing. And power relay feather wing has a big sugar cube Omron relay. And you know, you turn the relay on and it moves the latch, not the latch, the throw over to the other pole. And so the middle pad of this terminal block connects to the one or the other from the side. So either the left side or the right side. And that is a really nice, you know, high voltage friendly and we have a cutout and it can handle pretty high voltages and high currents. So this is 1200 watts. So it can do, you know, five, 10 amps of current for like a small appliance. Like it can definitely handle a small toaster oven or lamps, you know, fans, lamp, sort of think like small heaters, televisions, et cetera. Appliances, things around your house that use wall power, motors, et cetera. So this feather wing though, the relay that it uses is a power relay, but you have to hold the pin high to keep it connected. So basically like if you want it to only turn on when you turn the relay on or you want the contacts to close to create the circuit when the relay's on, you have to hold that pin high, which does use a lot of current. And the alternative to a normal relay is a latching relay. And a latching relay, you don't have to hold the pin high or energize the relay, you just toggle it and it sort of like clicks and latches the relay into one position and move another pin and it latches the relay the other direction. And we do have a latching relay feather wing, but it's only this mini-style. So you see there's a latching and non-latching. So this one is latching and it takes two pins, a set and unset pin. But this relay is much smaller than a telecom relay. So it's only like maybe a half an amp, not very high power, it takes 70 watts or so. Compared to the other one, which is over 1200 watts. So this person had a good point though. I was like, oh yeah, you know, I never made a latching power feather wing. I should do that. I should go didgy key and search for that because finding a latching relay that'll fit on a feather is not necessarily easy and a lot of relays are really big. We want something that's really slim. So let's go to the overhead real fast and I'll show that. So this is, I happened to have at home, a relay feather wing. So you can see it's kind of chunky. And this is about 15 and a half millimeters by 21 and a half millimeters by about 18 millimeters. And it's solid onto this feather wing. There's a, you know, it's not taking up all the space. I do have a little bit of room over here. So it can be a little longer, but I really don't want it any wider because you see it's already, it's kind of getting near the edges of the feather wing. I want it to be narrow. So let's make sure that we get one that's narrow enough that it fits onto the feather. It can be a little tall. I mean, this is, it's a sugar cube, it's chunky, but I don't want it to, I don't want it to be longer than the feather wider than the feather wing. All right, so let's go to, did you key? So let's go to the computer and I'll do the searching. Okay, so let's search for relay to start. And you're like, why don't you search for latching? Cause if it's latching and non latching, I'd rather just go to the relay section and then search. So, okay. So there's power relays and signal relays. So signal relays, again, like telecom relays, they're less expensive. They're smaller, they're often surface mount, but they don't do high power. They do less than two amps. Now granted that two amps is, you know, is it at low voltage or high voltage? Either way, I definitely want it to be five to 10 amps of current at 240 volts AC. So I definitely don't want this signal. I want the power relay. There's also solid state relays, which are going to be, they're different. I think, you know, they're not gonna be latching. They work in a different way and they don't work for DC voltage. So I want to make sure that this works for AC or DC. I want the mechanical relay. So let's go to the power relay section. So I have 22,000 relays. Let's go with active. Cause I want only ones that are, you know, I'm going to make a product out of this. Now normally it would only click in stock, but because there's been so many shortages of components, I'm going to click normally stocking instead. Because I want to make sure it's, there could be good ones that are just not in stock right now, but they will be in a couple of weeks or maybe a month or two. I want to find what's the best possible thing because relays are very specific. You know, they're not like some components, like some silicon components where it's like, well, you know, the op amp or the transistor, I'll just find whatever it is that's good enough. I really, you know, when I pick the one I pick, the mechanical shape is going to be very specific and it's not going to be easy to go to a different one. So I really kind of get all like, pick one and be done with it. Okay, so next up, the next thing that's really important is the coil voltage. So because it's going to be on a feather wing, it has to be powerable from feathers and feathers are three volt logic. They get five volt power from a USB, but they can also run off of a battery. And in fact, this person who posted said they want to run off a battery, which is why they wanted to be latching. And the batteries, you know, they're four volts or so, but really, you know, they can start dipping down to 3.5. So I don't want to require a five volt relay. I want it to be less than that. I think three volts or less to even 4.5 volts is a bit high because remember, LiPo batteries, 4.2 volts is about as high as they get. And there's no, I don't want to risk it. You know what I mean? I want to make sure that if they want the relay to activate or deactivate, that they can do so. The coil voltage, I'm going to select 2.2 to 3 volts. And you'll see this really reduces the number right now. I think I have like six, you know, we went from 12,000 or sorry, 20,000 to 6,000. To now it's only going to be 89 options. Okay, so next up, there's, you know, a couple of things like, you know, coil insulation or must operate voltage, but none of these are important. So there's now two kind of coil types, right? Because this is, why are we here? We're here for the latching. So the next question is, do you want single coil or double coil? And I actually didn't know the difference between double coil and single coil until I had to go look for this because I usually didn't matter. But when I picked, double coil is a little simpler to use. Double coil is you have two pins and when one goes high, it, you know, the relay latches one way and the other pin goes high, it latches the other way, right? So each pin, when you toggle it high. However, what I noticed was is that I had a lot fewer options with double coil and single coil. There's more single coil options because I think the mechanical construction's a little simpler. So I will select both, but I will note that single coil is a little easier and I'll show you how we're actually going to use it. When we get to use it. Okay, so the next step is, do we want single pole, single throw? Not really, single pole, single throw means really either connected or not connected. I like to have it be normally open, normally closed if you have two options. So I'm going to pick the SPDTs and the DPDTs. So you can have something open or closed and they pick which way they want it. Surface mat or through hole, I don't think I'm going to find it as surface mat one, but I'll leave it all selected. Okay, so going down here, I'm going to view the prices at 500 pieces just because relay is like, one relay is usually like $10 and they quickly go down in price. So I want to compare them at a reasonable price. So yeah, there's a couple of good options. So this one is sneaky. So there's this telecom relay and you're like, wow, there's a surface mat that looks pretty small and there's 10,000 stock. This looks great. The problem is, is that I actually went and looked at the data sheet, which hopefully will load. Oh no, data sheet, you can do it. Okay, so when you look at the data sheets, even though it's in the power relay section, if you look at the load, the current is five amps max, but only at low DC voltages. And as you get higher, it definitely goes down. And so I think it doesn't seem to, it said it was okay for AC, which is probably going to be the same as DC, but basically it's a single relay, it's not a power relay. So even though it's in this category, and it was like very attractive looking. So the next one I looked at was the Omron one. The reason is it's like, well, I know Omron ones and they look pretty good. It's inexpensive. It's not in stock right now, but I know that they're normally in stock. I also looked at this TE one. Although the pricing is significant, like this one's 250, and then they get to $4 quite quickly. And I do have to keep it price sensitive. So let's look at this relay. So this is a high switching current general purpose launching relay. That's kind of what I want. So this one, this is a confusing day sheet. Okay. So for this one, the thing I noticed was for the single winding latching type, you'll see it goes down to three volts rated, you know, 200 milliamps. But if you are five volts, hold on, this is like the trickiest day sheet. Sorry, this day sheet has like this weird page that kind of flips it. Okay, let me move this over. Okay, you see it, great. So for the double latching type over here, you'll notice that it doesn't have a waiting voltage lower than five volts. And that's why, you know, I was having this issue where I was like, oh, I want the double coil because it's easier to deal with. However, when I did double coil, I didn't get to find this particular relay because only the single coil version has three volt activation. So the difference in how you use it, and let's, again, hold on. Yeah, see this day sheet, it's like wide. It's kind of an interesting day sheet. So this is the single, this is a single coil relay. This is what it looks like. So you see there's, you know, one coil and there's one pin. And then when one pin is high and one is low, it pulls the armature one way. And then when you invert it, so the other pin is high and one pin is high and the other is low, the armature moves the other way. Why does this matter? So normally if you have a double coil latching relay, of course you have to put a lot of current through the coil. So you're gonna have two transistors, one for each coil, to help you drive that current through the coil, right? So you have, it's basically like a solenoid. You have two power transistors of some sort, you know, enough that it can pass 200 or 300 milliamps. And then you can turn each one, you know, high or low using that power transistor. With a single coil relay, you have to treat this like a motor. It's kind of like a bi-directional motor. So you actually need an H-bridge for this because you need to pull a lot of current through one way and then pull a lot of current the other way. It's basically an H-bridge, a little more annoying because now you need four transistors of two. However, it's a $1.50 lesson price and there's more options available. So, you know, I'm probably gonna do that. It's not a big deal. I can, you can make an H-bridge out of, you know, four transistors. It doesn't have to be a great H-bridge because it doesn't have to stay on for a very long time. It doesn't have to be on long enough to latch the relay one way or the other. So it can be like a kind of like, you don't have to use the best transistors in the world. It doesn't have to be good enough. And then you just have to have four diodes as well as flyback diodes for this. But basically you have to treat it like a motor. This is the single-pull, single-throw version and oh man, this data sheet, man. It's kind of weird. And this is the double-pull, double-throw. So this is the one I'm actually gonna get. And you can do, I'm just gonna double up. You know, it has, you know, multiple pins for some reason. I don't know why. They're actually inside. It's only one relay. So I don't know, twice as much current goes through. Maybe, I don't know. And then, again, this sneaky, I don't know, like this data sheet. Okay, and then over here you can see the, if it's a single double coil, you see how it works inside. There's actually, it is one coil, but it's doubled up and then you have to kind of like pull it from the middle. And that's why I guess you can't use it with three volts as a five volt only. Five to three volts. So, you know, my decisions have kind of been made for me. That's okay. It's a pretty good relay. So let's, next step, you know, now that we've decided, okay, we're gonna do the H bridge. We're gonna do the single coil. Let's look at the mechanics. Oh my God, this data sheet. Okay, well these are all mechanically the same size. So this is 29 by 12.7 by 15.5. Let's go to the overhead again real fast and we'll see how this compares. So length wise, this one is 22 or so and the one we're picking is about 29. So I like to sort of measure it. So it's gonna be a little bit longer, but not, you know, it's not longer than the feather. So that's fine. And then for width, it's gonna be 12.7 and this is 15.5. So it's gonna be thinner. It's gonna be kind of significantly thinner. It's gonna be a little thinner. It's gonna be skinny like that. And then for height, it's 15.5 millimeters, which is, I'll just do it here and you'll just have to trust me. So that's about 15.5. So it's actually a little bit shorter as well. So it's basically gonna be a longer, thinner, shorter version of this relay, but it'll definitely fit on the feather ring. So I'm gonna order some of these samples. Since the three-volt one isn't in stock, I'm probably gonna pick up like the six-volt one and then I'll just like jury rig it to, you know, get it tested and trying it out. I'll feed it six volts and then get the mechanics set up and then I'll be able to convert it to the three-volt version when I finish and then do the testing before I release the product. So the winner of this week's The Great Search is this Omron G5R LU1E DC3. You're the winner of this week's Great Search. Thank you. Okay, so here's some questions that came in and someone's typing one up. So we'll get to a few of them. Someone said I could see using SBC, using the U2, I have to talk to one or more microcontroller, small cameras, mind blown, yep. So does that relay need an H-bridge? It does need an H-bridge. So I'm gonna have to craft one up, but I'm just gonna use some, you know, FET transistors, whatever I've got and I'll just lay them out. I'm not gonna use like a motor driver. I don't need something that good because again, it only needs to pulse on. I don't have to worry about low resistance in the transistors because they're only pulsed. They're not gonna be on for a long time. And then one of our friends from Show & Tell who had their open source power chair, they're starting with a richer kit to get old power chairs functioning again. Basically, strip the chair down to the frame, seat and motors, putting it in the kit and the kit will be completely open source. So there's a couple of biz questions and I'm going to always say we can't give legal advice but we'll give you advice and we'll share our stories. I can give some entry advice but I can't give you advice. So here it is. I was talking with a lawyer, the volunteer's a scorer and he kept talking about a lot of pressing to get my thing patent and what I'm doing is intentionally, easily duplicatable. So to me, what's the point of patenting? Then I wonder if he's just really just practiced what true open source is and product production. So here's what I will tell you. At this stage and, you know, based on the project you're doing, it doesn't make sense to do a patent at all because you should start to get it out there. You should publish, you should put it up online and I know a lawyer will say things like, well, once it's out, someone can be first to file. It's much better to get it out than to worry about this. It's kind of like spending forever or thinking of a company name or a logo. Get that idea out of your head. And you have a year to file. Well, even if you didn't, it's okay. And the things that are worth protecting are not things like modifying a power chair and making it open source and reusing it. That's not something that's like, oh my gosh, I need to protect this because this is such a big idea that we've invested millions in. Get it out there and do the best you can and publish it and use that to move on to your next big idea, your next big project. At some point, if you get enough things going and you have an electronics company or if you have an invention company, then try your hand at patents. We are not a hypocrite. We decided to try to get one patent just so we could give it away. And the process was so long, by the time it was finished, we already released like 1,500 new products. So if we only sat around and waited for it a patent to get approved, we would just be 1,500 products short. Yeah, lawyers will always tell you to patent stuff because they make money when you patent stuff. Yeah, and it for patents is great for some people, but for what you're doing and you wanna be in the open source community and you wanna do all this stuff, do this and you can always think about it later, but trust me, this isn't the thing to spend that time and that money on. It's not. It is a lot of time. You're basically repairing. I haven't seen any company succeed or fail because of a patent thing. And at the higher level, they just use these nuclear arms. It's like, well, we're gonna nuke you with our patents and if you nuke us with your patents. Yeah, it's cold war patent diplomacy. So, any other questions? No, this is what we're here for, to give this type of advice. Let's see. Oh, this is the first dream of y'all is a properly caught. Just received my DigiQ orders that I made thanks to the power of LadyAid and the Great Search. Yeah, this works. Great power. You watch the Great Search. From great power comes great searching. You save a lot of time and then you place orders. Okay, I assume resetting a latch is why you need an H-bridge. So, could you use a non-latching and avoid the H-bridge or do you need latching for some reason? If you check out the post, the person really wanted a latching relay. So, that was not debatable. I'll see if this other question comes in. And then, are the T-shirts like LadyAid is wearing available now? No, and Pedro made this for me. You can check. They have video streams where they show how to make your own T-shirts. Ironing. They do iron-on. So, we have this thing where we just don't put the Adafruit logo on stuff besides our circuit boards. Yeah, we don't do that. I've seen a lot of tech companies put their logo on shock glasses and all sorts of stuff. This is not our vibe. We're just like, well, let's keep the logo special. So, when you see it on something like a circuit board, you're like, oh, this is... Yes, you can buy a circuit. You can buy our logo. But we do appreciate it. I want to do something. You can make a T-shirt out of circuit boards. Yeah, I want to do something in the lines of clothing and apparel, but, you know. Yeah. All right, so I think that's it. All right, thanks, everybody. We'll continue to answer business questions each chat. So just save them up for you. And that's it. All right, thanks, everybody. All right, see you during the week. And once again, please continue to check our site. We're going to try to check in with our friends, family, community in India and see what we can do. Oh, let me just make sure this one just came in. The whole idea is making this to get people mobile to me and say, are you helping people who want to produce a consumer? Yeah, yeah, okay, great. So lawyers will tell you, here's my outfit in closing. You won't get a direct answer out of a lawyer either. If you say, should I patent this? They'll say, well, if you don't patent it, then you can't patent it. Like they won't tell you, they won't look at your idea and say, hey, just doesn't make sense for this. Lots of other things totally make sense for patents. Even as flawed as a patent system is, there are reasons for certain things. And I would always say like, oh yeah, like you just invested like $45 million in this thing that helps discover the X and the Y and the Z and it's a process and you wanna recoup your money over a course of a few years. Sure, patent is kind of like a thing. But there's other things too. There's copyright and trademarks. Even when you do open source, people have to abide by the license because it's copyright. And for trademarks, you can trademark the name of your invention. You can just say, hey, copy it all you want, just don't call it that. So you have lots of, lots and lots of things. And then someone's gonna ask, can I put the logo, the show and tell on shirts? Yeah, sure, you can do that. Just do it for you, don't sell them because that'd be weird. Just do it for yourself. Okay. All right, thanks everybody. All right, bye everybody. Bye.