How low can the power consumption of the Cat Feeder Unreminder go? Parker and Stephen discuss leakage current on this episode of the podcast!
The quest for the right connector for a project! The right of passage for any hardware electrical engineer starts with a connector catalog.
This is the last installment of Stephen's 'Adventures in Injection Molding'. We are going to recap the entire two year sage and close the book on it.
Parker
Stephen
Parker is an Electrical Engineer with backgrounds in Embedded System Design and Digital Signal Processing. He got his start in 2005 by hacking Nintendo consoles into portable gaming units. The following year he designed and produced an Atari 2600 video mod to allow the Atari to display a crisp, RF fuzz free picture on newer TVs. Over a thousand Atari video mods where produced by Parker from 2006 to 2011 and the mod is still made by other enthusiasts in the Atari community.
In 2006, Parker enrolled at The University of Texas at Austin as a Petroleum Engineer. After realizing electronics was his passion he switched majors in 2007 to Electrical and Computer Engineering. Following his previous background in making the Atari 2600 video mod, Parker decided to take more board layout classes and circuit design classes. Other areas of study include robotics, microcontroller theory and design, FPGA development with VHDL and Verilog, and image and signal processing with DSPs. In 2010, Parker won a Ti sponsored Launchpad programming and design contest that was held by the IEEE CS chapter at the University. Parker graduated with a BS in Electrical and Computer Engineering in the Spring of 2012.
In the Summer of 2012, Parker was hired on as an Electrical Engineer at Dynamic Perception to design and prototype new electronic products. Here, Parker learned about full product development cycles and honed his board layout skills. Seeing the difficulties in managing operations and FCC/CE compliance testing, Parker thought there had to be a better way for small electronic companies to get their product out in customer's hands.
Parker also runs the blog, longhornengineer.com, where he posts his personal projects, technical guides, and appnotes about board layout design and components.
Stephen Kraig began his electronics career by building musical oriented circuits in 2003. Stephen is an avid guitar player and, in his down time, manufactures audio electronics including guitar amplifiers, pedals, and pro audio gear. Stephen graduated with a BS in Electrical Engineering from Texas A&M University.
Special thanks to whixr over at Tymkrs for the intro and outro!
Hello and welcome to the macro fab engineering podcast. We are your hosts Parker
Dolman. And Steven Gregg.
This is episode 256. I did not edit the intro from last week. So I was reading that I'm like, wait, I am not Josh Rocher. Pause myself real quick.
This one document has been copied 256 times to over this time step.
I'm wondering if I wonder if Google doc suffers from bit rot?
Probably not. I think I think our brains suffer from
Oh, for sure.
I don't remember 90% of the stuff from college.
I actually I up
in my garage I have. Okay, so I kept every homework assignment, every exam I ever took. Like, legitimately. There's, in fact, out of all all of my exams, I think there's only two that I wasn't able to keep because the professor kept them and just told us our grades or whatever. So I literally have every document super
secret.
Yeah, yeah, he did. He didn't want you to know they're actually passed. He just gave you a 17 because he loved it.
Your he had he had top men grading your your
Tuchman TAs, underpaid TAs.
Well, the top men grading your papers. And so they're in the, the area 51 in the Jones vault, next to next to
the Ark of the Covenant.
And the alien head or alien? Christmas. Christmas goal? Yeah.
Well, so So it's funny, I opened up this this box just the other day and looked in there. And I was just looking at some of the papers. I was like, Wow,
I did this at some point in time. You know, the fact that you brought it from Houston to Colorado as well that like Denver. Its commitment.
You know, what's funny, sitting on top of all the papers is my diploma in its original tube. I just never took it out of
it. Just kept all that paper and never like, framed the actual important thing, this
thing that all that paper means I did it sitting on top of
all that. Yeah.
Yeah, I will. I mean, luckily, no one's ever asked to prove that I have the diploma. I'd have to go fishing.
I got my that was actually when I graduated, my my grandmother bought me the the frame for my diploma. So really nice. Mike's ut one with like, this like emblem on it. It's pretty nice, actually, oh, you've seen it? Because I had it up at the fab.
You know, okay, so not that I need any more projects, I'm opening my drawer, I think I have this
thing he's gonna make. It's gonna He's gonna take, you know, 20 months to build a PCB frame for his diploma.
No, I wanted to do a nice wooden frame. I've shown Parker this before. But in school, I got to make a semiconductor die with a bunch of semiconductor parts. I wanted to build like a little circular spot into a frame with my diploma and embed this into it. Because I think that that would just, that would be a reason to frame. Yeah, the
idea is what kind of magnification? Would you see? Would you need to see your school's logo that you put on there?
Gosh, if I remember, right, the
the size of the logo is 150 micron. No, that seems really, that actually seems big. Maybe it's 15 micron. So I don't I don't know what kind of magnification it like. You can if you look really hard, you can see you can't see it, but you can see where it is
you can see the dot because that wouldn't be cool. If you could figure out how to frame it where like, you could see that logo like because like Oh mine my diploma at the bottom. You know, it's got like the seal of the school, right? Yeah, that'd be really cool. If you could have your your dye that you made, and then like a magnified version that showed like the logo of your school below it. Yeah, that'd be really cool. It would be might be hard to do. Like it'd be more like a shadowbox so you'd have to like have a focal length and optics. How do they work? How do they work?
I can't you know Okay, so did you ever did
you just scope creep
your Oh, you totally did. But but but one thing the one thing that I never undo stood, optics fell under Electrical Engineering at Texas a&m. Interesting. Did that was that true for UT?
Some? There were some optics under it. And then some most of it was under physics.
Yeah, yeah. There's there's a lot of crossover. But But somehow, if you wanted to become an optical engineer that fit under the electrical engineering path? Yeah, I
don't think at UT. Basically, if you want to do optical stuff you were physicists.
Which I guess makes sense. Yeah.
But there was like, there was a optical lab at at the basement, like level four of the electrical engineering building. But
did they have all like the air tables to reduce vibration and stuff?
The best thing was moving like pallets to their own, like people telling you to move pallets slower.
Few 1000 pound pallets
that Oh, that was I don't know if I talked about this on the podcast. I worked on campus at University of Texas in Austin, with electrical engineering. What college it's not college, but department I guess. And the, my main job was to do like mail sorting, and like, make sure the copy machines are full of paper. But then, then, you know, at the end of each semester, there's the surveys that all students have to take, I don't know how many schools do surveys, but ut surveys were a big deal for for all the classes. And they were paper. They weren't well, when I was there, they were paper my last year they turned to digital, okay, or at least our department, but they were paper. And so you had to move them from the the way your department was that up to the there was a office that was like sorting and counting office. It's not like they had some very wank prestigious name on the front of their their limestone built building. And, but their their engineering was like at the bottom of the hill. And this is at the very top of the hill on the campus like Yeah, exactly the flows downhill. And so that's actually how I became friends with so many people was learning where what buildings had lifts. So you can go at the ground floor and like exit at another floor that was higher up on the hill. And so you make friends with like, you know, the janitor or like someone in the lab that had lift access. And so you would do that. And the best one was the the biology building, because you could, because they had a lift that would go ups like four storeys. And so you would actually traverse like 50 feet vertical, through their, their building. But on the way, the person I knew who had the lift access, also had access to their, like, freezing room, where it's like the route the really cold room to do experiments. So it'd be like 110 degrees outside, you're pushing this cart all the way across campus sweating like crazy. And you pop in there, it's negative 20 Fahrenheit.
The sweat freezes.
Oh, yeah. And the thing about that, though, is that it was that lab. It was biology lab. And so I don't know exactly all the experience. But like, whenever you walk in there, like don't touch this stuff. This you basically walk in there with your hands in your pocket. See, make sure you don't touch anything. Like even by accident, you might contaminate it, or know just get infected by something. Oh, yeah, they've COVID strains all all 19 of them in there actually was aids. Oh. They did aid research on yeast. Apparently, yeast was that you can grow cultures yeasts really quickly, and then have genetic mutations really quickly in yeast. And so for some reason, that saves me it's probably because you can get a lot of gene mutations with the virus in yeast. I think it's probably why there's probably a biologist that's listening that's probably like to call it that details wrong. But that's true. Jump in our Slack channel and explain better to me then I remember from 10 years ago,
I got I got an interesting tangent about yeast real quick. So I Parker and I had been brewing for a long time. Let's put it that way. And something that I've been doing since the day one is when you go to put your yeast in the beer, you I've gone through great lengths to make sure that The temperature of the beer and the temperature of the yeast are spot on perfect. That's it, you're pouring yeast into, like, it's perfect temperature environment, kind of like, you know, when you buy a goldfish from Walmart, you put it in a bag and you you let the water warm up. Because if you throw a goldfish into like, you shock it with 30 degree change, like, it can't get away from the temperature, right, like, so the goldfish dies. Well, the caught that concept has been true. In my mind, at least with yeast for a long time. This last batch of beer, I got a packet. And the packet of yeast just said, Take from fridge dumping in beer, and it doesn't say like any like, like recommendations about temperature or anything like that. And I went and did some research on this company that that put this claim on their back because it was so ridiculous in my mind. And their engineers are what a biologist that they have the geese people, these engineers, they said on their on this YouTube thing. They were like, Yeah, you don't need to worry about that, like, first of all, yeas are really, really small. So they come up to temperature, they they the temperature change happens instantly. It doesn't necessarily kill the yeast. And they're like, we take yeast out of our fridge, which is negative 180 degrees Fahrenheit, and we put them directly in like 90 degree Fahrenheit stuff and you don't like some might die, but it's not going to kill like half your entire culture, which totally changes the entire way. I handled that now because like I like I said, I went to great lengths to make sure that they were at the same temperature. You kind of don't have to do that. And one
of it is because like organism size and then hardiness. Like we think a goldfish is a pretty hardy animal. Yeast can live anywhere. Yeah, yeah. Almost anywhere, I should say. Anywhere is a very all encompassing word. Can it live in the middle of the sun?
Probably not. Probably not. Tardigrades kid. No. Yeah. Okay, we should probably actually start this.
Start the broadcast. Right? Right. Yeah. Okay. So the caffeine I saw someone mentioned in I can't remember who in the Slack channel was mentioning the cat feeder and reminder. This is not a forgotten project. It is just I'm still waiting on the components test them with and none
of our projects are forgotten project. They're just not active.
Just not active. on pause,
yeah. Okay, so where are we at on it? Are you
I'm just waiting on components and the currently the pennant tars taking all my electrical time. Like, all my after work, electrical time is an independent tire right now. It might be for next couple of months as well, just from all the other stuff have to do with Penta because like, oh, yeah, sure. The board's mostly designed by Hall now there's a power supply, auxiliary systems specking out bill of materials, all that good stuff. The good thing is actually the worst of it is done. Like the Main Boards done. The light boards are done. The why all those wiring diagrams are done.
Oh, did you end up getting quotes for those?
fuck boy, you forgot. That's easy. That's an email and PDFs. That's fine. That's completely forgot.
I have remembered. We're gonna do that. And we're gonna talk of talking about it. Yeah.
We forgot about it. Yeah, that put it
on the list.
Yeah, I'm actually gonna add this to the notes so that when tomorrow when I do the blog post, I look at it like, Okay, I mean
you know, I run into the exact same thing. 99% of the time, the hardest thing to do is that last 1% Like literally putting the PDF on an email and clicking Send, like can take longer than drawing all the all the PDF. It only took
about about an afternoon. Three hours, three hours to do all those drawings. So it wasn't too bad. But yeah, I still haven't. I am I did all the research for I got a couple of vendors to use all that good stuff.
Hit me up tomorrow. I'll give you my cable guy and you can get a quote from that guy. He's awesome.
Okay, is it not China? Taiwan perfect. That's what I want. Because I feel like I have a US people and I have a couple Taiwan and then I think this one Malaysia so eventually they all come back within like pennies. Like okay, so an actual project I have been doing a lot of thinking about though, is long, long time ago and a podcast Far, far away. Oh, I love this. The Jeep fan project revisits Nice. So I built. So the what the Jeep fan project was was a Jeep fan controller that also was able to take sensor inputs from all over the Jeep like temperature. And I think it was it could do, it could read the speed sensors, stuff, all that kind of stuff, and then display it on a VFD or vacuum, vacuum fluorescent display. I actually got that working and use it for a bit and ended up and not using it in the final. At that current stage of the Jeep, I ended up not using it. Because it just it didn't do much to what I needed at the point at at that point in time. Well, I'm rebuilding the Jeep, again, the red Jeep, and I'm at this stage where I've got a couple like individual gauges I have now but I'm taking the gauge cluster out. And I want to just put one gauge in. So like a universal multi gauge is what I want, which is kind of what the Jeep fan did, it took a bunch of different sensors and just displayed it on one display. So I started actually looking at like, Okay, is there any off shelf stuff? Because I'm like developing the Jeep fan into this. It would be it wouldn't be that expensive. It just would take time. Right. But I'd rather pay a couple of dinero. Couple of dollars. I was about to say those are the same thing. Yeah, get something that works. But the two, there's two products out there that kind of work like this. But they kind of suck for different for different reasons. So one is, can't remember, I'm owning the companies, but one is the ad it's actually called like, multi gauge, or something like that. And the problem is it's more of a for like racecar drivers and stuff like that. And they don't have a lot of different inputs you can put into it. And the Customize customizability for doing the inputs and like naming the inputs like I want like I mostly you want a lot of temperature sensors, like I want to measure oil temp, transmission, oil temp, all that good stuff, where it can only do like, two, it got tender can do four, but two or four like air and two or four fluids. And like I want for fluids. Why can't I do for fluids? In the end, it's a K type thermocouple Come on, guys. And so and the thing is, it's reasonably priced this system is it's not too bad. It's like a couple 100 bucks, you're in good go. Well, for for gauges for automotive nice automotive gauges is like inexpensive, like the gauge cluster I put in the wagon is like 800. So for a couple 100 I'm like, okay, that's not bad at all. Now, there's another company that it's actually a company with a cluster, I put it in the wagon, they have a really good expandable system. So you can add whatever temperature modules you want. You can name them whatever you want, you can add as many basically any modules you want. And any kind of sensor inputs and all that good stuff. The problem is they don't have a, they don't make it like a multi gauge, you have to buy like a whole set of gauges. And then you have a screen that is like a multi gauge. And I'm like, I just want this green. That's the multi gauge. That's all I want. I don't want all these other gauges. And so it's very expensive to go that way. Right? It's like, yeah, just oodles of money. It's like almost an order of magnitude more than the other other company. Which makes sense because it's a lot of mechanical gauges and parts in that kit. So I was thinking, Man, what a what if I got that Jeep fan running again, because I have all the code written all the hardware is that I actually could basically delete a lot of the hardware off of it. And I don't need the VFD display. I was thinking so I'm like man, my my rear view mirror it has a screen in it for the backup camera and it has another input that you can display whatever you want. So what if and these are like sensors and stuff that you don't need to have an active display for. It's mainly for what I want it is basically like okay, breathe a transmission temperature. And let me know when it starts getting hot. Like it goes over a certain temperature and then do something about it. Absolutely. Yeah, I don't need no. Oh, it's 162 degrees Fahrenheit right now. Oh, yeah, no, that's like normal operating temperature, I don't care. I care if it gets to 200. Because now I'm like starting to cook the transmission oil. And so I'm like, that'd be perfect, because then I can have the the Jeep fan controller thing, he will rename it because it's not gonna be that anymore. But it can, when it has a you can set alerts and you can do all the Configuring to the screen in the in the rearview mirror. And you just have Hey, Windows alert and make a buzzing noise and then display it on the rearview mirror. And then that way that works great, because then you don't have like, a bazillion gauges around your your, your car. Too much data. It just doesn't mean anything. Yeah, it doesn't mean anything. It's like I only care if it gets over if it gets too hot. I only care if if the the tilt sensor is at whatever, what something like that. Yeah. And so I started looking into it. And so I'm like, Okay, now I need to, let's say basically what it turns into is I need to build composite video output, because that's what the screen accepts is composite video input. And so the first thing I do is, especially with these kind of projects, I go Arduino composite video to Google. Don't lie. You've done this before. I do it all the time. Yeah. Because if you get a lot of good introductory topics, you get the basics about about this kind of stuff. Yeah. And so you get this with this, but an Arduino, especially like the UNO or whatever, you can't generate a lot of resolution. And it's pretty blocky. I'm like, I need to actually display some information, especially if I'm doing configuration screens through composite, I need to have some fine resolution. But there's also same thing with Arduino. There's the was the SAM D series for Arduino, which is the higher end core ARM core processors for it. You still can't it still has limitations on like the resolution the output, it's so like only like a quarter of the resolution or half the resolution of a normal, like composite video screen.
So wait for ADP. Can I take a guess here? You go into a teensy No.
Parallax propeller? Of course, because the parallax propeller actually is the was the has hardware to drive. These kind of analog outputs really? Yeah. Or at least the video generation part. So you can actually generate like, a a 1600 by 1200 display out of a parallax propeller? No, really? Yeah. At what frame
rate enough. Five Hertz.
But so yeah, I can do NTSC out of it at 481 486 was a foreign ad. I think people say four ad but it's actually 486 With overscan lines. So I got to actually do proper NTC NTSC output out of the parallels pillar. I guess what the Jeep fan was originally a parallax propeller project. Ah, that's something like, Man, this is awesome. I just have to hook it up. And bam, Wiz, you're also the moon. Okay, so
you're sort of, you're, in a way you're creating a new project here. I understand you're resurrecting one. But you're creating a new one here. But I think you can actually, I'm going to feature creep a slight bit here, you can throw in another project and get two projects for the price of one here. Because you could put the octo prober into this and make this the octo Jeep prover. And then two projects are done.
Yeah, I like it. Actually. We'll do that with throw in.
It did not take much to convince you all it is, is just
copy pasting the front end Octo prover onto. Yeah, because then that way you just use K type terminals and K type thermocouples and
then suck them right into the processor. Yep. Yeah. So there you go. That I don't like Octo Jeep prover. But I so we'll have to come up with a different name for it. Yeah, it is. It is nice to get both of those done and like, like checkmarks done. Yeah. Yeah, what's the october i? If it was available? I would buy it right now. Because there's I have uses for the October Yeah. If it was let me preface that. If it was available and done.
That is available.
I couldn't build it right now. It just wouldn't mean anything.
Yeah, the actually I was thinking about this is looking at a lot of universal multi gauge stuff out there. They're they are fairly expensive to run. So this could actually be a way maybe to offer a product. I don't know yet if I want to, because especially when you start offering products that are cheaper than other products, you start getting it the support starts racking up because you get more beginners using your product. Is that necessarily a bad thing? Good thing? I don't know. I just know from supporting other inexpensive products is support ends up being your biggest expense. So I don't know if I want to do that. But it's something I might I might explore, depending on how it goes.
I would say you're already working on a product right now. That's not competitor. Yeah. Independent tire. So make the chief driver and
I offer you I think you offer it as open source. Yeah. And be like, if you want to build this and take it, go for it. Yeah. And keep it open source. I probably open under I think it was the ShareAlike is what I usually do it Sure. Well, it's the was open. Can't remember open hardware. I haven't actually looked at the licenses for hardware in a long time. Because I used to do stuff under Oh, that project doesn't even have a license. Oops. Sorry for the air silence people I'm looking at put the five years ago. So I usually I used to do creative commons, noncommercial sharealike which is basically you can take this and learn from it, and use it for your own personal projects. And if you but you can't commercialize it, and if you modify it and sell it, you have to publish the the that's what the ShareAlike means you have to publish your changes. So I really liked that license from the fact that how I view open source is I view open source as open source means you're you're using it to learn is not there to cost not what everyone else thinks no, because everyone copies Yeah, it's not there to copy, it's there to learn. And that's what I really appreciate about open source communities is open source. Everyone wants is doing it to learn. But then you get a couple of people who are just using it to copy and make clones. And so all that does is basically cuts the designer out of their hard work or the designers of the hard work. Hopefully I don't get any Twitter flack for that. Because, man four or five years ago, that was a hot topic in the hardware industry. No, I don't know if it is anymore because I haven't been so stuck in or stuck is the wrong word sucked into the heart of the manufacturing world where it doesn't really matter. Right? Well,
other than, and you've probably run into this before, but I've certainly had people submit projects and say like, Hey, I would like this bill, blah, blah, blah. And I asked her questions about it. And then they linked me to a GitHub repository. Oh, yes. Yeah. And that's what I'm like, Oh, yeah. I mean that before. That's cool. And I'm glad that you're pursuing this, but like, I think you need to have more information than someone else's GitHub.
Yeah. I do agree there. So but yeah, I don't that's topic for another day. And a couple more open source and cloning verse learning. But yeah, I do. I do think the biggest thing about that's, that's why I chose this license for my stuff back in the day. I don't know if there's a better license now. It also doesn't prevent people from other countries like China to just copy your shit.
So it's all open. Anyone can download
it. Yep. Cool, neat.
So I've got a I've got a fun, a fun thing that I actually kind of got working yesterday and I was spending some more time on it today. So it's something I kind of posted on the Slack channel for some for some help and got some got some good aid from some people. So I was in need a handful of months ago have a negative in negative out switch mode power supply, which although that sounds like pretty straightforward. It isn't necessarily as straightforward as you think. So basically, it's a switch mode power supply, that personally I needed a buck converter, so I needed a buck converter that would take a negative 12 volt in and produce negative five 2.5 out. And that's not as common as you think, well, first of all negative voltage rails are not as common as, as positive rails, for obvious reasons.
But so so you just flip your multimeter leads around.
I wish it was that easy. I really, really wish it was that easy.
It's topics like this that really show my lack of depth knowledge on this kind of stuff. Because, again, I do mostly digital. So if you're below zero, holy shit is that number two, that's a different dimension is a negative is an imaginary one.
And that's just the thing, like, You are the standard. You are the electrical engineer that all other electrical engineers are like, and I'm not trying to like water you down or anything like that. But what I'm saying is like, the world knocking me down a peg, the world lives from zero to 3.3, right? Like that's, that's actually 1.8 volts. Okay, now it is sure 1.8 volts. But so like, I'm one of those weirdos that has to work with negative voltages, which by the way, I freaking hate negative voltages, because they just I know they're the same thing upside down. And that's obviously a joke, but like, they're confusing they because they they confuse your normal train of thought.
So you have you have positive rail and Australian rail.
Yes. Northern, the southern hemisphere. Right? Oh, the electrons go counterclockwise in this in the southern hemisphere, right. On the toilet, though.
They have a different spin on him. Yeah, something
like that. So have filled G and Tom Anderson from the Slack channel kind of weighed in on my problem and gave me some suggestions. And specifically Phil turned me on to a particular part, which is the LTC 3805 by
analog delays. Oh, well, that's right. They got bought,
they got bought, right. So it what Yeah, that's really Lt.
So you said LTC, I'm like, efficently, near near tech device.
So So here's the thing. There is a really simple solution to this that I really wish I could have done, but oh my god, they just copied. So I'm gonna copy and pasted the pictures from
the linear tech
website. Well, I mean, if you're adopting an entire company, and you have to, like consume them quickly, like you're gonna find a way to make that work. Right. But it's just like
the same. It's okay, whatever. You're sorry. No, no, it's just that it's just that linear Tech had a a specific look to the PDFs. They had a Oh, yeah, whoever ran there, like it's not marketing. But there's a stand there standards departments for like, their look and feel. did an amazing job, because like, they have special fonts they use for their data sheets that has special colors like that. Like, can peach color that they use everywhere? Yeah. So what the thing is, the moment you see like that font or that color, you're like, linear tech. Right. Great. Great. Sorry about that. And I opened up the PDF, and it was like,
yeah, yeah, well, still kind of there. It's just like a mismatch of everything now.
Yeah, I mean, look up like everyone that's that's listening, go to the link that we're gonna post in the in the podcast notes. And like, so you pull it up as analog devices website, which is analog devices is like, gray, blue, gray blue. You have all these images that are copying phases and linear text website, which are this like, black and peach, peach and tan.
Yeah, it's it's a it's a copy pasta for sure. Yeah, it's great. So it's good, it's good. One of the the easiest, it's actually not that hard to derive a negative voltage rail. From a positive voltage rail, you can actually use a significant amount of controllers out there, and you just flip everything on the output and take off the negative side and you can get a negative rail. I really wish I could do that. And in my situation, the reason why I can't is because my application accepts both positive and negative input power. And if I were to derive my negative rail from the power rail, or the UPS right from the positive rail, now I would have both my positive and my negative rail being derived from my my positive input power and my load would be unbalanced. So okay, yeah, so most people who use our equipment purchase a box, a rack, and they purchase a power supply for that rack that has a rated positive and ready to negative output on it. And in general, you try to keep your load equal between the two, it's never going to be equal. And it's almost always tip towards the positive. But if I were to derive my own internal negative rail from the positive, I would put the entire load of my entire unit on the positive. And that would easily consume the entire power of the entire users case. So unfortunately, I can't do that. And that that's like, easy mode and cheap mode. I wish I could do that. So I needed a book converter that accepts a negative voltage and spits out a negative voltage, which is surprisingly hard to find. Just because, like, who needs it other than me, it's weird. So put some big diodes in there. Yeah. Well, that's actually that is the reason why I'm going with this switch mode is because I don't want to just dump everything is dump it all. Yeah, because the thing is, I designed a new product that has a lot of features and a lot of cool stuff going on under the hood. And so much so that it doesn't make sense to run it on the high rails, I need to drop my voltages because I'm burning so much in heat that the thing gets so hot. Like, it just gets like, you can't touch it, it gets so hot, because I've got so much crap going on in there. So I need a switch mode to get my voltage down without just burning it all as heat in regulators and, and, and everything. So
you're gonna have the first module, it's liquid cooled, man.
Yeah, right. And that's the only module you have, because I consume your entire power supplies available to our bank. Yeah, so So okay, so So Phil JY, from the Slack channel suggested design note 1022, from analog devices, which if you go to the link we send you for the LTC 3805. Down Under like design notes in their little section, they actually show how to derive a negative voltage from a negative voltage on this, which I've got a strange conspiracy here for a second. Like, two hours ago, I was at work. And I knew I had the podcasts and it was the end of the day. So I was just like, Oh, I'm going to talk about this part. So I go to Analog Devices website. And there's two design notes, both of which talk about deriving a negative voltage from a negative voltage. And these two design notes have been on the website for ever, like I mean, and what I mean by forever is like the last couple of months, every time I've gone, they've been there. And it's weird, because they both say the exact same thing. They were just written at different times by the same guy. One was design note 1022. And one was a en 486 or something like that. So design note or application note, whatever. So I drive home and I'm writing the notes for this podcast, and I go there, and one of them is missing. One of them was gone. They knew I downloaded it and they wouldn't check. Got rid of one. I don't know. It's a conspiracy. So regardless, they both say the same thing. So who cares? So yeah, this design note, it's a little bit. It's a it was a glitch in the matrix that got fixed. Oh, yeah. Deja vu. That's what it is. And actually wasn't ever two. I just it was in my mind the entire turbine the entire
time. So we're in the lamest version of the matrix.
Oh, for sure. We're in the 1999 version. This version after
that?
Oh, 2020 version 2020 Version of the matrix. If you don't know what we're talking about, go listen to the last week's episode, The Star Wars podcast. There's a lot about this. So So it's funny because this this design note is just a two page design note with like a handful of paragraphs just saying like, Oh, it works, basically. So I threw this whole thing together and just adjusted my resistors because they were trying to derive negative 3.3 volts, which I have no idea what you would ever use negative 3.3 volts for but parallax propeller and reverse. Like, yeah, right. It goes backwards instead of forwards. So so I just adjusted because I needed 5.5 volts. No, because you can
tell if you just wired up as the ground is your positive rail. Chip.
That doesn't always work. You have to you have to be careful about that.
Maybe it's isolated DC should work.
Yeah, if like I said, it doesn't always work. There are situations. In fact, this chip doesn't necessarily work that way. If you look at the design note on this, it The actual icy is the power is connected to ground. And ground is connected to negative input, right? So it is it's working the way you expect it to. But in terms of deriving its feedback voltage, it actually uses a current mirror in a, like a double transistor configuration or in order to drop current across the resistor for feedback applications. Because this particular I see, you can't just like, throw a resistor on it like you would in the power ups right in the positive range. So it's kind of gotta like a, I don't know, it's fun to look at this circuit, because you have to jump through some hoops just to make it negative. It's, it's weird. Regardless, I built this entire thing up and fired up. And lo and behold, it actually works. Go figure an application note is correct, correct? Yeah, the funny thing is the first time I fired it up, it was spitting out around 5.5 volts, which is what I wanted. I say around because I think my resistor tolerances are, I didn't use fantastic plus or minus 10%. Better than that, but not much. So. So yeah, so I'm getting I'm getting one on one. And I had a load attached to it that was like a one milliamp load or something like that just really, really light. So it was like, okay, cool. So I attached like a two milliamp or three milliamp load to it, and it just dies. So it works enough to just push like a 10k load. But as soon as I put like 5k on or something like that it dies. And what the interesting thing is there's there's a current sense resistor and or maybe current sense with the application of current limiting resistor in it. And in the datasheet, they call out like 15 milli ohms, or 60 milli ohms. Or you can calculate it to be whatever you want for current limiting. Well, I didn't have like a Oh, 402 15 milli ohm resistor just lying around. So I throw a one ohm resistor because that's the lowest I had.
So that's current limiting, like current currently,
which, which is super. Its current limiting, but I got so lucky with a one ohm resistor that worked for the load that I had just default attached to it. But as soon as I went anything above that at current limited and crapped the bed,
did you run the calculation on like, what a one ohm for the current limit
feedback would be? Not yet? No,
I bet you it's really close to what that that I think that threshold is like right there, right there.
So I got lucky because it would have been shitty if I turned it on. And it didn't do anything, because I'd be like, Oh, what did I do wrong? So you know, instead of instead of trying to like derive a really low value resistor, I just made like a long length of wire, and just soldered that in with the hope that it was like a handful of milli ohms. And that totally works. And I put like a 200 milliamp load on it. It worked just fine. So I am going to go in back calculate everything because really this this initial turn on was just to make sure that it was actually functioning. Yeah, I actually have to go and tune it and do things because like, in the standard analog or even LT way, like if you go read the datasheet, it has like 4000 different modes. And you can do all these different things from like, six pins, depending on how you configure it. So it's like, oh, great, okay, I got it functioning. Now I gotta go and actually, like, spend the time and work it all out. So. But
when the moon is waning, and your resistors tombstone, the chip will do this, oh,
I Analog Devices, their products are awesome, because they do so much stuff. They're amazing. But they're really, they're
really good. They have really good applications that are, I think it's because they sell there, or they they market their devices as applications instead of just things like they can do stuff. That's kind of a weird way to put it, but that's what it feels like. Yeah, like, you open up the datasheet. And it's like, this is the specs, but look at 90 pages of application notes in the datasheet and that's like oh, yeah,
you know, you know who Okay, slide digit. The one company that I feel has some of the weirdest offerings is microchip and the reason what like microchip has like when I say microchip, you think PIC processor, right? You think like that's their bread and butter or I guess AVR and I'll say AVR now yeah, that's their bread and butter. But like if you go look at Microchip bill, they'll have like, one random transistor. That is like, why so I don't so LBOs and stuff too. Yeah, they sell deals. They actually sell like MCP is what they start with? Isn't I think MCP is their DAX you might be right I think yeah, they they sell like a An eight a 10 and a 12 bit DAC that I'll start with MCP. So we actually use a voltage reference. That is a no
M cpws op amps is it? Yeah.
So I think we actually both had it wrong. Microchip puts MCP in front of like all of their products Oh god. Sorry, I got something down the wrong pipe there. So So actually, funnily enough, we we use MCPS or abuse MCP op amps in a in a particular way, the MCP 6001 6002, they're low power op amps that work really well, in single supply mode and a single op amp, we can we can use a single op amp to accept a wide input range, and we power it from zero volts ground and 3.3 volts. And the op amp does the job of clamping to zero to 3.3. So we don't have to have external components doing it, we just let the op amp like hard clamp the rails. And what's great about that is it makes a really great input buffer to a processor. So you don't know what the user is going to plug into your device. But the MCP makes sure that it's between zero and 3.3. Ah, and we just use the rail to rail operation of the
device you don't just use like a 3.3 volt zener
Xenos are not great for clamping because they have a really soft knee. And so the curve off the very end of things. op amps are great because they have a super hard knee on top of that, because op amps are you know, if you haven't been in inverting configuration or or whatever they work, they they have
her in your inverting configuration repeating the course Oh, of
course, they so since since they basically work in a in a current configuration, you can weigh over voltage it and it'll still work fine. You know, because most of the time we have input impedances and like the 100k range. So yeah, so it's like 24 volts on the input, you'll still just get 3.3 on the output
that you want. You're not going to Well, if you spec your op amp, right you you're not going to blow your input. Your op amp.
Yeah. Yeah, cute little trick
there is I've learned something I actually probably won't use them the Jeep fan controller.
Yeah, it's a it's a cheater way for getting specific voltage. And for some reason, the MCP 6000 series clamps, the rails super close to the actual root like really, really close. So it's like passable. And if there's any offset about being close to 3.3 volts, you can just calibrate that out in code. And it works out fine. So where were we at? Oh, yeah. So the, you know, it's funny, because when it comes to these negative SMPS configurations, all of my searching have been components that are made by Analog Devices, or Lt. In fact, LT specifically. And there's like, there's application notes for these other LT parts are like, yeah, we can do this, but they're like $9 apiece, you know, or in quantity. They're like, still like six bucks or something like that. Yeah. So the LTC 3805 is not particularly cheap in quantity, it's still like Buck 75, Buck 50, something like that.
But not too bad for specialized active component,
especially of active component right. And, and it solves a big problem of ours of generating way too much heat by just dumping it in resistively. So, thanks to Phil G and Tom Anderson, from the Slack channel for helping out with that I'm super happy that it works. I mean, I still have to tune it. But I've already proven that I can make it happen with loads in the 200 milliamp range, which like that will already service the thirstiest of our of our modules, you know, because we are talking about the negative voltage rail, which we use a lot less power from the negative voltage rail for a lot of reasons.
Well, I mean, LEDs don't run on negative rails.
No. Could though you could, but I mean, since we do so much op amp work. Almost all of our negative rail is dedicated to just bipolar supplies on op amps. So that's usually where it gets all consumed. So sort of in the same realm, I kind of have a cool chip that I want to
do we have a whole segment like this back in the day.
Like the cod What did you call it like the Indiana Jones ancient mystery Chinese chips or something? That
was different segments. Okay, we did have that was the ancient Chinese semiconductors.
Yeah. Bagel and stuff. Bagel pin
kind of stuff. I'm still on the lookout for that kind of stuff. I still haven't I you have to keep one upping yourself. Oh, yeah. And I had Bayeux pin is that? Yeah. So this one was that segment called?
chip of the chip of the week, wasn't it?
I think it was Tip of the Week. Yeah,
I'm gonna look up an old episode. See, once we probably did it like three or four, total. So okay, well, well, Parker's looking that up. This, this ran our
old podcast notes are hard to read.
There's this cool chip, called the SSI 2130, which is made by what sound semiconductor is the name of the company. So this is a VCO core all in one package, VCO meeting voltage controlled oscillator in one package, which VCOs are fairly common if you go to Mouser or something and look at them. But they're, they're not low frequency VCOs. They're, they're all like, I don't know, in the in the megahertz and above. So if you need something that is lower down in scale, it's a lot more difficult to find, pretty much to the point where you have to make your own VCO, which is its own exercise by itself. But SSI created the 2130, which is the the entire core of the of a VCO in one chip, which, if you've ever designed a VCO, like that, like whoa, that's kind of cool, actually, because it has the exponentiation inside, it has temperature compensation inside, it has all the signal conditioning and and wave shaping all done inside of one chip. And in singles, if you just find it off the shelf, this chip is like eight, nine bucks. But in quantity, you can get it for just a few bucks, if not down towards one or $2. And that's pretty significant. This, this IC allows a lot of really cool features like the data sheets, like 30 something pages, because you can configure it in so many ways. It's in like a small, either like a three millimeter by three millimeter or four by four QFN 32 package. But it gives you triangle output, it gives you a saw output, it gives you a square output a sign output, it also gives you a duty cycle, adjustable pulse output. And all of these, like I said, are temperature compensated with the exponential or linear voltage input. So if you want to control this, with a microcontroller, it's really easy to do and it runs off of it actually has really strange rails, it runs off of a positive five volt rail, and then anywhere from negative five to negative 18. On the bottom side, so it's lopsided in a weird way. It does require a really accurate voltage reference, which microchip has some pretty awesome voltage references that are point 1% initial accuracy and like 50 PPM temperature, so and they're not that cheap, I mean, not that expensive. So one of those plugged into this thing gets you kind of in a really good spot, you can get somewhere in the range of 10 octaves with it. If it was 1000 to one, maybe I'm wrong on that 10,000 to one range on it. So you can typically cover like your entire audio range all the way up to like 60 kilohertz with this thing. And then if you're willing to throw negative voltages into the front, you can get it down to as slow as you want it basically. So it makes for a really cool, easy to implement VCO that has a lot of applications and audio. But if you're wanting to do something like I don't know, if you want to do like a network analyzer in low frequency range, this could make an awesome sweep generator, which is easily controllable with a processor. And like I said, you get all of those functions out of it. One of the things though, that it includes, and one of the reasons why it's so attractive is it has what's called linear through zero as a function. So all oscillators oscillate in the positive frequency range. So if you think about zero hertz is DC, and then anything above that is positive frequency. There's,
do you have an offset of your zero
point, right? Yeah, yeah. And like in general, you would have an offset. But what if you kept going negative from that offset? Eventually you'd reach zero frequency, right? You'd effectively reach zero time, and you'd have a flat line right? Well, this this chip offers a feature that is really cool in synthesis maybe not particularly cool for Anything else, but it's called through zero, where you actually reverse time and you go into the negative frequency realm. So effectively what it what it means is you're just flipping the phase, you're going 180 degrees out, but it allows you to actually hit zero and then continue past zero and go negative. So you can do frequency modulation, you can modulate up and down without actually stopping the oscillator. Because a lot of problems with with VCOs that you run into is if you're, if you're parking at a particular frequency, and you go up and down from there, well, down has the problem of there being like a stopping point for the oscillator. But if you have this through zero function, you can continue into negative time realm. And so you don't get tuning issues, and you don't get this weird, like chopped off frequency. If you keep going negative, and this icy with a handful of extra parts, just a few op amp stages, will allow you to implement through zero, which traditionally has been very difficult to implement in VCOs. So that's a really cool feature of it. So if you want to check out a cool chip, check out the SSI 2130. I see lots of really cool applications outside of audio with this A B like the name of it. The 2130. No, it's the fat keys,
it's through its trademark, to fat keys, trademarked voltage controlled oscillator,
it has a lot of other stuff like, you know, if you want to know a lot of stuff in here, if you want to control the range or the scaling of it. So how many volts it takes to double your frequency, like all of that is controllable. I mean, there's just there's a lot of stuff behind this. They really sort of what's they geared it towards people who want to make like a keyboard synthesizer and have most of the stuff integrated.
That's all I was looking at. It's a lot of like, you could build a keyboard synthesizer. Almost with just this one chip, almost Yeah, all the other functions
like the amplifier and the and the filter and the envelopes and stuff, you'd have to do separate. But this does like 90% of the work. And one of the biggest attractive features of it is that it's all inside of one package. It's all temperature compensated. So as the die raises in temperature, all of its features go up and down with it as as you're slamming
those hot licks, those fat keys. And Edie the frequency stays steady.
Here's the thing, the sine wave generator in it is temperature compensated as well. So like I said earlier, if you wanted to make a network analyzer and you wanted to do a sine wave sweep, and and look at an FFT kind of thing, you can totally do that with with this and a processor. So I don't know that. I almost thought of doing something like that. I was like, I've got too much already. I don't want I don't need to add stuff.
I do like how it's this it's funny. He's sound semiconductor can't be that large of a company. They still suffer from the marketing on their first page of their data sheet. I mean, they all mainly, they all do. But especially they're like the citizens here. The s s 1213. Zero requires minimal external components. And it's available in a diminutive four by four millimeter QFN package. It's like, why do you need that adjective? They're just say, in a four by four millimeter QFN package?
You know, I can I have an answer to that. And the reason why I have an answer to that is because I called SSI the other day. I called them to talk about some of the things in the in their data sheet because their data sheets really good, but it doesn't have all the information I wanted. So I called them up to get some clarification on a handful of things. And one of the guys was saying, like, man, a lot of our clients are smaller companies that don't normally deal with QFN style packages. So they are so they have to kind of like build up like, Hey, man, QFN is cool, you should start to use this stuff because like not everything comes in a package. So that's exactly what they're trying to do. In fact, what's funny is they're actually releasing a dev board which is just a QFN to dip adapter. So you can breadboard with it. It's funny because like this is this is a chip that is like in my industry is like way high tech like awesome stuff, but they're still dealing with like hobbyist level guys. We're like I don't like it because it's QFN and I can't solder those. And I get it like they're hard to solder.
But like you know those people can buy prototypes. tobacco fab
that's a good segue
but cigarettes the outro of this episode
yeah
I don't know. This is not available directly at like Mauser anything like that so it might be hard for macro fab to go and get I know hard maximum could do it but you'd have to work a deal.
Oh yeah, we have to go to amazing smith.com
Buy him at nine bucks apiece
or amplifier parts.com Oof
I have an account with them.
Oh yeah. Do you get discounts? I do.
But their shipping is brutal. So he just gets rid of any kind of
are they from Tucson? Ah, that would make sense.
They have three different websites that are all almost identical. They have amplified parts.com ce distribution and tubes and more.com all three of them are almost identical. And they're all the same company. Just different branding, different branding for different people. Yeah.
That baffles my mind but it makes sense. So Well,
that was the Mac fab engineering podcast. We were your host Stephen Craig and Parker Dolman Take it easy.
Let everyone thank you. Yes, you our listener for downloading our podcast. Have you have a cool idea project or topic? Let Stephen I know. Tweet us at Vancouver EP I can't keep a straight face John do that voice at Longhorn engineer or analog in ng or emails that podcast at macro fed.com. Also check out our Slack channel. You can find it at macro fab.com/slack Seriously,
hit us up we'll probably talk about it on the budget.
The quest for the right connector for a project! The right of passage for any hardware electrical engineer starts with a connector catalog.
How low can the power consumption of the Cat Feeder Unreminder go? Parker and Stephen discuss leakage current on this episode of the podcast!
This is the last installment of Stephen's 'Adventures in Injection Molding'. We are going to recap the entire two year sage and close the book on it.