Solar Freaking Cat Feeders

Hello and welcome to the macro fab engineering podcast. We are your host, Stephen Craig and Parker Dolman. This is episode 235.

Okay, so I've been stretching this project out as much as I can.

Yeah, this is like this is like plastic deformation of project all the way to the breaking point.

I was thinking about Stretch Armstrong, but Okay, same thing. Yeah, same material. Alright. Okay, so the cat feeder and reminder, I actually had some time and I sat down and started actually picking parts and going through kind of what I want to do now. So again, this is the, like a timer, so that you are reminded to not feed the cats, because cats will meow and want to be fed all the time. So you don't want your cat to get to fat, especially in a household that has multiple people that will just feed cats to get the cat to shut up. Actually, since I came up with this idea, my cat has gained half a pound. This is definitely a problem.

The severity of needing to complete this project is is growing

is growing as the days go on. So the idea is it will have a timer in it. And then every you know, has an LED on it. And then every like 18 hours, it lets you know, by an LED that, Hey, it's okay to feed the cat. You feed the cat, you press the button on it, the timer resets. And then that way no one feeds the cat multiple times in a day. These that's the idea.

You know, I'm gonna pause you real quick. Parker and I talked about this. Oh, gosh, we've talked about this almost every week now for like two months.

Like, no, what am I about to say is what we've talked about. Okay. In terms of in terms of this podcast, and in terms of like the projects that we talked about on this podcast, we're not like, trying to curate every little thing that we talked about most of the time, we're talking about things that just interest us at the moment, or something that we're doing in our own time. But we also like do projects, and then we talk about like, Hey, did people like this? Did did, like were people talking about this? Or were they not talking about it at all? You know, we do a little bit of feedback. And for some reason, this project has gained like so much traction, and people are talking about it all the time on our Slack channel. And it's like, why but cool. Like, okay, like, what about this one that is like, so much more unique than all the others,

I think because we all share the shared experience of a cat bugging the shit out of you.

Everyone wants to have their hands in this. No, I mean, I love it. I love how random it is. And I love that like this is the project that has been like has got people fired up like it is they're engaging in the cat feeder and reminder.

Okay, so so the idea of with this project too, is I want to do something that I can learn from because I could easily just make a microcontroller with an RTC and then slap a lithium battery on it or like a double A battery and then call it good fun being there. i We did that with the macro watch years ago. It's it basically retool the macro watch, which is a binary PCB based watch that that Steve and I worked on. It was like our first project at macro fab way, way long ago,

today. And also is that not like sort of the MO of 99% of the digital projects. They're just reconfigurations of something you've already done.

Yes. So this was trying to do something I've never done before. And now the Mac watch was low power, but it was a microphone. So I'm like, okay. We talked about using a microcontroller a couple of weeks ago or last week, but I went back to okay, no microcontroller.

Oh, okay. I'm curious because I haven't heard about this.

So the biggest thing is I want it to last a long time and and so either of his battery powered or whatever it needs to last for at least a year or two. So the biggest thing was okay has to have indicators to let you know to feed the cat and also know it's working. And I we talked about different ways of indication last week like E Ink displays. I actually went through all my spare parts I did not find an E Ink display. So I decided not to use one because I didn't want to spend 30 bucks on one. Are they really that pricey? And singles for decent ones. Yeah. And then the, we talked about like a flip dot display which is a mechanical device. The problem with those is you have to have a large surge current to flip them because it's electromechanical coil. And I didn't really want to deal with that. So I went back to LEDs. And then I started looking at, you know, driving LEDs with pulse width modulator modulation instead of having on all on all the time, and so you can cut your current capacity down. And so I was looking at, I put some links here in our show notes. But it's basically driving how the drive LEDs using like an op amp and an oscillation mode. And so you can get down to like micro amps of current for the LED, but you're spiking them with like a milli amp. And you you can basically, your average current is like, I think it was like 30 micro amps or something which I'm like, basically, you can run a LED off like a coin cell for like, two years.

You know, the thing about it is like they the human eyeball with persistence, what is it persistence of vision, like that, you can exploit that to your advantage quite a bit like a small little pause here. I will catch and hang on to it for Yeah, that's that's the idea behind these. I like how they call this I went to this link, the 4049 abuser.

Yeah, that's one of them. If you scroll past that one, there's a better one which uses a op amp. And that second link is used as a MC seven to 15 op amp. And that one is like from three volts, it pulls six micro amps. So you can basically run a LED for two years. Now, it won't be super bright, but it doesn't need to be super bright. It just needs to be a something that you can see with your eyeballs. And you usually use the color green because your your vision has more green cones. That's right. Yeah. Rods are for luminosity cones or for colors. Thinking your eyeballs.

What's what's interesting is Yeah, so actually speaking about that, it takes less green for us to think something is green. Yes, like if you have you ever seen an image of like color corrected colors for the for the human eye where they basically take the spectrum and they make it flat? Like everything looks so dull, like greens are it? There's so little green to make something look green,

green and blue? Yeah, yeah.

I think I've heard plenty of evolutionary like reasons behind that we're like green is life. And green is like growth, and green is food and things. So like our eyes, were tuned to be able to see that to be able to interesting look at that at the same time. Like I've heard plenty of things about like men and women's ears are tuned to different frequencies to be able to hear the opposite sex, which is kind of interesting, you know,

like frequency content that fits within particular ranges. Interesting. We might hear things differently.

Is that why you don't listen to your wife?

No, like, I think it's actually an argument as to why like, her voice is more loud than other people's. And I'm not saying that in a bad way. Like a good way. Yeah, that was totally, we had like a whole chapter that in. I can't remember what class I took. There was some like non engineering class I took in college where we talked about that quite a bit.

So yeah, I'm gonna use I'm gonna use a circuit like this race to use an op amp to cycle the duty cycle that led to get the power down for the LED indicators. Basically, I can run my two LEDs I want one is a like, I'll have like a red, which is like power good. It's running, then a green, that's a time to feed the cat. Right. And that will be well within my power requirements for my device using this setup. And then for the timer. I'm going to use a LTC 2956 which is a really cool part. Unfortunately, I couldn't you we can't use the bagel pin part from couple weeks ago because it does not have a long enough timer on it. I think it was only like went like 30 minutes or something. But this is a a timer that goes from 250 milliseconds to 39 days.

And that's the second bullet point on their features of the first

I made sure to look at the bullet points assigned first see if they mentioned 39 days that's great. So this will this is basically a wake up slash watchdog timer circuit for low power devices like in a field application. So like if you have like a camera or or a IoT device that needs just like wake up every hour, and then do something and then go back to sleep, you would use one of these. Hmm.

So how does this How does this chip function?

I haven't read that far into that yet. But apparently you said it, you said it was some external, like the cycles, you said it with external resistors. And then some capacitors, oh,

that's actually really convenient. And I'm looking at the typical application right now. And there's three pins, it looks like one's called Long one's called range and one's called period. And somehow the combination of resistance that you apply to those sets the timing on it. Correct. Which that's actually kind of convenient. It's I really like it whenever, in fact, I was dealing with this earlier this week, an LED driver chip, where you can set the current of all of the LEDs that this thing drives based off of one resistor that's off of it. I know that's kind of common, but it's super nice when you get a a, an IC, where you get that external combination, or you get that x squirrel with the right word.

configuration based off of resistance. That's just super easy.

And the cool thing with this is it also supports a button in press to reset the timer.

Oh, that really kills it right there that just Yeah, so it

automatically has a built in, like reset functionality that I can use

that okay. Yeah, you win on this one? Yeah. This is like the perfect chip. It is a Linear Technology. Chip. Where are you going to find it from?

Do they? There's like 1000 in stock on like, Mouser. Okay, nice. Yeah, I, like immediately checks. Can

I buy this part? You're like, oh, this is Lt. But not this part.

You can still buy it. And it's like, it's, they, instead of calling it the wakeup timer with push button control. I see. They should call it the cat feeder under reminder. I see. It's perfect.

Yeah. Look at that. Okay, so they have some linear always has cool stuff like this. This is just a not just this is a timer. I remember remember, they they're not linear anymore. It's analog device law. That's true. You're right, it is analog. But they they're using this as a switchboard controller. You can set that up, although, I wouldn't think that it doesn't have any feedback. So it'll just run rampant if you just let it go. So but I mean, I love these kinds of things where you can look at this and see a ton of different applications on this.

Oh, yeah. Linear and analog are really good at that. And

Massimo at this. Yeah.

Yeah, it's super flexible. It has all the features I need. So yeah, I want to try this chip out. And then Okay, so that's the timer. So we got the timer, and we got how we're going to drive the LEDs. Now any power management? So I'd like the idea of going solar, because then my runtime is technically infinite, right? till the sun blows up, I guess right? 4 billion years or however long that's gonna take

it's longer than the life of the cat. Yes.

So I want to go solar and then for the temporary because you know, sometimes, you know, sometimes you have to turn the lights off in your house. So I'm gonna go supercapacitors. I haven't done the math yet. But I the napkin, thought, I guess napkin math. I did do some math. The napkin math shows that the super capacitors can make it live for a couple days. I gotta actually do some hard math to see if that's actually true. But we have a solar solar panel. Probably those those Panasonic B S G Series solar panels you know, what does that stand for? That'd be awesome. The bastar Galactica solar panels

you know, the one problem with super capacitors, which super capacitors are awesome. But you know, they're great for discharging, you still have to charge them. So the question is, like, do you have to have this thing on for a couple of days before you start it's

Well, I what I mean by ion is like the you have to like let it charge up and then you can begin operation.

Well, you hit it with well, so to that you usually have a fast charge, energy harvesting style IC. And so I'm going to be using the ATM 10 941 And it's a really cool IC that is it's a energy harvesting IC that's designed for solar and allow holes for multiple different kinds of technologies for storage, like lithium, super capacitors, bunch of other stuff on there. And it also has a built in LDU. So I can run my timer and my my pulse width LEDs off the 3.3 volts ldeo. It has. And this is actually very interesting question I came up with when I was researching, like energy harvesting ICS is What does energy harvesting mean? When you see that on a datasheet? Because it says it's a marketing term. It's a it's a it's an SEO thing. Yeah, it's an SEO term that you Google stuff that you're looking for shows up, it's what it is. But what does it actually mean? Like devices that have that on them, like integrated circuits that have that terminology for them,

if I were to take a guess at that, what I would say is excess energy that is being burned in some other mode can be extracted and utilized in some way, via this technology.

Pretty, pretty close. So the best thing I came up with this could be wrong. This is what I came up with is basically gathering energy from energy sources that are intermittent. Okay, so there's not you're not getting constant, I say constant. Because we know reliable, you're not getting reliable energy. Yeah, not reliable. Because like when you plug in to your wall outlet? Sure, it's not exactly 110 or 120. It varies as a percentage, but you're reliably going to get power out of that sockets. Yeah, if you stick

a fork in there, it's going to hurt reliably, reliably.

So yeah, the best I could come up with was energy harvesting is from sources that are intermittent or are dirty, I would say power sources. So they have really good rejection. Or they can also handle like, if you took the light switch and flipped it on and off really fast. And your dad always yelled at you because you might blow up the light bulbs.

Which that's not that's

totally not true. I don't know if it's true or not. But that's what I remember when I was a kid.

Like that argument was based off of the inrush of of current into a cold filament in the light bulb. And so if you turn it on, and you let it warm up, and then you flicker it like the there's not a ton of inrush because it's already warm, right? You lightbulbs get the majority of their stress from when they're going cold to hot. So they actually get the most amount of stress from daily use not from flickering. So there we go.

There Yeah. Until my dad something

don't flicker it?

So yeah, that's that's basically my part list for the cat feeder under reminder, and I'm gonna throw that together in a schematic tomorrow, and maybe we're on the board. We'll

see. I like I like how it has kind of evolved in multiple different ways. And it's gotten like, somewhat larger in scope and smaller, but but really what you've landed on here is a few like purpose built ICS that have multiple functions that will work together. I think that's, that's kind of the hallmark of inelegant design.

It's not a cheap design. But again, I'm making one, it's it's like, oh, how much is this LTC chip.

Although I honestly I could see something like this being on tindy you know, who are these like $8 chips?

No, it's not that bad. But that timer is a $5 $5.93.

Oh, yeah. I mean, okay, so a few weeks ago, we were talking about 7555 timers, which those in and of themselves are not particularly cheap, but they're not $5.

Yeah, but I kind of want to do it this way. Because these are all ICS I've never used before. Yeah, yeah. And it'd be fun to use these things.

So what's your so you're doing okay, you're doing a solar cell into a power management IC that feeds into a super capacitor, but also has the ldeo So basically, the output of that is chip is the power is the power, right? It's what you're looking for. And then that is going into your timer, which also has the reset button stuff. And that is going to feed a low power comparator system that illuminates LEDs, correct? Pretty, pretty straightforward. What I like about that is that it's linear. Like, yeah, like that concept of there's not a lot of parallelism in there. No. Yeah,

I think the only parallel thing is the output, because I'll have, I might try to find a low, a dual, low power op amp, so I can use one package for it.

Well, I was looking at that example circuit you showed, it's actually technically a comparator, which you could use an op amp for, but comparators will, you can find competitors that have much, much lower output. The one thing about this comparator is it has a traditional push pull output, as opposed to a open collector, which, you know, you want to really start losing power and things like that. You could go that route, but I think what what they were discussing there is pretty low power for push pull output.

Comparators not cheap, either. How much $2.42 It's gonna be like a $40.

In singles. Yeah, yeah. Well, look at the LM 339. That's, that's a pretty let me go look, right now, before I put my foot in my mouth, how much that cost? Because I use I use the LM three, nine a lot at work. In fact, we Okay, so we have a really cool project. I can't really talk too much about it.

I'm not complaining about the price. I just started looking at it. Because I'm like, I'm gonna build one. It doesn't matter if it's $3 or $1. Well, but

But yeah, if you were gonna buy more like you can get a quad comparator lm 339, for 35 cents and singles, and then no eight cents sales where you see an eight cents. I'm on mouse or whatever. Yeah, there's a cheap, you know, but I frankly, I don't normally have to care too much about low power applications. So your special comparator might need that. Yeah. Micro power comparator.

So yeah, I'm gonna, I'm gonna try this out.

You know, okay, real quick sidenote in, I don't even remember what class it was, it was my, either my sophomore year or junior college, we got into the first like, what was called electronics. And it was like less about building like robots that crawled across the floor, and more about learning about transistors, and op amps and things like that. And we learned that op amps could be configured as a comparator. And that's where we stopped. Like, in terms of like, they said, You could do that. And then they moved on to the next thing. I didn't realize at that point in, in life, how important comparing is actually in signal processing. You know, like, later on, I'm like, wow, you use that all over the place. And thanks for spending half a day on it and being like, you can do this. Let's talk about the next topic. You know, that's not enough. You know,

a lot of one on one classes are like that. Oh,

I think I think 90% of college is that where it's like, these are things that can happen and here's the theory behind them. Now you can go design a spaceship that goes to space right like good luck

hasn't stopped us yet.

No, no. What that means is we decide cat feeder on exactly what it does.

Super over engineered but simplistic. Low Power cat feeder under reminder.

Id Yeah, right. Because if you just wanted to design a cafe to run Romani you could buy an Arduino and have it done in?

Oh, no, you would buy that $10 a timer that goes for 12 hours?

Well, yeah, if you want to go fully mechanical, you could Okay. With an with a $35 Arduino and the LED that's on it. You could you could have this done effectively and a push button. You know, it wouldn't be fun, though. No, of course not.

Well, you you wouldn't even need the push button. You would just use the Reset on the Arduino. You could seriously use the board by itself use the board by itself. Yeah. And use a cell phone charger to plug into the wall. And then you're done. Yep, yep. Again, not fun.

So okay, so $5 for the cell phone charger, which is so add $35 for an Arduino and $5 for the charging at 40 bucks. And you said your design is about 40 bucks. So there you go. You're you're doing our due Arduino solution.

Well, mine is 40 bucks plus, like 100 hours of engineering time.

Yeah. I don't know. I didn't use like I really I in fact, I'm sitting at my desk right now and I pull an open a drawer. I've got I've got a what is an Arduino Mega. I think the mega that's the big one. That's the big one that 2562 processor or whatever. 64 Yeah, yeah. So like, I love having something like an Arduino available at my fingertips right now because like if I have a harebrained idea that I want to try for something, it takes a few seconds to prove it. I think that's where Arduino really excels. Not necessarily for like, hey, I want to make a cat feeder, I'd reminder and I want it to be able such that anyone can make it right now for, you know, really quickly. Like, it's more about like, I've got this idea like, how do I do this in five minutes? Well, I opened my drawer and I find my Arduino right here. Like, I don't know, it's my thoughts. So room remember, earlier this year, I kind of rebuilt my entire CNC and got it up and running.

So it is running, and it's doing what it needs to do. But I, I've got that itch, you know that it's for that little bit more accuracy, and that little bit more precision. Because percorsi my CNC, it's not made of like, space age technology. It's made of like stone age technology, as well as a caveman

technology.

But here's the thing that's funny is like I so I've literally spent God a ton of time on this machine just because like I really, there's some nostalgia in this machine. For me. It's like it is arguably a awful machine it is arguably terrible. But there's there's some nostalgia that keeps me coming back to him being like, I love you even though you have a lot of well, you built a lot of projects with it, I built a ton of projects on this thing. And I've built some production on this thing when I was running my company like it's, it's functional. Let's just put it that way like that is that is the one word that describes this it is purely functional. And, and right now I am easily able to hold less than 10 thousandths of an inch in one axes and less than 20 thousandths of an inch in the other axes. First one being x second one being y and then the Z axis I'm actually holding less than five thousandths of an inch consistently on that thing. 10 thousandths of an inch is fine in my mind. 20 makes me a little upset. And five is like I'm cool with this. So the thing that's funny though, is like when it comes down to what I'm cutting with it, which is almost entirely wood. All of those numbers are okay, for the next year.

When I do woodworking, I know someone out there is gonna be like oh, like, if you get Vineet, within a 30 seconds of an inch. Yeah, I'm like, you're good.

You know, and okay, so here's Yeah, I was about to make that exact same point. And I think it's funny, like, the most fine resolution for 99% of woodworkers out there is a 32nd of an inch. And the reason why that is theirs is because that's what their tape measure shows

that well tape measure shows six teeth, and then you can go,

Okay, some tape measure show a 32nd, you can pull out a ruler, right? So I'm two thirds, I'm within two thirds of a 32nd. Right, as far as I'm way within reason on that. And it was funny because like, I got the CNC up and running and I was pissed off at it because I wasn't getting as good of accuracy as I wanted on every axis. So I kind of like shelved it for a moment. And then like a few weeks ago, I was like, you know, I this is fine. For everything I'm doing I need to stop being such a little crybaby about

my so you're comparing it to your seen see you had it work?

Well, there's a difference between $2,000.02 $150,000 You don't like? And that's exactly the point that I want to get at here. Like, what's funny is the difference between 2250 $1,000 What that gets you is an extra few 1000s of accuracy. Like that's what it affords, you

know, as I'd say is those are two orders of magnitude often prices give you two orders of magnitude better resolution.

Just about

so that actually holds up,

it does hold up. And but the funny thing is, well, okay, we're also comparing apples and oranges because I work I hold about a half 1000 in 6061 aluminum, and at home I'm holding 20 1000s in MDF. So we're talking about a couple of orders of magnitude. But but in general like I like I mean I'm I'm cutting very different materials regardless, basically the stuff that I work on at home, and that matter for the CNC, I just build in the fat that works for that, you know, I just know that my accuracy is 20 thousandths of an inch. And the thing is 99% of the time it's better than 20 1000s of an inch. It's just the worst haste I've ever measured is 20 thousandths of an inch, which, frankly, is not actually that bad in the grand scheme of things, especially because we're talking about wood here. Like I could cut it perfectly, like no tolerance whatsoever today and come back a week later and it's swelled by 10 thousandths of an inch. Right? Like, I I've measured MDF and plywood where I cut it, and then it swelled within minutes. You're like, well, that's because you live in Houston. Well, right, exactly. I've mentioned that. So in fact, we, we even had to catch what was it? David gunness on from fulcrum acoustics, and we talked about manufacturing with wood. And he was basically saying, like, yeah, if you're, if you're planning on producing something today, you better cut it today. Like, you can't just like cut it and put it on the shelf, it's gonna, it won't work in a week or two weeks or whatever. So when we're talking about, you know, cutting organic material, regardless, all of this to be said, so I've been doing a handful of small projects on the CNC to kind of like, get my confidence back to being like, yeah, this thing actually works fairly well. So I cut what's called a cradle. Basically, I designed up this rotisserie. Right? Oh, yeah, spit, right.

Yeah, I would call it you built over history.

I think I think that's not not a terrible name for it. Like, basically, this is a cradle rotisserie that holds a chassis for whatever you're designing. And basically, you can like any audio gear you want, and you can, like scoot it in or out. It's for like, audio gear, like amplifiers and stuff. Like they weren't for that. But I mean, it's just it holds a chassis. Um, and I could, I could have used it for my brew rig box that I made, you know, so yeah, it's just Traditionally you see this in more audio production format. In fact, the funny thing was, I was actually designing it to send it off to my buddy, but I was because of COVID. And because of a whole bunch of other stuff, like I cut this entire thing, and I made it, it's gorgeous. It's absolutely beautiful. And then I took it to UPS, actually yesterday, and I was like, hey, I want to ship this to my buddy. And they're like, yeah, it's $190 ground shipping. I called my buddy, I was like, I guess you're not getting this. Like, I'm gonna go pay $190 to send my friend a couple of pieces of MDF. Yeah. How gorgeous they look. Yeah, right, right. Like I even like engraved logos and stuff and and stuff. It looks fantastic. So I basically I took all the pieces of MDF that I cut and I built it into the actual cradle and then sent him pictures and I was like, This is what you would have had if UPS wasn't so ridiculous. So okay, kind of on a side note, one of the things I was playing around with was okay, if I if I were to upgrade my CNC which I'm actually not trying to right now but if I were say, what would I What are the things that could happen? What are the things I could do to close that gap between say 20 1000s of an inch accuracy to half a 1,000th of an inch accuracy? Three inches per second in MDF? A depth of cut I don't know let's say a quarter of an inch

something I remember a long time ago this is like when we first started like like we weren't even friends at this point yet

right like I think you just started working at macro fab and you were talking about going rack and pinion on on your CNC and that never happened. So this is the next maybe evolution of your CNC says go for Yeah, perhaps I've really liked the rack and pinion idea, by the way record. Okay, so rack and pinion. And, okay, this opens up kind of a weird conversation that or a deep hole that I've been kind of diving down, which is interesting. So when you get into hobby level mechanics, there's something I've found and maybe you found this too when it comes to like 3d printing and things like that. There's like you can search eBay and Amazon all day long for 1000s and 1000s of like these Chinese grade parts, right. So one of the things I found is there's just Amazon is just littered with ball screws, if you want if you want linear motion from a rotational motor, you can buy a whole system from Amazon and have general accuracy I say general as in like, I don't know, however much money you want to spend you can spend at Amazon and get accuracy based off of ball screws, servo motors and whatever. And I guess drivers and things like that so. So I actually ended up buying a short 300 millimeter ball screw that, if I ever wanted to, I could convert it into a Z axis for my CNC, because I'm curious about these Chinese ball screws that you see all over the place. Like if you go to Amazon right now, and you type in S, F, EU, and then four digits, you can find all kinds of ball screws. And so if you, and then the first two digits will be the diameter of the ball screw, and then the second two digits will be its pitch. So like takes SFU 1605 would be a 16, diameter, bolt screw, and then oh five would be five millimeter pitch, you'll find tons of stuff, or say 2010 and B 20, diameter, 10, pitch, things like that. They're all over the place. And what's interesting is, if you really start digging into it, you look at they're, they're all they all have the exact same grade. In other words, they have the same kind of error, across length. So you don't get to pick their accuracy, really, you get to pick their pitch. And you get to pick the length of things. So anywhere from like 100 millimeters up to like 2000 millimeters, you get all of these things. And if you want to go step outside of like the hobbyist grade, like the prices are astronomical, like this stuff goes up so fast. With these bolts, screws, most of the accuracy grades you see are what's called C seven, and C seven is I think I'm getting this right, it's two thousandths of an inch every 300 millimeters, I think is correct, like that's what it will hold. So if you travel 600 millimeters, you can be sure you're accurate to within four thousandths of an inch.

That's pretty good.

It's not nice, not bad, but in turn, but it's like hobby grade. Because these things go up to like see one grade, which is like, point, you know, it's like a 10th It's like a 10,000 over 300 millimeters as opposed to 2000 somewhere. So like, but but every time you go up a grade on these things, the price like triples, as

well. It's what we said earlier, basically you go up a grade order magnitude gets tacked on.

It's really fast, like the money you can spend on this stuff just blows through the roof. And actually one of the reasons why he was even talking or looking at Paul screws is we had, so I had a our yearly maintenance on our big $250,000 CNC at work. The guy flew in to come and do all the jazz on this thing. Your machine has a butler? Oh, yeah, absolutely. And he is expensive. Let's just put it that way. But we figured we figured you know, we'll pay for one or two years of the service thing. Actually, you know, a side note, like this guy came in, and he was like, this is one of the best machine machines I've ever seen. And it's only a year old. But it's like, this is this thing is in great condition. What do you do to this, and I literally looked at the guy was like, I'm, I'm doing the bare minimum. I'm doing exactly what you guys told me to do. And I'm doing nothing more than that. Like you told me grease it on these dates. And like every Monday clean, you know, this part and that part. And I just do that. And he's like, worked out. He's like, Oh, nobody ever does that. It's like got a quarter million dollar machine like, I'm going to do the regular maintenance that you tell me to do?

Yeah. Well, you understand that? I understand that. Remember, remember when we were trying to get machines to be maintained at the fabric a long time ago? Yeah. And this is when we were buying used equipment a long time ago, Mac fab. And like, so you didn't have plans where someone would fly out to do maintenance on your machine. You had to do it yourself. And so we would have maintenance plans for operations to do it. They would never do it. Right.

I remember I remember one time was my machine broken.

It's like, have you gotten the maintenance and like the maintenance sheets blank. It's like, well, that is why

Right, right? Do that. I remember I remember one time we were we were we were out on the manufacturing floor and we're like, man, the reflow oven is really like squeaking a lot. And the chain rails that hold the PCBs hadn't been oiled in some time. I don't even know how long it was but it was like come on guy was or that was the old Bravo. Oh, do you right that was and it was it was screaming how loud the electro vert wasn't it?

Yeah, electro vert Bravo 4050. I think what what it was Yeah,

yeah. Oh, man. Oh, man. And it's like, Come on, guys. You got to actually have oil this stuff for it to work. So

I think I think are my solution to that one. Was I because I was the one who was doing all the maintenance documents is I basically double, like, doubled how often maintenance had to be performed. So if they missed one, they still are they still on time and actually started the work. Yeah. So now we have Micronics and a Howler and all that stuff. And those at those were all brought new. Yeah. And they come with bottlers that show up and maintain everything for you.

You know, okay, so and that's the thing, like this guy came in to maintain our CNC. And I was there for the whole day that he was watching, you know, going over things. He didn't do a whole lot of things outside of what I normally do, basically, he was there to make sure that what I'm supposed to do is was actually done. Don't get me wrong. He did. He did a handful of things under the under the hood. Like he made sure that the head was in tram and made sure that everything was squared. Those are things that I'm not capable of doing.

He's there to make sure you're still under warranty. Yep,

yep. Yep. But I mean, he he said he came the previous week from a place that had nine of the machines that I I was running, and none of them are maintained. Zero, like they've never been greased. Yeah, well, okay. So boil back this all up. So ball screws, right, you can buy these off of Amazon or eBay, right? So yeah, that's what we're talking when this when this guy had my entire CNC at work at shift apart. It's, it's a cast iron frame with ball screws on it. And these bolt screws are the same. You know, they look identical to the ones you buy on Amazon, they're quite a bit more expensive. I guarantee you that. But they they look almost identical. They even have the same end blocks and bearing housings and things like that. So, you know, yeah, sure. On the big CNC, they just bought the C one grade as opposed to the C seven grade.

But do you think they actually build them all and then bin them by how accurate they are in the end?

Well, so both groups are done in two ways. You can either grind them, or you could roll them. So if you're grinding them, you're inherently going to get better accuracy. But if you just roll them, you're inherently going to get kind of crappy your accuracy. So they they kind of split the line and then bin them that way. You know,

what I'm saying is? Do they actually measure them at the end? Oh, I

don't know. I mean, I

doubt it is then you could bin them? Because you'd be like, Okay, this one, actually the tolerances lined up? And it's a see one?

I don't know, I'd have to I'd have to dig further, deep, deeper into that. That's something that I think you'd have to go to the ball screw manufacturer themselves. But I bet you that they have the phrase guaranteed by design or what? You know, I really doubt that they're testing every single one. Yeah, you're probably right. Yeah. Because just because it takes so long to do. Oh, yeah. But you're gonna be surprised if somebody liked the manufacturer of my CNC at work, they install the screws, and then they probably test them for ag

or if they're grinding them, like the first, you know, 100 that come off the off the setup, RC ones, and then the next 1000 or SI toos? Because as you as your tooling wears out.

Oh, yeah. That seems pretty advanced. But maybe I mean, that's, that's a good manufacturing right there. If you know your yield to do that knows the numbers there that I mean, so I don't know. That'd be the you know, your stuff pretty good. You're at that point.

So, imagine if you're making like a million ball screws, like a month, you probably have that down? Probably, but no, or it can be done in a couple of different ways. No, I do.

Well, and that's just the thing like so every machine I actually walked around to all of our machines at work, they all function on ball screws. Like there's none of them that are rack and pinion. None of them that are Mag Drive, which the the chronic micronic is Mag Drive, which that is super cool. But frankly, I've never seen a C and C B mag drive because I don't think it has the actual I don't think it has the torque to drive a bit doesn't have the torque metal but it has plenty of acceleration to be able to Yeah, throw ahead across things. Mag Drive would be super cool, but that's well beyond hobby level. So I was thinking like, okay, so if let's just pretend I had some money to throw at my CNC. If I were to slap some ball screws on it, would I be able to get better accuracy than I'm already getting? And the answer is yes. Like it's it's difficult because like it all depends on how deep you're willing to go with it. And and so I've been trying to calculate like okay, so take a particular diameter All screws take a servo or a stepper motor that's connected to it, drive it at a particular RPM, what kind of linear force can you apply on a bit through a material. And like this is this is total electrical engineer trying to pretend to be a mechanic. As soon as you start really looking into the equations on this, it's like, oh, God, this is awful. This looks so bad, like, really determining, like, I don't know, moments of inertia on ball screws to be able to determine acceleration rates of a cutting bit through whatever material you're going through, like, Man, I, for all the mechanical engineers listening like mad props to you guys like this stuff is awesome. It's really cool. But it's also like, Oh, crap. I mean, I'm assuming you guys say the same thing about circuits. Whereas like, we can just like talking about cat feeder on reminders like it's not a big deal, right? Where that's probably black magic to some people. So So yeah, like, what are some of the calculations I've done? Taking some around 600 ounce inch stepper motors, deriving some ball screws that are 10 millimeter pitch rough calculations that I've seen, I can apply 300 kilograms of force to my cutting bit, which like for MDF, like think of yourself standing on your router, let's think of one and a half of you standing on your router and driving it through mbf. Pretty sure that would cut through it. Alright, there we go right through it. So like, these are the kinds of like tricks that I'm playing in my mind. Because I don't know. Like, I don't have a gut feel for like, what is 300 kilo kilograms where the force mean to MDF? Well, like, yeah, if there was going to be fine.

A part would go through it without it spinning. It punch a quarter inch hole.

Yeah. So okay, cool. Great. So I found out that like, the torque on my motor is enough for this, right? Yeah. I don't know. It's been. It's been interesting. So

like, but the tolerance tolerant back to tolerances would you have better tolerances

would I had better tolerance, and that's just the thing, like, okay, so most of the ball screws that you find on like Amazon, you're gonna find almost all five millimeter pitch tolerance, which five millimeter pitch means if you rotate it one time, you've literally moved five millimeters. So that's great, and all but if you want to go quickly, you now have to spin it much faster to actually move linearly fast. So for the kind of material that I'm looking to cut, speed matters more than say accuracy. Because you'll actually lose accuracy. If you go too slow. Like MDF and plywood you want from everything that I found, if you take your spindle and you rotate it, like between 10 and 20,000 RPM, you're looking at feed rates of two to four inches a second look, which is pretty fast. Geez, what is rapid, it's ripping, it's ripping. But think about it. If you have 10 millimetres pitch on your ball screw and you want to move something linearly. How many times do you have to rotate it to make it go four inches a second? Well, you're talking about like, you're, you're up in the 900 to 1200 RPM on that ball screw.

I was gonna say it'd be like, what? 12 rotations?

It starts to get out there, right? Well, yeah. Okay, so I'm 60 Anyone who's ever played with stepper motors might already have their ears like perking up a little bit when they were they're starting to think like 1200 RPM, that's pretty fast for a stepper motor, right. And I apologize, maybe my math was wrong, maybe I was thinking about a five millimeter pitch there. Regardless, the majority of the calculations that I've started to work out is for things like plywood and MDF, if I'm sitting at somewhere in the 450 to 500 RPM range on a 10 millimeter ball pitch, I can get three inches a second of travel, if like this is where mechanical engineering is so ridiculous. Like you guys have so many derived things where like, you have to follow this all the way to like CMCS are such a exercise in madness of Cartesian coordinates, where it's like, at the end of the day, all you're looking for is XY and Z. But you have to live but like to make a move in an x direction. You have to do like 15 things to get that to happen. And they all are like derived from each other. I don't know like this has been pretty fun to be honest. But regardless, okay, so take take a stepper motor that's doing say like microstepping 800 pulses per revolution or something of that range on a 10 millimeter ball pitch, if you just do the math straight out, you'll get something less than a 1,000th of an inch worth of accuracy. Now, anyone who's ever played with microstepping knows that like microstepping is first of all, it's not linear. And it's not always accurate. So, the So my rule of thumb when I'm doing microstepping is like, you never actually believe what microstepping is, you multiply it by a handful of times, and just hope that you're better than that. And I think that's, that's okay. For most applications, so like, like, say, if I was an eighth, eighth of a Thau, microstepping, and I multiplied by, say, four or something like, I'm still less than 1000. I'm still okay, on my CNC. So the short answer to all of this is, if I if I get a handful of fairly torquey stepper motors, and I do a 10 millimeter ball screw pitch, I can get less than 1000 accuracy in general, which is about 20 times better than my 20,000 error. Right? So the answer the answer along around is, yes, I can get more accurate if I spent some money on my machine. Which, who knows? Maybe that's a 2021 kind of thing. Yep. Whichever you're I would like to be able to cut some aluminum on this thing. Or every cool. Yeah. So yeah, so ball screws. Technically, the bulk screw was supposed to arrive today. I was hoping to have it in hand so we can play with it on the podcast. But USPS is not always accurate with their delivery dates. So too bad. Maybe it'll show up tomorrow.

It happens. It actually probably just got left on the truck. Probably. Usually what happens.

Cool. So there is one other thing that I just want to mention about something I've been doing. That is way off topic, but something really cool. I had a guy I

was reading your notes on this. I'm like, Oh, that reminds me of other things.

Oh, okay. Well, maybe we'll have to tie this in somehow. So randomly last week, just some a guy local here in Denver contacted me, and he's like, Hey, I've got an ARP 2600, which an ARP 2600 is a vintage synthesizer. They don't say Atari 2600. You know, same era, basically. And if you looked at the box and everything like that, you'd be like, Well, yeah, it looks about that time it has the fake wood on it. So the ARP 2600 is sort of like, I don't know, the vintage versions of everyone's first synthesizer. In a way it was legendary in terms of its sound. And it was like libraries used to have ARP 2600s, where you could like rent time to learn about synthesis. And things cool. It's super cool. Here's the thing that's incredible about this. You don't you're

sorry. It looks like those like, like 501, like electrical educational things you got from like, from RadioShack RadioShack. Oh, yes.

Like it's, it's very much that version of synthesis. Like colleges used to have this. Like I said, libraries used to have this. This was like the way you learn synthesis because it was an all in one unit. Like you could play it without patching. In other words, you could play it without connecting cables together. It has a built in amplifier with speakers in it. So you could play it by itself and things. But like the whole idea of signal routing, this thing, this thing would teach you signal routing and electronics and stuff. I just looked at what what costs on eBay. Oh, yeah. Yeah. That's crazy. This guy hits me up in an email last week. He's like, Hey, I'm picking up an ARP 2600 And would you be willing to do some repairs on I'm like, What's wrong with it? He's like, probably nothing. But I want you to just go and clean it up and do some stuff.

Make sure when it gets turned on, it doesn't go it doesn't

Yeah, it doesn't go boom. Right. So I was like, well, what's the history of this thing the guy said the original owner purchased it played on it for some time and then put it in his closet in 1984. And it has not left that closet until last weekend. And that's cool. He bought he bought this thing from this guy. And it is like factory fresh. This like so

Denver Colorado smells like the smoke that was in that race got

that old rosin smell to all the solder joints and everything. It is classic. No get this. It is so dry up here that nothing has rusted. The screws look like the you bought them from Home Depot right now like they they are perfect. This whole thing is flawless. And it almost entirely works. There's like one or two things that is like a little bit off and It's probably just because like, I don't know, solder joint cracked or something like that, I don't know, it's literally sat in a closet for years. And the guy just gave it to me. It is gorgeous. The keyboard is not like discolored from UV light or anything. It's still white keys, not the like yellow tobacco keys. And the funny thing was, like, we got the story from this original owner of like, okay, so what is this, so this guy was kind of a synth nut back in the 70s and early 80s. And that wasn't a thing in Denver, that wasn't something that like people were into. So he joined up in a country band, but he played a synthesizer in a country band. And they played some, they played some original tracks that they recorded. And it's, it was synth music, it was straight up synth music, but they just called it country, which is kind of ridiculous. Like they had to like sneak it in some way. So I'm gonna take a bunch of pictures and stuff and post it up in our Slack channel, if you want to come take a look at it. I'm mainly mainly what I've been going through is just looking at like solder joints that have just been kind of crappy. And I'm cleaning those up. Just lubricating all the pots and things like that. It's funny because like with with a repair like this, because I've done a bunch of vintage equipment from the 60s and 70s. Like, a lot of times you have to put some elbow grease into them, you got to make them work. This thing is like a joy. I don't have to do like I do like the minimum amount of stuff and it's working. So I am going to replace all the all the capacitors on it because right now it it takes five minutes to turn on. Like you turn it on and it like really slowly comes up. So I'm not even turning it on right now. I'm just going to replace all the caps before that. Because I know like they're just like they've puked and

dried out. Yeah, they're super dry. So this was what remind me of this is a barn fine. synthesizer. Oh, for sure. So that's a huge thing in the car community is finding barn find cars and stuff. Oh, yeah. My checker is the barn fine. Yeah, it was it was stored away in like the early 80s. partially restored, and then someone pulled it out and was in a barn for 31 years. And then someone pulled it out. Brought it to Houston. And then I bought it. Yeah, you got lucky on that. Yeah. So it's like, it's it's really cool. Finding stuff like that, that's just was just put away and forgotten about

this thing looks, there's not an ounce of corrosion on any of the hardware on this thing. Like this thing looks like it came off the factory line and someone just put it in a time capsule. Again, I mean, like this is a gem to find for sure. And an ARP 2600. In this kind of condition. I mean, we're talking five digits easy on this thing.

I was looking at what they call it an eBay, I'm like, immediately closed it out, like don't want to think about I

mean, you can you can get close to five digits for one that's in crappy shape, you know, the upholstery covering that's on the outside has some where and the previous owner put their own, like, they screwed in their own, like lock into the side. So they could you know, take it to gigs and lock it up. But other than that it is factory fresh. So super, super cool. It'd be fun. I told the I told the guy like, now, now that I'm not doing repairs as often, like I used to be pretty quick at my repairs. I tried to get everything done within a week. I told this guy I was like, I'm gonna hang on to this thing for a handful of weeks and I'm gonna have fun with it. He's like, cool, you just, he's not. He's good. Lunch is nice. So So yeah, like I'm just having this as like a fun side project when I have an extra hour or two here and there. But I'll take some pictures of it because it is gorgeous.

So yeah. Cool. That was the Mac fab engineering podcast.

We're your hosts Parker Dolman Steven Craig litter everyone. Take it easy.